Letter Cite This: Nano Lett. 2017, 17, 6808-6815
pubs.acs.org/NanoLett
Critical Role of Ultrathin Graphene Films with Tunable Thickness in Enabling Highly Stable Sodium Metal Anodes Huan Wang, Chuanlong Wang, Edward Matios, and Weiyang Li* Thayer School of Engineering, Dartmouth College, 14 Engineering Drive, Hanover, New Hampshire 03755, United States S Supporting Information *
ABSTRACT: Sodium (Na) metal has shown great promise as an anode material for the next-generation energy storage systems because of its high theoretical capacity, low cost, and high earth abundance. However, the extremely high reactivity of Na metal with organic electrolyte leads to the formation of unstable solid electrolyte interphase (SEI) and growth of Na dendrites upon repeated electrochemical stripping/plating, causing poor cycling performance, and serious safety issues. Herein, we present highly stable and dendrite-free Na metal anodes over a wide current range and long-term cycling via directly applying free-standing graphene films with tunable thickness on Na metal surface. We systematically investigate the dependence of Na anode stability on the thickness of the graphene film at different current densities and capacities. Our findings reveal that only a few nanometer (∼2−3 nm) differences in the graphene thickness can have decisive influence on the stability and rate capability of Na anodes. To achieve the optimal performance, the thickness of the graphene film covered on Na surface needs to be meticulously selected based on the applied current density. We demonstrate that with a multilayer graphene film (∼5 nm in thickness) as a protective layer, stable Na cycling behavior was first achieved in carbonate electrolyte without any additives over 100 cycles at a current density as high as 2 mA/cm2 with a high capacity of 3 mAh/cm2. We believe our work could be a viable route toward high-energy Na battery systems, and can provide valuable insights into the lithium batteries as well. KEYWORDS: Sodium metal anode, graphene film, tunable thickness, stable cycling
T
is also suffering from severe inherent problems due to its intrinsic reactive nature.12 Na can easily react with organicsolvent-based electrolyte, leading to the formation of unstable solid electrolyte interphase (SEI) upon repeated stripping/ plating during charge/discharge cycles, which results in Na dendrite growth (Figure 1a). The uncontrolled Na dendrite propagation and continuous breaking and reformation of SEI on high-surface-area dendritic structures cause the penetration of separator and rapid loss of both electrolyte and working Na, resulting in low Coulombic efficiency and short-circuit of the batteries. Therefore, a SEI layer with high uniformity and stability, as well as high ionic conductivity and mechanical strength is essential to ensure stable cycling performance of Na anode. Among previous efforts in constructing a stable SEI layer for metallic Na anode, it was found that that a uniform and compact SEI could be formed in ether-based electrolyte, such as glymes (mono-, di-, or tetraglyme) and dimethoxyethane.13−15 However, with respect to traditional carbonatebased electrolyte, which possesses wider electrochemical window, lower volatility, and is more industrially compatible than ether-based electrolyte, it is very challenging to realize a
he rapid development of renewable energy sources (such as solar and wind) and their intermittent nature stimulate the demand of grid-scale energy storage systems to meet the ever-increasing energy consumption.1,2 Rechargeable lithium (Li)-ion batteries currently dominate the market of portable consumer electronics and electric-powered vehicles, representing the highest performance batteries.3,4 However, the scarcity and geographical constraint of Li resource on a global scale as well as its increasing cost over time have largely impeded their implementation for grid-scale energy storage, motivating the development of Li-alloy or other cost-effective metal-based batteries.5−8 Among all promising alternative candidates, sodium (Na) is most appealing as a replacement for Li in battery chemistry because of its high similarity in chemical and physical properties as Li, low cost (raw materials for electrodes and salts for electrolyte), high natural abundance (over 1000 times more abundant than Li in earth crust), and accessibility. Metallic Na, as an anode material, has a high theoretical specific capacity (1166 mAh/g) and low electrochemical potential (−2.714 V vs standard hydrogen electrode).9 When paired with high capacity cathode materials, for example, sulfur (S) or oxygen (O2) at room temperature, the resulting Na−S and Na−O2 batteries can deliver high theoretical specific energies of 1274 and 1605 Wh/kg,10,11 respectively, which are 3 to 4 times higher than that of the Li-ion batteries. Sharing the long-standing challenges that metallic Li anode faces, Na anode © 2017 American Chemical Society
Received: July 18, 2017 Revised: September 12, 2017 Published: October 17, 2017 6808
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
Figure 1. (a,b) Schematic diagram showing the difference between bare Na and graphene film protected Na anodes. (a) Mossy and dendritic Na formation during repeated electrochemical stripping and plating. (b) Illustration of (i to iii) transferring free-standing graphene film onto Na metal surface, and (iv to vi) the high stability of graphene-coated Na anode during stripping/plating without the formation of Na dendrites. (c) UV−vis transmittance spectra of the graphene films with different number of layers on quartz substrate. The intrinsic graphene was grown on the Cu foil and then transferred onto the quartz substrate for UV−vis characterization. The black, red, and blue lines show the transmittance curves of graphene with single-layer, ∼7 layers, ∼15 layers, respectively, corresponding to a thickness of ∼0.33, ∼2.3, and ∼5 nm, respectively. Inset: photographs of the corresponding free-standing graphene/PMMA films. The shiny yellow window is the Cu residue, which was purposely protected by tape before etching to support the graphene/PMMA film. (d) Extensive Raman spectra for the graphene films with single-layer, ∼7 layers, and ∼15 layers. The red, black, cyan, and blue lines represent Raman spectra recorded at four different spots of each sample.
difficult to achieve. To the best of our knowledge, there were few reports on improving the cycling stability of Na anode that can work at relatively high current (>1 mA/cm2) and high capacity (>1 mAh/cm2) in conventional carbonate-based electrolyte. Graphene, a representative of two-dimensional (2D) Dirac materials, has attracted extensive attention since 2004 due to its extraordinary chemical and physical properties.18 The many unique features of graphene, including chemical inertness, strong mechanical strength,19 high flexibility and elasticity,20 impenetrability against molecules,21 and easy access for ion
stable SEI by merely tuning the electrolyte and Na salt composition. Recent research showed that the employment of a layer of coating on Na surface as an artificial SEI would be a more intriguing approach to stabilize Na metal in carbonate electrolyte.16,17 Hu and Kim’s groups separately reported the methods of stabilizing Na metal anode by coating a thin film of pure Al2O3 or Al2O3-particle−polymer composite as a protective layer on Na surface via atomic layer deposition or roll-pressing.16,17 While the cycling stability of Na anode was enhanced through these ceramic-based films, a stable Na anode working at high current density and high capacity is still quite 6809
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
Figure 2. Electrochemical performance of symmetric cells with FL-G/Na and bare Na electrodes. Galvanostatic cycling of symmetric cells made from FL-G/Na (red line) and bare Na (blue line) electrode at a current density and a cycling capacity of (a) 0.5 mA/cm2 and 0.5 mAh/cm2, respectively; (b) 0.25 mA/cm2 and 1 mAh/cm2, respectively; (c) 1 mA/cm2 and 1 mAh/cm2, respectively. (d) Enlarged voltage profile of the 1st, 50th, and 100th cycle at a current density of 0.5 mA/cm2 and a capacity of 0.5 mAh/cm2. (e) Nyquist plots of the symmetric cells with FL-G/Na electrode (red circle) and bare Na electrode (blue square) after 100 cycles.
diffusion via defects,22 make it an attractive candidate as an artificial SEI layer. However, it was reported that when using graphene film as a coating material on Li metal surface to stabilize Li, its effect was not quite satisfying: Cui et al. showed that a two-layer graphene grown on copper substrate improved the Coulombic efficiency for Li deposition/stripping to some extent, but the enhancement was not as prominent as that of boron nitride;23 Choi’s group also showed that multilayer graphene was not good enough as an interfacial layer to stabilize Li metal unless accompanied by Cs+ as an additive into the electrolyte.24 Thus, using pristine graphene film alone seems not appealing with elusive reasons, and the role of graphene on Na metal remains unexplored and is worth
investigating. Here in this paper, we present highly stable and dendrite-free Na metal anodes over a wide current range and long-term cycling by directly applying free-standing graphene films with tunable thickness (through controlling the number of graphene layers) on Na metal surface. To reveal the influence of graphene in detail, we systematically investigated the dependence of Na anode stability on the number of graphene layers at different current densities and capacities. We found that the optimal cycling performance of Na anode highly depends on the thickness of the graphene film on Na surface: a thickness difference in only a few nanometers (∼2−3 nm) can have decisive influence on the stability and rate capability of Na anode. We demonstrate that with a multilayer graphene film 6810
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
each sample. All spectra displayed three characteristic peaks of D (∼1350 cm−1), G (1590−1600 cm−1), and 2D bands (2690−2700 cm−1), which correspond to the defects, E2g vibrational mode of sp2 bonded carbon, and the second-order vibration (caused by scattering of phonons), respectively.27 Moreover, as the graphene thickness increases from a single layer to few layers or multilayers, the peak intensity of G band increases, and the 2D peaks become broader and blue-shifted due to the splitting electronic band structure of graphene layers. But the differences in G and 2D peaks between few-layer graphene and multilayer graphene are not very obvious, which is consistent with the previous report. 28 Notably, we deliberately lowered the graphene growth temperature to 800°C so as to reduce the conductivity of graphene to prevent Na deposition on the graphene surface (conventional graphene growth temperature is at 850−1000 °C).29,30 Additionally, lowering the growth temperature also introduced more point and line defects (proved by the prominent peaks of D band in Figure 1e), serving as channels to facilitate fast Na ions diffusion, which will be discussed in detail later. To evaluate the electrochemical performance of the asobtained G/Na electrode, Na metal electrodes coated by graphene with the above three different thicknesses were assembled as symmetric cells using two identical G/Na electrodes (assembled in 2032-type coin cell) for repeated stripping and plating experiment. The electrolyte used contains 1 M NaPF6 in ethylene carbonate/diethyl carbonate (EC/DEC, 1:1 by volume) without any other electrolyte additives. For comparison, symmetric cells using bare Na metal electrodes were assembled under the same experimental condition. Figure 2a compares the voltage profiles of the symmetric cells made from few-layer graphene (FL-G, ∼7 layers) covered Na (FL-G/ Na) electrode and bare Na electrode at a current density of 0.5 mA/cm2 with a cycling capacity of 0.5 mAh/cm2, respectively. It can be observed that the FL-G/Na electrode exhibited a much more stable voltage profile and a much lower overpotential over 100 cycles than that of the bare Na electrode. It is noted that the bare Na electrode displayed increasing voltage hysteresis over the first 44 cycles due to the increased impedance caused by the thick and excessive SEI formation, and then a voltage drop was observed at the 45th cycle, possibly attributed to the dendrite formation.31 Such voltage fluctuation was frequently shown for bare Na electrode, revealing the highly reactive nature between Na and the electrolyte. The comparison in the voltage hysteresis is more clearly presented in the enlarged voltage profile at the first, 50th, 100th cycle, as shown in Figure 2d. Moreover, single-layer graphene covered Na (SL-G/Na) electrode was investigated under the same cycling condition and revealed little effect on stabilizing Na compared to bare Na electrode (see Figure S3 in Supporting Information). This could be due to the insufficient thickness and the very high concentration of defects of singlelayer graphene film, which was unable to block the electrolyte from penetrating through. These results show that the number of graphene layers, or in other words, the graphene thickness plays a decisive role in improving the stability of Na anode. When the cycling capacity was increased to 1 mAh/cm2 and the current density was changed to 0.25 (Figure 2b) and 1 mA/cm2 (Figure 2c), the FL-G/Na electrode still displayed an excellent cycling stability and constantly low voltage hysteresis. In contrast, the bare Na electrode exhibited highly fluctuating and unstable voltage profiles and increased voltage hysteresis over cycling, indicating uncontrolled SEI formation and dendrite
(∼5 nm in thickness) as a protective layer on Na metal surface, stable Na cycling behavior was first achieved to the best of our knowledge in carbonate electrolyte without any additives over 100 cycles at a relatively high current density of 2 mA/cm2 with a high cycling capacity of 3 mAh/cm2. Graphene films were synthesized in a home-built chemical vapor deposition (CVD) system using a tube furnace. Typically, the precursor benzoic was sublimated upstream and transported by the carrier gas to the Cu foil (placed at the hot center of the tube furnace), on which the graphene film was grown (see Methods in Supporting Information for details). By adjusting the growth pressure and the flow of forming gas (95% of argon and 5% of hydrogen), the number of layers for the graphene films can be readily tuned from single layer to multilayers. Figure 1b illustrates the process that how the graphene film was transferred onto the Na metal surface and serves as an artificial SEI. In a typical procedure, poly(methyl methacrylate) (PMMA) film was first spin-coated on the Cu foil. Then, the Cu foil was etched away by ammonium persulfate ((NH4)2S2O8), followed by the cleaning and drying process, and thus a free-standing PMMA/graphene film was formed. The PMMA layer is critical in maintaining the integrity of graphene during the transferring process. Owing to the low melting point of Na metal, the as-obtained PMMA/graphene film was able to tightly attach to the Na surface while the Na was slightly heated at around 90 °C. When the PMMA/ graphene-coated Na metal was eventually used as electrode assembled in batteries, the PMMA film can be easily dissolved in the carbonate electrolyte,25 leaving the graphene layer remained on the Na metal surface. This graphene layer then serves as a protective artificial SEI to keep the Na surface smooth and stable without dendrite growth upon repeated stripping/plating during charge/discharge (Figure 1b, iv to vi). The graphene-coated Na is denoted by “G/Na” in the following content. Typical photographs of as-obtained PMMA/graphene films with different thicknesses are shown in the insets of Figure 1c. When the graphene is just a single layer, the PMMA/graphene film is almost transparent. As graphene thickness increases, the color of the PMMA/graphene film becomes darker, indicating the decrease of graphene transparency. Meanwhile, it can be observed that the graphene films are uniform, which could be further confirmed in the scanning electron microscope (SEM) images (Figure S1 in Supporting Information) and optical microscopy images (Figure S2 in Supporting Information). Ultraviolet−visible spectroscopy (UV−vis) was conducted to reveal the transparency and estimate the thickness of the graphene films. Taking the same blank quartz substrate as the reference, the UV−vis spectra show the optical transmittance of approximately 98.1%, 82.5%, and 62.0% at 550 nm, respectively (Figure 1c). Since a single-layer graphene gives an absorption of 2.3%,26 the number of graphene layers can be estimated to be single-layer, ∼7 layers, and ∼15 layers, corresponding to a thickness of approximately 0.33, 2.3, and 5 nm, respectively. This indicates that the as-synthesized graphene films tested as artificial SEI are ultrathin and the difference in their thicknesses is only within very small nanometer range. We further performed Raman spectroscopy to characterize the numbers of layers and defects of graphene. Extensive Raman spectra were recorded for the as-prepared graphene films with the above three different thicknesses on quartz substrates (Figure 1d). The red, black, cyan, and blue lines represent the Raman spectra taken at four different spots of 6811
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
Figure 3. Surface morphology characterization of few-layer graphene coated Na (FL-G/Na) electrode (left column, a−c) and bare Na electrode (right column, d−f) after 100 cycles of stripping/plating at a current density of 0.5 mA/cm2. (a,d) Photographs of (a) FL-G/Na electrode and (d) bare Na electrode. (b,c) SEM images of the FL-G/Na surface after 100 cycles. (e,f) SEM images of the bare Na surface. Scale bars in (b,e) and (c,f) are 200 and 20 μm, respectively.
growth. The difference in the electrode−electrolyte interface between the G/Na and the bare Na electrode was also confirmed by electrochemical impedance spectroscopy (EIS) measurement. In general, the high-frequency semicircle radius in the Nyquist plots is related to the charge transfer resistance at SEI.32 We found that the charge transfer resistances of the symmetric cells using bare Na and FL-G/Na electrode were almost the same before cycling (see Figure S4 in Supporting Information). This also indicates that the PMMA film had been dissolved into the electrolyte upon assembling the cell and did not increase the interfacial impedance. After 100 cycles, the charge transfer resistance for the symmetric cell made from FLG/Na could still remain the same as that before cycling, while the symmetric cell using bare Na electrode exhibited an extremely large interfacial impedance (Figure 2e), which was about three times higher than that before cycling. This comparison clearly suggests continuous buildup of SEI on bare Na electrode surface due to direct exposure to the electrolyte, whereas FL-G can act as a protective layer to prevent the electrolyte from attacking the Na electrode during the long-term cycling.
To examine and compare the surface condition of Na electrodes with/without graphene protection over cycling, the symmetric cells made from bare Na and FL-G/Na electrodes were dissembled after 100 cycles of electrochemical deposition/ stripping, and then the Na electrodes were taken out for further surface characterization. During the dissembling process, we noticed that the electrolyte added into the cell made from G/ Na electrode was well preserved, while little amount of electrolyte was observed in the cell made from bare Na. This indicates that the graphene film was able to effectively prevent the reaction between Na and the electrolyte. In addition, the graphene film was transferred to the surface of the separator during the cell dissembling, allowing us to clearly see the protected Na metal surface beneath the graphene film. As shown in the photographs of Na electrodes with/without graphene protection (Figure 3a,d), the surface of bare Na electrode shows a dark and blackish color, while the G/Na electrode still presents the color of a metallic luster. SEM characterization was further performed to investigate the surface morphology in detail at different magnifications. As shown in Figure 3e,f, mossy and fiber-like Na dendrites can be 6812
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
Figure 4. Cycling performance of the symmetric cells made from ML/G-Na electrodes (red) and bare Na electrodes (blue) at a current density 2 mA/cm2 with a cycling capacity of (a) 1 mAh/cm2, and (b) 3 mAh/cm2, respectively. (c,d) Enlarged voltage profiles of (c) the 7th to 9th cycle, and (d) the 58th to 60th cycle at a current density of 2 mA/cm2 with a capacity of 3 mAh/cm2. (e,f) SEM images of (e) ML-G/Na and (f) bare Na electrode surface after 100 cycles at a current density of 2 mA/cm2 with a capacity of 1 mAh/cm2. Scale bars, 200 μm. Insets: corresponding highmagnification SEM images. Scale bars: 10 μm.
clearly observed for bare Na electrode, which is a typical phenomenon in carbonate-based electrolyte. In comparison, for G/Na electrode (Figure 3b,c) the Na surface was very smooth without the growth of any dendritic and mossy structures (the darker spots in Figure 4c are the graphene residues on Na surface). Moreover, the surface of bare Na exhibits many obvious cracks (Figure 3e), while the G/Na still maintains high integrity over cycling (Figure 3b), revealing the protective effect of the graphene film. To evaluate the current limit at which the G/Na anode can work, a high current density of 2 mA/cm2 with a cycling capacity of 1 mAh/cm2 was applied to the symmetric cells using FL-G/Na and bare Na electrodes, respectively. We found that no obvious improvement in the cycling stability could be
observed (see Figure S5 in Supporting Information), indicating that such graphene thickness (∼7 layers with a thickness of ∼2.3 nm) could not prevent the Na dendrite growth and block the electrolyte from attacking the Na anode under such a high current density. However, a Na anode that can effectively work at both high current density and capacity is critical for practical implementation. To this end, we applied multilayer graphene (ML-G) with ∼15 layers, corresponding to a thickness of ∼5 nm as the protective film for Na anode. Figure 4 shows the comparison of the voltage profiles of the symmetric cells using ML-G covered Na (ML-G/Na) electrode and bare Na electrode at a current density of 2 mA/cm2 with different cycling capacities. When the capacity was set at 1 mAh/cm2 (Figure 4a), the ML-G/Na electrode displayed a more stable 6813
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters
conducted at a much lower temperature (800 °C) than that of regular graphene growth (>850 °C) so that high density of defects are present in the graphene layers, which serve as the channels for Na-ion diffusion. With respect to point defects (typically vacancies or interstitial atoms),35 their sizes are at a subnanometer scale, which could be too small to allow enough accumulation of Na to reach the critical nucleation concentration for Na dendrite growth. As to the line defects, they are mainly originated from the tilt grain boundaries separating two different lattice orientations due to the multiple graphene nucleation sites and random growth on polycrystalline Cu foil.34,36 Because the graphene growth on Cu substrate follows the surface catalysis mechanism, the upper-layer graphene can grow across the boundaries of the layer beneath,37 which means that the boundaries among different graphene layers are not completely overlapped. As a result, it is believed that increasing the number of graphene layers could decrease the possibility of line defects overlapping at a larger scale. Hence, it is reasonable to extrapolate that on one hand Na ions could diffuse via a tortuous path through a ML-G film (∼5 nm in thickness) at a relatively high current density (2 mA/cm2) to reach the Na metal beneath the graphene film during the plating process (see schematic shown in Figure S8 in Supporting Information). Similar diffusion path was also reported in previous study of multilayer graphene oxide.38 On the other hand, the nonoverlapping feature of the line defects over multiple graphene layers makes it difficult for the Na dendrites from penetrating through. In comparison to the ML-G coverage on Na metal, FL-G was not able to effectively protect the Na surface at relatively high current densities and cycling capacity. This could be due to the reason that it is easier for the Na deposits rapidly formed at the Na/graphene interface at high current density to penetrate through the thinner FL-G via adjacent line defects compared to ML-G, leading to unsatisfactory electrochemical performance (Figure S5 in Supporting Information). The surface morphology of FL-G/Na electrode was also investigated by SEM after 100 cycles at a current density of 2 mA/cm2 with a capacity of 1 mAh/cm2 (see Figure S9 in Support Information). Lots of random spherical Na deposits can be observed on FL-G covered Na surface, but the condition was still better than that of bare Na electrode (Figure 4f). These results and deductions clearly elucidate the critical roles of graphene thickness (in other words, the number of graphene layers) in the electrochemical behavior of G/Na anodes: a thickness difference in the range of only 2−3 nm can have significant influence on the rate capability of Na anodes. Therefore, in order to achieve the optimal performance, the thickness of the graphene film used to protect the Na metal anode needs to be meticulously selected based on the applied current densities. In summary, we have successfully presented highly stable Na anodes by employing free-standing graphene films with tunable thickness on Na metal surface. We discovered that the rate capability of Na anode is critically dependent on the thickness of graphene film, in other words, the number of graphene layers, which could be readily tuned via controlling the experimental parameters of the CVD system. Our results showed that a relatively thin graphene film (∼2.3 nm in thickness) is well suited for Na anode working at relatively low current density (≤1 mA/cm2) whereas a thick graphene film (∼5 nm in thickness) would be the best option to protect Na anode at high current density (2 mA/cm2). Moreover, to the best of our knowledge this is the first demonstration of a stable
voltage profile and lower voltage hysteresis compared to those of bare Na. These differences are more significantly revealed when the capacity was increased to 3 mAh/cm2 (Figure 4b). The magnified voltage profiles of the seventh to ninth cycle and the 58th to 60th cycle are shown in Figure 4c,d, respectively. At the eighth cycle, the voltage hysteresis of bare Na electrode reached 1 V and continuously increased to 3.5 V at the 58th cycle, displaying a seriously poor stability. In contrast, the MLG/Na electrode maintained a high cycling stability, demonstrating that the ML-G film can effectively protect and stabilize Na anode working at high current density and capacity. To further investigate the performance dependence on graphene thickness, the symmetric cells made from ML-G/Na electrode were also tested at a low current density of 0.25 mA/cm2 with a capacity of 1 mAh/cm2 (see Figure S6 in Supporting Information). The voltage profile was not very stable and exhibited a few overshoots. This phenomenon may be due to the reason that a current density as low as 0.25 mA/cm2 could not provide enough driving force for Na ions to diffuse through the thick layers of ML-G. In this case, some Na might deposit in the graphene interlayers rather than beneath the graphene film. SEM characterization was further performed to examine the morphology of Na electrode surface with or without ML-G protection after cycling at a higher current density. The Na electrodes were dissembled from the symmetric cells after 100 cycles at a current density of 2 mA/cm2 with a capacity of 1 mAh/cm2. As to ML-G/Na electrode (still with graphene covered on Na surface; no graphene transferred to the separator), it exhibited a smooth surface with very few random Na deposits (Figure 4e). On the other hand, this shows that the graphene film we synthesized are not conductive enough for Na to deposit on its surface, which excludes the possibility that the graphene may serve as a conductive network to smooth the current distribution on the electrode. Notably, the graphene wrinkles could be clearly observed (inset of Figure 4e), which are common features obtained during the graphene growth because of the difference in the coefficient of thermal expansion between Cu foil and graphene.26 In comparison, the surface of bare Na electrode is extremely rough with mossy/dendritic Na growth (Figure 4f). Furthermore, Na electrode with a partial ML-G coverage on the surface were employed to assemble the symmetric cell, and SEM image of the electrode was taken after cycling at the same condition. A clear comparison could be observed (Figure S7 in Supporting Information), where the bare Na portion showed mossy and dendritic Na, while the portion with ML-G coverage showed a much smoother surface. All these electrochemical and surface characterizations explicitly demonstrate that the employment of ML-G film as an artificial SEI layer could stabilize the Na electrode and effectively suppress Na dendrite growth at relatively high current densities and cycling capacities. To consolidate our conclusion, we further investigated the possible reasons for the dependence of Na anode stability on the thickness of graphene film (the number of layers). For an ideal graphene film without any defects, it would be difficult for Na ions to diffuse through the hexagonal rings of graphene layers.21 Previous experimental and theoretical works both proved that the introduction of defects to graphene could decrease the barrier energy of ion diffusion.22,23,33 For CVD-grown graphene, it is well-known that both point and line defects are widely distributed within the graphene layers,34 while the graphene can still maintain a high strength. In our experiment, the graphene growth was 6814
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
Letter
Nano Letters Na anode at a current density as high as 2 mA/cm2 with a cycling capacity up to 3 mAh/cm2. We believe that our method could be a viable route toward high-energy Na battery systems and can provide valuable insights into the Li batteries as well.
■
(14) Lee, J.; Lee, Y.; Lee, J.; Lee, S. M.; Choi, J. H.; Kim, H.; Kwon, M. S.; Kang, K.; Lee, K. T.; Choi, N. S. ACS Appl. Mater. Interfaces 2017, 9, 3723−3732. (15) Cao, R.; Mishra, K.; Li, X.; Qian, J.; Engelhard, M. H.; Bowden, M. E.; Han, K.; Mueller, K. T.; Henderson, W. A.; Zhang, J. Nano Energy 2016, 30, 825−830. (16) Luo, W.; Lin, C.; Zhao, O.; Noked, M.; Zhang, Y.; Rubloff, G. W.; Hu, L. Adv. Energy Mater. 2017, 7, 1601526. (17) Kim, Y. J.; Lee, H.; Noh, H.; Lee, J.; Kim, S.; Ryou, M. H.; Lee, Y. M.; Kim, H. T. ACS Appl. Mater. Interfaces 2017, 9, 6000−6006. (18) Geim, A. K. Science 2009, 324, 1530−1534. (19) Kim, K. S.; Zhao, Y.; Jang, H.; Lee, S. Y.; Kim, J. M.; Kim, K. S.; Ahn, J. H.; Kim, P.; Choi, J. Y.; Hong, B. H. Nature 2009, 457, 706− 710. (20) Lee, C.; Wei, X. D.; Kysar, J. W.; Hone, J. Science 2008, 321, 385−388. (21) Bunch, J. S.; Verbridge, S. S.; Alden, J. S.; van der Zande, A. M.; Parpia, J. M.; Craighead, H. G.; McEuen, P. L. Nano Lett. 2008, 8, 2458−2462. (22) O’Hern, S. C.; Boutilier, M. S. H.; Idrobo, J. C.; Song, Y.; Kong, J.; Laoui, T.; Atieh, M.; Karnik, R. Nano Lett. 2014, 14, 1234−1241. (23) Yan, K.; Lee, H. W.; Gao, T.; Zheng, G.; Yao, H.; Wang, H.; Lu, Z.; Zhou, Y.; Liang, Z.; Liu, Z.; Chu, S.; Cui, Y. Nano Lett. 2014, 14, 6016−6022. (24) Kim, J. S.; Kim, D. W.; Jung, H. T.; Choi, J. W. Chem. Mater. 2015, 27, 2780−2787. (25) Cao, Z. Y.; Xu, P. Y.; Zhai, H. W.; Du, S. C.; Mandal, J.; Dontigny, M.; Zaghib, K.; Yang, Y. Nano Lett. 2016, 16, 7235−7240. (26) Nair, R. R.; Blake, P.; Grigorenko, A. N.; Novoselov, K. S.; Booth, T. J.; Stauber, T.; Peres, N. M. R.; Geim, A. K. Science 2008, 320, 1308−1308. (27) Ferrari, A. C.; Meyer, J. C.; Scardaci, V.; Casiraghi, C.; Lazzeri, M.; Mauri, F.; Piscanec, S.; Jiang, D.; Novoselov, K. S.; Roth, S.; Geim, A. K. Phys. Rev. Lett. 2006, 97, 187401. (28) Ni, Z.; Wang, Y.; Yu, T.; Shen, Z. Nano Res. 2008, 1, 273−291. (29) Li, X.; Cai, W.; An, J. H.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. S. Science 2009, 324, 1312−1314. (30) Gao, T.; Song, X.; Du, H.; Nie, Y.; Chen, Y.; Ji, Q.; Sun, J.; Yang, Y.; Zhang, Y.; Liu, Z. Nat. Commun. 2015, 6, 6835. (31) Lin, D.; Liu, Y.; Liang, Z.; Lee, H. W.; Sun, J.; Wang, H.; Yan, K.; Xie, J.; Cui, Y. Nat. Nanotechnol. 2016, 11, 626−632. (32) Peled, E.; Golodnitsky, D.; Ardel, G. J. Electrochem. Soc. 1997, 144, L208−L210. (33) Tian, H.; Seh, Z. W.; Yan, K.; Fu, Z.; Tang, P.; Lu, Y.; Zhang, R.; Legu, D.; Cui, Y.; Zhang, Q. Adv. Energy Mater. 2017, 7, 1602528. (34) Lee, G.-H.; Cooper, R. C.; An, S. J.; Lee, S.; van der Zande, A.; Petrone, N.; Hammerberg, A. G.; Lee, C.; Crawford, B.; Oliver, W.; Kysar, J. W.; Hone, J. Science 2013, 340, 1073−1076. (35) Banhart, F.; Kotakoski, J.; Krasheninnikov, A. V. ACS Nano 2011, 5, 26−41. (36) Wofford, J. M.; Nie, S.; McCarty, K. F.; Bartelt, N. C.; Dubon, O. D. Nano Lett. 2010, 10, 4890−4896. (37) Hao, Y.; Wang, L.; Liu, Y.; Chen, H.; Wang, X.; Tan, C.; Nie, S.; Suk, J. W.; Jiang, T.; Liang, T.; Xiao, J.; Ye, W.; Dean, C. R.; Yakobson, B. I.; Jiang, T.; Yakobson, B. I.; McCarty, K. F.; Kim, P.; Hone, J.; Colonmbo, L.; Ruoff, R. S. Nat. Nanotechnol. 2016, 11, 426−431. (38) Perreault, F.; Faria, A. F.; de Elimelech, M. Chem. Soc. Rev. 2015, 44, 5861.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.7b03071. Details about material synthesis, characterization, and electrochemical performance of symmetric cells made from G/Na electrode and bare Na electrode (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Weiyang Li: 0000-0001-5827-415X Author Contributions
H.W. and C.W. contributed equally to the work. H.W. and W.L. conceived the idea. H.W. conducted the graphene growth and characterization. H.W. and C.W. carried out the electrochemical tests. H.W. and W.L. cowrote the paper. All the authors discussed the results and commented on the manuscript. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
■
REFERENCES
The authors greatly acknowledge the support from Young Investigator Program funded by U.S. Air Force Office of Scientific Research under award FA9550-17-1-0184 and support from the start-up funds at Thayer School of Engineering, Dartmouth College. The authors also would like to thank Professor Jifeng Liu and Mr. Sidan Fu for their help on the UV−vis measurement.
(1) Larcher, D.; Tarascon, J. M. Nat. Chem. 2014, 7, 19−29. (2) Yang, Z.; Zhang, J. L.; Kintner-Meyer, M. C. W.; Lu, X. C.; Choi, D. W.; Lemmon, J. P.; Liu, J. Chem. Rev. 2011, 111, 3577−3613. (3) Goodenough, J. B.; Park, K. S. J. Am. Chem. Soc. 2013, 135, 1167−1176. (4) Xu, K. Chem. Rev. 2014, 114, 11503−11618. (5) Zhao, J.; Zhou, G.; Yan, K.; Xie, J.; Li, Y.; Liao, L.; Jin, Y.; Liu, K.; Hsu, P.-C.; Wang, J.; Cheng, H.-M.; Cui, Y. Nat. Nanotechnol. 2017, 12, 993. (6) Kim, H.; Kim, H.; Ding, Z.; Lee, M. H.; Lim, K.; Yoon, G.; Kang, K. Adv. Energy Mater. 2016, 6, 1600943. (7) Liang, Y.; Yoo, H. D.; Li, Y.; Shuai, J.; Calderon, H. A.; Hernandez, F. C. R.; Grabow, L. C.; Yao, Y. Nano Lett. 2015, 15, 2194−2202. (8) Ren, X.; Wu, Y. J. Am. Chem. Soc. 2013, 135, 2923−2926. (9) Luo, W.; Hu, L. ACS Cent. Sci. 2015, 1, 420−422. (10) Manthiram, A.; Yu, X. W. Small 2015, 11, 2108−2114. (11) Hartmann, P.; Bender, C. L.; Vracar, M.; Durr, A. K.; Garsuch, A.; Janek, J.; Adelhelm, P. Nat. Mater. 2012, 12, 228−232. (12) Yadegari, H.; Sun, Q.; Sun, X. Adv. Mater. 2016, 28, 7065−7093. (13) Seh, Z. W.; Sun, J.; Sun, Y.; Cui, Y. ACS Cent. Sci. 2015, 1, 449− 455. 6815
DOI: 10.1021/acs.nanolett.7b03071 Nano Lett. 2017, 17, 6808−6815
