Cobalt-free compositionally complex nanocomposites with superior performance and stability for application as oxygen electrodes in solid oxide cells

Abstract

Cobalt-free oxide electrode materials have received increasing attention due to the interest in replacing critical raw materials in clean energy technologies. Among other compositions, La_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ (LSFCu)-based perovskites were proposed in the past for Solid Oxide Fuel Cell (SOFCs) applications due to their mixed ionic-electronic conductivity and low polarisation resistance. In this work, novel LSFCu-based compositions have been explored for their use as oxygen electrodes in SOFC and Solid Oxide Electrolysis Cells (SOECs). Single-phase LSFCu and fluorite–perovskite nanocomposites with (La,Ce)O_2−δ–(La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ composition were studied, the latter resulting from a one-pot co-synthesis of LSFCu and ceria with intermediate compositions reached by substantial cerium–lanthanum cation intermixing after self-organised phase separation. The best performances were achieved for compositionally complex nanocomposites based on 20% and 50% ceria with full-cell power densities of 1.15 and 1.26 W cm⁻² in SOFC mode (0.7 V), and current densities of 1.41 and 1.75 A cm⁻² in SOEC mode (1.3 V), at 800 °C. Durability tests performed in SOFC mode at 800 °C under high current densities (1 A cm⁻²) during nearly 900 h confirmed the benefits of optimal ceria incorporation, showing a remarkable reduction (7× lower) in the degradation rate compared to single-phase LSFCu. This work presents a straightforward approach to fabricating cobalt-free oxygen electrodes for SOC technology, opening a new avenue for developing highly complex composition (high entropy) self-organised nanocomposites.

---

1. Introduction

Solid Oxide Cells (SOCs) have emerged over recent decades as a promising technology to mitigate the intermittency of renewable energy sources. They play a critical role in the transition toward a sustainable, low-carbon energy matrix by facilitating decarbonisation of the transport and industrial sectors and enhancing grid stability through hydrogen production and utilisation. A key advantage of SOCs lies in their reversible operation: they can function either as Solid Oxide Fuel Cells (SOFC) or as Solid Oxide Electrolysis Cells (SOEC) within the same device, known as Reversible Solid Oxide Cells (RSOCs). This dual functionality offers high conversion efficiency, fuel flexibility, and potential for large-scale integration into future energy infrastructures.^1–3

Extensive research efforts and technological advancements by both industry and the scientific community have brought this technology to the initial stages of commercialisation.⁴ However, several challenges persist and must be addressed to enable their full development at different scales and industrial sectors. Among these, durability remains one of the most critical factors.⁵ SOCs typically operate at high temperatures (650–900 °C), which impose strict stability requirements on cell components over time; failure to meet these can compromise both performance and lifetime.^4–8

Regarding SOCs components, the design of fuel electrodes and electrolytes has reached a relatively mature stage.^2,9,10 Common electrolytes such as yttria-stabilised zirconia (YSZ) or gadolinia-doped ceria (GDC), as well as fuel electrodes like Ni-YSZ cermet, are well established in terms of materials selection and processing. Nevertheless, degradation of the Ni-YSZ fuel electrode remains a major contributor to overall cell degradation, particularly under SOEC operation.^11–13 In contrast, the oxygen electrode represents a critical bottleneck primarily in terms of materials sustainability, long-term stability, and compatibility with other cell components. State-of-the-art (SoA) oxygen electrode materials for intermediate-temperature SOCs are mainly based on perovskite-structured oxides.^14,15 Over the last decades, cobalt-containing perovskites such as La_1−xSr_xCo_1−yFe_yO_3−δ (LSCF) and La_1−xSr_xCoO_3−δ (LSC) have dominated the scenario due to their excellent electrocatalytic activity and mixed ionic-electronic conductivity (MIEC). However, despite their high initial performance, cobalt-based electrodes are prone to phase instability, chemical interaction, and mechanical compatibility with other cell components, which can contribute to performance degradation over long-term operation.¹⁶ While these effects are not always the dominant source of degradation in SoA SOCs, they remain relevant, particularly under demanding operating conditions. To mitigate these limitations, composite electrodes incorporating ionic conductors, such as GDC or other doped-CeO₂ materials, have been widely investigated as a general strategy to improve stability and compatibility.^5,7,17

Beyond performance and durability considerations, increasing attention has been directed towards the sustainability and supply risk associated with cobalt-based compounds. Cobalt is classified as a critical raw material due to its limited geographical availability, price volatility, and environmental and ethical concerns related to its extraction and refining. In addition, the high cobalt content of SoA oxygen electrodes poses challenges for large-scale manufacturing and long-term deployment of SOC technologies, particularly when considering cost reduction targets and recycling strategies. These factors have motivated growing research efforts towards cobalt-free oxygen electrodes that rely on more abundant elements, while preserving key functionalities.^18–20 To achieve optimal performance, various material engineering strategies have been explored, including exsolution, infiltration, composite formation, cation non-stoichiometry and increase of the compositional complexity towards high entropy.^17,21,22 Among them, the development of heterostructured composites incorporating pure CeO₂ has demonstrated effective mitigation of delamination, reduced strontium segregation, and significant improvements in both electrode performance and long-term stability.^23,24 In this regard, Baiutti et al.²⁵ recently demonstrated that Ce–La cation intermixing in La_0.8Sr_0.2MnO₃–Ce_0.8Sm_0.2O_2−δ (LSM–SDC) nanocomposites results in Sr–O over-bonding relative to the La–O polyhedral, thereby reducing the elastic driving force for Sr surface segregation. A similar effect has been reported for compositionally complex materials, also referred to as high-entropy oxides, which exhibit excellent stability as SOC electrodes, likely due to limited cation diffusion induced by local lattice distortions.^26,27

In this context, cobalt-free perovskite materials based on Cu-doped ferrites have emerged as attractive alternatives due to their mixed ionic-electronic conductivity, competitive electrocatalytic activity and lower TECs compared to their Co-based analogues.^18,28–37 Among the different electrode formulations reported to date, the La_1−xSr_xFe_1−yCu_yO_3−δ perovskite with x = 0.4 and y = 0.2 (LSFCu) has shown great potential as an oxygen electrode candidate for SOFC devices.^29,32,38,39 More recently, Draz et al.⁴⁰ further demonstrated the versatility of this material by confirming its effectiveness as both oxygen and fuel electrode in symmetric cells for CO₂ electroreduction.

In this work, we develop compositionally complex nanocomposite materials based on LSFCu and CeO₂ (LSFCu–Ce) as oxygen electrodes for SOCs. Using a one-pot combustion route, pure LSFCu and the nanocomposites containing 20 and 50 wt% ceria (denoted LSFCu–20Ce and LSFCu–50Ce, respectively) were co-synthesised. This strategy was designed to maximise heterogeneous interfaces and improve the catalytic activity while ensuring small particle size and uniform distribution.^21,23,41 The prepared compounds were characterised in terms of morphology, structure, and composition, and electrochemically evaluated in symmetric mode. Their performance in full-cell configuration was assessed in both SOFC and SOEC modes. Long-term durability was evaluated under SOFC conditions for approximately 900 h, whereas SOEC operation was limited to initial electrochemical characterisation. Despite this limitation, to the best of our knowledge, this is the first report of pure LSFCu being tested under SOEC conditions, and further characterisation will be reported elsewhere. Moreover, the prepared LSFCu–ceria nanocomposites constitute a new family of materials. The results demonstrate that the co-synthesised materials exhibit enhanced performance and stability, highlighting the beneficial effects of ceria when incorporated in optimal amounts.

2. Experimental

2.1. Materials preparation

La_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ (LSFCu), LSFCu–CeO₂ 20% wt% (LSFCu–20Ce) and LSFCu–CeO₂ 50% wt% (LSFCu–50Ce) compounds were synthesised via a combustion method (modified Pechini), previously described elsewhere by the authors.^42,43 For this, aqueous solutions containing stoichiometric amounts of La(NO₃)₃·6H₂O, Sr(NO₃)₂, Fe(NO₃)₃·9H₂O, Cu(NO₃)₂·3H₂O and Ce(NO₃)₂·6H₂O (all Sigma Aldrich, USA, >99.9%) were prepared according to the desired composition. Ethylenediaminetetraacetic acid (EDTA, Sigma Aldrich, USA, >99.4%) and ammonium nitrate (NH₄NO₃, Sigma Aldrich, USA, >99.0%) were added as complexing agent and combustion promoter, respectively, in a molar ratio of 1.1 : 3 : 1 to the metal concentration (EDTA : NH₄NO₃ : Mⁿ⁺). The final solution was heated on a hot plate at 120 °C with continuous stirring while maintaining pH = 10 until a gel formed. The temperature was then increased to 350 °C, causing the gel to self-ignite. The resulting ashes were ground and sintered at 850 °C in air for 8 h to obtain the desired materials as fine powders.

2.2. Fabrication of LSFCu-based cells

Button cells of 20 mm diameter were electrochemically characterised under symmetric and full-cell configurations. Symmetric cells were prepared using YSZ electrolyte supports (Kerafol, GE) of 250 µm thickness. Fuel-electrode-supported half-cells (SolydEra SpA, IT) consisting of a ∼250 µm thick Ni-YSZ fuel electrode and an 8–10 µm thick YSZ electrolyte were used to assemble the full-cells.

A ∼600 nm Ce_0.8Gd_0.2O_2−δ (GDC20) dense barrier layer was first deposited onto the electrolyte by pulsed laser deposition (PLD, PVD5000 system, PVD Products, USA) to prevent the well-documented reaction between YSZ and Sr-containing oxygen electrodes.^6,44–46 Details of the deposition conditions are available in previous publications from our group.^47,48 To enhance adhesion between the electrode and the dense PLD layer, a thin layer of Ce_0.9Gd_0.1O_2−δ (GDC10, Kceracell, SK) was air-brushed on top of the barrier layer and sintered in air at 1250 °C for 2 h. Slurries of the oxygen electrodes were prepared by dispersing 200 mg of the powders and 2 mg of polyvinyl pyrrolidone (PVP, Sigma Aldrich, USA) in 400 mg of α-terpineol (96%, Sigma Aldrich, USA) and stirring overnight. After calcination of the roughness layer, the electrode inks were brush-painted onto the surface and sintered in air at 950 °C for 4 hours. The active area of the deposited oxygen electrodes was 2.00 ± 0.12 cm², as determined from calibrated optical measurements.

2.3. Electrochemical characterisation

The electrochemical characterisation of the cells was carried out using a commercial ProboStat (NorECs AS, NW) test station with alumina sample housing and gas pipes, placed inside a high-temperature vertical tubular furnace.⁴⁹

For symmetric cells, gold mesh and gold paste were used to ensure a proper current collection. Electrochemical impedance spectroscopy (EIS) spectra were acquired with a Novocontrol (Novocontrol Technologies GmbH & Co., DE) frequency response analyser in the 10⁻¹ to 10⁶ Hz frequency range (50 mV amplitude). Measurements were conducted in synthetic air (100 NmL min⁻¹) between 600 °C and 850 °C on cooling with a 50 °C per step.

In the case of the full-cell characterisation, gold mesh/paste and nickel mesh/paste were used to collect the current on the oxygen and fuel electrodes, respectively. Ceramabond sealant (Aremco, USA) was employed to ensure gas tightness between the oxygen and fuel chambers. A Maynuo M9812 electronic load was used to control the current flow, and a Velleman LABPS3005D power source connected in series compensated for the voltage losses of the wiring system. Linear current sweep experiments with a 50 mA s⁻¹ rate were conducted to obtain polarisation curves in both fuel cell and electrolysis modes. Current density and power density values were calculated from the collected voltage and current data. EIS data was gathered with a PARSTAT 2273 potentiostat in galvanostatic mode with a 50 mA amplitude in the 10⁻¹ to 10⁵ Hz frequency range, at OCV, 700 and 850 mV (SOFC), and 1150 mV (SOEC).

The Zview software (Scribner Associates Inc., USA) was used to analyse the obtained data. The analysis of the EIS results was also supported by a distribution of relaxation times (DRT) tool, which aided the interpretation of the main resistive contributions.⁵⁰ The temperatures and gas flows used during the experiments are described in the text accordingly.

2.4. Compositional and microstructural characterisation

X-ray powder diffraction (XRD) was employed to determine the structure of the synthesised powders and the post-mortem characterisation of the surface of aged full cells. The measurements were carried out using a Bruker D8 Advance diffractometer in Bragg–Brentano geometry, with CuKα sealed-tube radiation (λ = 1.5418 Å) operating at 40 kV and 40 mA, and a Ni filter. A Lynxeye 1D detector was used for data collection at room temperature in the 2θ range between 20–120° (0.02° per step and 10 s per step). The data was fitted through the Rietveld method using the GSAS-II software suite.⁵¹

Scanning Transmission Electron Microscopy (STEM) was utilised to examine the microstructure of the prepared electrodes, while elemental analysis was carried out through Electron Energy Loss Spectroscopy (STEM-EELS) mapping. Pristine full cells containing the electrode were embedded in epoxy resin, and their surface was mechanically ground and polished to expose the cross-section. Lamellae containing the region of interest were prepared by FIB with a thickness of 70 nm, allowing for electron transparency. High-Resolution TEM (HR-TEM), High-Angle Annular Dark-Field (HAADF) STEM imaging, and STEM-EELS analysis of the lamellas were performed using a JEOL JEM-2010F and a JEOL JEM-ARM200, both operated at 200 kV.

Scanning electron microscopy (SEM) was used to study the surface morphology and microstructure of the electrodes in cross-sectional view. Digital micrographs were acquired in the 1.5–20 kV energy range with a ZEISS Auriga microscope equipped with SE2, Inlens, and ESB detectors. An EDS probe coupled to the microscope was used to obtain qualitative elemental compositional analyses on selected zones of interest.

3. Results and discussion

3.1. Structural characterisation of the materials

The structure of the LSFCu and LSFCu–Ce powders obtained through the self-combustion route and subsequent heat treatment in air at 850 °C was studied by XRD; the diffractograms collected at room temperature are shown in Fig. 1a. Acquired data were fitted using the Rietveld method to extract information about the samples' composition and relevant structural parameters about the phases present in each of them, as depicted in Fig. 1b for the LSFCu–20Ce (and Fig. S1 of the SI for LSFCu and LSFCu–50Ce). The analysis of the LSFCu pattern confirmed the presence of a single phase that matches with rhombohedral LSFCu perovskite, with space group R3̄c. The same was observed for the other mixed materials, which present only the phases corresponding to LSFCu and CeO₂ cubic fluorite, with space group Fm3̄m, as confirmed by the Rietveld fits. None exhibited impurities except for the LSFCu–50Ce sample, where a small quantity of CuO (monoclinic, C2/c) was detected (peaks marked with an asterisk in Fig. 1a). The amount of CuO determined by the Rietveld fit is ∼1.3% wt (Fig. S1b).

Fig. 1 (a) Room temperature XRD patterns of the powders and (b) Rietveld fit of the LSFCu–20Ce sample.

Table 1 provides an overview of the main structural and goodness-of-fit parameters obtained from the Rietveld refinements. These results allowed us to confirm the composition of the samples, obtaining weight percentages for the ceria of 19.3 and 54.7% for the samples with 20 and 50% nominal percentages, respectively. The observed differences fall within the expected error range for the synthesis process, given that initial calculations were based only on the final cation content.

Table 1 Rietveld fit parameters for the LSFCu and LSFCu–Ce materials

Parameter	LSFCu	LSFCu–20Ce		LSFCu–50Ce
Phase	La0.6Sr0.4Fe0.8Cu0.2O3	(La,Ce)0.6Sr0.4Fe0.8Cu0.2O3−δ	(La,Ce)O2−δ	(La,Ce)0.6Sr0.4Fe0.8Cu0.2O3−δ	(La,Ce)O2−δ	CuO
Space group	R3̄c	R3̄c	Fm3̄m	R3̄c	Fm3̄m	C2/c
a (Å)	5.50845(20)	5.4911(4)	5.45610(29)	5.5104(5)	5.44101(10)	4.673(4)
b (Å)	—	—	—	—	—	3.4299(9)
c (Å)	13.4288(16)	13.5315(10)	—	13.4371(11)	—	5.118(4)
V (Å^3)	352.881(16)	353.347(31)	162.423(26)	353.343(29)	161.079(9)	80.936(23)
Wt fraction (%)	100	80.7	19.3	44.0	54.7	1.3
R wp (%)	2.672	3.029		3.237
GOF	1.20	1.42		1.31

---

Focusing on the individual phases, the a-axis parameters of the LSFCu perovskite are quite similar, showing no difference between the original and mixed materials. On the other hand, a larger variation of the c-axis values is observed between the LSFCu–20Ce and the other two samples. However, this is probably due to experimental error in the refinement caused by peak broadening (small crystallite size) and consequent peak overlap, since the volume for all of them is almost the same.

Moving to the fluorite phase, the refined a-axis values (5.46 Å and 5.44 Å for the LSFCu–20Ce and LSFCu–50Ce, respectively) are noticeably larger than that of pure CeO₂ (5.41 Å).⁵² This deviation, well beyond the standard uncertainty, indicates the incorporation of a larger cation, such as La³⁺, into the CeO₂ lattice, suggesting that cation intermixing occurred during the synthesis process. According to the relationship reported by Artini⁵² between La content and the CeO₂ lattice parameter, the observed expansion corresponds to an estimated La doping level of approximately 15%. Such La–Ce interdiffusion has been reported in other perovskite–fluorite systems, including LSM–SDC nanocomposites.²⁵ In this scenario, La diffusion from the perovskite phase could be compensated by three concurrent mechanisms:

(i) Incorporation of Ce³⁺ ions into the perovskite lattice, as previously observed for LSFCu, LSC, and LaCoO₃ although only in small atomic fractions (1–7 at%).^53–55

(ii) The accommodation of A-site non-stoichiometry in the perovskite, which has been associated with enhanced structural stability through the suppression of Sr surface segregation.^56,57

(iii) Partial phase decomposition into simpler oxides, if the first two effects are insufficient to balance the La loss.

To examine the extent of cation intermixing in our system, EELS analysis was performed on a pristine LSFCu–50Ce sample. The results, shown in Fig. 2, reveal a clear migration of La³⁺ into the CeO₂ phase, where it appears homogenously distributed. Conversely, Ce³⁺ diffusion into LSFCu particles is also evident but occurs heterogeneously, localised in specific regions (highlighted with yellow rectangles in Fig. 2). This behaviour likely arises from the limited solubility of Ce into the perovskite lattice and its intrinsic capacity to tolerate A-site non-stoichiometry, both of which contribute to the enhanced structural stability of these nanocomposites. Overall, the here employed synthesis route proved useful for obtaining perovskite–fluorite nanocomposites with highly complex compositions (due to strong cation intermixing) following a one-step approach.

Fig. 2 HAADF-STEM image (left) and corresponding quantitative EELS maps of La, Ce, Fe, and O collected from a lamella of the LSFCu–50Ce electrode material, showing the elemental distribution across a cluster of particles.

The morphology and microstructure of the nanocomposite electrodes were studied by scanning electron microscopy. Fig. 3 shows cross-sectional images of the electrode materials attached to YSZ tapes. From bottom to top, the cell architecture consists of the dense YSZ electrolyte (dark grey), a thin GDC barrier layer deposited by PLD (light grey), a porous GDC roughness layer (light grey, 1–2 µm, varying slightly between samples), and the corresponding oxygen electrode (dark grey, porous). A detailed schematic diagram of these layers is presented in Fig. S2a–d of the SI. The electrodes exhibit uniform and continuous microstructure with open porosity and particle sizes ranging from 100 to 500 nm, as shown in Fig. 3a–c for the LSFCu–20Ce electrode. SEM was also employed to corroborate the homogeneous distribution of the two crystallographic phases throughout the samples. For this, the backscattered electrons signal (ESB) was studied on the acquired micrographs (2 kV, 2.8 mm WD) of pure LSFCu and the LSFCu–20Ce and LSFCu–50Ce nanocomposites, shown in Fig. 3d–f for the cell cross-sections and Fig. S2e–g in a close-up view of the oxygen electrodes. The ESB detector allows the collection of maximum compositional contrast information, and the brightness level in the images is related to the electrons' energy of each phase, enabling compositional contrast imaging with enhanced sensitivity to atomic number differences. As expected, the pure LSFCu sample presents only one type of particle, whereas the nanocomposite materials display two types with different mean grain size distributions, bigger particles corresponding to the (La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ perovskite phase (dark) and smaller particles corresponding to the (La,Ce)O_2−δ fluorite (bright). The two phases are closely associated, with grains often appearing in direct contact. This high degree of connectivity is likely originated from the cation diffusion occurring at the synthesis level, as indicated by the XRD and TEM results.

Fig. 3 SEM micrographs of the obtained powders. Microstructure of the LSFCu–20Ce electrode as deposited on a YSZ tape (a–c), and backscattered electrons micrographs of the LSFCu (d), LSFCu–20Ce (e) and LSFCu–50Ce (f) samples.

3.2. Electrochemical characterisation of La_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ and (La,Ce)O_2−δ–(La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ compositionally complex nanocomposite materials in symmetrical cells

EIS measurements of the electrodes implemented in symmetrical cells (LSFCu|GDC|YSZ|GDC|LSFCu) were performed under air in symmetric configuration to assess their electrocatalytic activity in the 600–850 °C temperature range. The Nyquist plots for the LSFCu, LSFCu–20Ce and LSFCu–50Ce samples are presented in Fig. 4 (between 750 and 850 °C) and Fig. S3 (for the whole temperature range). As expected for thermally activated phenomena, the total area specific resistance (ASR) of the cells diminished with increasing temperature. The differences observed in the ohmic resistance among the three cells are attributed to variations in the electrode–electrolyte interfacial contact and microstructural characteristics of the electrode layers (brush-painted), roughness layer thickness and homogeneity, as well as minor differences in current collection during testing. These effects are frequently reported in symmetric-cell measurements and mainly influence the absolute value of the ohmic contribution, without affecting the comparative trends in polarisation resistance associated with the different oxygen electrode materials.

Fig. 4 Nyquist plots of the EIS spectra for the LSFCu (a–c), LSFCu–20Ce (d–f) and LSFCu–50Ce (g–i) compounds collected in the 750–850 °C temperature range. The experimental data is represented with scatter, while the results of the fit done by the ECM method are shown as full lines. The equivalent circuit used to adjust each set of data is included as an inset.

To separate the ohmic (ASR_s) and polarisation (ASR_pol) contributions to the total ASR and validate the appropriate equivalent circuit for modelling, DRT analyses were performed on the EIS data collected at 750, 800, and 850 °C. The results of DRT as a function of operation temperature for the LSFCu, LSFCu–20Ce and LSFCu–50Ce compounds are reported in Fig. S4. In general, three main contributions, corresponding to distinct peaks in the DRT analysis, were identified for all the samples. For the pure LSFCu, their corresponding associated characteristic frequencies are ∼10⁰, 10¹ and 10² Hz, with the peak at ∼10¹ Hz showing a markedly reduced intensity at the higher temperatures (800 and 850 °C). In the mixed materials, all the peaks shift slightly to higher frequencies with respect to those of LSFCu and are in good agreement among them. The main difference is observed for the peak at ∼10¹ Hz, clearly visible for LSFCu–20Ce at all the measured temperatures, and only at 750 °C for LSFCu–50Ce, which shifts to higher frequencies when the temperature increases, overlapping with the peak(s) at f ∼ 10² Hz.

Guided by the DRT results, the components of the equivalent circuit model (ECM) were fitted to the EIS data. The circuits employed consisted of an inductance (L) in series with a serial resistance (R_s), followed by ZARC elements (resistance in parallel with a constant phase element, R_i‖CPE_i) representing the polarisation processes contributing to ASR_pol.

Three arcs were primarily identified in the ECM fitting. However, at some temperatures, the high-frequency (HF) and medium-frequency (MF) contributions overlapped, appearing as a single arc (high-medium frequency, HMF, f ∼ 10¹ to 10² Hz) in the Nyquist plots.⁵⁸ For this reason, the spectra were fitted using either two or three ZARC elements, i.e. L–R_s–(R₁‖CPE₁)–(R₂‖CPE₂) or L–R_s–(R₁‖CPE₁)–(R₂‖CPE₂)–(R₃‖CPE₃), depending on the degree of separation between the HF and MF processes.⁵⁸ The characteristic capacitances associated with both arcs lie in the 10⁻¹ to 10⁻³ F cm⁻² range and were ascribed to charge transfer processes.⁴⁰ In contrast, the low frequency (LF) contribution (f ∼ 1 Hz) showed minimal dependence on temperature, as illustrated in Fig. 4, where the size and shape of the contribution remain unchanged while temperature varies. This trend makes the feature more visible at high temperatures, as the rest of the thermally activated phenomena reduce their total contribution. As widely discussed in the literature, the LF arc was attributed to gas conversion processes.^18,59,60

Arrhenius plots obtained from the fitted impedance data are presented in Fig. 5a (LSFCu–20Ce) and S5 (LSFCu and LSFCu–50Ce). The total polarisation resistance (ASR_pol = R_HMF + R_LF) generally increases with decreasing temperature, following an Arrhenius-like temperature dependence. However, deviations are observed at higher temperatures (≥750 °C) where non-thermally activated gas diffusion phenomena dominate the overall resistance. To isolate the thermally activated contributions, the resistance of the low-frequency arc (R_LF) was subtracted from the total ASR_pol. The remaining high- and mid-frequency component (R_HMF) displays a consistent Arrhenius behaviour over the studied temperature range. This approach allowed a reliable extraction of activation energies (see Fig. 5a, S5 and Table S1), ensuring they reflect only the intrinsic electrode processes without interference from gas conversion-related phenomena. The resulting total polarisation resistances for all the tested materials are compiled in the bar plot of Fig. 5b. For the pure LSFCu material, the ASR_pol values obtained are 11, 23 and 50 mΩ cm² at 850 °C, 800 °C and 750 °C, respectively, while the activation energy obtained is E_a = 1.42 eV. Furthermore, increasing the ceria content seems to diminish the electrode's activity (increase in ASR_pol) but also to lower the activation energy of the nanocomposite materials when compared to the pure LSFCu. The polarisation resistance values reached were 29 mΩ cm² (850 °C), 60 mΩ cm² (800 °C), and 110 mΩ cm² (750 °C) for LSFCu–20Ce, and 25 mΩ cm² (850 °C), 48 mΩ cm² (800 °C), and 96 mΩ cm² (750 °C) for LSFCu–50Ce. The activation energies obtained in this case were 1.26 and 1.25 eV for LSFCu–20Ce and LSFCu–50Ce, respectively.

Fig. 5 (a) Arrhenius plots (as ASR_polvs. 1000/T) for LSFCu–20Ce showing the different contributions to the polarisation resistance of the electrode. The spheres represent the total resistance (R_HMF + R_LF arcs), while the crossed circles represent the data without the diffusion-associated contribution (R_HMF). The data corresponding to the diffusion process is presented as squares (R_LF). (b) Bar plot showing ASR_pol values without the LF contribution (R_LF) as a function of temperature in the 750–850 °C range for all the tested materials.

Overall, the tested materials showed activation energies (1.25–1.42 eV) in agreement with those expected for charge and non-charge transfer processes in MIEC^61,62 while their ASR falls within the range reported in the literature for similar oxygen electrodes.^21,61–63 Therefore, in terms of performance, these findings support the potential of LSFCu, LSFCu–20Ce and LSFCu–50Ce nanocomposites for SOC applications.^64,65

3.3. Full-cell test of La_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ and (La,Ce)O_2−δ–(La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2CuO_3−δ compositionally complex nanocomposite materials

The electrochemical characterisation of a complete fuel-electrode-supported cell with LSFCu as oxygen electrode is shown in Fig. 6. Voltage and power density vs. current density curves (V–I and P–I, respectively) measured with 50 NmL min⁻¹ cm⁻² of dry H₂ (fuel side) and 125 NmL min⁻¹ cm⁻² of synthetic air (oxygen side) are presented in Fig. 6a. The Open Circuit Voltage (OCV) values obtained of 1.12–1.14 V for the whole temperature range (700–800 °C) align with the expected ones, confirming the appropriate gas tightness of the system. In terms of cell performance, the power densities obtained (P, at 0.7 V) were 0.79 W cm⁻² (800 °C), 0.51 W cm⁻² (750 °C) and 0.40 W cm⁻² (700 °C).

Fig. 6 Electrochemical characterisation of the full cell with LSFCu oxygen electrode in the 700–800 °C temperature range. Voltage vs. current density (V–I, unfilled symbols) and power vs. current density (P–I, filled symbols) curves in SOFC mode (a), V–I curves in SOEC mode (b), Nyquist plots of EIS measurements performed under bias, and therefore under different current loads, at 700 mV in SOFC mode (c) and 1200 mV in SOEC mode (d).

After this, the cell was tested in SOEC mode. Fig. 6b shows the V–I curves collected in the 700–800 °C temperature range. In this case, a mixture of hydrogen and steam (90 : 10) was used on the fuel side with flows of 45 NmL min⁻¹ cm⁻² H₂O + 5 NmL min⁻¹ cm⁻² H₂ + 12.5 NmL min⁻¹ cm⁻² Ar (carrier); 62.5 NmL min⁻¹ cm⁻² of synthetic air were used on the oxygen side. OCV values of 0.83, 0.87 and 0.89 V were obtained at 800 °C, 750 °C and 700 °C, respectively, in good agreement with the theoretical ones. As expected, the results show that higher current densities were attained at higher temperatures, with values of 1.25 A cm⁻² (800 °C), 0.70 A cm⁻² (750 °C), and 0.55 A cm⁻² (700 °C), at 1.3 V (also summarised in Table S3 for both operating modes). To the best of our knowledge, this is the first time LSFCu was tested under steam electrolysis conditions and demonstrated excellent performance in both SOFC and SOEC modes, highlighting its potential as a promising oxygen electrode material.^24,63

Additionally, EIS measurements were conducted in the same experimental conditions as the V–I curves. The Nyquist plots of EIS carried out under voltage bias (and therefore, under different current loads) for all the temperatures in SOFC (0.7 V) and SOEC (1.2 V) modes are displayed in Fig. 6c and d. Bode plots in OCV and under bias conditions are shown in Fig. S6a. The spectra were fitted using the ECM with electric circuits composed of an inductance (L) connected in series with a serial resistance (R_s) and three (SOEC) or four (SOFC) R‖CPE elements (shown as insets of each figure) that represent the total ASR_pol. It should be noted that all the fittings were affected by a significant inductive contribution at high frequencies, which influenced the shape of the spectra in that region. The serial resistance values obtained are similar in both operation modes and vary accordingly with temperature. In SOFC mode, four processes with associated characteristic frequencies (f) (and capacities, C) of f ∼ 10³ to 10⁴ Hz (R₁, C ∼ 10⁻³ to 10⁻⁴ F cm⁻²), f ∼ 10² to 10³ Hz (R₂, C ∼ 10⁻² to 10⁻³ F cm⁻²), f ∼ 10¹ to 10² Hz (R₃, C ∼ 10⁻² F cm⁻²), and f ∼ 10⁰ Hz (R₄, C ∼ 10⁻¹ to 10⁰ F cm⁻²) were fit. As no gas changes were performed, the processes were ascribed based solely on the literature.^66–69 The high-frequency processes were assigned to oxygen ionic transport in the interfaces (f ∼ 10³ to 10⁴ Hz) and charge transfer reactions within the fuel electrode (f ∼ 10² to 10³ Hz). The mid-frequency (f ∼ 10¹ to 10² Hz) and low-frequency (f ∼ 10⁰ Hz) processes were assigned to oxygen surface exchange reaction within the oxygen electrode and gas diffusion within the fuel electrode, respectively. In SOEC mode, the higher frequency process assigned to oxygen ionic transport cannot be distinguished clearly, and only the three processes with f ∼ 10³ Hz, f ∼ 10¹ Hz and f ∼ 10⁰ Hz are observed. It is worth mentioning that EIS spectra collected in SOEC mode presented a high level of noise at low frequencies (f < 10 Hz), besides inductance problems, further affecting the accurate ECM fitting. Despite this, in both modes, the high-frequency contributions (R₁ and R₂ in SOFC, R₁ in SOEC) increase and become more limiting as the temperature decreases. An increase in the mid- and low-frequency arcs (R₃ and R₄ in SOFC, R₂ and R₃ in SOEC) is also observed, but in general, the main factor affecting the increase in ASR_pol (sum of the individual resistances) comes from the high-frequency processes, namely charge transfer processes. Normalised residual vs. frequency plots for representative fittings and detailed values of the ECM fitting results obtained (resistances, capacities and characteristic frequencies) are presented in Fig. S6d, e and Table S3 of the SI.

Once the pure LSFCu was characterised as oxygen electrode with promising results, the alternative compositions produced by the addition of different amounts of ceria in the synthesis process were also tested, aiming for an improvement in stability. The electrochemical characterisation of the LSFCu–20Ce and LSFCu–50Ce oxygen electrodes in complete fuel-electrode-supported devices is presented in Fig. 7. V–I and P–I curves, measured in the same conditions as for the cell with LSFCu, are shown in Fig. 7a for LSFCu–20Ce and Fig. 7b for LSFCu–50Ce. The OCV values follow the expected values, confirming the gas-tightness of the setup and the reliability of the measurements. The power densities achieved in SOFC mode at 0.7 V were 1.15 W cm⁻², 0.96 W cm⁻² and 0.72 W cm⁻² for LSFCu–20Ce, and 1.26 W cm⁻², 0.99 W cm⁻² and 0.72 W cm⁻² for LSFCu–50Ce, at 800, 750 and 700 °C, respectively (see Table S3). Overall, LSFCu–50Ce was the best-performing material, although both LSFCu–20Ce and LSFCu–50Ce have similar performances, especially at lower temperatures (700 °C). These results agree with those extracted from the EIS spectra, presented in Fig. 7c and d (as Nyquist plots) and Fig. S6b and c (Bode plots) of the SI. Normalised residual vs. frequency plots for representative fittings and detailed values of the ECM fitting results obtained (resistances, capacities and characteristic frequencies) in SOFC mode for the LSFCu–20Ce and LSFCu–50Ce are presented in Fig. S6f, g, Tables S4 and S5 of the SI. Note that, in this case, the DC bias used (0.85 V) is slightly higher than the one used for the LSFCu electrode (0.7 V) due to equipment limitations. Once again, it should be noted that the fittings were affected by an inductive contribution at high frequencies (SOFC and SOEC), and noise at low frequencies (SOEC only). The ohmic resistances agree with the expected ones and vary accordingly with temperature. Using the same ECM applied to the pure phase electrode, the fitting of the impedance spectra of the composite electrodes also revealed the presence of four distinct processes, consistent with those observed for pure LSFCu. The equivalent circuit used (shown as insets of the EIS spectra) is the same as for the LSFCu in SOFC mode and is composed of an inductance in series with an ohmic resistance and 4 ZARC elements.

Fig. 7 Electrochemical characterisation of the full cells with the LSFCu–20Ce and LSFCu–50Ce mixed materials as oxygen electrodes in the 700–800 °C range. SOFC mode: voltage (V–I, unfilled symbols) and power (P–I, filled symbols) density curves vs. current density for the LSFCu–20Ce (a) and LSFCu–50Ce (b) compounds. Corresponding Nyquist plots of EIS measurements performed under bias at 850 mV (c and d). SOEC mode: voltage vs. current density curves for LSFCu–20Ce (e) and LSFCu–50Ce (f). Corresponding Nyquist plots recorded under bias at 1150 mV (g and h).

Based on their associated characteristic frequencies and capacities, f ∼ 10³ to 10⁴ Hz (R₁, C ∼ 10⁻³ to 10⁻⁴ F cm⁻²), f ∼ 10² to 10³ Hz (R₂, C ∼ 10⁻² to 10⁻³ F cm⁻²), f ∼ 10¹ to 10² Hz (R₃, C ∼ 10⁻² F cm⁻²), and f ∼ 10⁰ Hz (R₄, C ∼ 10⁻¹ to 10⁰ F cm⁻²), they were assigned once again to oxygen ionic transport in the interface, charge transfer reactions within the fuel electrode, oxygen surface exchange reactions within the oxygen electrode and gas diffusion within the fuel electrode, respectively.^13,66–70 In general, polarisation resistance decreased with the addition of ceria, although the change observed between LSFCu–20Ce and LSFCu–50Ce is almost negligible. This agrees with the V–I curves and the results obtained in symmetric mode. In the case of LSFCu–20Ce, the growth in polarisation resistance seems to derive from an increase in the high-frequency contributions (R₁–R₂), clearly perceptible as temperature varies. For the LSFCu–50Ce, the high-frequency process (R₁) also seems to contribute to an increase in ASR_pol but appears to be more affected by the inductive contribution (see Fig. S6c, Bode plots at 850 mV). This is also reflected in the characteristic frequencies of the processes. However, the mid- and low-frequency arcs are the main contributors to the increase in polarisation resistance, suggesting a change in the governing process depending on ceria content. Further testing and analysis would be necessary to clarify these attributions.

Electrolysis measurements were carried out using the same conditions described above for the pure LSFCu cell. The V–I curves acquired in the 700–800 °C temperature range are shown in Fig. 7e and f for LSFCu–20Ce and LSFCu–50Ce, respectively. The OCV values obtained, between 0.83 and 0.89 V, are consistent with the theoretical ones. Moreover, the maximum current density values (at 1.3 V) achieved were 1.41 A cm⁻², 1.07 A cm⁻², and 0.81 A cm⁻², for LSFCu–20Ce; and 1.75 A cm⁻², 1.25 A cm⁻², and 0.88 A cm⁻², for LSFCu–50Ce; at 800, 750 and 700 °C, respectively. These values agree with those obtained in SOFC mode, exhibiting the same performance trend for the tested materials.

Nyquist plots of the EIS data collected at 1.15 V (with different current loads) for the LSFCu–20Ce and LSFCu–50Ce samples are presented in Fig. 7g and h, respectively, and Fig. S6b and c (Bode plots) of the SI. The serial resistances obtained are in line with those obtained in SOFC mode and the theoretical values. For the EIS fittings, the equivalent circuit used is the same as for the pure LSFCu in SOEC mode (an inductance in series with an ohmic resistance and 3 ZARC elements, shown in the inset). Normalised residual vs. frequency plots for representative fittings and detailed values of the ECM fitting results obtained (resistances, capacities and characteristic frequencies) for the LSFCu–20Ce and LSFCu–50Ce are presented in Fig. S6f, g, Tables S4 and S5. The same difficulties with the inductance at high frequencies and noise at low frequencies were experienced during the fitting. As for the pure LSFCu, three arcs with f ∼ 10³ Hz (C ∼ 10⁻⁴ F cm⁻²), f ∼ 10² to 10¹ Hz (C ∼ 10⁻¹ to 10⁻² F cm⁻²), and f ∼ 10⁰ Hz (C ∼ 10⁻¹ F cm⁻²) were discerned. In this mode, the polarisation resistance of LSFCu–50Ce is smaller than that of LSFCu–20Ce, which coincides with the difference in performance observed in the V–I curves. Regarding the processes contributing to the ASR_pol, they agree in general with those presented for pure LSFCu, with matching in capacity and frequency, and therefore suggest that the same processes are governing the electrochemistry of the system despite the addition of ceria to the material. For LSFCu–20Ce, the characteristic frequencies for all processes are slightly shifted to lower frequencies. However, in both composites, the arc dominating the change in ASR_pol with temperature is the one at high frequency (R₁), with a marked influence at all temperatures for LSFCu–20Ce.

A summary of the ASR contributions for LSFCu, LSFCu–20Ce, and LSFCu–50Ce cells in both SOFC and SOEC modes is presented in Fig. 8, which compares the individual components of the total cell resistance at 800 °C using a bar plot. When considering the total ASR (ASR_pol + ASR_s) in both SOFC and SOEC mode, the performance of both ceria-containing materials improved compared to pure LSFCu, indicating that the addition of ceria enhances the electrocatalytic activity of the electrode. Among the nanocomposite materials, LSFCu–50Ce exhibited the best performance in both modes, following the trend LSFCu–50Ce > LSFCu–20Ce > LSFCu. Compared to the pure LSFCu electrode, LSFCu–20Ce and LSFCu–50Ce demonstrated improvements of 47% and 60% in SOFC performance, and 13% and 40% in SOEC performance, respectively (values obtained at the mentioned voltages and 800 °C). Taking into account the different contributions, this seems to be influenced mostly by the difference in serial resistance between composite and non-composite samples, since the polarisation resistance (sum of R_i) is similar among them. In general, the differences in SOFC mode can be assigned to changes in R₃ and R₄ (R₁ + R₂ are similar in all cases), while in SOEC mode to a combination of all three.

Fig. 8 Bar plot summarising the area specific resistance contributions at 800 °C in SOFC and SOEC modes of full cells containing LSFCu, LSFCu–20Ce and LSFCu–50Ce as oxygen electrode.

Overall, the results obtained in both SOFC and SOEC modes indicate that increasing the compositional and nanostructural complexity of the single-phase LSFCu perovskite using ceria enhances its electrocatalytic activity in full-cell operation. This improvement may stem from enhanced mixed ionic-electronic conductivity in the composite material, likely due to improved grain connectivity, creating an extended TPB network where electrochemical reactions can take place.⁷¹ Note that tests with lower ceria content (10%) were also carried out, but no improvement over LSFCu was observed below the threshold of 20%. The V–I curves and EIS spectra recorded for this composition are shown in Fig. S5 of the SI.

3.4. Durability tests in SOFC mode and post-degradation characterisation of the cells

Durability tests in SOFC mode were conducted on all three fabricated samples with the different synthesised oxygen electrode materials to verify their stability in operating conditions at 800 °C. The obtained results are presented in Fig. 9. For the cell with the LSFCu oxygen electrode, a current density of 0.8 A cm⁻² was applied. During testing, gas flows of 50 NmL min⁻¹ cm⁻² of dry H₂ to the fuel electrode and 125 NmL min⁻¹ cm⁻² of synthetic air to the oxygen electrode were supplied. As displayed in the plot, the cell performance diminished rapidly with time, exhibiting a degradation rate of 202 mV kh⁻¹ when calculated considering the first 340 h. This accelerated degradation may have been induced by the relatively harsh testing conditions (current density value above typical SOFC operation, ∼0.5 A cm⁻²), which resulted in a degradation rate significantly higher than the values typically acceptable for SOFCs,⁶ and lead to the finalisation of the test upon reaching voltages around 0.6 V. This lack of stability for LSFCu had been previously observed by Cademartori et al.³³ in symmetric cell configuration at temperatures as low as 600 °C, attributing it to Cu and Fe interdiffusion into the electrolyte, structural evolution within the bulk of the La_0.7Sr_0.3Fe_0.85Cu_0.15O_3−δ electrode, and partial delamination.

Fig. 9 Durability tests at 800 °C in SOFC operating conditions of three different button cells using the LSFCu, LSFCu–20Ce and LSFCu–50Ce materials as oxygen electrode.

Ageing tests carried out for the cells containing the LSFCu–20Ce and LSFCu–50Ce nanocomposite materials are also displayed in Fig. 9. The plot shows the voltage evolution for about 900 h at 800 °C and 1 A cm⁻² (0.8 V). Despite the slightly different initial voltages, both cells tend to stabilise and match their voltages at ∼20 h, exhibiting nearly identical ageing behaviour for the remainder of the test. The calculated degradation rates were 3.9% kh⁻¹ (31.7 mV kh⁻¹) for the LSFCu–20Ce compound and 3.8% kh⁻¹ (30.7 mV kh⁻¹) for LSFCu–50Ce. Based on the linearity of the degradation rates along the whole duration of the tests, no conditioning period was considered. These results indicate the great stability of the here presented materials, especially when considering the high current densities applied (commercial devices typically operate around 0.5 A cm⁻²).^9,72 Both rates are substantially lower (7×) than the one obtained for pure LSFCu, even under higher current densities (1 A cm⁻² for the mixed materials vs. 0.8 A cm⁻² for LSFCu). These results indicate that increasing the compositional complexity of LSFCu and engineering nanocomposite oxygen electrodes not only enhances the electrocatalytic performance, as previously discussed, but also significantly improves long-term stability. This improvement can be attributed, firstly, to well-established advantages of composite electrodes, such as reduced thermal expansion coefficient (TEC) mismatch and increased mechanical robustness, and secondly, to the enhanced structural stability against Sr segregation observed in ceria/lanthanum-based perovskite heterostructures.^5,25 Building on the versatility of previously reported systems, we demonstrate here that ceria addition in LSFCu–Ce nanocomposites promotes La migration, which contributes to the stabilisation of highly degradable Cu-doped ferrite electrodes via the formation of A-site deficiency and cerium incorporation. This finding further highlights the potential of LSFCu–Ce nanocomposites as robust and durable cobalt-free oxygen electrodes for SOFC applications. Further testing in SOEC mode would be necessary to fully assess the stability of the material.

Post-testing characterisation of the tested cells was performed to identify potential sources of the different degradation rates observed in the durability experiments for the LSFCu vs. the LSFCu–20Ce and LSFCu–50Ce mixed materials. SEM micrographs of the cross-section of LSFCu and nanocomposite cells are presented in Fig. 10. An interesting feature observed in the SEM micrographs for the LSFCu electrode, which did not occur for the LSFCu–20Ce and LSFCu–50Ce materials, is partial delamination of the electrode (Fig. 10a and d). This was also evidenced by comparing the EIS spectra collected at the beginning and end of the ageing tests, as displayed in Fig. S6a for pure LSFCu (initial and 340 h) and LSFCu–20Ce (initial and 900 h, representative of the mixed materials). The impedance data reveal a marked increase in both serial and polarisation resistances of the LSFCu electrode after only 340 h, suggesting significant delamination. In contrast, the EIS spectrum of the LSFCu–20Ce electrode after ageing shows variation only in the ASR_s, while the polarisation contribution remains unchanged. This could indicate that the degradation observed in the nanocomposite electrodes is primarily associated with the electrolyte or electrolyte/interface, rather than the oxygen electrode itself. Moreover, Fig. 10a displays highly dense regions throughout the LSFCu electrode, being especially evident in the top part, where it is in contact with the current collection layer (Au paste + Au mesh). These denser zones could have been caused by cation migration (Sr, Cu) within the electrode or absorption of contaminants (S, Si), resulting in the formation of secondary phases on the surface. Such species, including oxides, carbonates, and hydroxides, among others, have been shown to negatively impact the performance of the cell.^6,27,73 These regions were also present in the LSFCu–20Ce and LSFCu–50Ce cells, but to a lesser extent, which could be explained by the lower overpotentials experienced by these two cells in comparison with the LSFCu one. Moreover, the addition of CeO₂ to form composites could also play a role by mitigating cation interdiffusion, as mentioned above.

Fig. 10 SEM micrographs of the cross sections of LSFCu (a and d), LSFCu–20Ce (b and e) and LSFCu–50Ce (c and f) full cells aged in SOFC mode. Magnified views of the electrolyte–oxygen electrode interface of the three cells are presented in (d–f), where a clear delamination of the LSFCu electrode can be seen (d).

To delve deeper into the phase inhomogeneities observed, preliminary non-quantitative examinations of the cross-section were conducted by EDS in line-scan mode. The analysis for LSFCu–20Ce is presented as a representative case in Fig. S6b of the SI section, where the elemental profiles acquired for La, Sr, Fe, Cu, Ni, Zr, Y and Ce across the line illustrated in the micrograph are shown. However, the tests did not reveal any unexpected findings; all elements were in the expected regions, and no secondary phase segregation resulting from metal intermixing was observed at this scale. The presence of small quantities of SrO and CuO was indeed detected in XRD patterns collected from the top of the aged cells (Fig. S6c–e), but their absence in the EDS scan might indicate that they are only superficial impurities. Additional impurities, such as alumina and silicates, were also identified and originate from the remaining ceramic sealant on the edge of the cell.

All in all, microstructural evolution and secondary phase formation at the surface were observed to some extent for all the cells and might explain the degradation seen in the durability tests. The effects are more pronounced for the cell with LSFCu, where partial delamination of the oxygen electrode layer occurred. This behaviour results from a combination of operating conditions, including current load, test duration, and electrochemical overpotential. Notably, the more severe degradation of LSFCu cannot be explained solely by either longer testing time or higher current density, as this cell was operated for only 340 hours at 0.8 A cm⁻², compared to 900 hours at 1 A cm⁻² for LSFCu–20Ce and LSFCu–50Ce cells. Instead, the higher overpotential experienced by the LSFCu electrode during operation likely contributed to the observed degradation, as elevated overpotentials are generally associated with increased electrochemical stress at the electrode–electrolyte interface. Overall, the tested cells present great stability behaviour in the degradation tests, considering the high current densities applied. While there is still room to improve the electrode's microstructure and its integration with other cell components, the results clearly demonstrate the high potential of (La,Ce)O_2−δ–(La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ materials as promising Co-free oxygen electrodes for SOC applications.

4. Conclusions

In this study, La_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ perovskites and ceria co-doped perovskite–fluorite nanocomposites with formula (La,Ce)O_2−δ–(La,Ce)_0.6Sr_0.4Fe_0.8Cu_0.2O_3−δ were investigated as oxygen electrodes for solid oxide cells. These materials were successfully synthesised using an innovative one-pot co-synthesis method that facilitated the desired cation intermixing. Comprehensive electrochemical characterisation of pure LSFCu was conducted in both SOFC and SOEC modes, reaching power and current density values of 0.9 W cm⁻² (0.6 V) and 1.25 A cm⁻² (1.3 V) at 800 °C, respectively. However, under SOFC operating conditions, the material exhibited poor stability. In contrast, ceria-containing nanocomposites exhibited both enhanced performance and improved stability. LSFCu–20Ce reached power density values of 1.15 W cm⁻² (SOFC at 0.7 V) and current density of 1.41 A cm⁻² (SOEC at 1.3 V) at 800 °C, while LSFCu–50Ce achieved peak values of 1.26 W cm⁻² and 1.75 A cm⁻² under the same conditions. Additionally, both compositions demonstrated degradation rates below 4% kh⁻¹ under high current densities for experiments of 900 h duration. These findings highlight the strong potential of LSFCu–20Ce and LSFCu–50Ce as cobalt-free oxygen electrodes and indicate that La–Ce intermixing and A-site reorganisation via (La,Ce)O_2−δ addition can enhance electrode stability compared to similar non-Ce-containing oxygen electrodes, suggesting this approach as a promising design strategy for the development of potentially stable and high-performing SOC electrode materials.

Author contributions

NDB: conceptualisation, methodology, investigation, formal analysis, writing – original draft. JG: investigation, writing – review & editing. AMA: methodology, formal analysis, writing – review & editing. KK: methodology, writing – review & editing. LB: methodology, writing – review & editing. LY: investigation, writing – review & editing. SE: investigation, writing – review & editing. FP: investigation, writing – review & editing. MT: conceptualisation, supervision, funding acquisition, writing – review & editing. LS: conceptualisation, supervision, funding acquisition, writing – review & editing. AT: conceptualisation, supervision, funding acquisition, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

Data supporting the findings of this study, including raw data from XRD, EIS, and current­–voltage experiments, are available on Zenodo at https://doi.org/10.5281/zenodo.18631320.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5ta08955g.

Acknowledgements

NDB gratefully acknowledges Comisión Académica de Posgrado (CAP, Uruguay) for the financial support received through a Master's Thesis Scholarship (BDMX_2020_1#47241316). AMA would also like to acknowledge Juan de la Cierva fellowship (JDC2023-052569-I, funded by MCIU/AEI/10.13039/501100011033 and by FSE+). The authors also acknowledge Dario Montinaro (SolydEra S.p.A., Italy) for the fuel-electrode-supported half-cells employed in this work. This research has been funded by the European Union GA No. 101101418 (24_7 ZEN). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or Clean Hydrogen Joint Undertaking. Neither the European Union nor the granting authority can be held responsible for them. LY, SE, and FP work is supported by the Spanish Project PID2022-138543NB-C21 funded by MCIU/AEI/10.13039/501.10001103 and projects 2021SGR00242 and 2023 CLIMA 00011 from Generalitat de Catalunya. LY acknowledges support from the Spanish Ministry for Science and Innovation through the Ramón y Cajal Fellowship RYC2022-037722-I. FP acknowledges the ICREA Academia 2022 grant.

References

1. J. T. S. Irvine, et al., Evolution of the electrochemical interface in high-temperature fuel cells and electrolysers, Nat. Energy, 2016, 1, 15014.
2. A. Hauch, et al., Recent advances in solid oxide cell technology for electrolysis, Science, 2020, 370, eaba6118.
3. P. Boldrin and N. P. Brandon, Progress and outlook for solid oxide fuel cells for transportation applications, Nat. Catal., 2019, 2, 571–577.
4. M. B. Mogensen, Materials for reversible solid oxide cells, Curr. Opin. Electrochem., 2020, 21, 265–273.
5. G. Sassone, et al., Effect of the operating temperature on the degradation of solid oxide electrolysis cells, J. Power Sources, 2024, 605, 234541.
6. J. Laurencin, et al., 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.
7. S. He, Y. Zou, K. Chen and S. P. Jiang, A critical review of key materials and issues in solid oxide cells, Interdiscip. Mater., 2022, 2, 111–136.
8. S. P. Jiang, Solid-State Electrochemistry and Solid Oxide Fuel Cells: Status and Future Prospects, Electrochem. Energy Rev., 2022, 5, 1–47.
9. L. Bernadet, et al., Enhanced diffusion barrier layers for avoiding degradation in SOFCs aged for 14000 h during 2 years, J. Power Sources, 2023, 555, 232400.
10. L. Blum, et al., Long-term operation of solid oxide fuel cells and preliminary findings on accelerated testing, Int. J. Hydrogen Energy, 2020, 45, 8955–8964.
11. M. Trini, et al., Comparison of microstructural evolution of fuel electrodes in solid oxide fuel cells and electrolysis cells, J. Power Sources, 2020, 450, 227599.
12. Q. Fang, C. E. Frey, N. H. Menzler and L. Blum, 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–F45.
13. V. Bilalis, X. Sun, H. L. Frandsen and M. Chen, Quantifying Galvanostatic Degradation of Solid Oxide Electrolysis Cells: The onset of accelerated degradation of Ni-yttria stabilized zirconia electrode, J. Power Sources, 2024, 606, 234490.
14. S. E. Wolf, et al., Solid oxide electrolysis cells - current material development and industrial application, J. Mater. Chem. A, 2023, 11, 17977–18028.
15. P. Kaur and K. Singh, Review of perovskite-structure related cathode materials for solid oxide fuel cells, Ceram. Int., 2020, 46, 5521–5535.
16. S. P. Jiang, Development of lanthanum strontium cobalt ferrite perovskite electrodes of solid oxide fuel cells – a review, Int. J. Hydrogen Energy, 2019, 44, 7448–7493.
17. A. K. Yadav, S. Sinha and A. Kumar, Advancements in composite cathodes for intermediate-temperature solid oxide fuel cells: a comprehensive review, Int. J. Hydrogen Energy, 2024, 59, 1080–1093.
18. A. Niemczyk, et al., Effective oxygen reduction on A-site substituted LaCuO3-: δ: toward air electrodes for SOFCs based on perovskite-type copper oxides, J. Mater. Chem. A, 2019, 7, 27403–27416.
19. C. Zhang, X. Zhao, R. Sacchi and F. You, Trade-off between critical metal requirement and transportation decarbonization in automotive electrification, Nat. Commun., 2023, 14, 1616.
20. A. Niemczyk, et al., Efficient and Economically Favorable Co-Free Air Electrodes for Solid Oxide Cells, ECS Meet. Abstr., 2021, 1497–1504.
21. S. Chen, et al., Review of SOFC Cathode Performance Enhancement by Surface Modifications: Recent Advances and Future Directions, Energy Fuels, 2023, 37, 3470–3487.
22. S. Biswas, et al., A Review on Critical Metals Used in Solid Oxide Cells for Power ↔ X Applications and Materials Recyclability, ACS Sustain. Chem. Eng., 2023, 12, 6037–6058.
23. J. Cui, et al., Suppressing Structure Delamination for Enhanced Electrochemical Performance of Solid Oxide Cells, Small Methods, 2024, 8, 2400178.
24. W. Zhang, et al., La0.6Sr0.4Co0.2Fe0.8O3−δ/CeO2 Heterostructured Composite Nanofibers as a Highly Active and Robust Cathode Catalyst for Solid Oxide Fuel Cells, ACS Appl. Mater. Interfaces, 2019, 11, 26830–26841.
25. F. Baiutti, et al., A high-entropy manganite in an ordered nanocomposite for long-term application in solid oxide cells, Nat. Commun., 2021, 12, 1–11.
26. M. Xiao, et al., High-entropy materials for solid oxide cells: synthesis, applications, and prospects, J. Energy Chem., 2025, 104, 268–296.
27. J. Sirvent, et al., Blocking Sr-Segregation in Perovskite Cathodes for Solid Oxide Cells by Mn Codoping, ACS Appl. Energy Mater., 2025, 8, 7022–7037.
28. Q. Sun, et al., Insights into the oxygen reduction reaction on Cu-doped SrFeO3-δ cathode for solid oxide fuel cells, J. Power Sources, 2021, 497, 229877.
29. F. Zurlo, A. Iannaci, V. M. Sglavo and E. Di Bartolomeo, Copper-based electrodes for IT-SOFC, J. Eur. Ceram. Soc., 2019, 39, 17–20.
30. F. Zurlo, et al., Copper-doped lanthanum ferrites for symmetric SOFCs, Acta Mater., 2016, 112, 77–83.
31. K. Li, et al., Co-free triple perovskite La1.5Ba1.5Cu3O7±δ as a promising air electrode material for solid oxide fuel cells, J. Power Sources, 2022, 532, 1–11.
32. Q. Zhou, et al., La 0.6Sr 0.4Fe 0.8Cu 0.2O 3-δ perovskite oxide as cathode for IT-SOFC, Int. J. Hydrogen Energy, 2012, 37, 11963–11968.
33. D. Cademartori, et al., A High-Performance Co-Free Electrode for Solid Oxide Cells: La0.7Sr0.3Cu0.15Fe0.85O3-δ Synthesis and Characterisation, J. Alloys Compd., 2023, 965, 171334.
34. J. Grassi, N. Di Benedetto, M. A. Macías, J. Basbus and A. Serquis, Power density vs ionic radii correlations in REBa2Cu3O6+δ (RE = La , Nd , Sm , Gd and Y) IT – SOFC potential cathodes, Electrochim. Acta, 2023, 464, 142931.
35. Y. Tian, et al., Cobalt-Free Perovskite Oxide La0.6Sr0.4Fe0.8Ni0.2O3−δ as Active and Robust Oxygen Electrode for Reversible Solid Oxide Cells, ACS Appl. Energy Mater., 2019, 2, 3297–3305.
36. J. H. Zhang, F. Z. Han, C. X. Li and S. L. Zhang, A-site deficient Sr0.9Ti0.3Fe0.7O3−δ perovskite: a high stable cobalt-free oxygen electrode material for solid oxide electrochemical cells with excellent electrocatalytic activity and CO2 tolerance, J. Eur. Ceram. Soc., 2022, 42, 5801–5812.
37. S. S. Hashim, F. Liang, W. Zhou and J. Sunarso, Cobalt-free perovskite cathode for solid oxide fuel cells: a review, ChemElectroChem, 2019, 6, 3549–3569.
38. S. Vázquez, et al., Synthesis and characterization of La0.6Sr0.4Fe0.8Cu0.2O3-δ oxide as cathode for intermediate temperature solid xide fuel cells, J. Solid State Chem., 2015, 228, 208–213.
39. M. Hussain, et al., Cobalt free LaxSr1-xFe1-yCuyO3-δ (x= 0.54, 0.8, y = 0.2, 0.4) perovskite structured cathode for SOFC, Ceram. Int., 2020, 46, 18208–18215.
40. U. Draz, et al., Copper-Enhanced CO2 Electroreduction in SOECs, ACS Appl. Mater. Interfaces, 2024, 16, 8842–8852.
41. Z. Yue, et al., Facile co-synthesis and utilization of ultrafine and highly active PrBa0.8Ca0.2Co2O5+δ-Gd0.2Ce0.8O1.9 composite cathodes for solid oxide fuel cells, Electrochim. Acta, 2022, 403, 139673.
42. N. Di Benedetto, et al., Influence of the Thermal Processing and Doping on LaMnO3 and La0.8A0.2MnO3 (A = Ca, Sr, Ba) Perovskites Prepared by Auto-Combustion for Removal of VOCs, Catalysts, 2022, 12, 865.
43. J. Grassi, et al., Synthesis and Characterization of High Temperature Properties of YBa2Cu3O6+δ Superconductor as Potential Cathode for Intermediate Temperature Solid Oxide Fuel Cells, J. Mater. Sci. Technol. Res., 2021, 8, 81–92.
44. F. Monaco, et al., 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.
45. Z. Pan, et al., 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.
46. K. Develos-Bagarinao, et al., Evolution of cathode-interlayer interfaces and its effect on long-term degradation, J. Power Sources, 2020, 453, 227894.
47. L. Bernadet, et al., Thickness effect of thin-film barrier layers for enhanced long-term operation of solid oxide fuel cells, APL Energy, 2023, 1, 036101.
48. M. Morales, et al., Enhanced Performance of Gadolinia-doped Ceria Diffusion Barrier Layers Fabricated by Pulsed Laser Deposition for Large-Area Solid Oxide Fuel Cells, ACS Appl. Energy Mater., 2018, 1, 1955–1964.
49. S. Anelli, Advanced Strategies for Solid Oxide Electrolysis Cells, Unversidad Autónoma de Barcelona, Bercelona, 2020.
50. D. Clematis, et al., On the stabilization and extension of the distribution of relaxation times analysis, Electrochim. Acta, 2021, 391, 138916.
51. B. H. Toby and R. B. Von Dreele, GSAS-II : the genesis of a modern open-source all purpose crystallography software package, Appl. Crystallogr., 2013, 46, 544–549.
52. C. Artini, Rare-Earth-Doped Ceria Systems and Their Performance as Solid Electrolytes: A Puzzling Tangle of Structural Issues at the Average and Local Scale, Inorg. Chem., 2018, 57, 13047–13062.
53. F. Giannici, et al., Cation Diffusion and Segregation at the Interface between Samarium-Doped Ceria and LSCF or LSFCu Cathodes Investigated with X - ray Microspectroscopy, ACS Appl. Mater. Interfaces, 2017, 9, 44466–44477.
54. E. Konysheva, R. Blackley and J. T. S. Irvine, Conductivity behavior of composites in the La0.6Sr 0.4CoO3±δ-CeO2 system: function of connectivity and interfacial interactions, Chem. Mater., 2010, 22, 4700–4711.
55. A. A. Ansari, et al., Catalytic performance of the Ce-doped LaCoO3 perovskite nanoparticles, Sci. Rep., 2020, 10, 15012.
56. W. Lee and B. Yildiz, Factors that Influence Cation Segregation at the Surfaces of Perovskite Oxides, ECS Trans., 2013, 57, 2115–2123.
57. D. Kim, A. Hunt, I. Waluyo and B. Yildiz, Cation deficiency enables reversal of dopant segregation at perovskite oxide surfaces under anodic potential, J. Mater. Chem. A, 2023, 11, 7299–7313.
58. J. Nielsen, T. Jacobsen and M. Wandel, Impedance of porous IT-SOFC LSCF:CGO composite cathodes, Electrochim. Acta, 2011, 56, 7963–7974.
59. A. Flura, et al., Identification and modelling of the oxygen gas diffusion impedance in SOFC porous electrodes: application to Pr2NiO4+δ, Electrochim. Acta, 2015, 174, 1030–1040.
60. D. Cademartori, et al., Elucidating the impact of gas transport on high-performing air electrodes: a 1D physically based model unveiling the correlation between microstructure and impedance response, J. Phys.: Energy, 2025, 7, 015005.
61. A. Maria Asensio, F. R. Bianchi, D. Clematis, B. Bosio and A. Barbucci, A kinetic study on oxygen redox reaction of a double-perovskite reversible oxygen electrode—Part I: experimental analysis, J. Phys.: Energy, 2024, 6, 015011.
62. Z. Gao, L. V. Mogni, E. C. Miller, J. G. Railsback and S. A. Barnett, A perspective on low-temperature solid oxide fuel cells, Energy Environ. Sci., 2016, 9, 1602–1644.
63. R. Vinoth Kumar and A. P. Khandale, A review on recent progress and selection of cobalt-based cathode materials for low temperature-solid oxide fuel cells, Renew. Sustain. Energy Rev., 2022, 156, 111985.
64. S. Pang, et al., Systematic evaluation of cobalt-free Ln0.5Sr0·5Fe0·8Cu0·2O3−δ (Ln = La, Pr, and Nd) as cathode materials for intermediate-temperature solid oxide fuel cells, J. Power Sources, 2016, 326, 176–181.
65. G. Xue, et al., Synthesis and characterization of fibrous La0.8Sr0.2Fe1-xCuxO3-δ cathode for intermediate-temperature solid oxide fuel cells, Ceram. Int., 2022, 48, 25940–25948.
66. H. S. Ur Rehman, et al., Gadolinium-Doped Ceria Room-Temperature Sputtered Thin Barrier Layers in Large-Area Solid Oxide Fuel Cells: Influence of Their Thickness and Thickness Gradient on the Cathodic Processes, ACS Appl. Energy Mater., 2025, 8, 4281–4287.
67. P. Caliandro, A. Nakajo, S. Diethelm and J. Van herle, Model-assisted identification of solid oxide cell elementary processes by electrochemical impedance spectroscopy measurements, J. Power Sources, 2019, 436, 226838.
68. A. Leonide, Y. Apel and E. Ivers-Tiffee, SOFC Modeling and Parameter Identification by Means of Impedance Spectroscopy, ECS Trans., 2009, 19, 81–109.
69. A. Hauch, K. Brodersen, M. Chen and M. B. Mogensen, Ni/YSZ electrodes structures optimized for increased electrolysis performance and durability, Solid State Ionics, 2016, 293, 27–36.
70. F. Kukk, S. Pylypko, M. S. Skjøth-Rasmussen, E. Lust and G. Nurk, Influence of Chemical Modifications with Fe, Co, and GDC on the Electrochemical Performance of the Ni-8YSZ Hydrogen Electrode in Solid Oxide Electrolysis Cells, J. Electrochem. Soc., 2025, 172, 110515.
71. E. P. Murray, M. J. Sever and S. A. Barnett, Electrochemical performance of (La,Sr)(Co,Fe)O3-(Ce,Gd)O3 composite cathodes, Solid State Ionics, 2002, 148, 27–34.
72. N. H. Menzler, D. Sebold, Y. J. Sohn and S. Zischke, Post-test characterization of a solid oxide fuel cell after more than 10 years of stack testing, J. Power Sources, 2020, 478, 228770.
73. K. Chen and S. P. Jiang, Surface Segregation in Solid Oxide Cell Oxygen Electrodes: Phenomena, Mitigation Strategies and Electrochemical Properties, Electrochem. Energy Rev., 2020, 3, 730–765.

---