The performance of present generation Li-ion batteries is not sufficient to completely electrify the car fleet. This disruptive step requires new ways to identify the best materials combinations in the battery cell. Simulations can contribute estimates of the density and ion conductivity for arbitrary electrolyte formulations [1] as well as the open cell voltage (OCV) data for new cathode materials [2]. However, it is tedious to test every combination. A Materials Acceleration Platform (MAP) uses active learning to sample the materials space and request data from both experimental and simulation sources [3]. This approach makes the sampling more efficient and significantly speeds up the development of next generation battery cells.
Dassault Systèmes (3DS) has been participating in creating the Materials Acceleration Platform for the Battery Interface Genome (BIG-MAP)*. This is an internationally distributed MAP, which has been set up to optimize one [4] or two tasks simultaneously [5]: the formulation of a three component electrolyte with two organic solvents and LiPF6 with respect to density and ion conductivity and the End-of-Life (EOL) predictions for coin cells using the same electrolyte system.
The Technical University of Denmark (DTU) was running two independent adaptive Gaussian optimizers to optimize the electrolyte formulation and EOL prediction. The optimizer was retraining the model for each generation and requesting new data from either experimental or simulation sources. Karlsruhe Institute of Technology (KIT) delivered experimental data for ion conductivity and cycling data. 3DS provided data for density and ion conductivity calculated by molecular dynamics (MD). The communication between the tenants was controlled by the FINALES broker server.
The results of the optimization run for the electrolyte, showed that the ion conductivity is strongly dependent on the LiPF6 concentration, but only a limited dependency on the mixture of the two solvents.
*This research is part of the BIG-MAP project, which has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 957189.
References
[1] Hanke et al., J. Electrochem. Soc. 167, 013522 (2020)
[2] J. Carlsson et al., US Patent US 2021/0286015 A1.
[3] S. Stier et al., Adv. Mat. 2024, 2407791.
[4] M. Vogler et al., Matter 6, 2647 (2023).
[5] M. Vogler et al., Adv. Ener. Mat. 2024, 2403263.