Designing Electrolytes With Controlled Solvation Structure for Fast‐Charging Lithium‐Ion Batteries

Recharging battery‐powered electric vehicles (EVs) in a similar timeframe as those used for refueling gas‐powered internal combustion vehicles is highly desirable for rapid penetration of the EV market. It is well known that the electrolyte in a battery plays a critical role in fast‐charging capability of the battery because it determines the rate of ion transport together with its derived electrode/electrolyte interphases on both cathode and anode of the battery. In this study, the effects of contents of salt, coordinating solvent, and noncoordinating diluent on salt dissociation degree and electrolyte ionic conductivity are investigated, and a controlled solvation structure electrolyte is developed to improve the lithium ion mobility and conductivity in the electrolyte and to enhance the kinetics and stability of the electrode/electrolyte interphases in the battery. This electrolyte enables fast‐charging capability of high energy density lithium‐ion batteries (LIBs) at up to 5 C rate (12‐min charging), which significantly outperforms the state‐of‐the‐art electrolyte. The controlled solvation structure sheds light on the future electrolyte design for fast‐charging LIBs.


Introduction
[6][7] One approach to achieve high-energy density LIBs is the utilization of high-voltage layered oxide cathode materials such as nickel-rich LiNi x Mn y Co 1−x−y O 2 (NMC) with a Ni content of ≥80% (NMC811) for its increased specific capacity.[13] Therefore, it is critical to design advanced electrolytes to have high-voltage stability and oxidative protection for Ni-rich NMC cathodes to mitigate the instability challenges and also enable the practical application of Ni-rich NMCs based fast-charging LIBs.
In recent years, there have been significant efforts to develop novel electrolytes to improve stability for high-voltage cathodes as well as graphite (Gr) or Li metal anodes.[20][21] However, due to the unique compact solvation nature of LHCEs, the salt dissociation degree in LHCEs was still very low as it was in HCEs, which drastically impaired the ion transport properties of LHCEs with low carrier mobility.The reduced ionic characteristics of LHCE prevented it from practical high current density applications.Baird et al. reported a systematic study of formulating locally superconcentrated electrolytes (LSCEs) and investigated the effect that the coordination environment and ionic properties of their LSCEs had on the fast-charging capabilities in Li||NMC622 cells to facilitate 80% SOC within 5-15 min with good cycle life. [21]Nevertheless, the design principle for these fast-charging electrolytes was still not clear.
Herein, we developed a series of electrolytes based on lithium bis(fluorosulfonylimide) (LiFSI) as the conducting salt, dimethyl carbonate (DMC) as the solvating solvent, ethylene carbonate (EC) as the additive, and 1,1,2,2-tetrafluoroethyl-2,2,3,3-tetrafluoropropyl ether (TTE) as the diluent for fastcharging capabilities by tuning the molar ratios of the solvating solvent to the diluent in the electrolytes.By carefully evaluating the solvation structure, coordination environment, viscosity, and ionic conductivity, a designing rule for electrolytes with high ionic conductivity was disclosed.Along with the fast-charging electrochemical performance in LIBs with NMC811 cathode and Gr anode, an electrolyte with a controlled solvation structure stood out.It had a high salt dissociation degree to enable a high ionic conductivity compared to regular LHCEs.Meanwhile, it had minimum free solvents to be decomposed on the cathode and anode in comparison with the SOA electrolyte.As a result, this controlled solvation electrolyte (CSE) exhibited excellent compatibility with both NMC811 cathode and Gr anode due to the formation of a thin and robust passivation layer that enabled enhanced ion mobility and prevented structural degradation and cracking of the cathode particles.This investigation reveals that the unique coordination structure of the CSE is critical to improve the ionic carrier mobility of Li + and sustain a fast-charging rate of 4 C over 300 cycles.

Determination of Factors Affecting Electrolyte Ionic Conductivity
The kinetics of the battery charge/discharge reactions is mainly determined by ion mobility through the bulk electrolyte and the electrode/electrolyte interphases.In our previous work, an advanced LHCE with the formulation of LiFSI-2DMC-0.2EC-3TTE(by mol) was developed for high voltage Gr||NMC811 batteries. [22]he cells employing this LHCE demonstrated significantly improved battery performances in a wide-temperature range, in terms of cycling stability at room temperature and high temperature, discharge rate capability, and low-temperature discharge performance compared to the SOA electrolyte by generating a highly ionic conductive, robust, and protective solid electrolyte interphase (SEI) on Gr anode and cathode electrolyte interphase (CEI) on NMC811 cathode. [22]Nevertheless, its fast-charging capability above 4 C (less than 15 min' charge) was hindered by its low ionic conductivity of 1.07 mS cm −1 due to the compact solvation nature of the LHCE, where LiFSI was surrounded by very limited solvating solvent molecules, making it difficult to dissociate; so, it had a very low dissociation degree of 7.4%. [23] facilitate the fast-charging capability of the LIBs, high ionic conductivity of the electrolyte is essential, especially in practical applications where high loading electrodes and lean amount of electrolyte are used to obtain high energy density of the batteries.As the solvating solvent can enable high salt dissociation degree and increase the electrolyte conductivity, while the HFE diluent behaves opposite, the new electrolytes are designed to have more solvating solvent but less HFE diluent so as to improve the ionic conductivities of the electrolytes.Then, a series of electrolytes represented by LiFSI-xDMC-TTE (by mol.) are designed and investigated in this work to achieve the comparable ionic conductivity with the SOA electrolyte, where x is varied from 2.2 to 5 molar equivalents.LiFSI-9DMC (i.e., 1 m LiFSI in DMC) without any non-coordinating TTE is also used as a reference for these electrolytes to understand the intrinsic properties of the electrolytes.
Figure 1 summarizes the contributions of the different factors to the electrolyte ionic conductivity.With the increase of the Li + coordinating solvent DMC from 2.2 to 3 and 4, the LiFSI dissociation degree is significantly improved from 4.2% to 27.1% and 30.74%, respectively, while the dissociation degree has minimal variation when further increasing DMC from 4 to 5, and even to 9. It clearly shows that 4DMC is sufficient to dissociate the LiFSI to the comparable level of a dilute electrolyte of 9DMC, indicating the dissociation degree of the LiFSI is the key factor determining the ionic conductivity when the molar ratio of DMC to LiFSI is below 4. As the overall LiFSI concentration decreases with the increase of DMC, the amount of the absolute dissociated LiFSI, which is calculated by multiplying the LiFSI content and the LiFSI dissociation degree, is almost the same in 3DMC-TTE and 4DMC-TTE, and it decreases in 5DMC-TTE and 9DMC because of the continuous reduction of salt concentration.However, the ionic conductivity is continuously increased in 5DMC-TTE and 9DMC.This time, another important factor, viscosity, starts to play a dominant role in the ionic conductivity, and a lower viscosity means a higher conductivity when the sufficiently dissociated LiFSI exists in the electrolyte with the DMC to LiFSI ratio above 4.
Detailed numbers of electrolyte compositions, LiFSI concentrations, dissociated degrees, absolute dissolved Li + values, viscosities, and ionic conductivities are given in Table S1, Supporting Information.Pulsed-field gradient nuclear magnetic resonance (PFG-NMR) measurements are performed to determine the self-diffusion coefficients of Li + and FSI − ; and then, the dissociation degree of the Li salt in electrolytes is calculated using the self-diffusion coefficient and measured ionic conductivity. [24]n addition to DMC variation in the electrolytes, TTE variations are also included in Table S1, Supporting Information, which further evidences the dominant role of dissociated LiFSI and viscosity in the electrolytes.
To understand the effects of increasing molar content of DMC on the solvation structure of the electrolytes, classical molecular dynamic (CMD) simulations were performed.The simulated snapshots of the electrolytes are represented in Figure 2a-d together with their corresponding radial distribution function (RDF) curves of the O atoms in anions and solvents to Li + ion, as shown in Figure 2e-h.Within the solvation cluster of the LiFSI salt in all the four electrolyte systems, the Li + ions were found to be coordinated with both O atoms in the FSI − anion and the DMC solvent, which are consistent with Raman spectra (Figure S1, Supporting Information).Li + ions were predominantly coordinated with the O atoms in DMC at the prominent peak at 1.95 Å when the ratio of DMC was greater than 3.There were free DMC molecules in the solution, which were both validated by simulation and Raman spectra.In all cases, Li + was barely coordinated with TTE due to its non-coordinating nature.The number of LiFSI cluster increased and the cluster size decreased with more DMC, but the Li + -FSI − cluster still could be observed in LiFSI:DMC:TTE = 1:5:1 system.This was in agreement with the Raman spectrum.In addition, as the molar concentration of DMC increased, the peak intensity (coordination number) of Li-(O-FSI) significantly decreased (Figure 2e,f) signifying that Li + ion became preferentially coordinated with the DMC.A more DMC-rich coordinated environment had a faster ionic kinetics that effectively facilitated Li + transport under high current density. [18]Therefore, a DMC-LiFSI molar ratio above 4 is preferable for a highly conductive electrolyte.

Electrochemical Performance of Gr||NMC811 Cells
In addition to the high ionic conductivity of the bulk electrolyte, highly conductive SEI and CEI have the equal importance to the fast-charging capability to the batteries.Whichever has the lowest ion transport determines the overall kinetics of the battery.As reported in our previous work, LHCEs without free DMC lead to anion-derived SEI and CEI on the Gr and NMC811, which are thin yet robust to protect the batteries from continuous electrolyte/electrode side reactions. [21]In addition, the coordinated DMC has higher oxidative stability at high voltage compared to the free DMC due to the reduced electron density on the O in DMC when it coordinates to Li + , as shown in linear sweep voltammetry (LSV) testing (Figure S2, Supporting Information).Therefore, in order to minimize the potential side reactions of the LiFSI-xDMC-TTE electrolytes on the electrode surfaces, the DMC amount needs to be controlled to a highly coordinated level  so that a minimum free DMC exists, while LiFSI is at highly dissociated level to ensure a high ionic conductivity.Based on the LiFSI-xDMC-TTE solvation structures obtained in Figure 2 and literature results of Li + -DMC solvation, [25,26] the full coordination number of DMC to Li + is up to 4-5.15.Therefore, the x in LiFSI-xDMC-TTE is limited to 5 in this study to maintain robust SEI and CEI while facilitating improved ion separation that can promote high ion mobility for fast-charging capabilities.
Charge rate capability testing of the Gr|| NMC811 cells was performed on these LiFSI-xDMC-TTE electrolytes to narrow down the most capable electrolytes for further electrochemical analyses (Figure 3a).The charge rates ranged from C/10 to 5 C while the discharge rate was constant at C/5.Out of the four new electrolytes, LiFSI-5DMC-TTE showed superior charge rate capabilities, delivering up to 155 mAh g −1 at 4 C and 145 mAh g −1 at 5 C, which were 84.7% and 79.2% of its capacity at C/5 (183 mAh g −1 ), respectively.When the charge rate returned to C/5, it also retained high discharge capacity and stable cycling profile; and thus, LiFSI-5DMC-TTE was selected for further performance evaluation.
The selected electrolyte was further modified by substituting 0.2 mol DMC with an equivalent 0.2 mol EC for improving the robust SEI forming property and the electrochemical cycling stability, and compared to a conventional SOA LiPF 6 /carbonates electrolyte (noted as SOA electrolyte). [22,27]he new EC-containing electrolyte, LiFSI-4.8DMC-0.2EC-TTE(by mol.), was named as controlled solvation electrolyte (CSE).Under the same testing protocol, this CSE showed even better rate performance of 165 mAh g −1 at 4 C and 160 mAh g −1 at 5 C, which are 89.2% and 86.5% of its capacity at C/5 (185 mAh g −1 ), respectively (Figure 3b).The cell with the SOA electrolyte initially had higher discharge capacities during the slower C-rates (<1 C), for example, 193 mAh g −1 at C/10 and 188 mAh g −1 at C/5, but it had a more significant drop in capacity starting at 1 C and gradually having inferior rate capability to the CSE with each incremental increase in C-rate.The capacity became 158 mAh g −1 at 4 C and 153 mAh g −1 at 5 C, which were 7 mAh g −1 less than the CSE at related C-rates.When the charge rate was returned to C/5, the discharge capacities of cells with the two electrolytes were similar, but the SOA electrolyte had the bigger drop in capacity from its initial C/5 rate.In addition, the charge/discharge voltage profiles at increasing C-rate from 1 C to 5 C showed the CSE had lower degrees of polarization than the SOA electrolyte during charge/discharge (Figure 3c), which is beneficial for the energy output of the cell at fast-charging process.
These results indicate that the cell using the CSE has a faster overall ion mobility than the cell using the SOA electrolyte, which is possibly because it has a more conductive electrode/electrolyte interphase on both NMC811 and Gr electrodes due to the CSE's more optimized solvation structure that has a minimal amount of free DMC that reduces the DMC decomposition.Therefore, the electrode/electrolyte interphases could be salt derived in CSE as in LHCEs and enhance the ion transport kinetics.
The long-term fast-charging cycling performance was evaluated at 4 C charging rate and C/3 discharging rate to simulate real world practical fast-charging and slow-discharging scenarios of EVs.As shown in Figure 3d, the CSE had a significantly higher discharge capacity and capacity retention over the 300 cycles compared to the SOA electrolyte although it had similar discharge capacities to the SOA electrolyte during the initial three formation cycles at C/10 and C/5 rates.When charged at 4 C, the CSE had a high initial discharge capacity of 179 mAh g −1 while the SOA electrolyte had a low initial capacity of 130 mAh g −1 .The capacity retention of the cell using the CSE was 69% after 300 cycles at 4 C charge and C/5 discharge.Although the SOA electrolyte had a higher capacity retention of 92.7% to initial 4 C capacity and 82.3% to its highest capacity after 300 cycles at 4 C charge and C/3 discharge, it continually led to lower discharge capacity than the CSE over the whole 300 cycles, reaching 120.3 mAh g −1 for the SOA electrolyte and 140.2 mAh g −1 for the CSE at the 300 th cycle.
Electrochemical impedance spectroscopy (EIS) was carried out to monitor the resistance of the electrode/electrolyte interface for the Gr||NMC811 cells cycled with the two electrolytes after 100, 200, and 300 cycles (Figure 3e-g).The cell with the CSE showed a smaller resistance for the associated resistance that is related to the electrode-electrolyte interphase (R SEI ) and the charge-transfer process (R ct ) and remained fairly constant throughout the 300 cycles, compared to the SOA electrolyte cell which had a dramatic increase in its R ct between 200 and 300 cycles (Table 1).The increase in the charge-transfer resistance of the SOA electrode was indicative of the electrolyte's instability.The overall lower, and more stable, protective interface for the CSE cells can provide a partial explanation on the improved electro-chemical performance and performance observed over the 300 fast-charge cycles.

NMC811 Structure and Interphase Morphology and Composition After 300 Cycles
To help understand the origin of the improved fast-charging cycling stability of the CSE over the SOA electrolyte in Gr||NMC811 cells, the bulk of the cathodes was characterized by plasma focused ion beam scanning electron microscopy (PFIB-SEM) imaging, and the surface layer was characterized by high resolution transmission electron microscopy (HRTEM) imaging, X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) analyses after 300 cycles at 4 C charging and C/3 discharging.As shown by the PFIB-SEM images in Figure 4a-c, the cross-section views of the NMC811 particles showed more severe cracking in the bulk of the secondary particles cycled in the SOA electrolyte while well-maintained structural integrity was observed for those cycled in the CSE.The observed cracking in NMC811 particles from the SOA electrolyte can be attributed to shifting in the crystalline parameters of the cycled NMC811 indicated in the peak shifts of (003) and ( 108)/(110) in the XRD patterns (Figure 4h,i). [28,29]The observed peak shifts can be indicative of anisotropic volume changes within the NMC particles due to Li + loss from the crystalline layers of the NMC811 particles cycled in the SOA electrolyte causing strain during the charge/discharge of the cell and leading to the observed microcracks in the SEM images (Figure 4b). [30,31]On the contrary, the NMC811 cathode cycled with the CSE had no peak shift observed.
In addition to the bulk structure of the NMC811 particles, the CEI on its surface played an important role in the rate capability and cycling stability.The composition of the electrolyte was a dominant factor for enabling sufficient protection and charge transfer kinetics of the CEI, inhibiting undesired side reactions and suppressing undesired degradation.As shown in Figure 4df, the surface of the pristine NMC811 particle was clean, but the cathode CEI from the SOA electrolyte was found to be inhomogeneous, with a thick CEI built up in one area of the particle, and very little, if any, CEI in other areas of the particle (Figure 4e) which would cause poor protection of the cathode particle and lead to severe particle damage and fast capacity fading from continuous side-reactions with the electrolyte.On the other hand, the CSE led to a more homogeneous CEI layer with a thickness around 5 nm on the NMC811 particle (Figure 4f).34] The composition of the CEI on the NMC811 electrode surface after 300 fast-charging cycles was investigated with XPS analysis and is summarized in Figure 4j-l; Figure S3, Supporting Information.While the C1s spectra were similar in the SOA electrolyte and CSE, higher amount of LiF was found in the CEI from the CSE than that from the SOA electrolyte.In addition, the observed S─O x and N─O x peaks in the O 1s (Figure 4k) and S 2p spectra (Figure S4, Supporting Information) on the CEI from the CSE indicate the decomposition of the LiFSI salt, which has been previously reported to be beneficial to the enhanced robustness and more uniform structure of the CEI layer and the improved cathode cycling stability. [14,19,35,36]The improved CEI from the CSE could more effectively protect the NMC811 particles, as indicated by the minimum M─O peak in the Figure 4k for the CSE while the M─O peak still existed in the cycled NMC811 electrode in the SOA electrolyte suggesting that the represented NMC811 particle was not fully covered by CEI in the SOA electrolyte.This result is consistent with the CEI morphologies, as shown in Figure 4e,f.The insufficient CEI protection in the SOA electrolyte resulted in continuous electrolyte/electrode side reactions, which could impede Li + migration of the NMC particles and cause a charge discrepancy, especially during high current density, such that that the de-lithiation potentially leads to cracking within the bulk of the NMC particles from strain of inhomogeneous volume fluctuations. [32]The insufficient CEI structure observed on the NMC811 surface, along with the cracking within the bulk of the NMC811 particles, could be attributed to the poor performance and the quick capacity fading in the SOA electrolyte during the fast charge cycling.

Gr Structure and Interphase Morphology and Composition After 300 Cycles
Analogous to the characterization performed for the NMC811 CEI layers, HRTEM and XPS analyses of the SEIs on cycled Gr anodes was investigated and the results are summarized in Figure 5. Similar to the cathode side, a homogenous SEI was obtained on Gr from the CSE with a thickness of 2.5 nm while the SEI from the SOA electrolyte on Gr was much thicker and more inhomogeneous.When analyzing the XPS spectra (Figure 5d-f; Figure S5, Supporting Information), the SEI from the SOA electrolyte had a higher content of C containing species, indicating a more organic-rich SEI layer derived from continuous decomposition of the solvent (Figure 5d).Contrarily, the SEI from the CSE was found to have more inorganic-rich species, which were derived from the decomposition of the FSI − anion (Figure 5e).Akin to what was found in the CEI results discussed above, the LiFSIderived SEI in the CSE was beneficial to stabilizing the Gr anode surface and promoting the reversible cycling with the Gr anode even though it had lower LiF content than solvent-derived SEI in the SOA electrolyte (Figure 5f).In addition, this more inorganic passivation layer more effectively suppressed further solvent decomposition and enabled highly efficient reversible Li + transport to and from the Gr anode during the fast charging and slow discharging cycling.

Conclusion
In this work, a controlled solvating electrolyte (LiFSI-4.8DMC-0.2EC-TTE) was developed for fast-charging capability and longterm cycling stability for high-voltage Gr||NMC811 cells.With a high dissociation degree of LiFSI in the CSE and a low viscosity, the CSE achieved a high ionic conductivity that is comparable with the SOA electrolyte.In addition, minimal free DMC in the CSE enabled an anion FSI-derived SEI similar to those typically found in the LHCEs, which facilitated a more robust and thinner SEI on the Gr anode and CEI on the NMC811 cathode with improved Li + charge transfer kinetics.Taking advantage of both high ionic conductivity of the bulk electrolyte and the enhanced ion transfer kinetics from the SEI and CEI layers, the CSE enabled an excellent fast-charging capability with high specific capacity of 165 mAh g −1 at 4 C (15 min charging) and 160 mAh g −1 at 5 C (10 min charging), which were 89.2% and 86.5% of its capacity at C/5 (185 mAh g −1 ), respectively.Moreover, the enhanced passivation layer on the cathode surface formed in the CSE effectively suppressed continuous electrolyte oxidation decomposition reactions and protected the Ni-rich NMC811 cathode particles from the structural degradation and cracking within the bulk of the particles.Meanwhile, the inorganic enriched SEI layer protected the Gr anode from continuous SEI growth.Therefore, the Gr||NMC811 cells with the CSE exhibited an improved stability in long-term performance over 300 cycles and maintained a highcapacity retention at a fast-charge rate.The findings of this work reveal the importance of the solvation environment and disclose a principle rule of designing advanced electrolytes to enable the practicality of fast-charging and long-term stable cycling of highvoltage LIBs with Ni-rich NMC cathodes.

Experimental Section
Electrolyte and Electrode Preparation: A conventional LiPF 6 -organic carbonate electrolyte, 1.0 m LiPF 6 in EC-EMC (3:7 by weight) + 2 wt% VC was employed as the benchmark, SOA, electrolyte.The CSEs were prepared by dissolving the LiFSI salt in DMC (as the solvating solvent), EC (as an additive), and TTE (as the diluent) in the designed molar equivalents in an argon (Ar)-filled glovebox (MBraun, H 2 O < 0.1 ppm, O 2 < 0.1 ppm).Laminates of Gr and NMC811 electrodes were obtained from the Cell Analysis, Modeling, and Prototyping (CAMP) Facility at Argonne National Laboratory (ANL) and their corresponding areal capacities were 1.8 and 1.5 mAh cm −2 , respectively.Electrode disks of Gr (16.0 mm in diameter) and NMC811 (12.7 mm in diameter) were punched, dried at 110 °C under vacuum for at least 12 h, and subsequently transferred into the glovebox.
Electrochemical Tests: Electrolyte conductivities were measured on a Bio-Logic MCS 10 fully integrated multichannel conductivity spectroscopy in the temperature range of −10 °C to 60 °C.CR2032 coin cell kits were ordered from MTI Corporation.Each coin cell was assembled with a piece of NMC811 disk, a piece of polyethylene separator (20 μm thick, Asahi Hi-Pore, Japan), a piece of Gr disk, and 75 μL of the electrolyte.To avoid the anodic corrosion of stainless steel in electrolytes at high voltages, the aluminum (Al)-clad positive case was employed and an additional Al foil of 19.0 mm diameter was placed in between the positive cathode disk and the Al-clad positive case during the cell assembly.The cells performed two formation cycles at C/20 with a cutoff voltage range of 2.5-4.4V, followed by rate capability and cycling tests in the same voltage range, where 1 C was 200 mAh g −1 on the basis of weight of the NMC811 active material.CSE (c,f).g-i) XRD patterns of NMC811 particles before and after 300 cycles of full spectra (g), (003) peak (h), and (108)/(110) peaks (i).j-l) XPS analysis of NMC811 cathode surfaces before and after 300 cycles.C 1s (j), O 1s (k), and F 1s spectra (l) for pristine cathode and cathodes after 300 cycles in SOA and CSEs (bottom to top, respectively).Characterizations: For postmortem analyses, including SEM, TEM, and XPS measurements, the cycled cells were carefully disassembled inside the glovebox to collect the Gr anodes and NMC811 cathodes.These cycled electrodes were rinsed with pure anhydrous EMC solvent to remove residual electrolyte, dried, and then sealed in the glovebox before being transferred for characterizations.SEM measurements were carried out on a Helios focused ion beam (FIB)-SEM at an accelerating voltage of 5 kV and a current of 86 pA.The cathode TEM samples were performed on an FEI Helios Dual Beam system.A randomly selected secondary particle of NMC811 was coated with an ≈2 μm Pt layer.The particle was then extracted along with the capping layers and welded to the TEM grid.The FIB processes were performed at 30, 5, and 2 kV to remove the damaged layers and polish the surface.Copper TEM grids (200 mesh, Tedpella, Inc.) were used to prepare the graphite samples by drop casting method in an Ar-filled glovebox.The as-prepared samples were characterized by a FEI Titan monochromated (scanning) transmission electron microscope ((S)TEM) equipped with a probe aberration corrector at 300 kV.XPS measurements were conducted on a Physical Electronics Quantera scanning X-ray microprobe with a focused monochromatic Al K X-ray (1486.7 eV) source for excitation and a pass energy of 69.0 eV for high-energyresolution spectra collection.For the electrolyte solvation structure, the Raman spectra were measured using a minimal power of 50 W from a 633 nm laser source and recorded using an inverted oprical microscope (Nikon Ti-E) coupled to a Raman spectrometer (LabRam HR, Horiba).
CMD Simulation: The CMD simulations were carried out with the GROningen MAchine for Chemical Simulations (GROMACS) simulation package.The simulation systems included 100 LiFSI molecules, 100 TTE molecules, and 220-500 DMC molecules, which followed the ratio in experiment.The salt, solvent, and diluent molecules were randomly inserted into the simulation box initially.The force field parameter of FSI anion was from a previous paper [37] and the other parameters were from OPLS-AA force field. [38]The F and H interaction parameters were tuned [39] to improve the excess volume and interfacial enthalpies of fluoroether.To include the electronic polarizability effect, the partial charges of ions were treated by electronic continuum correction (ECC) method. [40]The refractive index of DMC was obtained from a previous paper. [41]he steepest descent method was used to minimize the energy of the system.The systems were pre-equilibrated in an isothermal-isobaric (NPT) ensemble with 20 ns at 298 K and 1 bar.The temperature and pressure were controlled by V-rescale thermostat [42] and Berendsen barostat [43] with a time constant of 0.2 and 1 ps.Then, 10 ns production simulations were performed at 298K in canonical (NVT) ensemble.The temperature was controlled by the Nose-Hoover thermostat [44] with a time constant of 0.2 ps.The cutoff of the Lennard-Jones potential was 1.2 nm.The particle mesh Ewald method [45] with a Fourier spacing of 0.15 nm and a 1.2 nm real-space cutoff was used for calculating electrostatic interactions.Periodic boundary conditions were used in all three directions.The time step was 2 fs.The bonds between H and other atoms were constrained by the LINCS algorithm. [46]ulsed-Field Gradient Nuclear Magnetic Resonance (PFG-NMR): An Agilent 600 MHz spectrometer coupled with a 5 mm HX z-gradientprobe (Doty Scientific Inc. USA) was employed to perform the PFG-NMR experiments at 25 °C.Larmor frequencies of 599.8, 564.4,and 233.1 MHz were used for obtaining the PFG-echo profiles of 1 H, 19 F, and 7 Li, respectively.The self-diffusion coefficients of different electrolyte species were estimated from the PFG-echo profiles obtained as a function of the gradient strength using the Stejskal-Tanner equation S (g) = S (0) exp [−D (g) (Δ − ∕3)] (1) where S(g) and S(0) are PFG-echo intensities with the gradient strengths of g and 0, respectively., Δ, and  are the gyromagnetic ratios of the observed nuclear spin. and Δ are the gradient length and distance between the two pairs of bipolar gradients in the 13-interval stimulated echo pulse sequence (Dbppste in VNMRJ, Agilent).

Figure 1 .
Figure 1.Structure-property relationship plot of salt-solvent ratio, salt concentration, LiFSI dissociation, viscosity, and ionic conductivity of the LiFSI-xDMC-TTE electrolytes.The axes are shown in gradient height and each axis corresponds to the line chart with the same color.

Figure 3 .
Figure 3. Electrochemical performances of Gr||NMC811 cells.a,b) Rate capability of the cells using LiFSI-xDMC-TTE electrolytes (a) and SOA electrolyte and controlled solvation electrolyte (LiFSI-4.8DMC-0.2EC-TTE) at varying charge rates and C/5 discharge rate (b).c) Voltage profiles of the cells using SOA and CSE at different charge rates.d) Long-term cycling stability at 4 C charge rate and C/3 discharge rate.e-g) EIS spectra after 100 (e), 200 (f), and 300 cycles (g).All the cells were tested between 2.5 and 4.4 V, and 1 C rate corresponded to a current density of 1.5 mA cm −2 .

Figure 4 .
Figure 4. Characterization of NMC811 cathodes in pristine condition and after 300 fast-charging cycles in SOA and CSEs.a-c) Cross-sectional PFIB-SEM images of NMC811 particles and d-f) HR-TEM images of the surfaces of the cathode particles for pristine NMC811 (a,d) and cycled NMC811 in SOA electrolyte (b,e) andCSE (c,f).g-i) XRD patterns of NMC811 particles before and after 300 cycles of full spectra (g), (003) peak (h), and (108)/(110) peaks (i).j-l) XPS analysis of NMC811 cathode surfaces before and after 300 cycles.C 1s (j), O 1s (k), and F 1s spectra (l) for pristine cathode and cathodes after 300 cycles in SOA and CSEs (bottom to top, respectively).

Figure 5 .
Figure 5. Characterization of SEI morphology and composition with HRTEM and XPS analyses of Gr anodes before and after 300 cycles.a-c) SEI morphology, d) atomic ratio of the SEI composition of the SOA and CSEs, e) S 2p spectra of SEI in the CSE, and f) F 1s spectra for pristine, SOA, and CSEs (bottom to top, respectively).

Table 1 .
Summary of the fitted parameters for Nyquist plots in Figure3.