Research ArticleMATERIALS SCIENCE

Exceptional oxygen evolution reactivities on CaCoO3 and SrCoO3

See allHide authors and affiliations

Science Advances  09 Aug 2019:
Vol. 5, no. 8, eaav6262
DOI: 10.1126/sciadv.aav6262

Abstract

We investigated the roles of covalent bonding, separation of surface oxygen, and electrolyte pH on the oxygen evolution reaction (OER) on transition metal oxides by comparing catalytic onset potentials and activities of CaCoO3 and SrCoO3. Both cubic, metallic perovskites have similar CoIV intermediate spin states and onset potentials, but a substantially smaller lattice parameter and shorter surface oxygen separation make CaCoO3 a more stable catalyst with increased OER activity. The onset potentials are similar, occurring where H+ is removed from surface -OH, but two competing surface reactions determine the catalytic activity. In one, the surface -O is attacked by electrolyte OH to form the surface -OOH; in the other, two -O form a surface peroxide ion and an oxygen vacancy with electrolyte OH attacking the oxygen vacancy. The second pathway can be faster if the surface oxygen separation is smaller.

INTRODUCTION

The air electrodes of electrochemical water-splitting systems and metal-air rechargeable batteries require a high catalytic activity at room temperature of the oxygen evolution reaction (OER) and long-term structural stability of the catalyst in an alkaline medium (15). Traditional OER catalysts of acceptable catalytic activity at room temperature such as carbon-supported noble metals and RuO2 or IrO2 are too expensive and/or unstable under anodic (charging) conditions (68). Low-cost transition-metal oxides have been demonstrated to be promising OER catalysts (915). AMO3 perovskites are attractive for the study of the OER because their electrochemical potential and electronic properties can be systematically controlled by cation substitutions on both the A and M sites with different valences and ionic sizes (1618), but the relative merits of π* antibonding versus σ* antibonding orbitals of d-electron symmetry, the degree of metal d and oxygen 2p orbital mixing at the active sites, the separation of the surface oxygens, and the pH dependence of the catalytic activity have been little explored.

Here, we report a comparative study of the catalytic activities of two isostructural ACoO3 (A = Ca, Sr) perovskites that have been synthesized under high pressure and characterized physically (19, 20). Each perovskite is cubic and metallic; they exhibit long-range magnetic order, indicating a t4σ*1 intermediate spin state on a CoIV formal valence state. The π* antibonding t4 electrons are localized in SrCoO3; they are at the crossover from localized to itinerant behavior in CaCoO3, which has a notably shorter Co─O bond length than that of SrCoO3 (1.87 Å versus 1.92 Å). The itinerant character of the σ* electrons is responsible for the metallic conductivity and the lack of a cooperative Jahn-Teller distortion of the CoO6/2 sites that would be expected with localized e1 electrons.

The oxygen emitted by the OER comes from the alkaline medium either directly or indirectly via an extraction of oxygen from the catalyst that is replaced from the medium. Which route is taken depends on the separation of the surface oxygen, the zeta point of the catalyst, and the pH of the alkaline medium; at the zeta point; the adsorbed water or OH on oxide particles immersed in an aqueous medium does not change the pH of the medium (1). Acidic oxides immersed in pH 7 water make the water acidic by losing surface H+ to medium H2O forming H3O+ ions; basic oxides attract H+ from a pH 7 medium, -O2− + H2O = -OH + OH, to render the medium alkaline. In a strongly alkaline medium, the ACoO3 perovskites are acidic, i.e., they lose surface H+ to the medium: -CoOH + OH = -CoO2− + H2O. In oxidation in an anodic current, -CoO2−-e = -CoO, the O ion is vulnerable to attack by a medium OH to form -CoOOH or, alternatively, to the reaction -2CoO = -Co(O2)2− + Co-□ followed by occupancy of the surface oxygen vacancy □ either by medium OH or by a lattice oxygen. The creation of the surface CoO species may determine the onset potential, but the subsequent OER activity may depend on which subsequent route is faster and, therefore, on the separation of the surface oxygen. These competing interactions would give rise to different ratios of the catalyst and medium oxygen in the evolved O2. By using the 18O isotope in the catalyst, Grimaud et al. (21) have found zero, one, or two 18O in the evolved O2, which indicates that the two routes are both operative and that the separation of the surface oxygen may determine the activity of the OER following the onset potential.

RESULTS

Crystal structure of ACoO3 (A = Ca, Sr)

The Rietveld refined powder X-ray diffraction (XRD) patterns of high-pressure ACoO3 (A = Ca, Sr) and their structural information are shown in Fig. 1 and table S1. Both CaCoO3 and SrCoO3 have a cubic perovskite structure (space group: Pm-3m) with corner-shared CoO6/2 octahedra and Ca2+ or Sr2+ on the 12-fold coordinated A sites. CaCoO3 has the smaller lattice parameter (table S1) because of the smaller ionic radius of Ca2+ than Sr2+ ions, and the nearest Co─O bond length is decreased by Δr ≈ 0.05 Å from SrCoO3 to CaCoO3. The particle sizes of the ACoO3 (A = Ca, Sr) were about 10 μm (fig. S1), and energy-dispersive X-ray spectroscopy mapping in fig. S2 revealed a uniform distribution of Ca, Sr, Co, and O elements in the particles. A high-resolution transmission electron microscopy (TEM) image of ACoO3 in Fig. 1 (B and C) confirmed a homogeneous and well-crystallized structure of the surface and bulk of ACoO3 (A = Ca, Sr). The oxygen stoichiometry of each ACoO3 (A = Ca, Sr) was studied with cyclic voltammetry (CV); Co3O4 and LaCoO3 with no oxygen vacancies were tested for comparison (fig. S3). The reversible reduction/oxidation peaks at 1.0 to 1.2 V versus a reversible hydrogen electrode (RHE) in the CV of SrCoO3 correspond to the insertion (oxidation) reaction, SrCoO3−x + 2xOH → SrCoO3 + xH2O + 2xe, indicating that some oxygen vacancies existed; the absence of these peaks in CaCoO3, Co3O4, and LaCoO3 confirmed their stoichiometric composition.

Fig. 1 The structure of ACoO3 (A = Ca, Sr).

(A) Observed, calculated, and difference patterns for the Rietveld refinement from the XRD of cubic ACoO3. a.u., arbitrary units. (B and C) TEM images of ACoO3 (A = Ca, Sr).

OER performance of (A = Ca, Sr)

The OER performances of ACoO3 (A = Ca, Sr, La), commercial Co3O4, and RuO2 are shown in Fig. 2 (A to D) and table S2; the current density of all samples in the linear sweep voltammograms was normalized to the electrochemically active surface area to exclude geometric effects (fig. S4 and table S2). CaCoO3 has the highest OER activity with an onset potential of 1.48 V versus RHE and a small overpotential of 0.26 V at 10 mA cm−2. SrCoO3 shows almost the same onset potential (1.51 V) as that of RuO2 (1.50 V). The smaller Tafel slopes of CaCoO3 and SrCoO3 in Fig. 2C compared to those of RuO2, LaCoO3 with intermediate-spin t5e1 at 25°C of the surface Co(III), and Co3O4 indicate their superior OER kinetics. The perovskite LaCoO3 and spinel Co3O4 with larger surface O─O separation at 25°C exhibit a negligible catalytic activity compared with those of ACoO3 (A = Ca, Sr).

Fig. 2 OER performance of ACoO3 (A = Ca, Sr), RuO2, LaCoO3, and Co3O4.

(A) Linear sweep voltammograms at 1600 rpm in 0.1 M KOH. (B) Overpotential required at 10 mA cm−2. (C) Tafel slopes. (D) Chronoamperometric curves in an O2-saturated 0.1 M KOH electrolyte at 1.6 V versus RHE.

The electronic structure of (A = Ca, Sr)

The Co─O bond lengths of the perovskites CaCoO3, SrCoO3, and LaCoO3 are 1.867, 1.915, and 1.930 Å, respectively. The 180° Co─O─Co interaction in cubic ACoO3 (A = Ca, Sr) is strong enough to form a half-filled antibonding σ* itinerant-electron up-spin band (Fig. 3). The temperature dependence of resistivity in Fig. 3A of ACoO3 (A = Ca, Sr) shows a metallic behavior down to 2 K, and the electronic conductivity of CaCoO3 is one order of magnitude higher than that of SrCoO3. The higher electronic conductivity of CaCoO3 than SrCoO3 reduces the charge transfer resistance (Rct) and the ohmic potential drop of CaCoO3 during the OER (fig. S5A), and the Rct values of CaCoO3 and SrCoO3 are one order of magnitude smaller than those of the electronic insulators LaCoO3 and Co3O4 at potentials above 1.4 V (figs. S5A and S6). The magnetic properties of CaCoO3 and SrCoO3 were investigated to study the electron configuration of CoIV ions (Fig. 3A), which affects their interaction with the reaction intermediates and their OER activity. An effective magnetic moment μeff = 4.1 μB was obtained for CaCoO3 by fitting with the Curie-Weiss law, indicating that the CoIV ions of CaCoO3 have the intermediate spin state (t24σ*1) with S = 3/2 (Fig. 3B). SrCoO3, which also adopts the intermediate-spin electron configuration, has a ferromagnetic transition at 300 K in the magnetization. For comparison, the spin state of CoIII ions in LaCoO3 (Co3+: d6) can be described by a d6 mixed-spin scenario in which the ratio of the low-spin (LS) ground state (t6e0, S = 0) and the high-spin (HS) excited state (t4e2, S = 2) is roughly unity at room temperature (fig. S7) (22), while the CoIII ions in spinel Co3O4 are in the LS state (t6e0, S = 0) (Fig. 3B). Moreover, the π* and σ* bandwidths are greater in CaCoO3 with the shorter CoIV─O2− bond than in SrCoO3 (Fig. 3C).

Fig. 3 Magnetic and transport properties of ACoO3 (A = Ca, Sr).

(A) Temperature dependence of resistivity measured at zero field [CaCoO3 (red curve), SrCoO3 (blue curve digitized from (20)] temperature dependence of magnetization in both zero-field cooling (ZFC) and field cooling (FC). (B) Electronic spin states of the octahedral site Co ions of ACoO3 (A = Ca, Sr) and Co3O4. (C) Schematic band diagrams of ACoO3 (A = Ca, Sr). IS, intermediate spin.

Figure 4 and fig. S8 compare the onset potentials and the OER activity of CaCoO3 versus SrCoO3, Co3O4, LaCoO3, and RuO2 in O2-saturated KOH, with pH varying from 12.5 to 14. The onset potentials of the two perovskites are similar, but the activity of the OER is much higher on CaCoO3 than on SrCoO3. The activities of the two CoIV perovskites are substantially higher than those of the two CoIII oxides. In each case, the catalytic activity and onset potential show a dependence on the pH of the KOH solution.

Fig. 4 OER performance of ACoO3 (A = Ca, Sr) in alkaline solutions with different pH.

(A to D) CV measurements in O2-saturated KOH with pH 12.5 to 14, showing the pH-dependent OER activity of ACoO3 (A = Ca, Sr) on the RHE scale.

The stability of ACoO3 (A = Ca, Sr)

The durability of all samples tested at 1.6 V in an O2-saturated 0.1 M KOH solution is shown in Fig. 2D and fig. S5B. CaCoO3 had the best durability. It retained more than 90% of its initial current density after 20,000 s and still kept about 89% of its initial current density after 50,000 s. The same OER activity of CaCoO3 after the measurement as the fresh sample (Fig. 2D, inset) indicates a good stability of CaCoO3 in alkaline solution. The XRD (Fig. 5A) and Raman results (Fig. 5B) of ACoO3 (A = Ca, Sr) after the OER testing show that the samples retain the same cubic perovskite structure, and the TEM image of ACoO3 (A = Ca, Sr) before and after OER testing confirms that the surfaces of the particles retain a crystalline structure (Fig. 5, C and D). The strong Co─O bond of CaCoO3 increases its stability in alkaline solution. The oxygen deficiency in SrCoO3 is the result of a weaker Co─O bond.

Fig. 5 The durability of ACoO3 (A = Ca, Sr).

(A to D) XRD, Raman spectra, and TEM results of ACoO3 (A = Ca, Sr) after OER testing.

DISCUSSION

The energy of the CoV state is below the top of the O-2p bands. Therefore, the removal of an electron from an ACoO3 perovskite surface during charge cannot oxidize CoIV to CoV; electrons are removed from surface oxygen. In KOH solution at low pH, the electron removal reaction isCo(OH)+OHe=CoO+H2O(1)because the strong attraction of the Co for the O-2p electron weakens the attraction of surface oxygen for the surface proton relative to the attraction of the medium OH. The charging voltage needed to create a surface -CoO species is the onset potential of the OER. At high pH, the surface proton is removed chemically to form -CoO2− in the reactionCoOH+OH=CoO2+H2O(2)and electron removal during charge at the onset potential isCoO2e=CoO(3)

The surface -CoO may be attacked by a solution OH in the reactionCoO+OHe=CoOOH(4A)that is followed byCoOOH+OH=Co(O2)2+H2O(4B)

The activity of the OER after its onset depends on the rates of the reactions (4) and an alternative mechanism for the formation of the peroxide ion (O2)2− by a direct O─O interaction, which would be faster if the surface O─O separation is shorter2CoO+OH=Co(O2)2+Co(OH)(5)

Reaction 5 may compete with reaction 4. Reaction 1 is the step that determines the onset potential. The shorter Co─O bond length of CaCoO3 relative to that of SrCoO3 may not only lower the onset potential of reaction 1 but also increase the activity by reducing the surface oxygen separation to increase the rate of reaction 5 relative to that of reaction 4. Once the peroxide ion (O2)2− is formed, the reaction -Co(O2)2−−2e + OH = -Co(OH) + O2↑ should not be rate limiting.

The CoIII ions in Co3O4 and LaCoO3 can be oxidized to CoIV, but the onset potential for reaction 1 is necessarily higher starting from CoIII rather than CoIV. Figure S3 confirms that a greater removal of electrons before deprotonation occurs in Co3O4 than in ACoO3, with A = Ca or Sr.

Increasing the concentration of CoIV in the chemically/electrochemically delithiated Li1−xCoO2 has been reported to increase the OER activity. A metallic layer on the Li1−xCoO2 phase with LS CoIV/CoIII ions was obtained at x = 0.5, and at x > 0.5, the peroxide (O2)2− forms on the surface of Li1−xCoO2; as x increases, oxygen gas is released. As the formal valence of the cobalt increases, its Fermi level approaches and crosses the top of the O-2p bands in the layered Li1−xCoO2, which lowers the energy of back electron transfer from the oxygen to the cobalt and increases the O-2p character in the hybrid orbitals of Co-d symmetry. On oxidation, removal of electrons creates surface electron holes in the hybrid orbitals and, therefore, increasingly on the oxygen as the formal Co valence is increased. The top of the O-2p bands in the ACoO3 perovskites is lower with respect to the hybrid antibonding orbitals of d-wave symmetry than in the layered Li1−xCoO2, which is why the CoIV valence is stable in the ACoO3 perovskites, but the shorter the Co─O bond, the larger the O-2p and Co-3d overlap integrals and their hybridization.

Oxygen species (water or OH) are adsorbed on the perovskite to complete the sixfold coordination of the cobalt, and the surface protons reach equilibrium with the medium pH. The proton transfer from the catalyst to the medium can be either a single chemical transfer reaction 2 or a proton-electron transfer step as shown in reactions 1 and 4 above. Two possible OER mechanisms are shown in Fig. 6; electron transfer from the electrocatalyst at an applied onset potential creates an O intermediate with an electron hole that can be attacked by the medium OH to form OOH (step 2 in Fig. 6D) or two neighboring O may react as in reaction 5. The former involves a second removal of H+, whereas the latter does not, which may make the latter the faster mechanism.

Fig. 6 Possible OER mechanisms in ACoO3 (A = Ca, Sr).

(A) Conventional OER mechanism involving four proton-electron transfer steps on the active surface metal sites. (B) OER mechanism with surface lattice oxygen activated for OER to form peroxide and evolve O2. (C and D) Two proposed OER mechanisms based on the Co4+─O2− bond; the deprotonation process by OH is finished by a chemical step (step 5) rather than a proton/electron-coupled transfer step.

CONCLUSIONS

The ACoO3 (A = Ca or Sr) cubic perovskites synthesized under high pressure have been studied as OER electrocatalysts in KOH solution. The high-pressure synthesis stabilizes the CoIV ion in an intermediate spin-state configuration t4σ*1 in which the σ* antibonding orbitals of d-wave symmetry are itinerant, whereas the π* antibonding orbitals of d-wave symmetry are localized in SrCoO3 and are at the crossover to itinerant behavior in CaCoO3. Both perovskites are cubic, but CaCoO3 has a much shorter Co─O bond length and larger σ* bandwidth indicative of a greater O-2p character in the states of d-wave symmetry. The smaller onset potentials with CoIV than CoIII oxides show the importance of admixing O-2p orbitals in the states of d-wave symmetry for the OER. Nevertheless, the larger σ* bandwidth of CaCoO3 versus SrCoO3 has little influence on the onset potential of the OER, but it does stabilize the catalyst. The rate-limiting step responsible for the OER is the formation of surface peroxide ions (O2)2−. The substantially greater OER activity on CaCoO3 than SrCoO3 indicates a larger component of a faster mechanism than on SrCoO3, which suggests that the shorter separation of surface oxygen in CaCoO3 may be favoring reaction 5 relative to reaction 4. This suggestion warrants further experimental verification.

MATERIALS AND METHODS

The precursors of oxygen-deficient perovskites CaCoO3-δ and SrCoO3-δ were prepared by solid-state reaction. The mixture of CaCO3 (or SrCO3) (Alfa Aesar, 99.99%) and Co3O4 (Alfa Aesar, 99.99%) with a stoichiometric ratio was thoroughly ground and sintered at 1000°C for 48 hours. High-pressure syntheses of the cubic CaCoO3 and SrCoO3 perovskites were performed under 7 GPa and at 1200°C with a Walker-type multianvil module (Rockland Research Co.). The precursors were pressed into pellets and sealed in a platinum crucible with the oxygen-releasing agent KClO4. Each crucible was contained in a Mo heater surrounded by a LaCrO3 sleeve for thermal insulation. Cr-doped MgO octahedra with pyrophyllite gaskets were used as the pressure medium. The phase purity of the obtained products was examined by powder XRD at room temperature with a Philips X’Pert diffractometer (Cu Kα radiation). Structural parameters were obtained by refining the XRD patterns with the software FULLPROF. DC magnetic susceptibility was measured with a commercial Superconducting Quantum Interference Device (SQUID) magnetometer (Quantum Design). Resistivity data were collected using the Physical Property Measurement System (PPMS; Quantum Design).

The morphology and microstructures of the samples were investigated using field-emission scanning electron microscope (Quanta 650) and TEM. Raman spectroscopy was carried out with a WITec alpha 300 Raman microscope instrument (WITec, Germany) equipped with a 532-nm wavelength laser for excitation.

All electrochemical measurements were carried out using an Autolab electrochemical workstation with a conventional three-electrode system in O2-saturated KOH solution. The Hg/HgO electrode, Pt, and glassy carbon electrode (GCE; 5 mm in diameter) were used as the reference, counter, and working electrodes, respectively. All the potential values were calibrated to the RHE. The catalyst inks were prepared as follows: 5 mg of the catalyst and 20 μl of 5 weight % Nafion were dispersed in 1000 μl of isopropanol. Ten microliters of the catalyst ink with a catalyst loading of 0.25 mg cm−2 was dropped onto the GCE and dried at room temperature. The OER characterizations were carried out using a glassy carbon rotating disc electrode at a rate of 1600 rpm. The CV and linear sweep voltammetry measurements were performed at a rate of 50 and 5 mV s−1, respectively. The chronoamperometric responses were conducted on a fixed potential of 1.55 V in O2-saturated 0.1 M KOH solution. The electrochemical impedance spectroscopy (EIS) was measured using an Autolab workstation with an applied frequency range from 106 to 0.1 Hz. Titanium foil was used as the substrate while testing the materials after cycling.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaav6262/DC1

Fig. S1. SEM images and particle size distribution of CaCoO3 and SrCoO3.

Fig. S2. Elemental distribution of CaCoO3 and SrCoO3.

Fig. S3. CV curves of Co-based catalysts.

Fig. S4. Double-layer capacitance measurement to determine the electrochemically active surface area of CaCoO3 and RuO2.

Fig. S5. The impedance plots and cycling stability of the catalysts studied in this work.

Fig. S6. Response of charge transfer resistance (Rct) to applied potential of the catalysts studied in this work to applied potential in the solution of pH 12.5, 13, 13.5, and 14.

Fig. S7. The intermediate spin state of CoIII in LaCoO3 (t4e1, S = 1) at 25°C.

Fig. S8. pH-dependent OER activity of RuO2 in O2-saturated KOH with pH 12.5 to 14.

Table S1. Structural refinements of CaCoO3 and SrCoO3 obtained at room temperature.

Table S2. OER activity and ECSA of all studied catalysts.

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was supported by the Department of Energy, Office of Energy Efficiency and Renewable Energy (EERE) under award no. DE-EE000762 and Robert A. Welch Foundation of (Houston, TX) grant F-1066. X.L. also acknowledged the Beijing Institute of Technology Research Fund Program for Young Scholars. Y.Lo. was supported by the National Key R&D Program of China (grant 2018YFA0305700). Author contributions: J.B.G. supervised the project. Y.Li conceived the idea and designed the research. X.L., J.Z., Y.Lo., and C.J. prepared the sample and characterized the physical properties. H.W., Y.Li, Z.C., and S.X. did the other characterization and analyzed the data. Y.Li and J.B.G. wrote the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Navigate This Article