Dimensionality and Compositional Effects on Sr–Fe-Based Ruddlesden–Popper Oxides for Oxygen Catalysis

Abstract

Understanding how structural and compositional features influence the Oxygen Reduction Reaction (ORR) and Oxygen Evolution Reaction (OER) in oxygen electrocatalysis is crucial for the rational design of efficient catalysts. The O p-band center, obtained from density functional theory (DFT) calculations, serves as a predictive electronic descriptor linking the composition and structure of the oxide catalyst to ORR and OER activity. Ruddlesden–Popper oxides Sr _n+1Fe _n O_3n+1 (1 < n < ∞) provide a versatile platform for tuning this descriptor. Here, we systematically evaluate the effects of dimensionality, Fe substitution, and oxygen nonstoichiometry in the Sr _n+1Fe _n(1–x)M _nx O_3n+1−δ series (n = 1, 2, ∞; M = 3d-metal; x = 1/8; δ = 0, 1/8). Both increasing slab thickness (n = 1 → ∞) and Fe substitution with more electronegative transition metal elements enhance metal–oxygen hybridization, shifting the O p-band center toward the Fermi level by up to 0.2 and 0.45 eV, respectively, whereas 12% oxygen deficiency shifts it downward by up to 0.45 eV. Across the series, the combined effects of composition and structure span a ∼0.7 eV range in the O p-band center, implying only modest intrinsic variations in ORR/OER activity, often surpassed by extrinsic factors such as morphology and microstructure.

1. Introduction

The development of efficient and durable oxygen catalysts is a critical challenge in advancing sustainable energy technologies including water electrolysis, rechargeable metal–air batteries, and regenerative fuel cells. Extensive efforts have been dedicated to the design of novel oxide-based materials that can serve as active, stable, and earth-abundant oxygen electrocatalysts. − Despite significant progress, the catalytic performance of many state-of-the-art materials remains insufficient for large-scale applications, highlighting the need for advanced compositions and structure–property paradigms that can guide the discovery of next-generation oxygen catalysts. Among oxide families, perovskite-related structures such as Ruddlesden–Popper (RP) phases have attracted growing attention. These materials, with general formula A _n+1B _n O_3n+1, consist of perovskite-like blocks of corner-sharing BO₆ octahedra separated by rock-salt-type AO layers (Figure ). By varying the structural parameter n, RP oxides offer a platform to systematically tune the dimensionality of the B–O network from quasi-two-dimensional (low n) to fully three-dimensional (as n → ∞). Several RP systems have already demonstrated promising activity for the Oxygen Reduction Reaction (ORR), and the Oxygen Evolution Reaction (OER), including La_0.5Sr_1.5Ni_1–x Fe _x O_{4 ± δ}, Sr _n+1Fe _n O_3n+1−δ, or Sr _n+1(Co_0.8Fe_0.1Nb_0.1) _n O_3n+1−δ, underscoring the central role of dimensionality in governing oxygen electrocatalysis. ,− Together with dimensionality, the tunable composition of A _n+1B _n O_3n+1 has profound implications for the electronic structure, charge transport, and defect chemistry, all of which are intimately linked to catalytic behavior. However, a systematic understanding of how composition and dimensionality jointly influence these properties remains an open challenge in rational catalyst design.

1.

Crystal structures of the supercells used to model Sr _n+1(Fe_0.875 M_0.125) _n O_3n+1 (M = 3d metal): (a) Sr₂Fe_7/8M_1/8O₄, (b) Sr₃Fe_14/8M_2/8O₇, and (c) SrFe_7/8M_1/8O₃. O1, O2, and O3 denote the sites for oxygen-vacancy incorporation. Color code: Fe in brown, M in blue, Sr in green, and O in red.

Density functional theory (DFT) has become indispensable for establishing structure–composition–activity relationships in oxygen catalysts. Electronic descriptors derived from the projected density of statesparticularly the O p-band center relative to the Fermi levelhave emerged as robust indicators of catalytic potential, correlating with overpotential, oxygen-exchange kinetics, and oxygen-vacancy formation energies. − Systematic evaluation of the O p-band center across RP compounds of different dimensionality and composition can therefore provide insight into their catalytic suitability and enable predictive guidelines for designing more efficient oxygen electrocatalysts. With this aim, the present work focuses on the RP-family Sr _n+1Fe _n O_3n+1, − a series of materials of interest for oxygen electrocatalysis. The perovskite end-member (n = ∞), SrFeO_3‑δ is a mixed ionic–electronic conductor which demonstrated potential for the ORR at high temperature and the OER in alkaline media and at high temperature. − Its performance can be further improved through low-level Fe-substitution in SrFe_1–x M _x O_3−δ (x < 0.25). ,− This study includes the lower dimensional RP members Sr₂FeO₄ (n = 1) and Sr₃Fe₂O₇ (n = 2), that also show promise for electrocatalytic applications, although the number of available experimental and theoretical studies remains considerably more limited. ,−

The Sr _n+1Fe_(1–x)n M _xn O_3n+1 (n = 1, 2, ∞) phases provide a well-defined structural platform to investigate how dimensionality and composition influence the energy position of the O p-band center and, by extension, the material’s activity for oxygen catalysis. To independently evaluate the effect of chemical composition, Fe is partially substituted with a small fraction (1/8) of a 3d transition metal (TM), yielding Sr _n+1 Fe_(1–1/8)n M _n/8O_3n+1 (M = 3d metal), preserving the formal oxidation state of the B-site cations while introducing modifications to the electronic structure. In addition, the impact of oxygen nonstoichiometry is examined for the n = 1 and n = ∞ members, by introducing oxygen vacancies (SrFe_7/8M_1/8O_2.875, Sr₂Fe_7/8M_1/8O_3.875) and evaluating their effect on the O p-band center. Altogether, this approach enables a systematic evaluation of how structural dimensionality and targeted chemical tuning govern the electronic descriptors relevant for oxygen electrocatalysis.

2. Methodology

Calculations were performed using the Vienna ab initio simulation program (VASP). , The interaction of core electrons with the nuclei is described by the Projector Augmented Wave (PAW) method. Previous investigations have shown that the Strongly Constrained and Appropriately Normed (SCAN) meta-GGA exchange–correlation functional correctly reproduces the structural, magnetic, and electronic properties of complex oxides, − and it is therefore used in this study. Moreover, Jacobs et al. reported that SCAN yields lower errors than other functionals when comparing the predicted O p-band center values with the actual X-ray emission spectroscopy (XES) O p-band center. The energy cutoff for the plane wave basis set was set to 600 eV throughout the calculations. The integration in the Brillouin zone was done on the appropriate k-point meshes determined by the Monkhorst–Pack scheme. Parameters for the calculations are listed in Table S1.

The initial crystal structures of Sr₂FeO₄ and Sr₃Fe₂O₇ were obtained, respectively, from ICSD entries 69849 and 74437, both crystallizing in the I4/mmm space group. To model the M-substituted compounds, the tetragonal I4/mmm symmetry was reduced to triclinic. Specifically, a 2 × 2 × 2 supercell of the primitive unit cell was used for the n = 1 RP phase (Sr₁₆Fe₈O₃₂), while a 2 × 2 × 1 supercell was employed for the n = 2 phase (Sr₁₂Fe₈O₂₈). In each case, one of the eight Fe atoms was replaced with a dopant M atom, resulting in the compositions Sr₂Fe_0.875M_0.125O₄ and Sr₃Fe_1.75M_0.25O₇, respectively (Figure a,b). The starting atomic positions for the cubic perovskite SrFeO₃ were taken from ref . A 2 × 2 × 2 supercell of the primitive cubic structure (Sr₈Fe₈O₂₄, Figure c) was constructed, and M substitution was simulated by replacing one Fe atom, yielding SrFe_0.875M_0.125O_3.

To calculate the oxygen-vacancy formation energies of Sr₂Fe_0.875M_0.125O₄ and SrFe_0.875M_0.125O₃, one oxygen atom was selectively removed from the respective supercells Sr₁₆Fe₇MO₃₂ and Sr₈Fe₇MO₂₄. This procedure yields the oxygen-deficient compositions Sr₁₆Fe₇MO₃₁ (Sr₂Fe_7/8M_1/8O_3.875) and Sr₈Fe₇MO₂₃ (SrFe_7/8M_1/8O_2.875); in both cases, the formal oxidation state of B-site cations is +3.75, which enable evaluation of the influence of the M dopant on the vacancy formation energy. Note that for the perovskite series, this level of oxygen vacancies is close to that experimentally observed at RT (for instance, SrFeO_2.84 in ref or SrFe_0.9Cu_0.1O_2.71 in ref ).

The oxygen-vacancy formation processes can be described by the following reactions:

$$n=\mathrm{\infty }:{\mathrm{S}\mathrm{r}}_{16}{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{32}\to {\mathrm{S}\mathrm{r}}_{16}{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{31}+\frac{1}{2}{\mathrm{O}}_2$$ 1

$$n=1:{\mathrm{S}\mathrm{r}}_8{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{24}\to {\mathrm{S}\mathrm{r}}_8{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{23}+\frac{1}{2}{\mathrm{O}}_2$$ 2

From the computed total energies, the oxygen-vacancy formation energies (E _ovf) can be determined as

$$E_{\mathrm{o}\mathrm{v}\mathrm{f}}=\mathrm{E}({\mathrm{S}\mathrm{r}}_{16}{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{31})+\frac{1}{2}\mathrm{E}({\mathrm{O}}_2)-\mathrm{E}({\mathrm{S}\mathrm{r}}_{16}{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{32})$$ 3

$$E_{\mathrm{o}\mathrm{v}\mathrm{f}}=\mathrm{E}({\mathrm{S}\mathrm{r}}_8{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{23})+\frac{1}{2}\mathrm{E}({\mathrm{O}}_2)-\mathrm{E}({\mathrm{S}\mathrm{r}}_8{\mathrm{F}\mathrm{e}}_7{\mathrm{M}\mathrm{O}}_{24})$$ 4

Here, E(O₂) denotes the total energy of an isolated oxygen molecule in the gas phase, computed in its spin-polarized ground state after full structural relaxation.

Following previous studies on O p-band centers, all materials were simulated as ferromagnetic to ensure a consistent and tractable set of calculations across the investigated oxides. ,, The local magnetic moments are taken from the difference between projected electron densities of up and down spins onto a sphere of 1 Å radius. Bader charge analysis was performed on the charge density files using the Pymatgen package.

The O p-band center has been calculated as the centroid of the projected density of states (PDOS) of oxygen, including occupied and unoccupied states (up to a maximum energy, E _max), according to eq , where E is the electron energy and D _Op (E) is the DOS projected onto the p orbitals of O atoms. The band center is referred to the Fermi level, which in the case of semiconducting oxides is usually set at the conduction band minimum. Note that bulk Op-band centers correlate, and can be used to describe, the surface O p-band centers ,

$$\mathrm{O}\mathrm{p}-\mathrm{b}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}=\frac{{\int }_{E_{\min }}^{E_{\max }}E{D}_{\mathrm{O}\mathrm{p}}(E)\mathrm{d}E}{{\int }_{E_{\min }}^{E_{\max }}{D}_{\mathrm{O}\mathrm{p}}(E)\mathrm{d}E}-E_{\mathrm{F}\mathrm{e}\mathrm{r}\mathrm{m}\mathrm{i}}$$ 5

The choice of an upper energy, E _max, is ambiguous because the quantity of unoccupied states present in the calculated PDOS depends on the system and the number of bands used in the calculation (Figure S1). A good criterion is to choose E _max considering how many of the unoccupied states are chemically relevant. Hence, a reasonable E _max cutoff would be the vacuum level since above that, the states are actually unstable with respect to the electron leaving the material. Since this is typically around 4 eV for perovskite oxides, , this value is adopted as E _max in the present work.

3. Results

In the general formula A _n+1B _n O_3n+1, the B–O framework is quasi-two-dimensional for low n values (n = 1 and n = 2), whereas it becomes fully three-dimensional as n → ∞. As dimensionality increases, the stronger orbital overlap leads to enhanced bond hybridization (i.e., covalency). ,, As previously noted, , the value of the O p-band centers gets closer to the Fermi level with enhanced covalency and, therefore, in the RP series, a shallower O-p band center is expected with increasing dimensionality. Figure a–c shows the calculated DOS for the RP-Sr _n+1Fe _n O_3n+1 phases, highlighting the contribution from Fe, O, and Sr states. The Fe and O states are strongly hybridized, forming a broad band that extends from −8 up to approximately 2 eV above the Fermi level. Consistent with the dimensionality argument, the extracted O p-band center for Sr₂FeO₄ (−2.96 eV) is below that of the SrFeO₃ perovskite (−2.85 eV).

2.

Calculated atom-projected density of states for Sr _n+1Fe _n O_3n+1 with (a) n = 1, (b), n = 2, and (c) n = ∞. The Fermi level is set as the zero of energy. Up spin (or majority) and down spin (or minority) contributions are shown. Color code: Fe: navy, O: red, and Sr: green. (d) O p-band centers of Sr _n+1Fe_7n/8M _n/8O_3n+1. Pink squares, open circles, and green triangles denote, respectively, the n = 1, n = 2, and n = ∞ members.

In simple ABO₃ perovskites, it is well-established that B–O bond covalency increases across the 3d series, leading to a progressively shallower O p-band center as the 3d electron count increases. , Building on this known trend, we investigated how partial substitution of Fe by other 3d-TMs affects the electronic structure of Sr _n+1Fe _7n/8M _n/8O_3n+1. Results show that the expected increase in B–O covalency across the 3d series is preserved (Figure d). For a given TM, when the effect of dimensionality is considered, the n = 1 RP members systematically exhibit a deeper O p-band center than their perovskite (n = ∞) counterparts. However, the n = 2 compositions display a distinct and unexpected deviation from this dimensionality-driven trend. Rather than presenting O p-band-center values intermediate between those of the n = 1 and perovskite phases, the n = 2 members lie above the perovskite values for early 3d-elements, while for late 3d-elements, they more closely resemble the n = 1 compounds. This anomalous behavior disappears when the E _max in eq is extended to include higher-energy unoccupied states (Figure S2). Furthermore, the calculated Bader charges for Fe and O in the n = 2 compositions fall between those of the n = 1 and n = ∞ phases (Figure S3). A comprehensive mechanistic understanding of this distinctive regime for the n = 2 members lies beyond the scope of the present work and will be addressed in a dedicated follow-up study.

Oxygen nonstoichiometry is of paramount relevance for the catalytic properties of TM oxides, as it directly governs their defect chemistry, electronic structure, and oxygen transport. ,, Regarding defects energetics, perovskite SrFeO₃ is typically obtained in a nonstoichiometric form (SrFeO_3−δ), while stoichiometric Sr₂FeO₄ can be more easily synthesized. ,,, Accordingly, in this work, the calculated energy of vacancy formation for SrFeO₃ is of 1.97 eV, consistent with previous reports, ,, whereas that of Sr₂FeO₄ raises to 2.6 eV. These benchmarks serve as a reference for the subsequent exploration of oxygen-vacancy formation in the Fe-substituted n = 1 and n = ∞ series.

For Sr₂FeO₄, calculations indicate that the removal of equatorial oxygen atoms is energetically more favorable than that of apical oxygen, by approximately 0.5 eV per formula unit. Therefore, only equatorial oxygen sites are considered for the investigation in M-substituted Sr₂Fe_7/8M_1/8O₄ oxides. These sites are illustrated in Figure a: O1, shared between M and Fe, and O2, coordinated exclusively to Fe. In SrFe_7/8M_1/8O₃ perovskites, there are also oxygen sites shared between Fe and M (labeled as O1) and oxygen sites not bonded to M. For the latter, two distinct chain motifs can be distinguished, O_vac–Fe–O–Fe and O_vac–Fe–O–M, which give rise, respectively, to the oxygen vacancies labeled as O2 and O3 in Figure c. The calculated oxygen-vacancy formation energies for all compounds are positive (Figure a), indicating that their reduction is not thermodynamically favored at 0 K.

3.

(a) Calculated energy of oxygen-vacancy formation according to the reactions (3) and (4) for Sr₂Fe_7/8M_1/8O₄ (pink symbols) and SrFe_7/8M_1/8O₃ (green symbols). The different oxygen sites are indicated by asterisks (O1), triangles (O2), and circles (O3). (b) Calculated energy of oxygen-vacancy formation as a function of the O p-band center for Sr₂Fe_7/8M_1/8O₄ (pink symbols) and SrFe_7/8M_1/8O₃ (green symbols) for the different O-sites. (c,d) Calculated Bader charges for Fe/M ions of Sr₂Fe_7/8M_1/8O₄ (filled pink) and SrFe_7/8M_1/8O₃ (filled green), compared to their respective oxygen-deficient phases Sr₂Fe_7/8M_1/8O_3.875 (open pink) and SrFe_7/8M_1/8O_2.875 (open green).

Building on the analysis of the distinct oxygen sites and their local environments, the calculated energies reveal clear trends in oxygen-vacancy formation. It is observed that the formation energy is significantly lower in perovskites than that in the n = 1 RP counterparts, highlighting the role of dimensionality in defect energetics. Regarding compositional effects, a clear trend emerges across the 3d-series, with the propensity for oxygen-vacancy formation increasing from early to late TMs by as much as 0.8 eV. This tendency is primarily driven by the progressive weakening of the M–O bond across the series, coupled to the stability of the lower oxidation states of the TM cations, which energetically favor vacancy formation. The differences in the formation energies for the distinct oxygen sites for a given TM and structure range from a minimum of 0.05 eV for Co (n = 1) to a maximum of 0.62 eV for Sc (n = ∞). In general, consistent with expectations based on the strength of the Fe–O and M–O bonds being broken, the most favorable sites for early TMs are the O2 or O3 sites, which involve the breaking of two Fe–O bonds, whereas for late TMs, the O1 site, involving the breaking of one Fe–O and one M–O bond, becomes more favorable. Additionally, for SrFe_7/8Cu_1/8O₃ and SrFe_7/8V_1/8O₃ perovskites, the tendency of Cu and V to adopt 5-fold coordination may further promote oxygen-vacancy formation at the O1 site. These general trends are further confirmed by the calculated formation energies for late 3d TMs, which are in good agreement with those of previous works. Specifically, the reported formation energy of SrFeO₃ decreases from 2.0 to 0.90 eV in Cu-doped SrFe_0.725Cu_0.25O₃, and from 2.1 to 1.85 eV and 1.7 eV, in SrFe_7/8Co_1/8O₃ and SrFe_7/8Ni_1/8O₃, respectively. Furthermore, the previously observed correlation between oxygen-defect formation energies and O p-band centers in ABO₃ perovskites and (La_1–x Sr _x )₂MO_4±δ (M = Mn, Co, Ni, and Cu) is also noted in the present case (Figure b). In general, oxides with shallower O p-band centers exhibit a higher propensity for oxygen-vacancy formation.

Oxygen vacancies induce an electronic density redistribution that leads to the reduction of cations to lower oxidation states. In both Sr₂Fe_7/8M_1/8O₄ and SrFe_7/8M_1/8O₃, the formal oxidation state at the B site (i.e., for Fe and M) is +4 in the stoichiometric oxides and +3.75 in the oxygen-deficient models. To assess the effect of oxygen vacancies, Figure c,d compares the calculated Bader charges of M and Fe for the most stable structures with vacancies (open squares) and without vacancies (filled squares). The general trend is that introducing oxygen vacancies decreases the Bader charge on Fe, indicating Fe reduction, with the effect being more pronounced in the perovskite structure. Previous studies have debated whether, in SrFeO_3−δ, the reduced Fe ions reside in square-pyramidal or octahedral coordination environments. , In the present calculations for perovskites, the reduction is localized on the 5-fold-coordinated Fe ions. As expected, for the late 3d elements, the charge-density redistribution affects both Fe and M, consistent with the tendency of these TM to adopt lower oxidation states. The SrFe_7/8V_1/8O_3−δ perovskites are an interesting case. In the stoichiometric phase, the calculated magnetic moments indicate the presence of V⁵⁺ ions (Figure S3d), while Fe exhibits the most reduced character among the perovskite series (Figures S3 and c). Upon oxygen-vacancy incorporation, the formation of V-centered square pyramids leads to a substantial reduction of V⁵⁺ (Figure d) and, concurrently, the smallest change in Fe charge compared to that of the other perovskites (Figure c).

Figure a shows the O p-band centers extracted for the most stable oxygen-deficient models. As in the stoichiometric oxides, increasing dimensionality shifts the O p-band center closer to the Fermi level. In addition, when compared to the stoichiometric phases, the oxygen-deficient oxides exhibit a clear downshift of the O p-band center (Figures a and S4–S5). This behavior can be rationalized by the reduction of the B-site cations induced by oxygen-vacancy formation, which diminishes the Fe-3d/O-2p hybridization, reduces the covalent character of the Fe–O bond, and consequently lowers the O p-band center relative to the Fermi level. ,

4.

(a) The O p-band centers of Sr _n+1Fe_7n/8M _n/8O_3n+1−δ extracted using E _max = 4 eV in eq of the main text. Filled symbols denote the oxygen-stoichiometric compounds, while open symbols correspond to their oxygen-deficient counterparts (n = 1 pink squares, n = ∞ green triangles). For the oxygen-deficient phases, only the O p-band center associated with the most stable oxygen-vacancy configuration is shown. (b) Range of O p-band center values shown in panel (a), with bars indicating the maximum and minimum variation due to structure (purple), oxygen content (olive), and TM (blue).

4. Discussion

In Sr _n+1Fe _7n/8M _n/8O_3n+1−δ, both structure (dimensionality, n) and composition (nature of M and oxygen content) influence the O p-band center. From an application standpoint, bringing the O p-band center closer to the Fermi level enhances catalytic activity, as reflected by lower area-specific resistance (ASR) in high-temperature ORR measurements in solid oxide systems and by reduced overpotentials for the OER in aqueous electrolytes. ,,, Following this guiding principle, a quick inspection of Figure a shows that the most active Sr _n+1Fe _7n/8M _n/8O_3n+1−δ compounds are perovskites containing late 3d-TM with a high oxygen content, whereas the least active materials correspond to the n = 1 phases incorporating early TMs and oxygen deficiency. Across the full data set, the O p-band center spans from −2.6 eV for Sr₂Fe_7/8Zn_1/8O₃ (an unrealistic material) to −3.3 eV for Sr₂Fe_7/8V_1/8O_3.875, yielding a total variation of ∼0.7 eV. Figure b disentangles how dimensionality and composition jointly shape this band-center landscape.

For dimensinality, the structural change from n = 1 to perovskite generally induces a variation in the band center, on the order of ∼0.2 eV (e.g., SrFe_7/8Ti_1/8O_2.875 vs Sr₂Fe_7/8Ti_1/8O_3.875), presuming a higher activity as dimensionality increases. Benchmarking with experiments is difficult, since systematic investigations of dimensionality are scarce in the literature, in part due to the difficulties in the synthesis of the single phase for some RP phases. ,− Cao et al. succeeded to prepare La _n SrNi _n O_3n+1 (n = 1, 2, 3, ∞), where Ni adopts a formal oxidation state of +3. These nickelate oxides displayed a remarkable enhancement in the OER activities as the dimensionality increased with n. While this supports the general trend observed in Figure , it should be remarked that for some oxygen-deficient compounds (e.g., SrFe_7/8Ni_1/8O_2.875 vs Sr₂Fe_7/8Ni_1/8O_3.875), there is no difference in O p-band center between n = 1 and n = ∞. Therefore, one cannot categorically claim that perovskites are intrinsically more active than n = 1 phases; intrinsic activity depends critically on chemical composition. Indeed, according to the O p-band centers, the oxygen-stoichiometric n = 1 phases with M = Cu or Ni could outperform the intrinsic activity of oxygen-deficient perovskites with M = Ti or V.

The DFT results evidence that perovskites form oxygen vacancies more readily than their n = 1 counterparts (Figure a,b). Oxygen vacancies are crucial for the ORR at high temperature, as they promote faster oxygen adsorption, dissociation, and incorporation into the lattice, thereby enhancing overall electrocatalytic performance. The greater tendency of perovskites to accommodate oxygen vacancies, as revealed by the present DFT calculations, is consistent with prior experimental reports showing that oxygen-deficient perovskites often display enhanced catalytic activity, albeit accompanied by larger thermal expansion and reduced structural stability compared to RP phases. , Moreover, in perovskites with high oxygen-vacancy concentrations and strong metal–oxygen covalency, lattice oxygen might directly participate in O₂ evolution from alkaline electrolytes, in the so-called lattice oxygen mechanism (LOM). In this mechanism, which operates alongside the conventional adsorbate evolution mechanism (AEM), water reacts with lattice oxygen instead of metal sites, and oxygen vacancies are formed while O₂ is generated. ,, At the same time, it is important to note that the LOM is not universal for perovskites; its activation depends on the nature of the TM and oxygen content. To date, only a few catalysts have been reported to adhere to the LOM, and in particular, Sr _x Ca_1–x FeO_3−δ perovskites are found to evolve oxygen predominantly via the conventional AEM. Figure a,b reveals that the LOM in iron-based perovskites could be enhanced if Fe is partially replaced by more electronegative TM elements. Interestingly, some n = 1 members containing late TM elements exhibit competitive oxygen-vacancy formation energies, comparable even to some perovskites. This suggests that the LOM pathway may also be activated in low-dimensional RP terms, as has indeed been demonstrated for RP-Sr₃(Co_0.8Fe_0.1Nb_0.1)₂O_7−δ and La_0.5Sr_1.5Ni_1–x Fe _x O_4±δ.

As shown in Figure , oxygen content is a key determinant of the electronic structure, shifting the O p-band center by up to 0.44 eV in some cases (e.g., SrFe_7/8V_1/8O₃ vs SrFe_7/8V_1/8O_2.875), while in others, its effect is minimal (e.g., Sr₂Fe_7/8Cr_1/8O₄). Likewise, the identity of M can vary the band center by as much as 0.44 eV (SrFe_7/8V_1/8O_2.875 vs SrFe_7/8Cu_1/8O_2.875) and as little as 0.01 eV (e.g., Sr₂Fe_7/8Cr_1/8O₄ vs Sr₂Fe_7/8Mn_1/8O₄). Importantly, these factors are interdependent as the nature of M largely dictates the achievable oxygen content. Cations stable at low oxidation states (e.g., Co²⁺, Cu²⁺) rarely form stoichiometric perovskites, whereas high-valence cations (e.g., V⁵⁺, Mo⁶⁺) favor perovskites with fewer oxygen vacancies.

It is legitimate to question whether the variations of the O p-band center identified in this work have quantitative repercussions on catalytic activity. To address this, one can examine the correlations reported between DFT-O p-band centers and measured catalytic activities in the OER and ORR. It is important, however, to emphasize that catalytic activity measurements inevitably reflect a convolution of intrinsic catalyst properties and extrinsic factors such as the morphology, microstructure, and electrode architecture. Consequently, the following discussion should be interpreted primarily in terms of semiquantitative trends rather than absolute catalytic metrics. For the OER in alkaline solution, although the O p-band center is a well-established electronic descriptor and numerous perovskite oxides have been studied experimentally, a direct and systematic correlation between this descriptor and measured overpotentials remains largely unexplored. Nonetheless, in the study of (Ln_0.5Ba_0.5)­CoO_3−δ double perovskites, a shift of the O p-band center by ∼0.5 eV corresponds to a modest change of ∼0.07 V in overpotential. A similar O p-band center shift upon fluorine incorporation in La_0.5Ba_0.5Co_3–x F _x (x = 0, 0.1) translates into a measured OER overpotential difference of 0.05 V at 100 mA/cm². More combined experimental and computational investigations are needed to assess the relationship between the O p-band center and catalytic activity in aqueous solutions.

In contrast, at high temperature, a more comprehensive correlation has been demonstrated and a database covering dozens of perovskites shows a statistically significant linear relationship (R ² ≈0.86) between the bulk O p-band center and the logarithm of the surface exchange coefficient (k*). − According to the linear relationship reported by Morgan et al., the O p-band centers observed for realistic materials in our systems (Table S2) would yield only moderate variations of less than two factors in k*. While such variations are not negligible in principle, their impact becomes minor once extrinsic factors such as porosity, particle size, layer thickness, surface reconstruction, and other morphological parameters are considered. Moreover, when RP oxides are employed as very thin layers or films, substrate-induced effects such as epitaxial strain or interfacial interactions may further influence the catalyst performance, an aspect not explicitly addressed in the present bulk DFT analysis. These extrinsic factors are critical in solid oxide cell devices (fuel cells, SOFCs; and electrolyzers, SOECs) working at intermediate to high temperatures with gases as reactants and products. Indeed, several studies report quantitative improvements due to microstructural and morphology optimization. It is very common in the literature to report the catalytic properties of composite electrodes consisting of the material under study and electrolyte in different proportions. In addition, various strategies are used both to synthesize the materials and to prepare the composites to ensure high specific surface area and high porosity. This maximizes the triple phase boundary (TPB) (contact surface between O₂ molecules, electrocatalytic material, and electrolyte through which oxide anions must diffuse) where the ORR and OER take place. Sophisticated strategies can be employed, such as the construction of electrodes by single-step spray pyrolysis with simultaneous deposition of the catalyst and electrolyte and even decoration with metal nanoparticles. ,,−

Taken together, the above examples suggest that in the Sr _n+1Fe _n(1–x)M _nx O_3n+1−δ series (n = 1, 2, ∞; M = 3d metal; x = 1/8; δ = 0,1/8), extrinsic factors can induce catalytic activity enhancement comparable to, or even larger than, those expected from electronic tuning alone (e.g., modest shifts in the O p-band center). Therefore, while compositional tuning and its effect on the O p-band center may contribute to performance of the extensively studied Sr–Fe–O-based RP series, morphology, microstructure, and electrode conformation are likely to play an equal, or even dominant, role.

In the present study, the focus is on bulk electronic descriptors for catalytic activity, while operando surface evolution and long-term stability remain open questions for further investigation. Future work could extend the present DFT analysis by considering the behavior of these oxides under realistic OER/ORR conditions, where surface reconstruction or partial amorphization may influence the structure of the catalytically active region. Moreover, the O p-band center represents only one of several relevant descriptors of electrocatalytic activity. , Correlating this bulk descriptor with surface-specific quantities, such as the adsorption energies of key OER/ORR intermediates, and performing explicit computational surface studies, therefore represents an important avenue for future work to better assess the catalytic potential of this family of oxides.

5. Conclusions

Systematic evaluation of the O p-band center of the RP-Sr _n+1Fe_7n/8M _n/8O_3n+1−δ series provides insight into their catalytic potential and enables predictive guidelines for designing more efficient oxygen electrocatalysts. Both 12% Fe-substitution by late 3d-TM and increasing dimensionality (n → ∞) enhance B–O covalency, yielding shallower p-band and often lower oxygen-vacancy formation energies. The oxygen content can shift the O p-band center by up to 0.44 eV, comparable to the effect of Fe-substitution by 12% TM, showing that oxygen stoichiometry is a critical element of composition–activity relationships. Because the nature of the substituent –TM dictates the achievable oxygen stoichiometry, electronic tuning via B-site substitution cannot be decoupled from defect chemistry.

Defect energetics further distinguish between n = 1 and perovskite structures. Perovskites form oxygen vacancies more readily, suggesting enhanced ORR/OER activity. Noteworthily, certain n = 1 compositions with late 3d metals exhibit vacancy formation energies close to those of perovskites, indicating that layered RP phases can also reach vacancy-rich regimes compatible with LOM pathways.

The electronic trends align with the heuristic that shallower p-band O centers favor enhanced electrocatalytic activity; however, literature correlations indicate that the ∼0.7 eV range spanned by the entire RP system yields only modest activity gains. These improvements can easily be overshadowed by extrinsic factorssuch as morphology, microstructure, porosity, and surface areathat may induce equal or larger changes in measured performance. Overall, dimensionality, partial Fe-substitution, and oxygen content provide meaningful but moderate intrinsic levers for tuning the activity in Sr–Fe–O-based RP materials. These findings underscore that rational catalyst design must integrate electronic-structure engineering with careful morphological and microstructural control to unlock the activity in oxygen electrocatalysis.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemmater.5c03458.

• Details of the computational method, analysis of O p-band centers, calculated Bader charges, local magnetic moments, and DOS (PDF)

The authors declare no competing financial interest.

Published as part of Chemistry of Materials special issue “Bifunctional Electrocatalysts”.