Deciphering multi-dimensional interfacial mechanisms via organic cosolvent engineering for sustainable zinc metal batteries

Abstract

Introducing organic cosolvent is a common and cost-effective electrolyte engineering for aqueous Zn-battery, reshaping the solvation environment of electrolyte and modulating the interfacial electrochemistry on Zn-metal electrode. Clarifying the mechanisms governing interfacial dynamic evolution and electrochemical performance is essential for guiding cosolvent selection. However, the absence of direct visualization for dynamic interfacial evolution during Zn plating/stripping has impeded mechanistic understanding of cosolvent-mediated effects in electrolyte engineering. Here, we combine advanced in-situ spectroscopy with theoretical calculation to decouple the interfacial evolution at the molecular level. We find that cosolvents not only weaken the connectivity of the interfacial hydrogen-bond network between water molecules, thereby hindering the H⁺ transfer, but also accelerate the interfacial dynamic transition of Zn²⁺-(de)solvation from transient to steady state. Additionally, we observe a dynamic adsorption substitution between cosolvent and water, which weakens the electric field intensity exerted on interfacial water. Furthermore, we demonstrate that cosolvents can modify the components content and distribution of the passivation-layer via indirect regulation pathway, rather than a typical self-decomposition mechanism. These multidimensional insights bridge the knowledge gap in cosolvent functionality, offering rational principles for tailoring solvation structures and interfacial dynamics in next-generation aqueous batteries.

Introducing cosolvent is a common approach in aqueous zinc metal batteries but lack of direct visualization of dynamic interface during Zn-plating/stripping makes mechanistic understanding difficult. Here, authors decouple the interfacial evolution at molecular level and decipher multi-dimensional mechanisms experimentally and theoretically.

Introduction

Aqueous batteries possess research value and development potential due to their safe and convenient operating environment, as well as their long cycle life. Among these, aqueous zinc-ion batteries (AZBs) exhibit higher chemical stability and theoretical energy density, making them important candidates for research and development of aqueous secondary batteries^1,2. However, challenges, including the hydrogen evolution reaction (HER), oxygen evolution reaction (OER), and subsequent side reactions caused by interfacial water ionization, seriously affect the battery performance^3–5, hindering their further development. The primary scientific problem faced by Zn-metal electrodes is the competition for electrons between HER and Zn deposition reaction (ZDR). The dynamics of the two reactions directly affect the stability and reversibility of Zn-plating/stripping on the substrates.

Numerous electrolyte modification strategies have been developed to regulate the dynamics between HER and ZDR, such as high-concentration electrolytes^6,7, deep eutectic electrolytes⁸, weakly solvating electrolytes⁹. Among these approaches, cosolvent-based electrolytes have received widespread attention due to their economically efficient^10–13. However, a multitude of cosolvents classified as sulfoxides (dimethyl sulfoxide/DMSO)^14–16, nitriles (acetonitrile/AN)^17–19, alcohols (polyethylene glycol/PEG)²⁰ and amides (N-methyl formamide/NMF)²¹ based on organic functional groups manifest diverse electrochemical behaviors, heightening the analysis complexity. Meanwhile, researching the cosolvent from different dimensions reveals its diversity. Taking AN as an example, it was reported to influence Zn-deposition morphology through adsorption on the electrode surface rather than by altering the Zn²⁺ solvation structure¹⁹. It can regulate the conductivity of electrolytes by different AN-H₂O volume ratios²². In addition, in the pure AN solvent system, it can reduce the cycling overpotential and improve the reversibility of Zn-plating/stripping¹⁸. Taking DMSO as an example, it was reported that it can inhibit the decomposition of H₂O by changing the solvation structure of Zn²⁺ and forms a functionalized solid electrolyte interphase (SEI) on the electrode¹⁴. Moreover, the adsorption of DMSO on the electrode surface and reconstructing the H-bond environment with H₂O also affect the behavior of HER and ZDR^23,24. Nevertheless, many related researches and explanations remain in the theoretical level, lacking direct experimental evidence and establishing their connection to battery performance. Especially, visualization of the interfacial dynamic-evolution during Zn²⁺-(de)solvation, and evidence for cosolvent self-decomposition to participate in the SEI formation are challenging.

Herein, this work follows the research approach that structure determines properties, and properties determine behaviors. Using DMSO and AN as examples, we conduct a systematic and comprehensive study of the similarities and differences in the structure-property-behavior between the cosolvents, with the aim of guiding us in selecting appropriate cosolvent to design electrolytes based on specific scenario requirements. By combining multi-characterization techniques, including in situ Fourier transform infrared spectroscopy (FTIR), in situ online electrochemical mass spectrometry (OEMS), and theoretical calculations, etc., the regulatory mechanisms of cosolvents at the electrode/electrolyte interface are comprehensively analyzed, further establishing their connection to battery performance. Under the prerequisite that cosolvent and anion located in the second solvation shell of Zn²⁺, we decouple the synergistic effect on Zn²⁺-electrodeposition processes in organic cosolvent-related electrolytes from the following (1)-(5) dimensions. (1) In the H-bond section, the introduction of cosolvent disrupts the H-bond between H₂O and H₂O and reconstructs a stronger/weaker one between H₂O and cosolvent, which both weaken the connectivity of the H-bond network in H₂O molecules to impede H⁺ transfer, thereby suppresses the HER kinetics. (2) In the (de)solvation section, we reveal that cosolvents enhance the strength of the interfacial H-bond under the instantaneous potential at the abrupt-change stage. Also, they both promote the average (de)solvation degree of Zn²⁺ and accelerate the interface transition from transient to steady state during the gradual-change stage. (3) In the adsorption section, we observe a dynamic adsorption substitution between H₂O and cosolvent at the interface, which weakens the electric field intensity exerted on interfacial water, lowering its ionization probability to suppress HER. (4) In the passivation section, we demonstrate that cosolvents promote the formation of a Zn²⁺-rich layer on the electrode surface. DMSO and AN follow a regulation mechanism rather than a self-decomposition mechanism, modulating the properties and distribution of the original components (ZnO/Zn(OH)₂ and Zn₄(OH)₆SO₄·xH₂O (ZHS)). (5) In the charge-transfer section, cosolvents control the competition for electrons between HER and ZDR. DMSO accelerates ZDR kinetics more than AN. Multi-dimensional investigation on the structure-property-behavior of cosolvent provides valuable insights and guidance to design the cosolvent-based aqueous electrolyte according to specific application scenarios.

Results

Prerequisite and principle of cosolvent selection

The most important aspect of electrolyte design is determining the composition and corresponding content. Structure determines properties, and properties determine behaviors. For the design of organic cosolvent-based electrolytes, among thousands of organic compounds, cosolvent selection can be guided by the principle that different functional group structures (e.g., -O-H, -S = O, -C ≡ N) exhibit different properties (e.g., stability, polarity, solubility), which in turn can regulate specific behaviors (H-bond, desolvation, adsorption, passivation, charge-transfer) (Fig. 1a). After selecting an appropriate cosolvent, determining its additive amount becomes a key parameter, serving as the investigation prerequisite. Different mass/molar/volume ratios of a solvent-cosolvent system significantly influence the interactions among species in the electrolyte, thereby resulting in different solvation environments.

Fig. 1. Prerequisite and principle of cosolvent selection.

a Structures and properties of different cosolvents, and their multi-behaviors during Zn²⁺-electrodeposition. b ¹H NMR spectra of solvent-cosolvent system with different volume ratios of DMSO/AN mixed with H₂O. The interaction between solvent and cosolvent is changed at the reconstruction range. c The Zn²⁺-O/DMSO radial distribution function of 15DMSO and 50DMSO electrolytes (solvent-cosolvent-salt system) that 1 mol ZnSO₄ was dissolved in 1 L solvent-cosolvent system with different volume ratios. DMSO gradually enters the first solvated shell (P1) of Zn²⁺ with volume ratio increase. The inset shows the main solvation clusters of 15DMSO electrolytes.

To gain a comprehensive understanding of the similarities and differences in the structure-property-behavior of cosolvents, we conducted a systematic investigation by using DMSO and AN cosolvents as examples due to their simple molecular structures, good solubility in water, and significant advantages in performance enhancement (Supplementary Tables 1, 2 and Supplementary Notes 2, 3). Firstly, through analyzing the FTIR spectra of solvent-cosolvent systems with different volume ratios of cosolvent (V_(cosol.)), it was found that the interaction between solvent and cosolvent is altered (Supplementary Figs. 1, 2). In addition, the NMR spectra further reveal the covalent environment between them, as V_(cosol.) increased, the chemical shift of ¹H NMR initially shifts to a downfield, and the dipole-dipole cluster of 2solvent-1cosolvent is dominant. Inversely, the ¹H NMR chemical shift moves to upfield with a predominance cluster of 1solvent-2cosolvent as V_(cosol.) continues to increase. For DMSO, the reconstruction range is at 15% < V_(cosol.) < 30%, whereas for AN, it is at 0% < V_(cosol.) < 15%. (Fig. 1b and Supplementary Fig. 3). It indicates that the interaction between the solvent and the cosolvent is nonlinear, with a turning point occurring at the reconstruction range. Considering the inherent advantages of aqueous electrolytes, such as low viscosity, high safety, and non-flammability, along with the regulatory mechanisms of cosolvents (discussed in five dimensions below), we conducted the research with V_(cosol.) = 15%, at which point H₂O and DMSO predominantly exist in a 2H₂O-1DMSO structure, while H₂O and AN mainly form a 1H₂O-2AN structure.

Radial distribution function/g(r) was employed to investigate the solvation structure in a solvent-cosolvent-salt system. As the proportion of cosolvent increases, the cosolvents gradually enter the first solvation shell of Zn²⁺ (Fig. 1c, Supplementary Figs. 4, 5 and Supplementary Data 1). At V_(cosol.) = 15%, that 1 mol ZnSO₄ was dissolved in 1 L solvent-cosolvent with a volume ratio of 15(cosolvent):85(solvent) (denoted as 15DMSO or 15AN), DMSO/AN has little effect on the solvation structure of Zn²⁺ due to the slight change of the ⁶⁷Zn NMR chemical shift (Supplementary Fig. 6). DMSO/AN is located outside the first solvation shell of Zn²⁺ (P1, 0.20 nm) but is involved in the second one (P2, 0.42 nm). And the primary solvation clusters in the electrolyte are [Zn(H₂O)₆]²⁺, [Zn(H₂O)₅]²⁺, and [Zn(H₂O)₄]²⁺ (Fig. 1c, Supplementary Figs. 7–9 and Supplementary Data 1). Thus, the force field between Zn²⁺ and cosolvents is weak, which indicates the influence of Zn²⁺ solvation structure regulation on battery performance is nonsignificant. With this foundational understanding, we proceed to investigate the regulation mechanisms of organic cosolvent on the stability and reversibility of Zn-plating/stripping during Zn²⁺-electrodeposition from five dimensions.

Relation between H-bond and HER dynamics

According to the proton donor-acceptor mechanism, cosolvents as proton-acceptors can form H-bonds with proton-donor/H₂O^25,26. The H-bond energy between DMSO and H₂O, H₂O and H₂O, AN and H₂O is – 0.47 eV, – 0.24 eV, and – 0.20 eV, respectively (Fig. 2a), which indicates that the introduction of DMSO can enhance the H-bond strength of electrolyte environment, but AN can weaken that. Furthermore, using cosolvents as the center of dipole-dipole interaction, the distances between DMSO and H₂O, AN and H₂O are 4.8 and 4.0 Å, with coordination numbers of 24 and 17, respectively. It means that DMSO can bind more H₂O to form an H₂O-rich cluster, and more significantly disrupt the H-bond network among H₂O (Fig. 2b). According to the fitting results of stretching vibration of the O-H band (ν(O-H)), the proportions of strong H-bonds/S-H are 58.6% (Bare), 60.5% (15DMSO), and 56.3% (15AN), respectively (Fig. 2c, d and Supplementary Figs. 10, 11). It further demonstrates that DMSO forms the stronger H-bond with H₂O, while AN forms weaker. In addition, ¹⁷O NMR spectra were conducted by an external standard method to compare the noncovalent/covalent environment in different binary/ternary systems (Fig. 2e). The chemical shift observed in the H₂O-DMSO system (15% DMSO) shifts to a downfield (4.1 ppm) compared to the pure water system (3.5 ppm), while that in the H₂O-AN system (15% AN) shifts to an upfield (3.3 ppm). In stark contrast, after adding the solute (ZnSO₄), a significant shift to downfield is observed in both H₂O-DMSO/AN-ZnSO₄ (15DMSO (5.4 ppm)/15AN (4.6 ppm)) systems, which suggests that Zn²⁺ and SO₄^2- profoundly disrupt the H-bond network of H₂O, exceeding the effect caused by the cosolvents alone. Specifically, Zn²⁺ forms a coordination bond with H₂O, and SO₄^2- forms S-H with H₂O simultaneously, which markedly reduces the electron cloud density around the oxygen nucleus of H₂O. As a result, difference H-bond environment can further affect HER dynamics. As shown in Fig. 2f, the HER evolution in the Zn | |Na₂V₆O₁₆·xH₂O cell at 1 A g^–1 was investigated using in situ OEMS. For the Bare electrolyte, the hydrogen intensity is relatively low and gradually increases during Zn-stripping (Zn²⁺-solvation), while the hydrogen intensity significantly increases during Zn-plating (Zn²⁺-desolvation). As an obvious contrast, in the 15DMSO and 15AN systems, no hydrogen signal was detected throughout the entire discharge/charge process due to HER dynamics being strongly suppressed by the reconfigured H-bond environment.

Fig. 2. Relation between H-bond and HER dynamics.

a The H-bond interaction energy between different donors and acceptors. b The radial distribution function (g(r)) and coordination number (N(r)) of cosolvent to H₂O. c The ex situ FTIR spectra and the corresponding (d) proportion of weak H-bond/W-H and strong H-bond/S-H. e ¹⁷O NMR spectra in different solution systems. 15%DMSO/15%AN is a solvent-cosolvent system with the volume ratio of 85(solvent):15(cosolvent), and 15DMSO/15AN is a solvent-cosolvent-salt system that 1 mol ZnSO₄ was dissolved in 1 L 15%DMSO/15%AN. The inset is the diagram of the external standard method for NMR. f The H₂ evolution in Zn | |NVO cells with the three electrolytes at the first cycle. g Evolution schematic of the H-bond between water molecules after the introduction of cosolvent and salt.

As shown in Fig. 2g, H₂O molecules associate through H-bond in pure water solution (H-bond I), and H⁺ is conducted through the H-bond network via a hopping mechanism (Grotthuss mechanism)^27,28. The introduction of cosolvents disrupts the original H-bond and reconstructs a stronger/weaker one between water and cosolvent (H-bond II), and DMSO can form H₂O-rich cluster to disrupt the original H-bond network in a greater extent. Furthermore, with the addition of solute, the coordination environment between Zn²⁺ and H₂O is established, and the H-bond between H₂O and SO₄^2– (H-bond III) is reconstructed, heavily obstructing the H⁺ transmission. Typically, it is an entropy increase process as cosolvent (molecules) and solute (anions and cations) are introduced into a pure water system, accompanying a disorder-increased and connectivity-reduced of the original H-bond, which is also the key factor in inhibiting the HER dynamics during ZDR and is not only determined by the reconstructed H-bond strength.

(De)solvation evolution under interfacial electric field

In situ FTIR is employed to analyze the dynamic interface evolution and reveals two distinct stages during Zn²⁺-electrodeposition (Fig. 3a and Supplementary Figs. 12–14). The first stage (denoted as abrupt change) corresponds to the interface reconstruction upon potential application, quantified by the absorbance difference of ΔA(t_x) = A(t_x) – A(t₀) (relative absorbance). The second stage (denoted as gradual change) reflects Zn²⁺-(de)solvation evolution, expressed by ΔA(t_y) = A(t_y) – A(t_x)²⁹. Through the potential/time-dependent FTIR differential spectra, the abrupt change stage is first analyzed in Fig. 3b. Upon instantaneous potential application, the interface has been greatly disturbed. In the relative absorbance of ΔA(t_x), the O-H and S-O stretching vibration bands of H₂O and SO₄^2– (ν(O-H) and ν_as(S-O)) both exhibit a pair of alternating peaks, corresponding to W-H and S-H, solvent-separated ion pair (SSIP) and contact ion pair (CIP)^30,31. In the Bare electrolyte, the H-bond strength is significantly decreased with W-H dominating. The solvation effect of SO₄^2- on Zn²⁺ is relatively weak, primarily manifesting as SSIP, resulting in a weaker covalent interaction at the interface. The instantaneous interfacial state in the 15DMSO system is similar to that in the Bare electrolyte, but with stronger H-bond interaction, and also stronger electrostatic interaction between SO₄^2– and Zn²⁺. However, in the 15AN system, the interface exhibits a completely different state. That is, the interfacial H-bond strength is enhanced with a distinct reversal (even though the H-bond strength in the bulk is the weakest), and SO₄^2– has a greater influence on Zn²⁺-solvation, resulting in stronger covalent interaction at the interface.

Fig. 3. (De)solvation evolution under interfacial electric field.

a Schematic of electrochemical in situ FTIR to monitor the dynamic interface evolution during Zn-plating/stripping. The process can be decoupled into two stages, abrupt change of interface reconstruction (expressed as ΔA(t_x) = A(t_x) − A(t₀)) and gradual change of (de)solvation evolution (expressed as ΔA(t_y) = A(t_y) − A(t_x)). b The relative absorbance (ΔA(t_x)) comparation of three electrolytes under Zn deposition potential. The ΔA(t_y) evolution of ν(O-H) and ν_as(S-O) during (c) Zn²⁺-desolvation and (d) Zn²⁺-solvation. Insets in (c) and (d) represent the Zn²⁺-(de)solvation degree at t₂₇₀ and t₁₂₉₀. e The voltage-time profile during Zn²⁺-(de)solvation. And the ΔA(t_y) evolution of ν(O-H) obtained from (c) and (d). f The distribution probability of Zn²⁺-H₂O clusters with different coordination numbers at the interface and the LUMO energy level of the clusters. Inserting the stepwise desolvation process of Zn²⁺ and the corresponding desolvation energy.

The reconstructed interfacial state under instantaneous potential is significant to subsequent desolvation evolution. Following the abrupt change stage, the gradual change stage is further analyzed. During the desolvation (0 ~ 900 s), the ΔA(t_y) of ν(O-H) and ν_as(SO₄^2–) shows a clear downward trend. At the same desolvation time (e.g., 270 s), the average desolvation degree varies in different systems (Fig. 3c). Specifically, in cosolvent systems, ΔA(t₂₇₀) related to ν(O-H) decreases more significantly, and the decreased values are 0.015 (Bare), 0.031 (15DMSO) and 0.051 (15AN), respectively. It corresponds to a greater extent of H₂O to be desolvated and disperse from the interface to bulk electrolyte, suggesting that cosolvents facilitate Zn²⁺-desolvation. Simultaneously, SO₄^2– is also repelled away from the interface. Note that species (e.g., Zn²⁺) persist to be supplied (S) and consumed (C) at the dynamic interface simultaneously (Fig. 3e, top). Even though the desolvation process continues, the state at the interface would not change continuously. Instead, it would transfer from a transient state (I, S ≠ C) to a steady state (II, S = C), where multi-forces (maybe species amount) acting on the species reach a balance. This phenomenon can be shown by the change of relative absorbance (e.g., ΔA(t_y) of ν(O-H)) decreased to a certain extent, and then changed slightly and maintained stability (Fig. 3e, t ≤ 900 s). During the solvation (900 ~ 1800 s), the changes in relative absorbance of ν(O-H) and ν_as(S-O) exhibit a reversal and gradual increase compared to the desolvation process (Fig. 3d). After returning to the initial state (ΔA(t_y) ≈ 0) (III, C ≠ S), ΔA(t_y) continues to increase and eventually reaches a new steady-state (IV, C = S) (Fig. 3e, 900 s <t). Similar to the desolvation process, the average solvation degree varies in different systems at the same solvation time (e.g., 1290 s). Specifically, in the presence of cosolvent, the ΔA(t₁₂₉₀) of ν(O-H) increases more significantly, indicating that a greater amount of H₂O migrates/gathers to the electrode surface to participate in the Zn²⁺-solvation, forming solvated H₂O. In addition, SO₄^2– is attracted closer to the interface by the positive electrode potential and pre-solvated Zn²⁺ (Fig. 3d)³². So, these results demonstrate that cosolvents accelerate the Zn²⁺-solvation.

To further understand the dynamic interfacial evolution, theory calculation (density functional theory (DFT) and molecular dynamics (MD)) is employed to demonstrate that the solvated Zn²⁺ is not completely desolvated, with the primary solvation structures at the interface being [Zn(H₂O)₆]²⁺, [Zn(H₂O)₅]²⁺ and [Zn(H₂O)₄]²⁺ (Fig. 3f and Supplementary Figs. 15, 16). The [Zn(H₂O)₅]²⁺ cluster has the highest lowest unoccupied molecular orbital (LUMO) energy level (0.49 eV), resulting in forming the most stable structure to maximally inhibit H₂O ionization under the interfacial electric field, thereby preventing HER. It also indicates that the removal of one H₂O to form [Zn(H₂O)₅]²⁺ cluster during the desolvation process is beneficial to the whole Zn²⁺-electrodeposition. Moreover, by calculating the desolvation energy for the stepwise removal of H₂O, it evidences that removing the first H₂O is most difficult, needing the desolvation energy of 0.60 eV, which is the rate-determining step of the whole desolvation process. As a result, cosolvents facilitate the transition of H₂O towards a more stable state with more [Zn(H₂O)₆]²⁺ converts to [Zn(H₂O)₅]²⁺ in 15DMSO and 15AN, promoting the slowest step of the stepwise desolvation processes and further inducing subsequent rapid desolvation steps.

Dynamic adsorption substitution

The adsorption energy evidences that DMSO, AN, and H₂O all exhibit the adsorption property on the Cu-electrode surface to a certain extent. Notably, DMSO is the most easily absorbed with the S = O bond facing downward being the most favorable (– 1.23 eV) (Supplementary Fig. 17 and Supplementary Data 2). Figure 4a and Supplementary Fig. 18 show the amount and distribution of each species within the adsorption layer under Zn deposition potential (– 0.15 V vs. Zn²⁺/Zn), which suggests that H₂O is the most abundant species in the adsorption layer. Moreover, the contact angles of three electrolytes on the electrode surface are 82.3° (Bare), 87.2° (15DMSO), and 65.3° (15AN), indicating that 15AN is the most hydrophilic with the highest amount of H₂O at the interface (Supplementary Fig. 19). These results seem to suggest that more interfacial H₂O is ionized to produce H₂ in 15AN, showing an inconsistent result with Fig. 2f. Apparently, the difficulty of interfacial H₂O to be ionized is not only determined by the amount of it, but is closely related to its properties (orientation, bond-strength) (Supplementary Fig. 20), which associates deeply with the force field that H₂O is exerted under intrinsic and external electrical field. The electric field intensity applied to interfacial H₂O is composed of interactions with the electrode, as well as with various species in the adsorption layer (cation, anion, solvent, and cosolvent)^33–35. The field intensity in 15AN is 26 V nm^–1, much less than that in 15DMSO (38 V nm^–1) and Bare (40 V nm^–1) (Fig. 4b, Supplementary Fig. 21 and Supplementary Data 1). A weaker field intensity means that H₂O is exerted with a smaller electric force, making it more difficult to be ionized and undergo HER. Overall, less interfacial water (e.g., 15DMSO) and weaker electric field intensity (e.g., 15AN) are both conducive to inhibiting HER.

Fig. 4. Dynamic adsorption substitution.

a The number of species in the adsorption layer at the Cu/electrolyte interface, insets are the contact angles of three electrolytes on the electrode surface. b The calculated total electric field intensity applied on interfacial water. The field intensity exerted on water is influenced by interactions with the electrode, and various species in the adsorption layer (cation, anion, solvent, and cosolvent). The inset in (b) represents that the low field intensity is benefit to suppress HER. The time-dependent relative absorbance (ΔA(t_y)) evolution of ν(S = O) and ν(C ≡ N) in (c) 15DMSO and (d) 15AN system during Zn²⁺-(de)solvation. e The network of the intrinsic and external electric field in the electrolyte, and a schematic of the Zn²⁺-(de)solvation process and subsequent dynamic adsorption substitution at the interface. The inset in (e) shows the ΔA(t_y) evolution of ν(O-H) and ν(C ≡ N) during the Zn²⁺-desolvation, which represents the dynamic adsorption substitution of H₂O and cosolvent.

In situ FTIR tests further research the evolution of interfacial species that exhibit varying degrees of surface activity at the interface under intrinsic and external electrical fields during Zn²⁺-(de)solvation. During the desolvation process, the relative absorbance ΔA(t_y) of the S = O stretching vibration in DMSO (ν(S = O), including ν_as(S = O) and ν_s(S = O)) and the C ≡ N stretching vibration in AN (ν(C ≡ N), including ν_as(C ≡ N) and ν_s(C ≡ N)) gradually intensify (positive absorption) and tend to stabilize. It indicates that cosolvents are attracted and gradually gather at the electrode/electrolyte interface, and eventually reach saturation/equilibrium. During the solvation process, the absorption intensities of ν(S = O) and ν(C ≡ N) gradually weaken (ΔA(t_y) turns from positive to negative), crossing the initial state and continuously decreasing until the absorption intensity is unchanged. It represents that the cosolvents are repelled and gradually disperse until to reach a new dynamic equilibrium at the interface (Fig. 4c, d). The aforementioned process is similar to the four stages described in Fig. 3e, where both desolvation and solvation processes undergo a transition from a transient state to a steady state.

Figures 3c, d illustrate that, in the cosolvent system, the H₂O at the interface decreases during Zn²⁺-desolvation and increases during Zn²⁺-solvation, manifesting an opposite variation trend to that of the cosolvents in Figs. 4c–4d. The above intriguing phenomenon of mutual increase and decrease may be due to the presence of adsorption substitution. When cosolvents adsorb onto the electrode surface, they displace some adsorption sites where H₂O originally adsorbed, resulting in a decrease in interfacial H₂O and an increase in cosolvent (DMSO/AN). In addition, as the number of interfacial species dynamically changes, the interactions of the active species within the adsorption layer also continuously evolve. The infrared-inactive symmetric stretching vibration of SO₄^2– (ν_s(S-O)) also exhibits a trend similar to ν_as(S-O) upon the (de)solvation processes (Fig. 4c, d). During Zn²⁺-desolvation, part of SO₄^2- moves away from the electrode surface. Owing to the significant repulsive force between SO₄^2– and the cosolvent, the nearby cosolvent molecules are repelled, moving in the opposite direction of the SO₄^2– movement. Conversely, during Zn²⁺-solvation, the SO₄^2– migrating back to the electrode surface repels the cosolvent away from the electrode surface.

Although the chemical properties of the coordination bond and H-bond differ, they are essentially electrostatic interactions resulting from the intrinsic electric field generated between the species^36,37. When the electrolyte is exposed to an external electric field, it further disrupts the intrinsic electric field (Fig. 4e, left). Meanwhile, as the interfacial reactions proceed and the electric field continuously changes, it leads to the dynamic evolution of the interfacial environment from a transient state to a steady state during both the Zn²⁺-desolvation and solvation processes (different (de)solvation degree). Moreover, the adsorption substitution between solvent and cosolvent is beneficial to reduce the electric field intensity applied to interfacial H₂O, thereby affecting the reaction rate that H₂O undergoes ionization to generate hydrogen.

Passivation-layer regulation and charge-transfer competition

In ZnSO₄ electrolyte, inactive compounds such as ZnO and ZHS generate and accumulate on the electrode surface to form a loose, porous, and disordered passivation layer, which cannot effectively block the contact between the electrolyte and the electrode surface to inhibit subsequent side reactions³⁸. Generally, organic cosolvents with weak reduction stability are introduced to reduce preferentially at the electrode surface before Zn²⁺ reduction to form a favorable passivation layer that protects the Zn-metal electrode³⁹. However, experimental evidence that the cosolvent self-decomposition so that the passivation layer contains its decomposition products is lacking (self-decomposition mechanism). Alternatively, the cosolvent simply regulates the composition and structure of the original passivation layer (regulation mechanism).

In order to explore the influence mechanism of DMSO/AN on the passivation layer, firstly, Supplementary Fig. 22 shows that all the LUMO energy level of molecules/ions decreases after interacting with H₂O, demonstrating that the presence of H₂O increases the reduction possibility. Then, the reduction processes were studied using linear sweep voltammetry (Fig. 5a and Supplementary Fig. 23). A slow scan rate of 0.1 mV s^–1 was employed to eliminate the influence of the electrical double layer (EDL) charging current⁴⁰. The electrochemical behavior of the three electrolytes was essentially consistent throughout the entire potential scan range, revealing three distinct reduction stages. Take the 15DMSO system as an example, stage I corresponds to the adsorption of interfacial species⁴¹. When the potential reaches 1.17 V (vs. Zn²⁺/Zn), a reduction current of 5 µA cm^–2 appears and gradually increases to 48 µA cm^–2 at 0.31 V (vs. Zn²⁺/Zn), representing a dynamic change in the adsorption number of interfacial species as the potential varies. Stage II primarily involves the Zn²⁺-electrodeposition⁴², as the potential continues to scan negatively from 0.31 V (vs. Zn²⁺/Zn), the reduction current significantly increases, with the 15DMSO system showing the fastest increase, reaching a maximum value of 1.36 mA cm^–2 at – 0.08 V (vs. Zn²⁺/Zn), indicating the facilitation of ZDR. Stage III is the coupling potential region of HER and ZDR. When the potential continues to negative sweep, HER is gradually activated. The 15DMSO system has the smallest coupling reduction current, corresponding to the HER inhibition maximally. Throughout the entire potential scan range, no obvious reduction current due to cosolvents self-decomposition is observed, but the introduction of cosolvents clearly promotes the competition of Zn²⁺ for electrons and inhibits the ability of H⁺ to acquire electrons.

Fig. 5. Passivation-layer regulation and charge-transfer competition.

a The LSV curves using Ti as the working and counter electrodes, and Ag/AgCl as the reference electrode under 0.1 mV s⁻¹. b High-resolution Zn2p XPS spectra with 0 and 40 nm sputtering of the passivation layer on Zn-electrode of Zn | |Zn symmetric cell after 100 cycles. c The 2D view of ZnO^– (for ZnO) and ZnSO₄OH^– (for ZHS) on Zn-electrode of Zn | |Zn symmetric cell after 100 cycles tested by TOF-SIMS. d The corrosion current density evolution on Zn-electrode, the inset is the corresponding polarization curves. e Comparison of E_a,ct at Zn/electrolyte interface before cycle and after 2 cycles. f Schematic of cosolvent regulation mechanism for passivation-layer.

Zn²⁺-related species are the main components of the interfacial passivation layer. X-ray photoelectron spectroscopy (XPS) analysis of the 0 and 40 nm etch on the Zn-metal surface after 100 cycles reveals a Zn²⁺-rich layer in 15DMSO, with Zn²⁺ content approximately twice that of the Bare (Fig. 5b). In the X-ray diffraction (XRD) spectra, the Zn-metal electrode after 100 cycles exhibits distinct peaks at yellow and purple regions, corresponding to the diffraction peak of ZHS and ZnO/Zn(OH)₂, respectively (Supplementary Fig. 24). The strong peak intensity in 15DMSO further confirms the existence of a Zn²⁺-rich layer, which is mainly composed of ZHS and ZnO. Moreover, AN can regulate the electro-crystallization orientation of deposited-Zn, promoting Zn²⁺ to preferentially orient along the [002] crystal orientation, resulting in parallel deposition morphology on the substrate (Supplementary Figs. 25, 26)⁴³. Time of flight secondary ion mass spectrometry (TOF-SIMS) shows the ZnO^– and ZnSO₄OH^– slices on the electrode surface, in the Bare system, the content of ZnO and ZHS at the X-Y plane is lower, and the distribution of them at X-Z direction is deep, randomly and highly uneven (Fig. 5c and Supplementary Figs. 27, 28). In contrast, for 15DMSO and 15AN, the content of ZnO and ZHS significantly increases and both the X-Y and X-Z planes show a uniform and dense distribution of them. In addition, in the X-Z direction, ZnO is distributed deeper than ZHS, with ZHS mainly concentrated in the outer layer, while both the inner and outer layers contain a substantial amount of ZnO. Notably, the passivation-layer thickness in 15AN is almost twice as thick as that in 15DMSO. These results indicate that the presence of cosolvents regulates the architecture (thickness, density, uniformity) of the original passivation layer in the Bare system.

As cycling progresses, the corrosion current density (j_corr) in the Bare system gradually increases, corresponding to the HER rate acceleration. In contrast, in the cosolvent systems, HER corrosion is suppressed and remains stable due to the protection of the optimized passivation layer (Fig. 5d and Supplementary Fig. 29)⁴⁴. These results show that Zn²⁺ has a stronger competitive ability for electrons than H⁺ in the cosolvent systems. The interfacial charge-transfer activation energy (E_a,ct) before initial Zn deposited and after two cycles were further analyzed (Fig. 5e, Supplementary Figs. 30, 31, Supplementary Note 1 and Supplementary Tables 3, 4)⁴⁵. Owing to the interfacial passivation-layer formation, the activation energy increases after cycling. According to the TOF-SIMS results, the passivation layer in the 15AN system is thicker, which increases the transmembrane barrier and results in higher E_a,ct compared to 15DMSO, thereby presenting more pronounced polarization in 15AN (Supplementary Fig. 32).

Based on the above research, DMSO and AN follow the regulation mechanism rather than the self-decomposition mechanism (Fig. 5f). The presence of cosolvents does not introduce new components of the passivation-layer but mainly regulates the properties (size, orientation) and distribution (density, uniformity, thickness) of the original components (ZnO and ZHS). In the Bare system, the generated H₂ continuously damages the porous passivation layer, accelerating interfacial corrosion. In 15DMSO and 15AN systems, the dense and uniform passivation layer effectively suppresses HER dynamics. Notably, the thicker passivation layer creates a larger energy barrier for charge transfer and reduces its kinetics in 15AN, which is also an important factor in selecting the appropriate cosolvent.

Application scenario and battery performance

By decoupling the dynamic Zn²⁺-electrodeposition process, this work clarifies the interrelation between interfacial behaviors (H-bond, desolvation, adsorption, passivation, and charge-transfer) and battery performance around HER and polarization effect. The impact degree of different dimensions on the three electrolyte systems is shown in Supplementary Fig. 33, it indicates that the optimization degree of DMSO on the bare electrolyte is higher than that of AN through comprehensive and systematic analysis. As a result, 15DMSO exhibits notable stability and reversibility of Zn-plating/stripping than Bare electrolyte (Fig. 6a and Supplementary Figs. 34–37).

Fig. 6. Application scenario and battery performance.

a The stability and reversibility testing of Zn-plating/stripping at 1 mA cm^–2 and 1 mAh cm^–2. The insets are the diagrams of Zn-plating/stripping on Zn-electrode and Cu-electrode. b The battery performance test of Zn | |Zn symmetric cells at low-temperature (– 20 °C). Inset the local amplification of the initial cycles. The charging/discharging performance test of (c) Zn | |YEC and (d) Zn | |MnO₂ cells assembled with different electrolytes.

In addition, we further evaluate the different electrolyte systems under specific application scenarios, such as low-temperature and wide-voltage. At – 20 °C, 15DMSO still exhibits strong Zn-plating/stripping stability, while 15AN experiences severe polarization, leading to cycle failure (Fig. 6b and Supplementary Fig. 38). This is primarily due to the lower conductivity of 15AN under low-temperature conditions (Supplementary Fig. 39 and Supplementary Table 5), the weaker H-bond (H-bond dimension) formed between AN and water, and the formation of a thicker passivation-layer (passivation dimension). These results indicate that DMSO is beneficial to design low-temperature aqueous electrolytes. In the wide-voltage application scenario, we pair the electrolyte with an active carbon (YEC-8A) positive electrode. At a charging voltage of 1.9 V and current density of 0.5 A g^–1, the 15DMSO system shows higher initial capacity compared to the 15AN system, but 15AN exhibits higher cycling stability. After 1000 cycles, the capacity retention of 15AN reached 85.9%, whereas the 15DMSO system failed after 830 cycles (Fig. 6c). Moreover, even at a lower current density (0.1 A g^–1) and face more severe HER and OER possibility, 15AN still shows better cyclic stability (Supplementary Fig. 40). Therefore, AN is more suitable for the design of wide-voltage aqueous electrolytes. In addition, we also match the manganese dioxide (MnO₂) positive electrode, and contrary to the performance of the YEC positive electrode, 15DMSO showed lower initial capacity but more stable cycling (Fig. 6d). According to the results above, when designing cosolvent-based electrolytes, we need to select the appropriate cosolvent based on the specific application scenario, considering the compatibility between the electrolyte and the electrode.

Discussion

By comparing the structure-property-behavior relationship of cosolvents (DMSO and AN), this work decouples the electrodeposition process of Zn²⁺ in terms of five dimensions referring to H-bond, desolvation, adsorption, passivation, and charge-transfer. Comprehensive spectroscopy characterizations and theoretical calculations establish a direct comparison between the bare and cosolvent-participated electrolytes. We uncover decisive mechanisms for tunning interfacial chemistry under intrinsic and external electric fields. Firstly, we reveal that the reconstructed H-bond between water and different cosolvents is inconsistent, but they both disrupt the water-formed H-bond connectivity and impede H⁺ transmission. In situ FTIR demonstrates that cosolvents facilitate Zn²⁺-(de)solvation, accelerating the interfacial state transformation from a transient state to a steady state. Theoretical calculation further reveals that cosolvents promote the rate-determining step of the stepwise desolvation processes, inducing subsequent rapid desolvation steps. Notably, DMSO is more effective than AN in reducing the amount of interfacial active water, thereby lowering the ionization probability. Significantly, we observe an intriguing adsorption substitution phenomenon between water and cosolvent, where the potential-dependent mutual behaviors minimize the force field exerted on interfacial water to inhibit its ionization. Moreover, cosolvents promote the competition of Zn²⁺ for electrons and kinetically regulate the properties and distribution of the ZnO and ZHS components, rather than induce new components through self-decomposition. Based on these results, we visualize the interfacial dynamic evolution during Zn-plating/stripping and establish the correlation between the complex synergistic effect and battery performance. This multidimensional analysis provides fundamental insights into electrolyte engineering strategies for Zn-metal electrode stabilization, proposing feasible options for rational cosolvent selection in aqueous battery systems.

Methods

Chemicals and materials

Zn foil (thickness: 100 µm, 99.9%) and Cu foil (thickness: 100 µm, 99.9%) were purchased from Shenzhen Kejing Star Technology. Cu mesh (pore size: 148 µm, thickness: 100 µm) was purchased from the Nilaco Corporation. Zinc sulfate (ZnSO₄, > 99.0%) was purchased from Sigma-Aldrich Chemical Co., Ltd. Deuteroxide (D₂O, > 99.8%) was purchased from J&K Scientific. Vanadium oxide (V₂O₅, > 99%), dimethyl sulfoxide (DMSO, > 99.9%,) and acetonitrile (AN, > 99.8%) were purchased from Aladdin (Shanghai) Development Co., Ltd. N-methyl-2-pyrrolidinone (NMP, > 95%), Super P (SP), polyvinylidene fluoride (PVDF), α-manganese dioxide (α-MnO₂), carbon cloth (HCP330N, thickness: 320 µm, areal density: 14.1 mg cm^–2), and carbon paper (TGPH060T, thickness: 190 µm, areal density: 8.1 mg cm^–2) were obtained from Canrd Technology Co. Ltd. Stainless steel mesh (304, pores size: 26 µm, areal density: 22.4 mg cm^–2) was purchased from Tianjin Weixin Chemical Technology Co., LTD. Active carbon (AC, YEC-8A) was purchased from Fuzhou Yihuan Carbon Co., LTD. Ethanol (AR) was purchased from Adamas-beta. Deuterated chloroform (CDCl₃, > 99.8%) was purchased from Qingdao Tenglong Microwave Technology Co., Ltd. The separator (Whatman GF, pore size: 0.7 µm, thickness: 420 µm) was purchased from the Xiamen Xincheng Biotechnology Co., LTD. The case and related components of the coin cells were purchased from Teensky Technology Co., LTD (type: CR2032, material: 316 stainless steel, spring: Ф15.4 × 1.1 mm, gasket: Ф15.8 × 1.0 mm).

Electrolyte and electrode preparation

Deionized water was used to prepare all aqueous electrolytes. Specifically, Bare electrolyte contained 1 mol L^–1 ZnSO₄ in deionized water. The 15DMSO and 15AN electrolytes were prepared by adding the 15% volume ratio of DMSO/AN to the Bare electrolyte as cosolvents (The volume ratio of DMSO/AN and H₂O is 15:85.). For the Na₂V₆O₁₆·xH₂O (NVO) material, 3 g V₂O₅ was dissolved in 45 mL 2 mol L^–1 NaCl aqueous solution, followed by stirring for 72 h. An orange-red gel was then obtained, which was washed with deionized water and ethanol several times to remove the excess V₂O₅. The obtained product was dried in an oven at 60 °C for 12 h before use. For the positive electrode slurry, NVO/MnO₂ were mixed with SP and PVDF based on a weight ratio of 7:2:1, and the weight ratio of YEC, SP, and PVDF was 8:1:1. NMP was the dissolving solvent for the slurry. The slurry was mixed by an automatic mixer, which was homogeneously coated by an automatic coating machine onto the current collector of carbon cloth (for NVO), SS (for YEC), and carbon paper (for MnO₂). The blade height used to apply the slurry was 200 µm (for MnO₂) and 400 µm (for NVO and YEC). Then, the current collector with slurry was dried at 80 °C for 12 h in an oven. Finally, the current collector was cut into a positive electrode with a diameter of 11 mm (area: 0.95 cm⁻²) to assemble the cell in an open environment. The mass loading of NVO and YEC electrodes were 2.6 ~ 3.6 mg cm⁻² and 0.8 ~ 1.3 mg cm⁻² for MnO₂ electrode. Note that the metal electrodes and current collector were prepared directly before cell assembly without any treatments.

Cell assembling and testing

The Zn | |Zn symmetric cells and Zn | |Cu asymmetric cells were assembled by Zn foil (diameter: 1.1 cm, area: 0.95 cm²), Cu foil (diameter: 1.3 cm, area: 1.33 cm²), and separator (diameter: 19 mm) with 100 μL electrolyte transferred by pipette gun. The Zn | |NVO, Zn | |MnO_2, and Zn | |YEC full cells were assembled by Zn foil (diameter: 1.2 cm, area: 1.13 cm²), separator (diameter: 1.9 cm), and positive electrode (diameter: 1.1 cm) with 100 μL electrolyte transferred by pipette gun. For the Zn | |Zn cells, the current density of 1 mA cm⁻²/10 mA cm⁻² for 1 h to deposit and then stripped for 1 h. For Zn | |Cu cells, a fixed capacity of Zn (Q_p: 1 mAh cm⁻²/10 mAh cm⁻²) was plated on a Cu foil, and then stripped at a current density of 1 mA cm⁻²/10 mA cm⁻² at the cut-off potential of 1 V (corresponding to charge capacity (Q_s)). For Zn | |YEC cells, direct charge and discharge was applied at the current density of 0.5 A g⁻¹ at the voltage of 0.1 ~ 1.9 V (0.1 ~ 1.8 V for 0.1 A g⁻¹). For Zn | |MnO₂ cells, they were activated for 10 cycles at a current density of 0.1 A g⁻¹, and then for a long cycle at a current density of 0.5 A g⁻¹ at the voltage of 0.8 ~ 1.8 V. For Zn | |NVO cells (for OEMS tests), direct charge and discharge was applied at the current density of 1 A g⁻¹ at the voltage of 0.3 ~ 1.3 V for the initial cycle. The mass basis in the specific current and specific capacity is based on the active material. The 20^th cycle is used to calculate capacity retention. The coulombic efficiency is calculated using: $\text{CE}=\frac{{\text{Q}}_{\text{s}}}{{\text{Q}}_{\text{p}}}$. The galvanostatic cycling measurements were conducted on a Neware battery testing system (CT-4008Tn-5 V10 mA-164, Shenzhen, China). Before charge/discharge tests, the symmetric/asymmetric cells were subjected to a 0.5 h resting period under open-circuit conditions and the full cells for 5 hours. All cells were placed in the battery temperature test chamber (Shanghai Qixin Technology Instrument Co., LTD) for charge and discharge testing (average 25 °C ± 1, average – 20 °C ± 1 for low-temperature test).

Electrochemical measurements

The Linear Sweep Voltammetry (LSV), Polarization Curve (PC), Electrochemical Impedance Spectroscopy (EIS), and chronopotentiometry (CP) techniques were tested on a CHI 760E electrochemical workstation (Chenhua, China). LSV was tested at 0.1 mV s^–1 form – 1.6 V ~ 0.8 V (vs. Ag/AgCl) by Swagelok cell (three-electrode) using Ti foils as working electrode (diameter: 0.8 cm, area: 0.50 cm²) and counter electrode (diameter: 1.0 cm, area: 0.79 cm²), Ag/AgCl as the reference electrode. The electrode material of the Swagelok cell was titanium alloy, and the sleeve was made of polytetrafluoroethylene. PC was tested from – 0.15 V to + 0.15 V at 1 mV s^–1. The EIS for conductivity tests, the positive and negative electrodes of the coin cells are titanium foil with a diameter of 1.1 cm (area: 0.95 cm²). The positive and negative electrodes are separated by an insulating ring (thickness: 0.1 cm). The ionic conductivity is calculated by the formula: σ$=\frac{\text{l}}{R_s\text{S}}$. The EIS for activation energy tests, the positive and negative electrodes of the coin cells are Zn foil with a diameter of 1.1 cm. The activation energy is calculated by the formula: $\frac{1}{{\text{R}}_{\text{ct}}}=\text{A}\exp \left(\frac{\text{-}{\text{E}}_{\text{a}}}{\text{RT}}\right)$. The applied signal for EIS measurement was potentiostatic (single frequency) with an amplitude of 0.005 with a frequency range extended from 0.01 to 100,000 Hz until reaching a total of 85 data points for Ti | |Ti symmetric cells and 73 data points for Zn | |Zn symmetric cells. EIS data were fitted using ZView2 software. All cells undergoing electrochemical measurements were standing for 0.5 h before testing.

Characterizations

In/Ex-situ Fourier Transform infrared spectroscopy (FTIR) spectra were acquired in the range 4000 ~ 650 cm⁻¹ using a Nicolet iS50 FTIR spectrometer (Thermo Fisher Scientific, America). Scanning Electron Microscope (SEM) images were obtained on the Zeiss GeminiSEM 500. X-ray Diffraction (XRD) measurements were performed on Ultima IV (Rigaku Corporation, Japan) fitted with Cu-Kα X-rays (λ = 1.5406 Å) radiation at a scan rate of 5° min^–1. The ¹H Nuclear Magnetic Resonance (NMR) and ¹⁷O NMR spectra were accumulated 128 and 5000 times, respectively, by AVANCE NEO 500 MHz digital FT-NMR Spectrometer (Bruker). The ⁶⁷Zn NMR spectra were accumulated for two hours by AVANCE NEO 600 MHz digital FT-NMR Spectrometer (Bruker). All the NMR tests were performed using the external standard method, with CDCl₃ as the standard solution. Online Electrochemical Mass Spectrometry (OEMS) was tested by the mass spectrometer (HPR-20 R&D). X-ray Photoelectron Spectroscopy (XPS) characterization was tested by Thermo Scientific ESCALAB Xi⁺. Time of Flight Secondary Ion Mass Spectrometry (TOF-SIMS) measurements were conducted with an IONTOF M6 spectrometer. 30 keV bunched Bi³⁺ ion beam was used as the primary ion beam (Ion current 0.4 pA). The Zn electrodes used for ex situ characterizations (XPS, XRD, SEM, and TOF-SIMS) are all negative electrodes (Zn-plating state) from Zn | |Zn symmetric cells. The specific preparation method of the characterization electrode is as follows. The cells were disassembled after the cycle. The negative electrodes were washed with a large amount of deionized water, and then the water on the electrode was dried with filter paper, and the electrodes were dried in an oven at 60 °C for 0.5 hours. Then, stored in an oven at 25 °C for subsequent characterizations. The sampling and transport of electrodes and electrolytes for characterization without other temperature and atmosphere requirements.

In situ FTIR measurements

The in situ FTIR measurements were conducted by Thermo Fisher Scientific Nicolet iS50 spectrometer at room temperature. The system was equipped with a liquid nitrogen-cooled broad-band mercury-cadmium-telluride (MCT) detector. Experimental parameters included a fixed incident angle of 60° and spectral resolution of 4 cm⁻¹, which were maintained throughout the measurement protocol. Spectral data were recorded in absorbance mode, calculated according to the formula – log(I/I_o), where I and I_o correspond to the spectral intensities obtained from the sample and reference measurements, respectively⁴⁶.

The in situ FTIR technique coupled with chronopotentiometry was employed for monitoring the desolvation/solvation process of Zn²⁺ on the electrode surface with a home-built spectra-electrochemical cell. A zinc selenide (ZnSe) trapezoid prism was used as the reflection element. Before the in situ FTIR experiments, the basal plane of the ZnSe prism was mechanically polished with alumina suspensions of varied particle sizes (1.00, 0.30, and 0.05 μm) to a mirror finish and rinsed with deionized water and ethyl alcohol simultaneously. Finally, the ZnSe prism was transferred to ultrasonic cleaning in deionized water and ethyl alcohol for 20 mins, respectively.

In the in situ FTIR cell, the Cu mesh was the working electrode, and Zn foil was the counter and reference electrode. The Zn was plated on a Cu mesh electrode under 10 mA for 900 s and stripped at 10 mA for a 1.0 V cut-off voltage. Note that the background was first collected in the atmosphere, then the first spectrum after filling the electrolyte (A(t₀), without current), and the second spectrum (A(t_x), with current), as well as subsequent spectra (A(t_y), with current) were collected. We divided the whole Zn²⁺-desolvation and solvation process into two stages, denoted as abrupt change and gradual change, corresponding to the relative absorbance of ΔA(t_x) = A(t_x) – A(t₀) and ΔA(t_y) = A(t_y) – A(t_x), respectively.

DFT calculation

To ascertain the stability of water and solvent and its interaction with Zn²⁺ ions, DFT calculations were performed with the Gaussian 09 program (revision, D.01)⁴⁷. Initially, geometry optimizations for both water/solvent and water-solvent pair configurations were performed with the M06-2X/def2-TZVP level. It is worth noting that no imaginary frequencies were detected. Then the hydrogen bond energy ($E_{HB}$) between molecular A and B was estimated with the basis set superposition error (BSSE) correction⁴⁸, as:

$$E_{HB}=E_{AB}-E_A-E_B+E_{BSSE}$$

Where, $E_{AB}$ is the total energy of the AB pair, $E_A$ and $E_B$ are the single-point energy for molecular A and B, respectively. $E_{BSSE}$ is the basis set superposition error correction⁴⁸.

Subsequently, the lowest unoccupied molecular orbitals (LUMO) levels and the interaction between solvent molecules and Zn²⁺ are calculated with the SMD solvation model to account for the implicit solvation effects of aqueous solution.

The adsorption energy of solvent on the Cu(200) electrode was evaluated through the Vienna ab initio simulation package (VASP)⁴⁹. The Perdew-Burke-Ernzerhof (PBE) exchange-correlation functions of generalized gradient approximation (GGA) were employed in DFT calculations⁵⁰. The projector augmented wave (PAW) method⁵¹ with a cutoff energy of 500 eV was used to describe the interaction between nuclei and electrons. The convergence of energy and force were employed as 10^–5 eV and 0.03 eV Å^–1. The dipole correction and spin polarization were added for the calculations. The Γ-centered k-point meshes of 3 × 3 × 1 was adopted. In addition, for a correct treatment of physisorption interaction, Grimme’s D3 dispersion correction was employed in DFT calculations⁵². For each solvent, three different configurations were performed for the structure optimization. Subsequently, single-point energy calculations were conducted for these optimized structures. The adsorption energy was then determined as:

$$\mathrm{\Delta }E=E\left(surface+solvent\right)-E\left(surface\right)-E\left(solvent\right)$$

Molecular dynamics simulation

Molecular dynamics (MD) simulations were utilized with the MD package GROMACS¹. Initially, we conducted MD simulations of various electrolytes in their bulk states to elucidate the change in the solvation structure. Subsequently, the same electrolytes were confined between two electrode surfaces to investigate the additive solvent effect on the EDL structure. The SPC/E model was used to describe the interaction between water molcules⁵³. The ions in the system were modeled with an all-atom force field, which has been proved to reproduce the experimentally measured properties⁵⁴. The force field parameters of solvents are taken from OPLS⁵⁴ and assigned with RESP charge^55,56, which has been proved to reproduce the experimentally measured properties. The electrodes were represented by the Zn(002) metal facet^57,58. Temperature was controlled through the Nosé-Hoover thermostat^59,60 at 298 K. Long-range electrostatic interactions were calculated using the Particle Mesh Ewald (PME) method⁸. Reciprocal space computation was performed using a 0.1 nm fast Fourier transform (FFT) grid spacing combined with a cubic interpolation algorithm, while real-space electrostatic interaction was evaluated with a 1.2 nm distance cutoff. Particular attention was given to electrode polarization modeling in electrolyte environments through the implementation of the constant potential method (CPM). This method enables dynamic charge fluctuations on electrode atoms via instantaneous charge equilibration^61–64. Each simulation was systematically executed in three stages: initial thermal equilibration at 350 K for 3 ns, followed by gradual temperature reduction to 300 K over 2 ns, and subsequent 5 ns relaxation to achieve equilibrium. Thereafter, a 10 ns production was conducted for trajectory analysis³³. Five independent replicas with distinct initial configurations for each case were simulated, thereby guaranteeing the reliability of computational outcomes.

Electric field calculation

The electric field of interfacial water was quantitatively determined through systematic analysis of MD trajectories. Our computational framework implemented the PME algorithm, replacing the traditional distance-cutoff approximation, to achieve a precise evaluation of long-range electrostatic interactions. This method resolved the collective electrostatic contributions arising from four key components: electrode, water, ions, and solvent, thereby establishing a comprehensive interfacial potential model³³. The computational procedure was systematically executed in two sequential phases. Initially, van-der Waals interactions were turned off. Subsequently, the total electric field ($E_{total}$) was calculated through rerunning the MD-obtained trajectories. A systematic decomposition protocol was then implemented to resolve the field contributions from distinct sources: electrode effect ($E_{H_{2}O-electrode}$), intermolecular water dipole interaction ($E_{H_{2}O-H_{2}O}$), ionic contribution ($E_{H_{2}O-ZnS{O}_4}$), and solvent dipole interaction ($E_{H_{2}O-solvent}$). To calculate the interfacial electric field of water molecules, the computational strategy preserved inter-water molecular electrostatic interactions. Meanwhile, it deactivated all van der Waals forces and electrostatic contributions originating from the electrode, ZnSO₄, and solvent. This controlled simulation configuration enabled the precise determination of water-mediated electric field effects, denoted as $E_{H_{2}O-H_{2}O}$ ($E_{H_{2}O-H_{2}O}=E_1$). Similarly, to determine the contribution from ZnSO₄, the van der Waals forces of all particles and electrostatic interactions from both electrode and solvent were turned off. MD simulations were rerun on an existing trajectory to obtain the electric field experienced by interfacial water ($E_2$). Thus, the electric field induced by ions is denoted as: $E_{H_{2}O-ZnS{O}_4}=E_2-E_{H_{2}O-H_{2}O}$. For the contribution from the solvent, the van der Waals interaction of all particles and the electrostatic interaction from the electrode and ZnSO₄ were similarly turned off, yielding the electric field ($E_3$) experienced by interfacial water. The solvent-induced electric field was then determined as: $E_{H_{2}O-solvent}=E_3-E_{H_{2}O-H_{2}O}$. Finally, the contribution from the electrode was calculated by subtracting all other contributions from the total electric field: $E_{H_{2}O-electrode}=E_{total}-E_{H_{2}O-H_{2}O}-E_{H_{2}O-ZnS{O}_4}-E_{H_{2}O-solvent}$. This systematic decomposition ensures an accurate assessment of the individual contributions to the electric field experienced by interfacial water.

Author contributions

X.Y. and M.C. contributed equally to this work as co-first authors. X.Y., M.C., G.F, Y.Z., and Y.Q. contributed to the design of the research and performed the experimental data analysis. M.C., S.L., and G.F. perform theory calculation operations and analysis. X. Y., J. W., Q. Z., and Y. Q. performed in situ FTIR experiments and analysis. X.Y., H.Y.L., and H.Z. performed the battery performance testing and various characterizations. Y.D., H.F.L., J.Z., F.W., D.C., and S.-G.S. provided theoretical guidance. Y.Z, G.F., and Y.Q. supervised the work. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks Lei Wei Sanja Tepavcevic, Kuk Young Cho, and the other anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The source data for this study are provided as a Source Data file. Source data are provided in this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-59069-7.