Metal electrode potential diverges with ion additions Authors: Qiu Zhang, Seongjae Ko, Taisei Sakata, Shin-ichi Nishimura, Norio Takenaka, Atsushi Kitada, Chunsheng Wang, Atsuo Yamada Published: 2026 ABSTRACT The electrode redox potential, which varies substantially with the electrolyte, governs both the direction and thermodynamic driving force of electrochemical reactions. However, a comprehensive understanding and effective control of redox potential shifts remain unestablished. Here we report the pronounced effect of coordination-shell ions on the redox potential of metal electrodes. Cooperative interactions between the metal ion and surrounding anions and cations with large differences in hardness/softness amplify the redox potential shift, as corroborated by liquid Madelung potential. Specifically, in the Zn/Zn2+ system, the addition of soft anions paired with hard cations destabilizes the Zn2+ solution structure and upshifts the redox potential, whereas hard anions with soft cations stabilize the structure and induce a downshift. Following this principle, a potential gap exceeding 0.6 V was achieved—a remarkably large but reasonable value for a divalent system. This ion hardness/softness-based strategy also demonstrates practical benefits in Zn plating/stripping tests, where electrolytes with upshifted redox potentials enable an average Coulombic efficiency exceeding 99.9%. Main Progress towards a carbon-neutral society and the advancement of green energy technologies depend on transformative innovations in electrochemical devices. At their core, electrode redox potential governs the thermodynamic feasibility, efficiency and selectivity of redox processes, serving as the foundation of these technologies1–4. Therefore, a deeper understanding of the mechanisms behind potential shifts and their strategic application to enhance active reaction dynamics but suppress side reaction kinetics is crucial for maximizing active reaction efficiency and driving advancements across diverse electrochemical applications, including energy storage and conversion, catalysis, environmental remediation and biomedical technologies. The origin of electrode potential lies in the Gibbs free energy changes during electrode reactions. In a typical metal ion reduction process, Mn+(sol) + ne− = M, the chemical state of metal ions in an electrolyte and reduced metal as a solid determines the fundamental free energy change, while the solvation environment of metal ions further influences this energy change5–10. Previous studies have shown that the deposition potential of Li metal is affected by solvents with different solvation capabilities11. On a microscopic level, the competitive coordination between anion and solvent to Li+ leads to potential shifts. The liquid Madelung potential has recently been proposed as a concept that extends beyond classical Debye–Hückel theory to explain this phenomenon in a specific salt concentration regime12,13. However, the role of ions in the coordination shells, as key components of electrolytes, in potential shifts has not been thoroughly investigated. The microscopic understanding of the interactions between metal ions and the surrounding anions and cations, as well as their effects on potential shifts, remains elusive. Therefore, establishing the relationship among anion/cation characteristics, solvation structures and potential shifts is crucial. In addition, earlier research was limited to the binary system of Li+ and solvents. Research on the potential shift of redox reactions involving divalent metal ions (for example, Zn2+) remains unexplored. Here, we hypothesize that the hard/soft character of coordinating anions and cations strongly influences the redox potential of metal ions. By combining electrochemical measurements with molecular dynamics (MD) simulations and introducing a liquid Madelung potential framework, we aim to quantitatively capture the electrostatic environment of metal ions in electrolytes containing hard/soft ions. This study establishes a fundamental understanding of how ion-mediated solvation structures modulate electrode potentials and provides electrolyte design principles for next-generation metal batteries. Results and discussion Metal electrode redox potential shift The redox potential of a metal anode (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${E}_{{M/M}^{n+}}$$\end{document}EM/Mn+) is governed by the chemical potential of a metal ion (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mu }_{{{{M}}}^{n+}}$$\end{document}μMn+) in the electrolyte, following the equation, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mu }_{{{{M}}}^{n+}}={\mathrm{FE}}_{{{M}}/{{{M}}}^{n+}}$$\end{document}μMn+=FEM/Mn+. The chemical potential of Mn+, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mu }_{{{{M}}}^{n+}}$$\end{document}μMn+, is primarily influenced by its Coulombic interactions with the surrounding ions (both anions and cations)8. To understand how the nature of ions in coordination shells modulates overall interactions in the electrolyte and contributes to potential shifts, their inherent physicochemical properties—namely, hardness and softness—were thoroughly considered. Hard ions are generally composed of a single or small number of atoms with a small radius and localized charge density, resulting in strong Coulombic interactions14–16. Soft ions consist of multiple atoms with a large ionic radius and delocalized charge density, leading to weaker Coulombic interactions14–16. For example, when soft anions and hard cations occupy the inner and outer solvation shell, respectively, in the Zn2+ solvation structure, the weak attractive interactions from the soft anion, combined with the strong repulsive interactions from the hard cation, destabilizes Zn2+. Meanwhile, the presence of hard anions in the inner solvation shell and soft cations in the outer solvation shell stabilizes Zn2+. Consequently, the destabilization of Zn2+ in the former scenario causes a potential upshift, while its stabilization in the latter leads to a potential downshift (Fig. 1).Fig. 1Three typical ion coordination effects on the stability of Zn2+ ions with different ion additions.a, Solvation structures depicting anions and cations in the inner shell and outer shell, respectively. b, Site potential of Zn2+ in different solvation structures with the addition of different ions. The bars represent the hardness/softness of anions (red) and cations (blue) and the Zn/Zn2+ redox potential (black). A greater disparity in the hardness/softness between anions and cations provides a larger potential shift. For instance, adding a soft anion (pale red) with a hard cation (deep blue), or a hard anion (deep red) with a soft cation (pale blue), leads to a considerable upshift (pale black) or downshift (deep black) in redox potential. Three typical ion coordination effects on the stability of Zn2+ ions with different ion additions. a, Solvation structures depicting anions and cations in the inner shell and outer shell, respectively. b, Site potential of Zn2+ in different solvation structures with the addition of different ions. The bars represent the hardness/softness of anions (red) and cations (blue) and the Zn/Zn2+ redox potential (black). A greater disparity in the hardness/softness between anions and cations provides a larger potential shift. For instance, adding a soft anion (pale red) with a hard cation (deep blue), or a hard anion (deep red) with a soft cation (pale blue), leads to a considerable upshift (pale black) or downshift (deep black) in redox potential. Measurement of Zn electrode potential shift Ten commonly used ions—acetate anion (Ac−), chloride anion (Cl−), iodide anion (I−), trifluoromethanesulfonate anion (OTf−), bis(trifluoromethanesulfonyl)imide anion (TFSI−), magnesium cation (Mg2+), lithium cation (Li+), potassium cation (K+), ammonium cation (NH4+) and choline cation (Ch+)—were selected for systematic investigation of fundamental local correlations in coordination shells. The distribution and value of electrostatic potentials (ESPs) in each ion, which can be quantitatively estimated through density functional theory calculation, serve as an effective descriptor reflecting the strength of electrostatic interaction17,18 (Supplementary Figs. 1 and 2). The ESP is defined as the work required to bring a unit positive charge from infinity to a point on the surface with an electron density of 0.001 a.u. (ref. 19). A more negative value for anions or a more positive value for cations reflects a stronger ability to engage in electrostatic interactions. For example, uniform ESP mapping with large negative values of −6.00 and −5.36 eV was calculated for Cl− and I− anions, respectively, highlighting their strong affinity to Zn2+ (Fig. 2a). Ac−, OTf− and TFSI− anions, containing multiple atoms, exhibit uneven and distributed ESP mappings. Regions near oxygen atoms show the most negative values of −6.63, −5.26 and −4.88 eV for Ac−, OTf− and TFSI−, where Zn2+ preferentially binds. However, these anions also include areas with less negative ESP values of −3.19, −3.03 and −2.25 eV, respectively, which correspond to weaker interactions with Zn2+. The gap between maximum and minimum ESP values of Ac− is larger than that of OTf− and TFSI− anions, reflecting the highly localized electron density in its COO− functional group. The calculated changes in interaction strength with Zn2+ by different anions were experimentally confirmed by Raman spectroscopy (Supplementary Fig. 3). The addition of Zn2+ to 10 mol kg−1 (m) LiAc, LiCl or LiI aqueous electrolyte introduces a new peak due to the formation of Zn−Ac, Zn−Cl or Zn−I bonds20–22, demonstrating strong interactions between Zn2+ and Ac−, Cl− or I− anions. Meanwhile, negligible peak shifts are observed when Zn2+ ions were added into LiOTf- and LiTFSI-containing electrolytes23,24, suggesting weaker interactions with Zn2+. All the computational (ESP distributions and values) and experimental findings confirmed that Ac−, Cl− and I−, which exhibit highly localized and strongly negative ESP values, interact strongly with Zn2+ and are therefore classified as hard anions, whereas OTf− and TFSI−, which display more delocalized and less negative ESP values, interact more weakly and are thus categorized as soft anions.Fig. 2Charge distribution in ions and their effect on the potential shift of Zn/Zn2+ redox reaction.a, Calculated ESP extremum values for anions and cations. Bar colours represent the hardness/softness of each ion; deep colours indicate hard ions, while lighter colours represent soft ions. The structures of the anions and cations are displayed near each bar, with surrounding colours illustrating the ESP distribution—red for negative regions and blue for positive regions. The areas on the anions or cations that match the bar colour indicate the most negative or positive (that is, ESP extremum) positions, highlighting the regions with the highest electron density or deficiency. b, Anion effect for Zn/Zn2+ potential shift in the electrolytes consisted of 1 m ZnX2 and 10 m LiX (X represents different anions). c, Cation effect for Zn/Zn2+ potential shift in the electrolytes consisted of 1 m Zn(TFSI)2 and 4 m MTFSI or 1 m ZnCl2 and 10 m MCl (M represents different cations). For Mg2+, the salt concentration is halved. Deeper/paler colours represent harder/softer ions. Note that the Zn/Zn2+ potential shift in the presence of K+ was investigated only within a limited concentration range, constrained by the solubility limits of KTFSI (~1 m) and KCl (~4 m) (Supplementary Figs. 5 and 6).Source data Charge distribution in ions and their effect on the potential shift of Zn/Zn2+ redox reaction. a, Calculated ESP extremum values for anions and cations. Bar colours represent the hardness/softness of each ion; deep colours indicate hard ions, while lighter colours represent soft ions. The structures of the anions and cations are displayed near each bar, with surrounding colours illustrating the ESP distribution—red for negative regions and blue for positive regions. The areas on the anions or cations that match the bar colour indicate the most negative or positive (that is, ESP extremum) positions, highlighting the regions with the highest electron density or deficiency. b, Anion effect for Zn/Zn2+ potential shift in the electrolytes consisted of 1 m ZnX2 and 10 m LiX (X represents different anions). c, Cation effect for Zn/Zn2+ potential shift in the electrolytes consisted of 1 m Zn(TFSI)2 and 4 m MTFSI or 1 m ZnCl2 and 10 m MCl (M represents different cations). For Mg2+, the salt concentration is halved. Deeper/paler colours represent harder/softer ions. Note that the Zn/Zn2+ potential shift in the presence of K+ was investigated only within a limited concentration range, constrained by the solubility limits of KTFSI (~1 m) and KCl (~4 m) (Supplementary Figs. 5 and 6). Source data For cations, a higher positive ESP value signifies a stronger (repulsive) interaction ability. Mg2+ and Li+ are characterized by highly localized positive charge density on their nuclei, with high ESP values of +24.60 and +14.83 eV, classifying them as hard cations. Meanwhile, the ESP of weakly Lewis acidic K+ is calculated to be +8.27 eV. In addition, large NH4+ and Ch+ molecules possess delocalized positive charge density over their multiple atoms. Their maximum ESP values are +7.69 and +5.27 eV, while the minimum ESP values are +7.23 and +1.91 eV, respectively, both of which are considerably lower than those of Mg2+ and Li+. This indicates that K+, NH4+ and Ch+ are categorized as soft cations, exhibiting weak interactions with Zn2+. After confirming the hardness/softness of each anion and cation, the potential shift of the Zn/Zn2+ redox reaction was experimentally investigated in electrolytes containing various types of anion and cation. Given the high solubility of Li salt, the effect of different anions on the potential shift was evaluated by measuring the Zn/Zn2+ potential in aqueous electrolytes composed of 1 m ZnX2 + 10 m LiX (X represents different anions). As consistent with our original considerations (Fig. 1), hard X anions, such as Ac−, Cl− and I−, caused a potential downshift, while soft X anions, such as OTf− and TFSI−, led to a potential upshift (Fig. 2b and Supplementary Figs. 4–6). Note that the experimentally measured potential can deviate from the intrinsic thermodynamic Zn/Zn2+ redox potential owing to contributions from side reactions, including the hydrogen evolution reaction (HER), and liquid-junction potentials associated with the reference electrode. In the present electrolytes, however, the measured Zn electrode potential remains highly stable over the measurement timescale (Supplementary Fig. 5), and the potential offsets arising from HER and junction effects are small enough (Supplementary Fig. 4) to provide a reliable basis for tracking ion-dependent potential shifts across different electrolytes. The potential shifts of Zn electrodes in two types of anion electrolyte were measured. One electrolyte (1 m ZnCl2 + 10 m MCl) contains a hard anion of Cl−, and the other electrolyte (1 m Zn(TFSI)2 + 4 m MTFSI) contains a soft anion (TFSI−), where M in both electrolytes represents different cations of Ch+, NH4+, K+, Li+ and Mg2+, and the concentration of divalent Mg2+ is half of the monovalence cation (Fig. 2c and Supplementary Figs. 5 and 6). All the experiments validated that hard cations (Mg2+ and Li+) increase the potential, while soft cations (K+, NH4+ and Ch+) decrease the potential. The ion effect on Zn/Zn2+ potential is highly robust, being consistently observed across mixed-cation (Li+/Ch+) and mixed-anion (Cl−/TFSI−) aqueous electrolytes (Supplementary Fig. 7), a wide temperature range from 0 °C to 50 °C (Supplementary Fig. 8) and diverse solvent systems including aprotic solvents (dimethyl sulfoxide), ionic liquids (1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide) and deep eutectic (ChCl/urea) systems (Supplementary Figs. 9 and 10). Importantly, the effect of ion addition can be extended to other redox couples, such as Cu/Cu2+ (Supplementary Fig. 11), highlighting the general validity of the concept for understanding electrochemical potentials. Figure 3a summarizes the measured Zn/Zn2+ redox potentials in various electrolytes, which emphasize the critical role of the types of ion (hardness/softness) and salt concentration in potential shifts. Increasing the concentration of salts, composed of ions with contrasting hardness/softness, such as soft TFSI−–hard Li+ or hard Ac−–soft K+, results in substantial potential upshifts or downshifts. By contrast, salts with ions of similar properties, such as soft TFSI−–soft Ch+ or hard I−–hard Li+, provide minor potential shifts even at high concentrations. Notably, the observed redox-potential gap for divalent Zn/Zn2+ between 1 m Zn(TFSI)2 + 20 m LiTFSI and 1 m Zn(Ac)2 + 30 m KAc was 0.6 V, comparable to the value observed in the monovalent Li/Li+ system (Fig. 3b). Because the redox potential of divalent ions is determined by dividing the Gibbs free energy change by two charges, this underscores the critical role of enhanced Coulombic interaction by double-positive charge as well as ion hardness/softness in potential shifts.Fig. 3A substantial Zn/Zn2+ potential difference across various electrolytes.a, Potential shifts observed in diverse electrolytes comprising different anion–cation combinations. Increasing the concentration of added salts with contrasting properties (hard–soft pairs) induces large potential shifts, whereas salts with similar properties (soft–soft or hard–hard pairs) result in small potential shifts. Circles and boxes represent anions and cations, respectively; deep colours denote hard ions, while lighter colours indicate soft ions. Specific potential values for the given electrolytes are provided in Supplementary Fig. 5 and Supplementary Table 1. b, Schematic illustration of a remarkably large potential gap (≥0.6 V) in divalent Zn/Zn2+ systems highlights the pivotal role of ion hardness/softness in influencing potential shifts11,13. Here, R is the gas constant, T is the temperature, F is the Faraday constant, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${a}_{{M}^{n+}}$$\end{document}aMn+ is the activity of the metal ion.Source data A substantial Zn/Zn2+ potential difference across various electrolytes. a, Potential shifts observed in diverse electrolytes comprising different anion–cation combinations. Increasing the concentration of added salts with contrasting properties (hard–soft pairs) induces large potential shifts, whereas salts with similar properties (soft–soft or hard–hard pairs) result in small potential shifts. Circles and boxes represent anions and cations, respectively; deep colours denote hard ions, while lighter colours indicate soft ions. Specific potential values for the given electrolytes are provided in Supplementary Fig. 5 and Supplementary Table 1. b, Schematic illustration of a remarkably large potential gap (≥0.6 V) in divalent Zn/Zn2+ systems highlights the pivotal role of ion hardness/softness in influencing potential shifts11,13. Here, R is the gas constant, T is the temperature, F is the Faraday constant, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${a}_{{M}^{n+}}$$\end{document}aMn+ is the activity of the metal ion. Source data Coordination structure reflecting hard/soft ion interactions As evident from the discussion earlier, a deep understanding of Zn2+ solvation structure is essential to uncover the influence of anions and cations in coordination shells on the Zn/Zn2+ potential shifts. The local environment and the solution structure of the electrolyte were characterized using X-ray absorption fine structure (XAFS) analysis (Supplementary Fig. 12), coupled with MD simulations25,26. Two distinct cases with contrasting ion hardness/softness at high salt concentrations are demonstrated in Fig. 4a,b. In the combination of hard anion Cl−/soft cation Ch+ (1 m ZnCl2 + 10 m ChCl), the Cl− and Ch+ are primarily located at around 2 Å and 6 Å from the central Zn2+, forming the inner and outer solvation shells (Fig. 4a,c and Supplementary Figs. 13 and 14). Meanwhile, in the soft anion TFSI−/hard cation Li+-based electrolyte (1 m Zn(TFSI)2 + 10 m LiTFSI), the weaker attractive interaction with the bulky, soft TFSI− in the inner solvation shell and the stronger repulsive interaction with the hard Li+ in the outer shell lead to longer coordination distances (approximately 4–7 Å for TFSI− and 5–10 Å for Li+) (Fig. 4b,d and Supplementary Fig. 13). This leads to an upward shift in the Zn/Zn2+ redox potential with a shallower liquid Madelung potential of Zn2+, consistent with the predicted and measured hardness/softness ion effects on potential shifts (Figs. 1–3).Fig. 4Zn electrolyte solution structure with various ions.a,b, Amplitudes of k2-weighted Fourier transform (FT) of the Zn K-edge EXAFS, compared with the simulated results based on MD trajectory of 1 m ZnCl2 + 10 m ChCl (a) and 1 m Zn(TFSI)2 + 10 m LiTFSI (b) electrolytes. The insets show the corresponding electrolyte solvation structures. The simulated EXAFS spectra with MD trajectory align well with the experimental data, validating the accuracy of simulated Zn2+ solvation structures. c,d, The calculated Coulombic energy in 1 m ZnCl2 + 10 m ChCl (c) and 1 m Zn(TFSI)2 + 10 m LiTFSI (d) electrolytes, based on the XAFS-fitted MD simulations. The influence of ion softness/hardness in inner (first) and outer (second) solvation shells offers intuitive insight into potential shifts.Source data Zn electrolyte solution structure with various ions. a,b, Amplitudes of k2-weighted Fourier transform (FT) of the Zn K-edge EXAFS, compared with the simulated results based on MD trajectory of 1 m ZnCl2 + 10 m ChCl (a) and 1 m Zn(TFSI)2 + 10 m LiTFSI (b) electrolytes. The insets show the corresponding electrolyte solvation structures. The simulated EXAFS spectra with MD trajectory align well with the experimental data, validating the accuracy of simulated Zn2+ solvation structures. c,d, The calculated Coulombic energy in 1 m ZnCl2 + 10 m ChCl (c) and 1 m Zn(TFSI)2 + 10 m LiTFSI (d) electrolytes, based on the XAFS-fitted MD simulations. The influence of ion softness/hardness in inner (first) and outer (second) solvation shells offers intuitive insight into potential shifts. Source data Validation by liquid Madelung potential While identifying the effect of ions in inner (first) and outer (second) solvation shells provides intuitive insight into potential shifts (Fig. 4 and Supplementary Fig. 15), it is limited in qualitative analysis and overlooks long-range interactions. To better account for overall Coulombic interactions within dynamically fluctuating liquid systems, the liquid Madelung potential (ELM)12 was calculated using XAFS-fitted MD simulations (Figs. 4 and 5 and Supplementary Figs. 16 and 17). The concept of liquid Madelung potential has already been reported and applied to calculate the potential change in lithium batteries with different organic solvents and salt concentrations12. It is defined as the sum of the electrostatic energies of all atoms surrounding the central atom, thereby offering a comprehensive energetic perspective beyond the local solvation structure of Zn2+. ELM is given by\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${E}_{\mathrm{LM}}=\frac{1}{{N}_{{\mathrm{Zn}}^{2+}}}\left\langle \mathop{\sum }\limits_{i}\frac{{q}_{i}}{4{\rm{\pi }}{\varepsilon }_{0}{r}_{i}}\right\rangle ,$$\end{document}ELM=1NZn2+∑iqi4πε0ri,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${N}_{{\mathrm{Zn}}^{2+}}$$\end{document}NZn2+, ε0, q and r are the number of Zn2+, the permittivity of vacuum, the atomic charge of the atoms and the distance r of the atoms of the surrounding atoms from the central Zn2+, respectively. Based on MD trajectories, the liquid Madelung potential (ELM) was calculated as a time-averaging and space-averaging value, incorporating the shielding effects of all electrolyte components. As shown in Fig. 5a,b, the liquid Madelung potential closely matches the experimentally obtained values, whereas the classical extended Debye–Hückel model (Supplementary Note 1) and Pitzer equations (Supplementary Note 2) fail to predict the potential shifts with salt addition. This discrepancy arises because these traditional frameworks are either limited to dilute electrolytes or rely on empirically parameterized ion-interaction terms with limited transferability and thus cannot resolve specific ion–solvent coordination and the spatially heterogeneous electrostatic environment associated with high-charge-density divalent ions. By contrast, the liquid Madelung potential quantitatively captures the potential shifts by explicitly accounting for instantaneous solvation structures and long-range Coulombic interactions among all electrolyte species, thereby reflecting the cooperative effects of ion hardness/softness in practical electrolytes.Fig. 5Liquid Madelung potential (ELM) quantitatively explains potential shifts, incorporating ion hardness/softness effects.a,b, Comparison of experimentally obtained potential shifts (circles) with those calculated using the traditional extended Debye–Hückel (triangles) and Pitzer equations (diamonds), as well as the liquid Madelung potential (hexagons) in the given electrolytes with different anions (a) and cations (b). Liquid Madelung potential accurately reproduced the experimental data. The point colours represent the hardness/softness of each ion; deep colours indicate hard ions, while lighter colours represent soft ions, consistent with Figs. 2 and 3.Source data Liquid Madelung potential (ELM) quantitatively explains potential shifts, incorporating ion hardness/softness effects. a,b, Comparison of experimentally obtained potential shifts (circles) with those calculated using the traditional extended Debye–Hückel (triangles) and Pitzer equations (diamonds), as well as the liquid Madelung potential (hexagons) in the given electrolytes with different anions (a) and cations (b). Liquid Madelung potential accurately reproduced the experimental data. The point colours represent the hardness/softness of each ion; deep colours indicate hard ions, while lighter colours represent soft ions, consistent with Figs. 2 and 3. Source data Enhancing the reversibility of Zn-metal anode by potential upshift Strategic control of metal redox potentials in electrolytes contributes to the stability of a wide range of electrochemical devices. We used aqueous Zn-metal batteries as a model system, which are garnering considerable attention as post-Li-ion technologies due to their intrinsic safety and high theoretical energy density (theoretical capacity of Zn metal: 820 mAh g−1 and 5,854 mAh cm−3) yet suffer from poor reversibility caused by severe HERs27–31. As illustrated in Fig. 6a, the reversibility of Zn plating/stripping strongly depends on the Zn/Zn2+ redox potential in the electrolyte, where a higher potential enables stable plating/stripping with suppressed HERs, beyond the conventional considerations of Zn deposit morphology, orientation, kinetics and surface passivation (Supplementary Figs. 18−20). Impressively, an exceptionally high average Coulombic efficiency exceeding 99.93% was achieved using an electrolyte with a maximally upshifted Zn/Zn2+ redox potential, composed of multiple soft anions and a hard cation (1 m Zn(TFSI)2 + 20 m LiTFSI + 6 m lithium bis(pentafluoroethanesulfonyl)imide (LiBETI); Fig. 6a and Supplementary Figs. 21–23). Moreover, Zn||Mo6S8 and Zn||LiMn2O4 full cells were assembled as proof-of-concept systems to examine the compatibility of the electrolyte in full-cell configurations (Fig. 6b and Supplementary Fig. 24). This underscores the importance of electrolyte design based on ion hardness/softness, as it allows a striking redox potential shift of ≥0.6 V even in a divalent Zn/Zn2+ system, paving the way for the development of highly sustainable and reliable next-generation electrochemical devices.Fig. 6Highly reversible Zn plating/stripping reactions.a,b, Coulombic efficiency during Zn plating/stripping as a function of the Zn/Zn2+ redox potential (a) and Zn||Mo6S8 full cell (b), measured at a current density of 100 mA g−1. An increase in the Zn/Zn2+ redox potential enables stable Zn plating/stripping under a reduced thermodynamic driving force for electrolyte decomposition (HER), thereby allowing an exceptionally high average Coulombic efficiency exceeding 99.9% (Supplementary Figs. 22 and 23) and stable long-term cycling of full cells (Supplementary Fig. 24). The grey line serves as a visual guide. The insets in a and b show the cyclic voltammetric curves of Zn plating/stripping and the charge/discharge curves of the full cell in the given electrolytes, respectively. To obtain a maximal redox potential upshift, a high concentration of multiple soft anions (TFSI− and BETI−) paired with a hard cation (Li+) was added to a 1 m Zn(TFSI)2 solution, resulting in an even higher Zn/Zn2+ redox potential than that of 1 m Zn(TFSI)2 + 20 m LiTFSI solution (−0.621 V and −0.685 V versus Ag/AgCl, respectively) (Supplementary Fig. 5 and Supplementary Table 1).Source data Highly reversible Zn plating/stripping reactions. a,b, Coulombic efficiency during Zn plating/stripping as a function of the Zn/Zn2+ redox potential (a) and Zn||Mo6S8 full cell (b), measured at a current density of 100 mA g−1. An increase in the Zn/Zn2+ redox potential enables stable Zn plating/stripping under a reduced thermodynamic driving force for electrolyte decomposition (HER), thereby allowing an exceptionally high average Coulombic efficiency exceeding 99.9% (Supplementary Figs. 22 and 23) and stable long-term cycling of full cells (Supplementary Fig. 24). The grey line serves as a visual guide. The insets in a and b show the cyclic voltammetric curves of Zn plating/stripping and the charge/discharge curves of the full cell in the given electrolytes, respectively. To obtain a maximal redox potential upshift, a high concentration of multiple soft anions (TFSI− and BETI−) paired with a hard cation (Li+) was added to a 1 m Zn(TFSI)2 solution, resulting in an even higher Zn/Zn2+ redox potential than that of 1 m Zn(TFSI)2 + 20 m LiTFSI solution (−0.621 V and −0.685 V versus Ag/AgCl, respectively) (Supplementary Fig. 5 and Supplementary Table 1). Source data Conclusion The effect of ions in electrolytes on the potential shift of the Zn/Zn2+ redox reaction is systematically investigated, revealing the importance of their hardness/softness. Soft anion/hard cation pairs induce a substantial potential upshift, while hard anion/soft cation pairs lead to a downward potential shift. Liquid Madelung potential calculations accurately capture these trends by accumulating Coulombic interactions among all constituent atoms around Zn2+ in the electrolyte, which is oversimplified by traditional extended Debye–Hückel and Pitzer equations. A large difference in the Zn/Zn2+ redox potential over 0.6 V is readily observed between soft anion/hard cation and hard anion/soft cation combinations, with a more pronounced difference in ion hardness/softness, enabling a highly reversible Zn/Zn2+ redox process. This work highlights the critical role of ion properties in modulating electrode potential and their fundamental importance for better electrolyte design of diverse electrochemical systems. Methods Electrochemical measurements The potential of Zn metal (Nilaco) was tested in a three-electrode cell with Zn metal as the working electrode, Pt as the counter electrode, and Ag/AgCl electrode with saturated KCl solution as the reference electrode. All electrochemical tests were conducted using a VMP3 potentiostat (Bio-Logic). Zn plating/stripping on Ti foil (Fuchikawa Rare Metal) was evaluated using 2032-type coin cells assembled with Zn metal (Φ = 12 mm), Ti foil (Φ = 16 mm), glass fibre separator (Whatman GF/F) and 40 μl of electrolyte. Full cells were assembled with Zn metal (Φ = 12 mm), cathode (Φ = 10 mm), glass fibre separator (Whatman GF/F) and 40 μl of electrolyte. The tests were conducted using a LAND CT2000A charge–discharge machine at a current density of 0.5 mA cm−2 with an upper cut-off voltage of 0.5 V. Characterizations Raman spectroscopy for electrolyte was conducted on an NRS-5100 spectrometer JASCO with a 532-nm excitation laser. All the Raman peaks of the electrolytes were calibrated with a standard Si peak (520.7 cm−1). The XAFS of the Zn K edges were collected on the beamline 5S1 in AichiSR in the transmission mode. The radiation was monochromatized by a Si (111) double-crystal monochromator. The Athena and Artemis modules of the Demeter software package were used to process extended XAFS (EXAFS) data according to the standard procedures32. The DiffPy-CMI software package was conducted to simulate the R space of EXAFS based on the MD trajectory of the Zn2+ solvation shell33. MD simulations MD simulations simulations of the electrolyte structures including the cation coordination shell and water structure were performed using the GROMACS package with the AMBER03 force field34. The MD parameters for Cl−, I−, Zn2+, Mg2+ and Li+ were in the in-built force field parameters. Water molecules were simulated with the TIP3P model to obtain the electrolyte structure. The MD parameters for TFSI−, OTf−, NH4+ and Ch+ were generated by ACPYPE35, and the corresponding atom charges were based on restrained electrostatic potential charges36. An empirical continuum correction was applied with a scaled charge to account for polarizable effects. First, NPT simulations were performed at 300 K for 40 ns to ensure system equilibration, followed by an additional 10-ns NPT simulation for analysis. The snapshot of MD simulation was produced by VMD software37. Quantum chemistry calculations Quantum chemistry calculations were carried out with the Gaussian 09 software package. The structure optimization and ESP were performed with the B3LYP38 density functional and the aug-cc-pvdz39 basis set. The ionic potential was defined as the ionic charge divided by the ionic radius, where the ionic radius was calculated using Multiwfn40,41. The water–ion binding energy was calculated with basis set superposition error correction to accurately capture the weak interactions between anions and water molecules. The solvation energies in the implicit aqueous model were calculated as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta {G}_{\mathrm{solvation}}={{G}_{\mathrm{ion}}}^{(\mathrm{aq})}-{{G}_{\mathrm{ion}}}^{({\rm{g}})},$$\end{document}ΔGsolvation=Gion(aq)−Gion(g),where Gion(aq) is the Gibbs free energy of ion in the implicit model, and Gion(g) is the energy of ion in vacuum. The solvation energies using the cluster–continuum model were calculated as\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta {G}_{\mathrm{solvation}}={{G}_{\mathrm{ion}-\mathrm{water}}}^{(\mathrm{aq})}-{{G}_{\mathrm{water}}}^{(\mathrm{aq})}-{{G}_{\mathrm{ion}}}^{({\rm{g}})},$$\end{document}ΔGsolvation=Gion−water(aq)−Gwater(aq)−Gion(g),where Gion−water(aq) is the Gibbs free energy of the ion−water cluster embedded in the implicit solvent, and Gwater(aq) is the Gibbs free energy of the corresponding water cluster in the implicit solvent, averaged over six water molecules. Online content Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at 10.1038/s41557-026-02150-5. Supplementary information Supplementary InformationSupplementary Figs. 1–24, Notes 1 and 2 and Tables 1 and 2. Supplementary Figs. 1–24, Notes 1 and 2 and Tables 1 and 2. Source data Source Data Fig. 2Source data for Fig. 2. Source Data Fig. 3Source data for Fig. 3. Source Data Fig. 4Source data for Fig. 4. Source Data Fig. 5Source data for Fig. 5. Source Data Fig. 6Source data for Fig. 6. Source data for Fig. 2. Source data for Fig. 3. Source data for Fig. 4. Source data for Fig. 5. Source data for Fig. 6.