Enhancement of oxygen evolution reaction activity and durability of Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ by CO₂ thermal treatment

1. Introduction

The perpetual growth of energy demand and the use of fossil fuel-based resources as well as the related environmental issues have stimulated the development of more sustainable energy storage and conversion technologies that rely upon renewable resources such as solar, wind, and tidal resources [1]. The energy supply from these three resources however is subject to time and geographical location uncertainties; mostly of intermittent nature. Advanced electrochemical energy storage and conversion technologies such as water splitting, fuel cells, and metal-air batteries are considered as one of the key solutions to address this issue [2], [3], [4], [5], [6]. Central to these technologies is an oxygen evolution reaction (OER), which has sluggish kinetics given its complex four-electron oxidation process [7]. The presence of an efficient electrocatalyst becomes essential to accelerate the rate of OER. Precious metal-based iridium and ruthenium oxides (i.e., IrO₂ and RuO₂) display the highest OER activity but their high cost and low durability have hampered their large scale application [8], [9], [10], [11]. It becomes essential to find an alternative lower cost catalyst that can provide high OER activity and long term stability.

Perovskite oxides with the composition formula of ABO₃, have shown great potential as OER electrocatalyst for their high catalytic activity, low cost, and abundant availability [12,13]. In perovskite oxide structure, alkaline-earth or rare-earth metal cation occupies the A-site with 12-fold oxygen coordination and transition metal cation occupies the B-site with 6-fold oxygen coordination. Several perovskite oxides, i.e., LaBO₃ perovskites (e.g., LaNiO₃, LaCoO₃, and LaMnO₃) [13], SrNb_0.1Co_0.7Fe_0.2O₃ [2], (Ln_0.5Ba_0.5)CoO_3-δ (Ln = Sm, Gd and Ho) [14] were reported to show high OER activities in an alkaline media. Suntivich et al. correlated OER performances of eleven perovskite oxides to transition metal orbital filling principle [12]. In their work, BSCF with an e_g-filling electron of 1.2 (close to 1) exhibited the highest intrinsic OER activity among the tested perovskite oxides; the performance of which surpasses that of IrO₂ benchmark. During the repeated OER cycling tests nonetheless, barium cations (Ba²⁺) and strontium cations (Sr²⁺) diffused out from Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (BSCF) surface, resulting in the conversion of crystalline BSCF phase to amorphous phase. The increase in the amorphous layer thickness eventually led to the OER performance deterioration [15,16]. The change in the surface and bulk crystal structures and the modification in the morphology of the perovskite oxide may also cause significant difference in OER activity. In this context, Cho group has performed a series of studies that focuses on BSCF [17], [18], [19]. They applied thermal treatment in an oxygen atmosphere at 950 °C with an optimized time to tailor the surface structure of BSCF [19]. This thermal treatment process promotes the formation of the cubic structure and at the same time eliminates the homogeneous spinel surface layer formation; thus resulting into an enhanced OER activity relative to the non-treated BSCF catalyst.

This work reports an interesting phenomenon where the activity and durability of BSCF catalyst appears to be improved by the formation of an amorphous carbonate layer with a controlled thickness; the presence of which is hypothesized to contribute positively to the active self-reconstruction of BSCF surface during OER as reported elsewhere (Fig. 1) [20]. A simple conventional thermal route that induces the formation of an in situ decorative carbonate layer on BSCF surface was used to achieve such phenomenon. BSCF subjected to heating at 600 °C in CO₂ atmosphere (BSCF-600) displayed superior electrochemical performance compared to BSCF subjected to same treatment at 500 °C and 700 °C.

View original graphic| Download| PPT slide

Fig. 1.   Schematic illustration of the effect of heat treatment under CO₂ exposure on Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (BSCF).

2. Experimental

2.1. Catalyst synthesis

2.1.1. Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (BSCF) synthesis

Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (denoted as BSCF) was synthesized using a standard combined citrate and ethylenediaminetetraacetic acid (EDTA) method. Stoichiometric amounts of barium nitrate, strontium nitrate, cobalt nitrate, and iron nitrate (Sinopharm Chemical Reagent Co., Ltd.) were mixed and dissolved in deionized water. EDTA and citric acid were then added under stirring to form a homogeneous colored transparent solution at a mole ratio of 1:1:2 for total metal ions:EDTA:citric acid. To obtain complete complexation, an aqueous NH₃ solution was added to adjust the pH of the solution to 7. A caramel gel was formed by heating at 90 °C under stirring. The gel was heated at 250 °C for 5 h in air, resulting into the precursor. BSCF was finally obtained from the precursor by calcination at 1000 °C for 5 h in air; the ground powder or which was used for the following characterization and electrochemical tests.

2.1.2. BSCF-500, BSCF-600 and BSCF-700 syntheses

To obtain in situ carbonate modified BSCF, 0.5 g of BSCF powder was heated for 5 h at 5 °C min^-1 ramping rate up to 500 °C, 600 °C, or 700 °C in carbon dioxide atmosphere, respectively, which is denoted as BSCF-500, BSCF-600, BSCF-700, respectively. The flow rate of the gas through the heating system was 150 mL min^-1. Following the heat treatment, BSCF powders were milled again for further characterization and tests.

2.1.3. BSCF-S synthesis

Strontium acetate (99%) was dissolved in deionized water to form the impregnation liquid with a 0.1 mol L^-1 Sr²⁺ concentration. Approximately 10 mL of the impregnation solution was dropped on BSCF surface. Then, the impregnated powder was heated at 750 °C for 2 h, resulting into SrCO₃ nanoparticles-decorated BSCF (BSCF-S).

2.1.4. BSCF-6-HCl synthesis

Hydrochloric acid (HCl, 36% of mass fraction) solution was diluted with distilled water to 0.1 M. BSCF-600 sample was immersed into a 20 ml 0.1 mol L^-1 HCl solution, which was stirred to remove the carbonate from BSCF-600 surface, resulting into BSCF-6-HCl.

2.2. Characterization

To identify the phase of the catalyst, powder X-ray diffraction (XRD) was performed at room temperature on a Rigaku Smartlab device with CuKα radiation at 40 kV (wavelength = 1.5418 Å). The diffraction was carried out at 2θ range of 10° to 90°, at 0.02° scanning step, and at 20 min^-1 scan rate. The surface morphology and the particle size of the catalyst were evaluated using scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM). High angle annular dark field scanning transmission electron microscopy (HAADF-STEM) was performed. The distribution of the catalyst constituents was obtained using the energy dispersive mapping over the same sample. The chemical composition and oxidation state of the catalyst constituents were analyzed using X-ray photoelectron spectroscopy (XPS, PHI5000 VersaProbe) with an Al-Kα X-ray radiation source. The XPS data were processed and analyzed using XPSPEAK software. Nitrogen sorption (BELSORP II) was used to measure the specific surface areas of the catalyst, which was calculated by Brunauer-Emmett-Teller (BET) method. The pore size distribution was obtained from the desorption data using Barrett-Joyner-Halenda (BJH) method. The carbon dioxide temperature-programmed desorption (CO₂-TPD) test was performed using mass spectrometer. Approximately 50 mg of catalyst powder was pre-treated by vacuuming to remove contaminant. The powder was then heated from 50 to 1000 °C at 10 °C min^-1 ramping rate under flowing argon at 20 mL min^-1 flow rate.

2.3. Electrochemical tests

Electrochemical tests were performed in a typical three-electrode system (Pine Research Instrumentation) with a rotating disk electrode (RDE) using an electrochemical workstation (CHI 760E). A platinum (Pt) wire, an Ag|AgCl (3.5 mol L^-1 KCl) and a glassy carbon electrode (GC-RDE, 5 mm diameter, 0.196 cm²) were used as the counter, reference, and working electrodes, respectively. Catalyst suspension was also prepared in which 10 mg of catalyst, 10 mg of conductive carbon (Super P Li), 0.1 mL of Nafion solution (5 wt%, Sigma-Aldrich), and 1 mL of ethanol were mixed and sonicated for 1 h. Then, a 5 μL aliquot of this catalyst suspension was dropped onto the surface of the GC electrode, resulting into an approximate catalyst loading of 0.232 mg_catcm^-2, which was left to dry overnight in air for the electrochemical tests.

The electrolyte solution was an alkaline solution, i.e., 0.1 mol L^-1 KOH solution, which was saturated with O₂ for 30 min before each test to provide an oxygen-saturated atmosphere. Cyclic voltammetry (CV) tests were carried out for at least 5 times until overlapping curves were obtained, which removed air bubbles on the surface of RDE and activated the catalyst on the electrode. Each CV was performed from 0.2 V to 0.8 V (vs. Ag|AgCl (3.5 mol L^-1 KCl) at a 100 mV s^-1 scan rate. Linear sweep voltammetry (LSV) was carried out from 0.2 V to 1.0 V (vs. Ag|AgCl (3.5 mol L^-1 KCl)) in an O₂-saturated 0.1 mol L^-1 KOH solution at 5 mV s^-1 scan rate and 1600 rpm rotation rate. All the electrochemical tests were performed at room temperature. Note that all corrected potential values (E_iR-corrected) presented in this work were calibrated to the reversible hydrogen electrode (RHE, see more details in the Fig. S1 in supporting information) and were iR-corrected to compensate for the effect of iR-drop. The iR correction calculation follows E_iR-corrected = E-iR, where E is potential calibrated to the Ag|AgCl electrode, and i is the current and R is the resistance of the electrolyte (around 45 Ω) obtained using high-frequency alternate current (AC) impendence spectroscopy.

Electrochemical impedance spectra (EIS) were recorded at 0.6 V vs. Ag|AgCl (3.5 mol L^-1 KCl) with frequencies ranging from 10⁵ Hz to 0.1 Hz at a bias voltage of 5 mV. Electric conductivity tests were carried out by four-probe meter at room temperature. Electrochemically active surface area (ECSA) was determined by measuring the electrochemical double layer capacitance (C_dl) via cyclic voltammetry (CV) in an O₂-saturated 0.1 mol L^-1 KOH solution. CVs between 1.15 V and 1.25 V vs. RHE at different scan rates of 20, 40, 60, 80, 100, 120, 140, and 160 mV s^-1.

The LSV-based stability tests were carried out after CV tests at a scan rate of 100 mV s^-1 for 1, 500, 1000 and 1500 continuous cycles. In addition, chronopotentiometry (CP) tests were also performed at a 10 mA cm^-2 current density. In brief, the CP test was performed in three-electrode configuration where a platinum (Pt) wire, an Ag|AgCl (3.5 mol L^-1 KCl), and a self-supporting carbon-cloth loaded with catalyst with dimension of 0.5 cm × 1.2 cm were used as the counter, reference, and working electrodes, respectively. The homogenous slurry was deposited uniformly on carbon cloth; leading to the catalyst loading of $ \widetilde{2}$ mg_oxidecm^-2. The slurry was dried under air for two hours and then tested in an O₂-saturated 0.1 mol L^-1 KOH solution.

3. Results and discussion

3.1. Structure, morphology, and CO₂-TPD profile

Fig. 2(a) shows the powder X-ray diffraction (XRD) patterns of 1000 °C calcined Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (BSCF) and the CO₂-exposed BSCF at 500 °C, 600 °C, and 700 °C (BSCF-500, BSCF-600, and BSCF-700) powders. All powders exhibit characteristic peaks of primitive cubic perovskite lattice that can be indexed according to space group Pm-3m (#221) (PDF#01-075-0426) [12]. No distinct characteristic peaks of carbonate compounds can be identified in the powder XRD pattern of BSCF-500; most probably due to the very low amount of carbonate compounds in this sample that is below the detection limit of XRD. CO₂ exposure at higher temperature of 600 °C nonetheless led to the appearance of characteristic peaks of SrCO₃ (PDF#01-074-1623) and Sr_0.5Ba_0.5CO₃ (PDF#00-047-0224) in BSCF-600 pattern. In BSCF-700 case, these carbonate peaks appear to be more intense while the characteristic peaks of the cubic perovskite phase become weaker; indicating enhanced CO₂ reaction with the perovskite oxide at higher temperature. Another powder sample denoted as BSCF-S was prepared by wet impregnation of strontium acetate (SrC₄H₆O₄) into 1000 °C calcined BSCF powder followed by calcination at 750 °C for 2 h, resulting into strontium carbonate (SrCO₃)-decorated BSCF, which will be used in the following electrochemical tests. The presence of strontium carbonate characteristic peaks (PDF#00-005-0418) on the powder XRD pattern of BSCF-S confirmed the successful impregnation of SrCO₃ on the surface of BSCF (Fig. S2). X-ray photoelectron spectroscopy was then used to determine the constituents of BSCF, BSCF-500, BSCF-600, and BSCF-700; the survey scan spectra of which are displayed in Fig. 2(b). In the spectra of all four samples, intense peaks characteristics of Ba, Co, and O are present. There are marked peaks of Ba 3p_3/2 peaks at binding energies of 775 eV and 792 eV, which are hard to distinguish due to their overlapping with the Co peaks. The other less intense peaks at 130 eV and 256 eV and at 711 eV come from Sr and Fe, respectively. In addition, the signature peak of the carbon can be observed at 283 eV on all four spectra. These carbon signals likely come from different carbon species, i.e., surface carbon in BSCF case and carbonate in BSCF-500, BSCF-600, and BSCF-700.

View original graphic| Download| PPT slide

Fig. 2.   (a) Powder XRD patterns and (b) X-ray photoelectron spectroscopy (XPS) spectra of BSCF, BSCF-500, BSCF-600, and BSCF-700.

SEM images of the surfaces of BSCF-500, BSCF-600, BSCF-700, and BSCF are displayed in Fig. 3(a)-(d), respectively. BSCF clearly features dense surface as is the case for BSCF-500 (Fig. 3(a) and (d)). In BSCF-600 and BSCF-700, on the other hand, rougher surfaces from perovskite decomposition and aggregates of particles that correspond to carbonates can be observed (Fig. 3(b) and (c)). The average size of the aggregates in BSCF-700 is larger relative to BSCF-600.

View original graphic| Download| PPT slide

Fig. 3.   SEM images of (a) BSCF-500, (b) BSCF-600, (c) BSCF-700, and (d) BSCF.

The HRTEM image of BSCF-600 is shown in Fig. 4(a). Two different lattice fringes with crystal plane spacing (d) of 0.32 nm and 0.29 nm were observed, which correspond to SrCO₃ and Sr_0.5Ba_0.5CO₃, respectively, consistent with powder XRD results above (Fig. 2(a)). Combined HAADF-STEM image and energy dispersive X-ray (EDX) mapping profiles of Ba, Sr, Co, Fe, O, and C reveal homogeneous distribution of these components over the surface of BSCF-600 (Fig. S3).

View original graphic| Download| PPT slide

Fig. 4.   (a) HRTEM images of BSCF-600 and (b) CO₂-TPD profiles of BSCF, BSCF-500, BSCF-600, and BSCF-700 when heated from 50 °C to 1000 °C at 10 °C min^-1 ramping rate in flowing argon gas at 20 mL min^-1 flow rate.

The specific surface areas of BSCF, BSCF-500, BSCF-600, and BSCF-700 were also determined using nitrogen (N₂) sorption measurements; the isotherms of which are presented in Fig. S4. All four isotherms display hysteresis profiles that are characteristic of mesoporous material. The BET specific surface areas of BSCF, BSCF-500, BSCF-600, and BSCF-700 were 0.44 m² g^-1, 0.61 m² g^-1, 1.47 m² g^-1, and 3.17 m² g^-1, respectively. Pore size distributions obtained from these isotherms using BJH method reveal similar trend for all four samples where the majority of the pores have size that lies between 3 nm and 4 nm.

Carbon dioxide temperature programmed desorption (CO₂-TPD) profiles of BSCF, BSCF-500, BSCF-600, and BSCF-700 are presented in Fig. 4(b). The peak area and its position on such profile indicate how much CO₂ was released from carbonate decomposition and the extent of the carbonate bonding strength. It is apparent that BSCF-700 released the largest amount of CO₂ at the highest peak temperature of 843 °C. The carbonate amount and bonding strength increases in the order of BSCF, BSCF-500, BSCF-600, and BSCF-700, which is consistent with the carbonate content trend observed in their powder XRD patterns and SEM (Figs. 2(a) and 3). The largest carbonate amount in BSCF-700 is anticipated to restrict the contact between the OER reactive species, i.e., OH^- and the active surface sites required for the OER.

3.2. Oxygen evolution reaction performance

Fig. 5(a) shows LSVs over the OER portion for BSCF, BSCF-600, and IrO₂ obtained using RDE at 5 mV s^-1 scan rate and 1600 rpm rotation rate in an O₂-saturated 0.1 mol L^-1 KOH solution. LSV of SrCO₃ is also shown to evaluate its possible contribution to OER performance. SrCO₃ displays negligible OER signal up to 1.9 V vs. RHE; indicating its negligible OER contribution. BSCF-600 exhibited almost identical OER onset potential of around 1.46 V vs. RHE to the IrO₂ catalyst benchmark. This onset potential is lower than those for BSCF and BSCF-S; suggesting higher OER activities of BSCF-600 and IrO₂. Note that BSCF-S is strontium carbonate (SrCO₃)-decorated BSCF prepared previously by wet impregnating strontium acetate (SrC₄H₆O₄) into BSCF powder and calcination at 750 °C for 2 h. It becomes clear that SrCO₃ presence on the surface by wet impregnation did not contribute to OER performance enhancement as effectively as SrCO₃ presence obtained in situ by heat treatment in CO₂ atmosphere. This is likely attributed to the presence of additional oxygen vacancies on the CO₂-exposed BSCF powder from the dissolution of some Ba²⁺ and Sr²⁺ from BSCF perovskite lattice that react with CO₂ to form carbonates on such CO₂-exposed BSCF powder. For CO₂-exposed BSCF samples, the OER activity increased in the order of BSCF-500 < BSCF-700 < BSCF-600 (Fig. S5). To evaluate the effect of removing carbonate compounds from the surface of BSCF-600, BSCF-600 sample was acid washed in a 0.1 mol L^-1 hydrochloric acid (HCl) solution; the carbonate-free sample of which is denoted as BSCF-6-HCl (Fig. 5(c)). BSCF-6-HCl displayed OER performance that is higher than BSCF but is lower than BSCF-600. This is likely attributed to the simultaneous removal of carbonate and amorphous BSCF layers.

View original graphic| Download| PPT slide

Fig. 5.   (a) LSVs over OER portion for BSCF, BSCF-600, BSCF-S, IrO₂, BSCF-6-HCl, and SrCO₃ obtained using rotating disk electrode at 5 mV s^-1 scan rate and 1600 rpm rotation rate in an O₂-saturated 0.1 mol L^-1 KOH solution. All profiles were taken from the second scan profiles, (b) Tafel plots for BSCF-600, BSCF and IrO₂ between 0.1 and 0.5 mA cm^-2, (c) powder XRD patterns of BSCF, BSCF-6-HCl, and (d) powder XRD patterns of BSCF-600 before and after long term CP test.

An overpotential (η) required to achieve the current density of 10 mA cm^-2 (based on 10% solar water-splitting conversion efficiency) represents an important metric of relevance to solar fuel synthesis [21], thus comparing the overpotential of the catalyst samples at this current density is of significant interest. Note that the overpotential here is defined as the difference between the OER potential at 10 mA cm^-2 and the theoretical reversible potential (1.23 V vs. RHE). BSCF-600 can deliver 10 mA cm^-2 current density at a low overpotential of 0.36 V, which is significantly lower than 0.45 V, 0.46 V, and 0.47 V for BSCF, IrO₂, and BSCF-S, respectively (Fig. 5(a)). Under identical experimental condition, the overpotential of BSCF-600 at 10 mA cm^-2 current density was also lower than those of BSCF-500 and BSCF-700 (Fig. S5). Table 1 summarizes the mass activity (MA, J_MA=J/m, J is the current density and m is the mass loading) and the specific activity (SA, J_SA=J/(m·S_BET), S_BET is the specific surface area) of BSCF, BSCF-500, BSCF-600, and BSCF-700. At η of 0.4 V, BSCF-600 has a mass activity of 74.14 $ Ag_{cat}^{-1}$ (mass activity of perovskite oxide catalyst) and a specific activity of 5.04 mA $cm_{cat}^{-2}$, which are $\widetilde{3}$.9 and $\widetilde{1}$.2 times higher than those for BSCF (18.78 Agcat-1 and 4.27 mA cmcat-2), indicating the high OER activity of BSCF-600. Tafel plots show the relationship of the overpotential to the current density and provide an important insight into the OER kinetics. Fig. 5(b) displays the Tafel plots for BSCF-600, BSCF, and IrO₂. Tafel slope of BSCF-600 (63 mV dec^-1) was smaller than those of BSCF (73 mV dec^-1) and the IrO₂ benchmark catalyst (78 mV dec^-1); confirming BSCF fastest kinetics among the three samples. Table 1 provides an overview of Tafel slopes for BSCF, BSCF-500, BSCF-600, and BSCF-700.

Besides the OER activity, long term performance stability is required for practical application. The performance stability of BSCF-600 was evaluated against BSCF and IrO₂ using CP test in an O₂-saturated 0.1 mol L^-1 KOH solution. The CP test was performed in three-electrode self-supporting configuration, and we deposited the catalysts ink uniformly into a carbon cloth of 0.5 cm × 1.2 cm dimension, leading to a final loading of 2 mg cm^-2. Fig. 6(a) displays the time-dependent potential profile of BSCF-600, BSCF, and IrO₂ obtained upon maintaining a constant current density of 10 mA cm^-2 for up to 800 min. Over this period, BSCF-600 can retain the potential at a relatively stable value of $\widetilde{2}$.61 V; which highlights its performance durability. IrO₂, in contrast, exhibited rapid potential increase over the first 60 min. BSCF showed higher potential than that shown by BSCF-600 with more significant fluctuation over the first 400 min. BSCF-6-HCl, which is free from carbonate and amorphous BSCF layers, also displayed analogous behavior to BSCF-600, although at a slightly higher potential. After acid treatment, amorphous layer and carbonate particles were removed, then BSCF with more oxygen vacancy was exposed, which accounts for the better stability in alkaline media. The same durability behavior of BSCF-600 and BSCF-6-HCl hints toward almost identical efficiency obtained by having amorphous layer and after such layer is removed. Fig. 6(b) presents the LSVs of BSCF-600 at the 1 st, 500th, 1000th, and 1500th cycle scans during the continuous 1500 CV scans using RDE at the same experimental condition. Despite the more significant increase of the potential profile from the 1 st to the 500th scan, the subsequent LSVs profiles up to 1500th scan lie almost next to each other, supporting the performance durability observed in CP test result. Moreover, powder XRD patterns of BSCF-600 before and after the CP test feature similar diffraction peaks, with the exception of lower peaks intensities after the test (Fig. 5(d)). This indicates the reduction in the BSCF phase crystallinity following the long term CP test. The long term stability performance discrepancy between BSCF-600 and BSCF implies the positive contribution of in situ formed carbonate layer into the performance durability of BSCF.

View original graphic| Download| PPT slide

Fig. 6.   (a) Time-dependent potential profile from CP test of BSCF-600, BSCF, and IrO₂ obtained using carbon cloth as catalysts support upon maintaining a constant current density of 10 mA cm^-2 in an O₂-saturated 0.1 mol L^-1 KOH solution and (b) LSVs of BSCF-600 using RDE at 1600 rpm rotation rate after CVs at the 1st, 500th, 1000th and 1500th cycle scans.

The higher OER performance of BSCF-600 relative to BSCF is attributed to several factors. Firstly, compared to BSCF, BSCF-600 has larger electrochemically active surface area (ECSA). The ECSAs of BSCF-600 and BSCF were determined by measuring the electrochemical double layer capacitance (C_dl) via cyclic voltammetry tests between 1.15 V and 1.25 V vs. RHE at different scan rates of 20, 40, 60, 80, 100, 120, 140, and 160 mV s^-1 (Figs. 7(a) and S6). The C_dl value is obtained from the slope of the plot of half of the current density difference obtained at 1.2 V vs. RHE vs. scan rate (Fig. 7(b)) [22]. The C_dl value of BSCF-600 of 2.68 m F cm^-2 is higher than that of BSCF of 0.98 m F cm^-2, which signifies larger ECSA for BSCF-600 relative to BSCF (Fig. 7(b)). Larger ECSA is anticipated to promote the OER process. Secondly, BSCF-600 showed faster charge transfer rate compared to BSCF. Electrochemical impedance spectra of BSCF-600 and BSCF obtained in an O₂-saturated 0.1 mol L^-1 KOH solution are shown in Fig. 7(c). The semicircular diameter of the spectra reflects electrochemical impedance R_ct value. At η of 0.7 V, the R_ct value of BSCF-600 is lower than that of BSCF, indicating the faster charge transfer resistance of BSCF-600 relative to BSCF. Such faster charge transfer resistance of BSCF-600 is consistent with the higher electrical conductivity of BSCF-600 relative to BSCF (Table S1) that contributes to the enhancement in OER performance. The electrocatalytic activity of oxide catalysts depends on the electrical conductivity. Lastly, we further analyze the oxidation state of Co in BSCF and BSCF-600. Co 3p XPS spectra of BSCF-600 and BSCF are presented in Fig. S7. Determining the surface oxidation of Co however becomes a challenging task given the overlapping of Co 2p and Ba 3d cardinal lines [23,24]. There are Co 2p_3/2 and Ba 3d_5/2 at binding energy of ~780 eV and Co 2p_1/2 and Ba 3d_3/2 at binding energy of ~795 eV (Fig. 7(b)). Nonetheless, it is clear that the peaks positions for Co 2p_3/2 and 2p_1/2 are shifted to higher binding energies for BSCF-600 relative to BSCF, which suggests that the Co²⁺ content in BSCF-600 is larger than that in BSCF. The larger amount of Co²⁺ should lead to a shift toward higher binding energies [25,26]. Study by Fabbri et al. confirmed that BSCF with larger Co²⁺ content exhibits enhanced electrochemical activity towards OER [14]. In this work, we show that the superior OER activity of BSCF-600 relative to BSCF may also be associated with the lower Co oxidation state in BSCF-600. Recent operando OER mechanism study by Fabbri et al. on BSCF correlated its high OER activity to the formation of a self-assembled metal oxy (hydroxide) surface film, which is triggered by the lattice oxygen evolution reaction (LOER) [20]. During the OER, an increase in the Co oxidation state and a secondary CoO(OH) layer formation were observed in BSCF, which Fabbri et al. defined as “dynamic surface self-reconstruction” phenomena. They attributed high OER activity to such phenomena. Although we have shown that the formation of a thin amorphous carbonate layer on BSCF-600 surface enhance its OER activity relative to BSCF, which suggests the beneficial effect of such layer to the “self-reconstruction” of BSCF during OER, more detailed operando OER mechanism study is required to reveal the mechanism, which lies outside the scope of this work.

View original graphic| Download| PPT slide

Fig. 7.   (a) CV profiles of BSCF-600 in an O₂-saturated 0.1 mol L^-1 KOH solution at different scan rates of 20, 40, 60, 80, 100, 120, 140, and 160 mV s^-1, (b) plot of the half of current density difference at 1.2 V vs. RHE vs. scan rate for BSCF-600 and BSCF, (c) EIS of BSCF-600 and BSCF obtained at an overpotential of 0.7 V and AC bias potential of 10 mV, and (d) Co 2p and Ba 3d XPS spectra of BSCF-600 and BSCF (Z′: real part of impedance; Z′′: imaginary part of impedance; Sat: satellite peak).

4. Conclusion

It has been demonstrated that conventional heat treatment in CO₂ atmosphere can create in situ carbonate layer on the surface of Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3-δ (BSCF) that eventually enhanced its OER activity in an O₂-saturated 0.1 mol L^-1 KOH solution relative to the original BSCF. Heat treatment in CO₂ was performed at three different temperatures, i.e., 500 °C, 600 °C, and 700 °C; the samples of which are denoted as BSCF-500, BSCF-600, and BSCF-700. Powder XRD results reveal that the carbonates (i.e., SrCO₃ and Sr_0.5Ba_0.5CO₃) content increased with temperature rise. Nitrogen sorption results indicates increasing specific surface area with temperature rise. These trends are consistent to the trends observed in SEM images and CO₂-TPD profiles. Electrochemical tests nonetheless show that BSCF-600 exhibited the highest OER activity among the three samples. BSCF-600 also showed lower Tafel slope than BSCF and the benchmark IrO₂ catalyst, which suggests its fastest OER kinetics. Notably, BSCF-600 displayed excellent operational durability throughout 800 min chronopotentiometry at a constant current density of 10 mA cm^-2 and during 1500 cycle continuous CV scans. The factors contributing to the enhanced OER activity for BSCF-600 relative to BSCF have been discussed, which are larger electrochemically active surface area, faster charge transfer rate and higher electrical conductivity, and modified oxidation state for cobalt ion. The high durability shown by carbonate-decorated BSCF appears to be related to the suppression of barium and strontium ions from its surface. More comprehensive future studies should be performed to provide more insights into the mechanism. What we have shown here may be extended to other high performance OER electrocatalysts.

Acknowledgments

This work was supported financially by the National Natural Science Foundation of China (Nos. 51502138 and 51506085) and the Natural Science Foundation of Jiangsu Province (Nos. BK20150738 and BK20150742).

Appendix A. Supplementary data

Supplementary material related to this article can be found, inthe online version, at doi:https://doi.org/10.1016/j.jmst.2019.01.005.

The authors have declared that no competing interests exist.