Oxygen Triclusters in Aluminosilicate Glasses: A Computational Approach
Open Access
- Author:
- Welch, Rebecca
- Graduate Program:
- Materials Science and Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- June 10, 2024
- Committee Members:
- John Mauro, Program Head/Chair
Seong Kim, Outside Unit & Field Member
John Mauro, Chair & Dissertation Advisor
Ismaila Dabo, Major Field Member
Wesley Reinhart, Major Field Member - Keywords:
- glass structure
oxygen triclusters
molecular dynamics
modeling
aluminosilicate - Abstract:
- Modern glasses are highly engineered materials that play a critical role in cutting-edge science and technology. Specialty glasses, with diverse applications such as liquid crystal display substrates and chemically durable pharmaceutical vials, are ubiquitous across several industries, including information technology, healthcare, and more. The demand for new glasses with tailored properties continues to increase with emerging technological needs. However, given the immense complexity of possible glass chemistries, traditional empirical approaches to glass design are often insufficient. As such, the development of computational models, such as atomistic simulations, topological constraint theory, and the statistical mechanics model for glasses, provides new approaches for understanding the glassy state and designing new glassy materials. However, their full potential and applicability are yet to be realized. This work explores how computational techniques in glass science can be used to elucidate several conflicting interpretations of glass structure in the literature. The primary focus is on oxygen tricluster formation in aluminosilicate glasses. Triclusters, or three-coordinated oxygens, are prevalent in alumina-based minerals, including mullite and andalusite, but are often considered to be simulation artifacts in glass systems. As such, the extent of tricluster formation and its influence on glass properties is severely limited. This is primarily due to difficulties in interpreting experimental results from spectroscopy methods, such as nuclear magnetitic resonance, which has led to conflicting reports of aluminosilicate glass structure when compared to simulations. With the continued improvement of atomistic simulations and various computational tools developed recently in the glass literature, a comprehensive understanding of tricluster formation and its implications on glass-property relationships can be utilized. Since these oxygen species have only been directly observed in simulations, it has been argued that their formation is caused by the parameters utilized with molecular dynamics. These criticisms are investigated herein, and results from this work conclude that triclusters are neither a simulation artifact nor a byproduct of the fast timescales used in computational approaches. In addition, we provide the first comprehensive understanding of tricluster formation across a wide compositional range including various types of triclusters that can form. To elucidate the structural impact of triclusters on glass properties, topological constraint theory was used, which determined that triclusters can have an appreciable impact on the hardness. As such, manipulating tricluster formation could have a beneficial impact on improving the mechanical performance of glass. Moreover, this work critically evaluates traditional theories and methodologies in glass science such as Zachariasen’s rules for glass formation, which directly contradicts tricluster formation. In addition, IR simulations from this work of aluminosilicate glasses and andalusite minerals challenges preconceived notions regarding peak assignments used in vibrational spectroscopy. The traditional method of using spectral features found in mineral counterparts and applying them to glasses has recently been heavily criticized in the literature. Results from this work validate these criticisms, highlighting the need for new approaches towards glass structural analysis. As such, computational modeling can be beneficial in providing a more accurate understanding of glass structure. The potential consequence is the need for a rigorous reevaluation of established paradigms which is beneficial for the future of glass design and research.