The industrial scale production of high-performance and long term stable lithium ion batteries (LIBs) relies on the usage of precisely tailored materials of high purity. Even small deviations in the material properties or contaminations may lead to detrimental effects regarding the battery’s stability, safety and rate capability. For example, the commonly used conducting salt LiPF₆ hydrolyses in the presence of trace amounts of water in the organic solvent, generating HF and further decomposition products [1, 2]. These products can degrade battery components and, consequently, lead to a bad performance of the battery [3].

The components which can contribute to the overall water contamination level inside a LIB are the electrodes, the separator, the housing and the electrolyte itself. Therefore all these components of a LIB have to be dried carefully before assembly. These time and energy intensive processes, contribute to the overall costs of LIBs. In order to ensure a sufficiently low moisture level inside the battery components, while, at the same time, keeping the drying costs low, a detailed understanding of the water sorption and drying behaviour of those battery components is required. While the trace water concentrations in commercially available LIB electrolytes (usually 15 ppm or lower) is commonly precisely monitored, a systematic and comparative study of the hygroscopic and drying behaviour of different electrodes and separators has not been performed yet.

This work focusses on the investigation of hygroscopic and drying behaviour of various common anodes, cathodes and separators. Therefore, samples were stored in a defined atmosphere (25 °C, 40 %rh) and their moisture sorption as well as the subsequent drying behaviour (at 120 °C) were studied. The moisture content and drying behaviour of battery components was analysed by a coulometric Karl Fischer titrator. During the sample drying, the evaporated water was transferred into the titrator by a dry argon stream and the water detection was recorded.

The results show that the investigated cathodes and anodes differ strongly in their drying behaviour (Fig. 1). Cathodes with LiMn₂O₄ as active material absorb a high amount of water from the atmosphere and show and a delayed water release while drying. The LiFePO₄ cathode also shows strong water uptake, but to a lesser extent than the LiMn₂O₄ cathode. In contrast, the LiNi_1/3Co_1/3Mn_1/3 cathode shows an amount of water sorption, which is even lower than the water sorption of the commonly used graphite anode. In accordance with the hydrophobic surface behaviour of the polypropylene separator (Celgard PP2500), its water uptake is very low. The glass fibre separator (Whatman GF/C) on the other hand has a large and hydrophilic surface, which results in a very high water content.

The water emitted from the samples during the drying process is continuously quantified in the Karl Fischer Titrator. As a consequence time dependent moisture emission rates for the different electrodes and separators are obtained. Fitting and extrapolating the moisture emission rates allows to draw conclusions beyond the runtime of the drying experiments. For example the drying times until reaching threshold levels for the remaining water concentration in the sample can be calculated (Fig. 2).

[1]      D. Aurbach,  a. Zaban, Y. Ein-Eli, I. Weissman, O. Chusid, B. Markovsky, M. Levi, E. Levi,  a. Schechter, and E. Granot, J. Power Sources, vol. 68, no. 1, pp. 91–98, Sep. 1997.

[2]      A. V. Plakhotnyk, L. Ernst, and R. Schmutzler, J. Fluor. Chem., vol. 126, no. 1, pp. 27–31, Jan. 2005.

[3]      U. Heider, R. Oesten, and M. Jungnitz, J. Power Sources, vol. 81–82, pp. 119–122, Sep. 1999.