Thermal Stability and Electrochemical Performances in SOEC Mode of Mixed La/Pr n=2 Ruddlesden-Popper Nickelates as Oxygen Electrodes

Due to their high ionic and electronic conductivities as well as electrocatalytic activity, Ln n+1 Ni n O 3n+1± δ phases with Ruddlesden-Popper (RP) structure type are considered as oxygen electrode materials of substantial interest for SOFC & SOEC applications. Such RP layered compounds present n- Ln NiO 3 perovskite layers inter-grown with one Ln O rock-salt layer. The electronic conductivity is ensured by the Ni 2+/3+ mixed valence the perovskite NiO 6 octahedra. For the n=1 member a particularly high anionic diffusion is observed which mainly takes place through oxygen interstitial sites in the rock-salt slab. Thus, the number of perovskite layers increasing with the n-value, the n = 2 members present a balanced electronic/ ionic conductivity among the RP members, which could be beneficial to the electrochemical performances. 1–4 In the past, n=1 RP materials have shown promising properties as oxygen electrodes, especially with Ln = Pr 5 . First, La 2 NiO 4+δ has shown a good stability over time but a too high polarization resistance (R p ) at working temperatures. On the contrary, Pr 2 NiO 4+δ shows a low R p , but a crippling phase stability. The substitution of Pr with La in Pr 2 NiO 4+δ was brought as alternative and La 1.5 Pr 0.5 NiO 4+δ was found as the best compromise between thermal and electrochemical performances stabilities 6 . As second alternative to the n=1 compound, this study explores substituting La with Pr in La 3 Ni 2 O 7±δ , the n=2 counterpart, which, to date, has only been reported with La in the literature. La 3-x Pr x Ni 2 O 7± δ compounds with x = 0, 0.375, 0.75, 1.125, 1.5 were successfully synthetized, and their purity was confirmed through Rietveld refinement ( Amam space group). For the higher Pr contents, the chemical synthesis need to be performed under pure oxygen, while compounds with substitution levels above 50 % of Pr could not be stabilized. The phases remain thermally stable at 600, 700 and, 800 °C for 3 months. No cell parameter evolution was noticed by Rietveld refinement. When the powders were intimately mixed with GDC (used as a barrier layer in the SOCs devices) and aged for 3 months at the SOEC operating temperatures, a Pr and La migration into the ceria structure was evidenced with the help of XRD and SEM/EDS analysis. By XRD, the formation of a La 1-x Pr x NiO 3-δ perovskite phase was seen, while SEM/EDS experiments showed a higher extent of Pr diffusion compared to La. This led to the formation of a Pr-doped ceria (PrGDC). Local grain subregions containing up to 10 % of Pr into GDC where found while La diffusion was limited to 1 – 2 %. Pr doped ceria being a MIEC compound with electrocatalytic activity, this reactivity is not expected to negatively impact the electrochemical performances of the device. On the contrary La-doped ceria was seen by Vibhu et al. ( La 2 NiO 4+δ deposited on GDC) to dramatically decrease the electrochemical performances 6 acting as an oxide ion blocking layer. As-prepared powders were attrited to prepare terpineol based inks. Symmetrical cells were prepared by the screen-printing technique. Co-sintered electrodes with GDC were deposited on 8YSZ supports and sintered at 950 °C for 2 h as previously optimized in our lab 7 . Indeed, as showed in the past co-sintering the electrode improves the electrode interface with GDC and, favours smaller grain with a reduced sintering temperature. Hence, a lowered polarization resistance (R p ) is expected. The thickness was optimized: for 12, 24 and, 40 µm, polarization resistances of about 100, 85 and 60 mΩ·cm² at 700 °C for x = 1.5 were obtained, respectively. Thick electrodes were then selected for this study. A second sintering route was optimized, the conventional sintering in 2-steps of the electrode and GDC layers. It implies a higher temperature hence bigger grains and higher R p but the GDC layer reached is denser hence a lowered series resistance is expected. Here, several sintering temperatures (1000 to 1200 °C) and times (1h to 4h) were investigated. 1050 °C for 4 h stood out as the best sintering conditions. Indeed, for high temperature (1100 – 1200 °C) too large grains were obtained (0.8 – 1.2 µm), while for lower temperature (< 1100 °C), grains were much smaller (0.4 – 0.7 µm). The smaller grains were observed for 1000 °C / 4 h and 1050 °C / 1 h though a poor interface resulted from this sintering. The best sintering conditions were found at 1050 °C for 4 h. The co-sintered electrodes showed the best performances with polarization resistances about 30, 55 and 110 mΩ.cm² for x = 1.5, 0.75 and, 0 respectively. The 2-steps sintered electrodes presented higher R p , about 60, 90 and, 160 mΩ·cm² for x = 1.5, 0.75 and, 0 respectively. As explained before, this difference in polarization resistance can be explain by the difference of grain size between the two methods. Thus, with the co-sintering process, grains around 0.4 µm are seen by SEM while when using the 2-steps sintering methods, grains are around 0.6 µm. This is in good agreement with the Adler-Lane-Steele model 8 . Ageing of symmetrical cells was performed at 800 °C under i dc =0 A. Co-sintered and 2-steps sintered electrodes were aged. In both cases, no crippling loss of performances was obtained. For instance, La 1.5 Pr 1.5 Ni 2 O 7±δ 2-steps sintered electrode showed a polarization resistance increase from 57 to 71 mΩ·cm² after 1500 h and, La 3 Ni 2 O 7±δ (2-steps) from 160 to 200 mΩ·cm². For both sintering methods, most of the R p degradation is observed within the 500 first hours. Then, the polarization resistance progressively stabilizes. The co-sintered electrodes showed a percentage of R p degradation between 30 and 100 % while it was only 20 and 40 % for the 2-steps sintered electrode. This degradation cannot be explained by the phase degradation as in both cases, the RP n=2 phase is preserved, and only a small amount of La 1- x