Designing Complex Alloys and Oxides With Targeted Mechanical and Optical Properties Using First Principles Calculations and Machine Learning 10 views

The advancement of next-generation aerospace propulsion and power generation systems is fundamentally constrained by the thermal and mechanical limits of materials operating in extreme environments. This dissertation addresses these limitations by establishing integrated computational frameworks for the design of high-performance alloys and ceramic coatings, specifically focusing on refractory high-entropy alloys (RHEAs) and thermal barrier coatings (TBCs). The first half of the dissertation addresses the mechanical property challenges of RHEAs, where the design of high-strength alloys is often hindered by the pervasive scarcity of high-temperature experimental data and over-simplified linear models. To overcome the limitations of linear empirical models (e.g., Vegard’s Law) and "black-box" artificial intelligence (AI), I developed a hybrid workflow that integrates machine learning (ML) with mechanistic solid-solution strengthening models, centering on three fundamental physical descriptors: the elastic constant, the lattice parameter, and the atomic misfit volume. To overcome the bottleneck of data scarcity, an automated workflow leveraging Large Language Models (LLMs) was developed to curate experimental datasets for BCC RHEAs, which enabled the training of surrogate models for reliable lattice constant prediction. Temperature-dependent elastic constants were determined with quantified uncertainty by combining ML and Bayesian inference with the phenomenological Varshni model. Furthermore, an active learning (AL) framework coupled with Density Functional Theory (DFT) calculations was employed to capture the non-linear electronic interactions critical for predicting atomic misfit volumes. By synthesizing these high-fidelity descriptors into a parameter-free edge dislocation model, this study provides a robust methodology to bridge experimental data gaps and enable the accelerated design of high-strength alloys across both FCC and BCC crystal structures. The second half of this work targets on blocking thermal radiation in thermal barrier coatings (TBC). As operating temperatures exceed 1800 C, near-infrared (NIR) thermal radiation becomes a significant heat transfer mode; however, traditional wide band gap insulators, such as yttria-stabilized zirconia (YSZ) and Gd2Zr2O7 are ineffective because they are transparent to such radiation. Using DFT and many-body perturbation theory (MBPT), a systematic framework is developed to investigate transition metal substitutions in YSZ and YTaO4 as a means of inducing NIR absorption. The results reveal that transition metal substitutions facilitate broadband NIR absorption through the creation of localized d-d transitions enabled by crystal field splitting. The emergence of NIR absorption originates from strongly localized excitonic effects, which allow for electronic transitions within the NIR energy range without significantly reducing the wide band gap of the host lattice. Furthermore, this work establishes that local bond distortions dictate the resulting spectral intensity and peak positioning, providing a fundamental mechanism for tuning the optical response. Collectively, these findings provide a predictive, Dieke-type "optical absorption selection map" for rare-earth-free YSZ to engineer next-generation TBC coatings with integrated thermal radiation shielding.

PHD (Doctor of Philosophy)

High Entropy Alloys; Density Functional Theory; Machine Learning; Many-Body Perturbation Theory; Thermal Barrier Coatings; Refractory High Entropy Alloys

English

All rights reserved by the author (no additional license for public reuse)

2026-04-15