Self-Transforming Configuration Based on Atmospheric-Adaptive Materials for Solid Oxide Cells

Solid oxide cells (SOC) with a symmetrical configuration have been focused due to the practical benefits of such configurations, such as minimized compatibility issues, a simple fabrication process and reduced cost compared to SOCs with the asymmetrical configuration. However, the performance of SOCs using a single type of electrode material (symmetrical configuration) is lower than the performance of those using the dissimilar electrode materials (asymmetrical configuration). Therefore, to achieve a high-performance cell, we design a ‘self-transforming cell’ with the asymmetric configuration using only materials of the single type, one based on atmospheric adaptive materials. Atmospheric-adaptive perovskite Pr0.5Ba0.5Mn0.85Co0.15O3-δ (PBMCo) was used for the so-called self-transforming cell electrodes, which changed to layered perovskite and metal in the fuel atmosphere and retained its original structure in the air atmosphere. In fuel cell mods, the self-transforming cell shows excellent electrochemical performance of 1.10 W cm−2 at 800 °C and good stability for 100 h without any catalyst. In electrolysis mode, the moderate current densities of −0.42 A cm−2 for 3 vol.% H2O and −0.62 A cm−2 for 10 vol.% H2O, respectively, were observed at a cell voltage of 1.3 V at 800 °C. In the reversible cycling test, the transforming cell maintains the constant voltages for 30 h at +/− 0.2 A cm−2 under 10 vol. % H2O.

Like this, the SOC has a configuration dilemma; specifically, there is a trade-off relationship between the optimizing catalytic activity and the practical applicability. Thus, we introduce the 'transforming cell' based on the atmospheric-adaptive materials to take full advantage of both asymmetrical and symmetrical configuration. The atmospheric-adaptive materials refer to the material which changes the structural properties (e.g. phase and composition) by themselves to have the high catalytic activity depending on the atmosphere. Thus, the transforming cell can be fabricated with symmetrical configuration and can be operated with the asymmetric configuration, as illustrated in Fig. 1, providing excellent electrochemical performance and few thermal problems with a simple fabrication process.
As a candidate for the electrode material, Pr 0.5 Ba 0.5 MnO 3-δ (PBMO) was considered due to its excellent chemical and structural stability and the mixed valence Mn 4+ /Mn 3+ /Mn 2+ which can accept electrons to promote the reaction 10,16 . Also, the catalytic activities can be tailored by A-and B-site doping because the perovskite oxide can choose its elements flexibly. Recently, Kwon et al. reported the exsolution phenomenon in PBMO systems, with the substitution of transition metal for Mn. Based on previous studies, among the transition metals, cobalt is the most versatile as a dopant to both air electrodes and fuel electrodes with doping and exsolution. Generally, cobalt doping on air electrode material improves the electrical conductivity by increasing the oxygen vacancy concentration 17,18 . For the fuel electrode, the exsolved cobalt plays the role of a catalyst, enhancing the electrochemical performance [19][20][21] . In summary, Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3-δ (PBMCo) has simple perovskite structure with mixed cubic and hexagonal layers 19,[22][23][24][25][26] . Exposed to the hydrogen, the structure is changed to the cation-ordered perovskite with Co exsolution.
Thus, Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3-δ (PBMCo) was chosen as the atmospheric-adaptive material to obtain the cobalt doping effect on air electrodes and the cobalt exsolution effect on fuel electrodes. To increase the surface area and corresponding electrochemical performance of the electrodes, the infiltration technique was applied as a fabrication method. As expected, with the derived electrode materials, the transforming cell shows competitive electrochemical performance and promising stability compared to other symmetrical cells reported in the literature.

Results and Discussions
The operating conditions of the transforming cell were determined by conducting in-situ X-ray diffraction (XRD) under hydrogen with the pre-sintered Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3-δ (PBMCo)-LSGM electrode. As shown in Fig. 2, PBMCo has simple perovskite mixed cubic and hexagonal layers at temperatures under 600 °C. A layered perovskite structure was successfully with the disappearance of the hexagonal peaks in the temperature range from 700 °C to 800 °C. The metal peak was observed at around 2θ = 45° at 800 °C, indicating the exsolved cobalt. Taking these observations into consideration, the fuel electrode was reduced at 800 °C for subsequent measurements. Figure 3a illustrates the phase transition and metal exsolution of atmospheric-adapted material PBMCo. During fabrication, the air and fuel electrodes sintered in air show simple perovskite structure, implying cobalt-doped PBMO (Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3-δ , PBMCo). In operation, the fuel electrode was exposed to an H 2 atmosphere and thus, specialized by phase transition and cobalt metal exsolution (PrBaMn 1.7 Co 0.3-x O 5+δ + xCo, S-PBMCo), while the air electrode retained the simple perovskite structure. To confirm the phase transition and the cobalt exsolution, X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) were conducted. In Fig. 3b,c, XRD patterns identify the phases of the as-sintered samples before and after reduction in H 2 at 800 °C, representing PBMCo-LSGM and S-PBMCo-LSGM, respectively. Before the reduction, the XRD patterns can be well-indexed to the crystal planes of simple perovskite and LSGM without any undesired reaction. After reduction, the patterns involve the phases of cation-ordered perovskite and LSGM, combined with peaks of cobalt metal caused by the exsolution phenomenon. These results are in good agreement with the expectation that the cobalt is doped in PBMO at the air-electrode side and is exsolved at the fuel-electrode side during operation.
In Figure 3d,e, the surface electronic states of cobalt were analyzed with XPS profiles of Co2p for PBMCo-LSGM and S-PBMCo-LSGM. The binding energy (BE) was calibrated with the C1s peak at 284.5 eV. Considering the overlap between the peaks of the Ba 3d and Co 2p, the Ba 3d peaks of Pr 0.5 Ba 0.5 MnO 3 and PrBaMn 2 O 5 were analyzed to observe the change of Ba 3d during reduction and the corresponding phase  Fig. S1, implying that the Ba 3d peaks are independent of the phase transition. In this regard, changes in the XPS peak at a BE range of 770-810 eV demonstrate a change in the Co 2p peak during the phase transition of PBMCo and S-PBMCo. Based on the fitted results, the Co 3+ /Co 2+ increases from 0.56 for PBMCo to 1.43 for S-PBMCo can be attributed to lattice defects and corresponding cobalt oxidation (Co 2+ → Co 3+ ). Additionally, the cobalt oxidation was also supported by the magnesium reduction (Mn 3+/4+ → Mn 2+ ) as confirmed by the XPS results of the Mn 2p 3/2 peaks for PBMCo and S-PBMCo in Fig. S2. The appearance of cobalt metal peaks in S-PBMCo demonstrates that the cobalt exsolution occurs under an operating condition (H 2 , 800 °C). A cross-sectional view of the tri-layer cell (PBMCo-LSGM/ LSGM/S-PBMCo-LSGM) is shown in Fig. S3. The dense LSGM electrolyte was about 60 μm thick and appeared to be in good connection with the electrode layer. The scanning electron microscopy (SEM) images of the electrodes ( Fig. 3f-h) show that the nano-sized particles are formed and uniformly distributed on the LSGM, retaining its morphology even after reduction. These microstructural properties of the cell with the nanostructured electrodes and the thin electrolyte layer are expected that the self-transforming cells will show high electrochemical performance.
For the electrochemical analysis, the impedance spectra (Fig. S4) of the transforming cells were measured at 700, 750 and 800 °C. The ohmic loss of the cell was measured by the high-frequency intercept with the real axis to be 0.147, 0.187, and 0.244 Ω cm 2 at 800, 750, 700 °C, respectively. The polarization resistance originated from the electrodes was determined by the difference between the two intercepts on the real axis to be 0.139, 0.273, and 0.530 Ω cm 2 at 800, 750, 700 °C, respectively. To estimate the cathodic loss, we conducted the half-cell test using the PBMCo-LSGM/LSGM/PBMCo-LSGM cell. Figure S5a shows the Nyquist plots of PBMCo-LSGM at operating temperatures of 700, 750, and 800 °C, respectively, with the fitted line by the software EC-Lab. The ohmic resistance was eliminated to direct comparison of cathodic polarization resistance at the various operating temperature. The cathodic R p values of the PBMCo-LSGM were 0.033, 0.073, and 0.189 Ω cm 2 at 800, 750, and 700 °C, respectively. The fitted lines were in good agreement with the Nyquist plots on the basis of the equivalent circuit in the inset of Fig. S5b. In the equivalent circuit, L is an inductance caused by the cables; R 1 is ohmic resistance; R 2 is associated with the charge transfer during the migration and diffusion of oxygen ion from the reaction site into electrolyte lattice; R 3 is induced by the non-charge transfer related to oxygen surface exchange, solid-state diffusion, and gas-phase diffusion. The specific values of R 2 , R 3 and R p (R 2 + R 3 ) are also summarized in Table S1.  Figure 4a shows the power density and voltage as functions of current density for the transforming cell with humidified H 2 as fuel and air as the oxidant. As expected from low polarization resistance, the transforming cell presents remarkably high maximum power density of 1.10 W cm −2 at 800 °C without any additional catalyst in comparison with the experimental values of the symmetrical cell (Table S3) 3,[10][11][12]14,16,[27][28][29][30][31] . Consequently, the use of the atmospheric-adaptive electrode (PBMCo) provides excellent catalytic activity for each atmosphere by tailoring the structure itself. Such excellent performance should be attributed from the novel structure of infiltration and the thin electrolyte. Moreover, the transforming cell shows potential as an electrolysis cell, as shown in Fig. 4b. The electrochemical performance was recorded by scanning current density at the voltage range from 0 to 1.5 V with 3% H 2 O and 10% H 2 O (H 2 safe gas included), respectively. Even the small amount of H 2 O was applied, the transforming cell shows the reasonable electrochemical performance of −0.42 A cm −2 for 3 vol.% H 2 O and −0.62 A cm −2 for 10 vol.% H 2 O, respectively, at a cell voltage of 1.3 V at 800 °C. To evaluate the reversibility of operation as SOCs, the cycling between electrolysis mode at −0.2 A cm −2 and fuel cell mode at +0.2 A cm −2 was performed at 700 °C for 3 and 10 vol. % H 2 O, respectively. At both H 2 O amount, the cell shows a stable performance in electrolysis and fuel cell modes without detectable degradation (Fig. 4c). The cell under 10 vol. % H 2 O maintained the constant voltages in both electrolysis and fuel cell modes for 30 h (Fig. 4d). After reversible cycling for 6 h, the cell under 3 vol. % H 2 O was operated under SOFC modes with 3 vol. % H 2 O without any critical degradation during ~100 h (Fig. 4e). These results indicate that the transforming cell is a promising system for solid oxide cells as both fuel cells and electrolysis cells.

Methods
Preparation of the transforming cell. For the infiltration process, the preparation is just scaffold and pre-  Measurements. For the structural identification, X-ray diffraction (XRD) (Bruker D8 Advance) were conducted in a 2θ range from 20° to 60°. The microstructure was observed using scanning electron microscopy (SEM) (Nova SEM) and transmission electron microscopy (TEM) (JEOL, JEM-2100F). X-ray photoelectron spectroscopy (XPS) spectra were taken using ESCALAB 2 50XI from Thermo Fisher Scientific with a monochromated Al-Kα (ultraviolet He1, He2) X-ray source. To measure the electrochemical performance, the standard four-probe technique was employed with a BioLogic Potentiostat. The sample was prepared by attaching silver wires to both electrodes of the electrolyte supported cell as current collectors with silver paste. Then the cell was sealed onto a 10 mm diameter alumina tube with ceramic adhesive (Aremco, Ceramabond 552). For the electrochemical performance test on the fuel cell mode, the humidified hydrogen (3 vol.% H 2 O) and air were fed to the fuel electrode side and the air electrode side, respectively. On the electrolysis cell mode, the hydrogen containing 10 vol. % H 2 O was applied to the fuel electrode side. At the same time, area specific resistance (ASR) was measured by AC impedance spectroscopy at open circuit voltage (OCV) in the 10 −1 −10 4 Hz frequency range and was calculated with an active electrode area of 0.36 cm 2 .

Data Availability
The data that support the findings of this study are available from the corresponding authors upon reasonable request.