<
p>
Chromium (Cr) poisoning remains a significant issue in long-term solid oxide fuel cell (SOFC) operation. While the addition of Cr in the interconnect and balance-of-plant (BOP) materials is effective in improving their resistance to oxidation, it also causes the deposition of resistive phases in the air electrode and thus cell performance loss. Here we investigated a new, in-situ Cr poisoning mitigation strategy.<
/p>
<
p>
The most prevalent Cr vapor species concentration is CrO3 in the absence of water vapor, and much higher concentrations of CrO2(OH)2, and CrO2(OH) vapor species in the presence of water vapor. The equilibrium cell-potentials for the reduction of these oxide vapor species to the lower valent Cr2O3(s) are all near 1.12 V at 800 �C, which is close to the open-circuit potential of the cell, indicating that the dissociation of Cr vapor species into lower valent oxides with the onset of current flow. This was demonstrated in our earlier work, where we studied the mechanism of Cr-poisoning as a function of current density, humidity, microstructure, and distribution of electronic/ionic/mixed conducting oxide phases in the air electrode.<
/p>
<
p>
Our previous results have shown that when fuel electrode-supported SOFCs with La1-xSrxMnO3 (LSM)-based air electrodes and yttria-stabilized zirconia (YSZ) electrolyte were subjected to galvanostatic testing for 150 hours at 800 �C in contact with Crofer22APU/H mesh, under different air electrode atmospheres and current conditions, the performance degradation was most severe under galvanostatic conditions. Introduction of water vapor (10%) in the air further increased the rate of performance degradation. In comparison, under open circuit conditions (no current) for 150 hours at 800 �C, the cell performance did not degrade in dry air nor when 10% water vapor was introduced in the air supply. The majority of Cr deposition was observed near the air electrode/electrolyte interface under galvanostatic conditions with 10% humidified air followed by dry air. Under open circuit conditions in dry air and 10% humidified air, Cr-containing oxide deposits were not detected in the air electrode. Thus, current appears to play a primary role and humidity with current an ancillary accelerating role in the Cr-poisoning phenomena suggesting an electrochemically dominant degradation mechanism. Similar cell tests were also performed with fuel electrode supported SOFCs with YSZ electrolyte, Gd2O3 doped ceria (GDC) barrier layer and mixed ionic electronic La1-xSrxFeO3 (LSF)-based air electrodes (LSF-GDC active layer and LSF current collector) [2]. However, they showed negligible performance degradation during the 120 hours of galvanostatic testing. Microstructural investigation of the post-test LSF air electrodes indicated higher amounts of Cr- containing deposits on the electrode surface, much lesser amounts near the air electrode/electrolyte interface (compared to the LSM-based cells) and negligible amount in the bulk electrode. This indicates that the mixed conducting phases and their distribution determines the location and progression of the chromium deposit. Based on all our observations, two self-cleaning and performance recovery processes for the air electrode were proposed for this project: chemical cleaning and electrochemical cleaning. Since electrochemical cleaning was observed to be faster and safe for continued cell operation, this project focused on electrochemical cleaning.<
/p>
<
p>
Achievements of this project include:<
/p>
<
ol>
<
li>
Demonstration of electrochemical cleaning using LSM-based cells employing a mild electrolytic bias condition (-15 mA/cm2) with air containing 2-5% water vapor (H2O(g)). Changes in performance were compared using maximum power density and corroborated by post-test Cr quantification using EDS. Scanning electron microscopy (SEM) imaging determined that Cr2O3 type deposits are removed as result of electrochemical cleaning, whereas Cr, Mn spinel deposits are not. However, a lower temperature hold in oxygen can dissociate the Cr-Mn spinel into chromium oxide and manganese oxide. The chromium oxide can be electrochemically cleaned, <
/li>
<
li>
Comparison of Cr content post-test utilizing SEM and EDS analysis, demonstrating Cr removal because of electrochemical cleaning. The chromium quantification procedure was refined to better capture Cr content present in the entirety of the cell. <
/li>
<
li>
Demonstration of repeated electrochemical cleaning using LSM-based cells without effecting the stability of any of the fuel cell components. In conjunction with other poisoning mitigation strategies, we conclude that electrochemical cleaning can greatly improve the cell lifetime for LSM-based cells. <
/li>
<
li>
Development of framework to understand Cr poisoning and cleaning for other MIEC-type electrodes.<
/li>
<
/ol>