The micro-and nanostructure of a Lithium-ion battery electrode greatly influences its properties. Battery optimization must begin by understanding the electrode structure at these scales. An essential piece in this puzzle are insights about the distribution of the carbon-black and binder domain (CBD) in cathodes and of the binder domain in anodes.

The CBD plays an important role for the electrical conductivity of the cathode, since generally the active material is a poor electrical conductor. In anodes, the binder mainly provides mechanical stability for the electrode and, nevertheless its distribution determines the transport possibilities of electrons, as well as the Li+-ions. In both electrodes, a reduction of the amount of binder potentially increases the capacity of the electrode.

Over the last years, the identification of binder in 3D scans, as well as the digital separation of binder from the active material, has improved greatly. Since 2019, the module FiberFind-AI of GeoDict allows the intelligent separation of binder and fibers in fibrous materials. These are used e.g. for gas-diffusion layers in fuel cells or for filter media in filters. Two years later, GrainFind-AI added the possibility to identify binder in structures made of grains of different shapes, although initially this was implemented for binder surrounding mainly spherical particles. An example of these spherical-particle structures is NMC in cathodes of Li+-batteries.

Now, with the latest version of GrainFind-AI in GeoDict 2023, we added a trained Neural Network (NN) capable of reliably identifying binder around flake-like particles. This is often the case in graphite anode structures of Li+-ion batteries. Now the GrainFind-AI module of GeoDict is shipped with two different NNs trained for binder identification – one specifically suited to identify binder in anodes, and the other, explicitly trained to identify the CBD in cathodes. With these two NNs, a variety of scientific issues and questions can be clarified and solved.