Tian
Xia
*a,
Ziwei
Dong
a,
Yingnan
Dou
a,
Qiang
Li
a,
Jingping
Wang
b,
Liping
Sun
a,
Ruiping
Deng
c,
Lihua
Huo
a and
Hui
Zhao
a
aKey Laboratory of Functional Inorganic Materials Chemistry, Ministry of Education, School of Chemistry, Chemical Engineering and Materials, Heilongjiang University, Harbin 150080, Heilongjiang, P. R. China. E-mail: xiatian@hlju.edu.cn
bKey Laboratory of Superlight Materials and Surface Technology, Ministry of Education, College of Materials Science and Chemical Engineering, Harbin Engineering University, Harbin 150001, Heilongjiang, P. R. China
cState Key Laboratory of Rare Earth Resource Utilization, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, Jilin, P. R. China
First published on 17th September 2025
To realize a sustainable and clean society, highly efficient and eco-friendly energy conversion and production technologies have been developed, such as solid oxide cells (SOCs) and water electrolysis. However, most electrocatalytic reactions are normally driven by scarce noble metal-based catalysts. Thus, the exploration of active, stable, and cost-effective electrocatalysts remains an essential challenge. Herein, we summarize a series of layered perovskite oxides (LnBaCo2O5+δ) demonstrating attractive oxygen reduction, oxygen evolution, and hydrogen evolution electrolysis activities. Their variable crystal structures, flexible elemental compositions, and regulated electronic structures endow them with satisfactory activity for electrocatalytic water electrolysis and SOCs (fuel cells and electrolysis cells). This review summarizes the advances in LnBCO-based electrocatalysts and addresses several issues in their development and usage, including structural evolution, fundamental electrical properties, electrocatalytic mechanism and performance, structure–property–performance relationship, activity descriptors, and real merits/demerits. Finally, we outline the opportunities and challenges of layered perovskite oxides for practical applications, encouraging future research on next-generation electrocatalysts applied in an environmentally friendly society.
The produced H2 is fed to SOFCs, which directly convert chemical potential to electric energy. Representative oxygen-conducting SOFCs (O2−-SOFCs) consist of a porous anode, dense electrolyte, and porous cathode with a sandwich-like configuration, as schematically illustrated in Fig. 1a. The cathode side is fed with an oxidant gas, such as O2 and air. The dissociated O2− ions are transported from the cathode side to the anode side through the dense electrolyte (oxygen–ion conductor). The O2− ions are finally oxidized to H2O or CO2 when reacting with the fuel gas (H2 or CO) at the anode side, while releasing electrons through an external circuit.23–25 Utilizing the reverse reactions, H2O or CO2 can be decomposed into H2/CO fuel and O2 oxidant, i.e., solid oxide electrolysis cells (SOECs) (Fig. 1b).26–28 The gas steam (H2O or CO2) is supplied to the cathode side, and then the dissociated O2− ions are transported to the anode through the electrolyte component, driven by an external voltage (>Nernst voltage). The O2− ions finally evolve into O2 molecules at the anode, accompanied by electron release. As electrochemical devices, SOECs have a number of advantages, as follows: (1) producing hydrocarbon fuel from CO2 or H2O; (2) their reaction process is easy to control by regulating the voltage and reaction temperature; and (3) clean energy sources can be used to drive the process, such as solar, wind, geothermal, tidal, as well as surplus electricity from nuclear and hydroelectric sources. Differing from SOFCs, the high efficiencies of SOECs are attributed to their high operating temperatures. The total energy demand is almost constant from room temperature to 1200 °C for both H2O and CO2 electrolysis. The electrical energy decreases with an increase in temperature, and thus the energy difference is compensated for by increasing the heat supply. Because of the lower cost of heat than electricity, high-temperature electrolysis is more economically favored. Furthermore, increasing the operating temperatures can ensure a sufficient electrode performance from a dynamic perspective. There is a reasonable prospect that the conversion efficiency may exceed 60% for SOECs under desirable operating conditions. To enhance the efficiency of SOECs, studies have been conducted to develop electrode materials. The surface atoms in the cathode and anode act as active sites for the electrocatalytic reactions, the reduction of CO2/H2O and the oxygen evolution, respectively.
![]() | ||
| Fig. 1 Schematic of the operating principles of (a) solid oxide fuel cell (H2 is used as the fuel) and (b) solid oxide electrolysis cell (H2O electrolysis). | ||
Nowadays, simple perovskite oxides (ABO3) have been evaluated as potential electrocatalysts for SOFCs, SOECs, and water electrolysis,12–14,29–33 such as La(Sr)Co(Fe)O3−δ (LSCF) and Ba0.5Sr0.5Co0.8Fe0.2O3−δ (BSCF).34,35 The following aspects should be emphasized for an ideal electrocatalyst: (1) sufficient electrical conductivity (ionic, electronic, or mixed conductivity) for the charge transfer reaction; (2) effective active sites for the adsorption, dissociation, and desorption of reactants; (3) excellent long-standing working stability and durability under rigorous practical conditions; and (4) reasonable production cost. In the past two decades, much interest has been paid to a series of layered perovskite oxides, LnBaCo2O5+δ (LnBCO, Ln = lanthanide and Y).36–42 Owing to the difference in ionic radius between Ln3+ and Ba2+, LnBCO crystallizes into A-site cation-ordered structures, following the sequence of 〈Ln–Oδ〉–〈Co–O〉–〈Ba–O〉 along the c-axis direction. Naturally, intrinsic oxygen vacancies are formed in the 〈Ln–Oδ〉 layers, apparently influencing electrocatalytic oxygen-containing reactions. A cubic–tetragonal–orthorhombic phase transition can be identified by regulating the A-site Ln3+ ions, accompanied by oxygen vacancy disorder–order transformation. Both the average cobalt state and oxygen content increase due to an increment in the oxygen-coordinating number of Ln3+, whereupon charge transfer, oxygen surface, and bulk transport processes are improved.37,43 Initially, LnBCO can be used as promising oxygen electrodes for SOFCs, with certain intermediate-temperature ORR activity. It should be noted that their ORR activities are higher than that of the LSCF electrode and comparable to the popular BSCF electrode. In recent years, LnBCO has shown multi-functional electrocatalytic characteristics, e.g., OER and HER activity in alkaline media.44–51 These property diverse properties are attributed to the distinctive structures of layered perovskite with various elemental compositions. LnBCO belongs to a large material library concerning partial cation substitution, anion doping, and elemental stoichiometry, where its A-site comprises lanthanide and alkaline-earth elements, and its B-site almost covers the vast majority of transition metals in the periodic table. Also, its structural and compositional selectivity render a chance for different application demands. More importantly, recent research on perovskite materials unveils some crucial facts for electrocatalysis, such as the correlation between electronic structure and catalytic performance, catalytic mechanisms, and structural evolution. Hence, we deem that layered perovskite oxides (LnBCO) can be employed as perfect electrocatalytic models for strengthening scientific theories.
To date, systematic overviews of “112”-type layered perovskite electrocatalysts have been lacking. Most of the relevant summaries introduce their progress as classified by contained lanthanides (e.g., La, Pr, and Nd-based layered perovskite oxides),36 properties (e.g., electrical conductivity, dilatometric behavior, cathode performance, and chemical stability),36,37 design strategies (e.g., doping, defect engineering, nanostructure, and surface modification),39 and potential applications (e.g., energy and environmental applications).39 However, some scientific questions related to layered perovskite oxides in the electrocatalysis field are rarely summarized, which is required to guide rational prediction and future development directions.
In this review, the advances in a family of layered LnBCO perovskite oxides are systematically discussed toward SOCs and water electrolysis. Firstly, we outline their structural evolution under internal and external conditions, which ensures that LnBCO electrocatalysts exhibit controllable properties. Secondly, we interpret how to regulate their electrical properties (electrical conductivity, surface exchange, and bulk diffusion processes), and then focus on their electrocatalysis mechanism and progress. When exploring the structure–property–performance relationship of the perovskite oxides, their representative activity descriptors (e.g., eg orbital occupancy, O 2p-band center, electronegativity, tolerance factor, charge transfer energy, and Lewis acidity), applicability and inadequacies, are studied. Subsequently, from the perspective of practical applications, their cost, stability, and scalability are compared to that of noble-metal-based catalysts. We highlight their merits and demerits for their rational utilization in definite electrochemical devices and technologies. Finally, their opportunities, challenges, and applicable prospects in the electrocatalysis field are proposed. We hope that this review will contribute to the development of next-generation electrocatalysts for readers.
With an increase in the ion radius of Ln3+, the crystal structure transforms from orthorhombic to tetragonal, and to cubic perovskite with order–disorder transition of oxygen vacancies, including orthorhombic structure (space group Pmmm) for Ln = Sm, Eu, and Gd, tetragonal structure (space group P4/mmm) for Ln = Pr and Nd, and cubic structure (space group Pm
m) for Ln = La.37,43,44 The nonstoichiometry (δ) values are ∼0.5, 0.75, and 1.0 for orthorhombic, tetragonal, and cubic structures (Fig. 2a), respectively. Moreover, the experimental conditions of the synthetic route control the superlattice perovskite structure, such as the surrounding atmosphere, calcining temperature, and annealing time. In the case of larger lanthanides (i.e., from Pr to Tb), doubling of one parameters is observed (an ap × 2ap × 2ap supercell, and ap refers to the lattice parameter of cubic perovskite), whereas in the case of smaller lanthanides (i.e., Ho and Dy), tripling of two parameters is obtained (a 3ap × 3ap × 2ap supercell).41 These crystal data are deduced from air-synthesized perovskite oxides. However, annealing the air-prepared GdBCO (δ = 0.4) in an Ar flow leads to an ap × ap × 2ap superlattice (δ = 0). When reversibly calcining the Ar-annealed GdBCO in an O2-rich atmosphere, the oxygen atoms re-enter the lattice, leading to GdBCO (δ = 0.6) with an ap × 2ap × 2ap supercell. Similarly, annealing of HoBCO (δ = 0.3) in O2 leads to HoBCO (δ = 0.5), and the 3ap × 3ap × 2ap supercell is replaced by an ap × 2ap × 2ap one. In the Mn-doped GdBaCo2−xMnxO5+δ series, GdBaCo1.8Mn0.2O5+δ can be indexed to a tetragonal unit cell of ap × ap × 2ap (P4/mmm). In contrast, a high doping fraction induces a cubic disordered perovskite of GdBaCo0.2Mn1.8O5+δ, and the structures of all the Ar-prepared compounds are similar to that of GdBaCo1.8Mn0.2O5+δ. Under an Ar atmosphere, single-phase GdBaCo2−xMnxO5+δ (0 ≤ x ≤ 2) materials are obtained, while the GdBaCo2O5+δ composition is only obtained in Ar.52 To characterize the A-site cation-ordered structures of layered perovskites, the X-ray diffraction (XRD) technique is employed as well as Rietveld refinement analysis. Multiplet characteristics (peak splitting) are typical for layered perovskite oxides, and the doubling of the unit cell is also confirmed by the small peak at 2θ ≈ 11–12°. It should be noted that this diffraction intensity is very weak for the larger Ln3+ ions, especially Pr3+ and Nd3+. Advanced in situ neutron powder diffraction (NPD) is a more powerful tool to accurately probe crystal structures. The NPD patterns of PrBCO can be fit to a tetragonal layered perovskite structure (P4/mmm), agreeing well with the XRD results.53,54 According to high-resolution transmission electron microscopy (HRTEM) images and selected area electron diffraction (SAED) patterns, we visually observe the supercells of layered perovskite oxides. As studied in PrBa0.95Gd0.05Co2O5+δ, the clear fringes with an interplanar spacing of 0.763 nm are assigned to the (001) plane of tetragonal layered perovskite oxides, and the superlattice diffraction spots are found in the SAED pattern along the [110]P zone axis. Atomic-level energy-dispersive X-ray (EDX) elemental mappings manifest an ap × ap × 2ap supercell, as evidenced by the doubling of the lattice parameter c along the [001]p zone axis.55 The specific crystal structures of perovskite oxides influence their properties. In layered perovskite oxides (LnBCO), oxygen ions migrate through the 〈Ln–Oδ〉 and 〈Co–O〉 planes (ab-plane oriented),56–58 which is conductive to understanding their ionic transport mechanism.
![]() | ||
| Fig. 2 (a) Schematic of the structural evolution of LnBCO perovskite oxides: orthorhombic (Ln = Sm, Eu, and Gd), tetragonal (Ln = Pr and Nd), and cubic (Ln = La) structures with a vacancy ordered-disordered transition. (b) Conductivity (600–800 °C) and (b) k* and D* values (800 °C) of the LnBCO perovskite oxides, along with a structural evolution from orthorhombic to tetragonal, and to cubic structures. The k* and D* data of classic LSFC perovskite oxide are shown in Fig. 2b for comparison. | ||
Another method to induce structural evolution in layered perovskite oxides is ionic substitution. Lanthanide/alkaline earth and transition metal ions can be introduced into their A- and B-sites, respectively. Considering the B-site substitution, Zn2+, Ni2+, Fe3+, Sc3+, Mn3+, Zr4+, Nb5+, Ta5+, and Mo6+ have been substituted for Co3+.59–73 Commonly, the B-site Co3+ ions prefer to be replaced by Fe3+ ions due to their similar ionic radii, forming a complete solid solution in a doping fraction range of 0–2.67 At a low-doping level, the solid solutions possess tetragonal layered perovskite structures (P4/mmm), whereas a structural evolution from tetragonal perovskite to cubic perovskite is identified with an increase in the Fe content. As displayed in high-temperature XRD patterns, Fe-doped PrBCO is thermodynamically stable upon heating at 1000 °C, which is beneficial for practical long-term operation.74 Moreover, Fe ion substitution induces a variation in the supercell of layered perovskite oxides. Compared with air-prepared GdBCO (an ap × ap × 2ap supercell), GdBaCo1.4Fe0.6O5+δ shows an ap × 3ap × 2ap superstructure.75 In addition to the Bragg reflections of cubic perovskite, the extra reflections at Gp ± 1/2(001)p and Gp ± 1/3(010)p in the SAED patterns confirm the 3ap × 2ap periodicity. The contrast difference reveals the 2ap and 3ap periodicities along the [001]p and [010]p directions, respectively. The novel perovskite-related supercell of √2ap × √2ap × 10ap is observed in (GdBa)0.8Ca0.4Co0.6Fe1.4O5.41.76 The oxygen content, oxygen-vacancy concentration, and average Co valence state in these supercells can be modulated by controlling the crystal structures, further inuencing electrocatalysis reactions.
A-site alkaline earth ionic substitutions lead to the formation of a highly symmetric cubic structure. After substituting Ca2+ for Ba2+ in PrBCO, a tetragonal–cubic phase transition appears with an increase in the doping fraction, i.e., from a tetragonal structure (x = 0–0.2) to a cubic structure (x = 0.4). Mixed tetragonal and cubic perovskite oxides coexist at x = 0.3.77 Similar structural evolutions are also discovered in Ba-site Sr-doped LnBCO.78–80 Noticeably, these mixed perovskite oxides possess improved electrocatalytic activities arising from the synergistic effects between them.78 We put forward an effective Ca/Sr co-doping strategy, demonstrating a high-performance cubic perovskite oxide (Pr0.94Ba0.6Sr0.2Ca0.2Co2O5+δ) as an oxygen electrode for SOFCs.81 However, in the case of individual Sr doping, tetragonal layered perovskites (NdBa1−xSrxCo2O5+δ and SmBa1−xSrxCo2O5+δ) can be retained even at x = 0.5, which is attributed to the similar ionic radius between Ln3+ and Ca2+.79,80
In general, the structural evolution of LnBCO can be explained from the viewpoint of oxygen content. Taking NdBCO as an example, an orthorhombic layered perovskite oxide (Pmmm) is obtained after calcining at 1200 °C.82 If annealing at 1100 °C in air, NdBCO has a tetragonal ap × ap × 2ap supercell (P4/mmm).83 The lattice oxygen atoms are released with a decrease in oxygen content at elevated temperature. The (212) diffraction peak is split into (142) and (124) peaks, corresponding to an orthorhombic layered perovskite oxide. More lattice oxygen atoms prevent the formation of a vacancy-ordered structure, resulting in tetragonal or cubic perovskite oxides. Surrounding air, O2, and Ar atmospheres also tune the structural evolution by injecting or releasing oxygen atoms. Cation-ordered/disordered perovskite structures can be switched by controlling their annealing temperature. A-site cation-ordered LaBCO is stable at 1000 °C in air, but is transformed into a cation-disordered phase at 1100 °C.84,85 Briefly, structural evolution is an essential and effective strategy for designing perovskite oxide electrocatalysts.
To accelerate the charge transfer kinetics, the substitution of Ca2+/Sr2+ for Ba2+ has been considered an effective strategy up to now. Recently, we found that Ca2+/Sr2+ co-doping extremely enhances the electronic conductivity, and improves oxygen surface exchange and bulk diffusion processes.81 The electrical conductivity reaches as high as ∼1100 S cm−1 at 500 °C. The ionic radius of Ca2+ (r = 1.34 Å) and Sr2+ (r = 1.44 Å) is much smaller than that of Ba2+ (r = 1.61 Å) in a 12-fold oxygen-coordinating environment, which relatively matches the ionic radius of Pr3+ (r = 1.126 Å).86 The suitability of ionic size hinders the formation of an A-site cation-ordered structure, with an increase in oxygen content and average Co valence state. More highly oxidative Co4+ species are generated in doped perovskite oxides. As the “3d” hole carriers, Co4+ ions are favorable for electron transfer. After introducing the Ca/Sr dopants, the O–Co–O bond angle is straightened to the ideal 180°, enlarging the bandwidth and Co–O covalency. This is another reason for the increased electrical conductivity of Ca2+/Sr2+-doped LnBCO. In terms of B-site transition metal doping, many choices have been made in LnBCO, such as Zn2+, Ni2+, Sc3+, Fe3+, Mn3+, Nb5+, Ta5+, and Mo6+.59–73 These substitutions inhibit the electronic hopping through the O–Co–O pathway, and the electrical conductivity is reduced. Even though B-site doping is adverse for the charge transfer reaction, a trace of Ta dopants can enhance the electrical conductivity, e.g., ∼234 and 675 S cm−1 for PrBa0.94Co2O5+δ and PrBa0.94Co1.96Ta0.04O5+δ at 700 °C,87 respectively. This enhancement is attributed to the charge carriers provided by the acceptor dopants and stabilized layered perovskite structure. Moreover, cation-defect engineering can also facilitate the charge transfer reaction. As is known, cobalt-based perovskite oxides belong to mixed ionic and electronic conductors (MIECs), in which the charge carriers are electronic holes and oxygen vacancies. Two compensating mechanisms, the generation of extrinsic oxygen vacancies and the oxidation of Co3+ to Co4+, can respond to negatively charged A-site defects. Regarding the oxidation mechanism, enhanced hole carriers (Co4+) lead to an increase in electronic conductivity. An oxygen vacancy-compensated mechanism is observed in A-site-deficient LnBCO (Ln = La, Pr, Nd, and Eu) perovskite oxides.84,88–91 Consequently, the electrical conductivity decreases because the oxygen-ionic mobility is much lower than the electronic mobility in MIECs. The electrical conductivity of layered perovskite oxides is listed for reference in Table 1. Anyway, the charge transfer kinetics is one of the most important aspects for electrocatalysis reactions.
| Conductivity (S cm−1) | k* (cm s−1) | D* (cm2 s−1) | TEC (10−6 K−1) | Ref. | |
|---|---|---|---|---|---|
| LaBaCo2O5+δ | ∼50 (700 °C) | — | — | — | 40 |
| PrBaCo2O5+δ | ∼145 (700 °C) | — | — | — | 40 |
| NdBaCo2O5+δ | ∼170 (700 °C) | — | — | — | 40 |
| SmBaCo2O5+δ | ∼200 (700 °C) | — | — | — | 40 |
| GdBaCo2O5+δ | ∼90 (700 °C) | — | — | — | 40 |
| YBaCo2O5+δ | ∼30 (700 °C) | — | — | — | 40 |
| LaBaCo2O5+δ | 936 (600 °C) | 2.58 × 10−3 (800 °C) | 4.73 × 10−6 (800 °C) | — | 43 |
| PrBaCo2O5+δ | 853 (600 °C) | 2.58 × 10−3 (800 °C) | 4.82 × 10−6 (800 °C) | — | 43 |
| NdBaCo2O5+δ | 603 (600 °C) | 2.15 × 10−3 (800 °C) | 3.28 × 10−6 (800 °C) | — | 43 |
| SmBaCo2O5+δ | 430 (600 °C) | 2.07 × 10−3 (800 °C) | 2.20 × 10−6 (800 °C) | — | 43 |
| GdBaCo2O5+δ | 360 (600 °C) | 3.05 × 10−4 (800 °C) | 2.93 × 10−6 (800 °C) | — | 43 |
| EuBaCo2O5+δ | 341 (600 °C) | 6.12 × 10−4 (800 °C) | 2.54 × 10−6 (800 °C) | — | 43 |
| PrBaCo2O5.5 | — | — | 1.9 × 10−8 (500 °C) | — | 56 |
| PrBaCo2O5+δ | — | 4.6 × 10−7 (675 °C) | 1.7 × 10−8 (675 °C) | — | 57 |
| NdBaCo2O5+δ | — | ∼2.78 × 10−4 (700 °C) | ∼3.7 × 10−5 (700 °C) | — | 70 |
| NdBaCo2O5+δ | — | — | — | 23.1 (25–1100 °C) | 83 |
| SmBaCo2O5+δ | — | — | — | 20.8 (25–1100 °C) | 83 |
| PrBaCo2O5+δ | ∼2000–500 (25–812 °C) | 6.9 × 10−5 (500 °C) | 3.6 × 10−7 (500 °C) | — | 95 |
| GdBaCo2O5+δ | — | 2.8 × 10−7 (575 °C) | 4.8 × 10−10 (575 °C) | — | 96 |
| LaBaCo2O5+δ | — | — | — | 24.3 (80–900 °C) | 98 |
| NdBaCo2O5+δ | — | — | — | 19.1 (80–900 °C) | 98 |
| SmBaCo2O5+δ | — | — | — | 17.1 (80–900 °C) | 98 |
| GdBaCo2O5+δ | — | — | — | 16.6 (80–900 °C) | 98 |
| YBaCo2O5+δ | — | — | — | 15.8 (80–900 °C) | 98 |
| PrBaCo2O5+δ | — | — | — | 24.03 (100–800 °C) | 161 |
| NdBaCo2O5+δ | — | — | — | 21.5 (80–900 °C) | 162 |
| GdBaCo2O5+δ | — | — | — | 19.9 (80–900 °C) | 162 |
| EuBaCo2O5+δ | — | — | — | 14.9 (100–800 °C) | 163 |
| YBaCo2O5+δ | — | — | — | 14.5 (200–800 °C) | 164 |
| SmBaCo2O5+δ | — | 3.5 × 10−3 (800 °C) | 5.4× 10−6 (800 °C) | — | 185 |
![]() | (1) |
![]() | (2) |
βm tan βm = Lx; γn tan γn = Ly; δp tan δp = Lz; | (3) |
| Lx = x(k*/D*); Ly = y(k*/D*); Lz = z(k*/D*) | (4) |
As a classic intermediate-temperature oxygen electrode for SOFCs, LSCF is characterized by excellent mixed conducting properties and electrocatalytic ORR activity. The electrical conductivity of LSCF is ∼200–300 S cm−1 at 600–900 °C,34 but the maximum value of LaBCO exceeds 1000 S cm−1.43 The convenient charge transfer in layered perovskite oxides might be ascribed to the hopping of d electrons from the high-spin Co3+ to low-spin Co3+ with itinerant electrons. However, the high-spin Co3+ states in LSCF can be stabilized by Fe ions, and thus the thermally activated hopping of d electrons becomes more difficult. The prominent oxygen surface exchange and bulk diffusion rates of LSCF manifest its ORR characteristics, i.e., k* = 1.2 × 10−3 cm s−1 and D* = 6.70 × 10−8 cm2 s−1 at 800 °C, respectively.34 It is noteworthy that the k* and D* values of layered perovskite oxides surpass that of LSCF, e.g., enhanced D* by two orders of magnitude (∼10−6 cm2 s−1) (Fig. 2c).43 The as-collected k* and D* values of layered LnBCO perovskite oxides are listed in Table 1. Their high electrical conductivities, k*, and D* can contribute to the electrocatalytic ORR process.
| O2(g) → O2,ads., molecular oxygen adsorption | (5) |
| O2,ads → 2Oads., dissociation of adsorbed oxygen | (6) |
![]() | (7) |
![]() | (8) |
To further promote the ORR activity, the most extensive strategy is ion substitution, including A, B, and O-site substitution. At the A-site in LnBCO, Ln3+ or Ba2+ can be replaced by other lanthanide or alkaline earth ions.76–80,101–127 When substituting Ca2+ or Sr2+ for Ba2+, the ORR activity is significantly promoted due to the enhanced electron transfer, as discussed in Section 3.1. Furthermore, Ca2+/Sr2+ co-doping is more effective for improving the ORR activity. We developed a Ca2+/Sr2+ co-doped Pr0.94Ba0.6Sr0.2Ca0.2Co2O5+δ oxygen electrode with the ASR of 0.025 Ω cm2 at 700 °C, approximately reduced by ∼60%.81 This A-site ion substitution is one of the best-performing methods for improving the electrocatalytic properties.
In the case of B-site ion substitution, a variety of transition metals can be considered as dopants, such as Zn, Fe, Mn, Cu, Ni, Sc, Nb, Ta, and Mo.59–73,128–155 It should be noted that A and B-site co-doping is more beneficial to the ORR activity.136,143,156–160 The synergistic effects between two types of dopants have been identified in layered perovskite oxides, LnBa0.5Sr0.5Co1.5Fe0.5O5+δ (Ln = La, Pr, and Nd) and Pr0.8Nd0.2BaCo1.6Fe0.4O5+δ, leading to oxygen vacancy-rich characteristics, fast oxygen surface processes, and excellent durability. A fabricated single cell with an LaBa0.5Sr0.5Co1.5Fe0.5O5+δ-based composite oxygen electrode delivered a high peak power density (PPD) of ∼2200 mW cm−2 at 600 °C,143 suggesting the commercial possibility of this co-doped electrode for SOFCs. However, this family of layered perovskite oxide has a drawback, i.e., thermal expansion coefficients (TECs) of ∼15–24 × 10−6 K−1 (Table 1).83,98,161–164 Their high TECs are ascribed to the configuration transition of Co3+ from low-spin (r = 68.5 pm) to intermediate-spin, and then to high-spin (r = 75 pm) filling states, resulting in an increased volume of CoO6 octahedrons and corresponding lattice expansion. Thereupon, B-site ion substitution can suppress this dilatometric behavior. For example, the TEC value of 30 mol% Ni2+-doped NdBCO decreases from 19.1 × 10−6 K−1 to 16.7 × 10−6 K−1 in the temperature range of 80–900 °C, with an ASR of ∼0.5 Ω cm2 at 700 °C.60 However, in most cases, although Co-site doping restrains high-temperature expansion, it has a negative effect on the ORR activity, for example, in GdBaCo2−xNixO5+δ,61 PrBaCo2−xFexO5+δ,64 PrBaCo2−xNbxO5+δ,68 YBaCo2−xFexO5+δ,71 and YBaCo2−xCuxO5+δ.165
Incorporating the electrolyte components into the electrode backbone can extend the TPB region (electrode–gas–electrolyte), boosting the ORR kinetics at the electrode interfaces.166–193Fig. 3b–d schematically illustrate the possible electrocatalytic ORR zones of pure electronic conductor (PEC), MIEC, and MIEC/electrolyte composite materials, respectively. If a PEC is used as the oxygen electrode, the reaction zone is confined to TPB (Fig. 3b). In the case of an MIEC oxygen electrode, the oxygen reduction process happens not only at TPB but also at the electrode/gas interface, i.e., two-phase boundary (2PB). The MIEC surface can also provide extra active sites for adsorbing molecular O2, and then the oxygen ions migrate from the MIEC to the electrolyte due to oxygen-ionic mobility of MIEC (Fig. 3c). With respect to the composite oxygen electrode (MIEC and electrolyte), the electrochemically active zone is extended due to the expanded TPB length (Fig. 3d). The sum of TPB dominates the electrochemical performance of the oxygen electrode, and thus composite electrocatalysts should be ideal candidates for SOFCs. The fluorite-type oxides Ce0.9Gd0.1O1.95 (GDC) and Ce0.8Sm0.2O1.9 (SDC) are used as dense electrolyte components. Actually, the addition of electrolyte hugely improves the ORR activity of layered perovskite oxides. After adding the GDC electrolyte, the EuBCO/10 wt% GDC oxygen electrode shows an ASR of 0.055 Ω cm2 at 700 °C.166 Moreover, the TEC is lowered to 12.8 × 10−6 K−1, which is compatible with pure GDC electrolyte (11.9 × 10−6 K−1).167 Afterwards, the composite oxygen electrode is made up of Ba-deficient EuBCO and SDC, resulting in a low ASR of 0.028 Ω cm2 at 700 °C.168 In the same way, a single fuel cell with the oxygen electrode (EuBa0.5Sr0.5Co2O5+δ/SDC) can deliver a PPD of ∼1200 mW cm−2 at 700 °C.159 By restructuring the electrolyte and anode components, a PPD of ∼1100 mW cm−2 is achieved in a fuel cell with the configuration of Ni/GDC|GDC| NdBa0.5Sr0.5Co2O5+δ/GDC.170 Besides, elaborately designed oxygen electrodes are explored with simple perovskite oxides, Ruddlesden–Popper oxides, and metals, such as PrBa0.5Sr0.5Co1.5Fe0.5O5+δ/La2NiO4+δ,171,172 PrBCO/LSCF,173 PrBa0.5Sr0.5Co1.5Fe0.5O5+δ/Pr0.9Fe0.7Co0.3O3,174 PrBa0.5Sr0.5Co1.5Fe0.5O5+δ/Ag,175 and PrBa0.8Ca0.2Co2O5+δ/BaCoO3−δ.176 Interestingly, perovskite oxide heterostructures have been successfully developed, e.g., PrBa(Co1−xFex)2O5+δ/Pr0.5Ba0.5Co1−xFexO3−δ and PrBa0.94Co2O5+δ/Pr0.5Ba0.47CoO3−δ.177,178 Tetragonal layered and cubic perovskite oxides provide abundant hetero-interfaces, possessing dramatic oxygen adsorption and dissociation abilities. Benefiting from the high electron transfer efficiency and fast oxygen surface rate, an ASR of 0.026 Ω cm2 is obtained for the heterostructure electrode at 700 °C.178 The ASR can be as low as 0.094 Ω cm2 at 600 °C, which meets the demand of an ideal oxygen electrode for SOFCs (0.1 Ω cm2). The related fuel cell delivers a PPD of 1100 mW cm−2 at 700 °C, along with extremely stable power output over a period of 120 h. Furthermore, we develop BaO nanoparticle-decorated Ln0.94BaCo2O5+δ (e-Ln0.94BCO) perovskite oxides.100 The e-La0.94BCO oxygen electrode exhibits an ASR of 0.03 Ω cm2 at 700 °C, which is reduced by ∼43% relative to pristine La0.94BCO. According to DFT calculations, the interfacial BaO compositions lead to the movement of the bulk O 2p-band center to the Fermi level (EF), meaning a lower energy barrier of oxygen/vacancy hopping and easier oxygen dissociation. Briefly, externally added active compositions improve the ORR activity in some ways, motivating us to continue this work.
Cation-deficient engineering is considered a useful route for designing high-performance perovskite oxide electrocatalysts. Regarding the A-site-deficient status, the Co ions will be partially oxidized to maintain electroneutral balance, which is helpful for the ORR process.88,194,195 The ASRs of the PrBa0.94Co2O5+δ and Pr0.94BaCo2O5+δ oxygen electrodes are 0.042 and 0.11 Ω cm2 at 700 °C, respectively.88,196 Their performance is comparable to that of the popular BSCF electrode (0.094 Ω cm2). In B-site-deficient PrBCO, electron hopping is impeded by the introduced Co defects.197 Owing to the stable valence states of the A-site cations, extrinsic oxygen vacancies are generated in the lattice, facilitating oxygen-ionic transport and oxygen reduction kinetics. These positive impacts on electrochemical performance have been clarified in Nd1−xBaCo2O5+δ,89,90 EuBa1−xCo2O5+δ,91 NdBa1−xCo2−yFeyO5+δ,198 Sm1−xBaCo2O5+δ,199 and Y1−xBaCoCuO5+δ.200 Besides, morphology-controllable perovskite oxides with different shapes also possess excellent ORR activities. Hollow nanofibers of PrBa0.5Sr0.5Co2O5+δ are assembled into a fibrous porous oxygen electrode, with facile channels for gas transport and charge transfer.201 The experimental ASRs and PPDs of LnBCO-based oxygen electrodes are summarized in Table 2.
| Cathode | ASR (Ω cm2) | Electrolyte | PPD (mW cm−2) | Anode|electrolyte | Ref. |
|---|---|---|---|---|---|
| a BZCYYb: BaZr0.1Ce0.7Y0.1Yb0.1O3. b LSGMC: La0.9Sr0.1Ga0.8Mg0.115Co0.085O2.85. c ScSZ: Zr0.9Sc0.1O1.95. d LSGM: La0.9Sr0.1Ga0.8Mg0.2O2.85. e LDC: La-doped CeO2. f PBM: PrBaMn2O5+δ. g BSCZGY: Ba0.5Sr0.5Ce0.6Zr0.2Gd0.1Y0.1. h BZCY: BaCe0.7Y0.2Zr0.1O3. | |||||
| PrBaCo2O5+δ | ∼0.213 (600 °C) | SDC | — | — | 40 |
| PrBaCo2O5+δ-GDC | ∼0.1 (600 °C) | SDC | — | — | 40 |
| LaBaCo2O5+δ | 0.048 (700 °C) | GDC | 1212 (750 °C) | Ni-YSZ|YSZ|GDC | 43 |
| PrBaCo2O5+δ | 0.054 (700 °C) | GDC | — | — | 43 |
| NdBaCo2O5+δ | 0.069 (700 °C) | GDC | — | — | 43 |
| SmBaCo2O5+δ | 0.079 (700 °C) | GDC | — | — | 43 |
| EuBaCo2O5+δ | 0.089 (700 °C) | GDC | — | — | 43 |
| GdBaCo2O5+δ | 0.096 (700 °C) | GDC | 744 (750 °C) | Ni-YSZ|YSZ|GDC | 43 |
| PrBa0.95Gd0.05Co2O5+δ | 0.038 (700 °C) | GDC | 1230 (700 °C) | Ni-YSZ|YSZ|GDC | 55 |
| PrBa0.9Ca0.1Co1.85Zn0.15O5+δ-BZCYYb | 0.09 (700 °C) | BZCYYba | 876 (750 °C) | Ni-BZCYYb|BZCYYb | 59 |
| GdBaCo1.7Ni0.3O5+δ | 0.54 (600 °C) | GDC | — | — | 61 |
| PrBa0.5Sr0.5Co1.9Ni0.1O5+δ | 0.297 (800 °C) | YSZ | 120 (800 °C) | Pt|YSZ | 63 |
| PrBaCo1.5Fe0.5O5+δ | 0.091 (700 °C) | SDC | — | — | 64 |
| PrBaCo1.6Fe0.4O5+δ | 0.07 (750 °C) | SDC | 446.4 (700 °C) | Ni-SDC|SDC | 65 |
| GdBa0.5Sr0.5Co1.5Fe0.5O5+δ-GDC | — | — | 250 (800 °C) | Ni-YSZ|GDC | 66 |
| YBaCo2O5+δ | 0.11 (700 °C) | LSGMCb | 873 (800 °C) | Ni-SDC|SDC|LSGMC | 68 |
| PrBaCo1.5Sc0.5O5+δ | 0.022 (750 °C) | GDC | — | — | 69 |
| NdBaCo1.95Zr0.05O5+δ | 0.024 (700 °C) | SDC | 1012 (800 °C) | Ni-ScSZ|ScSZc | 70 |
| PrBaCo1.75Nb0.25O5+δ | 1.034 (600 °C) | SDC | — | — | 71 |
| PrBaCo1.75Ta0.25O5+δ | 0.101 (700 °C) | BZCY | 755 (700 °C) | Ni-BZCY|BZCY | 72 |
| PrBa0.8Ca0.2Co2O5+δ-GDC | 0.033 (700 °C) | LSGMd|LDCe | 460 (700 °C) | PBMf|LSGM|LDC | 77 |
| Pr0.94Ba0.7Sr0.3Co2O5+δ | 0.031 (700 °C) | GDC | 1077 (700 °C) | Ni-YSZ|YSZ|GDC | 78 |
| NdBa0.5Sr0.5Co2O5+δ-GDC | 0.112 (600 °C) | GDC | ∼1000 (600 °C) | Ni-GDC|GDC | 79 |
| SmBa0.25Sr0.75Co2O5+δ-GDC | 0.138 (600 °C) | GDC | 1039 (600 °C) | Ni-GDC|GDC | 80 |
| Pr0.94Ba0.6Sr0.2Ca0.2Co2O5+δ | 0.025 (700 °C) | GDC | 1194 (700 °C) | Ni-YSZ|YSZ|GDC | 81 |
| LaBa0.9Co2O5+δ | 0.118 (600 °C) | GDC | — | — | 84 |
| PrBa0.94Co1.96Ta0.04O5+δ | 0.02 (700 °C) | GDC | 1050 (700 °C) | Ni-YSZ|YSZ|GDC | 87 |
| PrBa0.94Co2O5+δ | 0.042 (600 °C) | GDC | ∼1030 (700 °C) | Ni-GDC|GDC | 88 |
| Nd0.95BaCo2O5+δ-Bi2O3 | 0.026 (800 °C) | LSGM | 720 (800 °C) | Ni-SDC|SDC|LSGM | 90 |
| EuBa0.98Co2O5+δ | 0.107 (700 °C) | GDC | 505 (700 °C) | Ni-YSZ|YSZ|GDC | 91 |
| PrBaCo2O5+δ | 0.15 (600 °C) | GDC | — | — | 95 |
| GdBaCo2O5+δ | ∼0.1 (700 °C) | GDC | — | — | 96 |
| Pr1.1Ba0.9Co2O5+δF0.1 | 0.033 (700 °C) | GDC | 1102 (700 °C) | Ni-YSZ|YSZ|GDC | 97 |
| LaBaCo2O5+δ-LSGM | — | — | 516 (800 °C) | Ni-GDC|LDC|LSGM | 98 |
| e-La0.94BaCo2O5+δ | 0.03 (700 °C) | GDC | 1050 (700 °C) | Ni-YSZ|YSZ|GDC | 100 |
| LaBa0.7Ca0.3Co2O5+δ | 0.033 (700 °C) | GDC | 1050 (700 °C) | Ni-YSZ|YSZ|GDC | 101 |
| Pr1.1Ba0.4Sr0.5Co2O5+δ | 0.05 (650 °C) | GDC | 1506 (750 °C) | Ni-YSZ|YSZ|GDC | 102 |
| YBaCo2O5+δ | 0.44 (650 °C) | GDC | — | — | 103 |
| Eu0.8Ca0.2BaCo2O5+δ | 0.028 (700 °C) | GDC | 1210 (700 °C) | Ni-YSZ|YSZ|GDC | 104 |
| Pr0.94Ba0.7Ca0.3Co2O5+δ | 0.022 (700 °C) | GDC | 1202 (700 °C) | Ni-YSZ|YSZ|GDC | 105 |
| Pr0.7Ca0.3Ba0.94Co2O5+δ | 0.027 (700 °C) | GDC | 1114 (700 °C) | Ni-YSZ|YSZ|GDC | 106 |
| LaBa0.5Sr0.25Ca0.25Co2O5+δ | 0.075 (800 °C) | SDC | 662 (800 °C) | Ni-SDC|SDC | 107 |
| LaBa0.5Sr0.5Co2O5+δ | 0.084 (800 °C) | SDC | 580 (800 °C) | Ni-SDC|SDC | 107 |
| LaBa0.8Ca0.2Co2O5+δ | 0.039 (700 °C) | GDC | 1063 (700 °C) | Ni-GDC|GDC | 108 |
| LaBa0.5Sr0.5Co2O5+δ | ∼0.25 (800 °C) | SDC | — | — | 109 |
| PrBa0.8Ca0.2Co2O5+δ | ∼0.024 (750 °C) | SDC | ∼949 (700 °C) | Ni-BZCYYb|SDC | 110 |
| Pr0.9Ca0.1BaCo2O5+δ | 0.081 (700 °C) | SDC | 646.5 (800 °C) | Ni-SDC|SDC | 111 |
| PrBa0.7Ca0.3CoCuO5+δ | 0.052 (650 °C) | GDC | 2040 (800 °C) | Ni-YSZ|YSZ|GDC | 112 |
| Pr0.7Y0.3BaCo2O5+δ | — | — | 200 (750 °C) | Ni-GDC|LDC|LSGM | 113 |
| PrBa0.8Ca0.2CoCuO5+δ | 0.051 (800 °C) | LSGM | — | — | 114 |
| NdBa0.75Ca0.25Co2O5+δ | 0.066 (600 °C) | GDC | 2114 (600 °C) | Ni-GDC|GDC | 115 |
| Nd0.9La0.1BaCo2O5+δ | 0.083 (700 °C) | GDC | 1045 (700 °C) | Ni-YSZ|YSZ|GDC | 116 |
| NdBaCo2O5+δ-GDC | 0.038 (700 °C) | GDC | — | — | 117 |
| SmBa0.8Ca0.2Co2O5+δ | 0.019 (700 °C) | SDC | — | — | 120 |
| Sm0.8Ca0.2BaCo2O5+δ | 0.075 (700 °C) | SDC | 753 (800 °C) | Ni-SDC|SDC|LSGM|SDC | 121 |
| SmBa0.5Sr0.5Co2O5+δ | 0.57 (750 °C) | SDC | 304 (700 °C) | Ni-SDC|SDC | 122 |
| Gd0.5Pr0.5BaCo2O5+δ | 1.9 (700 °C, wet air) | BSCZGYg | — | — | 123 |
| GdBa0.5Sr0.5Co2O5+δ | — | — | 725 (700 °C) | Ni-SDC|SDC | 124 |
| YBa0.5Sr0.5Co2O5+δ | — | — | 468 (800 °C) | Ni-SDC|SDC|LSGMC | 126 |
| Y0.8Ca0.2BaCo2O5+δ | 0.068 (700 °C) | LSGM | — | — | 127 |
| LaBaCuCoO5+δ | 0.11 (700 °C) | SDC | 603 (800 °C) | Ni-SDC|SDC | 128 |
| PrBa0.5Sr0.5Co2O5+δ-GDC | 0.093 (600 °C) | GDC | 1220 (600 °C) | Ni-GDC|GDC | 130 |
| PrBaCo2/3Fe2/3Mn2/3O5+δ-SDC | 0.023 (800 °C) | SDC | 621 (800 °C) | Ni-SDC|SDC | 131 |
| PrBa0.92CoCuO6−δ | 0.017 (750 °C) | GDC | 1228 (750 °C) | Ni-YSZ|YSZ|GDC | 132 |
| PrBa0.8Ca0.2Co1.5Fe0.5O5+δ | 0.08 (600 °C) | GDC | 1890 (600 °C) | Ni-GDC|GDC | 133 |
| PrBaCuCoO5+δ | 0.047 (700 °C) | SDC | 791 (700 °C) | Ni-SDC|SDC | 135 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ | 0.07 (800 °C) | LSGM | 697 (850 °C) | Ni-SDC|SDC|LSGM | 136 |
| PrBaCo2/3Fe2/3Cu2/3O5+δ | 0.038 (800 °C) | GDC | 659 (800 °C) | Ni-GDC|GDC | 137 |
| PrBaCo1.6Fe0.2Nb0.2O5+δ–PrBaCo1.6Fe0.2Nb0.2−xO5+δ | 0.079 (750 °C, wet air) | BZCYYb | 1059 (650 °C) | Ni-BZCYYb|BZCYYb | 139 |
| PrBa0.8Ca0.2(Co0.95Fe0.05)2O6−δ | 0.017 (700 °C) | GDC | 1270.1 (700 °C) | Ni-YSZ|YSZ|GDC | 143 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ-GDC | ∼0.056 (600 °C) | GDC | 2160 (600 °C) | Ni-GDC|GDC | 144 |
| NdBa0.5Sr0.5Co2O5+δ-GDC | 0.105 (600 °C) | GDC | 1477 (600 °C) | Ni-GDC|GDC | 145 |
| NdBaCo2/3Fe2/3Cu2/3O5+δ | 0.077 (700 °C) | LSGM | 736 (800 °C) | Ni-GDC|GDC|LSGM | 146 |
| NdBaCo1.8Sc0.2O5+δ | 0.035 (700 °C) | SDC | 1188 (800 °C) | Ni-ScSZ|ScSZ|GDC | 147 |
| NdBa0.9Co1.9Fe0.1O5+δ | 0.14 (700 °C) | GDC | — | — | 148 |
| SmBaCuCoO5+δ | — | — | 355 (700 °C) | Ni-BZCY|BZCY | 150 |
| SmSrCo2O5+δ | 0.073 (700 °C) | GDC | — | — | 151 |
| SmBa0.5Sr0.5CoCuO5+δ | 0.0263 (850 °C) | LSGM | 857 (850 °C) | Ni0.9Cu0.1-GDC|LSGM | 152 |
| SmBaCo1.7Ni0.3O5+δ | 0.0464 (800 °C) | SDC | 536 (800 °C) | Ni-SDC|SDC | 153 |
| SmBaCo0.5Mn1.5O5+δ | 0.081 (900 °C) | LSGM | 1060 (900 °C) | Ni-GDC|LDC|LSGM | 154 |
| GdBaCuCo0.5Fe0.5O5+δ-GDC | 0.118 (750 °C) | GDC | — | — | 155 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ nanofiber | ∼0.025 (750 °C) | — | ∼2539 (750 °C) | Ni-YSZ|YSZ|SDC | 156 |
| NdBa0.5Sr0.5Co1.5Fe0.5O5+δ-GDC | 0.263 (800 °C) | GDC | 886.44 (800 °C) | Ni-YSZ|YSZ|GDC | 159 |
| Pr0.8Nd0.2BaCo1.6Fe0.4O5+δ | 0.0279 (600 °C) | GDC | 1345 (600 °C) | Ni-GDC|GDC | 160 |
| EuBaCo2O5+δ | 0.098 (750 °C) | GDC | — | — | 163 |
| YBaCo2O5+δ–Ag | 2.03 (780 °C) | YSZ | — | — | 164 |
| YBaCo1.4Cu0.6O5+δ | 0.076 (750 °C) | LSGM | 479 (750 °C) | Ni-GDC|GDC|LSGM | 165 |
| EuBaCo2O5+δ-GDC | 0.055 (700 °C) | GDC | 810 (700 °C) | Ni-GDC|GDC | 166 |
| EuBa0.98Co2O5+δ-SDC | 0.028 (700 °C) | SDC | 980 (700 °C) | Ni-YSZ|YSZ|GDC | 168 |
| EuBa0.5Sr0.5Co2O5+δ-SDC | 0.061 (700 °C) | GDC | 1185 (700 °C) | Ni-YSZ|YSZ|GDC | 169 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ–La2NiO4+δ | 0.11 (700 °C) | GDC | 940 (750 °C) | Ni-YSZ|YSZ|GDC | 171 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ–La2NiO4+δ | 0.51 (700 °C) | GDC | 710 (700 °C) | Ni-YSZ|YSZ|GDC | 172 |
| PrBaCo2O5+δ-LSCF | 0.086 (750 °C) | GDC | — | — | 173 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ–Pr0.9Fe0.7Co0.3O3 | 0.09 (700 °C) | BZCYYb | 1080 (650 °C) | Ni-BZCYYb|BZCYYb | 174 |
| LaBa0.8Ca0.2Co2O5+δ–BaCoO3−δ | ∼0.03 (750 °C) | SDC | 940 (750 °C) | Ni-YSZ|ScSZ|SDCf | 176 |
| PrBaCo1.4Fe0.6O5+δ–Pr0.5Ba0.5Co0.7Fe0.3O3+δ | 0.039 (700 °C) | BZCYh | 1304 (700 °C) | Ni-BZCY|BZCY | 177 |
| PrBa0.94Co2O5+δ/Pr0.5Ba0.47CoO3−δ | ∼0.025 (700 °C) | GDC | 1100 (700 °C) | Ni-YSZ|YSZ|GDC | 178 |
| PrBaCoFeO5+δ-YSZ | 0.075 (700 °C) | YSZ | 910 (700 °C) | Ni-YSZ|YSZ | 180 |
| Pr0.9Y0.1BaCo1.8Ni0.2O6−δ-SDC | 0.11 (700 °C) | SDC | — | — | 181 |
| PrBaCo2O5+δ-SDC | 0.252 (600 °C) | SDC | 1150 (650 °C) | Ni-SDC|SDC | 182 |
| NdBaCo2O5+δ-GDC | 0.034 (700 °C) | GDC | 1000 (700 °C) | Ni-YSZ|YSZ|GDC | 183 |
| SmBaCo1.7Ni0.3O5+δ-SDC | 0.0272 (800 °C) | SDC | 628 (800 °C) | Ni-SDC|SDC | 184 |
| SmBaCo2O5+δ-SDC | 0.159 (700 °C) | LSGM | 408 (700 °C) | Ni-GDC|LDC|LSGM | 185 |
| SmBa0.5Sr0.5Co2O5+δ-GDC | 0.013 (700 °C) | GDC | 1310 (800 °C) | Ni-YSZ|YSZ|GDC | 186 |
| SmBa0.5Sr0.5Co2O5+δ-SDC | 1.28 (600 °C) | SDC | 823 (800 °C) | Ni-SDC|SDC | 187 |
| SmBa0.5Sr0.5Co2O5+δ-GDC | 0.031 (700 °C) | GDC | — | — | 188 |
| SmBa0.5Sr0.5Co2O5+δ-ScSZ | 0.054 (700 °C) | ScSZ | 1250 (700 °C) | 430L-YSZ|ScSZ | 189 |
| SmBa0.5Sr0.5Co2O5+δ-GDC | 0.019 (700 °C) | YSZ | — | — | 191 |
| PrBaCo2O5+δ-SDC | 0.073 (700 °C) | LSGM | 758 (800 °C) | Ni-SDC|SDC|LSGM | 192 |
| PrBaCoFeO5+δ-SDC | 0.037 (800 °C) | LSGM | 960 (800 °C) | Ni-SDC|SDC|LSGM | 193 |
| PrBa0.92Co2O5+δ | 0.093 (600 °C) | GDC | — | — | 194 |
| PrBaCo2O5+δ | 0.0594 (700 °C) | GDC | — | — | 195 |
| Pr0.94BaCo2O5+δ | 0.11 (600 °C) | GDC | 1050 (600 °C) | Ni-GDC|GDC | 196 |
| PrBaCo1.94O5+δ | 0.059 (700 °C) | GDC | 889 (650 °C) | Ni-YSZ|YSZ|GDC | 197 |
| Sm0.95BaCo2O5+δ | 0.038 (750 °C) | GDC | — | — | 199 |
| Y0.93BaCoCuO5+δ | 0.029 (800 °C) | LSGM | 643 (800 °C) | Ni-SDC|SDC|LSGM | 200 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ | — | — | 986 (800 °C) | Ni-YSZ|YSZ|GDC | 225 |
| H2O → HO* + H+ + e− | (9) |
| HO* → O* + H+ + e− | (10) |
| O* + H2O → HOO* + H+ + e− | (11) |
| HOO* → O2 + H+ + e− | (12) |
![]() | ||
| Fig. 4 Electrocatalytic mechanisms of the (a) OER AEM, (b) OER LOEM, (c) alkaline HER, and (d) acid HER. | ||
The OER investigations of layered perovskite oxides began with PrBCO in alkaline media.47 The structural evolution from disordered to ordered perovskite oxide is determined by the annealing temperature, i.e., disordered (Pr0.5Ba0.5)CoO3−δ, A-site disordered (Pr0.5Ba0.5)2Co2O5.5+δ, A-site ordered PrBaCo2O5+δ, and A-site ordered PrBaCo2O5.5+δ. Almost unchanged OER activity, at the current density (J) = 10 mA cm−2 at ∼1.8 V vs. reversible hydrogen electrode (RHE), is attained in all materials, revealing the negligible effect of the structural bulk phase on OER electrocatalysis. A and B-site co-doping has been performed in NdBCO to improve the OER activity. The overpotentials follow the order of NdBa0.5Sr0.5Co1.5Fe0.5O5+δ (361 mV) < SmBa0.5Sr0.5Co1.5Fe0.5O5+δ (372 mV) < GdBa0.5Sr0.5Co1.5Fe0.5O5+δ (390 mV) at 10 mA cm−2, with the smallest Tafel slope of 88 mV dec−1 for NdBa0.5Sr0.5Co1.5Fe0.5O5+δ.49 Based on the foundation of co-doping, the oxygen defects of PrBa0.5Sr0.5Co1.5Fe0.5O5+δ can be tailored carefully, and it is inferred that excessive oxygen vacancies promote the OH− affinity and reduce the formation energy of O* on the surface.203 We presented a series of EuBa0.5Sr0.5Co2−xFexO5+δ electrocatalysts for OER, gaining the optimal EuBa0.5Sr0.5Co1.6Fe0.4O5+δ one.204 Importantly, its OER activity is higher than that of the commercial IrO2 at η > 460 mV, clearly indicating its potential application at a large J. After that, we designed a composite electrocatalyst of EuBa0.5Sr0.5Co1.6Fe0.4O5+δ/20 wt% RuO2, which outperformed the high-performance RuO2 benchmark. This composite electrode-based electrolyzer delivered a low voltage (U) of 1.47 V for overall water splitting at 10 mA cm−2. DFT computation implies the presence of active Fe and Ru sites for the AEM process. The free energies of the intermediates on the Co, Fe, and Ru sites are calculated for EuBa0.5Sr0.5Co1.6Fe0.4O5+δ and RuO2 at different voltages (U = 0 and 1.23 V). The theoretical η values for the Fe and Ru atoms are 0.43 and 0.78 V, respectively. The results indicate that octahedral Fe atoms are highly active for OER. Conversely, the larger theoretical overpotentials of the Co and Sr atoms (1.73 and 2.01 V) exclude their probabilities as the reaction sites, respectively. In the case of an ideal oxide electrocatalyst, ΔEHOO* and ΔEHO* are predicted to be 3.69 and 1.23 eV, respectively, and ΔEO* is located in the middle at 2.46 eV.202 The octahedral Fe atom has a ΔEO* of 2.60 eV, and the ΔEHOO* and ΔEHO* of the Ru atom are 3.68 and 1.03 eV, respectively, testifying the cooperation between Fe and Ru for water oxidation. Analogous composite electrocatalysts, PrBa0.5Sr0.5Co2O5+δ@FeOOH-20, PrBa0.5Sr0.5Co2O5+δ@CoP, Pr(Ba0.5Sr0.5)0.95Co1.5Fe0.5O5+δ@N-doped graphene, and (PrBa0.8Ca0.2)0.95(Co1.5Fe0.5)0.95Co0.05O5+δ@Co/CoOx, have been explored for alkaline OER.205–208 To find a solution for the low surface areas of perovskite oxides, mesoporous nanofibers (PrBa0.5Sr0.5Co1.5Fe0.5O5+δ and SmBa0.5Sr0.5Co2O5+δ) have been prepared through the electrospinning technique.50,209 Their well-regulated B-site transition metal ratios and high surface areas (∼20 m2 g−1) resulted in a better OER performance than IrO2. Jo et al. found that the reconstruction of perovskite oxides promotes the OER, demonstrating stable and high-efficiency electrocatalysts.210 To expound the function of B-site tuning of perovskite oxides, PrBa0.5Sr0.5Co2−xFexO5+δ with various x values was studied. It was found that the layered perovskite structure and randomly disordered perovskite structure coexisted, altering the intrinsic properties with efficient OER activity and prolonged stability (at 100 mA cm−2 over 2000 h). Oxygen-vacancy defects have been deemed to be crucial for electrocatalytic reactions. Upon treatment in Ar and H2 plasma, oxygen vacancies are generated in PrBa0.5Sr0.5Co1.5Fe0.5O5+δ with tunable concentrations, which are associated with significantly enhanced OER activity.211 The OER properties of layered perovskite oxides are still inferior to that of commercial IrO2 and RuO2 (Table 3), especially operating at a low J (10 mA cm−2). We conclude that the determination of the real active sites is meaningful for the development of OER electrocatalysts, such as surface reconstruction, composition, size, and dimension modulation.
| OER overpotential (mV) | Medium | HER overpotential (mV) | Medium | Ref. | |
|---|---|---|---|---|---|
| LaBaCo2O5+δ | — | — | −156 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| PrBaCo2O5+δ | — | — | −255 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| NdBaCo2O5+δ | — | — | −389 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| SmBaCo2O5+δ | — | — | −456 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| GdBaCo2O5+δ | — | — | −495 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| EuBaCo2O5+δ | — | — | −551 (@10 mA cm−2) | KOH (1.0 M) | 44 |
| PrBaCo2O5+δ | — | — | −245 (@10 mA cm−2) | KOH (0.1 M) | 45 |
| PrBaCo2O5.8 | — | — | −240 (@10 mA cm−2) | KOH (1.0 M) | 46 |
| PrBaCo2O5+δ | 570 (@10 mA cm−2) | KOH (1.0 M) | — | — | 47 |
| NdBa0.5Sr0.5Co1.5Fe0.5O5+δ | 361 (@10 mA cm−2) | KOH (0.1 M) | — | — | 49 |
| SmBa0.5Sr0.5Co1.5Fe0.5O5+δ | 372 (@10 mA cm−2) | KOH (0.1 M) | — | — | 49 |
| GdBa0.5Sr0.5Co1.5Fe0.5O5+δ | 390 (@10 mA cm−2) | KOH (0.1 M) | — | — | 49 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ nanofiber | 300 (@10 mA cm−2) | KOH (0.1 M) | — | — | 50 |
| GdBa0.6Sr0.4Co2O5+δ | 460 (@1 mA cm−2) | KOH (0.1 M) | — | — | 125 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ–Ag | 340 (@10 mA cm−2) | KOH (1.0 M) | ∼−120 (@10 mA cm−2) | KOH (1.0 M) | 175 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ | 313 (@10 mA cm−2) | KOH (1.0 M) | — | — | 203 |
| RuO2 | 290 (@10 mA cm−2) | KOH (1.0 M) | — | — | 204 |
| IrO2 | 400 (@10 mA cm−2) | KOH (1.0 M) | — | — | 204 |
| EuBa0.5Sr0.5Co1.6Fe0.4O5+δ | 420 (@10 mA cm−2) | KOH (1.0 M) | — | — | 204 |
| EuBa0.5Sr0.5Co1.6Fe0.4O5+δ–RuO2 | 290 (@10 mA cm−2) | KOH (1.0 M) | — | — | 204 |
| PrBa0.5Sr0.5Co2O5+δ–FeOOH | 390 (@10 mA cm−2) | KOH (0.1 M) | −280 (@10 mA cm−2) | KOH (0.1 M) | 205 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ–CoP | 378 (@10 mA cm−2) | KOH (0.1 M) | — | — | 206 |
| (PrBa0.5Sr0.5)0.95Co1.5Fe0.5O5+δ-N-doped graphene | 320 (@10 mA cm−2) | KOH (0.1 M) | −230 (@10 mA cm−2) | KOH (0.1 M) | 207 |
| (PrBa0.8Ca0.2)0.95(Co1.5Fe0.5)0.95Co0.05O5+δ@Co/CoOx | 410 (@10 mA cm−2) | KOH (1.0 M) | ∼−210 (@10 mA cm−2) | KOH (1.0 M) | 208 |
| SmBa0.5Sr0.5Co2O5+δ nanofiber | 370 (@10 mA cm−2) | KOH (0.1 M) | — | — | 209 |
| PrBa0.5Sr0.5Co1.5Fe0.5O5+δ | 470 (@64.36 mA cm−2) | KOH (1.0 M) | — | — | 211 |
| Gd0.5La0.5BaCo2O5+δ | — | — | −240 (@338 mA cm−2) | KOH (1.0 M) | 237 |
| Pr0.5La0.5BaCo2O5+δ | — | — | −240 (@317 mA cm−2) | KOH (1.0 M) | 237 |
| LaBaCo2O5+δ | — | — | −233.2 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Gd0.2La0.8BaCo2O5+δ | — | — | −226.8 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Gd0.4La0.6BaCo2O5+δ | — | — | −209.8 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Gd0.5La0.5BaCo2O5+δ | — | — | −210.4 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Gd0.8La0.2BaCo2O5+δ | — | — | −236.2 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| GdBaCo2O5+δ | — | — | −240.4 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Pr0.5La0.5BaCo2O5+δ | — | — | −209.9 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Sm0.5La0.5BaCo2O5+δ | — | — | −227.9 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Pr0.5Gd0.5BaCo2O5+δ | — | — | −239.9 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Gd0.5La0.4Sr0.1BaCo2O5+δ | — | — | −229.5 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| PrBa0.5Sr0.5Co2O5+δ | — | — | −242.1 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| Ho0.8Ba0.6Sr0.6Co2O5+δ | — | — | −238 (@10 mA cm−2) | KOH (1.0 M) | 237 |
| P-doped Pr0.5La0.5BaCo2O5+δ nanofiber | — | — | −208 (@10 mA cm−2) | KOH (1.0 M) | 238 |
| P-doped Pr0.5La0.5BaCo2O5+δ nanofiber | — | — | −307 (@500 mA cm−2) | KOH (1.0 M) | 238 |
| Pt/C | — | — | −646 (@1000 mA cm−2) | KOH (1.0 M) | 240 |
| (LaBa)0.95Co2O5+δ | — | — | ∼−215 (@10 mA cm−2) | KOH (1.0 M) | 240 |
| (LaBa)0.95Co2O5+δ | — | — | −415 (@1000 mA cm−2) | KOH (1.0 M) | 240 |
| Pt/C | — | — | ∼−974 (@2000 mA cm−2) | KOH (1.0 M) | 241 |
| Pr0.4La0.5Co2O5+δ | — | — | ∼−636 (@2000 mA cm−2) | KOH (1.0 M) | 241 |
| Pt/C | — | — | −35 (@10 mA cm−2) | KOH (1.0 M) | 242 |
| Pt/C | — | — | −433 (@500 mA cm−2) | KOH (1.0 M) | 242 |
| PrBa0.94Co2O5+δ/Pr0.5Ba0.47CoO3−δ | — | — | −186 (@10 mA cm−2) | KOH (1.0 M) | 242 |
| PrBa0.94Co2O5+δ/Pr0.5Ba0.47CoO3−δ | — | — | −364 (@500 mA cm−2) | KOH (1.0 M) | 242 |
Besides alkaline water electrolysis, H2O or CO2 can be decomposed into H2/O2 or CO/O2via SOECs. ABO3 perovskite oxides have been used as anode materials for oxygen evolution, such as La(Sr)MnO3−δ, LSCF, BSCF, and composite electrocatalysts.3,26,212–220 However, their electrochemical efficiencies are insufficient at high temperatures (≥800 °C). In an 80%H2O/20%H2 atmosphere, the BSCF/SDC anode-based electrolysis cell shows a relatively acceptable J of ∼0.75 A cm−2 at 1.4 V towards H2O decomposition at 850 °C.221 However, the J is merely ∼0.2 A cm−2 at 1.4 V towards the CO2 reduction reaction (CORR) at 800 °C.222 A vanadium-doped BSCF perovskite oxide was developed for the CORR, with a J of ∼0.3 A cm−2 under the same conditions.222 By reason of their superior charge transfer reaction and intrinsic oxygen defects, layered perovskite oxides are considered up-and-coming anode alternatives for the CORR. It should be noted that their efficiencies are much higher than that of simple perovskite oxides, suggesting their more prominent OER activity. A family of layered perovskite oxides (Ln = La, Pr, Nd, Sm, Eu, and Gd) has been assessed as OER electrocatalysts for SOECs. The J values are 1.50, 174, and 1.91 A cm−2 for the GdBCO, NdBCO, and LaBCO anode-based electrolysis cells at 1.8 V towards the CO2RR at 800 °C, respectively.43 The efficiency ordering is the same as their ORR activities, highlighting the relevance of lanthanide modulation with electrocatalytic activity. The doping strategy is carried out to search for high-standard anodes for SOECs. A considerable anode (PrBa0.95Gd0.05Co2O5+δ) is developed by abnormal Ba-site Gd3+ substitution, displaying a large J of 2.63 A cm−2 at 1.8 V towards the CO2RR at 750 °C.58 Given its high-performance LaBCO composition, the J can reach an ultra-high 3.62 A cm−2 for the LaBa0.7Ca0.3Co2O5+δ anode at 1.8 V (CO2RR at 800 °C).102 Yao et al. devised a perovskite oxide anode (PrBaCo0.3Fe1.7O5+δ) by controlling the Co/Fe ratios, possessing a superior OER activity with a J of ∼3.9 A cm2 at 1.8 V (CO2RR at 800 °C).223 The co-doping strategy is still effective in enhancing the oxygen evolution performance. A Ca and Co co-doped PrBa0.8Ca0.2Co1.96Ta0.04O5+δ anode shows robust efficiency, including J = 1.44 A cm−2 at 1.6 V with attractive stability over 100 h for the CO2RR at 800 °C.224 A representative perovskite oxide electrocatalyst, PrBa0.5Sr0.5Co1.5Fe0.5O5+δ, can also be used as the anode material for SOECs. The maximum J of 3.694 A cm−2 is achieved at 2.0 V for water electrolysis at 850 °C, with a stable operation for 120 h with reversibility,225 and this anode stabilizes a long-term J of 1.31 A cm−2 at 1.3 V (800 °C).226 The intrinsic properties can be optimized via dual cation-defect engineering, i.e., balanced oxygen vacancies and oxygen-ionic conductivity. The electrolysis cell with the Pr0.97Ba0.97Co1.5Fe0.5O5+δ/GDC composite anode actualizes a J of 2.61 A cm−2 at 1.5 V for the water electrolysis at 800 °C.227 Layered perovskite oxides are also applied to proton ceramic electrochemical cells. An A-site deficient (PrBa0.8Ca0.2)0.95Co2O6−δ oxygen electrode-based proton electrolysis cell has a J of 0.72 A cm−2 at 1.3 V for water electrolysis at 600 °C.228 It is known that the cathode (fuel electrode) exposed to H2O/CO2 dominates the performance of SOECs. However, our results suggest that the anode also contributes to the performance of SOECs to a certain degree.102 In fact, another series of Fe/Mn-based layered perovskite oxides, LnBa(Fe/Mn)2O5+δ, can serve as cathode materials to absorb the H2O/CO2 reactants,223,229–235e.g., Pr0.95Ba0.95Fe1.6Ni0.2Nb0.2O5+δ.235 Accordingly, future work should aim at developing bifunctional electrodes for ORR and OER, which are used in reversible solid oxide cells (RSOCs).
| H2O + e− → H* + OH−, H* + H* → H2 (Volmer–Tafel) | (13) |
| H2O + e− → H* + OH−, H* + H2O + e− → OH− + H2 (Volmer–Heyrovsky) | (14) |
| H+ + e− → H*, 2H* → H2 (Volmer–Tafel) | (15) |
| H+ + e− → H*, H* + H* → H2 (Volmer–Heyrovsky) | (16) |
Layered LnBCO perovskite oxides have assured HER activity in alkaline media. Sun et al. systematically investigated PrBCO electrocatalysts with cubic, tetragonal, and orthorhombic structures, designating the optimal orthorhombic structure (δ ≈ 0.52) with an HER η of 245 mV at 10 mA cm−2 in 0.1 M KOH.45 Constantly progressive reaction steps are involved in the HER process in computational studies, including H2O adsorption on the surface, H2O dissociation into H* and –OH, formation of H* intermediates, and the combination of H* to form H2 (Fig. 5a).236 The RBaCo2O5.5, RBaCo2O5.75, and RBaCo2O6 models are constructed to calculate the free energies of these steps (Fig. 5b), in which a combined interplay between ordered oxygen vacancies (pyramidal high-spin Co3+ sites) and O 2p ligand holes (metallic octahedral intermediate-spin Co4+ sites) is assumed in RBaCo2O5.75. These two active sites produce a near-ideal reaction path to adsorb H2O and release H2, respectively. Experimentally, as-synthesized (Gd0.5La0.5)BaCo2O5.75 is superior to even the commercial Pt/C catalyst in terms of HER activity and stability. The rate-limiting step is determined to be the combination of adsorbed H* to form H2 (Fig. 5c), providing a theoretical basis for hydrogen-evolving electrocatalysis. Moreover, Guan et al. introduced A-site ionic electronegativity (AIE) as an efficient unifying descriptor to predict the HER activities of 13 cobalt-based perovskite oxides,237i.e., LaBaCo2O5.5+δ (Gd0), (Gd0.2La0.8)BaCo2O5.5+δ (Gd0.2), (Gd0.4La0.6)BaCo2O5.5+δ (Gd0.4), (Gd0.5La0.5)BaCo2O5.5+δ (Gd0.5), (Gd0.6La0.4)BaCo2O5.5+δ (Gd0.6), (Gd0.8La0.2)BaCo2O5.5+δ (Gd0.8), GdBaCo2O5.5+δ (Gd1), (Pr0.5La0.5)BaCo2O5.5+δ (Pr0.5La0.5), (Sm0.5La0.5)BaCo2O5.5+δ (Sm0.5La0.5), (Pr0.5Gd0.5)BaCo2O5.5+δ (Pr0.5Gd0.5), PrBa0.5Sr0.5Co2O5.5+δ (PrBa0.5Sr0.5), Ho0.8Ba0.6Sr0.6Co2O5.5+δ (Ho0.8Ba0.6Sr0.6), and Gd0.5La0.4BaSr0.1Co2O5.5+δ (Gd0.5La0.4Sr0.1). Their AIE values endow them with HER activity (η at 10 mA cm−2 and Tafel slope), as shown in the volcano curves. A fully identical layered perovskite oxide, (Gd0.5La0.5)BaCo2O5.5+δ, with an AIE value of ∼2.33, is predicted to have the peak HER activities (Fig. 5d and e). With the aid of X-ray adsorption spectroscopy and DFT computation, this performance can be attributed to the most appropriate electronic states with inductive effects in the perovskite structure, i.e., ∼+3.25 of Co valence state, medium Co–O covalency, band gap (0.14 eV), and O 2p-band position. The charge-transfer energy (Δ) will decrease from 6–7 eV for Co2+ to −2 eV for Co4+ (Fig. 5f), and the Co 3d–Co 2p covalency is enlarged with a shift in the electronic band states close to the EF. However, a perovskite oxide (LnBaCo2O5.75) with a moderate Co valence state (∼+3.25) satisfies the Sabatier principle. When the Co valence is too high (>3.25), the insufficient oxygen vacancies hamper H2O adsorption. If the Co valence is too low, the deficient Co4+ sites are not favorable for desorbing H2. Meanwhile, the O 2p-band center is −2.38 eV for Co3.25+, which is neither too close nor too far from the EF (−2.74 and −2.01 eV for Co3+ and Co3.5+), respectively. In terms of the AIE rule, two perovskite oxides with an AIE value of ∼2.33, Gd0.4La0.6Ba0.4Ca0.6Co2O5.5+δ and Pr0.5La0.5Ba0.5Ca0.5Co2O5.5+δ, have been synthesized for alkaline HER. Their activities are located at the top of the volcano plot and is comparable to that of Gd0.5La0.5BaCo2O5.5+δ. We discovered that the HER performance is related to the structural evolution of layered perovskite oxides.44 At tf ≈ 1.0, cubic perovskite oxide (LaBCO) shows an η of −156 mV at 10 mA cm−2 with a Tafel slope of 64.4 mV dec−1, which is much better than that of reported LnBCO-based electrocatalysts (Table 3). At a large current density (>230 mA cm−2), it even outperforms the Pt/C benchmark. The figure of merit for comparing the intrinsic activities of different electrocatalysts is the turnover frequency (TOF), which is defined as the number of hydrogen molecules evolved per second per active site. The TOF value of LaBCO is 4.1 H2 s−1 at η = 300 mV, significantly surpassing other perovskite oxides and Pt/C. Computational results reveal the rate-limiting step of H2 desorption, and the first-best energy collocation is identified for LaBCO including H2O adsorption (−0.43 eV), H2O dissociation (−0.06 eV), –OH desorption (−0.88 eV), and H2 desorption (1.12 eV). We perceive the necessity of design principles, opening up possibilities for the application of perovskite oxide electrocatalysts. Alternatively, tailored compositions and controlled morphologies ensure exceptional electrocatalytic properties. Li et al. prepared a P-doped Pr0.5La0.5BaCo2O5+δ electrocatalyst with a fiber-like shape, showing an η of −307 mV at 500 mA cm−2 with a Tafel slope of 32.9 mV dec−1.238 This fast hydrogen-evolving action exceeds the commonly used Pt/C. If the H* intermediates are insufficient, the rate-limiting step is the Volmer reaction with a theoretical Tafel slope of 120 mV dec−1. When the Heyrovsky or Tafel reaction is the rate-limiting step, the theoretical Tafel slope is 40 or 30 mV dec−1.239 The smaller value reflects the Volmer–Tafel mechanism with fast dynamics. We explored several perovskite oxide electrocatalysts via cation-defect engineering towards high-rate HER. A-site co-deficient (LaBa)0.95Co2O5+δ merely requires an η of −415 mV at 1000 mA cm−2.240 Most importantly, Pr-deficient Pr0.4La0.5BaCo2O5+δ can be operated under industrial-grade conditions, i.e., 500–2000 mA cm−2, resulting from its enhanced charge transfer and H2 desorption abilities.241 Additionally, layered perovskite oxide-based composite electrocatalysts are also active for the alkaline HER. Hetero-structured Pr0.94BaCo2O5+δ/Pr0.47Ba0.5CoO3−δ (tetragonal/cubic structure) affords a η of −364 mV at 500 mA cm−2.242 Heterostructures have a large surface area, more hydrophilic surface, and increased electrical conductivity arising from in situ exsolved cubic perovskite oxide, while improving the H2O adsorption/dissociation and H2 desorption processes.
![]() | ||
| Fig. 5 (a) Crystal structures of oxygen vacancy-ordered RBaCo2O5.5 and RBaCo2O5.75, as well as RBaCo2O6 without oxygen vacancies. In the ROδ planes, the hollow, grey, and orange circles represent Copyr3+ (with oxygen vacancies), octahedral Co3+ (Cooct3+), and Cooct4+ configurations, respectively. (b) Computational predictions for the alkaline HER processes of H2O adsorption, H2O dissociation, formation of H* intermediates, and subsequent recombination of two H* to form H2 on the surface of RBaCo2O5.75. (c) The calculated free energies of the H2O adsorption (grey), H2O dissociation to adsorbed OH− and H* (blue), OH− desorption (orange), and hydrogen desorption (red) in the alkaline HER on RBaCo2O5.5, RBaCo2O5.75, and RBaCo2O6 surfaces. Reproduced with permission.236 Copyright 2019, Wiley. (d) The HER activity trends of pure-phase cobalt-based perovskite oxides at the overpotential at 10 mA cmoxide−2 as a function of A-site ionic electronegativity. (e) The plots of Tafel slope values as a function of A-site ionic electronegativity (the rate-determining steps estimated from the Tafel slope are shown in pink and purple dashed lines). (f) The charge-transfer models of Co2+, Co3+, and Co4+ in the covalent systems. Reproduced with permission.237 Copyright 2019, Nature. | ||
Motivated by the success of the d-band theory as an activity descriptor for metal surfaces, a distinct OER activity design principle, i.e., the near-unity occupancy of the eg orbital of surface transition metal ions, can promote the intrinsic OER activity of perovskite oxides in alkaline media.243 This principle is based on a molecular orbital bonding framework. Because the eg orbital participates in σ-bonding with the anion adsorbates on the surface, the binding ability of oxygen-containing intermediate species at the B-sites affects the OER activity. In the case of La/Ba-based perovskite oxides, plotting their activities (voltage vs. RHE at a surface area-normalized current density of 50 µA cmox−2) as a function of the eg-orbital filling of the surface B-site cation can lead to a volcano plot (Fig. 6a). A highly active OER catalyst, BSCF, is obtained, which is the most popular oxygen electrode for SOFCs and oxygen permeation membranes. Moreover, BSCF has the highest OER activity among the studied perovskite oxides, as predicted by the eg descriptor. The electronic configuration of Co2.8+ in BSCF can be assigned to t52ge∼1.2g given that the Co ions are in the intermediate-spin states on the surface. A low eg-filling in the perovskite oxides is associated with the B-site cation–O bonding, which is too strong for O2 to be desorbed. Instead, high eg-filling prefers weak O2 adsorption. The best OER activity is correlated with a medium eg-filling value of ∼1. Notwithstanding, we take note of the eg occupancy from an ionic model, where the metal atoms are adopted as the active sites, and the participation of active lattice oxygen sites is ignored. Suntivich et al. probed some perovskite oxides with the same eg occupancy, e.g., LaMnO3, LaNiO3, and LaCoO3, but they observed different electrocatalytic activities.244 Consequently, it is difficult to explain these facts using the eg orbital occupancy.
![]() | ||
Fig. 6 (a) Relation between the OER catalytic activity, defined by the η at 50 µA cmox−2 and occupancy of the eg-symmetry electron of the transition metal (B in ABO3). The data symbols vary with the type of B ions (Cr, red; Mn, orange; Fe, beige; Co, green; Ni, blue; and mixed compounds, purple), where x = 0, 0.25, and 0.5 for Fe. The error bars represent the standard deviations of at least three independent measurements. The dashed volcano line is shown for guidance only. Reproduced with permission.243 Copyright 2011, Science. (b) Activity trend towards oxygen evolution plotted for perovskite oxides. The negative theoretical η is plotted against the standard free energy of the step. The low coverage regime is considered and the calculated values are used to show the activity of each perovskite oxide. The volcano curve is established by using the scaling relation between and . Reproduced with permission.202 Copyright 2011, Wiley. (c) Experimental ASRs at ∼1000 K vs. the calculated bulk O p-band center of perovskites with simulated composition: 1. Ba0.5Sr0.5Co0.75Fe0.25O3 (BSCF), 2. Sm0.5Sr0.5CoO3 (SSC), 3. PrBaCo2O6 (PBCO), 4. GdBaCo2O6 (GBCO), 5. La0.625Sr0.375Co0.25Fe0.75O3 (LSCF), 6. La0.75Sr.25CoO3 (LSC), 7. LaCoO3 (LCO), 8. La0.75Sr0.25MnO3 (LSM), and 9. LaMnO3 (LMO). The values with vertical errors bars are the average of multiple data values with an error bar equal to the standard deviation in the mean of the ASR. (d) Schematic electronic structure plots illustrating the correlation of the ORR energetics vs. the O p-band center based on a rigid band model. The shaded areas (red and blue in color) represent the occupied transition metal 3d and oxygen 2p states, and the empty areas represent the unoccupied states, respectively. The bold and dotted lines are the Fermi energy level before removing O2− and before adding O2−, respectively. Reproduced with permission.245 Copyright 2011, the Royal Society of Chemistry. (e) Plots of log k* of typical perovskite oxides as a function of calculated O p-band center. The blue symbols are the experimental data. The k* data for the purple symbols are obtained using the reported ASR data. The green diamonds are plotted using predicted log k* values based on the linear fit of the experimental data. Reproduced with permission.246 Copyright 2018, Wiley. (f) Evolution of the iR-corrected potential at 0.5 mAoxide−2versus the O p-band center relative to EF (eV) of (Ln0.5Ba0.5)CoO3−δ with Ln = Pr, Sm, Gd and Ho, for LaCoO3 (LCO), La0.4Sr0.6CoO3−δ (LSC46), Ba0.5Sr0.5Co0.8Fe0.2O3−δ (BSCF82), Ba0.5Sr0.5Co0.4Fe0.6O3−δ (BSCF46) and SrCo0.8Fe0.2O3−δ (SCF82). The error bars represent the standard deviations from at least four independent measurements. Reproduced with permission.247 Copyright 2013, Nature. Experimental plots of (g) log J and (h) log k* of typical perovskite oxides vs. calculated bulk O p-band enter for PBE. In the order of increasing J value for OER, the materials in (g) correspond to LaCrO3, LaFeO3, LaMnO3, LaCoO3, La0.75Sr0.25CoO3, LaNiO3, GdBaCo2O5.5, SmBaCo2O5.5, La0.5Sr0.5CoO3, and PrBaCo2O5.5. In the order of increasing k* value, the materials in (h) correspond to LaMnO3, LaCoO3, LaFeO3, La0.75Sr0.25CoO3, SrFeO2.75, La0.5Sr0.5CoO3, GdBaCo2O5.5, PrBaCo2O5.5, and Ba0.5Sr0.5Co0.75Fe0.25O2.625. Reproduced with permission.248 Copyright 2019, the American Chemical Society. | ||
The adsorption free energies of the intermediates can directly convey the electrocatalytic activity. Progress in the DFT calculations makes it possible to use them as accurate descriptors. Considering the OER intermediates as O*, HO*, and HOO*, the scaling relationship can be established between the binding energies of these species on the oxide surfaces. Man et al. established the formal scaling relationship between the HO* and HOO* binding energies on rutile, spinel, rock salt, bixbyite, and perovskite oxides.202 The linear correlation implies that HO* and HOO* normally prefer the same type of binding site, and the difference of ΔEHOO* − ΔEHO* is almost kept at a constant of 3.2 eV for all the studied oxides. Given the constant difference between the HO* and HOO* levels, the difference of
is a unique descriptor for OER activity. The theoretical overpotential (ηOER) under standard conditions is calculated using eqn (17), as follows:
![]() | (17) |
for a series of perovskite oxides, resulting in a universal volcano relationship (Fig. 6b), in which SrCoO3 with
of 1.48 eV is close to the very top of the volcano.
Electronic structure-type descriptors are also proposed for understanding the structure–property–performance relationship of perovskite oxides towards high-temperature ORR. According to Morgan's inference, the experimentally measured ASRs and k* values of the perovskite oxide cathodes for SOFCs are strongly correlated with the DFT-calculated O 2p-band center and oxygen vacancy formation energy, including layered perovskite oxides, PrBCO and GdBCO.245 As shown in Fig. 6c, the log of the ASRs is linearly correlated with the calculated bulk O 2p-band centers, suggesting the effectiveness of the O 2p-band center as a descriptor for ORR activity. The BSCF cathode with the lowest ASR is located at an O 2p-band center of ∼1.5 eV. A united trend is observed in the log of the k* values vs. O 2p-band centers. The O p-band descriptor can be qualitatively understood in terms of a rigid band model, which is used to describe some electron-rich perovskite oxides. Upon the addition of O to the material, electrons move from the EF to the O p-band, and oxygen removal corresponds to the electrons moving in the opposite direction (Fig. 6d). Because the O adsorption/desorption processes govern many aspects of ORR, the O p-band center is considered to be an effective descriptor for ORR and its relevant steps. As correlated between the calculated O 2p-band centers and predicted k* values, 2145 distinct perovskite oxides have been screened as highly active and stable SOFC cathodes via high-throughput DFT computation.246 The screening method qualitatively reproduces the experimental activity, stability, and conduction nature of well-studied cathode materials, among which 52 potential cathode materials with good predicted stability under SOFC operating conditions and predicted k* on par with first-class perovskite oxide cathodes are obtained. The calculated O p-band center is also used as a first principle-based descriptor of k*, which in turn correlates with the ORR activity (Fig. 6e), further verifying the reliability of the O p-band center. Grimaud et al. found that the alkaline OER properties of double perovskite oxides, (Ln0.5Ba0.5)CoO3−δ (Ln = Pr, Sm, Gd, and Ho), can be correlated likewise with the O p-band center.247 The high activity and stability can be explained by having the O p-band center neither too close nor too far from the EF. The η values (0.5 mA cmox−2) of these double perovskite and leading pseudocubic perovskite oxides are well correlated with the computed O p-band centers, as depicted in Fig. 6f. Moving the O p-band center close to the EF from LaCoO3 to (Pr0.5Ba0.5)CoO3 promotes the intrinsic OER activity. However, further lifting the O p-band center gives rise to a decrease in the activity and stability. No visible changes are observed in the perovskite oxides on the left branch during the OER process, whereas rapid amorphization in the near-surface regions occurs for the ones on the right branch, accompanied by leaching of the A-site ions. As a consequence, an active and stable (Pr0.5Ba0.5)CoO3 OER catalyst is obtained upon water oxidation in alkaline solution. However, this prediction largely depends on the DFT computation, given that discrepancies exist among the predictions made using different DFT exchange and correlation functionals and from experiments. Jacobs et al. compared five different DFT exchange and correlation functionals including Perdew–Burke–Ernzerhof (PBE), PBEsol, PBE+U, strongly constrained and appropriately normed functional (SCAN), and Heyd–Scuseria–Ernzerhof (HSE).248 The best correlations for all the measured ORR and OER activities, experimental log
J (1.6 V vs. RHE for OER) and log
k*, are made with PBE-level calculations with strong observed linear correlations as a function of bulk O p-band (Fig. 6g and h), respectively. In an account, Shao-Horn et al. reminded us of the potential limitations of this approach in practice.249 A specific limitation of the bulk O p-band center descriptor is that it is more applicable to predicting surface termination-sensitive phenomena. Given that the bulk and surface O p-band center are directly correlated with each other, the bulk descriptor provides a robust trend with surface-sensitive phenomena. More prevalent material databases and machine learning methods offer a chance to explore the descriptor-based mode for discovering and designing electrocatalysts.
Employing soft X-ray emission and absorption spectroscopy, the partial density of states of perovskite oxides are analyzed on an absolute energy scale, and the energy barriers for electron transfer and surface deprotonation are correlated with their OER activity.250,251 The charge-transfer energy is defined as the energy difference between the orbital center of the O 2p orbital and the unoccupied metal 3d orbital,252,253 which plays a pronounced role in adjusting the properties of oxides, e.g., delocalization of electrons and the redox couple of oxides.254–256 According to the plot of J (1.6 V vs. RHE for OER) as a function of change-transfer energy, reducing Δ greatly enhances the OER activity, and the trend passes through all materials from semiconducting to semi-metallic perovskite oxides (Fig. 7a), including layered perovskite oxides (PrBCO, SmBCO, and GdBCO).251 The Fermi level lies in forbidden energies and is poorly defined for semiconducting and insulating oxides. Therefore, the O p-band center is less reliable to describe the OER activity than the charge-transfer energy of all types of perovskite oxides. From semiconducting to semi-metallic perovskite oxides, a decrease in Δ corresponds to an increase in their linear slope relative to their OER activity (Fig. 7a). The change in slope is a typical indicator of a shift in the rate-limiting step. Perovskite oxides with high Δ values have lower activation energies for proton transfer, indicating that the concerted pathway instead becomes the electron-transfer step. When Δ decreases, the perovskite oxides employ a decoupled proton–electron transfer pathway. As Δ decreases further in the case of semi-metallic perovskite oxides, the rate of the OER is limited by a proton-transfer step. It can be seen that a decrease in Δ promotes the electrocatalytic activity, with a change in the OER mechanism from electron-transfer-limited to proton–electron-coupled and proton-transfer-limited reactions. The partial density of states (PDOS) of these 10 perovskite oxides are examined on an absolute energy scale to grasp the function of Δ on physicochemical properties and OER activity. All the perovskite oxides have similar O 2p energy levels, and their Δ values are from the unoccupied 3d-band of their transition metal (Fig. 7b).251 The 3d-band-relevant descriptors (e.g., oxidation state and number of d electrons) should correlate with OER activity because they have a strong effect on the 3d-band position.257 As dictated by Δ, the band positions can decouple the electronic and chemical interactions at the electrode/electrolyte interface for the OER, i.e., electron-transfer kinetics, hydroxide affinity, and adsorbate binding (Fig. 7c and d).251 The electron affinity of the oxide to the electrolyte corresponds to the Schottky barrier for an electron-transfer step. The perovskite oxides with a higher Δ value have a higher Schottky barrier for electron transfer at the electrode/electrolyte interface (Fig. 7c and d). With a decrease in Δ, the unoccupied energy level falls below the OER potential energy without the energy barrier for electron transfer (Fig. 7d). Clearly, semi-metallic perovskite oxides with low energy barriers for electron transfer converge to a Tafel slope of ∼60 mV dec−1. In contrast, semiconducting perovskite oxides with high Δ values have high Tafel slopes of 100–180 mV dec−1.
![]() | ||
| Fig. 7 (a) Correlation between the OER J at 1.6 V vs. RHE and the charge-transfer energy. The charge-transfer energy is defined differently for conductors (open) and semiconductors (closed). Oxide abbreviations: La0.8Sr0.2CoO3−δ (LSCO20), La0.5Sr0.5CoO3−δ (LSCO50), GdBaCo2O5+δ (GBCO), SmBaCo2O5+δ (SBCO), and PrBaCo2O5+δ (PBCO). (b) Electron transfer energy at the electrolyte-oxide interface, hydroxide affinity of the oxide, and oxide metal–oxygen bond strength estimated from X-ray spectroscopic data under rigid band assumption. Schematic of the features used to calculate the electron transfer energy, hydroxide affinity, and charge-transfer energy of the oxides, and trends in the band positions of the perovskite oxides (relative to vacuum) from the lowest to highest OER activity are shown on the right. The arrows are drawn to show a positive difference. The standard thermodynamic sign convention (negative energy is stronger affinity) is used. For conductors, the EF and unoccupied transition metal 3d-O 2p states are identical. The dashed line indicates the O2 and H2O/OH− redox level at pH 13. (c) Surface properties-Tafel slope, pH of zero charge, and adsorbate binding energy of HO* intermediates relative to O* on the transition metal site-relative to the electron transfer barrier, hydroxide affinity, and charge-transfer energy, respectively. (d) Trends in the electron transfer energy, hydroxide affinity of the perovskite oxide surfaces extracted from the band positions as a function of charge-transfer energy, as well as relative adsorbate binding energy of intermediates. Reproduced with permission.251 Copyright 2017, the Royal Society of Chemistry. | ||
Currently, a universal descriptor-based approach has not been established to correlate the HER activity. Guan et al. bridged the relationships between OER/HER activities and over ten representative material properties of 12 3d metal-based perovskite oxides (Fe, Co, Ni, and Mn).258 Among these property parameters, the charge-transfer energy can serve as an ideal unifying descriptor, where neither too high nor too low Δ (∼1 eV) is related to the great electrocatalytic activity, fulfilling Sabatier's principle. This principle states that the most efficient electrocatalyst binds the reaction intermediates neither too strongly nor too weakly.250,259 The correlations between macroscopic physicochemical properties (hydrophilicity, aerophobicity, and conductivity) and the OER/HER activities are first studied in 1.0 M KOH solution. The OER/HER η values at 200 µA cmox−2 are used for comparison. The methodologies of significance analysis and regression analysis are applied,260,261 where the degree of correlation between properties and activities can be quantified by the obtained p value. A small p value (≤0.05) means a significant property–activity relationship. The correlation degrees exclude the possibilities of solid–liquid contact angle and electron transfer as universal activity descriptors. However, the solid–gas contact angle can be a potential descriptor. Secondly, the correlations between molecular-level structural properties (bond length, cell volume, and strain) and the OER/HER activities are discussed. The low correlation degrees indicate that these structural features are unsuitable as ideal descriptors. Then, the electronic structural properties of the selected perovskite oxides are correlated with their intrinsic OER/HER activities, such as 3d metal valence state, charge-transfer energy, O 2p-band center, and conduction band gap. The correlations between Δ and OER/HER activities are significant. Systematic experiments and computations untangle that the La0.5Sr0.5CoO3−δ perovskite oxide (Δ ≈ 1 eV) has metal-like high-valence configurations with an LOEM for the OER/HER. However, 12 representative cation-disordered perovskite oxides are considered in this work. Thus, it is hoped that the universality of the Δ descriptor for layered perovskite oxides will be further validated.
To figure out the structure–property–performance relationship, the proper descriptors should be established to guide the design of needed electrocatalysts. As mentioned above, the correlations between the structure/properties of perovskite oxides and their electrochemical performance (ORR, OER, and HER) are discussed with several descriptors, such as eg occupancy, adsorption free energy, O 2p-band center, A-site ionic electronegativity, and charge-transfer energy. Although these parameters are referred to in some studies, they are not rigorously validated across a variety of compositions or experimental conditions. Each is directed against selected materials (specific perovskite series) and has merits and demerits. The accuracy of the proposed descriptors needs to be further verified. The individual proposed descriptor has a limitation in the performance statistics of all perovskite materials. Thus, a unified descriptor is lacking for different perovskite series. Considering the wide physicochemical space of perovskite oxides, machine learning, high-throughput DFT computation, and artificial intelligence represent powerful tools for unveiling their structure–property–performance relationship and searching for new perovskite catalysts.246,262,263 The conventional trial-and-error process for sample synthesis, physico-chemical characterization, and performance testing takes very long. Alternatively, an experimentally proven machine-learning-driven method is demonstrated to accelerate the discovery of efficient air electrodes for SOFCs, where ionic Lewis acid strength (ISA) is adopted as an effective descriptor for the ORR activity of perovskite oxides.262 Four perovskite oxides, Sr0.9Cs0.1Co0.9Nb0.1O3 (SCCN), Ba0.4Sr0.4Cs0.2Co0.6Fe0.3Mo0.1O3 (BSCCFM), Ba0.8Sr0.2Co0.6Fe0.2Nb0.2O3 (BSCFN), and Sr0.6Ba0.2Pr0.2Co0.6Fe0.3Nb0.1O3 (SBPCFN), screened from a huge amount of perovskite compositions (6871 distinct materials), are experimentally synthesized and have superior ORR activities. The overall workflow for the identification of high-performance oxygen electrodes is illustrated in Fig. 8a, including machine learning model training, materials screening, experimental verification, and computational analysis.262 According to the experimental characterization, decreased A-site and increased B-site ISAs in the perovskite oxides can improve their surface oxygen exchange kinetics. Among the four as-synthesized perovskite oxides, SCCN in particular has excellent ORR activity with extremely low ASR of 0.088 Ω cm2 at 550 °C and the lowest Ea (107 kJ mol−1) (Fig. 8b). The Ni/SDC cermet-anode-supported single cell delivers the PPDs of 2.05 W cm−2 at 650 °C (Fig. 8c). According to closed-loop experiments and active learning, a champion four-metal perovskite oxide, Ca0.8Pr0.2Co0.8Fe0.2O3−δ (CPCF), emerged from 10101 perovskite oxide candidates for oxygen-evolving electrocatalysis.263Fig. 8d shows the predicted η values of the ABO3 structures with Ca/Pr at the A-site and Co/Fe at the B-site, namely, CaxPr1−xCoyFe1−yO3. The most promising composition, CPCE, is chosen from the list of candidates. This OER electrocatalyst exhibits an intrinsic η of 391 mV at 10 mA cmoxide−2 (Fig. 8e), which is the lowest value among the four-metal perovskite oxides. However, the prediction of electrocatalytic properties with data-driven machine learning is still in the primary stage, and only a few studies used this technique to predict the ASR, η, and oxygen-ionic conductivity.262–266 Jacobs et al. developed a machine learning model to predict the catalytic properties of perovskite oxides for SOFCs/SOECs, including oxygen surface exchange rate, bulk diffusivity, and ASR. Compared with the models based on ab initio-derived features, these developed models are based on trivial-to-calculate elemental features, which are more accurate and faster.267 A schematic outline of the present work is illustrated in Fig. 9a, in which more mature data-centric approaches can be utilized to discover new materials and understand high-performing materials for SOFCs/SOECs, i.e., electrocatalytic ORR and OER applications. For example, the developed machine leaning model is used to calculate the ASRs of the perovskite oxides, together with calculations of materials cost and stability, screening new promising catalysts from a total of 19072821 materials. To find novel promising perovskite catalysts, the screening criteria are set to be values below a threshold of ASR, cost, and stability, including 1.33/0.21 Ω cm2 (LSCF and BSCF at 500 °C), 133.67 dollars kg−1 (LSCF), and 93.3 meV atom−1 (SCCN at 500 °C), respectively. Fig. 9b–d show the distributions of predicted log
ASR (500 °C), cost, and stability. It can be seen that 2135396, 2453872, and 1393424 materials separately pass the screening criteria of ASR, cost, and stability, respectively. Finally, 9135 (∼0.05%) materials meet all the screening criteria. From the list of screened materials, the most highly active, cheapest, and most stable candidates are SrCo0.75Nb0.125Ta0.125O3 (log
ASR = −0.43 Ω cm2), BaFe0.75Cu0.125Zr0.125O3 (1.15 dollars kg−1), and BaFe0.5Co0.25Mo0.25O3 (18.0 meV atom−1), respectively.267 As shown in Fig. 9e, SrZr0.125Nb0.125Co0.625Cu0.125O3 (SZNCCu) (log
ASR = −0.37 Ω cm2), K0.25Sm0.125Sr0.625Nb0.125Ta0.125Co0.75O3 (KSmSCNT) (log
ASR = −0.33 Ω cm2), and Bi0.125Sr0.875Y0.125Ni0.125Co0.75O3 (BiSYNC) (log
ASR = −0.25 Ω cm2) are predicted to outperform high-performing BSCF and SCCN, which are consistent with the experimental ASR values. Furthermore, they have lower Ea values than BSCF, indicating their performance will continue to be utilized <500 °C.267 In summary, to establish a universal electrocatalysis descriptor for perovskite oxides, efforts need to be devoted to all-type perovskite oxides.
![]() | ||
| Fig. 8 (a) The overall workflow diagram: machine learning model training, material screening, experimental verification, and DFT analysis. θ is X-ray incidence angle. (b) Arrhenius-type plots of ASR values of SCCN, BSCCFM, BSCFN, and SPBCFN. (c) Current density–voltage-power density curves of a single cell with the configuration Ni/SDC|SDC|SCCN at 450–650 °C. Reproduced with permission.262 Copyright 2022, Nature. (d) The predicted overpotential at 10 mA cmoxide−2 of CaxPr1−xCoyFe1−yO3 (0 ≤ x, y ≤ 1) oxides according to the precursor mixture ratio of Ca, Pr, Co and Fe. (e) Quantification of the electrochemical properties of CPCF compared to standard catalysts: linear sweep voltammetry curves. The inset shows Tafel slopes. Reproduced with permission.263 Copyright 2023, Nature. | ||
![]() | ||
Fig. 9 (a) Schematic of the present work of using data-centric machine learning approaches for predicting catalytic properties. Violin plots showing the distributions of screened materials: (b) screened ASR, (c) screened materials cost, and (d) screened materials stability. The numbers above each distribution denote the number of materials passing the given screening combination. The high, middle, and low colored ticks denote the maximum, median, and minimum of the distribution, respectively, while the black ticks denote the mean of the distribution. (e) Machine learning-predicted ASR temperature dependence for key materials. The solid lines are machine learning predictions using predicted log ASR value at 500 °C together with the machine learning model of predicted ASR barrier to scale the prediction to other temperatures. The error bars are the calibrated one standard deviation error bars from the machine learning model. Data points are experimental ASR values. Reproduced with permission.267 Copyright 2024, Wiley. | ||
Comparatively, the η values of layered LnBCO perovskite oxides are ∼−120–550 mV for alkaline HER at 10 mA cm−2, while η ≈ 310–570 mV for alkaline OER (Table 3). These activities are still much lower than that of the RuO2 (IrO2) and Pt/C benchmarks (Table 3). In the case of the HER, the computational H* adsorption energy on the Pt (111) surface (ΔGH* = −0.09 eV) is near to the ideal 0 eV,271 which is consistent with Pt being the most efficient HER electrocatalyst. However, this compares with the higher adsorption energy (0.62 eV) at the Co site on the PrBaCo2O6 (100) surface.242 Surprisingly, layered perovskite oxides outperform Pt/C upon a high current density.238,240–242 A-site cation-defect engineering demonstrates a highly active perovskite oxide (Pr0.4La0.5BaCo2O5+δ).241 The HER η is ∼−636 mV at 2000 mA cm−2, whereas η ≈ −974 mV for the commercial Pt/C catalyst. This tendency is in accordance with the OER experiments with the IrO2 benchmark.238,241 Furthermore, it preserves its stability even when operated at 500 mA cm−2 and undergoes a mild degradation rate of ∼0.5 mA cm−2 h−1, whereas Pt/C suffers from severe activity decline of ∼20.7 mA cm−2 h−1. In the ABO3 perovskite oxides, the A-site Ba2+ ions are easily leached from the surface region to the electrolyte due to the ionic bonding characteristic of Ba–O.272 The progressive amorphization breaks the charge transfer and suppresses the electrocatalytic activity. In the case of layered perovskite oxides, no elemental segregations are observed on the surface of Pr0.4La0.5BaCo2O5+δ after a continuous 10 h HER, maintaining a crystalline structure without any amorphous films.240 We infer that A-site Ln3+ ions can aid the stabilization of the Ba–O network. Regarding practical water splitting, the NiFe-LDH‖Pr0.4La0.5BaCo2O5+δ electrolyzer requires a voltage of 2.15 V to drive an ultrahigh J of 2000 mA cm−2, outperforming the IrO2‖Pt/C couple (2.45 V).240 As-reviewed layered perovskite oxides probably solve the problem of low efficiency at high current densities faced by noble metal catalysts. Aside from the electrocatalysts, the practical devices should fully take into account realistic production conditions, such as high current densities, long-term operation, high temperatures, and pressures.273 In laboratory evaluation, the stability test is continued for tens of hours at 10 mA cm−2 at 20–30 °C, and the industrial production is conducted at 200–2000 mA cm−2 under an operating pressure of 0.4–5.0 MPa for thousands of hours. These rigorous conditions probably cause catalyst failure, such as ion leaching, structural evolution, and electrode exfoliation, which cannot be monitored in laboratory-level tests. Therefore, despite the high catalytic performance (activity and stability) actualized in the laboratory, the performance evaluation should be supplied under industrial-grade conditions.
When assessed as oxygen electrodes for SOCs, layered perovskite oxides still possess strong stability at high temperatures. The classic LSCF oxygen electrode has poor durability due to Sr segregation on its surface,274,275 which is caused by the electrostatic attraction of positively charged oxygen vacancies.276 The Sr-enrichment induces the creation of other phases, such as SrCO3 and Sr(OH)2, giving rise to degradation of the electrode performance. Proven by experiments, elemental segregations appear in A-site cation-deficient PrBCO perovskites, forming (PrBa)CoO3 and BaO phases on their surface.100,178,242 Despite in situ exsolution on their surface, these composites influence the ORR activity in a positive way. Exsolved (PrBa)CoO3 perovskite oxides enhance the specific surface areas, electrical conductivity, oxygen adsorption/desorption, and transport abilities.178,242 As an insulator, amorphous BaO nanoparticles can physically adsorb the reaction substances, and then react with surface defects to participate in electrocatalysis.277 Compared to the stability of Ln0.94BaCo2O5+δ, the degradation rates of these oxygen electrode-based fuel cells follow the sequence of Sm0.94BaCo2O5+δ > Pr0.94BaCo2O5+δ > La0.94BaCo2O5+δ.100 After heating treatment, the BaO nanoparticles are in situ exsolved from the bulk. The Ba-segregation on the surface might be associated with the instability of Ln0.94BaCo2O5+δ. Long-term and stepwise tests confirm the steady operation of the LaBCO oxygen electrode in a fuel cell and electrolysis cell, surpassing the NdBCO and GdBCO oxygen electrodes. We deem that the degradation mechanism of layered perovskite oxides is elemental (barium) segregation. Cubic–tetragonal–orthorhombic structural evolution is identified with a decrease in the ionic size of Ln3+, accompanied by a decrease in the Co valence state and increase in oxygen vacancies. These two atomic-scale factors directly affect the catalytic pathways. In Co-containing perovskite oxides, the small polaron hopping mechanism dominates the charge transfer, i.e., the hole carriers of Co4+. Among the components, LaBCO shows the highest conductivity with an average Co valence state of ∼+3.36 and oxygen content of ∼5.86.43 The highest Co valence state corresponds to the best ORR activity of LaBCO, i.e., the highest electronic (hole) conductivity. The performance gradually weakens with a decrease in the ionic size of Ln3+. The general activity follows the sequence of cubic > tetragonal > orthorhombic perovskite oxides. As is known, oxygen-ionic conductivity is key to the ORR activity of MIECs, which is usually controlled by the vacancy mechanism. However, the interstitial oxygen and lattice oxygen mechanisms are also observed in perovskite oxide-like electrocatalysts.278 Although LaBCO has a lower oxygen-vacancy concentration, the oxygen bulk-diffusion process involves hole migration.279 The calculated oxygen-ionic conductivity of LaBCO is 1.65 S cm−1 at 800 °C, which is higher than that of the popular BSCF perovskite oxide (0.96 S cm−1).43,280 Accordingly, the vacancy mechanism is not supposed to be the principal pathway, and the active lattice oxygen mechanism may exist in the electrocatalysis reactions.278 Moreover, the more positive Co4+ sites prefer to adsorb oxygen species due to strong electrostatic affinity. As the initial step, abundant adsorbed oxygen species facilitate the ORR kinetics. A consistent performance trend is also discerned in the LnBCO series for alkaline HER.44 Therefore, advantageous oxygen-ionic conductivity, adequate electronic conductivity, and high average Co valence state play critical roles in the electrocatalytic activity of LnBCO. In all rare earth ions, the ionic radius of La3+ is closest to that of Ba2+. Given the strongest stability of LaBCO, the smaller size mismatch between La3+ and Ba2+ can reduce the segregation level due to the elastic energy difference, enabling more a stable electrode surface.276
In addition to sufficient ORR activity, the reasonable durability of the oxygen electrode is vital in practical SOFC and SOEC devices. If the oxygen electrodes are exposed to ambient air, they inevitably encounter some contaminants, such as ∼1 vol% CO2. Chen et al. reported a Ca-doped PrBCO oxygen electrode (PrBa0.8Ca0.2Co2O5+δ) with excellent ORR activity and strong CO2 tolerance.110 In air with ∼1 vol% CO2, this electrode can retain its ASR value after a 1000 h operation at 750 °C. According to impedance spectroscopy and in situ surface-enhanced Raman spectroscopy, the surface of PrBa0.8Ca0.2Co2O5+δ is more active for oxygen exchange and more robust against CO2 than that of LSCF. As supported by DFT calculations, PrBa0.8Ca0.2Co2O5+δ exhibits much weaker adsorption of CO2 than LSCF (−0.73 eV versus −1.25 eV), suggesting its CO2-tolerant characteristic. A highly active and CO2-tolerant oxygen electrode, Pr0.94Ba0.7Ca0.3Co2O5+δ, has been developed by our group.105 After annealing in air with 10 vol% CO2 for 12 h at 700 °C, the layered perovskite structure is still retained without any secondary phases. In turn, the carbonate of BaCO3 can be detected in undoped Pr0.94BaCo2O5+δ. Several influencing factors contribute to the CO2 tolerance of the oxides, such as average metal–oxygen bond energy (ABE), acidity of cations and defects, and oxygen-vacancy concentration.281–283 The ABE values can be calculated using the following equation:284
| 〈ABE〉 = 〈A–O〉 + 〈B–O〉 | (18) |
| 〈A–O〉 = (xA/nNA) × (ΔHAnOm − nΔHA − mDO2/2) | (19) |
| 〈B–O〉 = (xB/nNB) × (ΔHBnOm − nΔHB − mDO2/2) | (20) |
In summary, layered perovskite oxides exhibit much lower HER (OER) activities than that of the state-of-the-art Pt/C and RuO2 (IrO2) catalysts. However, their performance exceeds noble metal-based catalysts at large current densities, and can be maintained for a long working time. Not evident cation leaching and amorphous film are observed in the as-tested LnBCO electrocatalysts. Compared with Pt/C and RuO2 (IrO2), their intrinsic activity and stability preliminarily meet the requirement of industrial-grade water electrolysis (rapid and large-scale hydrogen production). Next, the studies should be concentrated on water electrolysis under practical conditions. Regarding the applications of SOCs, the electrocatalytic ORR activities of layered perovskite oxides are closely related to the structural evolution. Highly symmetrical cubic perovskite oxides are favorable for the ORR activity and stability, with a higher average Co valence state and mixed ionic and electronic conductivity. More highly oxidative Co4+ species mean more hole carriers, leading to higher electrical conductivity. An active lattice–oxygen mechanism with fast charge transfer rate boosts the ORR kinetics. During long-term operation, the performance degradation originates from the segregation of Ba on the surface. The smaller size mismatch between La3+ and Ba2+ reduces the segregation level, producing a more stable LaBCO oxygen electrode.
000 tons, of which 98% is distributed in South Africa and Russia. The reserves of ruthenium and iridium are even scarcer, only ∼5000 and 1600 tons, respectively. At the current mining velocity, the amount of iridium will satisfy 15 years of mining. Thus, this scarcity, coupled with their harsh extraction conditions (high temperatures and high pressures), determines the high costs of noble metal catalysts. At present, the international prices of Pt, Ru, and Ir are ∼42.5, 31.5, and 51.5 dollars g−1, respectively. The progress in large-scale water electrolysis has been limited by the trade-off between the high cost of platinum-group-metal catalysts and the limited performance of earth-abundant metal alternatives. Considering the production costs and industrial-grade efficiency, this review summarizes a family of layered LnBCO perovskite oxides for addressing cost-effective, high-performance, and durable electrocatalysts in sustainable and clean energy fields.
The final stage of hydrogen production is “green hydrogen”, which is produced through the usage of renewable energy sources. For example, relatively mature water electrolysis for hydrogen production can realize complete decarbonization. Depending on their membrane materials, electrolyzers can be divided into alkaline water electrolysis (AWE), proton exchange membrane water electrolysis (PEMWE), anion exchange membrane water electrolysis (AEMWE), and SOEC water electrolysis (SOECWE). KOH electrolyte is used in alkaline water electrolysis cells. Differing from AWE, PEMWE uses a perfluorosulfonic acid proton exchange membrane as a solid electrolyte with good chemical stability, proton conductivity, and gas separation. AEMWE is technology for hydrogen production using an anion exchange membrane, with decreased dependence on noble metals and equipment costs. High-temperature SOECWE uses solid oxides as the electrolyte materials, with a working temperature of 700–1000 °C. The electrochemical performance of electrocatalysts is significantly improved, and the efficiency is higher. The electrodes of SOECs are non-noble metal catalysts. The cathode and the anode use porous metal ceramic (Ni/YSZ) and perovskite oxide, respectively, and the electrolyte uses an oxygen-ionic conductor (YSZ). An all-ceramic solid structure avoids material corrosion problems. According to laboratory-level experiments, LnBCO electrocatalysts are more adapted to AEMWE and SOECWE. In the case of AEMWE, their HER/OER performance (activity and stability) is greater than that of Pt/C and IrO2 (RuO2) catalysts. In high-temperature SOECs, LnBCO-based oxygen electrodes outperform the popular ABO3 perovskite oxides, demonstrating excellent OER activity between 700–800 °C. The electrolysis J can reach ∼2–3 A cm−2 at 1.8 V for the CO2RR at 800 °C, along with satisfactory long-term and cycling stabilities. However, these perovskite oxides are ordinarily prepared at high temperatures (∼1000 °C). The as-prepared electrocatalysts are on a micron scale with irregular shapes, and thus it is difficult to produce a homogeneous electrode ink for AWE and PEMWE. In industrial-grade water electrolysis, the production rate of hydrogen is very rapid, with massive bubbles on the electrode surface. Under extreme production conditions, the perovskite oxide electrocatalysts might be peeled off from the metal substrate. In this context, uniform nanomaterials have an unique merit. We should be dedicated to exploring nanoscale perovskite oxides, while maintaining their high electrochemical performance. In the case of SOECWE application, the practical working environment (high temperature and high humidity) limits the choices and efficiencies of materials for the electrolysis cells, and also restricts large-scale promotion. Presently, the electrochemical test is at the primary experimental stage, and the scalability of these electrocatalysts still has no clear conclusion. Encouraged by the production costs, electrocatalytic activity, and operating stability, we anticipate the expanding the scale of layered perovskite oxides for practical usage.
The diversity and flexibility of the elemental components, crystal structures, and electronic structures of layered perovskite oxides bring new opportunities to extend their electrocatalysis scope. Besides the well-investigated ORR for O2−-SOFCs, recent work demonstrates highly active PrBCO-based oxygen electrodes for oxygen-conducting SOECs (O2−-SOECs), proton-conducting SOFCs (H+-SOFCs), and proton-conducting SOECs (H+-SOECs), such as Pr1.5Ba0.5Co2O5+δ, (La0.25Pr0.25Nd0.25Sm0.25)Ba0.5Sr0.5Co1.5Fe0.5O5+δ, PrBa0.5Sr0.5Co1.8Ni0.2O5+δ, PrBa0.5Sr0.5Co1.5Fe0.5O5+δ, and PrBaCo1.9Hf0.1O5+δ.288–292 Regarding extended electrocatalysis reactions, some perovskite oxides have been designed carefully, such as LaMnO3-based perovskite oxides for MOR,293,294 La0.6Sr0.4CoO3 for MOR,295 Eu2FeCoO6 for EOR and benzyl alcohol oxidation reaction (BOR),296 Sr2CuWO6 for CO2RR,297 LaMn0.6Co0.4O3, LaFe0.9Cu0.1O3, and (Ba0.5Sr0.5)0.85Co0.8Fe0.2O3−δ for NO3RR.298 In the reverse water–gas shift reaction, low-temperature CO2 conversion to CO can be catalyzed by the synergistic effect of binary Ni–Fe sites in double perovskite oxides (La2NiFeO6).299 To the best of our knowledge, there is little literature on LnBCO electrocatalysts for extended electrocatalysis and non-electrocatalysis reactions. As reported by Su et al., PrBCO can catalyze the generation of radicals from peroxymonosulfate for the oxidative degradation of organic waste in aqueous solution.300 In the PrBCO catalyst, the easy valence-state change of the B-site Co ions and intrinsic oxygen vacancies could mediate a redox process and be bonded with peroxymonosulfate, respectively. This work puts forward a new catalysis application for layered perovskite oxides. Heterogeneous noble-metal-based catalysts show a superior performance to noble metal-free ones in the above-mentioned reactions, which are adopted as benchmarks for direct performance comparison. Several characteristics of LnBCO make them more suitable for an extended electrocatalysis scope compared with noble-metal-based ones. (1) Their elemental compositions can be tailored vs. facile ion substitution, actualizing specific active sites for catalysis selectivity, e.g., complex MOR and EOR. (2) The synergistic effects of various metal sites in medium-entropy (high-entropy) perovskite oxides might realize multifunctional electrocatalytic activities, e.g., Fe–Co-based LnBCO electrocatalysts for ORR, OER, and HER. (3) Their crystal and electronic structures can be tuned via composition, size, and dimension modulation. Their structural evolution, electron-filled state, Co valence state, and oxygen vacancy defects could be tuned for a variety of electrocatalysis reactions. This goal would be given attention to realize new catalysis properties in layered perovskite oxides.
| This journal is © the Partner Organisations 2026 |