Use of Tween Polymer To Enhance the Compatibility of the Li

Jun 1, 2018 - Open Access ... (5,6) It is well-known that the solid electrolyte interphase (SEI) layer on the lithium ... Because of its high theoreti...
0 downloads 0 Views 6MB Size
Letter Cite This: Nano Lett. XXXX, XXX, XXX−XXX

pubs.acs.org/NanoLett

Use of Tween Polymer To Enhance the Compatibility of the Li/ Electrolyte Interface for the High-Performance and High-Safety Quasi-Solid-State Lithium−Sulfur Battery Jie Liu, Tao Qian, Mengfan Wang, Jinqiu Zhou, Na Xu, and Chenglin Yan* Soochow Institute for Energy and Materials InnovationS, College of Physics, Optoelectronics and Energy & Collaborative Innovation Center of Suzhou Nano Science and Technology, and Key Laboratory of Advanced Carbon Materials and Wearable Energy Technologies of Jiangsu Province, Soochow University, Suzhou 215006, China S Supporting Information *

ABSTRACT: Lithium metal batteries have attracted increasing attention recently due to their particular advantages in energy density. However, as for their practical application, the development of solid-state lithium metal batteries is restricted because of the poor Li/electrolyte interface, low Li-ion conductivity, and irregular growth of Li dendrites. To address the above issues, we herein report a high Li-ion conductivity and compatible polymeric interfacial layer by grafting tween-20 on active lithium metal. Sequential oxyethylene groups in tween-grafted Li (TG-Li) improve the ion conductivity and the compatibility of the Li/electrolyte interface, which enables low overpotentials and stable performance over 1000 cycles. Consequently, the poly(ethylene oxide)-based solid-state lithium−sulfur battery with TG-Li exhibits a high reversible capacity of 1051.2 mA h g−1 at 0.2 C (1 C = 1675 mA h g−1) and excellent stability for 500 cycles at 2 C. The decreasing concentration of the sulfur atom with increasing Ar+ sputtering depth indicates that the polymer interfacial layer works well in suppressing polysulfide reduction to Li2S/Li2S2 on the metallic Li surface even after long-term cycling. KEYWORDS: Quasi-solid-state Li−S battery, Li/electrolyte interface, in situ XRD, polysulfide diffusion

L

immobilizing the organic solvents such as ethylene carbonate (EC), diethyl carbonate (DEC), DME, and 1,3-dioxolane (DOL), etc., in a polymer network. The polymer network ensures the mechanical stability, and the organic electrolyte makes contributions to the high ionic conductivity.14 As a stable polymer with sequential oxyethylene groups (−CH2CH2O−), PEO exhibits good compatibility with electrodes and is considered suitable as the polymer matrix of GPEs.16 The oxyethylene groups can coordinate with lithium ions and help to dissociate and dissolve Li salts, which further enhance the ionic conductivity of the solid-state battery.14,17 However, although PEO-based GPEs are extensively researched in the solid-state battery, their progress has decelerated because of some remaining tough problems. The embedded organic electrolyte improves the conductivity of GPEs but also threatens the battery security because of the inevitable decomposition of organic electrolyte after long-term cycling.18,19 The continuous generation and breaking of the SEI on bare Li lead to a tough interphase.7,20 Moreover, the issue of Li dendrite still remains, which resulted in a safety hazard and

ithium metal batteries (LMBs) are considered as one of the optimal energy storage systems with high energy density because the Li metal anode has an ultrahigh specific capacity of 3860 mA h g−1, a very low redox potential (−3.040 V versus standard hydrogen electrode), and a small gravimetric density of 0.534 g cm−3.1−4 However, the poor cyclability and potential safety hazard have always seriously constrained the development of LMBs.5,6 It is well-known that the solid electrolyte interphase (SEI) layer on the lithium surface, resulting from the spontaneous reaction between reactive lithium metal and most organic electrolytes, always continuously generates and breaks during repeated cycling.7−9 As a result, both the lithium metal and electrolyte suffer from a constant loss, which leads to low Coulombic efficiency and rapid capacity decay. Meanwhile, there are always potential safety issues resulting from the instability and flammability of organic electrolytes, as well as the terrible growth of lithium dendrites.7,10,11 As a potentially practical application of LMBs, the solid-state LMBs partly solve the above obstacles, especially the security issue, and thus, the development of solid-state LMBs has drawn increasing attention in recent years. Because of the high ionic conductivity, good chemical/electrochemical stability, high mechanical strength, and close adhesion to electrodes, gel polymer electrolytes (GPEs) are extensively investigated in solid-state LMBs.12−15 GPEs can be obtained by © XXXX American Chemical Society

Received: May 9, 2018 Revised: May 29, 2018 Published: June 1, 2018 A

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. (a) Schematic diagram of tween-grafted lithium metal. The polymer interfacial layer can provide high ion conductivity in the PEO-based Li−S battery, efficiently keep lithium polysulfides away from active metallic lithium, and isolate their frequent reduction. Moreover, this polymer layer can improve the compatibility between the Li electrode and electrolyte because of the polymer structure of tween. (b) DFT calculations on the affinity between the alkyl chain and Li2S8, as well as Li2S6, to evaluate the resistance of alkyl chains to polysulfides.

delivers a high specific capacity of 552.1 mA h g−1 when the current rate is increased to 4 C, which is attributed to the high ionic conductivity of the interfacial layer. The prominent cycle performance can be demonstrated by the remarkable capacity retention of 98.4% at 1 C over 150 cycles and the long-term cycling for 500 cycles with a low decrement rate of 0.058% per cycle based on the initial discharge capacity. X-ray photoelectron spectroscopy (XPS) of TG-Li at different depths of the interfacial layer made evident that the polymer layer worked well in obstructing polysulfide diffusion in the resultant PEObased Li−S battery. Undoubtedly, this new strategy achieves high Li-ion conductivity and remarkable compatibility between solid electrolyte and anode, as well as a long-term and stable interface via inhibiting the shuttle effect, which provides a promising prospect in the advanced solid-state Li−S battery system. For the fabrication of a uniform polymer interfacial layer on the lithium surface, the lithium wafer was immersed into tween/anhydrous THF solution. Then, the lithium plate was taken out and rinsed by THF to remove the residual tween, followed by drying at room temperature. To exclude the impact of solvent, we put a bare Li wafer into anhydrous tetrahydrofuran (THF) solution (denoted to THF-Li), and almost no morphology change can be seen, as shown in Figure S1a. In contrast, after the immersion of bare Li in tween-20 for the same time, the Li surface presents an enormous change (Figure S1b), which indicates that, in the competing reaction system of tween-20/THF solution, the inactive reaction between THF and Li metal can be ignored. All of the above processes were carried out in a glovebox (Ar atmosphere, O2 < 0.1 ppm and H2O < 0.1 ppm). The cross-section scanning electron microscopy (SEM) images of TG-Li (Figure S2a,b) and the elemental mapping (Figure S3), showing the homogeneous distribution of carbon that derives from tween molecular, indicate that the polymer interfacial layer forms on the metal surface with an average thickness of ∼5 μm, which is significant to keep the polysulfides away from metallic Li. XPS results in Figure S4a demonstrate that the tween molecule

limited long-lifespan cycle ability for these batteries matching with metallic lithium. These negative effects are magnified in varying degrees when GPEs are used in the lithium−sulfur (Li− S) battery. Because of its high theoretical specific capacity (1675 mA h g−1) and energy density (2600 W h kg−1) relative to conventional lithium-ion batteries, the Li−S battery attracts a lot of attention nowadays.21−25 The Li−S battery based on the PEO system is expected to perform with the advantage of GPEs and improve the safety of the Li−S battery. However, the unexpected deposition of electron-/ion-insulated Li2S on the Li anode resulting from the dissolution and diffusion of lithium polysulfides decreases the ion conductivity of the resultant Li− S batteries and deteriorates the electrolyte−anode interface.26−28 The worsening compatibility between electrolyte and Li anodes renders low reversible capacities and poor cyclability to the batteries. Herein, we fabricated a compatible polymer interfacial layer on the surface of lithium metal to achieve high Li-ion conductivity and the resistance to polysulfides. This “smart” layer makes a good interfacial connection and helps to improve the compatibility between the Li anode and electrolyte. Tween20, an organic polymer with sequential oxyethylene groups and alkyl groups, is employed to complete this novel pattern. The −CH2CH2O− structure is expected to facilitate the continuous ion transport, and the long alkyl chain in tween molecular can effectively dampen the polysulfide diffusion, which was revealed by density functional theory (DFT) calculations. As a result, this polymer layer can efficiently keep polysulfides from frequent reduction by active lithium so that a reliable and compatible interface between the anode and electrolyte is obtained as exhibited in Figure 1a. The symmetric cell test shows that the as-prepared tween-grafted Li (TG-Li) has low overpotential and can be stable for 1000 cycles. In situ XRD and pressure measurements demonstrate that TG-Li can efficiently prevent the decomposition of electrolyte and thus improve the safety of batteries. When using TG-Li as the anode, the PEObased solid-state Li−S battery shows a high reversible capacity of 1051.2 mA h g−1 at 0.2 C. Importantly, the battery still B

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. (a) Schematic exhibits the mechanisms of the Li stripping/plating behaviors of bare Li and TG-Li. (b) Voltage profiles of bare Li foil symmetric cells and TG-Li symmetric cells at a current density of 0.5 mA cm−2, and all the electrochemical characterizations were carried out at 25 °C. The cross-section SEM images of (c) bare Li and (e) TG-Li before stripping/plating and (d, f) their changes in morphologies after tripping/ plating 50 times. Scale bar, 20 μm. All the electrochemical characterizations were carried out at 25 °C.

reacted with lithium metal by giving the compositional analysis of TG-Li. The binding energies are calibrated with respect to the C 1s peak at 284.8 eV. The results confirm the chemical composition of the polymer layer with the presence of C and O. In addition to the strong hydrocarbon peak (284.6 eV), the high-resolution C 1s spectra show two main C peaks at 289.6 and 286.1 eV, corresponding to O(CO) and CO, respectively, as shown in Figure S4b.29,30 DFT calculations were used to evaluate the resistance of alkyl chains to polysulfides, and the results are shown in Figure 1b. The affinity between the carbon chain and Li2S8 is as low as −0.12 eV, which is much lower than that of the common electrolyte solvents 1,3-dioxolane (DOL) and dimethoxyethane (DME) to Li2S8 (−0.75 and −0.60 eV, respectively). The binding energies between alkyl chains and Li2S6 (−0.13 eV) were also computed to manifest the weak interaction between alkyl chains and Li2S6. Figure 2a simulates the different procedures when polysulfides shuttle and get close to the bare Li and TG-Li surface. It is well-known that a SEI layer forms when lithium reacts with most organic electrolyte solvents.7,31 However, the SEI layer is easily broken when Li unevenly plates/strips. What is worse, the shuttled polysulfides will be irreversibly reduced to lithium sulfide on the metallic Li surface. Both of these facts eventually result in the formation of dendrites and “dead Li”, which results in low Coulombic

efficiency and rapid capacity degeneration, as well as potential safety hazards.32 After being grafted with tween molecular, the lithium surface is full of alkyl chains, which sufficiently suppress the contact and reaction between polysulfides and lithium metal. To reveal the suppressed redox, we directly put TG-Li and bare Li into 2 mM Li2S8 solution and then characterize their surface by XPS. Li2Sx (1 ≤ x ≤ 4) signals with high intensities are observed on the bare Li surface, which is attributed to the active reaction between metallic Li and Li2S8 (Figure S5a). In comparison, the TG-Li surface shows a weak signal of S 2p spectra, attributed to the residual Li2S8 that blocked in the interfacial layer (Figure S5b). The barely visible Li2Sx (1 ≤ x ≤ 4) signals indicate that the polymer layer can efficiently keep polysulfides from frequent reduction by active Li metal.33 We further conducted the UV/vis spectroscopy investigation to observe the reflection variation of Li2S8 solution before and after immersing bare Li or TG-Li (Figure S6). The UV/vis spectroscopy of TG-Li is almost overlapped, which means that Li2S8 is not involved in any redox reaction. In comparison, the reflection curves of bare Li show a noticeable reflection change and transformation to Li2Sx (1 ≤ x ≤ 6), indicating that Li2S8 is easily reduced by active bare Li.25,34 Sequential oxyethylene groups on TG-Li impel the rapid delivery of the Li ion, which can facilitate even Li plating/ stripping. For an estimation of the electrochemical performance C

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. (a) Schematic of in situ XRD investigation. Contour maps of in situ XRD patterns collected during the initial cycle of the (b) bare Li//S and (c) TG-Li//S battery at a rate of 0.3 C. Li2S phase evolution in the (d) bare Li//S and (e) TG-Li//S battery, demonstrating that Li2S cannot deposit on the metallic Li surface of TG-Li. (f) The pressure measurement exhibits a much smaller pressure change in the TG-Li//S battery, which indicates that TG-Li can efficiently improve the safety of the electrolyte.

and stripping processes, which is attributed to the durability of the polymer interfacial layer and the prominent Li-ion conductivity that arose from sequential oxyethylene groups. The overpotential at a current density of 1.0 mA cm−2 is shown in Figure S9, which also exhibits a low overpotential and better stability of TG-Li. After discharging/charging 200 h, the symmetrical cells are disassembled. It can be found that the surface of bare Li became more rugged, and the bulk dead Li was deposited on it (Figure 2c,d), which seriously decrease the ion conductivity and deteriorate the battery stability. However, TG-Li has little change in morphology (Figure 2e,f), indicating the splendid Li-ion transport capacity of TG-Li. In situ XRD was performed to directly monitor the variation of the lithium anode in the PEO-based Li−S battery during the charge−discharge process (Figure 3a). Figure 3b,c shows the contour maps of in situ XRD patterns collected during the initial cycle of Li−S cells using TG-Li (noted as TG-Li//S) and bare Li foil (noted as bare Li//S) as the anodes, respectively, at a rate of 0.3 C, where blue color represents high intensity. It is obviously seen that the XRD peaks at 21.2° and 23.4° appear in the bare Li//S battery, which are identified as the Li2CO3 phase decomposing from the ether-based electrolyte,35,36 but cannot be detected in the TG-Li//S battery. This result convinces us that TG-Li can efficiently prevent the decomposition of electrolyte, which can be made further evident by subsequent

of TG-Li, quasi-solid-state symmetrical cells were assembled with two identical TG-Li or bare Li foil. Figure S7 plots the dependence of the conductivity on 1/T for the PEO-based electrolyte in the temperature range 25−60 °C, showing the conductivity of 5.01 × 10−5 S cm−1 at room temperature, and the transference number of the PEO-based electrolyte was ∼0.32. Figure 2b exhibits the voltage hysteresis of two symmetrical cells at a current density of 0.5 mA cm−2. For the symmetrical bare Li cell, the crude oxide layer and spontaneously formed SEI hinder the transport of the Li ion, which results in a difficult Li plating/stripping and contributes to the high overpotential. The continuous generation of a new SEI layer leads to the rough change of overpotential. A significant voltage increase was found in the bare Li symmetrical cell after 200 h of plating/stripping, which is ascribed to the inherent instability of spontaneously formed SEI during discharging and side reactions. The same situation was observed in the THF-Li symmetrical cell, which exhibits an increasing overpotential after several plating/stripping cycles, indicating that THF-Li cannot maintain the stable and fast Li+ transfer (Figure S8). The TG-Li symmetrical cell exhibits stable voltage profiles at a low overpotential of ∼17 mV, and the flat voltage plateaus can be maintained for a prolonged lifetime over 1000 h without an obvious “bump” during cycling. This result indicates a lower energy barrier for both its nucleation D

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 4. (a) Rate performances of TG-Li//S and bare Li//S cells. (b) Rate capability of the TG-Li anode tested in PEO-based electrolyte. (c) Representative voltammograms of the TG-Li obtained at different scan rates and (d) linear relationship of Ip and ν1/2. (e) Nyquist plots of the impedance spectra of the TG-Li//S and bare Li//S cells before/after 50 cycles at a current rate of 1 C. (f) Cycling performance comparison between the TG-Li//S cell and bare Li//S cell at current rate of 1 C for 150 cycles. (g) Cycling performance of the TG-Li//S cell with high sulfur loading of 2.8 and 4.9 mg cm−2 at a current density of 500 μA cm−2. (h) Specific capacity and Coulombic efficiency of the TG-Li//S cell at the 2 C rate. All the electrochemical characterizations were carried out at 25 °C.

suggests that the long-term and stable interface of the TG-Li anode can evidently prevent the decomposition reactions of the electrolyte, which is quite significant for battery security. Because of the shuttle effect, polysulfides will diffuse from cathode to anode and eventually reduce to Li2S when they come into contact with lithium metal. For the bare Li//S battery, the Li2S phase evolution can be obviously seen at ∼13.6° as shown in Figure 3d, and the ascending intensity indicates that Li2S deposits on the lithium surface ceaselessly.38 However, no peak related to Li2S is observed during TG-Li//S battery cycles as shown in Figure 3e, revealing that the Li2S cannot be deposited on the lithium anode, which can be further confirmed by the comparison of Li2SOx (3 ≤ x ≤ 7) phase at 17.4° and 25.6° as shown in the XRD pattern.10,36 It is clearly seen that the Li2SOx (3 ≤ x ≤ 7) phase in the bare Li//S battery (Figure 3b), resulting from the oxidation of Li2S and Li2S2 by LiNO3, cannot be observed in the TG-Li battery (Figure 3c) during the discharge−charge procedure, which is due to the fact that the polymer interfacial layer of TG-Li can effectively suppress the reduction reaction between polysulfides and lithium metal. Using the stable TG-Li as the anode, we further demonstrate their prominent electrochemical performances in PEO-based solid-state Li−S batteries. Figure 4a clearly exhibits the rate

pressure measurements. In addition, after immersing the TG-Li and bare Li in the electrolyte for 24 h, we directly characterized their surface. The XRD profiles of TG-Li show similar peaks before and after immersing, and no new signal appeared (Figure S10). However, after dipping bare Li into the electrolyte for same time, we can observe the signals of Li2CO3 and Li2SOx (3 ≤ x ≤ 7) in corresponding XRD pattern,10,35,36 which were ascribed to the decomposition products of electrolytes. It is well-known that some byproducts, such as CH4, H2, and CO2, etc., would form upon battery cycling, especially during charge, because of the decomposition of electrolyte when it comes into contact with the fresh lithium.37 For the elimination of interference, diglyme using lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) as the electrolyte salt was chosen to replace the more commonly used DOL and DME solvent because of its lower volatility. Figure 3f shows the pressure trend when the battery using bare Li and TG-Li as the anode, respectively, is cycled at 0.1 C. For the battery with bare Li as the anode, the pressure always increases, and the tendency goes up faster after cycling 18 h. The pressure increases by 0.8 psi after 50 h. In contrast, the pressure−time curve of the TGLi//S battery shows a much smaller slope, demonstrating a much lower pressure increase during 40 h of cycling. This result E

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 5. (a) Schematic of XPS analysis at the different depths achieved by etching the TG-Li anode with Ar+-ion sputtering. (b) XPS depth profiles of Li 1s, where the curves go from green to red corresponding to increasing detection depths (0−3000 nm). The Li−S peak gradually shifts to Li−O as the detection depth increases, indicating that the polysulfides are on the decline. (c) XPS spectra of S 2p at different etching depths. The highresolution S 2p spectra of TG-Li at the etching depths of (d) 0, (e) 1500, and (f) 3000 nm, demonstrating the efficient rejection of the polymer layer to polysulfides. (g) Summary of the atomic concentration of Li, S, and C on the TG-Li anode with the etching depth increasing. (h) Zoomed image of the atomic sulfur concentration in part g; the decreasing trend indicates that the interfacial layer plays a significant role in restricting polysulfides.

7.16 × 10−8 cm2 s−1 and DLi+‑A = 1.40 × 10−7 cm2 s−1. In contrast, the bare Li foil exhibits a lower diffusion coefficient (DLi+‑C = 3.36 × 10−8 cm2 s−1 and DLi+‑A = 7.53 × 10−8 cm2 s−1) as shown in Figure S12a,b. This result suggests that the polymer layer of TG-Li can facilitate Li+ transport, and its conductivity is even better than the spontaneous SEI, which is of quite significance to determine the poor interface issue between GPE and bare Li metal. The electrochemical impedance spectroscopy (EIS) measurements of PEO-based TG-Li//S and Li//S cells before and after 50 cycles were carried out as shown in Figure 4e. Previous studies showed that the semicircle in the low-frequency range could reflect the charge transfer process, while the semicircle in the high-frequency region could be assigned to the surface film resistance, which is associated with the SEI and insulating layer of solid Li2S/Li2S2.39,41 It is clearly seen from Figure 4e that, before cycling, the TG-Li//S cell has a considerably lower charge transfer resistance (∼20.4 Ω) compared to the bare Li// S cell (∼40.3 Ω), which indicates the poor Li/electrolyte interface in the bare Li//S cell and the fast lithium-ion diffusion in the TG-Li//S cell.39 After 50 cycles, the TG-Li//S cell shows a much lower interfacial resistance value in the high-frequency region (6.8 Ω) than the bare Li//S cell (14.2 Ω). This is because the compatible interfacial layer of TG-Li can suppress the deposition of Li2S/Li2S2 on the metallic lithium surface. This result is in good agreement with the analysis of the lithium-ion diffusion coefficients. Furthermore, we characterize the impedance of both TG-Li//S and bare Li//S cells in liquid electrolyte before cycling as shown in Figure S13a,b. Both the bare Li//S and TG-Li//S cell in the PEO-based electrolyte have larger interfacial resistance than those in liquid electrolyte. However, the interfacial resistance of the TG-Li//S cell is smaller than that in the bare Li//S cell with the PEO-based electrolyte, which means that the TG-Li//S cell achieves a

performance differences between solid-state TG-Li//S and bare Li//S cells. For the TG-Li//S cell, the specific capacity of 1051.2 mA h g−1 can be obtained at 0.2 C, and the capacity still remains over 578.6 mA h g−1 even when the current rate is up to 4 C. In contrast, a much lower specific capacity retention is observed for the bare Li//S cell as the capacity dropped significantly from ∼982.6 mA h g−1 at 0.2 C to less than 285.1 mA h g−1 at 4 C, which is attributed to the high resistance and the difficult Li+ transfer especially at high current densities. Figure 4b and Figure S11 show the rate capability differences of the TG-Li anode tested in liquid ether-based electrolyte and PEO-based electrolyte, respectively. The brilliant rate capacity indicates the excellent compatibility between TG-Li and gel electrolyte. The greatly improved battery reversible capacity of the PEObased cell with TG-Li as the anode is found to be closely associated with Li-ion transport in the electrode. Cyclic voltammetry (CV) analysis is implemented to estimate the Li+-ion diffusion coefficients (DLi’s) of the cathodes according to the Randles−Sevcik equation as follows: Ip = (2.69 × 105)n3/2ADLi1/2 C Liν1/2

(1)

where Ip is the peak current, n is the number of reaction electrons (n = 2 for the Li−S battery), A is the area of the electrode (A = 1.13 cm2), DLi is the lithium-ion diffusion coefficient, CLi is the concentration of Li+ (CLi = 1.16 × 10−3 mol cm−3), and ν is the scanning rate.39,40 This equation reveals that the electrochemical process of a cell is affected by Li+ diffusion. Figure 4c shows the CV curves of TG-Li at different scan rates from 0.1 to 0.5 mV s−1. From the linear relationship of Ip and ν1/2 (Ip‑A = 0.3672ν1/2 − 3.01 × 10−4, Ip‑C = −0.2668ν1/2 − 2.93 × 10−4) as shown in Figure 4d, the lithiumion diffusion coefficients (DLi+‑C, cathodic peak at ∼2.05 V; and DLi+‑A, anodic peak at ∼2.30 V) are obtained, where DLi+‑C = F

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

5g,h, indicating that the polymer interfacial layer plays an important role in restricting polysulfides. In conclusion, we constructed a stable polymer interfacial layer with high ion conductivity on active lithium metal, which can achieve excellent compatibility between the Li anode and solid electrolyte. The alkyl chains in the polymer layer refuse polysulfides from going near the active metallic lithium and thus prevent the reaction between polysulfides and metallic lithium in the Li−S battery as confirmed by DFT calculations. This interfacial layer efficiently suppresses the growth of Li dendrite and induces the TG-Li anode to a low overpotential. It is noted that TG-Li can improve the battery security by inhibiting the decomposition of electrolyte as demonstrated by in situ XRD and pressure measurements. The high conductivity of the polymer layer and compatible Li/electrolyte interface guarantee the solid-state Li−S battery a high reversible capacity of 1051.2 mA h g−1 at 0.2 C. The prominent cycle performance can be demonstrated by the remarkable capacity retention of 98.4% at 1 C over 150 cycles and the long-term cycling for 500 cycles with a low decrement rate of 0.058% per cycle. After discharging−charging 100 times, the TG-Li was used to investigate the S 2p spectra at different depths of the interfacial layer by Ar+-ion sputtering. The decreasing sulfur atom concentration with increasing Ar+ sputtering depth indicates well that the polymer layer is sufficient in resisting the diffusion of polysulfides to the surface of the metallic lithium.

compatible and high-ion-conductivity interface in the PEObased electrolyte. Figure 4f compares the cycling performances between the TG-Li//S cell and bare Li//S cell at the current rate of 1 C for 150 cycles. The TG-Li//S cell delivers a high initial discharge capacity of 814.7 mA h g−1. After 150 cycles of charge− discharge, a reversible capacity of 801.3 mA h g−1 can be obtained (capacity retention of 98.4%). In contrast, the bare Li//S cell shows a relatively faster decline of capacity, decreasing from 731.0 to 385.2 mA h g−1 (corresponding to a capacity retention of 52.6%). The fast capacity decay can also be observed in the THF-Li//S battery, and the inferior performance demonstrates that the THF-Li made no impact on the improvement of battery cycling (Figure S14). The TG-Li// S cell exhibits a high initial capacity of 1002.1 mA h g−1 when the sulfur loading is 2.8 mg cm−2, and it still retains an excellent stability even with the loading as high as 4.9 mg cm−2 as shown in Figure 4g. The long-term performance of TG-Li//S demonstrates the excellent stability and Coulombic efficiency, as well as high specific capacity, for 500 cycles at 2 C with a low decrement rate of 0.058% per cycle (Figure 4h). Figure S15 shows the long-term cycling of the TG-Li//S batteries without LiNO3 additive in the electrolyte, which demonstrates the similar stability with that using LiNO3 additive. A relatively higher Coulombic efficiency in the cell with LiNO3 additive is because LiNO3 can catalyze the conversion of highly soluble polysulfides to slightly soluble elemental sulfur on the cathode.42,43 Additionally, TG-Li is proven to be also outstanding in improving the electrochemical performance in the commercial LiFePO4 (LFP) battery. As shown in Figure S16, the TG-Li//LFP cell maintains the capacity over 132.5 mA h g−1 after 200 cycles at the rate of 1 C (1 C = 178 mA h g−1) and keeps the Coulombic efficiency at around 99.4% all the time, while the cell using bare Li as the anode presents an obviously decreased capacity and efficiency. For insights into the restriction of TG-Li to polysulfides in long-term, an XPS investigation of TG-Li was applied to investigate the ratio variation of different elements at the different depth of polymer interfacial layer after the anode is discharged−charged 100 times, which can be achieved by etching the polymer layer with Ar+-ion sputtering (Figure 5a). Figure 5b shows the XPS depth profiles of Li 1s, where the curves go from green to red corresponding to increasing detection depths. The Li−S peak at ∼55.6 eV is detected at the early stage of sputtering.44 As the sputtering goes on, the detection region goes deep into the polymer layer. As a result, the Li−S peak gradually shifts to Li−O at ∼54.4 eV, which indicates that the polysulfides are on the decline as the detection depth increases.45 The signal of Li0 at ∼52.7 eV arises when the detection region is near the surface of the metallic Li foil underneath the polymer layer. For a further illustration of the resistance of the polymer layer to polysulfide diffusion, the high-resolution S 2p spectra of TG-Li at different etching depths are exhibited as shown in Figure 5c. The strong peaks at 160.2, 161.5, and 162.7 eV, corresponding to Li2S, Li2S2, and Li2Sx (4 ≤ x ≤ 8), respectively, are observed when the etch depth is 0 nm, which is attributed to the diffusion of highly soluble polysulfides (Figure 5d). The peak intensities decreased significantly when the etching depth is up to 1500 and 3000 nm (Figure 5e,f), demonstrating the very low level of lithium sulfides near metallic lithium. The decreasing concentration of atomic sulfur as the etching depth increases is shown in Figure



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.8b01882. Additional experimental and computational details and figures including SEM results, XPS results, and UV/vis spectroscopy of Li2S8 solution (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Jie Liu: 0000-0002-6970-7691 Tao Qian: 0000-0001-7252-8224 Mengfan Wang: 0000-0003-3370-6395 Chenglin Yan: 0000-0003-4467-9441 Author Contributions

J.L. and T.Q. contributed equally to this work. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We acknowledge the support from the National Natural Science Foundation of China (51622208 and 21703149) and Natural Science Foundation of Jiangsu Province (BK20150338).



REFERENCES

(1) Xu, W.; Wang, J. L.; Ding, F.; Chen, X. L.; Nasybutin, E.; Zhang, Y. H.; Zhang, J. G. Energy Environ. Sci. 2014, 7, 513−537. (2) Kim, H.; Jeong, G.; Kim, Y. U.; Kim, J. H.; Park, C. M.; Sohn, H. J. Chem. Soc. Rev. 2013, 42, 9011−9034. (3) Lin, D.; Liu, Y.; Cui, Y. Nat. Nanotechnol. 2017, 12, 194−206. G

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (4) Bruce, P. G.; Freunberger, S. A.; Hardwick, L. J.; Tarascon, J. M. Nat. Mater. 2012, 11, 19−29. (5) Cheng, X. − B.; Zhang, R.; Zhao, C. − Z.; Zhang, Q. Chem. Rev. 2017, 117, 10403−10473. (6) Zhang, W.; Zhuang, H. L.; Fan, L.; Gao, L.; Lu, Y. Sci. Adv. 2018, 4, 4410−4417. (7) Chi, S. − S.; Liu, Y.; Song, W. − L.; Fan, L. − Z.; Zhang, Q. Adv. Funct. Mater. 2017, 27, 1700348. (8) Zhang, Y. H.; Qian, J. F.; Xu, W.; Russell, S. M.; Chen, X. L.; Nasybulin, E.; Bhattacharya, P.; Engelhard, M. H.; Mei, D. H.; Cao, R. G.; Ding, F.; Cresce, A. V.; Xu, K.; Zhang, J. G. Nano Lett. 2014, 14, 6889−6896. (9) Park, C. M.; Kim, J. H.; Kim, H.; Sohn, H. J. Chem. Soc. Rev. 2010, 39, 3115−3141. (10) Aurbach, D.; Pollak, E.; Elazari, R.; Salitra, G.; Kelley, C. S.; Affinito, J. J. Electrochem. Soc. 2009, 156, A694−A702. (11) Fan, X.; Chen, L.; Ji, X.; Deng, T.; Hou, S.; Chen, J.; Zheng, J.; Wang, F.; Jiang, J.; Xu, K.; Wang, C. Chem. 2018, 4, 174−185. (12) Li, C.; Guo, Z.; Yang, B.; Liu, Y.; Wang, Y.; Xia, Y. Angew. Chem., Int. Ed. 2017, 56, 9126−9130. (13) Liu, M.; Zhou, D.; Jiang, H. R.; Ren, Y. X.; Kang, F. Y.; Zhao, T. S. Nano Energy 2016, 28, 97−105. (14) Zhang, S.; Ueno, K.; Dokko, K.; Watanabe, M. Adv. Energy Mater. 2015, 5, 1500117. (15) Hu, P.; Chai, J.; Duan, Y.; Liu, Z.; Cui, G.; Chen, L. J. Mater. Chem. A 2016, 4, 10070−10083. (16) Tao, X.; Liu, Y.; Liu, W.; Zhou, G.; Zhao, J.; Lin, D.; Zu, C.; Sheng, O.; Zhang, W.; Lee, H. − W.; Cui, Y. Nano Lett. 2017, 17, 2967−2972. (17) MacGlashan, G. S.; Andreev, Y. G.; Bruce, P. G. Nature 1999, 398, 792−794. (18) Chai, J.; Zhang, J.; Hu, P.; Ma, J.; Du, H.; Yue, L.; Zhao, J.; Wen, H.; Liu, Z.; Cui, G.; Chen, L. J. Mater. Chem. A 2016, 4, 5191−5197. (19) Zhou, W.; Wang, S.; Li, Y.; Xin, S.; Manthiram, A.; Goodenough, J. B. J. Am. Chem. Soc. 2016, 138, 9385−9388. (20) Li, N. − W.; Yin, Y. − X.; Yang, C. − P.; Guo, Y. − G. Adv. Mater. 2016, 28, 1853. (21) Ji, X.; Lee, K. T.; Nazar, L. F. Nat. Mater. 2009, 8, 500−506. (22) Xin, S.; Gu, L.; Zhao, N. − H.; Yin, Y. − X.; Zhou, L. − J.; Guo, Y. − G.; Wan, L. − J. J. Am. Chem. Soc. 2012, 134, 18510−18513. (23) Wang, J.; He, Y. − S.; Yang, J. Adv. Mater. 2015, 27, 569−575. (24) Chen, L.; Fan, L. − Z. Energy Storage Mater. 2018, 15, 37−45. (25) Liu, J.; Qian, T.; Wang, M.; Liu, X.; Xu, N.; You, Y.; Yan, C. Nano Lett. 2017, 17, 5064−5070. (26) Wei, S.; Ma, L.; Hendrickson, K. E.; Tu, Z.; Archer, L. A. J. Am. Chem. Soc. 2015, 137, 12143−12152. (27) Xu, R.; Lu, J.; Amine, K. Adv. Energy Mater. 2015, 5, 1500408. (28) Sun, Z.; Zhang, J.; Yin, L.; Hu, G.; Fang, R.; Cheng, H. − M.; Li, F. Nat. Commun. 2017, 8, 14627−14634. (29) Zheng, J.; Engelhard, M. H.; Mei, D.; Jiao, S.; Polzin, B. J.; Zhang, J. − G.; Xu, W. Nat. Energy 2017, 2, 17012−17019. (30) 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, 993−999. (31) Cheng, X.-B.; Zhang, R.; Zhao, C. − Z.; Wei, F.; Zhang, J. − G.; Zhang, Q. Adv. Sci. 2016, 3, 1500213. (32) Xu, K. Chem. Rev. 2014, 114, 11503−11618. (33) Yang, C. − P.; Yin, Y. − X.; Guo, Y. − G.; Wan, L. − J. J. Am. Chem. Soc. 2015, 137, 2215−2218. (34) Patel, M. U. M.; Dominko, R. ChemSusChem 2014, 7, 2167− 2175. (35) Song, S.; Xu, W.; Zheng, J.; Luo, L.; Engelhard, M. H.; Bowden, M. E.; Liu, B.; Wang, C. − M.; Zhang, J. − G. Nano Lett. 2017, 17, 1417−1424. (36) Nagao, K.; Nose, M.; Kato, A.; Sakuda, A.; Hayashi, A.; Tatsumisago, M. Solid State Ionics 2017, 308, 68−76. (37) Jozwiuk, A.; Berkes, B. B.; Weiß, T.; Sommer, H.; Janek, J.; Brezesinski, T. Energy Environ. Sci. 2016, 9, 2603−2608.

(38) Lin, D.; Liu, Y.; Chen, W.; Zhou, G.; Liu, K.; Dunn, B.; Cui, Y. Nano Lett. 2017, 17, 3731−3737. (39) Ghazi, Z. A.; He, X.; Khattak, A. M.; Khan, N. A.; Liang, B.; Iqbal, A.; Wang, J.; Sin, H.; Li, L.; Tang, Z. Adv. Mater. 2017, 29, 1606817. (40) Kim, H.; Lee, J.; Ahn, H.; Kim, O.; Park, M. J. Nat. Commun. 2015, 6, 7278−7287. (41) Xu, N.; Qian, T.; Liu, X.; Liu, J.; Chen, Y.; Yan, C. Nano Lett. 2017, 17, 538−543. (42) Ding, N.; Zhou, L.; Zhou, C.; Geng, D.; Yang, J.; Chien, S. W.; Liu, Z.; Ng, M. − F.; Yu, A.; Hor, T. S. A.; Sullivan, M. B.; Zong, Y. Sci. Rep. 2016, 6, 33154−33163. (43) Zhang, S. S. J. Power Sources 2016, 322, 99−105. (44) Zhang, J.; Shi, Y.; Ding, Y.; Peng, L.; Zhang, W.; Yu, G. Adv. Energy Mater. 2017, 7, 1602876. (45) Guo, J.; Du, X.; Zhang, X.; Zhang, F.; Liu, J. Adv. Mater. 2017, 29, 1700273.

H

DOI: 10.1021/acs.nanolett.8b01882 Nano Lett. XXXX, XXX, XXX−XXX