Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Superconcentrated electrolytes for a high-voltage lithium-ion battery

Superconcentrated electrolytes for a high-voltage lithium-ion battery ARTICLE Received 30 Nov 2015 | Accepted 24 May 2016 | Published 29 Jun 2016 DOI: 10.1038/ncomms12032 OPEN Superconcentrated electrolytes for a high-voltage lithium-ion battery 1, 1,2, 2,3,4 1 2,4 Jianhui Wang *, Yuki Yamada *, Keitaro Sodeyama , Ching Hua Chiang , Yoshitaka Tateyama 1,2 & Atsuo Yamada Finding a viable electrolyte for next-generation 5 V-class lithium-ion batteries is of primary importance. A long-standing obstacle has been metal-ion dissolution at high voltages. The LiPF salt in conventional electrolytes is chemically unstable, which accelerates transition metal dissolution of the electrode material, yet beneficially suppresses oxidative dissolution of the aluminium current collector; replacing LiPF with more stable lithium salts may diminish transition metal dissolution but unfortunately encounters severe aluminium oxidation. Here we report an electrolyte design that can solve this dilemma. By mixing a stable lithium salt LiN(SO F) with dimethyl carbonate solvent at extremely 2 2 high concentrations, we obtain an unusual liquid showing a three-dimensional network of anions and solvent molecules that coordinate strongly to Li ions. This simple formulation of superconcentrated LiN(SO F) /dimethyl carbonate electrolyte inhibits the dissolution of 2 2 both aluminium and transition metal at around 5 V, and realizes a high-voltage LiNi Mn O /graphite battery that exhibits excellent cycling durability, high rate capability 0.5 1.5 4 and enhanced safety. 1 2 Department of Chemical System Engineering, University of Tokyo, 7-3-1, Hongo, Tokyo 113-8656, Japan. Elements Strategy Initiative for Catalysts and 3 4 Batteries, Kyoto University, 1-30, Goryo-Ohara, Kyoto 615-8245, Japan. JST PRESTO, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan. International Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1, Namiki, Tsukuba 305-0044, Japan. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to A.Y. (email: [email protected]). NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ithium-ion batteries, having received great commercial Independent of the solvents used, the viscosity increases expo- success in the portable power source market, are being nentially with increasing the LiFSA mole fraction (X ). LiFSA Laimed for large-scale energy-storage application in electric Among the three groups of solutions, the group with DMC as the 1–3 vehicles . To approach the high energy-density requirements solvent shows the lowest viscosity because pure DMC has a lower for automobiles, a pragmatic approach is to elevate the operating viscosity than pure EC or mixed EC:DMC. For electrolytes with voltage of batteries, from the present 4 V to around 5 V (refs 4,5). similar solvation radiuses of mobile ions, the ionic conductivity is This allows the direct application of the mature fabrication proportional to the number of mobile ions and inversely pro- technology of 4 V-class lithium-ion batteries, the well-developed portional to the viscosity of the medium . As shown in Fig. 1c, at negative electrodes (for example, graphite and graphite/silicon), dilute concentrations of X o0.14 (below 1.5 mol dm ), the LiFSA and high-voltage positive electrodes (for example, spinel use of the EC:DMC mixture shows the highest ionic conductivity LiNi Mn O and some layered oxides). However, new owing to a synergistic effect: the high-dielectric-constant EC 0.5 1.5 4 challenges—which mainly arise from the electrolyte—hinder the increases the mobile ion number by promoting salt dissociation; practical application of the next-generation 5 V-class battery. the low-viscosity DMC increases the ion mobility by decreasing One major problem is metal dissolution from the positive the solution viscosity. This is why the mixed solvents of EC and electrode at high voltages, which poses a serious dilemma in linear carbonates are generally adopted in conventional designing an electrolyte. In state-of-the-art lithium-ion electro- electrolytes of the lithium-ion battery . However, when X LiFSA lyte, chemically unstable LiPF is an essential component to is above 0.14, the solution with DMC as the sole solvent shows an suppress anodic (oxidative) dissolution of an aluminium current even higher ionic conductivity than that with EC:DMC, which collector because its hydrolysis product of hydrofluoric acid (HF) should result from the much lower viscosity of the former at high 6,7 contributes to an insoluble AlF passivation film . However, the concentrations. This result suggests that the viscosity becomes the generated HF accelerates the dissolution of transition metals from decisive factor on the ionic conductivity for a concentrated the active electrode materials, which causes severe capacity decay solution, wherein intensive ionic association exists independent of upon cycling, especially at high voltages and elevated the solvents used, showing a distinct departure from the 8,9 temperatures . Using LiNi Mn O as an example, the conventional electrolyte design strategy on the basis of dilute 0.5 1.5 4 2þ 2þ dissolved Mn and Ni ions, albeit o1% of the total concentrations. For the LiFSA/DMC solution, a commercially amount, deposit on the surface of the graphite negative acceptable ionic conductivity of 1.12 mS cm (30 C) is obtained electrode, which thicken the solid electrolyte interphase (SEI) even at a ‘super-high’ concentration with salt-to-solvent molar by catalysing the reductive decomposition of the electrolyte, and ratio of 1:1.1 despite a high viscosity of 238.9 mPa s. Although the consume the limited lithium reserve in the battery to result in a ionic conductivity is lower than that of the commercial dilute 10,11 450% capacity loss in 100 charge/discharge cycles . electrolyte, it does not compromise the rate capability of the Diversified functional additives and/or alternative solvents have battery (shown later). 12–17 been explored but improvements are still unsatisfactory. On the other hand, the drawbacks of the high volatility and Efforts have tried more stable salts (less tendency to generate HF) high flammability of linear carbonate solvents can be overcome to to replace LiPF , such as lithium perfluorosulfonylamide a large degree owing to the much lower content of organic (shortened to ‘amide’) . However, the chemically stable amide solvents in the concentrated solutions. Thermogravimetry does not participate in the reaction with Al to form a stable measurements (Supplementary Fig. 1) show that the weight loss passivation film, thus causing severe anodic dissolution of the Al of the superconcentrated 1:1.1 LiFSA/DMC solution is only current collector at 44 V (refs 19–22). As a result, it remains a 1.5 wt% after elevating the temperature to 100 C, which is dilemma for electrolyte design to suppress both the Al dissolution considerably lower than those of a dilute 1:10.8 LiFSA/DMC (requiring an unstable salt) and transition metal dissolution solution (65.5 wt%, corresponding to 1.0 mol dm ) and a (avoiding an unstable salt). Recently, increasing the concentration commercial electrolyte (28.7 wt%). As demonstrated in the flame 23–25 of amide salts was reported to alleviate anodic Al dissolution , tests (Fig. 1d,e), the 1:1.1 LiFSA/DMC solution does not burn as but the operating voltage of a half-cell is still limited below 4.3 V, fiercely as the commercial dilute electrolyte. The superior thermal presumably owing to some or all of the following reasons: stability and flame retardant ability of the concentrated electro- 23 24 insufficient salt concentration , too low ionic conductivity and lytes contribute to the remarkably improved safety properties as too low oxidative stability of the solvent . compared with the dilute electrolytes. In this work, we report an electrolyte system to resolve the dilemma. We select stable yet dissociative lithium bis(fluorosul- fonyl)amide (LiFSA) as the salt and oxidation-stable carbonate Reversible reaction of a 5 V-class electrode. Anodic dissolution esters as the solvent. We demonstrate an unusual liquid with a of the Al current collector and/or oxidative decomposition of peculiar three-dimensional structural network obtained at solvent may be encountered in the high-voltage application of extremely high salt concentrations. The superconcentrated amide-based electrolytes. To exclude the possible influence of the electrolyte not only effectively suppresses the anodic Al dissolu- anodic Al dissolution, we initially used platinum foil as the tion but also remarkably inhibits the transition metal dissolution current collector for the LiNi Mn O electrode in a three- 0.5 1.5 4 and, thus, realizes a safe, stable and fast-rate high-voltage electrode cell (Supplementary Fig. 2). The results showed that LiNi Mn O |graphite battery. both dilute (1:10.8) and superconcentrated (1:1.1) LiFSA/DMC 0.5 1.5 4 electrolytes enabled a reversible Li de-intercalation/intercala- tion on the LiNi Mn O |Pt electrode, indicating a reasonably 0.5 1.5 4 Results good compatibility between the present electrolyte system and Physicochemical properties. LiFSA salt was dissolved at various LiNi Mn O material at B5V. 0.5 1.5 4 concentrations into three different carbonate ester solvents: However, when applied in a coin cell using the conventional Al dimethyl carbonate (DMC), ethylene carbonate (EC) and mixed current collector, low concentrations of LiFSA/DMC electrolytes EC:DMC. All the mixtures are transparent liquids at room encountered problems, confirming the critical drawback of temperature (see Fig. 1a as an example). Their basic physico- anodic Al dissolution for the amide-based electrolytes. As shown chemical properties are presented in Supplementary Table 1. in Fig. 2a, the first charge on the LiNi Mn O |Al electrode is 0.5 1.5 4 Figure 1b shows their viscosity as a function of salt concentration. impossible in the dilute 1:10.8 LiFSA/DMC electrolyte owing to 2 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE LiFSA-to-solvent molar ratio a b 1,000 30 °C EC EC:DMC 100 DMC LiFSA/DMC 1 : 10 1 : 2.0 1 : 1.3 1 : 1.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 X mole fraction LiFSA LiFSA-to-solvent molar ratio de –3 1 mol dm 1 : 1.1 LiPF / EC:DMC LiFSA / DMC 30 °C DMC EC:DMC EC 0.0 0.1 0.2 0.3 0.4 0.5 0.6 X mole fraction LiFSA Figure 1 | Physicochemical properties dependent on solution concentration. (a) Images of various salt-to-solvent molar ratios of LiFSA/DMC solutions. Viscosity (b) and ionic conductivity (c) for solutions of LiFSA in DMC, EC and EC:DMC (1:1 by mol.) at 30 C. The X mole fraction is the molar amount LiFSA of LiFSA salt divided by the total molar amount of the salt and solvents. The LiFSA-to-solvent molar ratios of the solutions are shown on the upper axis. (d) Flame tests of a commercial dilute electrolyte of 1.0 mol dm LiPF /EC:DMC (1:1 by vol.) and (e) the lab-made superconcentrated electrolyte of 1:1.1 LiFSA/DMC. the continuous Al dissolution at 4.3 V. In the concentrated 1:1.9 conductivity of the former. These results demonstrate that LiFSA/DMC electrolyte (Fig. 2b), the charge/discharge cycling the salt-superconcentrated strategy is a simple, effective and becomes possible up to the cutoff voltage of 4.9 V, but the large fruitful approach to various safe and stable high-voltage irreversible capacity indicates the parasitic Al dissolution remains. electrolytes. To the best of our knowledge, this is the first time The Al dissolution subsequently deteriorates the electrical that stable and fast charge/discharge cycling of a 5 V-class contacts between the LiNi Mn O material and the Al current electrode using amide salt-based organic electrolytes has been 0.5 1.5 4 collector, and results in a fast capacity decay (Fig. 2d). Actually, achieved. the poor cycling performance on the LiNi Mn O electrode The progressive inhibition of anodic Al dissolution with 0.5 1.5 4 was generally observed in other concentrated 1:2 LiFSA/ increasing salt concentration is further proved by linear carbonate ester electrolytes, such as LiFSA in EC, propylene sweep voltammetry (LSV) of an Al electrode and the subsequent carbonate (PC), ethyl methyl carbonate (EMC) and diethyl scanning electron microscopy observation on the polarized Al carbonate (DEC; Supplementary Fig. 3), indicating this concen- surface (see Fig. 3 for details). This notable concentration tration is not sufficient to fully inhibit Al dissolution at 5 V. effect was recently reported but with a debate on whether a In contrast, the superconcentrated 1:1.1 LiFSA/DMC electro- stable surface film on Al (ref. 23) or the elimination of þ 24,25 lyte enables a reversible Li de-intercalation/intercalation uncoordinated (free) solvents of electrolyte plays the key reaction on the LiNi Mn O electrode even at a high voltage role. We conducted a surface analysis of the Al electrodes 0.5 1.5 4 of 5.2 V (Fig. 2c). In a charge/discharge cycling test, the capacity polarized in various concentrations of LiFSA/DMC electrolytes retention after 100 cycles was over 95% (Fig. 2d), and the by X-ray photoemission spectroscopy (XPS) as well as a coulombic efficiency was close to 100% (Supplementary Fig. 4), comparative LSV study between fresh and polarized Al evidencing an effective inhibition of anodic Al dissolution. electrodes (see Supplementary Figs 5 and 6 for details). We Similarly, using the super-high concentration of 1:1.3, all the were unable to obtain any essential evidence to support the LiFSA/carbonate electrolytes enabled stable charging/discharging existence of a stronger surface film generated in the cycling of the LiNi Mn O electrode (see Fig. 2e for example). concentrated electrolyte. Instead, we found that the LiFSA salt 0.5 1.5 4 Especially, the electrolytes using low-dielectric-constant and low- readily decomposes and produces LiF upon Ar etching viscosity linear carbonate solvents (for example, DMC, EMC and (Supplementary Fig. 7). The previous observation in ref. 23—a DEC) showed a faster rate capability as compared with those much thicker surface film of LiF produced in a higher using high-dielectric-constant and high-viscosity cyclic carbonate concentration of electrolyte—is likely to arise from the solvents (for example, EC, PC and their corresponding mixtures), decomposition of un-rinsed amide salt induced by Ar which is at least partly owing to the much higher ionic etching in the XPS measurement. NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 3 1:22.7 1:10 1:4.9 1:2.9 1:2.0 1:1.3 1:22.7 1:10 1:4.9 1:2.9 1:2.0 1:1.3 –1 Conductivity (mS cm ) Viscosity (mPa S) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 a 5.5 5.5 1:10.8 LiFSA/DMC 1:1.9 LiFSA/DMC 1st 2nd 10th 50th 5.0 5.0 1st 4.5 4.5 1st 50th 10th 4.0 4.0 3.5 3.5 0 50 100 150 200 250 300 0 50 100 150 200 –1 –1 Capacity (mAh g ) Capacity (mAh g ) cd 5.5 LiNi0.5Mn1.5O4 LiFSA/DMC 1:1.1 LiFSA/DMC 2nd 1st 100th 1:1.1 5.0 4.5 4.0 1:1.9 1:10.8 3.5 0 50 100 150 0 20406080 100 –1 Capacity (mAh g ) Cycle number 1/5C 1/2C 1/5C 1:1.3 LiFSA/solvent 1C 2C 5C 10C DMC EMC DEC EC:DMC PC EC 0 1020304050 Cycle number Figure 2 | Performance of 5 V-class LiNi Mn O electrode in a half-cell. Charge–discharge voltage curves of LiNi Mn O |lithium metal half-cells 0.5 1.5 4 0.5 1.5 4 using (a) dilute 1:10.8, (b) moderately concentrated 1:1.9 and (c) superconcentrated 1:1.1 LiFSA/DMC electrolytes at a C/5 rate. Some/all curves of 1st, 2nd, 10th, 50th and 100th cycles are shown. (d) Discharge (Li intercalation) capacity retention of the half-cells using different concentrations of LiFSA/DMC electrolytes at a C/5 rate. (e) Rate capacity and subsequent cycling retention of the half-cells using 1:1.3 LiFSA-based electrolytes with different carbonate solvents. Charge–discharge tests were conducted at 25 C with a cutoff voltage of 3.5–4.9 V and a maximum-time restriction of 10 h except for that using the 1:1.1 LiFSA/DMC electrolyte whose cutoff voltage was 3.5–5.2 V. The 1C-rate corresponds to 147 mA g on the weight basis of the LiNi Mn O 0.5 1.5 4 electrode. We now study the solution structure of the electrolytes using concentration increases, the population of free DMC decreases þ þ Raman spectroscopy observation and density functional theory and that of Li -coordinated DMC increases; the Li -FSA molecular dynamics simulation (DFT-MD). As shown in the association simultaneously intensifies through the formation of Raman spectra (Fig. 4b left), a free DMC molecule exhibits an contact ion pairs (CIPs, FSA coordinating to one Li ) and O-CH stretching vibration band at 910 cm (ref. 27). This aggregate clusters (AGGs, FSA coordinating to two or more 1 þ þ band shifts up to 930–935 cm when DMC participates in Li Li ). The latter is evidenced from a remarkable upshift of the solvation. In dilute 1:10.8 LiFSA/DMC, the majority of DMC FSA band (700–780 cm , Fig. 4b right), which is typically 24,25,28–31 molecules exist in a free state because the solvent-to-salt molar observed in the amide-based concentrated solutions . ratio (10.8) is much larger than a typical four- or fivefold For the moderately concentrated 1:2 LiFSA/DMC solution, the coordination of Li in aprotic solvents. As the LiFSA Raman band corresponding to free DMC is remarkably 4 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications Voltage (V) Voltage (V) –1 Capacity (mAh g ) –1 Capacity (mAh g ) Voltage (V) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE finally, the diffusion of the solvated metal cation to the bulk 0.08 Al electrode electrolyte . At high voltages of B5 V, the first step proceeds 1:1.9 in LiFSA/DMC more rapidly and extensively than at the conventional operating 1:2.9 1:10.8 voltage of 4 V. Thereby, the subsequent coordination and 0.06 diffusion must be strongly inhibited by the nature of electrolyte solutions to suppress the metal ion dissolution. In the moderately concentrated 1:2 LiFSA/carbonate electrolytes, the presence of 0.04 significant free solvents and CIPs (with two or more coordination sites remaining vacant) could coordinate to Al cations and fail to inhibit Al dissolution completely at 5 V. In contrast, the superconcentrated 1:1.1 LiFSA/DMC electrolyte effectively 0.02 inhibits Al dissolution even over 5 V, which can be ascribed to 1:1.1 its peculiar AGGs-predominant solution structure: (i) all DMC solvents and all FSA anions strongly coordinate to Li cations 0.00 and thus have a much lower probability of coordinating to other metal cations; (ii) the resulting reinforced three-dimensional 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 network further retards the diffusion rate of the metal cations, Potential (V) vs Li/Li particularly, those with multiple charge. Figure 3 | Oxidation stability of an aluminium electrode. LSV of an aluminium electrode in various concentrations of LiFSA/DMC electrolytes Stable cycling of a 4.6 V LiNi Mn O |graphite battery.In 0.5 1.5 4 in a three-electrode cell. The scan rate was 1.0 mVs . The insets are addition to the excellent performance achieved on the scanning electron microscopy images of the Al surface polarized in the LiNi Mn O -positive electrode, the superconcentrated 1:1.1 0.5 1.5 4 dilute 1:10.8 (left of panel) and superconcentrated 1:1.1 (right of panel) LiFSA/DMC electrolyte also realized ultra-stable charge/discharge electrolytes. Many corroding pits cover the surface of the Al electrode cycling on the natural graphite-negative electrode despite the polarized in the dilute electrolyte, showing a severe anodic Al dissolution. In absence of conventional SEI-forming agent of EC (Supplementary contrast, no corroding pits appear on the surface of the Al electrode Fig. 9): the application in a graphite|Li half-cell exhibits a capacity polarized in the superconcentrated electrolyte, indicating a good inhibition retention of 99.6% after 100 cycles with coulombic efficiency of anodic Al dissolution. The white scale bar represents 20mm. of B99.8%, and with rate capability comparable with that using a commercial dilute electrolyte. Accordingly, the super- concentrated electrolyte was further applied in the high-voltage weakened, suggesting that the majority of DMC are solvating to LiNi Mn O |graphite full cell, a much harsher condition than 0.5 1.5 4 Li . This is consistent with the DFT-MD simulation, which in the half-cell, because the active lithium resource is limited and shows ca. 90% DMC are coordinating to Li with the rest as free a new underlying problem arises from the transition metal solvent (marked as light blue in Fig. 4d). Moreover, the dissolution from the LiNi Mn O especially at high voltages 0.5 1.5 4 simulation illustrates that all FSA anions are coordinating to and elevated temperatures. Figure 5a,b shows charge/discharge Li with ca. 20% as CIPs and ca. 80% as AGGs (marked as voltage profiles of LiNi Mn O |natural graphite full cells at 0.5 1.5 4 orange and dark blue in Fig. 4d, respectively). The coordination 40 C using a state-of-the-art commercial electrolyte and the environment is shown in Supplementary Fig. 8. For the lab-made superconcentrated electrolyte, respectively. The cell superconcentrated 1:1.1 LiFSA/DMC solution, both DMC and with the commercial electrolyte suffers from a severe capacity FSA bands further upshift substantially, indicating both Li - decay during cycling, that is, only 18% of the initial capacity left DMC and Li -FSA interactions enhanced compared with after 100 cycles (Fig. 5a,c), which is consistent with previous 11,33 those in 1:2 LiFSA/DMC. The DFT-MD simulation reveals that reports . In contrast, the capacity retention using the 1:1.1 all DMC molecules, together with all FSA anions, are LiFSA/DMC electrolyte is over 90% after 100 cycles, exhibiting a coordinating to Li (no free solvent or anion). Interestingly, remarkably improved cycling durability (Fig. 5b,c and besides oxygen, significant amount of nitrogen on FSA anions Supplementary Figs 10 and 11). Notably, the superiority of the also participate in the coordination with Li , which is hardly superconcentrated electrolyte becomes even more marked at observed at the lower concentrations. The contribution of 55 C (Supplementary Fig. 12). It is generally accepted that the nitrogen coordinating to Li is shown in Supplementary Fig. 8. poor cycling performance of the LiNi Mn O |graphite battery 0.5 1.5 4 More importantly, almost all FSA anions remain in AGG states originates from the dissolution of transition metals from during the whole DFT-MD simulation time (0.1 fs 100,000 LiNi Mn O into the electrolyte, as introduced at the 0.5 1.5 4 steps), and a CIP state is rarely observed, demonstrating the beginning of this article. Indeed, energy dispersive X-ray unusual solution structural feature with AGG clusters as the spectroscopy (EDS) observation shows a much lower content of predominant components in the superconcentrated LiFSA/DMC Mn and Ni on the graphite electrode of the full cell cycled in the solution. It is noteworthy that each FSA anion coordinates to superconcentrated electrolyte than that in the commercial þ þ 2–3 Li and each Li is coordinated by 2–3 FSA in 1:1.1 electrolyte, which provides evidence for the effective inhibition LiFSA/DMC. Hence, FSA anions in the superconcentrated of transition metal dissolution in the former. There are two main solution connect with each other through the intensive reasons for the improved performance: (i) LiFSA is less reactive to association with Li , leading to a reinforced three-dimensional produce HF as compared with LiPF , which alleviates the network (shown in Fig. 4e). This feature is different from the less corrosion of LiNi Mn O and, thus, reduces the formation of 0.5 1.5 4 2þ 2þ 2þ 2þ concentrated solutions, wherein significant amount of CIPs and soluble Mn and Ni ; (ii) even if some Mn and Ni are free solvents divide the solution structure into relatively small-size formed on the surface of LiNi Mn O , they can hardly dissolve 0.5 1.5 4 parts. in and transport through the AGGs-predominant super- Generally, the anodic metal dissolution requires three steps: concentrated electrolyte owing to the same functional manner first, oxidation of the metal to a metal cation; second, for the inhibition of Al dissolution. Moreover, it is worth coordination of the metal cation by solvents or anions; and noting that the rate performance of the full cell using the NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 5 –2 Current (mA cm ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 O c S S F F O O Free DMC Free FSA Li Li O O Li + – CIP: 1Li ···FSA S S + − AGG-I: 2Li ···FSA F F O O + − AGG-II: 3Li ···FSA O O Li + + − Li -coordinated DMC Li -associated FSA (three main positions) LiFSA:DMC 1 : 1.1 1 : 1.9 1 : 2.9 1 : 10.8 Pure DMC 960 940 920 900 780 760 740 720 700 –1 Raman shift (cm ) Figure 4 | Li salt solvent coordination structure dependent on salt concentration. (a) The several main species in the LiFSA/DMC solutions. (b) Raman spectra of LiFSA/DMC solutions with various salt-to-solvent molar ratios in the range of 890–900 cm (O-CH stretching mode of the DMC solvent) and 700–780 cm (S-N stretching mode of the FSA anion). Snapshots of typical equilibrium trajectories obtained by DFT-MD simulations: (c) 3  3 dilute solution (1 LiFSA/25 DMC, o1 mol dm ), (d) moderately concentrated solution (12 LiFSA/24 DMC, ca. 4 mol dm ), and (e) superconcentrated 3 þ þ solution (10 LiFSA/11 DMC, ca. 5.5 mol dm ). The coordination of Li  DMC and Li  FSA is supposed to build up when the involved atoms locate þ þ within 2.5 Å from Li . The coordination numbers of solvents and anions to Li are shown in Supplementary Fig. 8. Li cations are marked in purple. Free and coordinated DMC molecules are marked in light blue and grey, respectively. Free, CIP and AGG states of FSA anions are marked in red, orange and dark blue, respectively. Hydrogen atoms are not shown. superconcentrated 1:1.1 LiFSA/DMC is comparable with that Discussion using the commercial electrolyte (Fig. 5d), although the former The conventional dilute LiPF /EC-based electrolytes have domi- shows an ionic conductivity one-order lower than the latter. The nated the electrolyte market of 4 V-class lithium-ion batteries mechanistic understanding on the high-rate capability of the over the past 25 years; however, they show difficulties in satisfying superconcentrated electrolyte is underway in our laboratory. To the requirements of next-generation 5 V-class batteries in terms the best of our knowledge, this is the first case that, an electrolyte of both safety and stability. Our work demonstrates a number of with such an ultra-simple formulation—a single salt and a single electrolytes with a reinforced three-dimensional network that are solvent without any additive—realizes stable cycling of a high- obtained by simple mixing of a stable salt with a conventional voltage lithium-ion battery. carbonate solvent at ‘super-high’ concentrations. Owing to its 6 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications Coordinated DMC Free DMC AGG-II AGG-I CIP Free FSA Intensity (a.u.) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE ab 40 °C 40 °C 1:1.1 LiFSA/DMC Commercial electrolyte 5.0 5.0 100th 10th 2nd 100th 50th 2nd 4.5 4.5 4.0 4.0 3.5 3.5 0 30 60 90 120 0 30 60 90 120 –1 –1 Capacity (mAh g ) Capacity (mAh g ) c d 40 °C 2C 5C C/5 C/2 1C C/5 1:1.1 LiFSA/DMC 1:1.1 LiFSA/DMC Mn Kα Ni Kα 5.4 6.0 6.6 7.2 7.8 Commercial Energy (keV) Commercial 25 °C 0 30 0 20 40 60 80 100 0369 12 15 18 Cycle number Cycle number Figure 5 | Performance of a high-voltage LiNi Mn O |natural graphite battery. Charge–discharge voltage curves of LiNi Mn O |graphite full cells 0.5 1.5 4 0.5 1.5 4 using (a) a commercial 1.0 mol dm LiPF /EC:DMC (1:1 by vol.) electrolyte and (b) lab-made superconcentrated 1:1.1 LiFSA/DMC electrolyte at a C/5 rate and 40 C. The curves of 2nd, 10th, 50th and 100th cycle are shown. (c) Discharge capacity retention of the full cells at a C/5 rate. The inset shows EDS spectra on the graphite electrode surface (200 200mm area) after 8-day cycling tests, which is equivalent to the operating time of 100 and 20 cycles for the battery using the commercial and superconcentrated electrolytes, respectively. (d) Discharge capacity of the full cell at various C-rates and 25 C. All charge-discharge cycling tests were conducted with a cutoff voltage of 3.5–4.8 V. 1C-rate corresponds to 147 mA g on the weight basis of the LiNi Mn O electrode. 0.5 1.5 4 peculiar structural characteristics, the superconcentrated electro- superconcentrated electrolytes do not contain any additives, lytes overcomes the longstanding challenge faced by the unstable signifying the potential to further enhance the performance. LiPF -based electrolytes at high voltages (passivating the Al These desirable features above outperform the conventional dilute current collector versus accelerating the transition metal dissolu- electrolytes; meanwhile, the wide-temperature window of the tion of the active material), thus, enables a stable operation of a liquid state (ensuring a good contact with the electrode 5 V-class battery. Emphasis is on the fact that the peculiar materials), as well as the convenience of the approach, surpass solution structure and functionalities are unique to such super- the solid-state electrolytes. Therefore, the superconcentrated high concentrations (solvent/saltE1.1), and cannot be achieved electrolytes might offer opportunities to build safe and stable in moderately high concentrations (solvent/salt41.8) as in high-voltage batteries that are not limited to the lithium-ion. 24,30,31,34,35 previous reports . Besides the positive electrode side, the superconcentrated electrolytes also show a good compatibility Methods with the natural graphite-negative electrode even in the absence Preparation of electrolytes and electrodes. LiFSA (Nippon shokubai) and all of EC. It breaks through the limitation of a general requirement of solvents (DMC, DEC, EMC, EC and PC, Kishida Chemical Co. Ltd) were lithium EC for a SEI formation for a lithium-ion electrolyte, and battery grade and used without purification. Electrolyte solutions were prepared by mixing a given amount of LiFSA with solvents in an Ar-filled glove box. The diversifies the electrolyte formulation towards various EC-free commercial electrolyte of 1.0 mol dm LiPF /EC:DMC (1:1 by vol) was electrolytes. Different from the conventional electrolyte design purchased from Kishida Chemical Co. Ltd and used as the reference. Both the that requires a high-dielectric-constant (usually high-viscosity) lab-made LiFSA-based electrolytes and as-received commercial LiPF -based solvent, the superconcentrated electrolyte prefers a low-viscosity electrolyte were dried by molecular sieve before tests. The water content was less than 2 p.p.m., as detected by a coulometric Karl Fischer Titrator. solvent. Although the ionic conductivity of the superconcentrated The electrodes were fabricated by first well mixing the active materials of electrolyte is lower than that of the conventional dilute electrolyte, LiNi Mn O (Hosen Corp., mean particle size R¼ 5mm, no surface treatment) 0.5 1.5 4 it does not necessarily compromise the rate capability of the and natural graphite (SEC Carbon Ltd., R¼ 10mm), polyvinylidene difluoride battery. Clarification of the corresponding mechanism would be (PVdF) and/or Denka black (AB, HS-100) in N-methylpyrrolidone with weight ratios of 80:10:10 (LiNi Mn O :PVdF:AB) and 90:10 (graphite:PVdF). The enlightening for developing novel high-power batteries. 0.5 1.5 4 resultant slurry was cast on the Al or Pt foil (20mm thickness) for the Furthermore, the superconcentrated electrolytes show superior LiNi Mn O electrode and on the Cu foil (10mm thickness) for the graphite 0.5 1.5 4 thermal stability and flame retardant ability, alleviating the safety electrode using a 50mm doctor blade. All those electrodes were dried at 120 C risk for a high-voltage battery using conventional dilute under vacuum for 12 h. The active material mass loading was 0.7–2 mg cm with electrolytes. Finally, it is noteworthy that our reported a thickness of B15–20mm, unless otherwise mentioned. The use of relatively low NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 7 –1 Capacity (mAh g ) Voltage (V) –1 Capacity (mAh g ) Voltage (V) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 mass loading was to spotlight the critical issue of anodic Al dissolution as the 11. Pieczonka, N. P. W. et al. Understanding transition-metal dissolution behavior content ratio of metallic Al components (Al current collector and Al positive cap) in LiNi Mn O high-voltage spinel for lithium ion batteries. J. Phys. Chem. C 0.5 1.5 4 to the active electrode material becomes much higher in a coin cell. Nevertheless, 117, 15947–15957 (2013). thick electrodes with a high mass loading of B10 mg cm were also tested. The 12. Zhang, Z. et al. Fluorinated electrolytes for 5 V lithium-ion battery chemistry. results are shown in Supplementary Figs 11 and 12. Energy Environ. Sci. 6, 1806–1810 (2013). 13. Hu, L., Zhang, Z. & Amine, K. Fluorinated electrolytes for Li-ion battery: An FEC-based electrolyte for high voltage LiNi Mn O /graphite couple. Electrochemical measurements. LSV was performed by VMP-3 (BioLogic) in a 0.5 1.5 4 Electrochem. Commun. 35, 76–79 (2013). beaker cell with an Al belt (1 4cm , 0.6 cm soaked in the electrolyte) as a working electrode and lithium metal as the reference and counter electrodes (shown in 14. Pieczonka, N. P. W. et al. Impact of lithium bis(oxalate)borate electrolyte Supplementary Fig. 5 inset). LiNi Mn O |Li half-cells and LiNi Mn O |- additive on the performance of high-voltage spinel/graphite Li-ion batteries. 0.5 1.5 4 0.5 1.5 4 graphite full cells were assembled in the standard 2032-type coin cell hardware in J. Phys. Chem. C 117, 22603–22612 (2013). an Ar-filled glove box. A combined separator, composed of cellulose separator 15. Song, Y.-M., Han, J.-G., Park, S., Lee, K. T. & Choi, N.-S. A multifunctional (Nippon Kodoshi, placed on the positive electrode side) and glass fibre separator phosphite-containing electrolyte for 5 V-class LiNi Mn O cathodes with 0.5 1.5 4 (Advantec GB-50, placed on the negative electrode side), was used. The amount of superior electrochemical performance. J. Mater. Chem. A 2, 9506–9513 (2014). electrolyte in a coin cell was ca 160ml to fully wet the separators and electrodes. 16. Xu, M. et al. Improved performance of high voltage graphite/LiNi Mn O 0.5 1.5 4 In the full cells, the weight ratio of LiNi Mn O :graphite was B2:1, which 0.5 1.5 4 batteries with added lithium tetramethyl borate. ECS Electrochem. Lett. 4, corresponds to B1:1.3 of their theoretical capacity ratio. Galvanostatic charge/ A83–A86 (2015). discharge cycling and rate capability tests were conducted on a charge/discharge 17. Xu, K. Electrolytes and Interphases in Li-Ion Batteries and Beyond. Chem. Rev. unit (TOSCAT). Charge and discharge were conducted at the same C-rate without 114, 11503–11618 (2014). using a constant-voltage mode at both ends of the charge and discharge. 18. Xu, K. Nonaqueous liquid electrolytes for lithium-based rechargeable batteries. Chem. Rev. 104, 4303–4417 (2004). 19. Krause, L. J. et al. Corrosion of aluminum at high voltages in non-aqueous Characterization. The density and viscosity of solution samples were evaluated electrolytes containing perfluoroalkylsulfonyl imides; new lithium salts for with a DMA 35 density meter and a Lovis 2000 M viscometer, respectively. lithium-ion cells. J. Power Sources 68, 320–325 (1997). The ionic conductivity was measured by AC impedance spectroscopy at 1 kHz 20. Wang, X., Yasukawa, E. & Mori, S. Inhibition of anodic corrosion of aluminum (Solartron 147055BEC) in a symmetric cell (Pt|electrolyte|Pt). The flammability cathode current collector on recharging in lithium imide electrolytes. was tested on an electrolyte-soaked glass fibre filter (Advantec GB-100). Electrochim. Acta 45, 2677–2684 (2000). The solution structure was studied by a Raman spectroscopy with an exciting 21. Li, L. et al. Transport and electrochemical properties and spectral features of laser of 514 nm (NRS-5100). The samples were sealed in a quartz cell in the glove box to avoid any contamination from the air. non-aqueous electrolytes containing LiFSI in linear carbonate solvents. The morphology of Al electrodes after LSV tests were observed by a field- J. Electrochem. Soc. 158, A74–A82 (2011). emission scanning electron microscopy at 2.0 kV. The transition metals deposited 22. Kramer, E. et al. Mechanism of anodic dissolution of the aluminum current on the graphite electrode in the LiNi Mn O |graphite full cells after charge/ collector in 1 M LiTFSI EC:DEC 3:7 in rechargeable lithium batteries. 0.5 1.5 4 discharge cycling were estimated by an EDS. The cells were disassembled in the J. Electrochem. Soc. 160, A356–A360 (2013). glove box. The obtained electrodes were rinsed in DMC and dried in the glove box. 23. Matsumoto, K. et al. Suppression of aluminum corrosion by using high The sample was exposed in air for o1 min at sample loading. concentration LiTFSI electrolyte. J. Power Sources 231, 234–238 (2013). The experimental details for thermogravimetric analysis and XPS 24. McOwen, D. W. et al. Concentrated electrolytes: decrypting electrolyte measurements are shown in Supplementary Fig. 1 legend and Supplementary properties and reassessing Al corrosion mechanisms. Energy Environ. Sci. 7, Methods, respectively. 416–426 (2014). 25. Moon, H. et al. Solvent activity in electrolyte solutions controls electrochemical reactions in Li-ion and Li-sulfur batteries. J. Phys. Chem. C 119, 3957–3970 Simulations. Car-Parrinello type DFT-MD simulations were carried out using 36 (2015). CPMD code . LiFSA/DMC solutions with salt-to-solvent molar ratios of 1:25, 1:2 26. Yoshida, K. et al. Oxidative-stability enhancement and charge transport and 1:1.1 were calculated in cubic supercells with 15.05, 17.03 and 14.34 Å linear mechanism in glyme-lithium salt equimolar complexes. J. Am. Chem. Soc. 133, dimensions, respectively. A fictitious electric mass of 500 a.u. and a time step of 4 13121–13129 (2011). a.u. (0.10 fs) were chosen. The temperature was controlled using a Nose´ thermostat 27. Katon, J. E. & Cohen, M. D. The vibrational spectra and structure of dimethyl with a target temperature of 30 C. After 5 ps equilibration steps, statistical averages carbonate and its conformational behavior. Can. J. Chem. 53, 1378–1386 (1975). were computed from trajectories of at least 10 ps in length. The electronic wave 28. Seo, D. M., Borodin, O., Han, S.-D., Boyle, P. D. & Henderson, W.A. Electrolyte function was quenched to the Born-Oppenheimer surface approximately every 1 ps solvation and ionic association II. Acetonitrile-lithium salt mixtures: highly to maintain adiabaticity. The energy cutoff of the plane wave basis is set to 90 Ry. dissociated salts. J. Electrochem. Soc. 159, A1489–A1500 (2012). Goedecker–Teter–Hutter type norm-conserving pseudopotentials for C, H, O, N, S, 29. Yamada, Y., Yaegashi, M., Abe, T. & Yamada, A. A superconcentrated ether F and Li were used . electrolyte for fast-charging Li-ion batteries. Chem. Commun. 49, 11194–11196 (2013). Data availability. The data that support the findings of this study are available 30. Yamada, Y. et al. Unusual stability of acetonitrile-based superconcentrated from the corresponding author upon request. electrolytes for fast-charging lithium-ion batteries. J. Am. Chem. Soc. 136, 5039–5046 (2014). 31. Yamada, Y. & Yamada, A. Review-Superconcentrated electrolytes for lithium References batteries. J. Electrochem. Soc. 162, A2406–A2423 (2015). 1. Goodenough, J. B. & Kim, Y. Challenges for rechargeable Li batteries. Chem. 32. Yamada, Y. et al. Corrosion prevention mechanism of aluminum metal in Mater. 22, 587–603 (2010). superconcentrated electrolytes. ChemElectroChem 2, 1687–1694 (2015). 2. Etacheri, V., Marom, R., Elazari, R., Salitra, G. & Aurbach, D. Challenges in the 33. Glatthaar, S., Dorit, N., Binder, J. R. & Andres, H. Evidence of loss of active development of advanced Li-ion batteries: a review. Energy Environ. Sci. 4, lithium in titanium-doped LiNi Mn O /graphite cells. J. Power Sources 274, 0.5 1.5 4 3243–3262 (2011). 1267–1275 (2015). 3. Amine, K., Kanno, R. & Tzeng, Y. Rechargeable lithium batteries and beyond: 34. Jeong, S.-K., Inaba, M., Iriyama, Y., Abe, T. & Ogumi, Z. Electrochemical Progress, challenges, and future directions. MRS Bull. 39, 395–401 (2014). intercalation of lithium ion within graphite from propylene carbonate 4. Patoux, S. et al. High voltage spinel oxides for Li-ion batteries: From the solutions. Electrochem. Solid State Lett. 6, A13–A15 (2003). material research to the application. J. Power Sources 189, 344–352 (2009). 35. Suo, L., Hu, Y.-S., Li, H., Armand, M. & Chen, L. A new class of solvent-in-salt 5. Croy, J. R., Abouimrane, A. & Zhang, Z. Next-generation lithium-ion batteries: electrolyte for high-energy rechargeable metallic lithium batteries. Nat. the promise of near-term advancements. MRS Bull. 39, 407–415 (2014). Commun. 4, 1481 (2013). 6. Myung, S.-T., Sasaki, Y., Sakurada, S., Sun, Y.-K. & Yashiro, H. Electrochemical 36. CPMD. http://www.cpmd.org:81/manual/node4.html, Copyright IBM Corp behavior of current collectors for lithium batteries in non-aqueous alkyl carbonate (1990-2015), Copyright MPI fu¨r Fesko¨rperforschung Stuttgart (1997-2001). solution and surface analysis by ToF-SIMS. Electrochim. Acta 55, 288–297 (2009). 37. Goedecker, S., Teter, M. & Hutter, J. Separable dual-space Gaussian 7. Zhang, X. & Devine, T. M. Identity of passive film formed on aluminum in Li- pseudopotentials. Phys. Rev. B 54, 1703–1710 (1996). ion battery electrolytes with LiPF6. J. Electrochem. Soc. 153, B344–B351 (2006). 8. Aurbach, D. et al. Review on electrode–electrolyte solution interactions, related to cathode materials for Li-ion batteries. J. Power Sources 165, 491–499 (2007). Acknowledgements 9. Zhan, C. et al. Mn(II) deposition on anodes and its effects on capacity fade in This work was partially supported by JSPS Grant-in-Aid for Young Scientists (A) spinel lithium manganate-carbon systems. Nat. Commun. 4, 2437 (2013). (No. 26708030) and JSPS Specially Promoted Research (No. 15H05701). The calculations 10. Kim, J.-H. et al. Understanding the capacity fading mechanism in were carried out at the super-computer centres of National Institute for Materials LiNi Mn O /graphite Li-ion batteries. Electrochim. Acta 90, 556–562 (2013). Science, the University of Tokyo, and the K computer at the RIKEN through the HPCI 0.5 1.5 4 8 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE System Research Projects (hp150209). We thank Keisuke Kikuchi, Reiko Kawakami Competing financial interests: The authors declare no competing financial interests. and Dr Kouhei Okitsu for their assistance in the experiments, and specially thank Reprints and permission information is available online at http://npg.nature.com/ Dr Sai-Cheong Chung for his valuable suggestions on the manuscript. reprintsandpermissions/ How to cite this article: Wang, J. et al. Superconcentrated electrolytes for a high-voltage Author contributions lithium-ion battery. Nat. Commun. 7:12032 doi: 10.1038/ncomms12032 (2016). J.W. and Y.Y. contributed equally to this work. Y.Y. and A.Y. proposed the concept. J.W. and Y.Y. designed the experiments. J.W. and C.H.C. carried out the experiments and analysed the data. K.S. and Y.T. designed and conducted the theoretical calculations. This work is licensed under a Creative Commons Attribution 4.0 J.W., Y.Y. and A.Y. wrote the manuscript. A.Y. supervised the whole project. International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, Additional information users will need to obtain permission from the license holder to reproduce the material. Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 9 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Superconcentrated electrolytes for a high-voltage lithium-ion battery

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/superconcentrated-electrolytes-for-a-high-voltage-lithium-ion-battery-6QmYSoKgJE

References (39)

Publisher
Springer Journals
Copyright
Copyright © 2016 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/ncomms12032
Publisher site
See Article on Publisher Site

Abstract

ARTICLE Received 30 Nov 2015 | Accepted 24 May 2016 | Published 29 Jun 2016 DOI: 10.1038/ncomms12032 OPEN Superconcentrated electrolytes for a high-voltage lithium-ion battery 1, 1,2, 2,3,4 1 2,4 Jianhui Wang *, Yuki Yamada *, Keitaro Sodeyama , Ching Hua Chiang , Yoshitaka Tateyama 1,2 & Atsuo Yamada Finding a viable electrolyte for next-generation 5 V-class lithium-ion batteries is of primary importance. A long-standing obstacle has been metal-ion dissolution at high voltages. The LiPF salt in conventional electrolytes is chemically unstable, which accelerates transition metal dissolution of the electrode material, yet beneficially suppresses oxidative dissolution of the aluminium current collector; replacing LiPF with more stable lithium salts may diminish transition metal dissolution but unfortunately encounters severe aluminium oxidation. Here we report an electrolyte design that can solve this dilemma. By mixing a stable lithium salt LiN(SO F) with dimethyl carbonate solvent at extremely 2 2 high concentrations, we obtain an unusual liquid showing a three-dimensional network of anions and solvent molecules that coordinate strongly to Li ions. This simple formulation of superconcentrated LiN(SO F) /dimethyl carbonate electrolyte inhibits the dissolution of 2 2 both aluminium and transition metal at around 5 V, and realizes a high-voltage LiNi Mn O /graphite battery that exhibits excellent cycling durability, high rate capability 0.5 1.5 4 and enhanced safety. 1 2 Department of Chemical System Engineering, University of Tokyo, 7-3-1, Hongo, Tokyo 113-8656, Japan. Elements Strategy Initiative for Catalysts and 3 4 Batteries, Kyoto University, 1-30, Goryo-Ohara, Kyoto 615-8245, Japan. JST PRESTO, 4-1-8 Honcho Kawaguchi, Saitama 332-0012, Japan. International Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1, Namiki, Tsukuba 305-0044, Japan. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to A.Y. (email: [email protected]). NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ithium-ion batteries, having received great commercial Independent of the solvents used, the viscosity increases expo- success in the portable power source market, are being nentially with increasing the LiFSA mole fraction (X ). LiFSA Laimed for large-scale energy-storage application in electric Among the three groups of solutions, the group with DMC as the 1–3 vehicles . To approach the high energy-density requirements solvent shows the lowest viscosity because pure DMC has a lower for automobiles, a pragmatic approach is to elevate the operating viscosity than pure EC or mixed EC:DMC. For electrolytes with voltage of batteries, from the present 4 V to around 5 V (refs 4,5). similar solvation radiuses of mobile ions, the ionic conductivity is This allows the direct application of the mature fabrication proportional to the number of mobile ions and inversely pro- technology of 4 V-class lithium-ion batteries, the well-developed portional to the viscosity of the medium . As shown in Fig. 1c, at negative electrodes (for example, graphite and graphite/silicon), dilute concentrations of X o0.14 (below 1.5 mol dm ), the LiFSA and high-voltage positive electrodes (for example, spinel use of the EC:DMC mixture shows the highest ionic conductivity LiNi Mn O and some layered oxides). However, new owing to a synergistic effect: the high-dielectric-constant EC 0.5 1.5 4 challenges—which mainly arise from the electrolyte—hinder the increases the mobile ion number by promoting salt dissociation; practical application of the next-generation 5 V-class battery. the low-viscosity DMC increases the ion mobility by decreasing One major problem is metal dissolution from the positive the solution viscosity. This is why the mixed solvents of EC and electrode at high voltages, which poses a serious dilemma in linear carbonates are generally adopted in conventional designing an electrolyte. In state-of-the-art lithium-ion electro- electrolytes of the lithium-ion battery . However, when X LiFSA lyte, chemically unstable LiPF is an essential component to is above 0.14, the solution with DMC as the sole solvent shows an suppress anodic (oxidative) dissolution of an aluminium current even higher ionic conductivity than that with EC:DMC, which collector because its hydrolysis product of hydrofluoric acid (HF) should result from the much lower viscosity of the former at high 6,7 contributes to an insoluble AlF passivation film . However, the concentrations. This result suggests that the viscosity becomes the generated HF accelerates the dissolution of transition metals from decisive factor on the ionic conductivity for a concentrated the active electrode materials, which causes severe capacity decay solution, wherein intensive ionic association exists independent of upon cycling, especially at high voltages and elevated the solvents used, showing a distinct departure from the 8,9 temperatures . Using LiNi Mn O as an example, the conventional electrolyte design strategy on the basis of dilute 0.5 1.5 4 2þ 2þ dissolved Mn and Ni ions, albeit o1% of the total concentrations. For the LiFSA/DMC solution, a commercially amount, deposit on the surface of the graphite negative acceptable ionic conductivity of 1.12 mS cm (30 C) is obtained electrode, which thicken the solid electrolyte interphase (SEI) even at a ‘super-high’ concentration with salt-to-solvent molar by catalysing the reductive decomposition of the electrolyte, and ratio of 1:1.1 despite a high viscosity of 238.9 mPa s. Although the consume the limited lithium reserve in the battery to result in a ionic conductivity is lower than that of the commercial dilute 10,11 450% capacity loss in 100 charge/discharge cycles . electrolyte, it does not compromise the rate capability of the Diversified functional additives and/or alternative solvents have battery (shown later). 12–17 been explored but improvements are still unsatisfactory. On the other hand, the drawbacks of the high volatility and Efforts have tried more stable salts (less tendency to generate HF) high flammability of linear carbonate solvents can be overcome to to replace LiPF , such as lithium perfluorosulfonylamide a large degree owing to the much lower content of organic (shortened to ‘amide’) . However, the chemically stable amide solvents in the concentrated solutions. Thermogravimetry does not participate in the reaction with Al to form a stable measurements (Supplementary Fig. 1) show that the weight loss passivation film, thus causing severe anodic dissolution of the Al of the superconcentrated 1:1.1 LiFSA/DMC solution is only current collector at 44 V (refs 19–22). As a result, it remains a 1.5 wt% after elevating the temperature to 100 C, which is dilemma for electrolyte design to suppress both the Al dissolution considerably lower than those of a dilute 1:10.8 LiFSA/DMC (requiring an unstable salt) and transition metal dissolution solution (65.5 wt%, corresponding to 1.0 mol dm ) and a (avoiding an unstable salt). Recently, increasing the concentration commercial electrolyte (28.7 wt%). As demonstrated in the flame 23–25 of amide salts was reported to alleviate anodic Al dissolution , tests (Fig. 1d,e), the 1:1.1 LiFSA/DMC solution does not burn as but the operating voltage of a half-cell is still limited below 4.3 V, fiercely as the commercial dilute electrolyte. The superior thermal presumably owing to some or all of the following reasons: stability and flame retardant ability of the concentrated electro- 23 24 insufficient salt concentration , too low ionic conductivity and lytes contribute to the remarkably improved safety properties as too low oxidative stability of the solvent . compared with the dilute electrolytes. In this work, we report an electrolyte system to resolve the dilemma. We select stable yet dissociative lithium bis(fluorosul- fonyl)amide (LiFSA) as the salt and oxidation-stable carbonate Reversible reaction of a 5 V-class electrode. Anodic dissolution esters as the solvent. We demonstrate an unusual liquid with a of the Al current collector and/or oxidative decomposition of peculiar three-dimensional structural network obtained at solvent may be encountered in the high-voltage application of extremely high salt concentrations. The superconcentrated amide-based electrolytes. To exclude the possible influence of the electrolyte not only effectively suppresses the anodic Al dissolu- anodic Al dissolution, we initially used platinum foil as the tion but also remarkably inhibits the transition metal dissolution current collector for the LiNi Mn O electrode in a three- 0.5 1.5 4 and, thus, realizes a safe, stable and fast-rate high-voltage electrode cell (Supplementary Fig. 2). The results showed that LiNi Mn O |graphite battery. both dilute (1:10.8) and superconcentrated (1:1.1) LiFSA/DMC 0.5 1.5 4 electrolytes enabled a reversible Li de-intercalation/intercala- tion on the LiNi Mn O |Pt electrode, indicating a reasonably 0.5 1.5 4 Results good compatibility between the present electrolyte system and Physicochemical properties. LiFSA salt was dissolved at various LiNi Mn O material at B5V. 0.5 1.5 4 concentrations into three different carbonate ester solvents: However, when applied in a coin cell using the conventional Al dimethyl carbonate (DMC), ethylene carbonate (EC) and mixed current collector, low concentrations of LiFSA/DMC electrolytes EC:DMC. All the mixtures are transparent liquids at room encountered problems, confirming the critical drawback of temperature (see Fig. 1a as an example). Their basic physico- anodic Al dissolution for the amide-based electrolytes. As shown chemical properties are presented in Supplementary Table 1. in Fig. 2a, the first charge on the LiNi Mn O |Al electrode is 0.5 1.5 4 Figure 1b shows their viscosity as a function of salt concentration. impossible in the dilute 1:10.8 LiFSA/DMC electrolyte owing to 2 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE LiFSA-to-solvent molar ratio a b 1,000 30 °C EC EC:DMC 100 DMC LiFSA/DMC 1 : 10 1 : 2.0 1 : 1.3 1 : 1.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 X mole fraction LiFSA LiFSA-to-solvent molar ratio de –3 1 mol dm 1 : 1.1 LiPF / EC:DMC LiFSA / DMC 30 °C DMC EC:DMC EC 0.0 0.1 0.2 0.3 0.4 0.5 0.6 X mole fraction LiFSA Figure 1 | Physicochemical properties dependent on solution concentration. (a) Images of various salt-to-solvent molar ratios of LiFSA/DMC solutions. Viscosity (b) and ionic conductivity (c) for solutions of LiFSA in DMC, EC and EC:DMC (1:1 by mol.) at 30 C. The X mole fraction is the molar amount LiFSA of LiFSA salt divided by the total molar amount of the salt and solvents. The LiFSA-to-solvent molar ratios of the solutions are shown on the upper axis. (d) Flame tests of a commercial dilute electrolyte of 1.0 mol dm LiPF /EC:DMC (1:1 by vol.) and (e) the lab-made superconcentrated electrolyte of 1:1.1 LiFSA/DMC. the continuous Al dissolution at 4.3 V. In the concentrated 1:1.9 conductivity of the former. These results demonstrate that LiFSA/DMC electrolyte (Fig. 2b), the charge/discharge cycling the salt-superconcentrated strategy is a simple, effective and becomes possible up to the cutoff voltage of 4.9 V, but the large fruitful approach to various safe and stable high-voltage irreversible capacity indicates the parasitic Al dissolution remains. electrolytes. To the best of our knowledge, this is the first time The Al dissolution subsequently deteriorates the electrical that stable and fast charge/discharge cycling of a 5 V-class contacts between the LiNi Mn O material and the Al current electrode using amide salt-based organic electrolytes has been 0.5 1.5 4 collector, and results in a fast capacity decay (Fig. 2d). Actually, achieved. the poor cycling performance on the LiNi Mn O electrode The progressive inhibition of anodic Al dissolution with 0.5 1.5 4 was generally observed in other concentrated 1:2 LiFSA/ increasing salt concentration is further proved by linear carbonate ester electrolytes, such as LiFSA in EC, propylene sweep voltammetry (LSV) of an Al electrode and the subsequent carbonate (PC), ethyl methyl carbonate (EMC) and diethyl scanning electron microscopy observation on the polarized Al carbonate (DEC; Supplementary Fig. 3), indicating this concen- surface (see Fig. 3 for details). This notable concentration tration is not sufficient to fully inhibit Al dissolution at 5 V. effect was recently reported but with a debate on whether a In contrast, the superconcentrated 1:1.1 LiFSA/DMC electro- stable surface film on Al (ref. 23) or the elimination of þ 24,25 lyte enables a reversible Li de-intercalation/intercalation uncoordinated (free) solvents of electrolyte plays the key reaction on the LiNi Mn O electrode even at a high voltage role. We conducted a surface analysis of the Al electrodes 0.5 1.5 4 of 5.2 V (Fig. 2c). In a charge/discharge cycling test, the capacity polarized in various concentrations of LiFSA/DMC electrolytes retention after 100 cycles was over 95% (Fig. 2d), and the by X-ray photoemission spectroscopy (XPS) as well as a coulombic efficiency was close to 100% (Supplementary Fig. 4), comparative LSV study between fresh and polarized Al evidencing an effective inhibition of anodic Al dissolution. electrodes (see Supplementary Figs 5 and 6 for details). We Similarly, using the super-high concentration of 1:1.3, all the were unable to obtain any essential evidence to support the LiFSA/carbonate electrolytes enabled stable charging/discharging existence of a stronger surface film generated in the cycling of the LiNi Mn O electrode (see Fig. 2e for example). concentrated electrolyte. Instead, we found that the LiFSA salt 0.5 1.5 4 Especially, the electrolytes using low-dielectric-constant and low- readily decomposes and produces LiF upon Ar etching viscosity linear carbonate solvents (for example, DMC, EMC and (Supplementary Fig. 7). The previous observation in ref. 23—a DEC) showed a faster rate capability as compared with those much thicker surface film of LiF produced in a higher using high-dielectric-constant and high-viscosity cyclic carbonate concentration of electrolyte—is likely to arise from the solvents (for example, EC, PC and their corresponding mixtures), decomposition of un-rinsed amide salt induced by Ar which is at least partly owing to the much higher ionic etching in the XPS measurement. NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 3 1:22.7 1:10 1:4.9 1:2.9 1:2.0 1:1.3 1:22.7 1:10 1:4.9 1:2.9 1:2.0 1:1.3 –1 Conductivity (mS cm ) Viscosity (mPa S) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 a 5.5 5.5 1:10.8 LiFSA/DMC 1:1.9 LiFSA/DMC 1st 2nd 10th 50th 5.0 5.0 1st 4.5 4.5 1st 50th 10th 4.0 4.0 3.5 3.5 0 50 100 150 200 250 300 0 50 100 150 200 –1 –1 Capacity (mAh g ) Capacity (mAh g ) cd 5.5 LiNi0.5Mn1.5O4 LiFSA/DMC 1:1.1 LiFSA/DMC 2nd 1st 100th 1:1.1 5.0 4.5 4.0 1:1.9 1:10.8 3.5 0 50 100 150 0 20406080 100 –1 Capacity (mAh g ) Cycle number 1/5C 1/2C 1/5C 1:1.3 LiFSA/solvent 1C 2C 5C 10C DMC EMC DEC EC:DMC PC EC 0 1020304050 Cycle number Figure 2 | Performance of 5 V-class LiNi Mn O electrode in a half-cell. Charge–discharge voltage curves of LiNi Mn O |lithium metal half-cells 0.5 1.5 4 0.5 1.5 4 using (a) dilute 1:10.8, (b) moderately concentrated 1:1.9 and (c) superconcentrated 1:1.1 LiFSA/DMC electrolytes at a C/5 rate. Some/all curves of 1st, 2nd, 10th, 50th and 100th cycles are shown. (d) Discharge (Li intercalation) capacity retention of the half-cells using different concentrations of LiFSA/DMC electrolytes at a C/5 rate. (e) Rate capacity and subsequent cycling retention of the half-cells using 1:1.3 LiFSA-based electrolytes with different carbonate solvents. Charge–discharge tests were conducted at 25 C with a cutoff voltage of 3.5–4.9 V and a maximum-time restriction of 10 h except for that using the 1:1.1 LiFSA/DMC electrolyte whose cutoff voltage was 3.5–5.2 V. The 1C-rate corresponds to 147 mA g on the weight basis of the LiNi Mn O 0.5 1.5 4 electrode. We now study the solution structure of the electrolytes using concentration increases, the population of free DMC decreases þ þ Raman spectroscopy observation and density functional theory and that of Li -coordinated DMC increases; the Li -FSA molecular dynamics simulation (DFT-MD). As shown in the association simultaneously intensifies through the formation of Raman spectra (Fig. 4b left), a free DMC molecule exhibits an contact ion pairs (CIPs, FSA coordinating to one Li ) and O-CH stretching vibration band at 910 cm (ref. 27). This aggregate clusters (AGGs, FSA coordinating to two or more 1 þ þ band shifts up to 930–935 cm when DMC participates in Li Li ). The latter is evidenced from a remarkable upshift of the solvation. In dilute 1:10.8 LiFSA/DMC, the majority of DMC FSA band (700–780 cm , Fig. 4b right), which is typically 24,25,28–31 molecules exist in a free state because the solvent-to-salt molar observed in the amide-based concentrated solutions . ratio (10.8) is much larger than a typical four- or fivefold For the moderately concentrated 1:2 LiFSA/DMC solution, the coordination of Li in aprotic solvents. As the LiFSA Raman band corresponding to free DMC is remarkably 4 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications Voltage (V) Voltage (V) –1 Capacity (mAh g ) –1 Capacity (mAh g ) Voltage (V) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE finally, the diffusion of the solvated metal cation to the bulk 0.08 Al electrode electrolyte . At high voltages of B5 V, the first step proceeds 1:1.9 in LiFSA/DMC more rapidly and extensively than at the conventional operating 1:2.9 1:10.8 voltage of 4 V. Thereby, the subsequent coordination and 0.06 diffusion must be strongly inhibited by the nature of electrolyte solutions to suppress the metal ion dissolution. In the moderately concentrated 1:2 LiFSA/carbonate electrolytes, the presence of 0.04 significant free solvents and CIPs (with two or more coordination sites remaining vacant) could coordinate to Al cations and fail to inhibit Al dissolution completely at 5 V. In contrast, the superconcentrated 1:1.1 LiFSA/DMC electrolyte effectively 0.02 inhibits Al dissolution even over 5 V, which can be ascribed to 1:1.1 its peculiar AGGs-predominant solution structure: (i) all DMC solvents and all FSA anions strongly coordinate to Li cations 0.00 and thus have a much lower probability of coordinating to other metal cations; (ii) the resulting reinforced three-dimensional 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 network further retards the diffusion rate of the metal cations, Potential (V) vs Li/Li particularly, those with multiple charge. Figure 3 | Oxidation stability of an aluminium electrode. LSV of an aluminium electrode in various concentrations of LiFSA/DMC electrolytes Stable cycling of a 4.6 V LiNi Mn O |graphite battery.In 0.5 1.5 4 in a three-electrode cell. The scan rate was 1.0 mVs . The insets are addition to the excellent performance achieved on the scanning electron microscopy images of the Al surface polarized in the LiNi Mn O -positive electrode, the superconcentrated 1:1.1 0.5 1.5 4 dilute 1:10.8 (left of panel) and superconcentrated 1:1.1 (right of panel) LiFSA/DMC electrolyte also realized ultra-stable charge/discharge electrolytes. Many corroding pits cover the surface of the Al electrode cycling on the natural graphite-negative electrode despite the polarized in the dilute electrolyte, showing a severe anodic Al dissolution. In absence of conventional SEI-forming agent of EC (Supplementary contrast, no corroding pits appear on the surface of the Al electrode Fig. 9): the application in a graphite|Li half-cell exhibits a capacity polarized in the superconcentrated electrolyte, indicating a good inhibition retention of 99.6% after 100 cycles with coulombic efficiency of anodic Al dissolution. The white scale bar represents 20mm. of B99.8%, and with rate capability comparable with that using a commercial dilute electrolyte. Accordingly, the super- concentrated electrolyte was further applied in the high-voltage weakened, suggesting that the majority of DMC are solvating to LiNi Mn O |graphite full cell, a much harsher condition than 0.5 1.5 4 Li . This is consistent with the DFT-MD simulation, which in the half-cell, because the active lithium resource is limited and shows ca. 90% DMC are coordinating to Li with the rest as free a new underlying problem arises from the transition metal solvent (marked as light blue in Fig. 4d). Moreover, the dissolution from the LiNi Mn O especially at high voltages 0.5 1.5 4 simulation illustrates that all FSA anions are coordinating to and elevated temperatures. Figure 5a,b shows charge/discharge Li with ca. 20% as CIPs and ca. 80% as AGGs (marked as voltage profiles of LiNi Mn O |natural graphite full cells at 0.5 1.5 4 orange and dark blue in Fig. 4d, respectively). The coordination 40 C using a state-of-the-art commercial electrolyte and the environment is shown in Supplementary Fig. 8. For the lab-made superconcentrated electrolyte, respectively. The cell superconcentrated 1:1.1 LiFSA/DMC solution, both DMC and with the commercial electrolyte suffers from a severe capacity FSA bands further upshift substantially, indicating both Li - decay during cycling, that is, only 18% of the initial capacity left DMC and Li -FSA interactions enhanced compared with after 100 cycles (Fig. 5a,c), which is consistent with previous 11,33 those in 1:2 LiFSA/DMC. The DFT-MD simulation reveals that reports . In contrast, the capacity retention using the 1:1.1 all DMC molecules, together with all FSA anions, are LiFSA/DMC electrolyte is over 90% after 100 cycles, exhibiting a coordinating to Li (no free solvent or anion). Interestingly, remarkably improved cycling durability (Fig. 5b,c and besides oxygen, significant amount of nitrogen on FSA anions Supplementary Figs 10 and 11). Notably, the superiority of the also participate in the coordination with Li , which is hardly superconcentrated electrolyte becomes even more marked at observed at the lower concentrations. The contribution of 55 C (Supplementary Fig. 12). It is generally accepted that the nitrogen coordinating to Li is shown in Supplementary Fig. 8. poor cycling performance of the LiNi Mn O |graphite battery 0.5 1.5 4 More importantly, almost all FSA anions remain in AGG states originates from the dissolution of transition metals from during the whole DFT-MD simulation time (0.1 fs 100,000 LiNi Mn O into the electrolyte, as introduced at the 0.5 1.5 4 steps), and a CIP state is rarely observed, demonstrating the beginning of this article. Indeed, energy dispersive X-ray unusual solution structural feature with AGG clusters as the spectroscopy (EDS) observation shows a much lower content of predominant components in the superconcentrated LiFSA/DMC Mn and Ni on the graphite electrode of the full cell cycled in the solution. It is noteworthy that each FSA anion coordinates to superconcentrated electrolyte than that in the commercial þ þ 2–3 Li and each Li is coordinated by 2–3 FSA in 1:1.1 electrolyte, which provides evidence for the effective inhibition LiFSA/DMC. Hence, FSA anions in the superconcentrated of transition metal dissolution in the former. There are two main solution connect with each other through the intensive reasons for the improved performance: (i) LiFSA is less reactive to association with Li , leading to a reinforced three-dimensional produce HF as compared with LiPF , which alleviates the network (shown in Fig. 4e). This feature is different from the less corrosion of LiNi Mn O and, thus, reduces the formation of 0.5 1.5 4 2þ 2þ 2þ 2þ concentrated solutions, wherein significant amount of CIPs and soluble Mn and Ni ; (ii) even if some Mn and Ni are free solvents divide the solution structure into relatively small-size formed on the surface of LiNi Mn O , they can hardly dissolve 0.5 1.5 4 parts. in and transport through the AGGs-predominant super- Generally, the anodic metal dissolution requires three steps: concentrated electrolyte owing to the same functional manner first, oxidation of the metal to a metal cation; second, for the inhibition of Al dissolution. Moreover, it is worth coordination of the metal cation by solvents or anions; and noting that the rate performance of the full cell using the NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 5 –2 Current (mA cm ) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 O c S S F F O O Free DMC Free FSA Li Li O O Li + – CIP: 1Li ···FSA S S + − AGG-I: 2Li ···FSA F F O O + − AGG-II: 3Li ···FSA O O Li + + − Li -coordinated DMC Li -associated FSA (three main positions) LiFSA:DMC 1 : 1.1 1 : 1.9 1 : 2.9 1 : 10.8 Pure DMC 960 940 920 900 780 760 740 720 700 –1 Raman shift (cm ) Figure 4 | Li salt solvent coordination structure dependent on salt concentration. (a) The several main species in the LiFSA/DMC solutions. (b) Raman spectra of LiFSA/DMC solutions with various salt-to-solvent molar ratios in the range of 890–900 cm (O-CH stretching mode of the DMC solvent) and 700–780 cm (S-N stretching mode of the FSA anion). Snapshots of typical equilibrium trajectories obtained by DFT-MD simulations: (c) 3  3 dilute solution (1 LiFSA/25 DMC, o1 mol dm ), (d) moderately concentrated solution (12 LiFSA/24 DMC, ca. 4 mol dm ), and (e) superconcentrated 3 þ þ solution (10 LiFSA/11 DMC, ca. 5.5 mol dm ). The coordination of Li  DMC and Li  FSA is supposed to build up when the involved atoms locate þ þ within 2.5 Å from Li . The coordination numbers of solvents and anions to Li are shown in Supplementary Fig. 8. Li cations are marked in purple. Free and coordinated DMC molecules are marked in light blue and grey, respectively. Free, CIP and AGG states of FSA anions are marked in red, orange and dark blue, respectively. Hydrogen atoms are not shown. superconcentrated 1:1.1 LiFSA/DMC is comparable with that Discussion using the commercial electrolyte (Fig. 5d), although the former The conventional dilute LiPF /EC-based electrolytes have domi- shows an ionic conductivity one-order lower than the latter. The nated the electrolyte market of 4 V-class lithium-ion batteries mechanistic understanding on the high-rate capability of the over the past 25 years; however, they show difficulties in satisfying superconcentrated electrolyte is underway in our laboratory. To the requirements of next-generation 5 V-class batteries in terms the best of our knowledge, this is the first case that, an electrolyte of both safety and stability. Our work demonstrates a number of with such an ultra-simple formulation—a single salt and a single electrolytes with a reinforced three-dimensional network that are solvent without any additive—realizes stable cycling of a high- obtained by simple mixing of a stable salt with a conventional voltage lithium-ion battery. carbonate solvent at ‘super-high’ concentrations. Owing to its 6 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications Coordinated DMC Free DMC AGG-II AGG-I CIP Free FSA Intensity (a.u.) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE ab 40 °C 40 °C 1:1.1 LiFSA/DMC Commercial electrolyte 5.0 5.0 100th 10th 2nd 100th 50th 2nd 4.5 4.5 4.0 4.0 3.5 3.5 0 30 60 90 120 0 30 60 90 120 –1 –1 Capacity (mAh g ) Capacity (mAh g ) c d 40 °C 2C 5C C/5 C/2 1C C/5 1:1.1 LiFSA/DMC 1:1.1 LiFSA/DMC Mn Kα Ni Kα 5.4 6.0 6.6 7.2 7.8 Commercial Energy (keV) Commercial 25 °C 0 30 0 20 40 60 80 100 0369 12 15 18 Cycle number Cycle number Figure 5 | Performance of a high-voltage LiNi Mn O |natural graphite battery. Charge–discharge voltage curves of LiNi Mn O |graphite full cells 0.5 1.5 4 0.5 1.5 4 using (a) a commercial 1.0 mol dm LiPF /EC:DMC (1:1 by vol.) electrolyte and (b) lab-made superconcentrated 1:1.1 LiFSA/DMC electrolyte at a C/5 rate and 40 C. The curves of 2nd, 10th, 50th and 100th cycle are shown. (c) Discharge capacity retention of the full cells at a C/5 rate. The inset shows EDS spectra on the graphite electrode surface (200 200mm area) after 8-day cycling tests, which is equivalent to the operating time of 100 and 20 cycles for the battery using the commercial and superconcentrated electrolytes, respectively. (d) Discharge capacity of the full cell at various C-rates and 25 C. All charge-discharge cycling tests were conducted with a cutoff voltage of 3.5–4.8 V. 1C-rate corresponds to 147 mA g on the weight basis of the LiNi Mn O electrode. 0.5 1.5 4 peculiar structural characteristics, the superconcentrated electro- superconcentrated electrolytes do not contain any additives, lytes overcomes the longstanding challenge faced by the unstable signifying the potential to further enhance the performance. LiPF -based electrolytes at high voltages (passivating the Al These desirable features above outperform the conventional dilute current collector versus accelerating the transition metal dissolu- electrolytes; meanwhile, the wide-temperature window of the tion of the active material), thus, enables a stable operation of a liquid state (ensuring a good contact with the electrode 5 V-class battery. Emphasis is on the fact that the peculiar materials), as well as the convenience of the approach, surpass solution structure and functionalities are unique to such super- the solid-state electrolytes. Therefore, the superconcentrated high concentrations (solvent/saltE1.1), and cannot be achieved electrolytes might offer opportunities to build safe and stable in moderately high concentrations (solvent/salt41.8) as in high-voltage batteries that are not limited to the lithium-ion. 24,30,31,34,35 previous reports . Besides the positive electrode side, the superconcentrated electrolytes also show a good compatibility Methods with the natural graphite-negative electrode even in the absence Preparation of electrolytes and electrodes. LiFSA (Nippon shokubai) and all of EC. It breaks through the limitation of a general requirement of solvents (DMC, DEC, EMC, EC and PC, Kishida Chemical Co. Ltd) were lithium EC for a SEI formation for a lithium-ion electrolyte, and battery grade and used without purification. Electrolyte solutions were prepared by mixing a given amount of LiFSA with solvents in an Ar-filled glove box. The diversifies the electrolyte formulation towards various EC-free commercial electrolyte of 1.0 mol dm LiPF /EC:DMC (1:1 by vol) was electrolytes. Different from the conventional electrolyte design purchased from Kishida Chemical Co. Ltd and used as the reference. Both the that requires a high-dielectric-constant (usually high-viscosity) lab-made LiFSA-based electrolytes and as-received commercial LiPF -based solvent, the superconcentrated electrolyte prefers a low-viscosity electrolyte were dried by molecular sieve before tests. The water content was less than 2 p.p.m., as detected by a coulometric Karl Fischer Titrator. solvent. Although the ionic conductivity of the superconcentrated The electrodes were fabricated by first well mixing the active materials of electrolyte is lower than that of the conventional dilute electrolyte, LiNi Mn O (Hosen Corp., mean particle size R¼ 5mm, no surface treatment) 0.5 1.5 4 it does not necessarily compromise the rate capability of the and natural graphite (SEC Carbon Ltd., R¼ 10mm), polyvinylidene difluoride battery. Clarification of the corresponding mechanism would be (PVdF) and/or Denka black (AB, HS-100) in N-methylpyrrolidone with weight ratios of 80:10:10 (LiNi Mn O :PVdF:AB) and 90:10 (graphite:PVdF). The enlightening for developing novel high-power batteries. 0.5 1.5 4 resultant slurry was cast on the Al or Pt foil (20mm thickness) for the Furthermore, the superconcentrated electrolytes show superior LiNi Mn O electrode and on the Cu foil (10mm thickness) for the graphite 0.5 1.5 4 thermal stability and flame retardant ability, alleviating the safety electrode using a 50mm doctor blade. All those electrodes were dried at 120 C risk for a high-voltage battery using conventional dilute under vacuum for 12 h. The active material mass loading was 0.7–2 mg cm with electrolytes. Finally, it is noteworthy that our reported a thickness of B15–20mm, unless otherwise mentioned. The use of relatively low NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 7 –1 Capacity (mAh g ) Voltage (V) –1 Capacity (mAh g ) Voltage (V) ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 mass loading was to spotlight the critical issue of anodic Al dissolution as the 11. Pieczonka, N. P. W. et al. Understanding transition-metal dissolution behavior content ratio of metallic Al components (Al current collector and Al positive cap) in LiNi Mn O high-voltage spinel for lithium ion batteries. J. Phys. Chem. C 0.5 1.5 4 to the active electrode material becomes much higher in a coin cell. Nevertheless, 117, 15947–15957 (2013). thick electrodes with a high mass loading of B10 mg cm were also tested. The 12. Zhang, Z. et al. Fluorinated electrolytes for 5 V lithium-ion battery chemistry. results are shown in Supplementary Figs 11 and 12. Energy Environ. Sci. 6, 1806–1810 (2013). 13. Hu, L., Zhang, Z. & Amine, K. Fluorinated electrolytes for Li-ion battery: An FEC-based electrolyte for high voltage LiNi Mn O /graphite couple. Electrochemical measurements. LSV was performed by VMP-3 (BioLogic) in a 0.5 1.5 4 Electrochem. Commun. 35, 76–79 (2013). beaker cell with an Al belt (1 4cm , 0.6 cm soaked in the electrolyte) as a working electrode and lithium metal as the reference and counter electrodes (shown in 14. Pieczonka, N. P. W. et al. Impact of lithium bis(oxalate)borate electrolyte Supplementary Fig. 5 inset). LiNi Mn O |Li half-cells and LiNi Mn O |- additive on the performance of high-voltage spinel/graphite Li-ion batteries. 0.5 1.5 4 0.5 1.5 4 graphite full cells were assembled in the standard 2032-type coin cell hardware in J. Phys. Chem. C 117, 22603–22612 (2013). an Ar-filled glove box. A combined separator, composed of cellulose separator 15. Song, Y.-M., Han, J.-G., Park, S., Lee, K. T. & Choi, N.-S. A multifunctional (Nippon Kodoshi, placed on the positive electrode side) and glass fibre separator phosphite-containing electrolyte for 5 V-class LiNi Mn O cathodes with 0.5 1.5 4 (Advantec GB-50, placed on the negative electrode side), was used. The amount of superior electrochemical performance. J. Mater. Chem. A 2, 9506–9513 (2014). electrolyte in a coin cell was ca 160ml to fully wet the separators and electrodes. 16. Xu, M. et al. Improved performance of high voltage graphite/LiNi Mn O 0.5 1.5 4 In the full cells, the weight ratio of LiNi Mn O :graphite was B2:1, which 0.5 1.5 4 batteries with added lithium tetramethyl borate. ECS Electrochem. Lett. 4, corresponds to B1:1.3 of their theoretical capacity ratio. Galvanostatic charge/ A83–A86 (2015). discharge cycling and rate capability tests were conducted on a charge/discharge 17. Xu, K. Electrolytes and Interphases in Li-Ion Batteries and Beyond. Chem. Rev. unit (TOSCAT). Charge and discharge were conducted at the same C-rate without 114, 11503–11618 (2014). using a constant-voltage mode at both ends of the charge and discharge. 18. Xu, K. Nonaqueous liquid electrolytes for lithium-based rechargeable batteries. Chem. Rev. 104, 4303–4417 (2004). 19. Krause, L. J. et al. Corrosion of aluminum at high voltages in non-aqueous Characterization. The density and viscosity of solution samples were evaluated electrolytes containing perfluoroalkylsulfonyl imides; new lithium salts for with a DMA 35 density meter and a Lovis 2000 M viscometer, respectively. lithium-ion cells. J. Power Sources 68, 320–325 (1997). The ionic conductivity was measured by AC impedance spectroscopy at 1 kHz 20. Wang, X., Yasukawa, E. & Mori, S. Inhibition of anodic corrosion of aluminum (Solartron 147055BEC) in a symmetric cell (Pt|electrolyte|Pt). The flammability cathode current collector on recharging in lithium imide electrolytes. was tested on an electrolyte-soaked glass fibre filter (Advantec GB-100). Electrochim. Acta 45, 2677–2684 (2000). The solution structure was studied by a Raman spectroscopy with an exciting 21. Li, L. et al. Transport and electrochemical properties and spectral features of laser of 514 nm (NRS-5100). The samples were sealed in a quartz cell in the glove box to avoid any contamination from the air. non-aqueous electrolytes containing LiFSI in linear carbonate solvents. The morphology of Al electrodes after LSV tests were observed by a field- J. Electrochem. Soc. 158, A74–A82 (2011). emission scanning electron microscopy at 2.0 kV. The transition metals deposited 22. Kramer, E. et al. Mechanism of anodic dissolution of the aluminum current on the graphite electrode in the LiNi Mn O |graphite full cells after charge/ collector in 1 M LiTFSI EC:DEC 3:7 in rechargeable lithium batteries. 0.5 1.5 4 discharge cycling were estimated by an EDS. The cells were disassembled in the J. Electrochem. Soc. 160, A356–A360 (2013). glove box. The obtained electrodes were rinsed in DMC and dried in the glove box. 23. Matsumoto, K. et al. Suppression of aluminum corrosion by using high The sample was exposed in air for o1 min at sample loading. concentration LiTFSI electrolyte. J. Power Sources 231, 234–238 (2013). The experimental details for thermogravimetric analysis and XPS 24. McOwen, D. W. et al. Concentrated electrolytes: decrypting electrolyte measurements are shown in Supplementary Fig. 1 legend and Supplementary properties and reassessing Al corrosion mechanisms. Energy Environ. Sci. 7, Methods, respectively. 416–426 (2014). 25. Moon, H. et al. Solvent activity in electrolyte solutions controls electrochemical reactions in Li-ion and Li-sulfur batteries. J. Phys. Chem. C 119, 3957–3970 Simulations. Car-Parrinello type DFT-MD simulations were carried out using 36 (2015). CPMD code . LiFSA/DMC solutions with salt-to-solvent molar ratios of 1:25, 1:2 26. Yoshida, K. et al. Oxidative-stability enhancement and charge transport and 1:1.1 were calculated in cubic supercells with 15.05, 17.03 and 14.34 Å linear mechanism in glyme-lithium salt equimolar complexes. J. Am. Chem. Soc. 133, dimensions, respectively. A fictitious electric mass of 500 a.u. and a time step of 4 13121–13129 (2011). a.u. (0.10 fs) were chosen. The temperature was controlled using a Nose´ thermostat 27. Katon, J. E. & Cohen, M. D. The vibrational spectra and structure of dimethyl with a target temperature of 30 C. After 5 ps equilibration steps, statistical averages carbonate and its conformational behavior. Can. J. Chem. 53, 1378–1386 (1975). were computed from trajectories of at least 10 ps in length. The electronic wave 28. Seo, D. M., Borodin, O., Han, S.-D., Boyle, P. D. & Henderson, W.A. Electrolyte function was quenched to the Born-Oppenheimer surface approximately every 1 ps solvation and ionic association II. Acetonitrile-lithium salt mixtures: highly to maintain adiabaticity. The energy cutoff of the plane wave basis is set to 90 Ry. dissociated salts. J. Electrochem. Soc. 159, A1489–A1500 (2012). Goedecker–Teter–Hutter type norm-conserving pseudopotentials for C, H, O, N, S, 29. Yamada, Y., Yaegashi, M., Abe, T. & Yamada, A. A superconcentrated ether F and Li were used . electrolyte for fast-charging Li-ion batteries. Chem. Commun. 49, 11194–11196 (2013). Data availability. The data that support the findings of this study are available 30. Yamada, Y. et al. Unusual stability of acetonitrile-based superconcentrated from the corresponding author upon request. electrolytes for fast-charging lithium-ion batteries. J. Am. Chem. Soc. 136, 5039–5046 (2014). 31. Yamada, Y. & Yamada, A. Review-Superconcentrated electrolytes for lithium References batteries. J. Electrochem. Soc. 162, A2406–A2423 (2015). 1. Goodenough, J. B. & Kim, Y. Challenges for rechargeable Li batteries. Chem. 32. Yamada, Y. et al. Corrosion prevention mechanism of aluminum metal in Mater. 22, 587–603 (2010). superconcentrated electrolytes. ChemElectroChem 2, 1687–1694 (2015). 2. Etacheri, V., Marom, R., Elazari, R., Salitra, G. & Aurbach, D. Challenges in the 33. Glatthaar, S., Dorit, N., Binder, J. R. & Andres, H. Evidence of loss of active development of advanced Li-ion batteries: a review. Energy Environ. Sci. 4, lithium in titanium-doped LiNi Mn O /graphite cells. J. Power Sources 274, 0.5 1.5 4 3243–3262 (2011). 1267–1275 (2015). 3. Amine, K., Kanno, R. & Tzeng, Y. Rechargeable lithium batteries and beyond: 34. Jeong, S.-K., Inaba, M., Iriyama, Y., Abe, T. & Ogumi, Z. Electrochemical Progress, challenges, and future directions. MRS Bull. 39, 395–401 (2014). intercalation of lithium ion within graphite from propylene carbonate 4. Patoux, S. et al. High voltage spinel oxides for Li-ion batteries: From the solutions. Electrochem. Solid State Lett. 6, A13–A15 (2003). material research to the application. J. Power Sources 189, 344–352 (2009). 35. Suo, L., Hu, Y.-S., Li, H., Armand, M. & Chen, L. A new class of solvent-in-salt 5. Croy, J. R., Abouimrane, A. & Zhang, Z. Next-generation lithium-ion batteries: electrolyte for high-energy rechargeable metallic lithium batteries. Nat. the promise of near-term advancements. MRS Bull. 39, 407–415 (2014). Commun. 4, 1481 (2013). 6. Myung, S.-T., Sasaki, Y., Sakurada, S., Sun, Y.-K. & Yashiro, H. Electrochemical 36. CPMD. http://www.cpmd.org:81/manual/node4.html, Copyright IBM Corp behavior of current collectors for lithium batteries in non-aqueous alkyl carbonate (1990-2015), Copyright MPI fu¨r Fesko¨rperforschung Stuttgart (1997-2001). solution and surface analysis by ToF-SIMS. Electrochim. Acta 55, 288–297 (2009). 37. Goedecker, S., Teter, M. & Hutter, J. Separable dual-space Gaussian 7. Zhang, X. & Devine, T. M. Identity of passive film formed on aluminum in Li- pseudopotentials. Phys. Rev. B 54, 1703–1710 (1996). ion battery electrolytes with LiPF6. J. Electrochem. Soc. 153, B344–B351 (2006). 8. Aurbach, D. et al. Review on electrode–electrolyte solution interactions, related to cathode materials for Li-ion batteries. J. Power Sources 165, 491–499 (2007). Acknowledgements 9. Zhan, C. et al. Mn(II) deposition on anodes and its effects on capacity fade in This work was partially supported by JSPS Grant-in-Aid for Young Scientists (A) spinel lithium manganate-carbon systems. Nat. Commun. 4, 2437 (2013). (No. 26708030) and JSPS Specially Promoted Research (No. 15H05701). The calculations 10. Kim, J.-H. et al. Understanding the capacity fading mechanism in were carried out at the super-computer centres of National Institute for Materials LiNi Mn O /graphite Li-ion batteries. Electrochim. Acta 90, 556–562 (2013). Science, the University of Tokyo, and the K computer at the RIKEN through the HPCI 0.5 1.5 4 8 NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12032 ARTICLE System Research Projects (hp150209). We thank Keisuke Kikuchi, Reiko Kawakami Competing financial interests: The authors declare no competing financial interests. and Dr Kouhei Okitsu for their assistance in the experiments, and specially thank Reprints and permission information is available online at http://npg.nature.com/ Dr Sai-Cheong Chung for his valuable suggestions on the manuscript. reprintsandpermissions/ How to cite this article: Wang, J. et al. Superconcentrated electrolytes for a high-voltage Author contributions lithium-ion battery. Nat. Commun. 7:12032 doi: 10.1038/ncomms12032 (2016). J.W. and Y.Y. contributed equally to this work. Y.Y. and A.Y. proposed the concept. J.W. and Y.Y. designed the experiments. J.W. and C.H.C. carried out the experiments and analysed the data. K.S. and Y.T. designed and conducted the theoretical calculations. This work is licensed under a Creative Commons Attribution 4.0 J.W., Y.Y. and A.Y. wrote the manuscript. A.Y. supervised the whole project. International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, Additional information users will need to obtain permission from the license holder to reproduce the material. Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ NATURE COMMUNICATIONS | 7:12032 | DOI: 10.1038/ncomms12032 | www.nature.com/naturecommunications 9

Journal

Nature Communications – Springer Journals

Published: Jun 29, 2016

There are no references for this article.