Analysis of the binder and carbon black distribution in graphite electrodes for lithium ion batteries using electron dispersive X‐R ay spectroscopy and energy selective backscatter electrons

Abstract

Lithium ion batteries (LIBs) as mobile storage systems become more important in the future. At the same time, the demands on LIBs, such as high capacity in combination with high dis‐/charge rates, low weight, long life time and cycle stability, are rising. This presents great challenges to the internal structure of the LIB, especially to the electrodes. Here, graphite is the most common material, which is used within an anode. The anode itself consists also of carbon black for the electro conductivity and binder for cohesion (particle and carbon black) and adhesion (electrode on current collector). But both are electrochemical inactive. Hence the amount must be as small as possible. However, the distribution of the binder and the carbon black has not been discussed in detail in the literature, in particular for low amounts. Therefore the distribution of two different types of binder poly vinylidene fluoride (PVDF) and a mixture of styrene‐butadiene rubber (SBR) and sodium salt of carboxymethylcellulose (Na‐CMC), within graphite anodes is investigated by scanning electron microscopic (SEM) methods in this work. The investigated electrodes are about 350 µm in thickness. To facilitate analysis in the SEM smooth cross‐sections were prepared by argon ion polishing. In case of the PVDF binder the distribution can be visualized based on its fluorine content. Therefore electron dispersive X‐Ray spectroscopy (EDXS) mapping was used and drying related gradients in the binder distribution could be identified (Figure 1b). High binder concentrations were usually found in areas of high carbon black concentration. However, the rather poor lateral resolution of EDXS (about 1 µm) impedes more detailed investigations. On the other hand the detection of energy selective backscattered electrons (ESB) can be used to obtain element specific information at lateral resolutions comparable to conventional secondary electron images. By optimizing the grid voltage of the ESB – detector and the electron energy, it was possible to obtain high resolution images, in which fluorine rich regions appears bright (Figure 1a). It can be shown that not only the binder but also the carbon black differs in contrast compared to the graphite particles (Figure 2b). The obtained ESB – images were evaluated with image manipulation software to mark the particles and the carbon black (Figure 2d). Due to an about 100 times faster acquisition compared to EDXS, images spanning a large part of the cross‐sections could also be obtained in a rather short time, but the evaluation is not that simple and cannot be automated like EDXS ‐ mappings. Since the SBR – binder as such does not contain a suitable element for mapping, osmium tetroxide (OsO4) staining was used to selectively mark the double bonds of the SBR – binder. Thereafter, localization of the binder was possible by mapping the Os distribution with EDXS. Unfortunately the osmium concentration was too low to visualize the SBR – binder by ESB.