Hyperbranched β-Cyclodextrin Polymer as an Effective

Letter pubs.acs.org/NanoLett

Hyperbranched β‑Cyclodextrin Polymer as an Effective Multidimensional Binder for Silicon Anodes in Lithium Rechargeable Batteries You Kyeong Jeong,†,⊥ Tae-woo Kwon,†,⊥ Inhwa Lee,‡ Taek-Soo Kim,‡,∥ Ali Coskun,*,†,§,∥ and Jang Wook Choi*,†,∥ †

Graduated School of EEWS, ‡Department of Mechanical Engineering, §Department of Chemistry, and ∥Center for Nature-inspired Technology (CNiT), KAIST Institute NanoCentury, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, 305-701, Republic of Korea S Supporting Information *

ABSTRACT: Polymeric binders play an important role in electrochemical performance of high-capacity silicon (Si) anodes that usually suffer from severe capacity fading due to unparalleled volume change of Si during cycling. In an effort to find efficient polymeric binders that could mitigate such capacity fading, herein, we introduce polymerized β-cyclodextrin (β-CDp) binder for Si nanoparticle anodes. Unlike one-dimensional binders, hyperbranched network structure of β-CDp presents multidimensional hydrogen-bonding interactions with Si particles and therefore offers robust contacts between both components. Even the Si nanoparticles that lost the original contacts with the binder during cycling recover within the multidimensional binder network, thus creating a selfhealing effect. Utilizing these advantageous features, β-CDp-based Si electrode shows markedly improved cycling performance compared to those of other well-known binder cases, especially when combined with linear polymers at an appropriate ratio to form hybrid binders. KEYWORDS: Binder, cyclodextrin, cycle life, lithium ion battery, self-healing, silicon anode

T

It has been suggested26 that ideal binders for Si electrodes hold stiff polymeric backbones as well as efficient interactions with native silicon oxide layers on the Si surfaces and copper (Cu) oxides on the Cu current collectors. In particular, focusing on the latter aspect recent investigations have revealed that the polymeric binders containing carboxylic acid and/or hydroxyl functional groups can result in superior cycling performance utilizing hydrogen bonding interactions with Si active materials as compared to that of the conventional poly(vinylidene fluoride) (PVDF) binders functioning based mainly on van der Waals interactions.27 The representative binders possessing the hydrogen bonding capabilities are carboxymethyl cellulose (CMC),27−31 poly(acrylic acid),27,31−33 and alginate (Alg).34 Even within these series of binders, the number of carboxylate sites per given molecular weight of polymer directly affects their interactions with Si and thus cycle life.32 Although the improved cycle life based on these polymeric binders represent a significant progress in the Si anode research, the linear, onedimensional (1D), backbones of these polymers limit their interactions with Si to only point or linear contacts. Therefore,

he battery community has paid significant attention to silicon (Si) as a lithium ion battery (LIB) anode material on account of its exceptional theoretical gravimetric capacity (∼4200 mAh g−1), that is approximately 10 times larger than that of the conventional graphite (∼370 mAh g−1).1−4 High gravimetric capacity of Si is expected to fulfill an important role in bringing many of the key LIB applications5−7 to a reality. Good examples along this direction are advanced portable electronic devices requiring high power consumption and electrical vehicles with longer driving distance per each charge. Despite the advantageous gravimetric capacity, Si suffers from short cycle life originating from its large volume change during repeated Li (de)insertion. Upon full lithiation forming an alloy of Li4.4Si, Si undergoes volume expansion of ∼300% with respect to its initial state.1−3 Such unparalleled volume change of Si during battery operations triggers fatal fading mechanisms such as pulverization of active materials, loosened contacts between Si and carbon conductive agents, film delamination from current collectors, and destabilized solidelectrolyte interphase (SEI) formation.8−11 While diverse smart electrode designs12,13 involving various nanostructured Si have been demonstrated2,14−21 to be effective in mitigating these ruinous mechanisms, the choice of binder has also turned out to affect cell performance significantly.22−25 © 2014 American Chemical Society

Received: November 14, 2013 Revised: January 7, 2014 Published: January 13, 2014 864

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

Letter

Figure 1. (a,b) Structural formulas and graphical representations of (a) the Alginate (Alg) and (b) β-CD polymer (β-CDp) binders. The side chains of β-CDp could be dihydroxypropyl moieties in monomer or dimer or trimer form. The bridges could be glyceryl moieties in each form. (c,d) Schematic representations of Si-binder configurations for (c) SiAlg and (d) Siβ‑CDp during lithiation/delithiation of Si (sphere). Hyperbranched structure of β-CDp offers multidimensional hydrogen bonding interactions, which enable the binder to maintain the interactions with Si during continuous volume change of Si via self-healing process.

CDp and more uniform distribution of both binders due to the series of noncovalent interactions. β-CDp can be obtained by reacting of β-CD with EPI under the strong basic conditions. Hyperbranched β-CDp consists of the β-CD units containing various hydroxyl groups in the side chains and bridges due to the undiscriminating reactions of hydroxyl groups in β-CD with EPI. As a result, upon polymerization the hyperbranched β-CDp network structure incorporating a series of hydroxyl and ether groups can be attained. β-CDp could therefore generate multidimensional contacts with Si particles and provide superior mechanical strength against the volume expansion of Si when compared against the conventional linear binders. Furthermore, the hydrogen bonding interactions between hyperbranched polymeric network and Si particles can create a self-healing effect36 for the Si particles missing the initial contacts, thus contributing to the improved cyclability of the electrode. In addition, β-CDp is insoluble in the carbonate-based electrolyte solvents (Supporting Information Figure S1), supporting its stable character in the most commonly used LIB electrolytes. From the structural viewpoint, the binding mechanisms of Si electrodes with Alg and β-CDp are expected to be different (Figure 1c,d). During the Li insertion, Alg strands are forced to

it is very desirable to develop new polymeric binders, which can function beyond such interactions. From the binder design perspective, a natural extension for further improvement in the cycle life is construction of the increased number of contacts between Si and binders. On the basis of this conceptual motivation, herein, we introduced for the first time hyperbranched β-cyclodextrin polymer (β-CDp) as a new Si anode binder. β-CD is a 7-membered sugar macrocycle incorporating glucose as monomeric units. In the practical viewpoint, the use of β-CD is especially meaningful because it can be easily synthesized from ordinary starch by established35 enzymatic processes. Presence of primary and secondary alcohols renders β-CD as an ideal backbone for further functionalization with epichlorohydrin (EPI) in order to form hyperbranched β-CDp with various hydroxyl groups. Unlike conventional 1D polymeric binders, β-CDp presents multiple noncovalent interactions at local spots of Si (Figure 1) while its interconnected hyperbranched network structure forms a stable electrode film, resulting in a marked improvement in both capacity and cycle life. We have also demonstrated a hybrid binder approach where addition of an optimal amount of 1D Alg binder further improves electrochemical performance due to decreased agglomeration of β865

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

Letter

stretch or move by the expanded Si particles compared to their initial state. Up to this point, the Si-Alg interactions could be maintained. Upon delithiation in the same cycle, the Si particles shrink back to their original state; the Alg binder, however, cannot fully follow the shrinkage of Si, thus leading to a contact loss between both components. This issue becomes more prominent over extended cycling. On the contrary, β-CDp constitutes a polymeric network in which the fractal-like structure of hyperbranched main-backbone of β-CDp and its various bulky side groups are physically entangled and grasp each other, leading to reinforced binding ability with Si particles via purely noncovalent interactions, for example, hydrogen bonding and the van der Waals force. Thus, in this polymeric network the β-CDp binder provides multidimensional noncovalent interactions with the Si surfaces through the congested bulky side groups (Figure 1d). These interactions not only allow the β-CDp binder to accommodate the massive volume expansion of Si during lithiation but also keep the Si-binder interactions even during delithiation. In particular, once again the Si particles that lose contact with the binder at certain points of the delithiation and following cycles could recover the interactions at other points in the hyperbranched binder network. Hence, self-healing effect could come into play throughout battery operations, resulting in superior cycling performance. In order to assess the binding strength of β-CDp and Alg to the Si surface, the Si/binder blends were characterized by X-ray photoelectron spectroscopy (XPS) before and after washing the samples with water (Supporting Information Figures S2−S3). The C1s spectrum (Supporting Information Figure S2a) of Si/ β-CDp exhibited the peaks associated with the C−C, C−O, and O−C−O bonds at 285.0, 286.8, and 287.9 eV, respectively. Similarly, the C1s spectrum (Supporting Information Figure S2b) of Si/Alg indicates the presence of four main peaks assigned to the C−C (284.9 eV), C−O (286.8 eV), O−C−O (287.9 eV), and COONa (289.0 eV) bonds. Unexpectedly, in the case of Si/Alg the C−C bond peak was intensified after the washing, which can be attributed to the degradation and/or chemical reactions of Alg with hydroxyl groups on the Si surface.34,37 But, further investigation may be required for indepth understanding of such abnormal trend. For each binder, after the washing the integrated area of these C1s peaks decreased due to the loss of binders from Si surfaces, but this effect was more pronounced in the case of Alg, reflecting its weaker interactions with Si. More quantitative results were obtained by engaging thermogravimetric analyses (TGA). The TGA profiles (Supporting Information Figure S4) indicate that after the same washing step the Si/β-CDp and Si/Alg blends retained 64.2 and 19.9% of the original amounts of the binders (Figure 2a), respectively, supporting the enhanced Si-binder adhesion in the Si/β-CDp on account of the multidimensional noncovalent interactions between β-CDp and Si. It should be noted that the solubility of binder could affect the binder-toparticle interaction during the washing process. As shown in Supporting Information Figure S1, despite the substantial solubility of β-CDp in water, the strong interactions of β-CDp with Si particles are able to maintain the binder-to-particle interactions during the aggressive washing process. In order to investigate the binder effect on the mechanical stability of the bulk-scale electrode film, the conventional 180° peeling tests 11,26,38 were carried out (see Supporting Information). While the mass ratio was Si/binder = 3:1, the film thickness of each electrode was 15 μm on average. The

Figure 2. (a) The comparative amounts of the binders in the Si/binder blends calculated based on the TGA analyses before and after washing the blends with water. (b) Bulk-scale peeling characterizations for the Si/binder films based on both binders.

force versus displacement graph (Figure 2b), which is an indication of the adhesion between Si, binder, and copper foil, showed clearly distinctive behavior for each binder case, as the maximum force (denoted as *) was measured to be 2.00 and 1.72 N for the Si/β-CDp and Si/Alg films, respectively. The average force required for the peeling of the Si/β-CDp film (1.73 N) was also significantly higher when compared to that of the Si/Alg film (1.01 N). These values provide further strong evidence that utilization of the multidimensional hydrogen bonding interactions in the hyperbranched polymeric network gives higher mechanical stability to the Si/β-CDp film when compared to the Si/Alg film. In addition, the Si/Alg film exposed a significantly larger portion of Cu on its top surface of the electrode after the peeling test compared to the Si/β-CDp film (Supporting Information Figure S5). It can be concluded that the adhesion of the electrode film to the Cu current collector is a bottleneck in the overall mechanical stability of the Si/Alg film. On the contrary, the Si/β-CDp film has enhanced adhesion at such interface, keeping most of the electrode film (instead of Cu) on its top surface after the same peeling test, which suggests clearly that the β-CDp improves the adhesion between the electrode and the Cu current collector along with the adhesion between Si particles and binder within the electrode film. The binder effect on the battery performance was also investigated by a series of galvanostatic measurements. For this testing, coin-type half-cells incorporating Li metal as both the reference and counter electrodes were prepared. Lithium hexafluorophosphate was (1.15 M LiPF6) dissolved in the cosolvent mixture of ethylene carbonate (EC), ethyl methyl carbonate (EMC), and diethyl carbonate (DEC) (3:5:2, v/v/v) and used as the electrolyte. The electrolyte also contains 5 wt % 866

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

Letter

Figure 3. (a) The first voltage profiles of Siβ‑CDp, Siβ‑CD, and SiAlg, and inset shows the ICE for each sample. The cells were measured at C/10 (420 mA g−1) in the potential range of 0.01−1.0 V (versus Li/Li+). (b) The cycle performance of Siβ‑CDp and Siβ‑CD measured at 1C (4200 mA g−1). For both cases, the compositions of the electrode were Si/Binder/SuperP = 40:40:20 by weight, and the loading amounts of Si were 0.3 mg cm−2. (c) Comparison of the cycle performance of Siβ‑CDp with commonly used 1D binder-based electrodes (SiAlg and SiPVDF) when measured at 1C (4200 mA g−1). (d) The cycle performance of the Si electrodes based on a hybrid binder approach employing different ratios of β-CDp and Alg. In (c,d), the electrode compositions were Si/Binder/SuperP = 60/20/20 by weight, and the loading amounts of Si were 0.6 mg cm−2.

its specific capacity dropped severely to 460 mA g−1 after the same number of cycles, which corresponds to 25.4% capacity retention, thus reconfirming the importance of the hyperbranched polymer network and self-healing ability for stable cycling. From control experiments (Supporting Information Figure S1), it was found that unlike β-CD, β-CDp is insoluble in the electrolyte. This solubility difference may play an important role in the distinctive cycling performance between these two electrodes. The Coulombic efficiency (CE) of Siβ‑CDp reached 99.6% at the 200th cycle and the average CE in the cycling range of 2−200 was 99.0%. Further increase of CEs is desirable for use of β-CDp in practical full-cell applications and could be achieved by engaging optimal electrolyte additives and composite electrode structures including carbon. As depicted in Figure 3c, the cycling performance of Siβ‑CDp was also compared with that of the same Si electrodes but incorporating conventional 1D binders including Alg and PVDF (denoted as SiAlg and SiPVDF). All of the electrodes were measured at 1C for both charge and discharge over repeated cycles. While the capacity of SiPVDF dropped in the early cycling period below 5 cycles, Siβ‑CDp and SiAlg exhibited capacity retentions of 50.6 and 27.1%, respectively, after 150 cycles. The poor performance of SiPVDF can be explained by its weak van der Waals interactions between the binder and Si surface that cannot accommodate significant volume change of Si during the lithiation/delithiation. Even between β-CDp and Alg that function based on hydrogen bonding interactions, Siβ‑CDp showed better cycling performance compared to that of SiAlg,

of fluoroethylene carbonate (FEC) as an additive. In a typical test, the cells were cycled at C/10 (420 mA g−1) in the precycles and were then tested at various higher C-rates in the following cycles. All of the C-rates in the current study were defined based on the 1C value (4200 mA g−1), not actual charging/discharging durations. Also, in order to clearly elucidate the importance of hyperbranched network structure, β-CD itself (not polymerized) was also tested as a control binder. Hereafter, the Si electrodes based on β-CDp, β-CD, and Alg are denoted as Siβ‑CDp, Siβ‑CD, and SiAlg, respectively. While the first galvanostatic profiles of the three samples exhibited characteristic plateaus1,39,40 of Si at 0.1 V for lithiation and 0.45 V for delithiation (Figure 3a), the initial Coulombic efficiency (ICE) of the given samples were different such that Siβ‑CDp, Siβ‑CD, and SiAlg showed 85.9, 75.6, and 84.7%, respectively. The superior ICE of Siβ‑CDp compared to those of the other two electrodes suggests that the hyperbranched polymeric network of β-CDp promotes the reversibility during the first (de)lithiation perhaps by forming more stable SEI layers. The enhanced reversibility was also reflected in the cycle life. Figure 3b comparatively displays the cycling performance of Siβ‑CDp and Siβ‑CD over 200 cycles when measured at 1C for both charge and discharge in each cycle. Both cycle life and specific capacity were significantly better for Siβ‑CDp compared to those of Siβ‑CD. The specific capacity of Siβ‑CDp started at 2142 mA g−1 in the first cycle and ended at 1471 mA g−1 at the 200th cycle, corresponding to 68.7% capacity retention. By contrast, Siβ‑CD started at a lower capacity of 1810 mA g−1 and 867

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

Letter

Figure 4. Proposed mechanism of synergistic effect in the β-CDp/Alg hybrid binder. (a) The agglomerative property of β-CDp due to the intramolecular hydrogen bonding. (b) The stretched feature of Alg due to the electrostatic repulsions between carboxylates. (c) The configuration of the hybrid binder. The positively cooperative effects promote a more homogeneous polymeric network. Moreover, the Na+ originally from Alg also contributes to the good distribution and stability of the hybrid binder by creating a “glue effect”.

Figure 5. SEM images of Siβ‑CDp and SiAlg at different stages during the first full cycle. Left-column, Siβ‑CDp; right-column, SiAlg. (a,b) Before any electrochemical process; (c,d) after full lithiation; (e,f) after delithiation. 868

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

Letter

(also see gel permeation chromatography data in Supporting Information Figure S9). In addition, sodium ions from Alg can coordinate with oxygen units of β-CDp, thus contributing further to good distribution of the hybrid binder. Hence, the optimal point of 13 wt % can be understood that at this point Alg strands can suppress the aggregation of β-CDp most effectively without impairing the advantageous characteristics of β-CDp. Other electrochemical properties of Siβ‑CDp including rate performance, self-discharge, and loading dependence are presented in Supporting Information Figures S10−S12. The surface morphology of each Si electrode was characterized using scanning electron microscopy (SEM) to see the mechanical stability of the electrode film (Figure 5). Both Siβ‑CDp and SiAlg exhibited uniform particle distributions in their initial states before cycling (Figure 5a,b) as well as similar SEI formation filling the void space between Si particles after the first lithiation (Figure 5c,d). The difference between the samples, however, became evident after a full cycle such that Siβ‑CDp still preserved the uniform particle distribution to a large extent (Figure 5e), whereas SiAlg clearly showed micrometerscale cracks over the entire area of the film (Figure 5f), indicating the β-CDp binder is better at preserving the original film morphology utilizing its superior multidimensional binding capability based on the hyperbranched polymeric network. In addition, the electrochemical impedance spectroscopy measurements conducted at every 50 cycle (Supporting Information Figure S13) indicate that the interfacial resistance increased in the first 100 cycles but became saturated thereafter, suggesting stable interfacial stability of Siβ‑CDp during cycling. In conclusion, we have introduced a multidimensional hyperbranched β-CD polymer as an excellent binder for Si anodes in LIBs. The hyperbranched polymeric network originating from the unique macrocycle structure of β-CD and its polymerization in all of the three dimensions facilitates enhanced Si-binder interactions as well as the self-healing ability, making substantial progress in improving mechanical stability of the electrode film and therefore resolving the chronic insufficient cycle lives of Si anode. Furthermore, the versatile use of β-CD in a wide range of other conventional products and its exceptionally easy polymerization process render the current β-CDp binder more attractive for its practical adoption in actual battery manufacturing processes. Also, the current β-CDp binder should be readily applicable to other high capacity LIB active materials with oxide surfaces that suffer from large volume changes during battery operations.

which verifies again the significance of the unique multidimensional hydrogen bonding interactions and the self-healing ability in Siβ‑CDp. The superior performance of Siβ‑CDp was also observed when compared against linear CMC binder (Supporting Information Figure S6). The initial capacity decay of Siβ‑CDp is speculated to be associated with the dynamic environments around Si particles in the beginning cycling period. During the initial cycles where the electrode continuously finds more stable structure, the originally agglomerated Si particles could get segregated into smaller secondary particles during volume expansion of Si, resulting in increased number of interactions between Si and the binder. As a consequence, Li ion diffusion around the Si particles could be more hindered by the presence of the binder, causing the decrease in the capacity. The initial capacity decay is also expected to be from a certain portion of Si particles that are detached from the agglomeration and thus permanently lost, again during the volume expansion of Si. However, once such microscopic electrode structure becomes stabilized, most Si particles are electrochemically activated via accessible Li ions through their established diffusion pathways while the overall electrode structure and given capacity are well preserved during successive cycling. To understand the initial capacity decay further, the same experiment was conducted at a lower current density (3250 vs 4200 mA g−1), and such capacity decay disappeared (Supporting Information Figure S7). At lower current density, Li diffusion takes place during a longer duration of charging process, and the specific capacity in each cycle is less affected by surrounding binder around Si particles, thus leading to normal capacity retention during cycling and also supporting our suggested mechanism on the initial capacity decay. Also, the smaller specific capacities in Figure 3c compared to those in Figure 3b are attributed to the distinctive mass loadings of Si (0.3 vs 0.6 mg cm−2). In addition, from a series of control experiments (Supporting Information Figure S8), it turned out that the specific capacity of β-CDp alone is only 2% of that of Si, thus indicating that the β-CDp contribution to the overall capacity of Siβ‑CDp is very minor. In an effort to further improve the electrochemical performance of the Si electrode, a hybrid binder approach in which hyperbranched β-CDp was mixed with 1D Alg at different weight ratios was investigated (Figure 3d). The Si electrode showed the best cycling performance when the amount of the Alg binder was 13 wt %. The binder amount above or below this value resulted in inferior performance. The capacity retentions of the Si electrodes containing 0, 13, and 27 wt % were 54.4, 70.6, and 63.7%, respectively, after 100 cycles. The presence of the optimal binder amount might be related to conformation of β-CDp (Figure 4). Generally, the neutral polymer that has the attractive interaction sites tends to be entangled by itself due to strong intramolecular interactions. For this reason, β-CDp could form aggregate through the intramolecular hydrogen bonding (Figure 4a), which could deteriorate its interaction with other neighboring polymer chains and thus the overall mechanical stability of the Si electrode. In contrast, the conformation of Alg could be stretched because its carboxylates (COO−) impart the Coulombic repulsion (Figure 4b). Thus, the combination of these two polymers could exhibit a mutually beneficial outcome in a way that while the addition of Alg into β-CDp diminishes the agglomeration of β-CDp, β-CDp can neutralize electrostatic repulsion between the carboxylate units in Alg (Figure 4c), resulting in a more uniformly distributed polymeric network

■

ASSOCIATED CONTENT

S Supporting Information *

Binder solubility, XPS, TGA, FT-IR, peel test, electrochemical data of other control samples, gel permeation chromatography, and particle size analysis. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: (A.C.) [email protected]. *E-mail: (J.W.C.) [email protected]. Author Contributions ⊥

Y.K.J. and T.K. contributed equally to this work.

Notes

The authors declare no competing financial interest. 869

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870

Nano Letters

■

Letter

(29) Bridel, J. S.; Azaïs, T.; Morcrette, M.; Tarascon, J. M.; Larcher, D. Chem. Mater. 2010, 22 (3), 1229−1241. (30) Hochgatterer, N. S.; Schweiger, M. R.; Koller, S.; Raimann, P. R.; Wohrle, T.; Wurm, C.; Winter, M. Electrochem. Solid-State Lett. 2008, 11 (5), A76−A80. (31) Koo, B.; Kim, H.; Cho, Y.; Lee, K. T.; Choi, N.-S.; Cho, J. Angew. Chem., Int. Ed. 2012, 51 (35), 8762−8767. (32) Magasinski, A.; Zdyrko, B.; Kovalenko, I.; Hertzberg, B.; Burtovyy, R.; Huebner, C. F.; Fuller, T. F.; Luzinov, I.; Yushin, G. ACS Appl. Mater. Interfaces 2010, 2 (11), 3004−3010. (33) Yabuuchi, N.; Shimomura, K.; Shimbe, Y.; Ozeki, T.; Son, J.-Y.; Oji, H.; Katayama, Y.; Miura, T.; Komaba, S. Adv. Energy Mater. 2011, 1 (5), 759−765. (34) Kovalenko, I.; Zdyrko, B.; Magasinski, A.; Hertzberg, B.; Milicev, Z.; Burtovyy, R.; Luzinov, I.; Yushin, G. Science 2011, 334 (6052), 75−79. (35) Biwer, A.; Antranikian, G.; Heinzle, E. Appl. Microbiol. Biotechnol. 2002, 59 (6), 609−617. (36) Wang, C.; Wu, H.; Chen, Z.; McDowell, M. T.; Cui, Y.; Bao, Z. Nature Chem. 2013, 1−7. (37) Papageorgiou, S. K.; Kouvelos, E. P.; Favvas, E. P.; Sapalidis, A. A.; Romanos, G. E.; Katsaros, F. K. Carbohydr. Res. 2010, 345 (4), 469−473. (38) Xu, Y. H.; Yin, G. P.; Ma, Y. L.; Zuo, P. J.; Cheng, X. Q. J. Power Sources 2010, 195 (7), 2069−2073. (39) Hatchard, T. D.; Dahn, J. R. J. Electrochem. Soc. 2004, 151 (6), A838−A842. (40) Wen, C. J.; Huggins, R. A. J. Solid State Chem. 1981, 37 (3), 271−278.

ACKNOWLEDGMENTS J.W.C. acknowledges the financial support by the National Research Foundation of Korea (NRF) Grant funded by the Korea government (MEST) (NRF-2010-C1AAA001-0029031 and NRF-2012-R1A2A1A01011970). A.C. acknowledges the support from Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT, & Future Planning (2013R1A1A1012282).

■

REFERENCES

(1) Boukamp, B. A.; Lesh, G. C.; Huggins, R. A. J. Electrochem. Soc. 1981, 128 (4), 725−729. (2) Chan, C. K.; Peng, H.; Liu, G.; McIlwrath, K.; Zhang, X. F.; Huggins, R. A.; Cui, Y. Nat. Nanotechnol. 2007, 3 (1), 31−35. (3) ograve, A. S. A.; Bruce, P.; Scrosati, B.; Tarascon, J.-M.; van Schalkwijk, W. Nat. Mater. 2005, 4 (5), 366−377. (4) Huggins, R. A. J. Power Sources 1999, 81, 13−19. (5) Armand, M.; Tarascon, J. M. Nature 2008, 451 (7179), 652−657. (6) Tarascon, J. M.; Armand, M. Nature 2001, 414 (6861), 359−367. (7) Whittingham, M. S. Chem. Rev. 2004, 104 (10), 4271−4302. (8) Wu, H.; Chan, G.; Choi, J. W.; Ryu, I.; Yao, Y.; McDowell, M. T.; Lee, S. W.; Jackson, A.; Yang, Y.; Hu, L.; Cui, Y. Nat. Nanotechnol. 2012, 7 (5), 310−315. (9) Tamura, N.; Ohshita, R.; Fujimoto, M.; Kamino, M.; Fujitani, S. J. Electrochem. Soc. 2003, 150 (6), A679−A683. (10) Oumellal, Y.; Delpuech, N.; Mazouzi, D.; Dupré, N.; Gaubicher, J.; Moreau, P.; Soudan, P.; Lestriez, B.; Guyomard, D. J. Mater. Chem. 2011, 21 (17), 6201−6208. (11) Chen, Z.; Christensen, L.; Dahn, J. R. Electrochem. Commun. 2003, 5 (11), 919−923. (12) Liu, B. R.; Soares, P.; Checkles, C.; Zhao, Y.; Yu, G. H. Nano Lett. 2013, 13 (7), 3414−3419. (13) Wu, H.; Yu, G. H.; Pan, L. J.; Liu, N. A.; McDowell, M. T.; Bao, Z. A.; Cui, Y. Nat. Commun. 2013, 4, 1943−1946. (14) Magasinski, A.; Dixon, P.; Hertzberg, B.; Kvit, A.; Ayala, J.; Yushin, G. Nat. Mater. 2010, 9 (4), 353−358. (15) Hwang, T. H.; Lee, Y. M.; Kong, B.-S.; Seo, J.-S.; Choi, J. W. Nano Lett. 2012, 12 (2), 802−807. (16) Liu, N.; Wu, H.; McDowell, M. T.; Yao, Y.; Wang, C.; Cui, Y. Nano Lett. 2012, 12 (6), 3315−3321. (17) Choi, J. W.; McDonough, J.; Jeong, S.; Yoo, J. S.; Chan, C. K.; Cui, Y. Nano Lett. 2010, 10 (4), 1409−1413. (18) Jung, D. S.; Hwang, T. H.; Park, S. B.; Choi, J. W. Nano Lett. 2013, 13 (5), 2092−2097. (19) Zhou, S.; Liu, X.; Wang, D. Nano Lett. 2010, 10 (3), 860−863. (20) Golmon, S.; Maute, K.; Lee, S.-H.; Dunn, M. L. Appl. Phys. Lett. 2010, 97 (3), 033111. (21) Jung, D. S.; Ryou, M. H.; Sung, Y. J.; Park, S. B.; Choi, J. W. Proc. Natl. Acad. Sci. U.S.A. 2013, 110 (30), 12229−12234. (22) Liu, G.; Xun, S.; Vukmirovic, N.; Song, X.; Olalde-Velasco, P.; Zheng, H.; Battaglia, V. S.; Wang, L.; Yang, W. Adv. Mater. 2011, 23 (40), 4679−4683. (23) Liu, W.-R.; Yang, M.-H.; Wu, H.-C.; Chiao, S. M.; Wu, N.-L. Electrochem. Solid-State Lett. 2005, 8 (2), A100−A103. (24) Yue, L.; Zhang, L.; Zhong, H. J. Power Sources 2014, 247, 327− 331. (25) Murase, M.; Yabuuchi, N.; Han, Z.-J.; Son, J.-Y.; Cui, Y.-T.; Oji, H.; Komaba, S. ChemSusChem 2012, 5 (12), 2307−2311. (26) Ryou, M. H.; Kim, J.; Lee, I.; Kim, S.; Jeong, Y. K.; Hong, S.; Ryu, J. H.; Kim, T. S.; Park, J. K.; Lee, H.; Choi, J. W. Adv. Mater. 2013, 25 (11), 1571−1576. (27) Komaba, S.; Shimomura, K.; Yabuuchi, N.; Ozeki, T.; Yui, H.; Konno, K. J. Phys. Chem. C 2011, 115 (27), 13487−13495. (28) Li, J.; Lewis, R. B.; Dahn, J. R. Electrochem. Solid-State Lett. 2007, 10 (2), A17−A20. 870

dx.doi.org/10.1021/nl404237j | Nano Lett. 2014, 14, 864−870