Microstructural Changes in Ni-YSZ Electrodes Operated in Fuel Cell and Electrolysis Modes: Effect of Gas Diffusion Limitations

Abstract

Lifetime performance stability is a key issue for the commercialization of solid oxide cells. Ni migration in Ni-YSZ electrode supported cells is an important degradation mechanism. Here we present the results of life tests on symmetric Ni-YSZ electrode-supported cells. These symmetric cells have similar processing and microstructure as conventional Ni-YSZ-supported cells, but the symmetric structure provides information on the Ni-YSZ electrode operating as both an anode and a cathode in the same test. Life tests of up to 1000 h were carried out at 800 ˚C in 50-50 H2-H2O at current densities of 0, 0.75, 1.00, and 1.50 A/cm2. Total applied voltage to each cell was tracked over the lifetime, and small portions of the cells were removed during the life tests to study the time-dependent changes in microstructure.Figure 1 summarizes the microstructural results from a Ni-YSZ cell operated at 1.0 A/cm2, observed at 100, 500, and 1000 h. 2D and 3D microstructure characterization was used to show the volume fractions of pore and active (electrically connected) Ni versus position. Somewhat surprisingly, the cathode microstructure remains relatively unchanged compared to a non-polarized cell, with no evidence of Ni migration or isolation. In contrast, Ni migration and isolation was observed to increase with time at the anode in polarized cells, with the thickness of the Ni deactivated region growing from ~5 mm at 500 h to ~15 mm at 1000 h for 1.0 A/cm2 and ~2 μm at 500 h and ~8 μm at 1000 h for 0.75 A/cm2. In addition, total cleavage at the anode-electrolyte interface occurs at 1.5 A/cm2 by 100 h, and at 1.0 A/cm2 by 1000 hr. The increase of porosity in the altered zones clearly shows where Ni is depleted. There is no evidence of Ni enrichment adjacent to the depleted region, as might be expected if Ni was moving via surface diffusion.Although there have been numerous recent reports of Ni migration/isolation in Ni-YSZ cathodes during electrolysis cell operation1, Ni migration has also been reported in Ni-YSZ fuel cell anodes2,3. Here we suggest that such results can be explained by a relatively high steam content in the Ni-YSZ anode functional layer. Conversely, the lack of Ni migration in the electrolysis cathode can be explained by a relatively low steam content in the cathode functional layer. To quantitatively assess the gas compositions, one-dimensional modeling of the electrochemical and gas diffusion processes of these cells was performed using a finite difference method (FDM) with modified Butler-Volmer kinetics and the dusty-gas model respectively. Using diffusivity values calculated via electrochemical impedance spectroscopy and microstructural measurements, the modeling reveals that electrochemically active Ni sees significantly different gas composition than the inlet, creating steam-rich anodes (PH2O = 0.73, 0.8, and 0.95 atm for 0.75, 1.0 and 1.5 A/cm2 respectively) and steam-depleted cathodes (PH2O = 0.27, 0.2, and 0.05 atm for 0.75, 1.0 and 1.5 A/cm2 respectively). This is in accord with results and models suggesting that Ni migration is important mainly under high steam conditions, probably due to vapor transport. Note that these effects are exacerbated by the relatively low porosity and small pore size in the present electrodes that lead to asymmetric Knudsen diffusion, wherein H2O diffuses at ~1/3 the rate of H2. These results suggest that the nature of the porosity in Ni-YSZ supports can lead to significant variations in the extent and directionality of Ni migration.Figure 1: (a, b, c) Polished cross sectional backscatter electron (BSE) images and (d,e,f) low voltage secondary electron (LV-SE) images of an electrode-supported symmetric Ni-YSZ cell at 100, 500, and 1000 h of galvanostatic operation at 1.00 A/cm2 in 50-50 H2-H2O and T = 800 ˚C. BSE images reveal a depletion of total Ni at the anode-electrolyte interface (AEI) by a net increase in porosity, and eventual cleavage at that line. LV-SE images reveal that near total deactivation of Ni occurs near the AEI as well. (g,h,i) Quantitative image analysis reveals the extent of porosity increase (via Ni depletion) and Ni deactivation at 100, 500, and 1000 hours. The regimes for both events exactly overlap, with the depletion/deactivation extending 0, 5, and 15 μm from the AEI, respectively. M. B. Mogensen et al., Fuel Cells, 21, 415–429 (2021).J. Geng et al., J. Power Sources, 495, 229792 (2021).Z. Jiao and N. Shikazono, J. Power Sources, 396, 119–123 (2018). Figure 1