Nanocrystalline SrTi 1 -x Fe x O 3- δ perovskites as high-performance electrodes for intermediate-temperature symmetric SOFCs

Nanocrystalline SrTi 1 − x Fe x O 3 − δ perovskites (x = 0.20–0.50) represent a compositionally tunable platform for cobalt-free symmetric electrodes in intermediate-temperature solid oxide fuel cells. Synthesized via a gel-combustion route, all as-prepared compositions form single-phase cubic perovskites with nanometric domains, while the Fe/Ti ratio governs lattice contraction, mixed Fe + 3 / F e + 4 surface chemistry, and redox responsiveness. Electrochemical impedance spectroscopy reveals that Fe-rich materials (x = 0.35 and x = 0.50) exhibit markedly lower polarization resistances than the x = 0.20 composition under both oxidizing and reducing atmospheres. Distribution of relaxation times analysis reveals that Fe incorporation markedly attenuates the polarization contributions across the intermediate- and high-frequency ranges, confirming accelerated charge-transfer and surface-reaction kinetics under both oxidizing and reducing conditions. Under reducing conditions, STF50 undergoes adaptive structural evolution, including Fe exsolution and the emergence of Ruddlesden–Popper-type domains, highlighting a dynamic interplay between high electrochemical activity and redox-induced restructuring. Preliminary button-cell tests confirm the functional viability of STF50 as a symmetric electrode within the intermediate-temperature range. Overall, nanocrystalline SrTi 1 − x Fe x O 3 − δ emerges as a promising and structurally adaptive electrode family, in which controlled Fe substitution modulates defect chemistry, redox behavior, and electrochemical performance. These findings demonstrate the potential of Fe-rich titanate perovskites as next-generation cobalt-free electrodes for efficient and adaptable IT-SOFC architectures.

Solid oxide fuel cells (SOFCs) are among the most promising electrochemical energy-conversion technologies due to their attractive characteristics, including high efficiency, fuel flexibility, and low environmental impact (Singhal, 2000; Minh, 2004). Nevertheless, their practical application remains limited by material degradation and the relatively high operating temperatures typically required (Singhal, 2000; Sunarso et al., 2008). Consequently, considerable efforts have been devoted to the development of intermediate-temperature SOFCs (IT-SOFCs), operating below 800 ° C, since lower operating temperatures can reduce degradation phenomena, improve durability, and decrease manufacturing costs (Brett et al., 2008; Sunarso et al., 2008). However, reducing the operating temperature generally results in slower electrode kinetics and increased polarization losses, making the development of efficient electrode materials a critical challenge (Steele and Heinzel, 2001; Shao and Haile, 2004).

Among the different strategies proposed to improve electrode performance and durability, symmetric electrodes have attracted increasing attention. In these systems, the same material is employed as both anode and cathode, simplifying cell fabrication while minimizing compatibility issues associated with differences in thermal expansion behavior, chemical interactions, and degradation mechanisms between the two electrodes (Su et al., 2015; Suenaga et al., 2025). However, the identification of materials capable of maintaining high electrochemical activity and structural stability under both oxidizing and reducing atmospheres remains a major challenge (Tian et al., 2022; Liu et al., 2025).

Perovskite oxides with general formula ABO 3 have attracted considerable attention as symmetric electrode materials because of their structural flexibility and the possibility of tailoring their transport and catalytic properties through cation substitution (Goodenough, 2003; Sunarso et al., 2008). Among these oxides, SrTiO 3 -based materials are particularly attractive owing to their excellent chemical stability under reducing conditions, and compositional versatility (Phoon et al., 2019; Kujawska et al., 2024; Winterhalder et al., 2025; Reinke et al., 2026). Furthermore, the remarkable redox stability of these perovskites enables their use across oxidizing and reducing atmospheres, preserving the perovskite structure, a fundamental requirement for symmetric electrode applications. This behavior contrasts with that of several conventional Co-containing oxygen electrodes, in which the progressive reduction of cobalt species under reducing conditions often leads to phase instability (Radovic et al., 2008; Santaya et al., 2025).

Fe has attracted considerable interest as a dopant for SrTiO 3 because partial substitution of Ti by Fe yields SrTi 1 − x Fe x O 3 − δ (STF), a mixed ionic-electronic conductor whose transport properties strongly depend on the Fe concentration (Rothschild et al., 2006; Jung and Tuller, 2009; Kim et al., 2025). Studies have shown that the incorporation of Fe promotes the formation of oxygen vacancies and improves electronic conductivity through the presence of mixed-valence Fe species, thereby enhancing oxygen transport and surface-reaction kinetics (Jung and Tuller, 2009; Li et al., 2022; Wei et al., 2026). As a result, STF materials have attracted considerable interest for a wide range of electrochemical applications, including oxygen electrodes for solid oxide cells, oxygen-permeation membranes, gas sensors, and model perovskites for mixed ionic-electronic conduction (Mroziński et al., 2020; Li et al., 2022; Geisendorfer et al., 2025; Lacharme and Donazzi, 2025).

Most studies have focused on their performance as SOFC cathodes. For instance, Jung and Tuller investigated dense SrTi 1 − x Fe x O 3 − δ thin-film electrodes (x = 0.05 –0.80) deposited on yttria-doped zirconia (YSZ) electrolytes and demonstrated typical mixed ionic-electronic conducting behavior governed by surface oxygen-exchange kinetics. The electrodes exhibited low area-specific resistance values between 570 and 650 ° C and surface-exchange coefficients on the order of 10 − 5 cm s − 1, comparable to those reported for reference cathodes such as (La,Sr) F e O 3 − δ (LSCF) perovskites (Jung and Tuller, 2009). More recently, Qiu et al. enhanced the cathodic performance of STF-based materials by introducing A-site Sr deficiency in Sr 0.95 Ta 0.1 Fe 0.9 O 3 − δ. The optimized composition exhibited low area-specific resistance values of 0.095 and 0.42 Ω cm 2 at 600 ° C on Sm 0.2 Ce 0.8 O 1.9 (SDC) and BaZr 0.1 Ce 0.7 Y 0.1 Yb 0.1 O 3 − δ (BZCYYb) electrolytes, respectively, together with stable operation for 200 h. The improved electrochemical performance was attributed to the increased concentration of oxygen vacancies and enhanced oxygen-reduction kinetics (Qiu et al., 2024).

While significant progress has been achieved in improving the performance of STF-based oxygen electrodes, their application as symmetric electrodes remains relatively unexplored. Among the limited reports available, Mroziński et al. investigated porous STF35 symmetric electrodes deposited on CGO (Ce 0.9 Gd 0.1 O 2 − δ) electrolyte substrates. The authors reported polarization resistances close to 100 m Ω c m 2 at 800 ° C in air and identified oxygen adsorption and surface exchange as the dominant rate-limiting processes through impedance spectroscopy and DRT analysis (Mroziński et al., 2020). Although these findings demonstrated the potential of STF35 for symmetric SOFC applications, the influence of Fe content on the electrochemical behavior of STF-based materials remains largely unexplored. Furthermore, little attention has been devoted to STF nanomaterials synthesized through gel-combustion routes, despite their potential advantages in terms of surface area, gas accessibility, and shortened diffusion pathways.

Nanostructuring has been recognized as an effective approach to improving electrode performance by increasing the density of active reaction sites and facilitating mass- and charge-transport processes (Varma et al., 2016; Li et al., 2015). Various wet-chemical synthesis routes, including sol-gel (Navas et al., 2021), co-precipitation (Haron et al., 2017; Tien et al., 2026), hydrothermal methods (Walton, 2020), and gel-combustion synthesis (Carlos et al., 2020), have been explored to produce nanostructured perovskite oxides with enhanced microstructural control and electrochemical performance. Among these approaches, gel-combustion synthesis offers a simple and scalable route for producing highly porous nanostructured oxide powders while limiting grain growth through relatively low calcination temperatures (Varma et al., 2016).

Therefore, in the present work, nanostructured SrTi 1 − x Fe x O 3 − δ powders with x = 0.20, 0.35, and 0.50 were synthesized via a gel-combustion route using citric acid and glycine as fuel and complexing agents, followed by calcination at 800 ° C. Their structural and microstructural properties were investigated by X-ray diffraction, Rietveld refinement, and energy-dispersive spectroscopy. The electrochemical performance of the materials as symmetric electrodes was evaluated under oxidizing and reducing atmospheres up to 750 °C using electrochemical impedance spectroscopy and polarization measurements. Furthermore, distribution of relaxation times (DRT) analysis was used to resolve the main contributions to electrode polarization and to assess the influence of Fe content on electrode kinetics.

Synthesis of STF powders and cell preparation

Nanostructured SrTi 1 − x Fe x O 3 − δ powders with x = 0.20, 0.35, and 0.50 were synthesized via a gel-combustion route using citric acid (C 6 H 8 O 7) and glycine (C 2 H 5 NO 2) as fuels and complexing agents. Titanium isopropoxide (Ti[OCH(CH 3) 2 ] 4 was dissolved in isopropyl alcohol under continuous magnetic stirring. In parallel, citric acid and glycine were dissolved in deionized water, followed by the addition of strontium nitrate (Sr(NO 3) 2 and iron nitrate (Fe(NO 3) 3 ⋅ 9 H 2 O) according to the desired stoichiometry. The aqueous precursor solution was subsequently added dropwise to the titanium-containing solution under continuous stirring. The resulting mixture was heated at 150 ° C until gel formation was achieved. Further heating at 350 ° C induced a self-sustained combustion reaction, yielding a porous precursor powder. The obtained powder was ground and calcined at 800 ° C for 3 h using a heating rate of 5 ° C min − 1.

Symmetric cells were fabricated using dense La 0.8 Sr 0.2 Ga 0.8 Mg 0.2 O 3 − δ (LSGM) electrolyte substrates prepared from a commercial powder (Fuel Cell Materials; specific surface area: 4 m 2 g − 1). For impedance measurements, the electrolyte pellets were uniaxially pressed at 250 MPa for 2 min and subsequently sintered at 1450 ° C for 10 h in air, yielding dense substrates with a final thickness of approximately 300 μ m. Conversely, pellets designated for current density–voltage (I – V) measurements were pressed at 100 MPa and sintered under identical conditions, resulting in thicker LSGM electrolytes of approximately 400 μ m. The surfaces of the sintered substrates were polished to obtain a smooth and homogeneous finish prior to electrode deposition.

Then, the as-prepared STF powder was dispersed in an alcoholic medium and mixed with Decoflux (WB41, Zschimmer & Schwarz) to obtain a homogeneous slurry. The slurry was subsequently applied to both sides of the LSGM substrates. Each symmetric cell was heated to 900 ° C for 2 h to remove the organic binder and ensure proper bonding between the electrode layers and the electrolyte. For impedance measurements, the resulting electrodes exhibited an active area of 0.196 cm 2, and Ag paste was applied as the current collector. For current–voltage measurements, a larger active area of 0.283 cm 2 was employed, and Pt paste was used as the current collector.

Compositional, structural, and microstructural analysis

X-ray diffraction (XRD) measurements were performed at room temperature using a Malvern Panalytical Empyrean diffractometer equipped with Cu K α radiation (λ = 1.5406 Å). Diffraction patterns were recorded within the 2 θ range of 10–120 ° using a step size of 0.015 °. Phase identification was conducted using the PDF-5+ database (ICDD), while structural refinement was performed by the Rietveld method using the FullProf Suite. The apparent crystallite size and microstrain were estimated from the profile-broadening analysis implemented in the Rietveld refinement. In this approach, the diffraction peaks are described by a Voigt function that accounts for both instrumental and sample-related broadening contributions, allowing the extraction of the coherent diffraction domain size and lattice microstrain from the full diffraction pattern (Rodríguez-Carvajal, 2001; Rodríguez-Carvajal and Roisnel, 2004).

X-ray photoelectron spectroscopy (XPS) data were acquired using a PHI 5000 VersaProbe II instrument featuring a monochromatic Al K α X-ray source (1486.6 eV) and a spherical capacitor analyzer, with a base pressure maintained below 1 × 1 0 − 10 Torr. The X-ray beam was focused to a 100 μ m spot size and operated at 25 W and 15 kV. The energy resolution was verified using an Ar-sputtered silver standard, yielding an Ag 3 d 5 / 2 FWHM of 0.69 eV. A dual-beam charge-neutralization method was continuously applied to prevent sample charging. All XPS spectra were energy-calibrated using the adventitious carbon C 1s peak (C–C/C–H bond) at a binding energy of 285.0 eV. Data processing and spectral fitting were performed using CasaXPS software (version 2.3.18).

Morphology was analyzed by scanning electron microscopy (SEM) using a Zeiss EVO MA10 variable-pressure scanning electron microscope equipped with an Oxford X-act energy-dispersive X-ray spectroscopy (EDS) detector. EDS data were processed using the AZtec 6.0 software package. SEM and EDS measurements were performed at an accelerating voltage of 20 kV and a working distance of approximately 8.7 mm. Elemental mapping was acquired using a dwell time of 200 μ s and three scanning passes. Powder samples were mounted directly on conductive carbon adhesive tape attached to the SEM sample holder for the measurements.

For higher-magnification imaging and characterization of the primary particle morphology and size, the STF powders were further examined using an AURIGA Compact field-emission scanning electron microscope (FE-SEM, ZEISS), equipped with scanning transmission electron microscopy (STEM) and focused ion beam (FIB-SEM) capabilities. Secondary-electron images were acquired using the InLens detector. For sample preparation, approximately 10 mg of STF nanopowder was dispersed in 1 mL of analytical-grade ethanol (Merck) and ultrasonicated for 1 min in a Sonorex TK 30 ultrasonic bath (Bandelin) operating at 50 kHz. Subsequently, 100 μ L of the suspension was deposited onto a 1 × 1 cm 2 silicon wafer (PureWafer) and allowed to dry overnight in air at room temperature (22 ° C). The dried specimens were mounted on SEM stubs using conductive carbon tape (EMS) prior to FE-SEM analysis.

Reduction behavior was investigated by temperature-programmed reduction (TPR) using a Micromeritics ChemiSorb 2720 system. The samples were pretreated at 300 ° C for 1 h under He flow (50 mL min − 1). Subsequently, approximately 50 mg of powder was heated up to 900 ° C at a rate of 10 ° C min − 1 under a 5 vol% H 2 /Ar flow of 50 mL min − 1. Hydrogen consumption was continuously monitored using a thermal conductivity detector (TCD).

Electrochemical evaluation

Electrochemical impedance spectroscopy (EIS) measurements were carried out using an Autolab PGSTAT302N system equipped with an FRA2 module. Spectra were collected with an AC perturbation amplitude of 20 mV over the frequency range of 0.1- 10 6 Hz. All samples were first evaluated in air over the temperature range of 750–550 ° C, using temperature decrements of 50 ° C. Subsequently, the same cells were tested under a humidified diluted reducing atmosphere of 7 vol% H 2 / N 2 following the same temperature sequence. In all cases, a total gas flow rate of 50 cm 3 min − 1 was maintained. Moreover, EIS measurements under reducing atmospheres were additionally carried out at different humidified hydrogen partial pressures (p H 2) of 7, 9.7, 19.4, 48.5, and 97%, using a total gas flow rate of 80 cm 3 min − 1 to evaluate the effect of fuel concentration on the electrochemical response. To limit the reducing strength of the atmosphere and prevent potential cell short-circuiting, all gas mixtures were humidified to approximately 3 mol% H 2 O. The obtained spectra were analyzed using appropriate equivalent-circuit models with ZView software. The resulting resistance values were normalized to the electrode surface area and divided by two to account for the symmetric cell configuration.

Distribution of Relaxation Times (DRT) analysis was performed using the DRTtools software developed by Ciucci and co-workers (Wan et al., 2015; Wang et al., 2025).

Current–voltage measurements were carried out using button-type cells. During testing, humidified concentrated H 2 was supplied to the anode at 100 mL min − 1, whereas air was fed to the cathode at 100 mL min − 1. Power density data were acquired using an AUTOLAB VIONIC potentiostat/galvanostat (INTELLO software) by stepping the potential down from the initial open-circuit voltage (OCV) in 20 mV increments.

STF nanopowders

Figure 1 displays the powder X-ray diffraction patterns of the nanostructured STF series calcined at 800 ° C, confirming the formation of single-phase materials without detectable secondary phases. A slight peak broadening is observed for all compositions, which is consistent with the nanostructured nature of the powders obtained by gel-combustion synthesis.

Detailed structural information was obtained by Rietveld refinements. The refinement profiles and the corresponding structural parameters are provided in Supplementary Figures S1–S3 and Supplementary Table S1 of the Supplementary Material, respectively. All diffraction patterns were satisfactorily refined using a cubic perovskite structure with space group Pm 3 ̄ m (ICDD PDF-5+ 04–027–7272), consistent with previous studies (Fuentes et al., 2015; Mroziński et al., 2020; Hasan et al., 2022). The refinements yielded low residual factors and χ 2 values below 2.6, indicating a satisfactory agreement between the calculated and experimental diffraction patterns. Increasing the Fe content from x = 0.20 (STF20) to x = 0.50 (STF50) results in a monotonic decrease in the lattice parameter from 3.90588(12) Å to 3.89699(10) Å. Fe substitution occurs at the 1b site, which is shared by Ti and Fe cations in octahedral coordination, leading to a progressive contraction and local relaxation of the perovskite lattice, as reflected in the structural parameters summarized in Supplementary Table S1. Similar structural evolution has been reported for Fe-substituted SrTiO 3 systems and has been associated with the combined effects of cation substitution, defect chemistry, and modifications of the local bonding environment rather than with ionic size alone (Fuentes et al., 2015; Mroziński et al., 2020). Although Ti + 4 in octahedral coordination has an ionic radius of 0.605 Å, the observed lattice contraction cannot be explained solely by the replacement of Ti + 4 by Fe cations (Shannon, 1976). Previous studies have shown that increasing Fe incorporation promotes the coexistence of Fe + 3 / F e + 4 species together with oxygen-vacancy formation for charge compensation, leading to significant changes in the local Fe–O bonding environment and the electronic structure of the perovskite (Rothschild et al., 2006; Da Silva et al., 2014). These changes can modify the covalent character of the Fe–O bonds and promote local distortions and relaxation of the BO 6 octahedra, resulting in a shortening of the average B–O bond distance and, consequently, contributing to a gradual contraction of the unit cell with increasing Fe content (Filatova et al., 2017; Nqayi and Sondezi, 2025).

To gain further insight into the microstructure of the STF powders, Rietveld refinement was employed to evaluate the crystallite size and lattice microstrain (Supplementary Table S1). The apparent crystallite sizes, corresponding to coherent diffraction domains, were found to lie within the nanometric range, varying from approximately 47–51 nm. Furthermore, a slight decrease in lattice microstrain was observed with increasing Fe content, indicating that the perovskite framework can effectively accommodate increasing Fe concentrations without introducing substantial additional lattice strain.

The surface chemical composition and oxidation states of the constituent elements in SrTi 1 − x Fe x O 3 − δ (STF) samples were investigated by XPS. Survey spectra (Supplementary Figure S4a) confirmed the presence of Sr, Ti, Fe, O, and adventitious carbon on all sample surfaces. High-resolution spectra of the Fe 2p, Ti 2p, O 1s, and Sr 3d regions are presented in Figure 2, whereas the Sr 3p and C 1s spectra are shown in (Supplementary Figure S4b), providing complementary information regarding the surface chemistry, oxidation states, and segregation phenomena occurring in the STF series. Detailed fitting parameters are also given in Supplementary Table S2 of the Supplementary Material.

The Fe 2p spectra (Figure 2a) consist of the characteristic Fe 2 p 3 / 2 and Fe 2 p 1 / 2 spin–orbit doublets accompanied by pronounced shake-up satellite features. The dominant Fe 2 p 3 / 2 component is centered at 710.9 ± 0.2 eV and is associated primarily with Fe + 3, while a secondary contribution at slightly higher binding energy at 712.7 ± 0.2 eV corresponds to Fe + 4 (Ghaffari et al., 2012). Corresponding contributions are observed in the Fe 2 p 1 / 2 region at 724.5 ± 0.2 eV and 726.3 ± 0.2 eV, respectively. The fitting therefore indicates mixed Fe + 3 / F e + 4 character in all compositions. The relative contribution of the Fe + 4 component increases progressively from STF20 to STF50, giving average Fe oxidation states of approximately 3.2, 3.3, and 3.4 for x = 0.20, 0.35, and 0.50, respectively. This evolution indicates an increasing contribution of Fe + 4 with Fe substitution and is consistent with charge compensation involving changes in Fe valence and oxygen non-stoichiometry. The well-defined satellite structures observed between the spin–orbit components further support the predominance of high-spin Fe + 3 species within the perovskite lattice. No evidence for metallic iron or segregated iron oxide phases was detected, indicating that iron remains structurally incorporated within the STF framework.

The Ti 2p spectra (Figure 2b) display well-resolved Ti 2 p 3 / 2 and Ti 2 p 1 / 2 peaks located at 458.3 ± 0.1 and 464.0 ± 0.1 eV, respectively, characteristic of Ti + 4 in an octahedral oxide environment (Baek et al., 2020). The peak positions, widths, and spin–orbit splitting remain essentially unchanged across the compositional series, demonstrating that the electronic environment of titanium is largely unaffected by Fe substitution. No low-binding-energy peak attributable to Ti + 3 was detected. Because X-ray irradiation can induce reduction of transition-metal species (Sayers and Armstrong, 1978; Baek et al., 2020; Bagus et al., 2019), measurements were performed using a protocol in which the X-ray beam was moved to a fresh position after acquisition to minimize prolonged irradiation of the same area. Under these measurement conditions, Ti is therefore considered to remain predominantly in the Ti + 4 state. The results suggest that the charge compensation associated with Fe incorporation occurs mainly through changes in Fe valence and oxygen non-stoichiometry rather than through changes in the oxidation state of titanium.

The O 1s spectra (Figure 2c) were fitted using two principal components. The dominant low-binding-energy contribution at 529.9 ± 0.1 eV is assigned to lattice oxygen within the perovskite structure, corresponding to oxygen coordinated with the Ti and Fe cations. A second, higher-binding-energy contribution at 532.2 ± 0.4 eV range is associated with surface oxygen-containing species. This contribution decreases progressively from STF20 to STF50. Although the higher-binding-energy O 1s component cannot by itself uniquely identify a carbonate species, its systematic evolution is consistent with the surface chemistry indicated by the Sr 3d and C 1s spectra discussed below.

The Sr 3d spectra (Figure 2d) contain the characteristic Sr 3 d 5 / 2 /Sr 3 d 3 / 2 doublet at 132.9 ± 0.1 and 133.9 ± 0.1, respectively, associated with Sr + 2 occupying the A-site of the perovskite lattice (Baek et al., 2020). In addition, a second Sr 3d doublet at 134.7 ± 0.6 and 135.7 ± 0.4, respectively, is observed at higher binding energy. Its inclusion is supported by independent spectral features observed in the Sr 3p and C 1s regions (Supplementary Figure S4b), indicating the presence of Sr in a chemically distinct surface environment. The higher-binding-energy Sr component is therefore assigned to surface-associated Sr species in a different chemical environment from lattice Sr + 2, consistent with carbonate-containing or hydroxide Sr species.

The relative intensity of this surface-related Sr component is greater for STF20 and STF35 and substantially lower for STF50. This systematic variation, together with the corresponding decrease in the higher-binding-energy O 1s contribution, is consistent with a reduction in the amount of surface-associated Sr species with increasing Fe content. These observations are consistent with the tendency of Sr-containing perovskite oxides to undergo Sr surface segregation, followed by interaction of the segregated Sr species with atmospheric gases during sample handling, preparation, and storage (Cai et al., 2012; Opitz et al., 2018; Koo et al., 2018; Chen et al., 2012; Zheng et al., 2023; Kubicek et al., 2014; Baeumer et al., 2021). The present XPS results therefore provide evidence for a surface-enriched Sr chemical environment consistent with Sr segregation and carbonate formation, although the precise depth distribution of the segregated species cannot be determined from conventional XPS alone.

Further information is obtained from the Sr 3p and C 1s spectra presented in (Supplementary Figure S4b), which confirm the above interpretation of the Sr 3d XPS regions. The Sr 3p spectra contain the expected Sr 3 p 3 / 2 and Sr 3 p 1 / 2 components together with the contributions associated with the surface Sr chemical environment, see Supplementary Table S2 of the Supplementary Material. The C 1s spectra contain a dominant component near 285 eV that is attributed primarily to adventitious carbon, together with a higher-binding-energy contribution around 289–290 eV that is characteristic of carbonate-containing species. The carbonate assignment on the surface is based on the correlation between the C 1s carbonate-related contribution and the additional Sr components observed in the Sr 3d and Sr 3p regions, together with the corresponding surface-related contribution in the O 1s spectrum.

The particle morphology was first examined by SEM at relatively low magnification, using the regions analyzed by EDS mapping. The micrographs reveal a highly porous microstructure composed of interconnected agglomerates with irregular morphology. The presence of abundant open porosity is characteristic of powders synthesized by the gel-combustion method, in which the rapid evolution of gaseous species during the combustion reaction promotes the formation of a porous framework and limits extensive densification during calcination (Tarragó et al., 2015; Varma et al., 2016). Higher-magnification SEM images and the corresponding particle-size distributions are presented in Supplementary Figure S5. The agglomerates are composed of nanosized primary particles, with mean particle sizes of 52 ± 5 nm, 43 ± 4 nm, and 53 ± 4 nm for STF20, STF35, and STF50, respectively. These values are consistent with the coherent diffraction domain sizes obtained from Rietveld refinement, providing complementary evidence of the nanoscale nature of the STF powders synthesized via the gel-combustion route.

The elemental distribution of the STF powders was investigated by EDS mapping. The analyses show a homogeneous distribution of the constituent elements in all samples, consistent with their nominal stoichiometries. Figure 3 presents the SEM micrograph and EDS elemental maps of the STF50 powder, whereas the corresponding results for STF20 and STF35 are provided in the Supplementary Material (Supplementary Figures S6–S7). The maps for the three samples reveal a homogeneous spatial distribution of Sr, Ti, Fe and O throughout the analyzed region, with no evidence of elemental segregation or local enrichment. A minor carbon contribution was also detected in all compositions. Although this observation is consistent with the presence of carbonate-containing species inferred from the XPS analysis, part of the detected carbon signal may originate from the conductive carbon tape used during sample preparation.

Redox behavior and structural evolution under reducing conditions

The redox behavior of the STF series was investigated by H 2 -TPR, and the resulting reduction profiles are shown in Supplementary Figure S8 of the Supplementary Material. Peaks centered at approximately 356 ° C and 407 ° C are consistent with the partial reduction of Fe cations through the Fe + 4 / F e + 3 and Fe + 3 / F e + 2 redox couples, respectively (Barbero et al., 2009; Santaya et al., 2019). A broad feature is observed above 650 ° C, a temperature range in which the reduction of Ti + 4 to Ti + 3 has previously been reported (Kayaalp et al., 2019). However, this region has also been associated with the onset of Fe reduction toward metallic iron (Barbero et al., 2006; Merino et al., 2006). In the present case, the progressive increase in the intensity of this feature with increasing Fe content suggests that it is predominantly associated with the initial stages of Fe + 2 reduction to Fe 0. An additional reduction feature appears above 750 ° C, particularly for the Fe-rich samples, which can be assigned to the further reduction of Fe + 2 toward Fe 0.

To obtain a deeper understanding of these reduction processes, H 2 -TPR experiments were performed up to 750 and 900 ° C. After cooling to room temperature under the same reducing atmosphere, the resulting samples were characterized by ex situ X-ray diffraction, as shown in Supplementary Figure S9. The diffraction patterns obtained after reduction at 750 ° C show that the perovskite phase remains dominant in all samples, although metallic Fe (PDF 00–006–0696) is detected in every composition. In addition, STF50 exhibits the emergence of a Ruddlesden–Popper (RP) layered phase, identified as Sr 3 Ti 1.5 Fe 0.5 O 7 (PDF 04–014–6009).

After reduction at 900 ° C, the structural transformation becomes significantly more pronounced. In addition to the exsolution of metallic Fe, both STF35 and STF50 exhibit clear reflections associated with the RP phase, indicating that the reduction process promotes a progressive reorganization of the initial single-phase perovskite. The extent of this transformation increases with Fe content, as evidenced by the more intense RP reflections and the larger fraction of exsolved metallic Fe observed in the Fe-rich compositions.

The phase fractions determined by Rietveld refinement are summarized in Table 1. Although all compositions were initially obtained as single-phase perovskites, the reduction treatments induce the formation of metallic Fe and RP secondary phases. For STF20 and STF35, the perovskite phase remains the major constituent after reduction. In contrast, STF50 undergoes a much more extensive transformation, leading to the formation of an RP-dominated multiphase system after treatment at 900 ° C. These results indicate that increasing Fe content enhances the tendency of the perovskite lattice to accommodate reduction through structural reorganization, generating multiphase systems.

Electrochemical performance as symmetric SOFC electrodes

Electrochemical impedance spectroscopy (EIS) measurements were performed on symmetric cells under both air and reducing atmospheres while lowering the temperature from 750 to 550 ° C in 50 ° C intervals. Figures 4a,b show the comparative Nyquist plots of the STF20, STF35, and STF50 symmetric electrodes measured at 700 and 750 ° C under oxidizing conditions. The complete set of impedance spectra is provided in Supplementary Figure S10 of the Supplementary Material. To facilitate a direct comparison of the electrochemical performance of the electrodes, the ohmic resistance (R s) was excluded from the real-axis scale.

The spectra of all samples are dominated by a depressed polarization response that can be satisfactorily described by the equivalent-circuit model shown in the inset of Figure 4. The circuit includes an inductive element (L 1) accounting for the contribution of the measurement leads, the ohmic resistance (R s), mainly associated with the electrolyte, a R p 1 – CPE 1 contribution describing faster electrode processes, and a Warburg element (W 1) accounting for the characteristic transport-related response of the mixed ionic–electronic conducting electrode. According to previous studies on MIEC oxygen electrodes, including STF-based materials, the impedance response of the oxygen reduction reaction may involve partially overlapping contributions from oxygen surface exchange, charge transfer, oxide-ion incorporation, and oxide-ion transport through the electrode and across the electrode–electrolyte interface. Within this framework, the R p 1 – CPE 1 contribution is associated mainly with faster electrochemical processes involving charge transfer and oxide-ion transfer at or near the electrode–electrolyte interface, whereas the Warburg-type response is mainly associated with oxygen-transport processes within the mixed-conducting electrode, potentially coupled with surface-exchange processes (Adler, 2004; Sacanell et al., 2007; Mroziński et al., 2020; Zhao et al., 2020; Vivier and Orazem, 2022).

For the STF20 sample, the total R p decreases from approximately 0.21 Ω cm 2 at 700 ° C to nearly 0.125 Ω cm 2 at 750 ° C. STF35 and STF50 exhibit nearly overlapping impedance responses throughout the investigated temperature range, indicating comparable electrochemical performance. At 700 ° C, both compositions present R p values close to 0.04–0.05 Ω cm 2, which decrease to approximately 0.02–0.03 Ω cm 2 at 750 ° C. The decrease in R p with increasing temperature reflects the thermally activated nature of the oxygen reduction reaction, whereas the consistently lower R p values of STF35 and STF50 indicate improved electrode kinetics with increasing Fe content. Importantly, the fitted contributions show that this improvement is particularly pronounced in the transport-related response. At 750 ° C, W 1 − R decreases from 0.1024 Ω cm 2 for STF20 to only 0.0040 and 0.0060 Ω cm 2 for STF35 and STF50, respectively. A similar trend is observed at 700 ° C, where W 1 − R decreases from 0.1335 Ω cm 2 for STF20 to 0.0200 and 0.0174 Ω cm 2 for STF35 and STF50, respectively. This marked attenuation indicates that the improved performance of the Fe-rich compositions is strongly associated with a reduction in the transport-related contribution to the overall electrode polarization. The R p 1 contribution also decreases with increasing Fe content, although less dramatically, indicating that faster interfacial electrochemical processes are also favored.

This behavior is consistent with the defect chemistry of Fe-substituted SrTiO 3, for which increasing Fe incorporation is associated with mixed Fe + 3 / F e + 4 states and oxygen-vacancy formation, both of which can favor mixed ionic–electronic transport and oxygen-exchange kinetics. These mechanisms provide a physically plausible framework for interpreting the lower polarization resistance of STF35 and STF50, although their individual contributions cannot be unambiguously separated from the impedance spectra alone. The nanometric character of the STF powders may additionally favor the electrode response by providing a high density of active surface sites and relatively short transport pathways (Gómez et al., 2019; Mroziński et al., 2020).

For STF35 and STF50 at the highest temperatures, a small additional very-low-frequency contribution was included in the fitting model to account for a minor gas-transport-related process (Adler, 2004; Sacanell et al., 2017). Although its contribution to the total polarization resistance was negligible, its inclusion improved the description of the low-frequency response and allowed a more reliable determination of the Warburg contribution. Thus, gas-phase oxygen transport does not appear to constitute a significant limitation under the investigated conditions. The relevant resulting fitting parameters are summarized in Supplementary Table S3.

Figures 5a,b compare the impedance spectra of the STF20, STF35, and STF50 electrodes measured at 700 and 750 ° C under humidified diluted hydrogen (7 vol% H 2 / N 2). The complete temperature-dependent dataset (550–650 ° C) is presented in Supplementary Figure S11 of the Supplementary Material, while the corresponding equivalent-circuit fitting parameters for all temperatures are summarized in Supplementary Table S4.

STF20 exhibits the largest polarization arc at both temperatures, indicating slower electrode kinetics under reducing conditions. Increasing the Fe concentration in STF35 substantially decreases the R p, while a further increase in Fe content in STF50 results in the smallest impedance response among the investigated samples across the entire temperature range. At 750 ° C, STF50 displays a highly compressed arc with an R p of approximately 0.56 Ω cm 2.

The lower R p exhibited by STF35 and, particularly, STF50 is consistent with their enhanced reducibility, as evidenced by the H 2 -TPR profiles and ex situ X-ray diffraction patterns. Higher Fe contents favor oxygen-vacancy formation and increase the redox activity of the lattice, which may contribute to the electrochemical processes occurring at the fuel electrode (Merkle and Maier, 2008; Mroziński et al., 2020). Under reducing conditions, Fe-rich STF compositions can accommodate oxygen loss through structural rearrangements involving Fe exsolution and the formation of RP-type domains (Santaya et al., 2022). Similar redox-driven transformations have been reported for other Fe-containing perovskites, where metallic Fe exsolution was accompanied by the formation of Ruddlesden–Popper domains as the parent lattice reorganized to accommodate oxygen loss (Wang et al., 2022). In STF-based materials, RP phases formed under reducing conditions have also been shown to reversibly transform back into the parent perovskite structure upon reoxidation, together with partial reincorporation of the exsolved nanoparticles into the oxide lattice (Schwiers et al., 2024). These observations highlight the ability of Fe-rich STF compositions to undergo dynamic structural and redox adaptations under changing atmospheres.

To further assess the influence of fuel concentration on the electrochemical response of the STF50 electrode, EIS measurements were carried out at 750 ° C under a constant humidity level, while systematically varying the hydrogen partial pressure (p H 2) from 7% to 97% using N 2 as the balancing gas. The corresponding Nyquist spectra are shown in Figure 6, together with the fitted responses represented by the solid red lines. The corresponding impedance spectra and equivalent-circuit fitting results obtained for STF50 under the 97% H 2 atmosphere at the remaining temperatures are provided in the Supplementary Material (Supplementary Figure S11d; Supplementary Table S4).

As can be seen in Figure 6, increasing the hydrogen concentration progressively reduces the arc diameter. At an intermediate level of 19.4% p H 2, the semicircle diameter is markedly reduced to nearly 0.33 Ω cm 2, indicating that the electrochemical processes become increasingly favored as the hydrogen concentration increases. Under a highly concentrated hydrogen atmosphere (97% p H 2), the spectrum exhibits the smallest arc diameter, corresponding to an R p of approximately 0.16 Ω cm 2. This progressive decrease in R p with increasing hydrogen partial pressure is consistent with the enhanced reducing character of the atmosphere, which promotes Fe reduction and oxygen-vacancy formation and may further induce Fe exsolution and RP-type structural rearrangements, as discussed above under diluted reducing conditions.

A Distribution of Relaxation Times (DRT) analysis was performed to further resolve the individual processes contributing to the overall electrode polarization. The resulting DRT spectra obtained under oxidizing and reducing atmospheres are presented in Figure 7. A comparison between the DRT profiles derived from the experimental and fitted impedance spectra is provided in Supplementary Figure S12 of the Supplementary Material. The close agreement between the two profiles supports the reliability of the equivalent-circuit models proposed in Figures 4–6 for describing the frequency-dependent response of the STF electrodes.

Under oxidizing conditions (Figure 7a), three main relaxation contributions, denoted as P1, P2, and P3, are identified at 750 ° C. The P1 contribution, located in the low-frequency region at approximately 1 Hz, shows a magnitude comparable to the response reported for porous SrTi 0.3 Fe 0.7 O 3 − δ electrodes, where it was attributed to gas-phase diffusion limitation within the porous electrode (Mroziński et al., 2023), in agreement with observations reported for other related perovskite electrodes (Orozco Gil et al., 2024; Arcentales-Vera et al., 2026). Its relatively small magnitude is consistent with the minor gas-transport contribution identified by the equivalent-circuit analysis. The higher-frequency P2 and P3 contributions extend from the HF (1 0 2 – 1 0 3 Hz) to the UHF (> 1 0 3 Hz) regions, where electrochemical processes related to charge and oxygen-ion transport have been reported (Mroziński et al., 2020; Santaya et al., 2022; Han et al., 2023). The relatively pronounced high-frequency response may also be influenced by the electrode–electrolyte firing conditions. Here, the electrodes were fired at 900 ° C to promote their adhesion to the electrolyte while preserving the nanometric character of the STF powders. In comparison, similar electrolyte-supported configurations based on STF electrodes have been processed at higher firing temperatures, for example, at 1100 ° C for STF and STF-Ni electrodes deposited on GDC-coated ScSZ electrolytes Lacharme et al. (2026). Consequently, the lower firing temperature used in this work may therefore be one factor contributing to the higher resistance associated with the electrode–electrolyte response.

Among the three samples, STF20 exhibits the highest DRT amplitudes over the entire temperature range, in agreement with its higher polarization resistance obtained from the EIS analysis. In contrast, STF35 and STF50 display remarkably similar DRT profiles in terms of both shape and magnitude, together with lower overall contributions. Upon cooling, the relaxation contributions progressively shift toward lower frequencies, consistent with the thermally activated nature of the electrode processes, while additional high-frequency processes become increasingly apparent Orozco Gil et al. (2024). Although these additional HF processes become more clearly identifiable upon cooling, their contribution remains smaller than that of the dominant P2 process. The lower DRT response observed for STF35 and STF50 is consistent with their lower polarization resistance and with previous reports showing enhanced electrochemical activity with increasing Fe content in STF materials (Mroziński et al., 2020).

For Fe-rich STF compositions, previous studies have related the enhanced electrochemical performance to changes in the defect chemistry and electronic structure induced by the substitution of Ti + 4 by Fe. Depending on the atmosphere and electrochemical conditions, Fe can coexist in the Fe + 3 and Fe + 4 oxidation states, while charge compensation associated with the formation of Fe + 3 can occur through electronic holes and, predominantly, oxygen-vacancy formation (Merkle and Maier, 2008; Mroziński et al., 2020). These changes have been proposed to favor mixed ionic–electronic transport. In Fe-rich compositions, a pronounced reduction of the effective band gap and an electron-rich character have also been reported, which may increase the availability of electronic carriers for oxygen reduction (Mroziński et al., 2020). These characteristics have been linked to enhanced oxygen exchange kinetics and electrochemical activity (Rothschild et al., 2006; Merkle and Maier, 2008).

Under humidified reducing conditions (Figure 7b), the relaxation contributions are broader and more difficult to resolve than under oxidizing conditions, particularly in the MF and UHF regions. At 750 ° C, four main relaxation contributions (P1–P4) are observed. The P1 contribution is the dominant process and is located in the MF region at approximately 10–30 Hz. Similar LF–MF responses have been reported for Fe-rich and Ni-doped STF-based anodes, where these contributions have been associated with H 2 transport and H 2 adsorption–dissociation processes (Santaya et al., 2022). The remaining P2–P4 contributions are located at higher frequencies, extending from the HF to the UHF region. Similar high-frequency processes have been reported for STF-based anodes and have been associated with faster steps of the hydrogen oxidation reaction, including charge transfer and oxygen-ion transport across the electrode–electrolyte interface (Mroziński et al., 2020; Santaya et al., 2022).

The STF20 composition exhibits the broadest and most intense response, dominated by the MF contribution, whereas STF35 and STF50 show lower amplitudes. As shown in Figure 7, for STF50, the dominant P1 contribution remains pronounced at 600 ° C, but its intensity progressively decreases with increasing temperature, resulting in a marked contraction of the overall DRT response toward 750 ° C. This evolution is consistent with the decrease in R p observed with increasing temperature and occurs within the temperature range where Fe exsolution and the gradual formation of RP-type domains are observed under reducing conditions, suggesting that these structural changes may be associated to the evolution of the electrode response under hydrogen. In addition, although STF35 and STF50 exhibit similar dominant MF contributions, STF50 shows a stronger attenuation of the secondary processes in the 1 0 2 – 1 0 5 Hz range, contributing to its lower overall R p. Notably, this improved performance is achieved within the intermediate-temperature range, where the nanostructured character of the combustion-synthesized powders can be largely preserved. Operating at higher temperatures would promote particle coarsening and surface-area loss, reducing the benefits associated with the nanoscale microstructure and potentially accelerating microstructural degradation (Gómez et al., 2019; Haron et al., 2017).

Furthermore, increasing the hydrogen partial pressure to 97% in STF50 further attenuates the MF contribution and markedly reduces the secondary processes, indicating that higher hydrogen availability mitigates fuel-side polarization losses and improves the overall electrochemical performance.

The temperature dependence of the electrochemical response was further evaluated through Arrhenius plots, yielding apparent activation energies (E a) for the overall area specific resistance (ASR). Figure 8a shows the linear fits and the resulting activation energies measured in air. The apparent activation energy increases with increasing Fe substitution, rising from 0.95 eV for STF20 to 1.19 eV for STF50. However, it is important to note that this higher thermal activation barrier does not translate into poorer overall performance; rather, STF50 consistently exhibits the lowest ASR values in air across the evaluated temperature range.

Under reducing atmospheres (Figure 8b), the activation energy follows the same compositional trend observed in air, systematically increasing with increasing Fe content. For the measurements conducted in 7% H 2, the activation energy rises from 0.94 eV for STF20 to 1.20 eV for STF50. Furthermore, evaluating the STF50 electrode under a highly concentrated fuel stream (97% H 2) yields an activation energy of 1.31 eV. Despite presenting higher apparent activation energies than its lower-Fe counterparts, the STF50 electrode demonstrates superior electrochemical activity, displaying the lowest polarization resistance in both oxidizing and reducing environments. This indicates that the higher Fe incorporation provides a highly favorable overall electrochemical response, where the enhanced mixed ionic-electronic conductivity and/or the larger density of active sites may offset the effect of the higher activation energy barrier.

Figure 9 presents the current density–voltage (I – V) and power density curves of the STF50 single cell employing an LSGM electrolyte, measured under humidified H 2 fuel and air over the operating temperature range of 700–850 ° C. The cell exhibits open-circuit voltage (OCV) values close to 1.05–1.08 V across all temperatures, indicating adequate gas separation and high electrolyte density during measurements. Upon increasing the current density, the voltage drop is progressively reduced at higher temperatures, reflecting a significant decrease in the total cell resistance and enhanced thermal activation of the electrode activity. As a consequence, the STF50 configuration reaches maximum power densities of approximately 60 mW cm − 2 at 700 ° C and 110 mW cm − 2 at 750 ° C. Furthermore, this electrochemical performance is notably enhanced at elevated temperatures, delivering approximately 210 mW cm − 2 at 800 ° C and reaching up to approximately 330 mW cm − 2 at 850 ° C.

This work demonstrates that the Fe/Ti ratio is a key parameter governing the structural, surface-chemical, redox, and electrochemical behavior of nanocrystalline SrTi 1 − x Fe x O 3 − δ electrodes for IT-S-SOFCs. Increasing Fe substitution produced a systematic evolution in the crystal structure, surface chemistry, reducibility, and electrochemical response of the STF series. Surface chemical analysis indicated that Ti remained predominantly in the Ti + 4 state, whereas the Fe 2p spectra were consistent with mixed Fe + 3 / F e + 4 surface species. In addition, the lower intensity of carbonate-related Sr and O components in the Fe-rich samples suggests that increasing Fe content modifies the surface chemistry and decreases the abundance of carbonate-containing surface species.

The electrochemical results demonstrate that STF35 and STF50 exhibited markedly lower polarization resistances than STF20. Under oxidizing conditions, equivalent-circuit analysis showed that this improvement is particularly associated with a strong reduction in the transport-related contribution to the electrode polarization, together with a smaller decrease in the faster interfacial electrochemical contribution. DRT analysis revealed substantially lower contributions from the dominant MF and secondary HF–UHF processes for STF35 and STF50 at intermediate temperatures, with the differences being particularly pronounced at 650 ° C. Under reducing atmospheres, increasing the hydrogen partial pressure to 97 % in STF50 further attenuated the MF contribution and markedly reduced the secondary relaxation processes, indicating a more favorable electrochemical response. Preliminary button-cell measurements further support the feasibility of STF50 as a symmetric electrode material for intermediate-temperature operation.

Nevertheless, the electrochemical advantage of the Fe-rich compositions is accompanied by a greater sensitivity to reducing conditions. Ex situ reduction treatments revealed an increasing tendency toward Fe exsolution and Ruddlesden–Popper-type phase formation as Fe content increases, with STF50 exhibiting the most pronounced structural transformation. Therefore, STF50 should be regarded as the most electrochemically active composition within the studied series, rather than as an unequivocally optimal formulation for long-term operation.

The performance of STF50 was further validated in a button-cell configuration employing humidified H 2 as the fuel and air as the oxidant. Maximum power densities of approximately 60, 110, 210, and 330 mW cm − 2 were achieved at 700, 750, 800, and 850 ° C, respectively, demonstrating the excellent electrochemical viability of this composition over the intermediate-temperature range.

Overall, nanocrystalline SrTi 1 − x Fe x O 3 − δ perovskites constitute a promising cobalt-free platform for intermediate-temperature symmetric SOFCs. Their compositional simplicity, tunable Fe/Ti ratio, and nanoscale character make them attractive candidates for further development. Future work should focus on long-term operation, redox cycling, and post-test structural and microstructural characterization to establish the stability limits of the most active compositions and clarify the role of reduction-induced phase transformations during cell operation.