Lithium Plating and Stripping on Carbon Nanotube Sponge - Nano

Dec 5, 2018 - (21) elongated the cycling life of the lithium metal anodes. Wang et al.(22) stabilized the lithium metal anode by regulating lithium ...
2 downloads 0 Views 5MB Size
Letter Cite This: Nano Lett. 2019, 19, 494−499

pubs.acs.org/NanoLett

Lithium Plating and Stripping on Carbon Nanotube Sponge Gaojing Yang,†,‡ Yejing Li,† Yuxin Tong,†,‡ Jiliang Qiu,†,§ Shuai Liu,†,§ Simeng Zhang,†,§ Zhaoruxin Guan,∥ Bin Xu,∥ Zhaoxiang Wang,*,†,‡,§ and Liquan Chen†,‡,§ †

Nano Lett. 2019.19:494-499. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/09/19. For personal use only.

Key Laboratory for Renewable Energy, Chinese Academy of Sciences, Beijing Key Laboratory for New Energy Materials and Devices, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China ‡ School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100190, China § College of Materials Science and Opto-Electronic Technology, University of Chinese Academy of Sciences, Beijing 100190, China ∥ State Key Laboratory of Organic−Inorganic Composites, Beijing Key Laboratory of Electrochemical Process and Technology for Materials, Beijing University of Chemical Technology, Beijing 100029, China S Supporting Information *

ABSTRACT: Lithium metal is an ideal anode material due to its high specific capacity and low redox potential. However, issues such as dendritic growth and low Coulombic efficiency prevent its application in secondary lithium batteries. The use of three-dimensional (3D) porous current collector is an effective strategy to solve these problems. Herein, commercial carbon nanotube (CNT) sponge is used as a 3D current collector for dendrite-free lithium metal deposition to improve the Coulombic efficiency and the cycle stability of the lithium metal batteries. The high specific surface area of the CNT increases the density of the lithium nucleation sites and ensures the uniform lithium deposition while the “pre-lithiation” behavior of the porous CNT enhances its affinity with the deposited lithium. Meanwhile, the lithium plating/stripping on the sponge maintains high Coulombic efficiency and high cycling stability due to the robust structure of graphitic-amorphous carbon composite in the ether-based electrolyte. Our findings exhibit the feasibility of using CNT sponge as a 3D porous current collector for lithium deposition. They shed light on designing and developing advanced current collectors for the lithium metal electrode and will promote the commercialization of the secondary lithium batteries. KEYWORDS: Lithium metal anode, carbon nanotube, deposition, lithium nucleation, porous current collector

W

the feasibility to tune the electrochemical plating and stripping behaviors by modifying the geometrical properties of the electrode surface. In recent years, three-dimensional (3D) porous current collectors attract wide attention because they can provide a basically uniform electric field and homogeneous charge distribution along the 3D skeleton. In addition, the 3D porous network provides a large amount of space to accommodate the large volumetric variation of the electrode upon lithium plating and stripping. By electrochemical dealloying, Zhao et al.20 prepared 3D copper foil with uniform pores as a dendrite-free current collector. With 3D porous copper foil as the current collector, Yang et al.21 elongated the cycling life of the lithium metal anodes. Wang et al.22 stabilized the lithium metal anode by regulating lithium plating/stripping in vertically aligned microchannels. However, the complication of the fabrication and high cost of such porous current collectors hinder their commercialization.23

ith the increasing demand for high energy-density storage devices, secondary lithium batteries receive extensive attention.1,2 Metallic lithium is an ideal anode material for its low redox potential (−3.04 V vs SHE) and high specific capacity (3860 mAh g−1).3,4 However, the secondary lithium batteries cannot be commercialized before their safety and cyclability issues can be well addressed due to the aggressive growth of the lithium dendrites during cycling.4,5 The inhomogeneous distribution of the local current density and the electrolyte concentration is believed to be responsible for the formation and growth of the lithium dendrites.6 Many efforts have been made to combat these issues, such as application of the electrolyte additives,7−9 optimization of the electrolyte compositions,10−12 fabrication of artificial interfaces on the lithium metal electrodes,13−15 and development of separators with functional polar groups.16 Surface modification to the lithium metal foil was reported to be an effective strategy to improve the performance by way of microneedle pretreatment for lithium metal foil.17 Park et al.18 pressed some patterns on the lithium foil and effectively hindered the growth of the lithium dendrites. We successfully controlled the lithium deposition by fabricating electrodes with appropriate micro/ nanoarchitectures on various substrates.19 These demonstrate © 2018 American Chemical Society

Received: October 31, 2018 Revised: November 26, 2018 Published: December 5, 2018 494

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499

Letter

Nano Letters

Figure 1. Physical characterization of the CNT sponge. (a and b) SEM images (inset, photograph), (c) nitrogen adsorption−desorption isotherm (inset, the corresponding pore size distribution), (d and e) HRTEM images, and (f) comparison of the Raman spectrum of the CNT sponge before and after 20 cycles of lithium plating/stripping.

production, agreeing with the energy dispersive spectroscopic (EDS) mapping (Figure S1). The strong G band at 1571 cm−1 in the Raman spectrum (Figure 1f) demonstrate the graphitic structure of the CNTs, but the presence of the strong D band at 1339 cm−1, the broadened G band and its weak shoulder at around 1610 cm−1 demonstrate that the CNT sponge contains much amorphous carbon as well, agreeing with the above nitrogen adsorption−desorption isotherm analysis. Such graphitic-amorphous carbon composite structure is critical in ensuring the structural and cycling stability of the current collector because cointercalation of the solvated lithium ions occurs in graphitic carbons in ether-based electrolyte and destroys the graphitic structure (Figure 1f vs Figure S2). The metallic lithium plating/stripping processes on the CNT sponge were characterized by cyclic voltammetry (CV) between −0.1 and 2.0 V (Figure S3). The cathodic peak at −0.09 V and the anodic peak at around 0.1 V correspond to the lithium plating and lithium stripping, respectively. The cathodic peak around 1.5 V is attributed to the electrolyte decomposition and the formation of the solid electrolyte interphase (SEI) layer on the electrode surface; it does not appear any longer in the subsequent cycling. Figure 2 shows the electrochemical performance of the CNT sponge. The areal discharge (or lithium plating) capacity was set to 2.0 mAh cm−2 for all the tested cells unless otherwise specified. The Coulombic efficiency is defined as the ratio of the lithium stripping capacity to the lithium plating capacity in each cycle. At a current density of 1.0 mA cm−2, the Coulombic efficiency is 72.1% in the first cycle but sharply increases to 97.5% and 98.8% in the second and 10th cycles, respectively (Figure 2a). It maintains above 98.5% in the subsequent 90 cycles (cycling still in process). The Li||CNT sponge cell runs very stably for a cycling duration of as long as 400 h (Figure S4). The stable structure of the graphiticamorphous carbon composite of the CNT sponge is clearly a precondition for the stable cycling behavior of lithium on it. Figure 1f compares the Raman spectrum of the CNT sponge before and after 20 cycles of 2.0 mAh cm−2 lithium plating/ stripping. No detectable spectral changes can be observed on it, verifying the structural stability of the CNT sponge upon

Carbon is an important type of lithium storage materials in lithium ion batteries. It was also used as current collectors for lithium metal batteries because carbon is among the lightest materials available for scaffold construction.24 Emerging carbon materials have high surface areas with excellent mechanical strength.24 Among them, the carbon nanotubes (CNTs), highly conductive and commercially available, have moderate specific surface areas and lithium storage capacities.25 Sun et al.26 used mechanically robust CNT paper for lithium metal anodes and demonstrated that its robust and expandable nature is distinguished in comparison with that of other 3D scaffolds and is key to the improved electrochemical performance of the Li/CNT anodes. However, as symmetric (Li/CNT) cells were used in their study (the lithium preloading on the Li/CNT working electrode was 11 mAh cm−2 while the highest lithium stripping capacity was 10 mAh cm−2), some important issues were circumvented (such as lithium nucleation) or even covered up (such as cointercalation-induced exfoliation of graphitic carbon in ether-based electrolyte). With its commercial availability and high specific surface area, especially its graphitic-amorphous carbon composite feature as shown in Figure 1, the CNT sponge is herein used as a 3D porous current collector for dendrite-free lithium metal deposition. The improved cycling stability of the lithium plating and stripping on it demonstrates the feasibility of using CNT sponge as a current collector for the secondary lithium batteries. The scanning electron microscopic (SEM) imaging shows that the sponge is composed of intersected carbon nanotubes with diameters ranging from 30 to 50 nm (Figure 1a,b). The nitrogen adsorption−desorption isotherm (Figure 1c) indicates that the Brunauer−Emmett−Teller (BET) surface area of the sponge is 113.8 m2 g−1. The pore size ranges from 1 to 100 nm, but macropores are dominant. This means that each carbon nanotube is porous in addition to the 3D porous network of the sponge as a whole. The presence of the equal thickness interference fringes in the high-resolution transmission electron microscopic (HRTEM) image (Figure 1d) reveals the graphene-layer structure of the wall of the CNT. The HRTEM imaging (Figure 1e) further shows the presence of iron as the catalyst inside the CNT for the sponge 495

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499

Letter

Nano Letters

(Figure 2d and Figure S6; more description and discussion can be found in the Supporting Information). The lithium nucleation potential on the CNT sponge is much higher than on the planar copper foil and considerable nucleation capacity is required before lithium plating on the sponge. These differences are believed to be originated from the distinctions of these two current collectors in surface properties. The CNT sponge is porous and has a reversible lithium storage capacity of 150 mAh g−1 above 0.0 V (Figure 2b). The microporous feature of our CNT sponge enables it to adsorb a large amount of lithium, and condensed lithium is formed on its surface.27 Therefore, the “pre-lithiated” CNT sponge can have a much higher chemical affinity with the metallic lithium, leading to reduced nucleation overpotential. On the other hand, prelithiation of graphitic carbon in etherbased electrolyte is accompanied by unavoidable cointercalation of the solvated lithium ions and exfoliation or pulverization of the layer-structured graphite (Figure S2). Therefore, a graphitic-amorphous carbon composite structure is critical before the chemical affinity of the prelithiation can make any sense. In contrast, no lithiation occurs on the copper foil before the first lithium atom deposits on it. In order to examine the electrochemical performance of the CNT sponge at lower and higher current densities, the Coulombic efficiency and the potential profile at current densities of 0.5 and 5.0 mA cm−2 were evaluated (Figure S7). The Coulombic efficiencies of the lithium plating/stripping on the CNT sponge are 79.9% and 68.3% in the first cycle at current densities of 0.5 and 5.0 mA cm−2, sharply increase to 97.9% and 95.1% in the second cycle, and 98.4% and 98.2% in the 10th cycle, respectively (Figure S7a,b), similar to the case at 1.0 mA cm−2 as discussed above. The interesting relationship between the initial Coulombic efficiency and the current density (lower current density for higher initial Coulombic efficiency; Figure S7c) is attributed to the inevitable electrolyte decomposition in the first cycle (independent of the current density) and the currentdensity-dependent morphology of the deposited lithium. The nucleation potential becomes high with the increasing cycle number (Figure S7d,e). All these electrochemical features agree well with those at a current density of 1.0 mA cm−2. Comparison of the lowest nucleation potentials at the above three current densities shows that the overpotential becomes small as the current density increases. Actually the nucleation overpotential at 5.0 mA cm−2 is as small as negligible (inset of Figure S7e). In combination with the electrochemical performance at a current density of 0.5 mA cm−2, the number of the nuclei per unit area is believed to be increasing when the current density increases. As a result, the deposited lithium spreads out continuously rather than grows as isolated particles on the substrate.28 The increasing potential gap between lithium plating and stripping is attributed to the polarization of the cell at high current densities (∼135 mV at 5.0 mA cm−2 vs ∼28 mV at 0.5 mA cm−2). Little work has been conducted to investigate the lithium deposition process on the porous current collectors.29 Although excellent cycling stability was obtained in CNT paper in ref 19, important issues such as lithium nucleation on the CNT were circumvented as a symmetrically configured (Li/CNT)||Li cell was adopted in that work. In the present work, we exerted a constant current density of 0.5 mA cm−2 on the asymmetric Li||CNT cell and recorded the SEM images of the CNT sponge at different plating/stripping contents (Figure

Figure 2. (a) Coulombic efficiency and (c) the selected lithium plating/stripping potential profiles at a current density of 1.0 mA cm−2 (the inset in part a is for the first discharge potential profile; the inset in part c is for the enlarged potential profile), (b) the potential profile between 0.0 and 2.0 V at 1.0 mA cm−2 in the first 10 cycles, and (d) comparison of the evolution of the nucleation overpotential on the CNT sponge and on Cu foil with cycling.

lithium intercalation, adsorption, and deposition in the etherbased electrolyte. The low initial Coulombic efficiency of the cell is attributed to the irreversible reduction of the electrolyte on the CNT sponge. A discharge plateau appears at 1.5 V in the potential profile of the CNT sponge (insets of Figure 2a). This plateau does not appear any longer in the discharge profile of the second or subsequent cycles (the discharge capacity for the sharp slope before the inflection point for the lithium nucleation at ∼−0.01 V is less than 0.05 mAh cm−2 in the potential profile of the second discharge). Figure 2b further shows that the discharge capacity of the CNT sponge at 1.0 mA cm−2 is around 0.65 mAh cm−2 but only 0.20 mAh cm−2 (∼150 mAh g−1) of which is reversible between 0.0 and 2.0 V. Part of the decomposition products forms the solid electrolyte interphase (SEI) layer on the surface. The SEI layer protects the electrode from direct contact with the electrolyte and is beneficial for subsequent lithium storage and plating. Characterization by X-ray photoelectron spectroscopy (XPS) (Figure S5) indicates that, after 5 cycles of lithium plating/ stripping, the SEI layer on the CNT sponge is composed of ROLi, ROCOOLi, LiNxOy, and LiF, well consistent with the previous report.16 The charge/discharge profile (Figure 2c) exhibits two interesting lithium plating/stripping characteristics. With increasing cycling number, the potential for the lithium plating rises but that for the lithium stripping falls in the first 50 h of cycling (∼8 cycles; Figure S4a,b). They then keep stable until the end of the test (a duration of ∼400 h). In this process, the nucleation potential rises (from −33 mV in the second discharge to −21 mV in the 50th discharge), the corresponding areal energy for the nucleation (the integral area for the nucleation envelope) decreases, but the onset potential of the nucleation does not change (inset of Figure 2c). The above lithium plating/stripping features on the CNT sponge are clearly different from that on the planar copper foil 496

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499

Letter

Nano Letters

Figure 3. Morphology of the CNT sponge at different stages of lithium plating and stripping: discharged to (a) 0.0 V, (b) plating 5 min, (c) enlarged view of part b, (d) discharged for 2.0 mAh cm−2 lithium, and (e) charged to 0.1 V after discharged for 2.0 mAh cm−2. (f) CNT sponge charged to 0.1 V after 20 cycles. (g) Schematic diagrams of electrochemical plating/stripping process of lithium on the CNT sponges.

3). When the cell is discharged to 0.0 V, some of the lithium ions are intercalated in the interlayer of the CNT and some other lithium atoms are adsorbed in the nanopores on the surface of the CNT (Figure 2b and Figure 3a). This is the stage of lithiation, resulting in CNT surfaces affinitive with the later deposited lithium. As the cell is further discharged below 0.0 V, the deposited lithium atoms nucleate on the CNT surface. Due to the affinity of the prelithiated CNT surface and its high surface area, the lithium nucleus is small but the areal density of the nucleation sites is high, resulting in low nucleation potential. This corresponds to the nucleation stage. Further discharging of the cell leads to interconnection of the lithium nuclei and lithium plating on the CNT. Figure 3b,c demonstrate that the deposited lithium is plated uniformly on the skeleton of the CNT sponge. As more lithium is deposited, the voids (macropores) in the sponge network are partially filled. No lithium dendrites can be observed in this process (Figure 3d). The high stability of the deposited metallic lithium in the ether-based electrolyte and the inhibition of the lithium dendrites on the CNT sponge ensure the high Coulombic efficiency and high cycling stability of the lithium plating/stripping. As the cell is charged to 0.1 V (Figure 3e), the lithium metal was stripped from the CNT sponge and the 3D skeleton of the CNT sponge becomes clear again. No residual metallic lithium can be detected on the CNT sponge charged to 0.1 V (Figure S5d). Figure 3g schematically summarizes the above plating/stripping processes on the CNT sponge. After 20 cycles of plating/stripping, both the 3D skeleton of the CNT sponge (Figure 3f) and its graphiticamorphous carbon composite structure (Figure 1f) are well preserved. In these aspects, the 3D structured CNT sponge is much superior over the planar (copper foil as a) current collector (Figures S6, S8, and S9). The purpose of developing secondary lithium batteries is to increase the energy density of the battery. Therefore, the areal capacity of the lithium anode should not be lower than that of the graphite (∼4.0 mAh cm−2) or even the silicon/carbon composite anode (currently ∼7.0 mAh cm−2). Keep this in mind, metallic lithium as the anode is expected to provide a high areal capacity of lithium deposition. Figure 4 shows that the SEM images of the CNT sponge with areal capacities

Figure 4. Morphology of the CNT sponge covered with (a) 1.0 mAh cm−2, (b) 2.0 mAh cm−2, (c) 5.0 mAh cm−2, and (d) 10.0 mAh cm−2 of lithium metal at a current density of 0.5 mA cm−2.

ranging from 1 mAh cm−2 to 10 mAh cm−2 at a current density of 0.5 mA cm−2. It is seen that the surface of the lithium-plated CNT sponge is dendrite-free in all these samples. This evidently indicates that the CNT sponge can effectively suppress the growth of lithium dendrites. These also demonstrate that the 3D network of the CNT sponge is favorable for the deposition of lithium metal without dendrites with an ultrahigh areal capacity of 10.0 mAh cm−2 (Figure 4d). The deposition of dendrite-free lithium is ascribed to the large active surface area and the microstructural feature of the CNT sponge. The BET surface area of the CNT sponge reaches 113.8 m2 g−1 while that of the planar copper foil (11 μm thick) is only 0.01 m2 g−1. It was reported that the large active surface area plays an important role in lithium deposition.21,30 It ensures the good electric contact between the electrode and the electrolyte, thereby reducing the local current density. In addition, with the increase of the contact surface (via the SEI layer, of course) between the electrolyte and the electrode, the deposition sites of the metallic lithium rises up and dendrite-free lithium can grow uniformly. 497

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499

Letter

Nano Letters

Figure 5. Evolution of the resistance of lithium deposition on the CNT sponge. (a) Comparison of the impedance spectra of the cell at the discharge and charge states in some selected cycles (inset for the equivalent circuit). Fitting results of electrochemical impedance spectra in the initial 10 cycles of the (b) discharged and (c) charged CNT sponge.

onset potential of the nucleation does not change with the cycling number (inset of Figure 2c), it seems that the decreased polarization of the cell cannot be the main reason for the rising of the nucleation potential. In summary, the nucleation and plating/stripping behaviors of metallic lithium were studied on the CNT sponge as a 3D porous current collector. The lithium-storage capability of the porous CNT above 0.0 V makes it a “pre-lithiated” substrate, enhances its affinity with the subsequently deposited metallic lithium below 0.0 V, and lowers the lithium nucleation overpotential. The high surface area of the CNT sponge increases the density of the lithium nucleation sites, decreases the local current density on the CNTs, and ensures the uniform lithium deposition. Therefore, dendrite-free lithium plating was realized on the CNT sponge until a capacity of 10.0 mAh cm−2. Due to the stable graphitic-amorphous carbon composite structure of the CNT sponge, the lithium plating/ stripping on the sponge is stable for 90 cycles (cycling in progress) and with high Coulombic efficiencies. These results exhibit the feasibility of using the CNT sponge as the 3D porous current collectors for lithium deposition. These findings shed light on designing and developing advanced current collectors for the lithium electrode and will promote the commercialization of the secondary lithium batteries.

Moreover, both the graphitic layers and the graphene edges on the surface can store a large amount of lithium, assigning the CNTs with outstanding chemical affinity with the metallic lithium and low nucleation energy.27 Therefore, metallic lithium can be plated on the CNT sponge up to 10.0 mAh cm−2 without formation of dendrites. The electrochemical impedance spectrum (EIS) of the cell in the first 10 cycles was recorded (Figure 5) to understand the nucleation and plating/stripping behaviors of lithium on the CNT sponge. Figure 5a shows that, with increasing cycling number, the impedance of each cell at the charged (lithium stripping) state is lower than at the discharged (lithium plating) state but the potential gap of lithium plating/stripping decreases in the first 5 cycles. Further analysis shows that the values of the ohmnic resistance (Rs; the total resistance for the electrolyte, separator, and electrical contact resistances) and the SEI layer resistance (R1) keep roughly unchanged in the first 10 cycles (Figure 5b,c). The stable resistance for the SEI layer is ascribed to the negligible reactivity between the deposited lithium and the electrolyte, beneficial for the high Coulombic efficiency (after the first cycle) and long-term cyclability of the cell. However, different from that of the other lithium cells (the lithium ion cells, for example), the resistance for the charge transfer on the electrolyte/electrode (the lithium metal) interface, R2, decreases sharply with the cycling number (Figure 5b,c). This explains the reduced polarization (or gap of the lithium plating/stripping potentials) in the voltage profile of the CNT sponge and means that the deposited lithium metal becomes more compact with cycling. With the above EIS results, we can understand the two interesting plating/stripping characteristics in the potential profiles (Figure 2c), the decreasing plating/stripping potential gap and the decreasing nucleation overpotential, and attribute them to the decreasing polarization of the CNT sponge electrode (Figure 5). The rising nucleation potential is partially attributed to the decreasing polarization of the cell (Figure 5) and partially to the increasing nucleation density.28 As the



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.8b04376. Additional experimental and figures including details on experimental methods, XPS results, and additional SEM, Raman, EDS, and electrochemical tests (PDF)



AUTHOR INFORMATION

Corresponding Author

*Phone: +86-10-82649050. E-mail: [email protected]. 498

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499

Letter

Nano Letters ORCID

(21) Yang, C.-P.; Yin, Y.-X.; Zhang, S.-F.; Li, N.-W.; Guo, Y.-G. Nat. Commun. 2015, 6, 8058. (22) Wang, S.-H.; Yin, Y.-X.; Zuo, T.-T.; Dong, W.; Li, J.-Y.; Shi, J.L.; Zhang, C.-H.; Li, N.-W.; Li, C.-J.; Guo, Y.-G. Adv. Mater. 2017, 29 (40), 1703729. (23) Li, Z.; Huang, J.; Yann Liaw, B.; Metzler, V.; Zhang, J. J. Power Sources 2014, 254, 168−182. (24) Lin, D.; Liu, Y.; Liang, Z.; Lee, H.-W.; Sun, J.; Wang, H.; Yan, K.; Xie, J.; Cui, Y. Nat. Nanotechnol. 2016, 11 (7), 626−632. (25) Che, G.; Lakshmi, B. B.; Fisher, E. R.; Martin, C. R. Nature 1998, 393, 346−349. (26) Sun, Z.; Jin, S.; Jin, H.; Du, Z.; Zhu, Y.; Cao, A.; Ji, H.; Wan, L.J. Adv. Mater. 2018, 30 (32), 1800884. (27) Jiao, J.; Xiao, R.; Tian, M.; Wang, Z.; Chen, L. Electrochim. Acta 2018, 282, 205−212. (28) Pei, A.; Zheng, G.; Shi, F.; Li, Y.; Cui, Y. Nano Lett. 2017, 17 (2), 1132−1139. (29) Liu, L.; Yin, Y.-X.; Li, J.-Y.; Li, N.-W.; Zeng, X.-X.; Ye, H.; Guo, Y.-G.; Wan, L.-J. Joule 2017, 1 (3), 563−575. (30) Zuo, T.-T.; Wu, X.-W.; Yang, C.-P.; Yin, Y.-X.; Ye, H.; Li, N.W.; Guo, Y.-G. Adv. Mater. 2017, 29 (29), 1700389.

Bin Xu: 0000-0001-5177-8929 Zhaoxiang Wang: 0000-0002-1123-6591 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We acknowledge the financial support of this work by the National Natural Science Foundation of China (NSFC Grant No. 51372268) and the National Key Development Program of China (Grant No. 2015CB251100).



ABBREVIATIONS CNT, carbon nanotube; SEM, scanning electron microscopy; SEI, solid electrolyte interphase; XPS, X-ray photoelectron spectroscopy



REFERENCES

(1) Armand, M.; Tarascon, J. M. Nature 2008, 451 (7179), 652− 657. (2) Etacheri, V.; Marom, R.; Elazari, R.; Salitra, G.; Aurbach, D. Energy Environ. Sci. 2011, 4 (9), 3243−3262. (3) Goodenough, J. B.; Kim, Y. Chem. Mater. 2010, 22 (3), 587− 603. (4) Cheng, X.-B.; Zhang, R.; Zhao, C.-Z.; Zhang, Q. Chem. Rev. 2017, 117 (15), 10403−10473. (5) Xu, W.; Wang, J.; Ding, F.; Chen, X.; Nasybulin, E.; Zhang, Y.; Zhang, J.-G. Energy Environ. Sci. 2014, 7 (2), 513−537. (6) Brissot, C.; Rosso, M.; Chazalviel, J. N.; Lascaud, S. J. Power Sources 1999, 81-82, 925−929. (7) Zheng, J.; Engelhard, M. H.; Mei, D.; Jiao, S.; Polzin, B. J.; Zhang, J.-G.; Xu, W. Nature Energy 2017, 2, 17012. (8) Liu, S.; Li, G.-R.; Gao, X.-P. ACS Appl. Mater. Interfaces 2016, 8 (12), 7783−7789. (9) Ding, F.; Xu, W.; Graff, G. L.; Zhang, J.; Sushko, M. L.; Chen, X.; Shao, Y.; Engelhard, M. H.; Nie, Z.; Xiao, J.; Liu, X.; Sushko, P. V.; Liu, J.; Zhang, J.-G. J. Am. Chem. Soc. 2013, 135 (11), 4450−4456. (10) Li, X.; Zheng, J.; Ren, X.; Engelhard, M. H.; Zhao, W.; Li, Q.; Zhang, J.-G.; Xu, W. Adv. Energy Mater. 2018, 8 (15), 1703022. (11) Markevich, E.; Salitra, G.; Chesneau, F.; Schmidt, M.; Aurbach, D. ACS Energy Letters 2017, 2 (6), 1321−1326. (12) Chen, S.; Zheng, J.; Mei, D.; Han, K. S.; Engelhard, M. H.; Zhao, W.; Xu, W.; Liu, J.; Zhang, J. G. Adv. Mater. 2018, 30 (21), 1706102. (13) Luo, J.; Fang, C.-C.; Wu, N.-L. Adv. Energy Mater. 2018, 8 (2), 1701482. (14) Lin, L.; Liang, F.; Zhang, K.; Mao, H.; Yang, J.; Qian, Y. J. Mater. Chem. A 2018, 6 (32), 15859−15867. (15) Li, N.-W.; Yin, Y.-X.; Yang, C.-P.; Guo, Y.-G. Adv. Mater. 2016, 28 (9), 1853−1858. (16) Cheng, X.-B.; Hou, T.-Z.; Zhang, R.; Peng, H.-J.; Zhao, C.-Z.; Huang, J.-Q.; Zhang, Q. Adv. Mater. 2016, 28 (15), 2888−2895. (17) Ryou, M.-H.; Lee, Y. M.; Lee, Y.; Winter, M.; Bieker, P. Adv. Funct. Mater. 2015, 25 (6), 834−841. (18) Park, J.; Jeong, J.; Lee, Y.; Oh, M.; Ryou, M.-H.; Lee, Y. M. Adv. Mater. Interfaces 2016, 3 (11), 1600140. (19) Li, Y.; Jiao, J.; Bi, J.; Wang, X.; Wang, Z.; Chen, L. Nano Energy 2017, 32, 241−246. (20) Zhao, H.; Lei, D.; He, Y.-B.; Yuan, Y.; Yun, Q.; Ni, B.; Lv, W.; Li, B.; Yang, Q.-H.; Kang, F.; Lu, J. Adv. Energy Mater. 2018, 8 (19), 1800266. 499

DOI: 10.1021/acs.nanolett.8b04376 Nano Lett. 2019, 19, 494−499