Empirical models for cyclic voltammograms

Technological devices such as mobile phones and laptop computers have created an<br/>immense demand for efficient and long lasting power sources such as Lithium-ion<br/>batteries. Key to improving the current generation of batteries is the understanding<br/>of Lithium based materials that are suitable for use in batteries. Researchers investigating<br/>battery materials often plot the output from their experiments as a cyclic<br/>voltammogram. A voltammogram is simply a plot of Current against Potential.<br/>In this thesis we investigate a range of empirical models for cyclic voltammograms<br/>with a Bayesian perspective, using data from experiments carried out in the School<br/>of Chemistry, University of Southampton. This work is motivated by the lack of well<br/>formulated mathematical models for cyclic voltammograms involving a Lithium-ion<br/>compound. By setting the models within a Bayesian framework, we are able to<br/>obtain posterior predictive distributions for characteristics of the voltammogram of<br/>interest to chemists.<br/><br/>Markov Chain Monte Carlo sampling methods are used to explore the posterior<br/>distribution of the model parameters and to estimate the posterior predictive distributions.<br/>We investigate four methods of modelling the experimental data: multiple<br/>regression models for summary statistics, autoregressive models, sinusoidal models<br/>and stochastic volatility models. The application of Bayesian model choice techniques<br/>showed that the sinusoidal model provided the best description of the data.

Name dc:type.qualificationname

Ph.D.

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

University of Southampton

Year dc:date.issued

2010

Author dc:creator

• Samuel, Jeffrey J.

Advisor dc:contributor.advisor

• Sahu, J.K.