14 May 2012
Who or what inspired you to pursue a career in science?
There wasn't one specific person or thing that inspired me to become a scientist, but lots of little nudges. Science was something which fascinated me from a young age - I remember reading the Junior Encyclopaedia of Science and being blown away by the concept that everything was made of atoms, and similarly, later, by Stephen Hawking's A Brief History of Time.
I had various inspiring teachers throughout primary and secondary school who encouraged and nurtured my interest in the sciences. I think ultimately, what motivated me to pursue science as a career was curiosity about the world we live in. The world is so much richer and more exciting when you look at it through the lens of understanding.
What is your research topic?
My research focuses upon the development of methods for computational chemistry, specifically methods that explicitly model the repulsive interactions between pairs of electrons, called explicitly-correlated methods. Computational chemistry allows us to model chemical systems in great detail, using powerful computers and complex software to simulate mathematically the structure and interactions of molecules, helping us to understand the ways these systems behave in the physical world.
The methods I work on use quantum mechanics to simulate atoms and molecules in great detail, down to the interactions of the electrons on individual atoms. For accurate results, it is important to model the repulsive interactions of pairs of electrons, electron correlation, but this is computationally expensive. I am currently working on an approximate method of modelling electron correlation, which differs from currently established methods in the way that computationally-expensive many-electron integrals are approximated.
What is innovative about your most recent research project?
The current standard method for approximating the expensive many-electron integrals, which arise in explicitly-correlated methods, is called resolution of the identity (RI), whereby the many-electron integrals are factorised into more manageable 2-electron integrals. At this point, there is no competitive alternative to using RI in this way.
In my current research project, I am attempting to implement explicitly-correlated methods which approximate many-electron integrals without using RI at any stage, and the hope is that the resulting method will offer the same accuracy for a lower computational expense.
Are there any potential applications or any competitive advantages for industry as a result of your research?
Pharmaceutical companies are increasingly looking to theoretical methods to inform and support rational drug design, and to model the interactions of drugs with biological molecules. The design of novel materials can also be informed by computational studies.
Theory and mathematics are the foundation of these computational applications - better approximations and more effective implementations lead to better results, faster calculations and ultimately better science.
So, while method development in computational chemistry might seem a little esoteric and far removed from industry, I would argue that method development is increasingly vital.
What have been your proudest achievements so far?
Probably my proudest achievement has been receiving first class honours for my undergraduate degree at Bristol. It was very difficult, but absolutely worth the effort!
What is the next milestone in your career?
Completing my PhD. After that I am interested in gaining some experience outside of academia, using the skills and knowledge I've acquired during my research in an industrial/commercial setting. I hope to work in the field of high performance scientific computing, perhaps developing scientific software.
What does winning an SCI scholarship mean for you?
Aside from the financial award, the opportunities for networking and the membership of SCI should help me develop my career beyond my PhD. I am also enjoying being a member of SCI, with all the events and meetings that entails.