MnxCo3-xO4 spinel oxides as efficient oxygen evolution reaction catalysts in alkaline

3-x O 4 spinel oxides were synthesised via a soft chemistry method. (cid:1) The effect of Mn addition on structural and physicochemical properties was investigated. (cid:1) Synthesised oxides powders were examined as OER catalysts in alkaline medium. (cid:1) Synergistic effect of dual phase MnCo 2 O 4 improves the OER catalytic activity. for oxygen evolution reaction (OER) is an essential task in developing sustainable water splitting technology for the production of hydrogen. In this work, manganese cobalt spinel oxides with a general formula of Mn x Co 3-x O 4 (x ¼ 0, 0.5, 1, 1.5, 2) were synthesised via a soft chemistry method. Non-equilibrium mixed powder compositions were produced, resulting in high electrocatalytic activity. The oxygen evolution reaction was evaluated in an alkaline medium (1 M KOH). It was shown that the addition of Mn (up to x (cid:3) 1) to the cubic Co 3 O 4 phase results in an increase of the electrocatalytic performance. The lowest overpotential was obtained for the composition designated as MnCo 2 O 4 ,which exhibited a dual-phase structure(~30% Co 3 O 4 þ 70% Mn 1.4 Co 1.6 O 4 ): the benchmark current density of 10 mA cm (cid:4) 2 was achieved at the relatively low overpotential of 327 mV. The corresponding Tafel slope was determined to be ~79 mV dec (cid:4) 1 . Stabilities of the electrodes were tested for 25 h, showing degradation of the MnCo 2 O 4 powder, but no degradation, or even a slight activation for other spinels. ©


Introduction
Among different hydrogen production methods, water electrolysis seems to be a viable process to provide clean hydrogen [1,2], especially when coupled with renewable electricity production. The key point impeding the electrolysis process is the sluggish oxygen evolution reaction (OER) kinetics of the multi-electron charge transfer reaction resulting in high reaction overpotentials [3,4]. For this reason, effective electrocatalysts may be applied to accelerate the reaction and reduce the overpotential. The most active catalysts for water splitting reactions, hydrogen and oxygen evolution, are based on precious elements: Pt and IrO 2 , respectively. Their low abundance in the Earth's crust coupled with high costs significantly limits those elements' application for water splitting electrodes.
Recently  [21]. Cubic structure Co 3 O 4 spinel nanoparticles with average diameters of 5.9 nm were synthesised by Esswein et al. and demonstrated h of 328 mV at 10 mA cm À2 in 1.0 M KOH [22]. One of the promising spinel group materials is mixed-valent Mn x Co 3-x O 4 due to the possibility of tuning the cation distribution by changing the synthesis parameters and composition. Mn-Co spinels have interesting flexible properties and have been reported both for oxygen reduction [23] and evolution processes. Han et al. fabricated mixed phases of Coand Mn-based oxides which for OER demonstrated h of 450 mV at 10 mA cm À2 in 1.0 M KOH with a low Tafel slope of 35.8 mV dec À1 [24]. The hybrid nanostructure of manganese cobaltite/nitrogen-doped multi-walled carbon nanotubes (Mn x Co 3-x O 4 @NCNTs) was proposed by Zhao et al. and exhibited an overpotential of 470 mV at 10 mA cm À2 in 0.1 M KOH [25].
The physicochemical properties of the spinels can be tailored by manipulating different cations and also by altering their preparation methods. Spinels can be synthesised by different approaches, including physical and preferably chemical processes: hard-template [26], wet-chemical [27], hydrothermal [28,29], solvothermal [30], spray pyrolysis [31], co-precipitation [32], and sol-gel [22,33e35]. Depending on the thermal history, the cation distribution may vary resulting in data scattering and uncertainties.
In this work, series of Mn x Co 3-x O 4 (x ¼ 0, 0.5, 1, 1.5, 2) spinels were prepared by an ethylenediamine tetra acetic acid -citric acid -ethylene glycol (EDTA-CA-EG) process, and were tested as OER catalysts in an alkaline media (1M KOH). The purpose of this study was to explore the relationship between the Mn content in the Co-based spinel structure and its electrocatalytic activity.

Powder synthesis
The spinel powders evaluated in this work were synthesised via a sol-gel EDTA-citric acid-ethylene glycol method. The initial compounds with thermogravimetrically standardised cation content: Mn(NO 3 ) 2 $4H 2 O and Co(NO 3 ) 2 $6H 2 O (analytical grade) were dissolved in deionised water and added to a mixture of CA and EDTA. The solution was stirred for 10 min on a magnetic stirrer at room temperature. In order to avoid the risk of citrate protonation or metal hydroxide precipitation, the pH of the solution was sustained at a level of 6 by the addition of ammonia solution (25% solution, POCH, Poland). Ethylene glycol (EG, Aldrich) was added and the temperature of the solution was increased up to 80 C in order to initiate the transesterification reaction, and was kept overnight until all solvents evaporated, leaving a semi-solid gel. The specific molar ratio of the reagents was maintained, i.e. TMI:CA:EDTA:EG ¼ 1:2:1:2 (TMI -Total Metal Ions). The obtained gel was heated in a muffle furnace first at 130 C to ensure the evaporation of residual solvents. Subsequently, the gel was pre-calcined at 400 C, ground in an agate mortar and calcined at a final temperature of 600 C. The calcined powders were ball-milled in isopropanol with yttria-stabilised zirconia (YSZ) spherical grinding media (43 mm) for the purpose of agglomerate defragmentation with the intention of improving electrocatalytic performance [4]. After milling, powders were dried in a laboratory drier at 100 C in the air.

Powder analysis
Crystalline phases were identified by powder X-ray diffraction (pXRD), conducted at room temperature on a Bruker D2 Phaser diffractometer with CuK a radiation (l ¼ 1.5404 A) and a Lyn-xEye XE-T detector. The results were analysed by Rietveld refinement using the Fullprof software package [36]. The.cif files required to perform the analyses were downloaded from the Crystallography Open Database [37,38]. Calculation of MnOx-CoOx phase diagram in air was done utilising the Thermo-Calc software [39] with the thermodynamic description of Co-Mn-O developed by Weiland [40]. FEI Quanta 250 FEG, and Thermo Fisher Phenom XL scanning Electron international journal of hydrogen energy 45 (2020) 14867e14879 Microscopes (SEM), and a C s -corrected Titan Cubed G2 60-300 (FEI) Scanning Transmission Electron Microscope (S/TEM) were used to investigate the morphologies and chemical composition (using the ChemiSTEM EDX system based on 4 windowless Silicon Drift Detectors (Super X) of the asprepared powders). The powders for TEM investigations were prepared traditionally: a drop of a water suspension containing the powder was placed onto a copper grid followed by vacuum drying. The specific surface areas were determined by means of the 10-point BrunauereEmmetteTeller (BET) method: the N 2 adsorption isotherms were measured by a Quantachrome Instruments Autosorb iQ analyser. Assuming a spherical geometry, the mean sizes of the powders' particles were calculated by the following equation: where d is particle diameter, A BET is the BET specific surface area, and r is the theoretical density calculated based on data obtained from Rietveld.

Preparation of electrodes for OER
Highly porous Ni foam (1.6 mm thick, 110 ppi, Shanghai Tankii Alloy Material Co. Ltd., China) was cut into smaller samples in the shape of a square with a working (active) area of 1 cm 2 with an additional holding/electrical connection part. Before the deposition of the catalyst inks, the nickel was cleaned in deionised water and, subsequently in acetone for 5 min in an ultrasonic bath. Then, the Ni foam was dried in a laboratory drier at 100 C in the air for at least 1 h. The electrocatalyst inks were prepared by mixing 100 mg of the prepared spinel powders, 100 mg Timcal Super C45 Conductive Carbon Black (Imerys, Belgium), and 6 ml of Ethyl Alcohol Absolut 99.8% Pure (POCH, Poland). The prepared mixture was ball-milled for 24 h. Afterwards, the milling balls were rinsed with 4 ml of ethanol and the powder drying step was skipped to avoid reagglomeration. After the milling step, 1 ml of 5% Nafion 117 solution was added to the mixture as an ink dispersant and a catalyst binder [41,42]. At the end, the ink was sonicated in a water bath for 30 min to obtain a homogeneous colloid. The catalyst ink was pipetted onto the cleaned Ni foam to achieve a mass loading of 1.3 mg, and dried overnight in ambient conditions. For comparison, reference IrO 2 (PK Catalyst, FuelCellStore, USA) electrodes were prepared in the same manner as the other ones.

Electrochemical tests
The electrochemical measurements were performed in a three-electrode glass cell system in 1.0 M KOH aqueous solution. A HydroFlex reversible hydrogen reference electrode (RHE) (Gaskatel, Germany), a Pt sheet, and a coated Ni foam were used as the reference (RE), counter (CE), and working (WE) electrodes, respectively. The electrochemical tests were performed using an Ametek VersaSTAT 4 potentiostat/galvanostat/impedance meter. All measurements were carried out at 25 C, maintained by a Julabo F12 thermostat. Before each measurement, the electrolyte was purged with high purity argon for 20 min. After purging, the working electrode was cycled between 1.1 V and 1.6 V vs. RHE at a scan rate of 100 mV s À1 for 40 (example of activation procedure is shown Fig. S1). This treatment reduces the activity of the electrodes but allows to achieve stable performance and reproducible results. Cycling voltammetry (CV) scans were performed in the non-faradaic potential region from 1.15 V to 1.25 V (vs. RHE) at scan rates of 10, 20, 40, 60, 80, 100 mV s À1 in order to estimate the double-layer capacitance (C dl ). Then, the values of C dl were used to assess the electrochemical active surface area (ECSA) of the material [21,43]. To obtain the surface area values from the specific capacitances, a reference capacitance value of 40 mFc m À2 was assumed [44]. Linear sweep voltammetry (LSV) data were collected from 1.1 to 2.0 V (vs. RHE) at a scan rate of 5 mV s À1 . Electrochemical impedance spectroscopy (EIS) measurements were carried out to investigate the electrical properties of the materials, i.e. the charge transfer resistance R ct . Measurements were performed in the frequency range from 10 kHz to 0.1 Hz at 1.7 V vs. RHE with an amplitude of 10 mV. For the quantitative analysis, a fitting to a modified Randles equivalent circuit was carried out using the Zview software. To evaluate the temporal stability of the electrodes, chronopotentiometry was performed at 10 mA cm À2 for 25 h.
All potential values were iR-corrected to remove the effect of solution resistance according to the equation E iR-corrected- where i is the current, and R un is an uncompensated ohmic electrolyte resistance. The overpotential (h) for the oxygen evolution reaction was calculated by the following equation: h ¼ E (10 mA$cm À2 ) e 1.23 V (vs. RHE). The current density was normalised by the geometrical surface area of the working electrode (1 cm 2 ). The catalyst mass activity at specific overpotential values was calculated by current normalisation by the catalyst oxide loading (~1.3 mg).

Results and discussion
Catalyst characterisation Mn x Co 3-x O 4 (x ¼ 0, 0.5, 1, 1.5, 2) spinel powders were synthesised by a EDTA-CA-EG process. Utilisation of the softchemistry based synthesis method leads to powders with a small particle size and high specific surface area, which is beneficial for the catalysis. The powders' morphologies, observed by SEM, are presented in Fig. 1. Submicron particles were obtained for all compositions. It was observed that the introduction of Mn into the Co-based spinel results in finer particles (enlarged images are shown in Fig. S2). This observation is confirmed by specific surface area (SSA-BET) measurements. The specific surface area of the powders increases on the addition of Mn with x ¼ 0.5, and then decreases for a higher Mn content. The Mn 0.5 Co 2.5 O 4 powder has the highest specific area of 32.9 m 2 g À1 . Interestingly, for the Mn-rich Mn 2 CoO 4 composition, the SSA is similar to the pure Co 3 O 4 spinel, indicating a complex influence of the Mn on the particle size. Based on the SSA, particle sizes were calculated, and the values are presented in Table 1. The obtained values are consistent with the SEM images, denoting submicrometric grains, with particles in the 30e100 nm range.
international journal of hydrogen energy 45 (2020) 14867e14879 In order to evaluate the chemical compositions of the synthesised powders, EDS analyses were performed (quantitative results are included in Table 1 and the obtained EDS spectra are shown in Fig. S3). The Mn:Co ratios in the synthesised powders were close to the desired ones.
In order to determine the crystallographic structure and phase purity of the synthesised Mn x Co 3-x O 4 powders, X-ray diffraction (pXRD) characterisation was performed, followed by Rietveld refinement of the unit cell parameters. The pXRD patterns are shown in Fig. 2, and the obtained results are summarised in Table 2 [45].
The powder designated as MnCo 2 O 4 , which was expected to crystallise in a single cubic phase, shows the presence of two phases. Analysis of the peaks reveals the formation of two phases with similar structure (with the same space group). The phases differ by cation composition and thus a shift of the peaks is visible in the spectra. One of the phases (based on lattice constant value) is similar to the pure Co 3 O 4 phase, and the other must contain the Mn and Co. The elemental content, i.e. the Mn:Co ratio in the powder, has been confirmed by EDS, as presented in Table 1. For this particular cation ratio, a single cubic phase is expected as presented in the literature [46e48]. By rough analysis of the peaks' intensity, the amount of Co 3 O 4 phase in the mixed powder is estimated to be~30 vol%. Assuming that some of the Co atoms were consumed to form the pure Co 3 O 4 phase, the rough calculation of the stoichiometry of the second phase is Mn 1.4 Co 1.6 O 4 .
In order to verify the result, the synthesis of the spinel was repeated, and the same result was obtained. In order to exclude possible influence of a too-short annealing step, an additional thermal annealing step was carried out. The powders were subjected to a second heat treatment of 600 C for 2 h, yielding the same mixed structure (Fig. S4). Therefore the structure of the spinel is reproducible. For the spinel initially designated as MnCo 2 O 4 , to highlight the phase composition of the powder, the name "dual-phase MnCo 2 O 4 " spinel will be used in the remainder of the work. We postulate, that the mixed-phase composition of the dual-phase MnCo 2 O 4 spinel (Co 3 O 4 þMn 1.4 Co 1.6 O 4 ) results from the use of a modified EDTA-CA-EG powder synthesis method. When utilising a pure EDTA-CA synthesis method (with no EG), only a single cubic spinel was produced (results included in Fig. S5, full results will be reported elsewhere). Interestingly, the single phase MnCo 2 O 4 powder revealed much lower performance towards OER than the dual phase powder. Apparently, due to the different binding/chelating powers of the acids/EG, different powders were produced.
The powder with the intended stoichiometry of Mn 1.5 Co 1.5 O 4 consisted of two phases: cubic (85%) and tetragonal (15%), while the Mn 2 CoO 4 was entirely tetragonal. These powders have structures according to the reported ones.
Calculation of MnO x e CoO x phase diagram in the air was performed in order to compare structures of prepared powders with theoretical data. According to Fig. 2e [51]. The cation composition is thus quite similar to the one obtained for the Mn-containing phase in the MnCo 2 O 4 designated sample. For XRD, the powders were calcined at 800 C. Interestingly, depending on the precipitation agent used, either single or double phase materials were obtained. This highlights the complex nature of the spinels, and the possible effects on their structure. For the case where the pure cubic spinel phase was obtained (using EDTA as the precipitation/chelating agent), the lattice parameter was~8.20 A, so comparable with the one obtained in our study.
In order to further characterise the catalyst with the nominal composition of MnCo 2 O 4 , transmission electron microscopy (TEM) analyses were performed. Fig. 3 (and Fig. S6)  presents high-angle annular dark-field (HAADF) images, a TEM-EDS elemental analysis, and selected area electron diffraction (SAED) patterns. From the TEM images, it can be observed ( Fig. 3a and b) that the synthesised material consists of nanocrystals in two size ranges i.e. < 50 nm and >100 nm. EDS mapping shows that Co and Mn are mutually well dispersed in the powder volume in the case of the small particles (Fig. 3e). Additionally, the HRTEM image (Fig. 3j)  For the electrochemical study, the produced spinel oxides were deposited on highly porous nickel foam substrates (porosity > 95%). For the determination of the coating quality, i.e. the homogeneity of catalyst distribution, scanning   electron and visible light microscopy analyses were performed. Images of the Ni foam before the catalyst deposition and after the catalyst deposition are presented in Fig. 4. The nickel foam had visible grains within a size range of 5e20 mm. The prepared catalyst ink was distributed evenly over the active area of the porous foam, the surface coverage being very uniform. The total amount of the catalyst spinel powder was~1.3 mg. SEM images confirm the uniform microstructure of the catalyst, with no agglomerates visible.

Evaluation of electrocatalytic activity for OER
The electrocatalytic OER performance of the Mn x Co 3-x O 4 spinels was investigated in 1.0 M KOH solution. For comparison, pure Ni foam and Ni foam coated with commercial IrO 2 catalyst were also examined. Based on the LSV polarisation curves presented in Fig. 5a, it can be observed that all of the spinel catalysts outperformed the pure Ni foam electrode. Among the spinels, the lowest performance was determined for the tetragonal Mn 2 CoO 4 , whereas the best performance was found for the dual-phase MnCo 2 O 4 material (Co 3 O 4 / Mn 1.4 Co 1.6 O 4 ).
The electrode overpotential determined at 10 mA cm À2 , Tafel slope, R ct , and ECSA are the most common parameters determined and used to assess the catalytic performance of an electrode towards OER. These values obtained for the studied materials are summarised in Table 3. The results indicate that the lowest overpotential (h)a t1 0m Ac m À2 of 289 mV was exhibited by IrO 2 but its performance weakened at higher overpotentials compared to the other synthesised spinels. Pure tetragonal Mn 2 CoO 4 spinel showed very low catalytic activity, the overpotential was even slightly higher than the overpotential obtained for the pure nickel. Therefore, this phase can be considered inactive towards OER. The activity obtained for the pure Co 3 O 4 spinel was slightly better in comparison to the nickel foam. For the Mn 1.5 Co 1.5 O 4 , which was a mixture of the tetragonal and cubic phases, the performance was comparable to the Co 3 O 4 catalyst.
The dual-phase MnCo 2 O 4 exhibited the lowest h of 327 mV. Since it was a mixture of Co 3 O 4 (~30 vol%) and Mn 1.4 Co 1.6 O 4 (~70 vol%), with the former being only a mediocre catalyst (LSV curves obtained for dp-MnCo 2 O 4 and sp-MnCo 2 O 4 are compared in Fig. S7), either the latter has very high activity or there exist some synergistic interaction between the two phases. Also, as based on the TEM analyses, the Mn-Co phase in the mixed catalyst had a much finer structure, preferable for the surface-driven OER. In any case, the dual-phase spinel designated as MnCo 2 O 4 is an interesting alternative to the state-of-the-art IrO 2 in the role of OER catalyst.
From Fig. 5b, it can be seen that the Tafel slopes for all measured spinels are very similar (70e80 mV dec À1 ) and are slightly lower compared to IrO 2 . The results indicate that the  OER mechanism for all of the spinels is similar. The values of the Tafel slopes are characteristic for the limiting process of the electrochemical reaction. The value of 120 mV dec À1 represents the limiting process of the first electron transfer step, the slope of 60 mV dec À1 is caused by the chemical step, and finally, the value of 40 mV dec À1 is representative of the second electron transfer step [52e54]. The values obtained in this work point to a possible mixed mechanism, with a strong influence of the chemical step. Fig. 6a presents the relationship between the current density and scan rate obtained from cycling voltammetry (CV curves are shown in Fig. S8). Double layer capacitances were used for the calculation of the electrochemically active surface areas (ECSAs) of the catalysts. The collected data was used to calculate C dl , and consecutively ECSA (  [57]. Additionally, the EIS was employed in order to study the electrochemical properties of the material in more detail.  Nyquist plots obtained by EIS are shown in Fig. 6b. As might be predicted, the dual phase MnCo 2 O 4 demonstrates the semicircle with the smallest diameter, which indicates the lowest charge transfer resistance among all of the spinels, and results in superior OER activity. The R s (CPE-R p ) equivalent circuit was chosen based on the literature reports [3,30,58]. The trend of R ct (Table 3) correlates with the trend of h at 10 mA cm À2 i.e. the catalyst with a lower R ct reveals higher OER performance (lower h) due to the faster charge transfer kinetics. The relation suggests that improved charge transfer may be one of the main factors determining OER performance in Mn x Co 3-x O 4 spinels.
Based on the above discussion, it seems that the performance of the MnCo2O4 spinel, which is in reality a dual-phase compound consisting of Co 3 O 4 and Mn 1.15 Co 1.85 O 4 , is the highest one based on the metrics shown. In the following discussion, other performance metrics are shown, indicating the complex nature of catalyst benchmarking. Namely, the specific weight of the catalyst and its surface area (based on BET) are used. Fig. 7a and b present the specific mass activities of the catalysts determined at the h ¼ 327 mV and h ¼ 442 mV (values of h for dual-phase MnCo 2 O 4 needed to achieve a current density of 10 mA cm À2 and 150 mA cm À2 , respectively). In addition, the surface-specific activity was calculated based on the determined BET surface areas and the results are presented in Fig. 7c [60]. These very high performances were obtained on 1 nme20 nm thick films, so their scaling and thus direct comparison to the values with the powder-based catalysts are questionable.
When comparing the surface-specific (BET adjusted) catalyst performance, presented in Fig. 7c, the overpotential values determined at 25 mAc m À2 are quite similar, no large differences are noticed. The lowest overpotential values are obtained for the IrO 2 reference catalyst, followed by the MnCo 2 O 4 catalyst. In this case, it would mean that the performance gain is caused by the extended specific surface area of the powders obtained for Mn-containing compounds. However, this is not entirely true, as the ECSA does not scale linearly with the BET area, and the comparisons are more complex and require further studies.
The catalysts' durability was evaluated by chronopotentiometry tests at a 10 mA cm À2 current density performed for 25 h. From the plots presented in Fig. 8 a) it can be seen that all electrodes exhibited an initial large increase of potential in the first 1e2 h. This can be possibly caused by the formation of sticking O 2 bubbles on the electrode surface, and thus decreasing the effective surface coverage. After a stationary state was achieved, the electrode degradation could be discussed.
Up to~10 h, the spinel materials demonstrated quite stable behaviour. A segment of each plot corresponding to the last 5 h, where the degradation was relatively steady, has been used to estimate the degradation rate. The results plotted in Fig. 8 b) demonstrate a slight loss of performance with a rate of around 0.3e2.5 mV h À1 . MnCo 2 O 4 appears to be the least stable spinel with a performance drop-rate of 2 mV h À1 . The reference IrO 2 catalyst shows the lowest initial overpotential and a slow, but visible degradation rate. Notably, the spinels exhibit better stability than the pure Ni foam electrode.
For a broader comparison with the data obtained in this study, results recently reported in the literature for spinels and Mn/Co-based materials are summarised in Table 4. The presented data shows that the MnCo 2 O 4 on nickel foam synthesised in this work by a EDTA-CA-EG method exhibit improved or comparable electrocatalytic performance to reported spinel-based electrocatalysts. The reached overpotential of 327 mV at 10 mA cm À2 can be considered one of the lowest reported for spinels. The lowest overpotential value of 270 (at 20 mA cm À2 ) for a spinel/sulphide hybrid material in the form of nanosheet was reported by Du et al. [61]. The materials have however a somewhat higher Tafel slope of 130 mV dec À1 . The very low overpotential of this hybrid structure shows the great potential of combining spinels with other materials for improved performance, also supported by other good results [61e63]. Wang [68]. Interestingly, the Co-rich cubic compound outperformed the Mn-rich spinel in OER, but the trend was reversed in ORR. In addition, a study by Rios et al. reported that increasing the content of Co in Mn x Co 3-x O 4 (0 x 1) improved the OER rate, which was related to the higher Co 3þ content in the spinel structure [70]. In their case, the best performance was obtained for a pure Co 3 O 4 spinel. As reported by Wei et al. [47], tetragonal spinel Mn 2 CoO 4 could be expected to exhibit the highest OER catalytic activity, which was not the case in this study.
The differing and sometimes contradicting results obtained by different groups indicate the possible influence of the preparation of the catalyst and measurement procedures on determining the electrocatalytic properties. As pointed out by Wei et al. [47] (for the Mn-Co spinels), and Zhou et al. [71] (in the case of MnFe 2 O 4 ), changes in the annealing temperature of the powders can result in oxidation of Mn 2þ to Mn 3þ and transfer from tetrahedral to octahedral sites in the lattice. Thus the comparison and discussion are altered by the preparation, thermal history and measurements methodology. More studies are required in order to clarify the observed differences.
The good catalytic properties of the heterogeneous hybrid structure of Co 3 O 4 and Mn 1.4 Co 1.6 O 4 spinels (h of 327 mV at 10 mA cm À2 ) can result from various factors. Firstly, the addition of Mn causes the powder's grain refinement with a mean particle diameter smaller than 100 nm (determined by BET). Based on TEM, the smaller particles attributable to the mixed Mn-Co phase are smaller than 50 nm. The grain refinement is directly related to the increase of the BET surface area. Esswein et al. determined that the overpotential for water oxidation at a constant current density of 10 mA cm À2 decreases by~50 mV with the increase of catalyst's surface area by an order of magnitude [22]. The ECSA calculated from the electrode's double-layer capacitance is the highest for the electrode coated with dual-phase MnCo 2 O 4 catalyst (143 cm 2 ). When recalculated per weight of the catalyst (~1.3 mg), the mass-specific ECSA of the dual-phase MnCo 2 O 4 is~11 m 2 g À1 . Compared to other reported catalysts, this value seems relatively high.
Additionally, dual-phase MnCo 2 O 4 exhibits the lowest charge transfer resistance of 0.69 U cm 2 , which indicates a much higher conductivity compared to other Co-Mn-O catalysts [24]. A combination of a dual-phase structure, relatively high specific surface area and the high conductivity of MnCo 2 O 4 are the possible explanations of its remarkable activity compared to other stoichiometries [72]. international journal of hydrogen energy 45 (2020) 14867e14879 Conclusions Mn x Co 3-x O 4 spinel powders were fabricated via a facile EDTA-CA-EG method. The powders were analysed for their surface area, chemical and phase composition, and their electrocatalytic properties towards oxygen evolution reaction. The performed electrochemical studies indicated that the low content (x 1) of Mn in Mn x Co 3-x O 4 enhances its catalytic activity towards OER in an alkaline medium by increasing the number of active sites and lowering the charge transfer resistance. The highest initial performance has been reported for the spinel with the nominal MnCo 2 O 4 designation, which has been revealed to consist of two cubic phases: Co 3 O 4 (~30 vol%) and Mn 1.4 Co 1.6 O 4 (~70 vol%). This particular material shows very high activity, not explainable by simple additive properties of the two phases. The effects of BET and ECSA surface areas on the OER activity were investigated. Except for MnCo 2 O 4 , all prepared materials revealed relatively good stability with degradation rates similar to IrO 2. Future work will focus on further studies of the effects of the synthesis and microstructure of the spinels on their electrocatalytic activity. Studies of dual-phase materials seem a promising route for further lowering and understanding the complex OER phenomena.