Pressure-dependent structural, mechanical, thermodynamic, and electronic properties of oxide nanomaterials: A hybrid DFT–EOS–machine learning study

Revealing pressure-dependent structural, mechanical, thermodynamic, and electronic behaviors of oxide nanomaterials is of paramount significance toward their applications in high-pressure optoelectronic, catalytic, and energy-conversion systems. A detailed first-principles and data-driven characterization of nano-TiO₂, nano-ZnO, and nano-Fe₂O₃ is presented over hydrostatic pressures from 0 to 25 GPa. The equilibrium structures, elastic constants, thermodynamic parameters, and electronic band gaps in compression were determined using density functional theory (DFT) calculations. Pressure–volume behavior and associated thermodynamic quantities were also examined with a compression-dependent equation of state based on the Grüneisen formalism. To address the computational cost of repeated first-principles calculations, a supervised machine-learning (ML) framework was proposed that uses DFT data as training input. The ML-based model can identify the pressure-dependent lattice parameters, elastic constants, bulk modulus, thermal expansion coefficient, Debye temperature, Grüneisen parameter, and band gap with near-DFT accuracy. The excellent agreement among ML, DFT, EOS, and experimental data confirmed the high quality, reliability, and efficiency of the hybrid DFT–ML–EOS approach. The findings disclose different properties in relation to pressure among the three oxides and indicate high mechanical stiffness of nano-TiO₂, higher pressure sensitivity of nano-ZnO, while nano-Fe₂O₃ exhibits clear coupling among mechanical, thermal, and electronic behavior. This integrated approach provides a scalable framework for high-throughput prediction and design of pressure-adapted nanomaterials.