# Fileset

[version 7.7_final.docx](https://mdr.nims.go.jp/filesets/9b22d5fd-73f8-49cf-834a-486988cc9756/download)

## Creator

Yiyi Zheng, Tian Zhang, Pui-Kit Lee, Qiaohui Duan, Xin Li, [Shuyu Dong](https://orcid.org/0000-0002-6154-4533), Tian Tan, Yao Wang, [Denis Y.W. Yu](https://orcid.org/0000-0002-5883-7087)

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Boosting Li-ion transport for graphite electrodes with lithium bis(fluorosulfonyl)imide salt and methyl acetate additive for fast-charging Li-ion batteries](https://mdr.nims.go.jp/datasets/e21c6ba4-f9d3-4994-8aca-a7e55c5d924c)

## Fulltext

Boosting Li-ion transport for graphite electrodes with lithium bis(fluorosulfonyl)imide salt and methyl acetate additive for fast-charging Li-ion batteriesYiyi Zhenga,1, Tian Zhangb,1, Pui-Kit Leea, Qiaohui Duana, Xin Lic, Shuyu Donga, Tian Tana, Yao Wanga, Denis Y. W. Yua,d*a School of Energy and Environment, City University of Hong Kong, Tat Chee Ave, Kowloon, Hong Kong, China.b Department of Materials Science and Engineering and Center of Super-Diamond and Advanced Films, City University of Hong Kong, Tat Chee Ave, Kowloon, Hong Kong, China.c Department of Applied Biology and Chemical Technology, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, China.d Research Center for Energy and Environmental Materials (GREEN), National Institute for Materials Science, Tsukuba, Ibaraki 305-0044, Japan.*yu.denis@nims.go.jpAbstract: Lithium-ion battery (LIB) is now widely used in the world. However, its unsatisfactory fast-charging performance is limiting its applications. The fast-charging capability of LIB is highly affected by the surface composition of graphite electrode, which can be modified through electrolyte design. Herein we demonstrate that the combination of lithium bis(fluorosulfonyl) imide (LiFSI) salt with methyl acetate (MA) additive in carbonate electrolyte enables fast-charging Li-ion battery. Results from X-ray photoelectron spectroscopy and transmission electron microscopy indicate that the LiFSI-based electrolyte with MA forms a stable and less-resistive solid electrolyte interphase on graphite anode compared to electrolyte with LiPF6, facilitating the transport of Li-ion and achieving a superior fast-charging performance with a capacity of 230 mAh g-1 at 2 C. An electrolyte swapping experiment also indicates that the LiFSI salt forms a more stable SEI on graphite than LiPF6. This work provides a conducive electrolyte design for fast charging Li-ion battery with graphite anode. Key words: fast charging; graphite anode; electrolyte; solid electrolyte interphase; methyl acetateIntroductionIn the process of transitioning to a low-carbon economy, a number of emerging products have been developed, among which electric vehicles (EV) are regarded as a key component to de-carbonization. Even though there are some environmental impacts during the battery manufacturing process, EVs can reduce CO2 emission and fossil fuel consumption during usage, especially if the electricity can be generated from clean energy such as renewables.[1] Since their introduction, global sales of EVs have exceeded the 10 million mark. However, EV sales still account for only 13% of global vehicle sales in 2022, which reflects their low market acceptance. A major reason for this is the long charging times of EVs, which take up to an hour or more to charge even with a supercharger. Even though the U.S. Advanced Battery Alliance has proposed several development goals that can potentially increase the adoption of EVs, including to charge the EV batteries by 80% capacity in 15 minutes,[3] existing commercial lithium-ion batteries (LIB) with graphite anode cannot reach that requirement at the moment due to its sluggish kinetics. In a LIB, lithium ions are transported in the electrolyte, into the electrode, through the electrolyte-electrode interface and also within the active materials. Conventional electrolyte that containing LiPF6 and carbonated-based solvent cannot meet the requirement of fast charging due to high viscosity that hinder the ion transfer through electrolyte. To allow fast charging, it is therefore essential to reduce the resistance to ion transfer. As the electrolyte is the medium of ion transfer between the two electrodes and it can also affect the properties of the solid electrolyte interphase (SEI) on the active materials, optimizing it is a simple and effective method to realize faster kinetics. A number of works focusing on electrolyte design have been conducted to increase fast-charging performance of graphite anode. For example, Fan’s group designed two different electrolytes with low desolvation energy of Li+ using dioxolane (DOL) or fluorinated ethylene carbonate/acetonitrile (FEC/AN) as solvents to improve the fast-charging performance of batteries.[2] Dahn’s group added methyl acetate (MA), an organic compound with a low melting point of -98 C, into LiPF6-based carbonate electrolyte to increase the ionic conductivity across all salt concentrations of the electrolyte and the rate performance of the corresponding LiNi0.5Mn0.3Co0.2O2||graphite batteries.[3] Jiang et al. developed a localized high-concentration electrolytes (LHCEs) with diluent that can form a robust SEI and inhibit solvent co-intercalation into graphite.[4] In general, kinetics of lithium ion transport in a battery depends on the configuration and properties of the electrolyte such as the solvation structure and solvation energy of Li+, ionic conductivity and transference number, as well as the composition and thickness of the SEI on the surface of graphite. Though, the underlying mechanism may be more complicated than suggested. For example, Du et al. showed that LiNi0.8Mn0.1Co0.1O2||graphite full cells give better fast charging capability in electrolyte with lithium bis(fluorosulfonyl) imide (LiFSI) salt compared to LiPF6 salt, which was attributed to the better ionic conductivity and higher Li ion transference number of the electrolyte with LiFSI.[5] On the other hand, Kang et al. tested graphite||Li half cells with LiFSI and LiPF6-based electrolyte, showing that the cells have better high-temperature cycle stability with LiFSI, and suggested that LiFSI forms a more stable inorganic-rich SEI layer on graphite.[6] The difference in interpretation prompted us to conduct a systematic study here to clarify the effect of salt and electrolyte additive on the rate performance of graphite anode. We found that LiFSI salt facilitates the formation of a stable and less-resistive SEI with more Li2CO3 and less LiF on graphite compared to LiPF6, significantly improving Li ion transport. Addition of MA can further inhibit the decomposition of the LiFSI salt and increase ionic conductivity of the electrolyte. The graphite half-cell with LiFSI salt and MA additive can achieve a capacity of 230 mAh g-1 at a 2 C lithiation rate. This work emphasizes the importance of electrolyte design on the rate performance of graphite and provides an alternative electrolyte choice for fast-charging LIBs.ExperimentalMaterialsGraphite was purchased from SHANSHAN Advanced Materials Co., Ltd. Lithium chips with a thickness of 450 μm were obtained from DoDoChem. The solvents including fluoroethylene carbonate (FEC), diethyl carbonate (DEC) and MA were received from Sigma Aldrich. The commercially available Li salts, including LiPF6 and LiFSI were purchased from DoDoChem (purity ≥ 99.9%). Preparation of electrolyte and electrodeElectrolytes were prepared in an argon-filled glove box (MBraun) with less than 0.1 ppm O2 and H2O. Specifically, different Li salts were dissolving in FEC/DEC (1:1 by volume). For example, 1.0 M LiPF6 was dissolved in FEC/DEC mixture to obtain 1.0 M LiPF6 FEC/DEC. 1.5 M LiFSI was dissolved in FEC/DEC mixture to obtain 1.5 M LiFSI FEC/DEC. The electrolyte containing MA was prepared by adding 2 wt% MA into 1.5 M LiFSI FEC/DEC. The graphite electrode was prepared by coating a slurry mixture of 91 wt% graphite, 4 wt% acetylene black as the conducting agent and 5 wt% sodium carboxymethyl cellulose (Na-CMC in de-ionized water (Sigma Aldrich-low viscosity)) as the binder on copper foil. After drying at 80 C for 0.5 h, the Cu foil was punched into disks with a diameter of 16 mm. The average active mass loading of the electrodes is 3.0 mg cm-2 with an electrode density of 1.2 g cm-3 after pressing for half cell tests. A lower mass loading was used for half cells so that the influence of the electrolyte on the SEI can be explored with minimal effect from ion transport through the electrode. Before transferring into an Ar-filled glove box, electrodes were further dried at 110°C for 4 h in vacuum.Battery assembly2032-type coin cells were used to evaluate the electrochemical performance of different electrolytes. The dried electrode was assembled with lithium metal as counter electrode, followed by a piece of separator (Celgard 2500), and 120 μL of electrolyte. Cells were rested for 12 h after assembly to allow for electrolyte wetting. For full cell fabrication, LiFePO4 electrode and graphite electrode were purchased from Guangdong Canrd New Energy Technology Co., Ltd. LiFePO4 is chosen as the cathode here because it has a lower charge cutoff potential without corrosion of the Al current collector even with LiFSI-based electrolytes. The average active mass loading of the LiFePO4 electrodes is 10 mg cm-2, the average active mass loading of the graphite electrodes is 5 mg cm-2. The full cells were cathode limited, and the N/P ratio is around 1.2.Electrochemical measurementsThe cycle and rate tests in this study were performed in constant current (CC) mode rather than constant current-constant voltage (CC-CV) mode so that the differences can be observed easier. The cycling and rate performances were performed on a Neware battery testing system with a voltage range of 0.01-1.5 V at different current densities based on the mass of graphite active material. For cycle performance, the graphite electrodes were tested with a current rate of 0.1 C (1 C = 372 mA g-1) for 15 cycles followed by 2 C for subsequent cycles. For rate performance, delithiation rate (defined as charging here) is kept at 0.2 C while the lithiation rate (defined as discharging here) is varied. All the cycle and rate performance tests were conducted under constant current condition. Electrochemical impedance spectra (EIS) were acquired on a VMP3-potentiostat (Bio-Logic) from 10 mHz to 1 MHz with an AC amplitude of 10 mV at the discharged state. All the cells were held at 0.01 V for 1.5 h before EIS test. Cyclic voltammetry (CV) was performed on the same instrument and in similar cells, in the voltage range of 0.01–1.5 V vs. Li/Li+ at 0.05 mV s-1. All electrochemical tests were conducted at room temperature (25 C). The voltage range for full cell test is between 2.5-3.8 V, and the current rate calculation is based on the mass of cathode. CharacterizationsThe ionic conductivities of different electrolytes were measured by SevenCompact Cond meter S230, METTLER TOLEDO. For postmortem analyses, the test cells after three charge-discharge cycles at 0.1 C were disassembled in the glovebox and the extracted graphite electrodes were washed with dimethyl carbonate (DMC) solvent in the glovebox to remove the residue salt. X-ray photoelectron spectroscopy (XPS) was performed using Thermo Scientific K-Alpha on the electrodes, and the binding energy was calibrated based on the C 1s peak of 284.8 eV. High resolution transmission electron microscopy (HRTEM JEOL JEM-2100F) and scanning electron microscopy (JOEL JCM-6000 Versatile Benchtop SEM) were used to characterize the morphology of the cycled graphite electrode. Results and discussionsElectrochemical performances of graphite electrode with different electrolytesLiPF6 is the common salt that is used in lithium-ion batteries. As a benchmark, we used 1.0 M LiPF6 FEC/DEC = 1:1 as the starting electrolyte in our study, as FEC was previously shown to give better performance for graphite electrodes due to its excellent thermal stability, lower interfacial resistance of SEI and better fast charging capability than ethylene carbonate.[7, 8] To study the effect of salt concentration, 1.5 M LiPF6 FEC/DEC = 1:1 with the same solvent composition was prepared and tested with graphite electrode. To study the effect of the type of salt, we replaced LiPF6 with LiFSI to make an electrolyte with 1.5 M LiFSI FEC/DEC = 1:1. The influence of electrolyte was further investigated by adding 2 wt% MA into the electrolyte to form 1.5 M LiFSI FEC/DEC + 2 wt% MA. Rate performance and long-term cycle performance of the graphite electrode are first evaluated with different electrolytes. It should be specified that in the half-cells, discharge corresponds to the lithiation process of the graphite electrode while charge corresponds to the opposite. For all the tests of half-cells, the charge (delithiation) current was set at 0.2 C while the lithiation current were varied. The first discharge-charge curves of the graphite anode in different electrolytes are shown in Figure S1. The initial Coulombic efficiency (ICE) of graphite anode in 1.0 M LiPF6 FEC/DEC, 1.5 M LiPF6 FEC/DEC, 1.5 M LiFSI FEC/DEC and 1.5 M LiFSI FEC/DEC + 2 wt% MA electrolyte is about 88%, 86%, 87%, and 87%, respectively, which are similar to each other. For both the graphite electrodes tested in conventional electrolyte with LiPF6 salt (1.0 M LiPF6 and 1.5 M LiPF6 FEC/DEC), it can be seen from Figure 1a that they show poor rate performance, with a capacity of only about 60-80 mAh g-1 at 2 C. In comparison, graphite electrode tested in 1M LiFSI FEC/DEC electrolyte shows higher capacity of about 180 mAh g-1 at 2 C. With the addition of 2 wt% MA, the electrode exhibits even better rate capability with 230 mAh g-1 at 2 C rate, which amounts to a capacity utilization of 69.7% compared to the base capacity at 0.2 C. The corresponding charge-discharge curves at 0.5 C and 2 C are shown in Figure 1b. There is a clear potential plateau around 0.2 V at 2 C for 1.5 M LiFSI FEC/DEC + 2 wt% MA, showing the staging process during Li+ intercalating into graphite.[9] However, no obvious plateau can be seen for the electrodes tested in 1.0 M or 1.5 M LiPF6 FEC/DEC. The overpotential of the graphite electrode using LiFSI/MA-containing electrolyte is smaller than that in the other two LiPF6-based electrolytes, suggesting an easier lithiation process for the graphite electrodes using LiFSI/MA-containing electrolyte. Figure 1. Electrochemical performance of graphite electrodes in half cells. (a) Rate performance of the graphite electrodes using different electrolytes; (b) corresponding charge-discharge curves of the graphite electrodes using different electrolytes at selected discharge rate; (c) first cyclic voltammetry curves for graphite electrodes at a scan rate of 0.05 mV s-1; (d) cycle performance of the graphite electrodes using different electrolytes (cells were initiated at 0.1 C for the first 15 cycles followed by 2 C); (e) ionic conductivity of different electrolytes. Cyclic voltammetry (CV) was further conducted to investigate the differences between the electrolytes. It can be seen from Figure 1c, the peak currents during lithiation and delithiation for 1.5 M LiPF6 FEC/DEC are lower than that of 1.0 M LiPF6 FEC/DEC, while those of the cell with LiFSI/MA-containing electrolyte are significantly higher. Similar trend is also observed at different cycles of the CV plots (Figure S2), which implies that lithium ions have faster charge transfer kinetics in the battery using LiFSI/MA-containing electrolyte. In addition, the cycle performances of the graphite electrodes in the different electrolytes are also explored, as shown in Figure 1d. A similar trend to the rate performance is also observed in the cycle tests, where the capacity at 2 C of graphite electrode in 1.5 M LiPF6 FEC/DEC is the lowest, followed by 1.0 M LiPF6 FEC/DEC and 1.5 M LiFSI FEC/DEC, while that of 1.5 M LiFSI FEC/DEC + 2 wt% MA is the highest. For the graphite in LiFSI/MA electrolyte, even though there is a large drop in capacity when the current is increased to 2 C at the 16th cycle, the capacity gradually recovers with cycling probably due to cell activation. After 100 cycles, the cell with LiFSI/MA electrolyte shows a stable capacity of around 230 mAh g-1, which is much larger than the graphite electrode tested with 1.0 M LiPF6 FEC/DEC (~50 mAh g-1), 1.5 M LiPF6 FEC/DEC (~20 mAh g-1) and 1.5 M LiFSI FEC/DEC (~180 mAh g-1). Figure S3 shows the corresponding charge-discharge curves of the graphite electrodes at different cycle number. The graphite electrodes using LiPF6-based electrolytes show significantly larger polarization than those with LiFSI-based electrolytes. Specifically, the graphite electrode using LiFSI/MA electrolyte shows the smallest overpotential. This not only suggests that the LiFSI/MA electrolyte provides higher Li ion migration, but also forms a stable surface on the graphite for extended cycling. The above data shows that LiFSI/MA system is effective in improving the rate performance of the graphite electrode. To understand the reason, we first study the ionic conductivity of the different electrolytes, as shown in Figure 1e. The ionic conductivities of electrolyte with 1.0 M and 1.5 M LiPF6 are 8.8 and 3.3 mS cm-1, respectively. The large drop in ionic conductivity when the LiPF6 concentration is increased from 1.0 M to 1.5 M could be the reason why the performance of graphite in 1.5 M LiPF6 is worse than that in 1 M LiPF6. On the other hand, the ionic conductivity of electrolyte with 1.5 M LiFSI without MA is 8.4 mS cm-1. Even though it is smaller than the ionic conductivity of 1.0 M LiPF6 FEC/DEC, the LiFSI-based electrolyte shows much better rate performance than LiPF6-based electrolyte. This suggests that the improved rate capability from LiPF6 salt to LiFSI is not due to physical transport of lithium-ion in the electrolyte, but due to changes in the kinetics of the surface layer induced by LiFSI during charge-discharge, which will be further discussed in the next section. In comparison, the addition of 2 wt% MA increases the ionic conductivity of the electrolyte, which could be one of the reasons for the further improvement in rate performance. Since electrodes with 1.0 M LiPF6 FEC/DEC electrolytes gives better performance than that with 1.5 M LiPF6 FEC/DEC, we chose it as the benchmark in the following tests.Here, we have so far demonstrated the performances of electrolyte with 2 wt% MA. A comparison of LiFSI electrolyte with different MA amounts was also conducted, and the results are shown in Figure S4. Briefly, ionic conductivity is increased with the increasing amount of MA, to about 9.5-9.8 mS cm-1 for electrolyte with 2-5 wt% MA (Figure S4a). EIS tests of the graphite electrodes show that the electrolyte with 2 wt% MA has the smallest fitted value of RSEI among these three electrolytes after cycling (Figure S4b-e). In addition, rate performance of graphite using LiFSI electrolyte with different MA amounts was also investigated. The graphite electrode using 2 wt% MA electrolyte exhibits the highest capacity utilization at 2 C among these three electrolytes (Figure S4f). Thus, electrolyte with 1.5 M LiFSI FEC/DEC with 2 wt% MA is chosen here for further analyses.To further understand the effect of the salt and additive on the resistance of the battery, EIS test was conducted on the graphite half-cells using the different electrolytes. The Nyquist plots after different cycles are shown in Figure 2a and Figure S5a, b. In general, battery resistance is mainly composed of bulk resistance, surface layer resistance and charge transfer resistance. Based on the data, the plots obtained for the different electrolytes show the same characteristics, i.e., semicircles in the high and mid frequency regions and a sloping straight line in the low frequency region. The semicircle in the high frequency region represents the interface impedance, which is most likely governed by the SEI passivation film on the graphite surface (RSEI). The semicircle in the mid frequency range corresponds to charge transfer impedance (Rct), while the sloping line in the low frequency region is the diffusion impedance of Li+ inside the electrode.[10, 11] The impedance results were fitted with an equivalent circuit as shown in the inset of Figure 2a, where Re is the bulk resistance of the cell, RSEI and CSEI are the resistance and capacitance of the SEI, respectively, Cdl and Rct are the capacitance and resistance of the charge-transfer resistance, respectively, and W is the Warburg impedance.[12] Even though Rct is not the smallest for the graphite electrode tested in LiFSI/MA-containing electrolyte (Figure S5c), it shows the best rate performance. This implies that the charge transfer process is not the rate limiting step that affects the fast-charging performance. On the other hand, RSEI results (Figure 2b) show that the cells tested in electrolyte with LiFSI show a decreasing impedance with cycling, while that with LiPF6 shows an increase in impedance. This indicates the SEI of graphite in cells with LiFSI-based electrolytes is much more stable. In addition, the electrode tested with 1.5 M LiFSI FEC/DEC + 2 wt% MA shows the smallest overall impedance compared to the other two electrolytes, indicating the smallest resistance for Li+ transport through the graphite surface, which can be one possible reason for the observed excellent fast-charging performance. Figure 2. (a) Nyquist plots of graphite electrodes using different electrolytes after 50 cycles at 2 C discharge rate; (b) fitted value of RSEI of graphite electrodes using different electrolytes after different cycles.Effect of SEI on Li transportFrom the above data, it can be concluded that changing the electrolyte from LiPF6 to LiFSI/MA can improve the rate performance of graphite electrodes. To understand whether LiFSI or MA has larger effect, we conducted a control experiment by adding MA to LiPF6-based electrolyte (1.0 M LiPF6 FEC/DEC + 2 wt% MA). As shown in Figure S6a, compared to the graphite electrode using LiFSI/MA-containing electrolyte, the graphite electrode using LiPF6/MA-containing electrolyte shows a much worse rate performance (discharge capacity of 48 mAh g-1 at 2 C). In addition to the smaller capacity at higher lithiation rate, the overpotential of graphite electrode using LiPF6/MA-containing electrolyte is also much larger than that of LiFSI/MA-containing electrolyte (Figure S6b), implying higher resistance to Li+ intercalation for the graphite electrode tested in LiPF6/MA-containing electrolyte. As the ionic conductivity of the electrolyte is in fact increased with MA addition, as shown in Figure S6c, the result indicates that the rate performance of graphite is more affected by the salt than the MA additive.  To further explore the effect of the salt and additive, SEM was first conducted to study the difference in surface morphologies of the graphite electrodes. All the graphite electrodes were obtained from cells that had been cycled for 50 cycles at 2 C rate and finally cut off at 1.5 V (the cut off voltage for charge process). As shown in Figure S7, the graphite electrode tested in LiPF6-based electrolyte shows an irregular mossy morphology and the graphite particles are difficult to be distinguished from each other, whereas the graphite electrodes tested in LiFSI and LiFSI/MA electrolytes show flakes of graphite clearly visible on the surface. This suggests that the LiPF6-based electrolyte forms a thicker surface layer on graphite than the LiFSI/MA electrolyte. XPS was further conducted to explore the composition differences in SEI. The graphite electrodes used for the XPS test were removed from a graphite||Li half-cell that had been cycled at 0.1 C for 3 cycles, and the typical probing depth of XPS is 3-10 nm. Figure 3a and Figure 3b are the C 1s spectra and O 1s spectra, respectively of the cycled graphite electrodes. Figure S8 shows the F 1s spectra. The spectra of C 1s spectra can be deconvoluted into several peaks corresponding to C-C for graphite (284.8 eV), C-O (286.8 eV), C=O (288.7 eV) and Li2CO3 (290.3 eV),[13-15] while the spectra of O 1s spectra can be deconvoluted into two peaks corresponding to C═O (531.8 eV) and C-O (533.4 eV).[16, 17] For the F 1s spectra, the major peak around 685 eV can be attributed to LiF species, while the other small peak could be originated from the decomposed products of PF6- [18] and FSI- group.[19] Moreover, there are obvious signals of decomposition products of LiFSI at 165-172 eV from the S 2p spectra in Figure 3c. In general, all the graphite electrode surfaces contain an SEI layer with LiF, Li2CO3 and lithium carbonates with small portions of polymeric compounds.[20] A comparison of the XPS spectra indicates that the graphite electrodes tested with LiFSI-based electrolyte contains more C=O component and less LiF species than that tested with LiPF6-based electrolyte. LiF is known to be more resistive, so the reduction of the LiF amount on the surface with LiFSI salt can decreases resistance for lithium transport, which may explain why rate performance is increased when we replace LiPF6 salt with LiFSI salt. [21] Figure 3. Interfacial chemistry of graphite electrode in different electrolytes revealed by XPS spectra. (a) C 1s spectra; (b) O 1s spectra; (c) S 2p spectra; (d) HRTEM image of cycled graphite electrode using LiFSI/MA-containing electrolyte.MA additive further changes the composition of the SEI. The C=O and C-O peak signals may come from organic ROCO2Li species and inorganic Li2CO3 species.[22] For the LiFSI-based electrolyte, the C-O peak intensity is reduced while the C=O peaks and Li2CO3 peak intensities are increased with MA addition (Figure 3a). In addition, the amount of S-containing components is reduced (Figure 3c). The results suggest that MA addition has two effects, first is that it facilitates the production of Li2CO3 and second is that it reduces the decomposition of LiFSI. The synergistic effect between LiF and Li2CO3 can promote ion transport in the SEI.[23] In addition, Li2CO3 is endothermic and harder to dissolve in the electrolyte compared with metastable organic compounds (ROCO2Li species).[24] The induced SEI is therefore more stable, as consistent with EIS results, which improves cycle stability. HRTEM was further used to investigate the thickness of the SEI. It can be seen from the Figure 3d that the thickness of SEI on the graphite with LiFSI/MA-containing electrolyte is around 5 nm after 3 cycles at 0.1 C. Under the same conditions, as can be seen from Figure S9, the SEI thickness of the graphite electrode using LiPF6-based electrolyte is about 15 nm thick, which is thicker than that of the LiFSI/MA electrolyte. Overall, LiFSI/MA generates a thinner SEI with less LiF component compared to LiPF6, which explains why the rate performance of graphite with LiFSI/MA is better than that with LiPF6. To further understand how the LiFSI salt and MA additive affect the stability of the SEI layer on the graphite electrode, electrolyte swapping experiments were conducted. Specifically, the graphite||Li half cells were first cycled at 0.1 C in their respective electrolytes (A) initially for 3 times to form an SEI. Then, the cells were disassembled to obtain the cycled graphite electrodes, which were then re-assembled into new batteries with a new electrolyte (B), as shown in the schematic diagram in Figure 4a and Table 1. The new batteries were tested with a charge current rate of 0.2 C (delithiation process) and a discharge current rate of 2 C (lithiation process). Figure 4b-4e show the obtained discharge capacity and their corresponding charge-discharge curves from the electrolyte swapping experiments. For graphite electrodes tested only in A = B = 1.5 M LiFSI FEC/DEC + 2 wt% MA (cell 6), the 2 C discharge capacity is as high as 210 mAh g-1. In comparison, if the graphite electrode with FSI-SEI (A) is tested in 1.0 M LiPF6 FEC/DEC (B) (cell 2), the capacity is much worse and the overpotential is significantly increased at a higher current rate (Figure 4b and 4c). This indicates that the introduction of LiPF6 salt to the cell slows down the kinetics of Li transport, suggesting that the LiPF6 will form a new insulating layer on the surface. On the other hand, the graphite electrode tested only with 1.0 M LiPF6 FEC/DEC (cell 1) shows poor rate performance, but the performance improves if the electrolyte is changed to 1.5 M LiFSI/FECDEC + 2 wt% MA (cell 4). This is interesting because it suggests that the new salt is able to modify the existing SEI on the graphite surface and improve its ionic conductivity. Figure 4. (a) Schematic diagram of the electrolyte exchange experiments; (b) and (c) rate performance and its corresponding charge-discharge curves of graphite half cells first using 1.5 M LiFSI FEC/DEC + 2 wt% MA as SEI forming electrolyte (cell 2 and 6); (d) and (e) rate performance and its corresponding charge-discharge curves of graphite half cells first using 1.0 M LiPF6 FEC/DEC (cell 1 and 4). Table 1. Setup of the electrolyte swapping experiment and the summary of discharge capacity after 5 cycles at 2 C rate. Initial electrolyte (A)Electrolyte after swapping (B) 1.0 M LiPF6 FEC/DEC 1.5 M LiFSI FEC/DEC 1.5 M LiFSI FEC/DEC + 2 wt% MA 1.0 M LiPF6 FEC/DEC 75 mAh g-1(cell 1) - 25 mAh g-1 (cell 2) 1.5 M LiFSI FEC/DEC - 175 mAh g-1(cell 3) - 1.5 M LiFSI FEC/DEC + 2 wt% MA 148 mAh g-1(cell 4) 225 mAh g-1(cell 5) 230 mAh g-1(cell 6)To see the effect of MA, we also made a cell tested only with 1.5 M LiFSI FEC/DEC (cell 3) and compared to it another cell with FSI-SEI but swapped to 1.5 M LiFSI FEC/DEC + 2 wt% MA afterwards (cell 5). The results, as shown in Table 1 and Figure S10, indicate that the FSI-based SEI gives good rate-performance, while the addition of MA can further enhance Li-ion transport at high current rates.XPS tests were also conducted to study the SEI composition of the cells after the electrolyte swapping experiments for cell 2 (LiPF6 swapped to LiFSI/MA) and cell 4 (LiFSI/MA swapped to LiPF6). Figure 5a, 5b and 5c show the C 1s spectra, O 1s spectra and F 1s spectra, respectively, of the graphite electrodes extracted from the two cells after 2 C discharge. The XPS data show similar trend to previous results as described in Figure 3. After changing the electrolyte from 1.0 M LiPF6 FEC/DEC to 1.5 M LiFSI FEC/DEC + 2 wt% MA, the relative content of Li2CO3 increased, and the LiF content, on the contrary, reduced. This allows faster kinetics of Li-ions through the surface layer in cell 2, consistent with the rate performance results. In comparison, for the cell first tested in 1.5 M LiFSI FEC/DEC + 2 wt% MA and then replaced by 1.0 M LiPF6 FEC/DEC (cell 4), the relative content of LiF is increased, and the Li2CO3 content, on the contrary, is reduced. It is likely that the new LiPF6 salt decomposes on the surface of the graphite electrolyte, increasing the resistance of the SEI, which can explain the poorer high-rate capability as observed after electrolyte replacement.Figure 5. Interfacial chemistry of graphite electrode in different electrolytes revealed by XPS spectra. (a) C 1s spectra; (b) O 1s spectra; (c) F 1s spectra; (d) schematic diagram of the effects of LiFSI and MA on SEI composition.In general, the salt and additive in the electrolyte can affect the SEI on the surface of graphite, and a summary of the effects is shown in Figure 5d. In particular, LiFSI/MA reduces the thickness of the SEI, reduces the content of LiF on the surface and also reduces salt decomposition, thereby induces a more stable SEI for fast charging graphite electrode compared to LiPF6. This electrolyte swapping experiment also indicates that the SEI varies dynamically with cycling after changing the electrolyte, which is an area that is worth further investigations in the future. Electrochemical performance of LiFePO4||graphite full cellsLiFSI/MA-based electrolyte shows better rate performance than that with LiPF6 in half cells because of its ability to modulate the composition and thickness of SEI of the graphite anode. Here, we further demonstrate the effectiveness of the LiFSI/MA electrolyte with LiFePO4||graphite full cells with higher mass loadings. The specific capacity is calculated based on the mass of the LiFePO4 cathode, and 1 C was set to 150 mA g-1. The cells were first tested at 20 mA g-1 for 3 cycles followed by a current rate of 2.27 C (340 mA g-1) subsequently. As shown in Figure 6a, the LiFePO4||graphite batteries all show an initial capacity of about 140 mAh g-1, close to the available capacity of LiFePO4. With cycling, the battery with LiPF6-based electrolyte shows fast drop in capacity, which is attributed to the continuous growth of SEI on the graphite. On the other hand, the cycle performance of the full cells with LiFSI/MA electrolyte is  better than that of cell with LiPF6-based electrolyte. This is attributed to the more robust SEI formed on the graphite surface with LiFSI/MA electrolyte.Figure 6. (a) Cycle performance and (b) rate performance of LiFePO4||graphite full cells using the different electrolytes; (c) charge-discharge curves of for rate test at the current rate of 340 mA g-1. The effect of the electrolyte on the rate capability of the full cells is shown in Figure 6b. When the current is increased to above 100 mA g-1 (0.67 C), the capacity of the battery using 1.0 M LiPF6 FEC/DEC is significantly reduced. Specifically, when the current is increased to 340 mA g-1, the battery using 1.0 M LiPF6 FEC/DEC only shows a capacity of about 80 mAh g-1. Rate performance is improved with LiFSI/MA electrolytes. As can be seen in Figure 6c, the full cell with LiFSI/MA electrolyte also exhibits smaller overpotential. This indicates that the new LiFSI/MA electrolyte can potentially be used for full cell applications with LiFePO4 cathode. ConclusionsThe Li-ion transport properties and rate performance of graphite electrode are highly dependent on the salt and additive used. A combination of LiFSI salt and MA additive leads to a thinner and less resistive SEI layer compared to LiPF6 salt, thus improving the fast-charging capability of graphite anode for LIBs. Our electrolyte swapping experiments also indicate that the SEI is dynamic, and its composition can change if new electrolyte is introduced. This suggests that there are opportunities to tune the interaction of salt, additive and solvent with the material surface to further improve the rate and stability of the corresponding battery. AcknowledgementThe work described in this paper was supported by General Research Fund (CityU 11305220) and the Research Matching Grant Scheme (PJ9229008) from the University Grants Committee of the Hong Kong Special Administrative Region, China. References[1]. X. Xia, P. Li, A review of the life cycle assessment of electric vehicles: Considering the influence of batteries, Science of The Total Environment, 814 (2022) 152870.[2]. C. Sun, X. Ji, S. Weng, R. Li, X. Huang, C. Zhu, X. Xiao, T. Deng, L. Fan, L. Chen, X. Wang, C. Wang, X. Fan, 50C Fast-Charge Li-Ion Batteries using a Graphite Anode, Advanced Materials, 34 (2022) 2206020.[3]. E. Logan, E.M. Tonita, K. Gering, J. Li, X. Ma, L. Beaulieu, J. Dahn, A study of the physical properties of Li-ion battery electrolytes containing esters, Journal of The Electrochemical Society, 165 (2018) A21.[4]. L. L. Jiang, C. Yan, Y. X. Yao, W. Cai, J. Q. Huang, Q. Zhang, Inhibiting Solvent Co-Intercalation in a Graphite Anode by a Localized High-Concentration Electrolyte in Fast-Charging Batteries, Angewandte Chemie International Edition, 60 (2021) 3402-3406.[5]. Z. Du, D.L. Wood, I. Belharouak, Enabling fast charging of high energy density Li-ion cells with high lithium ion transport electrolytes, Electrochemistry communications, 103 (2019) 109-113.[6]. S. J. Kang, K. Park, S. H. Park, H. Lee, Unraveling the role of LiFSI electrolyte in the superior performance of graphite anodes for Li-ion batteries, Electrochimica Acta, 259 (2018) 949-954.[7]. H. Shin, J. Park, A.M. Sastry, W. Lu, Effects of fluoroethylene carbonate (FEC) on anode and cathode interfaces at elevated temperatures, Journal of The Electrochemical Society, 162 (2015) A1683.[8]. H.B. Son, M. Y. Jeong, J. G. Han, K. Kim, K.H. Kim, K. M. Jeong, N. S. Choi, Effect of reductive cyclic carbonate additives and linear carbonate co-solvents on fast chargeability of LiNi0.6Co0.2Mn0.2O2/graphite cells, Journal of Power Sources, 400 (2018) 147-156.[9]. H. Wang, M. Yoshio, Carbon-coated natural graphite prepared by thermal vapor decomposition process, a candidate anode material for lithium-ion battery, Journal of Power Sources, 93 (2001) 123-129.[10]. C. Wang, I. Kakwan, A.J. Appleby, F.E. Little, In situ investigation of electrochemical lithium intercalation into graphite powder, Journal of Electroanalytical Chemistry, 489 (2000) 55-67.[11]. A. Funabiki, M. Inaba, Z. Ogumi, S.i. Yuasa, J. Otsuji, A. Tasaka, Impedance Study on the Electrochemical Lithium Intercalation into Natural Graphite Powder, Journal of The Electrochemical Society, 145 (1998) 172.[12]. S.S. Zhang, Electrochemical study of the formation of a solid electrolyte interface on graphite in a LiBC2O4F2-based electrolyte, Journal of Power Sources, 163 (2007) 713-718.[13]. K.G. Saw, J. du Plessis, The X-ray photoelectron spectroscopy C 1s diamond peak of chemical vapour deposition diamond from a sharp interfacial structure, Materials Letters, 58 (2004) 1344-1348.[14]. X. Chen, X. Wang, D. Fang, A review on C1s XPS-spectra for some kinds of carbon materials, Fullerenes, Nanotubes and Carbon Nanostructures, 28 (2020) 1048-1058.[15]. D. Ensling, M. Stjerndahl, A. Nytén, T. Gustafsson, J.O. Thomas, A comparative XPS surface study of Li2FeSiO4/C cycled with LiTFSI- and LiPF6-based electrolytes, Journal of Materials Chemistry, 19 (2009) 82-88.[16]. R. Muzyka, M. Kwoka, Ł. Smędowski, N. Díez, G. Gryglewicz, Oxidation of graphite by different modified Hummers methods, New Carbon Materials, 32 (2017) 15-20.[17]. K. Jurewicz, K. Babeł, R. Pietrzak, S. Delpeux, H. Wachowska, Capacitance properties of multi-walled carbon nanotubes modified by activation and ammoxidation, Carbon, 44 (2006) 2368-2375.[18]. N. Kim, Y. Myung, H. Kang, J. W. Lee, M. Yang, Effects of Methyl Acetate as a Co-Solvent in Carbonate-Based Electrolytes for Improved Lithium Metal Batteries, ACS Applied Materials & Interfaces, 11 (2019) 33844-33849.[19]. Q. Li, W. Xue, J. Peng, L. Yang, H. Pan, X. Yu, H. Li, Synergistic Effect of Temperature and Electrolyte Concentration on Solid-State Interphase for High-Performance Lithium Metal Batteries, Advanced Energy and Sustainability Research, 2 (2021) 2100010.[20]. S.J. An, J. Li, C. Daniel, D. Mohanty, S. Nagpure, D.L. Wood, The state of understanding of the lithium-ion-battery graphite solid electrolyte interphase (SEI) and its relationship to formation cycling, Carbon, 105 (2016) 52-76.[21]. D. Aurbach, Y. Ein‐Eli, B. Markovsky, A. Zaban, S. Luski, Y. Carmeli, H. Yamin, The Study of Electrolyte Solutions Based on Ethylene and Diethyl Carbonates for Rechargeable Li Batteries: II . Graphite Electrodes, Journal of The Electrochemical Society, 142 (1995) 2882.[22]. S. Leroy, F. Blanchard, R. Dedryvère, H. Martinez, B. Carré, D. Lemordant, D. Gonbeau, Surface film formation on a graphite electrode in Li-ion batteries: AFM and XPS study, Surface and Interface Analysis, 37 (2005) 773-781.[23]. Q. Zhang, J. Pan, P. Lu, Z. Liu, M.W. Verbrugge, B.W. Sheldon, Y.-T. Cheng, Y. Qi, X. Xiao, Synergetic Effects of Inorganic Components in Solid Electrolyte Interphase on High Cycle Efficiency of Lithium Ion Batteries, Nano Letters, 16 (2016) 2011-2016.[24]. M. Lu, H. Cheng, Y. Yang, A comparison of solid electrolyte interphase (SEI) on the artificial graphite anode of the aged and cycled commercial lithium ion cells, Electrochimica Acta, 53 (2008) 3539-3546.image1.tiffimage2.tiffimage3.tiffimage4.tiffimage5.tiffimage6.tif