Integrated classical/quantum–classical neural networks, DFT, and experiments for predicting bandgap and color of CrxSbxTi1−2xO2 yellow-orange pigments
Year: 2026
Type: article
Abstract: The accurate prediction of the bandgap (\((\text{E}_{\text{g}})\)) and color of ceramic pigments is of considerable importance in materials science research. In this work, we investigate the \((\text{E}_{\text{g}})\) and color of Cr\((_x)\)Sb\((_x)\)Ti\((_{1-2x})\)O\((_2)\) rutile pigments by combining experimental, density functional theory (DFT and DFT+\((U)\)) calculations, and neural network approaches. The main novelty of this study is the application of classical/hybrid quantum--classical neural networks to the prediction of pigment color and \((\text{E}_{\text{g}})\). Cr/Sb co-doped rutile pigment was synthesized and characterized, and the electronic structures of undoped and doped \ce{TiO2} were analyzed using first-principles calculations. In parallel, classical and hybrid quantum--classical neural network models were developed and trained on datasets compiled from the Materials Project database and literature-reported rutile-based pigments. The models successfully captured the trends in \((\text{E}_{\text{g}})\) and chromaticity, with the hybrid architectures providing slightly improved predictive performance and more consistent agreement with experiment. Overall, this work demonstrates an effective framework that links experiment, first-principles calculations, and classical/quantum--classical neural networks for predicting key properties of ceramic pigments, and it points toward the future use of fully quantum machine learning models for materials design.
Source: Scientific Reports
Institution: Yazd University
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
Funder: Yazd University
OpenAlex ID: W7215037802