Integrating experiments and advanced machine learning to predict annealing-tunable optoelectronic properties of Ni0.6Mg0.2Co0.2FeCrO4 spinel ferrites

This study comprehensively investigates the impact of annealing temperature (850°C–950°C) on the structural and optoelectronic properties of sol–gel synthesized Ni 0.6 Mg 0.2 Co 0.2 FeCrO 4 spinel ferrites using an integrated experimental-AI approach. XRD and SEM analyses confirm a single-phase cubic spinel structure with significant crystallite growth (from 124 to 183 nm) and lattice expansion as temperature increases. UV–Vis-NIR spectroscopy reveals highly tunable optical absorption, penetration depth, and reflectance across the 200–2400 nm range. To model these complex non-linear relationships, seven machine learning algorithms were evaluated, with Least Squares Boosting (LSBoost-GBM) and Feedforward Neural Networks (FNN) achieving near-perfect predictive accuracy (R 2 ≈1.0000 and 0.9993, respectively). These results demonstrate that the integrated experimental-AI framework provides a robust and efficient pathway for predicting material behavior and accelerating the design of tailored spinel ferrites for targeted optoelectronic applications.