Scientists from Diamond Light Source and the ISIS Neutron and Muon Source have contributed to the development of the world's first open-access battery imaging library. The library was created through an international collaboration led by Dr Antony Vamvakeros, Royal Society Industry Fellow at Imperial College London and Research and Development Lead at Finden Ltd , which is co-located with both facilities on the Harwell Science and Innovation Campus.
The Battery Imaging Library (BIL) contains over 4.5 terabytes of experimental data acquired at various synchrotrons and national laboratories. The initiative brought together 48 scientists from 18 institutions around the world to acquire experimental battery data using 13 distinct imaging modalities at different length scales.
Using Diamond's I12-JEEP, the UK's flagship high-energy X-ray imaging and diffraction beamline, the team was able to acquire high resolution static and dynamic X-ray micro-CT measurements. The team scanned commercial cylindrical Li ion, LiFeS2 and alkaline batteries to reveal their internal structures. Prof Genoveva Burca, Principal Beamline Scientist at I12, said: "The I12-JEEP beamline offers unique capabilities for probing and investigating the internal morphological structure of batteries. We are delighted that these dynamic X-ray micro CT datasets are now freely available through the Battery Imaging Library so that scientists worldwide can use them to develop safer and more efficient energy storage devices."
The IMAT neutron imaging beamline at the ISIS Neutron and Muon Source was used to collect neutron-CT data, which is particularly valuable as it is rarely available to the battery community. It provides complementary contrast to X-rays for light components in Li-ion batteries, enabling the visualisation of Li-based electrolytes and the location of the Li species in general within the cell. By including data from multiple techniques, the library enables scientists to get the full picture required for their analysis.
Rather than keeping these valuable files hidden away on private hard drives, the project ensures that anyone can access both the raw data and reconstructed images. The collection is designed to help researchers validate new computational models and test image analysis software. In the accompanying publication, Battery Imaging Library: Multi-length scale and multi-modal synchrotron and laboratory battery imaging data for all , the team highlight the library's core aims:
widening access: making data freely available to researchers at all stages of their careers
improving method development: providing realistic datasets to test and improve imaging and analysis techniques
supporting learning: giving students and professionals practical data and workflows for training and education
Lead author Ronan Docherty, a PhD student in the Centre for Doctoral Training for the Advanced Characterisation of Materials at Imperial College London, took on the challenge of making this vast amount of information easy to browse by creating a dedicated website, designed to make data easier to find, access and re-use.
Ronan said: "A major challenge with existing experimental datasets is simply finding what you need. We developed the library's website to solve this problem, allowing users to browse the different modalities and quickly access the specific data relevant to their research."
Dr Sam Cooper, Professor of AI for Materials Design in the Dyson School of Design Engineering at Imperial College London and Chief Scientist at Imperial spin-out, Polaron, also a co-author of the work, said: "Open data is critical for training the AI models that will accelerate the next generation of scientific discovery. The necessary infrastructure has been available for sometime already, but the culture in the scientific community has lagged behind. I hope BIL will inspire other groups to follow suit."
Dr Antony Vamvakeros, corresponding author of BIL, serves as a Royal Society Industry Fellow at the Dyson School of Design Engineering at Imperial College London and R&D Lead Scientist at Finden Ltd. He created the library, coordinated the international team and performed most of the synchrotron and neutron experimental work with his former Imperial PhD students, Dr Sam Riley and Dr John Morley.
Antony said: "Together with Professor Sam Cooper at Imperial, our ambition was to open up these complex imaging datasets to the entire scientific community. By sharing this extensive collection, we hope to accelerate the development of machine learning methods for experimental data. For example, researchers can use this resource to train self-supervised neural networks for data denoising, super resolution and data fusion."
