Decoding Disorder: Structure–Property Relationships in Amorphous SiO₂–TiO₂ Hybrid Thin Films
Amorphous SiO₂–TiO₂ thin films are gaining attention as versatile materials for multifunctional coatings, offering a unique combination of optical transparency, photocatalytic activity, and mechanical durability. This DPhil project aims to develop and investigate these hybrid films at the atomic level, using a combination of advanced synthesis and computational modelling. The films are fabricated via sol–gel chemistry and thermal processing, allowing for precise control over porosity and shell architecture. These structural features are critical for enhancing light scattering, surface area, and wettability—key properties for applications in self-cleaning surfaces, photovoltaics, and environmental remediation. However, despite their promise, the fundamental structure–property relationships in amorphous SiO₂–TiO₂ systems remain poorly understood. To study the structure–property relationships, advanced analytical techniques will be employed to investigate structural features, e.g., porosity, shell architecture, and atomic disorder using advanced electron microscopy techniques and X-ray diffraction to confirm amorphous phases. Raman spectroscopy will help identify bonding environments, including Ti–O–Si linkages. Compositional analysis will be conducted using X-ray photoelectron spectroscopy and energy-dispersive X-ray spectroscopy, providing insights into elemental ratios and chemical states. Optical and surface properties will be assessed through UV-Vis spectroscopy, ellipsometry, contact angle measurements, and atomic force microscopy, revealing information about transparency, refractive index, wettability, and surface roughness.
A major challenge lies in the difficulty of linking atomic-scale disorder to macroscopic behaviour in amorphous materials. Unlike crystalline systems, which have well-defined periodic structures, amorphous materials remain challenging to simulate, let alone to “design” computationally [see, e.g., https://doi.org/10.1038/s41578-024-00754-2]. To address this challenge, we will leverage expertise in VLD’s group to fit machine-learning-based interatomic potential models using graph network models and model “distillation” into faster architectures. By correlating computed structural features with experimentally measured properties such as refractive index, hydrophilicity, and photocatalytic efficiency, the project aims to establish predictive design rules for tailoring the performance of amorphous hybrid films. This interdisciplinary approach—combining precise experimental synthesis with cutting-edge machine learning simulations—offers a novel pathway to understanding and engineering disordered materials.
While crystalline TiO₂ has been extensively studied, the amorphous phase, particularly in combination with SiO₂, remains underexplored both experimentally and theoretically. The atomic-scale interactions between silicon and titanium, especially the role of Ti–O–Si linkages, are not yet fully understood, limiting our ability to optimize these materials for real-world applications. By addressing this gap, the project not only advances the fundamental science of amorphous oxides but also contributes to the broader field of sustainable materials design, with potential impact across energy, environmental, and optical technologies.
