Due to the rapid development of consumer electronics and electric vehicles, an urgent and ever-increasing demand for advanced energy generation and storage systems with high energy and rate densities has emerged. Theoretically, rechargeable metal-air battery technologies offer extremely high energy capacities; zinc-air batteries (ZABs) in particular, are affordable, safe, and eco-friendly [1-3]. However, low efficiency and poor cycling performance hinder the practical application of these technologies. A key, but equally challenging, approach is to design an appropriate porous air electrode structure, with typical requirements of high surface area, high electrical conductivity, and superior bifunctional electrocatalytic activity, for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) [4, 5]. Conventionally, most air electrodes are fabricated by bonding noble metal or non-precious powder catalysts onto a current collector using a polymer binder, e.g., Nafion, polyfluortetraethylene polymer, or PVDF; this often involves complex procedures and has a negative impact on overall battery performance [6-8]. The resulting electrode usually lacks direct electric contact with the catalyst and substrate, which hinders ion and electron transport in the systems, causing a high-charge overpotential and low round-trip efficiency owing to electrode polarization [9]. Moreover, the addition of a binder to electrodes leads to undesirable inactive volume and increases side reactions. To address these issues, binder-free bifunctional electrodes that can be utilized directly for rechargeable metal-air batteries are desired.

Diversified metal-organic frameworks (MOFs) and MOF-derived materials have recently attracted considerable interest as promising alternatives to noble metal electrocatalysts, owing to their large surface areas, large pore volumes, uniform pore distributions, and tunable chemical structures [10-12]. Moreover, MOF structures can be either free-standing or supported on various substrates, and can be well-controlled in one-, two-, or three-dimensions simply by varying the constituent geometries of metal ions/centers and organic linkers [13, 14]. To develop MOF-based ORR/OER bifunctional electrodes, various efforts have been made to transfer intrinsic MOFs containing abundant carbon and nitrogen species into transition metal compounds or N-doped nanocarbon hybrid catalysts with improved catalytic performance [15]. Since most MOF-derived nanocarbon/metal oxide catalysts have exhibited unsatisfactory electrochemical activity [16], precise control of pyrolytic conditions to optimize the morphology, structure, and composition is required. However, MOF-derived electrocatalysts formed via high-temperature pyrolysis recently have demonstrated reduced OER performance compared to non-nitrogenous intrinsic MOFs [17]. Therefore, it remains a great challenge to design and fabricate highly active transition metal MOF-based and binder-free bifunctional oxygen electrodes with fast kinetics for four-electron (4 e^-) processes [18].

Herein, we develop a nickel-foam-supported NiCo-MOF-based oxygen electrode using a growth-pyrolysis-regrowth strategy with well-defined two-dimensional (2D) MOF/MOF derivative coupled arrays and high oxygen electrode catalytic bifunctionality performance. In contrast to traditional and current designs, we explore electrode activity by integrating various functional components and increasing the exposure of active electrochemical area. The 2D MOF/MOF derivative coupled arrays (R-NCM) exhibit an on-set potential of ~0.90 V for the ORR and 100 mA cm^-2 at an overpotential of 319 mV for the OER; these values are clearly enhanced relative to those of controlled samples of first-growth NiCo-MOFs (NCM) and their pyrolytic derivative (A-NCM). The overall superior ORR/OER performance results from a significantly increased electrochemical active surface area and enhanced reaction kinetics, as confirmed by double-layer capacitances (C_dl) and charge-transfer resistance measured by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS), respectively.

Nickel acetate (Ni(Ac)₂·4H₂O, 98%, Sinopharm Chemical Reagent), cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O, 99%, Sigma-Aldrich), 2, 6-naphthalenedicarboxylic acid dipotassium (C₁₀H₆(CO₂K), 95%, Sigma-Aldrich), hydrochloric acid (HCl, 37%, Sinopharm Chemical Reagent), and ethanol (C₂H₆O, 99.7%, Aladdin) were used as received without further purification.

A 2D NCM array on a Ni foam (NF) substrate was grown directly via a facile hydrothermal reaction. Typically, a piece of Ni foam (4 cm ×6 cm × 1.6 mm) was first cleaned to remove surface oils and the oxide layer, using repeated ultrasonic treatment in HCl solution (6 M), followed by acetone and deionized (DI) water. Next, the freshly-cleaned nickel foam was quickly immersed into a prepared reaction system containing 60 mL of DI-water (18 MΩ cm^-1), 160 mg of Ni(Ac)₂·H₂O, 40 mg of Co(NO₃)₂·6H₂O, and 200 mg of organic ligand, 2, 6-naphthalenedicarboxylate tetrahydrate. After further reaction at 60 ℃ for 20 h, the mixture was cooled to room temperature, and the final NCM samples were rinsed with excess DI-water with light ultrasonic treatment, then dried at room temperature.

The A-NCM samples were directly derived from as-prepared NMC using a typical pyrolysis method, where NMC was annealed in N₂ atmosphere at 650 ℃ for 2 h with a ramp rate of 2 ℃ min^-1. Next, R-NCM samples were prepared similar to A-NCM samples, with some modifications; the reaction system contained 60 mL of DI-water, 96 mg of Ni(Ac)₂·H₂O, 24 mg of Co(NO₃)₂·6H₂O, and 120 mg of organic ligand, 2, 6-naphthalenedicarboxylate tetrahydrate.

Sample morphologies were observed with transmission electron microscopy (TEM, JEM-2100F, Japan) and scanning transmission electron microscopy (FE-SEM, S-4800, Japan) equipped with energy dispersion spectra (EDS) and elemental mapping. X-ray diffraction (XRD) patterns were acquired using a D/MAX 2500 diffractometer (Rigaku, Japan) fitted with Cu K_α radiation. X-ray photoelectron spectroscopy (XPS) was conducted on a Thermo Scientific ESCALAB 250Xi with Al K_α (1487.6 eV) as the excitation source and C 1s peak (284.8 eV) as the calibration [19].

To evaluate bifunctional electrocatalytic activity, OER and ORR polarization curves were measured at room temperature in 1 M KOH aqueous solution with a typical three-electrode system on an electrochemical workstation (CHI760E, Shanghai, China). All polarization curves were iR-corrected and performed at a scan rate of 5 mV s^-1, with a carbon rod as the counter electrode, a saturated AgCl as the reference electrode, and the active material on nickel foam as the working electrode. All potentials were calculated with respect to the reversible hydrogen electrode (RHE), based on the formula: E_RHE = E_AgCl + 0.059 × pH + 0.197 V. Electrochemical stability was further evaluated by the accelerated durability test (ADT) using cyclic voltammetry (CV) in the ranges 1.1 to 1.7 V for the OER and 0.45 to 1.0 V for the ORR, with a scan rate of 100 mV s^-1. Cyclic voltammetry curves were measured under different scan rates to determine the electrochemical surface area (ECSA). Electrochemical impedance spectra (EIS) were measured at 1.6 V for the OER and 0.8 V vs. RHE for the ORR [20, 21].

The Zn-air battery tests were performed with a homemade cell configuration using a LANBTS BT-2016C system to carry out the cycling test (10 min for each discharge and charge period). The Zn-air battery was assembled using the as-prepared R-NCM, A-NCM, or NCM electrodes as cathodes, and a fresh polished Zn plate (~1 mm thick) as the anode. The alkaline aqueous electrolyte was composed of 0.2 M ZnCl₂ and 6 M KOH solutions.

As illustrated in Fig. 1, the synthesis of 2D R-NCM coupled arrays on Ni foam using a growth-pyrolysis-regrowth strategy with Ni²⁺, Co²⁺, and 2, 6-naphthalenedicarboxylate tetrahydrate as organic ligand is proposed. Briefly, we fabricate 2D NiCo-based MOFs according to a typical hydrothermal synthesis at 60 ℃, then yield 2D MOF derivatives annealing in N₂ atmosphere. During the process, A-NCM turns dark black and R-NCM becomes a mixed black and light green film (Fig. S1), demonstrating the corresponding changes in each step. The morphology of relevant samples was further characterized by SEM and TEM. Compared to the cellular-structured NCM samples formed by vertically grown ultrathin nanosheets (Figs. 2(a) and S2), thicker flakes of stacked nanosheets appeared in A-NCM after annealing (Figs. 2(b) and S3), while denser MOF nanosheets filled the interspaces by regrowth in R-NCM (Fig. 2(c)). The NCM nanosheets have a lateral size of several hundred nanometers (Fig. 2(d)) with a smooth, well-defined morphology. Uniformly distributed nanoparticles (dominated by metallic Ni and Co) were observed on the annealed MOF-derived nanoflakes (Figs. 2(e) and S4). TEM, HR-TEM, and selected-area electron diffraction (SAED) images were collected where NCM and A-NCM appeared simultaneously in order to further determine the configurations of R-NCM (Figs. 2(f)-(m), S5 and S6). In accordance with our design, we confirmed the coexistence of 2D MOFs and MOF derivative coupled nanosheets with metallic Ni, Co, and C phases, as shown in Fig. 2(f) and 2(g). Moreover, the crystalline metal nanoparticles are encased by few-layer graphene, due to Ni nanocrystal catalysis during the annealing process [22-24], with an average diameter of about 10 nm and a homogeneous distribution of various elements.

The configuration of R-NCM with pristine MOFs and NiCo@C 2D arrays was also confirmed by XRD patterns and XPS spectra, as shown in Fig. 3. After annealing, the original peaks belonging to MOF nanosheets disappear and several new peaks are observed, which are ascribed to the (111), (200), and (220) planes of metallic Ni and C (002), respectively. Consistent with TEM results, organic ligands in MOFs transform into graphene and metal nodes form the nanoparticles, while maintaining their 2D features in both the A-NCM and R-NCM samples. In addition, XPS results were further analyzed to investigate changes in the chemical state of the MOF structure. The XPS survey spectra, shown in Fig. S7, indicate the Ni, Co, C, and O are present, in good agreement with elemental mapping results. As shown in Fig. 3(b) and 3(c), additional characteristic peaks of metallic Ni 2p_1/2 (852.9 eV), Ni 2p_3/2 (870.4 eV), and Co 2p_3/2 (778.7 eV) are observed in R-NCM and A-NCM, which illustrates the reduction of Ni and Co nodes to metallic atoms during pyrolysis [25-30]. Significantly, two evident shifts: Co²⁺ 2p_3/2 peak to lower binding energy and Ni²⁺ 2p_3/2 peak to higher binding energy, are observed for R-NCM compared to NCM. This implies a strong interaction between the Co and Ni species and indicates that the local electronic structure of the Ni centers is modified via partial electron transfer from Ni²⁺ to Co²⁺ through the oxygen of the ligand [25-28]. The XPS spectrum of C 1s for NCM (Fig. 3(d)) shows two peaks at 284.8 and 288.4 eV, corresponding to the sp² hybridized C-C and O=C-O groups, respectively [31]. However, in the A-NCM and R-NCM samples, the peak related to O=C-O groups disappears and a peak assigned to C-O (286.2 eV) emerges [32], which confirms partial loss of oxygen due to carbonization. Disappearance of O=C-O is also observed in O 1s spectra (Fig. 3(e)) after annealing, where typical XPS peak corresponding to Ni-O bond at 530.1 eV is also confirmed for A-NCM and R-NCM [31, 33].

To test oxygen electrode performance, R-NCM and other controlled samples were immersed directly in 1 M KOH electrolyte as working electrodes. For OER, NCM requires an overpotential (η_OER) of only 313 mV to reach a current density of 100 mA cm^-2 (Fig. 4(a)), while A-NCM needs a larger overpotential of 366 mV. This illustrates that the OER performance of NiCo-based MOFs is degraded by annealing, but can be recovered by R-NCM (overpotential of 319 mV), due to the regrown MOF nanosheet array. The Tafel slopes for NCM and R-NCM are 86 mV dec^-1 and 78.2 mV dec^-1 (Fig. 4(c)), much smaller than those of A-NCM (89.1 mV dec^-1) and NF (194.8 mV dec^-1), suggesting favorable OER kinetics for R-NCM. During the ORR, A-NCM and R-NCM electrodes exhibit enhanced activity, with an onset potential of 0.90 V and current densities of 91.1 and 82.1 mA cm^-2 at 0.45 V_, respectively (Fig. 4(b)). This demonstrates that the electrochemical catalytic performance for the ORR improves dramatically after annealing.

In order to better understand the electrochemical reaction kinetics, the corresponding Tafel plots were further analyzed (Fig. 4(d)). A-NCM and R-NCM have much lower Tafel slopes (63.2 and 77.4mV dec^-1, respectively) compared to NCM (296.2 mV dec^-1) and NF (234.6 mV dec^-1), demonstrating their intrinsically favorable activity and kinetics for the ORR after annealing. Therefore, in general, the R-NCM electrode exhibits superior ORR/OER bifunctional electrocatalytic activity than A-NCM, NCM, and other reported results (Table S1). According to previous reports [17, 33], the proposed active sites can be attributed to oxidized Ni metal nodes in MOFs (such as NiO₆/NiOOH species); for OER, additional structural vacancies are introduced by trace Co moieties, while the transition metal particle-activated 2D carbon layers impart synergistic effects in MOF derivatives for ORR.

The stability of the R-NCM electrode was confirmed (presented in Fig. 4(e) and 4(f)); over 5000 CV cycles, negligible degeneration was observed, due to the stable morphology and structure (Fig. S8) [34]. To further confirm the stability of the materials, we also tested the catalytic stabilities of R-NCM electrodes for the ORR and OER using a typical chronoamperometry (CA) method. As shown in Fig. S9, the R-NCM sample exhibits minor degradation after 12 h CA testing, with reductions of ~9% and ~12% in the OER and ORR activity, respectively.

Inspired by the outstanding bifunctional activity and stability of R-NCM, we investigated its practical application by integrating it as an air electrode in a rechargeable zinc-air battery (Figs. S10 and S11). The Zn-air batteries with a R-NCM air cathode exhibited the best discharge ability and highest peak power density of 45.2 mW cm^-2. Zinc-air battery performance was evaluated by pulse discharge-charge tests using a current density of 10 mA cm^-2 with a fixed capacity of 17.7 mAh (20 min per cycle). The charging-discharging potential gap was maintained at approximately 0.9 V throughout the experiment, after full activation for 50 h, which indicates excellent stability and capacitance for the R-NCM-based battery. In addition, a green LED powered by two battery units integrated in series is displayed as well; this suggests great potential for energy storage applications.

For additional insight, the electrochemical surface area (ECSA) was measured and electrochemical impedance spectroscopy (EIS) were collected, as illustrated in Figs. S12 and S13. Double-layer capacitances (C_dl) were evaluated by cyclic voltammetry (CV) to measure ECSA, with a result of 72.5 mF cm^-2 for R-NCM, compared to 6.5 mF cm^-2 for A-NCM and 1.5 mF cm^-2 for NCM, indicating that R-NCM has more accessible active surface area. The increased ESCA of R-NCM may be attributed to the A-NCM/NCM interfacial structure, which provides more surface-active sites for the ORR and OER. Furthermore, R-NCM and other sample-based catalysts show two semicircles in the high- and low-frequency regions of Nyquist plots, which can be fitted with common equivalent circuits (inset of Fig. S13) with charge-transfer resistance (R_ct) and solution resistance (R_s), respectively [35]. Notably, R-NCM not only possesses a smaller R_ct of 4.31 Ω for OER, but also R_ct of 148 Ω for ORR, in contrast to other samples, which contributes to its optimized activity. Therefore, the prominent activity of R-NCM can be derived from synergistic effects of the 2D metal organic framework and its derivative arrays, favorable charge-transfer kinetics, and increased intrinsic active sites [36-38].

In summary, we fabricated an advanced bifunctional oxygen electrode with vertically grown 2D NiCo-based MOFs and their derivatives and evaluated each by their OER and ORR electrocatalytic performance. Due to the robust freestanding and hierarchical structure achieved by the growth-pyrolysis-regrowth process, the resultant R-NCM-based electrode exhibits significantly enhanced bifunctionality, kinetics, and stability as an air-cathode in rechargeable Zn-air batteries. Considering feasibility and simplicity, our design strategy can be extended to the fabrication of other efficient self-supporting electrocatalysts with advanced nanostructures.