Investigating the generation and properties of model network systems

Thesis

Understanding the structure and properties of thin film materials is an area of continued research, due to the experimental synthesis and characterisation of carbon, boron nitride, silica and aluminosilicate thin films. The structures of these films can be atomically resolved, providing a high level of structural detail for small sample areas. These experimental images show atoms arranged in network structures, interpreted as a percolating net of rings. This atomistic structure has a direct impact on mechanical and permeation properties, meaning the ability to characterise and control the structure has the potential to unlock a wide range of technologically useful properties.

The overall aim of this thesis is to improve the simulation and characterisation of experimentally realisable two dimensional disordered materials. With accurate means to simulate a range of two-dimensional structures, we can begin to understand the feasibility of using these materials effectively for a range of applications. Characterising known and potential structures is crucial to understanding how well our simulations fit with experimentally observed structures. This is accomplished in this work by examining a range of metrics, both empirical laws (namely Lemaître’s law and the Aboav-Weaire law) and graph theory metrics, evaluated by interpreting the atomic systems as mathematical networks.

In this thesis, low level potential models are developed for thin films of graphene, boron nitride and bilayer silica systems. Example structures are generated across a range of disorder using a bond switching algorithm, and evaluated against experimental samples where present, with the goal of evaluating the effectiveness of our potential models. These potential models are then applied to systems with ‘pores’, to generate structures predicted to have technologically useful permeation properties. Both periodic and aperiodic approaches are used to predict the structure and stability of graphene and silica bilayer pore structures with pore density. Continuing work on silica bilayers, this thesis then establishes an escalating potential model to generate and evaluate a broad range of bilayer configurations efficiently. Moving from computationally 'cheap' to 'expensive' potentials allows us to sequentially narrow the phase space of interest, providing an efficient high-throughput method. The data set generated allows for an understanding of energetic and structural trends with disorder. In addition to disordered structures, novel crystalline bilayer structures derived from zeolite structures are proposed and evaluated. These structures are predicted to approximate zeolite pore sizes, and have similar permeability characteristics. Novel techniques for using procrystal methods are developed and introduced, allowing for the generations of sample three-dimensional silica structures.

Authors: Whitaker, OP

• DOI: 10.5287/ora-5rn95mrbg
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: rigid truss; procrystals; Euler characteristic; diffusion; graphene; pore evaporation; boron nitride sheets; zeolites; silica bilayers; corneal endothelium; network theory
• Subjects: Zeolites--Absorption and adsorption; Silica; Corneal endothelium; Zeolites; Boron nitride; Chemistry, Physical and theoretical; Graphene; System analysis