Uniting statistical physics, quantum mechanics, and artificial intelligence to explore matter at the atomic level.
The Group Mission
The Data Driven Materials and Molecular Science (DDMMS) group is hosted within the Scientific Computing Department (SCD) of the Science and Technology Facilities Council (STFC), part of UK Research and Innovation (UKRI), based at Sci-Tech Daresbury.
Computational materials science has long faced a fundamental trade-off: high-accuracy quantum mechanical calculations (such as density functional theory and post-Hartree-Fock) are computationally expensive and limited to small systems, whereas classical empirical force fields scale to millions of atoms but suffer from fixed functional forms and limited chemical transferability.
Our core mission is to eliminate this trade-off by constructing robust, physics-informed machine-learned interatomic potentials (MLIPs), creating automated calculation workflows, and running extreme-scale simulations on high-performance computing facilities.
Methodological Pillars
Our research combines rigorous physics with modern data science across four interconnected layers:
1. Machine Learning & AI
Equivariant graph neural networks (MACE, SevenNet, CHGNet), polarizable foundation models, and active learning strategies.
2. Atomistic Simulation
First-principles DFT, nonadiabatic dynamics with CP2K, and massive-parallel molecular dynamics with DL_POLY 5.
3. Materials Discovery
Metal-organic frameworks (MOFs), molten salts for green energy, negative thermal expansion, and heterogeneous catalysts.
4. Open Science & FAIR
Modular, reproducible software pipelines published under permissive open-source licenses for the international community.
Collaborative Ecosystem & PSDI Data to Knowledge
We are an active development partner in the UKRI Physical Sciences Data Infrastructure (PSDI) under the Data to Knowledge program. Through projects such as Goldilocks (goldilocks.ac.uk), we develop tools, machine learning representations, and optimal parameter datasets to enhance the efficiency, reproducibility, and sustainability of electronic structure calculations across the UK research community.
We also work closely with the Collaborative Computational Project for computer simulation of condensed and materials phases (CCP5), the ISIS Neutron and Muon Source, the Diamond Light Source, and academic institutions worldwide.