Use this URL to cite or link to this record in EThOS:
Title: Modelling nanowires : crystals encapsulated in carbon nanotubes
Author: Brown, Samuel F.
ISNI:       0000 0004 6348 6909
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Nanotube-encapsulated ionic materials can be usefully modelled using ab initio methods to inform and validate experiment. This thesis models these materials using density functional theory, explores the specific challenges these materials pose, and presents results only visible once these challenges are addressed. Chemical potentials and ab initio calculations are used to predict tube radii at which filling occurs for a range of filling conditions, which informs our choice of tube. The practice of fitting both ionic core and encapsulating tube into the same unit cell causes artificial strain. A method of quantifying this mismatch is developed and the associated strain energy and distortions predicted. Current methods of approximating the nanotube treat it as a smooth cylinder. We find significant texture which can resist core movement, allow metastable twisted states, and be used to determine which supercells have physically realistic mismatch strain. Cores have previously been relaxed in vacuo with limited justification. We find that, in general, cores are not stable without constraints. We guide three bare silver iodide ansatze to metastable saddle-points by applying artificial symmetry constraints during in vacuo relaxation. An investigation into their unconstrained unstable phonon modes reveals them to be stabilised by an encapsulating tube. Two classes of stabilisation are identified, one due to radial encapsulation and a second due to mismatch-like effects. Charge of a few tenths of e per core ion is seen to transfer from core to tube, with magnitude depending on tube radius. In order to approximate the charge field by atomic charges, four population analysis schemes are investigated. The Bader and Density Derived Electrostatic and Chemical (DDEC) population analyses agree to physically sensible values while using different approaches. A machine-learned model is trained using local chemical-environment descriptors to transferrably predict atomic charges to within 0:004 e of DDEC values with an orders-of-magnitude speedup.
Supervisor: Not available Sponsor: Engineering and Physical Sciences Research Council
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: QC Physics