Atomic Structure Simulation and Properties’ Prediction using Machine Learning on Neodymium Oxide Nanoparticles Zinc Tellurite Glasses Aided by FTIR and TEM Analysis

Nazrin, S.N.
Zaman, Halimah Badioze
Jothi, Neesha
Jouay, Doha
Lahrach, Badreddine
Halimah, M.K.

JOIV : International Journal on Informatics Visualization
Vol 8, No 3 (2024)

The optical, structural, and physical characteristics of zinc tellurite glasses doped with neodymium oxide nanoparticles, which are produced by the melt-quenching method, were examined in this work. The amorphous character of the glasses was verified by XRD analysis. Using the Pair Distribution Function (PDF) and Monte Carlo simulations and visualisation for precise molecule distribution representation, an intuitive Python interface was created to emphasize these features. The density increased with increasing Nd2O3 concentrations, from 5346 to 5606 kg/cm2. Density data was used to infer the molar volume. The best projected density was achieved by the Gradient Boosting Regressor model, with a R2 of 0.9988 and an RMSE of 0.0032; the best predicted molar volume was achieved by linear regression, with a R2 of 1 and an RMSE of 2.67e-15. These models successfully represent the correlations between dopant concentration and glass properties, advancing our knowledge of the optical properties for further glass technology research.