Crystal structure prediction for rechargeable battery anodes

This thesis presents results of high-throughput crystal structure prediction calculations to predict the compositional phase diagrams of material systems that have relevance as conversion anodes in sodium- and potassium-ion batteries. The aim is to discover phase diagrams that can balance high capacity with low volume expansion (thus limiting electrode cracking and other degradation routes). By combining ab initio random structure searching, evolutionary algorithms and data mining approaches with high-throughput density-functional theory calculations, accurate predictions of the chemical and dynamical stability of possible atomic configurations are made across several promising phase diagrams. Firstly, the foundations of first-principles modelling of rechargeable batteries and the methods of crystal structure prediction are introduced. Results on phosphorous conversion anodes for sodium- and potassium-ion batteries are then presented and compared to the literature. In the sodium phosphide case (Na–P), nuclear magnetic resonance (NMR) calculations on predicted metastable structures allow for tentative assignment of local phosphorous environments observed in operando NMR measurements during electrochemical cycling. For the potassium phosphides, careful energetic comparisons and uncertainty estimates reveal several previously unknown compositions (KP₇, K₃P₇, K₅P₄) that are predicted to be stable, including the high capacity endpoint composition K₃P. Following this, the K–Sn–P ternary system is then studied in depth, with comparison to recent results from electrochemical cycling experiments. Several novel structures are predicted to be stable the K–Sn–P space, with compositions of KSnP, KSn₃P₃, K₅SnP₃ and K₈SnP₄, with many additional low-lying metastable phases (KSn₂P₂, K₂Sn₃P₃), as well as new binaries in the K-Sn space. Finally, a practical discussion of research data management and dissemination in materials science is presented through the lens of the OPTIMADE specification and related software endeavours.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2023

Author dc:creator

• Evans, Matthew L

Advisor dc:contributor.advisor

• Morris, Andrew

• crystal structure prediction
• density-functional theory
• materials databases
• nuclear magnetic resonance
• rechargeable batteries

Licence

• https://apollo8-f-pro.lib.cam.ac.uk/bitstreams/cbd65a1f-f173-40df-a434-087d46f9fa94/download
• https://creativecommons.org/licenses/by/4.0/

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.104811

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/362882