Introduction

Aqueous zinc sulfur (Zn||S) batteries have emerged as the highly competitive choice for large−scale energy storage systems and mobile electronics, attributing to the high intrinsic safety and cost−effectiveness (2.9 US$ kgZn−1 and 0.25 US$ kgS−1). Sulfur, as the positive electrode material, exhibits a extremely high theoretical specific capacity of 1675 mAh g−1, which is obviously superior to the reported intercalated positive electrodes for aqueous zinc ion batteries, such as manganese oxides, vanadium oxides, Prussian blue analogs and organic compounds1,2. The cathodic sulfur redox reactions in aqueous Zn||S batteries commonly involve the direct solid−solid conversion between elemental sulfur and zinc sulfide (ZnS), which can thoroughly circumvent the notorious polysulfides dissolution and shuttling issues in the non−aqueous lithium sulfur (Li||S) batteries3,4. However, aqueous medium also brings out several adverse effects5,6,7. First, the reaction kinetics is believed to be sluggish due to the high activation energy barrier required for the intricate structural decomposition and regeneration during sulfur reduction reaction and ZnS oxidation. Second, poor wettability of sulfur electrode by aqueous solution severely impedes the interfacial Zn−ions transport which is mainly ascribed to the inherent hydrophobicity of sulfur and ZnS, leading to high potential polarization. Moreover, the thermodynamically unstable ZnS could react with H2O via the disproportionation to form irreversible SO42− by−products in water solvent, especially under high oxidation potential during the charge process, resulting in the loss of active material, low coulombic efficiency and poor reversibility.

In the past few years, some preliminary efforts have been dedicated to design and optimize functional sulfur positive electrodes for aqueous Zn||S batteries, typically by exploiting the conductive host materials, such as porous cabon matrixes and conductive polymers, to enhance the charge and electrolyte transport in the positive electrodes8. Moreover, various catalyst materials have been explored to improve the slow reaction kinetics of solid−phase conversion between Li2S2 and Li2S in non−aqueous Li||S batteres9,10. Such the electrocatalytic strategy could also be adopted to effectively boost sulfur conversion reaction in Zn||S batteries11,12,13. For example, atomic iron catalysts with Fe−N4 coordination have been exploited as the bidirectional electrocatalyst to simultaneously accelerate the complete sulfur reduction reaction and reduce the energy barrier for ZnS decomposition during the discharging and charging processes13. Furthermore, electrolyte engineering is recognized to be an effective strategy for constructing electrode interfacial chemistry and improving the battery performance. Some electrolyte addtives, such as iodine−containing species14,15,16,17 and thiourea18,19,20, have been developed to strongly interact with the discharge product ZnS, which results in the atomtic rearrangement of ZnS and weakens the Zn−S bonding, thus promoting the electrochemical decomposition of ZnS. Another enhancement is the cosolvent strategy by introducing organic solvents, which not only endows the electrolyte with good wettability for sulfur electrode, but also weakens Zn2+−water coordination and inhibits water−induced interfacial side reactions21,22,23,24,25,26,27. However, there is still a lack of fundamental understanding of the intricate relations between the interfacial structure/properties of sulfur electrode and reaction kinetics of sulfur conversion.

Electric double layer (EDL), occuring at the sulfur electrode/electroyte interface, plays a pivotal role in the charge transfer and ion transport process during the sulfur conversion reaction28,29,30,31. Yet, the interfacial organizations (i.e., the arrangement of anion, cations, or solvent molecules) near sulfur electrode surface have not received much attention. According to the Grahame and Bockris−Devanathan−Muller model, the EDL structure consists of an inner Helmholtz plane (IHP), an outer Helmholtz plane (OHP) and a diffuse layer32,33. In general, when the sulfur electrode is brought into contact with the aqueous electrolyte, adsorbed anions and water molecules are located in the IHP, and positive−charged solvated Zn2+ are located in the OHP. The sulfur reduction reaction in Zn||S batteries involves the Zn2+ transfer from the OHP to sulfur electrode surface, simultaneously accompanied by the electron transfer from the electrode to sulfur nanoparticles for phase conversion into ZnS. Due to the poor interface wettability, the Zn2+ transport/transfer within the EDL region would dominate the kinetics of cathodic sulfur reduction reaction. Thus, regulating the interfacial EDL structure and Zn2+ arrangement provides an intrinsic design principle to facilitate the sulfur conversion.

In this work, we design a hybrid electrolyte, namely, a conventional aqueous electrolyte with introduction of ZnI2 additive and organic N,N−dimethylformamide (DMF) cosolvent, in which the iodide species are functionalized for catalytic oxidation of ZnS, while DMF cosolvent can effectively facilitate sulfur reduction reaction. By combining operando Raman spectroscopy with non−destructive electrochemical impedance spectroscopy and theoretical calculations/simulations, the dynamic interfacial EDL structure and Zn2+ charge properties of sulfur electrode are unveiled at the microscopic level. Specifically, DMF cosolvent molecules can preferentially adsorb on the sulfur electrode surface and exhibit strong interaction with Zn2+ due to the electron−donating −C=O group, thereby reconstructing interfacial EDL with H2O−poor IHP and Zn2+−rich OHP, which favors interfacial Zn2+ transfer and facilitates solid−phase conversion of sulfur with a voltage hysteresis of 0.326 V. Moreover, the strong H−bonding interaction between DMF and water molecules brings down the content of free water and reduces the water reactivity, thus effectively suppressing the side reaction of ZnS disproportionation. By utilizing an additional constant voltage charge procedure to avoid I−/I3− redox shuttle, the Zn||S battery exhibits a long−term cycling stability with a capacity fading of 0.034% per cycle after 1000 cycles at 2 C (i.e., 3.34 A g−1). A high areal capacity of 7.68 mAh cm−2 is also achieved with a sulfur loading of 6.4 mg cm−2. Even at a low temperature of −10 °C, the hybrid electrolyte still enables the Zn||S battery to deliver a reversible specific capacity of 500 mAh g−1 after 100 cycles.

Results and discussion

Electrochemical characterization of sulfur conversion reaction

Sulfur/Ketjen black (S/KB) composite material was fabricated via a melt−diffusion method due to the high electrical conductivity, large specific surface area (1287.53 m2 g−1) and abundant mesoporous structure of KB (Supplementary Fig. 1). The thermogravimetry (TG) analysis indicates that the sulfur content in the S/KB composite is about 53.8 wt%, while the X−ray diffraction pattern presents weak characteristic signals of sulfur, suggesting that sulfur has been melted and impregnated into the mesopores of KB (Supplementary Fig. 2). Subsequently, Zn||S cells were assembled in the Swagelok−type configuration which applied the S/KB composite and zinc foil as the positive electrode and negative electrode, with three types of 2 M ZnSO4 electrolytes, including ZnSO4−H2O, ZnSO4−ZnI2−H2O (with 0.2 M ZnI2), and ZnSO4−ZnI2−H2O−DMF (with 0.2 M ZnI2 and 30 vol. % DMF) hybrid electrolyte. As shown from the galvanostatic charge−discharge (GCD) curves in Fig. 1a−c, all the Zn||S cells with different electrolytes demonstrate the gradually reduced electrode reaction polarizations, especially the discharge reactions, indicating an interfacial activation phenomenon during the cycling process (see the detailed discussion in the following section). Compared with ZnSO4−H2O and ZnSO4−ZnI2−H2O electrolytes, the Zn||S cell with ZnSO4−ZnI2−H2O−DMF electrolyte performs the minimum voltage hysteresis of 0.326 V for the reversible transformation process between sulfur and ZnS. This is consistent with the cyclic voltammetry (CV) results (Supplementary Fig. 3), where the highest redox−peak current and the lowest potential polarization can be clearly observed in the ZnSO4−ZnI2−H2O−DMF electrolyte. In addition, the Tafel plots of the reduction and oxidation peaks are analyzed to explicitly illustrate the working mechanism of ZnI2 additive and DMF cosolvent (Supplementary Fig. 4). For the reduction process (from sulfur to ZnS), the fitted Tafel sloples in the ZnSO4−H2O and ZnSO4−ZnI2−H2O electrolytes are 355.8 and 348.1 mV dec−1, which can be greatly reduced to 179.3 mV dec−1 by introducing the additional DMF cosolvent. For the oxidation process (from ZnS to sulfur), the Tafel slope shows an obvious decrease from 113.1 mV dec−1 to 70.3 mV dec−1 with the addition of ZnI2, and presents a further decline to 56.7 mV dec−1 in the ZnSO4−ZnI2−H2O−DMF electrolyte. It can thus be concluded that ZnI2 additive could be mainly conducive to lower the activation energy barrier of ZnS oxidation, while DMF cosolvent can effectively accelerate the reaction kinetics of sulfur conversion into ZnS.

Fig. 1: Electrochemical reaction behaviors of sulfur conversion.
figure 1

a–c Galvanostatic charge−discharge curves of Zn||S cells with ZnSO4−H2O, ZnSO4−ZnI2−H2O, and ZnSO4−ZnI2−H2O−DMF electrolytes at 25 °C and 0.1 C. d, e CV profiles of Zn||S cells under different scan rates from 0.1 to 0.5 mV s−1 in ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes. f Proportional relation between the reduction/oxidation peak current density and scan rate in logarithmic plots. g Linear fitting of average current density and scan rate to obtain the EDL capacitance of sulfur electrodes in different electrolytes. h Relation plots between accumulated charge (Q) and impulse potential according to the pulse voltammetry. i Ratio of the EDL capacitance and pseudocapacitive contribution.

Full size image

With regard to the catalytic mechanism of ZnI2 for the charging process, previous report has proved that I− can exhibit the strong interaction with ZnS to weaken the Zn−S bonding, rather than serving as the so−called redox mediator due to the higher electrochemcial oxidation potential of I− than ZnS (Supplementary Fig. 5)16. As the concentration of ZnI2 increases, Zn||S cells exhibit higher specific capacity, lower polarization potential and better cycling stability (Supplementary Fig. 6). The optimized concentration of ZnI2 is identified to be 0.2 M. To definitely validate the catalytic effect of DMF cosolvent, the Zn||S cell was assembled using the electrolyte without ZnI2 additives. As displayed in Supplementary Fig. 7, the voltage hysteresis also could be markedly decreased from 0.903 V in the ZnSO4−H2O electrolyte to 0.695 V in the ZnSO4−H2O−DMF electrolyte. Moreover, the rising discharge capacity and reduced polarization potential of Zn||S cells can be achieved with a higher volume concentration of DMF in the ZnSO4−ZnI2−H2O−DMF electrolyte (Supplementary Fig. 8), further confirming that DMF cosolvent can facilitate the sulfur conversion reactions, particularly the sulfur reduction reaction. The optimized volume concentration of DMF is identified to be 30 vol%, which has been used in the subsequent experiments unless otherwise stated. Therefore, this work will mainly focus on the promotion mechanism of sulfur reduction reaction aroused by the DMF cosolvent, and further understand the relationship between electrode reaction kinetics and structure/properties of sulfur electrode interface.

The CV measurements at different rates were carried out to investigate the charge storage behavior of sulfur electrodes in different electrolytes. According to the equation of J = aνb, the b value can be figured out by the slope of linear fitting between the reduction/oxidation peak current density and scan rate in logarithmic plots, which quantitatively determines the capacitive and diffusive contributions in the charge storage process34. As presented in Fig. 1d−f and Supplementary Fig. 9, the calculated b values of reduction peak 1 and oxidation peak 2 in the ZnSO4−ZnI2−H2O electrolyte are 0.67 and 0.46, respectively. With the addition of DMF cosolvent, higher b value can be attained for the reduction peak 3 (0.71) and oxidation peak 4 (0.60), suggesting that the Zn2+ storage process in sulfur electrode can endow more capacitive contribution, but is still controlled by the diffusion behavior. That is to say, the interfacial capacitance of sulfur electrode would be improved by introducing DMF cosolvent, which should include EDL capacitance (Cdl) and pseudocapacitance (Cpseudo). Based on the method proposed by Pu et al.35, the Cdl can be determined by the slope of the linear fitting between the average current density and the scan rate within the potential range from 0.92 V to 0.98 V vs. Zn2+/Zn, where no faradic process has been involved. As exhibited in Fig. 1g and Supplementary Fig. 10, Cdl could be greatly increased from 2.22 to 55.25 mF cm−2 after the addition of DMF cosolvent. Meanwhile, the total charge accumulation on sulfur electrodes can be quantified through pulse voltammetry36. Specifically, the assembled Zn||S cells were tested by applying a pulse potential from the aimed reduction potential rapidly switching to the open circuit potential. Then the accumulated charge (Q) could be obtained by integrating the cathodic transient current peak (Supplementary Fig. 11). As revealed in Fig. 1h, the accumulated charge in the ZnSO4−ZnI2−H2O−DMF electrolyte is obviously higher than that in absence of DMF cosolvent. By fitting the linear slope between accumulated charge (Q) and the corresponding potential (E) before the onset potential of sulfur reduction reaction, the total capacitance value can be determined to be 4.24 mF cm−2 for ZnSO4−ZnI2−H2O electrolyte, and 168.59 mF cm−2 for ZnSO4−ZnI2−H2O−DMF electrolyte. As shown in Fig. 1i, the pseudocapacitive contributions are calculated to be 2.02 mF cm−2 in ZnSO4−ZnI2−H2O electrolyte, and 113.34 mF cm−2 in ZnSO4−ZnI2−H2O−DMF electrolyte, respectively. Furthermore, Zeta potentials of sulfur electrodes in different electrolytes were measured to experimentally disclose the surface charge states. As presented in Supplementary Fig. 12, the sulfur electrode in the ZnSO4−ZnI2−H2O−DMF electrolyte shows a lower Zeta potential of −1.37 mV compared with that of −0.11 mV in the ZnSO4−ZnI2−H2O electrolyte, signifying that much more negative charges are adsorbed at the electrode interface. The higher capacitance achieved in the ZnSO4−ZnI2−H2O−DMF electrolyte at different potentials could also be verified by the alternating current (A.C.) voltammetry (Supplementary Fig. 13)37. It is thus reasonable to anticipate that DMF cosolvent would induce more Zn2+ charge accumulation in the interfacial EDL region to balance the surface negative charge, which is in accordance with the increased interfacial capacitance.

Interfacial activation and mechanism analysis of sulfur electrodes

The wettability of different electrolytes to sulfur electrode was explored. As shown in Fig. 2a, the droplet formed by the ZnSO4−ZnI2−H2O electrolyte on the surface of sulfur electrode gives a contact angle of 108.4°, which is decreased to 92.2° for the ZnSO4−ZnI2−H2O−DMF electrolyte. This indicates that DMF cosolvent endows the sulfur electrode with better electrolyte penetration, thereby promoting the interfacial Zn2+ ions transport and electrode reaction kinetics, which could also be verified by the higher diffusion coefficient (D) value and lower polarization overpotential from the galvanostatic intermittence titration techniques (GITT) curves (Supplementary Fig. 14). After five cycles of sulfur electrode, the contact angles for the both electrolytes are reduced to 103.0° and 72.0°, respectively, further confirming the existence of interfacial activation phenomenon.

Fig. 2: Interfacial activation and mechanism analysis of sulfur electrodes.
figure 2

a Contact angle tests of the ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes on sulfur electrode before and after five cycles at 25 °C and 0.1 C. b Calculated adsorption energy values of various electrolyte solvent molecules and ions with S8 and ZnS. c, d Operando Raman spectra of sulfur electrode surface during the cycling process in different electrolytes at 25 °C and 0.1 C. e Adsoption configuration and electrostatic potential distribution of DMF molecules adsorbed on the surface of ZnS. f Calculated HOMO energy levels of various electrode and electrolyte components.

Full size image

Density fuctional theory (DFT) calculations were conducted to understand the interaction between positive electrode components and electrolyte solvents. As depicted in Fig. 2b, the adsorption energies (Eads) of DMF on S and ZnS (the discharging product) surfaces are calculated to be −0.48 and −1.06 eV, respectively, which are more negative than the Eads of free H2O on S and ZnS surfaces (−0.19 and −0.46 eV), suggesting that DMF molecules can preferentially adhere to the surface of both S ans ZnS to form (S8∙DMF) and (ZnS∙DMF) components. In addition, the calculated Eads of Zn2+ on (S8∙DMF) and (S8∙H2O) are −1.67 and −0.34 eV, respectively, which indicates that DMF cosolvent could enhance the Zn2+ adsorption on the sulfur electrode surface. Meanwhile, the Eads of (ZnS∙H2O) and (ZnS∙DMF) toward I− are −2.37 and −2.98 eV, respectively, demonstrating that I− is more inclined to adsorb on the surface of ZnS to further promote its electrochemcial oxidation with the aid of DMF.

Operando Raman spectroscopy was exploited to illuminate the interfacial activation process of sulfur electrode. The configuration diagram of the operando Raman cell is presented in Supplementary Fig. 15. As shown in Fig. 2c and Supplementary Fig. 16, in addition to the characteristic Raman peaks of surface−adsorbed SO42− (980 cm−1), D band (1345 cm−1) and G band (1600 cm−1) of conductive carbon materials, a weak peak emerges at 1438 cm−1 after the initial charging process in the ZnSO4−ZnI2−H2O electrolyte which can be assigned to the –COO group38, indicating that the carbon substrate would be slightly oxidized during the charging process. In contrast, various new Raman peaks appear in the ZnSO4−ZnI2−H2O−DMF electrolyte (Fig. 2d and Supplementary Fig. 17). The peak centered at 487 cm−1 indicates the Zn−O bond39. The peaks located at 1251, 1438, and 1515 cm−1 are ascribed to the –COO group40. The peaks emerging at 2688 and 2869 cm−1 are associated with the –CHx group41. All the newly generated components should be attributed to the oxidative decomposition of DMF, which also has been verified by DFT calculation. As mentioned above, DMF molecules would preferentially adsorb on the surface of ZnS. The electrostatic potential (ESP) distribution in Fig. 2e presents that the negative charges are concentrated on carbonyl oxygen atom of DMF, thereby resulting in the formation of Zn−O bond. Meanwhile, as presented in Fig. 2f, the calculations on the highest occupied molecular orbital (HOMO) energy levels demonstrate that DMF shows a slightly higher HOMO energy level than that of ZnS, which indicates that DMF would be subjected to decomposition along with the electrochemcial oxidation of ZnS, thereby resulting in the plentiful –COO and –CHx groups formed at the sulfur electrode/electrolyte interface. These results elucidate the intrinsic mechanism of the interfacial activation, simultaneously interpreting the better electrode wettability and improved interfacial Zn2+ transport after electrochemical redox reactions.

To investigate the electrolyte consumption due to DMF degradation, the Zn||S cells after different cycles were disassembled, and the Raman spectra of the wetted separators were immediately acquired. As presented in the Supplementary Fig. 18, the intensity of C–H stretching of DMF located at 2950 cm–1 shows a continuous decline over the first 20 cycles, but stabilizes in subsequent cycles, which suggests that DMF decomposition does not always exist in the long−term cycling process. This can be well explained by theoretical calculations of the adsorption energies (Supplementary Fig. 19). Eads of DMF on the carbon substrate would be reduced from −1.29 eV to −0.44 eV after introducing the plentiful –COO groups on the surface, which indicates that the –COO groups could inhibit the surface adsorption of DMF and subsequent decomposition. Moreover, the Eads of Zn2+ on the carbon substrate without and with rich –COO groups are calculated to be −0.21 eV and −1.53 eV, which suggests that the surface –COO groups can function as a substitute of DMF for Zn2+ enrichment at the sulfur electrode interface to accelerate sulfur conversion reaction, which also has been experimentally verified (Supplementary Fig. 20). All the above results evidently demonstrate that the –COO groups, as the decomposition products of DMF, would weaken the DMF adsorption on the sulfur electrode surface and suppress further degradation. Meanwhile, it can similarly enhance interfacial Zn2+ adsorption to facilitate sulfur conversion reaction in the long−term cycling process.

Dynamic interfacial properties of sulfur electrode

To investigate the dynamic interfacial properties of sulfur electrode during the electrochemical process, in situ galvanostatic electrochemical impedance spectroscopy (GEIS) combined with the distribution of relaxation time (DRT) analysis was carried out at different potentials, as shown in Fig. 3a−d. A three−electrode Swagelok cell was utilized with an additional zinc foil as the reference electrode to eliminate the coupled responces of the zinc counter electrode (Supplementary Fig. 21). Instead of the conventional circuit fitting methods, DRT method offers the advantage of directly conducting impedance spectrum analysis, discerning dynamic characteristics based on time constants, and preventing subjective analysis errors caused by manual circuit fitting42. According to the transmission line model (TLM) of porous electrodes, the relaxation time (τ) identifications of different electrochemical processes during sulfur conversion reaction are first conducted. As shown Supplementary Fig. 22, the high−frequency impedance arc (10−4 s < τ < 10−2 s) should be attributed to the contact resistance between the electrode and the current collector (Rcon)43. The presence of non−vertical line in the mid−frequency region (10−2 s < τ < 1 s) generally represents the ion transport in the pore structure of electrode (Rion)44. And the low−frequency arc (1 s < τ < 102 s) can be interpreted as the interfacial charge transfer resistance (Rint), which is demonstrated to be the largest and dominates the kinetics of sulfur conversion reactions45,46. Notably, under the premise of sulfur nanoparticles accommodated in the conductive host material with good access to electrons, the interfacial charge transfer process of sulfur conversion should be mainly controlled by the interfacial Zn2+ transfer instead of electron transfer. The raw impedance spectra in Supplementary Figs. 23 and 24 are converted into DRT functions in Supplementary Figs. 25 and 26, respectively. The interfacial resistance values (Rint) can be obtained by transforming horizontal coordinate into the logarithm of τ and calculating the integral area of the characteristic peak. The corresponding interfacial capacitance (Cint) also can be determinated according to the equation of \(\tau={RC}\) (Supplementary Tables 1 and 2).

Fig. 3: Dynamic interfacial properties of sulfur electrode during the cycling process.
figure 3

a, c Potential profiles of GEIS measurements with a three−electrode Swagelok cell at 25 °C and 1 C and b, d the corresponding DRT spectra at different potentials in ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes. e The interfacial capacitance and Zn2+ transfer resistance variation with the potential. f, g Schematic of interfacial EDL structure of sulfur electrodes in different electrolytes.

Full size image

Figure 3e exhibits the dynamic evolution of Cint and Rint of sulfur electrodes in different electrolytes during the discharge−charge processes of the initial three cycles. As the sulfur redox reactions proceed, there are obvious differences of Rint and Cint in the discharging process. An inflection point of potential can be obtained to observe the significant variation of Cint, which is close to the plateau potential in the charge−dicharge curves (Fig. 3a, b), representing the occurrence of phase conversion reaction. Moreover, Rint displays a sharp decrease before the inflection point, suggesting that Zn2+ ions are enriched at the sulfur electrode interface to promote interfacial Zn2+ transfer kinetics as the interfacial potential difference increases, which corresponds to the pseudocapacitance behavior of sulfur reduction reaction. As shown in the related Bode plots (Supplementary Figs. 27 and 28), the gradually reduced phase angle in the low−frequency range further confirms the enhanced charge transfer kinetics47. Therefore, an interfacial charge accumulation−induced phase conversion reaction mechanism can be concluded. Enough Zn2+ ions enrichment in the interfacial EDL region is requisite to activate sulfur nanoparticles to undergo the phase transition into ZnS. In the ZnSO4−ZnI2−H2O electrolyte, the extremely low Cint and high Rint in the initial stage should be contributed to poor electrode wettability, causing that the sulfur phase−conversion reaction occurs at 0.45 V vs. Zn2+/Zn. With the introduction of DMF cosolvent, the initial Cint is significantly increased due to the specific adsorption and improved wettability, which indicates that more Zn2+ ions would be accumulated in outer Helmholtz layer at the sulfur electrode/electrolyte interface, thus facilitating the occurrence of sulfur conversion at 0.75 V vs. Zn2+/Zn, much higher than that (0.45 V) in the ZnSO4−ZnI2−H2O electrolyte. Moreover, the rised discharge plateau potential in the subsequent cycles is companied by the increased initial Cint, which coincides with the interfacial activation phenomenon mentioned above. To further investigate the dynamics changes of interfacial EDL structure during the phase transition reaction of sulfur, in situ GEIS and DRT analysis were conducted at different states of charge (Supplementary Figs. 29−33). The Rint keeps constant during the discharging and charging processes in the ZnSO4−ZnI2−H2O−DMF electrolyte, while the obvious fluctuation of Rint can be observed in the ZnSO4−ZnI2−H2O electrolyte, suggesting the formation of more stable interfacial EDL structure in the presence of DMF cosolvent. Furthermore, higher Cint in the discharging process also demonstrates the faster charge transfer kinetics during sulfur conversion reaction in the ZnSO4−ZnI2−H2O−DMF electrolyte.

All the above results demonstrate the differentiated interfacial EDL structure of sulfur electrode in the ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes. As illustrated in Fig. 3f, g, the preferential adsorption of DMF molecules would extrude H2O dipoles to form the H2O−poor inner Helmholtz layer (IHP), thereby inhibiting the H2O−induced disproportionation of ZnS in the charging process. Meanwhile, the adsorbed DMF could facilitate Zn2+ charge accumulation in the outer Helmholtz layer (OHP) due to the strong interaction between the carbonyl group (−C=O) of DMF and Zn2+. Therefore, DMF molecules could lead to the production of H2O−poor IHP and Zn2+−rich OHP at the interface of sulfur electrode, which in turn promotes the interfacial Zn2+ transfer and enhances the sulfur conversion reaction kinetics.

The sulfur conversion reaction mechanism was further elucidated through ex situ X−ray photoelectron spectroscopy (XPS). In the XPS spectra of S 2p (Supplementary Fig. 34), the pristine sulfur electrode shows double peaks at binding energy of 164.1 and 165.3 eV, corresponding to the S−S bond of S8 molecules. Upon completely discharging to 0.3 V, new peaks emerge at 162.2 and 163.3 eV due to the generation of ZnS, at which point the peaks of S8 completely disappear. When charged to 1.2 V, the peaks of sulfur reappear without the signal response of ZnS. Moreover, ex situ X−ray diffraction (XRD) was also conducted to investigate the phase transition during sulfur conversion process. At the initial discharge to 0.3 V, diffraction peaks appear at 28.70°, 33.23°, 48.63° and 58.32°, corresponding to the characteristic peak of ZnS. When charged from 0.3 V to 1.2 V, the ZnS peaks are gradually decreased, which indicates the reversible transformation of sulfur and zinc sulfide during the battery cycling process, but with no observed polysulfides. An optical electrochemical cell was assembled to visually observe the discharging process of sulfur electrode, further confirming the absence of polysulfide dissolution (Supplementary Fig. 35). These results convincingly support a solid−solid conversion mechanism of sulfur electrode in Zn||S batteries. During the discharging process, Zn2+/e− would directly incorporate into reactive sulfur species at the sulfur electrode interface, enabling S−S cleavage and Zn−S bonding for ZnS nucleation. Thus, constructing the Zn2+−rich EDL structure could favor the interfacial Zn2+ transfer to facilitate sulfur conversion reactions.

Zn2+ solvation structure and interfacial adsorption

Zn2+ solvation structures of different electrolytes were investigated via molecular dynamics (MD) simulations. The snapshots of the simulated electrolyte structure (Fig. 4a, b) along with the radial distribution functions (RDFs) and coordination number (CN) analysis (Fig. 4c, d) reveal that an evident Zn2+−DMF pair can be observed at ~2.08 Å after introducing DMF. The CN of H2O in the solvation shell of Zn2+ decreases from 6.1 to 4.8, while the CN of DMF increases from 0 to 0.15, indicating that DMF molecules are participating in the Zn2+ solvation structure. Moreover, the CN of SO42− also rises from 0.5 to 1.1, suggesting that the SO42− is more disposed to enter into the solvation shell of Zn2+. As shown in Fig. 4e, the content of free H2O decreases from 76.67% to 74.68%, which can be attributed to the strong H−bonding interaction between DMF and H2O, implying a lower possibility for the occurrence of side reactions, such as the ZnS disproportionation on the sulfur positive electrode side and hydrogen evolution reaction on the Zn negative electrode side. According to MD simulations, the molecular−form solvation structures can be expressed as Zn2+−(H2O)6.1∙(SO42−)0.5 for ZnSO4−ZnI2−H2O electrolyte and Zn2+−(DMF)0.15(H2O)4.8∙(SO42−)1.1 for ZnSO4−ZnI2−H2O−DMF electrolyte, which could be further simplified as Zn2+−(H2O)6∙(SO42−) and Zn2+−(DMF)(H2O)5∙(SO42−), respectively, for the convenience of subsequent interfacial adsorption investigation.

Fig. 4: Zn2+ solvation structure of electrolyte and interfacial adsorption of sulfur electrode.
figure 4

a, b 3D snapshots of ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes from MD simulations and partially enlarged snapshot representing the Zn2+ solvation structures. c, d Radial distribution functions (RDFs) and the corresponding coordination number among Zn2+, SO42−, H2O and DMF. e Percentage of free H2O in different electrolytes from MD simulations. f, g The charge density difference for demonstrating the interfacial adsorption model of sulfur electrode for Zn2+ solvation structures (front view). The yellow area indicates electron accumulation region, while the blue area represents electron depletion region. The adsorption energy and the Bader charge of Zn are presented. h Arrhenius plots showing the proportional relation between logarithmic values of the reductive current density at 0.75 V vs. Zn2+/Zn and the reciprocal of the absolute temperatures.

Full size image

Since the distinct solvation structure would inevitably differentiate electrode interfacial adsorption behavior, specific adsorption models on the pure sulfue electrode surface were constructed with the simplified solvation structures, including the anions, cations and solvent molecules48. The optimized geometry and corresponding differential charge density distributions of two adsorption configurations are manifested in Fig. 4f, g. The adsorption energy of the Zn2+−(H2O)6∙(SO42−) solvation structure at the sulfur electrode surface is determined as −0.57 eV, and the Bader charge of Zn indicates that Zn2+ ion would donate 0.0055 e to the sulfur electrode (Fig. 4f). By contrast, as presented in Fig. 4g, the solvation structure of Zn2+−(DMF)(H2O)5∙(SO42−) demonstrates a stronger adsorption (Eads = −0.78 eV) and more charge transfer (0.0061 e from Zn) toward sulfur electrode, indicating that Zn2+ ion is easier to participate in the sulfur conversion reaction. Moreover, the temperature−dependent CV tests were performed to experimentally determine the activation energy of sulfur conversion reaction in the discharging process according to the classic Arrhenius equation. As resulted in Fig. 4h and Supplementary Fig. 36, the activation energy of sulfur reduction reaction is decreased from 47.4 kJ mol−1 to 42.3 kJ mol−1 after introducing the DMF cosolvent.

Depressed water−induced side reactions

Various spectroscopic characterization techniques, including nuclear magnetic resonance (NMR), Raman spectroscopy, and Fourier transform infrared (FTIR) spectroscopy, were conducted to investigate the molecular−level interactions in hybrid electrolytes. As shown in Supplementary Fig. 37, the characteristic 1H peaks of H2O located at ~4.73 ppm would shift to the lower field with higher DMF concentration in the electrolyte. In addition, three characteristic 1H peaks are centered at 7.48 and 2.37/2.60 ppm, which correspond to the aldehydic hydrogen and methylic hydrogen of DMF, respectively. As the content of DMF increases, the signals of aldehydic hydrogen and methylic hydrogen shift to the high field. These shifts should be resulted from the H−bonding networks formed between the O atom in DMF and the H atom in H2O molecules (C=O∙∙∙H−O), which decreases the electron density of the hydroxyl hydrogen of H2O, and enhances the shielding effect on the aldehydic hydrogen and methylic hydrogen of DMF49. Fig. 5a, b show the regional Raman and FTIR spectra of ZnSO4−ZnI2−H2O electrolyte with different DMF volume fractions (the full Raman and FTIR spectra are displayed in the Supplementary Fig. 38). The Raman spectra of H2O can be differentiated into three bands located at 3228, 3425, and 3541 cm−1, which are assigned to the symmetric and asymmetric OH stretching and the free OH of water, respectively (Fig. 5a). The weakened OH stretching indicates that the addition of DMF significantly breaks the strong H−bonding networks between H2O and H2O. And the H−bonding networks between DMF and H2O molecules are constructed, which is confirmed by the blue shift of C−H stretching of DMF50. In the FTIR spectra, the O−H stretching vibration recorded within the region of 2800−3800 cm−1 can be separated into three deconvoluted local environments for H2O molecules: 3223, 3367 and 3592 cm−1, corresponding to the bulk, cluster and isolated H2O, respectively (Fig. 5b). Obviously, the relative content of bulk and cluster H2O are decreased, while the isolated H2O keeps almost unchanged with increasing DMF volume fractions, further demonstrating that the DMF addition disrupts the original H−bonding networks via the construction of DMF−H2O interactions51. This also could be illustrated by the adjacent characteristic peaks between the C=O bending vibration of DMF (1647 cm−1) and the O−H bending vibration of H2O (1633 cm−1). Besides, the stretching of SO42− shows an obvious downward trend upon DMF introduction (Supplementary Fig. 38), implying that SO42− is more inclined to incorporate into the solvation structure of Zn2+, which conforms to the results of MD simulation in Fig. 4. The above results demonstrate that the DMF addition could undermine the original H−bonding networks and stabilize H2O molecules with depressed reactivity via strong intermolecular DMF−H2O interaction.

Fig. 5: Depressed water reactivity and inhibited side reactions.
figure 5

a Raman and b FTIR spectra of ZnSO4−ZnI2−H2O electrolytes with different concentrations of DMF. c SEM images of sulfur positive electrodes and zinc negative electrodes after 150 cycles in ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes at 25 °C and 1 C. d Cycling performance of Zn||Zn symmetric cells in different electrolytes at 1 mA cm−2 and 1 mAh cm−2. XPS profiles of e S 2p, f O 1 s and g Zn 2p peaks for sulfur positive electrodes before and after 150 cycles in ZnSO4−ZnI2−H2O and ZnSO4−ZnI2−H2O−DMF electrolytes at 25 °C and 1 C.

Full size image

To investigate the regulation of hydrogen evolution reaction (HER) through DMF addition, linear sweep voltammetry (LSV) tests were conducted in the configuration of three–electrode Swagelok cells, which utilized Ti mesh as the working electrode, Zn foil as the counter electrode, and Ag/AgCl as the reference electrode, respectively. As shown in Supplementary Fig. 39a, the onset potential of zinc deposition and HER shift negatively after the addition of DMF cosolvent, which suggests that the reaction kinetics of both zinc deposition and hydrogen evolution are inhibited. To decouple the hydrogen evolution reaction from zinc deposition, the three–electrode cells were also assembled with 0.2 M KI−H2O and 0.2 M KI−H2O–DMF electrolytes, which applied Zn foil as the working electrode, carbon paper as the counter electrode, and Ag/AgCl as the reference electrode, respectively. As presented in Supplementary Fig. 39b, the obviously decreased current density could further confirm the suppressed HER by introducing DMF cosolvent into the electrolyte. Moreover, direct evidence for inhibited HER was obtained from in situ optical visualization experiments. As displayed in Supplementary Fig. 39c, the initially flat Zn surface in the ZnSO4−ZnI2−H2O electrolyte would become uneven and rough after 90 min of zinc deposition at 1 mA cm–2, along with a visible gas bubble on the surface. In comparison, uniform and compact zinc deposition without gas bubble formation can be observed for the Zn foil in the ZnSO4−ZnI2−H2O−DMF electrolyte under the same condition for over 120 min. All the above results clearly prove that the hydrogen evolution reaction is suppressed with the introduction of DMF in the electrolyte.

The surface morphologies of sulfur positive electrodes and zinc negative electrodes after cycling in different electrolytes were characterized via scanning electron microscopy (SEM) images. As revealed in Fig. 5c and Supplementary Fig. 40, both the sulfur positive electrode and zinc negative electrode cycled in the ZnSO4−ZnI2−H2O electrolyte present a rough surface with the emergence of scattered micro/nano flakes, while the uniform and compact morphology can still be maintained for the sulfur positive electrode and zinc negative electrode cycled in ZnSO4−ZnI2−H2O−DMF electrolyte. In addition, the power X−ray diffraction (XRD) patterns evidentially demonstrate that the flakes are ascribed to the formation of zinc sulfate hydroxide by−products (Supplementary Fig. 41), thus suggesting that the side reactions on the electrode surface can be significantly suppressed due to the reduced H2O reactivity by introducing DMF cosolvent, which is conducive to improve the cycling performance of sulfur positive electrode and zinc negative electrode. X−ray photoelectron spectroscopy (XPS) was also conducted to further explore the interface components of cycled zinc negative electrodes. Compared with cycled zinc negative electrode in ZnSO4–ZnI2–H2O electrolyte, the zinc negative electrode after cycled in ZnSO4–ZnI2–H2O–DMF electrolyte shows a significantly weakened SO42– signal (Supplementary Fig. 42), implying the suppressed generation of zinc sulfate hydroxide by–products, which is in line with the result from the XRD patterns of cycled zinc negative electrodes.

Subsequently, the cycling stability of Zn||Zn symmetric cells in different electrolytes under 1 mA cm−2 and 1 mAh cm−2 were evaluated. As shown in Fig. 5d and Supplementary Fig. 43, the symmetric cell assembled with ZnSO4−ZnI2−H2O electrolyte exhibits a lower voltage hysteresis and better cycling ability of ~6300 h in comparison with that in the ZnSO4−H2O electrolyte (<1000 h). The same result has also been verified by previous reports52,53,54, which demonstrate that the iodide ions can adsorb in the zinc surface, not only suppressing the water−induced side reactions and by−product formation, but also enhancing the kinetics of zinc deposition. More excitingly, the addition of DMF into the iodide−containing electrolyte could further prolong the cycling life of Zn||Zn symmetric cell to over 10,000 h despite of the slightly increased voltage hysteresis, which should be attributed to the reduced ion conductivity of hybrid electrolyte and depressed desolvation kinetics of Zn2+.

X−ray photoelectron spectroscopy (XPS) was carried out to further elucidate the cyclic reversibility of sulfur positive electrode. In the XPS spectra of S 2p (Fig. 5e), two peaks at the low binding−energy region between 164 eV and 166 eV represent the S−S bond of elementary sulfur. After cycled in the ZnSO4−ZnI2−H2O electrolyte, the intensity response of S−S bond significantly decreases, and two additional peaks can be observed at 169.5 eV and 170.8 eV, corresponding to the SO42− species formed on the surface of sulfur electrode, which should be resulted from the H2O−participated ZnS disproportionation55. In contrast, the sulfur positive electrode after cycled in the ZnSO4−ZnI2−H2O−DMF electrolyte demonstrates almost unchanged intensity response of S−S bond and inappreciable SO42− signal. Moreover, as shown in Fig. 5f, the XPS spectra of O 1 s for pristine sulfur positive electrode can be separated into two characteristic peaks located at 530.9 eV and 533.2 eV, which are ascribed to the C−O and C=O bonds on the surface of carbon materials. After cycled in the ZnSO4−ZnI2−H2O electrolyte, an additional peak at 532.4 eV corresponding to the SO42− species could be observed, whereas the sulfur positive electrode displays a significantly weakened SO42− signal after cycled in the ZnSO4−ZnI2−H2O−DMF electrolyte. In the XPS spectra of Zn 2p (Fig. 5g), the cycled sulfur positive electrode in the ZnSO4−ZnI2−H2O electrolyte exhibits a stronger intensity response than that in the ZnSO4−ZnI2−H2O−DMF electrolyte. All the above results further confirm that DMF addition could suppress the generation of zinc sulfate hydroxide by−products, and improve the cyclic reversibility of sulfur positive electrode due to the reduced H2O reactivity and restrained side reactions of ZnS disproportionation.

Electrochemical performance of Zn||S cells

The electrochemical performance of the assembled Zn||S cells with different electrolytes was investigated through utilizing an additional constant voltage charge procedure to avoid I−/I3− redox shuttle. As illustrated from the rate capabilities in Fig. 6a and Supplementary Fig. 44, the conventional ZnSO4−H2O electrolyte is incompetent for the normal operation of Zn||S cell due to the continuous capacity fading and rapid failure. Moreover, in comparison with ZnSO4−ZnI2−H2O electrolyte, the Zn||S cell in the ZnSO4−ZnI2−H2O−DMF electrolyte can exhibit higher specific capacity at low rate, but similar specific capacity at high rate due to the increased viscosity and diminished ion conductivity, which delivers the considerable reversible specific capacities of 1643, 1534, 1372, 1198, 1000 and 837 mAh g−1 at different rates from 0.1 C (i.e., 0.167 mA gs−1) to 3 C (i.e., 5.01 mA gs−1) after subtracting the I− contribution in the discharge potential range between 1.1 V and 1.2 V. While the Zn||S cell in ZnSO4−ZnI2−H2O−DMF electrolyte shows a greater capacity attenuation with increasing rate due to the larger concentration polarization, the significantly decreased voltage hysteresis can be observed even at high rates (Fig. 6b and Supplementary Fig. 45). This suggests that DMF cosolvent can markedly lower the activation polarization, which still dominates the sulfur conversion reaction process in Zn||S battery at high rates. Notably, the specific capacity and reaction polarization of aqueous Zn||S cells achieved in our work are highly competitive compared with the previously reported state−of−art aqueous Zn||S systems (Supplementary Fig. 46).

Fig. 6: Electrochemical performance of Zn||S cells with different electrolytes.
figure 6

a Rate performance at 25 °C and different specific currents. b Galvanostatic discharge−charge curves at 25 °C and different specific currents. c Cycle performance at 25 °C and 1 C (i. e., 1.67 A gs−1). d Long−term cycling stability at 25 °C and 2 C (i. e., 3.34 A gs−1). e Cycle performance in the ZnSO4−ZnI2−H2O−DMF electrolyte with high sulfur loading of 6.4 mgs cm−2 and E/S ratio of 10 μL mg−1 at 25 °C and 0.2 C (i. e., 0.334 A gs−1). f, g Galvanostatic discharge−charge curves at various temperatures. h Cycle performance at −10 °C.

Full size image

As presented in Fig. 6c, high cycling stability can be achieved in the ZnSO4−ZnI2−H2O−DMF electrolyte. After 150 cycles at 1 C (i.e., 1.67 A g−1), the specific discharge capacity can still be maintained at above 1080 mAh g−1, and the coulombic efficiency of each cycle is nearly 100%. In contrast, the Zn||S cell with DMF−free electrolyte shows an evident capacity fading, especially in the conventional ZnSO4−H2O electrolyte, suggesting that DMF cosolvent is beneficial to the improvement of cycling stability. Remarkably, as shown in Fig. 6d, the specific capacity of Zn||S cell with ZnSO4−ZnI2−H2O electrolyte is rapidly declined to almost no capacity release after the long−term cycles, while the Zn||S cell with ZnSO4−ZnI2−H2O−DMF electrolyte could still release a specific capacity of around 610 mAh g−1 with a capacity decay of 0.034% per cycle after 1000 cycles at 2 C (i.e., 3.34 A g−1), which is highly rivalrous in contrast with the recently reported state−of−art aqueous Zn||S batteries (Supplementary Table 3). As a demonstration (inset in Fig. 6d), three Zn||S swagelok cells assembled in series−wounded configuration successfully power the light−emitting diode (LED) lamps.

To meet the requirement of high−energy density for the practical application of Zn||S cells, the sulfur positive electrodes were explored in the ZnSO4−ZnI2−H2O−DMF electrolyte at high sulfur content and high sulfur loading. The S/KB composite material with high sulfur content of 80 wt % can still show a relatively low voltage polarization of 0.583 V and a reversible specific capacity of 1497 mAh g−1 (Supplementary Fig. 47). Moreover, as shown in Fig. 6e, when the sulfur loading is increased to 6.4 mgs cm−2 (the ratio of electrolyte volume to sulfur mass, E/S = 10 μL mg−1), the Zn||S cell exhibits a high discharge capacity of 1200 mAh g−1 with a coulombic efficiency of above 97% per cycle after 30 cycles at 0.2 C (i.e., 0.334 A g−1), achieving a high areal capacity of 7.68 mAh cm−2. Furthermore, the electrochemical performance of Zn||S cells in different electrolytes was preliminarily studied under a wide temperature range. As displayed in Fig. 6f, g and Supplementary Fig. 48, the voltage hysteresis is obviously lowered, whereas the decreased reversible capacity also can be perceived with the addition of DMF cosolvent under different temperature, which should be caused by the favorable interfacial ions transfer, but depressed ionic conductivity of electrolyte (Supplementary Fig. 49). Under the relatively low temperature of −10 °C (Fig. 6h), the Zn||S cell with ZnSO4−ZnI2−H2O−DMF electrolyte delivers a reversible capacity of 500 mAh g−1 after 100 cycles at 0.2 C (i.e., 0.334 A gs−1), while the cell is subjected to rapid failure in the ZnSO4−ZnI2−H2O electrolyte, meaning that DMF cosolvent also shows the great potential for cryogenic Zn||S systems.

In summary, we have developed a hybrid electrolyte containing the ZnI2 additive and organic N,N−dimethylformamide (DMF) cosolvent, where the iodide species is functionalized to promote the electrochemical oxidation of ZnS, while DMF cosolvent can accelerate sulfur reduction reaction. By combining operando Raman spectroscopy with in situ GEIS and theoretical calculations/simulations, the dynamic EDL structure and interfacial Zn2+ charge properties of sulfur electrodes are unveiled at the microscopic level, which demonstrates an interfacial Zn2+ charge accumulation−induced sulfur conversion reaction mechanism. That is to say, enough Zn2+ enrichment in the interfacial EDL region is essential to activate sulfur nanoparticles for the phase transition reaction into ZnS. In this regard, due to the surface preferential adsorption and strong interaction with Zn2+, DMF cosolvent contributes to reconstructing the interfacial EDL of sulfur electrode with H2O−poor IHP and Zn2+−rich OHP, which favors the interfacial Zn2+ transfer and facilitates the sulfur reduction reaction. Moreover, the strong H−bonding interaction between DMF and water molecules could decrease water reactivity and effectively inhibit the side reactions of ZnS disproportionation. With such the optimized electrolyte design, the Zn||S batteries could show a voltage hysteresis of 0.326 V at 0.1 C (i.e., 0.167 A g−1), and a long−term cycling stability with a capacity fading of 0.034% per cycle after 1000 cycles at 2 C (i.e., 3.34 A g−1). In addition, a high areal capacity of 7.68 mAh cm−2 at high sulfur loading (6.4 mg cm−2) and a stable low−temperature performance also can be achieved, which demonstrates the great potential for practical application. This work sheds light on the fundamental understanding of sulfur electrode/electrolyte interface, and provides rational scientific principles to modulate the interfacial EDL region for advanced aqueous zinc sulfur batteries with low polarization and long−term cycling life. To enhance the commercialization of aqueous zinc sulfur batteries, future improvements must concentrate on novel electrode materials and electrolytes, cost optimization, and safety strategies.

Methods

Reagents and materials

All chemicals were used directly without any purification treatment. Sublimed sulfur powder (S, ≥99.5%), zinc sulfate (ZnSO4, ≥99%), zinc iodide (ZnI2, ≥98%), N,N−dimethylformamide (DMF, ≥99.5%) were purchased from Alladin. Ketjen black (KB, ECP-600JD), polytetrafluoroethylene (PTFE) and glass fiber (Whatman GF/D, 1823-090, 2.7 Micron, 2.6 s/100 mL Flow Rate, 9.0 cm Diameter, 675 μm Thickness) were obtained from Guangdong Canrd New Energy Technology Co., Ltd. Zn foil (99.99%, 100 μm Thickness) and Titanium mesh (99.99%, 100 mesh, 0.1 mm Wire Diameter) were acquired from Non-ferrous Metal Material Co., Ltd.

Preparation of electrolytes

The preparation of ZnSO4−ZnI2−H2O−DMF electrolyte was carried out by adding DMF into water (volume ratio 3:7), and then dropping ZnSO4 and ZnI2 salts into the above solution with concentrations of 2 M and 0.2 M, respectively. Similarly, no DMF and only ZnI2 was added to obtain ZnSO4−ZnI2−H2O electrolyte, no ZnI2 and only DMF was added to get ZnSO4−H2O−DMF electrolyte, and no DMF and ZnI2 was added to get ZnSO4−H2O electrolyte. Meanwhile, the ZnSO4−ZnI2−H2O electrolytes with different concentration of DMF (10 vol %, 20 vol % and 30 vol %) were also prepared.

Synthesis of sulfur/Ketjen black (S/KB) composites

The S/KB composite materials were fabricated via a melt−diffusion method. Typically, the KB and sublimated sulfur powder were mixed and grinded evenly with a mass ratio of 1:1. Then the mixture was sealed in a reactor, followed by the heat treatment at 160 °C for 12 h. After cooled to room temperature, 50% S/KB composite material was obtained. In addition, the S/KB composites with sulfur contents of 66.7% and 80% were also obtained by controlling the mass ratio to be 2:1 and 4:1, respectively.

Cell assembly and electrochemical measurements

The 50% S/KB composite, Ketjen black and a polytetrafluoroethylene binder were mixed in isopropyl alcohol with a mass ratio of 7:2:1 to form a mixed slurry. This slurry was pressed on the Ti mesh current collector using a roll−press machine with a targeted thickness of 20 μm. After being dried at 60 °C for 12 h, the as−prepared electrode was cut into several disks (diameter of 10 mm, geometric surface area of 0.785 cm2), which were used as the sulfur positive electrodes. Zn||S cells were assembled in the Swagelok−type configuration, which consisted of the sulfur positive electrodes, zinc negative electrodes (diameter of 12 mm, thickness of 100 μm) and glass fiber separator (Whatman GF/D). ZnSO4−H2O, ZnSO4−ZnI2−H2O, and ZnSO4−ZnI2−H2O−DMF were applied as the electrolytes, respectively. The amount of electrolyte injected was kept at 50 μL. The areal sulfur loading of the positive electrode was controlled at appropriately 1.5−2.0 mg cm−2 (E/S ratio: 31.8−42.5 μL mg−1), while the high−sulfur−loading cycling test was conducted with the sulfur loading of 6.4 mg cm−2 (prepared by using 66.7% S/KB composite with a roll−press thickness of 60 μm) and the E/S ratio of 10 μL mg−1.

All the electrochemical tests were conducted in a climatic/environmental chamber at an average temperature of 25 °C unless otherwise stated. Galvanostatic charge/discharge tests for Zn||S cells were carried out in a voltage ranging from 0.3 V to 1.2 V vs. Zn2+/Zn with an additional charge procedure at the constant voltage of 1.2 V for 1 h (Land CT3002A battery testing system), which could guarantee the complete decomposition of ZnS without excessive electrochemical oxidation of I−. Cyclic voltammetry (CV) was characterized by the Ivium−n−Stat electrochemical workstation (Ivium, Netherlands), and the CV profiles were collected at different scan rates (0.1−0.5 mV s−1) from 0.3 V to 1.2 V vs. Zn2+/Zn. Alternating current (A.C.) voltammetry was performed in the negative−sweep potential regions between 0.5 V and 1.0 V vs. Zn2+/Zn where interfacial adsorption occurred in order to evaluate the interfacial capacitance. The amplitude of the applied sinusoidal voltage of the A.C. signal was 5 mV, with a frequency of 6 Hz. GITT were performed by a series of galvanostatic discharge and charge processes of 30 min at 0.1 C (i.e., 0.167 A g−1) followed by 120 min of rest at each step. Besides, Zn||Zn symmetrical cells were assembled using the CR2032−type coin cells with glass fiber separator and 200 μL of electrolyte, which were measured with galvanostatic methods to detect Zn2+ deposition/stripping properties.

Material characterizations

Zeta potential analysis was performed by Zetasizer (NanoPlus−3) to obtain the surface charge properties of the material samples. Thermogravimetric analysis (NETZSCH, STA449C) was conducted with a heating rate of 10 °C/min under Ar atmosphere. XRD patterns were recorded using a Bruker D8 Discover X−ray diffractometer using Cu Kα X−ray source at 3 kW with a goniometer radius of 250 mm. FT−IR measurements were carried out in the transmission mode of a Nicolet iS50 FTIR spectrometer (Thermo Fisher Scientific) boasting a MCT−A detector with a resolution ratio of 0.09 cm−1. Raman spectra were probed with an excitation wavelength of 532 nm on an alpha 300 R confocal Raman microscope (WITec, Germany). The specific surface areas were calculated using the Brunauer–Emmett–Teller (BET) equation based on the N2 adsorption–desorption measurement (ASAP2460). The pore−size distribution was analyzed by the Barret–Joyner–Halenda method. SEM and transmission electron microscopy (TEM) images were captured by Verios G4 (Thermo Fisher Scientific) and Tecnai G2 F20 (FEI, America), respectively. Surface−element analysis was conducted by means of an energy−dispersive X−ray fluorescence spectrometer. XPS measurements were implemented on ESCALAB QXi (Thermo Fisher Scientific) using monochromatic Al Kα (1486.8 eV) X−ray source at 24 W and 16 KV with a beam spot size of 100 μm. The contact angle between the electrolyte and sulfur electrode was measured through OCA 25 goniometer (Dataphysics, Germany). The ion conductivity of electrolyte was obtained from a portable DZB−718 L detector. 1H Nuclear magnetic resonance (NMR) spectra were acquired using an AVANCE NEO 600 MHz instrument. The NMR samples were prepared by using the pure D2O instead of H2O. For characterizing the cycled positive/negative electrodes, cells were disassembled in a climatic/environmental chamber at an average temperature of 25 °C, followed by thoroughly washing the electrodes with deionized water to remove the residual electrolyte salts. Then the dried electrodes were sealed in glass bottles and transferred quickly for XRD, SEM, and XPS measurements.

Double layer capacitance (C dl) and pseudocapacitance (C pseudo) of sulfur electrodes

CV profiles were collected with Zn||S cells at different scan rates (0.1, 0.2, 0.3, 0.4, 0.5, 0.7 and 1.0 mV s−1) in the non−faraday potential range of 0.92−0.98 V (vs. Zn2+/Zn). The capacitive current (Jc) was calculated by the half value of current difference during the anodic scan and cathodic scan at 0.95 V according to the equation of Jc = (Janodic – Jcathodic)/2. The Cdl values were determined through the linear relationship between capacitive current (Jc) and scan rate (ν), which can be obtained from the linear slope.

For the pulse voltammetry, it refers to the method from previous publications36,56. It was demonstrated that the applied bias did not directly drive the electrochemical reaction, but affected the electrocatalytically generated current and reaction kinetics through charge accumulation in the catalyst. Thus, the charge stored in the catalyst at a given potential can be quantified by integrating the current response over time. Specifically, the assembled Zn||S cell was initially set at the aimed reduction potential of 0.8 V (vs. Zn2+/Zn) for 30 s, and then abruptly switched to the open circuit potential (OCP, 1.0 V vs. Zn2+/Zn) for 10 s. The corresponding current density vs time profile was acquired, in which the current density at 29 s was recorded as J, and the cathodic transient current peak was integrated to determine the accumulated charge (Q). The J and Q values were measured point−by−point through repeating the above procedures except for varying the aimed reduction potential to 0.5−0.8 V vs. Zn2+/Zn. The total capacitance (Ctotal) can be obtained from the linear slope (∂Q/∂E) between accumulated charge (Q) and the corresponding potential (E), subsequently determining the pseudocapacitive contribution (Cpseudo = Ctotal−Cdl).

Operando Raman spectroscopy measurements

Zn||S cells were assembled with the operando Raman configuration, in which the sulfur electrode was placed close to the quartz window (see the details of architecture and assembly in Supplementary Fig. 14). The cell was cycled at the constant current density of 0.1 C (i.e., 0.167 A g−1). During the discharge and charge processes, Raman spectra were collected once every 600 s from the sulfur electrode surface. Each Raman spectrum was acquired over a collection time of 30 s with 10 accumulative scans.

In situ GEIS measurements of sulfur electrode during cycling

In situ GEIS measurements were carried out on an Autolab 302 N electrochemical workstation with a three−electrode Swagelok cell (Supplementary Fig. 17), which utilized sulfur electrode as working electrode, two pieces of zinc foils as counter electrode and reference electrode, respectively. A current density of 0.1 C (i.e., 0.167 A g−1) was used for the galvanostatic charge−discharge cycling of the cell. The EIS spectra were recorded at different potentials or different states of charge (SOC) with peak−to−peak amplitude of 10 mV, ranging the frequency from 1000 kHz to 0.01 Hz with 10 data points per decade of frequency. DRT analysis of the EIS data were conducted using the pyDRTtools software proposed by Ciucci et al.57,58, which is publicly available on GitHub: https://github.com/ciuccislab/pyDRTtools.

Density Functional Theory (DFT) calculations

The DFT calculations with periodic models were performed by Vienna Ab initio Simulation Package59,60. The generalized gradient approximation (GGA)61 and the Perdew−Burke−Ernzerhof (PBE) functional with projector augmented wave62 method were applied to better simulate the electronic structures and the core−electron interaction. The Gamma scheme was used to generate the k−point sampling grids within the Brillouin zone. The cutoff energy for all the computations was set 500 eV, which was accompanied with a total energy convergence set at 10−5 eV for self−consistent iterations. And the force threshold of optimization convergence was 0.05 eV Å−1. For the adsorption calculation, a vacuum layer of 15 Å was set on the slab structure for erasing the effect of periodic condition. The optimized computational models were provided in Supplementary Data 1–5.

To calculate the adsorption energies of the obtained molecules on the considered slab, the following formula was used:

$${E}_{{{{\rm{ads}}}}}={E}_{{{{\rm{slab}}}}+{{{\rm{molec}}}}}-{E}_{{{{\rm{slab}}}}}-{E}_{{{{\rm{molec}}}}}$$
(1)

where Eslab+molec means the total energy of the slab with species adsorbed, Eslab indicates the energy of the slab and Emolec represents the energy of the molecule in vacuum.

Charge density difference analysis was employed in this work to investigate the charge transfer during species adsorption. In this scenario the following formula was used:

$$\Delta \rho={\rho }_{{{{\rm{a}}}}+{{{\rm{b}}}}}-{\rho }_{{{{\rm{a}}}}}-{\rho }_{{{{\rm{b}}}}}$$
(2)

where ρa+b indicates the overall charge density of the system containing a and b, ρa or ρb represents the charge density of system a or b, respectively.

The ESP and HOMO energy levels were obtained by Dmol3 program Materials Studio. The DNP basis set was used to perform the structure optimizations. The exchange and correlation terms were determined using the GGA in the form proposed by PBE. Moreover, the Gamma scheme was used to generate the k−point sampling grids within the Brillouin zone, and the orbital cutoff sets as 12.5 Å.

Molecular dynamic (MD) simulations

The MD was conducted by the Forcite module in Materials Studio to explore the influence of DMF cosolvent on the Zn2+ solvation structure comprising 500 water molecules, 50 DMF molecules, 25 ZnSO4 units and 4 ZnI2 units. Moreover, the COMPASS II force field of Forcite module was used within the simulation system. First, the geometry optimization algorithm was used with a combination of steepest descent, adopting basis Newton−Raphson, and quasi−Newton methods based on the convergence of the total energy (0.0001 kcal mol−1) with a force of 0.05 kcal mol−1 Å−1. After geometric optimization, the NPT (250 picoseconds) was performed to optimize the density of the solution under a pressure of 0.1 MPa, a constant temperature of 298 K. The final configurations from the simulations were subject to additional equilibration simulations under NVT ensemble (T = 298 K) for a 0.1 nanoseconds. The RDF, coordination number, and the percentage of free H2O were obtained by the results of MD. The initial and final configurations were provided in Supplementary Data 6.