Dealing with degradation in solid oxide electrochemical cells: novel materials and spectroscopic probes

Robles Fernández, Adrián
Merino Rubio, Rosa Isabel (dir.) ; Orera Utrilla, Alodia (dir.)

Universidad de Zaragoza, 2022

Abstract: In this PhD thesis, we have focused on two of the main issues regarding solid oxide fuel cells and electrolysers. On the one hand, the high temperatures at which they work (800-1000ºC) is detrimental for their long-term performance, and novel combinations of electrolyte and oxygen electrode materials have been tested in order to establish their suitability to work in intermediate temperature (600-800ºC) solid oxide fuel cells. On the other hand, degradation issues affect these devices greatly when working in the electrolyser mode, often assigned to the development of high oxygen partial pressures within the electrolyte. Regarding this topic, we have developed an analytical procedure to monitor the oxygen activity inside a YSZ electrolyte using redox dopants as spectroscopic probes and used it in cells tested in different conditions in the electrolysis mode.
First, an aluminium-doped lanthanum silicate compound (LSAO) with the apatite structure was chosen as electrolyte, and eight different strontium and cobalt-free compounds with a perovskite structure with the general formula LaMxN1-xO3 (where M: Fe, Mn, Cr; and N: Ni, Cu) were selected to be tested with the apatite electrolyte. The solid-state synthesis of the apatite and perovskite-type compounds was optimised, achieving perovskite compounds free from secondary phases or decomposition, both in powder form and also after sintering them in pellet form, as it was proven by X-ray diffraction and Raman spectroscopy. The chemical compatibility of the electrolyte and electrode materials was tested by mixing and heating the powders at temperatures above the operational and sintering ones, and no reaction between the compounds took place, as proven again by X-ray diffraction. The thermomechanical compatibility between sintered materials was tested by dilatometry, and no big differences could be found regarding the thermal expansion behaviour of the apatite and perovskites in a wide range of temperatures.
Then, the electrochemical performance of the compounds was tested by electrochemical impedance spectroscopy. The LSAO ionic conductivity at intermediate temperatures (9.4·10-3 S·cm-1 at 800ºC) was close to the one shown by a conventional YSZ electrolyte, and the electrical conductivity of the oxygen electrode materials ranged from 10 to 100 S·cm-1 at 800ºC, with activation energies in the high temperature range between 0.1 and 0.3 eV. Among the perovskite materials, the ones containing manganese and copper showed the highest electrical conductivities and the lowest activation energies. The iron-containing compounds (LFN and LFC) exhibited a different activation behaviour with temperature than the rest of the compounds.
The following step consisted on manufacturing symmetrical cells with the apatite electrolyte and perovskite electrodes. For that purpose, slurries of the electrode materials were prepared and the electrolyte pellet was coated with them by dip-coating and sintered. The microstructure of the cells was checked in terms of electrode thickness, porosity, particle size, and adherence of the electrodes. Among the compositions tested, LFC showed the lowest ASR with just 4.3 Ω·cm2 at 700ºC, value comparable to the state-of-the-art oxygen electrodes. When applying a small DC bias, the activation energies of the electrodes decreased, as well as their polarization resistances. Promising results were found in this thesis about novel electrolyte/electrode combinations for IT-SOFC, with room for improvement regarding electrode microstructure and the fabrication of composite electrodes with the LFC material.
In order to examine the degradation issues concerning SOEC devices, the research began with finding a suitable spectroscopic probe that allowed us to track the oxygen activity in the cells. First, the optical signals in a YSZ electrolyte doped with redox ions were investigated. The objective was to select the ones suitable to track the oxygen activity, allow for a detection in the backscattering configuration and be operative at high temperatures. The samples tested in this part of the thesis were either commercial or solidified on purpose YSZ single-crystals doped with Ce, Mn, Mn-Nd, V or Tb; or polycrystalline ceramics of Tb or Pr-doped YSZ.
Among the commercial samples, it was found that YSZ-Ce showed strong change in its optical signal upon redox treatment, and the Ce3+ backscattering signal could be used to monitor the oxygen activity, although the signal disappeared at temperatures above 300ºC. In the case of YSZ-Mn and YSZ-V, even though a change in the optical signal upon oxidation/reduction could be found, there were no backscattering signals that could be used for tracking the oxygen activity. The luminescence of minority rare earth dopants in these samples (Pr3+, Er3+ or Nd3+) was measured and a change in the backscattering signal could be observed upon redox treatment. Nevertheless, the quantification of these signals would have been complicated due to a possible interaction with the major dopants, and these commercial samples were not used to track the oxygen activity within the electrolyte.
In the case of praseodymium-doped YSZ, a change in optical signal was observed by diffuse reflectance, and bands due to Pr3+ and Pr4+ could be found. In the case of backscattering signal, Pr3+ ions exhibited an intense luminescence band which decreases upon oxidation to Pr4+, and this signal held up to 700ºC. This could be useful to make in-situ or in-operando measurements of the oxygen activity inside the electrolyte. Nevertheless, ion-ion interaction and concentration quenching of the luminescence band prevented an easy quantification of the oxygen activity, and this probe was also discarded.
Terbium-doped YSZ was found suitable in order to track the oxygen activity within YSZ. Changes in optical and luminescence signals could be observed and attributed to different oxidation states of terbium (Tb3+ and Tb4+) upon redox treatments. Tb3+ was not affected by concentration quenching and a quantitative analysis could be carried out. It was found that Tb4+ absorbance was proportional to PO2^(1⁄4), as expected for the electron trapping model. A relation between the Tb3+ luminescence intensity and the oxygen partial pressure could be found, and it proved to be useful in the high PO2 range (10-4-100 bar). Terbium was therefore the selected probe in order to carry out the electrochemical experiments to detect degradation mechanisms in electrolysers.
3%Tb-doped 8YSZ shows appropriate oxide ion conductivity to be used as the electrolyte in solid oxide cells. Then, an electrolyte-supported solid oxide cell was prepared using a LSM/YSZ composite for the oxygen electrode and a NiO/YSZ composite for the fuel electrode, and its electrochemical properties were tested in a bicameral cell at 800ºC. Using EIS and changing the atmosphere in the oxygen side, the electrochemical properties of the system were described, and the polarization resistance of each electrode was assigned. After those measurements, several experiments were carried out polarising different cells using a range of biases in the electrolyser mode. When a steady-state was reached around 48 hours after applying a constant voltage, cells were quenched to freeze the high temperature polarization state.
The post-mortem cells were analysed in terms of the Tb3+ luminescence across the electrolyte thickness. The luminescence values were transformed into oxygen partial pressures using the relation mentioned above and profiles of the oxygen activity within the electrolyte could be obtained. These measurements were noisy and a couple of corrections were made in order to obtain a suitable signal. An edge-correction due to the loss of signal near the electrodes and a saturation correction due to microstructural aspects of the cell were applied. The method presented here has potential to visualize PO2 profiles in SOEC. Further experiments should be done in order to achieve higher accuracies.
Finally, numerical solutions to the transport equations for describing the oxygen activity within the electrolyte were found and compared with the experimental results. It was found that the simulations assuming polarization resistances as derived from EIS spectra at the beginning of the CA experiments did not agree with the oxygen activity profiles obtained from the luminescence experiments. By analysing the SEM micrographs of the post-mortem cells, we could find that the most degradation had occurred near the fuel electrode. The nickel particles tended to agglomerate, especially for high polarization biases, and the porosity of the electrode decreased with applied bias. Besides, cracks within the electrolyte were found near this electrode and even a complete delamination of the fuel electrode was observed for the highest polarization experiment. These observations allowed us to assign a higher polarization resistance to the fuel electrode and then the numerical model results were closer to the results of the luminescence measurements. In order to get better insights of the degradation conditions of the cell while working on electrolyser mode, more experiments should be done.

Abstract (other lang.): 

Pal. clave: electroquimica ; materiales ceramicos ; preparacion y caracterizacion de materiales inorganicos ; propiedades opticas de materiales

Titulación: Programa de Doctorado en Física
Plan(es): Plan 488
Nota: Presentado: 03 06 2022
Nota: Tesis-Univ. Zaragoza, , 2022

Creative Commons License

 Record created 2022-09-14, last modified 2022-09-14

Download fulltext

Rate this document:

Rate this document:
(Not yet reviewed)