Computational framework for metastable materials discovery: integrating kinetics and accurate properties

Computational crystal structure prediction has led to the discovery and experimental realization of many new materials with exceptional properties and technological promise. One of the remaining challenges in computational materials discovery is the prediction of metastable materials: long-lived materials with an atomic structure that is not the global minimum of the relevant thermodynamic potential. Identification of such materials requires understanding their statistics and thermodynamics to understand their realizability, their phase transformation kinetics to determine if they are long-lived, and their properties to understand their technological relevance. The realizability and thermodynamics are assessed via ab initio random structure sampling, while kinetics are assessed using atom-to-atom structure mapping and solid-state nudged elastic band. These coherent transformations effectively classify whether a structure will be long-lived. Herein, each of these key steps is performed across the family of group-IV carbides (SiC, GeC, and SnC). A common family of tetrahedral polymorphs is identified, along with a high-pressure rocksalt phase. The tetrahedral polymorphs are long-lived, while the rocksalt transforms rapidly to the ground state. SnC is predicted to have exceptional electron mobility via accurate electron-phonon coupling calculations. This computational approach also reveals a thermodynamic and kinetic cause for the absence of polymorphism in compounds with the rocksalt ground state across different chemistries. Given that many other polymorphs transform quickly to the rocksalt structure in these compounds, this structure's probability is far greater than that of any other. This description of a single crystallographic structure with kinetics that prevent the metastability of other polymorphs offers a new tool to classify whether compounds will exhibit polymorphism. This tendency also explains the dominance of the rocksalt structure in high-entropy ceramics and ternary nitrides with a 1-to-1 metal-to-nonmetal ratio. Beyond polymorphism, amorphous materials are important metastable states that can be represented via an ensemble average of structures from random sampling. We show that this representation offers a method to accurately compute electronic structure of amorphous materials. A band-like conduction state and wave-like electron transport are accurately reproduced in amorphous In2O3.

Name thesis:degree_name

Doctor of Philosophy (Ph.D.)

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Metallurgical and Materials Engineering

Grantor dc:publisher

Colorado School of Mines. Arthur Lakes Library

Year dc:date.issued

2025

Author dc:creator

• Jankousky, Matthew Charles

Advisor dc:contributor.advisor

• Stevanovic, Vladan

Committee members dc:contributor.committeemember

• Brennecka, Geoffrey
• Toberer, Eric
• Ciobanu, Cristian V.

• computational
• disorder
• kinetics
• materials
• metastable
• polymorph

Statement dc:rights

• Copyright of the original work is retained by the author.

Language dc:language.iso

eng, English

Identifier

T 9995

OAI identifier oai:identifier

oai:repository.mines.edu:11124/181193