The use of ‘water-in-salt’ electrolytes has considerably expanded the electrochemical window of aqueous lithium-ion batteries to 3 to 4 volts, making it possible to couple high-voltage cathodes with low-potential graphite anodes1,2,3,4. However, the limited lithium intercalation capacities (less than 200 milliampere-hours per gram) of typical transition-metal-oxide cathodes5,6 preclude higher energy densities. Partial7,8 or exclusive9 anionic redox reactions may achieve higher capacity, but at the expense of reversibility. Here we report a halogen conversion–intercalation chemistry in graphite that produces composite electrodes with a capacity of 243 milliampere-hours per gram (for the total weight of the electrode) at an average potential of 4.2 volts versus Li/Li+. Experimental characterization and modelling attribute this high specific capacity to a densely packed stage-I graphite intercalation compound, C3.5[Br0.5Cl0.5], which can form reversibly in water-in-bisalt electrolyte. By coupling this cathode with a passivated graphite anode, we create a 4-volt-class aqueous Li-ion full cell with an energy density of 460 watt-hours per kilogram of total composite electrode and about 100 per cent Coulombic efficiency. This anion conversion–intercalation mechanism combines the high energy densities of the conversion reactions, the excellent reversibility of the intercalation mechanism and the improved safety of aqueous batteries.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The data that support the findings of this study are available from the corresponding author on request.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Suo, L. et al. “Water-in-salt” electrolyte enables high-voltage aqueous lithium-ion chemistries. Science 350, 938–943 (2015).
Yang, C. et al. 4.0 V aqueous Li-ion batteries. Joule 1, 122–132 (2017).
Suo, L. et al. Advanced high-voltage aqueous lithium-ion battery enabled by “water-in-bisalt” electrolyte. Angew. Chem. Int. Ed. 55, 7136–7141 (2016).
Yang, C. et al. Unique aqueous Li-ion/sulfur chemistry with high energy density and reversibility. Proc. Natl Acad. Sci. USA 114, 6197–6202 (2017).
Larcher, D. & Tarascon, J. M. Towards greener and more sustainable batteries for electrical energy storage. Nat. Chem. 7, 19–29 (2015).
Radin, M. D. et al. Narrowing the gap between theoretical and practical capacities in Li-ion layered oxide cathode materials. Adv. Energy Mater. 7, 1602888 (2017).
Zhan, C. et al. Enabling the high capacity of lithium-rich anti-fluorite lithium iron oxide by simultaneous anionic and cationic redox. Nat. Energy 2, 963–971 (2017).
Lee, J. et al. Reversible Mn2+/Mn4+ double redox in lithium-excess cathode materials. Nature 556, 185–190 (2018).
Bruce, P. G., Freunberger, S. A., Hardwick, L. J. & Tarascon, J. M. Li-O2 and Li-S batteries with high energy storage. Nat. Mater. 11, 19–29 (2012); erratum 11, 172 (2012).
Seel, J. A. & Dahn, J. R. Electrochemical intercalation of PF6 into graphite. J. Electrochem. Soc. 147, 892–898 (2000).
Milliken, J. W. & Fischer, J. E. Ionic salt limit in graphite–fluoroarsenate intercalation compounds. J. Chem. Phys. 78, 5800–5808 (1983).
Wu, F. & Yushin, G. Conversion cathodes for rechargeable lithium and lithium-ion batteries. Energy Environ. Sci. 10, 435–459 (2017).
Axdal, S. H. A. & Chung, D. D. L. Kinetics and thermodynamics of intercalation of bromine in graphite—I. Experimental. Carbon 25, 191–210 (1987).
Erbil, A., Dresselhaus, G. & Dresselhaus, M. S. Raman scattering as a probe of structural phase transitions in the intercalated graphite–bromine system. Phys. Rev. B 25, 5451–5460 (1982).
Anthonsen, J. The Raman spectra of some halogen gas hydrates. Acta Chem. Scand. A 29a, 175–178 (1975).
Stammreich, H. & Forneris, R. The Raman frequency of bromine monochloride. J. Chem. Phys. 21, 944–945 (1953).
Heald, S. M. & Stern, E. A. Extended-x-ray-absorption-fine-structure study of the Br2-graphite system. Phys. Rev. B 17, 4069–4081 (1978).
Huggins, F. E. & Huffman, G. P. Chlorine in coal: an XAFS spectroscopic investigation. Fuel 74, 556–569 (1995).
Heald, S. M. & Stern, E. A. EXAFS study of Br2-graphite intercalation compounds. Synth. Met. 1, 249–255 (1980).
Selig, H. & Ebert, L. B. Graphite intercalation compounds. Adv. Inorg. Chem. Radiochem. 23, 281–327 (1980).
Dresselhaus, M. S. & Dresselhaus, G. Intercalation compounds of graphite. Adv. Phys. 30, 139–326 (1981).
Cao, J. et al. Two-step electrochemical intercalation and oxidation of graphite for the mass production of graphene oxide. J. Am. Chem. Soc. 139, 17446–17456 (2017).
Lin, M. C. et al. An ultrafast rechargeable aluminium-ion battery. Nature 520, 324–328 (2015).
Leung, S. Y. et al. Structural studies of graphite intercalation compounds using (00l) x-ray diffraction. Phys. Rev. B 24, 3505–3518 (1981).
Furdin, G., Bach, B. & Herold, A. Contribution à l’étude du système ternaire graphite–brome–chlore. C. R. Acad. Sci. Paris C 271, 683 (1970).
Eeles, W. & Turnbull, J. The crystal structure of graphite-bromine compounds. Proc. R. Soc. Lond. A 283, 179–193 (1965).
Sasa, T., Takahashi, Y. & Mukaibo, T. Crystal structure of graphite bromine lamellar compounds. Carbon 9, 407–416 (1971).
Chung, D. Structure and phase transitions of graphite intercalated with bromine. Phase Transit. 8, 35–57 (1986).
Placke, T. et al. In situ X-ray diffraction studies of cation and anion intercalation into graphitic carbons for electrochemical energy storage applications. Z. Anorg. Allg. Chem. 640, 1996–2006 (2014).
Titantah, J. T., Lamoen, D., Schowalter, M. & Rosenauer, A. Density-functional theory calculations of the electron energy-loss near-edge structure of Li-intercalated graphite. Carbon 47, 2501–2510 (2009).
Wen, C. J., Boukamp, B. A., Huggins, R. A. & Weppner, W. Thermodynamic and mass transport properties of “LiAl”. J. Electrochem. Soc . 126, 2258–2266 (1979).
Barsoukov, E. & Macdonald, J. R. Impedance Spectroscopy: Theory, Experiment, and Applications (John Wiley & Sons, Hoboken, 2018).
Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 12, 537–541 (2005).
Newville, M. EXAFS analysis using FEFF and FEFFIT. J. Synchrotron Radiat. 8, 96–100 (2001).
Bunker, G. Introduction to XAFS: A Practical Guide to X-ray Absorption Fine Structure Spectroscopy (Cambridge Univ. Press, Cambridge, 2010).
Starovoytov, O. N., Borodin, O., Bedrov, D., & Smith, G. D. Development of a polarizable force field for molecular dynamics simulations of poly (ethylene oxide) in aqueous solution. J. Chem. Theory Comput. 7, 1902–1915 (2011).
Borodin, O. et al. Liquid structure with nano-heterogeneity promotes cationic transport in concentrated electrolytes. ACS Nano 11, 10462–10471 (2017).
Borodin, O. Polarizable force field development and molecular dynamics simulations of ionic liquids. J. Phys. Chem. B 113, 11463–11478 (2009).
Zaytsev, I. D. & Aseyev, G. G. Properties of Aqueous Solutions of Electrolytes. (CRC Press, Boca Raton, 1992).
Dünweg, B. & Kremer, K. Molecular dynamics simulation of a polymer chain in solution. J. Chem. Phys. 99, 6983–6997 (1993).
Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953 (1994).
Kresse, G. & Hafner, J. Ab initio molecular-dynamics simulation of the liquid-metal–amorphous-semiconductor transition in germanium. Phys. Rev. B 49, 14251 (1994).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).
Klimeš, J., Bowler, D. R. & Michaelides, A. Chemical accuracy for the van der Waals density functional. J. Phys. Condens. Matter 22, 022201 (2009).
Ghosh, D. & Chung, D. D. L. Two-dimensional structure of bromine intercalated graphite. Mater. Res. Bull. 18, 1179–1187 (1983).
Henkelman, G., Arnaldsson, A. & Jónsson, H. A fast and robust algorithm for Bader decomposition of charge density. Comput. Mater. Sci. 36, 354–360 (2006).
Momma, K. & Izumi, F. VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data. J. Appl. Cryst. 44, 1272–1276 (2011).
Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H–Pu. J. Chem. Phys. 132, 154104 (2010).
Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem. 32, 1456–1465 (2011).
Zhang, Y. & Yang, W. Comment on “Generalized gradient approximation made simple”. Phys. Rev. Lett. 80, 890 (1998).
VandeVondele, J. & Hutter, J. Gaussian basis sets for accurate calculations on molecular systems in gas and condensed phases. J. Chem. Phys. 127, 114105 (2007).
Goedecker, S., Teter, M. & Hutter, J. Separable dual-space Gaussian pseudopotentials. Phys. Rev. B 54, 1703–1710 (1996).
Hartwigsen, C., Gœdecker, S. & Hutter, J. Relativistic separable dual-space Gaussian pseudopotentials from H to Rn. Phys. Rev. B 58, 3641 (1998).
Frisch, M. J. et al. Gaussian 09, revision A. 1 http://gaussian.com/glossary/g09/ (2009).
Francl, M. M. et al. Self-consistent molecular orbital methods. XXIII. A polarization-type basis set for second-row elements. J. Chem. Phys. 77, 3654–3665 (1982).
The principal investigators (C.W. and K.X.) received financial support from the US Department of Energy (DOE) through ARPA-E grant DEAR0000389. O.B. and T.P. acknowledge support from ARL Enterprise for Multiscale Modelling. This research used resources of the Advanced Photon Source, an Office of Science User Facility operated for the US DOE Office of Science by Argonne National Laboratory, and was supported by the US DOE under contract number DE-AC02-06CH11357 and by the Canadian Light Source and its funding partners.
Nature thanks Chao Gao, Jintao Zhang and the other anonymous reviewer(s) for their contribution to the peer review of this work.
Extended data figures and tables
a, Hydration of anhydrous LiBr/LiCl mixed salts (20 mg, molar ratio 1:1) in 2 g (left) and 5.1 g (right) of WiBS liquid electrolyte. b, Hydration of LiBr/LiCl monohydrate mixed salts (20 mg, molar ratio 1:1) in 1.5 g (left) and 3.0 g (right) of WiBS liquid electrolyte. The solid residue is indicated by the dashed circle. c, Immiscibility (clear phase separation) of an as-prepared mixture of an aqueous solution (top) of LiBr·3H2O (0.8 g) and LiCl·3H2O (0.4 g) in WiBS liquid electrolyte (bottom; 3 g). d, MD simulation cell snapshots, showing the initial configuration (left) and the final configuration (right) after a 9-ns run for a solution of 18 mol kg−1 LiBr and 21 mol kg−1 LiTFSI in H2O at 363 K; the Br− anions are highlighted in red. e, Only trace concentrations (<35 p.p.m.) of Br− and Cl− are detected in the WiBS liquid electrolyte by anion-exchange liquid chromatography after 500 h of exposure to the LiBr/LiCl solution.
a–c, Scanning electron microscope (a) and energy-dispersive X-ray spectroscopy mapping (b, c) images of an LBC-G composite cathode, showing the morphology and elemental distributions of Br (b) and Cl (c) in the cathode layer after 5 full cycles. The distributions of Br and Cl are overlapping, indicating that the two salts are well mixed as a result of their close association during co-intercalation/de-intercalation. d, Ex situ XRD patterns of LBC-G cathodes collected from disassembled cells after the 5th charge and discharge. The disappearance of the LiBr and LiCl peaks and the appearance of the GIC peaks of the LBC-G cathode confirm the BrCl intercalation reaction at the fully charged state, whereas the typical patterns of crystalline LiBr and LiCl at the fully discharged state suggest that solid LiBr and LiCl are reformed after de-intercalation of halogen anions from graphite. The (002) peak of graphite, which has very high intensity, is cut off to show the other peaks. Theta, diffraction angle. e, The potential of the LBC-G cathode during discharge, open-circuit relaxation during a 40-h rest, and charging at 0.2 C. The complete recovery of the charge capacity in the next cycle shows that all of the active LiBr and LiCl material was well confined in the LBC-G cathode and there was no capacity loss during the long rest. f, The open-circuit voltage decays in the 40-h rest of the LBC-G cathode at the fully charged state of 4.5 V at 0.2 C. Self-discharge was evaluated by comparing with the Coulombic efficiency and the capacity loss after resting.
Extended Data Fig. 3 Absence of corrosion in the current collector and oxidation of graphite and water at the operation potential.
a, Linear sweep voltammetry of a pure graphite electrode (with only PTFE binder) on a Ti-mesh current collector in LiBr·3H2O, LiCl·3H2O and WiBS electrolyte with a Ag/AgCl electrode as reference at 1 mV s−1. The results show absence of side reactions, such as corrosion of the current collector and oxidation of graphite and water, before the onset of the large increase of the current density at about 4.0 V, 4.5 V and 5.0 V versus Li/Li+, in accordance with the oxidation of Br−, Cl− and water, respectively. b, c, C 1s X-ray photoelectron spectra (b) and overall spectra (c; binding energy of 0–293 eV) of the LBC-G cathode before and after 10 full cycles. LiBr and LiCl were removed to avoid interference. No carbon–oxygen or carbon–halogen bonds were observed. Only traces of Br were detected as intercalation residual.
a, Equivalent circuit used for fitting the Nyquist plots in Fig. 1f, consisting of ohmic resistance R1 and a constant-phase element (CPE1) parallel to a resistor (R2) connected in series with a finite-diffusion Warburg impedance (Ws1). We note that the open Warburg impedance was not included, whereas the data at low frequencies were truncated accordingly during fitting. b, Comparison of Nyquist plots of LBC-G cathodes containing graphite hosts with different average flake sizes (KS4, about 4 μm; ‘Big graphite’, about 800 μm), showing the independence of the diffusion kinetics from the halogen diffusion length inside the graphite interlayer. c, Practical volumetric energy density of the LBC-G cathode compared with those of other representative state-of-the-art cathodes when paired with Li metal anodes. For a fair comparison, the potential of a unit stack (the smallest cell unit) comprising a 100-μm-thick cathode, a 9-μm separator and a Li metal anode (average discharge voltages referred to Li/Li+) was calculated using capacity matching. The volume fraction of the active material in each electrode was considered to be 70 vol% for intercalation materials and 60 vol% for conversion-type cathodes. The material properties in the fully expanded (lithiated) state were used to calculate the volumetric capacity and inactive volume within each electrode. The areal capacities of the anodes and cathodes were matched at 1:1 and no extra capacity was considered for formation losses. d, e, Schematic of the full cell configurations with an LBC-G composite cathode in WiBS aqueous-gel electrolyte during charging (d) and discharging (e).
Galvanostatic charge and discharge profiles of different composite cathodes at a current density of 80 mA g−1 in WiBS gel electrolyte. a, LiBr–graphite (mass ratio of about 1:1) cathode in the potential range 3.20–4.62 V. Without the presence of Cl−, there were no further oxidation reactions of Br0 until the potential was raised to above 4.55 V versus Li/Li+, where Br0 was further irreversibly oxidized into BrO−. b, Composite of (LiBr)0.5(LiCl)0.5 and titanium nanopowder (mass ratio 1:20), showing a charge capacity of 85% of the theoretical value for halogen anion redox reactions and negligible discharge capacity. The higher overpotential might be due to the lack of carbon catalysis for the redox reactions. c, (LiBr)0.5(LiCl)0.5/graphitized carbon black (mass ratio 1:3). d, (LiBr)0.5(LiCl)0.5/active carbon (mass ratio 1:3). e, (LiBr)0.5(LiCl)0.5/KS4 (mass ratio 6:4). f, g, N2 absorption/desorption isotherm of a graphite (KS4) electrode (f) and an active-carbon electrode (g) with 5 wt% PTFE binder. The results indicate that, unlike active carbon, the graphite host cannot provide a large surface area and small size pores to store halogens by adsorption. h, Ex situ XRD intensity of LiBr/LiCl/active-carbon cathodes at fully charged and discharged states. After adsorbing halogen (Br2 and BrCl) during charging, a relatively strong peak appears in the (002) peak area, and (100) weakens. This might imply the reformation of randomly oriented small graphitic zones with the help of halogen integration, which indicates a minor contribution of intercalation-like behaviour to halogen storage into nano-graphitized grains.
a–f, XRD patterns of chemically intercalated Br2 (a–c) and BrCl (d–f) GICs used as reference samples. These GICs were prepared by exposing the graphite flakes in high-concentration Br2 vapour and BrCl gas for 2 h (more synthesis details in Methods). The spontaneous slow de-intercalations of the XRD peaks that appear at 48 h were observed using the θ–2θ scan mode with Cu Kα radiation (1.5418 Å) in reflection geometry. g, Raman spectra (50–500 cm−1) of chemically intercalated Br2 and BrCl GICs used as reference samples.
Extended Data Fig. 7 Representative structures of stage-I [Br0.5Cl0.5]C3.5 complex obtained from DFT simulations.
All structures have intercalation voltages within 0.02 V per ion of a structure with homogenous Br–Cl–Br–Cl bond lengths of 2.45 Å (top left). The bottom right structure is simulated on the basis of the reported Br2 structure28. Quantum chemistry calculations performed on a Cl–Br∙∙∙Cl–Br cluster surrounded by conductive polarized continuum also yielded a zig-zag configuration for the Cl–Br∙∙∙Cl–Br complex with a Cl–Br∙∙∙Cl–Br angle of around 110°, which has lower energy than the linear Cl–Br∙∙∙Cl–Br configuration by 0.1 eV according to MP2/aug-cc-pvTz and PBE/aug-cc-pvTz calculations. The most stable geometry obtained from these cluster calculations is similar to that found in the stage-I complexes shown above.
Extended Data Fig. 8 Stage-I [Br0.5Cl0.5]C3.5 complex structures obtained from ab initio MD simulations.
Results from 30 ps of NVT ab initio MD simulations using the CP2K package and starting from a structure with homogenous –Br–Cl– bond lengths, as this structure was the most computationally efficient. a, Radial distribution function g(r) of stage-I [Br0.5Cl0.5]C3.5 from 30 ps of MD simulations at 333 K (left) and final snapshot of the trajectory (right). b, DFT results for stage-I [Br0.5Cl0.5]C3.5 from 30 ps of simulations at 333 K, following initial annealing at 633 K to accelerate the appearance of disorder (left) and final snapshot of the trajectory (right), with a close Br–Br contact highlighted in red. No close Cl–Cl contacts form at this voltage, as evidenced by the absence of features near the gas-phase Cl–Cl bond length in the radial distribution function. NVT simulations used the Langevin thermostat with an associated time constant of 10 fs and average box dimensions obtained from the equilibration runs performed in the NPT ensemble for 100 ps. A 1 fs timestep was used throughout. No signs of gassing and subsequent graphite exfoliation were observed over 100 ps of additional simulations under constant-pressure conditions, even after brief annealing at 633 K and relaxation back to 333 K.
Extended Data Fig. 9 Representative in-plane configurations of the cathode structure from DFT calculations.
a–d, Stage-II C7[Br] (a), stage-I C7[BrCl] (b), stage-II C8[Br] (c) and stage-I C8[BrCl] (d) cathodes obtained from DFT simulations. Only a single set of bond lengths can be obtained with this C4[X] stoichiometry (X, halogen). e–h, Comparison of scattering paths calculated using FEFF6 with different DFT structures to determine the best modes for fitting the experimental XAFS data: stage-II C7[Br] (e) and C8[Br] (f), and stage-I C3.5[Br0.5Cl0.5] (g) and C4[Br0.5Cl0.5]) (h). The nearest (black), second-nearest (red) and third-nearest (blue) scattering paths around the Br centre are shown. The absence of the second-nearest Br–Br (R ≈ 2.6 Å) and Br–Cl (R ≈ 2.6 Å) scattering paths for the C4[X] stoichiometry suggests that the stoichiometry of C3.5[X] would be the dominant one in the real materials.