Embedded Files
The Hill research group works in the field of theoretical chemistry. This falls into two main strands: using cutting-edge technology to study the electronic structure of molecules, and the use of artificial intelligence (AI) in self-driving labs.
We aim to answer fundamental questions at the molecular level by developing and applying theory and computational tools. We collaborate with several experimental groups, looking at systems ranging from the spectroscopy of small molecules to the dynamics of biomolecules. The work on self-driving labs is usually in collaboration with non-academic partners and has a focus on sustainable chemistry.
Our research
Our research
Methods and basis sets
Methods and basis sets
To be able to study interesting chemistry, new theories and tools need to be developed, implemented and improved. This work goes on to be used by other research groups, making computation in chemistry possible.
Automation of both computational chemistry workflows and basis set optimisation are also of interest.
Machine learning & data science
Machine learning & data science
We are investigating ways to improve current quantum chemistry approaches using the latest techniques from computer science and statistics.
We also apply data science techniques to uncover trends and find physical insights in large chemical datasets (including a collaboration with CCDC).
Self-driving labs
Self-driving labs
Working with other research groups and non-academic partners, we are using theoretical chemistry, AI and computer control to help develop self-driving labs. We take a "closed-loop" approach where automated experimental results are used to re-train models that predict which experiment to perform next.
Find out more
Find out more
Research outputs
Research outputs
We have developed a Python package for the automation of basis set development tasks. The accompanying publication demonstrates how it can be used in several common basis set development tasks.
We have developed a new method for calculating intermolecular interactions. It is fast, linear scaling (in time and storage) and similar accuracy as CCSD.
We have published correlation consistent basis sets for the heavy group 1 (K–Fr) and 2 (Ca–Ra) elements. Sets for both valence and outer-core correlation are presented.
New, efficient schemes for the prescreening and evaluation of effective core potentials. Large speedups can be achieved realtive to quadrature-based methods.
We developed a simple statistical model for predicting equilibrium interaction energies of halogen bonds. Remarkably, the model requires only one fitted parameter per molecule, yet outperforms some of the best DFT functionals.
Halogen bonds are less affected by secondary interactions than hydrogen bonds, F12 calculations and SAPT are used to provide insights into this behaviour.
Page updated
Report abuse