Our team is currently engaged in a fascinating collaboration with Fraunhofer IKTS on a project centered around battery research. Chimaera’s contribution is the application of AI to automate the complex evaluation of cathode images.

The battery’s analysis involves the utilization of focus ion beam (FIB) tomography, resulting in the acquisition of a large number of images. These scanning electron microscope (SEM) images show a cross-sectional view of the battery's cathode, where different classes of cathode active material particles, a mixed binder and carbon phase, and pores can be recognized as distinct phases. By annotating the different classes on the individual images, they can then be combined into a comprehensive 3D model facilitating simulations of battery properties. These simulations contain critical factors like energy density and potential charge rate, which strongly depend on the quantity, size distribution, and morphology of individual phases within the cathode structure.

Chimaera therefore trains an AI model on manually annotated datasets to perform this time-consuming step for the vast amount of images required to generate a reliable 3D model.