The complex operating behaviour of PEM Fuel Cells is heavily influenced by different non-linear loss processes, which limit the cell performance. Depending on the operating conditions, the different losses change in magnitude and characteristic time constants [1]. During technical operation of large-sized PEM Fuel Cells, gradients in operating conditions occur along the gas channels, resulting in an inhomogeneous current density distribution. This is caused by the locally varying electrochemical activity due to in-plane gradients in concentrations of reactants and reaction products, gas pressure and temperature. Consequently, each loss process occurs locally distributed to a varying degree. Therefore, a profound knowledge about the spatial distribution of losses is essential for optimizing the operating strategy as well as the cell and stack components.

The commonly measured I/V characteristics provide only integral information about the overall performance. Thus, the data obtained represent merely the sum of all loss processes averaged over the entire cell area and do not allow detailed conclusions about the respective spatial distribution along the channel.

Inevitably, in order to deconvolute the loss processes in the entire area of large sized cells, an approach is required that provides deeper insight in the processes and their interdependencies. For this purpose, an impedance-based methodology is developed that enables a spatially resolved deconvolution of loss processes. It is based on electrochemical impedance spectroscopy (EIS) and an impedance data analysis by the distribution of relaxation times (DRT) [2,3] that is applied to a segmented cell [4] with almost gradient-free segments. By systematically varying operating conditions the different loss processes are identified, quantified and separated with respect to their characteristic frequencies.

In this contribution the design of the test bench and the segmented cell will be presented. Furthermore, first results regarding the electrochemical characterization and the impedance-based deconvolution of loss processes in a segmented cell will be discussed.

1. M. Heinzmann et al., J. Power Sources 402, pp. 24-33 (2018).
2. H. Schichlein et al., J. Appl. Electrochem. 32, pp. 875-882 (2002).
3. E. Ivers-Tiffée et al., J. Ceram. Soc. Japan 125, pp. 193-201 (2017).
4. T. Reshetenko et al., J. Electrochem. Soc. 163, pp. F1100-F1106 (2016)