Metal-organic frameworks (MOFs) have become a widely studied class of porous materials over the last 30 years. They have been, and continue to be, extensively researched for applications requiring porous and adsorptive properties of materials. This thesis focuses on three main topics which sit at the forefront of MOF research: reliable topological characterisation of crystalline materials, machine learning and data driven manufacturing, and the development of new computational tools.
Firstly, we investigate topological characterisation of these crystalline materials, describing the methods and algorithms by which they can be categorised through the
medium of a perspective review, followed by the integration of a newly developed open-source software with the Cambridge Crystallographic Data Centre’s (CCDC) existing crystallographic data suite and Python API.
This is succeeded by the introduction of state-of-the-art machine learning (ML) and digital manufacturing techniques with the view that they can be applied to the future of the field. In this work we discuss the use of ML in solid state materials development which is followed up by work in which we developed a new method of abstracting existing synthesis information published in thousands of previous MOF studies.
Lastly, we apply new augmented reality techniques to visualise the results of topological deconstruction and adsorption studies of MOFs. The result of this work over the last 4 years has enabled researchers to use the Cambridge Structural Database (CSD) to maximum effectiveness when searching for synthesis conditions, precursors, linker types, topologies, and more whilst also integrating ML techniques such as Natural Language Processing (NLP) for data mining and introducing new ways of visualising the results.
