Steric Hindrance-Driven Closed-Loop Conversion of Acceptor Enables Long-Life and High-Capacity Fluoride-Ion Batteries Authors: Meng Lei, Decheng Li, Keyi Chen, Zhe Han, Guyue Li, Xianhui Nie, Jingxian Zhang, Chilin Li Published: 2025 ABSTRACT Due to the ultrahigh theoretical energy density, no dendrite issue, and abundant resources, fluoride ion batteries (FIBs) have received a lot of attention. Regarding the issue of dissolving inorganic salt CsF in aprotic organic solvents, some anion acceptor (AA) strategies have been proposed. However, the strong binding force greatly constrains F– ion desolvation, resulting in short lifespan, low specific capacity, and poor reversibility of FIBs. Herein, we propose a concept of steric hindrance-driven closed-loop acceptor to address these problems, by using tetraphenylphosphonium chloride (Ph4PCl) with appropriate Lewis acidity to dissolve CsF and prepare a dynamic fluoride ion electrolyte based on the F–Cl exchange reaction, with a high ionic conductivity of 4.1 mS/cm at room temperature. The steric hindrance effects of chlorine and phenyl can accelerate the desolvation kinetics of F– ions. The excellent kinetics of the Ph4PCl-based electrolyte endows FIBs with long-term cycling stability, and the Sn@SnF2 symmetric cells can cycle for 500 h at 100 μA/cm2 and tolerate a critical current density as high as 1250 μA/cm2. Due to the potential dissociation ability of five-coordinated acceptor central phosphorus for fluorides and the closed-loop conversion effect of chlorine, the fluorination and defluorination reaction proceeds in a dissolution-deposition mode. The CuF2//Sn@SnF2 cell (under a high cathode loading of 4.2 mg/cm2) exhibits the highest reversible capacity up to 717.7 mAh/g and remains 316 mAh/g after 65 cycles with a small voltage polarization of only 11 mV. This work points out the novel design concept of AAs for developing high capacity and long lifespan FIBs. Introduction The rapid development of electric vehicles and low-altitude aircraft is driving the need for new chemical energy storage systems that can provide low-cost and sufficient power for these high-energy consumption applications. Considering the inherent high specific capacity advantage of multielectron transfer reaction, some conversion-type systems such as lithium–sulfur, lithium–metal fluoride, and fluoride ion batteries (FIBs) are regarded as the next generation high specific energy batteries. − In theory, a fluoride-ion battery (FIB) has ultrahigh energy density (3500–5000 Wh/L) that can surpass a lithium–metal battery, while further reducing the dependence on rare mineral resources (such as Li and Co) and lowering costs. Also, FIB also has the advantages of abundant F-containing resources, high oxidative stability (∼5.9 V vs Li/Li+), and low charge carrier mass. However, FIB suffers from several challenges that block their practical development, such as poor reversibility of metal fluoride cathodes and a lack of suitable electrolytes; even the design concept and guiding rule of FIB electrolytes are vague. In the early stages, FIB research has focused on solid-state battery configurations. Limited by the relatively low F-ion conductivity of solid-state electrolytes at room temperature (such as La0.9Ba0.1F2.9 et al. with a value of 10–7–10–6 S/cm), the batteries are usually required to operate under high temperature conditions (>140 °C). Meanwhile, due to the insufficient solid–solid contact between the electrode and electrolyte and the reversibility issue of the metal fluoride electrode conversion reaction, solid-state FIBs cannot achieve appealing electrochemical performance. In recent years, some superior solid-state electrolytes have been developed with the higher ionic conductivity of 10–4–10–3 S/cm at room temperature, such as PbSnF4, CsPb0.9K0.1F2.9, KSn2F5, et al., but the released capacities of corresponding FIBs are relatively low, attributed to the complex electrode–electrolyte interface problem. − Different from the poor and limited contact in solid-state systems, liquid fluoride ion electrolytes have inherently excellent interfacial wetting ability, an ideal ion diffusion environment at room temperature, and potential compatibility with large-scale manufacturing. These advantages attract researchers to focus on liquid electrolyte FIBs, which are expected to achieve superior electrochemical performance and great potential for developing high-specificity energy storage systems. Unfortunately, compared to the development of solid-state FIBs, the reports about liquid FIBs are relatively rare, attributed to the lack of electrolyte formulation and an unclear reaction mechanism. Thus, designing and preparing advanced liquid electrolytes with outstanding properties are the primary issues for FIBs. Given that large anion lithium salts (such as LiPF6, LiTFSI, and LiFSI) with excellent performance have long been a consensus choice for researchers and industry, and have been expanded to other battery systems (such as Na+ and Mg2+), so that fluoride salts with large-sized cationic groups were first used in fluoride ion electrolytes. This large-sized ion cluster with compensating charge can greatly weaken the ionic bonds between charge carrier ions (such as Li+, Na+, Mg2+, and F–) and cluster-type counterions, thus increasing their solubility in organic solvents. Some large-sized cationic fluoride salts were proposed, such as N,N,N-trimethyl-N-neopentylammonium fluoride (Np1F), tetramethylammonium fluoride (TMAF), and 1-methyl-1-propylpiperidinium fluoride (MPPF), and they can be quickly dissolved in organic solvents (with a solubility of 0.35–2 M at room temperature). The corresponding prepared electrolytes display superior ionic conductivity (2.351–9.53 mS/cm). − However, due to the β-H elimination reaction of these salts, their stability is poor, and the assembled FIBs show a rapidly decaying capacity, short working lifespan, and high cost. Ionic liquid (such as 1,3-dimethylimidazolium bis­(trifluoromethanesulfonyl)­imide [MMIm]­[TFSI]) has also been used to dissolve the fluoride salts mentioned above, and the electrochemical window has been improved (4.65 V). However, the capacity and reversibility are still poor, and the cost is also further increased. , Also, the synthesis processes of organic cationic fluoride salts are complicated and involve toxic hydrofluoric acid (HF), which is relatively dangerous; thus, structurally simple inorganic fluoride salts are another option (e.g., cesium fluoride, CsF). Due to the strongest electronegativity and nucleophilicity of F– ions, crystallized CsF is easily hydrolyzed and possibly leads to the formation of corrosive HF. On the other hand, it is hardly dissolved in aprotic solvents. To alleviate the HF formation, a salt-in-water strategy was proposed (37 M CsF solution), but this cannot fundamentally eliminate the potential HF formation. Therefore, how to improve the solubility of CsF in anhydrous aprotic solvents becomes a new challenge for FIBs. In order to address the CsF insoluble challenge in anhydrous aprotic solvents, anion acceptors (AAs) were used to weaken the CsF lattice energy, significantly enhancing the dissolution of CsF. The Lewis acid–base interaction between the AAs with a Lewis acidity and the F– ion with a Lewis basicity can effectively cleave the Cs–F bond and promote the dissolution of CsF, meanwhile achieving the F– ion solvation through the covalent bond formation between the acceptor center atom and F– ion. This strategy reduces the F– ion basicity through an AA-F– solvation complex and avoids its nucleophilic attack on solvent molecules with certain acidity. Some boron-based AAs such as fluorobis (2,4,6-trimethylphenyl)­borane (FBTMPhB), triphenylboroxin (TPhBX), and 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)­pyridine (DiOB-Py) have been successfully used to promote the dissolution of CsF in ether electrolytes, but the corresponding FIBs displayed rapid capacity decay and poor reversibility. This inferior effect should be attributed to the interaction between acceptor center B and F– ion, which is too strong to allow easy desolvation. − Compared with boron-based acceptors, group 15 elements-based AAs (e.g., N, P, As, and Sb) exhibit a relatively lower acidity, and their valence sphere is five-coordinated, therefore allowing for more flexible ligand design and tailoring of more suitable AAs for FIBs. In this work, a phosphorus-based AA, tetraphenylphosphonium chloride (Ph4PCl), is used to assist the dissolution of crystallized CsF in dimethyl sulfoxide (DMSO) and prepare a novel fluoride ion electrolyte DMSO-Ph4PCl-CsF (with 0.5 M Ph4PCl and 0.5 M CsF) (denoted as DPF0.5) through the halogen exchange reaction, with a measured F– ion concentration of 0.487 M. In this electrolyte, the F–Cl exchange occurring on the acceptor central atom P5+ is not only the key rate-determining step for the dissolution of CsF, but also is an important medium for the (de)­solvation of F– ions during cycling. The formation of Ph4PX2 – (X = F– or Cl–) in a form of six-coordinate valence-sphere corresponds to the change of P orbital hybridization, and it substitutes the unsolvated F– ions in the electrolyte, avoiding some potential side reactions such as Hofmann elimination and achieving the shielding effect on F– ions. The steric hindrance effects of chlorine and phenyl can effectively accelerate the desolvation kinetics of F– ions, and the Cl– ions enable a closed-loop conversion effect during the (de)­fluorination process. Also, the unique F–Cl exchange pathway for releasing F– enables the great reduction of desolvation barrier of F– ions on the PhPX2 – complex and the acceleration of electrode–electrolyte interface kinetics, therefore allowing the existence of higher valence state Cu (+2.5 and +3) and improving the specific capacity of FIBs. The five-coordinated acceptor central phosphorus is potentially capable of dissociating fluorides, and the (de)­fluorination reaction proceeds in a dissolution-deposition mode. The Sn/SnF2 symmetric cells could stably operate for at least 500 h with a small overpotential of 20 mV at 100 μA/cm2 and keep a low overpotential of 100 mV at 1250 μA/cm2. The CuF2–Sn/SnF2 battery configuration displays a reversible capacity as high as 717.7 mAh/g even under a high cathode loading of 4.2 mg/cm2, and the capacity is preserved at 316 mAh/g after 65 cycles. This unexpected specific capacity originates from the formation of high valence state Cu products, which is attributed to the fast ion migration and F– desolvation kinetics in the DMSO-Ph4PCl-CsF electrolyte. Results and Discussion Preparation and Characterization of the Electrolyte For the electrolyte preparation, DMSO is chosen as a polar aprotic solvent with a high boiling point, good thermal stability, and great miscibility, which endows most organic compounds with good solubility. Therefore, we added CsF and Ph4PCl acceptor powders alone or in various molar ratios into DMSO, which was continuously stirred for 12 h, and finally obtained a series of electrolytes. As shown in Figure A, there is an obvious solid–liquid phase separation in the DMSO-CsF system, with transparent supernatant and crystal precipitate at the bottom of the bottle. The conductivity of DMSO-CsF is only 0.17 mS/cm, indicating that DMSO hardly dissolves CsF. In the DMSO-Ph4PCl solution, a clear and transparent solution is observed, without any precipitation, suggesting good dissolution of Ph4PCl in DMSO. The corresponding conductivity of DMSO-Ph4PCl is as high as 6.25 mS/cm (Figure S1). For the DMSO-Ph4PCl-CsF system, the light-yellow supernatant and white precipitate at the bottom represent the occurrence of the F–Cl exchange reaction and the dissolution of CsF. The corresponding conductivity of DMSO-Ph4PCl-CsF is as high as 4.1 mS/cm when the concentration ratio of Ph4PCl-CsF is 0.5–0.5 M. In addition, when we change the concentration ratios of Ph4PCl-CsF to 0.5–1 M (denoted as DPF1.0) and 0.5–1.5 M (denoted as DPF1.5), the corresponding ionic conductivities decrease to 3.19 and 2.91 mS/cm (Figure B). This evolution may be related to the desolvation barrier of the F– ions. As shown in Figure S1, the ionic conductivity remains basically unchanged, and no precipitation is observed after 30 days, proving the long-term stability of DPF0.5, DPF1.0, and DPF1.5 electrolytes. We chose the optimized DPF0.5 electrolyte for subsequent discussion. The precipitate at the bottom for DPF0.5 is characterized by X-ray diffraction (XRD) after drying, as displayed in Figure S2. There are the diffraction peaks corresponding to CsCl (Pm̅3m) and no peaks of CsF (Fm̅3m), demonstrating the occurrence of F–Cl exchange and the dissolution of CsF. The nuclear magnetic resonance (NMR) spectra are used to characterize the chemical coordination environment of the DPF0.5 electrolyte and other control electrolytes. In the 19F NMR spectra (Figure C), no F signal is detected in the DMSO-CsF electrolyte even though its ionic conductivity is 3 orders of magnitude higher than that of pure DMSO solvent (0.6 μS/cm), which may be attributed to the nucleophilic attack of CsF on DMSO. Under the same scale, NMR spectra from DPF0.5 electrolyte exhibit the broad overlapping-peaks located at −188.2 and −171.6 ppm and a bulge located at −146.7 ppm, which respectively correspond to the different F– coordination environments, including Ph4PF (with a peak area ratio of 46.4%), Ph4PF2 – (49%), and Ph4PFCl– (4.6%). This obvious contrast between DMSO-CsF and DPF0.5 electrolytes suggests that Ph4PCl AA plays an outstanding role in assisting the dissolution of CsF. Dissolution of inorganic fluoride salt and demonstration of acceptor–anion interactions. (A) Dissolution comparison and optical photos of CsF in different solvents after stirring for 12 h. (B) Ionic conductivity comparison of different solvent systems and various CsF addition amounts. (C) 19F NMR spectra of DPF0.5 and DMSO-CsF. (D) ESP mapping on van der Waals surfaces of CsF, DMSO, and acceptors with different charge states, a color scale of blue-white-red reflecting local negative, neutral, and positive areas, respectively. (E) CsF dissociation energy barriers in the cases without (cyan line) and with (orange line) acceptor Ph4P+ based on DFT calculation. (F) Adsorption energy calculation and corresponding geometry optimization of the Ph4PX and Ph4P+ acceptor after capturing different anions. (G) 31P and (H) 1H NMR spectra comparison between DPF0.5 and DMSO-Ph4PCl. (I) FTIR and (J,K) Raman spectra of DPF0.5 and DMSO-Ph4PCl electrolytes at room temperature. The poor solubility of CsF in DMSO and the assisted dissolution mechanism by Ph4PCl AA are revealed through the molecule electrostatic potential (ESP) analysis based on the ESP value mapping on the van der Waals isosurface. Generally, the ESP of a molecule reflects the charge distribution in different regions on the molecule surface, which directly affects the electrostatic interaction between salts and solvents. The lowest negative ESP (ESPmin) or the highest positive ESP (ESPmax) ensures the nucleophilic or electrophilic ability of the solvation solvent, respectively, allowing the dissociation prediction of lithium salt. As shown in Figure D, the blue and red regions stand for the negative charge and positive charge areas, respectively. There are the well-defined positive and negative charge areas for CsF, and its ESPmin and ESPmax values are, respectively, −58.65 and 47.5 kcal/mol. This ESP range exceeds that of the DMSO molecule (from −46.6 to 29.31 kcal/mol). It can be seen that the active site on the DMSO molecule can hardly split the Cs–F ionic bond, thus CsF is almost insoluble in DMSO solvent. The ESP of Ph4PCl has a range of −70.23–29.31 kcal/mol (Figure S3A), and its ESPmax is close to that of DMSO. Ph4PCl is a nonpolar molecule, and the phenyls are prone to bonding to the methyl of DMSO by van der Waals interaction, allowing DMSO to penetrate the molecular interspace of Ph4PCl and dissolve it. The appearance of the Ph4P+ cation reflects the dissociation of Cl– ions from Ph4PCl. Note that the electrostatic interaction between Ph4P+ and SO groups is relatively weak; thus, DMSO can only partially dissociate Cl– ions, as discussed later from the NMR result of the DMSO-Ph4PCl electrolyte without CsF (Figure G). Although the ESP value of Ph4PCl AA is also lower than that of CsF, the Ph4P+ cation displays an ESPmax value of 95.69, far exceeding that of CsF. Therefore, this AA is expected to cleave the Cs–F bond and trap F– into a coordination state after its dissolution in DMSO, thus promoting the dissolution of CsF in DMSO. Meanwhile, the dissociated free Cs+ can bond with Cl– to form the CsCl precipitate, and it is a F–Cl exchange reaction process. Furthermore, the five-coordinated central P (e.g., in Ph4PF and Ph4PCl) can still coordinate extra F–/Cl– to form a six-coordinated sphere with negative charge, namely, one P can match up to two F–/Cl– ions. In the DPF0.5 electrolyte, there is a coexistence of five-coordination and six-coordination in the central P chemical environment, which provides the vacancies and desolvation space for the migration of F– ions. Combined with the formation of CsCl precipitate, a clear chemical landscape about F–Cl exchange is formed in DMSO solvent, Ph4PCl + CsF → Ph4PF + CsCl↓, and the reaction is driven by the bonding of AA central P and F as well as the precipitation of CsCl. The CsF dissociation under different conditions is plotted in Figure E. There is a high dissociation barrier of 5.83 eV when it dissociates directly into Cs+ and F– ions. With the introduction of Ph4P+ cation, Ph4PF and Cs+ ion are formed, and the corresponding barrier significantly decreases to 1.99 eV, consistent with the ESP prediction that Ph4P+ cation can dissociate CsF. Furthermore, we note that in this case the energy barrier is a positive value (>0), indicating that this dissociation with Ph4P+ cation is not spontaneous. However, when Cl– and Cs+ combine to form a CsCl precipitate, the corresponding reaction barrier decreases to −0.45 eV (<0), reflecting that the reaction proceeds spontaneously (Figure S4). The ESPmin value of six-coordinated cluster Ph4PF2 – is −6.52 kcal/mol, at the midpoint of two F atoms on this AA solvate, preventing the nucleophilic attack of F– on other molecular hydrogen in the DPF0.5 electrolyte. The adsorption energies of Ph4P+ cations for various F– and Cl– are summarized in Figure F. The electrically neutral Ph4PCl and Ph4PF possess high energies of 3.00 and 4.43 eV, respectively, and they are relatively stable. In other negatively charged clusters, the energies of Ph4PF2 –, Ph4PFCl–, and Ph4PCl2 – are 1.59, 1.44, and 0.11 eV, respectively. It indicates that as the amount of Cl– increases, the energy gradually reduces and the stability decreases, which may be attributed to the significant steric hindrance of large-sized Cl in the six-coordination sphere. Further, the 31P and 1H NMR spectra were conducted to analyze the chemical environment of the electrolyte. As a comparison, the 31P NMR spectrum of DMSO-Ph4PCl electrolyte shows a triplet peak located at 22.09, 22.31, and 22.53 ppm, corresponding to Ph4P+, Ph4PCl, and Ph4PCl2 – respectively. Also, we note that the two peaks of Ph4P+ and Ph4PCl2 – are reflectionally symmetrical in terms of position, intensity, and area, which is a manifestation of electroneutrality and illustrates the excellent electron-accepting ability of the AA central P atom. These same peaks are also observed in the 31P NMR spectrum of the DPF0.5 electrolyte (Figure G), but the difference is that their intensities are lower than those of the DMSO-Ph4PCl electrolyte, reflecting the low Cl– concentration in the DPF0.5 electrolyte. Also, in the downfield region, there is a new triplet peak located at 25.74, 25.5, and 25.23 ppm, respectively, corresponding to Ph4PFCl–, Ph4PF2 –, and Ph4PF. The strongest peak located at 25.5 ppm represents the high content Ph4PF2 – cluster that is the most stable form with a six-coordinate, distorted octahedral coordination sphere around the central P atom. In Ph4PF2 –, limited by the steric hindrance of large-sized phenyl, one F and phenyl occupy the axial position, and the others occupy the equatorial position (Figure S5). In addition, a weak peak appears at the high-field region, located at 18.38 ppm, and it may be two five-coordinated central P atoms co-coordinated with a F– ion (e.g., Ph4PF···F–···Ph4PF). In the 1H NMR spectra of DPF0.5 and DMSO-Ph4PCl electrolytes (Figure H), there is the same strongest peak located at 2.55 ppm, and it is derived from DMSO-d 6. The other weak peak appearing only in DPF0.5 electrolyte represents the hydrogen bond interaction between the F of Ph4PF2 – and the alkyl hydrogens of DMSO. Moreover, Fourier transform infrared spectra (FTIR) and Raman spectra are used to characterize the bonding properties of electrolytes. As displayed at Figure I, the DPF0.5 and DMSO-Ph4PCl electrolytes both possess the same absorbance peaks, corresponding to the –CH3 stretching vibration (2993.5 and 2910.5 cm–1) and SO band (895–1112 cm–1) from DMSO molecules and the C–C/CC (1310.9 and 1405.8 cm–1), C–H in phenyl (553–676 cm–1) and P–C vibration (1435.3 cm–1) from Ph4P+. , The P–F peak (1193.2 cm–1) exists only in DPF0.5. For Raman spectra (Figure J,K), there are many of the same signals from DMSO solvent, including the –CH3 asymmetric (2914.1 cm–1) and symmetric (2997.1 cm–1) stretching, SO vibration peaks at 1047.4 and 1061.0 cm–1 (respectively corresponding to DMSO cluster and free DMSO, reflecting the molecular aggregate feature of DMSO), and another peak at 1030.0 cm–1 assigned to the Ph4P+-DMSO coordination structure. , Also, the P-Ph vibrations of both electrolytes are in the same position at 1099.4 cm–1. The P–C stretching (999.4 cm–1) and Ph4P+ vibration (3060.7 cm–1) in DPF0.5 are red-shifted by 2 cm–1 relative to those of DMSO-Ph4PCl (respectively at 1001.4 and 3062.7 cm–1), but the Ph4P+ aromatic in-plane stretching vibration (1589.3 cm–1) is blue-shifted by 2 cm–1 relative to that of DMSO-Ph4PCl (1587.3 cm–1). Due to the introduction of F– ions with a stronger electronegativity in the DPF0.5 electrolyte, the Ph4PFCl–and Ph4PF2 – configurations are produced, the phosphorus-halogen bond strengths are enhanced, and the corresponding P–C bond stretching degree increases, resulting in a red shift of P–C and P-Ph peaks. Correspondingly, the interaction exerted by the acceptor central P on phenyls is weakened, and the Ph4P+ aromatic in-plane stretching vibration is slowed, leading to a blue shift of this signal. DFT Calculations and Interaction Region Indicator Analysis of the Acceptor In order to get insight into the interaction between central P and coordinated groups (such as phenyl, F, and Cl) and further understand the relationship between bonding and electronic structure, projected crystal orbital Hamilton population (pCOHP) analysis (pCOHP) was carried out by using the LOBSTER package. In a typical −COHP plot (Figure A), the hybridized energy levels are split into two parts, the antibonding states on the left (e.g., the negative part) are near the Fermi level, and the bonding states on the right (e.g., the positive part) are far away from the Fermi level. This result is consistent with the projected density of states (PDOS) of Ph4PCl (Figure S3B), where the overlapping PDOS between P and Cl indicates the bonding states, but only the PDOS peak from Cl appears near the Fermi level, indicating the antibonding state below the Fermi level. Generally, the integrated COHP value (ICOHP) is calculated by integrating the energies of the highest occupied bands below the Fermi level to quantitatively analyze the bonding strength. Herein, we use the –ICOHP value to describe the bond strength. The low occupancy of bonding states in Ph4PCl causes a weak P–Cl bond (0.9 eV), suggesting that Ph4PCl is easy to dissociate into Ph4P+ and Cl–. Different from the weak P–Cl bond, the P–F bonds exhibit the higher –ICOHP values of 3.91, 4.25, and 5.89 eV in Ph4PF, Ph4PFCl– and Ph4PF2 –, respectively, and these P–F bond strengths are significantly stronger than that of P–Cl bond, demonstrating the F–Cl exchange reaction in DPF0.5 electrolyte. From these pCOHP plots of P–F bonds (Figure B–D), the five-coordinated Ph4PF has the relatively weaker P–F bond, and the six-coordinated Ph4PF2 – exhibits the strongest P–F bond. The antibonding states in Ph4PF and Ph4PF2 – are close to the Fermi level, but the antibonding state in Ph4PFCl– is positioned far from the Fermi level, likely attributed to the preferential contribution of the P–Cl bond to the antibonding state. Atomic-level and quantitative insight into the interaction between phosphorus-centered acceptor and halogen anions. (A–D) Phosphorus–halogen bond crystal orbital Hamilton population analysis (COHP) between acceptor and halogen atoms under different charge states, Ph4PCl, Ph4PF, Ph4PFCl–, and Ph4PF2 –. (E) P–F and (F) P–Cl bond strength and length relationship evaluated through the integrated COHP (–ICOHP) analysis. (G) P–C bond and (H) C–H bond strength and length summary of optimization configurations of Ph4P+ acceptor and F–/Cl– ions. The circles represent bond lengths, the squares represent –ICOHP values with the area size reflecting the value. The subscripts (B) and (M) represent bulk material and single molecule system, respectively, and the shaded area also represents the single molecule system. The negative correlation between bond length and bond strength (e.g., –ICOHP value) is summarized in Figure E,F. When the bond length is shorter, the –ICOHP value is larger, and the bonding is stronger through a comprehensive comparative analysis on the bonds between different AA central P atoms and various F–/Cl– ions. The P–F/P–Cl bond lengths and –ICOHP values of electrically neutral Ph4PCl (2.74 Å, 0.9 eV) and Ph4PF (1.805 Å, 3.91 eV) are close to those of the corresponding bulk materials Ph4PCl (2.78 Å, 0.87 eV) and Ph4PF (1.803 Å, 4.47 eV). For the negatively charged configurations, the P–Cl bond length of Ph4PCl2 – is 2.97 Å, with a corresponding –ICOHP value of only 0.02 eV, and the P–Cl bond length of Ph4PFCl– is as high as 3.8 Å with a negligible –ICOHP value (0.01 eV), attributed to the large size of Cl– and the steric hindrance of phenyl. Interestingly, in view of the small size of the F– ion, the P–F bonding is abnormally enhanced in the Ph4PFCl– (1.77 Å, 4.25 eV) and Ph4PF2 – (1.67–1.7 Å, 5.47–5.89 eV) configurations, forming a covalent bond, which may be related to the orbital hybridization change for AA central P. Note that in the Ph4PF2 – configuration, two P–F bond strengths are different, indicating that the P–F distribution is not even, which is further demonstrated by charge density difference (CDD) and Bader charge analysis (Figure S6). The central P atom loses electrons (−2.350 e), and a large amount of electrons accumulate around the F atom at the axis position, while the other F atom occupying the equatorial position has only a small amount of electrons, even though their Bader charges are close (0.777–0.786 e), indicating that the latter F is more easily desolvated. Similarly, in Ph4PFCl–, a lot of electrons accumulate around the F atom (0.784 e), and Cl (0.341 e) is far away from the central P (−2.212 e), exhibiting an isolated charge distribution and weak P–Cl bond strength. Overall, compared with the weak P–Cl bonds, the strong P–F bonds can constrain the nucleophilic attack of F– ions on the potential active sites in the DPF0.5 electrolyte. In addition, the bond strengths (i.e., –ICOHP values) of P–C (4.1–6.6 eV), C–H (5.6–5.9 eV), and C–C (8.6–10.1 eV) bonds (Figure S7) in molecules are slightly weaker than those of Ph4PCl and Ph4PF bulk materials (Figures G,H and S8). Note that the P–C bonds of Ph4PF2 – (4.1–4.87 eV) are significantly reduced, which is consistent with abnormally enhanced P–F bonds. These robust P–C, C–H, and C–C bonds suggest the thermodynamic stability of AA with various coordinated environments, which ensures the DPF0.5 electrolyte has a wide electrochemical window and excellent electrode compatibility. The change in chemical coordination environment directly affects the orbital hybridization for the AA central P, and the F– ion migration kinetics depends on the transformation mechanism between P–F and P–Cl bondings. In the simple PX n (X = F, Cl, n = 4, 5, and 6) molecule systems (Figure S9), almost without steric hindrance from the ligand, the P–F bond length range is 1.47–1.63 Å, and the P–Cl bond length range is 1.94–2.15 Å. These bond lengths are much shorter than those of the AA-F clusters. There are the tetrahedron (C 3v ), trigonal bipyramid (D 3h ), and octahedron (O h ) geometry configurations in PX n , and the corresponding central P orbital hybridization modes are sp3, sp3d, and sp3d2, respectively. In fact, the central P coordination polyhedrons (such as Ph4PFCl–, Ph4PF2 –, Ph4PCl2 –, and Ph4PCl) have been distorted when four halogens are substituted by phenyls, resulting in the disappearance of inversion symmetry, which is caused by the steric hindrance between the large-sized phenyls. For the Ph4P+ cation with four ligands, there is a sp3 orbital hybridization for P (Figure S10), and the same hybridization also exists in Ph4PCl2 –. Because the large-sized Cl– ions are blocked by phenyls, it is difficult for Cl– ions to bond with the central P, and they can only be coordinated through electrostatic interaction (Figure A). The same steric hindrance effect is also observed in Ph4PFCl–, the central P has five-coordination and sp3d hybridization, and the lone-pair electron of the F– ion fills the empty P 3d orbital (Figure B). Similarly, in Ph4PF2 –, two long-pair electrons from F– ions both fill the empty P 3d orbital, forming a sp3d2 hybridization and P–F covalent bond (Figure C), leading to the abnormal enhancement of P–F bond strength. The varied hybridization and distorted coordination polyhedrons will lead to changes in charge distribution of the central P5+ and the coordinated halogen, resulting in different bonding modes and bonding strengths, which are consistent with the CDD analysis results (Figure S6). The influence of steric hindrance of phenyl ligand on the phosphorus–halogen bond is striking, and we performed COHP analysis on the central phosphorus bonding with various number of phenyls. The corresponding bond strength evolutions of P–F and P–Cl bonding are summarized in Figure S11. In the PX n (X = F, Cl, n = 4, 5, and 6) system, the number of phenyls is zero, there are very robust covalent bonds, and the –ICOHP values of P–F and P–Cl are in the ranges of 6.87–8.6 eV and 4.79–6.15 eV, respectively. When the number of phenyls increases to three, the –ICOHP values of P–F and P–Cl are in the ranges of 5.62–6.19 eV and 3.46–3.65 eV, respectively, suggesting these bonds are slightly affected by the steric hindrance from phenyls. Furthermore, when the number of phenyls increases to four, e.g., with the formation of Ph4PCl system as used, there is a significant steric hindrance effect. The –ICOHP values of P–F bonds range from 3.9 to 5.89 eV, the –ICOHP value of P–Cl bond decreases to only 0.9 eV for electrically neutral Ph4PCl, and the –ICOHP values of P–Cl bonds are only 0.01–0.02 eV, especially in the cases of a six-coordinated environment. Orbital hybridization changes of AA central P and the correlation between steric hindrance of different groups and F–desolvation kinetics. Orbital hybridization modes of different central P: (A) four-coordinated P displaying sp3 hybridization (Ph4PCl2 –), (B) five-coordinated P displaying sp3d hybridization (Ph4PFCl–), and (C) six-coordinated P displaying sp3d2 hybridization (Ph4PF2 –). The scatter maps between IRI and Sign­(λ2)­ρ for (D) Ph4PCl2 – and (F) Ph4PF2 – and the coloring maps of Sign­(λ2)­ρ on IRI = 1 isosurface for (E) Ph4PCl2 – and G Ph4PF2 –. The bottom color scale is the standard coloring method and chemical explanation of Sign­(λ2)­ρ on IRI isosurfaces. (H) Desolvation barriers of F– from the six-coordinated P complex with negative charge in different desolvation pathways. To intuitively understand the effect of steric hindrance, a real space function, namely, the interaction region indicator (IRI), is used to clearly reveal the chemical bonds and interactions between any atoms. , In the scatter maps, the nature of interactions can be easily identified; the blue, green, and red regions stand for the strong attraction, van der Waals interaction, and strong repulsion. Comparing these scatter maps (Figure D,F), Ph4PCl2 – shows an inverted parabola blue region, which corresponds to the binding force of Cl nuclei on electrons on the IRI = 1.3 isosurface (Figure S12). In the IRI = 1.0 isosurfaces, there are strong attractions on the C–C, C–H, and P–C bonds (blue regions), strong repulsions focused on the center of phenyl (red cone regions), and van der Waals interactions appearing between phenyls (green regions). In the IRI plot of Ph4PCl2 – (Figure E), two Cl– ions are too far away from the central P to form P–Cl bonds, Cl– and Ph4P+ are interconnected by electrostatic force, and Cl– and nearby H atoms produce weak attraction. In addition to the similar strong attractions mentioned above, in Ph4PF2 – (Figure G), new strong attractions appear on the P–F bond regions, indicating the formation of P–F covalent bonds. There is a red repulsive force between the two F– ions. Similarly, the P–F covalent bond and van der Waals interaction near Cl– both appear at the Ph4PFCl– molecule (Figure S13). These stable P–F covalent bonds can effectively enhance the stability of the electrolyte by constraining F– nucleophilic attack on other molecules, which provides a potentially wide electrochemical window. While electrolyte stability needs to be emphasized, the desolvation of F– from the stable Ph4PF2 – structure is a more important aspect of the DPF0.5 electrolyte. There are two possible F– desolvation pathways in the DFP0.5 electrolyte, as displayed in Figure H and Table S1. The direct defluorination process (Path 1) requires the cleavage of the P–F covalent bond, first undergoing Ph4PF2 – → Ph4PF (with a barrier of 1.21 eV), and then the energy barrier from Ph4PF to Ph4P+ becomes as high as 3.84 eV. For the most stable neutral geometric configuration Ph4PF (Figure F), the F– ion desolvation barrier from six-coordinated Ph4PF2 – to five-coordinated Ph4PF is low and 1.21 eV, but the secondary desolvation is limited by electrostatic forces, with a barrier as high as 3.84 eV. When the Ph4PF is exchanged to Ph4PCl (i.e., Path 1–2, based on F–Cl exchange), the secondary desolvation barrier decreases from 3.84 to 1.39 eV. For the F–Cl exchange defluorination pathway (Path 2), the energy barriers of Ph4PF2 – → Ph4PFCl– and Ph4PFCl– → Ph4PCl2 – are 1.55 and 1.21 eV, respectively, and during this desolvation process, the charges of six-coordinated structures remain constant. When the Ph4PFCl– directly defluorinates to form Ph4PCl (i.e., Path 2–1, based on direct defluorination), the secondary desolvation barrier is 1.05 eV. From the perspective of nucleophilicity, although F– ions are covalently anchored on the Ph4P+ acceptor, the P–F bonds can still be easily cleaved, even by a weak nucleophile. In the DPF0.5 electrolyte, the nucleophilicity of Cl– can promote the cleavage of the P–F bond and accelerate the desolvation kinetics of F– ions. The F–Cl exchange not only plays an important role in the dissolution of CsF during electrolyte preparation but also has a promoting effect in reducing the energy barrier during the F– desolvation process. More importantly, from the perspective of the entire battery, the five-coordinated Ph4PF produced by the direct defluorination pathway is electrically neutral and cannot continue to migrate in the direction of the electric field, thereby interrupting the transport of F– ions. In contrast, the F–Cl exchange process ensures some constant six-coordinated anion clusters (such as Ph4PF2 –, Ph4PFCl–, and Ph4PCl2 –) that can migrate in the direction of the electric field and become the F– ion transport carriers, therefore improving battery kinetics. Based on the above discussion, we clearly demonstrate that the large-sized Cl– plays a significant role in reducing the desolvation barrier, especially in the secondary desolvation stage. Combined with Figure B, the CsF content in electrolyte increases from 0.5 to 1.5 M, but the corresponding conductivity decreases from 4.1 to 2.91 mS/cm, also suggesting that Cl– helps F– transport and F–Cl exchange is the main defluorination pathway. In DPF1.5 electrolyte, the excessive CsF addition is prone to causing the Ph4PF2 – to dominate six-coordinated acceptor P and a loss of five-coordinated acceptor P (Ph4PF and Ph4PCl). The different electrolyte environments lead to the evolution of F– ion desovlation pathways in DPF0.5 and DPF1.5, from F–Cl exchange to direct defluorination. The macroscopic ionic conductivity results are consistent with the theoretically different thermodynamic energy barriers generated by different F– ion desolvation pathways. The steric hindrance of Cl– causes strain on the phenyl ligand that can be released in the electrolyte environment, resulting in the increase of phosphorus-halogen distance and promotion of F– desolvation. For example, the P–F bond length in Ph4PF is 1.82 Å under vacuum (Figure S14), and therein, there are obvious red steric hindrance areas around F–. The P–F distance becomes 4.36 Å in the implicit solvent DMSO, and the surrounding red steric hindrance areas disappear, instead of the only existence of van der Waals areas (green areas). Similar bond length changes also occur in Ph4PCl, and the P–Cl distance in DMSO is as high as 5.0 Å. The other P–F and P–Cl bond lengths are summarized in Table S2. The steric hindrance effect of phenyl is demonstrated through the direct comparison of IRI maps between four- and three-phenyl systems (Figure S15). There are six steric hindrance areas (e.g., brown-green regions within black dotted lines) in the Ph4P+ cation, but the Ph3P2+ cation only displays three steric hindrance areas, and their color is lighter. In its corresponding neutral molecule triphenylphosphonium dichloride (Ph3PCl2), there are two P–Cl covalent bonds, and the bond distance (2.33 Å) is close to those of PCl5 and PCl6 without phenyls. In addition, in the six-coordinated environments, including Ph3PCl3 –, Ph3PF2Cl–, Ph3PF3 –, and Ph3PFCl2 – (Figure S16), the steric hindrance effect is enhanced with the corresponding area increase and color deepening. Therein, the P–Cl bond lengths range from 2.25 to 2.4 Å, and the P–F bonds are always constant at 1.66 Å, close to that of PF6 without phenyls (1.63 Å). This difference of only one phenyl on the AA central P atom leads to a significant change in steric hindrance, which enables a huge effect on the phosphorus-halogen bonds. In the prepared electrolyte DMSO-Ph3PCl2-CsF (0.5–0.5 M), there is a large amount of CsCl precipitation, and the corresponding room temperature conductivity is 2.5 mS/cm (Figure S17). Note that the reported electrolyte prepared by triphenylantimony dichloride (Ph3SbCl2) has a lower conductivity of 0.74 mS/cm, likely owing to the formation of stronger Sb–F covalent bonds and the robust Lewis acidity of Sb5+. This comparison indicates that, in addition to the role of steric hindrance from ligands, the acidity of the AA central element is also an important factor. Herein, the design guidelines for AAs centered on phosphorus group elements are proposed. First, the central element of AA has sufficient acidity to form a covalent bond with the nucleophilic F–, forming a stable solvation structure to guarantee the chemical stability of the electrolyte. Second, the appropriate number of stable ligands contributes to the steric hindrance effect and the promotion of F– desolvation, endowing the electrolyte with fast F– transport ability. Electrochemical Performance of the DPF0.5 Electrolyte To evaluate the F– ion transport in this DPF0.5 electrolyte and the cycling stability of the fluorination/defluorination processes, the symmetric cell configuration is assembled using Sn@SnF2 electrodes. As shown in Figure A, in the Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cells, there is a small overpotential of 6.8 mV at 20 μA/cm2 at the precycle stage, and the polarization overpotentials are 20 and 50 mV at the higher current densities of 100 and 500 μA/cm2, with stable cycling for 500 and 170 h, respectively. The electrochemical impedance spectroscopy (EIS) of a symmetric cell based on the DPF0.5 electrolyte is carried out in a frequency range of 1 Hz–10 MHz (Figure B). The resistance of the symmetric cell before cycling is 59.7 Ω, and the resistances are 63, 74.7, and 81.5 Ω after cycling for 10, 100, and 200 h, respectively. The resistance arc and corresponding tail slope gradually increase, corresponding to the formation of a solid electrolyte interphase (SEI) at the electrode–electrolyte interface. The fluorination and defluorination voltage profiles of the symmetric cell at different current densities are displayed in Figure C. When the current density exceeds 100 μA/cm2, the voltage curves gradually become oblique. With the current density gradually increasing to 1000 and 1250 μA/cm2, the corresponding polarization values are 85 and 106 mV, respectively. When the current density returns to 100 μA/cm2, the polarization remains at 16.5 mV with further cycling for 350 h. Note that the symmetric cell based on Pb@PbF2 electrodes exhibits a high polarization overpotential up to 500 mV, and after cycling for 150 h, the overpotential gradually becomes asymmetric (Figure S18). This polarization difference may be caused by the different lattice energies of SnF2 and PbF2; thus, we use Sn@SnF2 as the electrode in the following text. Moreover, both DPF1.0 and DPF1.5 electrolytes are introduced into symmetric cells for comparison (Figure S19). There are high voltage polarization values of 110–400 mV, and the curve profiles are asymmetric. The increase of CsF content not only reduces the conductivity but also strongly retards the kinetics at the electrolyte–electrode interface, consistent with the low desolvation barrier of F– ions in the F–Cl exchange pathway. The long-term cycle stability, high current tolerance, and low polarization of symmetric cells illustrate the superior F– transport performance of DPF0.5 electrolyte. Figure D displays the electrochemical window of the DPF0.5 electrolyte, which ranges from −1.75 to 2.65 V (vs Sn@SnF2). This wide operating window of 4.4 V is attributed to the stability of the DMSO solvent and solvated structure, especially Ph4PF2 –. Considering the presence of the Cl element in the DPF0.5 electrolyte, the molar ratio of F/Cl = 1:0.3 is quantitatively determined by ion chromatography, based on the F– and Cl– ion concentrations of 0.487 and 0.147 mol/L, respectively (Figure S20). The F– ion transference number is estimated to be 0.42 from the chronoamperometry curve of the Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cell (Figure E,F). Theoretically, each acceptor central P atom can coordinate up to two halogen atoms, and about 70% of the chlorine element is precipitated as CsCl; thus, the molar ratio of phosphorus to halogen in the DPF0.5 electrolyte is expected to be 1:1.3. This effect indicates that there are many five-coordinate species (e.g., Ph4PCl and Ph4PF) that can be regarded as F– ion vacancies to promote F– ion migration. Symmetric cell performance of DMSO-Ph4PCl-CsF electrolyte. (A) Cycling performance of Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cells at 100 and 500 μA/cm2 and (B) its Nyquist plots at different cycle stages at room temperature. (C) Rate performance of symmetric cells at various current densities in a range of 25–1250 μA/cm2. (D) Electrochemical window of DPF0.5 through linear scanning voltammetry (LSV) test. (E) Chronoamperometry curve of the Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cell under an applied voltage of 10 mV and (F) corresponding Nyquist plots of the symmetric cell before and after polarization at room temperature. In order to assess the practical application of DPF0.5–1.5 electrolytes, we incorporated them into the cells with a conversion-type CuF2 cathode (with a theoretical energy density of 382 Wh/kg based on cathode mass and a theoretical specific capacity of 528 mAh/g for CuF2). For the CuF2-based cells, the EIS spectra are also used to evaluate the electrode–electrolyte interfacial resistance. The highest resistance value of 1907 Ω appears in the case of DPF1.5 electrolyte (Figure S21), and the resistances of DPF1.0 and DPF0.5 systems are 92.6 and 66.6 Ω, respectively. At a current density of 40 μA/cm2, the initial discharge specific capacity of CuF2/DPF1.5/Sn@SnF2 cell (with a cathode loading of 1 mg/cm2) is 207.3 mAh/g (Figure A), and the initial charge specific capacity is 323.7 mAh/g, corresponding to an initial columbic efficiency as high as 149.9%. The voltage polarization during the first cycle is 156 mV. After the early activation cycles, the reversible specific capacity gradually increases to 478 mAh/g and remains at 214 mAh/g after 120 cycles, and the voltage polarization value gradually decreases from 110 mV and stabilizes at 30 mV (Figure B). The corresponding columbic efficiencies are in the range of 98–102%. This phenomenon of columbic efficiency exceeding 100% is due to the occurrence of Cu in the higher valent state (+2.5 or +3) on the cathode surface during the charging stage, which contributes to the extra reversible capacity. Note that, after about 20 cycles, the components in DPF1.5 electrolyte self-optimize so that the proportion of Ph4PF2 – decreases and the proportion of five-coordinated Ph4PF as an acceptor vacancy increases; correspondingly, the cell polarization decreases from 85 to 45 mV, and the capacity slightly increases from 260 to 356 mAh/g. The high electrode–electrolyte interface resistance (1907 Ω) leads to high voltage polarization, as shown in Figure S22. The resistance values after the 10th and 50th cycles are 1210.1 and 1419.2 Ω, respectively. In the DPF1.5 electrolyte, the excessive addition of CsF causes CsCl to almost precipitate, and the Cl content reduces. Therefore, the stable Ph4PF2 – occupies the vast majority of the AA coordination environment, and the content of five-coordinated Ph4PF is very low. As a result, F– ion migration vacancies are reduced, and F– desolvation becomes difficult in view of the dominance of the direct defluorination pathway, leading to the high polarization and resistance of cells. Also, the excessive F may lead to the emergence of a lot of CsF in the SEI layer, which would cause the repeated fluctuations in specific capacity. For the DPF1.0 electrolyte (Figure S23), the initial specific capacity is 387.6 mAh/g with a voltage polarization of 51.9 mV, and the reversible capacity is only 40 mAh/g. Its cycling stability is poor. Electrochemical performance of CuF2//Sn@SnF2cells and kinetics evaluation of electrolyte–electrode interface. Cycling performance of CuF2//Sn@SnF2 cells based on (A) DPF1.5 and (C) DPF0.5 electrolytes at 40 μA cm–2 and (B,D) corresponding galvanostatic curves in a voltage range of −0.2–0.2 V. (E) Comparison of electrochemical performance between this CuF2/DPF0.5/Sn@SnF2 battery and other reported liquid electrolyte FIBs. (F) Galvanostatic discharge/charge profiles of CuF2//Sn@SnF2 cell based on DPF0.5 electrolyte. (G) In situ EIS evolution, and corresponding distribution of relaxation time (DRT) analysis at (H) charge and (I) discharge stages. The CuF2/DPF0.5/Sn@SnF2 cell exhibits superior electrochemical performance at 40 μA/cm2 (even based on a much higher cathode loading of 4.2 mg/cm2, Figure C). The initial discharge and charge specific capacities are 473.9 and 720.5 mAh/g, respectively, with a Coulombic efficiency of 152%, and the corresponding voltage polarization is as small as 26 mV. The highest reversible specific capacity is 717.7 mAh/g and the capacity remains at 316 mAh/g after 65 cycles, with the stable voltage polarization preserved at only 11 mV (Figure D). The comparison of capacity and polarization voltage between DPF1.5 and DPF0.5 electrolytes is plotted in Figure S23. In the DPF0.5 electrolyte, the suitable F–Cl ratio enables the F–Cl exchange dominating the desolvation process, with the low energy barrier and fast kinetics. In contrast, in the DPF1.5 electrolyte, the small Cl content and dominant Ph4PF2 – coordination environment result in the dominant direct defluorination pathway, with the high energy barrier and sluggish kinetics. The higher slope of capacity decay for DPF0.5 is also attributed to the much higher loading mass. Benefiting from the fast defluorination kinetics dominated by the F–Cl exchange pathway in our designed DPF0.5 electrolyte, the CuF2/DPF0.5/Sn@SnF2 battery exhibits the smaller voltage polarization and excellent reversibility compared with the already reported liquid electrolyte FIBs (Figure E), which illustrate both the outstanding F– transport and conversion reaction kinetics. ,,− The small voltage polarization of the voltage–capacity curve is related not only to the fast F– ion transport kinetics but also to the dissolution-deposition fluorination mechanism on the electrode surface. After initial discharge, the acceptor can also dissolve the fluoride species (such as CuF2, SnF2, and PbF2) on the electrode surface in the same way as dissolving CsF, and it would cause a dissolution-deposition fluorination reaction. In terms of the thermodynamic dissociation energies of these fluorides (Figure S24), the highest barrier of PbF2 is 7.2 eV; thus, its symmetric cell polarization overpotential is higher than that of SnF2 (with a barrier of 6.9 eV), consistent with the symmetric cell measurement. Note that the dissociation barrier of CuF2 is 4.5 eV, close to that of CsF (4.3 eV); thus, the dissolution-deposition reaction mainly occurs on the CuF2 surface. Thus, the polarization value of the CuF2/DPF0.5/Sn@SnF2 cell is low and ranges from 26 to 11 mV, close to that of the Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cell (12 mV) at 50 μA/cm2. Moreover, a trace amount of free Sn2+ ions in the electrolyte can diffuse into the cathode side driven by the electric field and concentration polarization, finally leading to the formation of some metallic alloys at the cathode side (as discussed later). When the equilibrium of concentration and the formation of the SEI layer, the diffusion and alloying processes are constrained. The scanning electron microscopy (SEM) images of the CuF2 cathode surface are observed after 5 cycles (Figure S25). There are continuously distributed Cu nanoparticles/alloy in the discharged (defluorination) state, and the number of Cu nanoparticles/alloy in the charged (fluorination) state is significantly reduced, but they do not completely disappear. The small amount of irreversible Cu or alloy nanoparticles may be the reason for the gradual decline of specific capacity. To address the potential CuF2 dissolution loss causing capacity decay, we conducted inductively coupled plasma-optical emission spectrometer (ICP-OES) measurement on the DPF0.5 electrolyte after 5 cycles (Figure S25c,d). The dissolved Cu only accounts for 0.001% of the total mass of the Cu element in the cathode, and the almost negligible CuF2 dissolution amount hardly dominates the battery capacity decay. To understand the charge transfer kinetics of the CuF2/DPF0.5/Sn@SnF2 cell, in situ EIS measurement with DRT technique analysis is carried out in a frequency range from 106 to 100 Hz. , The typical voltage–capacity curve of CuF2/DPF0.5/Sn@SnF2 cell is shown in Figure F, and its in situ AC impedance spectra are correspondingly recorded in Figure G, in which a constant real axis intercept is observed, representing the internal resistance of the cell. This cell displays a small resistance ranging from 45 to 56 Ω, corresponding to the semicircle part located at high/medium frequencies of 1 × 106–102 Hz. These resistance semicircles are overlapping, and only the slope of the diffusion tail is different, causing the ambiguous resistance comparison. Therefore, the DRT analysis is used to decouple these EIS data. As displayed in Figure H,I, the P1 peak (located at 10–6 s) maintains an almost constant intensity, and it belongs to the internal resistance of the cell, contributed by the ionic and electronic resistances of electrodes and electrolyte. The P2 peak (10–4–10–3 s) is attributed to the electrical double layer relaxation and the formation of an SEI layer at the Sn@SnF2 anode. The P3 peak (10–3–10–2 s) has a low intensity down to about 5 Ω, and it represents the charge transfer at the dissolution-deposition interface on the CuF2 cathode. The P4 peak (10–2–10–1 s) stands for the charge transfer behavior in the cathode electrolyte interface (CEI) layer during the fluorination/defluorination reaction. The P5 peak (10–1–100 s) represents the resistance of the ion diffusion process in the cell. At the end of the discharge state (i.e., complete defluorination of CuF2 cathode), the corresponding P3 peak almost disappears and overlaps with the broadened P2 and forward P4 peaks. It means that the resistance at the CuF2 cathode side is the lowest, and a lot of Cu nanoparticles are generated, and that the resistance of the Sn@SnF2 anode is the highest, and a large amount of Sn is fluorinated. At the other discharge and charge states, these P2, P3, and P4 peaks maintain similar peak profiles and stable resistances. In view of the resistance value, these P5 peaks are generally 1 order of magnitude higher than those of other peaks (Figure S26), and the strongest P5 peak appears at the discharge end state. Therein, the stable Ph4PF2 – dominates the electrolyte where the Cl– ions with assisted desolvation role are precipitated as CsCl, leading to the constrained F– migration kinetics and thus the maximal resistance. The small P5 peak resistances appear in the early stages of discharge and the late stages of charge, corresponding to the active ion transfer environment. Therein, the Ph4PCl2 – and Ph4PFCl– species are formed in electrolyte, the CsCl precipitate is dissolved, and the F– desolvation becomes easier than at the discharge end state, resulting in the fast F– and Cl– ion migration kinetics. (De)­fluorination Reaction Mechanism of the Cathode in the DPF0.5 Electrolyte High-resolution transmission electron microscopy (TEM) (HRTEM) and selected area electron diffraction (SAED) are used to reveal the microstructure and phase evolution of the CuF2 cathode at charge and discharge states. For the charge state (after 5 cycles), there is an overall cluster morphology (Figure A), consisting of CsF, CsCl, CsCuF4, CsCu2F6, and CuF2 phases (Figure B). Among those, CsF and CsCl are formed by the spontaneous crystallization of electrolyte residues after drying. The CuF2, CsCuF4, and CsCu2F6 phases are typical fluorination reaction products, especially the latter two demonstrate that the extra capacity comes from the high oxidation state Cu compounds of CsCuF4 (+3) and CsCu2F6 (+2.5). As shown in Figure C,D, the fluorination reaction phases present a gradient distribution based on Cu valence states, i.e., the inner CuF2 region, the middle CsCu2F6 region, and the outer discrete CsCuF4 region. The formation of the gradient region of CuF2–CsCu2F6–CsCuF4 depends on the diffusion distance of Cs+ in the CuF2 crystal domains, thus forming a CEI layer with a thickness of 10–30 nm. The fluorination mechanism of the CuF2/DPF0.5/Sn@SnF2 battery at the charged state is depicted in Figure E. There are large and integrated fluoride regions, and the Ph4PX2 – acceptor complex migrates to the cathode side and releases F– to achieve Cu fluorination. For the discharge state (after 5 cycles), the overall morphology is layered and discrete (Figure F,H). There are CsCl, CsF, Cu3Sn, Cu, CsCu2F6, and CsCuF4 phases, as seen from the SAED pattern (Figure G); among them, Cu and Cu3Sn stand for the defluorination reaction of the CuF2 cathode. The free Sn2+ diffuses to the cathode surface under the concentration difference and electric field to form the Cu3Sn alloy phase. Note that the fraction of Sn is much lower than that of Cu on the cathode surface, and there is no SnF2 in the charge state, illustrating that the diffused Sn does not contribute to the capacity. In addition, compared to the charge state, the finer and irregular crystal domains such as CsF, Cu, CsCu2F6, and CsCuF4 are observed in the HRTEM of the discharge state (Figure I). The adjacent CsF and Cu are the defluorination reaction products of CsCu2F6 and CsCuF4, and the smaller CsCu2F6 and CsCuF4 crystal domains are the residual phases after incomplete defluorination. Note that the observed CsCl crystal domain not only is the electrolyte residue after drying but also includes the CsCl precipitated from the Ph4PCl2 – and Ph4PFCl– acceptors. After comparing the SAED patterns of charge and discharge states (Figure B,G), we find that both have dispersion diffraction rings, reflecting the existence of an amorphous phase. There are some different diffraction planes of the same species at charge and discharge states, such as for CsF, CsCl, CsCu2F6, and CsCuF4 fluorides. Combined with the phase distribution differences in HRTEM images (Figure C,D,I), we can infer that these crystal phases are involved in the fluorination and defluorination reactions. At the discharge state (Figure J), the acceptor central P captures F– ions and migrates to the anode side under an electric field; meanwhile, the desolvated Cl– ions are precipitated as CsCl on the cathode surface. Also, the CsCu2F6 and CsCuF4 fluorides are defluorinated to form CsF and Cu, under the attraction of the acceptor central P to F– ions. In terms of the fluoride dissociation (Figure S27), the barriers of CuF2, CsCuF4, and CsCu2F6 are 4.5, 6.1, and 9.6 eV, respectively. The difficult dissociation of CsCuF4 and CsCu2F6 leads to their residual phases during electrochemical cycling. (De)­fluorination reaction mechanism of the CuF2cathode in the DMSO-Ph4PCl-CsF electrolyte. (A) TEM image of CuF2 cathode at the charge state, (B) corresponding SAED pattern, and (C,D) HRTEM images at different regions presenting the phase distribution gradient. (E) Schematic illustration of the fluorination process of the cathode at the charge stage and F– and Ph4PF2 – cluster migration process. (F,H) TEM images, (G) SAED pattern, and (I) HRTEM image of the CuF2 cathode at the discharge state, displaying a relatively discrete phase distribution. (J) Schematic illustration of the defluorination process of the cathode at the discharge stage, F– extraction mechanism, and phase separation. The chemical composition and valence state evolution of the CuF2 cathode are further analyzed by X-ray photoelectron spectra (XPS). For the Cu 2p spectra (Figure A), there are clear fluorination and defluorination signals. The highest Cu–F peak (933.1 eV) exists in the charge state after etching for 30 s, and it represents the fluorination of Cu. This peak intensity is reduced at the discharge state, representing the defluorination reaction. The Cu0 (930.2 eV) signal exists in charge state even with a minor intensity, and it may be caused by the residual of nonfluorinated metallic Cu. In the clearer Cu LMM spectra (Figure B), the ever-present Cu0 (568.5 eV) and boundary Cu (566.0 eV) signals may originate from Cu3Sn alloy, and the alloying Cu is difficult to fluorinate and is prone to remaining on the cathode surface to serve as a component of CEI layer and contribute to excellent charge transfer. The Cu2O peak (570.6 eV) is related to the inevitable oxygen intrusion during the ball-milling preparation process of the CuF2 cathode. The constant presence of Cu–F peaks (571.5 eV) even under the discharge state indicates that complete defluorination is difficult. At the charge state after etching for 30 s, the Cu–F peak shifts to a higher energy position at 572.7 eV, meaning the formation of CsCuF4 and CsCu2F6 fluorides with higher Cu-oxidation states. For the F 1s spectra (Figure C), these peaks are deconvoluted into CsF (682.4 eV), CuF2 (684.8 eV), and F–C (687.6 eV) from PVDF binder under charge and discharge states. Note that, at charge state, CsCuF4 (685.7 eV) and CsCu2F6 (686.4 eV) appear at the surface and at the subsurface after etching for 30 s, respectively, consistent with the HRTEM analysis (Figure C). For the Cl 2p spectra, the peak located at 198.3 eV is identified as CsCl (Figure D). As shown in Figure E, the Cs 3d5/2 spectra are separated into Cs–Cu–F (725.3 eV), CsF (724.7 eV), and CsCl (722.8 eV). The CsCl signal intensity at the discharge state is significantly stronger than that at the charge state, indicating that more CsCl is precipitated on the CuF2 cathode surface during the discharge stage, while CsCl is dissolved in the DPF0.5 electrolyte during the charge stage. The change in CsCl content on the cathode surface at different states proves that it is involved in the solvation process. In the P 2p spectra (Figure F), there are two obvious peaks belonging to Ph4PF (133.2 eV) and Ph4PCl (130.8 eV) at the charge state, but without corresponding signals at the discharge state. This comparison indicates that these acceptor-F complexes (including Ph4PF2 –, Ph4PFCl–, and Ph4PCl2 –) migrate to the cathode side under an electric field during the charge stage. Combined with the changes in CsCl signals and P 2p spectra, we find an interesting species evolution phenomenon. In the discharge stage, the defluorination reaction of CuF2 releases capacity, and the free F– ions are captured by the AA central P atom, forming many P–F covalent bonds. At this time, the DPF0.5 electrolyte is dominated by Ph4PF2 – and Ph4PFCl–, which migrate toward the anode side. The desolvated Cl– bonds with Cs+ in the electrolyte and precipitates on the cathode surface, corresponding to the solidification of chloride. Thus, we observe that the CsCl signal is enhanced and there is no P 2p signal at the discharge state. In the charge state, Ph4PF2 – and Ph4PFCl– transport to the cathode side and release F–, so obvious P 2p signals are observed on the charge-state cathode surface. In addition, the AA dissociates CsCl precipitate, and Cl– ions re-enter the solvation state (i.e., resolvation of Cl–), leading to reduced CsCl and promoted F– desolvation. At this time, the metal Cu is refluorinated to CuF2 and Cs+ infiltrates to form the higher oxidation state CsCu2F6 and CsCuF4. Also, the increase of CsCl content in the discharge state is relatively consistent with the result of multiple diffraction rings of CsCl in SAED analysis. Surface composition evolution of the CuF2cathode after five cycles. XPS spectra of (A) Cu 2p3/2, (B) Cu LMM, (C) F 1s, (D) Cl 2p, (E) Cs 3d5/2, and (F) P 2p core levels at discharge and charge states with and without the etching process. For the alloying phenomenon on the cathode surface, the Sn 3d XPS spectra at both cathode and anode sides are also analyzed after cycling (Figures S28 and S29). Different from the Sn0 peak positions (Sn 3d5/2 at 484.7 eV and Sn 3d3/2 at 493.2 eV) in the Sn@SnF2 anode, the corresponding peak positions in the CuF2 cathode shift toward the higher energy direction, with Sn 3d5/2 at 485.2 eV and Sn 3d3/2 at 493.6 eV belonging to the Cu3Sn alloy phase. The Sn metal (Sn 3d5/2 at 484.5 eV and Sn 3d3/2 at 492.8 eV) only appears on the cathode surface, and the Sn 3d5/2 and Sn 3d3/2 peaks located at 487 and 495.4 eV should correspond to the formation of SnO2. Correspondingly, the boundary Cu peak position in Cu 2p spectra is lower than that of the Cu0 peak (Figure A), which may be attributed to the formation of Cu3Sn alloy phase. The Cu3Sn alloy phase causes an increase in the binding energy of the Sn 3d peak and a decrease in the binding energy of the Cu 2p peak, which has been reported previously. In addition, the peak intensity of Sn 3d spectra changes with the discharge and charge states of the CuF2 cathode. The Cu3Sn peaks in Sn 3d in the charged state are significantly weaker than those in the discharged state, and it is closely related to the morphological evolution of the CEI layer on the CuF2 cathode surface. As shown in Figure , the CEI layer formed in the charged state would shield the Sn 3d signals, and the phase separation generated by the defluorination reaction in the discharged state would expose the Cu3Sn particles, resulting in an enhancement of the Sn 3d signal. The persistent presence of a Cu3Sn signal illustrates the loss of Cu bound by alloying, leading to a capacity decay. On the other hand, from Figure S29, the main components of the SEI layer include CsF, CsCl, and a few of the acceptor (Ph4PF). In these components, the Ph4PF acceptor has fluorine ion vacancies and can play a role in F– ion transport, and it can also promote the dissociation of fluoride, which likely undergoes a dissolution-deposition mechanism and reduces the battery polarization voltage. CsF in the SEI layer is expected to block the growth of SnF2 grains and has a certain confinement effect. Also, CsCl can transform into the resolvated Cl– ions and bind to the acceptor central P, achieving the F–Cl exchange and promoting the efficient F– ion desolvation at the electrode–electrolyte interface. Closed-Loop Conversion Mechanism of Chlorine Herein, we build a chemical landscape of fluoride ion electrolyte DPF0.5 and the reaction mechanism (Figure ). Since the aprotic solvent DMSO does not dissolve CsF and CsCl salts, the AA Ph4PCl is used to assist in dissolving CsF through the F–Cl exchange reaction, and correspondingly, a lot of CsCl is precipitated. The F– and Cl– ions exist in the electrolyte in the form of solvates, such as Ph4PCl, Ph4PF, Ph4PF2 –, Ph4PCl2 –, and Ph4PFCl–, and free Cs+ exists in the electrolyte to maintain its electrical neutrality. Among them, these five-coordinated acceptors not only serve as an active site to promote F– ion conduction, but also accelerate the dissolution of fluoride (CuF2, CsCuF4, and CsCu2F6) in defluorination reaction and enhance the conversion reaction kinetics. Due to the significant steric hindrance effect of large size Cl–, the desolvation barrier of F– in the F–Cl exchange pathway is significantly decreased, and thus the solvated Cl– facilitates the fluorination–defluorination reaction kinetics. During the discharge (defluorination reaction) process, Cl– ions are solidified as the CsCl crystal phase and precipitated on the cathode surface, and the six-coordinated acceptors Ph4PF2 – and Ph4PFCl– carry F– to the anode side. When in the charging (fluorination reaction) state, these six-coordinated acceptors migrate to the cathode side and therein release F–. The resulting five-coordinated acceptors dissociate the CsCl precipitate, and the resolvated Cl– ions accelerate the F– desolvation and fluorination reaction kinetics. Further, Cs+ ions infiltrate into the formation of CuF2 and finally lead to the coformation of CsCuF4 and CsCu2F6, contributing extra reversible capacity. In the above discussion, we find that Cl– ions are interconvertible between solid-phase chlorine (e.g., in CsCl) and solvated chlorine (e.g., in Ph4PFCl– and Ph4PFCl2 –) during charge and discharge. Therefore, Cl– with steric hindrance is regarded as a closed-loop conversion mediator, and it can promote the dissociation and solvation of fluorides and the desolvation of F– ions, which are sustainable rather than one-time consumed, accelerating the kinetics of the (de)­fluorination reaction. Schematic diagram of acceptor-accelerated F– migration and F–Cl exchange mechanism at the acceptor center phosphorus, and chemical landscape of the closed-loop conversion mechanism of Cl– ions between solid and solvated chlorine. Conclusions In summary, we designed an aprotic fluoride ion electrolyte, which uses DMSO as a solvent and the AA central P as the reaction site. Through the F–Cl exchange reaction and CsCl preferential precipitation in DMSO, many fluorine vacancies are constructed on the acceptor P, providing some active sites for the fast migration of F– ions, with a high ionic conductivity of 4.1 mS/cm and a high F– transference number of 0.42 at room temperature. Benefiting from the significant steric hindrance effect of large Cl– ions, the F– desolvation kinetics is enhanced. The five-coordinated acceptor P can accelerate the dissociation of fluoride through the attraction of P to F, which eliminates the hysteresis issue of the (de)­fluorination reaction at conversion-type cathode CuF2. Also, the fast electrode–electrolyte kinetics generates the high oxidation state fluorides CsCuF4 and CsCu2F6, providing extra reversible capacity. The Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cells display a long cycling lifespan of over 500 h at 100 μA/cm2 with a voltage polarization as small as 20 mV and exhibit excellent rate performance with a high current density up to 1250 μA/cm2. Due to the dissolution-deposition fluorination mode, the CuF2/DPF0.5/Sn@SnF2 cell exhibits the highest reversible specific capacity up to 717.7 mAh/g and remains a capacity of 316 mAh/g after 65 cycles at 40 μA/cm2 (under a high cathode loading of 4.2 mg/cm2), and the voltage polarization remains stable at only 11 mV. Further, based on the HRTEM and XPS results, the (de)­fluorination mechanism of electrolyte DPF0.5 is revealed, and a concept of closed-loop conversion of chlorine is proposed. This work not only provides a novel aprotic AA-based fluoride ion electrolyte but also points out the design concept of AAs for developing high capacity and long lifespan FIBs. Experimental Section Preparation of the Electrolyte First, 0.3748 g of tetraphenylphosphonium chloride (marked as Ph4PCl, Macklin, 98%) was dissolved in 2 mL of dimethyl sulfoxide (DMSO, Aladdin, 99.9%) after continuous stirring for 30 min, and a homogeneous and transparent solution (marked as DMSO-Ph4PCl) was obtained. Then, 0.1519 g of cesium fluoride powder (CsF, Aladdin, 99.9%) was added to the solution and stirred for one night. The upper layer is a light yellow transparent solution (i.e., DPF0.5 electrolyte, with a measured F– ion concentration of 0.487 M), and the bottom is white precipitate (CsCl). When the CsF powder was increased to 0.3038 and 0.4557 g, we obtained the DPF1.0 and DPF1.5 electrolytes. Also, 0.1519 g of CsF powder was also added into 2 mL of DMSO (marked as DMSO-CsF), and the CsF is slightly dissolved. Preparation of the CuF2 Cathode First, 0.8 g of commercial CuF2 (anhydrous, Innochem, 99.5%) and 0.2 g of Ketjen Black (KB-C, Kejing) were mixed in a mass ratio of 8:2, and 45 g of zirconium dioxide ball mill beads with a diameter of 2 mm and 4 mL of 1-methyl-2-pyrrolidinone (NMP, C5H9NO, Aladdin, 99.9%) were also added as wet ball milling media. After high-energy ball milling at 600 rpm for 5 h, the obtained black slurry was rinsed with acetone (C3H6O, Aladdin, 99%) and dried in vacuum at 80 °C to obtain the primary composite powder CuF2–C. Then, the mixed slurry consisting of CuF2–C, Super P (Kejing), and poly­(vinylidene fluoride) (PVDF, Aladdin, average Mw ∼400,000) in NMP was doctored onto the carbon-coated aluminum foil, followed by drying in a vacuum oven at 80 °C for 24 h. Finally, the dried aluminum foil was cut into electrodes with a diameter of 8 mm. Preparation of the Composite Anode The Sn@SnF2 electrode was prepared by blending Sn powder (Aladdin, 99.99%), SnF2 powder (Macklin, 99%), Super P, and PVDF in NMP with a mass ratio of 4:4:1:1 and then coated onto Cu foil and dried at °C for 24 h under vacuum. Also, the Pb@PbF2 electrode was prepared using the same process as above. These dried Cu foils were cut into electrodes with a diameter of 10 mm. The active material loading of the anode electrode is excessive relative to that of the cathode in order to completely release the capacity of the CuF2 cathode. Electrochemical Performance Measurement of Electrolyte and Battery The ionic conductivity of an electrolyte was tested by a conductivity meter (INESA, DDSJ-308A) at room temperature. The 2025-type coin cells of Sn@SnF2//Sn@SnF2, Pb@PbF2//Pb@PbF2, and CuF2//Sn@SnF2 configurations were assembled in an Ar-filled glovebox (Vigor, O2, and H2O < 0.1 ppm), and a glass fiber membrane was used as the separator. Electrochemical performance experiments were carried out using a battery testing system (LAND-CT2001A). For the Sn@SnF2//Sn@SnF2 and Pb@PbF2//Pb@PbF2 symmetric cells, the cycling program was set at the current density of 100 or 500 μA/cm2 with a fixed plating/stripping interval of 0.5 h, and the current density was set from 25 to 1250 μA/cm2 in the rate measurement. The CuF2//Sn@SnF2 cells were tested at a current density of 40 μA/cm2 in a voltage range of −0.2–0.2 V. The AC impedance spectra of symmetric and asymmetric cells were recorded by a Solartron frequency analyzer (1260–1296) in a frequency range from 10 MHz to 1 Hz with a fixed amplitude of 10 mV. LSV was used to evaluate the electrochemical window of the DPF0.5 electrolyte in the Sn@SnF2//stainless steel asymmetric cell architecture on the electrochemical workstation (VersaSTAT 3, AMETEK Scientific Instruments), and it was conducted from 0 to 3 V and from 0 to −3 V, with the positive and negative scan rate of 1.0 mV/s. The ionic transference number (t –) of the DPF0.5 electrolyte was obtained from AC impedance and DC polarization measurement by using a Sn@SnF2/DPF0.5/Sn@SnF2 symmetric cell architecture. The AC impedance was first recorded before the DC polarization test, and then the symmetric cell was performed on a VersaSTAT 3 electrochemical workstation with a voltage bias of 10 mV, and finally, AC impedance was recorded again to obtain the EIS data of the symmetric cell after steady-state current. The t – value is calculated according to the following equation t−=Is(ΔV−I0R0)I0(ΔV−IsRs) where the ΔV is the voltage bias (10 mV), I 0 and I s stand for the initial and steady-state currents, respectively, and R 0 and R s represent the resistances of the symmetric cell before and after the DC polarization test. In situ EIS spectra were recorded by the chronopotentiometry method (with a constant current of 10 μA and a sampling interval of 10 min) based on the CuF2/DPF0.5/Sn@SnF2 cell at a VersaSTAT 3 electrochemical workstation under a frequency range of 1 MHz to 1 Hz and a fixed voltage amplitude of 10 mV. The DRT as a model-free analysis method was used to deconvolute the AC impedance data through an infinite number of relatively small, different RC parallel circuit components (i.e., τ = RC), converting the frequency-domain EIS spectra to the relaxation time-domain function γ­(τ) according to the following mathematical transformation Z(ω)=R0+Zpol(ω)=R0+Rpol∫0∞γ(τ)1+jωτdτ where Z(ω) is the impedance data as a function of angular frequency ω, τ is the relaxation time, and R 0 and R pol represent the ohmic and polarization resistances, respectively. The DRT analysis was conducted with RelaxIS 3 software. Characterizations For the characterization of an electrolyte, NMR measurements were carried out on a Bruker Advance HD 400 MHz spectrometer with a deuterated reagent of DMSO-d 6. The 19F NMR spectra of DPF0.5 and DMSO-CsF were acquired to prove the solubility of CsF salt assisted by Ph4PCl and the chemical environment of the fluorine element. The 31P and 1H NMR spectra of DPF0.5 and DMSO-Ph4PCl were obtained to distinguish the chemical coordination environment of the central P atom in different electrolytes. Fourier transform infrared spectroscopy (FTIR) measurements of DPF0.5 and DMSO-Ph4PCl were performed on the Nicolet iS5 equipment within a range of 400–4000 cm–1. Raman spectra were acquired using a miniature laser confocal Raman spectrometer (DXR2xi) with a 532 nm laser at room temperature. The anion content and ratio were quantitatively determined by using ion chromatography (Dionex Aquion ICS 2100). For the ion chromatography measurements, 0.5 mL of DPF0.5 electrolyte was diluted to 10 mL to meet the equipment test requirement. For the ICP-OES (PE Avio200) measurement, a large amount of electrolyte is required. We disassembled the coin cell after 5 cycles and soaked the separator in 7 mL of DMSO for 12 h under an Ar atmosphere. XRD (Bruker D8 Advance diffractometer with Cu Kα radiation) was used to detect the phase of dried white powder at the bottom of the DPF0.5 electrolyte. The microstructures and surface morphologies of the CuF2 cathode after 5 cycles were observed by a field-emission electron scanning microscope (FE-SEM, Magellan 400 L, FEI). The TEM (JSM-6700F, JEOL) and SAED were operated at 200 kV to characterize the microstructure and phase distribution of the CuF2 cathode at discharged and charged states after 5 cycles. The cycled CuF2/DPF0.5/Sn@SnF2 coin cell was disassembled in an Ar-filled glovebox, and the CuF2 cathode powder was carefully scraped off with a new scalpel. The cycled sample was then placed in 0.5 mL of NMP and ultrasonically vibrated to obtain a cathode component suspension, and finally, the TEM sample was prepared. Similarly, the CuF2 cathodes at discharged and charged states were directly peeled off for XPS measurement to investigate the chemical state of elements on the CuF2 cathode surface after 5 cycles. These XPS data were obtained from an ESCAlab-250 instrument with an Al target. Computational Methods Density functional theory (DFT) calculations for this work were performed using the Vienna Ab initio simulation package (VASP), within the projector-augmented-wave approach, and the Perdew–Burke–Ernzerhof (PBE) generalized gradient approximation functional was used. For all the geometrical optimizations, a kinetic energy cutoff of 450 eV was used, and the convergence criterion was set as 10–6 eV in energy and 0.02 eV/Å in force, respectively. The van der Waals interaction was corrected by the Grimme D3 method. The following static self-consistent calculations used a finer energy convergence criterion of 10–7 eV. For the charged species, a spin-polarized calculation with the self-consistent potential correction method was conducted to define the energy of charged species. For all single molecule systems, such as Ph4PF, Ph4PCl, Ph4PF2 –, Ph4PFCl–, Ph4PCl2 –, Ph4P+, PCl4, PCl5, PCl6, PF4, PF5, PF6, Ph3PCl2, Ph3PCl2F–, Ph3PClF2 –, Ph3PCl3 –, and Ph3PF3 – (Figure S30), they were placed into a large cell with a volume of 25 × 26 × 27 Å3, to avoid the interactions between adjacent molecules, and with a gamma-centered k-points mesh of 2 × 2 × 2, to meet the subsequent analysis. The adsorption energies (E ads) of Ph4PX or Ph4P+ for halogen ion (X–) are calculated based on the following equations Eads=−(Etotal−EPh4P+−EX−)⁣forPh4P−ClandPh4P−Fsystems Eads=−(Etotal−EPh4F−EX−)⁣forPh4PF−Fsystem Eads=−(Etotal−EPh4Cl−EX−)⁣forPh4PCl−Cl/Fsystems where E total, EPh4F , EPh4Cl , EPh4P+ , and EX− are the energies of adsorption configuration, Ph4PF, Ph4PCl, Ph4P+, and halogen ion, respectively. In the F– ion desolvation pathway calculation, the total amount of substances in the entire reaction pathway is always constant, consisting of 1 P, 24 C, 20 H, 2 F, and 2 Cl. The Gibbs free energy of F– ion desolvation was calculated based on optimized geometry and thermodynamic correction according to the following equation (Table S3)­ ΔG=ΔE+ΔZPE−TΔS where ΔE is the reaction energy of the initial and final molecules, and the zero-point energy (ΔZPE) and entropy (TΔS) were acquired from the vibration frequencies calculation. The Gibbs energy barriers of CsF dissociation with and without AA Ph4PCl and F–Cl exchange reactions are calculated according to the following reaction equations CsF→Cs++F−⁣directdissociation CsF+Ph4P+→Ph4PF+Cs+⁣withPh4P+ CsF+Ph4PCl→Ph4PF+CsCl↓⁣F−Clexchange These energy barriers are calculated by subtracting the total Gibbs free energies of reactants from the total free energy of products, e.g., ΔE = G (Cs+) + G (F–) – G (CsF). Generally, a negative value means the reaction is spontaneous, a positive value means the reaction is nonspontaneous and requires extra energy. For the dissociation energies of CsF, PbF2, SnF2, CuF2, CsCuF4, and CsCu2F6, the energies of Cs, Cu, Sn, Pb, and F atoms were used based on bulk Cs, Cu, Sn, and Pb metals and F2 gas phase. The k-point mesh used for bulk materials of Ph4PF and Ph4PCl is 3 × 2 × 3. The pCOHP analysis was carried out by using the LOBSTER package. For the dissociation energy calculation, the bulk phases CsF (6 × 6 × 6), CsCl (8 × 8 × 8), PbF2 (5 × 9 × 4), SnF2 (7 × 6 × 4), CuF2 (10 × 7 × 7), CsCuF4 (6 × 6 × 8), CsCu2F6 (7 × 7 × 8), Sn (12 × 12 × 8), Pb (7 × 7 × 7), Cu (9 × 9 × 9), and Cs (5 × 5 × 5) were optimized (Figure S31), and their sampling K-spacing value was kept at 0.03. The quantum chemistry calculations, including geometry optimization and single point energy, were carried out by the Gaussian 09 software package at the M062X/def2TZVP basis set level. An implicit solvent model was used to evaluate the solvent effect for the AA by a solvation model based on density (SMD). The key descriptors for DMSO solvent include its dielectric constant (ε = 46.826), refractive index (n 25 = 1.4783), bulk surface tension (γ = 61.78), and Abraham’s hydrogen bond acidity and basicity (0 and 0.88, respectively). The molecular electrostatic potential (MESP) and interaction region indicator (IRI) maps were calculated and visualized by Multiwfn 3.8 and VMD software. Supplementary Material