Thinner-than-paper and broad-temperature-adaptive zinc-iodine batteries enabled by nanophase separated deep-eutectic hydrogel electrolytes

Abstract

Hydrogel electrolyte based secondary batteries are promising for wearable electronics, yet face challenges including limited mechanical resilience, and narrow temperature range. Herein, we report a robust deep-eutectic hydrogel electrolyte fabricated via synergistic interplay of dual nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks. The interwoven nanophase separation architecture, integrating hydrophilic polyvinyl alcohol phases and hydrophobic polyacrylonitrile phases, realizes high fracture-strength (4.1 MPa) and toughness (13.66 MJ m⁻³). Meanwhile, deep-eutectic chemistry modulates Zn²⁺ solvation structures and leverages cyano-coordination channels of polyacrylonitrile to achieve high Zn²⁺ ionic conductivity (28.2 mS cm⁻¹) and transference number (0.65) at 20 °C. Concurrently, abundant hydrogen bonds induced by multiple donor sites of hydrophilic phases, urethane, and Zn(ClO₄)₂ immobilize active H₂O to ensure broad-temperature durability. This tripartite synergy directs planar Zn deposition along (002) planes and suppresses dendrite growth, enabling Zn||I₂ batteries with a thinner-than-paper thickness (42 μm) and high flexibility. The assembled Zn||I₂ batteries demonstrate high specific energy (108.99 Wh kg⁻¹) and cycling stability (over 36,000 cycles under −40 to 80 °C). In this work, the convergence of molecule design, phase modulation, and process engineering establishes a feasible methodological framework for developing advanced flexible batteries that integrate high energy density and harsh environment tolerance.

Flexible batteries often fail under harsh conditions. Here, authors developed a hydrogel electrolyte via synergism of nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks, enabling flexible Zn||I₂ batteries with high specific energy and long lifespan across −40 to 80 °C.

Introduction

Aqueous zinc-metal batteries have emerged as promising candidates for next-generation energy storage, owing to their inherent safety, cost-effectiveness, and environmental compatibility^1–9. However, conventional liquid electrolytes (LE) suffer from critical limitations, including leakage susceptibility, narrow temperature range, and uncontrolled zinc dendrite growth, which severely restrict their practical deployment^10–14. Hydrogel electrolytes (HE), particularly polyvinyl alcohol (PVA)-derived HEs, can somewhat lock up water molecules by hydrogen bonds and inhibit dendrite growth, but still remain plagued by intrinsic trade-offs^15–21. Specifically, HE systems often restricted by the reconcile mechanical robustness with electrochemical performance, the excessive free water content that compromises temperature adaptability, and the thickness usually exceeding 100 μm that diminish energy density^22–29. Although phase-separated HE, as demonstrated in our prior work, have shown enhanced modulus and dendrite suppression, it remains a significant challenge to achieve simultaneous advances in thin geometries, broad-temperature resilience, and electrochemical stability³⁰.

Herein, we present the fabrication of a robust deep-eutectic HE (deep-eutectic-HE) via synergistic integration of nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks. The interwoven nanophase separation architecture, engineered through solvent-exchange, establishes a dual-domain structure comprising hydrophilic PVA phases and hydrophobic polyacrylonitrile (PAN) phases, endowing the hydrogel with good mechanical properties (4.1 MPa fracture strength and 13.66 MJ m⁻³ toughness). Concurrently, hydrogen-bond interactions between hydrophilic phases, urethane (UR) moieties, water molecules, and ClO₄⁻ ions immobilize active water, enabling stable operation across a broad temperature range (−40 to 80 °C). Deep-eutectic chemistry and cyano-coordination pathways of PAN collaboratively reconfigure Zn²⁺ solvation structures, achieving high ionic conductivity (28.2 mS cm⁻¹) and Zn²⁺ transference numbers (0.65), while suppressing polyiodide shuttling. This tripartite synergy directs planar Zn deposition on (002) crystallographic planes and mitigates dendrite formation, enabling Zn||I₂ full batteries with a thinner-than-paper thickness (42 μm) and high flexibility that deliver stable power under mechanical deformation (180° bending). The as-assembled Zn||I₂ pouch battery achieves high specific energy and volumetric energy density (108.99 Wh kg⁻¹ and 189.41 Wh L⁻¹), rate capability, and long lifespan under broad temperatures (with 92.31%, 75.28%, and 81.00% capacity retention after 36,000 cycles at −40, 20, and 80 °C, respectively). By synergistically optimizing molecular compositions, phase structures, and fabrication processes, a practical systematic approach is built to engineer advanced flexible batteries that simultaneously achieve good energy storage capacity and resistance to harsh environments.

Results

Design, preparation, and characterizations of deep-eutectic-HE

In this work, we propose a synergistic architectural strategy to construct a robust deep-eutectic-HE capable of reconciling thin geometries with good mechanical properties, electrochemical performances, and temperature adaptability. The stepwise fabrication strategy is realized by integrated manipulation of nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks in the HE matrix, as presented in Fig. 1a. Typically, the fabrication paradigm commences with dimethyl sulfoxide (DMSO) mediated dissolution of polyvinyl alcohol (PVA) and acrylonitrile (AN). The strong hydrogen-bond disrupting capability of DMSO enables nearly complete dissociation of intra- and inter-chain interactions, yielding a homogeneous polymeric precursor³¹. Subsequent heating initiates the polymerization of AN, driving the in-situ formation of a dual-network PAN-PVA composite through progressive crosslinking and chain entanglement. Finally, the DMSO removal and Zn²⁺ electrolyte exchange step establishes uniform polymer-electrolyte integration, and also optimizes zinc ion solvation structure via competitive coordination between unsaturated cyano groups and water molecules³⁰. During this phase evolution process, hydrophilic phases (rooted in the hydrophilic hydroxyl groups on PVA chains) and hydrophobic phases (formed via dipole-dipole interactions of PAN segments) spontaneously organize into a dual-nanophase separated architecture. This architecture, in synergy with the crystalline domains formed by a part of PVA segments, significantly enhances the mechanical strength of deep-eutectic-HE. Simultaneously, the hydrophilic hydroxyl groups of PVA contribute to expanding the polymer network, increasing the content of hydrated eutectic solvent in the hydrogel and forming abundant hydrogen bonds with H₂O to reduce the activity of H₂O. Moreover, the hydrophobic cyano groups on PAN chains construct a rapid transfer channel for Zn²⁺, thereby improving the electrochemical kinetics of deep-eutectic-HE^30,32.

Fig. 1. Design, preparation, and characterizations.

a The flowcharts for preparing deep-eutectic-HE through a solvent-exchange process. b ATR-FTIR and c Raman spectra of deep-eutectic-HE, deep-eutectic-LE, non-eutectic-HE, non-eutectic-LE, UR, and Zn(ClO₄)₂, respectively. d SEM image, e AFM image, f XRD patterns, and g SAXS patterns of deep-eutectic-HE. The inset in g is 2D SAXS patterns of deep-eutectic-HE. Source data for (b, c, f, g) are provided as a Source Data file.

To benchmark the advantages of deep-eutectic-HE across mechanical and electrochemical properties, we chose the non-gelated deep-eutectic liquid electrolytes (H₂O/Zn(ClO₄)₂/UR, namely deep-eutectic-LE), the non-eutectic liquid electrolytes (H₂O/Zn(ClO₄)₂, namely non-eutectic-LE), and gelated non-eutectic-LE (namely non-eutectic-HE) as the control groups. The attenuated total reflection Fourier transform infrared (ATR-FTIR) spectra (Fig. 1b) revealed critical molecular interactions among UR, Zn(ClO₄)₂, and trace H₂O. Compared with non-eutectic-HE and non-eutectic-LE, the N–H asymmetric stretching vibrations (3325 and 3415 cm⁻¹) of deep-eutectic-HE and deep-eutectic-LE underwent a 54–66 cm⁻¹ blue shift accompanied by peak broadening upon eutectic formation, indicating hydrogen-bond reorganization between UR molecules, water clusters, and ClO₄⁻ anions due to the addition of UR. Simultaneously, the C = O stretching vibration of deep-eutectic-HE and deep-eutectic-LE red-shifted by 6 cm⁻¹ (1687 → 1681 cm⁻¹), confirming Zn²⁺ coordination to carbonyl oxygen atoms within UR³². Complementary Raman analyses (Fig. 1c) demonstrated vibrational mode modifications consistent with this interaction. The C–N and N–H stretching frequencies of UR in deep-eutectic-HE and deep-eutectic-LE shifted by +17 and +10 cm⁻¹ respectively, reinforcing evidence for three-dimensional hydrogen-bond network formation³².

Scanning electron microscopy (SEM) image of deep-eutectic-HE (Fig. 1d) revealed a well-defined nanophase-separated structure with phase dimensions of 10–30 nm, substantially below the wavelengths of visible light (400–700 nm). This contrasts sharply with non-eutectic-HE (Supplementary Fig. 1), where the salting-in effect induced by excess ions suppresses phase separation, resulting in a porous morphology with limited hierarchical structure³³. Specifically, the poorly hydrated chaotropes anions (ClO₄⁻) will induce hydrophilic effects of the polymer chain segments, while weakening the hydrogen bonds between polymer chain segments, causing the polymer to dissolve (i.e., salting in). Consequently, the compaction degree of the PVA network is significantly reduced, making it difficult to form a nanoscale phase separation structure with PAN³⁴. The atomic force microscopy (AFM) mapping (Fig. 1e and Supplementary Figs. 2–3) further validated the nanophase-separated structure of deep-eutectic-HE, revealing its phase size of 20 nm approximately³³. The small-scale nanophase separation structure not only significantly enhanced the toughness of deep-eutectic-HE, but also improved its transparency. The optical implications of this nano-structuring become apparent in Supplementary Fig. 4, where both deep-eutectic-HE and non-eutectic-HE exhibit >90% visible-light transmittance – a direct consequence of subwavelength phase-separation dimensions³⁰. X-ray diffraction (XRD) patterns (Fig. 1f) provided crystallographic evidence for enhanced structural order, with deep-eutectic-HE displaying intensified (101) and (200) diffraction peaks of PVA at 18.0° and 21.3°, respectively. This improved crystallinity established a structural foundation for good mechanical performance through optimized polymer chain packing³⁵. Given the complexity of structural domains of deep-eutectic-HE, small-angle X-ray scattering (SAXS) profiles (Fig. 1g) further quantitatively resolved the nanophase-separated structure, revealing two characteristic distances at 25.7 nm and 10.9 nm corresponding to hydrophobic and crystalline domain spacing, respectively³⁶. These dimensions, which aligned with SEM-observed phase separation, also explained the enhanced optical clarity and mechanical robustness of deep-eutectic-HE through nanoconfined polymer dynamics.

Mechanical properties and temperature tolerance of deep-eutectic-HE

The mechanical properties of deep-eutectic-HE were systematically evaluated to elucidate the synergistic effects of nanophase-separated architecture, hydrated eutectic solvation, and hydrogen-bond network optimization. Tensile testing (Fig. 2a, b) demonstrated good mechanical properties in deep-eutectic-HE relative to non-eutectic-HE, with enhancements in fracture strain (752%), fracture strength (4.09 MPa), elastic modulus (1.84 MPa), and toughness (13.66 MJ m⁻³). These good mechanical properties of deep-eutectic-HE stem from two complementary mechanisms. First, ClO₄⁻, as chaotropic ions positioned on the right side of the Hofmeister series, typically induce polymer dissolution through salting-in effects^34,37. In deep-eutectic-HE, however, UR-mediated hydrogen bonds with ClO₄⁻ mitigates this destabilizing interaction while reinforcing the hydrogel network. Second, replacing DMSO (a good solvent) with deep-eutectic-LE (a poor solvent) in deep-eutectic-HE shifts thermodynamic equilibria toward polymer chain association, promoting crystalline domain formation as evidenced by XRD analysis (Fig. 1f)³⁸. This nanophase separation engineering strategy enabled fabrication of thin deep-eutectic-HE membranes down to 6 μm thickness (Fig. 2c), which maintain strong load-bearing capacity by sustaining 2.2 kg weights without fracture (Fig. 2d)—a 44-fold improvement over non-eutectic-HE.

Fig. 2. Mechanical properties and temperature tolerance.

a Tensile-strain curves, b elastic modulus, and toughness of deep-eutectic-HE and non-eutectic-HE. The error bars in (b) indicate the standard deviation (SD). c, d Photographs of deep-eutectic-HE (with a thickness of 6 μm) lifting heavy objects. e, f DSC curves of various electrolytes in (e) low and (f) high temperature ranges. g Temperature dependence of G′ and G″ for deep-eutectic-HE and non-eutectic-HE. h MD simulation snapshot and i corresponding Zn²⁺ solvation structure of deep-eutectic-HE. j Number of hydrogen bonds between H₂O molecules, k RDF, coordination number, and l Zn²⁺ diffusion coefficient in various electrolytes calculated via MD simulation. Source data for (a, b, e–g, j–l) are provided as a Source Data file. Source data for (h–i) are provided as Supplementary Data 1–2.

Differential scanning calorimetry (DSC) profiles in Fig. 2e also revealed distinct phase-transition behaviors among the four electrolyte samples. Deep-eutectic-HE and non-eutectic-HE demonstrated glass transition temperatures of −85.7 °C and −80.5 °C, whereas deep-eutectic-LE and non-eutectic-LE exhibited exothermic crystallization peaks at −43 and −81.8 °C, respectively. Benefiting from the high concentration of Zn(ClO₄)₂, non-eutectic-LE and non-eutectic-HE exhibit similar low freezing points³⁹. For deep-eutectic-LE, the addition of UR increases hydrogen bonds with H₂O, restricting movements such as vibration or rotation of the H₂O molecules, resulting in an elevation in the glass transition temperature⁴⁰. For deep-eutectic-HE, the dense network structure effectively confines H₂O molecules within nanoscale pores, thereby inhibiting the nucleation and growth of ice crystals⁴¹. Compared to non-eutectic-HE, the additional 5.2 °C glass transition temperature reduction in deep-eutectic-HE underscored the critical role of eutectic chemistry in optimizing low-temperature adaptability.

Thermal endurance characterization via DSC (Fig. 2f) also highlighted contrasting high-temperature responses. Neither deep-eutectic-HE nor deep-eutectic-LE displayed endothermic transitions between 20–120 °C, confirming their structural integrity under relatively high temperatures⁴². This stability arises from the capacity of deep-eutectic system to minimize free water content through UR-mediated hydrogen-bond networks and nanoconfined water states. In contrast, non-eutectic-HE and non-eutectic-LE underwent a phase transition at 72.3 and 74.0 °C due to unbound water evaporation, as evidenced by its larger number of hydrogen bonds among H₂O. Rheological profiling (Fig. 2g) corroborated this thermal endurance, showing stable storage modulus (G′) and loss modulus(G″) for deep-eutectic-HE during heating cycles versus progressive stiffening in non-eutectic-HE due to free water evaporation. Water retention studies (Supplementary Fig. 5) quantified the hydration dynamics underpinning this thermal stability: deep-eutectic-HE retained 88.4% water content after 48 h at 25 °C/15% RH, higher than non-eutectic-HE (78.5%) through UR-mediated water immobilization.

Molecular dynamics (MD) simulations (Fig. 2h and Supplementary Fig. 6) quantitatively resolved water confinement states in these electrolyte systems. Figure 2j demonstrated that non-eutectic-HE and non-eutectic-LE contain the most hydration bonds between H₂O molecules, rendering them susceptible to both ice crystallization at low temperatures and dehydration-induced structural collapse at elevated temperatures. Conversely, deep-eutectic-HE and deep-eutectic-LE exhibited reduced number of hydrogen bonds, confirming effective water activity suppression through the combined effects of hydrogen-bond networks and eutectic solvent interactions. Moreover, deep-eutectic-HE exhibited an even lower number of hydrogen bonds compared to deep-eutectic-LE, attributed to the hydrogen-bond adsorption of H₂O by the polymer network of hydrophilic PVA phases. These MD results aligned well with DSC-observed thermal behaviors, indicating the restricted water mobility directly correlates with enhanced cryogenic performance and evaporation resistance.

MD simulation simultaneously unveiled alterations of Zn²⁺ solvation structures (Fig. 2i and Supplementary Fig. 6, with source data provided as Supplementary Data 1–8). For non-eutectic-LE, radial distribution function (RDF, g(r)) analysis confirmed a first solvation shell coordination number of 6 for H₂O around Zn²⁺ (Fig. 2k). Post-gelation, the emergence of coordination interactions between Zn²⁺ and cyano in PAN chains led to their incorporation into the first solvation shell, thereby diminishing the hydration number to 5.37. Strikingly, the introduction of UR induced a pronounced reduction in the hydration number within the first solvation shell of Zn²⁺, dropping below 3 in both deep-eutectic-HE and deep-eutectic-LE systems. This substantial perturbation of the Zn²⁺ solvation structure correlated with an enhancement in the diffusion coefficient of Zn²⁺, exhibiting a 9-fold increase in deep-eutectic-HE relative to the non-eutectic LE (Fig. 2l), thereby enabling its higher ionic conductivity.

Electrochemical performances of deep-eutectic-HE

The electrochemical performances of deep-eutectic-HE were systematically investigated through Zn||Zn and SS (stainless steel)||SS symmetric batteries. Temperature-dependent ionic conductivity profiling (Supplementary Fig. 7 and Table 1) demonstrated efficient charge transport capabilities of deep-eutectic-HE across −40 to 80 °C. The deep-eutectic-HE maintained 28.2 mS cm⁻¹ ionic conductivity at 20 °C and retained 1.7 mS cm⁻¹ at −40 °C. This is a direct consequence of hydrogen-bond-mediated water confinement that suppresses ice nucleation while preserving ion mobility. Comparative analyses (Fig. 3a, Supplementary Fig. 8 and Table 2–4) showed deep-eutectic-LE, non-eutectic-HE and non-eutectic-LE exhibited inferior room-temperature conductivities (11.0, 8.0, and 6.2 mS cm⁻¹, respectively), with deep-eutectic-LE becoming non-conductive below −40 °C due to crystalline phase formation. The structure of non-eutectic-HE collapsed at 80 °C (Supplementary Fig. 8c), further highlighting the critical role of UR-enhanced hydrogen-bond networks in stabilizing hydrated eutectic architectures.

Fig. 3. Electrochemical performances.

a Ionic conductivities, b Zn²⁺ transference numbers, and c interfacial impedances of deep-eutectic-HE, deep-eutectic-LE, non-eutectic-HE, and non-eutectic-LE at different temperatures, respectively. The error bars in (a, b) indicate SD. Points, lines and the inset in (c) are experimental data, fitting results and equivalent circuit, respectively. R₀, R_el, and CPE in equivalent circuit are body resistance, interface impedance of electrode/electrolyte, and constant phase element, respectively, with corresponding values detailed in Supplementary Table 5. d Rate performances, e Tafel curves, and f–h galvanostatic cycling stabilities of Zn||Zn symmetric batteries at various temperatures of f 20 °C, g −40 °C, and h 80 °C, respectively. i Performance comparisons between the deep-eutectic-HE and other recently reported HEs in terms of fracture strength, ionic conductivity, electrochemical stability window, and low-/high-temperature tolerance. The source of the literature data shown in (i) can be found in Supplementary Table 6–7. Source data for (a–i) are provided as a Source Data file.

Zn²⁺ transference number (t_Zn²⁺) analyses (Fig. 3b and Supplementary Fig. 9) quantified Zn²⁺ transport kinetics of deep-eutectic-HE (t_Zn²⁺ = 0.62–0.68 across −40 to 80 °C), higher than deep-eutectic-LE (t_Zn²⁺ = 0.36–0.38), non-eutectic-HE (t_Zn²⁺ = 0.43–0.46) and non-eutectic-LE (t_Zn²⁺ = 0.41–0.46). This enhancement arises from synergistic effects: cyano groups on PAN chains coordinate with Zn²⁺ ions to establish rapid ionic migration channels while hydrogen-bond UR-ClO₄⁻ complexes restrict anion mobility, collectively restructuring the Zn²⁺ solvation structure (Fig. 2i)^43–46. Electrochemical impedance spectroscopy (EIS) measurements (Fig. 3c and Supplementary Table 5) revealed interfacial resistance values of 238.22 Ω for Zn|deep-eutectic-HE|Zn battery, significantly lower than those of deep-eutectic-LE (465.98 Ω), non-eutectic-HE (663.73 Ω) and non-eutectic-LE (599.18 Ω) counterparts, attributable to optimized electrode-electrolyte adhesion.

Rate performance evaluations (Fig. 3d and Supplementary Fig. 10) revealed that deep-eutectic-HE maintains a voltage polarization below 200 mV across a broad current density range (0.2–5.0 mA cm⁻²), contrasting sharply with non-eutectic-HE, which exhibited large polarization (>5.0 V) at 5.0 mA cm⁻². This improved electrochemical performance of deep-eutectic-HE primarily stems from its optimized Zn²⁺ ion transport kinetics and reduced interfacial impedance⁴⁷. Tafel analyses (Fig. 3e) corroborated these findings, showing its low corrosion current (0.06 μA cm⁻²) and suppressed polarization behavior, again verifying its ability to achieve stable Zn²⁺ ion plating/stripping behavior.

Long-term cycling testing (Fig. 3f) demonstrated the improved electrochemical stability of deep-eutectic-HE, sustaining operation for 4200 h with minimal polarization (<100 mV at 0.5 mA cm⁻² and 0.5 mAh cm⁻²). This starkly contrasts with the rapid failures in deep-eutectic-LE (100 h), non-eutectic-HE (200 h), and non-eutectic-LE (230 h) where Zn dendrite penetration caused the short circuits of batteries. The Zn|deep-eutectic-HE|Zn symmetric battery exhibited highly reversible Zn plating/stripping behavior, unequivocally validating the synergistic advantages of deep-eutectic-HE: rapid Zn²⁺ ion transport kinetics, robust dendrite suppression capability, and prolonged cyclability.

Broad-temperature evaluations (Fig. 3g, h) validated the broad temperature adaptability of deep-eutectic-HE. The critical working temperature tests of Zn||Zn symmetric batteries assembled with different electrolytes revealed that the Zn|deep-eutectic-HE|Zn symmetric battery could maintain stable electrochemical performance up to ~80 °C, which is better than those of deep-eutectic-LE ( ~ 70 °C), non-eutectic-HE ( ~ 50 °C), and non-eutectic-LE ( ~ 60 °C) counterparts (Supplementary Fig. 11). When the operation temperature was raised higher than the critical working temperature, significant ionic insulation would occur due to the precipitation of salts from solvent evaporation for LE samples or the collapse of gel structure for HE samples. The Zn|deep-eutectic-HE|Zn symmetric battery maintained stable cycling for more than 800 h with polarizations of 153.4 mV (at −40 °C) and 56.6 mV (at 80 °C), enabled by UR-H₂O hydrogen bonds that suppress H₂O-H₂O network formation. In contrast, non-eutectic-HE exhibited large voltage polarization (>1.0 V) at −40 °C, and failed at 80 °C due to structural collapse. Meanwhile, deep-eutectic-LE and non-eutectic-LE suffered rapid short-circuiting under high temperatures. In conclusion, deep-eutectic-HE displayed the ability to optimize ion transport kinetics, inhibit Zn dendrite growth, and maintain structural stability under demanding conditions.

The electrochemical stability window of deep-eutectic-HE was also investigated via linear sweep voltammetry. Supplementary Fig. 12 demonstrated that deep-eutectic-HE achieved an expanded electrochemical stability window (2.69 V), surpassing deep-eutectic-LE (2.57 V), non-eutectic-HE (2.49 V), and non-eutectic-LE (2.44 V). This enhancement arises from restricted water molecule mobility within the hydrogen-bond eutectic network, which suppresses parasitic reactions at elevated potentials. To contextualize these results, Fig. 3i and Supplementary Table 6–7 compared deep-eutectic-HE against other recent reported HEs across several critical performance parameters: fracture strength (4.09 MPa), ionic conductivity (28.2 mS cm⁻¹), electrochemical stability window (2.69 V), low-temperature tolerance (−40 °C), high-temperature tolerance (80 °C), film thickness (6 μm), coordination number of Zn-H₂O (2.71), Zn²⁺ transference number (0.65), cycling stability, and battery lifespan^{16,30,48–58}. Overall, the as-proposed deep-eutectic-HE combines these attributes, establishing it as a multifunctional electrolyte platform.

Zinc electro-plating/stripping behaviors guided by deep-eutectic-HE

To comprehensively investigate the capacity of deep-eutectic-HE to regulate zinc deposition behavior, multi-scale analyses of Zn||Cu half batteries were conducted. Due to the inferior electrochemical performances of non-eutectic-HE and non-eutectic-LE, deep-eutectic-LE was chosen as a control group for subsequent tests. The average Coulombic efficiency, a critical metric for evaluating zinc utilization reversibility, was measured over 20 electro-plating/stripping cycles using the Aurbach method (Fig. 4a)⁵⁹. The deep-eutectic-HE achieved an average Coulombic efficiency of 98.47%, higher than deep-eutectic-LE (96.99%) by 1.48%. Voltage-time profiles (Fig. 4b) further revealed the reduced polarization of deep-eutectic-HE, with a 33% reduction in hysteresis compared to deep-eutectic-LE during the initial deposition phase.

Fig. 4. Zinc electro-plating/stripping behaviors.

a Potential-time curves and b their enlarged views of Zn||Cu half batteries assembled with deep-eutectic-HE and deep-eutectic-LE, respectively. c Voltage profiles during the initial Zn deposition process and d CV curves of Zn||Cu half batteries assembled with deep-eutectic-HE and deep-eutectic-LE, respectively. e XRD patterns and f, g SEM images of Zn electrodes retrieved from Zn||Zn symmetric batteries assembled with f deep-eutectic-HE and g non-eutectic-LE after Zn electro-plating/stripping processes. Source data for (a–e) are provided as a Source Data file.

Detailed overpotential analyses (Fig. 4c and Supplementary Fig. 13) quantified the kinetic advantages of deep-eutectic-HE. The nucleation overpotential (42.6 mV vs. 66.5 mV) and mass-transfer overpotential (22.0 mV vs. 46.4 mV) of deep-eutectic-HE relative to deep-eutectic-LE were reduced by 35.9% and 52.6%, respectively. These improvements stem from synergistic effects between molecular-scale coordination chemistry and macro-scale mechanical properties. At the atomic level, cyano groups along the polymer backbone in PAN form transient coordination complexes with Zn²⁺, creating low-energy migration pathways that lower activation barriers for Zn²⁺ transport³⁰. Macroscopically, the high elastic modulus of deep-eutectic-HE generates substantial mechanical resistance against the growth of zinc dendrites during zinc deposition, physically constraining vertical growth while promoting lateral spreading of Zn across the electrode surface^30,60,61.

Cyclic voltammetry (CV) provided additional evidences on optimized reaction kinetics (Fig. 4d). The deep-eutectic-HE exhibited a peak response current density of 15.9 mA cm⁻² with minimal polarization (33 mV) compared to deep-eutectic-LE (9.2 mA cm⁻² and 68 mV). The 72.8% current density enhancement directly results from the coordinated action of accelerated Zn²⁺ mobility and spatially confined deposition geometry. This dual mechanism ensures rapid ion replenishment at the electrode-electrolyte interface while maintaining planar deposition morphology through stress distribution.

Crystallographic orientation analyses via XRD revealed fundamental differences in zinc deposition patterns (Fig. 4e). After 50 cycles (100 h), deep-eutectic-HE enhanced the (002) plane diffraction intensity by 40.97% relative to pristine zinc foil, indicating preferential horizontal growth along this low-surface-energy crystallographic orientation^62–64. In sharp contrast, deep-eutectic-LE increased (101) plane dominance, promoting vertically zinc dendrite formation, ultimately leading to severe Zn dendrite growth. SEM analyses of cycled zinc negative electrodes in deep-eutectic-HE and deep-eutectic-LE revealed stark morphological contrasts. Figure 4f demonstrates that deep-eutectic-HE enables horizontally aligned, layered zinc deposition, maintaining negative electrode surface uniformity over prolonged cycling. This planar growth mode ensures stable ion flux and interfacial integrity. In contrast, deep-eutectic-LE promotes vertical dendrite formation within 50 cycles, with progressive sharpening leading to severe surface roughness (Fig. 4g). After 200 cycles in deep-eutectic-LE, interconnected protrusions were observed on the Zn electrode, heightening the risks of separator penetration. XRD data (Fig. 4e) corroborate these findings, showing preferential horizontal (002) plane deposition in deep-eutectic-HE versus longitudinal crystallographic growth in deep-eutectic-LE. The zinc deposition behavior in different electrolytes was observed by in-situ optical microscopy (Supplementary Fig. 14). In deep-eutectic-HE, the Zn surface is flat and smooth, without obvious protrusions throughout the whole deposition process. In contrast, in deep-eutectic-LE, Zn protrusions emerge on Zn surface after consecutive deposition for 20 min, and obvious Zn dendrites are observed after 40 min. These results are consistent well with the SEM observations (Fig. 4f, g), indicating that the deep-eutectic-HE is effective in inhibiting the growth of Zn dendrites. The combined XRD, SEM and in-situ optical microscopy evidences conclusively established the dual functionality of deep-eutectic-HE in sustaining electrochemical stability and enforcing dendrite-suppressive deposition, critical for long-term battery reliability.

Electrochemical performances of zinc-metal batteries

For practical validation, we constructed full batteries using zinc foil negative electrodes and iodine positive electrodes. To ensure compatibility with the thin deep-eutectic-HE (6 μm), I₂/carbon nanotubes (CNTs) composite positive electrode was developed by a convenient filtration and adsorption method (Supplementary Fig. 15). Zn||I₂ full batteries were then assembled with thin I₂/CNTs positive electrode (20 μm thick). The robust nanoarchitecture enables fabrication of 6-μm-thick deep-eutectic-HE membranes while resisting dendrite penetration (Supplementary Fig. 16 and 17). By integrating Zn foil negative electrode (10 μm), deep-eutectic-HE (6 μm), I₂/CNTs positive electrode (20 μm), and polyimide (PI) package (3 μm × 2), the Zn||I₂ full battery achieved high compactness (42 μm total thickness, Fig. 5a, g), representing 55% reduction compared to standard A4 paper (104 μm, ISO 216). The achieved thickness is competitive compared to some recent relevant reports^{16,30,48–55}.

Fig. 5. Electrochemical performances of zinc-metal batteries.

a Schematic configuration of thinner-than-paper Zn||I₂ battery based on deep-eutectic-HE. b Rate performances of Zn|deep-eutectic-HE|I₂ and Zn|deep-eutectic-LE|I₂ batteries at 0.5 to 10 A g⁻¹ with a negative/positive (N/P) ratio of 18.50–55.50. c–e Long-term cycling stability of Zn||I₂ batteries at the current density of 5 A g⁻¹ with a negative/positive (N/P) ratio of 18.50–55.50 at various temperature of c 20 °C, d −40 °C, and e 80 °C. f Comparisons between Zn|deep-eutectic-HE|I₂, Zn|deep-eutectic-HE|V₂O₅, and other recent reported zinc-metal batteries in terms of specific energy and volumetric energy density (calculated based on the entire soft-packaged pouch battery assembly: discharge energy divided by the total mass or total volume encompassing the polyimide encapsulation material, positive electrode, deep-eutectic hydrogel electrolyte, and negative electrode). The source of the literature data shown in (f) can be found in Supplementary Table 8. Source data for (b–f) are provided as a Source Data file.

During the discharging process of Zn||I₂ batteries, I₂ in the positive electrode undergoes a reduction reaction and converts into I⁻, which combines with Zn²⁺ in the electrolyte to form ZnI₂ and adsorbs in the positive electrode. Oppositely, a reverse pathway occurs during the charge process. Notably, the tailored eutectic chemical environment induces a paradigm shift in iodide ion chemistry through multi-scale confinement effects (Fig. 5a). The I⁻ solubility undergoes a three-order-of-magnitude reduction from 9.22 M in conventional aqueous systems to merely 0.79 M in deep-eutectic-LE (Supplementary Fig. 18), fundamentally inhibiting the loss of active substances and altering the iodine redox equilibrium (I₂ + I⁻ ⇌ I₃⁻)⁶⁵. This dramatic solubility suppression originates from synergistic steric hindrance and electronic modulation. The hydrogen-bond network between UR, H₂O and ClO₄⁻ creates molecular-scale barriers that physically restrict I⁻ accommodation⁶⁵. Density functional theory (DFT) simulations quantitatively verified this dual confinement mechanism. As illustrated in Supplementary Fig. 19a (with source data provided as Supplementary Data 9–12), the diffusion energy barrier for I⁻ migration increases exponentially from 1.98 eV in non-eutectic system to 3.69 eV in the deep-eutectic system. This is an 86% enhancement that effectively mitigates I₂ shuttling at the molecular level. Concurrently, the formation energy of I₃⁻ complexes rises by 32% (0.280 eV in deep-eutectic system vs. 0.173 eV in non-eutectic system, Supplementary Fig. 19b), indicating thermodynamic disfavoring of polyiodide generation. Moreover, the coordinated evolution of deep-eutectic-HE simultaneously locks activity water molecules via hydrogen bonds with UR and ClO₄⁻ ions while reducing Zn²⁺ hydration number, collectively conferring resistance to both low and high temperatures without sacrificing ionic transport efficiency⁴².

Rate capability tests in Fig. 5b revealed that the Zn|deep-eutectic-HE|I₂ battery maintains a specific discharge capacity of 198.39 mAh g⁻¹ when returning to 0.5 A g⁻¹ after high-rate cycling at 10 A g⁻¹, demonstrating 99.80% capacity retention relative to initial low-rate value. This improved reversibility contrasts sharply with the deep-eutectic-LE system, which exhibits rapid capacity degradation, delivering only 97.86 mAh g⁻¹ at 10 A g⁻¹ with 79.25% capacity recovery upon rate reduction (Fig. 5b and Supplementary Fig. 20). The enhanced rate performance of deep-eutectic-HE benefits from its faster Zn²⁺ transfer kinetics and higher zinc electro-deposition/stripping homogeneity^20,66,67. In contrast, the inferior rate performance of deep-eutectic-LE control sample originates from three synergistic limitations. Firstly, sluggish Zn²⁺ transport kinetics of deep-eutectic-LE arise from the unoptimized Zn²⁺ solvation structure, causing insufficient dissociation and reduced ionic conductivity under rapid ion flux demands at high rates. Secondly, the absence of a robust hydrogen-bond network in deep-eutectic-LE induces elevated interfacial impedance and uncontrolled Zn dendrite, both of which consuming active Zn and electrolyte, resulting in irreversible capacity loss. Thirdly, the low Zn²⁺ transference number of deep-eutectic-LE exacerbates a concentration polarization across electrodes at high rates, further limiting positive electrode capacity utilization. These factors collectively result in the poor rate performance of deep-eutectic-LE control sample.

Long-term cycling stability of Zn|deep-eutectic-HE|I₂ battery was assessed subsequently (Fig. 5c), maintaining a Coulombic efficiency of >99.5% over 36,000 cycles at 5 A g⁻¹, an initial specific capacity of 146.15 mAh g⁻¹, and a capacity retention of 75.28% (average decay: 0.0007% per cycle). When cycled at 2 A g⁻¹, the Zn|deep-eutectic-HE|I₂ battery also displayed an initial specific capacity of 170.28 mAh g⁻¹ and a capacity retention of 65.77% (average decay: 0.0034% per cycle) after 10,000 cycles (Supplementary Fig. 21). Thermogravimetric analysis offered a clear explanation for the capacity degradation during long-term cycling (Supplementary Fig. 22)⁵⁸. After 10,000 cycles at 2.0 A g⁻¹, it was discovered that the I₂ content retained in the positive electrode had diminished to 71.39% of its initial level, which could be ascribed to tardive I⁻ dissolution and polyiodide shuttling behaviors. Comparatively, deep-eutectic-LE battery exhibited rapid capacity decay at 5.0 and 2.0 A g⁻¹, attributed to its lower Zn²⁺ transport kinetics, uncontrolled dendrite growth, and electrolyte decomposition. Long-term cycling tests were also conducted at a high I₂ loading of 4.0–5.0 mg cm⁻² (Supplementary Fig. 23). The Zn|deep-eutectic-HE|I₂ battery similarly demonstrated better electrochemical performances, boasting an initial specific capacity of 144.32 mAh g⁻¹ and a capacity retention of 88.35% (average decay: 0.0011% per cycle), which markedly surpasses those of Zn|deep-eutectic-LE|I₂ battery.

The comprehensive rate capability and cycling assessments validate that deep-eutectic-HE delivers high cycling stability and good rate performance in Zn||I₂ full batteries, surpassing conventional HEs. Self-discharge tests were also performed to evaluate the suppression of I⁻ dissolution and polyiodide shuttling effects in deep-eutectic-HE (Supplementary Fig. 24). After standing for 24 h, the capacity retention of Zn|deep-eutectic-HE|I₂ battery is 93.50%, much higher than that of Zn|deep-eutectic-LE|I₂ battery (80.37%), indicating the improved capability of deep-eutectic-HE to inhibit I⁻ dissolution and polyiodide shuttling.

Temperature resilience tests (Fig. 5d, e) further highlighted the broad temperature adaptability of deep-eutectic-HE. Under −40 °C conditions, the Zn|deep-eutectic-HE|I₂ battery enabled 36,000 cycles with a capacity retention of 92.31% (initial: 97.05 mAh g⁻¹; decay rate: 0.0002%/cycle). At 80 °C, the Zn|deep-eutectic-HE|I₂ battery delivered an initial specific capacity of 146.15 mAh g⁻¹ and a capacity retention of 81.00% after 36,000 cycles (decay rate: 0.0002%/cycle). In sharp contrast, deep-eutectic-LE counterparts failed catastrophically within 10 cycles under high temperatures due to frozen electrolyte channels (upon low temperature) and accelerated corrosion (upon high temperature). The broad temperature resilience demonstrated by deep-eutectic-HE under varied temperature conditions (−40 °C to 80 °C) confirms its robust thermomechanical stability and preserves ionic transport characteristics across a wide temperature range. This temperature-adaptive behavior, arising from its synergistic interplay of nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks, established its viability for deployment in geographically diverse environments spanning arctic to desert climates.

Soft-packaged pouch Zn|deep-eutectic-HE|I₂ batteries demonstrated high specific energy of 108.99 Wh kg⁻¹ and high volumetric energy density of 189.41 Wh L⁻¹ when paired with thicker electrodes, a performance competitive compared to some recent reported zinc-metal batteries (Fig. 5f, Supplementary Fig. 25 and Supplementary Table 8)^68–75. To validate the extensive and universal applicability of deep-eutectic-HE, we also fabricated and tested full battery based on V₂O₅ positive electrode material. The assembled Zn|deep-eutectic-HE|V₂O₅ full battery achieved good energy storage performance, delivering a specific energy of 76.89 Wh kg⁻¹ and a volume energy density of 236.62 Wh L⁻¹, both metrics significantly surpassing recent reported zinc-metal batteries. Furthermore, the practical feasibility was validated by three series-connected thinner-than-paper Zn||I₂ batteries that successfully powered a 5×10 light emitting diode (LED) array (3.0 V operational threshold) under demanding conditions, including: (1) severe mechanical deformation (e.g., repeated 180° folding into paper-airplane like origami structures), (2) complete fragmentation via scissor dissection, and (3) cycling under low and high temperatures (−40 °C to 80 °C), as visualized in Supplementary Fig. 26–27. Even under 90° and 180° bending states, the capacity retention trends of the thinner-than-paper Zn||I₂ pouch batteries during 3000 cycles are almost identical to the normal unbent state (Supplementary Fig. 28), confirming that the batteries maintain stable electrochemical performances even under significant mechanical deformation, highlighting its potential for practical flexible applications. Moreover, to better demonstrate practical viability, we have integrated three series-connected thinner-than-paper batteries into a watch band for powering a wristwatch, as well as charging a smartphone (Supplementary Fig. 29 and Movie 1–2). These demonstrations highlight the integrability and processability of our electrolyte in more complex and demanding flexible electronic devices.

Discussion

In summary, we report a design of robust deep-eutectic-HE that simultaneously achieves good mechanical stability, high Zn²⁺ conductivity, and broad temperature tolerance. The synergistic effects of nanophase separation, hydrated eutectic solvation, and hydrogen-bond networks endow deep-eutectic-HE with good mechanical properties (4.1 MPa fracture strength and 13.66 MJ m⁻³ toughness) and stable operation from −40 to 80 °C. It effectively regulates Zn²⁺ transport pathways, promoting uniform zinc deposition and suppressing dendrite formation. The resultant materials demonstrate strong mechanical resilience, high ionic conductivity, and broad tolerance to varied temperature conditions. Comprehensive electrochemical and morphological analyses confirm these mechanisms, revealing enhanced cycling durability (36,000 cycles at −40, 20, and 80 °C) and improved rate capability in Zn||I₂ full batteries. The thin thickness of deep-eutectic-HE enables high specific energy and volumetric energy density of 108.99 Wh kg⁻¹ and 236.62 Wh L⁻¹, coupled with good flexibility and mechanical durability under demanding conditions. Through the synergistic integration of molecular design, phase modulation, and process engineering, this work develops a practical methodological framework to advance flexible batteries with high energy density and harsh-environment durability for wearable electronics.

Methods

Chemicals and materials

Zn(ClO₄)₂.6H₂O, urethane (UR), dimethyl sulfoxide (DMSO, 99%), acrylonitrile (AN, 99%), N, N’-methylenebis(acrylamide) (MBA, 99%), ammonium persulfate (APS, 98%), activated carbon (AC, phosphotungstic acid modified), iodine (I₂, 99.99%), vanadium oxide (V₂O₅, 99.99%), N-Methylpyrrolidone (NMP, 99.9%), and lithium iodide (LiI, 99%) were purchased from Aladdin. Polyvinyl alcohol (PVA, M.W. 89,000–98,000, 99% hydrolyzed) was purchased from Sigma Aldrich. Ethanol and acetone were purchased from Wuxi Yasheng Chemical Co., Ltd. Triton X-100 surfactant (10%, Sterile) was purchased from Shanghai Biyun Tian Biotechnology Co., Ltd. Ketjen black (KB, ECP600JD) was purchased from Suzhou Sinero Co., Ltd. Polyvinylidene difluoride (PVDF, 99%) was purchased from Macklin (Shanghai, China). Zinc foils (99.999%, 10–20 μm) were purchased from Shanghai Wenheng Metal Materials Co., Ltd. Cu foils (99.99%, 16 μm) and stainless steel (SS, 316 L, 10 μm) foils were purchased from Hefei Kejing Material Co., Ltd. Glass fiber separators (aperture of 2.7 μm) were purchased from Hangzhou Situofan Biotechnology Co., Ltd. Aqueous microporous filtration membranes (0.22 μm) were purchased from Tianjin Jinteng Technology Co., Ltd. All materials were directly used without further purification. Amino-functionalized carbon nanotubes (amino-CNTs, 98%) were purchased from Shenzhen Suiheng Technology Co., Ltd. All metal ingots are used directly without any rolling, surface cleaning, or storing in inert environments treatments.

Preparation of deep-eutectic liquid electrolyte (deep-eutectic-LE) and non-eutectic liquid electrolyte (non-eutectic-LE)

The deep-eutectic-LE sample is prepared by mixing Zn(ClO₄)₂, deionized water, and UR in a molar ratio of 1:8:4.5 and stirring at 75 °C for 30 min. The non-eutectic-LE sample is prepared by mixing Zn(ClO₄)₂ and deionized water in a 1:8 molar ratio and stirring at 20 °C for 30 min.

Preparation of deep-eutectic hydrogel electrolyte (deep-eutectic-HE) and non-eutectic hydrogel electrolyte (non-eutectic-HE)

The deep-eutectic-HE and non-eutectic-HE samples are prepared through a convenient solvent-exchange process. Briefly, a certain amount of PVA powder is dissolved in DMSO under vigorous stirring at 95 °C for about 30 min to obtain a uniform 5 wt.% PVA solution. Subsequently, AN (3 M), MBA (3 mM), and APS (2 mM) are added to the PVA solution under an ice-water bath. Then, the above solution is transferred to a customized mold and copolymerized at 70 °C for 2 h in air isolation, leading to the pristine organogel. Finally, deep-eutectic-HE or non-eutectic-HE sample is prepared by immersing the organogel in deep-eutectic-LE or non-eutectic-LE respectively for 48 h to complete the solvent-exchange between DMSO and deep-eutectic-LE or non-eutectic-LE.

Preparation of positive electrode

The I₂/CNTs positive electrode is prepared by a filtration and adsorption method. Briefly, amino-CNTs (108 mg), AC (12 mg), KB (6 mg), and Triton X-100 surfactant (1 mL) are dissolved in deionized water. Then, the above solution is subjected to ultrasonic dispersion treatment in an ice-water bath for 200 min using an ultrasonic cell disruptor to obtain a uniform mixture solution. Next, the above solution is vacuum filtered via an aqueous microporous filtration membrane. Afterwards, the filtration membrane is dissolved and washed by acetone, and the resulting CNTs film with a thickness of 20–200 μm is dried in a vacuum oven at 80 °C for 24 h. The obtained self-supporting CNTs film is stamped into circular electrode plates with a diameter of 10 mm. Subsequently, the CNTs film electrode is immersed in an ethanol solution of 0.1 M I₂ for 24 h to fully adsorb I₂ into the CNTs film electrode (I₂/CNTs). Finally, the I₂/CNTs electrode is transferred to a vacuum oven and dried at 80 °C for 24 h to obtain the I₂/CNTs positive electrode. The loading amount of I₂ active material is varied to be about 0.5–1.5 mg cm⁻² (unless otherwise noted), 4.0–5.0 mg cm⁻² (for high I₂ loading tests), or 18–22 mg cm⁻² (for the assembly of soft-packed batteries), which is determined by the thickness of CNTs film. All I₂/CNTs positive electrodes are used directly without any rolling, surface cleaning, or storing in inert environments treatments.

The V₂O₅ positive electrode is prepared using a scratch coating method. Briefly, 140 mg of V₂O₅, 40 mg of KB, and 20 mg of PVDF (with a mass ratio of 7:2:1) were evenly ground with an agate mortar, and mixed with NMP, and stirred for 24 h to form a uniform slurry. The slurry is coated on a SS foil (thickness of 10 μm) with a stainless-steel scraper and dried in a vacuum oven at 60 °C for 24 h. After drying, the V₂O₅ positive electrode is rolled by a roller machine with a pressure of 10 MPa. The active mass loading is 19–22 mg cm⁻². All V₂O₅ positive electrodes are used directly without any surface cleaning or storing in inert environments treatments.

Material characterizations

Attenuated total reflection Fourier transform infrared (ATR-FTIR) spectra are collected using a Thermo Fisher Scientific Nicolet iS10. Raman spectra were collected using a Horiba JY Raman spectrometer equipped with a 473 nm laser light source. Scanning electron microscopy (SEM) images are obtained using FEI Nova NanoSEM-450 field emission scanning electron microscopy. The 2D and 3D topographies of the samples are observed with a Bruker Dimension Icon atomic force microscopy (AFM) instrument. The optical transmittance measurements are measured via a Shimadzu UV-2700 UV-vis-NIR spectrophotometer. X-ray diffraction (XRD) data is obtained using a Bruker D8 ADVANCE powder X-ray diffractometer. Small angle X-ray scattering (SAXS) data are measured on the Xeuss 2.0 SAXS system. The mechanical properties are measured using an Instron 3343 instrument. Differential scanning calorimetry (DSC) tests are conducted using a PerkinElmer Pyris 1 analyzer. The rheological tests are performed on a TA DHR-2 rheometer. The temperature-dependent sweeps are conducted ranging from 10 to 100 °C at 0.1% strain and 5 rad s⁻¹. The moisture retention rate is tested by drying method. Specifically, a standard cylindrical sample is exposed to an environment of 25 °C and 15% relative humidity (RH). During this process, the weight of the sample is regularly measured, and the water retention of the sample over time is calculated using Eq. (1), as provided below. The percentage content of I₂ in the electrode is evaluated via thermogravimetric (TG) analysis on a Netzsch STA 449 C thermal analyzer with a heating rate of 10 °C min⁻¹ from 30 to 500 °C under a dry air atmosphere. Some electrodes and electrolytes are directly characterized after preparation, while others are obtained from aqueous batteries after specific electrochemical tests. In the air environment, stainless-steel pliers are used for disassembling CR2032 batteries, and stainless-steel tweezers are used to transfer electrodes and electrolytes. After transferring from the battery, the negative electrodes are rinsed with deionized water and dried at 60 °C before further characterization, while positive electrodes and electrolytes are dried without rinsed. The inert atmosphere has never been used.

Fabrication of the batteries

The electrochemical performances of LE and HE samples are evaluated using CR2032 batteries or soft-packaged pouch batteries. The SS||SS symmetrical batteries are composed of two SS foils (diameter of 12 mm, thickness of 10 μm) and a specific electrolyte. The Zn||Zn symmetrical batteries are composed of two Zn foils (diameter of 12 mm, thickness of 10 μm) and a specific electrolyte. The Zn||Cu half batteries are composed of a Zn foil negative electrode (diameter of 12 mm, thickness of 10 μm), a Cu foil positive electrode (diameter of 12 mm, thickness of 16 μm), and a specific electrolyte. The Zn||SS half batteries are composed of a Zn foil negative electrode (diameter of 12 mm, thickness of 10 μm), a SS foil positive electrode (diameter of 12 mm, thickness of 10 μm), and a specific electrolyte. The Zn||I₂ CR2032 batteries are composed of a Zn foil negative electrode (diameter of 12 mm, thickness of 10 μm), a I₂ positive electrode (diameter of 12 mm, thickness of 20 μm), and a specific electrolyte. The single layer Zn||I₂ soft-packaged pouch batteries are composed of a layer of Zn foil negative electrode (5.0 cm × 4.0 cm, thickness of 10 μm), a layer of I₂ positive electrode (5.0 cm × 4.0 cm, thickness of 20 μm, active mass loading of 1.0–1.5 mg cm⁻² or 4.0–5.0 mg cm⁻²), and a layer of deep-eutectic-HE (6.0 cm × 5.0 cm, thickness of 10 μm). The multi-layer Zn||I₂ soft-packaged pouch batteries are composed of 11 layers of Zn foil negative electrodes (5.0 cm × 4.0 cm, thickness of 10 μm), 10 layers of I₂ positive electrodes (5.0 cm × 4.0 cm, thickness of 190 ± 10 μm, active mass loading of 18–22 mg cm⁻²), and 20 layers of deep-eutectic-HEs (6.0 cm × 5.0 cm, thickness of 10 μm). The multi-layer Zn||V₂O₅ soft-packaged pouch batteries are composed of 11 layers of Zn foil negative electrodes (5.0 cm × 4.0 cm, thickness of 15 μm), 10 layers of V₂O₅ positive electrodes (double-side coated, 5.0 cm × 4.0 cm, thickness of 100 ± 10 μm, active mass loading of 19–22 mg cm⁻²), and 20 layers of deep-eutectic-HEs (6.0 cm × 5.0 cm, thickness of 10 μm).

The diameter, thickness and shell material of the above-mentioned CR2032 battery are 20 mm, 3.2 mm, and stainless steel, respectively. The above-mentioned soft-packaged pouch batteries are packaged with polyimide films (thickness of 3 μm) under a vacuum condition, without applying external pressure during the charging/discharging cycling process. The above-mentioned electrolytes refer to deep-eutectic-HE (without any separator), deep-eutectic-LE (with glass fiber separators), non-eutectic-HE (without any separator), or non-eutectic-LE (with glass fiber separators). The thickness of deep-eutectic-HE, non-eutectic-HE and glass fiber separator are 10 μm, 100 μm and 250 μm, respectively. The size of deep-eutectic-HE, non-eutectic-HE and glass fiber separator used in CR2032 batteries are 16 mm in diameters. The size of deep-eutectic-HEs used in soft-packaged pouch batteries are 5.5 cm × 4.5 cm. All electrolytes are used within 7 days after preparation, storing in glass bottles (LEs) and cling films (HEs) under air atmosphere (20 ± 2 °C). LEs and HEs are transferred with a pipettor (polypropylene pipette tip) and a stainless-steel tweezer under air atmosphere (20 ± 2 °C), respectively. The glass fiber separators are cut with a stainless-steel circular slicer and are directly wetted by dripping electrolyte during battery assembly instead of prewetting. The negative/positive ratio (N/P) of the above-mentioned Zn||I₂ coin batteries, Zn||I₂ soft-packaged pouch batteries and Zn||V₂O₅ soft-packaged pouch batteries are calculated to be 18.50–55.50 or 5.55–6.94, 1.26–1.54 and 1.35–1.57 (based on the active mass loading of positive electrodes).

Electrochemical measurements

The cyclic voltammetry (CV), linear sweep voltammetry (LSV), alternating-current electrochemical impedance spectroscopy (EIS), chronoamperometry (CA), and Tafel curves are collected on RST-5200F and CHI-760 electrochemical workstations at 20 ± 2 °C. The CV and LSV tests are performed at a scan rate of 1.0 mV s⁻¹, with a potential range of −0.2–0.4 V and 1.0–3.0 V (vs. Zn/Zn²⁺), respectively. The EIS tests are performed with an amplitude of 5 mV in the frequency range between 100 kHz and 0.01 Hz at 20 ± 2 °C. The measurement is conducted after holding the batteries at the specified temperature for 1 h during open-circuit potential. Nyquist plots obtained from EIS tests are fitted using ZView software. The CA tests are conducted after EIS tests immediately with a polarization voltage of 10 mV and a discharging time of 600 s. The Tafel curves are collected at the scanning rate of 1 mV s⁻¹ across a voltage range of ±180 mV relative to the open-circuit voltage.

Galvanostatic charge/discharge cycling tests and rate performance measurements are conducted on a NEWARE CT-4008Tn battery analyzer using constant current techniques in a incubator (−40 ± 2 °C, 20 ± 2 °C or 80 ± 2 °C). The Zn||Zn symmetrical batteries are tested at the current density ranging from 0.5–5.0 mA cm⁻² with an area capacity of 0.5 mAh cm⁻². The average Coulomb efficiency is measured with the Zn||Cu half batteries using the following procedure: a discharging process is first performed at a current density of 0.5 mA cm⁻² for 1 h. Then, 20 charging/discharging cycles are performed at 0.5 mA cm⁻² with an area capacity of 0.5 mAh cm⁻². Finally, the battery is charged to 0.5 V (vs. Zn/Zn²⁺) at a current density of 0.5 mA cm⁻². The average Coulombic efficiency is then calculated via Eq. 4. The Zn||I₂ batteries are tested at the current density ranging from 0.5–10.0 A g⁻¹ with a discharging cut-off potential of 0.7 V (vs. Zn/Zn²⁺) and a charging cut-off potential of 1.5 V (vs. Zn/Zn²⁺). The Zn||V₂O₅ batteries are tested at the current density of 0.5 A g⁻¹ (based on the mass of V₂O₅ positive electrode active materials) with a discharging cut-off potential of 0.2 V (vs. Zn/Zn²⁺) and a charging cut-off potential of 1.6 V (vs. Zn/Zn²⁺). All electrochemical tests are conducted in a gas environment of air. No local temperature measurements are carried out. The specific capacities and current density of Zn||I₂ and Zn||V₂O₅ batteries are calculated according to the mass of positive electrode active materials.

All electrochemical measurements are repeated at least 3 times to ensure statistical reliability and reproducibility. The values of ionic conductivity (5 repetitions) and Zn²⁺ transference number (5 repetitions) reflect statistical analysis of all tested batteries in parallel experiments, as these tests allow high-throughput measurement under identical conditions, and enable rigorous error quantification critical for characterization. Conversely, the CV (3 repetitions), LSV (3 repetitions), Tafel (3 repetitions) and long-cycling performance (3–5 repetitions) data presented in the figures represent measurements from a specific representative battery selected from multiple identical experiments (exhibited median performance in several repetitions). This approach is adopted because a single battery can display its detailed electrochemical performance in a complete curve, while the statistical analysis inevitably loses some critical information, especially the large amount of data of galvanostatic charge/discharge cycling tests (over 1000 cycles).

Calculation equations

The water retention ($W_r$) is calculated according to the following equation:

$$W_r=\frac{W_s}{W_0}\times 100\%$$ 1

where $W_0$ and $W_s$ are the mass of the sample after exposure for 0 and s h, respectively.

The ionic conductivity ($\mathrm{\sigma }$) is calculated according to the following equation:

$$\sigma =\frac{l}{R\bullet S}$$ 2

where $l$ and $S$ represent the thickness and area of the electrolyte, respectively. $R$ is the bulk resistance of the electrolyte.

The Zn-ion transference number ($t_{{Zn}^{2+}}$) is calculated according to the following equation:

$$t_{{Zn}^{2+}}=\frac{I_S\left(\mathrm{\Delta }V-I_0{R}_{el}^0\right)}{I_0\left(\mathrm{\Delta }V-I_s{R}_{el}^s\right)}$$ 3

where $\mathrm{\Delta }V$ is the polarization voltage (30 mV), $I_0$ and $I_s$ are the initial current and steady current, respectively. $R_{el}^0$ and $R_{el}^s$ are the initial interfacial resistance and steady interfacial resistance, respectively.

The average Coulombic efficiency ($C{E}_{avg}$) during the Zn plating/stripping process in Zn||Cu half batteries is calculated according to the following equation:

$${CE}_{avg}=1-\frac{Q_1-Q_r}{Q_{c}N}$$ 4

where $Q_1$ is the initial Zn plating capacity (5.0 mAh cm⁻²), $Q_r$ is the capacity retention measured by the final Zn stripping process, $Q_c$ is the fixed plating or striping capacity (0.5 mAh cm⁻²) involved in a single plating or stripping process (half cycle), and $N$ is the cycling number of plating/stripping process.

The percentage content of I₂ in the electrode ($M_{{\mathrm{I}}_2}$) is calculated according to the following equation:

$$M_{{\mathrm{I}}_2}=\frac{M_{{\mathrm{I}}_2+\mathrm{C}+{\mathrm{H}}_2\mathrm{O}}-M_{\mathrm{C}+{\mathrm{H}}_2\mathrm{O}}}{1-M_{\mathrm{C}+{\mathrm{H}}_2\mathrm{O}}}$$ 5

where $M_{{\mathrm{I}}_2+\mathrm{C}+{\mathrm{H}}_2\mathrm{O}}$ and $M_{\mathrm{C}+{\mathrm{H}}_2\mathrm{O}}$ are the weight loss of electrodes containing I₂ and electrodes without I₂ in TG tests, respectively.

The percentage content of I₂ retained in the positive electrode (I₂ remained) is calculated according to the following equation:

$${\mathrm{I}}_2\mathrm{remained}=\frac{M_{\mathrm{After}}\left(1-M_{\mathrm{Before}}\right)}{M_{\mathrm{Before}}\left(1-M_{\mathrm{After}}\right)}$$ 6

where $M_{\mathrm{Before}}$ and $M_{\mathrm{After}}$ are the percentage content of I₂ in the positive electrode before and after charging/discharging cycling tests, respectively.

The standard deviation (SD) of the samples is calculated according to the following equation:

$$\mathrm{SD}=\sqrt{\frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{x}}_{\mathrm{i}}-\overline{\mathrm{x}}}{\mathrm{n}-1}}$$ 7

where ${\mathrm{x}}_{\mathrm{i}}$, $\overline{\mathrm{x}}$, and $\mathrm{n}$ are the i-th sample data, sample mean, and sample size, respectively. All error bars mentioned in the text represent SD.

The specific capacity of the batteries is calculated according to the following equation:

$$\mathrm{S}\mathrm{pecific}\mathrm{capacity}=\frac{C}{M_p}$$ 8

where $C$ and $M_p$ are the capacity (mAh) and mass (g) of the positive electrode active materials, respectively.

The specific energy of the batteries is calculated according to the following equation:

$$\mathrm{S}\mathrm{pecific}\mathrm{energy}=\frac{E}{M}$$ 9

where $E$ and $M$ are the total energy (Wh) and total mass (kg) of the battery, respectively.

The volumetric energy density of the batteries is calculated according to the following equation:

$$\mathrm{V}\mathrm{olumetric}\mathrm{energy}\mathrm{density}=\frac{E}{V}$$ 10

where $E$ and $V$ are the total energy (Wh) and total volume (L) of the battery, respectively.

Theoretical simulations

All theoretical calculations are performed using Materials Studio. The geometric structures and binding energies between PAN, PVA, UR, H₂O, and Zn(ClO₄)₂ are optimized using the DMol3 module. The gradient approximation (GGA) of the Perdew-Burke-Ernzerh (PBE) functional is used to describe electron exchange and associated energy. Density functional theory (DFT) - D3 method is used in the calculation to consider van der Waals forces. The plane wave cutoff energy is set to 520 eV, and the energy convergence value between two consecutive steps is less than 10⁻⁵ eV. The maximum allowable force is <0.002 eV/Å, and the maximum allowable displacement is <0.005 Å.

Four electrolyte models are constructed to simulate four as-prepared electrolytes: deep-eutectic-HE, deep-eutectic-LE, non-eutectic-HE, and non-eutectic-LE, respectively. Molecular dynamics (MD) simulations are performed on four electrolyte models using the Forcite module. The Condensed Phase Optimized Molecular Potential (COMPASS III) force field is selected for atomic simulation studies for all MD simulations. For each system, the isothermal isobaric ensemble (NPT) is first used to pre-equilibrate the system for 1 ns. Then, a 5 ns MD run is performed in the canonical ensemble (NVT) to achieve equilibrium, followed by a 5 ns NVT simulation to obtain the final electrolyte model. The time step of the MD simulation process mentioned above is set to 1 fs.

Author contributions

Z.J. conceived the idea of this study. T.Y.S. and Z.J.C. designed the experiments. T.Y.S., Z.J.C., J.Y.W., K.X.H., W.M., J.C.L., and H.G.L. performed sample preparations. T.Y.S., Z.J.C., Y.M.Y., and Z.X.T. performed electrochemical measurements, battery tests, and data analyses. T.Y.S. and Y.X.Y. performed the theoretical calculations. Z.J., C.H.L., T.Y.S., Z.J.C., and Q.C.Y. wrote the manuscript. Z.J. and C.H.L. supervised the project. All authors participated in the scientific discussion of this project.

Peer review

Peer review information

Nature Communications thanks Junnan Hao and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The authors declare that all data supporting the findings of this study are available within the paper and its Supplementary Information. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71312-3.