About this Abstract |
| Meeting |
MS&T26: Materials Science & Technology
|
| Symposium
|
Energy Materials for Sustainable Development
|
| Presentation Title |
Accelerated AI-Assisted Material Development Workflow for Solid Oxide Electrolyzer Cells |
| Author(s) |
Johanna Kramer-Schöggl, Fereshteh Falah Chamasemani, Andreas Egger, Edith Bucher, Roland Brunner |
| On-Site Speaker (Planned) |
Roland Brunner |
| Abstract Scope |
The performance of air electrodes in solid oxide electrolyzer cells strongly depends on compositional and morphological features. This dependence is rather complex and requires a fundamental understanding of the underlying structure-property relationships. We present an AI-assisted framework to optimize cells with La₀.₆Sr₀.₄Co₀.₂Fe₀.₈O₃₋δ (LSCF) – Ce₀.₉Gd₀.₁O₁.₉₅ (GDC) air electrodes that goes beyond our recent work [1]. In particular, we test possibilities towards the development of a Bayesian framework to optimize the LSCF-GDC ratio. 3D microstructural as well as electrochemical and processing data for different LSCF-GDC ratios are collected. Microstructural data is obtained on the nm-scale by focused-ion-beam scanning electron microscope tomography. The individual pore and material phases are derived from the reconstructed 3D images by semantic segmentation. A variety of microstructural features like tortuosity, triple phase boundary density with percolation-based activity classification etc. are extracted from the 3D images. The presented approach could replace empirical screening with physics-informed optimization.
[1] https://doi.org/10.1016/j.jpowsour.2025.238174 |