Fuel cells shall play a key role in the transition and establishment of a emission free energy matrix. The electrolyte is the heart and soul of a fuel cell, being the component that makes possible their functioning but also causing their highest losses.
Electrolytes have been studied and characterised up to date almost exclusively by means of mono-causal classical models, e.g. Ohm’s Law and Fick’s Law. These approaches have proved to be partially, when not entirely, unsatisfactory when compared to experimental data. Therefore it is proposed to study and measure the transport processes taking into consideration a multi-causal approach, known as Non Equilibrium Thermodynamics, as proposed by the Nobel prize award physicist Lars Onsager. The approach has been proved to fit better experimental data of membrane processes, as e.g. in PEM fuel cells.
The two widespread types of fuel cells have been considered, so that two electrolytes were subject to investigation: the solid oxide electrolyte and the polymer electrolyte.
Both electrolytes were investigated in terms of Non-Equilibrium Thermodynamics. Models were developed to experimentally determine the transport coefficients defined by this theory by means of Electrochemical Impedance Spectroscopy, the novelty of this thesis.
Two test benches were modified to apply gradients in temperature or concentration, so
that the coupled effects could be magnified thus making possible to measure the transport coefficients.
The measurements of the solid oxide electrolyte show large deviations compared to literature values, making possible only a qualitative analysis. The ionic transport coefficient was determined and the coupled transport coefficient was approximated. Possible improvements to increase the reliability of the measurements and the test bench were proposed.
The characterisation of the polymer electrolyte proved to be even more challenging because of the coupling between electrolyte and gas diffusion media. Moreover, the test bench introduced a distortion frequency, reducing the meaningfulness of the measurements. The ionic conductivity was calculated directly from impedance spectra while the coupled coefficients were approximated from the slopes of established curves. The results were compared toliterature values, estimated from mono-causal models.
|