1 Introduction
The rapid development of modern society poses a great challenge for power sources. But the massive consumption of fossil fuels aggravates energy crisis and results in severe environmental pollution. The concerns over these two issues trigger the intensive exploration of sustainable energy, such as solar power and wind energy [
1−
3]. However, the intermittent feature of these renewable energy needs energy storage systems to regulate the power capacity and connect to the grid. As a typical energy storage system, lithium-ion batteries (LIBs) have been extensively commercialized in the portable electronics and electric vehicles [
4,
5]. Moreover, low-cost sodium ion batteries (SIBs) have also attracted widespread attention of researchers due to the great prospect in the energy-storage market [
6−
8]. Despite the great progress on these techniques, the energy density and safety remain challengeable [
9], due to the limitation of electrode materials and utilization of organic electrolyte.
As an alternative, zinc−air batteries (ZABs) with high theoretical energy density holds the prominent advantages of good safety, low cost and environmental friendliness, due to the aqueous electrolyte and abundance of zinc metal [
10−
12]. In addition, zinc metal is less active than lithium and sodium, which leads to facile assembly of ZABs and manufacture procedure in ambient conditions [
13]. However, during the discharge and charge processes, sluggish kinetics of oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) at the cathode poses a critical challenge for the application [
14−
16]. The ORR and OER need an efficient electrocatalyst to overcome the energy barriers, i.e., lowering the overpotential. Currently commercialized electrocatalysts are based on precious metals, such as Pt/C and RuO
2, which highly raise the cost of ZABs and limit the widespread application [
17,
18]. Therefore, it is imperative to design non-precious materials with efficiently bifunctional activity toward OER and ORR, not only solving the issue of slow kinetics, but also lowering the whole cost of ZABs [
19,
20].
Among many non-precious materials, heteroatom-doped carbon materials demonstrate excellent activity toward ORR, while layered double hydroxides (LDHs) display good activity toward OER [
21−
23]. It was found that the doping of heterogeneous atoms into carbon matrix skeleton can effectively modulate the electronic structure and enhance the intrinsic properties of the catalyst. Especially, single-atom transition metal (TM) bonded with N-doped carbon materials, such as Fe−N−C, shows superior ORR activity to commercial Pt/C [
24,
25]. Meanwhile, the carbon framework also plays an important role in boosting the reaction kinetics, as the micropores and mesopores in the framework are beneficial for exposing more active sites and facilitating the diffusion of the reactants in the catalyst layer [
26−
28].
Metal−organic frameworks (MOFs), such as zeolitic imidazolate framework-8 (ZIF-8), are a type of ideal precursor for N-doped carbon matrix, due to containing abundant N species and tunable pore structure [
29−
32]. Moreover, organic ligands easily integrate with TM atoms to form chelates. After carbonization at high temperature, highly efficient electrocatalysts with single metal atoms are usually harvested [
33,
34]. On the other hand, two-dimensional (2D) LDHs, such as NiFe-LDHs not only possess large specific surface area, but also exhibit highly efficient OER activity due to the deprotonating effect of hydroxyl anions to form highly oxidized metal sites. But the poor electrical conductivity of LDHs limits their catalytic activities toward OER. It could be a feasible solution to integrate LDHs with heteroatom-doped carbon materials, and thus realizing good bifunctional electrocatalytic activity toward ORR and OER [
35−
38].
Herein, we report the successful coupling of Fe-doped hollow carbon dodecahedron (FeNC) derived from ZIF-8 with NiFe-layered double hydroxides (LDHs) to form nanocomposites (FeNC/LDHs). The synergetic effect arising from intimate contact between two components endows the nanocomposites with high-performance bifunctional electrocatalytic activity toward ORR and OER. The potential difference (ΔE = E1/2‒Ej = 10) between the half-wave potential of ORR (E1/2) and the potential at 10 mA·cm−2 in the OER (Ej = 10) is as low as 0.68 V, outperforming most reported non-precious metal bifunctional electrocatalysts. The liquid-state ZAB based on catalyst FeNC/LDH shows superior performance with a high-power density (85 mW·cm−2), good long-term cycling stability (over 160 h), and excellent specific capacity (810 mAh·g−1). Moreover, the assembled flexible solid-state ZAB demonstrates a large power density of 32.4 mW·cm−2 and excellent cycling stability at various bent angles (0‒180°), providing a feasible pathway for flexible electronic device applications.
2 Experimental section
2.1 Materials
Carbon paper (HCP020P) was purchased from Suzhou Sinero Technology Co., Ltd. Ferric nitrate nonahydrate (Fe(NO3)3·9H2O), RuO2, Pt/C (20 wt%) were purchased from Shanghai Macklin Biochemical Technology Co., Ltd. 2-methylimidazole (2-MIM), potassium hydroxide (KOH), sodium hydroxide (NaOH), zinc chloride (ZnCl2), ferric acetylacetonate (Fe(acac)3) and monohydrate sodium carbonate (Na2CO3·H2O) were purchased from Aladdin; Nickel nitrate hexahydrate (Ni(NO3)2·6H2O) and zinc nitrate hexahydrate (Zn(NO3)2·9H2O) were purchased from Shanghai LingFeng Chemical Reagent Co., Ltd. Nafion solution (5 wt%) were purchased from Sigma-Aldrich. All reagents were used without further purification.
2.2 Synthesis of FeNC
Briefly, 2-MIM (1.314 g) was immersed in a methanol solution (45 mL) containing Zn(NO3)2·9H2O (1.19 g), Fe(acac)3 (0.14 g). The above solution was stirred vigorously for 1 h and then transferred to a 100 mL hydrothermal reactor. After 4 h of reaction at 120 °C, the resulting reaction material was washed several times by centrifugation with methanol, and then dried under vacuum at 60 °C overnight. Finally, the dried materials were heated at 900 °C for 3 h in N2 atmosphere with a heating rate of 5 °C·min−1, to obtain FeNC.
2.3 Synthesis of FeNC/LDH-21 and FeNC/LDH-11
The FeNC (50 mg) was ultrasonically mixed with 10 mL of deionized water and poured into an aqueous solution (5 mL) containing NaOH (0.024 g) and Na2CO3 (0.022 g). The aqueous solution (5 mL) containing Ni(NO3)2·6H2O (0.0873 g) and Fe(NO3)3·9H2O (0.0464 g), respectively, was poured slowly into the above alkaline solution at the same time. After stirring at room temperature for 24 h, the samples were washed several times by centrifugation with ethanol and dried overnight at 60 °C under vacuum, to obtain FeNC/LDH-21. Preparation of FeNC/LDH-11 by the same method as halving the FeNC content.
2.4 Characterization
Scanning electron microscopy (SEM) was performed on a JEOL JSM-IT800 instrument. Transmission electronic microscopy (TEM), high-resolution TEM (HR-TEM), high-angle annular dark field (HAADF) scanning transmission electron microscopy (STEM), energy dispersive X-ray spectroscopy (EDS), and line scanning were carried out with a JEM-2100 F machine operating at 200 kV. X-ray diffraction (XRD) was performed with a Bruker D8 Advanced Twin using Cu Ka radiation. X-ray photoelectron spectroscopy (XPS) was performed on a ESCALAB Xi+ instrument. Raman tests were performed using a DXR microscope.
2.5 Density Functional Theory (DFT) calculations
All DFT calculations were performed using the Vienna ab initio simulation package (VASP) [
39]. In addition, the exchange-correlation functional was treated by the generalized gradient approximation with the Perdew‒Burke‒Ernzerh (GGA-PBE) functional [
40]. The supercells of NiFeOOH (LDH) nanosheet and FeNC were built, with the periodic surface of 11.9 Å × 11.9 Å and 12.1 Å × 12.1 Å, respectively. The heterostructure of FeNC/LDH was constructed with a lattice mismatch of 1.7%. A vacuum slab of 20 Å in thickness was built to separate the layer from its periodic images. The kinetic energy cut-off value was set to be 450 eV. A mesh of 3 × 3 × 1 was used for the k-point sampling obtained from the Gamma center. In geometry optimization, the atomic positions were fully optimized until the energy and forces were converged to 1×10
−5 eV and 0.02 eV·Å
−1, respectively.
2.6 Electrochemical measurements
All electrochemical measurements were conducted by using a CHI 760D electrochemical workstation in 0.1 mol/L KOH solution with a scan rate of 10 mV·s−1. Hg/HgO and carbon rod served as reference and counter electrodes, respectively. And a rotating disc electrode (RDE) coated with 20 μL of catalyst ink acted as the working electrode, and the ink was prepared by dissolving 5 mg of the catalyst in the mixed solution containing 0.48 mL of water, 0.48 mL of ethanol, and 0.04 mL of Nafion solution. All potentials were calibrated to the reversible hydrogen electrode (RHE): ERHE = EHg/HgO + 0.0591 × pH + 0.098. OER and ORR performances were tested in 0.1 mol/L KOH with 95%-IR compensation. For comparison, the electrochemical activity of commercial RuO2 and Pt/C catalyst were also tested following the same procedure.
2.7 Zinc–air battery test
In the rechargeable zinc−air batteries (RZAB) test, a zinc sheet with a thickness of 0.3 mm was used as the anode, a carbon paper loaded with catalyst as the air cathode, and 6 mol/L KOH solution containing 0.2 mol/L ZnCl2 was used as the electrolyte. The catalyst ink was prepared by the following process. First, 10 mg of the catalyst was dispersed in 1 mL of mixed solution (40 μL of Nafion, 480 μL of water and 480 μL of ethanol) and ground for 10 minutes. Subsequently, 600 μL of the catalyst ink was loaded on carbon paper over a loading area of approximately 4 cm2. The electrochemical performance was tested at room temperature using a CHI-760D electrochemical workstation.
In the flexible RZAB test, a zinc sheet with a thickness of 0.1 mm was used as the anode, a carbon cloth loaded with catalyst as the air cathode, and PVA-KOH gel was used as the electrolyte. The air cathode was prepared by loading the above catalyst ink on carbon cloth for 1.5 g·cm−2. The electrochemical performance was tested at room temperature using a CHI-760D electrochemical workstation.
3 Results and discussion
As schematically illustrated in Fig.1(a), ZIF-8 trapping ferric ions (Fe-ZIF-8) was firstly synthesized by using zinc nitrate hexahydrate, ferric acetylacetonate and 2-methylimidazole as raw materials in a methanol solution (Figs. S1−S4). Then, the as-obtained Fe-ZIF-8 was calcined at 900 °C for 3h in Ar atmosphere to form Fe-doped carbon dodecahedron (FeNC). Subsequently, the FeNC dodecahedrons were deposited by NiFe-layered double hydroxides (LDHs) to form FeNC/LDH nanocomposites at room temperature. The as-obtained sample with the weight ratio (2:1) of FeNC to LDHs was named FeNC/LDH-21, while FeNC/LDH-11 corresponds to the ratio (1:1) of FeNC to LDHs by adjusting the ratio of raw materials. SEM image in Fig.1(b) reveals that the surface of aggregated carbon dodecahedrons is decorated by nanosheet-like NiFe-LDHs. Some broken dodecahedrons exhibit the hollow feature derived from the graphitization of Fe-ZIF-8. TEM image in Fig.1(c) clearly displays the shape of FeNC with the size of 0.75 μm and confirms the hollow structure. Enlarged TEM image in Fig.1(d) further confirms the configuration of FeNC/LDH-21, where the FeNC dodecahedrons are connected by the layered LDHs. The joint area and surface of carbon dodecahedrons show obvious NiFe-LDH nanoflakes, indicating these two components are successfully interconnected. The HRTEM image [Fig.1(e)] reveals that the lattice fringe with a lattice spacing of 0.34 nm is attributed to (002) plane of the partially graphitized carbon skeleton and the lattice fringe with a lattice spacing of 0.25 nm assigned to (012) plane of NiFe-LDH. The selected area electron diffraction (SAED) pattern in the inset of Fig.1(c) also shows that the presence of polycrystalline NiFe-LDH with discernible diffraction rings. Furthermore, the scanning TEM image [Fig.1(f)] and corresponding elemental mapping images [Fig.1(g)] clearly show the uniform distribution of C, Fe, O and Ni, further illustrating the successful synthesis of FeNC/LDH.
XRD patterns in Fig.2(a) reveal that FeNC has a weak diffraction peak, indicating its low crystallinity. A peak assigned to Fe
3C (N0.35-0772) is also observed in addition to (002) peak of carbon, which is caused by the surplus iron raw materials. After loading NiFe-LDHs on FeNC dodecahedrons, the as-formed FeNC/LDH-21 and FeNC/LDH-11 show obvious XRD pattern assigned to NiFeOOH in addition to (002) peak of carbon (No. 40-0215). Raman spectra show two prominent peaks at 1349 (D band) and 1588 (G band) cm
−1 in Fig.2(b). The intensity ratio (
ID/
IG) of FeNC, FeNC/LDH-21, and FeNC/LDH-11 is 1.04, 1.0, and 1.01, respectively, indicating the similar graphitization. X-ray photoelectron spectroscopy (XPS) was conducted to further verify the chemical states of FeNC/LDH-21 and FeNC/LDH-11. The high-resolution C 1s spectra of FeNC/LDH-21 and FeNC/LDH-11 reveal the presence of C‒C, C‒O and C(O)‒O, which indicates the successful formation of carbon matrix and the introduction of oxygen-containing functional groups during the loading process of LDHs [Fig.2(c)]. FeNC/LDH-21 and FeNC/LDH-11 show obvious N 1s peaks, including pyridinic-N (Fe‒N
x) (398.3 eV), pyrrolic-N (399.6 eV), graphitic-N (401.1 eV), and oxidized N (402.6 eV) [Fig.2(d)]. Compared to FeNC/LDH-11, FeNC/LDH-21 exhibits more distinctive characteristic peaks, especially pyridinic-N and graphitic-N, suggesting more sites for the irons to be anchored by chemical bonding to form Fe‒N‒C active sites [
41]. The spectra of Ni 2p of FeNC/LDH-21 and FeNC/LDH-11 are shown in Fig.2(e). Two deconvoluted peaks at 873.93 and 856.23 eV are ascribed to Ni 2p
3/2 and Ni 2p
1/2 respectively, together with two corresponding satellite peaks (Sat.) located at 880.53 and 862.42 eV, indicating the existence of iron at two different geometrical positions (i.e., octahedral and tetrahedral geometry) [
42]. Similarly, the high-resolution Fe 2p spectra of FeNC/LDH-21 and FeNC/LDH-11 in Fig.2(f) show distinctive peaks at 724.53 and 711.48 eV and two satellite peaks are also observed at 726.52 and 713.38 eV, which are attributed to Fe 2p
3/2 and Fe 2p
1/2 respectively, confirming the presence of Fe
3+ oxidation state [
43,
44].
Electrochemical tests were carried out to evaluate the bifunctional performances of the as-prepared catalysts by using a three-electrode system in O2-saturated 0.1 mol/L KOH. LSV curves in Fig.3(a) show that FeNC/LDH-21 exhibits a half-wave potential of E1/2 = 0.82 V, which is consistent with the CV curve (Fig. S5), much better than NiFe-LDHs (0.6 V) and FeNC/LDH-11 (0.76 V), and comparable to commercial Pt/C (0.82 V), but inferior to FeNC (0.85 V). This confirms the outstanding contribution of FeNC sites to the ORR performance. Furthermore, the Tafel slope of FeNC/LDH-21 is 73.8 mV·dec−1 as shown in Fig.3(b), significantly better than that of Pt/C (93.5 mV·dec−1) and FeNC/LDH-11 (95.7 mV·dec−1), indicating fast reaction kinetics. The electron-transferred number of FeNC/LDH-21 is calculated to be about 4.0 by testing LSV curves at different rotational speeds [Fig.3(c)]. Furthermore, the rotating ring-disk electrode (RRDE) was used to confirm the electron-transferred number (Fig. S6). The electron transferred number is calculated to be n = 3.9 [Fig.3(d)], further confirming a four-electron reaction. The stability performance is also one of the important indicators to judge electrocatalysts. The current‒time (j‒t) test reveals that FeNC/LDH-21 achieves a current retention rate of 78% after 8 h, which is significantly better than the commercial Pt/C [Fig.3(e)].
To evaluate the bifunctional electrocatalytic performance, LSV curves of the as-prepared electrocatalyst toward OER were measured. Fig.3(f) shows that FeNC/LDH-21 exhibits a potential of Ej = 10 = 1.5 V at the current density of 10 mA·cm−2, which is comparable to NiFe-LDH (1.5 V) and significantly better than commercial RuO2 (1.66 V). Correspondingly, the Tafel slope of FeNC/LDH-21 derived from the LSV curve is 45.17 mV·dec−1 [Fig.3(g)], which is significantly better than RuO2 (104.35 mV·dec−1), suggesting the fast reaction kinetics. The difference of half-wave potential and potential at the current density of 10 mA·cm−2 (ΔE = Ej = 10 ‒ E1/2) is usually used to judge the bifunctional performance toward ORR and OER. It can be seen from Fig.3(h) that the ΔE of FeNC/LDH-21 is about 0.68 V, which is significantly better than the mixture of Pt/C+RuO2 (0.84 V) and most reported bifunctional electrocatalysts (Table S1). Moreover, the double-layer capacitance (Cdl) was calculated by testing CV curves in the voltage window from 1.03 to 1.13 V at different rates [Figs. S7(a, b)], to further evaluate the electrochemically active surface area (ECSA), because ESCA is an important indicator of active sites. The Cdl of FeNC/LDH-21 was 2.37 mF·cm−2, higher than that of FeNC/LDH-11 (1.44 mF·cm−2), indicating larger ECSA and more active sites [Fig. S7(c)]. Electrochemical impedance spectroscopy (EIS) shows that FeNC/LDH-21 has a smaller charge-transfer resistance (47.2 Ω) than FeNC/LDH-11 (54.3 Ω) [Fig.3(i)].
To gain insights into the improved bifunctional electrocatalytic activity, density functional theory (DFT) calculation was conducted by constructing the models of FeNC, NiFe LDH and FeNC/LDH (Fig. S8). Fig.4(a‒d) show the total density of states (DOS) of FeNC, NiFe LDH, FeNC/LDH (Fe‒O‒Fe) and FeNC/LDH (Fe‒O‒Ni), where Fe‒O‒Fe (or Fe‒O‒Ni) represents the heterostructure of FeNC/LDH is bonded by Fe in FeNC and ‒O‒Fe (or ‒O‒Ni) in LDH. The total DOS shows d-band center (
εd) of FeNC/LDH (Fe‒O‒Fe) and FeNC/LDH (Fe‒O‒Ni) is ‒1.502 and ‒1.515 eV, respectively, which is between ‒1.295 eV of FeNC and ‒1.567 eV of NiFe LDH. This suggests that the formation of coupling interface modulates the ε
d to be a moderate value, which is beneficial for the adsorption/desorption of oxygen-containing intermediates [
45−
47]. Differential charge density of FeNC/LDH (Fe‒O‒Fe) and FeNC/LDH (Fe‒O‒Ni) demonstrates the obvious charge redistribution along Fe‒O‒Fe [Fig.4(e)] bridge and Fe-O-Ni [Fig.4(f)] bridge, which effectively affects the activity of active sites.
Considering the bifunctional performance of FeNC/LDH-21 toward ORR and OER, the potential application in ZABs was evaluated by coating FeNC/LDH-21 on carbon cloth [Fig.5(a)] as air electrode, while a polished zinc sheet used as the anode and 1 mol/L KOH solution as electrolyte. A red light-emitting diode (LED) was able to be successfully lighted by connecting two ZABs in series [Fig.5(b)]. The discharge capability of ZABs was tested at different current density from 2 to 40 mA·cm−2 [Fig.5(c)], demonstrating good discharge performance of FeNC/LDH-21. In addition, the discharge voltage can recover when the discharge current is turned back to 2 mA·cm−2 again. The charge-discharge performance of FeNC/LDH-21 and Pt/C+RuO2 was tested as shown in Fig.5(d). FeNC/LDH-21 shows a lower polarization between the charge and discharge than Pt/C+RuO2 at the same current density, which is consistent with its outstanding bifunctional performance toward ORR and OER. Moreover, the ZAB with FeNC/LDH-21 as cathode achieves a power density of 85 mW·cm−2 [Fig.5(e)], which is obviously higher than that of commercial Pt/C+RuO2 (76 mW·cm−2). The specific capacity of the FeNC/LDH-21-based ZAB is 810 mAh·g−1 at the current density of 10 mA·cm−2, which is higher than that of Pt/C+RuO2 (786 mAh·g−1) [Fig.5(f)] and most values reported recently (Table S2). In addition, the FeNC/LDH-21-based ZAB exhibits stable cycling performance up to 160 h in charge-discharge process [Fig.5(g)]. The enlarged cycling plots in Fig.5(h) show that the round-trip efficiency of the ZAB with FeNC/LDH-21 as cathode ranges from 55.2% to 56.2%, which also indicates its good stability.
FeNC/LDH-21-based electrocatalyst was also assembled with a flexible rechargeable ZAB to illustrate its potential application in wearable devices. FeNC/LDH-21 with a loading of 1.5 g·cm−2 was coated on the carbon cloth surface as the air electrode, and nickel foam, PVA and zinc foil were used as the collector, electrolytes, and anode, respectively [Fig.6(a)]. In the open-circuit voltage test, the assembled flexible ZAB delivers a stable open-circuit voltage of 1.43 V [Fig.6(b)], which is comparable to the recently reported values (Table S3). Fig.6(c) demonstrates that the maximum power density of the flexible ZAB reaches 32.4 mW·cm−2, lower than the values in the literature (Table S3), which could be caused by the difference of assembly procedure. Through the test at different current densities, it is found that the flexible ZAB assembled by FeNC/LDH-21 has good rate capability with a retention rate of 94%, when the current density is returned to 1 mA·cm−2 [Fig.6(d)]. In the cycling test at the current density of 2 mA·cm−2, the flexible ZAB maintains a stable charge/discharge process for 20 h [Fig.6(e)]. In addition, Fig.6(f) demonstrates that two flexible ZABs connected in series can successfully light up a red LED. Moreover, at the bending or thumping situation, the LED keeps lighting, indicating a stable discharge state of the flexible ZABs at hostile conditions. The above battery performance tests also demonstrate the potential application of FeNC/LDH-21 in the energy supply for flexible wearable devices.
4 Conclusion
A bifunctional electrocatalyst has been successfully synthesized by coupling carbon dodecahedrons with Fe‒N‒C active sites toward ORR to active NiFe-LDHs with high-valence Ni/Fe sites toward OER. The unique layered structure formed by anchoring NiFe-LDH on the FeNC surface facilitates the transport of electrons in the reaction, effectively reducing the agglomeration of active sites and significantly improving the OER and ORR performance. DFT calculations reveal the modulated d-band center for FeNC/LDH and obvious charge redistribution along Fe‒O‒Fe and Fe‒O‒Ni bridges between FeNC and LDHs. As the optimized FeNC/LDH-21 exhibits a half-wave potential of 0.82 V toward ORR and 1.5 V at the current density of 10 mA·cm−2 toward OER. The ZAB with FeNC/LDH-21 as the cathode demonstrates a high-power density of 85 mW·cm−2 and steadily works for 160 h. The solid-state ZAB also shows good charge‒discharge performance with a power density of 32.4 mW·cm−2 and remains stable at different bending or hammering states. This bifunctional performance and stability are attributed to the synergistic interaction between FeNC and NiFe-LDHs. This study will promote the development of cost-effective bifunctional electrocatalysts toward ZABs.