1 Introduction

Due to the unique benefits of Zn metal anodes, such as their high abundance, aqueous solution reaction environment, suitable electrochemical potential (-0.762 V vs SHE), high volumetric capacity (5855 mA h cm−3), and specific capacity (820 mA h g−1), aqueous zinc-ion batteries (AZIBs) have recently attracted a lot of attention [1,2,3,4]. However, Zn anodes face the challenge of dendrite growth, which can lead to low coulombic efficiency (CE), poor cycling stability, and short-circuiting [5,6,7,8]. In addition to this, side effects (hydrogen precipitation), zinc corrosion, and formation of irreversible discharging substances during charging and discharging have seriously hindered the practical application of zinc anodes [9,10,11]. Worse still, dendrite formation and side reactions are mutually supportive during electrochemical reactions: side reactions change the distribution of electric fields and ions on the surface, resulting in new dendrite growth sites; localized dendrite growing areas have higher loading concentrations and more active sites, which can catalyze the activity of side reactions. Resolving zinc dendrites is an inevitable process for large-scale applications of rechargeable batteries, so proposing an effective strategy is a critical step.

So far, various strategies, such as the structural design of Zn metal anode [12,13,14,15], utilization of electrolyte additives [16,17,18] and incorporation of interfacial protective layers [19,20,21], have been proposed to optimize and stabilize the Zn metal anode. Among them, introducing a functional protective layer is considered the most straightforward and efficient way to improve the reversibility of Zn anode. By simply coating Zn anode with different materials, the direct contact between Zn and the aqueous electrolyte could be avoided, therefore alleviating the side reactions, such as corrosion and hydrogen evolution. More importantly, uniform deposition of Zn could be achieved by homogenizing the interfacial electric field or ion distribution with electron or ion-regulating materials, leading to dendrite-free Zn anodes [22]. For example, inorganic materials like CaCO3 [23], ZrO2 [24], kaolin [25], Zn-based montmorillonite [26], and Mg–Al layered double hydroxide [27], etc., have been demonstrated to prolong the cycling life of Zn anode, by producing uniform zinc ion flux on the surface of Zn metal [28,29,30]. In particular, fast ion conductor, NaTi2(PO4)3 (NTP) [31] has been reported as a solid-state electrolyte protection layer to enable internal transport/mobility of Zn2+ like an “ion passable fence”. The NTP layer prevents the zinc dendrites and exhibits superlong cycling life, which suggests that the fast ion conductor could be a promising material for the interfacial engineering of Zn metal anode. However, the work evolving fast ion conductors for Zn anode protection is still limited and needs to be further explored in stabilizing the Zn metal anode.

Inspired by the application of NTP, herein, a fast ion conductor Na5YSi4O12 (NYSO) was developed as a protection layer to promote the electrochemical properties of the Zn metal anode. With an NYSO coating of 20.3 µm, the Zn metal anode demonstrates an extended life span of 1896 h at the current density of 0.5 mA cm−2. In addition, the NYSO@Zn anode presents a favorable rate performance at a high current density of 10 mA cm−2. The superior reversibility of NYSO@Zn anode could be attributed to the elimination of Zn dendrites and side reaction by-products ZnSO4(OH)6·5H2O. Therefore, an aqueous NYSO@Zn|CNT@MnO2 full cell delivers superior capacity and stability to that of a bare Zn battery system.

2 Results and discussion

The schematic crystal structure of Na5YSi4O12 is shown in Fig. 1a, which is composed of SiO4 tetrahedra linked to form puckered Si12O36 rings stacking to form large rigid columns parallel to the c axis. Na5YSi4O12 is synthesized as in previous reports, and the detailed synthesis process is presented in the experimental section. The X-ray Diffraction (XRD) pattern shows that all the diffraction peaks are readily assigned to Na5YSi4O12 (JCPDS Card No. 32–1204) (Fig. 1b). The SEM image of Na5YSi4O12 (Fig. 1c) presents the morphology of the micro-sized bulk material. The Na5YSi4O12 coating layers are successfully prepared on the Zn metal anode by using a controllable thickness coater, here a high-energy ball mill is also used to reduce the particle size. The electrochemical performance of the bare Zn and NYSO@Zn electrodes with different coater thicknesses of 20 μm, 50 μm, 100 μm and 200 μm are investigated in the symmetric cell by repeated plating/peeling measurements at 1 mA cm−2 (Fig. 1d). After 130 h of stripping/plating, a sudden voltage drop was detected, which is attributed to a short circuit caused by the Zn dendrites, resulting in failure for the bare Zn cell. In contrast, all NYSO@Zn symmetric cells could sustain repeated deposition/dissolution processes without obvious significant fluctuations over 200 h, indicating improved anode stability. As displayed in Figure S1, NYSO@Zn symmetric cells show high dependence on the NYSO thickness and exhibit the smallest polarization voltage of 28 mV when the coater thickness is set to 50 μm. The lowest charge transfer resistance is also observed in the same coater thickness of 50 μm, suggesting efficient interfacial kinetics, which may promote a relatively uniform metal plating process (Fig. 1e). The scanning electron microscopy (SEM) image of Na5YSi4O12 layer with coater thickness of 50 μm shows a uniform and dense morphology of Na5YSi4O12 (Fig. 1f). The uniform distribution of the Na, Y, Si and O elements in energy-dispersive X-ray spectroscopy (EDS) mapping further confirms the compact coverage of Na5YSi4O12 covering the Zn anode (Fig. 1g-j). The cross-sectional SEM imaging shows a real thickness of 20.3 μm when the coater thickness is 50 μm, due to the evaporation of solvents (Fig. 1k).

Fig. 1
figure 1

a Schematic crystal structure, b XRD pattern, and c SEM image of the as-prepared Na5Ysi4O12 sample; d Galvanostatic cycling performances of bare-Zn and different thickness NYSO (i.e., 20 μm, 50 μm, 100 μm, and 200 μm) symmetrical cells at 1 mA cm−2 with 1 mAh cm−2. e Nyquist plots of NYSO symmetrical cells with different thickness. f SEM image of Na5Ysi4O12 and corresponding element distribution of g Na, h Y, i Si, and j O. k The cross-sectional SEM of optimal 50 μm-NYSO@Zn electrode

Full size image

To evaluate the anode stability, the long-term electrochemical performance of the bare Zn and NYSO@Zn symmetric cells were tested at different current densities of 0.5, 1, 2 and 5 mA cm−2 for 0.5 mAh cm−2 respectively. Cycled at 1.0 mA cm−2 with 0.5 mA h cm−2, the short-circuit induced premature failure occurs for the symmetric cell with bare Zn after 500 h. By contrast, the NYSO@Zn electrode shows ultralow voltage hysteresis (≈30 mV) with negligible voltage oscillation and ultralong lifespan of 1896 h (Fig. 2a). Under an elevated current density of 1 mA cm−2, the NYSO@Zn symmetric cell shows prolonged cycling life of 1270 h, which is 9 times that of the bare Zn anode (Figure S2). Besides, comparing the initial surface, the bare Zn suffers deep corrosion and chaos dendrite formation after 100 cycles, whereas the NYSO@Zn maintains even surface morphology, much attributing to the NYSO layer’s ability to restrain the attack of H2O and realize even Zn2+ transport flux (Figure S3). The bare Zn symmetric cells experience more rapid failure and show short circuits at cycling times less than 100 and 25 h in 2 and 5 mA cm−2, which might be due to the rampant Zn dendrites growth at high current densities. However, the NYSO@Zn anodes display smaller overpotentials and stable cycle life up to 1418 and 385 h at 2 and 5 mA cm−2 as shown in Figs. 2b and S4. In addition, the enlarged voltage–time curves of NYSO@Zn symmetric cells deliver a smaller polarization than those of bare Zn at 0.5 and 2 mA cm−2 (Figure S5). These results indicate that Na5YSi4O12 coating can achieve uniform Zn2+ deposition to avoid dendrite formation and extend battery life. In addition, to explore the electrochemical performance of symmetric cells at high current densities, the rate performance was carried out from 0.2 to 10 mA cm−2 with 1 mAh cm−2 (Fig. 2c). As expected, the NYSO@Zn anode presents a favorable rate performance with smaller voltage hysteresis at a high current density of 10 mA cm−2, whereas the bare-Zn anode shows severe short-circuit at 1 mA cm−2. The superior cycling life and rate capability can be ascribed to the relatively high Zn2+ ionic conductivity of the Na5YSi4O12 coating with a controllable interface (Figure S6). As calculated from the Nyquist plot (Fig. 2e), the charge transfer resistance of the NYSO@Zn electrode in the initial cycle was obviously lower than that of the Zn foil, indicating an efficient charge transfer kinetics. Upon cycling, the NYSO@Zn electrode exhibits similar impedances after 50 cycles, indicating stable interfacial Zn+ conduction. However, a drastic decrease in charge resistance is observed in the bare Zn cell, suggesting a fundamentally altered charge-conduction mechanism called soft shorts [32] (Fig. 2d). In this case, the Zn dendrites cause the small localized electrical connection between two electrodes, which allows the co-existence of direct electron transfer and interfacial reaction, resulting in rapid cell degradation.

Fig. 2
figure 2

Long-term cyclability of the bare-Zn and NYSO@Zn symmetric cells at a 0.5 mA cm−2/0.5 mAh cm−2 and b 2 mA cm−2/0.5 mAh cm−2. c Rate performances of the Zn symmetric cells from 0.2 to 10 mA cm−2. Nyquist plots of d bare-Zn and e NYSO@Zn symmetric cells during pristine and various cycling durations

Full size image

Coulombic efficiency (CE), the ratio of the stripping capacity and the plating capacity of the same cycle, is also measured in Cu|bare Zn and Cu|NYSO@Zn cells to evaluate the utilization and sustainability of Zn [33]. Specifically, the CE measurement is conducted by plating Zn onto a Cu substrate with a capacity of 1.0 mA h cm−2 at 1.0 mA cm−2, and then stripping to 0.3 V (Fig. 3a). The Cu|bare Zn cell short-circuit after less than 75 cycles and shows fluctuating CE around 50 cycles, implying severe side reactions on the bare Cu substrate during cycling. For the Cu|NYSO@Zn cell, the CE is remarkably stable up to 138 cycles with a high CE of 99.91%. Indeed, the CE stability of cells with Na5YSi4O12 protected anodes is significantly enhanced. The corresponding galvanostatic charge/ discharge (GCD) profiles of the bare Zn and NYSO@Zn anodes at 1 mA cm−2 are illustrated in Fig. 3b, c. It is found that bare Zn shows a large voltage gap of 73 mV, which falls to 40 mV after 70 cycles. In contrast, lower polarizations of 63 mV at the 1st cycle and 33 mV at the 70th cycle were observed, which confirms that NYSO@Zn could avoid the side reactions and maintain reversible and fast Zn2+ kinetics. Apart from this, the nucleation overpotential is further measured to highlight the effect of Na5YSi4O12 surface coating on Zn deposition. As shown in Figure S7, the nucleation overpotential of bare-Zn reaches as high as 62 mV, higher than that of NYSO@Zn (49 mV) at a current density of 1 mA cm−2. This result indicates that the surface of NYSO coating could prevent the growth of Zn dendrites and modify the homogeneous nucleation with a lower overpotential. The enhanced reaction kinetics with the Zn-Cu asymmetric cells is also discussed by CV testing of NYSO@Zn anode (Fig. 3d). The NYSO@Zn anode shows a higher current, indicating that more active sites on the Zn anode surface achieved faster reaction kinetics, which is in good agreement with the EIS result [34, 35]. The diffusion of Zn2+ on the Zn anode continues to deposit continuously at the most favorable nucleation position, and ultimately leads to dendrite forming. The current–time curve may actually reflect the morphological change on the Zn anode along with the electrodeposition process. Therefore, the chronoamperometry (CA) test in Fig. 3e was adopted to analyze the current change of bare Zn and NYSO@Zn anode under -150 mV overpotential within 100 s. The current of bare Zn exhibits a continuous increase within 100 s, indicating the uncontrollable deposition of Zn2+. The adsorbed Zn2+ continuously two-dimensional diffuses on Zn anode to achieve the minimization surface energy, and always deposit on the preferential nucleation position, resulting in the growth of dendrite. Different from bare Zn, the NYSO@Zn anode presents a stable and smaller current in a short time. With the assistance of the NYSO@Zn coating, two-dimensional diffusion is regulated during the process of deposition, so that the Zn2+ is uniformly deposited on the Zn anode [36, 37].

Fig. 3
figure 3

a Coulombic efficiency and corresponding voltage–time curves of b Cu|bare Zn and c Cu|NYSO@Zn half-cells at 1 mA cm−2. d CV and e CA curves of Zn nucleation of bare-Zn and NYSO@Zn electrodes at Cu|Zn half-cells and Zn symmetric cells, respectively. f XRD pattern of bare-Zn and NYSO@Zn after immersion in 2 M ZnSO4 electrolyte for 7 days and g, h corresponding evolution of surface morphology

Full size image

To investigate the protective effect of Na5YSi4O12 coating on Zn metal, we further observe the morphological evolution by SEM and XRD for the bare-Zn and NYSO@Zn in 2 M ZnSO4 electrode after 7 days respectively. For the bare-Zn, newly generated flake by-products, like ZnSO4(OH)6·5H2O (Fig. 3f), are found on the surface (Fig. 3g). In comparison with the bare-Zn, the overall morphology of NYSO@Zn remains as is (Fig. 3h), though there are still some minor by-products. In addition, the change of the electrode surface during electrochemical plating of bare-Zn and NYSO@Zn are recorded directly under in-situ optical microscope images. After 5 min of plating at 20 mA cm−2, several protrusions appear on the surface of bare-Zn, indicating a non-uniform Zn plating behavior and the formation of dendritic growth, which becomes more pronounced with increasing reaction time (Fig. 4g). In contrast, the NYSO@Zn electrode shows no significant surface changes even after 30 min, demonstrating the efficiency of the Na5YSi4O12 layer in inhibiting dendrite growth (Fig. 4h).

Fig. 4
figure 4

In-situ optical microscopy images of Zn plating at a bare-Zn and b NYSO@Zn symmetric transparent cells under current density of 20 mA cm−2. Scale bars: 100 μm

Full size image

The practical use of NYSO@Zn anodes in aqueous ZIBs was further evaluated in full cells. Carbon nanotube (CNT)@MnO2 composite was used to pair with bare Zn and NYSO@Zn due to the high theoretical capacity and low cost of MnO2 and the high electronic conductivity of CNT. The X-ray Diffraction (XRD) pattern in Figure S8 shows that all the diffraction peaks match well with MnO2 (JCPDS PDF 44–0141). As shown in Figure S9, SEM images show a nanotube morphology of CNT@MnO2 composite. The thermogravimetric analysis (TGA) of the CNT@MnO2 composite exhibits a 21.15% weight loss in the temperature range from 220 to 700 °C, suggesting a 21.15% weight ratio of CNT (Figure S10). The cyclic voltammetry (CV) profiles of bare Zn/MnO2 and NYSO@Zn|CNT@MnO2 full cells are illustrated in Fig. 5a within a voltage window of 0.8 to 1.8 V. The two pairs of redox peaks are assigned to the Zn2+ insertion and extraction in the CNT@MnO2 cathode. In the NYSO@Zn|CNT@MnO2 cells, CV curves demonstrate higher current responses of the redox peaks than that in the bare Zn|MnO2 cells, implying fast kinetics of Zn deposition/dissolution at NYSO@Zn anode. Figure 5b displays the rate performances of bare Zn|CNT@MnO2 and NYSO@Zn|CNT@MnO2 full cells at various current densities. In NYSO@Zn|CNT@MnO2 cells, discharge capacities of 250 mAh g−1, 224 mAh g−1, 176 mAh g−1, 123 mAh g−1, 107 mAh g−1 and 94 mAh g−1 were recorded at 0.2 A g−1, 0.5 A g−1, 1 A g−1, 2 A g−1, 3 A g−1 and 5 A g−1, respectively. When the current density returns to 0.2 A g−1, NYSO@Zn|CNT@MnO2 cell still retains a capacity of 280 mA h g−1. The superior rate capability of NYSO@Zn than bare Zn further confirms the rapid Zn2+ stripping and plating at the NYSO@Zn interface. Consequently, NYSO@Zn|CNT@MnO2 cell shows a higher capacity than bare Zn and a long cycling life of 660 cycles at 1 A g−1 (Fig. 5c).

Fig. 5
figure 5

The electrochemical performance comparisons of bare-Zn|CNT@MnO2 and NYSO@Zn |CNT@MnO2 full batteries. a CV curves, b rate performance from 0.2 to 5 A g−1, and c long cycling performance and corresponding coulombic efficiency at 1 A g.−1

Full size image

3 Conclusions

In summary, we developed a Na5YSi4O12 protection layer to inhibit the growth of zinc dendrites and stabilize the Zn metal anode. The 20.5 μm thickness of NYSO@Zn anode exhibits a superior lifespan over 1896 cycles at 0.5 mA cm−2 in symmetrical batteries. Meanwhile, the NYSO@Zn anode presents a favorable rate performance at a high current density of 10 mA cm−2. Additionally, the Na5YSi4O12 coating can successfully reduce the generation of by-products and prevent dendrites growth on the Zn metal anode/electrolyte interface. Consequently, this simple strategy enables the NYSO@Zn|CNT@MnO2 cell with a higher capacity and stability than bare Zn.