Introduction

In nature, liquids tend to reduce their surface area, analogous to a stretched elastic membrane, a phenomenon termed “surface tension”1,2. In microgravity environments, water droplets spontaneously adopt a spherical shape to significantly minimize their surface energy. Similarly, crystal grains within metallic negative electrodes exhibit varying growth rates3,4, forming irregular and disordered deposition morphologies, such as dendrites, due to differences in surface energy. This characteristic poses a significant challenge in practical battery applications because dendritic electrodeposits can proliferate and bridge between electrodes, resulting in short circuits and battery failure5,6,7.

According to the Gibbs-Curie-Wulff crystal growth theory, under isothermal and constant-volume conditions, the equilibrium morphology of a crystal is determined by the minimization of its total surface energy8,9. The linear growth rate of a crystal face family is directly proportional to its specific surface free energy: \({\sigma }_{i}\)/\({r}_{i}\) = constant, where \({\sigma }_{i}\) represents the specific surface free energy of the crystal face \(i\), and \({r}_{i}\) denotes the central distance from the equilibrium-formed crystal10,11. It can be succinctly stated that crystals with lower surface energy exhibit slower growth rate and become the primary exposed facets12. Extending the Gibbs-Curie-Wulff theorem to metal negative electrodes, during battery charging, grains with lower surface energy on the electrode surface exhibit significantly slower growth rates than those with higher surface energy, making them more susceptible to exposure and retention.

Tailoring the crystallographic features of metal negative electrodes to expose the close-packed crystal planes with the lowest surface energy has been recognized as an effective strategy in pursuit of highly reversible metal negative electrodes. This includes (110)-textured lithium (Li, body-centered cubic (BCC))13,14,15, (111)-textured aluminum (Al, face-centered cubic (FCC))16,17, and (0002)-textured zinc (Zn, hexagonal closed-packed (HCP))18,19,20. Horizontal migration of metal atoms predominantly occurs on these crystal planes due to their lower migration barrier21. Furthermore, these planes exhibit higher thermodynamic and electrochemical stability toward the electrolyte, contributing to suppress side reactions14,22. However, recent studies have indicated that textured metal negative electrodes with high-surface-energy crystal planes, such as (002)-textured Li23 and (10\(\bar{1}\)0)-textured Zn24,25, can still achieve uniform deposition morphology, contradicting the notion established in conventional metal negative electrodes. This discrepancy stems from the observation that although low-surface-energy crystal planes effectively suppress side reactions, they also result in sluggish plating rates and significant stripping polarization26,27. Consequently, despite advancements in crystallographic optimization and mechanistic comprehension, elucidating atom-level electrocrystallization behavior and establishing criteria for texture modulation remain challenging.

In this study, a single [11\(\bar{2}\)0]-oriented Zn (Zn(11\(\bar{2}\)0)) metal was selected to investigate its morphology and reversibility as a negative electrode in the battery system. Two important considerations in choosing the Zn(11\(\bar{2}\)0) metal include: (i) the low reactivity of Zn allows for focused exploration of atomic-scale processes using advanced characterization techniques28; (ii) the higher surface energy of the Zn(11\(\bar{2}\)0) crystal plane29, compared to other reported Zn crystal planes, including Zn(0002), Zn(10\(\bar{1}\)1), and Zn(10\(\bar{1}\)0), suggests a higher growth rate and lower thermodynamic stability, rendering it a representative substrate for investigating the influence of metal negative electrode orientation on electrocrystallization behavior. The electrodeposited Zn(11\(\bar{2}\)0) metal negative electrode, characterized by grains with consistent crystallographic orientation, forms a stable and well-ordered grain boundary (GB) network to effectively suppress side reactions and reduce local stress. Aberration-corrected scanning transmission electron microscopy (STEM) reveals that Zn deposits maintain the same crystalline orientation as the Zn(11\(\bar{2}\)0) substrate along the growth direction (GD), indicating perfect lattice alignment at the epitaxial interface. As expected, the Zn(11\(\bar{2}\)0) negative electrode demonstrates high reversibility over thousands of cycles even at elevated current densities. This concept can be extended to other oriented Zn negative electrodes, illustrating that any Zn metal negative electrode with a single crystallographic orientation guarantees uniform, dendrite-free deposition morphology through its dislocation-free homoepitaxial growth mechanism.

Results

Preparation and characterization of Zn metal negative electrodes

In HCP-structured Zn crystals, the surface energies of the stoichiometric Zn planes follow the sequence (0002) < (10\(\bar{1}\)1) < (10\(\bar{1}\)0) < (11\(\bar{2}\)0)30,31, suggesting that the (11\(\bar{2}\)0) crystal planes are prone to induce rapid crystal growth in their perpendicular direction (Supplementary Fig. 1a, b). According to the Bravais rule and Gibbs-Curie-Wulff principle, planes with lower crystal plane density and higher specific surface energy tend to diminish or even vanish in the final structure (Supplementary Fig. 1c)21,32. Therefore, the high-specific-surface-energy (11\(\bar{2}\)0) crystal planes are challenging to expose.

Typically, the lowest-surface-energy planes are predominantly exposed from a ZnSO4 electrolyte by suppressing side reactions during the electrodeposition process (Supplementary Fig. 2)20. Here, uniaxially oriented Zn metal negative electrodes, characterized by an ultra-strong (11\(\bar{2}\)0) texture, were prepared using a direct-current electrodeposition technique in a ZnSO4 electrolyte incorporating polyacrylamide (PAM; Supplementary Fig. 3) polymer33. During the deposition process, Zn nuclei rapidly covered the copper (Cu) substrate, followed by the adsorption of PAM polymer onto the deposited metal layer, which significantly reduced the surface energies of all Zn crystal planes (Supplementary Fig. 1b). The predicted equilibrium morphology of a Zn crystal after the adsorption of PAM polymer along the [0001] direction is depicted in Supplementary Fig. 4b. In contrast to the Wulff construction of the ground-state Zn crystal (Supplementary Fig. 4a), the proportion of the (11\(\bar{2}\)0) crystal plane increases from 11% to 28%, demonstrating the highest growth tendency among all crystal planes. Additionally, the (11\(\bar{2}\)0) plane exhibits a significantly smaller interlayer spacing of 0.133 nm and stronger bonding energy between its planes compared to the (0002), (10\(\bar{1}\)1), and (10\(\bar{1}\)0) planes (Supplementary Fig. 5 and Supplementary Table 1)34. Consequently, Zn deposition perpendicular to the (11\(\bar{2}\)0) plane, that is, along the [11\(\bar{2}\)0] direction, is more likely to penetrate the surface-adhered PAM membrane, leading to preferential generation of the (11\(\bar{2}\)0) plane.

The microstructure of the electrodeposited Zn metal negative electrodes was systematically analyzed using X-ray diffraction (XRD), scanning electron microscopy (SEM), precession electron diffraction (PED), and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) characterizations. The ~120-μm-thick electrodeposited Zn metal negative electrode displays a reflective and smooth surface morphology (Supplementary Fig. 6). In contrast to the multiple peaks observed in commercial bare Zn (Supplementary Fig. 7), the XRD pattern of the electrodeposited Zn metal presents a single diffraction peak located at 70.7o (Fig. 1a), corresponding to the (110) characterized peak of a Zn crystal35, indicating its single lattice orientation. The orientation image maps (OIMs) show a variety of colors for the bare Zn, suggesting randomly oriented grains (Supplementary Fig. 8a–c). The corresponding pole figures confirm that no strong texture can be detected in bare Zn (Supplementary Fig. 8d). Instead, the consistent color in the OIMs for the electrodeposited Zn from the Z-axis view (Fig. 1b), as further confirmed by the {11\(\bar{2}\)0} pole figure, indicating a single [1\(1\bar{2}\)0] orientation. The as-prepared Zn(11\(\bar{2}\)0) metal has an average grain size of 168 ± 52 nm (Fig. 1c) and low residual stress (10.2 MPa) compared to bare Zn (25.58 ± 0.55 μm and 22.8 MPa, Supplementary Fig. 9), indicating minimal lattice strain and fewer misfit dislocations36. Moreover, the Zn(11\(\bar{2}\)0) metal exhibits superior resistance to deformation compared to the bare Zn and Zn(0002) metals (Supplementary Table 2). The bright-field transmission electron microscopy (BF TEM) image from plane-view (Fig. 1d) shows that the grains are nanocrystals approximately 100-200 nm in size. The selected area electron diffraction (SAED) pattern (Fig. 1e) and high-resolution image (Fig. 1f) of the black grain highlighted in yellow reveal an ~ABABAB~ stacking sequence of (0002) planes. According to the Weiss zone law, the incident direction can be determined to be parallel to the zone axis [11\(\bar{2}\)0] (Supplementary Fig. 10), consistent with the PED results. The density contour in orientation distribution function (ODF) figures represents texture intensity. The Euler angle37 with the strongest intensity is located at (90o, 30o) (\(\phi\),\(\,\varphi\) 2) (Fig. 1g), a \(\{11\bar{2}0\}\) fiber texture can be defined. Accordingly, the electrodeposited Zn metal has only one [\(11\bar{2}0\)]-crystallographic axis, displaying an ultra-strong (\(11\bar{2}0\)) metal texture.

Fig. 1: Microstructural characterization of the electrodeposited Zn metal.
figure 1

a XRD pattern of the obtained Zn metal. The inset image displays the simulated crystallographic structure of Zn. b–g Detailed crystallographic observation of the electrodeposited Zn(11\(\bar{2}\)0) metal: b OIMs and their corresponding pole figures by PED. The texture strength scale bar ranges from 0 to 22.27. c Statistical distribution of grain sizes obtained from BF TEM images. The grain sizes range from 67.5 to 307.5 nm, with a Gaussian distribution yielding an average value of 168 ± 52 nm. d, e A typical BF TEM image from the plane-view (d) and the corresponding SAED pattern (e). f The HAADF-STEM image observed from the plane direction, and the incident crystal zone axis of [11\(\bar{2}\)0]. g ODFs images, showing a < 110> fiber texture.

Full size image

Morphological evolution during electroplating

Note that the surface energy of Zn(0002) crystal plane is significantly lower than that of other Zn crystal planes, indicating higher thermodynamic stability and excellent corrosion resistance35. However, this lower surface energy also suggests that Zn ions are difficult to detach from the crystal plane, resulting in sluggish kinetics behavior at the electrode interface25,26. As shown in Supplementary Fig. 11, the Zn(\(11\bar{2}0\))| |NH4V4O10 cells exhibit lower charge transfer resistance (\({R}_{{ct}}\)) both before and after cycling compared to the bare Zn| |NH4V4O10 cells. In the full-cell electrochemical impedance spectroscopy (EIS) measurement, a 5 mV voltage bias is applied, driving the system towards the Zn → Zn2+ oxidation (stripping) process. Under these conditions, the measured \({R}_{{ct}}\) primarily reflects the resistance associated with Zn dissolution (ion departure). Therefore, the lower \({R}_{{ct}}\) observed on the Zn(\(11\bar{2}0\)) surface can be attributed to its higher surface energy and lower Zn-ion departure energy compared to other crystal planes (Supplementary Fig. 12 and Supplementary Table 3). Additionally, cyclic voltammetry (CV) curves show that the onset potentials of Zn plating/stripping behavior in the asymmetric Zn(\(11\bar{2}0\))||Cu cell are located at −0.0217/ − 0.0144 V, superior to −0.0386/ − 0.0101 V for the bare Zn| |Cu and −0.0244/ − 0.0131 V for Zn(0002)||Cu cells (Supplementary Fig. 13). The relatively large enclosed area and high current intensity indicate the fast charge-transfer process and rapid reaction kinetics38,39. This is further supported by the temperature-dependent Nyquist plots and the corresponding activation energy (Supplementary Fig. 14). These results suggest improved electrochemical performance and favorable Zn plating/stripping kinetics for the Zn(11\(\bar{2}\)0) electrode.

To determine how a uniaxially oriented Zn metal electrode affects Zn deposition behavior, we observed the morphology of Zn deposits on bare Zn and Zn(11\(\bar{2}\)0) electrodes at a current density of 4 mA/cm2 after plating for 5, 10, 20, and 30 min using SEM (Supplementary Figs. 15a–c). On the bare Zn, a porous, moss-like structure composed of by-products and deposited Zn was observed. In contrast, uniform and compact Zn deposition formed on the Zn(11\(\bar{2}\)0) surface. Interestingly, low- and high-magnification SEM images of Zn deposits on Zn(11\(\bar{2}\)0) after 3 h of plating display that the compact Zn deposits are formed by vertically stacked Zn plates (Supplementary Figs. 15d, e). Considering the HCP structure of Zn crystal (Supplementary Fig. 15f)40, the deposited Zn exhibits a lamellar structure, indicative of coherent deposition. Subsequently, the morphologies and surface roughness of deposited Zn metal with a 2.0 mAh/cm2 areal capacity under current densities ranging from 1.0 mA/cm2 to 20.0 mA/cm2 were characterized using a confocal laser scanning microscope (CLSM). In contrast with glass fiber (GF)-wrapped Zn dendrites on bare Zn (Fig. 2a), Zn deposits on Zn(11\(\bar{2}\)0) show relatively flat and densely packed morphologies at all current densities (Fig. 2c). Furthermore, the surface roughness of bare Zn increases from 1.91 μm to 2.55 μm with higher current densities (Fig. 2b), which is derived from the enhanced growth of Zn dendrites under high current density. Electrodes with high surface roughness exhibit numerous protrusions that act as charge centers during subsequent reactions, triggering “tip effects” that intensify the uneven distribution of the electric field on the electrode surface41,42, leading to the accumulation of Zn2+ and eventually causing battery failure43. Regardless of the current density, the surface roughness of the Zn(11\(\bar{2}\)0) consistently remains lower than that of the bare Zn. Notably, under a high current density of 20.0 mA/cm2, its surface roughness of 1.43 μm is significantly lower than that of bare Zn (2.55 μm). This result was further confirmed by in situ optical microscopy analysis in coin-type Zn symmetric cells with an open configuration (Supplementary Fig. 16). The morphological studies demonstrate that Zn(11\(\bar{2}\)0) can greatly facilitate uniform Zn deposition and lead to a reduction in the surface area available for parasitic reactions between the Zn metal electrode and electrolyte.

Fig. 2: The morphology analysis of Zn deposits.
figure 2

a, c CLSM images of bare Zn (a) and Zn(11\(\bar{2}\)0) (c) electrodes in Zn symmetric cells with a fixed areal capacity of 2.0 mAh/cm2 under different current densities ranging from 1.0 mA/cm2 to 20.0 mA/cm2. The corresponding upper and lower limit values of the colormap are 80 μm and 0 μm, respectively. b The corresponding surface roughness of the Zn electrodes in (a) (blue) and (c) (red).

Full size image

Growth mechanism at the Zn (11\(\bar{{{\boldsymbol{2}}}}\)0) electrode-electrolyte interface

To reveal the atomic structure and growth mechanism of the deposited Zn-Zn(11\(\bar{2}\)0) growth interface, atomic-resolution STEM imaging was conducted. Figure 3a shows a typical cross-sectional TEM image of Zn(11\(\bar{2}\)0), in which a direct and clear coherent interface between the Zn deposits and the Zn(11\(\bar{2}\)0) electrode was observed (Supplementary Fig. 17). The corresponding nano-beam electron diffraction (NBED) patterns of grains selected from Zn(11\(\bar{2}\)0) substrate to the subsequently deposited Zn exhibit consistent diffraction spots along the GD, all of which match the standard electron diffraction pattern of the HCP crystal in the [0001] direction (Fig. 3b and Supplementary Fig. 18), indicating their consistent crystallographic orientation. Specifically, the (11\(\bar{2}\)0) diffraction spots are vertical to the X-axis, indicating that the (11\(\bar{2}\)0) crystal planes grow parallel to the growth interface, due to the 90o rotation angle between the reciprocal space (diffraction spots) and the real space (STEM images). Such atomic-level Zn growth at the epitaxial interface between the deposits and the substrate was intuitively observed using HAADF-STEM. Figure 3c and Supplementary Fig. 19 illustrate the atomic structures of the deposited Zn, Zn(11\(\bar{2}\)0) substrate, and the coherent interface, showing zero lattice mismatch. The crystal planes perpendicular to the GD, in both the Zn substrate and the subsequently grown Zn, exhibit a lattice spacing of 0.1331 nm, confirming that the (11\(\bar{2}\)0) planes are parallel to the growth interface. Across the coherent interface, the epitaxial Zn is bonded directly to the Zn(11\(\bar{2}\)0) substrate without dislocations, indicating no lattice distortion in this region. Figure 3d, e display the fast Fourier transform (FFT) patterns of the high-resolution HAADF-STEM images. The diffraction spot distributions in both patterns are identical, indicating that the Zn deposits share the same orientation as the Zn(11\(\bar{2}\)0) substrate. This confirms that the single crystallographic orientation of the Zn(11\(\bar{2}\)0) substrate enables homoepitaxy without misfit dislocations, facilitating uniform Zn deposition.

Fig. 3: Atomic epitaxial mechanism at the growth interface of Zn(11\(\bar{2}\)0) metal electrode.
figure 3

a Cross-sectional TEM image of Zn grown on a Zn(11\(\bar{2}\)0) substrate. The dashed box outlines the Zn substrate and the growth region near the growth interface. The Zn(11\(\bar{2}\)0) sample of the plating side was obtained in coin-type Zn symmetric cells after plating 2 h at 5.0 mA/cm2. b The corresponding NBED patterns of the labeled Zn grains from (a). c The magnified HAADF-STEM image of the epitaxial interface along the GD. The cross symbol (⊗) is the incident direction. d, e The FFT patterns of the deposited Zn (d) and the Zn(11\(\bar{2}\)0) substrate (e).

Full size image

Ex situ XRD was conducted to detect the structural evolution of Zn electrodes in sealed-type Zn symmetric cells. The multiple peaks located at 36o, 39o, 43o, 54o and 71o, corresponding to (002), (100), (101), (102) and (110) planes of Zn44, respectively, are observed in bare Zn electrode (Supplementary Fig. 20a). During the charging process, an emerging peak at approximately 7o was assigned to the corrosion byproducts Zn4SO4(OH)6·5H2O (ZHS)45. However, throughout the charging and discharging process, the Zn(11\(\bar{2}\)0) electrode consistently displays a single (110) characteristic peak, with only changes in its intensity (Supplementary Fig. 20b). Furthermore, to maintain a single crystallographic orientation, the applied current density for Zn(11\(\bar{2}\)0) electrodes should not exceed 40.0 mA/cm2 (Supplementary Fig. 21). This finding indicates that the uniaxially oriented Zn metal electrodes effectively suppress detrimental side reactions while maintaining homoepitaxial, dendrite-free Zn deposition.

Grain boundary stability between Zn grains

Despite the Zn(11\(\bar{2}\)0) crystal plane exhibiting lower thermodynamic stability and inferior corrosion resistance relative to other Zn crystal planes, the GBs demonstrate significantly higher reactivity and are more prone to undesirable parasitic reactions than the intragranular lattice46,47. This leads to low coulombic efficiencies (CEs), rapid depletion of electrolytes, reduced active Zn inventory and, ultimately, poor cycling life48. Therefore, it is imperative to prioritize the study of GB stability in Zn metal electrodes rather than that of the grain interior.

The calculated GB energies for Zn grains with identical and different crystallographic orientations are shown in Fig. 4a–d and Supplementary Figs. 22a–d. For Zn(10\(\bar{1}\)0), Zn(10\(\bar{1}\)1), and Zn(0002) oriented grains, the GB energies between grains with identical orientations are lower than those between grains with different orientations. Although the GB energy of Zn(11\(\bar{2}\)0)-Zn(11\(\bar{2}\)0) (0.84 J/m2) is slightly higher than that of Zn(11\(\bar{2}\)0)-Zn(0002) (0.74 J/m2), it remains lower than the GB energies of Zn(11\(\bar{2}\)0)-Zn(10\(\bar{1}\)0) and Zn(11\(\bar{2}\)0)-Zn(10\(\bar{1}\)1). According to the Bravais rule (Supplementary Fig. 23), the significant difference in surface energy between Zn(0002) and Zn(11\(\bar{2}\)0) crystal planes greatly reduces the probability of Zn(11\(\bar{2}\)0)-Zn(0002) GB formation. Therefore, these results demonstrate that the uniaxially oriented Zn metals present higher GB stability compared to randomly oriented Zn metals. Scanning Kelvin probe microscopy (SKPM) measurements performed on Zn metal electrodes reveal a lower contact potential difference between grain interiors and boundaries in Zn(11\(\bar{2}\)0) electrodes (0.12 V), compared to bare Zn electrodes (0.18 V), as shown in Supplementary Fig. 24. This finding implies that the GBs in Zn(11\(\bar{2}\)0) electrodes have fewer defects than those in bare Zn electrodes49. To further assess the GB stability of Zn metal electrodes, TEM/SEM-measured distribution maps of GB characteristics were employed to analyze their GB energies (Fig. 4e, g)50,51. Generally, low-angle grain boundaries (GBs) are characterized by a small misorientation angle between grains, typically less than 15°, and exhibit relatively low energy. In contrast, high-angle GBs, with a misorientation angle greater than 15°, possess higher energy52. Figure 4f shows that the proportion of low-angle GBs in Zn(11\(\bar{2}\)0) metal is higher than in bare Zn metal, increasing from 33.6% to 36.3%. Furthermore, compared with bare Zn, high-angle GBs in Zn(11\(\bar{2}\)0) are coincidence site lattice (CSL) GBs (Supplementary Fig. 25), which, despite having large misorientation angles, can exhibit lower energy due to their high degree of lattice coincidence53. Consequently, statistical results suggest that GBs in Zn(11\(\bar{2}\)0) electrodes generally consist of low-angle and CSL GBs, demonstrating higher stability than those in bare Zn electrodes.

Fig. 4: GB stability comparison for Zn grains with identical/different crystallographic orientations.
figure 4

a–d The relaxed structure and calculated GB energies between Zn(10\(\bar{1}\)0) (a), Zn(10\(\bar{1}\)1) (b), Zn(0002) (c), Zn(11\(\bar{2}\)0) (d) oriented grains and Zn(10\(\bar{1}\)0), Zn(10\(\bar{1}\)1), Zn(0002), Zn(11\(\bar{2}\)0) oriented grains, respectively. The atomic coordinates of the optimized structure of the computational models are provided in the Source Datafile. e, g Typical distribution maps of GB characteristics of Zn(11\(\bar{2}\)0) (e) and bare Zn (g) electrodes. Colored lines stand for different boundaries including low-angle GBs (gray) and ordinary high-angle GBs (black). f The corresponding misorientation distributions of bare Zn and Zn(11\(\bar{2}\)0), with the statistical results from (e) and (g).

Full size image

The high GB stability of the uniaxially oriented Zn metal electrodes not only effectively prevents harmful side reactions but also reduces the mismatch of Zn homoepitaxy at the GBs. The Zn(11\(\bar{2}\)0) electrode exhibits a more negative hydrogen evolution reaction (HER) onset potential and a lower current density than the bare Zn electrode (Supplementary Fig. 26). Due to the continuous generation of H2 by HER, cells based on bare Zn expand and detach, leading to electrolyte leakage and early battery failure54, while the HER can be significantly suppressed in Zn(11\(\bar{2}\)0) symmetric cells (Supplementary Fig. 27). Linear polarization curves of Zn electrodes show that the corrosion current density of the Zn(11\(\bar{2}\)0) electrode is greatly reduced to 1.56 mA/cm2 compared to that of bare Zn (3.33 mA/cm2). The corresponding corrosion potential increases from −0.967 V to −0.951 V, indicating that the Zn(11\(\bar{2}\)0) electrode effectively reduces the corrosion rate and mitigates the corrosion process (Supplementary Fig. 28). This is further corroborated by immersion experiments (Supplementary Fig. 29).

Electrochemical performance of Zn metal batteries

The Zn plating/stripping reversibility of Zn electrodes over a wide current density range was evaluated using Zn| |Cu asymmetric cells. At a low current density of 1.0 mA/cm2, a Zn(11\(\bar{2}\)0)||Cu cell shows a relatively steady voltage hysteresis and an average CE as high as 99.4% with excellent cycling stability (over 1500 cycles), whereas the bare Zn| |Cu cell is associated with a low CE of 95.3% and a short cycle life of 115 cycles (Fig. 5a and Supplementary Fig. 30a). The stability of the homoepitaxial growth mechanism in Zn(11\(\bar{2}\)0) electrodes was assessed by measuring CEs in Zn| |Cu cells at higher current densities of 10.0, 20.0 and 40.0 mA/cm2, respectively. Zn(11\(\bar{2}\)0)||Cu cells have a steady CE and a long cycling stability (over 4000 cycles) compared to bare Zn| |Cu cells (Fig. 5b and Supplementary Fig. 30b). Also, the Zn(11\(\bar{2}\)0)||Cu cell has a good electrochemical performance with a high first-cycle CE of 96.1%, a steady CE of 99.5%, a long cycle life and relatively stable voltage hysteresis even at a high current density/areal capacity of 40.0 mA/cm2 and 4.0 mAh/cm2 (Fig. 5c). However, under the same conditions, the bare Zn| |Cu cell works for only 120 cycles with a high voltage hysteresis (Supplementary Figs. 30c, 31), indicating sluggish reaction kinetics and slow mass transport processes at the electrode surface. These results show that the Zn(11\(\bar{2}\)0) electrode exhibits improved cycling stability and high reversibility under both low and high current densities.

Fig. 5: Electrochemical performance of Zn(11\(\bar{{{\bf{2}}}}\)0) metal negative electrodes.
figure 5

a Plating/stripping CE at a low current density of 1.0 mA/cm2. b CE at high current densities from 10.0 mA/cm2 to 40.0 mA/cm2. c CE at high current density of 40.0 mA/cm2 and high areal capacity of 40.0 mAh/cm2. d Long cycling performance at 2.0 A/g. The insets are the SEM images of the cycled bare Zn and Zn(11\(\bar{2}\)0) negative electrodes after 800 and 1000 cycles. Scale bar: 50 µm.

Full size image

The stability and durability of Zn metal negative electrodes were further assessed at a current density of 2.0 mA/cm2 with an areal capacity of 2.0 mAh/cm2 using the Zn| |Zn symmetric cells (Supplementary Fig. 32a). The voltage profiles of the Zn(11\(\bar{2}\)0)||Zn(11\(\bar{2}\)0) cell show a stable overpotential of 43 mV with a cycle life for 2300 h. In contrast, the bare Zn| |bare Zn cell exhibits a high initial overpotential of 210 mV and a limited lifespan for 73 h, followed by significant fluctuations at the final stage due to severe parasitic reactions. Even when the current density and areal capacity are further increased, the cycling stability of Zn(11\(\bar{2}\)0)||Zn(11\(\bar{2}\)0) symmetric cell is superior to that of the bare Zn| |bare Zn cells (Supplementary Fig. 32b, c). After 100 cycles, the surface of the Zn(11\(\bar{2}\)0) electrode remains flat with a dendrite-free morphology compared to bare Zn (Supplementary Fig. 33a, b). Additionally, the Zn(11\(\bar{2}\)0)||Zn(11\(\bar{2}\)0) cell shows superior rate performance than the bare Zn| |bare Zn cell (Supplementary Fig. 34). By comparison, in terms of Zn| |Cu asymmetric and Zn | |Zn symmetric cells, the Zn(11\(\bar{2}\)0) shows good cycling performances among previously reported protective Zn negative electrodes (Supplementary Fig. 35 and Supplementary Table 4). The full-cell performances were further evaluated using the NH4V4O10 positive electrode. Zn(11\(\bar{2}\)0)| |NH4V4O10 full cells achieve a capacity of 354.9 mAh/g at 0.5 A/g and 107.9 mAh/g at 5.0 A/g, showing excellent rate performance than bare Zn| |NH4V4O10 full cells (Supplementary Figs. 36, 37). The Zn(11\(\bar{2}\)0)| |NH4V4O10 full-cell functions with both a high cycling stability and a high CE at low (0.5 A/g) and high (2.0 A/g) current densities (Supplementary Fig. 38 and Fig. 5d). The bare Zn| |NH4V4O10 cell shows a rather low initial capacity of 353.2 mAh/g and 252.1 mAh/g at 0.5 A/g and 2.0 A/g, respectively, and the capacity rapidly decays to zero after 970 and 821 cycles. In sharp contrast, the Zn(11\(\bar{2}\)0)| |NH4V4O10 cell, at 0.5 A/g, exhibits stability with high capacity retention of 87.6% and a CE of 99.9% for 1000 cycles. At 2.0 A/g, 94.7% of the initial capacity is still retained after 2600 cycles, with a high initial capacity of 253.2 mAh/g. Such performance outperforms previously reported Zn-based full cells (Supplementary Fig. 39 and Supplementary Table 5). Furthermore, the normalized discharge/charge profiles of the cell of the bare Zn| |NH4V4O10 cell show larger voltage hysteresis than that of the Zn(11\(\bar{2}\)0)| |NH4V4O10 cell (Supplementary Fig. 40). It is worth noting that the morphology of bare Zn negative electrode is irregular, with sharp tips and corrosion pits, which are the main causes of short circuits. In contrast, the Zn(11\(\bar{2}\)0) electrode remains a relatively homogenous and dense morphology even after 1000 cycles, as shown in the insets of Fig. 5d, Supplementary Figs. 41, 42. These results verify the potential of the Zn(11\(\bar{2}\)0)| |NH4V4O10 cell for practical application, arising from its enhanced full-cell performance.

Discussion

In this study, a single [11\(\bar{2}\)0]-oriented Zn metal negative electrode with high surface energy demonstrates unexpectedly high reversibility and battery performance. This finding challenges the conventional crystallographic modulation concept, which typically favors exposing low-surface-energy crystal planes in metal electrodes. We found that uniaxially oriented Zn metal negative electrodes could guarantee spontaneous homoepitaxial growth during electrochemical plating due to their high GB stability and identical grain growth rates. The concept of uniaxial orientation can serve as a starting point to inspire researchers to explore other highly reversible metal negative electrodes in rechargeable batteries.

Methods

Materials

Polyacrylamide (PAM, AR), boric acid (H3BO3, AR), ammonium metavanadate (NH4VO3, AR), and N-methyl pyrrolidone (NMP, AR) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Zinc sulfate heptahydrate (AR), ethanedioic acid dihydrate (H2C2O4·2H2O, GR), absolute ethanol (analytical grade), perchloric acid (HClO4, AR grade), and zinc foil (AR) were obtained from Sinopharm Chemical Reagent Co., Ltd. Super P (battery grade) and polyvinylidene fluoride (PVDF, battery grade) were supplied by Soochow DoDoChem Technology Co., Ltd. All chemicals were used as received without further purification. Moreover, the separator used in this study was a commercially available glass microfiber filter (Whatman GF/C, CAT No. 1822-090, Lot No. 17387792, 90 mm diameter, thickness ~260 μm, porosity ~1.2 μm) from Cytiva. One piece (19 mm in diameter) was used per cell without further treatment.

Preparation of Zn metal electrodes

Copper (Cu) foil (100 μm, 99.9999% metal basis, Alfa) was rinsed with 1.0% sulfuric acid (H2SO4) to remove surface oxides and increase active sites before being used as an electrodeposition substrate. Zn metal electrodes with a thickness of approximately 120.0 μm were synthesized using direct-current electrodeposition. A Cu sheet with a diameter of 15.0 mm served as the positive electrode, and a pure Zn plate as the negative electrode. The distance between the negative and positive electrodes was approximately 5.0 cm. For the electrodeposited Zn(11\(\bar{2}\)0) metal electrodes, the electrolyte consists of 100.0 g/L ZnSO4·7H2O, 10.0 g/L boric acid (H3BO3) and 0.8 g/L PAM. The electrolyte bath was mechanically stirred at 800 rpm. The deposition was performed 1 h at 25 °C with a current density of 30 mA/cm2 and a pH value of 2. The Zn(0002) metal with a single (0002) texture was synthesized in the electrolyte with 100.0 g/L ZnSO4·7H2O and 20.0 g/L H3BO3, and other conditions remained the same.

Preparation of NH4V4O10 positive electrode

1.170 g NH4VO3 was dissolved in deionized water at 80 °C and then formed a light-yellow solution. Subsequently, 1.891 g of H2C2O4 · 2H2O solid powder was added to the solution under magnetic stirring until it turned black-green. The resulting solution was transferred to a 50 mL autoclave and kept in an oven at 140 °C and 3 MPa for 48 h. After cooling to room temperature, the products were collected, washed repeatedly with deionized water, and dried at 60 °C for 12 h to obtain the NH4V4O10 positive electrode materials. The positive electrode consists of NH4V4O10, conductive carbon black and PVDF mixed at a mass ratio of 7:2:1 in NMP solvent. The above mixture was uniformly spread on a 304 stainless steel current collector (400 mesh, wire diameter 0.03 mm) and dried in a vacuum oven at 90 °C for 12 hours. The effective mass loading of active materials is controlled at 1.2-1.5 mg/cm2.

Characterizations

Structural characterization and texture analysis of Zn metal electrodes during stripping/plating process were first conducted using an X-ray diffractometer (Rigaku Mini Flex 600 diffractometer) operated at 100 mA and 40 kV with Cu-Kα radiation at a step of 10o/min, and then by SEM, TEM and HAADF STEM. Thin foils for plane-view SEM, TEM and HAADF-STEM observations were peeled off from the Zn metals and cut from the growth plane, followed by mechanical grinding and subsequent electrochemical-polishing in a solution of alcohol and perchloric acid at 0 °C. Samples for side-view TEM and HAADF-STEM observations were prepared by using focused ion beam (FIB) using a FEI Nova 200 Nanolab FIB-SEM system operated at a voltage of 30 kV. TEM and HAADF-STEM observations were performed on a Tecnai G2 F20 and FEI Talos F200X operated at 200 kV. Grain sizes of bare Zn and Zn(11\(\bar{2}\)0) metals were evaluated by plane-view SEM and DF TEM images, respectively. Orientation mapping was operated on a Tecnai G2 F20 electron microscope at a voltage of 200 kV by using NanoMEGAS hardware and ASTARTM system. PED patterns are captured at predefined step intervals when the electron beam is scanned across the TEM samples. The crystal orientation information of grains is obtained and an orientation map is developed through ultrafast matching of these acquired PED patterns with the precalculated kinematical diffraction patterns. Pole figures were obtained from the measured crystal orientation map, ODFs were calculated likewise and presented in constant \({{{\rm{\varphi }}}}_{1}\,=\,{180}^{o}\) sections in the Euler space defined by three Bunge Euler angles: φ1, ϕ and φ2. Test samples for residual stress, hardness and Young’s modulus measurements were pretreated by electrolytic polishing at 12 V in a solution of perchloric acid and alcohol (1:9) at a temperature of −20 °C for 30 s. Residual stress was measured on a Rigaku D/max-2400 X-ray diffractometer using XRD with the classical sin2ψ method, which is expressed as: \({\sigma }_{\varphi }=K\cdot M\), where \(K=\frac{E}{2(1+\upsilon )}\tan {\theta }_{0}\cdot \frac{\pi }{180}\), \(M=\frac{\partial {(2\theta )}_{\psi }}{\partial {\sin }^{2}\psi }\), ψ is the angle between the coating surface normal and the bisector of the incident and reflected beams, \({\theta }_{0}\) is the diffraction angle of the lattice planes in the absence of stress, E the elastic modulus, and υ the Poisson’s ratio. The ψ angles used were ±25.00o, ±17.48o, ±11.80o, and 0o. Young’s modulus was measured using nanoindentation tests on a Hysitron G200 TriboIndenter equipped with a Berkovich geometry diamond tip.

Growth mechanism and GB energy characterizations

To assess the crystal structure and atomic arrangement of the Zn deposited on Zn(112̅0) substrates, HAADF-STEM was employed. Samples were electrodeposited at 5.0 mA/cm2 for 2 h, then sectioned parallel to the growth direction using a FIB. The cross-sections were polished mechanically and then electrochemically using a mixture of ethanol and perchloric acid at 0 °C. Only regions adjacent to the epitaxial interface were selected for imaging. GB information was extracted using EBSD and PED mapping techniques.

Theoretical calculations

First-principles simulations were carried out using the Vienna Ab initio Simulation Package (VASP) with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional under the generalized gradient approximation (GGA). The projected augmented wave (PAW) method was applied, and a plane-wave energy cutoff of 520 eV was chosen to ensure adequate description of the valence electrons. The Brillouin zone was sampled using k-point grids spaced at 0.03*2π/Å. Van der Waals interactions were included using the DFT-D3 dispersion correction. Gaussian smearing with a width of 0.05 eV was applied to account for partial orbital occupancies. Geometry optimizations were performed with force convergence thresholds below 0.01 eV/Å and electronic energy convergence set to less than 10−7 eV.

Owing to the restrictions of computing performance and efficiency, quantum mechanical methods are usually used to study small models containing 10 to 100 atoms. There are numerous amide groups in the PAM, which can react with many compounds to generate various polyacrylamide derivatives. Referring to the work of Zhang et al, the nonionic PAM structural unit molecule was constructed in this paper. The PAM was assembled using the Visualizer program in Materials Studio 7.0 software. Geometry optimization was performed on the PAM structure in a 10 × 10 × 10 Å3 cubic box using the CASTEP program.

Surface energies of Zn facets were determined using a slab model with a 15 Å vacuum layer to prevent inter-slab interactions. Each slab included more than 12 Å of atomic layers to isolate surface effects. Aqueous solvent effects were modeled using the VASPsol implicit solvation framework during energy and adsorption calculations. The surface energy (\(\sigma\)) was computed using the relation: \(\sigma=({E}_{s}-n{\mu }_{{Zn}})/2A\), where \({E}_{s}\) is the slab’s total energy, µZn is the chemical potential of bulk Zn, n is the atom count, and A is the surface area. Several slab models with 5, 7, 9, and 11 atomic layers were evaluated to fit σ-n relationship. Here, the above equation can be modified as: \({E}_{s}=2A\times \sigma+n{\mu }_{{Zn}}\). The surface energy can be fitted from the relationship between Es and n. Surface energy is susceptible to changes due to the adsorption of ions or molecules. The corresponding adsorption energy (\({E}_{{ads}}\)) is calculated as: \({E}_{{ads}}\,=\,{E}_{t}-\,{E}_{s}-n{\mu }_{i}\), where \({E}_{t}\) is the total energy of the system after adsorption, \({E}_{s}\) refers to the energy of the pristine slab, and \({\mu }_{i}\) represents the chemical potential of the adsorbed species. Following adsorption, the surface energy is modified and denoted as \({\sigma }^{*}\), which is expressed by: \({\sigma }^{*}\,=\,\sigma+\,{E}_{{ads}}/A\). To determine the equilibrium morphology of Zn nanoparticles, we applied Wulff’s construction, which utilizes surface energy anisotropy and crystallographic symmetry to predict the thermodynamically favored shape. From this, the surface coverage ratio (θ) for different crystallographic planes can also be derived. For interfaces, the interface energy (γ) is calculated using the formula: \(\gamma=({E}_{{total}}-n{\mu }_{{Zn}})/2A\), where \({E}_{{total}}\) denotes the total energy of the interface model, \({\mu }_{{Zn}}\) is the chemical potential of Zn in its bulk phase, n is the number of Zn atoms in the interfacial region, and A is the interfacial area.

Electrochemical measurements

All electrochemical evaluations were conducted at 25 ± 5 oC using CR2025 coin cells without environmental chamber control. Whatman GF/C glass fiber (260 μm thick) served as the separator. Coulombic efficiency was measured using asymmetric Zn| |Cu cells. Tafel curves were recorded in 3 M ZnSO4 electrolyte on a Multi Autolab/M204 station with a three-electrode setup: bare Zn or Zn(11\(\bar{2}\)0) as the working electrode, platinum (Pt) foil as the counter electrode, and Ag/AgCl as the reference electrode. Full cell tests using Zn| |NH4V4O10 configurations were conducted within a 0.4-1.4 V voltage window using 100 μL of 3 M ZnSO4 electrolyte. For symmetric Zn| |Zn and asymmetric Zn| |Cu cells, the applied current densities were normalized to the electrode surface area. Specific capacities and current densities in Zn| |NH4V4O10 cells were based on the active mass of the cathode. EIS and linear sweep voltammetry (LSV) tests were performed using a CHI660E workstation.