The superior rate performance and long cycle stability make LAN promising for Li metal practical batteries. Industry-level pouch cells were paired with thin Li foil (50 µm thickness, ~ 10 mAh/cm2) and high-loading NCM811 (~ 3.66 mAh/cm2), offering an N/P ratio of 2.7. To achieve high specific energy, we assembled pouch cells with lean electrolytes (E/C = 2.8 g/Ah). All the pouch cells underwent two formation cycles at 0.1 C (i.e., 0.366 mA/cm2), yielding an initial specific energy density of 310 Wh/Kg (Fig. 5a). Except for the mass of taps and packing foil, an innate energy density for cells was up to ~ 403 Wh/kg, which could be a valuable parameter for evaluating the energy density of larger size batteries. To determine a limited operating current density of our designed LAN, we first assessed performances of rate capacity with two different protocols: charging/discharging at the same C rate; constant charging at 1.46 mAh/cm2 but an increasing discharging rate. Under two different protocols, LAN presented excellent fast charging/discharging capability compared to RCE (Fig. 5a,b). For pouch cells cycled in RCE, a sudden capacity drop started with charging or discharging at 3.66 mA/cm2, and cells operated worse in following extended cycles. Surprisingly, LAN endowed pouch cells with superior cycling stability and high capacity even at the considerably large discharging current density of 1.8 C, i.e., 6.59 mA/cm2, and their corresponding power densities were shown in Fig. 5c. The maximum power density was calculated to be 408 W/kg (except taps and packing foil: 530 W/kg) at 1.8 C rate for LAN; contrastingly, RCE enabled a maximum value of 318 W/kg (except taps and packing foil: 414 W/kg) during cells cycling at 1.4 C, i.e., 5.12 mA/cm2. This power density of LAN is one of the largest among advanced Li metal cells (Fig. 5d, Supplementary Table 3).
Moreover, the long cycling stability of LAN was tested by pouch cells cycled under three conditions. As shown in Fig. 5e, cells delivered a total capacity of 400 mAh and maintained 90% capacity retention after 120 cycles at a slow charging rate of 0.2 C and fast discharging at 0.6 C rate (1 C = 3.66 mA/cm2). In contrast, cells with the control electrolyte, RCE, only survived 25 cycles with an initial capacity of 394 mAh under identical conditions. The extending cycle life indicated that LAN contributed to fast Li+ transportation under stringent conditions. Under higher charging/discharging current densities, LAN still made a high reversible Li+ plating/stripping possible for pouch cells cycling at 1.46 mA/cm2 (i.e., 0.4 C) for charging and 2.93 mA/cm2 (i.e., 0.8 C) for discharging, with an 86% capacity retention after 100 cycles (Supplementary Fig. 29). To achieve special battery applications, we compared the fast-discharging capability of cells at a current density of 3.66 /cm2 (i.e., 1.0 C) (Fig. 5f), and 81% capacity was preserved for LAN after 100 cycles, but an abrupt fade in both capacity and CE happened to RCE after ~ 50 cycles. These outstanding performance of fast operation enriched LAN as an advanced electrolyte for high-power-density electric devices.
“Self-cleaning” of SEI layer by the LFEA electrolyte
We carried out systematic studies to find out the mechanism of LiFEA in enabling the batteries with impressive fast charging/discharging performances. As previously reported, a multitude of factors limited the fast charging/discharging behavior in batteries35, such as ionic conductivity of electrolytes, the structure and composition of SEI layers, and tLi+. Due to LiFEA with an additive-level amount in LA, an ignored shift in conductivity was confirmed by EIS spectra (Supplementary Fig. 30). SEI structure and composition are generally believed to correlate with the solubility of inorganic/organic lithium salts and oligomers and the rate of approaching their saturation limit17. Among these reduced/decomposed species, Li2CO3, LiF and Li2O had been reported to be less soluble in carbonate electrolytes than organic-based species because of their larger lattice energies43, 44. Unexpectedly, LiFEA-based electrolytes with high-DN properties enhanced the solubility of these three representative lithium salts. Therefore, it is highly promising that the high-DN electrolytes are also able to boost the solubility of organic lithium salts and oligomers to a greater extent than the inorganics, e. g., Li2CO3, LiF and Li2O.
Considering that LiFEA possesses a unique, folded structure able to greatly increase the tLi+ and boost the solubility of organic/inorganic salts species, we hypothesized that the particular geometry of LiFEA both improves the Li+ transport property in the bulk electrolytes and elevates the quality of SEI layers by altering the solubility of inorganic/organic lithium salts and oligomers (Fig. 6a), thus actuating a combined improvement in the battery performance. As the Li+ transport property in the bulk electrolytes has been described from the view of tLi+ to conductivity, we will not discuss it anymore in the following context and lay more attention on the mechanism of the dissolving (“cleaning”) components from SEI layers.
In order to gain more insightful information, we combined quartz crystal microbalance analysis with electrochemical methods (in situ EC-QCM) to investigate mass variations of the SEI film soaked in different electrolytes45 (Supplementary Fig. 31). As the schematic diagram demonstrates (Fig. 6b), on the surface of Li metal deposited on Q-sensor grows in-situ pristine SEI. Once SEI components dissolve in electrolytes, Q-sensor will output a higher frequency record, which can be calculated into mass loss according to a formula: \(\varDelta m=-C\frac{\varDelta f}{\begin{array}{c}n\end{array}}\). Figure 6c exhibited shifts in frequency after refreshing RCE, and Q-sensor gradually accessed a balance stage within 40 min. During this recovering balance, Δf increased to 916.0 Hz, indicating that SEI film experienced a mass loss of 8.3 µg. Subsequently, LA was injected to exchange for the original, soaked electrolyte. Q-sensor underwent a remarkable frequency shift and took longer from the injecting timepoint to a steady balance stage. The mass of SEI film was decreased by 25.7 µg due to soaked LA, three times that of RCE (Fig. 6c inserted Figure). These distinct differences between SEI mass loss provided clear proof that LiFEA swept some components from the SEI layer.
X-ray photoelectron spectroscopy (XPS) analysis offered a deeper insight into the surface composition and spatial element distribution of SEI layers soaked with the LA or RCE. Figure 6d listed atomic ratios of SEIs in 0 and 60 s sputtering depth. SEI layer treated with LA contained higher Li content but slightly lower C, O and F atomic ratios on its surface than that of RCE, confirming organic, F-based species removed from SEI into the soaked LA. Considering that LiF mostly dominated F-based species in an SEI layer46, it is reasonable for this decreased F content because LiFEA contributed to dissolving some LiF. The dissolved organic-based species were proved to be -(CH2CH2O)n-, some organic carbonates (e.g., ROCO2Li) by comparing assigned peaks in O 1s and C 1s spectra22, 28, 47 (Fig. 7a,b, Supplementary Fig. 32). A quantitative comparison of these carbonate contents in SEI layers was shown in Supplementary Fig. 33, which makes the tendency of LiFEA “sweeping” organic lithium polycarbonates from the SEI layer more apparent.
More direct evidence for this differentiated SEI composition is provided by time-of-flight secondary-ion mass spectrometry (TOF-SIMS) (Fig. 7c,d, Supplementary Fig. 34). During sputtering, various secondary ion fragments were obtained, originating from the organics and inorganics of SEI layers. Among these signals, the C2H3O− is generally from the -(CH2CH2O)n- and lithium ethylene mono-carbonate components (LEMC), and the CO3− commonly belongs to ROCO2Li/Li2CO3. In the normalized TOF-SIMS depth profiles of C2H3O−, the signal from SEI soaked with LA arrived at its maximum normalized intensity earlier than that of RCE and experienced a steep drop afterward (Fig. 7c). This indicates LiFEA enhancing the solubility of ether oligomers and LEMC components, consent with XPS results. The two depth profiles of CO3− fragments presented similar changes to those of C2H3O− signals as the sputtering prolonged, and the SEI layer treated with LA showed a lower intensity of CO3− than that of RCE, verifying a carbonate-cleaning process induced by LA. Visual 3D images revealed a majority of C2H3O− and CO3− signals located in the outer layer of SEIs (Fig. 7d), with the SEI layer soaked in LA enabling weaker signals compared with that in RCE. This lowered signal intensity vividly showed that LiFEA could clean out more organic-species from the SEI layer and therefore, leaving more inorganic-rich components remaining in the SEI, especially in the inner layer.