Our Research
Our research focuses on understanding the relationships between structure, properties, performance and processing conditions, and the use of computational methods to support decisions in the molecular, materials and nanosciences.
The nodal surface of DNA.
Quantum Chemistry
Quantum chemical methods are the methods of choice when accurate properties and processes descriptions are required. Unfortunately these methods are computationally expensive, lack consistency, and fail to describe some fundamental properties such as interatomic and intermolecular forces.
Here in the MMM lab we are developing and testing the next generation of scalable quantum chemical methods; where geometries can be optimised, in the presence of solvents, at a level of accuracy you can trust.
More information can be found here.
Visualization of nanocatalyst evolution, in the cave.
Structure/Property Relations for an Imperfect World
When we talk about tailoring the properties of molecules, materials or nanostructures, we are actually drawing on structure/property relationships and our ability to manipulate matter. This strategy works very well when the structure/property relationships are well established and understood, and/or when our ability to control the structure has been refined. Perfect structures can delivery precise properties, but what do we do when imperfection is persistent, or actually enhanced performance?
Research is underway in the MMM lab to understand how imperfection arises in molecules, materials or nanostructures; how it can be controlled; and how it impacts properties and our ability to predict them.
More information can be found here.
Simulating entire ensembles of structures provides greater insights.
Big Data Challenges for the Science of Small Things
For the post decides the focus of much research in computational chemistry, physics and materials science has been been on making perfect structures, but perfection rarely exists outside the computer. Defects and polydispersivity is persistent in real systems, particularly nanomaterials, and are often ignored in our search for the optimal configuration; the one perfect result.
However, if treated consistently, reliably and reproducible, representing the entire configuration space does not hinder development of novel applications, and can actually help identify new industrial opportunities. The fault tolerance of structure-dependent variations in properties must also be predictable, and as important to applications as the properties themselves.
More information can be found here.
Pragmatic Materials Design
Accelerating materials discovery using computational methods can have huge pay offs, since the cost of systematically predicting new materials is orders of magnitude less costly than systematically fabricating and testing them. The problem, however, is that systematic prediction rarely works. For decades highly accurate computational methods have been used to identify optimal structure/property relationships for promising, even revolutionary, materials that (unfortunately) could never be made; stalling innovation.
A more pragmatic approach, that circumvents this problem, is to eliminate impossible, improbable or impractical candidate materials (the “non starters”) at an early stage, while retaining the cost effective advantages of computational materials design. In this project new methods combining aspects of physics, chemistry, economics and statistics will be developed, and combined in an automated workflow that can be universally applied to any material, and in any application domain.
More information can be found here.