Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2026 Sep 16;11(40):61921–61936. doi: 10.1021/acsomega.6c07898

Investigation of Secondary Phase Evolution in LSCF Electrodes during Long-Term SOEC Operation by X‑ray Diffraction and Electron Microscopy Characterizations

Surendra B Karki 1,*, Danny J Edwards 1,*, Matthew J Olszta 1, Christopher Coyle 1, Long Q Le 1, Nathanael D Royer 1, Lorraine M Seymour 1, Kerry D Meinhardt 1, Anthony D Guzman 1, Olga A Marina 1, John S Hardy 1
PMCID: PMC13670294  PMID: 42889901

Abstract

To clarify the origin of performance degradation in solid oxide electrolysis cells (SOECs), it is essential to distinguish between phases formed during fabrication and those that evolve under operation, and to quantify their impact on electrode functionality. (La0.6Sr0.4)0.95Co0.2Fe0.8O3‑δ (LSCF) is a widely employed oxygen electrode material, yet its long-term stability in SOEC mode remains a critical challenge. In this work, we combined X-ray diffraction (XRD), scanning electron microscopy (SEM), and transmission electron microscopy (TEM) with energy-dispersive X-ray spectroscopy (EDS) to systematically compare as-received powders, untested electrodes, and electrodes operated for 2800 h. We identify that Fe–Co spinel and related secondary phases are already present after powder synthesis and cell fabrication but undergo significant compositional and morphological evolution during operation. Montage-based SEM/EDS analysis of the post-test electrode showed increases in both the area fractions of spinel and La or Sr-rich secondary phases after long-term operation. TEM analysis confirmed that the spinel phase in the untested electrode adopted an inverse cubic structure, while analysis of the tested electrode revealed compositional heterogeneity and cation redistribution among the evolved phases. These findings provide postmortem evidence of substantial secondary-phase evolution in the LSCF oxygen electrode and indicate its potential contribution to overall cell degradation.

graphic file with name ao6c07898_0016.webp

graphic file with name ao6c07898_0014.webp

1. Introduction

Solid oxide electrolysis cell (SOEC) technology has the potential to significantly influence the energy economy as it can help fulfill the need for energy sources to meet the demand for fuel and power through its eco-friendly and highly efficient electricity-to-fuel conversion. , A typical SOEC consists of a porous oxygen electrode (anode), a dense electrolyte, and a porous hydrogen electrode (cathode) in a “sandwich” structure. The hydrogen electrode is fed a gas stream composed of H2O and/or CO2 that is reduced to H2, CO, or syngas fuel by electrolysis. Operated in reverse, the same cell functions as a solid oxide fuel cell (SOFC), electrochemically converting H2 or CO fuel back into electricity.

During SOEC operation, steam or CO2 produces dissociated oxygen ions that move toward the oxygen electrode through the dense electrolyte, driven by an external voltage that is higher than the Nernst potential. Oxygen ions are oxidized to oxygen molecules in the oxygen electrode, followed by the release of electrons. Many advantages of the SOEC include its ability to convert CO2 and water into fuels, its compact, modular, on-demand, and scalable design, and its efficiency in utilizing surplus electricity from clean energy sources such as solar and wind, as well as from nuclear and hydroelectric sources, to create fuels, which can be converted back to electricity when needed, thereby acting as a scalable energy storage technology. The stability of the electrode is of great significance for the production of potentially ultralow-cost and highly efficient hydrogen through ultra-high steam electrolysis current densities (>3 A/cm2).

An effective SOEC is required to be stable for several thousand hours, with a maximum degradation rate of 0.75 per 1000 h. La1–x Sr x Co1–y Fe y O3‑δ (LSCF) catalysts with a perovskite structure (ABO3) are the most investigated electrode candidates. Among several LSCF compositions, La0.6Sr0.4Co0.2Fe0.8O3‑δ is one of the highly studied electrode materials for both SOFC and SOEC. , LSCF compositions have superior features such as mixed ionic and electronic conducting properties, a compatible thermal expansion coefficient (TEC) with ceria, superior oxygen reduction or evolution reaction (ORR or OER) kinetics, lower reaction activation energy, , and a suitable sintering temperature (∼1000 °C). LSCF is used as the air electrode at intermediate temperature because of its excellent electrical and catalytic performance.

LSCF suffers from phase instability during long-term operation in SOECs. LSCF shows performance degradation in SOEC/SOFC modes due to Sr segregation/enrichment, , spinel phase formation over time, and polarization resistance increase. The LSCF electrode degrades predominantly due to cation diffusion and the formation of SrZrO3, which is electrocatalytically inactive. − Sr diffusion, in particular, could also trigger the formation of spinel Co3O4 and SrZrO3 in the anode/electrolyte interface, resulting in the valence state reduction of iron from Fe4+ to Fe3+ after long-term operation in SOEC mode. For example, Frey et al. reported the formation of multiple detrimental phases such as SrCrO4, SrO, and SrZrO3 in an LSCF-based SOEC stack after 20000 h of testing. Thus, to prevent the formation of resistive phases such as SrZrO3 and La2Zr2O7, a ceria barrier layer between the oxygen electrode and the YSZ electrolyte is inserted into the cell. −

Although spinel phase formation, Sr segregation, and LSCF degradation have been widely reported in recent SOEC studies, − their spatial distribution and evolution across the oxygen electrode remain insufficiently understood. Conventional analysis methods in SEM/EDS are often based on either low-magnification mapping or small, localized high-magnification regions. In this work, montage-based high-resolution SEM (JEOL IT800 with very high beam current) with 170 mm2 EDS Aztec system was used to analyze phase distributions of the starting LSCF powder and the untested and long-term-tested LSCF electrodes. The high-count EDS system enabled low-voltage mapping with sufficient X-ray counts to distinguish chemically distinct particles and estimate their area fractions over a larger and more representative electrode area. This approach provided improved statistical confidence for revisiting the established spinel formation pathway and revealed additional phase heterogeneities that could be missed by localized analysis. Thus, XRD was first used to monitor secondary-phase formation during cell fabrication and thermal aging at the SOEC operating temperature. However, minor phases below the practical XRD detection limit of approximately 0.1–1 wt % may not be reliably detected. The primary focus was the study of the LSCF-based composite oxygen electrode after SOEC operation at 1.3 V, 50% steam, and 750 °C for 2800 h. Although cell fabrication promoted spinel formation, other minor secondary phases were not obviously detected by XRD. In contrast, electrochemical loading during SOEC operation increased the spinel area fraction and promoted the formation of additional secondary phases, as supported by TEM analysis.

2. Experimental Section

2.1. Materials, Cell Fabrication, and X-ray Diffraction

Commercial LSCF (Trans-Tech Ceramics) and yttrium doped ceria (YDC; Fuel Cell Materials) powders were used to prepare screen-printing inks. Firstly, the hydrogen electrode support was fabricated as described previously. − Then, a 20–24 μm-thick LSCF/YDC or pure LSCF layer with an active area of 1.5 cm2 was screen-printed and sintered at 1000 °C for 2 h. An Ag/Pd contact layer was screen-printed on the oxygen electrode, and Ag mesh was attached using Ag/Pd ink, while on the hydrogen electrode, a NiO ink was used to attach Ni mesh for current collection. Contact layers were not added to the cell used for thermal aging.

The cells were tested at 750 °C in potentiostatic mode at a constant voltage of 1.3 V using a potentiostat (Arbin Instruments). The oxygen electrode was exposed to air, and the hydrogen electrode was supplied with a gaseous mixture of 50% H2O in H2. The electrochemical impedance spectroscopy (EIS) data were recorded in a frequency range of 100,000 to 0.1 Hz (Gamry Instruments) both at the open-circuit voltage (OCV) and at 1.3 V with an AC amplitude of 20 mA. Distribution of relaxation times (DRT) analysis of the time-dependent EIS spectra was used to interpret the electrochemical phenomena of the tested cells. The details of the oxygen electrode, barrier layer, and electrolyte configurations of cell specimens used in this study are provided in Table ST1.

X-ray diffraction (XRD) studies were performed using a Bruker D8 advance diffractometer with Cu Kα1 radiation (λ = 1.540590 Å). The diffractometer contained a furnace where LSCF powders or a button cell was mounted, and XRD data were measured at various temperatures. Rietveld refinement of the diffraction data was performed using the GSAS-II software package.

2.2. Electron Microscopy Characterization

Three conditions were examined to interrogate the sources of degradation in the LSCF-based oxygen electrode fabricated on a NiO/YSZ electrode-supported button cell operated for 2800 h. The first condition examined was the LSCF powder used for fabricating the electrode. The nominal composition of the powder with elements in atomic (at) % is La 11.5 at. %, Sr 7.7 at. %, Co 4.0 at. %, Fe 16.2 at. %, and O 60.6 at. %. The second condition was an untested button cell used to probe the starting microstructural state of the SOEC button cell. The untested cells contained an Ag/Pd current collector, an LSCF/YDC composite oxygen electrode, a YDC barrier, a YSZ electrolyte, and Ni/YSZ hydrogen electrode layers. The third condition was the tested cell under the conditions specified. The untested and tested button cells were broken in half, one piece was selected from each cell, vacuum-impregnated with epoxy, mounted on edge in epoxy, and then polished through successive grinding and polishing steps to the same final colloidal silica finish as the powder. For all three conditions, the final polishing was done using OP-U colloidal silica on a vibratory polisher for 2–3 h. The metallography samples were then coated with 5–6 nm of carbon to remove charging effects under the high-current electron probes used for this study.

High-resolution SEM was conducted using a JEOL IT800 field-emission gun SEM equipped with an Oxford Ultim Max 170 mm2 silicon drift detector (SDD) energy-dispersive spectrometer (EDS). Beam parameters were chosen to maximize the count rate of the detector at less than 30–40% dead time to ensure high X-ray count maps could be acquired in a reasonable time frame (<10 min per map), while balancing the image quality, interaction volume, and X-ray generation volume to achieve maximum sensitivity for the phases present in these porous oxide samples. The Oxford AZtec software package can acquire elemental maps that can then be analyzed by manually selecting areas or regions for composition, or by using their Phase Imaging module, a cluster-analysis routine that gathers pixels of similar X-ray spectra and clusters them into a discrete phase. With the proper choice of parameters to ensure 100 counts per pixel, the phase imaging module was able to consistently identify the phases in each sample over a wide area. Phase Map can find all of the pixels that belong to a particular phase, for example, the Fe–Co spinel particles, and provides a collective phase spectrum with high X-ray counts to improve upon the quantitative analysis. By comparison, manually selecting each small spinel particle will yield a reasonable but noisier spectrum due to a much smaller number of X-ray counts.

To provide additional confirmation of the structure and respective chemistry of different phases, TEM analysis was used to probe the microstructure of LSCF/YDC in the untested and tested cells. TEM lamellae were extracted from the oxygen electrode regions using a Thermo Fisher 3D Quanta FIB/SEM. The FIB lamella was thinned to ∼100 nm using a final Ga beam of 2 kV; however, the porous nature of the sample led to curtaining in some parts of the sample. Despite this, a sufficient thin area was present with uniform thickness to allow elemental mapping over areas not affected by curtaining. A cold-field emission gun, aberration-corrected JEOL ARM200CF transmission electron microscope, equipped with a high-angle annular dark-field detector and bright-field scanning transmission electron microscopy (STEM) detectors, was used to image the TEM samples, and elemental maps were acquired using the Centurio EDS detector, a silicon drift detector with a large collection angle of 0.98 Sr. A 0.18 nm probe with ∼1 nA was used to acquire elemental maps from selected regions. Thermo Fisher Pathfinder software was used to collect the elemental maps and to analyze the data obtained from the elemental maps using the principal component analysis module, Compass. Compass determines the principal components of the spectral maps, and its Phase module can be used to combine principal components as deemed necessary. The process is similar to the overall steps in the Phase Mapping module in the AZtec software, a clustering algorithm allowing pixels of similar spectral characteristics to be grouped in discrete phases or structural features (such as epoxy-filled holes or regions with colloidal silica) and validated by the user.

3. Results and Discussion

Under steam electrolysis at 1.3 V, 50% steam, and 750 °C, the cell with an LSCF/YDC oxygen electrode initially delivered ∼1.05 A/cm2; the cell rapidly continued to lose performance for 1500 h at a degradation rate of ∼18% per 1000 h and then stabilized at ∼0.75 A/cm2 over 2800 h (Figure S1a). The degradation rate after 1500 h was determined to be 0.65% per 1000 h. The impedance data show that this loss is dominated by an increase in polarization resistance from ∼0.28 to ∼0.44 Ω cm2 (∼58% increase), whereas the ohmic resistance remained low and nearly constant around 0.07–0.08 Ω cm2 (Figure S1b). These results suggest that the observed degradation is dominated by electrode-related processes, with minimal contribution from the electrolyte or contact resistance. This interpretation is further supported by the DRT analysis of the impedance spectra (Figure S1c), where the most pronounced time-dependent changes occur in the ∼102–103 Hz, a frequency range commonly attributed to electrode charge-transfer and interfacial electrochemical processes. The initially higher degradation rate may reflect concurrent changes at both electrodes. Under electrolysis conditions, the Ni/YSZ hydrogen electrode may undergo Ni oxidation, migration/agglomeration, and loss of triple-phase-boundary sites, whereas cation redistribution and Sr segregation in the LSCF-based electrode may reduce its electrochemical activity. The subsequent lower degradation rate may reflect a transition to slower microstructural evolution. Although the overall degradation likely involves contributions from both electrodes, the present study focuses on the microstructural changes of the oxygen electrode.

Having established the overall electrochemical performance and quantified the degradation rate of the LSCF/YDC oxygen electrode during long-term SOEC operation, we next investigated the origin of the rapid degradation behavior by correlating it with phase evolution and microstructural changes in the oxygen electrode. First, X-ray diffraction is used to identify secondary phases present in the as-received LSCF powder and in the as-fabricated cell (Section ). We then employ SEM-based phase mapping to track how these phases are distributed and how their fractions evolve in the untested cell with the YDC barrier layer and in the cell after 2800 h of operation (Section ). Finally, TEM analysis (Section ) provides high-resolution insight into the nanoscale morphology and chemistry of the LSCF/secondary-phase interfaces that are most relevant to the observed electrochemical degradation.

3.1. XRD Analysis of the LSCF Powder and As-Fabricated Cell

Figure a shows the overlay of XRD patterns of the as-received LSCF powder after heating over a range of temperatures from 700 to 900 °C. The high-temperature XRD data were collected by letting the powder dwell at each temperature for 2 h to monitor for secondary phase formation. The main purpose of this experiment was to document if the LSCF powder was decomposed into other phases such as spinel. Similarly, the powder was allowed to age for a week at 750 °C (Figure b), which was the SOEC operating temperature. Secondary phase formation was not observed in 2 h of thermal aging at these temperatures.

1.

1

Open in a new tab

Overlay of XRD patterns of thermally aged LSCF powders: (a) for 2 h at temperatures ranging from 700–900 °C in 100 °C increments, and (b) for 168 h at 750 °C, with the inset showing a magnified view of the main diffraction peak to highlight changes in unit cell size.

A button cell was fabricated on a YSZ/NiO-YSZ electrode-supported substrate with a screen-printed YDC barrier. The oxygen electrode was then fabricated by screen-printing a pure LSCF ink, followed by sintering at 1000 °C for 2 h. The cell was mounted in an X-ray diffractometer equipped with a furnace, and XRD data were collected at temperatures between 700 and 900 °C. The cell was not electrochemically tested. The XRD data (Figure a) revealed a peak at 35.5° corresponding to the Fe–Co spinel phase. The cell aged at 750 °C for 132 h, with XRD data collected every 12 h, which further confirmed the presence of the spinel phase (Figure b).

2.

2

Open in a new tab

XRD patterns of thermally aged LSCF electrode in an untested cell: (a) the peak of the spinel phase at different temperatures, and (b) monitoring of the spinel phase during thermal aging at 750 °C did not show obvious or detectable growth.

XRD showed that the LSCF peaks shifted to lower angles after thermal aging, indicating lattice expansion. Results of the Rietveld refinement confirmed that the cubic lattice parameter of the powder increased from 3.92266 Å after 2 h to 3.93189 Å after 168 h (Figure S2), consistent with literature values for thermally aged LSCF at 750 °C. This change may result from Sr loss, cation exsolution, and/or changes in oxygen nonstoichiometry. − Oxygen loss can increase oxygen-vacancy concentration and cause chemical expansion of the LSCF lattice. However, this study did not directly measure oxygen stoichiometry, so its effect cannot be distinguished from Sr loss or cation exsolution. Minor secondary phases, if present, may have remained below the XRD detection limit. The Fe–Co spinel phase formed during cell fabrication and persisted after 132 h at 750 °C. Its lattice parameter decreased from 8.4403 Å to 8.3091 Å (Figure S3), suggesting changes in composition or defect chemistry during aging. , The high sintering temperature of 1000 °C and subsequent thermal exposure may have promoted LSCF demixing and secondary-phase evolution. Because Sr excorporation is thermally activated, higher temperatures can further accelerate this process. , Such LSCF destabilization, together with Ni migration in the fuel electrode under electrolysis conditions, has been linked to the higher degradation rates of SOECs relative to SOFCs. ,

To further confirm the appearance of a spinel phase that was hinted at by the results of XRD studies and to investigate the presence of other phases, SEM/TEM analysis was assessed. The focus of this work is to determine if there were changes in the oxide phases and elemental distributions of the oxygen electrode that might lead to the slow degradation observed over 2800 h at 750 °C. Although these results were preliminary and require further study, we observed some significant differences that might be the source of the observed degradation. Sr exsolution from the structure of LSCF can potentially initiate several side reactions, such as the formation of sulfur (S)- and phosphorous (P)-containing phases, with S from the air or P from additives such as dispersants or other components used during cell fabrication. S poisoning has been previously reported in LSCF cathodes, with SO2 being the main source.

3.2. Scanning Electron Microscopy Studies

The approach and choice of parameters used for the elemental mapping are provided in Supporting Information and summarized in Table ST5. The overall intent of this work was to reveal the spatial distribution of the various phases that are often seen in bulk analysis techniques such as X-ray diffraction. To reach this goal of observing the different phases and particles, the accelerating voltage and beam currents were optimized to achieve better discrimination of the small, closely spaced particles of different phases.

3.2.1. Phase Distributions in LSCF Powders

Compositions were obtained from the large areas indicated by the white borders at various locations on the precursor powder, as shown in Figure , and the resulting compositions are reported in Table . The measured compositions differed somewhat from the stated composition provided with the purchased powders, notably in that the atomic percent (at %) of oxygen is low and that of Fe and La are higher than the nominal composition. This is attributed to the use of standardless quantification.

3.

3

Open in a new tab

BSE image of mounted and polished LSCF powders used for the electrode layer. Five regions were selected for EDS acquisition.

1. Compositions of LSCF Powders Obtained from Several Selected Areas of Mount.
Spectrum Label La at% Sr at% Co at% Fe at% O at%
Nominal Composition 11.5 7.7 4.0 16.2 60.61
1 15.5 8.2 4.7 18.5 52.2
2 14.7 9.0 4.6 17.6 54.2
3 14.9 9.2 4.6 18.4 54.9
4 14.7 9.0 4.5 17.6 54.3
5 15.1 9.2 4.7 18.3 52.7
Open in a new tab

In Figure , a montage of four elemental maps taken from adjacent areas in 2048 × 1536 pixels at 15 k × magnification and 10 kV is shown, highlighting elemental variation throughout the sample that leads to at least three different phases. The elemental maps for each element were thresholded to reveal the individual hot spots, which, in turn, was used for the color overlay in Figure a. These maps were taken at a beam current of ∼10 nA with a dead time of ∼38%, giving a total of 529 million counts at a 200 μs dwell time for all merged images. The pixel spacing is 4 nm with no binning, and 17 nm at 4 binning, yielding an average X-ray count at each pixel of 46 and 728, respectively. While this ensures that, image-wise, small particles aren’t missed, the interaction volume at 10 kV is likely around 500 nm in width, so the chosen parameters are sampling a much larger volume than the initial probe (<10 nm diameter).

4.

4

Open in a new tab

LSCF Powder: Elemental TruMaps were taken from four adjacent areas at 15 k× to create a montage. The color overlay map in (a) is generated from the individual elemental maps shown in the set of images in (c-–f). Maps were 4 × 4 binned to increase the X-ray count in each grouped pixel and thresholded to reveal the hotspots for each element.

The elemental maps show that there are hot spots for each element, which potentially indicate discrete particles of phases other than LSCF. In Figure , white borders were hand-drawn around several hundred particles using the tools within the AZtec software, and the EDS spectra from all of the particles within the borders were analyzed. The results of that composition analysis are summarized in the plot in Figure b. Many of the particles have a composition that revolves around the LSCF phase, but it appears from the maps and the selected compositions that there is chemical heterogeneity within the LSCF particles. This was primarily related to a shift to higher Fe and Sr contents, offset by a loss of Co and La, leading to Sr–Fe-enriched regions. Some of the particles clearly exhibited a composition of an iron-rich Fe–Co spinel phase, whose composition often fluctuated around 25 at. % Fe and 15–17 at. % Co, with minor concentrations of La and Sr. As shown in Figure , close examination of some of the largest Fe–Co spinel particles suggests that the La and Sr may come from the electron-beam sampling of surrounding particles of other phases, so they may be close to a pure Fe–Co spinel phase. Other phases found in addition to the Sr–Fe-rich LSCF and Fe–Co spinel are particles that appear to be some form of La­(FeCo) oxide phases, with the Fe and Co levels changing depending on the particle. One particle was found with 35% Co and 13% Fe, and very little Sr or La. The spinel particles ranged in size from 150 to 550 nm in diameter, and the La­(FeCo) oxide particles ranged in size 150–350 nm in diameter Table .

5.

5

Open in a new tab

LSCF Powders: Selected particles were chosen from the elemental map, as highlighted in (a) to analyze their composition, a summary of which is shown in (b). The compositions are plotted against their Fe/Co ratio, which proved a useful method to discriminate the various phase compositions. 57 individual particles were selected across the montage. See Supporting Information on the approach used to derive these plots.

6.

6

Open in a new tab

LSCF Powders: The phase map derived from the elemental maps shown in Figure reveals a small amount of Fe–Co spinel, La­(CoFe) oxide particles, and one Co–Fe spinel. The phase map cluster analysis also indicates a Sr-rich LSCF phase is present. The area fractions derived from the Phase Map analysis are provided in Table 2. See the text for an explanation.

2. Phases and the Area Fraction/Pixel Count for Each Phase.
Phase Fraction (%) Pixel Count
LSCF 92.9 672819
Sr-rich LSCF 5.7 41010
Fe–Co spinel 0.3 2446
La(CoFe) oxide 0.3 2179
Co–Fe Spinel 0.2 1090
Open in a new tab

Using the Phase Map routine within AZtec, the binned (4 × from original resolution) elemental maps were analyzed using a boundary tolerance of 5 and a grouping level of 1, which deviates from the default settings of 3 and 2, respectively. The initial cluster analysis revealed up to 10 phases, but many were simply variants of the same phase, such as the outer shell of a Co–Fe spinel particle. Subsequently, manual assessment led to many of the phases being merged together to mimic the hot spots observed in the elemental maps, which led to a final phase distribution of a Sr-rich (lower La and Co) phase, Fe–Co spinel particles, and a series of particles lumped into a La­(FeCo) oxide phase. Individually, each particle fluctuated from this mean value; only the Fe–Co spinel exhibited a narrow composition range. The Sr-rich phase may be a direct derivative of the loss of Co and La to the other phases, leaving behind some LSCF particles that are richer in Sr and possibly Fe.

3.2.2. Phase Fractions in an Untested Cell with a YDC Barrier

Six elemental maps of the LSCF untested electrode near the YDC barrier were taken at 10 kV and 15×, then stitched into a montage with a 15% overlap. The map had a total of 722 million counts. The color variations in the elemental map shown in Figure a reveal a non-uniform distribution of elements, with local hot spots that are Co-rich and La-rich, independently. A large YDC particle and the surrounding area rich in La and Fe are captured in this map and are highlighted by the yellow dashed box on both the elemental map and the phase map in Figure a and b, respectively. The phase map was able to show this clearly in Figure f, which has both regions merged together.

7.

7

Open in a new tab

Untested YDC Cell: 10 kV elemental map and resulting phase maps. The montage was taken from six adjacent areas and stitched together.

The cluster analysis routine in Phase Map yielded three phases in the electrode: LSCF particles, Fe–Co spinel particles, and La­(CoFe) oxide particles. One small particle of strontium zirconate (SrZrO3) was discovered by the cluster analysis and is shown in Figure f. Zr diffusion to the oxygen electrode is unlikely, and the particle most likely comes from Zr-based grinding media used during the oxygen electrode powder milling process. The composition, area fraction, and average diameter (the latter was not measured for the LSCF particles) of these three phases are provided in Table . The cluster analysis did not yield consistent evidence of a Sr-rich LSCF phase as was observed in the LSCF powders; however, the individual compositions from multiple selected areas indicated that it was occurring. Figure shows a plot of the compositions extracted from multiple discrete regions that were manually selected with great care from the colored overlay and elemental maps based on perceived differences in the composition of the various elements. The results indicate that there is a clustering of particles whose compositions correspond to the LSCF phase, with some variability in composition that may be due to edge effects and varying particle size or real variations in the Fe and Sr content as observed in the powders. The composition plots show clustering in the 20–25 Fe at. % range that corresponds to the Fe–Co spinel (Fe3O4 structure type). The region highlighted in the green box corresponds to the La­(CoFe) oxide with varying Co/Fe ratios.

3. Untested Cell: Details on Distributions and Chemistry (at. %) of the Three Primary Phases Identified in the LSCF Oxygen Electrode.
Label La Sr Co Fe O Area Fraction % Avg. Diameter (nm)
LSCF 17.1 7.9 5.6 19.2 50.2 98.13 Not measured
Fe–Co spinel 7.6 4.1 15.0 23.7 49.7 1.40 62 particles measured, 230 ± 137
La(CoFe) Oxide 18.8 5.1 12.0 10.4 53.7 0.47 61 particles measured, 137 ± 67.7
Open in a new tab
8.

8

Open in a new tab

Untested Cell: Selected particles were chosen from the elemental map in Figure a to analyze a random selection of particles. Their composition plotted against the Fe/Co ratio for each selection is presented in Figure b. 128 individual particles were selected across the montage. See Supporting Information on the approach used to derive these plots.

Both the discrete measurements taken from manually selected areas and the Phase Imaging results, which produce a single spectrum from all the X-ray counts collected in each pixel belonging to a single phase, show that the Fe–Co spinel is highly enriched in Fe. However, the phase imaging captures both large and small particles, hence, the varying particle size means the beam interaction volume may increasingly sample regions outside of the actual Fe–Co spinel particle as the particle size decreases, which may account for the La and Sr content. Selecting the center of the largest Fe–Co spinel particle yields a composition that is enriched in Fe and Co, with very little La and Sr.

3.2.3. Phase Distributions in a Cell Tested at 1.3 V, 50% Steam, and 750 °C for 2800 Hours

The tested cell shows a clear difference in the area fraction of secondary phases. Figure shows a montage of four elemental maps over a larger area using 10 kV and 15 k × for each image. Maps were taken in 2048 × 1536 pixels with a dwell time of 200 μs. The montage data has 602 million X-ray counts, providing over 740 X-ray counts per pixel on average when the maps were binned at 4, which was more than sufficient to reveal the spinel particles using the cluster analysis algorithm in Phase Map. Examination of the elemental maps, however, indicated that La­(FeCo) oxide particles should have been present, so the binning was increased to 8, which yielded 2900 counts per pixel on average. This allowed the Phase Map cluster analysis to detect both the Fe–Co spinel particles and the La­(FeCo) oxide particles, as shown in Figure c and d, respectively. The area fraction of spinel in this region is ∼5%, considerably higher than found in the untested cell. The chemistry, area fractions of the two phases, and their average sizes are presented in Table . The plot of the various particle compositions derived from hand-selected borders on the elemental-map overlays is shown in Figure e.

9.

9

Open in a new tab

Tested Cell: BSE image is shown in (a), and (b) shows the color overlay of the Co, Fe, and La elemental maps taken at 10 kV and 15 k×. Phase mapping revealed the phase distribution in (c) and (d) for the Fe–Co spinel and La­(FeCo) oxide phase, respectively. The plot in (e) was taken from 112 different particle regions corresponding to each of the three phases.

4. Tested Cell: Details on Distributions and Chemistry (at. %) of the Three Primary Phases Identified in the LSCF Oxygen Electrode.
Label La Sr Co Fe O Area Fraction % Avg. Diameter (nm)
LSCF 12.7 4.6 5.7 19.0 58.0 93.20 Not measured
Fe–Co spinel 7.8 2.6 11.4 22.5 55.0 5.0 152 ± 135
La(CoFe) Oxide 15.0 4.0 10.1 14.9 56.0 1.8 118 ± 76.4
Open in a new tab

3.3. TEM Analysis

The same cells used for SEM studies were also used to conduct limited TEM and STEM analyses on both untested and tested cells.

3.3.1. Untested Cell

For the untested cell used for SEM analysis, a FIB sample was extracted from the LSCF phase. Figure shows an elemental map taken from an area that includes a large Fe–Co spinel particle. The elemental maps show that there are two other phases in addition to the LSCF main phase and the Fe–Co spinel phase. Table shows the compositions of the four phases found in the STEM Elemental Map in an untested cell. The first phase is labeled La­(CoFe) oxide due to the higher levels of La and Co, and the second phase is labeled as LaFe oxide due to the high La and Fe compositions. The oxygen contents are highly variable and rarely match the expected phase composition unless the particle is large enough for the interaction volume from the X-ray beam to be completely confined within the particle analysis area, which is a difficult constraint for these materials with such fine particle sizes and overlapping particles. Oxygen is a difficult element to quantify without the use of elemental standards for the phase, which was not available for this work; hence the work was done using standardless quantitative routines with the EDS software. Variations with sample tilt, thickness, and adjacent or overlapping phases can all contribute to altering the oxygen content. Figure presents another region taken adjacent to the barrier layer, showing a spinel particle, an La­(FeCo) phase, and two other phases with varying levels of Fe, Sr, and La, as noted in Table . Note that one phase that is low in La and rich in Fe and Sr is observed, an observation confirming the SEM results. STEM/HAADF (High-Angle Annular Dark-Field) images of the crystal structure of the single grain of the spinel phase are shown in Figure , confirming its inverse cubic spinel structure (Fe3O4 prototype). The composition was extracted from the center of the particle where no overlap with nearby grains occurs, and it proved to be a pure Fe–Co spinel; no La or Sr were present in the STEM analysis.

10.

10

Open in a new tab

Elemental mapping of an untested cell reveals the presence of four different phases, including a large Fe–Co spinel particle (red phase). The Fe–Co spinel composition is an Fe-rich spinel - (FeCo)3O4 with no La or Sr in the larger particle.

5. Untested Cell: Compositions of the Four Phases Found in the above STEM Elemental Map in the LSCF Regions of the Electrode.
TEM Phase La at% Sr at% Co at% Fe at% O at%
La(CoFe) Oxide 28.4 2.9 20 6.1 42.6
LaFe Oxide 24.4 1.1 0.3 19.4 54.8
Fe–Co spinel - - 18.7 28.3 55
LSCF 21.3 12.8 4 26.6 35.3
Open in a new tab
11.

11

Open in a new tab

Untested cell with LSCF region adjacent to the YDC barrier. No evidence of any migration of Sr or La into the YDC. Two oxide phases were also found along with the cubic spinel Fe–Co spinel phase and LSCF.

6. Untested Cell: Compositions of the Four Phases Found in the above STEM Elemental Map in the LSCF Regions Adjacent to YDC Barrier.
Phase La at% Sr at% Co at% Fe at% O at%
Fe–Co Spinel 2.8 0.1 16.4 24.0 56.7
La(FeCo) oxide 18.5 5.0 9.9 10.2 56.4
Sr-rich LSCF 7.5 14.6 1.0 18.2 58.7
LSCF 21.7 11.6 3.8 24 39
Open in a new tab
12.

12

Open in a new tab

Untested cell: STEM­(HAADF) images from cell with a YDC barrier: a) Low magnification image showing the relation of the Fe–Co spinel (red dashed box). b) Medium magnification image showing the faceted surface ([111] family of planes) of the spinel. c) Atomic column imaging of the Fe3O4 inverse spinel structure in the ⟨110⟩ orientation confirming its structure.

3.3.2. Tested Cell

Elemental maps were taken from a TEM sample extracted from the tested cell, revealing a higher area fraction of small Fe–Co spinel particles that were not easy to find in the untested cell. Two maps are shown in Figure . Small spinel particles are present in each map, as well as two other oxide phases: La­(CoFe) and (LaSr)­Fe. The compositions and area fractions of these four phases are summarized in Table . The La-rich La­(CoFe) phase is similar to the phase found in the untested cell, but the Sr-rich (LaSr)Fe phase was not observed in the untested cell. The (LaSr)Fe phase appears to be a phase present at the interfaces among other phase particles.

13.

13

Open in a new tab

Tested cell: Elemental maps and HAADF imaging revealed the presence of a higher area fraction of the Fe–Co spinel phase, including the presence of a higher density of small particles.

7. Tested Cell: Compositions of the Four Phases Found in the above STEM Elemental Map.
TEM Phase Area% La at% Sr at% Co at% Fe at% O at%
La(CoFe) Oxide 21.8 20.9 1.3 10.2 7.2 60.3
Fe–Co spinel 16.0 5.3 0 13.1 23.7 63.0
(LaSr)Fe oxide - 6.1 14.9 0.9 16.4 61.8
LSCF 62.2 14.9 8.5 1.7 17.7 57.2
Open in a new tab

Fe–Co spinel phases were the primary phase of interest since they represented a significant volume fraction. Large-area montages that were collected to overcome the inhomogeneity of the microstructure provided evidence that the Fe–Co spinels increased in area fraction (from 1.5% in the untested condition to over 5% in the tested cells), nucleating small, 100 nm-sized spinel particles that were easily visible in the TEM images. SEM also resolved some of these particles. TEM confirmed that the spinel particles consistently had a composition of pure Fe–Co oxides with a nominal Fe-rich composition of 26% Fe and 13–15% Co, with no La or Sr present. Reports in the literature have confirmed the formation of Co3O4-type spinel within LSCF electrodes operated in SOEC mode through SEM and HAADF/STEM analyses.

In addition to the increased Fe–Co spinel population, the large-area montages revealed other compositionally distinct regions in the tested cell, including a La-rich La­(FeCo) oxide phase with variable Fe/Co ratios and an Sr-rich (LaSr)Fe oxide phase. Because EDS alone cannot determine the oxygen stoichiometry or crystallographic structure of these phases, it would be less accurate to present the exact oxygen content in the observed phases. In one examined area of the tested cell, La-rich oxide phases extending over several micrometers were observed. It is also difficult to establish whether these regions were associated with cracks or porosity. Additional targeted cross-sectional TEM analysis will be required to identify any secondary phases present at crack or pore interfaces. Similar degradation phenomena have been reported for La0.8Sr0.2MnO3-based oxygen electrodes, where cracking and delamination have been associated with interfacial thermal stresses and oxygen accumulation, and for La0.6Sr0.4CoO3‑δ electrodes, where closed pores can form during annealing or anodic polarization.

The compositional evolution observed in the LSCF-based oxygen electrode is consistent with local decomposition and cation demixing, which can be represented schematically as

La0.6Sr0.4Co0.2Fe0.8O3−δ→(1−x)La‐rich(LaSr)(FeCo)O3−δ+(y)Sr‐rich(LaSr)(Fe)O3−δ+(z)(CoFe)3O4

Here, the coefficients x, y, and z can be used as phase fractions to illustrate the compositional partitioning of the parent LSCF phase into La-rich, Sr-rich, and Fe–Co spinel regions. Under electrolysis operation, anodic polarization increases the local partial oxygen pressure near the electrode/barrier interface, accelerating Sr segregation and LSCF decomposition relative to SOFC operation. The oxygen-potential gradient can also drive kinetic cation demixing and electromigration. In addition, Sr segregation is favored by elastic and electrostatic interactions associated with the size mismatch of Sr2+ and its interaction with oxygen vacancies. Reaction with the barrier layer, including Sr uptake into GDC, can provide a thermodynamic sink for Sr and further sustain local Sr depletion from the LSCF matrix.

The parent LSCF composition typically exhibits an electronic conductivity of approximately 200–300 S/cm at 800 °C and a Thermal Expansion Coefficient (TEC) of approximately 15–17 × 10–6 K–1. In comparison, La-rich LSCF (La0.9Sr0.1Co0.2Fe0.8O3‑δ) and Sr-rich LSCF (La0.3Sr0.7Co0.2Fe0.8O3‑δ) exhibited electrical conductivities of approximately 60 and 46 S/cm, respectively, at 800 °C, with their corresponding TEC values of 16 and 27 × 10–6 K–1. In comparison, the Fe-rich spinel (CoFe2O4) exhibited a substantially lower electronic conductivity of approximately 0.93 S/cm at 800 °C and a TEC of approximately 14.9 × 10–6 K–1. Therefore, the formation of Fe-rich spinel and Sr-rich LSCF regions may reduce electronic transport and introduce local thermal-expansion mismatch within the electrode that could introduce pores or cracks. Therefore, the formation of Fe–Co spinel and Sr-rich LSCF regions can block electronic conduction pathways and introduce local thermal-expansion mismatch within the electrode, particularly for the Sr-rich regions.

On the basis of the microstructural, compositional, EIS, and DRT observations, the proposed degradation mechanism is that secondary-phase formation, initiated during fabrication and accelerated by electrolysis-mode polarization, depletes and partitioned the LSCF matrix. The resulting Fe–Co spinel, La-rich, and Sr-rich regions reduce electronic connectivity, alter local oxygen vacancy concentrations, and decrease oxygen surface exchange and bulk transport activities. Simultaneously, the high TEC of Sr-rich regions introduces localized mechanical stresses, which can promote pore formation or cracking. Together, the loss of electronic percolation, oxygen-electrode active area, and microstructural continuity contributes to the observed increase in polarization resistance during electrolysis operation. EDS does not establish the oxygen stoichiometry or crystallographic structure. Therefore, the phase assignments remain compositional, and the association of La-rich regions with cracks versus porosity is not resolved by the present montages.

4. Conclusion

From XRD, SEM, and TEM analyses of the LSCF electrode, we have confirmed the formation of the spinel phase during cell fabrication. The results demonstrated that the overall composition of the starting LSCF powders closely matched that of the vendor’s specifications. Moreover, both untested and tested cells showed spinel formation, which was from 1.4% in the untested to as high as 5% in the tested cells. The spinel phase was quantified as having a composition of 26% Fe and 13–15% Co, with no La or Sr present according to TEM. Other phases, La­(FeCo) and (LaSr)Fe oxide, were also observed in both SEM and TEM. However, the distributions of these phases need to be better investigated using both TEM and SEM to tease out their spatial distribution and area fraction.

Supplementary Material

ao6c07898_si_001.pdf (340.2KB, pdf)

Acknowledgments

This material is based upon work supported by the U.S. Department of Energy, Office of Critical Minerals and Energy Innovation (CMEI), specifically the Hydrogen Fuel Cell Technology Office under Agreement no. 37440. PNNL is a multiprogram national laboratory operated for the U.S. Department of Energy by Battelle under Contract DOE-AC05-76RL01830.

Data supporting the findings of this study can be obtained from the corresponding author upon reasonable request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c07898.

  • Long-term SOEC stability testing data (Figure S1); Rietveld refinement profiles of LSCF powders after heat treatment (Figure S2); spinel lattice evolution data (Figure S3); details of electrode specimens (Table ST1); refined structural parameters from Rietveld refinement analysis (Tables ST2–4); description of SEM parameter selection; and details of the elemental and phase maps (Table ST5) (PDF)

The authors declare no competing financial interest.

References

  1. Dau H., Limberg C., Reier T., Risch M., Roggan S., Strasser P.. The Mechanism of Water Oxidation: From Electrolysis via Homogeneous to Biological Catalysis. ChemCatChem. 2010;2:724–761. doi: 10.1002/cctc.201000126. [DOI] [Google Scholar]
  2. Jensen S. H., Larsen P. H., Mogensen M.. Hydrogen and synthetic fuel production from renewable energy sources. Int. J. Hydrogen Energy. 2007;32:3253–3257. doi: 10.1016/j.ijhydene.2007.04.042. [DOI] [Google Scholar]
  3. Zhang X., Song Y., Wang G., Bao X.. Co-electrolysis of CO2 and H2O in high-temperature solid oxide electrolysis cells: Recent advance in cathodes. J. Energy Chem. 2017;26:839–853. doi: 10.1016/j.jechem.2017.07.003. [DOI] [Google Scholar]
  4. Zheng Y., Wang J., Yu B., Zhang W., Chen J., Qiao J., Zhang J.. A review of high temperature co-electrolysis of H2O and CO2 to produce sustainable fuels using solid oxide electrolysis cells (SOECs): advanced materials and technology. Chem. Soc. Rev. 2017;46:1427–1463. doi: 10.1039/C6CS00403B. [DOI] [PubMed] [Google Scholar]
  5. Wang Y., Li W., Ma L., Li W., Liu X.. Degradation of solid oxide electrolysis cells: Phenomena, mechanisms, and emerging mitigation strategiesA review. J. Mater. Sci. Technol. 2020;55:35–55. doi: 10.1016/j.jmst.2019.07.026. [DOI] [Google Scholar]
  6. Wolf S. E., Winterhalder F. E., Vibhu V., de Haart L. G. J., Guillon O., Eichel R.-A., Menzler N. H.. Solid oxide electrolysis cells – current material development and industrial application. J. Mater. Chem. A. 2023;11:17977–18028. doi: 10.1039/D3TA02161K. [DOI] [Google Scholar]
  7. Teraoka Y., Zhang H. M., Okamoto K., Yamazoe N.. Mixed ionic-electronic conductivity of La1–xSrxCo1–yFeyO3−δ perovskite-type oxides. Mater. Res. Bull. 1988;23:51–58. doi: 10.1016/0025-5408(88)90224-3. [DOI] [Google Scholar]
  8. Kim D., Park J. W., Chae M. S., Jeong I., Park J. H., Kim K. J., Lee J. J., Jung C., Lee C.-W., Hong S.-T., Lee K. T.. An efficient and robust lanthanum strontium cobalt ferrite catalyst as a bifunctional oxygen electrode for reversible solid oxide cells. J. Mater. Chem. A. 2021;9:5507–5521. doi: 10.1039/D0TA11233J. [DOI] [Google Scholar]
  9. DiGiuseppe G., Sun L.. Electrochemical performance of a solid oxide fuel cell with an LSCF cathode under different oxygen concentrations. Int. J. Hydrogen Energy. 2011;36:5076–5087. doi: 10.1016/j.ijhydene.2011.01.017. [DOI] [Google Scholar]
  10. Tai L. W., Nasrallah M. M., Anderson H. U., Sparlin D. M., Sehlin S. R.. Structure and electrical properties of La1–xSrxCo1– yFeyO3. Part 2. The system La1–xSrxCo0.2Fe0.8O3 . Solid State Ionics. 1995;76:273–283. doi: 10.1016/0167-2738(94)00245-N. [DOI] [Google Scholar]
  11. Petric A., Huang P., Tietz F.. Evaluation of La–Sr–Co–Fe–O perovskites for solid oxide fuel cells and gas separation membranes. Solid State Ionics. 2000;135:719–725. doi: 10.1016/S0167-2738(00)00394-5. [DOI] [Google Scholar]
  12. Hong J., Bhardwaj A., Namgung Y., Bae H., Song S.-J.. Evaluation of the effects of nanocatalyst infiltration on the SOFC performance and electrode reaction kinetics using the transmission line model. J. Mater. Chem. A. 2020;8:23473–23487. doi: 10.1039/D0TA07166H. [DOI] [Google Scholar]
  13. Jiang S. P.. A comparison of O2 reduction reactions on porous (La,Sr)­MnO3 and (La,Sr)­(Co,Fe)­O3 electrodes. Solid State Ionics. 2002;146:1–22. doi: 10.1016/S0167-2738(01)00997-3. [DOI] [Google Scholar]
  14. Jiang S. P.. Development of lanthanum strontium cobalt ferrite perovskite electrodes of solid oxide fuel cells – A review. Int. J. Hydrogen Energy. 2019;44:7448–7493. doi: 10.1016/j.ijhydene.2019.01.212. [DOI] [Google Scholar]
  15. Yan Y., Fang Q., Blum L., Lehnert W.. Performance and degradation of an SOEC stack with different cell components. Electrochim. Acta. 2017;258:1254–1261. doi: 10.1016/j.electacta.2017.11.180. [DOI] [Google Scholar]
  16. Chen K., Jiang S. P.. ReviewMaterials Degradation of Solid Oxide Electrolysis Cells. J. Electrochem. Soc. 2016;163:F3070. doi: 10.1149/2.0101611jes. [DOI] [Google Scholar]
  17. Finsterbusch M., Lussier A., Schaefer J. A., Idzerda Y. U.. Electrochemically driven cation segregation in the mixed conductor La0.6Sr0.4Co0.2Fe0.8O3−δ . Solid State Ionics. 2012;212:77–80. doi: 10.1016/j.ssi.2012.02.006. [DOI] [Google Scholar]
  18. Chen H., Guo Z., Zhang L. A., Li Y., Li F., Zhang Y., Chen Y., Wang X., Yu B., Shi J.-M., Liu J., Yang C., Cheng S., Chen Y., Liu M.. Improving the Electrocatalytic Activity and Durability of the La0.6Sr0.4Co0.2Fe0.8O3−δ Cathode by Surface Modification. ACS Appl. Mater. Interfaces. 2018;10:39785–39793. doi: 10.1021/acsami.8b14693. [DOI] [PubMed] [Google Scholar]
  19. Bai J., Zhou D., Zhu X., Wang N., Chen R., Wang B., Yan W.. In-situ segregation of A-site defect (La0.6Sr0.4)0.90Co0.2Fe0.8O3‑δ to form a high-performance solid oxide fuel cell cathode material with heterostructure. Ceram. Int. 2023;49:5687–5699. doi: 10.1016/j.ceramint.2022.10.295. [DOI] [Google Scholar]
  20. Li J., Zhou X., Wu C., Zhao L., Dong B., Wang S., Chi B.. Self-stabilized hybrid cathode for solid oxide fuel cell: A-site deficient perovskite coating as solid solution for strontium diffusion. Chem. Eng. J. 2022;438:135446. doi: 10.1016/j.cej.2022.135446. [DOI] [Google Scholar]
  21. De Vero J. C., Develos-Bagarinao K., Kishimoto H., Ishiyama T., Yamaji K., Horita T., Yokokawa H.. Influence of La0.6Sr0.4Co0.2Fe0.8O3‑δ Microstructure on GDC Interlayer Stability and Cation Diffusion across the LSCF/GDC/YSZ Interfaces. J. Electrochem. Soc. 2016;163:F1463. doi: 10.1149/2.0021614jes. [DOI] [Google Scholar]
  22. Pan Z., Liu Q., Ni M., Lyu R., Li P., Chan S. H.. Activation and failure mechanism of La0.6Sr0.4Co0.2Fe0.8O3−δ air electrode in solid oxide electrolyzer cells under high-current electrolysis. Int. J. Hydrogen Energy. 2018;43:5437–5450. doi: 10.1016/j.ijhydene.2018.01.181. [DOI] [Google Scholar]
  23. Mahmoud A., Al Daroukh M., Lipinska-Chwalek M., Luysberg M., Tietz F., Hermann R. P.. A Mössbauer spectral study of degradation in La0.58Sr0.4Fe0.5Co0.5O3–x after long-term operation in solid oxide electrolysis cells. Solid State Ionics. 2017;312:38–43. doi: 10.1016/j.ssi.2017.10.003. [DOI] [Google Scholar]
  24. Frey C. E., Fang Q., Sebold D., Blum L., Menzler N. H.. A Detailed Post Mortem Analysis of Solid Oxide Electrolyzer Cells after Long-Term Stack Operation. J. Electrochem. Soc. 2018;165:F357. doi: 10.1149/2.0961805jes. [DOI] [Google Scholar]
  25. Laurencin J., Hubert M., Sanchez D. F., Pylypko S., Morales M., Morata A., Morel B., Montinaro D., Lefebvre-Joud F., Siebert E.. Degradation mechanism of La0.6Sr0.4Co0.2Fe0.8O3‑δ/Gd0.1Ce0.9O2‑δ composite electrode operated under solid oxide electrolysis and fuel cell conditions. Electrochim. Acta. 2017;241:459–476. doi: 10.1016/j.electacta.2017.05.011. [DOI] [Google Scholar]
  26. Schefold J., Brisse A., Poepke H.. 23,000 h steam electrolysis with an electrolyte supported solid oxide cell. Int. J. Hydrogen Energy. 2017;42:13415–13426. doi: 10.1016/j.ijhydene.2017.01.072. [DOI] [Google Scholar]
  27. Schefold J., Brisse A., Tietz F.. Nine Thousand Hours of Operation of a Solid Oxide Cell in Steam Electrolysis Mode. J. Electrochem. Soc. 2011;159:A137. doi: 10.1149/2.076202jes. [DOI] [Google Scholar]
  28. Wang C. C., Chen K., Jiang T., Yang Y., Song Y., Meng H., Jiang S. P., Lin B.. Sulphur poisoning of solid oxide electrolysis cell anodes. Electrochim. Acta. 2018;269:188–195. doi: 10.1016/j.electacta.2018.02.149. [DOI] [Google Scholar]
  29. The D., Grieshammer S., Schroeder M., Martin M., Al Daroukh M., Tietz F., Schefold J., Brisse A.. Microstructural comparison of solid oxide electrolyser cells operated for 6100 h and 9000 h. J. Power Sources. 2015;275:901–911. doi: 10.1016/j.jpowsour.2014.10.188. [DOI] [Google Scholar]
  30. Al Daroukh M., Tietz F., Sebold D., Buchkremer H. P.. Post-test analysis of electrode-supported solid oxide electrolyser cells. Ionics. 2015;21:1039–1043. doi: 10.1007/s11581-014-1273-2. [DOI] [Google Scholar]
  31. Tietz F., Sebold D., Brisse A., Schefold J.. Degradation phenomena in a solid oxide electrolysis cell after 9000 h of operation. J. Power Sources. 2013;223:129–135. doi: 10.1016/j.jpowsour.2012.09.061. [DOI] [Google Scholar]
  32. Reiner D., Marković S., Schröttner H., Boškoski P., Hochenauer C., Subotić V.. Long-term degradation and microstructural evolution of industrial-sized solid oxide electrolysis cells operated in co-electrolysis for power-to-X applications. Appl. Energy. 2026;417:128051. doi: 10.1016/j.apenergy.2026.128051. [DOI] [Google Scholar]
  33. Wen Y., Huang K.. Predicting the Rate of Degradation Related to Oxygen Electrode Delamination in Solid Oxide-Ion Electrolyzers. J. Electrochem. Soc. 2024;171:034510. doi: 10.1149/1945-7111/ad3414. [DOI] [Google Scholar]
  34. Chen W., Sun C.. Recent advances in high temperature solid oxide electrolytic cells. Energy Mater. 2025;5:500045. doi: 10.20517/energymater.2024.144. [DOI] [Google Scholar]
  35. Choi H., Yeon C., Park S.-Y., Lee J.-H., Yoon K. J., Lee C.-W., Chang H. J.. Nanoscale Degradation Mechanisms of Solid Oxide Electrolysis Cells with Electron-Conducting vs. Mixed Ionic-Electronic Conducting Air Electrodes. ECS Meet. Abstr. 2025;MA2025-03:295. doi: 10.1149/MA2025-031295mtgabs. [DOI] [Google Scholar]
  36. Lang M., Lee Y. S., Lee I. S., Szabo P., Hong J., Cho J., Costa R.. Analysis of electrochemical degradation phenomena of a 60-cell SOC stack operated in reversible SOFC/SOEC cycling mode. Appl. Energy. 2025;386:125565. doi: 10.1016/j.apenergy.2025.125565. [DOI] [Google Scholar]
  37. Yang W., Pan Z., Jiao Z., Zhong Z., O’hayre R.. Advanced microstructure characterization and microstructural evolution of porous cermet electrodes in solid oxide cells: A comprehensive review. Energy Rev. 2025;4:100104. doi: 10.1016/j.enrev.2024.100104. [DOI] [Google Scholar]
  38. Marina O. A., Pederson L. R., Coyle C. A., Thomsen E. C., Nachimuthu P., Edwards D. J.. Electrochemical, structural and surface characterization of nickel/zirconia solid oxide fuel cell anodes in coal gas containing antimony. J. Power Sources. 2011;196:4911–4922. doi: 10.1016/j.jpowsour.2011.02.027. [DOI] [Google Scholar]
  39. Hardy J. S., Coyle C. A., Bonnett J. F., Templeton J. W., Canfield N. L., Edwards D. J., Mahserejian S. M., Ge L., Ingram B. J., Stevenson J. W.. Evaluation of cation migration in lanthanum strontium cobalt ferrite solid oxide fuel cell cathodes via in-operando X-ray diffraction. J. Mater. Chem. A. 2018;6:1787–1801. doi: 10.1039/C7TA06856E. [DOI] [Google Scholar]
  40. Karki S. B., Le L., Seymour L., Springer R., Liu T., Marina O. A.. Stable Cu-Doped La2NiO4+δ Oxygen Electrodes for Solid Oxide Electrolysis Cells. ACS Appl. Mater. Interfaces. 2025;17:58270–58278. doi: 10.1021/acsami.5c15693. [DOI] [PubMed] [Google Scholar]
  41. Huang J., Papac M., O’hayre R.. Towards robust autonomous impedance spectroscopy analysis: A calibrated hierarchical Bayesian approach for electrochemical impedance spectroscopy (EIS) inversion. Electrochim. Acta. 2021;367:137493. doi: 10.1016/j.electacta.2020.137493. [DOI] [Google Scholar]
  42. Toby B. H., Von Dreele R. B.. GSAS-II: the genesis of a modern open-source all purpose crystallography software package. Appl. Crystallogr. 2013;46:544–549. doi: 10.1107/S0021889813003531. [DOI] [Google Scholar]
  43. Nechache A., Mansuy A., Petitjean M., Mougin J., Mauvy F., Boukamp B. A., Cassir M., Ringuedé A.. Diagnosis of a cathode-supported solid oxide electrolysis cell by electrochemical impedance spectroscopy. Electrochim. Acta. 2016;210:596–605. doi: 10.1016/j.electacta.2016.05.014. [DOI] [Google Scholar]
  44. Hardy J. S., Templeton J. W., Edwards D. J., Lu Z., Stevenson J. W.. Lattice expansion of LSCF-6428 cathodes measured by in situ XRD during SOFC operation. J. Power Sources. 2012;198:76–82. doi: 10.1016/j.jpowsour.2011.09.099. [DOI] [Google Scholar]
  45. Mineshige A., Izutsu J., Nakamura M., Nigaki K., Abe J., Kobune M., Fujii S., Yazawa T.. Introduction of A-site deficiency into La0.6Sr0.4Co0.2Fe0.8O3−δ and its effect on structure and conductivity. Solid State Ionics. 2005;176:1145–1149. doi: 10.1016/j.ssi.2004.11.021. [DOI] [Google Scholar]
  46. Świerczek K., Dabrowski B., Suescun L., Kolesnik S.. Crystal structure and magnetic properties of high-oxygen pressure annealed Sr1–xLaxCo0.5Fe0.5O3−δ (0<x<0.5) J. Solid State Chem. 2009;182:280–288. doi: 10.1016/j.jssc.2008.10.030. [DOI] [Google Scholar]
  47. Ebbesen S. D., Jensen S. H., Hauch A., Mogensen M. B.. High temperature electrolysis in alkaline cells, solid proton conducting cells, and solid oxide cells. Chem. Rev. 2014;114:10697. doi: 10.1021/cr5000865. [DOI] [PubMed] [Google Scholar]
  48. Ingram B. J., Eastman J. A., Chang K. C., Kim S. K., Fister T. T., Perret E., You H., Baldo P. M., Fuoss P. H.. In situ x-ray studies of oxygen surface exchange behavior in thin film La0.6Sr0.4Co0.2Fe0.8O3−δ . Appl. Phys. Lett. 2012;101:051603. doi: 10.1063/1.4739518. [DOI] [Google Scholar]
  49. Dupuy A. D., Chiu I. T., Shafer P., Arenholz E., Takamura Y., Schoenung J. M.. Hidden transformations in entropy-stabilized oxides. J. Eur. Ceram. Soc. 2021;41:6660–6669. doi: 10.1016/j.jeurceramsoc.2021.06.014. [DOI] [Google Scholar]
  50. Monaco F., Ferreira-Sanchez D., Hubert M., Morel B., Montinaro D., Grolimund D., Laurencin J.. Oxygen electrode degradation in solid oxide cells operating in electrolysis and fuel cell modes: LSCF destabilization and interdiffusion at the electrode/electrolyte interface. Int. J. Hydrogen Energy. 2021;46:31533–31549. doi: 10.1016/j.ijhydene.2021.07.054. [DOI] [Google Scholar]
  51. Fang Q., Frey C. E., Menzler N. H., Blum L.. Electrochemical Performance and Preliminary Post-Mortem Analysis of a Solid Oxide Cell Stack with 20,000 h of Operation. J. Electrochem. Soc. 2018;165:F38. doi: 10.1149/2.0541802jes. [DOI] [Google Scholar]
  52. Hoerlein M. P., Riegraf M., Costa R., Schiller G., Friedrich K. A.. A parameter study of solid oxide electrolysis cell degradation: Microstructural changes of the fuel electrode. Electrochim. Acta. 2018;276:162–175. doi: 10.1016/j.electacta.2018.04.170. [DOI] [Google Scholar]
  53. Wang F., Yamaji K., Cho D.-H., Shimonosono T., Kishimoto H., Brito M. E., Horita T., Yokokawa H.. Effect of strontium concentration on sulfur poisoning of LSCF cathodes. Solid State Ionics. 2012;225:157–160. doi: 10.1016/j.ssi.2012.05.026. [DOI] [Google Scholar]
  54. Chen K., Jiang S. P.. Failure mechanism of (La,Sr)­MnO3 oxygen electrodes of solid oxide electrolysis cells. Int. J. Hydrogen Energy. 2011;36:10541–10549. doi: 10.1016/j.ijhydene.2011.05.103. [DOI] [Google Scholar]
  55. Krammer M., Schmid A., Nenning A., Bumberger A. E., Siebenhofer M. U., Herzig C., Limbeck A., Rameshan C., Kubicek M., Fleig J.. Closed-Pore Formation in Oxygen Electrodes for Solid Oxide Electrolysis Cells Investigated by Impedance Spectroscopy. ACS Appl. Mater. Interfaces. 2023;15:8076–8092. doi: 10.1021/acsami.2c20731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Petric A., Ling H.. Electrical Conductivity and Thermal Expansion of Spinels at Elevated Temperatures. J. Am. Ceram. Soc. 2007;90:1515–1520. doi: 10.1111/j.1551-2916.2007.01522.x. [DOI] [Google Scholar]
  57. Jayapragasam P., Wen Y., Cook K., Wrubel J. A., Ma Z., Huang K., Jin X.. Crack Growth Rate at Oxygen Electrode/Electrolyte Interface in Solid Oxide Electrolysis Cells Predicted by Experiment Coupled Multiphysics Modeling. J. Electrochem. Soc. 2023;170:054509. doi: 10.1149/1945-7111/acd4f1. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao6c07898_si_001.pdf (340.2KB, pdf)

Data Availability Statement

Data supporting the findings of this study can be obtained from the corresponding author upon reasonable request.

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES