A lean-zinc anode battery based on metal–organic framework-derived carbon

Abstract

Improving zinc metal (Zn⁰) reversibility and minimizing the N/P ratio are critical to boosting the energy density of Zn⁰ batteries. However, in reality, an excess Zn source is usually adopted to offset the irreversible zinc loss and guarantee sufficient zinc cycling, which sacrifices the energy density and leads to poor practicability of Zn⁰ batteries. To address the above conundrum, here, we report a lean-Zn and hierarchical anode based on metal–organic framework (MOF)-derived carbon, where trace Zn⁰ is pre-reserved within the anode structure to make up for any irreversible zinc source loss. This allows us to construct low N/P ratio Zn⁰ full cells when coupling the lean-Zn anode with Zn-containing cathodes. Impressively, high Zn⁰ reversibility (average Coulombic efficiency of 99.4% for 3000 cycles) and long full-cell lifetime (92% capacity retention after 900 cycles) were realized even under the harsh lean-Zn condition (N/P ratio: 1.34). The excellent Zn reversibility is attributed to the hierarchy structure that homogenizes zinc ion flux and electric field distribution, as confirmed by theoretical simulations, which therefore stabilizes Zn⁰ evolution. The lean-Zn anode design strategy will provide new insights into construction of high-energy Zn⁰ batteries for practical applications.

INTRODUCTION

Safe, environmentally friendly, efficient, and economical rechargeable battery technologies are critical to enabling a sustainable and electrified future. Despite prevailing state-of-the-art lithium-ion batteries exclusively based on toxic and flammable organic electrolytes, these have questionable sustainability due to the limited lithium resource reservoir and intrinsic safety concerns (e.g., fires and even explosions).¹ In recent decades, rechargeable batteries using nonflammable and high-ionic-conductivity aqueous electrolytes have increasingly attracted more attention. Among those, aqueous zinc metal (Zn⁰) batteries (AZMBs) have been recognized as one of the most promising beyond-lithium technologies, which are primarily useful due to their excellent characteristics such as high theoretical energy density, cheap and easily available raw materials, high safety, low manufacturing cost, and so on.^2–4 To date, a considerable number of cathode materials have been developed and explored for AZMBs, including Prussian blue analogues, vanadium-based compounds, manganese-based compounds, and organic materials.^4,5 Yet, the large-scale implementation of AZMBs has been limited by the poor irreversibility on the anode side because of the parasitic reaction at the interface between the Zn⁰ anode and the electrolyte and the maligant dendritic/dead Zn⁰ evolution upon repeated electroplating/stripping cycling. Low Zn⁰ reversibility leads to continuous Zn source loss within the battery system. To sustain the necessary Zn cycling in AZMBs, a thick and heavy Zn⁰ foil is usually adopted as the anode. Such a Zn-excess configuration, however, decreases the energy density of whole cells.

Ideally, an anode-free full cell with zero-excess Zn is anticipated to minimize batteries’ overall mass or volume and maximize the energy density.^6–8 In this configuration, the Zn⁰ anode is in situ-formulated through Zn⁰ plating on the anode current collector after initially charging the anode-free cell, where the zinc source only comes from zinc-containing cathodes. Upon discharge, Zn⁰ on the anode side will transform into Zn²⁺, and then Zn²⁺ will dissolve into the electrolyte, migrate from the anode to the cathode through the electrolyte, and finally insert back into the active sites of the cathode. Thus far, a few anode-free ZMBs have been successfully demonstrated, involving anode-free Zn-MnO₂,⁹ Zn-LiMnO₂,¹⁰ Zn-ZnMn₂O₄,¹¹ and Zn-graphite¹² cells. However, in practice, the attainable cycling performance of these anode-free ZMBs is far from satisfactory, for example, 68.2% capacity retention after 80 cycles in anode-free Zn-MnO₂,⁹ which fails to compete with Zn-excess systems. The poor cyclability occurs as a result of the inevitable zinc loss during battery cycling, due to the irreversible Zn⁰ plating/stripping and side reactions between Zn⁰ and the electrolyte, particularly when planar current collectors (such as Ti and Cu foils) and routine electrolytes (such as ZnSO₄- and Zn(CF₃SO₃)₂-based ones) are used. The irreversible Zn⁰ plating/stripping behavior in anode-free cells is similar yet more severe than that in the Zn-excess system. Therefore, to realize long cycling of full cells without sacrificing the energy density, a useful and more practical solution is to construct full cells with low N/P ratio, that is, adopting a lean-Zn configuration.

In the lean-Zn battery configuration, the limited zinc reservoir can be generally achieved on the basis of the following two factors: (1) direct use of thin Zn foils and (2) inclusion of a designated amount of Zn metal with three-dimensional (3D) hosts using electroplating and molten infiltration methods. The former approach usually has the disadvantage of limited Zn nucleation sites that are unfavorable for smooth Zn plating. Comparatively, loading Zn⁰ on 3D hosts can better regulate Zn nucleation/growth behaviors to achieve a smooth Zn morphology owing to the high surface area, porous network, and stress release advantages of 3D structures. However, the process of loading Zn⁰ on 3D hosts is often complicated and difficult to implement, which is not conducive to practical application. It is well known that metal–organic frameworks (MOFs) or zeolitic imidazolate frameworks (ZIFs) have a porous and designable framework that contains a metal center and an organic ligand, thus making them promising hosts and/or coating materials to stabilize both cathodes and anodes in AZMBs.^4,5 For example, the introduction of ZIF-67-derived carbon into MnO₂ to form C-MnO₂ can effectively solve the problem of manganese-based cathode dissolution and stably run for 1000 cycles.¹³ Mn(BTC) MOF, where the zinc storage mechanism is realized through the conversion reaction from Mn(BTC) into MnO₂, has achieved high reversible capacity and excellent long-term cycling performance over 800 cycles.¹⁴ Coating ZIF-8 on a Zn anode has been proven to be effective in homogenizing charge and ion distributions at the anode–electrolyte interface, which can prolong the cycle life of the anode by eight times.⁷ In addition, a lot of other MOF-based materials, such as oxygen vacancy-enriched MOF-derived MnO/C hybrids,¹⁵ layered V-MOF nanorods,¹⁶ and V-MOF@graphene-derived 2D hierarchical V₂O₅@graphene,¹⁷ have also been studied in AZMBs. Considering that the metal source in MOFs and ZIFs can be regulated, a zinc-containing MOF or ZIF or their derivates could be potentially used as lean-Zn anode materials, which, however, has never been discussed before.

In this work, for the first time, we propose a novel anode structure, that is, a hierarchical host with trace Zn⁰, which enables the construction of a highly reversible, long-life, and anode-lean Zn⁰ battery in coupling with a zinc-containing cathode. The hierarchical and lean-Zn anode structure was realized by simultaneously carbonizing the conventional metal–organic framework (MOF-5) and reducing the trace amount of Zn²⁺ in MOF-5 into metallic Zn. The resultant MOF-5-derived carbon (denoted as MDC) retains the original porous structure and shows excellent wettability toward the aqueous ZnSO₄ electrolyte. More interestingly, the hierarchical structure yields massive Zn⁰ nucleation sites and can homogenize the electric field distribution, which promotes smooth and dendrite-free Zn plating, while residual Zn⁰ inside MDC functions as the backup Zn⁰ source to make up for the irreversible zinc loss during battery cycling. Consequently, high Zn⁰ plating/stripping reversibility (3000 cycles with an average Coulombic efficiency (CE) of 99.4% and a low voltage hysteresis of 47.42 mV) was achieved in half-cells. When pairing the MDC anode with a Zn-containing cathode (Zn/Mn-MOF@CNT), a high specific discharge capacity (398.8 mA h g⁻¹ at 1 C rate) and excellent cycling stability (92% capacity retention after 900 cycles at a 3 C rate, and 60% capacity retention after 1400 cycles at a 5 C rate) were demonstrated for the full cell.

EXPERIMENTAL SECTION Preparation of a lean-Zn anode

Zn(NO₃)_2·6H₂O (1, 1.4, 2, 2.5, 3 mmol) was dissolved in 20 mL of N,N-Dimethylformamide (DMF) to form solution A. 1 mmol of terephthalic acid was dissolved in 20 mL of DMF to form solution B. Solution B was slowly added to solution A and stirred for 10 min to ensure full mixing of the solutions. The mixed solution was placed into a hydrothermal reactor for a high-temperature and high-pressure reaction at 150°C for 7 h, followed by filtration four times to remove impurities. The extracted powder was soaked in chloroform for 24 h to remove the residual DMF, and then filtered again with chloroform as a solvent for two to three times and dried in vacuum at 160°C for 10 h to obtain MOF-5. The annealing in argon can be divided into two stages. First, the temperature is increased to 600°C at 5°C/min for 4 h, and the temperature is increased to 800°C, 850°C, 900°C, 950°C, and 1000°C at 3°C/min for 4 h, and then the MOF-5-derived carbon is obtained by natural cooling. The binder (polyvinylidene fluoride [PVDF]), conductive agent (acetylene black), and active substance (MOF-5-derived carbon) were ground in a mortar at the ratio of 8:1:1 for 5 min until they were well mixed. The mixture was transferred to a 25 × 40 weighing bottle, and N-methyl-2-pyrrolidone (NMP) was added dropwise while stirring until a fluid slurry was formed. The prepared slurry was coated on a copper foil with a coating thickness of 60 μm. Finally, the copper foil was dried in vacuum at 80°C for 12 h, and cut into 12 mm electrode pieces using a cutting machine. The difference between the mass of the pure copper foil and the corresponding electrode sheet is the mass of the mixed material. According to the proportion of each component, 80% of the mixed material is active material.

Preparation of Zn/Mn-MOF@CNT composites

0.02 mol of 2-methylimidazole was dissolved in 50 mL of methanol and denoted solution A. The solution of 2 mmol of Mn(NO₃)₂ with mass fraction of 50% and 1 mmol of Zn(NO₃)₂·6H₂O were co-dissolved in 50 mL of methanol, which was denoted solution B. Solution A was slowly added to solution B, stirred for 10 min to mix the two solutions completely, and a pretreated carbon nanotube paper was placed in the mixed solution. Zn/Mn-MOF (ZM) was grown on carbon nanotube paper (CNT) in situ under magnetic stirring; the reaction temperature was 35°C and the reaction time was 18 h. After the reaction was complete, Zn/Mn-MOF@CNT was removed from the mixed solution. The Zn/Mn-MOF@CNT cathode material was obtained by washing the surface with methanol 3 times and drying it in a blast drying box at 60°C for 12 h.

Materials' characterizations

The morphology and structure of the samples were investigated by scanning electron microscopy (SEM) (JEOL JSM-7200F) and transmission electron microscopy (TEM) (FEI Tecnai G2 F20). Elemental mapping images were recorded using an EDX spectroscope attached to a SEM. The crystal phases of the samples were identified by X-ray diffraction (XRD) on a Rigaku Smartlab SE with a Cu Kα radiation (λ = 1.5406 Å) at a voltage of 30 kV and a current of 10 mA. The X-ray photoelectron spectroscopy (XPS) spectra were obtained using a Thermo Scientific K-Alpha System. Thermogravimetric analysis (TGA) was performed using a TG DTA7200 thermal analyzer under air flow. N₂ adsorption–desorption isotherms and pore size distribution curves were measured using Micromeritics ASAP 2460. The degree of amorphousness of the carbon materials was analyzed by Raman spectroscopy (HR Evolution). The hydrophilicity of the carbon coating derived from MOF-5 was analyzed using a contact angle meter (OCA100).

Electrochemical measurements

CR-2032-type coin cells were assembled for electrochemical tests in an air atmosphere using a glass fiber filter (GF-D; Whatman) as the separator. The half battery was prepared using a zinc foil as the anode and MDC–Cu or Cu as the cathode. The full battery was prepared using Zn/Mn-MOF@CNT as the cathode and MDC–Cu as the anode. The electrolyte used in the above battery was 2 M ZnSO₄ + 0.1 M MnSO₄. MDC, carbon black, and PVDF were mixed thoroughly in an NMP solvent in an 8:1:1 mass ratio with magnetic stirring to obtain a slurry. The mixture was coated on a copper foil and dried in a vacuum oven at 80°C for 12 h. The loading mass of the MDC active material in the electrodes was about 1–1.3 mg cm⁻². The value of N/P ratio was calculated according to the following formula: [Image Omitted. See PDF]where C_1,anode is the initial discharge capacity of 0.1 C anode per unit area, C_1,cathode is the initial charge capacity of 0.1 C cathode per unit area, and η_1,anode, η_2,anode, and η_3,anode are the CEs of the first, second, and third cycles at 0.1 C, respectively. The LAND CT3001A battery-testing system was used to test the cycle performance and rate performance of the battery. The impedances were carried out from 0.01 Hz to 100 kHz with an amplitude of 5 mV using an electrochemical workstation (Modulab XM).

Finite element simulation based on a lean-Zn anode model

Simplified 2D Zn²⁺ flux and 3D electrodeposition models at the interface between the anode and the electrolyte based on COMSOL Multiphysics software were established to compare the proportional schematics of ion flux, current density, and electric field distribution. In the 2D Zn²⁺ flux model, the length and thickness of the electrode were set to 5 µm and 3 µm, respectively, and the diameter of the raised semicircle was 0.3 μm. The exchange current density was 1 A m⁻² and the ion diffusion coefficient was set to 1e⁻¹² m² s⁻¹. The initial concentration was set to 1000 mol m⁻³. In the 3D electrodeposition model, the length and thickness of the electrode were set to 80 and 10 µm, respectively. The diameter of the glass fiber was set to 2 μm. The ionic conductivity of the electrolyte was about 0.35 S m⁻¹. The gap between the glass fiber was about 1 μm. The scaffold height was 6 μm, with a diameter of 80 nm. An overpotential of 3.5 mV was used for voltage excitation between the anode side and the electrolyte side.

RESULTS AND DISCUSSION

In this work, MOF-5 material was selected as the lean-Zn anode precursor due to its unique composition (trace Zn metal source and porous organic ligand), highly ordered structure, and ultrahigh porosity. The pristine MOF-5 materials were synthesized using a solvothermal method, modified from the reported literature.¹⁸ Four peaks at 6.81°, 9.64°, 13.73°, and 15.31° in the XRD pattern (Figure S1A) correspond to the (200), (220), (400), and (420) crystal planes of MOF-5, respectively, indicating the successful synthesis of the MOF-5 materials. The MDC was thereafter obtained by annealing MOF-5 in an inert Ar atmosphere, during which Zn²⁺ within the MOF-5 skeleton was transformed into Zn⁰, while the organic ligand backchain was pyrolyzed into conductive carbon (Figure 1A).^19,20 To establish the relationship between the MDC's structure and ZMBs’ performance, in this work, we designed and fabricated 25 kinds of MDC by adopting different MOF-5 with varied organic ligand to metal source ratios and adjusting the thermal annealing temperature (see Section 2). Note that the original porous structure of MOF-5 could be well maintained in MDC and it also showed good structural robustness (Figure S2).²¹

Figure 1. (A) Schematic diagram of annealing and Zn plating and the stripping process of MOF-5-derived carbon. (B) Schematic diagram of the electroplating/stripping principle of Zn on MOF-5-derived carbon. MOF, metal–organic framework.

Therefore, it can be concluded that the use of MDC as a lean-Zn anode is capable of not only improving Zn⁰ reversibility but also retaining sufficient Zn source in the battery system, as shown in Figure 1B. Within MOF-5 precursors, there are a large number of metal sources. After annealing treatment, some metal sources protruded from the carbon skeleton driven by the air flow, leaving massive zinc storage sites. These sites can guide uniform zinc plating/stripping and suppress dendrite growth, thus producing high Zn⁰ reversibility. Furthermore, the existence of residual Zn⁰ within MDC can supplement the irreversible zinc loss in the cycling to achieve a long battery cycle life, even though bulk Zn foils are absent.

To evaluate the residual zinc content within MDC, TGA was first conducted on two representative MOF-5s: one had a high zinc content (MOF-5-3:1, Figure 2A) and the other had a low zinc content (MOF-5-1:1, Figure 2B). In general, the weight deviation can be divided into three stages^22,23: (I) weight decrease at 0-300°C due to the water evaporation, (II) weight change at 300-600°C due to the carbonization of organic ligands and the formation of C/ZnO composites, and (III) weight change at 600–1000°C owing to the ZnO reduction (by C) and gradual zinc evaporation with increasing temperature. The low Zn mass loss rate in a wide temperature range allows us to accurately control the zinc removal from MOF-5. In addition, according to the mass loss in phase III that is mainly due to Zn evaporation, we can further calculate the residual Zn content at any specific temperature. The calculation result is summarized in Figure 2C. As can be seen, the MOF-5-3:1 showed 37%, 31%, and 19% zinc residues when annealed at 800°C, 850°C, and 900°C, respectively, while the residual zinc contents in MOF-5-1:1 were 22%, 13.9%, and 2.5% after annealing at 800°C, 850°C, and 900°C, respectively.

Figure 2. Thermogravimetric diagram of (A) MOF-5-3:1 and (B) MOF-5-1:1. (C) Residual amount of zinc at different annealing temperatures. (D) Zn 2p3/2 XPS spectrum of MOF-5. (E) Zn 2p3/2 XPS spectrum of MDC (annealing temperature lower than 907°C). (F) EDS image of MDC. (G,H) TEM images of the Zn0 in the MDC. (I) Pore size distribution curve of partial MDC. EDS, energy-dispersive spectrum; MDC, MOF-5-derived carbon; MOF, metal–organic framework; TEM, transmission electron microscopy; XPS, X-ray photoelectron spectroscopy.

The existence of metallic Zn in MDC samples was confirmed by XPS. As shown in Figure 2D, the Zn 2p_3/2 peak of MOF-5 before annealing deconvoluted into two peaks: ZnO at 1022.19 eV and Zn(OH)₂ at 1024.19 eV. In comparison, the Zn⁰ signal (at 1021.6 eV) can be identified in the Zn 2p_3/2 peak of annealed MDC (derived from MOF-5-3:1 and annealed at 850°C, Figure 2E). The appearance of Zn⁰ in MDC is attributed to the thermal reduction of Zn²⁺ in the inert atmosphere.²⁴ The energy-dispersive spectrum mapping image shows the uniform distribution of Zn in MDC (Figure 2F). However, when the annealing temperature was higher than the boiling point (907°C) of Zn metal, only two broad characteristic peaks of 23.6° and 44.3° could be observed in the XRD pattern of MDC, which correspond to the (002) and (100) crystal planes of MDC, respectively, indicating that MOF-5 has been completely carbonized and Zn⁰ no longer exists (Figure S1B).²⁵ Consistent with the above results, obvious Zn 2p, C 1s, and O 1s characteristic peaks can be observed in the XPS patterns of MOF-5 and MDC (annealing temperature lower than 907°C), while Zn 2p characteristic peaks cannot been detected in the XPS pattern of MDC (annealing temperature higher than 907°C), which proves once again that Zn⁰ no longer exists in the fully carbonized MOF-5 (Figure S1C). The existence of Zn⁰ in MDC at an annealing temperature lower than 907°C was further confirmed by the TEM characterization. Figure 2G shows the typical particle morphology of Zn metal with a diameter of about 80 nm. In addition, clear lattice fringes with a lattice spacing of 0.239 nm, corresponding to the (101) crystal plane of Zn metal, can be identified (Figure 2H). Note that some zinc remains in the oxidized state (i.e., ZnO), which has been reported to be more zincophilic than metallic Zn and can promote dense nucleation and 2D growth in the early deposition process.²⁶

The specific surface area and pore structure of a 3D host are critical to the ionic migration and volume accommodation within metal anodes.^27,28 To reveal the structural property of the resultant MDC, a N₂ adsorption/desorption isotherm test was conducted. As shown in Figure S3, MDC samples obtained under different conditions all showed typical type IV isotherms with characteristic hysteresis rings, which is indicative of a mesoporous structure. The pore size distribution curve obtained using the BJH calculation method shows that the pore size in MDC varies between 0 and 20 nm and is concentrated around 3.5 nm (Figure 2I). Here, the hierarchical porous structure promotes fast ion transfer and enables stabilization of Zn⁰ plating/stripping within MDC through a multiscale spatial confinement effect.^29–32 In addition, a high specific surface area of 270.921 m² g⁻¹ was achieved for MDC, which enables provision of massive active sites for smooth Zn nucleation/plating, increases the contact area between the electrode and the electrolyte, and facilitates ionic transport.^33,34

To clarify the Zn⁰ reversibility in MDC, we carried out repeated zinc electroplating/stripping tests based on a Cu||Zn half-cell configuration, from which the CE can be obtained as the indicator. To demonstrate, MDC–Cu||Zn half-cells and Cu||Zn half-cells (control) were assembled for the following electroplating/stripping tests. During electroplating, Zn was plated at a current density of 1 mA cm⁻² to reach a total capacity of 0.5 mA h cm⁻². Each stripping process was followed by Zn plating using the same current density until the working potential reached 0.5 V vs. Zn/Zn²⁺. For a bare Cu foil-based cell, that is, a Cu||Zn cell, a low initial CE of 73.74% was obtained (Figure 3A), along with voltage hysteresis (between plating and stripping) of 93 mV (Figure S4B), and a nucleation overpotential (the difference between the bottom of the voltage tip at the beginning of nucleation and the flat part of the voltage platform) of 62.3 mV (Figure S4C). The Cu||Zn half-cell failed after 145 cycles, which may have been induced by the dendritic zinc due to the nonuniform Zn⁰ nucleation/plating on the Cu surface. Comparatively, on using the 3:1-850 MDC as the host, the voltage hysteresis was reduced to 47.42 mV and the nucleation overpotential was reduced to 14.3 mV. Similar low voltage hysteresis and nucleation overpotential were also achieved for MDC in other states (Figures S5 and S6, Tables S1 and S2). The significant decrease in the overpotential in MDC is attributed to the hierarchical porous structure that homogenizes Zn ion flux.^29–32 The stable Zn⁰ plating/stripping behavior in MDC was further confirmed by the minor charge-transfer resistance change upon cycling (Figure 3B), which was distinct from the marked increase in the impedance behavior when a bare Cu foil was used (Figure S4D). Impressively, the 3:1-850 MDC-based Cu||Zn cell could stably run for more than 3000 cycles at 1 mA cm⁻² and 0.5 mA h cm⁻², showing an average CE of 99.4% (Figure 3A). Extended cycling life and high average CE were also found at a relatively higher cycling capacity of 1 mA h cm⁻² (100 cycles, average CE: 98.2%), 2 mA h cm⁻² (95 cycles, average CE: 98.4%), and even 3 mA h cm⁻² (48 cycles, average CE: 98.7%), as shown in Figure 3C and Figure S7. In addition, the plating/stripping voltage curves upon cycling remain almost the same, further proving the excellent reversibility. The improved CE and stability are ascribed to minimized competitive reaction (i.e., HER) and dendritic/“dead” zinc formation.³⁵ To evaluate the rate performance of the MDC coating, we conducted an electrochemical plating/stripping test at stepwise increasing current densities from 0.2 to 2 mA cm⁻². The results in Figure 3D show that the MDC hosts have excellent adaptability to the current density change, and the voltage hysteresis remains relatively stable with the increase of current density. The Zn plating/stripping reversibility was also examined under different Zn⁰ loadings (Figure 3E). We found that MDC–Cu||Zn cells showed high CEs under all test conditions (from 0.25 to 3 mA h cm⁻²). These results strongly support our expectation that the high specific surface area and porous structure of MDC can stabilize the Zn⁰ anode by promoting a smooth Zn⁰ evolution morphology and facilitating charge transfer within the host.^29–34

Figure 3. Electrochemical Zn plating/stripping performance of the MDC–Cu electrode in 2 M ZnSO4 + 0.1 M MnSO4. (A) Low-loading (0.5 mA h cm−2) galvanostatic plating/stripping stability of the MDC–Cu and bare Cu electrodes at 1 mA cm−2. The inset shows the corresponding 300th, 800th, and 1100th plating/stripping profiles of MDC–Cu. (B) Nyquist plots of the MDC–Cu electrode after the initial 15 cycles. (C) High-loading (3 mA h cm−2) galvanostatic plating/stripping stability of the MDC–Cu electrode at 1 mA cm−2. The inset shows the corresponding 10th, 20th, and 35th plating/stripping profiles. (D) Galvanostatic Zn plating/stripping curves of the MDC–Cu electrode at different current densities (electrodeposition time: 30 min). (E) Galvanostatic plating/stripping curves of the MDC–Cu electrode recorded over a range of capacities at 1 mA cm−2. MDC, MOF-5-derived carbon.

As mentioned above, the Zn⁰ reversibility is highly dependent on the Zn⁰ evolution morphology. SEM characterization was thereafter performed to visualize the Zn⁰ morphology (Figure 4B and Figure S8). The comparison of morphologies of different hosts was conducted after Zn⁰ plating/stripping at varied capacities (0.5, 1, and 2 mA h cm⁻²). In sharp contrast to the large flower-shaped and spherical shape of Zn⁰ deposits on bare copper after the initial electroplating, the Zn⁰ deposits on the MDC–Cu electrode were well-arranged nano-flake arrays. In the subsequent stripping process, the plated Zn⁰ was reversibly removed from MDC–Cu and the original dense arrangement of particles was maintained. Comparatively, the surface of bare copper not only changed in shape but became rugged; in addition, many irregular flakes of “dead zinc” can be observed, which will affect the surface electric field distribution and result in more severe nonuniform Zn⁰ deposition.³⁶ As shown in Figure S8, at a low cycling capacity of 0.5 mA h cm⁻², a large number of flake arrays in different directions were deposited on the bare copper surface and stacked with each other to form flower-shaped and large-sized aggregates. The uneven and loose distribution of flake Zn⁰ dendrites may lead to an uneven interfacial electric field and Zn²⁺ ion concentration, which further promotes the growth of dendrites. On the contrary, after repeated Zn⁰ deposition/dissolution on the surface of the MDC coating for 500 cycles, ultra-thin hexagonal patches with a size of 1–2 μm were observed on the surface, consistent with previously reported results.³ After Zn exfoliation, the MDC particle still retained a uniform and dense arrangement. Similar morphological evolution was also observed under the cycling capacities of 1 and 2 mA h cm⁻², where well-arranged nano-sheet arrays were formed on the surface of the MDC coating while the Zn⁰ deposition on bare copper was much more disorderly (Figure S8). The reversible and stable Zn⁰ morphology changes on MDC are shown in Figure 4A, in which MDC shows a highly conductive network and a stable interface to induce uniform nucleation and deposition of Zn⁰. On the other hand, the nonuniform and dendritic Zn⁰ evolution pattern on bare copper is attributed to the limited surface area and randomly distributed Zn⁰ nucleation behavior (Figure S8). To explore the stability of MDC as a zinc storage substrate, we carried out Raman tests. As shown in Figure 4C–E, the MDC defect density ratio (I_D/I_G) increased from 1.093 to 1.166 after electroplating, and then decreased to 1.123 after stripping. After cycling, the defect density of MDC only increases by 2.7%, proving that MOF-5-derived carbon can be used as a stable zinc storage substrate in the electroplating/stripping process.³⁷

Figure 4. (A) Shape change of MDC–Cu in the electroplating/stripping process. (B) SEM images of MDC–Cu under different plating/stripping conditions. (C–E) Raman spectra of original MDC, MDC in the charging state, and MDC in the discharging state. MDC, MOF-5-derived carbon; SEM, scanning electron microscopy.

To further analyze the difference in deposition behaviors, the Zn²⁺ flux distribution of the bare Cu foil and the MDC–Cu electrode during deposition was numerically simulated. The conventional planar configuration of the Cu foil and the inherent characteristics of uneven surface often lead to a single nucleation site, and the flux of Zn²⁺ near the prominent nuclei is obviously enhanced, resulting in an uneven flux distribution.³⁸ Zn⁰ deposition “hot spots” (small isolated red areas in the blue background) inevitably occur near each independent nucleation site, resulting in considerable local Zn⁰ accumulation and the formation of dendrites (Figure 5A). However, after the MDC coating is done on the surface of bare Cu, the deposition of Zn⁰ can be confined in the cavity, instead of preferentially depositing on the peak of Zn⁰ protrusion. Therefore, there is no deposition “hot spot” near the protruding Zn⁰ core, which indicates that the uniform Zn⁰ deposition inhibits the formation of zinc dendrites (Figure 5B). In addition, the magnitude of Zn²⁺ flux is enhanced and homogenized at the same time, which proves that the uniformly distributed zinc storage sites and the site size matching with the size of the Zn⁰ core can maximize the spatial confinement effect, thus promoting the directional deposition of Zn⁰.^39–41

Figure 5. Numerical simulation of the Zn2+ flux distribution during deposition (A) on the Cu foil and (B) on the MDC–Cu electrode. (C) Simulation of current density distribution in MDC–Cu. (D) Simulation of the electric field distribution of MDC–Cu and a planar bare Cu foil. MDC, MOF-5-derived carbon.

To explain the spatial confinement effect of MDC theoretically, the electric field distribution and local current density of the MDC coating were further simulated using the finite element method. As shown in Figure 5C, the construction of a 3D porous MDC coating significantly improves the specific surface area of the electrode, provides high-speed channels for ions and electrons, and can effectively reduce the current density (from 1.4 mA cm⁻² down to 0.1 mA cm⁻²) and buffer volume change. According to Sand's time law,⁴² decreasing the local current density will delay the nucleation and growth rate of zinc dendrites, thus inhibiting the formation of “dead zinc” and the expansion of internal polarization in the process of repeated cycles. As we all know, once the Zn⁰ is heterogeneously nucleated on the Cu foil, the local electric field near the nucleation point will be enhanced, which tends to drive more Zn²⁺ nucleation (Figure 5D). The emergence of a large number of independent nucleation points will promote the uneven distribution of the electric field, showing a high-intensity gradient. On the contrary, MDC with a continuous conductive skeleton has a relatively more uniform electric field distribution and Zn⁰ nucleation sites, which is mainly due to the charge redistribution effect of the 3D conductive skeleton.^43–45

To confirm the feasibility of the MDC–-Cu electrode in practical applications, zinc metal batteries were assembled by coupling a zinc-containing cathode (Zn/Mn-MOF@CNT) with 25 kinds of MDC–Cu electrodes (Figure 6A). Zn/Mn-MOF@CNT belongs to the typical ZIF-8 structure, showing a uniform morphology of close-packed particle, which can provide a fresh zinc source for the system to maintain the normal operation of the full battery (Figure S9). Figure 6B shows the rate performance of the full cell based on MDC-3:1 at different annealing temperatures. Here, different annealing temperatures will lead to different residual zinc contents within the MDC, as shown in Figure 2A–C. When the annealing temperature was 850°C, the average discharge capacities of the full battery at 0.5, 0.7, 1, 1.5, 2, 2.5, and 3 C were 459.9, 437.5, 398.8, 345.9, 250.5, 167.9, and 132.7 mA h g⁻¹, respectively. The capacity retention rate was 99.9% on reverting to 0.5 C, which proves that our material has good rate performance. The rate performance of MDC under other conditions is shown in Figure S10A–D, and compared in Figure 6C and Figure S11, with detailed performance summarized in Tables S3–S9. It can be seen that the 3:1-850 MDC showed excellent and stable rate performance, which may be because the 3:1-850 MDC not only retains sufficient Zn to make up for the zinc loss during cycling but also leaves enough zinc storage sites to accommodate the migrated Zn²⁺. The voltage hysteresis upon cycling (defined as the voltage gap at semi-discharge capacity) is an important index to investigate the reversibility of the battery. Through the galvanostatic charging and discharging voltage curves at 1 C (Figure 6D), we can see that the voltage hysteresis in the MDC-based cell (637 mV) is lower than that of the bare copper-based cell (747.8 mV), which indicates that the polarization degree of the MDC battery is lower. As shown in Figure 6E and Figure S13, the high reactivity and small polarization of the full cell can be attributed to the low charge-transfer resistance of MDC, which is fitted according to the equivalent circuit diagram as shown in Table S10. The specific resistance values are summarized in Table S10.

Figure 6. (A) Full battery diagram. Electrochemical performance of a lean-Zn anode zinc-ion battery. (B) Rate performance of 3:1MDC–Cu under 5 different annealing conditions. (C) Capacity distribution range of MDC under different process parameters at a 1 C rate. (D) Constant current charge–discharge curve. (E) Electrochemical impedance of 3:1MDC–Cu under five different annealing conditions. (F,G) Cycle performance at 1 and 5 C rates. MDC, MOF-5-derived carbon.

We defined the capacity retention below 60% as a cut-off parameter and compared the cycling performance of different full cells. As shown in Figure 6F, in contrast to the short-circuit failure of bare copper electrodes after 93 cycles, both 3:1-850 MDC (Zn⁰ content: 31%) and 1.5:1-1000 MDC (Zn⁰ content: 0%) show a higher discharge specific capacity, larger CE, and extended cycle life at a cycling rate of 1 C. The capacity retention of 3:1-850 MDC after 200 cycles and 350 cycles was 93.23% and 67.73%, respectively. The specific performance comparison is shown in Table S11. The contact angle test (Figure 7D and Figure S12D) shows that the electrolyte wettability of MDC electrodes and the electrolyte is better than that of a bare Cu electrode, which allows MDC electrodes to effectively retain the electrolyte and enhance the uniform distribution of Zn²⁺ at the interface.³⁸ However, the full cells based on 1.5:1-1000 MDC, 1:1-1000 MDC, and 1:1-950 MDC were short-circuited after 200 cycles, 183 cycles, and 166 cycles, respectively (Figure 6F and Figure S12A). The relatively poor cycling performance of 1.5:1-1000 MDC, 1:1-1000 MDC, and 1:1-950 MDC could be attributed to the following two reasons: (1) the electrolyte can easily penetrate the electrode, further exacerbating the contact between free water and dissolved oxygen inherent in the electrolyte and corroding the electrode to cause a short circuit,^46–48 which was confirmed by the lower corrosion potential in the Tafel curve (Figure S12C). (2) There is no Zn⁰ reservoir in 1.5:1-1000 MDC, 1:1-1000 MDC, and 1:1-950 MDC, i.e., the as-constructed full cells are anode-free cells, which cannot replenish the irreversible zinc loss in time and will lead to carbon skeleton collapse and irreversible deformation during repeated charge and discharge cycling.²⁴ In addition, we further tested the cycling performance at high rates. As shown in Figure 6G and Figure S12B, the full battery based on 3:1-850 MDC can stably operate at a 3 C rate for 1332 cycles, and the maximum specific capacity reached as high as 200 mA h g⁻¹. At a 5 C rate, capacity retention of 60% was also achieved after 1420 cycles.

Figure 7. SEM images of the MDC anode surface of the 500th charge–discharge cycle of the battery: (A) original MDC anode surface, (B) MDC anode surface after the 500th charge, and (C) MDC surface after the 500th discharge. (D) Contact angle contrast. (E) Capacity contribution ratio of cathode, anode, and electrolyte in the first three cycles. MDC, MOF-5-derived carbon; SEM, scanning electron microscopy.

The MDC after the 500th charge and discharge of the full battery was characterized by SEM (Figure 7A–C). The results show that the platelet-shaped Zn⁰ deposits were distributed uniformly on the surface of MDC during the battery charging process, and the original morphology of MDC was restored in the subsequent discharge process. Uniform Zn⁰ deposition and minimal shape change contribute to maintaining stable and long-term battery cycling without failure.⁴⁹ Finally, to explore the individual capacity contribution from the cathode, the anode, and the electrolyte, we assembled three different full batteries using Zn/Mn-MOF@CNT and pure CNT as cathodes and MDC–Cu as the anode, where the amounts of electrolyte were fixed at 40 μL and 80 μL. From the calculation, it is concluded that in the first charge–discharge cycle, the capacity contribution rate of the cathode, the anode, and the electrolyte was 59%, 19.2%, and 21.8%, respectively (Figure 7E). In the second cycle, the capacity contribution rate of the cathode remained basically unchanged (around 59.2%), while the electrolyte's contribution decreased to 17.6%. Simultaneously, the capacity contribution rate from the anode increased to 23.2%. This may be due to the fact that the residual Zn⁰ in MDC could be activated and participated in the subsequent cycle as a supplementary zinc source. In the third cycle, the capacity contribution rate of the cathode decreased slightly to 51.9%, while the capacity contribution rate of electrolyte decreased rapidly to 7.8%, which was due to the irreversible Zn²⁺ loss caused by suppressed ion migration kinetics and inevitable side reactions at the solid–liquid interface.^50–52 Meanwhile, the capacity contribution rate of the anode reached 40.3%, which was because most Zn⁰ in MDC was used in the battery cycling after a sufficient electro-activation process. Because there is no excessive bulk Zn⁰ source on the anode side, the residual Zn⁰ in MDC, which is the backup zinc source, is one of the main reasons for prolonging the reversible cycle life of the battery.

CONCLUSION

In summary, for the first time, we report a lean-Zn and hierarchical anode material for zinc metal batteries with low N/P ratio. The innovative anode structure shows a large number of active sites and abundant ionic migration channels that jointly homogenize the ionic flux and electric field distribution, thus stabilizing Zn plating/stripping with high reversibility. The full cell with low N/P ratio delivers a high reversible capacity (398.8 mA h g⁻¹ at 1 C) and superior cycling stability (92% capacity retention after 900 cycles at 3 C). This work offers a simple and effective solution for practical high-capacity and durable Zn metal batteries. The hierarchical lean-Zn anode design strategy will inspire more practical designs for a host of aqueous metal anode batteries that also have the disadvantages of poor metal anode reversibility and low anode utilization.

ACKNOWLEDGMENTS

This work was supported by the National Natural Science Foundation of China with grant No. 21905304 and 52073305; the Natural Science Foundation of Shandong Province with grant No. ZR2020QE048; and the State Key Laboratory of Heavy Oil Processing with grant No. SKLHOP202101006.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.