Fraunhofer Institute for Ceramic Technologies and Systems IKTS, Forchheim, Germany Freie Universität Berlin, Department of Physics, Berlin, Germany Institute of Nanotechnology and Correlative Microscopy INAM, Forchheim, Germany
03.11.2026. u 11:00h
Institute of Physics, 1st wing lecture room & Zoom
The development of next-generation batteries increasingly depends on our ability to understand and control complex relationships between materials, processing, microstructure, interfaces, defects and electrochemical performance across multiple length and time scales. Conventional characterization approaches, however, often generate isolated datasets and provide only snapshots of this multidimensional materials space. The key challenge is therefore shifting from individual measurements towards correlated, data-centric and ultimately predictive materials intelligence.
In this presentation, we introduce a scale-bridging multimodal analytics framework for battery materials and cells, combining X-ray microscopy and tomography (µCT/nanoCT), analytical electron and ion microscopy (SEM, FIB-SEM with EDX, EBSD and ToF-SIMS), optical spectroscopies (stimulated and spontaneous Raman, IR, µXRF), scanning probe methods including conductive AFM, and electrochemical characterization. Precise sample relocation and registration using nanoGPS technology [1] enable information from complementary techniques to be correlated from the cell and electrode scale down to individual particles, interfaces and nanoscale defects. Such correlative approaches are particularly important for identifying local heterogeneities and degradation pathways that ultimately determine macroscopic battery performance.
Particular emphasis will be placed on defect formation and degradation in battery electrodes, including heterogeneous particle and pore networks, interfaces, local compositional changes, mechanical damage and transport limitations. Pristine, cycled and post-mortem states are correlated, while inert-transfer, glovebox-integrated and operando workflows preserve chemically sensitive states and provide access to dynamic processes under realistic operating conditions. The relevance of contamination-free and operando workflows for linking microstructural evolution to device performance has been established in aforementioned multimodal analytics framework.
The resulting heterogeneous 2D, 3D and, increasingly, 4D datasets are integrated within the CORRELYZE data platform and workflow engine [2]. Rather than treating analytical images and spectra as individual results, they are transformed into quantitative material descriptors and linked with metadata such as processing conditions and electrochemical performance. Machine-learning-assisted segmentation, defect classification, anomaly and rare event detection and pattern recognition support the extraction of statistically meaningful structure–property relationships. CORRELYZE thereby provides the digital backbone for converting multimodal characterization into reusable material data and predictive descriptors.
A central perspective is the transition from multimodal characterization to predictive materials intelligence. Experimentally derived descriptors such as porosity, tortuosity, particle connectivity, interface characteristics, compositional gradients and defect populations can be coupled to physics-based multiscale simulations and data-driven models. This closes the loop between processing → structure → properties → performance → degradation, allowing experimental observations not only to explain failure retrospectively, but increasingly to predict material and cell behavior and guide subsequent processing and material-design decisions.
Using novel battery electrode materials from Li and Na type and cells as representative examples, the presentation will demonstrate how this convergence of advanced analytics, structured material data, machine learning and physics-based modelling can transform characterization from a diagnostic tool into an active component of materials development. Such workflows provide a pathway towards digital twins and closed-loop optimization, reducing experimental iteration while improving reliability, lifetime, sustainability and resource efficiency of future battery technologies.
[1] O. Acher, et al., Meas. Sci. Technol., 2021, 32, 045402
Keywords: Battery materials; Multimodal characterization; Correlative microscopy; Materials data; Degradation; Machine learning; Predictive materials intelligence; Digital twin
Join Zoom Meeting:
https://us06web.zoom.us/j/5081440931
Meeting ID: 508 144 0931
Seminar hosts: Juraj Krsnik i Borna Radatović

