Water oxidation, known as the anodic oxygen evolution reaction (OER), is the essential half-reaction of water electrolysis for producing high-value hydrogen energy to replace traditional fossil fuels [1-4]. However, the OER is a sluggish multistep reaction due to the complex four-electron transfer reaction (2H₂O → 4H⁺ + O₂ + 4e^-), which usually requires a large overpotential of more than 1.23 V (standard equilibrium potential) to drive the effective water electrolysis [5-8]. Therefore, to realize efficient H₂ generation, one should search for effective electrode materials to reduce the overpotential for the OER [9, 10]. At present, platinum group metal oxides, such as ruthenium and iridium oxides, remain the most efficient and stable catalysts for the OER [11-14]. However, considering the scarcity and high cost of these noble metal materials, it is crucial to develop naturally abundant, non-noble metal electrocatalysts to catalyze the OER on a large scale to meet the global energy demand [15, 16].

First-row transition-metal (Mn, Fe, Co, Ni, etc.) based compounds, such as oxides [17], (oxy)hydroxides [18, 19], sulfides [20, 21], and phosphides [22, 23], are extensively investigated as efficient OER catalysts to replace noble metal catalysts. Among these catalysts, non-noble metal oxides have accumulated more attention because recent studies have demonstrated that the real active sites for the OER may be the oxides/(oxy)hydroxides, which are formed on the surface of catalysts during the electrochemical oxidation process [24, 25]. Concerning this problem, designing and tuning unique metal oxides with remarkable activities and stability for an enhanced OER have gained much focus. For example, Boettcher et al. [26] reported that cobalt oxides exhibited desirable activity in alkaline media. To maximize the activity of cobalt oxides, numerous strategies have been adopted to improve the intrinsic OER activities, including doping other transition-metals, designing unique nanostructures, and improving the conductivity of catalysts using ideal conductive materials. Hu's group reported that the intentional introduction of Ni or Fe into cobalt oxides can significantly increase their intrinsic activities, because Ni or Fe dopants can increase the activity of cobalt oxides related to the anodic shift in the nominally Co²⁺/³⁺ redox wave, indicating strong synergistic metal-metal electronic interactions between Ni and Co [27]. However, it is usually challenging to synthesize well-defined Ni-doped cobalt oxides with unique structures to expose more active sites and improve the conductivity of catalysts [28]. Therefore, employing an ideal carbon substrate with high conductivity, large surface area, and strong tolerance to harsh acidic or alkaline conditions is highly desired to derive efficient Co-based hybrid catalysts. Additionally, the N heteroatoms doped into carbon materials can influence the electronic structures due to the different electronegativities of N and C, thus activating the inert carbon surface. Organic polymers, a class of N-doped carbon materials, have accumulated more interests due to their tunable shape, low density, high conductivity, and porous structures after undergoing carbonization at a high temperature [29-32]. The application of organic polymers relies on the design of homogeneously distributed nanostructures with porous structures and high conductivity. Furthermore, controlling the size of carbon nanostructures below 100 nm is essential due to the many valuable nanoscale effects [29]. Therefore, designing unique metal/nonmetal-doped oxides strongly coupled with well-defined carbon nanostructures is urgently needed, although it is still challenging.

Based on the above analysis, we have designed and synthesized well-defined Ni/Co₃O₄@NC hybrids via a facile impregnation-calcination process (Fig. 1). The first step involving polymer impregnation with Ni and Co ions can ensure the homogeneous distribution of metals, thus guaranteeing the following in situ calcination reaction, which would produce well-dispersed Ni and Co₃O₄ nanoparticles. Moreover, the calcination process confined within the carbon matrix can effectively prevent the aggregation and excessive growth of active species. Benefiting from the synergistic effect between Ni, Co₃O₄, and NC species, the obtained Ni/Co₃O₄@NC hybrids exhibit enhanced OER activity in alkaline media, with a smaller overpotential of 350 mV to afford 10 mA cm^-2, lower Tafel slope of 52.27 mV dec^-1, smaller charge-transfer resistance, and higher double-layer capacitance of 25.53 mF cm^-2 compared to those of unary Co₃O₄@NC or Ni@NC metal hybrids. The present strategy may be also applicable in designing other heteroatom-doped oxides coupled with the ideal carbon materials to realize high OER electrocatalytic activities.

The well-dispersed polymer nanospheres were synthesized via the polymerization of aniline and pyrrole using Triton X-100 as a surfactant, according to the procedure employed in our previous study [33]. To synthesize Ni²⁺, Co²⁺@polymer nanospheres, the polymer nanospheres were dispersed in 20 mL of a solution containing the same amount of Ni(NO₃)₂ 6H₂O and Co(NO₃)₂∙6H₂O (n_Ni:n_Co = 1:1) with strong ultrasonication to form homogeneous ink. Subsequently, the Ni²⁺, Co²⁺@polymer nanospheres were obtained by evaporating the residual water at 80 ℃.

A 0.2 g portion of the as-prepared Ni²⁺, Co²⁺@polymer was calcined in a tube furnace at 800 ℃ for 4 h, followed by isothermal calcination at 500 ℃ for 2 h at a heating rate of 5 ℃ min^-1 in an inert atmosphere. After cooling to 25 ℃, the product was filtered with water and EtOH several times and subsequently dried in a vacuum. The obtained particles were denoted as NiCo@NC nanospheres.

The NiCo@NC nanospheres were heated in a muffle furnace to 400 ℃ at a rate of 2 ℃ min^-1, after which they were kept for 2 h to yield Ni/Co₃O₄@NC nanospheres.

The NCs were synthesized using the same procedure for preparing the Ni/Co₃O₄@NC nanospheres without introducing Ni(NO₃)₂·6H₂O and Co(NO₃)₂·6H₂O. The Ni@NC and Co₃O₄@NC nanospheres were also synthesized using the synthesis procedure for the Ni/Co₃O₄@NC nanospheres but without adding Co(NO₃)₂·6H₂O and Ni(NO₃)₂·6H₂O, respectively.

The surface morphology and size of the catalysts were observed using a Hitachi S-4800 scanning electron microscope (SEM) equipped with an energy-dispersive X-ray detector (EDX) and an FEI Tecnai G² transmission electron microscope (TEM) with an acceleration voltage of 200 kV. Powder X-ray diffraction (XRD) (Cu K_α, λ = 0.154178 nm) analysis was performed on an X'Pert PRO MPD diffractometer at a scanning rate of 3 ℃ min^-1. The compositions and valence states were determined by X-ray photoelectron spectroscopy (XPS) (ThermoFisher Scientific II) with an Al K_α (1486.6 eV) X-ray source.

The electrochemical measurements of all catalysts were conducted using a Gamry Reference 600 workstation in a standard three-electrode system. Hg/HgO and a platinum plate were used as the reference and counter electrodes, respectively. For the preparation of the working electrode, 5 mg portions of the catalysts were dispersed in 1 mL of a water/ethanol/Nafion aqueous solution by strong ultrasonication to form homogeneous ink. Subsequently, 5 μL of the ink was dropped onto a glassy carbon electrode (GCE) (0.1256 cm²) and then dried in a vacuum. The OER polarization curves of all samples were obtained in an O₂-saturated 1.0 M KOH solution at a scan rate of 2 mV s^-1. Electrochemical impedance spectroscopy (EIS) measurements were carried out at 1.4 V (vs. RHE) with a frequency in the range of 100000-0.1 Hz with an AC voltage of 5 mV. The estimation of the electrochemical surface area of the catalysts was carried out by cyclic voltammetry (CV) in the non-Faradaic region at various scan rates (40, 60, 80, 100, 120, and 140 mV s^-1). Long-term stability tests were conducted by continuous CV at a scan rate of 100 mV s^-1 from 1.3 V to 1.6 V (vs. RHE) for 1000 cycles.

SEM is firstly employed to examine the morphologies and compositions of the as-synthesized products, as illustrated in Fig. 2. As shown in Figs. 2(a) and 2(b), the polymer nanospheres are composed of well-dispersed nanospheres with a diameter of ~100 nm, and the surface is smooth. After impregnation with Ni and Co ions, followed by calcination at high temperatures, the obtained NiCo@NC nanospheres still preserve their spherical morphology but with a shrunken size of ~90 nm (Figs. 2(c) and 2(d)). Despite the harsh thermal condition, the final as-synthesized Ni/Co₃O₄@NC nanospheres maintain their uniform nanospherical structures, while their surfaces become rougher, indicating the stable properties of the polymer framework (Figs. 2(e) and 2(f)). To further confirm the existence and homogeneous distribution of elements, the SEM images and elemental mappings of Ni/Co₃O₄@NC are presented. Fig. 2(g) displays the homogeneous distribution of the Ni, Co, C, N, and O elements of the Ni/Co₃O₄@NC nanospheres, which may be because Ni²⁺ and Co²⁺, dispersed in the polymer in an atomic scale, promote the good dispersion of the elements in the Ni/Co₃O₄@NC nanospheres. The EDX spectrum of Ni/Co₃O₄@NC in Fig. S1 also demonstrates the existence and contents of Ni, Co, C, N, and O. For comparison, the SEM images of NC, Ni@NC, and Co₃O₄@NC reveal that these samples also maintain the spherical morphology, implying that both Ni and Co species have no notable effect on the spherical polymer framework (Figs. S2-S4). Figs. S5 and S6 exhibit the homogeneous elemental distribution and contents of the Ni@NC and Co₃O₄@NC catalysts, respectively.

The nanostructures of the as-prepared catalysts are further examined by TEM. Figs. 3(a) and 3(b) show that the polymer nanospheres exhibit the typical spherical morphology with a diameter of ~100 nm, which is consistent with the SEM analysis. After the calcination of the Ni²⁺, Co²⁺@polymer nanospheres, many nanoparticles are homogeneously dispersed throughout the nanospheres (Fig. 3(c)). Under high magnification (Fig. 3(d) and Fig. S7), numerous Ni and Co nanoparticles are embedded in several carbon layers. Benefiting from the homogeneous distribution of Ni and Co nanoparticles in NiCo@NC, the as-synthesized Ni/Co₃O₄@NC nanospheres deriving from NiCo@NC also maintain a spherical morphology with many nanoparticles well-dispersed throughout the carbon matrix, and the surface is evidently porous (Fig. 3(e) and Fig. 3(f)). It is worth noting that the carbon matrix can significantly confine the excessive growth of these nanoparticles and prevent their stacking and coalescence. Moreover, Fig. 3(g) exhibits the noticeable lattice fringes with spacings of 0.204 and 0.243 nm, which correspond to the (111) crystal plane of Ni and the (311) crystal plane of Co₃O₄, respectively. The XRD patterns of all the samples are also presented to further verify the crystallographic structures of the catalysts. As shown in Fig. S8, NC exhibits a broad peak at ~24°, corresponding to the (002) facets of graphitic carbon. For NiCo@NC, the typical peaks located at 44.4°, 51.7°, and 76.2° can be ascribed to the mixture of Ni (PDF No. 01-089-7128) and Co (PDF No. 01-089-4307). After annealing of NiCo@NC, the typical peaks at 44.5°, 51.7°, and 76.2° of the as-prepared Ni/Co₃O₄@NC correspond to the (111), (200), and (220) planes of Ni (PDF No. 01-089-7128), respectively. Moreover, the peaks at 37.1°, 59.2°, and 65.2° can be ascribed to the (311), (511), and (440) facets of Co₃O₄ (PDF No. 00-001-11152), respectively, which is consistent with the TEM results (Fig. 3(h)). This result indicates that the Co species are more easily oxidized during the annealing process. In addition, the XRD patterns of the reference samples show that the Ni@NC nanospheres are composed of metallic Ni and that the Co₃O₄@NC nanospheres consist of Co₃O₄ (Fig. S9).

XPS measurements are further carried out to investigate the composition and valence states of the Ni/Co₃O₄@NC catalyst. As shown in Fig. 4(a), the XPS survey demonstrates the existence of Ni, Co, C, N, and O in the Ni/Co₃O₄@NC catalyst, which is consistent with the SEM mapping analysis. For Ni 2p in Fig. 4(b), the peaks located at 855.7 and 873.2 eV are assigned to Ni 2p_3/2 and Ni 2p_1/2, respectively, which are ascribed to Ni⁰ [34, 35]. Moreover, the peaks located at 856.8 eV in the Ni 2p_3/2 region and 874.5 eV in the Ni 2p_1/2 region correspond to Ni³⁺, whose formation is due to the surface oxidation; the Ni oxides cannot be observed in the XRD pattern due to their amorphous state [36]. The existence of the low-valence Ni⁰ in further oxidation may play a crucial role in enhancing the OER performance, while the high-valence Ni³⁺ may be inactive for the OER due to the difficulties associated with its oxidation. The other two peaks located at 861.7 and 880.1 eV are satellite peaks [37, 38]. In addition, the high-resolution XPS spectra of the Co 2p region can be fitted with two different doublets (Fig. 4(c)): one doublet located at 780.6 and 795.4 eV, ascribed to Co₃O₄; and the other doublet located at 782.3 and 796.9 eV, assigned to Co³⁺ [39]. Two other peaks located at 786.4 and 803.4 eV are assigned to the satellite peaks of Co 2p [40]. The high-resolution XPS spectra of C 1s can be deconvoluted into four peaks, located at 284.6 eV for C=C, 285.0 eV for C-O, 286.0 eV for C-C, and 228.4 eV for C-N, respectively (Fig. 4(d)) [41]. This result demonstrates that the heteroatom N-doped into the carbon matrix successfully. This can also be verified in the XPS spectra of the N 1s region, which can be deconvoluted into three peaks of pyridinic-N (398.6 eV), graphitic-N (399.4 eV), and pyrrolic-N (400.4 eV) (Fig. 4(e)) [42]. Notably, introducing electron-rich N dopants is beneficial for optimizing the electron configuration, and thus accomplishing intrinsic enhanced OER performance [43]. Overall, the exact composition of the Ni/Co₃O₄@NC catalyst has been confirmed as a hybrid of Ni and Co₃O₄, as verified by HRTEM, XRD, and XPS analyses.

To evaluate the electrocatalytic OER performance of Ni/Co₃O₄@NC, we utilize a three-electrode configuration with the scan rate of 2 mV s^-1 in 1.0 M KOH. The OER performance of bare GCE, NC, NiCo@NC, Co₃O₄@NC, and Ni@NC are also presented as a comparison (Fig. 5(a)). The linear sweep voltammetry (LSV) curves of all the samples show that Ni/Co₃O₄@NC exhibits a much higher current density (j) at the same overpotential (η) compared with other reference samples (Fig. 5(a)). To reach a current density of 10 mA cm^-2, Ni/Co₃O₄@NC only requires a small overpotential of only 350 mV compared to those of NiCo/NC (510 mV), Co₃O₄@NC (400 mV), and Ni@NC (480 mV) (Fig. 5(b)), suggesting that the synergistic effect between Ni, Co₃O₄, and NC contributes significantly to the enhanced OER performance. Bare GCE exhibits nearly no current response even with a large overpotential, excluding the effect of bare GCE on the excellent OER performance of Ni/Co₃O₄@NC. As shown in Table S1, the OER performance of the Ni/Co₃O₄@NC hybrids is superior to those of many other Co-based materials, suggesting the advantages of employing Ni/Co₃O₄@NC hybrids as efficient OER catalysts. The Tafel slope simulated from the LSV curve via the Tafel equation (η = a + blogj, where a represents the intercept and b represents the Tafel slope) is an inherent property to elucidate the possible reaction mechanism [1]. As observed in Figs. 5(c) and 5(d), the Tafel slope of Ni/Co₃O₄@NC is 52.27 mV dec^-1, which is lower than those of NC (168.05 mV dec^-1), NiCo/NC (143.01 mV dec^-1), Co₃O₄@NC (62.55 mV dec^-1), and Ni@NC (114.89 mV dec^-1), demonstrating the favorable OER kinetics of the Ni/Co₃O₄@NC electrode. Therefore, compared with NC, Co₃O₄@NC, and Ni@NC, it can be speculated that Ni, Co₃O₄, and NC can synergistically boost the OER performance.

To evaluate the electrochemical surface area (ECSA) of the interface between the catalysts and electrolyte, the double-layer capacitances (C_dl) of the catalysts were obtained by CV at different scan rates (40, 60, 80, 100, 120, and 140 mV s^-1) due to the proportional relationship between the ECSA and C_dl (Fig. 6(a)) [3]. The C_dl value of Ni/Co₃O₄@NC was calculated to be 25.53 mF cm^-2, which is much higher than those of NC (1.81 mF cm^-2), NiCo/NC (1.78 mF cm^-2), Co₃O₄@NC (15.28 mF cm^-2), and Ni@NC (2.70 mF cm^-2) (Fig. 6(b)). Accordingly, the calculated ECSA values of all the samples are listed in Table S2. Such a high ECSA value of Ni/Co₃O₄@NC implies the existence of additional exposed active sites, which can be ascribed to the strong coupling of Ni/Co₃O₄ with the NC matrix, thus boosting the OER performance. To further reflect the intrinsic activities of all the samples, we calculate their electrocatalytic activities, where the current density is normalized to the ECSA. As is shown in Table S2 and Fig. S10, Ni/Co₃O₄@NC exhibits a higher current density (j) at the same overpotential (η) compared with other samples, which is consistent with the current density normalized to the geometric area of the electrode, demonstrating the superior performance of the Ni/Co₃O₄@NC hybrids.

To elaborate on the intrinsic activities of the catalysts toward the OER, the charge-transfer resistances (R_ct) were obtained by fitting the equivalent circuit according to the EIS plots (Fig. 6(c)). Ni/Co₃O₄@NC exhibits a much smaller semicircle compared with other reference catalysts, which implies a considerably small R_ct value, indicating an effective electron transfer process between the catalysts and electrolyte, effective Faradic process, and favorable OER kinetics. The heterostructured Ni/Co₃O₄@NC exhibits a highly decreased R_ct value compared with Ni@NC or Co₃O₄@NC, which may be related to the better contact among Ni, Co₃O₄, and NC. The metallic Ni combined with Co₃O₄ induces strong synergistic metal-metal electronic interactions between Ni and Co₃O₄, thus accelerating the charge-transfer rate. The charge-transfer rate of the heterostructured Ni/Co₃O₄@NC can be further decreased when coupling Ni/Co₃O₄ with the highly conductive NC materials. Durability is also an important criterion in evaluating the performance toward the OER. Therefore, the stability of Ni/Co₃O₄@NC is tested through continuous CV for 1000 cycles in 1.0 M KOH. As shown in Fig. 6(d), the LSV curve of Ni/Co₃O₄@NC remains nearly unchanged after continuous 1000 cycles. This may be due to the intimate combination among Ni, Co₃O₄, and NC that prevents the agglomeration and corrosion of active species.

In summary, Ni/Co₃O₄@NC hybrids were synthesized via a facile impregnation-calcination process. The impregnation of Ni and Co ions ensured the homogeneous distribution of metals and guaranteed the production of well-dispersed Ni and Co₃O₄ nanoparticles by the subsequent in situ calcination reaction. The subsequent two-step calcination process constructed synergistic Ni and Co₃O₄ species throughout the carbon nanospheres, which prevented the aggregation and corrosion of the obtained N-doped graphitic carbon and regulated the electronic configuration of the active species. Benefiting from the intimate contact of Ni, Co₃O₄, and NC and the efficient charge transfer among these active OER species, the obtained Ni/Co₃O₄@NC hybrids demonstrated enhanced OER activity with a smaller overpotential of 350 mV to drive a current density of 10 mA cm^-2, lower Tafel slope of 52.27 mV dec^-1, smaller charge-transfer resistance, and higher double-layer capacitance of 25.53 mF cm^-2 compared to those of the unary Co₃O₄@NC and Ni@NC metal hybrids. This work proposes a feasible route to design other heteroatom-doped oxides supported on ideal carbon materials to realize high electrocatalytic activities for the OER.