German Edition: DOI: 10.1002/ange.201805456Lithium-Metal Batteries Hot PaperInternational Edition: DOI: 10.1002/anie.201805456

Stabilizing Lithium Plating by a Biphasic Surface Layer Formed In SituQuan Pang, Xiao Liang, Ivan R. Kochetkov, Pascal Hartmann, and Linda F. Nazar*

Abstract: The dendritic growth of Li metal leads to electrodedegradation and safety concerns, impeding its application inbuilding high energy density batteries. Forming a protectivelayer on the Li surface that is electron-insulating, ion-conducting, and maintains an intimate interface is critical.We herein demonstrate that Li plating is stabilized by a biphasicsurface layer composed of a lithium-indium alloy and a lithiumhalide, formed in situ by the reaction of an electrolyte additivewith Li metal. This stabilization is attributed to the fast lithiummigration though the alloy bulk and lithium halide surface,which is enabled by the electric field across the layer that isestablished owing to the electron-insulating halide phase. Agreatly stabilized Li-electrolyte interface and dendrite-freeplating over 400 hours in Li jLi symmetric cells using an alkylcarbonate electrolyte is demonstrated. High energy efficiencyoperation of the Li4Ti5O12 (LTO) jLi cell over 1000 cycles isachieved.

The development of electric vehicles places great emphasison rechargeable batteries. Beyond lithium-ion technologies,potential candidates that can offer high energy density are Limetal batteries, that include Li–S and Li–O2 chemistries, inwhich a Li metal anode is the pivotal element.[1–5] However,the structural degradation and safety concerns of Li metalover continuous plating/stripping cycles hinders its applica-tion.[6–8] This originates from the interplay of the inherentdendritic growth of Li and its reactivity with the electrolytes.Based on ChazalvielQs model, anion depletion in low trans-ference number electrolytes intrinsically causes spatial chargeaccumulation. It incurs a large electric field that drivesdendritic growth at high currents.[9, 10] On the other hand,parasitic reactions of Li with the electrolyte lead to a non-uniform and ion-impeding solid electrolyte interphase (SEI)that causes localized heterogeneous current at high flux andaggravates dendrite growth (Figure 1a).[11,12]

Much effort has been devoted to tackle the dendritegrowth problem from different perspectives. AccommodatingLi in a 3D host framework lowers the effective current densityand alleviates volume expansion.[13–15] Artificial SEIs that aremechanically robust have been shown to effectively block the

vertical growth of Li dendrites.[16–18] Modifying the SEIcompositions via concentrated electrolytes,[19, 20] and electro-lyte additives[21]—mostly via increasing the fraction of LiF—show promise in stabilizing the Li metal anode performance.Relying on the reaction of Li with additives of polysulfides,LiNO3, organosulfides, fluoroethylene carbonate or a combi-nation of them, a chemically stable SEI forms and increasesthe plating efficiency.[22–27] However, formation of an ion-conducting but electronically insulating SEI that is stable inalkyl carbonate electrolytes—and which also maintainsintimate contact with Li metal—remains an ongoing chal-lenge.

We and others have recently reported that Li-In alloycomposite coated Li electrodes prepared by solution pre-treatment can sustain long-term dendrite-free plating.[28,29]

Herein, instead of relying on ex situ pre-treatment, wedemonstrate a simple approach to realize a biphasic protec-tive layer in situ (that is, in the assembled cell), by reaction ofLi with indium halides (InX3, X = F, Cl, Br, I) that are addedto the electrolyte. An immediate advantage is the locallyintimate contact of the biphasic layer with Li, irrespective ofthe rough Li foil surface, which also helps prevent layerdelamination during cell operation. Furthermore, the InX3

excess in the EC/DMC electrolyte allows self-healing of theprotection layer during cycling. While the alloy phase LixIny

exhibits a high bulk diffusion coefficient (DLi& 10@8–10@6 cm2 s@1,[30, 31] compared to 5.69 X 10@11 cm2 s@1 in bulklithium metal[32]) and the LiX halides enable fast surfacelithium migration,[21] the electron-insulating LiX phase isnecessary to establish an electric field to drive lithium ionsacross the layer (Figure 1b). We show that with a low

Figure 1. Illustration of the solid electrolyte interphase (SEI) formationwith general Li plating behavior (left) and the detailed ion/electrontransport scenario across the SEI that accounts for the different platingbehavior (right) in the blank (a) and InF3-added EC/DMC electroly-te (b).

[*] Dr. Q. Pang, X. Liang, I. R. Kochetkov, Prof. Dr. L. F. NazarDepartment of Chemistry and the Waterloo Institute for Nano-technologyUniversity of WaterlooWaterloo, Ontario N2L 3G1 (Canada)E-mail: lfnazar@uwaterloo.ca

Dr. P. HartmannBASF SELudwigshafen (Germany)

Supporting information and the ORCID identification number(s) forthe author(s) of this article can be found under:https://doi.org/10.1002/anie.201805456.

AngewandteChemieCommunications

9795Angew. Chem. Int. Ed. 2018, 57, 9795 –9798 T 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

concentration of 60 mm InF3 in a conventional alkyl carbon-ate electrolyte, Li plates in a non-dendritic and compactmorphology. A stabilized Li-electrolyte interface and highplating/stripping reversibility is obtained in Li jLi symmetriccells with InF3 added to the electrolyte, which is superior tocontrol cells using blank or LiF-added electrolytes.

The formation of the two phases—LixIny and LiF—at theLi-electrolyte interface occurs via the following reactions[Eq. (1),(2)]:

3 Li þ InX3 ! In þ 3 LiX ðX ¼ F, Cl, Br, IÞ ð1Þ

x Li þ y In ! LixIny ð2Þ

To examine the surface composition and morphologyupon cell operation, polished Li foil was treated in 60 mm InF3

carbonate electrolyte for 48 hours at 25 88C. With the InF3

treatment, the Li shows a dense particulate surface witha layer thickness of approximately 1 mm (Figure S1 in theSupporting Information). The Li-In alloy phase, whichappears bright in the back-scattered scanning electronmicroscopy (SEM) images, is distributed homogeneouslywithin the LiF phase, which makes the alloy-halide compositelayer electronically insulating as previously reported.[28] Theobserved finite thickness indicates that film growth becomeslimited by both mass transport and by its insulating nature. X-ray photoelectron spectroscopy (XPS) confirms the formationof the two targeted phases. The In 3d spectrum of InF3 treatedLi shows one major component corresponding to Ind@

characteristic of Li-rich alloys (3d5/2 : 442.5 eV, Figure S2a),along with a small fraction of In metal (3d5/2 : 443.1 eV).[25,33]

The F 1 s spectrum confirms a majority fraction of LiF,[11,22]

along with a marginal component of LixPFy as a result ofLiPF6 decomposition (Figure S2c).[34] The X-ray diffractionpattern of the InF3 treated Li shows the presence of crystallineLi13In3, LiF and In metal (Figure S3). Un-alloyed In metal wasabsent after Li plating (2 mAh cm@2), indicating that furtheralloying of In occurs over initial conditioning (Figure S2b andS3). The lower reactivity of InF3 than InCl3 yields a muchthinner layer than that previously demonstrated with InCl3.

[25]

As we will later show, the biphasic layer of this composition iseffective in suppressing dendrite growth over long-termcycling, as it can self-repair owing to the excess additivepresent in the electrolyte (see Note S1 in the SupportingInformation).

The Li plating behavior in a classic battery electrolyte (1mLiPF6 in ethylene carbonate/dimethyl carbonate, EC/DMC)with added InF3 is revealed by SEM imaging. Li electrodeswere retrieved after plating 2 mAhcm@2 Li at a current of0.5 mAcm@2 in Li jLi symmetric cells. The Li plated in theblank electrolyte shows a typical needle-shaped morphologyacross the thickness and exhibits high porosity (Figure 2a,d).The Li plated in the LiF-only electrolyte shows a similarmorphology, albeit with reduced porosity (Figure 2b,e). Incontrast, the plated Li in the InF3-added electrolyte showsa compact and relatively smooth surface, indicating film-likegrowth behavior (Figure 2 c,f). Confirmation of plating of Liunder the protective layer is afforded by SEM imaging, whicheasily discriminates between the underplated Li and the

protective layer (which is brighter due to its heavier Incomponent; Figure 2 f). This is further supported by theobservation of LixIny alloy and LiF on the surface after plating(XPS, Figure S2b,d). The Li surface morphology and SEIthickness after 50 cycles confirms the long-term effectivenessof the biphasic layer in isolating Li metal and suppressingparasitic reactions (Figure S4).

The evolution of electrochemical impedance spectra(EIS) upon open-circuit rest of the Li jLi symmetric cellsreveals the chemical stability of the Li-electrolyte interface.Upon a 30 minute rest, the InF3 cell shows a 10-fold lowerinterfacial charge transfer resistance (Rct) compared to theblank cell (25 vs. 300 Wcm2), as determined by the intercept ofthe semicircle with the abscissa(Figure 3a,b, Figure S5a,b).This demonstrates that the spontaneous formation of thebiphasic layer suppresses electrolyte decomposition whileserving as a Li-ion migration conduit. The Rct of the blank cellexhibited a continuous increase over 24 hours, whereas that ofthe cell with added InF3 remains almost constant, indicatinghigh chemical stability of the interphase.

It has been demonstrated that highly evolving potentialprofiles of a Li jLi symmetric cell signify microstructureformation/consumption upon plating/stripping.[35, 36] As shownin Figure 3c, the Li plating in the blank electrolyte experi-ences a large overpotential for Li nucleation. Upon Listripping, the voltage of the blank cell clearly evolves ina volcano curve, where the potential first increases and thendecreases, corresponding to the dissolution of plated den-drites and of the bulk Li forming surface pits, respectively.[32]

The voltage of the cell with the LiF-added electrolyte alsoevolves in a volcano-like curve, albeit with less potentialvariation (Figure 3c). In contrast, the cell in the InF3-added

Figure 2. a)–c) The surface and d)–f) cross-section SEM images of Lielectrodes after Li plating in the a),d) blank EC/DMC electrolyte,b),e) LiF-added EC/DMC electrolyte and c),f) InF3-added EC/DMCelectrolyte (current, 0.5 mAcm@2 ; capacity 2 mAhcm@2). Scale bars are2 mm in (a–c) and 5 mm in (d–f).

AngewandteChemieCommunications

9796 www.angewandte.org T 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2018, 57, 9795 –9798

electrolyte shows a rather flat potential profile, indicatingsuppressed surface area variation and dendrite formation.

The plating/stripping cycling stability on Cu foil wasevaluated by cyclic voltammetry (CV; Figure 3d, S5c). On the1st cycle, the cell using the blank electrolyte shows almostzero current. On the 5th cycle, the LiF and InF3 cells showmuch higher current than the blank cell (stripping: 12.5, 12.0and 5.0 mAcm@2, respectively), indicating that both the LiFand InF3 derived SEIs exhibit more favorable inter-facial kinetics. However, on the 10th cycle, in contrastto the stabilized current presented for the InF3 cell,the current of the LiF cell declines to about9 mAcm@2 (Figure S5c). This indicates inferior lith-ium conductivity and structural stability for the LiFderived SEI, which likely degrades over continuousplating/stripping.

Figure 4a shows the evolution of the voltageprofile with plating/stripping in Li jLi symmetriccells at a current of 1 mAcm@2. The cell usinga 60 mm InF3 carbonate electrolyte (denoted asInF3-60 mm) shows stable voltage profiles over400 hours of cycling, whereas the blank cell experi-ences fluctuating and increasing overpotential, indi-cating accumulation of an impeding SEI and ubiq-uitous local short-circuits. The cell with the LiF-60 mmelectrolyte also shows a voltage increase after cyclingfor 200 hours (Figure S6). Electrolytes with a lowerconcentration of InF3 additive were also examined.The InF3-15 mm cell shows a stable evolution of thevoltage profile, but the InF3-7.5 mm cell experiencesa potential increase after 150 cycles (Figure S7a).Furthermore, even at a higher current density of4 mAcm@2, the InF3-added cell also exhibits stablevoltage profiles over 100 cycles, in contrast to thefluctuating voltage of the blank cell (Figure S8). In

a glyme electrolyte (which presents better chemical stabilitywith Li) the InF3 added cell maintains exceptional plating/stripping over 1000 hours (Figure 4b).

We expanded the study to other indium halide additives(15 mm), and further elucidated the respective roles played bythe Li-In alloy and LiF. The cell using InCl3-added electrolyteexhibits as stable an evolution of the voltage profile as that ofthe InF3 added cell (Figure S7b), implying that both LiF andLiCl are effective in enabling lithium migration and main-taining structural stability. However, cells with InBr3 and InI3

added electrolytes underwent a dramatic voltage increaseafter 70 and 40 cycles, respectively (Figure S7c). We observedthat these electrolytes turned dark in color and gelled on theLi surface after cycling; we thus propose the occurrence ofunexpected electrolyte decomposition/polymerization initi-ated by the strong Lewis acids I@/Br@ and the heat generatedby the reaction shown in Equation (1).

To further demonstrate the practical significance of the insitu formed biphasic protective layer, Li metal cells wereassembled using zero-strain Li4Ti5O12 (LTO) as the workingelectrodes.[37] The LTO jLi cells were cycled at 5C (ca.1.9 mAcm@2) in EC/DMC-based electrolytes. The cell usinga blank alkyl carbonate electrolyte experienced sudden deathat 320 cycles, whereas the InF3-added cell exhibits stablecycling with a capacity retention of 67% over 1000 cycles(Figure 4c). Energy efficiency (EE) of the LTO jLi cell—defined as the discharge energy over charge energy—is animportant metric to monitor the cell polarization (note theLTO cells exhibit nearly 100 % coulombic efficiency overcycling (Figure S9)). The EE of the InF3-added cell reachesa maximum of 82.5 % after a few conditioning cycles, followedby slow fading to 73.8% over 1000 cycles, in contrast to the

Figure 3. a),b) Electrochemical impedance spectra for the Li jLi sym-metric cells using a) the blank electrolyte and b) the InF3-addedelectrolyte over 24 hours of rest at open-circuit; note that the scale ofthe abscissa in (a) and (b) are different by 20-fold. c) The first cyclevoltage profile of Li jLi symmetric cells using different electrolytes(current, 1 mAcm@2 ; capacity, 1 mAhcm@2). d) Cyclic voltammetry ofCu jLi cells using blank and InF3 added electrolytes for the 1st, 5th,and 10th cycles.

Figure 4. The evolution of voltage profiles of the symmetric cells over cyclingusing blank and InF3-added electrolytes based on a) 1m LiPF6 EC/DMC andb) 1m LiTFSI in DOL/DME (current, 1 mAcm@2 ; capacity, 1 mAh cm@2). c) Thedischarge capacity retention and energy efficiency (as calculated by dischargeenergy W 100/charge energy) of corresponding LTO jLi full cells over cycling at5C in InF3-added 1m LiPF6 EC/DMC electrolyte; d) the representative discharge/charge profiles at the 250th cycle.

AngewandteChemieCommunications

9797Angew. Chem. Int. Ed. 2018, 57, 9795 –9798 T 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org

rapid fading from 79.5% to 64.7 % over 320 cycles for theblank cell (Figure 4c). The voltage profiles for the 250th cyclein Figure 4d confirms the difference in polarization, indicat-ing favorable ion transfer in the alloy-LiF based SEI. We havealso coupled a LiFePO4 electrode with the protected Li metalanode using a EC/DMC based electrolyte in the window of2.2–4.2 V (Figure S10). While long-term cycling (as for LTO)has not yet been conducted at this preliminary stage, thepositive effect of the layer is still apparent. We observed bothhigher capacity retention and lower voltage polarization forthe cell using the InF3-added electrolyte versus the blank cell,confirming the validity of this approach to higher voltagebatteries.

In conclusion, we show that the in situ formation of anlithium-conducting, electron-insulating, and chemically stablebiphasic layer on the Li metal surface leads to dendrite-freeand stabilized Li plating in alkyl carbonate electrolytes. Thefunction of the biphasic layer is two-fold. The physicalisolation of Li leads to significantly reduced parasitic reac-tions with the electrolyte, thus creating a uniform SEI andalleviating current/nucleation heterogeneity. The Li-In alloyprovides efficient lithium-migration channels, allowing Liplating under the protecting layer, while the insulating LiFphase prevents electron transfer across the layer. Stable long-term plating/stripping of symmetric cells over 400 hours incarbonate electrolytes was demonstrated. This concept rep-resents a viable and scalable avenue for generating an SEIwith desired properties in situ using rationally designedelectrolyte additives.

Acknowledgements

This research was supported by the BASF InternationalScientific Network for Electrochemistry and Batteries. L.F.N.also thanks NSERC for generous support via their CanadaResearch Chair, and Discovery Grant programs.

Conflict of interest

The authors declare no conflict of interest.

Keywords: biphasic layer · dendrite prevention · indium ·lithium · lithium-metal batteries

How to cite: Angew. Chem. Int. Ed. 2018, 57, 9795–9798Angew. Chem. 2018, 130, 9943–9946

[1] W. Xu, J. Wang, F. Ding, X. Chen, E. Nasybulin, Y. Zhang, J.-G.Zhang, Energy Environ. Sci. 2014, 7, 513 – 537.

[2] X.-B. Cheng, R. Zhang, C.-Z. Zhao, Q. Zhang, Chem. Soc. Rev.2017, 117, 10403 – 10473.

[3] P. G. Bruce, S. A. Freunberger, L. J. Hardwick, J. M. Tarascon,Nat. Mater. 2012, 11, 19 – 29.

[4] D. Lin, Y. Liu, Y. Cui, Nat. Nanotechnol. 2017, 12, 194 – 206.[5] J. Zheng, M. H. Engelhard, D. Mei, S. Jiao, B. J. Polzin, J.-G.

Zhang, W. Xu, Nat. Energy 2017, 2, 17012.[6] D. Aurbach, E. Zinigrad, T. Cohen, H. Teller, Solid State Ionics

2002, 148, 405 – 416.

[7] K. J. Harry, D. T. Hallinan, D. Y. Parkinson, A. A. Macdowell,N. P. Balsara, Nat. Mater. 2014, 13, 69 – 73.

[8] X.-B. Cheng, C. Yan, J.-Q. Huang, P. Li, L. Zhua, L. Zhao, Y.Zhang, W. Zhu, S.-T. Yan, Q. Zhang, Energy Storage Mater. 2017,6, 18 – 25.

[9] J. N. Chazalviel, Phys. Rev. A 1990, 42, 7355 – 7367.[10] C. Brissot, M. Rosso, J. N. Chazalviel, S. Lascaud, J. Power

Sources 1999, 81 – 82, 925 – 929.[11] D. Aurbach, Y. Gofer, J. Electrochem. Soc. 1989, 136, 3198 –

3205.[12] Y. S. Cohen, Y. Cohen, D. Aurbach, J. Phys. Chem. B 2000, 104,

12282 – 12291.[13] C.-P. Yang, Y.-X. Yin, S.-F. Zhang, N.-W. Li, Y.-G. Guo, Nat.

Commun. 2015, 6, 8058.[14] D. Lin, Y. Liu, Z. Liang, H.-W. Lee, J. Sun, H. Wang, K. Yan, J.

Xie, Y. Cui, Nat. Nanotechnol. 2016, 11, 626 – 632.[15] Y. Zhang, W. Luo, C. Wang, Y. Li, C. Chen, J. Song, J. Dai, E. M.

Hitz, S. Xu, C. Yang, Y. Wang, L. Hu, Proc. Natl. Acad. Sci. USA2017, 114, 3584 – 3589.

[16] G. Zheng, S. W. Lee, Z. Liang, H.-W. Lee, K. Yan, H. Yao, H.Wang, W. Li, S. Chu, Y. Cui, Nat. Nanotechnol. 2014, 9, 618 – 623.

[17] W. Luo, C.-F. Lin, O. Zhao, M. Noked, Y. Zhang, G. W. Rubloff,L. Hu, Adv. Energy Mater. 2017, 7, 1601526.

[18] P. Bai, J. Li, F. R. Brushett, M. Z. Bazant, Energy Environ. Sci.2016, 9, 3221 – 3229.

[19] J. Qian, W. A. Henderson, W. Xu, P. Bhattacharya, M. Engel-hard, O. Borodin, J.-G. Zhang, Nat. Commun. 2015, 6, 6362.

[20] L. Suo, Y.-S. Hu, H. Li, M. Armand, L. Chen, Nat. Commun.2013, 4, 1481.

[21] Y. Lu, Z. Tu, L. A. Archer, Nat. Mater. 2014, 13, 961 – 969.[22] W. Li, H. Yao, K. Yan, G. Zheng, Z. Liang, Y.-M. Chiang, Y. Cui,

Nat. Commun. 2015, 6, 7436.[23] C. Yan, X.-B. Cheng, C.-Z. Zhao, J.-Q. Huang, S.-T. Yang, Q.

Zhang, J. Power Sources 2016, 327, 212 – 220.[24] Q. Shi, Y. Zhong, M. Wu, H. Wang, H. Wang, Proc. Natl. Acad.

Sci. USA 2018, 115, 5676 – 5680.[25] G. Li, Q. Huang, X. He, Y. Gao, D. Wang, S. H. Kim, D, Wang,

ACS Nano 2018, 12, 1500 – 1507.[26] X.-Q. Zhang, X. Chen, X.-B. Cheng, B.-Q. Li, X. Shen, C. Yan, J.-

Q. Huang, Q. Zhang, Angew. Chem. Int. Ed. 2018, 57, 5301 –5305; Angew. Chem. 2018, 130, 5399 – 5403.

[27] C. Yan, X.-B. Cheng, Y. Tian, X. Chen, X.-Q. Zhang, W.-J. Li, J.-Q. Huang, Q. Zhang, Adv. Mater. 2018, 30, 1707629.

[28] X. Liang, Q. Pang, I. Kochetkov, M. Safont-Sempere, H. Huang,X. Sun, L. F. Nazar, Nat. Energy 2017, 2, 17119 – 17124.

[29] Z. Tu, S. Choudhury, M. J. Zachman, S. Wei, K. Zhang, L. F.Kourkoutis, L. A. Archer, Nat. Energy 2017, 3, 310 – 316.

[30] M. Hiratani, K. Miyauchi, T. Kudo, Solid State Ionics 1988, 28 –30, 1406 – 1410.

[31] Z. Shi, M. Liu, J. L. Gole, Electrochem. Solid-State Lett. 2000, 3,312 – 315.

[32] E. Dologlou, Glass Phys. Chem. 2010, 36, 570 – 574.[33] S. A. Webb, L. Baggetto, C. A. Bridges, G. M. Veith, J. Power

Sources 2014, 248, 1105 – 1117.[34] Z. W. Seh, J. Sun, Y. Sun, Y. Cui, ACS Cent. Sci. 2015, 1, 449 –

455.[35] K. N. Wood, E. Kazyak, A. F. Chadwick, K.-H. Chen, J.-G.

Zhang, K. Thornton, N. P. Dasgupta, ACS Cent. Sci. 2016, 2,790 – 801.

[36] D. Lin, J. Zhao, J. Sun, H. Yao, Y. Liu, K. Yan, Y. Cui, Proc. Natl.Acad. Sci. USA 2017, 114, 4613 – 4618.

[37] T. Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc. 1995,142, 1431 – 1435.

Manuscript received: May 10, 2018Accepted manuscript online: June 26, 2018Version of record online: July 9, 2018

AngewandteChemieCommunications

9798 www.angewandte.org T 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2018, 57, 9795 –9798