Ordered planar plating/stripping enables deep cycling zinc metal batteries

Abstract

Metal anodes are emerging as culminating solutions for the development of energy-dense batteries in either aprotic, aqueous, or solid battery configurations. However, unlike traditional intercalation electrodes, the low utilization of “hostless” metal anodes due to the intrinsically disordered plating/stripping impedes their practical applications. Herein, we report ordered planar plating/stripping in a bulk zinc (Zn) anode to achieve an extremely high depth of discharge exceeding 90% with negligible thickness fluctuation and long-term stable cycling. The Zn can be plated/stripped with (0001)_Zn preferential orientation throughout the consecutive charge/discharge process, assisted by a self-assembled supramolecular bilayer at the Zn anode-electrolyte interface. Through real-time tracking of the Zn atoms migration, we reveal that the ordered planar plating/stripping is driven by the construction of in-plane Zn─N bindings and the gradient energy landscape at the reaction fronts. The breakthrough results provide alternative insights into the ordered plating/stripping of metal anodes toward rechargeable energy-dense batteries.

Ordered planar plating and stripping of zinc foil anodes achieves a high depth of discharge with negligible thickness fluctuation.

INTRODUCTION

The past decades have witnessed the sparkling times of lithium ion batteries in which the lithium ions are (de)intercalated between the cathode host and the anode one, enormously powering portable electronics, transportations, and emerging smart grids (1). However, because of the high weight percentage of the host in the traditional intercalation electrodes, their inherent capacity ceiling has been driving the exploration for “hostless” metal anodes featured by a fully electrochemical active nature. In contrast to other high–energy density metals such as lithium and sodium, Zn is chemically stable in air and is nonflammable, thus offering opportunities toward intrinsically safe and energy-dense aqueous batteries (2, 3). Nevertheless, unlike the ordered and reversible (de)intercalation under traditional intercalation chemistry, hostless Zn anodes suffer from severe irreversibility issues such as dendrite formation and side reactions (4), leading to low Coulombic efficiency and even short-circuiting of the battery in prolonged cycles. In previous reports, Zn mostly had to be used in notable excess to maintain the supply during irreversible consumption, which evidently declines the energy densities of the batteries and further impedes their widespread deployment (2, 3).

Extensive efforts have been devoted to tackling the irreversibility issues of Zn anodes from the point of view of electrode-electrolyte interface (5, 6), solvation sheath (7–10), crystallographic orientation (11, 12) and electrode composition (13–15), etc. For instance, formation of a high-quality solid-electrolyte interphase is an effective strategy to prolong the life span and prevent irreversible losses caused by nonuniform ions/atoms migration, side reactions, and gas generation (16, 17). Designing substrates with strong orientation relations, e.g., graphene or textured materials that have low lattice mismatch with Zn, can also improve the reversibility and, to some extent, mitigate the electrode volumetric change upon cycling (18, 19) However, the correlation between the electrode and/or its interface and the cell performance is notably complex, especially at deep (dis)charge state, with effects at multiple length and timescales. The disordered Zn plating/stripping has substantial impacts but is rarely studied. Zn plating/stripping is conducted in a hostless manner, that is, the electrochemical performance of the Zn anodes critically depends on the electrode microstructures. The active atoms contribute to the chemistry (electrochemical potentials, electron densities, locally structural environments, etc.) differently in time and space. In general, the atoms at the reactive locations (such as lattice defects and crystal plane edges) release and/or absorb as redox active ions notably faster than their counterparts, resulting in an nonplanar and disordered plating/stripping behavior (20). As the usable Zn utilization [often referred to as depth of discharge (DOD_Zn)] further increases to >50%, which is the stringent energy density requirement at the cell level (21), no effective strategy is currently capable of addressing the irreversibility issues.

Herein, we report an ordered planar Zn plating/stripping benefited from the dynamic formation of a self-assembled supramolecular bilayer (SAB) interphase by a custom-designed lipopeptide at the electrode-electrolyte interface. This SAB interphase builds the in-plane local motifs of the topmost Zn layer through formation of Zn_{(topmost layer)}─N_{(C-terminal amino group of lipopeptide)} bindings. As a result, the activation energy of the distinct Zn atom migration pathways is substantially lower compared to the bulky counterparts, dominating the reaction kinetics of the Zn/Zn²⁺ redox process and further driving rapid and continuous Zn plating/stripping under a planar mode with (0001)_Zn preferential orientation and negligible side reactions. On the basis of the bulk Zn foil anode, the highly ordered plating/stripping achieves a DOD_Zn of 90% and a cumulative capacity of up to 11,000 mAh cm⁻², the highest values reported to date, thus offering fundamental insights into ordered and reversible plating/stripping for deep cycling metal anodes.

RESULTS

Self-assembly of SAB on metallic Zn surface

Considering that the plating/stripping of the hostless Zn metal anodes generally occurs at the electrode-electrolyte interface, we aimed to confine the redox reaction front within the topmost Zn layer to enable the ordered planar plating/stripping (Fig. 1A). Thus, the critical steps in implementing the concept were (i) identifying large-scale, high-density, and dynamic interphase materials that have high affinities with the Zn surface and (ii) fabrication of a long-range-ordered interphase that rapidly sieves/desolvates and transports/homogenizes specific Zn²⁺ ions to prevent side reactions and/or dendrites during the (dis)charge process.

Fig. 1. Large-scale, uniform, and ultrathin SAB self-assembled by lipopeptides.

(A) Schematic depiction representing the ordered planar plating/stripping process. (B) Height statistical deviation (Sq) of the SAB self-assembled by RWW-containing lipopeptides with different aliphatic chain lengths. (C) Atomic force microscopy characterization of the SAB self-assembled by C₁₄-RWW-NH₂. Top: Morphological pattern of the SAB. Scale bar, 5 μm. Bottom: Height profile corresponding to the white line in the morphology image. The height of the SAB was measured to be 2.9 ± 0.3 nm, approximately twice the molecular length of a C₁₄-RWW-NH₂. (D) Neutron reflectivity profile of the SAB. The discrete circles and solid lines represent the experimental and fitted results, respectively. (E) Schematic representation of the molecular configuration inside the SAB fitted from the neutron reflectivity curves, showing the thickness of the SAB to be ~3.4 ± 0.2 nm. It is noted that because of the large roughness of the metallic Zn wafer, the morphological characterizations of the SAB films were conducted using the silica or mica surfaces. (F) Bonding geometry of the Trp residue (corresponding to the C-terminal moiety of C₁₄-RWW-NH₂ SAB) on the Zn(0002) surface as determined by density functional theory calculations.

In this regard, inspired by the conformation of natural cell membranes (22), in which the phospholipid molecules self-assemble into high-order bilayer (SAB) superstructures, we first developed a series of lipopeptides composed of a hydrophobic aliphatic chain and arginine (R), a positively charged amino acid with high affinity to metallic Zn (23), followed by a di-tryptophan (di-Trp, WW) motif capable of chelating with Zn and/or Zn²⁺ moieties (figs. S1 to S6) (24). Accordingly, driven by noncovalent driving forces, these bioinspired amphiphiles could self-assemble into a SAB membrane on the surface of the metallic Zn in aqueous ZnSO₄ electrolyte (25), as evidenced using Fourier transform infrared (FTIR) spectroscopy and energy-dispersive x-ray (EDX) spectra (figs. S7 and S8). By optimizing the length of the alkyl chains (C₁₀, C₁₄, C₁₈, etc.), a large-scale, uniform, and ultrathin SAB membrane (<5 nm, approximately twofold of a lipopeptide molecular length) with nearly 100% coverage (Sq <5 Å of C₁₄-RWW-NH₂) could be grown throughout the surface without facet selectivity (Fig. 1, B and C; also see fig. S9).

The bilayer structure of the SAB formed by C₁₄-RWW-NH₂ was further confirmed by both neutron reflectivity (Fig. 1D) and spectroscopic ellipsometry (table S1) measurements. The scattering length density correlation data showed that the C₁₄-RWW-NH₂–based SAB was indeed a sandwich-like configuration, comprising two peptide layers at two ends (exterior) and an interdigitated alkyl chain phase in the middle (interior) (Fig. 1E and table S2), with an apparent total thickness of 3.4 ± 0.2 nm, suggestive of an upright arrays alignment (26). On the Zn(0002) surface, we calculated the adsorption energy of the delocalized Zn adatoms by detachment from equilibrium position as −1.26 eV (Fig. 1F and fig. S10), indicating strong specific interactions between the C-terminal amino group on the exterior peptide layer and the Zn surface, which is expected to activate the topmost Zn layer in an energetically favorable manner (as discussed below). These results indicate that the C₁₄-RWW-NH₂ lipopeptide was successfully absorbed onto the surface of the Zn anode to generate a dynamic SAB with ordered molecular arrays.

Electrochemical analysis of the SAB-Zn anode

To confirm the effect of the SAB on the Zn anode, we first examined the stability of the SAB-Zn. The overall SAB-Zn–based electrochemical configuration is schematically illustrated in Fig. 2A. In sharp contrast to the pristine Zn, the incorporation of SAB significantly improved the chemical stability of Zn with no changes in the morphology and phase of SAB-Zn after soaking in the electrolyte solution for 7 days (figs. S11 and S12). The double-layer capacitance analyzed using cyclic voltammetry at various scanning rates (fig. S13) and zeta potential characterizations (fig. S14) affirmed the dynamic protection of the Zn surface by the SAB. Also, Raman spectroscopy characterizations demonstrated that in the presence of SAB, the [Zn(H₂O)₆]²⁺ peak (400 cm⁻²) decreased and became broad (fig. S15), indicating that the solvation effect had become weaker (27). The deuterium nuclear magnetic resonance (²H-NMR) spectra revealed that after SAB incorporation, the O─H stretching in the D₂O showed a downfield shift (fig. S16), indicating that more free water molecules had been released (28), thus demonstrating that the SAB contributed to the desolvation of the Zn²⁺. Upon the growth of SAB on the metallic Zn, the relatively low desolvation activation energies of Zn²⁺ sheath (58.75 kJ mol⁻¹, calculated on the basis of the temperature-dependent Nyquist plots) (fig. S17) and the high hydrophobic nature with a contact angle of up to 87.4° (fig. S18) could further restrain the water adsorption at the electrode-electrolyte interface toward inhibition of hydrogen evolution and other side reactions (mainly due to the presence of the hydrophobic interior) (29). Correspondingly, the corrosion current density was decreased from 0.68 to 0.30 mA cm⁻² (Fig. 2B) under the decreased initial hydrogen evolution reaction potential (fig. S19).

Fig. 2. Electrochemical performance of the SAB-Zn anode in (a)symmetric cells.

(A) Schematic illustration of the electrochemical configuration using SAB-Zn as the anode. (B) Linear polarization curves, (C) chronoamperometry curves, and (D) CE evolution curves at a current density of 1 mA cm⁻² with an areal capacity of 1 mAh cm⁻² on a Ti substrate of the Zn|Zn (Ti) (a)symmetric cells in the absence (black) or presence (red) of SAB. (E) Representative voltage profiles corresponding to the circled numbers in (D). (F and G) Cycling stability of the Zn|Zn symmetric cells in the absence or presence of SAB under current densities and capacities of (F) 1 mA cm⁻² and 1 mAh cm⁻² and (G) 20 mA cm⁻² and 10 mAh cm⁻², respectively. The insets represent the charge-discharge curves at different cycles of the cells. (H) Summary of previously published Zn plating/stripping CE and galvanostatic measurements. The utilization of Zn per cycle is labeled in blue, where available (see also table S3 for the reference, description, and values of each point).

Furthermore, the dendritic growth could also be inhibited with the introduction of SAB. Dense nucleation sites and stable Zn²⁺ ion three-dimensional (3D) diffusion are known to be beneficial for obtaining a compact deposition layer (6). Compared to the pristine Zn, the nucleation overpotential of Zn on a Ti electrode with SAB was increased by approximately 36 mV (fig. S20), suggesting the formation of uniform and refined nuclei. This can be attributed to the strong hindrance of the ordered lipopeptide molecular arrays of SAB, which guides the uniform Zn²⁺ ion flux on the metallic surface. Correspondingly, the 2D diffusion was well restricted to only ~5 s followed by a continuous 3D diffusion as verified by chronoamperometry measurement (Fig. 2C). Moreover, the exchange current density of SAB-Zn (1.8 mA cm⁻²) was higher than that of the pristine Zn (1.3 mA cm⁻²) (fig. S21), which indicates that the SAB could promote the Zn²⁺ transfer and reaction kinetics. These preliminary results imply the electrolyte blocking and rapid, homogeneous Zn²⁺ transport nature of SAB (Fig. 2A), thus constructing an ideal electrode-electrolyte interface toward an ordered and reversible Zn plating/stripping reaction.

We then investigated the effect of SAB on the cycling behaviors of the Zn anodes. The SAB-Zn displayed stable voltage profiles up to 1000 cycles (Fig. 2E) at 1 mA cm⁻² and 1 mAh cm⁻² on a Ti substrate, with an average Columbic efficiency (CE) of 99.6% (Fig. 2D). By contrast, a marked voltage increase along with CE attenuation were observed for the pristine Zn anode. Moreover, when increasing the deposition areal capacity to 10 mAh cm⁻², the Zn plating/stripping was maintained stable and exhibited a high average CE of 99.4% (fig. S22). Figure 2 (F and G) shows the galvanostatic charge/discharge voltage-time curves for the control (pristine Zn) and SAB-Zn–based symmetric cells under various test conditions, in which the concentration of C₁₄-RWW-NH₂ in the electrolyte was optimized to the critical micelle concentration (i.e., 0.4 mM) aiming to obtain the best performance (figs. S23 to S25) (30). Specifically, the symmetric cells composed of SAB-Zn electrodes showed a reproducible performance with a maximum life span of >3000 hours at 1 mA cm⁻² and 1 mAh cm⁻² (Fig. 2F). Similar results were obtained at higher current densities (10 and 20 mA cm⁻²) and areal capacities (5, 10 mAh cm⁻²) (fig. S26). Particularly, the SAB-Zn anode presented stable cycling for over 1100 hours with a low overpotential of 85 mV at 20 mA cm⁻² and 10 mAh cm⁻² (fig. S26B), achieving an impressive cumulative capacity of 11,000 mAh cm⁻² (Fig. 2H and table S3).

In an extreme case, ultrathin Zn foils only 18.9 μm in thickness were used to assemble SAB-Zn symmetric cells (fig. S27). When increasing the DOD_Zn to 90%, corresponding to an areal capacity of 10 mAh cm⁻², a stable cycling up to 300 hours was recorded, achieving the highest DOD_Zn among the Zn-based symmetric cells reported to date (Fig. 2, G and H; also see table S3). In sharp contrast, the controls degraded after 50 hours with a sudden voltage drop. The improved electrochemical performance proves the substantial advantages of the SAB-mediated Zn plating/stripping reaction process.

Ordered planar Zn plating/stripping

To further track the changes in packing density and uniformity of the Zn anodes upon successive deep (dis)charge steps, we recorded scanning electron microscopy (SEM) images of the Zn anodes before and after plating/stripping under 90% DOD_Zn (Fig. 3, A to D). Holes and dendrites were extensively detected across the pristine Zn anode after the first cycle (upper panel in Fig. 3C; also see figs. S28, A to F, and S29, A and B), probably because of the severe side reactions and nonplanar, disordered plating/stripping process (19). Instead, the SAB-Zn retained an obviously smooth morphology and intact electrode structure even after extreme discharging (lower panel in Fig. 3C and figs. S28, G to I, and S29, C and D). Unexpectedly, upon further charging back to the equilibrium potential, the majority of the Zn reverted to their original sites with a slight volumetric change of 16.7% (from 30 to 35 μm; 30-μm Zn foil was used for convenience in characterization) (Fig. 3, A and B), one order of magnitude lower than the pristine Zn (>300%; fig. S29B). We thus hypothesize that the SAB-Zn largely exhibits ordered planar plating/stripping under such ultradeep (dis)charging process. It is noted that during cycling, the SAB-Zn anode appeared to be “sliced” in parallel and showed intimately attached hexagonal texturing along the plane surface without dendrite growth (Fig. 3D), implying an in-plane synchronous and continuous Zn crystallographic reorientation induced by the SAB, as discussed below. EDX analysis of the charged SAB-Zn showed overlapping Zn and N signals, further verifying the strong dynamic correlation of Zn and SAB during the reaction process. Under a regular DOD_Zn of 1.07% following extensive cycling, damages and aggregations were observed within the pristine Zn (figs. S30, A and D). In comparison, no observable traces of microcracks or dendrites were found on the SAB-Zn anode over 100 cycles (fig. S30, B and E), showing excellent surface damage/dendrite inhibition and ordered planar plating/stripping induction (31), as also confirmed by in situ optical microscopy characterization at a larger observation scale (fig. S31). Furthermore, we found that the typical FTIR signals representing the interactions between the head group of the SAB and the Zn were detected after 10 or even 100 cycles, suggesting that the structure of the SAB did not substantially change during long-term battery cycling (fig. S32).

Fig. 3. Mechanism of the SAB-induced highly ordered plating/stripping of the Zn anode.

(A and B) Cross-sectional SEM images of the 30-μm-thick SAB-Zn electrode (A) before and (B) after first cycling until 90% plating/stripping in the symmetric cell. The Zn foil showed a slight thickness change of 16.7%, from 30 to 35 μm. (C) SEM images of the first stripped pristine Zn and SAB-Zn foil. (D) EDX analysis of the subsequently plated SAB-Zn electrode surface, showing dynamically uniform distribution of the Zn and N elements. (E) XRD patterns of SAB-Zn before cycling, after first stripping, first plating, and the following stripping process at DOD_Zn of 90%. Dashed line, plating or stripping at DOD_Zn of 50%. (F and G) (0001)_Zn pole figures of (F) the pristine Zn and (G) SAB-Zn anode after first cycling. The concentrated intensities in (0001)_Zn pole figure verified the planar orientation growth of Zn in the presence of the SAB. (H and I) Time-dependent MSD of the topmost Zn layer of (H) SAB-Zn and (I) pristine Zn among AIMD simulations. The overall trend demonstrated the tendency toward a synchronized behavior along the z axis for SAB-Zn. (J) Schematic geometry at the key migration positions, indicated with white numbers along the Zn slab. (K) Radial distribution function of the SAB-Zn among AIMD simulations, illustrating the strong interaction between Zn and N. (L) MSD of Zn²⁺ along the SAB and other directions among CMD simulations. (M) Schematic presentation showing the possible molecular mechanism of SAB for synchronous in-plane Zn/Zn²⁺ transfer and reaction. The exterior peptide layers contribute to coordination with Zn/Zn²⁺, and the hydrophobic interior forms directional desolvation/transport channels along the z axis. Test conditions in (A) to (F), 20 mA cm⁻² and 15 mAh cm⁻². Scale bar in (A) to (D), 20 μm.

The effect of SAB on the texture of metallic Zn during plating/stripping reaction process was analyzed by x-ray diffraction (XRD) characterization (Fig. 3E). After first stripping, the SAB-Zn presented a similar XRD pattern to the original Zn foil, confirming its uniform Zn dissolution on each orientation (27). Conversely, the (1011)_Zn facet visibly exceeded in the case of pristine Zn (fig. S33), indicative of a preference of uneven Zn dissolution that leads to the generation of surface cracks and/or holes (11, 12). After the first plating process, the SAB-Zn changed the texturing behavior and resulted in an evidently preferred (0001)_Zn orientation at the 2θ of ~42.4°. The concentrated intensities in the (0001)_Zn pole figure further verified the oriented growth of Zn (Fig. 3G). Further stripping led to the reversible structural evolution of the SAB-Zn anode (blue lines in Fig. 3E; also see fig. S35C). Notably, the (0001)_Zn plane was predominantly parallel to the basal plane of the electrode, thus providing a reliable textured Zn to guide a continuous ordered planar Zn plating/stripping reaction (19). On the contrary, the pristine Zn anode exhibited textureless characteristics (Fig. 3F). Clear XRD patterns at 28.2° and 18.6° in the pristine Zn indexed to the (006) and (004) plane of Zn₄SO₄(OH)₆·xH₂O by-product, respectively, indicating parasitic side reactions that trigger the Zn passivation and heterogeneity. Electrode structural degradation (holes, dendrites, by-products, etc.) was thus rapidly observed under most testing conditions. The intensity of the (0001)_Zn facet orientation for SAB-Zn was also observed to increase within the prolonged cycles (fig. S34). As expected, when reaching the 10th and 30th cycles, the SAB-Zn anode still sustained a flat and compact Zn surface, whereas the pristine Zn was completely degraded (fig. S35). As a further demonstration, we also performed electrochemical plating and stripping on a commercial Cu current collector substrate. The SAB indeed promoted the formation of evenly distributed, horizontal (0001)_Zn deposits, in contrast to the uneven, coarse dendritic Zn morphologies formed in the absence of SAB regulation (fig. S36). These findings clearly reveal the nearly full interfacial regulation effect of the SAB in circumventing the disadvantages of aqueous electrolyte and altering the plating/stripping mode of the Zn/Zn²⁺ upon the (dis)charge process.

Simulation of the SAB functional mechanism

Density functional theory (DFT) and ab initio molecular dynamics (AIMD) calculations were further carried out to shed light into the mechanism of ordered planar reaction of the SAB-Zn. AIMD simulations showed that whereas Zn atoms in the pristine Zn anode fluctuated at nearly the same positions (Fig. 3I, fig. S37, and movie S1), the entire topmost Zn layer anchored to the SAB was lifted along the z axis by 0.2 Ų [extracted from mean-square displacement (MSD); the total MSD of the fully delocalized Zn atom was 7.2 Ų) (Fig. 3H, figs. S38 and S39, and movie S2) during the equilibration process up to 20 ps, suggesting that the SAB-Zn may undergo a synchronous in-plane activation to facilitate the homogeneous electrochemical reaction over the electrode surface.

We attempted to understand the role that large-scale, high-density and dynamic SAB played in this process. Radial distribution function (Fig. 3K) and static Bader charge analysis (fig. S40) revealed that ~0.27 |e| charge was transferred from the Zn atom to the C-terminal N group of the exterior peptide layer, effectively breaking the interaction of Zn─Zn coupling (32). Chemical analysis of the ion-etched area using x-ray photoelectron spectroscopy (fig. S41) well affirmed the formation of these strong interactions. The absence of Zn─N bindings in the bulk resulted in retained Zn─Zn bonds. A representative configuration shown in Fig. 3J further demonstrated average elongated Zn─Zn bond lengths of ~3.46 Å at the surface, considerably longer than those in bulky atoms (~2.84 Å). In combination with the full coverage and dynamic distribution of the C₁₄-RWW-NH₂–based SAB on the Zn surface, these features guarantee synchronous in-plane activation at the electrode scale. Such “raised” Zn─Zn local motifs effectively modify the (electro)chemical characteristics along the depth, thus generating stepwise kinetic pathway among neighboring active layers, as inferred from the MSD profiles. The decreased activation energies of atom migration within the topmost layer (~0.25 eV) versus the bulky counterpart (~0.41 eV) were also obtained by climbing image nudged elastic band calculations (fig. S42). Accordingly, the topmost Zn layer attached to the SAB underwent a faster electrochemical reaction than the boundary and/or foundation layers enclosed by the intrinsic atoms, facilitating the planar stripping process. This simulated prediction was confirmed by morphological visualization and structural interpretation for the SAB-Zn, in which the electrode tended to form a flat and intact surface upon an extreme discharging process.

In addition to the effect of induced sequential stripping, the specific interaction between Zn²⁺ and ordered zincophilic group-terminated SAB that can promote the Zn²⁺ ions migration behavior was also taken into consideration (33). Binding energy calculations showed that the Zn²⁺ was strongly coordinated with the functional N groups from the exterior peptide layer compared H₂O molecules (fig. S43), which is consistent with the Raman results. The strong and directed Zn²⁺ coordination induced by the SAB facilitated the transport of more Zn²⁺ ions from the electrolyte environment to the electrode surface. The ~30% increased transference number of Zn²⁺ ions demonstrated that the SAB-Zn indeed showed enhanced Zn²⁺ transfer compare to the pristine Zn (fig. S44), indicating more extensive and faster migration of Zn²⁺ ions at the SAB-Zn surface, in good agreement with the calculation results and electrochemical kinetics analysis. To reveal the Zn²⁺ transport pathway, we traced all Zn²⁺ migration steps through classical molecular dynamics (CMD) simulation time. The MSD analysis revealed that the diffusion coefficient of Zn²⁺ (extracted from the slope of the MSD curves from 100 to 400 ps) along the SAB arrays was larger than along other directions (with a slope ratio of 1.5:1) (Fig. 3L and fig. S45), indicating the directed transport nature (34). The restricted 2D displacement along x and y axes is consistent well with the SEM and chronoamperometry results. Combined with the experimental observations and the calculations, a plausible molecular mechanism of SAB-mediated enhancement of Zn²⁺ transfer is depicted in Fig. 3M. The vertically ordered orientation of the exterior peptide layers with high Zn affinity on both the Zn surface and the electrolyte side enabled extensive and fast Zn²⁺ transport, whereas the hydrophobic interior served as directional channels that desolvated and carried a uniform Zn²⁺ ion flux across the bilayer interphase. It is possible that in-plane Zn─N bindings further facilitated the synchronous transfer and redox of Zn/Zn²⁺ under the electrical field, thus leading to the planar, (0001)_Zn-textured Zn growth (35). Therefore, considering the probably synchronous Zn/Zn²⁺ in-plane activation behavior, an ordered planar reaction in the case of SAB-Zn was achieved. On the one hand, the side reactions were inhibited with the construction of an ideal electrode-electrolyte interface. On the other hand, the planar Zn plating/stripping initiated from the topmost Zn layer was guaranteed, thereby promoting the reversibility of the reaction and lastly leading to the stable Zn operation even at a deeper DOD_Zn state.

Energy-dense aqueous SAB-Zn–based full cells

Next, we pursued the realization of full cells using the SAB-Zn anode with typical MnO₂ (fig. S46, A and B) as the cathode under realistic conditions. The low capacity ratio of the negative electrode to the positive electrode (N/P) was ~3.6 with lean electrolyte (10 μl mAh⁻¹), which was used to fabricate a 280 Wh liter⁻¹ aqueous zinc-ion battery (with a calculated energy density of 157.5 Wh kg⁻¹; see tables S4 and S5). At such a low anode loading, a Zn metal foil would decrease the cycling stability owing to the deepened plating/stripping process of Zn/Zn²⁺ ions (36). Between 0.8 and 1.8 V (versus Zn/Zn²⁺), the capacity of the pristine Zn|MnO₂ showed a rapid drop after 0.1 C activation, with only ~17.8% capacity retention at 1 C after 300 cycles (figs. S46, C and D). On the contrary, the SAB-Zn|MnO₂ delivered a significantly enhanced 246 mAh g⁻¹ discharge capacity at 1 C and retained 64.7% of its original capacity over 500 cycles (Fig. 4, A and B), indicative of a promising ordered planar reaction to boost high-energy density aqueous batteries. Even when the current density was further increased to 10 C, the full cell of SAB-Zn|MnO₂ showed marked cycling stability and CE, with little capacity loss over 200 cycles (fig. S47). Increased discharge-specific capacities and capacity retention were observed using rate (fig. S46, E and F) and self-discharge (Fig. 4C and fig. S48) tests, which are presumably assigned to the weakened electrolyte decomposition at an increased charge/discharge rate due to the SAB stabilization, further verifying its practicability. In addition, when replacing MnO₂ with NiHCF, another typically unstable cathode, the SAB-Zn|NiHCF full cell also showed an enhanced CE and prolonged cycle life span relative to the controls, generating an average CE of 99.7% along with a 75.6% capacity retention after 1000 cycles (fig. S49). Figure 4D and table S5 summarize the electrochemical performances and testing conditions of previously published rechargeable zinc-ion batteries. Apparently, the SAB-induced ordered and reversible plating/stripping electrochemistry supports a full cell with excellent cycle life, high CE, as well as unprecedented Zn anode utilization and energy density.

Fig. 4. Characterization of the Zn|MnO₂ full cell equipped with the SAB-Zn anode.

(A) The charge and discharge profiles of SAB-Zn|MnO₂ full cell. (B) Cycling stability and CE of the full battery composed of an ultrathin SAB-Zn anode (11.7 mAh cm⁻²), a high-mass loading MnO₂ cathode (3.2 mAh cm⁻²), and a MnSO₄-containing ZnSO₄ electrolyte (10 μl mAh⁻¹) at a current density of 1 C (1 C = 500 mA g⁻¹). (C) Storage performance of the SAB-Zn|MnO₂ full cell evaluated by resting for 48 hours at 100% state of charge, followed by full discharging. (D) Comparison of the energy density obtained in this work to previously reported values of Zn-ion batteries. The specific energies are calculated from the specific capacities of the cathode and anode and the average discharge voltage of the cell. Q_anode represents the utilization of a Zn anode. The goal of >100 Wh kg⁻¹ is shown as a dashed horizontal line (see table S5 for details of the data).

DISCUSSION

In summary, an ordered planar plating/stripping process in a hostless Zn foil anode is demonstrated, enabling an ultrahigh utilization exceeding 90% with negligible thickness fluctuation and long-term stable cycling. This process is based on SAB-induced synchronous in-plane activation involving the formation of interfacial Zn_{(topmost layer)}─N_{(C-terminal amino group of lipopeptide)} bonds at the Zn/Zn²⁺ redox reaction front. The associated low Zn activation barriers of the topmost layer reconcile the apparent contradiction between the high depth of (dis)charge capability of Zn and the poor reaction gradient of its intrinsically all-active phases, which further drives continuous planar plated/stripped Zn with (0001)_Zn preferential orientation and negligible side reactions. The ordered planar plating/stripping is expected to be extended to other metal anodes, such as Li, Na, K, Mg, Ca, and Al, to achieve ultrahigh DOD for practical high energy density. We anticipate that the mechanistic insights into the behavior of ordered and reversible plating/stripping will inspire and guide the future development of robust rechargeable metal batteries and realize energy-dense battery chemistry.

METHODS

Materials

The aliphatic acids were purchased from Alfa Aesar and N-terminal– and side chain–protected amino acids [Fmoc-Arginine(Pbf), Fmoc-Tryptophane(tBu)] from GL Biochem, O-(1H-benzotriazole-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU) and N-hydroxybenzotriazole anhydrous (HOBt) from Chem-Impex International (Wood Dale, IL, USA), N,N′-diisopropylethylamine (DIEA), triisopropylsilane (TIS), hexafluoroisopropanol, and Rink-amide MBHA resin from Merck, trifluoroacetic acid (TFA), dichloromethane (DCM) and N,N-dimethylformamide (DMF) of peptide synthesis purity grade from Bio-Lab. Other chemicals were purchased from Sigma-Aldrich and were used as received without further purification. Water was processed by a Millipore purification system with a minimum resistivity of 18.2 megohms cm⁻¹. Zinc foils (100, 20 μm) were purchased from Alfa Aesar. Specifically, 18.9- and 30-μm zinc foil was further produced by repeated self-folding and rolling of the commercial Zn foil. Glass fiber separator (GF/D) was purchased from Whatman Corporation. Cu, Ti foil, and coin-cell cases (CR2025-type) were purchased from Kelude Materials Corporation.

Lipopeptide synthesis

Lipopeptides were synthesized on a CEM Liberty Lite microwave peptide synthesizer (Matthews, NC, USA) using the standard Fmoc solid-phase synthesis strategy. Briefly, after deprotection of the Rink-amide MBHA resin with a 20% piperidine and 0.1 M HOBt in DMF solution, Fmoc-modified amino acids were introduced successively from the C terminus to N terminus, followed by introducing the aliphatic acid (37). The carboxylic groups were activated by treatment with HBTU/HOBt/DIEA, transforming the carboxylic acids into activated esters to react with the deprotected-amine groups. After synthesis, cleavage from the resin and deprotection of the side chains were coperformed using a mixture of TFA, TIS, and H₂O at a ratio of 95:2.5:2.5. The cleavage mixture and subsequent DCM washing solution were then purged with nitrogen.

The obtained concentrated solutions were added to water and lyophilized, and then subjected to reverse-phase high-performance liquid chromatography (HPLC) and mass spectrometry (MS) for analysis. The results showed that the products purity was >98% (figs. S1 to S6). The lyophilized lipopeptide powders were then dissolved in 2.0 M ZnSO₄ aqueous solution to prepare the electrolytes at lipopeptide concentrations of 0.1, 0.2, 0.4, 0.8, and 1.0 mM. Unless stated otherwise, the SAB was formed in 2.0 M ZnSO₄ with 0.4 mM C₁₄-RWW-NH₂.

MnO₂ and NiHCF cathode synthesis

MnO₂ was fabricated by a hydrothermal method. Typically, 1.103 g of Mn(CH₃COO)₂ was added into 40 ml of deionized water. After adding 10 ml of 0.1 M KMnO₄, the mixture was stirred for 2 hours and then transferred to an 80-ml Telflon-lined autoclave at 80°C for 4 hours. The obtained MnO₂ nanowires were collected by filtration, washed with water, dried, and further annealed at 200°C for 2 hours in air. NiHCF was synthesized through a facile coprecipitation method. A solution of nickel acetate tetrahydrate (0.249 g; 1 mmol) in water (17.5 ml) and DMF (2.5 ml) was added to a solution of hexacyanoferrate decahydrate (0.484 g, 1 mmol) and NaCl (0.7 g) in water (17.5 ml) and DMF (2.5 ml) with stirring. The obtained NiHCF was collected by filtration, washed with water and acetone, and dried at 60°C.

Mass spectrometry

MS measurements were performed using a Bruker Biflex III matrix-assisted laser desorption ionization-time of flight mass spectrometer equipped with a 337 nm nitrogen laser. α-Cyano-4-hydroxycinnamic acid was used as the matrix. The lipopeptide sample was dissolved within the matrix in a mixture of acetonitrile and water (1:1, v/v) containing 1% TFA. A ~0.8-μl sample solution was placed on a metal sample plate and allowed to air-dry at ambient temperature. Mass spectra were acquired in positive linear mode using an acceleration voltage of 20 kV. External mass calibration was performed using a standard lipopeptide mixture. Spectra were obtained by setting the laser power close to the threshold of ionization and 120 pulses were acquired and averaged. Figures S1 to S3 present the MS spectra of the synthesized lipopeptides. The observed molecular masses were well consistent with the calculated ones, as follows:

1) C₁₀-RWW-NH₂:

expected masses [M + H] + =685.6;

observed masses [M + H] + =699.9.

2) C₁₄-RWW-NH₂:

expected masses [M + H] + =741.8;

observed masses [M + H] + =756.0.

3) C₁₈-RWW-NH₂:

expected masses [M + H] + =797.6;

observed masses [M + H] + =784.0.

Except the distinguished single-charge molecular ion peaks, no other peaks as well as fragmental ion peaks were observed.

Reverse-phase high-performance liquid chromatography analysis

Reverse-phase HPLC analysis was performed using a Waters 2695 Alliance HPLC system at 25° ± 2°C. After filtration through a 0.22-μm filter, the lipopeptide aqueous solution was directly injected into a C18 reversed-phase column (4.6 mm by 150 mm). A gradient elution mode was used: eluent A, 0.1% (v/v) TFA in water, 0 → 1 min, 95% (A%), 1 → 20 min, 95% → 5% (A%), 20 → 25 min, 5% → 95% (A%); eluent B, 0.1% (v/v) TFA in acetonitrile. Other analytical conditions were as follows: monitoring at 214 nm; flow rate of 0.8 ml min⁻¹. The HPLC profiles of the synthesized lipopeptides are presented in figs. S4 to S6. The peaks at 11 to 14 min correspond to the lipopeptides, with a relative area of >98%.

Critical micelle concentration determination

The critical micelle concentrations of the lipopeptides were determined by ionic conductivity measurements of 2.0 M ZnSO₄ solutions of the lipopeptide at different concentrations using a conductometer (Model DDS-307 from Shanghai Leici Corporation, China) equipped with a model DJS-1C conductivity cell. The cell constant of 1.05 cm⁻¹ was obtained by calibration using potassium chloride standards (0.01 M). The solutions and the measurement cell were immersed in an electronically controlled water bath, with the temperature maintained at 25.0° ± 0.1°C.

Atomic force microscopy

Samples were prepared by dropping 10 μl of a 2.0 M ZnSO₄ solution of the lipopeptide onto a freshly cleaved mica surface for a few seconds. The mica surface was extensively rinsed with water and dried gently using an ultrapure nitrogen gas. To prevent impurity adsorption on the prepared surface, atomic force microscopy (AFM) imaging was performed immediately after all traces of solvent disappeared. AFM measurements were performed with a Nanoscope IVa MultiMode AFM (Digital Instruments, Santa Barbara, CA) in the tapping mode using a tapping mode^TM etched silicon probe (Veeco, Santa Barbara, CA) with a nominal spring constant of 42 N m⁻¹ and a typical frequency of 300 kHz, under ambient conditions, at 512 × 512-pixel resolution, integral and proportional gains of 0.1 to 0.2 and 0.2 to 0.3, respectively, and a scanning speed of 1 Hz.

Spectroscopic ellipsometry

The samples were prepared by immersing the clean silica substrate in a 2.0 M ZnSO₄ solution of the lipopeptide for 30 min, followed by gentle drying with ultrapure nitrogen gas. The measurements were performed using a Jobin-Yvon UVISEL spectroscopic ellipsometer over a typical wavelength range of 300 to 600 nm, with the incident light at 70°. The experimental data were analyzed to obtain the thickness using the DeltaPsi2 software developed by Jobin-Yvon (38).

Neutron reflection

Neutron reflectivity measurements were carried out on SURF at Rutherford Appleton Laboratory, Oxford, UK. The wavelength of the neutron beam used ranged from 0.5 to 6.5 Å. Measurements were performed by clamping a stainless-steel trough against the polished face of a silicon <111> block with dimensions of 6 cm by 5 cm by 1.2 cm. The sample cell required ~2 ml of the lipopeptide solution in water (H₂O) or heavy water (D₂O) containing 2.0 M ZnSO₄ to fill it. The sample cell was mounted on a goniometer stage linked to a control computer terminal. The neutron beam entered the small face of the silicon block, was reflected from the solid/solution interface, and exited from the opposite end of the small face. The neutron beam was defined by two sets of horizontal and vertical slits placed before the sample cell, providing a typical beam illuminated area of approximately 5 × 3.5 cm². Each reflectivity was carried out at three incidence angles of 0.35°, 0.8°, and 1.8° and the resulting reflectivity profiles were combined to cover the wave vector (κ) range between 0.012 and 0.5 Å⁻¹. Reflectivity profiles below the critical angle were theoretically equal to unity and all the data measured were scaled accordingly. Constant background was subtracted using the average reflectivity between 0.3 and 0.5 Å⁻¹. The background was found to be typically ~2 × 10⁻⁶ in H₂O or D₂O. The measured neutron reflectivity profiles were analyzed using model fits based on the optical matrix formula. The fitting usually started with the assumption of a model interfacial structure, followed by calculation of the reflectivity. The calculated reflectivity was then compared to the measured one. The process was iterated until a good fit was obtained.

Microscopy characterizations

For transmission electron microscopy measurements, a 10-μl aliquot of a 2.0 M ZnSO₄ solution of the lipopeptide was adsorbed onto a copper grid coated with carbon and formvar layer (400 mesh, Plano GmbH). After 2 min of drying, the grids were stained with 10 μl of 2–wt % uranyl acetate solution for 2 min and washed with water. Measurements were performed on a JEM2100F electron microscope with 120-kV acceleration voltage. The surface morphology of the pristine Zn and SAB-Zn before and after Zn plating/stripping was observed by SEM (Nova NanoSEM450, Japan) and optical microscopy (YUESCOPE YM10R, China).

Spectroscopy characterizations

FTIR spectra were recorded on a Nicolet iS20 FTIR spectrometer ranging from 4000 to 400 cm⁻¹ under ATR mode. Raman spectra were conducted on a Thermo Fischer DXR Microscope. The ²H-NMR spectra were obtained on a Bruker AVANCE III HD 500 NMR spectrometer with deuterated D₂O as the field frequency lock. X-ray photoelectron spectroscopy (XPS) measurements and depth profile XPS were performed on a Versa probe III (PHI 5000) spectrometer. Analysis was performed using CASA XPS. All the XPS spectra were calibrated to the adventitious hydrocarbon carbon peak at 284.8 kV.

X-ray diffraction

XRD spectra and pole figures were recorded using a Bruker D8 Advance x-ray powder diffractometer (Bruker) equipped with Co Kα radiation (λ = 1.7903 Å), at room temperature with a scan range 2θ of 10° to 90° and a count of 2 s. The Zn foil was placed on a standard flat sample reflection holder. Data collection and analysis were performed using the MDI Jade software.

Contact angle measurements

The OCA50AF Contact Angle System was used to measure the contact angle of the electrolytes on Zn foils.

Zeta potential

Zeta potential measurements were conducted using a Zetasizer Nano ZS, Malvern Instruments. Unless stated otherwise, the powdered Zn was dissolved in different electrolytes to a concentration of 1 mg ml⁻¹ and further dispersed for 30 min at room temperature. Just before the measurements, 600-μl solution was injected in 1-ml disposable folded capillary cells (DTS-1060, Zetasizer Nano series, Malvern). The zeta potential of the SAB-Zn was derived from the electrophoretic mobility based on the Smoluchowski formula. All measurements were averaged over two individually performed experiments and conducted in triplicates for each sample.

First-principle calculations

First-principle calculations were performed using the Vienna ab initio simulation package (VASP) (39, 40) based on DFT. Core electrons are described using the projector-augmented-wave method (41) with the standard Perdew-Burke-Ernzerhof (42) of generalized gradient approximation, as exchange-correlation functional. The plane-waves of kinetic energy were set at 400 eV. The first Brillouin zone of slab calculations was sampled using a Γ-centered k-point grid of 2 × 2 × 1 for geometry optimization. To consider the interactions between the molecules, the Grimme’s method (DFT-D3) (43) was used to account for the dispersion interactions. For the adsorption model, a supercell (13.85 Å by 15.99 Å) of Zn(0002) surface with four zinc layers keeping the two bottom layers fixed to describe the bulk properties and 15 Å vacuum layer were selected for simulation. Furthermore, the climbing image nudged elastic band (CI-NEB) method (44) implemented in VASP was used to determine the migration energy barrier for Zn atoms within the different layers. To observe the migration of zinc atoms within the bulk phase, only the bottom layer zinc atoms were fixed during CI-NEB method calculations. The adsorption energy (E_ads.) between Zn slab and different molecules was defined as follows

$$E_{\text{ads}.}=E_{\text{Zn}-\text{slab}+\text{molecules}}-E_{\text{Zn}-\text{slab}}-E_{\text{molecules}}$$

where E_{Zn−slab+molecules} is the total energy of the adsorption configurations, E_molecules is the energy of adsorbates and E_Zn−slab is the energy of the Zn(0002) surface.

Ab initio and CMD simulations

AIMD simulations were performed using the VASP code of a single Trp molecule on a Zn(0002) slab consisting of four layers of 6 × 6 atoms each, in a periodic unit cell with dimensions of 13.85 Å by 15.99 Å by 22.42 Å. Owing to the large size of the supercells, only the Γ point was used. The initial structures were statically relaxed at 0 K. For all of the AIMD trajectories, the time step was set to 1 fs and the initial 2 ps (2000 steps) was heated from 100 to 400 K at a constant rate by velocity scaling. NVT ensemble (fixed number of atoms, volume and temperature) using Nosé-Hoover thermostat (45, 46) was adopted in the equilibrium 400-K temperature simulation. AIMD equilibrium simulations were then performed for 20 ps (20,000 steps). VESTA (47) and VMD (48) were used to display all crystal structures and the post-processing VASPKIT (49) code was used for the VASP calculated data. CMD simulations were conducted by the Forcite module in Materials Studio to explore the migration of Zn ions comprising 498 water molecules, 50 ZnSO₄ units, and 2 C₁₄-RWW-NH₂ lipopeptides. The simulation was performed within a periodic box with dimensions of 55 Å by 30 Å by 30 Å. The COMPASS III force field was used within the simulation system (50). First, the geometry optimization algorithm was used with a combination of steepest descent, adopting basis Newton-Raphson, and quasi-Newton methods based on the convergence of the total energy (0.001 kcal mol⁻¹) with a force of 0.5 kcal mol⁻¹ Å⁻¹. Following geometric optimization, a 500-ps NPT (constant number of particles, pressure, and temperature) pre-equilibrium simulation was initiated under a pressure of 0.1 MPa, maintaining a constant temperature of 298 K using the Andersen thermostat (51). Pressure was controlled by the Berendsen barostat (52). Subsequently, a 500-ps equilibrated NVT (constant number of particles, volume, and temperature) simulation was conducted under a pressure of 0.1 MPa, using the Nosé-Hoover thermostat to sustain the system at 298 K (45, 46). Throughout the simulations, a time step of 1 fs was used to strike a balance between computational efficiency and accuracy. For MSD data acquisition, the region with high linearity (100 to 400 ps) was used in the equalized NVT simulation. The Ewald scheme was used to calculate long-range electrostatic interactions, and atom-based van der Waals interactions with a 12.5-Å cutoff distance in dynamic simulations.

Quantum chemical calculations

Quantum chemical calculations were performed using the Gaussian 16 software. Geometry optimizations were carried out with the B3LYP/6-311G(d,p) basis set (53) to allow accuracy in describing molecular structures. To account for noncovalent interactions, the Grimme D3 dispersion correction with Becke-Johnson damping (GD3BJ) was applied (54). Solvation effects were considered using the Solvation Model based on Density (SMD) (55) with water as the solvent, using appropriate solvent parameters. To mitigate the Basis Set Superposition Error (BSSE) (56), counterpoise-corrected calculations were conducted to calculate the binding energy as follows

$$E_{\text{BE}}=E_{\text{AB}}-(E_{\mathrm{A}}+E_{\mathrm{B}})-E_{\text{BSSE}}$$

where E_AB, E_A, and E_B are the total energies of the AB complexes, bare A, and bare B, respectively. E_BSSE is the BSSE correction energy, which is used to correct the energy of interaction in all the complexes.

Electrochemical measurements

A multichannel battery cycler (Neware) was used to evaluate the electrochemical performance. All coin cells (CR2025-type) were fabricated by sandwiching a GF/D separator (Φ 20 mm) between the anode and cathode chips (Φ 16 mm) with 30-μl electrolyte. A three-electrode system was fabricated using a 1.0 cm by 3.0 cm Ti working electrode with a larger area Zn counter electrode and a 3.0 M Ag/AgCl reference electrode at room temperature and pressure. Cells containing Zn|Ti, Zn|Zn, Zn|MnO₂, and Zn|NiHCF were tested in aqueous 2.0 M ZnSO₄ and 2.0 M ZnSO₄ with 0.4 mM C₁₄-RWW-NH_2. Linear sweep experiments were conducted for 0.4 mM C₁₄-RWW-NH₂ in 2.0 M Na₂SO₄ and 2.0 M Na₂SO₄. The half cells of Zn|Ti were operated with Zn plating capacity of 1 or 10 mAh cm⁻² followed by stripping Zn with a cutoff voltage (1 V) for every cycle at current densities of 1 or 10 mAh cm⁻², respectively. The Zn|Zn symmetric cells were fabricated with an areal capacity of 1, 5, or 10 mAh cm⁻² at a current density of 1, 10, or 20 mA cm⁻², respectively, at room temperature, 25°C. All full cells were cycled from 0.8 to 1.8 V with initial five activation cycles at 0.1 C and subsequent long cycling at 1 C (charging) and 1 C (discharging). MnO₂ was used as the cathode and Zn foil [20 μm; ~11.7 mAh cm⁻² for the cell with a low capacity ratio of the negative electrode to the positive electrode (N/P ratio)] was used as the anode in full cells. An additional 0.1 M MnSO₄ was added to the electrolyte to prevent Mn²⁺ dissolution. The cathode was composed of 95–wt % MnO₂, 1–wt % polyvinylidene fluoride (PVDF) binder, and 4–wt % Super P. The electrolyte used in each cell was controlled as 15 μl [for the cell with a low electrolyte weight to cathode capacity (E/C) ratio], and the areal mass loading of the MnO₂ cathode was ~12.7 mg cm⁻² (cell with low N/P ratio). The rate performance was examined through the Zn|MnO₂ full cells cycled at various rates ranging from 0.8 to 5 C (1 C = 0.5 A g⁻¹). In addition, full cells using NiCHF as the cathode and Zn foils (20 μm) as the anode were assembled to further evaluate the cycling stability. The cathode was composed of 80–wt % NiHCF, 10–wt % PVDF binder, and 10–wt % Super P. The electrolyte volume was 30 μl, and the mass loading of NiHCF cathode was ~2 to 3 mg cm⁻². All the full cell evaluations were implemented at room temperature. The Coulombic efficiency and galvanostatic discharge-charge performances of the half cells and full cells were evaluated using a Neware battery test system (CT-4008) at room temperature. Other electrochemical measurements were carried out using a CHI660D electrochemical workstation.

Supplementary Materials

This PDF file includes:

Figs. S1 to S49

Tables S1 to S5

Legends for movies S1 and S2

References

Other Supplementary Material for this manuscript includes the following:

Movies S1 and S2