Letter | Published:

Cryo-STEM mapping of solid–liquid interfaces and dendrites in lithium-metal batteries

Naturevolume 560pages345349 (2018) | Download Citation


Solid–liquid interfaces are important in a range of chemical, physical and biological processes1,2,3,4, but are often not fully understood owing to the lack of high-resolution characterization methods that are compatible with both solid and liquid components5. For example, the related processes of dendritic deposition of lithium metal and the formation of solid–electrolyte interphase layers6,7 are known to be key determinants of battery safety and performance in high-energy-density lithium-metal batteries. But exactly what is involved in these two processes, which occur at a solid–liquid interface, has long been debated8,9,10,11 because of the challenges of observing such interfaces directly. Here we adapt a technique that has enabled cryo-transmission electron microscopy (cryo-TEM) of hydrated specimens in biology—immobilization of liquids by rapid freezing, that is, vitrification12. By vitrifying the liquid electrolyte we preserve it and the structures at solid–liquid interfaces in lithium-metal batteries in their native state, and thus enable structural and chemical mapping of these interfaces by cryo-scanning transmission electron microscopy (cryo-STEM). We identify two dendrite types coexisting on the lithium anode, each with distinct structure and composition. One family of dendrites has an extended solid–electrolyte interphase layer, whereas the other unexpectedly consists of lithium hydride instead of lithium metal and may contribute disproportionately to loss of battery capacity. The insights into the formation of lithium dendrites that our work provides demonstrate the potential of cryogenic electron microscopy for probing nanoscale processes at intact solid–liquid interfaces in functional devices such as rechargeable batteries.

Access optionsAccess options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Wu, Y. & Yang, P. Direct observation of vapor−liquid−solid nanowire growth. J. Am. Chem. Soc. 123, 3165–3166 (2001).

  2. 2.

    Weiner, S. & Addadi, L. Crystallization pathways in biomineralization. Annu. Rev. Mater. Res. 41, 21–40 (2011).

  3. 3.

    Stamenkovic, V. R., Strmcnik, D., Lopes, P. P. & Markovic, N. M. Energy and fuels from electrochemical interfaces. Nat. Mater. 16, 57–69 (2017).

  4. 4.

    Tarascon, J.-M. & Armand, M. Issues and challenges facing rechargeable lithium batteries. Nature 414, 359–367 (2001).

  5. 5.

    Zaera, F. Probing liquid/solid interfaces at the molecular level. Chem. Rev. 112, 2920–2986 (2012).

  6. 6.

    Tikekar, M. D., Choudhury, S., Tu, Z. & Archer, L. A. Design principles for electrolytes and interfaces for stable lithium-metal batteries. Nat. Energy 1, 16114 (2016).

  7. 7.

    Cheng, X.-B. et al. A review of solid electrolyte interphases on lithium metal anode. Adv. Sci. 3, 1500213 (2016).

  8. 8.

    Yamaki, J. et al. A consideration of the morphology of electrochemically deposited lithium in an organic electrolyte. J. Power Sources 74, 219–227 (1998).

  9. 9.

    Aurbach, D. Review of selected electrode-solution interactions which determine the performance of Li and Li ion batteries. J. Power Sources 89, 206–218 (2000).

  10. 10.

    Bai, P., Li, J., Brushett, F. R. & Bazant, M. Z. Transition of lithium growth mechanisms in liquid electrolytes. Energy Environ. Sci. 9, 3221–3229 (2016).

  11. 11.

    Kushima, A. et al. Liquid cell transmission electron microscopy observation of lithium metal growth/dissolution: root growth, dead lithium and lithium flotsams. Nano Energy 32, 271–279 (2017).

  12. 12.

    Dubochet, J. et al. Cryo-electron microscopy of vitrified specimens. Q. Rev. Biophys. 21, 129–228 (1988).

  13. 13.

    Hu, Y.-Y. et al. Origin of additional capacities in metal oxide lithium-ion battery electrodes. Nat. Mater. 12, 1130–1136 (2013).

  14. 14.

    Heymann, J. A. W. et al. Site-specific 3D imaging of cells and tissues with a dual beam microscope. J. Struct. Biol. 155, 63–73 (2006).

  15. 15.

    Zachman, M. J., Asenath-Smith, E., Estroff, L. A. & Kourkoutis, L. F. Site-specific preparation of intact solid–liquid interfaces by label-free in situ localization and cryo-focused ion beam lift-out. Microsc. Microanal. 22, 1338–1349 (2016).

  16. 16.

    Cody, G. D. et al. Quantitative organic and light-element analysis of comet 81P/Wild 2 particles using C-, N-, and O-mu-XANES. Meteorit. Planet. Sci. 43, 353–365 (2008).

  17. 17.

    Yang, C. R., Wang, Y. Y. & Wan, C. C. Composition analysis of the passive film on the carbon electrode of a lithium-ion battery with an EC-based electrolyte. J. Power Sources 72, 66–70 (1998).

  18. 18.

    Zhuang, G. V., Xu, K., Yang, H., Jow, T. R. & Ross, P. N. Lithium ethylene dicarbonate identified as the primary product of chemical and electrochemical reduction of EC in 1.2 M LiPF 6/EC:EMC electrolyte. J. Phys. Chem. B 109, 17567–17573 (2005).

  19. 19.

    Metzger, M., Strehle, B., Solchenbach, S. & Gasteiger, H. A. Origin of H2 evolution in LIBs: H2O reduction vs. electrolyte oxidation. J. Electrochem. Soc. 163, A798–A809 (2016).

  20. 20.

    Szczȩśniak, M. M. & Ratajczak, H. Ab initio calculations on the lithium fluoride–ethylene complex. J. Chem. Phys. 67, 5400–5401 (1977).

  21. 21.

    Onuki, M. et al. Identification of the source of evolved gas in Li-ion batteries using #2#1 -labeled solvents. J. Electrochem. Soc. 155, A794–A797 (2008).

  22. 22.

    Aurbach, D. & Weissman, I. On the possibility of LiH formation on Li surfaces in wet electrolyte solutions. Electrochem. Commun. 1, 324–331 (1999).

  23. 23.

    Liu, D.-R. & Williams, D. B. The electron-energy-loss spectrum of lithium metal. Philos. Mag. B 53, L123–L128 (1986).

  24. 24.

    Oudriss, A. et al. The diffusion and trapping of hydrogen along the grain boundaries in polycrystalline nickel. Scr. Mater. 66, 37–40 (2012).

  25. 25.

    Islam, A. K. M. A. Lighter alkali hydride and deuteride. Phys. Status Solidi b 180, 9–57 (1993).

  26. 26.

    Settouti, N. & Aourag, H. Structural and mechanical properties of alkali hydrides investigated by the first-principles calculations and principal component analysis. Solid State Sci. 58, 30–36 (2016).

  27. 27.

    Choudhury, S. & Archer, L. A. Lithium fluoride additives for stable cycling of lithium batteries at high current densities. Adv. Electron. Mater. 2, 1500246 (2016).

  28. 28.

    Lu, Y., Tu, Z. & Archer, L. A. Stable lithium electrodeposition in salt-reinforced electrolytes. Nat. Mater. 13, 961–969 (2014).

  29. 29.

    Suo, L. et al. Fluorine-donating electrolytes enable highly reversible 5-V-class Li metal batteries. Proc. Natl Acad. Sci. USA 115, 1156–1161 (2018).

  30. 30.

    Zhao, Q. et al. Building organic/inorganic hybrid interphases for fast interfacial transport in rechargeable metal batteries. Angew. Chem. Int. Ed. 57, 992–996 (2018).

  31. 31.

    Tate, M. W. et al. High dynamic range pixel array detector for scanning transmission electron microscopy. Microsc. Microanal. 22, 237–249 (2016).

  32. 32.

    Egerton, R. F., Li, P. & Malac, M. Radiation damage in the TEM and SEM. Micron 35, 399–409 (2004).

  33. 33.

    Lin, F., Markus, I. M., Doeff, M. M. & Xin, H. L. Chemical and structural stability of lithium-ion battery electrode materials under electron beam. Sci. Rep. 4, 5694 (2014).

  34. 34.

    Miki, T., Ikeya, M., Kondo, Y. & Kanzaki, H. Reflectance spectrum of lithium hydride at the Li K-absorption edge. Solid State Commun. 39, 647–649 (1981).

  35. 35.

    Liu, D.-R. Electron energy loss spectroscopy of LiH with a scanning transmission electron microscope. Solid State Commun. 63, 489–493 (1987).

  36. 36.

    Parades Mellone, O. A., Ceppli, S. A., Arneodo Larochette, P. P. & Stutz, G. E. Excitación de electrones K del Li a baja transferencia de momento por dispersión inelástica de rayos X en LiH. Anal. Asoc. Fís. Argentina 26, 93–97 (2015).

  37. 37.

    Kesselman, E. et al. Cryogenic transmission electron microscopy imaging of vesicles formed by a polystyrene−polyisoprene diblock copolymer. Macromolecules 38, 6779–6781 (2005).

  38. 38.

    Cueva, P., Hovden, R., Mundy, J. A., Xin, H. L. & Muller, D. A. Data processing for atomic resolution electron energy loss spectroscopy. Microsc. Microanal. 18, 667–675 (2012).

Download references


We thank D. Muller, H. Abruña and J. Noble for discussions. M.J.Z. and L.F.K. acknowledge support by the NSF (DMR-1654596) and the Packard Foundation. Z.T., S.C. and L.A.A. acknowledge support from the Department of Energy, Advanced Research Projects Agency - Energy (ARPA-E) through award number DE-AR0000750. This work made use of the Cornell Center for Materials Research (CCMR) Shared Facilities with funding from the NSF MRSEC programme (DMR-1719875). Additional support for the FIB/SEM cryo-stage and transfer system was provided by the Kavli Institute at Cornell and the Energy Materials Center at Cornell, DOE EFRC BES (DE-SC0001086). The FEI Titan Themis 300 was acquired through NSF MRI-1429155, with additional support from Cornell University, the Weill Institute and the Kavli Institute at Cornell. This work made use of electrochemical characterization facilities in the KAUST-CU Center for Energy and Sustainability, supported by the King Abdullah University of Science and Technology (KAUST) through award number KUS-C1-018-02.

Author information


  1. School of Applied and Engineering Physics, Cornell University, Ithaca, NY, USA

    • Michael J. Zachman
    •  & Lena F. Kourkoutis
  2. Department of Materials Science and Engineering, Cornell University, Ithaca, NY, USA

    • Zhengyuan Tu
    •  & Lynden A. Archer
  3. Robert Frederick Smith School of Chemical and Biomolecular Engineering, Cornell University, Ithaca, NY, USA

    • Snehashis Choudhury
    •  & Lynden A. Archer
  4. Kavli Institute at Cornell for Nanoscale Science, Cornell University, Ithaca, NY, USA

    • Lena F. Kourkoutis


  1. Search for Michael J. Zachman in:

  2. Search for Zhengyuan Tu in:

  3. Search for Snehashis Choudhury in:

  4. Search for Lynden A. Archer in:

  5. Search for Lena F. Kourkoutis in:


M.J.Z. and L.F.K. conceived the project and wrote the manuscript. M.J.Z. prepared samples, performed the experiments and analysed the data. Z.T. constructed, cycled and disassembled the traditional electrolyte coin cell batteries. S.C. constructed, cycled, dissembled and measured the electrochemical properties of the full-fluoride and traditional electrolyte coin cell batteries displayed in Extended Data Figure 4b–f. Z.T. and L.A.A. provided assistance with interpretation of the data and made revisions to the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Lena F. Kourkoutis.

Extended data figures and tables

  1. Extended Data Fig. 1 Schematic and SEM images of the cryo-FIB lift-out sample preparation process, and examples of additional final lamellae.

    a, A buried structure or interface is identified for preparation, here a dendrite embedded in frozen electrolyte above the anode (indicated by the red arrow). In our coin-cell batteries, raised regions of electrolyte were used to localize buried dendrites. b, e, f, Trenches are then site-specifically milled around the site of interest, forming a vertical cross-sectional lamella containing the structure or interface. The sample is aligned in the microscope so that the electrode surface normal is parallel to the electron beam direction in e, and tilted by 52° to image the lithium anode–electrolyte interface and the electrolyte-embedded dendrite in f. c, g, A cooled nanomanipulator needle is then attached to the cryo-immobilized lamella by water vapour from a gas-injection system deposited as amorphous ice. The lamella is then cut free from the sample and lifted out. d, h, Finally, the lamella is attached to a TEM grid post with additional ice deposition, cut free from the nanomanipulator and thinned to electron transparency with the ion beam. i, j, Lamellae containing type II dendrites above lithium electrodes. The lamella in i contains a fluorine-rich structure as well. Different electrolyte thicknesses and milling parameters were used to prepare these lamellae, resulting in different final dimensions. k, A lamella produced from an uncycled electrode, used to obtain reference spectra. The increased signal of the uncycled electrode is due to different image-acquisition parameters, not a material difference.

  2. Extended Data Fig. 2 Elemental maps of the regions near both types of dendrite surface, carbon-bonding environment maps resulting from fitting of MCR spectra back to original data, and corresponding summary schematics of both dendrite types and their SEI layers.

    Carbon, oxygen and fluorine are shown in a composite map in Fig. 2. a, d, Individual elemental maps showing the full count range, excluding 0.2% of high- and low-intensity outliers, make it clear that there is a substantial concentration of oxygen in the type I dendrite and very little in the type II dendrite, and that there is increased oxygen in the type I SEI compared to the electrolyte. In addition, essentially no fluorine is present in the type I SEI, and the large fluorine-rich structure contains a higher fluorine concentration than the electrolyte. Nitrogen maps are included as well, and largely show noise with little spatial dependence. Corresponding count scale bars are shown next to each map. b, Individual maps determined by MCR corresponding to the spectra primarily located at (and labelled in Fig. 3 as) the SEI, electrolyte and fluorine-rich structure (top to bottom), displaying the original counts. c, e, Individual plasmon maps determined by MCR for LiH, lithium and the electrolyte (top to bottom), displaying the original counts. f, Top, type I dendrites consist of partially oxidized lithium metal with small LiH regions at the surface, and have an extended SEI layer consistent with lithium ethylene dicarbonate (LEDC) that contains bubbles, probably from ethylene, a by-product of the SEI formation. Large fluorine-rich structures are often found near the dendrites. Bottom, type II dendrites consist of uniform LiH and have a compact Li2O/LiOH · H2O SEI layer. Although not depicted, fluorine-rich structures were also observed near type II dendrites. Scale bars, 300 nm.

  3. Extended Data Fig. 3 Comparison of the type II SEI oxygen K-edge with reference spectra and an example of a bandpass-filtered spectrum.

    a, The O K-edge of the type II dendrite appears to be consistent with a combination of Li2O and LiOH · H2O. Spectra are offset vertically for clarity. b, A 0.6-eV bandpass (BP) filter was applied to the O K-edge spectra acquired on the F20 to remove high-frequency noise. This preserved the main features of the edge while eliminating those below the energy resolution of the instrument.

  4. Extended Data Fig. 4 Cryo-FIB, cryo-STEM EELS and electrochemical results comparing lithium deposition in cells using traditional and full-fluoride electrolytes.

    a, b, Cryo-FIB reveals that the dendrite density is much lower for the full-fluoride fluoroethylene carbonate (FEC) electrolyte (b) than with the traditional EC:DMC electrolyte (a). In the former case, nearly no LiH dendrites are present and the lithium deposition is modified, forming broad localized depositions. c, d, Cross-sections of these deposits reveal that they are composed of many smaller ‘blocks’ in contact, separated by SEI layers. e, f, A lamella of this type of deposition was prepared by cryo-FIB lift-out (e) and cryo-STEM EELS of the Li K-edge of the material revealed that it is composed of partially oxidized lithium metal (f), as was the type I dendrite in the traditional electrolyte. g, The Coulombic efficiency measured in a lithium versus stainless steel set-up using a constant current density of 1 mA cm−2 and capacity of 1 mAh cm−2 was greatly improved for the full-fluoride electrolyte compared to the traditional electrolyte. h, Cycling of a full cell comprising a lean lithium anode (50 μm) and a nickel manganese cobalt oxide (NCM) cathode (2 mAh cm−2) with the full-fluoride electrolyte resulted in a substantial decrease in capacity fade and improved Coulombic efficiency over the traditional electrolyte. The discharge capacity is plotted on the left axis, whereas the Coulombic efficiency is on the right axis. The operating voltage range was 4.3 V to 3 V. In all figures, the red lines and symbols represent results for the EC:DMC, 1 M LiPF6 electrolyte, whereas the black lines and symbols are for FEC, 2 M LiPF6.

  5. Extended Data Fig. 5 Full spectra recorded from the dendrites (intensities on a logarithmic scale).

    The spectra show clear differences in the plasmons and Li K-edges, as well as a large difference in oxygen content between the type I and type II dendrites. The small amount of oxygen on the type II dendrite is probably due to water molecules adsorbed on the surface of the sample in the microscope vacuum, which would typically react with materials such as lithium or sodium at room temperature. No nitrogen was present in either dendrite, confirming that no reaction with nitrogen in the air or liquid N2 had occurred.

  6. Extended Data Fig. 6 Example damage series profiles and initial/final spectra taken for lithium materials relevant to this study over a range of doses at which damage occurs, dark-field cryo-STEM images of various types of damage induced in a frozen organic electrolyte at different doses with corresponding spectra, and before and after images of the regions in which the EELS maps in the main text were taken.

    All spectra were recorded at cryogenic temperatures. a, We found all oxide materials convert to Li2O under large doses. Li2O and LiH are primarily affected by mass loss, with no substantial changes in fine structure. The maps presented in the main text were acquired at doses lower than the dose indicated by the red arrows shown at the bottom of the plots, of the order of 102 e Å−2. b, c, While some structural modification of the electrolyte material was present at low doses, probably due to liberation of hydrogen, a dose greater than 103 e Å−2 was required for substantial mass loss and modification of spectral fine structure. At 3 × 103 e Å−2, approximately 50% of the material remained after the map, as determined by the ADF signal. At 104 e Å−2, the material was completely removed in some areas, but the carbonate portion of the molecule remained. Doses applied during acquisition of the maps in the main text were less than the lowest dose shown here. Spectra are offset vertically for clarity. d, e, In the maps displayed in the main text, small structural changes were observed in the organic materials, which is expected given our damage analysis. This is probably due to liberation of hydrogen from the molecules, which occurs at low dose. The fine structure is not greatly affected until approximately an order of magnitude higher dose than was applied during these maps, which was of the order of 102 e Å−2. Scale bars, 200 nm, 30 nm and 60 nm (b, left to right), 300 nm (d, e).

  7. Extended Data Fig. 7 Charging profile from a symmetric lithium coin cell.

    A constant current of 1 mA cm−2 was applied to the cells for 24 h (bottom). The resulting voltage profile from one of the coin cells used is shown in the top panel.

  8. Extended Data Fig. 8 Amorphous diffraction pattern of the electrolyte recorded in a cryo-lamella produced by cryo-FIB lift-out.

    Cryo-TEM diffraction of the electrolyte on samples produced by cryo-FIB lift-out shows that it is frozen amorphously and does not recrystallize at any point during the preparation, storage, transfer or characterization.

  9. Extended Data Table 1 Comparison of the properties of type I and II dendrites
  10. Extended Data Table 2 Threshold electron doses and primary damage mechanisms observed for relevant materials

About this article

Publication history







By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.