Visagan Ravindran 

PhD Student

What do you do?

I am a PhD student working with Dr Nikitas Gidopoulos and Professor Stewart Clark to better understand and refine the theoretical framework of density functional theory (DFT).

My research focuses on implementing theoretical ideas and numerical algorithms within the CASTEP DFT code.

Outside the department, I enjoy exploring the great British outdoors and can often be found on various hills throughout the day with the university hill-walking society.

How are you involved in this area of science? 

DFT is ubiquitous in modelling materials at the level of first-principles and has enjoyed great success over the last few decades. DFT is an exact theory in principle but requires approximations for the exchange-correlation (XC) potential*. Although great strides have been made in understanding and refining these approximations, their behaviour when applied to different materials, particularly in so-called strongly-correlated systems which are challenging for DFT approximations to describe, can be, at times, mysterious.

We study the behaviour and devise/implement various DFT approximations and other theoretical ideas using the CASTEP code. One such technique is density inversion to reverse-engineer the XC term from an input density. When applied to other computational techniques such as Quantum Monte Carlo**, which gives nearly exact densities, an exact XC potential can be reverse-engineered, against which other approximations for XC can be benchmarked.

What do you love about this topic?

I feel a great sense of accomplishment when implementing algorithms within the CASTEP code. It is very satisfying when equations on whiteboards are turned into lines of code that when executed, yield the values for properties which experimentalists can then measure within the real world. CASTEP is also a code that is developed by various groups around the UK and used by even more worldwide; contributing to the code makes me feel like part of a much wider community.

How does this work deliver real-world impact?

Quantum mechanics is at the heart of physics and chemistry. DFT, in conjunction with robust implementation in codes, makes quantum mechanics for systems of many atoms and particles available for the masses.
The theoretical advancements over the last few decades in DFT have enabled an accurate description of more exotic functional material. These materials, designed to solve specific real-world problems, such as more powerful computer chips and more efficient energy-storage devices, can now be prototyped accurately on a computer first, potentially greatly saving in R&D costs.

*often referred to as "nature's glue", see Kurth and Perdew (2000). This is because it is a small portion of the total energy of a many-body system (atom, molecule, solid/material) yet turns out to be really important for a lot of the physical and chemical properties (we know this because if we ignore it, the predictions of the theory when we do calculations are completely wrong!) 

**Monte Carlo refers to a family of mathematical methods that utilise random sampling in some way to solve computational problems. In the case of QMC, this refers to essentially calculating integrals (effectively area under the curve of functions) which become prohibitively expensive in higher dimensions, e.g. on a grid.

Find out more:

• Watch a short video on 'How to simulate the universe using DFT'