Machine learning-driven prediction of optical and physical properties in lanthanum and gold-doped zinc borotellurite glasses for optoelectronic applications
Recent advancements in glass science have focused on optimizing optical and structural properties through the incorporation of rare oxide nanoparticles, such as lanthanum and gold oxide, to enhance transparency, refractive index, and band gap energy for photonic applications. Additionally, the integ...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/44999/ http://umpir.ump.edu.my/id/eprint/44999/1/Machine%20learning-driven%20prediction%20of%20optical%20and%20physical%20properties.pdf |
| Summary: | Recent advancements in glass science have focused on optimizing optical and structural properties through the incorporation of rare oxide nanoparticles, such as lanthanum and gold oxide, to enhance transparency, refractive index, and band gap energy for photonic applications. Additionally, the integration of machine learning techniques in material design accelerates the prediction and optimisation of these properties, enabling more efficient discovery and development of advanced glass systems for optoelectronic and photonic devices. The meltquenching technique was employed to fabricate zinc borotellurite glasses doped with lanthanum oxide and gold oxide. The glass’s amorphous nature was confirmed through structural analysis, which detected absorption bands corresponding to TeO3, BO4, and BO3 units. A drop in molar volume with an increase in gold oxide quantity prompted a suggestion for structural reorganisation. Optical investigations indicated an increase in the optical band gap energy and a blue shift in the absorption edge. Four machine learning models—Random Forest, Gradient Boosting, XGBoost, and Artificial Neural Networks—were employed to forecast critical optical and physical parameters, such as refractive index, density, molar volume, optical band gap, and Urbach energy, to expedite the discovery process. This data-driven approach illustrates the transformative capabilities of machine learning in materials science, facilitating the swift modification of glass compositions to enhance optical |
|---|