Regulating Interfacial Chemistry in Lithium-Ion Batteries by a Weakly Solvating Electrolyte**

The performance of Li-ion batteries (LIBs) is highly dependent on their interfacial chemistry, which is regulated by electrolytes. Conventional electrolyte typically contains polar solvents to dissociate Li salts. Herein we report a weakly solvating electrolyte (WSE) that consists of a pure non-polar solvent, which leads to a peculiar solvation structure where ion pairs and aggregates prevail under a low salt concentration of 1.0 M. Importantly, WSE forms unique anion-derived interphases on graphite electrodes that exhibit fast-charging and long-term cycling characteristics. First-principles calculations unravel a general principle that the competitive coordination between anions and solvents to Li ions is the origin of different interfacial chemistries. By bridging the gap between solution thermodynamics and interfacial chemistry in batteries, this work opens a brand-new way towards precise electrolyte engineering for energy storage devices with desired properties.


Introduction
TheN obel Prize in Chemistry 2019 finally rewarded the development of Li-ion batteries (LIBs). These lightweight, rechargeable,a nd ubiquitous energy storage devices have profoundly revolutionized our modern life during the past 30 years. [1] Thei ncreasing demands of electric vehicles and grid energy storage is gradually pushing the performance of LIBs to their limits,i ncluding high energy density,f ast-charging, high safety,long life,and low cost. [2] To meet these high bars, current LIBs must venture into more challenging territories such as Li/Si anodes, [3] high-voltage/capacity cathodes, [4] and aqueous LIBs. [5] Eventually,t he challenges for these aggressive battery chemistries are partially or completely passed on to designing advanced electrolytes. [6] Theelectrolytes in LIBs not only serve as an ionic conductor, but also largely determine the electrode/electrolytei nterfacial chemistry. [7] Theexploration of state-of-the-art electrolytes is essential to achieve to the high expectations of working rechargeable batteries since the performance of LIBs is strongly dependent on the electrode/electrolyte interfaces.
It is well-established that the interfacial chemistry on electrodes is closely correlated to the solvation structure of electrolytes.I nc onventional dilute electrolytes,L ii ons are usually solvated by strongly solvating polar solvents and most anions are excluded from the solvation sheath ( Figure 1a). [7,8] Since the primary solvation sheath is the precursor of solid electrolyte interphase (SEI), such solvation structure leads to solvent-derived interfacial chemistry. [9] Fore xample,t he indispensable role of ethylene carbonate (EC) in modern LIBs originates from its preferential solvation and reduction which creates an exclusive EC-derived SEI to support reversible Li + intercalation in graphite.One major innovation of unconventional electrolytes in the past decade is the concept of superconcentrated electrolyte (SCE), with salt concentration (> 3.0 M) far beyond conventional electrolytes ( % 1.0 M, required by the optimum conductivity). [10] Unlike the solvent-dominated solvation structure in dilute electrolytes,anions inevitably appear in the primary solvation sheath of Li + to form ion pairs or aggregates because of the scarcity of solvents and abundance of anions ( Figure 1a). [11,12] Such solvation structure leads to anion-derived SEI that enables high-rate and long-term cycling of graphite and Li metal electrodes. [13] Considering the high cost and viscosity of SCE, diluting SCE with non-polar solvents emerged in recent years as an alternative to mitigate these issues. [14,15] Thediluted SCE is termed localized superconcentrated electrolytes (LSCE) because the local solvation structure of LSCE is very similar to that of SCE, and therefore they belong to the same methodology.
Because solvent and anion can both serve as ligands to coordinate with Li + through ion-dipole or ion-ion interactions,t he actual solvation structure depends on the competitive coordination between them. [16] In dilute electrolytes, solvents usually outnumber anions and hence dominate the solvation sheath of Li + .T oa chieve anion-derived interfacial chemistry,the straightforward strategy is to increase the ratio of anion to solvent as in SCE or LSCE (Figure 1a). However, is this the only way towards anion-derived interfacial chemistry?
Am ore essential approach towards anion-derived interfacial chemistry involves tuning the intrinsic solvating power of solvents.B ecause solvents and anions are competing to enter the solvation sheath of Li + ,r educing the solvating power of solvents can theoretically allow more anions to coordinate with Li + .The ideal scenario (Figure 1a)isaweakly solvating electrolyte (WSE) that generates abundant ion pairs or aggregates under low salt concentrations.W hile SCE and LSCE are extensively studied, WSE is rarely visited because solvents with low solvating power usually can not even dissolve enough Li salts. [14,16] This contradiction therefore has left this area blank, with some potentially important concepts and theories of electrolyte undiscovered.
In this work, we successfully prepared aW SE and systematically studied its solvation structure as well as interfacial chemistry on electrodes.P articularly,u ltra-low solvating power and moderate Li salt solubility are simultaneously achieved in as pecific solvent (1,4-dioxane) despite the apparent contradiction. Spectroscopic results confirm that WSE exhibits apeculiar solvation structure,inwhich ion pairs and aggregates prevail under astandard Li salt concentration of 1.0 M. Such solvation structure leads to an anion-derived, inorganic-rich SEI on graphite electrode,which allows for fast Li + transport. First-principles calculations unravel af undamental rationale that the relative binding energy between anions/solvents and Li + dictates the electrode/electrolyte interfacial chemistry,w hich blazes an ew trail in precise electrolyte design for future batteries.

Results and Discussion
Model System Figure 1b lists the dielectric constant (e,a lso known as permittivity) of various solvents used in this study,which is an important indicator of the solvating power of solvents. Another frequently used indicator,d onor number (DN), is also provided for al ist of solvents (Supporting Information, Table S1) along with adetailed discussion on the applicability of these two parameters.I nt he carbonate family,E C possesses an extremely high e of 89.8 as as trongly solvating solvent and dominates the primary solvation sheath of Li + . TheL i + -coordinated EC is then reduced on graphite electrode to form ad esirable SEI, which is contributed by the typical solvent-derived interfacial chemistry.I nt his study, commercial electrolyte consisting of EC/ethyl methyl carbonate (EMC;1 :2, v/v) mixed solvents and 1.0 Ml ithium bis(fluorosulfonyl)imide (LiFSI) serves as the control sample and is denoted as EC/EMC.T hree ethers,d imethoxyethane (DME), 1,3-dioxane (1,3-DX), and 1,4-dioxane (1,4-DX) with e of 7.0, 13.0, and 2.2, respectively,a re chosen as the model system to induce atransition from solvent-derived interfacial chemistry to anion-derived interfacial chemistry based on solvating power regulation (Figure 1c). DME has the largest solvating power among the three solvents despite the moderate e,b ecause it has ah igh donor number (DN = 20.0) and chelating effect on Li + . [17] 1,3-DX exhibits lower solvating power due to the steric effect caused by its cyclic structure.T he most extreme case and the protagonist in this study,1,4-DX, possesses an ultra-low e even lower than that of benzene (e = 2.3), which are both typical non-polar solvents. Theoretically,1,4-DX should have an extremely weak solvating power. Actually,l ithium hexafluorophosphate (LiPF 6 )i s almost insoluble in 1,4-DX,and lithium bis(trifluoromethylsulfonyl)imide (LiTFSI) also exhibits avery limited solubility (< 0.3 M) although these two salts possess ah igh solubility and are commonly adopted in battery research and industrial applications (Supporting Information, Figure S1a,b). Interestingly,l ithium bis(fluorosulfonyl)imide (LiFSI) is found to be the only soluble Li salt in 1,4-DX, exhibiting am aximum solubility up to 2.0 M( Supporting Information, Figure S1c) and conceivably forming au nique solvation structure.M ore discussions on the solvating power of these solvents are given in the supporting information. To ensure that solvating power is the only controlled variable in this study,a ll electrolytes were formulated with 1.0 ML iFSI as Li salt and denoted as the name of their solvents.T he elaborately designed electrolyte series,n amely DME, 1,3-DX, and 1,4-DX, should represent ad ecreasing trend of solvating power and increasing trend of ion pair formation.

Solvation Structure
Thes olvation structures of the above-mentioned electrolytes were investigated by spectroscopic characterizations combined with first-principles calculations.R aman spectra were firstly obtained for three ether-based electrolytes (Figure 2a;S upporting Information, Figure S2). As shown in Figure 2a,T he S-N-S bending signal in FSI À anion can be classified into three distinctive bands:f ree anion (FA, 719.0 cm À1 ,n on-coordinated FSI À ), contact ion pair (CIP, 730.6 cm À1 ,one FSI À binding with one Li + ), and ion aggregate (AGG,742.3 cm À1 ,one FSI À binding with two or more Li + ). [18] Thes pecific portion of these three species were calculated from the peak area and listed in Figure 2b. [12] DME electrolyte contains 71.3 %o fF A, 28.7 %C IP,a nd no AGG, indicating that most anions are expelled from the primary solvation sheath due to the strong solvating power of DME. Thed issociation degree of LiFSI, a,i s7 1.3 %i nt his case. Accordingly,anew vibration band of DME solvent at 800-900 cm À1 arises (Supporting Information, Figure S2a), which signifies the abundant Li + -DME complexes.I n1 ,3-DX with less solvating power, the ratio of FA significantly reduces as the ratio of CIP and AGGi ncreases.T he additional band of 1,3-DX vibration (Supporting Information, Figure S2b 17 On uclei which is subsequently expressed by the upfield displacement of chemical shifts in NMR spectra. [19] Thec hemical shift of 17 On uclei in LiFSI molecules decreases in the order of DME > EC/EMC > 1,3-DX > 1,4-DX, indicating that the coordination strength between Li + and FSI À follows the reverse trend. Thea bove preliminary results show that aW SE is indeed constructed exactly as designed when LiFSI is dissolved in 1,4-DX. As ad irect correlation, when the solvating power of solvent reduces,solvents in the primary solvation sheath are gradually replaced by anions. To elucidate the origin of different solvation structures, first-principles calculations were further conducted to probe the molecular interactions between anions/solvents and Li + (Figure 2d-j;S upporting Information, Table S2), which are expressed in terms of binding energy.T he binding energy of Li + -solvent (E S )a nd Li + -anion (E A )c omplexes is primarily determined by two major factors:1)The chemical structure of the ligands.T ypically,c arbonyl Oe xhibits higher nucleophilicity than ethereal O; therefore,c arbonates usually have higher solvating power than ethers.M oreover,l igands with multiple coordination sites (also known as the chelating effect) exhibit stronger interaction with Li + than monodentate ligands.F or example,D ME and FSI À both have two O atoms to coordinate with Li + (Figure 2e and j), therefore exhibit larger binding energies (for example, E S = À1.43 eV for DME). Thec oulombic attraction between Li + and FSI À contributes to an even stronger interaction and thus larger E A compared to E S .2 )T he dielectric constant of the solution. Large dielectric constant of solvents weakens the Li + -anion and Li + -solvent interactions,w hich can be approximately described by classical physical models: [16] where e is the dielectric constant, q the charge of ion, m the dipole moment of dipole, r the distance between ion and ion or ion and the center of dipole,and q the dipole angle relative to the line r joining the ion and the center of the dipole.F or instance,the relatively low E S of EC (À0.67 eV,which seems contradictory to its high solvating power) is due to its large e (89.8), and the extremely high E A (À3.15 eV) in 1,4-DX is due to its small e (2.2) that inhibits salt dissociation. Interestingly,i ft he binding energy of Li + -EC is calculated in 1,4-DX environment (which practically means to add as mall amount of EC in 1,4-DX that does not change the solvent environment), the E S of Li + -EC (À1.38 eV) is significantly larger than the E S of Li + -1,4-DX (À1.13 eV). According to the above analyses,itisunreasonable to directly compare E S -E A in different electrolytes because it does not reveal direct information on the solvation structure of an electrolyte.Onthe other hand, analyzing the value of E S and E A in the same electrolyte environment affords fresh insights on the competitive coordination between anions and solvents with Li + .
Thedescriptor of E S -E A is further proposed to predict the actual solvation structure in different electrolytes.Alarger E S -E A indicates that the ion pair and aggregate are preferentially formed over Li + -solvent complexes,n amely that anions win the coordination competition over solvents.F igure 2d illustrates that the trend of E S -E A (DME < EC < 1,3-DX < 1,4-DX) is in perfect accordance with spectroscopic results,w hich strongly affirms the applicability of the descriptor of E S -E A .A sar ule of thumb,l arge numbers of Li + -solvent complexes and free anions are anticipated for E S -E A close to 0( such as in DME);i on pairs and aggregates prevail for extremely large E S -E A (> 2.0 eV,s uch as in 1,4-DX);Li + -solvent complex and ion pair jointly constitute the solvation structure for intermediate E S -E A (0.5-1.5 eV,s uch as in 1,3-DX). Themost striking significance to emerge from

Angewandte Chemie
Research Articles E S -E A is that it serves as aquantitative indicator to predict to what extent do anions intrude the primary solvation sheath of Li + .O ur theory reveals the underlying mechanism that different solvation structures originate from the competitive coordination between solvents and anions towards athermodynamically stable Li + solvation sheath.

Li + Intercalation Behavior in Graphite
To explore the effect of different solvation structures on the interfacial chemistry of electrodes,g raphite electrode is chosen as at ouchstone because the reversible Li + intercalation in graphite is highly sensitive to the solvation structure of Li + in bulk electrolyte. [20] Figure 3a,b exhibits the charge/ discharge curves and cyclic voltammetry (CV) curves of graphite during the first cycle in different electrolytes.E Ci s strongly coordinated with Li + and reduced at about 0.8 Vvs. Li/Li + to form as table SEI in the EC/EMC electrolyte (Supporting Information, Figure S4), which is atypical case of solvent-derived interfacial chemistry.G raphite lithiation/ delithiation in EC/EMC is highly reversible,w ith three voltage plateaus between 0.05-0.25 Vrepresenting the different stages of Li-graphite intercalation compounds.U nlike carbonates,e thers have long been regarded as unstable against graphite electrode. [7] DME electrolyte causes severe co-intercalation at 0.4-1.0 Vthat undermines the structure of layered graphite (Supporting Information, Figure S5) so that reversible lithiation cannot be achieved. This is because DME are also strongly coordinated with Li + but is unable to form stable SEI that prevents co-intercalation. This phenomenon is common for ether-based electrolytes,as1,3-DX also exhibits slight co-intercalation, sluggish lithiation kinetics,a nd an initial coulombic efficiency( ICE) of merely 68.03 %. Although the co-intercalation of 1,3-DX is milder than that of DME because of the weaker solvating power and higher degree of ion pair formation ( Supporting Information, Figure S5), the reversibility of graphite lithiation is still unsatisfactory.S urprisingly,1 ,4-DX electrolyte exhibits ah igh reversible capacity of 360.5 mAh g À1 and faster lithiation/ delithiation kinetics even exceeding the commercial EC/EMC electrolyte.The ICE of 1,4-DX (86.7 %) is close to that of EC/ EMC (88.94 %), implying that 1,4-DX electrolyte leads to as table SEI formation. To the best of our knowledge,t his is the first report of highly reversible lithiation of graphite in neat ether electrolytes without applying superconcentration or any additives.T his unexpected phenomenon is attributed to the unique solvation structure of the 1,4-DX electrolyte, where the prevailing ion pairs and aggregates leads to preferential reduction of anions (at about 1.0 V; Figure 3b; Supporting Information, Figure S4) to form an anion-derived SEI. To verify this postulation, adetailed investigation on the SEI of graphite is requested.

Interfacial Chemistry and Kinetics
Thes urface passivation film on graphite (SEI) is the key to reversible Li + intercalation. [21] TheL i + intercalation behavior of graphite indicates that only two electrolytes can form stable SEI and enable reversible lithiation:t he commercial EC/EMC electrolyte and the 1,4-DX electrolyte (also denoted as WSE). XPS is conducted to characterize the composition and structure of SEI on graphite in these two electrolytes and study the SEI formation mechanisms.T he deconvolution of C1sspectra reveals four peaks (Figure 4a), representing C À C(284.8 eV,from graphite), C À O(286.6 eV), C=O( 288.8 eV), and CÀF( 290.1 eV,f rom PVDF binder). Thep eak intensities of CÀOa nd C=Oi nW SE are significantly lower than that of EC/EMC,i ndicating as uppressed solvent decomposition in WSE compared to the EC decomposition in EC/EMC that generates abundant organic species in SEI.
Thea tomic concentration at different etching depths reveals the structure of SEI (Figure 4b,c). Theetching depth corresponds to the standard thermal oxidation of SiO 2 samples.F or EC/EMC,t he Ca nd Oc oncentrations sharply decrease from 0t o1 0nma st he Fc ontent increases,t hen  stabilize from 10 to 20 nm. This result is in accordance with the classic two-layer SEI model, in which the outer layer mainly consists of organic species at higher oxidation state (mainly Li alkyl carbonates) and the inner layer consists of various inorganic compounds (LiF,L i 2 CO 3 ,a nd N,S-containing species as shown in Figure 4d)that are more stable against reduction. [22] Therefore,SEI is mainly solvent-derived in EC/ EMC electrolytes,accompanied by partial anion reduction. In contrast, the atomic contents in WSE-derived SEI are almost constant from 0t o2 0nmw ith lower Cc ontent and more inorganic ingredients,i ndicating that the SEI is highly homogeneous along its depth and inorganic in nature.A closer examination reveals that the Oc ontent in WSEderived SEI is roughly twice of the Fcontent, which is exactly the stoichiometric ratio in FSI À .Therefore,inWSE the SEI is generated mainly through anion reduction that generates abundant inorganic species such as LiF,Li 2 O, Li 3 N, Li sulfide, and Li oxysulfide (Figure 4d), and so on. TheX PS results confirm that EC/EMC features solvent-derived interfacial chemistry and WSE features anion-derived interfacial chemistry.
Te mperature-dependent electrochemical impedance spectroscopy (EIS) were employed to determine the kinetics of different interfacial processes.Three-electrode setup using aL i@Cu reference electrode was implemented to accurately measure the impedance signal of graphite electrode without the complication of the Li counter electrode (Figure 5a; Supporting Information, Figure S6a-c). Based on aw ellestablished theory,t he semicircle at mid-frequencyr egion in the Nyquist plot represents the desolvation step of Li + (known as the charge-transfer impedance) and the semicircle at high-frequencyr egion represents Li + transport through SEI (Figure 5b). [23] TheE IS spectra were fitted according to the classic Arrhenius law and activation energies of each interfacial process are obtained (Figure 5b,c). WSE shows as lightly reduced Li + desolvation energy barrier (E a,ct = 48.2 kJ mol À1 )c ompared to EC/EMC (E a,ct = 54.7 kJ mol À1 ). Since the Li + -solvent interaction is much weaker in WSE than in EC/EMC as previously demonstrated, such reduction of Li + desolvation energy barrier may seem insignificant. However,s ince the Li + -FSI À interaction in WSE is much stronger than in EC/EMC,d esolvation is mainly contributed by the dissociation of ion pairs and aggregates which is also energy-consuming. [24] Most importantly,the activation energy for Li + transport through SEI in WSE (E a, SEI = 26.6 kJ mol À1 ) is significantly lower than in EC/EMC (E a, SEI = 44.7 kJ mol À1 ). This is because the inorganic species dispersed in anionderived SEI creates abundant phase boundaries and vacancies for rapid Li + diffusion, which prominently reduce the energy barrier.I nt he solvent-derived SEI, Li + undergoes pore diffusion in the outer layer, which requires ah igher activation energy and renders limited kinetics.T he kinetics analysis implies that the unique anion-derived interphase may potentially enable fast-charging characteristic.

Electrochemical Performance
To understand the role of different interfacial chemistries in the electrochemical performance of electrodes,b oth rate and cycling tests were conducted for graphite electrodes in EC/EMC and WSE electrolytes.The WSE exhibits aremarkable fast-charging performance even far exceeding the commercial EC/EMC electrolyte (Figure 6a,b), retaining 54 %o fi ts capacity even at ad emanding rate of 4.0 C. The charging process can be divided into 4s teps:1)Li + diffusion in the bulk electrolyte,e specially in the micropores of the Figure 5. Kinetics of interfacial processes at the graphite/electrolyte interface measured by EIS using a3-electrode setup. a) Cell configuration of 3-electrode setup for EIS measurements. b) Temperaturedependent EIS curves of cells containing EC/EMC and WSE. c) Arrhenius behavior of the resistance corresponding to Li + desolvation. d) Arrheniusb ehavioro fthe resistance corresponding to Li + transport through SEI. Figure 6. Electrochemical performance of graphite electrode in different electrolytes. a) Specific capacity of graphite electrodes in EC/EMC and WSE under various charge and discharge rates. b) The corresponding charge and discharge curves at selected rates. c) Long term cycling performance of graphite electrode in EC/EMC, WSE and WSE + 2% EC electrolyte at 1.0 Ccharge and discharge rate. Long term cycling tests were conducted after the rate tests without interval. graphite electrode;2 )Li + desolvation at the electrolyte/ electrode interfaces;3 )Li + transport through SEI;a nd 4) Li diffusion within graphite galleries.Itisobvious that process 4 is identical in EC/EMC and WSE. Because the ionic conductivity of WSE is nearly one-magnitude lower than that of EC/EMC (Supporting Information, Figure S7) due to the lack of Li salt dissociation, step 1cannot be the reason for its outstanding rate performance.C onsequently,t he exceptional rate performance of WSE is attributed to the accelerated Li + desolvation step induced by its unique solvation structure,a nd rapid Li + diffusion through the anion-derived SEI. This conclusion is supported by the interfacial kinetics analysis.I no ther words,e ven the conductivity of WSE is substantially smaller, its anion-derived interfacial chemistry induces rapid kinetics of Li + migration across interfaces and the impressive fast-charging capability.
Long-term cycling of graphite electrodes at 1Crate were carried out straight after the rate tests to examine the SEI stability (Figure 6c;Supporting Information, Figures S8a and  S8b). EC/EMC exhibits a7 8% capacity retention after 300 cycles,w hich is acceptable for routine EC-based electrolytes without any additive.H owever,W SE renders ar apid capacity decay during long term cycling and only retains 34 % of its initial capacity after 300 cycles.T his phenomenon is attributed to the fragile nature of inorganic-rich SEI derived from anion decomposition. Thea nion-derived SEI is broken under high stress due to the volume fluctuation of graphite during cycling. This leads to the repeated cracking and repair of SEI that gradually increase its thickness overtime,w hich finally result in agrowing resistance and capacity fade.Onthe contrary,t he organic SEI layer in EC/EMC possesses higher elasticity and is more resilient to mechanical deformations, therefore offers abetter protection of the graphite electrode.
Interestingly,t he fragility of anion-derived SEI and the poor cycling performance of WSE can be overcome by exploiting the competitive coordination between solvents and anions.Aspreviously shown, the binding energy between Li + and EC is large in 1,4-DX environment (À1.38 eV; Supporting Information, Figure S4). Simply by adding 2.0 wt %E C into WSE (denoted as WSE + 2% EC), some EC molecules will coordinate with Li + and replace asmall part of ion pairs and aggregates.C onsequently,t hese EC molecules are reduced on graphite electrodes to produce as mall number of organic compounds that infiltrate into the inorganic compounds,which serves as the glue to enhance the stability of SEI. As ar esult, WSE + 2% EC enables ultra-stable cycling of graphite electrode with 92 %c apacity retention after 500 cycles (Figure 6c;S upporting Information, Figure S8c), and retains as atisfactory rate performance (Supporting Information, Figure S9). If the cell was directly cycled at 1Cwithout the rate test, alonger life exceeding 840 cycles can be obtained with 80 %c apacity retention (Supporting Information, Figure S10). Such superior cycling performance is very rare for graphite electrodes in ether-based electrolytes, further demonstrating the huge potential of anion-derived interfacial chemistry achieved by solvating power regulation.
It is important to note that the aim of this work is not to demonstrate ap ractical electrolyte suitable for commercial LIBs,typically with high-voltage cathodes,high areal loading and wide-temperature range.F or instance,t he ether-based WSE is incompatible with high-voltage batteries,and the high melting point of 1,4-DX (11.8 8 8C) rules out low-temperature operation. Instead, an ew concept in electrolyte is proposed, in which the methodology and underlying mechanism may inspire future electrolyte innovation towards more practical applications.T herefore,although electrochemical tests under practical conditions (such as full cell/pouch cell) are not provided herein, these preliminary results suggest that the concept of WSE bears huge potential for next-generation electrolyte systems for advanced LIBs.F uture study may discover new solvents and lithium salts with better properties such as high anodic stability,w ide liquid range,i nhibition of Al dissolution and so on, for constructing WSEs with the potential to replace commercial EC-based electrolytes.

Conclusion
Acompletely new route towards anion-derived interfacial chemistry in LIBs is developed. Unlike superconcentrated electrolytes,t he essence of this methodology is constructing aw eakly solvating electrolyte by using an on-polar but saltdissolving solvent. WSE exhibits ap eculiar solvation structure where ion pairs and aggregates prevail under al ow salt concentration of 1.0 M. As ar esult, the anion-derived SEI exhibits superior interfacial charge transport kinetics and high stability,e nabling fast-charging and long-term cycling of graphite electrodes.F irst-principles calculations unravel the fundamental rationale that the competitive coordination between solvents and anions controls the transition from solvent-derived interfacial chemistry to anion-derived interfacial chemistry.F urthermore,asemi-empirical descriptor was put forward to predict the actual solvation structure in electrolytes.T his work constitutes the first step of an undiscovered way towards anion-derived interfacial chemistry,inwhich the methodology serves as an emerging principle for coming studies on precise electrolyte engineering towards next-generation energy storage devices.