Zinc-pyrophosphate chelated solution was obtained by adding ZnCl2 solution dropwise to K4PPi solution with constant stirring (See Methods). To investigate the complexation ratio of the chelated solution, electrospray ionization-high resolution mass spectrometry (ESI-HRMS) was conducted. As depicted in Fig. 1a, an anion fragment peak of [K5ZnP4O14]- (m/z = 606.5764) was clearly observed. Due to the presence of isotopes, the anion fragment peak of [K5ZnP4O14]- were also observed at other positions such as m/z = 608.5736 and m/z = 610.5722. The fragment peaks of [K7ZnP4O14]+ (e.g., m/z = 684.5015) and [K4HZnP4O14]- (e.g., m/z = 568.6188) have also been detected according to mass spectral data in Supplementary Fig. 1a and Fig. 1b. In addition, Cyclic voltammetry (CV) curves (Supplementary Fig. 2) of 0.05 M Zn2+ mixed with different concentrations of PPi4- indicated that the slope of the equilibrium potential vs. ln[P2O74-] was close to 2, corresponding to a cation-anion coordination ratio of 1:220. Based on the aforementioned findings, the chelated ion in the solution is proven to be Zn(PPi)26-, and its coordination form is similar to other metal chelates comprising P2O74- such as Mn(P2O7)26- and Cu(P2O7)6-221,22. The structure of Zn(PPi)26- ion was further analyzed by 31P nuclear magnetic resonance (NMR) spectra. As illustrated in Fig. 1b, a non-concentration-dependent single peak is observed at −5.8 ppm for the Zn(PPi)26- solution. For reference, the 31P NMR spectrum of K4PPi solution with a symmetrical structure also shows a single peak at −6.2 ppm. This means that all P-atoms in Zn(PPi)26- are chemically equivalent. Besides, the chemical shift of Zn(PPi)26- solution exhibits a lower field shifting (toward higher ppm) compared with K4PPi solution, indicating the enhanced role of magnetic susceptibility23, which is related to the decrease in charge density near the P atom after the introduction of Zn2+ ion24. Meanwhile, Raman spectroscopy and attenuated total reflectance-fourier transform infrared (ATR-FTIR) spectroscopy were also conducted to uncover the structure of Zn(PPi)26- solution. As shown in Fig. 1c, the Raman peaks of Zn(PPi)26- solution detected at 1024 cm−1 and 1140 cm−1 belong to symmetric and antisymmetric stretching vibrations modes of PO3 group25,26, positively shifting compared with K4PPi parental solution (located at 1015 cm−1 and 1090 cm−1, respectively). Figure 1d presents the ATR-FTIR spectra of ZnCl2, K4PPi and Zn(PPi)26-, respectively. The vibrational peaks observed at 1012 cm−1 and 1128 cm−1 correspond to symmetric and antisymmetric stretching vibrations modes of PO3 group in Zn(PPi)26-, whereas the related vibrational peaks in K4PPi solution are located at 900 cm−1 and 1078 cm−126. Combining the above experimental results and subsequent theoretical simulation in Fig. 2f, we depicted the formation process of Zn(PPi)26- in Fig. 1e. It is also found that the stability constant of Zn(PPi)26- is 1 × 1011.027. Based on Le Chatelier-Braun’s principle, excess PPi4- ligands promote the formation of the Zn(PPi)26- complex and improve its stability. When the concentration ratio of PPi4- ligand to Zn2+ is 3:1, the solubility of the Zn(PPi)26- solution reaches 0.9 M. As the molar ratio of the PPi4- to the Zn2+ increases from 3:1 to 10:1, the conductivity of the resulting saturated solution gradually increases, whereas the solubility decreases from 0.9 M to 0.5 M (Supplementary Fig. 3). In addition, if ZnBr₂ is used to prepare the complex solution with a ligand-to-zinc ion ratio of 3:1, the solubility of the resulting complex solution will decrease to 0.7 M.Fig. 1: Structural information of Zn(PPi)26- ions.a ESI-HRMS spectrum of the Zn(PPi)26-, the peak found at m/z = 606.5764 is assigned to [K5ZnP4O14]- (calcd: 606.5715). b 31P NMR of 3 M K4PPi, 0.3 M and 0.8 M Zn(PPi)26-, respectively. c Raman spectra of 0.8 M ZnCl2, 3 M K4PPi, 0.3 M and 0.8 M Zn(PPi)26-, respectively. d ATR-FTIR spectra of 0.8 M ZnCl2, 3 M K4PPi and 0.8 M Zn(PPi)26-, respectively. e The chelated process of Zn(PPi)26- ions.Fig. 2: Theoretical calculation results for Zn(H2O)62+ and Zn(PPi)26-.3D snapshot of a 0.2 M ZnBr2 system and d 0.2 M ZnCl2-K4PPi (1:3) system obtained from MD simulations. RDFs for b ZnBr2 and e ZnCl2-K4PPi system collected from MD simulations. The optimized molecular structures and corresponding binding energy of c Zn(H2O)62+ and f Zn(PPi)26-. ESP-mapped molecular van der Waals surface of g Zn(H2O)62+ and h Zn(PPi)26-54. i The LUMO and HOMO isosurfaces of Zn(H2O)62+ (left) and Zn(PPi)26- (right), respectively.Molecular dynamics (MD) simulations were then carried out to analyze the solvation structure of the ZnBr2 and ZnCl2-K4PPi (1:3) systems, with the Zn2+ concentration set at 0.2 M for both systems. Herein, we opted to utilize ZnBr2 solution for comparative purposes, given its widespread application in zinc-based flow batteries. The results show that when the whole system became stable in a pure ZnBr2 environment, six H2O molecules appeared in the primary solvation shell (PSS) of Zn2+ (Fig. 2a), in accordance with previous literature28. On the contrary, when PPi4- was introduced into ZnCl2, the PSS of Zn2+ changed significantly (Fig. 2d). Analysis of the corresponding radial distribution functions (RDFs) and coordination numbers in different electrolytes show that for ZnBr2 solution, a main peak of the Zn-O pair appeared at a distance of about 1.99 Å, which is attributed to H2O in the PSS (Fig. 2b), and the coordination number of Zn2+ is near to 6. Meanwhile, for the ZnCl2-K4PPi system, a sharp peak appeared at a distance of 1.7 Å from Zn2+ with the O atoms in PPi4-, and the coordination number of Zn-O in the first PSS is close to 4 (Fig. 2e), assuring the chelation of two PPi4- with one Zn2+. Based on the results, density functional theory (DFT) calculations were further conducted to gain insights into the interaction behavior between Zn2+ ion and H2O or PPi4-, respectively. The optimized structure of Zn(H2O)62+ is shown in Fig. 2c, and the binding energy of Zn2+ with H2O is estimated to be -60.00 kcal mol-1. The optimized structure of Zn(PPi)26- is shown in Fig. 2f, and the simulated infrared spectrum of Zn(PPi)26- is relatively close to the experimental spectrum (Supplementary Fig. 4). It is noteworthy that the bond length of P = O in PPi4- is 1.55 Å, whereas that of P = O in Zn(PPi)26- is 1.53 Å (Supplementary Fig. 5), which accounts for the blue-shifting in the infrared absorption after coordination (Fig. 1d)29. The binding energy of Zn(PPi)26- is estimated to be -138.7 kcal mol-1, much higher than that of Zn(H2O)62+, indicating a stronger interaction of Zn2+ with PPi4-. Besides, Zn(PPi)26- shows a larger molecular size (6.65 Å) than that of Zn(H2O)62+ (5.34 Å). To understand the charge distribution and electron density of two species, electrostatic potential (ESP) mapped molecular van der Waals surfaces of them were also calculated. The results exhibit total different electric inherent, i.e., positive ESP for Zn(H2O)62+ ion (Fig. 2g), while negative ESP for Zn(PPi)26- ion (Fig. 2h). In general, the ion with positive ESP is electrophilic30, and Zn(H2O)62+ has been shown to be susceptible to by-products (e.g., Zn(OH)2, ZnO, Zn5(OH)8Cl2 · H2O, Zn4SO4(OH)6, etc.) during Zn deposition process31,32,33. In contrast, Zn(PPi)26- with negative ESP is nucleophilic, which would contribute to the suppression of the by-product (e.g., Zn(OH)2 and ZnO) formation during zinc deposition32. Frontier orbital analyses were conducted for two species to gain insight into the coordinator effect on the redox potential of Zn2+, since the redox potential of an active molecule shows a substantial correlation with its LUMO energy level (Fig. 2i). The results show that Zn(PPi)26- owns a higher LUMO energy (-0.17 eV) than Zn(H2O)62+ (-1.62 eV), which is attributed to the stronger ligand basicity of PPi4- ions than water molecules34.To clarify the redox process of Zn(PPi)26-, CV curves of 0.1 M K6Zn(PPi)2 electrolyte and 0.1 M ZnBr2 electrolyte were measured on a carbon paper electrode (1 cm−2) at 100 mV s−1 for comparison, as shown in Fig. 3a. It is found that the plating/stripping potential of Zn(PPi)26-/Zn is apparently negatively shifted to -1.08 V (vs. SHE), as supposed to -0.76 V (vs. SHE) for Zn2+/Zn. Note that the concentration of Zn(H2O)62+ is 1 × 10-10 M in 0.1 M Zn(PPi)26- electrolyte based on the stability constant of Zn(PPi)26-. Correspondingly,\({\varphi }_{{{{{\rm{Zn}}}({{\rm{PPi}}})}_{2}}^{6-}/{{\rm{Zn}}}}\) is calculated to be -1.05 V (vs. SHE) using the Nernst equation, very close to our experimental value. This result indicates that the plating process of Zn(PPi)26- electrolyte consumes the free Zn2+, and simultaneously the dissociation of Zn(PPi)26- releases the free Zn2+. The kinetic rate constants of the two electrolytes were further investigated using the steady-state polarization method (Fig. 3b and Supplementary Fig. 6). The reduced rate constant (k0) of Zn2+ is determined to be 1.1 × 10-4 cm s-1. In contrast, the reduced rate constant (k0) of Zn(PPi)26- is calculated to be 6.1 × 10-5 cm s-1, slightly lower than that of Zn2+. In addition, the reduction peak of Zn(PPi)26- electrolyte is clearly visible compared to that of Zn2+ electrolyte. To figure out the reason, CV curves at different sweep rates of Zn(PPi)26- electrolyte were investigated (Fig. 3c). The reduction peak current is linearly related to the square root of the sweep rates (Fig. 3d), and the diffusion coefficient (D) is calculated to be 3.03 × 10-6 cm2 s-1 according to the Randles-Sevick equation35, apparently lower than that of Zn2+ (2.44 × 10-5 cm2 s-1)36. It is considered that the diffusion of Zn(PPi)26- to the electrode surface can’t compensate for the Zn2+ consumption, thus the cathodic current reaches its maximum value to form a reduction peak. This phenomenon also occurs in other Zn2+-complex electrolytes, such as ZnBr42- and Zn(NH3)42+19,37.Fig. 3: Electrochemical properties of Zn(PPi)26- electrolyte.a CV curves of 0.1 M Zn(PPi)26- and 0.1 M ZnBr2 solution on a carbon paper electrode at 50 mV s−1, respectively. b Tafel plots for Zn plating/stripping in 0.2 M Zn(PPi)26- solution at 0.1 mV s−1. c CV curves of 0.1 M Zn(PPi)26- at various scan rates ranging from 10 to 50 mV s−1. d Linear relationship between reduction peak current densities (ipc) with square root of the scan rate (ν1/2) derived from c.We firstly demonstrated two parallel near neutral ZIFBs, one with 0.2 M ZnBr2 negolyte (pH=5.6), and the other with 0.2 M K6Zn(PPi)2 negolyte (pH=9.2), respectively. Both ZIFBs employed a low-cost polyolefin cation exchange membrane (JCM-D membrane) due to its lower area resistance (0.99 Ω cm2) in 1 M KCl solution compared to Nafion 212 membrane (1.18 Ω cm2), as shown in Supplementary Fig. 7. Figure 4a presents the galvanostatic charge/discharge (GCD) curves in the initial cycle for the two cells at a current density of 40 mA cm-2. It is clearly observed that the ZIFB with Zn(PPi)26- negolyte not only exhibits higher cell voltage of 0.3 V than that of the ZIFB with Zn2+ negolyte, but also shows a higher CE of 98% (79% for Zn2+ negolyte). The fluctuation in the GCD curves of Zn(PPi)26- based ZIFB may be due to the disruption of the coordination equilibrium of Zn(PPi)26-. Linear sweep voltammetry (LSV) of the two negolytes was then performed, and the result showed that the hydrogen evolution reaction (HER) in the 0.2 M ZnBr2 negolyte was more severe than that in the 0.2 M Zn(PPi)26- negolyte (Supplementary Fig. 8). The CV curves of cycled KI posolytes (Supplementary Fig. 9) indicates that Zn2+ ions heavily penetrate the JCM-D membrane, whereas Zn(PPi)26- ions barely cross through the membrane (the reduction peak around -1.2 V is originated from KI electrolyte, as shown in Supplementary Fig. 10). Further crossover tests of Zn2+ and Zn(PPi)26- ions through the JCM-D cation membrane were performed by H-type cells (Supplementary Fig. 11–14). The results show that the permeation rate of Zn2+ ions is as high as 3.2 × 10-3 cm2 h-1, but Zn(PPi)26- ions could not be detected in the reference cell, indicating that the chelated Zn(PPi)26- with multiple negative charges and larger molecular size could be isolated by the membrane. In addition, unlike Zn2+ negolyte, which usually suffers from the hydrolysis side reaction in aqueous solution, Zn(PPi)26- negolyte exhibits an excellent chemical stability over three months, as revealed by the ATR-FITR spectra and GCD tests (Supplementary Fig. 15).Fig. 4: Electrochemical performance of 0.2 M Zn(PPi)26- based ZIFBs.a GCD profiles of the ZIFBs at 40 mA cm−2 using 0.2 M Zn(PPi)26- negolyte or 0.2 M ZnBr2 negolyte in the first cycle. The charge process ended with a cutoff voltage of 1.9 V and 1.6 V, respectively, while the discharge process ended with a cutoff voltage of 0.2 V. b Rate performance of 0.2 M Zn(PPi)26- based ZIFB with a charging capacity of 20 mAh cm−2 at various current densities, the discharge process ended with a cutoff voltage of 0.2 V. c Cycling performance of 0.2 M Zn(PPi)26- based ZIFB at 40 mA cm-2. The charging capacity was controlled to 20 mAh cm-2, while the discharge process ended with a cutoff voltage of 0.2 V.The electrochemical performance of Zn(PPi)26- based ZIFBs were further investigated. Supplementary Fig. 16a shows the relationship between open-circuit voltage (OCV) and state of charge (SOC) using 0.2 M Zn(PPi)26- negolyte, and the OCV gradually increases from 1.56 V to 1.68 V as the SOC increases from 10% to 100%. The best cell performance is achieved at 80% SOC in terms of energy efficiency and electrolyte utilization (Supplementary Fig. 17b). Subsequent rate and cycling performance tests were controlled with a deposited Zn areal capacity of 20 mAh cm-2 (78% SOC). As shown in Fig. 4b, a slight increase in polarization potential difference is observed as the current density increases from 40 to 100 mA cm-2, implying excellent energy efficiency of the cell. When the cell was cycled at 40 mA cm-2, an average energy efficiency of 85% was achieved with negligible discharge capacity degradation over 100 cycles (Fig. 4c).To investigate the plating/stripping behavior of the Zn(PPi)26- negolyte at high areal capacities, the assembled ZIFBs were charged and discharged at different depths using 0.8 M Zn(PPi)26- negolyte at a current density of 80 mA cm-2. As the charge duration gradually increases from 0.5 to 2.5 h (Fig. 5a), the corresponding deposited Zn areal capacity increases from 40 to 180 mAh cm-2. The discharge duration is always close to the charge duration, implying that the Zn(PPi)26- negolyte still maintains a high CE at high capacity deposition. In addition, the rate performance of the cell using 0.8 M Zn(PPi)26- negolyte was investigated (Fig. 5b), revealing a decrease in average energy efficiency from 87% to 65% as the current densities increased from 40 to 200 mA cm-2. When the cell was cycled at high current density of 200 mA cm-2 for 250 cycles, the representative GCD curves at 10th, 120th and 250th cycles showed a stable discharge voltage plateau near 1.4 V with no significant degradation in discharge capacity (Fig. 5c). The charging voltage polarization decreases gradually after a few cycles due to the close contact of the residual zinc with the carbon felt. The overall cycling performance of the cell is presented in Fig. 5d, which displays an average CE of over 97% and average energy efficiency around 70%. After the cycling tests, 31P NMR and ATR-FITR spectra of the Zn(PPi)26- negolyte were performed (Supplementary Fig. 17 and 18) and no new peaks were identified in either of the spectra, indicating its excellent electrochemical stability. The polarization curves of the cell at 30%, 50% and 80% SOCs conform a linear trend (Fig. 5e), which means that the voltage drop is dominated by the Ohmic polarization rather than the kinetic polarization. Benefiting from the high cell voltage of the ZIFB, it exhibits a maximum output power of 606.5 mW cm-2 at 80% SOC, which is superior among reported ARFBs with cell voltages over 1.5 V8,38,39,40,41. When the deposited Zn areal capacity was increased to 60 mAh cm-2, the cell still exhibited excellent performance with an average CE of 97% and an average energy density of 85% at 80 mA cm-2 (Supplementary Fig. 19).Fig. 5: Electrochemical performance of 0.8 M Zn(PPi)26- based ZIFBs.a The GCD profiles of the ZIFB with different Zn areal capacities at 80 mA cm-2. The charging capacity was controlled from 40 to 180 mAh cm−2, while the discharge process ended with a cutoff voltage of 0.2 V. b Rate performance of the ZIFB at various current densities. The charging capacity was controlled to 200 mAh, while the discharge process ended with a cutoff voltage of 0.2 V. c Representative GCD curves of the ZIFB at 200 mA cm−2. The charging capacity was controlled to 40 mAh cm-2, while the discharge process ended with a cutoff voltage of 0.2 V. d Overall cycling performance of the ZIFB at 200 mA cm-2. e Polarization and power curves of the ZIFB at different SOCs. f Performance comparison of several ZRFBs in terms of areal capacity and operating current density.To visualize the comprehensive performance of the Zn(PPi)26- based ZIFBs, the deposited Zn areal capacity vs. current density is compared with other high-performance Zn-based flow batteries (ZRFBs)10,15,18,19,31,42,43,44,45. Figure 5f presents the areal capacity of several ZRFBs at different current densities. The observed high rate (i.e., 200 mA cm-2) is superior to that of most reported ZRFBs. This phenomenon suggests that dissociation of Zn(PPi)26- (or the complexation of Zn2+ and PPi4-) occurs at a very rapid rate during the Zn-plating (or Zn-stripping) process. As previously mentioned, the stability constant (1011) of Zn(PPi)26- indicates that the free Zn2+ concentration is very low, at 10-10 M, in the equilibrium state, signifying significant stability of the complex under equilibrium condition. However, it should be noted that such a high stability constant does not necessarily imply a low dissociation rate under non-equilibrium conditions. During the Zn-plating process, as free Zn2+ is consumed, dissociation occurs rapidly to maintain the free Zn2+ concentration at 10-10 M. Similarly, during the Zn-stripping process, as the free Zn2+ concentration increases, complexation occurs rapidly to ensure the concentration remains at 10-10 M. Furthermore, the similar behavior has been observed in chelated ZnBr42- or Zn(OH)42- based zinc-iron flow batteries19,46. In addition, to demonstrate the ability of the Zn(PPi)26- electrolyte to be plated at high rate. We also performed a static three-electrode test (Supplementary Fig. 20), and it was found that the initial plating current density exceeded 140 mA cm-2 when the polarization potential was controlled at -1.5 V vs. Ag/AgCl (i.e., -1.29 V vs. SHE), indicating a high dissociation kinetic of Zn(PPi)26-. The rapid decrease in plating current density with time suggests that the rate of diffusion of Zn(PPi)26- ions to the electrode surface determines the plating rate.To gain insight into the mechanism of zinc deposition for Zn(PPi)26- negolyte, the Zn deposits on carbon felt using Zn2+ negolyte or Zn(PPi)26- negolyte were collected by charging the corresponding ZIFBs to 100 mAh at 40 mA cm-2 (Supplementary Fig. 21). Both of the zinc deposits were investigated by powder X-ray diffraction (PXRD), respectively (Supplementary Fig. 22, 23). The results show that the deposited Zn grows mainly along the (101) crystal plane for both negolytes. Nevertheless, the laser confocal scanning microscope (LCSM) image revealed an irregular Zn deposition morphology after Zn2+ negolyte plating (Fig. 6a), and plenty of flower-like dendrites on the surface of carbon felt could be observed in the scanning electron microscopy (SEM) images (Supplementary Fig. 24a–c). However, for Zn(PPi)26- negolyte, the Zn deposits uniformly grew along carbon felt fibers (Fig. 6b), and no zinc dendrites were found in the SEM images (Supplementary Fig. 24 d-f). The element mapping with EDX analysis results also confirmed the even distribution of metallic Zn on carbon felt (Supplementary Fig. 25). To unveil the Zn growth mechanism of the two negolytes, we assembled symmetrical cells with uncharged filter paper as separator. The purpose of using uncharged filter paper is to exclude the Coulombic interaction between the membrane and the electrolyte, which could play a role in inhibiting Zn dendrite38. As shown in Supplementary Fig. 26, Zn(PPi)26- negolyte undergoes a higher nucleation overpotential (NOP) (220 mV) compared to Zn2+ negolyte (30 mV), which is expected for a much smaller critical Zn nucleus radius (r), based on the relationship of r and NOP47:$$r=2\frac{\gamma {V}_{m}}{F{\mbox{|}}\eta {\mbox{|}}}$$
(1)
Where γ is the surface energy of the Zn–electrolyte interface, Vm is the molar volume of Zn, F is Faraday’s constant, and η is the NOP.Fig. 6: Investigation of Zn deposition mechanism of two negolytes.a, b Laser confocal scanning morphology of Zn deposits obtained by charging 0.2 M Zn2+ negolyte and 0.2 M Zn(PPi)26- negolyte in an unsymmetrical ZIFB with a JCM-D CEM, respectively. c, d SEM morphology of Zn deposits obtained by charging 0.2 M Zn2+ negolyte and Zn(PPi)26- negolyte in a symmetrical ZFB with a filter paper separator, respectively. e PXRD patterns of carbon felts for Zn(PPi)26- negolyte with deposition capacities ranging from 40 to 180 mAh cm-2. f Binding energy of H2O molecule and PPi4- ion on the surface of Zn (101) crystalline plane. g, h The proposed Zn deposition process for Zn2+ negolyte and Zn(PPi)26- negolyte, respectively.Subsequent SEM images of Zn deposits on carbon felt showed that the Zn2+ negolyte tended to grow flower-like dendrites readily during the deposition process (Fig. 6c), whereas no dendrites were formed for Zn(PPi)26- negolyte plating (Fig. 6d). Therefore, it is reasonable to believe that the absence of dendrite upon the deposition of Zn(PPi)26- negolyte is primarily due to its high initial NOP on the carbon felt rather than the Coulombic repulsion between it and the anion groups on the JCM-D membrane. The PXRD patterns of carbon felts for Zn(PPi)26- negolyte with deposition capacities of 40-180 mAh cm-2 (Fig. 6e) reveal the strongest Zn (101) crystalline diffraction peaks for all carbon felts, regardless of the deposition capacity. No zinc dendrites were observed on the surface of the carbon fibers despite the diameter of the fiber increased as the areal capacity increased (Supplementary Fig. 27). The binding energy for H2O and PPi4- on the Zn(101) crystalline plane was also carried out to analyze the dendrite-free growth mechanism for the Zn(PPi)26- electrolyte. It is found that the binding energy between a PPi4- ion and the uncharged Zn(101) surface is -1.05 eV (Fig. 6f), which is significantly higher than that between a H2O molecule and the uncharged Zn(101) surface (-0.32 eV). In addition, the zeta potentials of zinc powders in H2O, ZnBr2 and K4PPi solutions were evaluated. As depicted in Supplementary Fig. 28, the zeta potentials of Zn are -0.45 mV (in pure H2O), 1.19 mV (in 0.8 M ZnBr2), and -6.20 mV (in 0.8 M K4PPi), respectively. The negative potential (-6.20 mV) strongly supports the calculation result that Zn metal prefers to adsorb PPi4-. As mentioned above, the plating process of Zn(PPi)26- electrolyte consumes the free Zn2+, and simultaneously the dissociation of Zn(PPi)26- releases the free Zn2+. Unlike the dendrite deposition mode of conventional Zn(H2O)62+ ions on the Zn surface (Fig. 6g) due to high interfacial water activity48, the interfacial water activity of Zn(PPi)26- ions is effectively reduced, allowing the subsequent dissociated Zn2+ ions to plate on the Zn surface in an orderly manner in assistance with the PPi4- ions (Fig. 6h). As a result, the Zn(PPi)26- electrolyte can alleviate the undesired HER and facilitate smooth Zn plating.
High-voltage and dendrite-free zinc-iodine flow battery
