If a dislocation is present in the active volume of a light emitting device, it causes failure by acting as a carrier recombination pathway and grows through the material by emitting atoms, eventually quenching all luminescence. Despite the significant technological progress improved knowledge would generate, the atomistic mechanisms underlying this recombination-enhanced mechanism of dislocation climb and its interaction with vacancies and interstitials are poorly understood, with no first principles work reported to date. This PhD project will address this deficiency for the first time.

How do metals break? How can we make them stronger? What are the roles of defects and impurities? The strength of materials are ultimately determined by the microscopic interactions on the atomic level, which can be modelled accurately. However, the challenge is that computationally it is not possible to propagate information in one step from the nanometer to the millimeter scale. In this project, you will use combined Quantum Mechanics-Molecular Mechanics and Gaussian Approximation Potentials, a machine learning approach, to develop coarse-grain models of dislocations and to make quantitative predictions of plastic deformations in metals and alloys.

Understanding the crystallisation of polymers is crucial to improve their functional properties. For instance, the strength-to-weight ratio of Kevlar depends on its degree of crystallinity, which impacts its usage in composite materials such as F1 chassis. However, the current theoretical frameworks for the crystallisation kinetics of polymers suffer from severe limitations when dealing with non-isothermal conditions or the presence of nucleating agent. This project seeks to break new ground by using multi-scale simulations (from molecular dynamic simulations to lattice models) to build a machine learning model for predicting the non-isothermal crystallisation of heterogeneous polymer mixtures.

In-service performance of functional materials for energy storage-related applications depends on an intimate interplay between various physical-chemical processes, e.g. strain-dependent ionic transport in polymer electrolytes. Classical electrochemical-mechanical (ECM) models, used to predict the optimum performance of those materials, are based on ad hoc assumptions, omitting the origins of the ECM coupling at the nanoscale.

Therefore, the ambition of this research project is to develop new foundations for a holistic computational modelling framework that rigorously captures the ECM coupling by bridging vastly different length scales via multiscale simulations of the material and machine learning. Particularly, the transport of charged species taking place in the presence of stress, electrostatic, and chemical potential gradients will be explored from nano to macro scale.

Two thirds of all energy we use is lost into heat during conversion processes, a loss which puts enormous pressure on the planet and energy sustainability. Thermoelectric materials convert waste heat into electricity and can provide a solution towards this problem. This project studies the electronic and thermal properties of complex electronic structure materials subject to a large degree of nanostructuring, which is believed to be the most promising way to achieve extremely high conversion efficiencies. Density function theory methods, as well as semiclassical and quantum transport methods are merged to explore the design space of new generation thermoelectric materials.

Some topological magnetic structures can behave as quasiparticles and be manipulated to store information. The discovery of skyrmion-type particles was a huge leap forward opening up the prospect of new smaller and more efficient devices. In this project we will identify atomistic spin models by using machine learning tools on data from calculations of the fundamental quantum mechanics of the electrons and study them to discover what complex but stable magnetic objects can emerge in some choice materials. These developments have wider scope too – for modelling the thermal properties of magnetic phases and their response to applied fields for new solid state cooling applications. 

Functional materials for catalysis and magnetooptical applications often contain precious materials, the scarcity of which is a strong motivation for maximizing the efficiency of their use. Single atoms can act as catalysts (Single atom catalysts, SACs) or single atom magnets (SAMs) when stabilized on well-defined substrates. SAC/SAM materials are often studied with X-ray photoelectron spectroscopy and X-ray absorption spectroscopy, but the rich structure and overlapping features make these spectra hard to disentangle. This project will develop new simulation methodology to predict such complex spectroscopic signatures and, in collaboration with experimental partners, novel SAM and SAC materials will be characterized.

This project explores the exciting and novel physics of multilayer structures built from 2D materials. 2DMs can undergo dramatic changes in their fundamental physical properties when they are combined into heterostructures, particularly if their lattices are misaligned: an example is how graphene becomes superconducting when the alignment of two layers is “twisted” by specific angles of a few degrees. This new field of “twistronics” explores how properties of 2D materials can be tailored for specific applications by stacking them together. We will harness unique capabilities of Linear-Scaling DFT to design 2DM heterostructures “ab initio” for future application in electronic devices that combining high performance and ultra-low power usage. There will be opportunities to use Machine Learning tools to accelerate these simulations, and to develop theoretical spectroscopy methods that enable prediction and interpretation of state-of-the-art experimental results.

We are surrounded by situations that depend on a controlled outcome in our day-to-day lives, ranging from controlling the evacuation of crowds, to efficient drug delivery, or cooling systems inside computing centres. Most real-life scenarios rely on complicated models which are too complex to tackle analytically or computationally. Using the framework provided by a beautiful and rich physical problem – controlling nonlinear waves in falling liquid films – the project will provide opportunities to develop analytical and computational multi-physics tools. Acting in tandem for the first time, they become sufficiently powerful to translate robust theoretical strategies into realistic technological solutions.

Spatial and temporal fluctuations in fluid behaviour control mixing and reaction processes. Observations show that velocity fluctuations are correlated with mixing and reaction rates, and degree of these correlations vary across the scale. These provide compelling evidence that the key statistics of reaction parameters driving the transport processes are scale dependent and functions of the increments of porous media geometrical characteristics. This project aims to develope theoretical and computational frameworks to assimilate data associated with diverse variables collected at a range of scales and combine these to provide predictions of reactive solute dynamics and quantify associated uncertainties across the scales.

The candidate will benefit from HetSys CDT training on computational data science and geostatistical methods and will perform research in the field of reactive transport and turbulent mixing processes in porous media in collaboration with partners at Bayer A.G. (Germany).

Biological systems often need to switch behaviour due to environmental (e.g. temperature) and other stimuli (e.g. cell-cell interactions). Yet, understanding how cells can make reliable decisions given the inevitable noisiness of the intracellular conditions remains an open problem. Here, we will develop stochastic models of small biochemical networks that undergo stochastic switching in behaviour. The project involves building a robust software framework for exploring the role of spatial and temporal fluctuations in protein number within different boundary constraints. Working closely with experimentalists (Loose lab, IST-Austria), we will generate predictions that can then be directly tested.

The best pharmacologically active compound may have been found, but its function, properties and processability are heavily dependent on the solid form. Solid-state Nuclear Magnetic Resonance and diffraction experiments, coupled with quantum mechanical calculations are powerful tools to elucidate the atomic structure, which is needed to understand and control macroscopic properties. The challenge in amorphous systems is that experiments only provide globally averaged measurements, while large length scales are needed to capture the disorder, which render quantum mechanical calculations unfeasible. In this project, machine learning models will be used to bypass the need for these costly computations, allowing interpretation of experimental data and structure determination with unprecedented accuracy.