Research Article www.acsami.org
Cite This: ACS Appl. Mater. Interfaces 2017, 9, 38950-38958
Li+‑Permeable Film on Lithium Anode for Lithium Sulfur Battery Yan-bo Yang,† Yun-xia Liu,†,‡ Zhiping Song,† Yun-hong Zhou,† and Hui Zhan*,†,§ †
College of Chemistry and Molecular Sciences, Wuhan University, Hubei 430072, China Chemistry and Pharmaceutical Engineering, Chongqing Industry Polytechnic College, Chongqing 401120, China § Engineering Research Center of Organosilicon Compounds & Materials, Ministry of Education, Wuhan University, Hubei 430072, China ‡
S Supporting Information *
ABSTRACT: Lithium−sulfur (Li−S) battery is an important candidate for next-generation energy storage. However, the reaction between polysulfide and lithium (Li) anode brings poor cycling stability, low Coulombic efficiency, and Li corrosion. Herein, we report a Li protection technology. Li metal was treated in crown ether containing electrolyte, and thus, treated Li was further used as the anode in Li−S cell. Due to the coordination between Li+ and crown ether, a Li+permeable film can be formed on Li, and the film is proved to be able to block the detrimental reaction between Li anode and polysulfide. By using the Li anode pretreated in 2 wt % B15C5containing electrolyte, Li−S cell exhibits significantly improved cycling stability, such as∼900 mAh g−1 after 100 cycles, and high Coulombic efficiency of>93%. In addition, such effect is also notable when high S loading condition is applied. KEYWORDS: lithium sulfur battery, lithium anode, lithium protection, crown ether, Li-crown ether coordination wave of research on solid-state electrolyte16 and triggered the research on Li protection as well. Some new strategies, including the physical or chemical method, were proposed accordingly. A hybrid anode structure with electrically connected graphite to the Li metal in parallel was proposed to protect the Li anode from the attack of polysulfide.17 Besides, a protective layer was chemically formed on Li anode during the charge−discharge through the cross-linking reaction of the curable monomer in the presence of liquid electrolyte and photoinitiator.3 Except the direct protection of Li metal, some indirect ways of aiming to block the harmful reaction between Li and polysulfides were also proposed. Typical examples involve LiNO3 additive18,19 and Nafion-attached multilayer separator.20,21 The former can react with Li anode to form a protective film before polysulfide diffuses to Li, and the latter can stop the polysulfide ion from permeating through the separator. The application of Li+-permeable Nafion membrane in Li−S cell inspired us to use other ion-selective material, more specifically, Li+-selective material to prohibit polysulfide ions from accessing Li metal. Crown ether was chosen because of its exclusive Li+ permeability. Crown ether was first used as conductivity enhancer or anode additive in Li secondary battery.22 Shu et al.23 used crown ethers to reduce the electrolyte decomposition on
1. INTRODUCTION High specific energy and long cycle life are always the primary requirements imposed on the battery for portable electronic devices, power tools, and electric vehicles.1 Although Li-ion battery has made great progress since its first commercialization in the 1990s, the space for further improvement in capacity and energy seems to be quite limited.2,3 Thus, an alternate battery with elemental sulfur (S) cathode and Li metal anode has gained much attention in the recent decade. It has the merits of low cost, high capacity (theoretical capacity of sulfur is 1675 mAh g−1), and potential high-energy output (≥500 Wh kg−1). However, many issues still remain unsolved. Among them, the impact of liquid organic electrolytes on both cathode and anode has been recognized and emphasized. It causes dissolution of lithium polysulfide, which is believed to be the origin of the shuttle loop, dissolution loss of sulfur, and further Li degradation.4−7 Preventing the harmful reactions between Li anode and polysulfide is a good solution. For this purpose, using S/C, S/ polymer, or other multilayer composites instead of pure S as the cathode were proposed.8−10 Besides, constructing a buffer layer between S cathode and separator was also tried.11−14 All these mainstream approaches could effectively confine the dissolved polysulfide species within the cathode or “cathode region”. In the past few years, Li anode degradation in Li−S battery was recognized, and it was revealed as a more crucial issue that needed to be tackled when practical pouch cell instead of experimental coin cell was being used.15 It directly set off the © 2017 American Chemical Society
Received: July 15, 2017 Accepted: October 17, 2017 Published: October 17, 2017 38950
DOI: 10.1021/acsami.7b10306 ACS Appl. Mater. Interfaces 2017, 9, 38950−38958
Research Article
ACS Applied Materials & Interfaces
Figure 1. (a) FTIR spectra of B15C5-pretreated Li anode (before and after cycling) and B15C5 powder; (b) UV−vis spectrum of B15C5, 1 mol L−1 LiTFSI/DOL-DME-treated Li metal, and 2%-B15C5-added 1 mol L−1 LiTFSI/DOL-DME-treated Li metal; (c) impedance of B15C5-treated Li metal after being stored in 1 mol L−1 LiTFSI/DOL-DME electrolyte for different times. stored in Ar-filled glovebox. The S content in S/C composite was determined as 70% by thermogravimetric (TG) analysis. 2.3. Cell Assembly and Electrochemical Measurement. S electrodes were fabricated by mixing the S/C composite with polyvinylidene fluoride (PVdF) binder at a weight ratio of 85:15 in N-methyl-2-pyrrolidone (NMP) dispersant to form a homogeneous slurry. The slurry was coated on aluminum foil and vacuum-dried at 60 °C for 3 h.31 Then the electrode was punched and vacuum-dried at 60 °C for another 8 h before the cell assembly. Typical S loading was 1.2 mg cm−2. CR2016-type coin cells were fabricated in the glovebox. Untreated or pretreated Li anodes and the aforementioned S/C cathode were separated by a piece of Celgard 2400 separator. One mol L−1 LiTFSI/ DOL-DME (1:1 v/v) was used as the electrolyte. Galvanostatic charge and discharge was performed on a battery charge−discharge tester (CT2001A, Land, China) in the voltage range of 1.5−3.0 V, and the capacity calculation was based on the mass of S. Electrochemical impedance spectra (EIS) of Li anodes were collected on Autolab PGSTAT30 electrochemical workstation (Eco Chemie) with threeelectrode cells, in which another two fresh Li plates were used as the reference and counter electrode. (In the impedance measurement of the cycled Li anodes, Li anode, along with the separator, was taken out from the coin cell after 10 cycles and then used as the working electrode.) The frequency range was 100 kHz to 0.1 Hz. All the electrochemical tests were carried out at room temperature. 2.4. Morphology Observation and Surface Characterization of Li Anode. The morphology of Li anodes before and after cycling was observed on SEM (SEM 6095, Carl Zeiss). The surface components on Li were determined by FTIR spectra (IS10, Thermo). The cycled Li anode was rinsed with DME solvent. After being naturally dried in the glovebox, the surface layer of Li anode was carefully scraped off, and the scraping was collected, mixed with KBr, and then subjected to FTIR measurement. XPS (ESCALAB 250Xi, Thermo Fisher) measurement was carried out to compare the surface species on cycled Li anodes, and monochromatic Al Kα (1486.6 eV) was used as the light source. After 10 cycles, Li−S cells were dismantled in the glovebox and Li anodes were taken out and washed with DME. After being naturally dried in the glovebox, they were well-sealed in an airtight bag prior to XPS measurement.
graphite anode in Li-ion battery, and Morita et al. found that 12-crown-4 could improve the capacity and rechargeability of TiS2 anode.24 When being used in Li−air battery, adding 12crown-4 and 15-crown-5 could increase the electrolyte conductivity as well as the discharge capacity.25 The improvements are all related to the exclusive chelation between Li+ and macrocyclic ligand.26 In this work, Li metal was preconditioned by benzo-12crown-4 (B12C4), benzo-15-crown-5 (B15C5), and dibenzo18-crown-6 (DB18C6). The preconditioned Li was further used as the anode in Li−S cell. Electrochemical tests proved that Li pretreatment led to greatly enhanced columbic efficiency and cycling stability. The effect of crown ether on Li was further investigated in terms of scanning electron microscope (SEM), Fourier transform infrared spectrum (FTIR), and X-ray photoelectron spectroscopy (XPS) analysis.
2. EXPERIMENTAL SECTION 2.1. Li Pretreatment. Different amounts of crown ether (2 wt % for B12C4, 2 wt % for B15C5, and 1 wt % for DB18C6) were added to 1 mol L−1 bis(trifluoromethane)sulfonimide lithium (LiTFSI) in 1,3dioxolane (DOL) and 1,2-dimethoxyethane (DME) (1:1 v/v) electrolyte. Li foil was dipped into the above electrolyte for 5 min and then taken out and naturally dried in Ar-filled glovebox before use. Here, we mainly chose benzo-substituted crown ether in our study for two reasons. First, benzo group substitution could weaken the electron-donor ability of the ring oxygens, which means it could induce higher metal-ion mobility (specifically Li+ mobility in our work).26,27 Second, aryl group substitution could reduce the solubility of crown ether in water and other aprotic solvent,28−30 which could greatly help avoid the possible dissolution of crown ether in the electrolyte during the cycling. 2.2. Preparation of S/C Composite Cathode Material. S/C composite was obtained through a melt-infiltration method. S was mixed with conductive carbon (ECP-600JD, Ketjen Black) at the weight ratio of 7:3, and the mixture was heated at 155 °C for 10 h to ensure S was infiltrated into the pores of the conductive carbon. After being cooled to room temperature, the powder was collected and 38951
DOI: 10.1021/acsami.7b10306 ACS Appl. Mater. Interfaces 2017, 9, 38950−38958
Research Article
ACS Applied Materials & Interfaces
Figure 2. SEM images of (a) fresh Li foil and Li foils pretreated with (b) standard electrolyte and the electrolyte containing (c) B12C4, (d) B15C5, and (e) DB18C6.
Figure 3. (a) Cycle performance of Li−S cells with untreated Li anode and Li anodes pretreated in crown ether-containing electrolyte (voltage range = 1.5−3.0 V; current = 0.2 C); (b) impedance spectra of Li anodes after 20 cycles.
surface.35−37 Therefore, in our study, when Li metal was treated in crown ether containing electrolyte, a polymer-like film could form on the Li surface, being embedded with Li+-crown ether complex ([Li+-crown ether]TFSI−) from the coordination reaction, alkoxides from DOL reduction,38,39 and some Fcontaining species from the salt decomposition. The construction of the film is quite similar to the classic Li+-permeable film,40 except with DOL polymer replacing polyvinyl chloride (PVC) as the polymer base and TFSI− instead of Cl− as the counteranions (the existence of TFSI− was evidenced by the S2p XPS spectrum of B15C5-tretead Li metal in Figure S1). Similar Li+-selective coating film was formed on Li in Li−air battery, and capacity increase was observed. The Li-crown ether complex-containing surface film on Li metal was confirmed by FTIR measurement. In Figure 1a, an FTIR spectrum of B15C5-treated Li anodes (before and after 20 cycles) is presented, and B15C5 powder is used as the reference. According to previous literature about the FTIR characterization of crown ether,41,42 the characteristic bands at 2948 cm−1 should be assigned to aromatic ν(C−H)
UV−visible spectroscopy (UV-3600, Shimadzu) was employed to study the surface chemistry on Li metal. Li metal was dipped in standard 1 mol L−1 LiTFSI/DOL-DME electrolyte or 2%-B15C5added 1 mol L−1 LiTFSI/DOL-DME electrolyte. After being naturally dried in the glovebox, Li metal was subjected to UV−vis measurement and the reflection mode was adopted during the measurement. For better comparison, UV−vis measurement was also conducted on pure B15C5.
3. RESULTS AND DISCUSSION 3.1. Effect of Crown Ether Pretreatment on Li Anode. The application of crown ether in Li secondary battery can be summarized in two aspects. First, it is usually utilized as conductivity enhancer either in Li−air or Li-ion battery because the particular coordination of crown ether to Li+ can obviously reduce the cation−anion interaction and hence increase the ion conductivity.25,32,33 Also, thanks to the coordination, crown ether can enhance the dissociation of oxide generated in Li−air battery and reduce its precipitation.34 On the other side, DOL is usually considered as a good solvent for Li metal anode as it can easily polymerize and form an elastic barrier on Li 38952
DOI: 10.1021/acsami.7b10306 ACS Appl. Mater. Interfaces 2017, 9, 38950−38958
Research Article
ACS Applied Materials & Interfaces
weaker one at ∼303 nm. The interaction of the crown-ether fragment with the metal cation often causes a blue-shift in the UV−vis spectrum of crown ether, and it is reasonable to believe that the UV−vis absorbance in B15C5-treated Li metal originated from the Li-crown ether complex.42,44 EIS measurement also confirmed the stability of the surface film on B15C5-treated Li. B15C5-treated Li was stored in electrolyte for 20 days and then taken out for EIS measurement. Figure 1c shows the impedance evolvement during the storage. We found that, after 20 days, the semicircle indicative of the SEI film resistance just slightly enlarged, suggesting the chemical stability of the surface film on preconditioned Li.45 The surface morphology of bare Li and pretreated Li is compared in Figure 2. Untreated Li had a smooth metal surface with some scratches on it. On the other side, after being immersed into the electrolyte, no matter if crown ether was or was not added, Li metal was covered by a surface layer. The observation agrees with the previous study about the interaction between Li and electrolyte, in which the surface layer was also found on Li when it was exposed to DOL.43 The effect of crown ether pretreatment was first evaluated by charge−discharge tests. Figure 3a compares the cycling performance of the Li−S cells with preconditioned Li anodes (when treating Li anode, the mass concentration of B12C4 and B15C5 in 1 mol L−1 LiTFSI/DOL-DME electrolyte was 2%, and the concentration of DB18C6 was 1% because of its limited solubility in the electrolyte). As seen, using preconditioned Li anodes led to better cycling stability. For instance, the initial discharge capacity of Li−S cell with untreated Li anode was 1200 mAh g−1 and declined to 740 mAh g−1 after 50 cycles, but in the experimental group, the 50th capacity increased to 816, 878, and 992 mAh g−1 if Li anodes were pretreated by DB18C6, B12C4, and B15C5, respectively. Along with the improved cycling stability, the Coulombic efficiency was increased from ∼80% in the baseline to 90−95% in the experimental group. More detail can be revealed if carefully comparing the Li−S cells with preconditioned Li anodes. B15C5 pretreatment on Li anode resulted in the best cycling stability and the highest efficiency, while the effect of DB18C6 pretreatment was exactly the opposite: it had lower efficiency and poorer cycling stability. A similar difference was also observed in the previous study. When using crown ether as the additive in Li−air battery, adding 12C4 or 15C5 in electrolyte usually led to higher ionic conductivity and obvious capacity increase, while such effect of 18C6 was mostly negligible.25 Also, when preparing Li+selective membrane, 12C4 or 15C5 rather than 18C6 was used.46 All these were ascribed to the combining ability between different crown ether and Li+. In our study, the different effects of B12C4, B15C5, and DB18C6 could be explained similarly. It is well-known that the cavity size greatly determines the coordination between Li and crown ether, which further determines the Li+ selectivity of crown ether.26,27,47 Meanwhile, the solvent environment also has some influence on the Li+-crown ether coordination.48 Thus, in a classic selective membrane, B12C4 shows the best selectivity toward Li+, while in Li−air battery, B15C5 is usually found to have the strongest coordination with Li+. In any case, comparing with the weak coordination between DB18C6 and Li+, the cavity size of B12C4 or B15C5 better matches Li+, and they have higher chelating power toward Li+; in other words, they are much more permeable to Li+. Thus, when treating Li anode in the electrolyte added with B12C4 or B15C5 crown
Figure 4. XPS spectra of Li anodes in Li−S cell after being cycled at 0.2 C for 10 times: (a) untreated Li anode, (b) DB18C6-pretreated Li anode, and (c) B15C5-pretreated Li anode.
stretching vibration, the band at 1500 cm−1 should be ascribed to the ν(C−H) vibration of phenyl ring, the 1590 cm−1 band represented the ν(CC) vibration of phenyl ring, the 1256 cm−1 band was attributed to the ν(C−O) vibration, and the 1122 cm−1 band should be assigned to the ν(C−O−C) vibration. All these characteristic bands could be clearly seen in B15C5-treated Li, indicating the surface film containing crown ether moiety was formed on Li. Meanwhile, the band at 1630 cm−1 represents the ν(−OCO2Li) vibrationthat originated from the decomposition of DOL on Li anode.43 When the FTIR spectrum was collected again on Li anode that had been cycled 20 times, these characteristic bands were reserved, suggesting that, after repetitive charging−discharging, the crown ether moiety was well-maintained in the surface film. UV−vis spectroscopy further proved the existence of crown ether moiety in the surface film. In Figure 1b, the UV−vis spectra of Li metal dipped in standard electrolyte and 2%B15C5-added electrolyte are compared, and pure B15C5 sample was presented as the reference. If Li metal was only dipped in standard electrolyte, no UV−vis signal could be found on it. On the contrary, in pure B15C5, very strong absorbance at 281 nm and a weaker one at 310 nm was found. The result is quite understandable as UV−vis signals within the 200−400 nm near-ultraviolet range usually represent a conjugated molecular structure. The absorbance signal in B15C5 should be ascribed to the highly conjugated benzosubstituted crown ether, which is absent in standard 1 mol L−1 LiTFSI in DOL+DME electrolyte. However, if Li metal was dipped in 2%-B15C5-containing electrolyte, we could see similar UV−vis absorbance at ∼267 nm and also a relatively 38953
DOI: 10.1021/acsami.7b10306 ACS Appl. Mater. Interfaces 2017, 9, 38950−38958
Research Article
ACS Applied Materials & Interfaces
Figure 5. Front view (a−c) and cross section (d−f) of Li anodes in Li−S cell after 10 cycles: (a, d) untreated, (b, e) pretreated with 1 wt % DB18C6-containing electrolyte, and (c, f) pretreated with 2 wt % B15C5-containing electrolyte.
ether, the surface film formed on Li could be more permeable to Li+ and better block the entering of Sx2− anion, while such a blocking effect was negligible if treating Li in DB18C6-added electrolyte. EIS analysis supplies more proof of the relation between the Li+-crown ether coordination and the blocking effect. After being cycled 20 times at 0.2 C (1 C = 1675 mAh g−1), Li−S cells were dissembled in Ar-filled glovebox. Li anode, along with the separator, was taken out. Then a three-electrode cell was assembled with the cycled Li anode as the working electrode and another two Li plates as the reference and counter electrodes. Figure 3b compares the impedance spectra of Li anodes after cycling. The difference was mainly in the highfrequency range. The diameter of the semicircle in the highfrequency range decreased in the order of untreated Li anode ≈ DB18C6 pretreated Li anode > B12C4-pretreated Li anode >
B15C5-pretreated Li anode. Usually, the increase of Li impedance during the cycling comes from the accumulation of the insulating byproduct formed on Li anode (major component was Li2S2 or Li2S).49 Therefore, the biggest surface impedance on untreated Li or Li anode pretreated with DB18C6 is easily understood. As discussed above, DB18C6 barely can complex with Li+ and block the polysulfide ions; even with the surface film containing DB18C6, Li still cannot resist the Sx2− corrosion, and insulative Li2S2 or Li2S thus grows on Li and increases the surface impedance. On the other hand, the B12C4- or B15C5-containing surface layer can act as a Lipermeable “wall” on Li anode and reduce the generation of Li2S2/Li2S. The smallest semicircle observed on B15C5-treated Li should be ascribed to the larger Li+ conductivity of the B15C5-induced surface film, as widely observed in Li−air battery.50 38954
DOI: 10.1021/acsami.7b10306 ACS Appl. Mater. Interfaces 2017, 9, 38950−38958
Research Article
ACS Applied Materials & Interfaces
species, such as sulfone or sulfite, and a number of bands between 160.0 and 165.0 eV were most likely associated with sulfide species.51,52 All these species were found on standard Li and DB18C6-treated Li anode. However, on B15C5-treated Li, the energy band at the lower-energy region (1000 mAh g−1; as the current rate increased to 0.5 C, the capacity was reserved at 884 mAh g−1, and even at the current rate of 2 C, the capacity was 536 mAh g−1, still much higher than that of standard Li−S cell. Therefore, the possible impact of B15C5 on rate capability is not a big issue, as for Li−S cell, the “healthy” Li anode is much more important than the possibly slower Li+ transfer across the surface layer. We further tested the effect of the crown ether pretreatment on Li−S cell with “thick” S cathode (sulfur loading was 2 mg cm−2), and the result is shown in Figure 8b. With high S loading, the Li−S cell with bare Li anode had even lower efficiency of 95% accompanied with very stable cycling, and after 30 cycles, it started to show obvious capacity advantage over the standard Li−S cell.
on the cycled Li anode was determined as +1 by the XPS characterization).54 While on B15C5-treated Li anode, front view and cross section both evidenced that the surface film on Li was dense and thin and the surface layer was well-reserved after cycling. More importantly, the mossy-like morphology observed on bare or DB18C6-tretaed Li was barely found. By combining with the XPS result in Figure 4, it could be concluded that the protection layer on B15C5-treated Li will not be removed by repeated cycling, and it can greatly help to resist the corrosion from polysulfide. Summarizing the above results, the role of crown ethercontaining surface film on Li can be illuminated as follows (Figure 6). Without crown ether pretreatment, Sx2− can access Li anode and, thus, Li corrosion happens. With DB18C6 pretreatment, the protection effect was insufficient, so Sx2− can still access Li anode, although not as much as on bare Li. Only with B12C4 or B15C5 pretreatment could Sx2− be mostly blocked on the outside, and the detrimental reaction between Li and Sx2− was thus hindered. As B15C5 treatment induced a surface film with higher Li+ conductivity, in the following, we will further discuss the effect of B15C5 and focus on the influence of different contents of B15C5. 3.2. Effect of B15C5 Content. Li anodes were treated in B15C5-containing electrolyte with different B15C5 concentrations, i.e., 1, 2, and 5 wt %, respectively (we did not try other B15C5 concentrations; as in the preparation of Li-selective membrane, low crown ether content, usually
