In my research, I normally use computer simulations and classical statistical mechanics to study the behaviour of molecular and colloidal systems. In particular, I am interested in nucleation and self-assembly – and combinations of the two – and how thermodynamic and kinetic factors affect and control them.

Self-assembly is a feature of microscopic systems: we would be quite justifiably surprised if, having shaken a box containing the loose component parts of a model aeroplane, we ended up with the fully assembled model. At the microscopic scale, however, such processes are not particularly surprising, and the spontaneous formation of complex virus capsids from simple protein building blocks, for instance, is a well-known and well-studied example of self-assembly. Artificially constructed materials that can spontaneously self-assemble into their target structures have also been studied extensively, and some amazing structures have been shown to form from simple building blocks: from closed clusters to complex crystals, liquid crystals and even quasicrystals.

However, there is a limit to how far simple building blocks can go in terms of the complexity of the self-assembled target structures that are possible to form in practice. It came as something of a shock to the self-assembly community when a recent experimental study [Science 338, 1117 (2012)] showed that potentially thousands of distinct DNA molecules can reproducibly self-assemble into complex, fully addressable, nearly error-free target structures. This was not meant to happen, as it had been well-established that having such a large number of distinct components would lead to self-poisoning and in turn a kinetic arrest of the growth of the target structure.

Much of my recent research has focussed on why structures like these can form successfully whilst other systems have failed to achieve anything even approaching such complexity. Following our simulation-based and theoretical work on model systems with the same type of connectivity and specificity as can be achieved with short DNA molecules, we have begun to understand how nucleation plays a very significant role in why such systems can self-assemble so successfully. I am currently continuing this work to try and understand how we can exploit the system's properties to guide the nucleation pathway and design the target structure in a rational way, as well as to learn more about the specific properties that govern this type of self-assembly that might be applied to different building blocks – such as colloids – in the future.