Models of metal-doped zinc oxide for low-E glass coating applications

The main objective of the thesis was to investigate the structural effect of Al as a dopant for ZnO when used in the production of low-emissivity coatings. The work involved first developing a method to fit a (classical) variable charge reactive force-field potential for the Al-ZnO system which was later extended to Ti-ZnO. This was achieved using a ReaxFF model by carrying out a large number of ab initio calculations and fitting parameters of the model to match energies and forces in selected systems.

The particle swarm optimisation model was found to be the best method for fitting parameters in a high dimensional space with the best accuracy when only function evaluations and not derivatives were used. This was achieved through comparisons with other traditional optimisation models.

The developed model for the Al:ZnO system was used together with a previously developed model for Ag:ZnO to investigate energetic impacts on the surface of pure ZnO. This involved setting up a database of all potential interactions over an energy range typical for magnetron sputtering deposition with energies between 0.1 and 40 eV using molecular dynamics. Depositing both Al and Ag on the surface allowed study of outcomes such as adsorption, penetration or reflection at different angles. Energies lower than 10 eV showed that adsorption was the most likely outcome regardless of angle for both Ag and Al.

The final part of the work applied the Al:ZnO potential to investigate the effect of small amounts of Al dopant on the growth of ZnO using a combination of molecular dynamics and an adaptive Monte Carlo technique. This allowed the ability to model experimental time scales to improve the growth of the Al:ZnO. This preliminary investigation shows that Al will strongly bond with oxygen which helps pin other arriving atoms but that this pinning does not especially effect the structure of the growing ZnO film.