Porous calcium–manganese oxide microspheres for electrocatalytic oxygen reduction with high activity

Abstract

A series of calcium–manganese oxides (Ca–Mn–O) were prepared through thermal decomposition of carbonate solid–solution precursors and investigated as electrocatalysts for oxygen reduction reaction (ORR). The synthesized crystalline Ca–Mn–O compounds, including perovskite-type CaMnO₃, layered structured Ca₂Mn₃O₈, post-spinel CaMn₂O₄ and CaMn₃O₆, presented similar morphologies of porous microspheres with agglomerated nanoparticles. Electrochemical results, surface analysis, and computational studies revealed that the catalytic activities of Ca–Mn–O oxides, in terms of onset potential, reduction current, and transferred electron number, depended strongly on both the surface Mn oxidation state and the crystallographic structures. Remarkably, the as-synthesized CaMnO₃ and CaMn₃O₆ exhibited considerable activity and enabled an apparent quasi 4-electron oxygen reduction with low yield of peroxide species in alkaline solutions, suggesting their potential applications as cheap and abundant ORR catalysts.

---

Introduction

Achieving efficient catalysis of the sluggish oxygen reduction reaction (ORR) plays an important role in a variety of electrochemical energy conversion and storage technologies including fuel cells and metal–air batteries.^1–5 Precious metals such as Pt and Pt alloys are known to exhibit the best overall catalytic performance for the ORR, but their limited availability and high price hinder the widespread practical applications.^6–8 Thus, extensive research effort has been dedicated to developing alternative catalysts based on non-noble elements.^9–13 Among them, manganese oxides are particularly attractive because of high abundance, low cost, minimum environmental impact, and considerable ORR activity.^14–22 Additionally, the variable valence of Mn (e.g., +2, +3, +4, +6 and +7) and structural flexibility of oxides (e.g., available in binary oxide type or capable of incorporating another metal to form composite oxide structures such as perovskite and spinel) offer opportunity yet challenges to tailor the physical–chemical and catalytic properties of the manganese oxide family.^23,24

Manganese oxides can also serve as electrocatalysts for the oxygen evolution reaction (OER). In Nature, photosystem II (PS II) has the unique capability to oxidize water to molecular oxygen via the oxygen-evolving complex (OEC), which contains a cubane-like μ-oxido-Mn₄Ca cluster housed in a special protein environment.²⁵ Inspired from the biological OEC active site, much interest has been focused on functional Mn-based molecular mimics for the OER.²⁶ As there are parallels between abiotic and biotic MnO_x structures,²⁷ substantial progress has also been made in inorganic nanoscaled binary manganese oxides such as Mn₂O₃ nanostructured films,¹⁹ nanocrystalline λ-MnO₂,²⁸ and Mn oxide clusters.²⁹ Interestingly, recent studies demonstrated that incorporation of Ca into binary manganese oxides significantly improved their water oxidation activities.^30,31 Considering the inherent connection between the oxygen electrochemistry of ORR and OER, the Ca–Mn oxides might also act as efficient catalysts for the ORR. However, to the best of our knowledge, this has not been systematically investigated so far.

Aiming at developing new ORR catalysts composed of earth abundant and inexpensive elements, we report in this study the preparation of a series of calcium–manganese oxides (CaMnO₃, Ca₂Mn₃O₈, CaMn₂O₄ and CaMn₃O₆) and the detailed investigation of their electrocatalytic properties. This mixed metal oxide series are featured by different chemical compositions, crystal structures and oxidation states. The synthesis of these Ca–Mn–O compounds was achieved through a simple calcination route using Ca_1−xMn_xCO₃ solid–solution precursors. The parallel formation of highly crystalline Ca–Mn–O porous microspheres with similar textures allowed us to systematically investigate the structural and compositional effect on electrochemical properties. In addition, the ORR and OER catalytic activities of the synthesized Ca–Mn–O compounds were compared with MnO_x and benchmark Pt/C catalysts in alkaline conditions. Both experimental results and theoretical study demonstrated that the surface Mn oxidation state and crystal structure are influential factors determining the activity of Ca–Mn–O electrocatalysts.

Experimental section

Materials synthesis

Samples of CaMnO₃, Ca₂Mn₃O₈, CaMn₂O₄ and CaMn₃O₆ were synthesized by simply firing the corresponding Ca_1−xMn_xCO₃ (x = 0.5, 0.6, 0.67, and 0.75) solid–solution precursors (see ESI†, Fig. S1) in air.³² The Ca–Mn–O compounds were obtained by heating the solid–solution carbonate precursors in recrystallized alumina boats in air. Calcination parameters: 900 °C, 5 h for CaMnO₃, 700 °C, 1 h for Ca₂Mn₃O₈, 800 °C, 1 h for CaMn₃O₆, and 950 °C, 10 h for CaMn₂O₄. The benchmark Pt/C catalyst (30 wt% Pt nanoparticles supported on Vulcan XC-72 carbon powders, ESI†, Fig. S2) was prepared following the procedures of our previous report.³³

Materials characterization

Chemical composition of the as-synthesized samples was determined by atomic adsorption spectrometry (AAS, Hitachi 180-80 spectrophotometer) and inductively coupled plasma-atomic emission (ICP) spectrometer (ICP-9000, Thermo Jarrell-Ash). The oxidation state of Mn was analyzed by a chemical titration: samples were dissolved in dilute H₂SO₄ and H₃PO₄ solution with an excess of Fe(NH₄)₂(SO₄)₂, and titrated with calibrated KMnO₄ solution. Structures and morphologies were characterized by powder X-ray diffraction (XRD, Rigaku D/max-2500 X-ray diffractometer, Cu Kα radiation), scanning electron microscopy (SEM, FEI Nanosem 430), and transmission electron microscopy (TEM, Philips Tecnai F20) equipped with energy dispersive spectroscopy (EDS). X-Ray photoelectron spectroscopy (XPS) data were collected using a Perkin Elmer PHI 1600 ESCA system.

Electrochemical tests

Electrochemical measurements were conducted using a standard three-compartment electrochemical cell. A Pt foil and a saturated calomel electrode (SCE) were used as the reference electrode and counter electrode, respectively. The working electrode was a catalyst-coated glass carbon (GC) substrate, which is confined in a rotating disk electrode (RDE, diameter 4 mm) or a rotating ring-disk electrode (RRDE, Pt ring and GC disk, disk diameter 4.2 mm). The thin catalyst layer was obtained by naturally drying a catalyst ink, which was produced by ultrasonically dispersing a mixture of Ca–Mn–O sample, carbon powders (Carbot Vulcan XC-72) and neutralized Nafion solution (5 wt%, Sigma-Aldrich) in isopropyl alcohol. The oxide samples were mixed with carbon at a 3 : 7 mass ratio to overcome electronic conductivity limitation. Catalyst loading is approximately 0.09 and 0.12 mg cm⁻² for RDE and RRDE, respectively.

All tests were performed at room temperature on a computer-controlled potentionstat/galvanostat workstation. The supporting electrolyte comprising 0.1 M aqueous KOH solution was purged with Ar or O₂ (Air Product, purity 99.995%) for at least 30 min prior to testing and maintained under Ar or O₂ atmosphere during the test. In RDE measurement, the working electrode was scanned at a rate of 5 mV s⁻¹ in the potential range of 0.1 to −0.6 V (ORR) or 0.4–0.9 V (OER) versus SCE. In RDE measurement, the ring potential was set at 0.5 V versus SCE. Unless stated, all potentials were reported with reference to the reversible hydrogen electrode (RHE) potential scale. In 0.1 M KOH solution, the potential of SCE was calibrated as +0.990 V with respect to RHE.³⁴ All the ORR curves obtained in O₂-saturated solutions were corrected by subtracting the background capacitive current obtained in an Ar-saturated electrolyte.

Results

Structures and morphologies

Fig. 1 illustrates the crystal structures of carbonate calcite and composite oxides with increasing Mn : O ratio. The solid–solution calcite adopts the trigonal structure with space group of R-3c (no. 167).³⁵ In calcite, the Ca/Mn cations and carbonate groups are arranged in a distorted halite structure, in which cations are coordinated by six oxygen anions and each oxygen is bonded to two cations (Fig. 1a). The orthorhombic perovskite structure of CaMnO₃ is based on a framework of corner-sharing MnO₆ octahedra (Fig. 1b). The Mn(IV) cations locate at the center of the MnO₆ octahedra while each Ca²⁺ ion occupies the inter-octahedral sites and is surrounded by 12 oxygen anions.³⁶ The monoclinic layered structure of Ca₂Mn₃O₈ can be described as infinite sheets of Mn₃O₈⁴⁻ held together by prismatically coordinated Ca²⁺ ions, which separate the edge-sharing MnO₆ octahedral sheet (Fig. 1c).³⁷ Both CaMn₃O₆ and CaMn₂O₄ are built up with double chains of edge-sharing MnO₆ octahedra (Fig. 1d and e). The corner-linked double chains create the six-sided tunnels for Ca²⁺ ions to reside in. Compared to the post-spinel phase of CaMn₂O₄, one third of the Ca position remains vacant in the tunnels of CaMn₃O₆, which can be viewed as an equivalent notation of Ca_2/3Mn₂O₄.³⁸ The structural diversity of perovskite, layer and post spinel phases of the present Ca–Mn–O series is a great motivation for us to investigate their electrolytic properties.

Fig. 1 Schematic crystal structures of (a) Ca_1−xMn_xCO₃ calcite, (b) CaMnO₃, (c) Ca₂Mn₃O₈, (d) CaMn₂O₄ and (e) CaMn₃O₆.

The XRD patterns of the Ca_xMn_1−xCO₃ precursors could be readily indexed to a calcite-type structure (ESI†, Fig. S1). Fig. 2 displays the XRD patterns and corresponding Rietveld refinement of the obtained samples after firing the precursors. No peak from other phases was detected, indicating the formation of high-purity oxides. The four oxides could be readily assigned to orthorhombic CaMnO₃ (Joint Committee on Powder Diffraction Standards, JCPDS card no. 76-1132), monoclinic Ca₂Mn₃O₈ (no. 73-2290), orthorhombic CaMn₂O₄ (no. 70-4889) and monoclinic CaMn₃O₆ (no. 31-0285), respectively. The Rietveld refinement of CaMnO₃ gave calculated cell parameters of a = 5.270 Å, b = 7.447 Å and c = 5.287 Å, in good agreement with the standard values. Similarly, the refined results of Ca₂Mn₃O₈, CaMn₂O₄ and CaMn₃O₆ matched quite well with the experimental data. The EDS elemental analysis confirmed the corresponding Ca : Mn compositional ratio of each synthesized oxide (ESI†, Fig. S3).

Fig. 2 Rietveld refined XRD patterns of the synthesized (a) CaMnO₃, (b) Ca₂Mn₃O₈, (c) CaMn₂O₄ and (d) CaMn₃O₆, with experimental data (red dots), calculated profile (cyan lines), allowed positions of Bragg reflection (green vertical bars) and difference curves (blue lines).

The morphology and microstructure of the synthesized Ca–Mn–O compounds were analyzed by SEM and TEM imaging. The four samples presented similar shape of hierarchical microspheres, which consisted of aggregated nanoparticles (see typical SEM images in Fig. 3a, d, g and j). All oxide products maintained essential spherical shape of the carbonate precursors, with diameters of 1.0–3.0 μm. Compared to the dense calcite precursors (ESI†, Fig. S4), the product spheres were porous. This was possibly due to two aspects: the phase transition from low-density carbonate to high-density oxides accompanying the release of CO₂ from precursor decomposition, and the Ostwald ripening of aggregated particles as previously described in the formation of perovskite microstructures.³⁹ The disparity in surface morphology of obtained oxides could be ascribed to the difference in synthetic conditions. For Ca₂Mn₃O₈, the low calcination temperature and short heating time favored the formation of smaller nanoparticles, which aggregated into loose microspheres. In comparison, the relatively elevated temperature and prolonged time resulted in compact CaMn₂O₄ microspheres with more smooth surfaces and less porosity.

Fig. 3 SEM and TEM micrographs of synthesized CaMnO₃ (a–c), Ca₂Mn₃O₈ (d–f), CaMn₂O₄ (g–i) and CaMn₃O₆ (j–l) samples.

The formation of Ca–Mn–O microspheres with aggregated nanoparticles was also clearly revealed by typical TEM observation (Fig. 3b, e, h and k), which showed pores within each sample. The difference in porosity and primary particle sizes among the four hierarchical microspheres coincided with SEM images. To gain structural information, we performed HRTEM imaging on edges of isolated nanoparticles (Fig. 3c, f, i and l). All oxide products indicated high crystallinity with clear distinguished lattice fringes. Taking CaMnO₃ as an example, the measured neighboring interlayer distance was consistent with the spacing between the (121), (002), and (1̄2̄1) planes, confirming the XRD analysis. The well-defined points in the corresponding fast Fourier transform (FFT) diffraction pattern (inset of Fig. 3c) conformed to the orthorhombic crystal structure of perovskite CaMnO₃. Likewise, the measured interlayer distances and FFT patterns of Ca₂Mn₃O₈, CaMn₂O₄ and CaMn₃O₆ matched the neighboring separations of their related planes and the allowed Bragg diffraction of their phases, respectively.

Electrocatalytic results

To evaluate the ORR catalytic activities of the Ca–Mn–O microspheres, electrochemical tests were carried out using rotational disk electrodes in 0.1 M KOH solution saturated with Ar or O₂ (ESI†, Fig. S5). Compared with the featureless, negligible reduction current in Ar-saturated electrolyte, the detected current at O₂ atmosphere was attributed to the catalytic oxygen reduction. The linear sweeping voltammograms (LSVs) of all oxides displayed similar shape, with two regions of potential–current response. Scanning the potential cathodically, the reduction proceeded under mixed kinetic-diffusion controlled regime (0.9–0.6 V), where the detected currents increased rapidly. At potentials below ∼0.6 V, current plateaus were observed, with the appearance of diffusion-limiting currents (i_d). The reduction currents increased with rising rotation rates as a result of faster oxygen flux to electrode surface. Fig. 4a compares the LSVs of oxide microspheres, GC support, Vc-72 carbon additive and benchmark Pt/C at 1600 round per minute (rpm). Among the synthesized Ca–Mn–O series compounds, the level of catalytic activities follows an order of CaMnO₃ > CaMn₃O₆ > CaMn₂O₄ ≈ Ca₂Mn₃O₈, based on their onset potentials (ESI†, Fig. S6) and i_d values. Note that neat carbon obviously contributes to the ORR kinetics, consistent with reported results in alkaline media.⁴⁰ Since CaMn₂O₄ and Ca₂Mn₃O₈ do not show significant activity relative to neat carbon, we next focus on the highly active CaMnO₃ and CaMn₃O₆.

Fig. 4 (a) LSVs of CaMnO₃, Ca₂Mn₃O₈, CaMn₂O₄, CaMn₃O₆, GC, Vc-72 and Pt/C recorded at 1600 rpm in O₂-saturated 0.1 M KOH solution. (b) K–L plots of different catalysts at 0.5 V.

The rotational speed-dependent current can be theoretically applied to construct Koutechky–Levich (K–L) plots (i⁻¹versus ω^−1/2) according to the K–L equation:

where i_k is the kinetic current density, ω is the angular velocity of the disk (ω = 2πN, N is the rotation speed), n is the overall transferred electron number per molecular oxygen in the ORR, F is the Faraday constant, A is the geometric electrode area, k is the rate constant of the reaction, C⁰ is the saturated O₂ concentration in the electrolyte, D_O2 is the diffusion coefficient of O₂, and v is the kinematic viscosity of the electrolyte.⁴¹Fig. 4b shows the K–L curves, in which n can be determined from the slopes of the fitted linear line. The slopes of CaMnO₃ and CaMn₃O₆ were close to that of Pt/C. In comparison, CaMn₂O₄ and Ca₂Mn₃O₈ showed higher slopes. It has been recognized that the ORR on Pt-based catalysts mainly proceeds through a four-electron (4e) transfer mechanism while a two-electron (2e) reduction is the dominate process on carbonaceous materials.^3,5,6 Accordingly, CaMnO₃ and CaMn₃O₆ catalyzed the oxygen reduction via a quasi-4e pathway, whereas the apparent electron transfer numbers were between 2 and 4 for CaMn₂O₄ and Ca₂Mn₃O₈.

To further analyze the electrocatalytic ORR pathways, the Ca–Mn–O microspheres were further investigated using the rotating ring-disk electrodes (RRDE). Fig. 5a show the polarization curves recorded on the ring and disk. The shape of voltammetry on the disk and the catalytic trend (in terms of both potential and current) were consistent with RDE measurement. Intermediate peroxide species generated from the ORR electrocatalysis on the catalyst disk is detected on the Pt ring. Based on the disk and ring currents (i_d and i_r), the yield of peroxide species (y_peroxide, defined as percentage of formed peroxides with respect to the total oxygen reduction products) and the transferred electron number (n) can be determined according to the following equations:

where N is the current collection efficiency of RRDE.^15,22,41 As shown in Fig. 5b, the determined y_peroxide and n varied with disk potential. From 0.4 to 0.7 V, the average peroxide yields were measured to be ∼2.2% and ∼7% for CaMnO₃ and CaMn₃O₆, respectively. The corresponding electron transfer number was 3.95 and 3.86. Similar to Pt/C, the ORR catalyzed by CaMnO₃ and CaMn₃O₆ was mainly through a quasi-4e ORR. This was consistent with the results from the K–L determination (Fig. 4b). Thus, the RRDE data confirmed that CaMnO₃ and CaMn₃O₆ exhibited remarkable catalytic activities.

Fig. 5 Percentage of peroxide (solid line) and the electron transfer number (dotted line) of CaMnO₃, CaMn₃O₆ and Pt/C at different potentials.

Besides high activity, Ca–Mn–O microspheres also exhibited respectable catalytic stability in alkaline electrolyte. In a continuous polarization period of 60 000 s, the ORR currents maintained 93.6% and 93% of the initial values for CaMnO₃ and CaMn₃O₆, respectively (Fig. 6). The gradual decay of reduction current might be partly due to insufficient gas flux, and the slow catalyst peeling off the electrode surface during the prolonged test with applied constant electrode rotation. Irrespective of this possibility, CaMnO₃ and CaMn₃O₆ exhibited slightly better activity retention as compared to the Pt/C benchmark (92.2%). Activity degradation of Pt/C catalyst in alkaline electrolytes has also been observed previously and ascribed to dissolution/aggregation of Pt nanoparticles and corrosion of carbon support.^22,34 The morphologies of Ca–Mn–O oxides may provide some merit against particulate agglomeration. Furthermore, the electrochemical durability was also confirmed by elemental analysis that indicated negligible leaching of Ca and Mn cations in the electrolytes after cycles of voltammetry test (ESI†, Fig. S7).

Fig. 6 Chronoamperometric curves (percentage of retained current as a function of operation time) of CaMnO₃, CaMn₃O₆, and Pt/C electrodes maintained at 0.8 V versus RHE in O₂-saturated electrolyte (0.1 M KOH).

Fig. 7 displays the Tafel plots of CaMnO₃, CaMn₃O₆, and comparative Pt/C in alkaline electrolyte. Kinetic currents derived from the mass-transport correction (i_k = (i × i_d)/(i_d − i)) were employed to construct the Tafel curves. Two linear regions are observed at low and high overpotentials, respectively. The corresponding slopes are close to −2.303RT/F and −2.303(2RT/F) (i.e., −59 and −118 mV dec⁻¹ at 25 °C). This feature is similar to that previously reported for manganese oxides.^15,42 The variation of Tafel slopes indicates a possible change in the ORR reaction mechanism. It has been suggested for copper manganite spinels that at low applied potential the rate determining step (rds) is the conversion of S⋯O₂⁻ (S denotes active sites) into an adsorbed peroxide species while at high potential the reduction of the adsorbed O₂ to S⋯O₂⁻ is rate-limiting.⁴² The presence of two slopes could be also interpreted by structural effects of the catalyst layer or potential-dependent coverage of adsorbed species on catalyst surface.⁴⁰ Further elucidation of the underlying mechanism is still required.

Fig. 7 Tafel plots of CaMnO₃, CaMn₃O₆, and Pt/C derived by the mass-transport correction of voltammetry data.

Discussion

The main electrocatalytic characteristics of the Ca–Mn–O series compounds are summarized in Table 1. Comparing the catalytic properties, the four oxides can be divided in two groups of (1) CaMnO₃ and CaMn₃O₆ and (2) Ca₂Mn₃O₈ and CaMn₂O₄. The first group showed more positive onset and half-wave potentials, larger currents, and higher transferred electron numbers relative to the second group. Remarkably, CaMnO₃ was highly active, manifesting electrochemical parameters (apart from lower potentials) close to those of the benchmark Pt/C. Also, CaMn₃O₆ demonstrated comparable activities to CaMnO₃ despite lower reduction current. In contrast, Ca₂Mn₃O₈ and CaMn₂O₄ exhibited inferior performance. The rising question is what is responsible for the different activities of the Ca–Mn–O series. In regard to the heterogeneous ORR electrocatalysis on transition metal oxides, the catalytic mechanism involves oxygen adsorption, dissociation and reduction on catalyst surface, which is closely correlated with redox reactions of transition metal-containing species.^5,14–22 As a result, it is reasonable to hypothesize that chemical composition, elemental valence, crystal structure, and surface state are possible factors determining the catalytic activities.

Table 1 Summary of electrochemical results for the Ca–Mn–O series and the benchmark Pt/C^a

Catalyst	E onset (V)	E half (V)	I s (mA cm^−2)	I m (mA mg^−1)	n
a E onset, Ehalf, Is, Im, and n denote onset potential, half-wave potential, specific current, mass current density and electron transfer number, respectively. Is, Im, and n correspond to values determined at 0.5 V.
CaMnO3	0.96	0.76	−5.79	62.7	3.96
Ca2Mn3O8	0.85	0.70	−3.54	38.3	3.53
CaMn2O4	0.85	0.69	−3.71	40.2	3.50
CaMn3O6	0.95	0.78	−4.43	48.0	3.86
Pt/C	0.99	0.84	−6.17	66.8	3.98

---

For Ca–Mn–O series oxides, Mn possesses variable valences and thereby more likely acts as the active center. The nominal distribution of valence is Mn(IV), Mn(IV), Mn(III), and 1/3Mn(IV)2/3Mn(III) for CaMnO₃, Ca₂Mn₃O₈, CaMn₂O₄, and CaMn₃O₆, respectively. The difference in catalytic activities of the series compounds cannot be directly attributed to bulk composition and Mn valence. We performed XPS analysis to characterize the surface state of the oxides as it is a critical factor. The core-level XPS data are shown in Fig. 8 and Table 2. All the Mn 2p spectra (Fig. 8a) displayed two strong peaks centered at about 642 and 653 eV, which were attributed to Mn 2p_3/2 and Mn 2p_1/2 spin–orbit doublet, respectively. These broad peaks were asymmetrical towards higher binding energy side, agreeing with early studies.^43,44 The peak position of Mn 2p_3/2 moved from 641.5 eV in CaMn₂O₄ to 642.0 eV in Ca₂Mn₃O₈. This shift was ascribed to changes in the electrostatic energy at the Mn site, which was driven by the decrease in the 3d count.⁴⁴

Fig. 8 (a) Mn 2p and (b) Mn 3s XPS spectra of synthesized CaMnO₃, Ca₂Mn₃O₈, CaMn₂O₄ and CaMn₃O₆ microspheres.

Table 2 Mn 2p_3/2 and Mn 3s XPS data of four Ca–Mn–O oxides^a

Compound	Mn 2p3/2 (eV)	Mn 3s (eV)			Average oxidation state
		Peak 1	Peak 2	ΔeV
a Data of α-Mn2O3, γ-Mn2O3 and β-MnO2 are adapted from ref. 46–48 and listed for comparison. The parenthesis data correspond to the values determined by chemical titration.
CaMn2O4	641.5	88.57	83.23	5.34	2.9 (2.84)
CaMn3O6	641.6	88.30	83.25	5.05	3.4 (3.47)
CaMnO3	641.7	88.45	83.60	4.85	3.8 (3.81)
Ca2Mn3O8	642.0	88.77	84.03	4.74	4.0 (4.00)
α-Mn2O3	641.9	89.1	83.9	5.2	3.0
γ-Mn2O3	641.7	88.8	83.6	5.2	3.0
β-MnO2	642.2	89.4	84.7	4.7	4.0

---

Although the chemical shift of Mn 2p qualitatively reflected the oxidation state of Mn,⁴⁵ the magnitude was quite small to quantitatively deduce the valence. Fortunately, the splitting width of Mn 3s peaks (Fig. 8b) was found to be different among these four manganese oxides. The separation of peak energies (ΔE_b) of the Mn 3s increased from ∼4.7 eV for Ca₂Mn₃O₈ to ∼5.3 eV for CaMn₂O₄, which were close to those expected for Mn(IV) and Mn(III), respectively.^46–48 Moreover, there was a linear relationship between the energy separation of 3s splitting and the valence of Mn in the oxides.^46–49 Based on this relationship, the mean Mn oxidation states were estimated to be 4.0, 3.8, 3.4 and 2.9 for Ca₂Mn₃O₈, CaMnO₃, CaMn₃O₆ and CaMn₂O₄, respectively. These surface Mn valences were essentially consistent with but slightly lower than that of the expected values (except Ca₂Mn₃O₈). The lower oxidation state of Mn might originate from the presence of oxygen defects in the oxides synthesized at high calcination temperatures. Chemical analysis of the samples confirmed the slight compositional deviation from the stoichiometry and was consistent with the XPS evaluation (Table 2). Accordingly, the highly catalytically active Ca–Mn–O compounds (CaMnO₃ and CaMn₃O₆) possessed average surface Mn valence between +3 and +4. The pronounced inferior electrocatalytic performance of CaMn₂O₄ could be attributed to its low reducibility.

Recently, a variety of composite transition metal oxides have been investigated as catalysts for the oxygen electrochemistry.^24,50 It has been demonstrated that the intrinsic ORR activity of an oxide is strongly correlated to σ*-antibonding orbital occupancy and metal–oxygen covalency of surface transition-metal cations.²⁴ For Mn-based oxides, the most active catalyst generally has an average valence of Mn(III)–Mn(IV), which affords moderate bond strength between the catalyst surface and the reactant or product. Within this valence range, the compound can be viewed as mixed-valent manganese oxides. Due to the coexistence of multivalent Mn, cation redox reaction, surface charge storage as well as electronic and ionic properties are feasible to be enhanced.⁵¹ Additionally, our results suggest that the ORR activity increases with higher Mn oxidation state in the Ca–Mn–O oxides. Previous in situ XANES (X-ray absorption near edge structure) study has also revealed that higher content of Mn(IV) phase in MnO_x species was associated with better catalytic performance.⁵² Unlike noble metal catalysts, transition metal oxides mediate the ORR electrocatalysis through surface redox reactions concomitant with electron transfer to oxygen.^{5,14–22,24,50} Manganese oxides featured with multiple and high valence may favor oxygen activation and therefore allow better ORR kinetics.

Next, we consider the effect of crystallographic structure. In the layered structure of Ca₂Mn₃O₈ (Fig. 1c), the edge and corner-sharing MnO₆ octahedra sheet (like in birnessite-MnO₂) is much denser than that of other compounds, unfavourable for oxygen adsorption. The perovskite CaMnO₃, built upon corner-sharing MnO₆ octahedra (Fig. 1b), provides (1 × 1) tunnels with comparable size to the O–O bond length (∼0.12 nm) in molecular oxygen.⁵³ Hence, oxygen could be easily accommodated, favouring the cleavage of O–O bond and benefiting the ORR. This speculation explains the positive potential, large current, and high apparent electron transfer number of CaMnO₃. For CaMn₃O₆ and CaMn₂O₄, the open structures are advantageous to oxygen access, however, this benefit was negated by the large tunnel size. Their activities thus fell between that of Ca₂Mn₃O₈ and CaMnO₃. Furthermore, our preliminary density functional theory (DFT) studies indicate that the series Ca–Mn–O compounds lead to different configuration of surface oxygen adsorption and the corresponding strength of oxygen binding coincides with the order of their catalytic activities (ESI†, Fig. S8).

We have also investigated the catalytic activity of the series Ca–Mn–O compounds for the OER (ESI†, Fig. S9). Again, the mixed-valence composite oxides showed superior performance. Interestingly, CaMn₃O₆ clearly outperformed the other three oxides and the comparative Pt/C, resulting in lower overpotential and higher OER current. In regard to the oxygen electrochemistry on transition metal oxides, the ORR and OER may proceed via identical cycles but through opposite position.^24,50 The order of ORR and OER activities for Ca–Mn–O compounds should present dissimilar trends due to a different rds. For multivalent Ca–Mn–O oxides, lower Mn oxidation state leads to higher OER activity, which is consistent with previous reports.⁵⁴ Regardless of the difference in OER/OER characteristics, CaMnO₃ and CaMn₃O₆ can serve as non-precious bifunctional catalysts because of their prominent performance.

Lastly, the role of Ca in Ca–Mn–O composite oxides deserves to be noted. Our primary electrochemical results revealed that CaMnO₃ exhibited more positive reduction potential and larger current than did MnO₂, which had similar Mn oxidation state and open tunnel structure (ESI†, Fig. S10). We envisage that Ca plays an important role for the ORR electrocatalysis, although Ca is generally redox inactive. In the PSII OEC center, Ca is identified as an essential cofactor for the OER, probably as a site for the binding and activation of oxygen-containing species.^31,55 It is speculated that the incorporation of Ca in manganese oxides may affect behaviours of oxygen adsorption, activation and reduction. Furthermore, calcium cations feasibly stabilize the crystal structure with high-valence Mn, due to its flexible coordination with oxygen in oxide lattice frameworks. This can be validated by the fact that, in the absence of Ca only Mn(II)/Mn(III)-based binary oxides are thermodynamically stable at high temperatures,⁵⁶ while Ca–Mn–O oxides show chemical and structural robustness in solvothermal systems containing high concentration of KOH.^30,57 In contrast to binary manganese oxides that undergo dramatic changes in crystalline and chemical structure upon premature ageing,¹⁷ CaMnO₃ and CaMn₃O₆ exhibit remarkable stability in alkaline solution. Therefore, the catalytic durability of Ca–Mn–O microspheres is not surprising.

Conclusions

In summary, a series of porous Ca–Mn–O microspheres were synthesized by calcinating the solid–solution carbonate precursors and systematically investigated as electrocatalysts for the ORR. Catalytic properties of the Ca–Mn–O series compounds were found to be closely correlated with the surface oxidation state of Mn and the crystallographic structures that affects the extent of O₂ activation. The perovskite CaMnO₃ with open tunnels and multivalence exhibited the highest activities. The current density and electron transfer number were comparable to those of the benchmark Pt/C. Furthermore, CaMn₃O₆ and CaMnO₃ manifested superior OER performance and catalytic durability. The results may shed light on searching for efficient, cheap, and scalable transition-metal-oxide oxygen electrocatalysts that are widely used in metal–air batteries and alkaline fuel cells.

Acknowledgements

This work was supported by the National 973 (Grant no. 2011CB935902), 863 (Grant no. 2011AA050704), NSFC (Grant no. 21231005), 111 Project (Grant no. B12015), and Tianjin High-Tech (Grant no. 10SYSYJC27600).

References

1. B. C. H. Steele and A. Heinzel, Nature, 2001, 414, 345.
2. L. Malavasi, C. A. J. Fisher and M. S. Islam, Chem. Soc. Rev., 2010, 39, 4370.
3. P. A. Christensen, A. Hamnett and D. Linares-Moya, Phys. Chem. Chem. Phys., 2011, 13, 5206.
4. P. G. Bruce, S. A. Freunberger, L. J. Hardwick and J. M. Tarascon, Nat. Mater., 2012, 11, 19.
5. F. Y. Cheng and J. Chen, Chem. Soc. Rev., 2012, 41, 2172.
6. Y. H. Bing, H. S. Liu, L. Zhang, D. Ghosh and J. J. Zhang, Chem. Soc. Rev., 2010, 39, 2184.
7. D. F. Yancey, L. Zhang, R. M. Crooks and G. Henkelman, Chem. Sci., 2012, 3, 1033.
8. H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal., B, 2005, 56, 9.
9. B. Šljukić, C. E. Banks, S. Mentus and R. G. Compton, Phys. Chem. Chem. Phys., 2004, 6, 992.
10. Y. J. Feng and N. Alonso-Vante, Phys. Status Solidi B, 2008, 245, 1792.
11. F. Jaouen, E. Proietti, M. Lefèvre, R. Chenitz, J. P. Dodelet, G. Wu, H. T. Chung, C. M. Johnston and P. Zelenay, Energy Environ. Sci., 2011, 4, 114.
12. A. Morozan, B. Jousselme and S. Palacin, Energy Environ. Sci., 2011, 4, 1238.
13. Z. W. Chen, D. Higgins, A. Yu, L. Zhang and J. J. Zhang, Energy Environ. Sci., 2011, 4, 3167.
14. M. S. El-Deab and T. Ohsaka, Angew. Chem., Int. Ed., 2006, 45, 5963.
15. I. Roche, E. Chaînet, M. Chatenet and J. Vondrák, J. Phys. Chem. C, 2007, 111, 1434.
16. A. Debart, A. J. Paterson, J. Bao and P. G. Bruce, Angew. Chem., Int. Ed., 2008, 47, 4521.
17. I. Roche, E. Chaînet, M. Chatenet and J. Vondrák, J. Appl. Electrochem., 2008, 38, 1195.
18. F. Y. Cheng, Y. Su, J. Liang, Z. L. Tao and J. Chen, Chem. Mater., 2010, 22, 898.
19. Y. Gorlin and T. F. Jaramillo, J. Am. Chem. Soc., 2010, 132, 13612.
20. E. M. Benbow, S. P. Kelly, L. Zhao, J. W. Reutenauer and S. L. Suib, J. Phys. Chem. C, 2011, 115, 22009.
21. F. Y. Cheng, J. Shen, B. Peng, Y. D. Pan, Z. L. Tao and J. Chen, Nat. Chem., 2011, 3, 79.
22. Y. Liang, H. Wang, J. Zhou, Y. Li, J. Wang, T. Regier and H. J. Dai, J. Am. Chem. Soc., 2012, 134, 3517.
23. J. E. Post, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 3447.
24. J. Suntivich, H. A. Gasteiger, N. Yabuuchi, H. Nakanishi, J. B. Goodenough and Y. Shao-Horn, Nat. Chem., 2011, 3, 546.
25. Y. Umena, K. Kawakami, J. R. Shen and N. Kamiya, Nature, 2011, 473, 55.
26. C. W. Cady, R. H. Crabtree and G. W. Brudvig, Coord. Chem. Rev., 2008, 252, 444.
27. I. Saratovsky, P. G. Wightman, P. A. Pastén, J. F. Gaillard and K. R. Poeppelmeier, J. Am. Chem. Soc., 2006, 128, 11188.
28. D. M. Robinson, Y. B. Go, M. Greenblatt and G. C. Dismukes, J. Am. Chem. Soc., 2010, 132, 11467.
29. F. Jiao and H. Frei, Chem. Commun., 2010, 46, 2920.
30. M. M. Najafpour, T. Ehrenberg, M. Wiechen and P. Kurz, Angew. Chem., Int. Ed., 2010, 49, 2233.
31. M. Wiechen, I. Zaharieva, H. Dau and P. Kurz, Chem. Sci., 2012, 3, 2330.
32. H. S. Horowitz and J. M. Longo, Inorg. Synth., 1983, 22, 73.
33. J. Zhu, Y. Su, F. Y. Cheng and J. Chen, J. Power Sources, 2007, 166, 331.
34. Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier and H. J. Dai, Nat. Mater., 2011, 10, 780.
35. B. Fubini and F. S. Stone, J. Chem. Soc., Faraday Trans. 1, 1983, 79, 1215.
36. W. Paszkowicz, J. Pietosa, S. M. Woodley, P. A. Dluzewski, M. Kozlowski and C. Martin, Powder Diffr., 2010, 25, 46.
37. G. B. Ansell, M. A. Modrick, J. M. Longo, K. R. Poeppeimeler and H. S. Horowitz, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1982, 38, 1795.
38. J. Hadermann, A. M. Abakumov, L. J. Gillie, C. Martin and M. Hervieu, Chem. Mater., 2006, 18, 5530.
39. X. Yang, J. Fu, C. Jin, J. Chen, C. Liang, M. Wu and W. Zhou, J. Am. Chem. Soc., 2010, 132, 14279.
40. J. Perez, E. R. Gonzalez and E. A. Ticianelli, Electrochim. Acta, 1998, 44, 1329.
41. A. J. F. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Wiley, New York, 2000.
42. E. Ríos, S. Abarca, P. Daccarett, H. N. Gong, D. Martel, J. F. Marco, J. R. Cancedo and J. L. Gautier, Int. J. Hydrogen Energy, 2008, 33, 4954.
43. M. E. Melo Jorge, A. Correia dos Santos and M. R. Nunes, Int. J. Inorg. Mater., 2001, 3, 915.
44. M. Abbate, F. M. F. de Groot, J. C. Fuggle, A. Fujimori, O. Strebel, F. Lopez, M. Domke, G. Kaindl, G. A. Sawatzky, M. Takano, Y. Takeda, H. Eisaki and S. Uchida, Phys. Rev. B: Condens. Matter, 1992, 46, 4511.
45. B. J. Tan, K. J. Klabunde and P. M. A. Sherwood, J. Am. Chem. Soc., 1991, 113, 855.
46. M. Toupin, T. Brousse and D. Bélanger, Chem. Mater., 2002, 14, 3946.
47. M. Toupin, T. Brousse and D. Bélanger, Chem. Mater., 2004, 16, 3184.
48. M. Oku, K. Hirokawa and S. Ikeda, J. Electron Spectrosc. Relat. Phenom., 1975, 7, 465.
49. V. R. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N. A. Ovechkina, E. Z. Kurmaev, N. I. Lobachevskaya, Y. M. Mukovskii, J. Mitchell and D. L. Ederer, Phys. Rev. B: Condens. Matter, 2002, 65, 113102.
50. J. Suntivich, K. J. May, H. A. Gasteiger, J. B. Goodenough and Y. Shao-Horn, Science, 2011, 334, 1383.
51. M. K. Song, S. Cheng, H. Chen, W. Qin, K. W. Nam, S. Xu, X. Q. Yang, A. Bongiorno, J. Lee, J. Bai, T. A. Tyson, J. Cho and M. L. Liu, Nano Lett., 2012, 12, 3483.
52. F. H. B. Lima, M. L. Calegaro and E. A. Ticianelli, J. Electroanal. Chem., 2006, 590, 152.
53. Y. Wang and P. B. Balbuena, J. Phys. Chem. B, 2004, 108, 4376.
54. M. M. Najafpour, Dalton Trans., 2011, 40, 3793.
55. C. F. Yocum, Coord. Chem. Rev., 2008, 252, 296.
56. A. N. Grundy, B. Hallstedt and L. J. Gauckler, J. Phase Equilib., 2003, 24, 21.
57. J. Spooren and R. I. Walton, J. Solid State Chem., 2005, 178, 1683.

---

Footnote

† Electronic supplementary information (ESI) available: Additional characterization, electrochemical data, and computational details. See DOI: 10.1039/c2sc21475j

---