Students in park (c) Jonas Kron


Learning Multi-Scale Optimisation for SiO2-based Anode Materials

Short Facts

  • Duration: 01 Feb. 2021 to 31 Jan. 2024
  • Funded by: Bavarian State Ministry of Economic Affairs, Regional Development, and Energy
  • Funding programme: BayVFP materials
  • Joint project


In the development process of materials for lithium-ion batteries, the complex interaction of many parameters and burdensome synthesis steps, we aim to come up with a (hopefully) high-performance energy storage. As the requirements for their application increase steadily, it is essential to optimise such processes, materials and their results. In search for improvements, a clear set of possibilities regarding the composition of available materials becomes apparent quickly which, each for itself, must be processed, tested, and validated in time-consuming and cost-intensive processes. By means of artificial intelligence, highly complex relationships between numerous parameters can be identified and interpreted, letting them contribute to a significant improvement of the current state of technology.

Project goals and contents

Goal of this project is to develop and produce new SiO2 materials for replacing graphite, which is currently the preferred material in the anodes of lithium-ion batteries, as they bear the potential of significantly increasing the anode capacity. In addition, a number of technical problems require solving which arise from the material properties of silicon. Here, IDEE’s ambitions are focussed on investigating to what extent a drafted machine learning model can be integrated into the current process of material development and how its intelligent forecasts can specifically optimise the search-space of promising parameters and shorten working steps significantly. The set goals require extensive knowledge in the diverse areas of battery development, processes and key parameters for the development of nano-porous SiO2 materials and innovative methods for the high-throughput characterisation, as well as data-driven modelling through machine learning. Only due to this high degree of interdisciplinarity in the joint project, the combination of materials development, characterisation methods and AI is viable.

Project contact(s)

M.Sc. Philipp Seitz