Improving the electrochemical performance of high voltage spinel cathode at elevated temperature by a novel electrolyte additive

Journal of Power Sources 303 (2016) 41e48

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

Improving the electrochemical performance of high voltage spinel cathode at elevated temperature by a novel electrolyte additive Jiahui Chen, Hui Zhang, Mingliang Wang, Jianhong Liu, Cuihua Li*, Peixin Zhang** Department of Chemistry and Chemical Engineering, Shenzhen University, Shenzhen 518060, China

h i g h l i g h t s
AMSL is proposed as a novel electrolyte additive for high voltage spinel cathode.
The cycling performance of LiNi0.5Mn1.5O4 is dramatically improved by using AMSL.
A less resistive and high thermal stable SEI film is formed on the cathode surface.
The AMSL derived SEI film is essential to suppress the electrolyte decomposition.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 July 2015 Received in revised form 24 October 2015 Accepted 26 October 2015 Available online xxx

In this work, we report a novel electrolyte additive allyloxytrimethylsilane (AMSL) to improve the electrochemical performance of high voltage LiNi0.5Mn1.5O4 cathode. In presence of 0.5% AMSL, the discharge capacity retention of Li/LiNi0.5Mn1.5O4 cell is improved from 73.1% to 80.2% after 500 cycles at room temperature, and from 52.4% to 92.5% after 100 cycles at 55  C. Moreover, the Li/LiNi0.5Mn1.5O4 cell with AMSL delivers a superior discharge capacity of 95.6 mAh g1 at high rate of 3 C, whereas the cell without AMSL only remains 76.8 mAh g1. Theoretical calculation and experimental results reveal that AMSL is oxidized prior to the carbonate solvents during the first charge process and then creates a less resistive and high thermal stable SEI film on the surface of LiNi0.5Mn1.5O4 cathode. The AMSL derived SEI film, composed of organic silicon-based species, ether moieties and reduced LiF, is responsible for the suppression of serious electrolyte decomposition and dissolution of transition metal ions at high voltage, especially at high temperature. © 2015 Elsevier B.V. All rights reserved.

Keywords: Lithium ion battery Electrolyte additive Allyloxytrimethylsilane High voltage Spinel LiNi0.5Mn1.5O4

1. Introduction Li-ion batteries have been widely used as power sources for portable devices due to their characteristics of high energy and power densities, long cycle life and environmental friendliness [1e3]. However, conventional Li-ion batteries based on layered lithium metal oxides (e.g. LiCoO2) or olivine phosphates (e.g. LiFePO4) with poor energy density can hardly satisfy the increasing energy demand for electric vehicles, energy storage systems and other high power applications [4e7]. A high density in Li-ion battery can be obtained by increasing either the operating potential or the reversible capacity of the cathode. As the attempts of increasing

* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (P. Zhang).

(C.

http://dx.doi.org/10.1016/j.jpowsour.2015.10.088 0378-7753/© 2015 Elsevier B.V. All rights reserved.

Li),

[email protected]

the working potential, spinel LiNi0.5Mn1.5O4 was proposed as a high voltage cathode candidate with operating voltage at ~4.75 V vs. Li/ Liþ and energy density of 650 Wh kg1 [8,9]. In this respect, LiNi0.5Mn1.5O4 has been extensively investigated as a promising material for the next generation Li-ion battery. Although LiNi0.5Mn1.5O4 cathode possesses high energy density, its practical application to lithium ion batteries is still quite challenging. Firstly, the conventional electrolyte readily suffers from substantial oxidative decomposition at voltages above 4.5 V vs. Li/ Liþ and elevated temperature (for instance, 55  C), which may result in the formation of resistive and unstable solid/electrolyte interface (SEI) film on the cathode surface [10,11]. In addition, the commonly used LiPF6 salt in current commercialized electrolyte is very susceptible to hydrolysis even if trace amounts of water are present in the electrolyte [12], i.e.: LiPF6 þ H2O / POF3 þ LiF þ 2HF. The PF5, a thermal decomposition product of LiPF6, can also produce HF via many reaction paths [13]. This resulting HF attacks the

42

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

LiNi0.5Mn1.5O4 cathode and thus causes Mn and Ni ions dissolution. Moreover, the HF severely consumes active Liþ to form LiF and promotes useless SEI film growth. These undesired processes significantly scarify battery cycling life, especially at elevated temperature. To improve the stability of electrolyte, the use of novel electrolyte solvents and electrolyte additives has been extensively investigated. Ionic liquids [14], fluorinated ethers and carbonates [15e17], sulfones [18], and dinitriles [19] with high anodic stability have been explored as high voltage electrolyte solvents. Unfortunately, these solvents still remain problems to be solved, such as, high viscosity, poor compatibility with anodes and difficulties in synthesis. Zhang et al. [20] reported the combination of fluorinated cyclic and linear carbonates together with fluorinated ether as electrolytes to improve the electrochemical performance of Li4Ti5O12/LiNi0.5Mn1.5O4 full cell at elevated temperature. But these fluorinated electrolytes tend to decompose on graphite anode continuously. Dong et al. [21] employed a sulfone/carbonate mixed electrolyte for Li/LiNi0.5Mn1.5O4 cell and obtained improved capacity retention from 81% to 98% at 5 C rate. According to recent reports, the use of film forming electrolyte additives is an effective and economic way to improve the electrochemical performance of LiNi0.5Mn1.5O4 cathode [22]. These additives usually decompose preferentially to form protective film on the cathode surface and thus stabilize the cathode/electrolyte interface and suppress electrolyte decomposition. Yan et al. [23] reported that a stable SEI film can be formed on the LiNi0.5Mn1.5O4 cathode through the polymerization of 1-propylphosphonic acid cyclic anhydride (PACA), and consequently suppressed the self-discharge and transition metal ions dissolution, leading to improved cycling stability. By using tris(trimethylsilyl) phosphate (TSMP) as an electrolyte additive, Rong et al. [24] improved the capacity retention of Li/ LiNi0.5Mn1.5O4 cell from 70% to 94.9% after 70 cycles at 55  C (cycled between 3.0 and 4.9 V at 0.5 C). A detailed investigation from Song's group revealed that TMSP was capable of eliminating HF from the electrolyte and modifying the surface of LiNi0.5Mn1.5O4 cathode [25]. Very recently, Huang et al. [26] proposed phenyl trifluoromethyl sulfide (PTS) as additive to enhance the interfacial stability between LiNi0.5Mn1.5O4 and electrolyte. The Li/ LiNi0.5Mn1.5O4 cell with PTS containing electrolyte exhibited high capacity retention of 95% after 100 cycles at 55  C (cycled between 3.0 and 4.9 V at 1 C). To date, massive film forming additives have been explored, but the achievement is not satisfying yet. Density functional theory (DFT) calculation provides a more efficient and economic approach for discovering appropriate additives and finding mechanisms of their actions than the experimental trial and error testing. Bearing this in mind, a novel electrolyte additive, allyloxytrimethylsilane (AMSL), was screened out through DFT calculation. We expected that it will create a protective SEI film on the LiNi0.5Mn1.5O4 cathode surface that can alleviate electrolyte decomposition and transition metal ions dissolution. Based on a comprehensive understanding of the effect of AMSL, electrochemical performances, thermal stability and surface chemistry of LiNi0.5Mn1.5O4 cathode were systematically investigated. 2. Experimental 2.1. Calculation methods All calculations were performed by the Gaussian 09 package [27]. The equilibrium structures were optimized at the DFT/B3LYP level using 6-31 þ G (d, p) basis set. Conductor-variant polarized continuum model (CPCM) was employed to investigate the effects of solvents. A dielectric constant of 28.8 was adopted as a volume

average value of the dielectric constants of ethylene carbonate (EC: 89.2), ethyl methyl carbonate (EMC: 2.9) and diethyl carbonate (DEC: 2.8), since EC:EMC:DEC ¼ 3:5:2 (by volume) solution is used as an electrolyte solvent in this study. The frontier molecular orbital energy and Liþ binding energy were calculated at B3LYP/6-31 þ G (d, p) level. The calculated oxidation potential was converted from the free energy cycle for the oxidation reaction [28]. 2.2. Electrolyte and electrode preparation Battery grade carbonate solvents and LiPF6 were kindly provided by Optimum Nano Energy Co. Ltd. AMSL was purchased from Alfa Aesar and used as received. The blank electrolyte was consisted of 1 mol/L LiPF6 in a solvent mixture of EC, EMC, and DEC in 3:5:2 volume ratio. The additive containing electrolyte was obtained by adding 0.5 wt% AMSL into the blank electrolyte. The HF and water content in these electrolytes were controlled to less than 15 ppm. LiNi0.5Mn1.5O4 electrode was prepared with a mixture of 80 wt% LiNi0.5Mn1.5O4 (Shenzhen Tianjiao Co. Ltd), 10 wt% acetylene carbon black and 10 wt% polyvinylidene (PVDF) binder in N-methylpyrrolidone. The resulting slurry was cast onto an aluminum foil and dried overnight at 70  C. After drying, the electrode was punched into 14 mm diameter round discs and dried again at 110  C for 8 h. The active material loading is about 1.6 mg cm2. Graphite electrode was prepared with the same method as LiNi0.5Mn1.5O4 electrode, coating a mixture of 80 wt% artificial graphite (Shenzhen Kejing Co. Ltd), 10 wt% acetylene carbon black and 10 wt% PVDF on copper foil. For the electrochemical tests, CR2032 type coin cells were assembled in an Ar-filled glove box (Unilab, Mbraun) with LiNi0.5Mn1.5O4 cathode or graphite anode as working electrode, lithium foil as counter electrode, Celgard 2400 polypropylene membrane as separator and prepared electrolyte. Same amount of electrolyte of 50 mL was added for each cell. 2.3. Electrochemical measurements The fabricated Li/LiNi0.5Mn1.5O4 coin cells were galvanostatically precycled for three formation cycles at 0.1 C (1 C ¼ 147 mA g1) between 3.5 and 5 V on LAND battery test system (CT2001A, Wuhan Land Electronics Co. Ltd). Thereafter, the cells were cycled at different current densities between 3.5 and 5 V at room (25  C) and elevated (55  C) temperature. Rate capability test was conducted at various C rates: 0.1 C, 0.2 C, 0.5 C, 1 C, 2 C and 3 C. To investigate the impact of AMSL on the graphite anode, Li/graphite cells were galvanostatically cycled at 0.5 C (1 C ¼ 372 mA g1) between 0.01 and 3 V at 25  C after three formation cycles at 0.1 C. Electrochemical impedance spectroscopy (EIS) measurements were carried out on a potentiostat (1470E, Solartron) coupled with a frequency response analyzer (FRA 1260, Solartron) in frequency range from 106 Hz to 0.01 Hz with an amplitude of 10 mV. Cyclic voltammetry was also performed on the Solartron 1470E potentiostat between 0.01 and 2 V with a scan rate of 0.1 mV s1. 2.4. Material characterization After cycling at elevated temperature, the cells were disassembled in Ar-filled glove box. The cathodes were rinsed with anhydrous dimethyl carbonate (DMC) three times to remove residual electrolyte components followed by vacuum drying at 40  C for surface characterization and thermal stability test. Surface morphology of the cathodes was observed by field emission scanning electron microscopy (FE-SEM, S3400N, Hitachi). X-ray photoelectron spectroscopy system (XPS, Quantera-II, Ulvac-Phi) was used to analyze the surface compositions, using a focused monochromated Al Ka radiation under ultra-high vacuum. All XPS

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

spectra were calibrated by the hydrocarbon C1s line at 284.6 eV. Line syntheses of spectra were conducted using GaussianeLorentzian (80:20) curve fitting on XPSpeak 4.1 software. Attenuated total reflectance fourier transform infrared (ATR-FTIR) spectra were collected using an IR spectrometer (Nicolet 6700, Thermo Fisher) which has smart accessory with a ZnSe crystal. The thermal stability of the cycled cathode in electrolyte was carried out by differential scanning calorimeter (DSC, 200F3, Netzsch). 2 mg cathode material and 5 mL electrolyte were sealed in aluminum pan and transferred into the DSC sample holder. The pan was heated from 30 to 350  C at a rate of 5  C min1. For the Mn, Ni deposition test, cycled lithium foils were also collected from the dissembled cells and washed with anhydrous DMC. After vacuum drying, the lithium foils were dissolved in 5 mL 2 wt% HNO3 solution and then analyzed with inductively coupled plasma atomic emission spectrometry (ICP-AES, Optima 7000DV, Perkin Elmer).

3. Result and discussion 3.1. Oxidative stability of AMSL As widely reported, molecular orbital energy and Liþ binding affinity (Eb) play important role in forming successful SEI film and obtaining great battery performance [29e32]. For the former, it is often used to predict the oxidation and reduction tendency of organic molecules, while the latter is a predominant characteristic attracting the molecules to the cathode surface. The highest occupied molecular orbital (HOMO)/lowest unoccupied molecular orbital (LUMO) energies and calculated oxidation potentials of the three carbonate solvents and AMSL additive are listed in Table 1. The HOMO energy of AMSL is higher than that of the carbonate solvents, indicating that AMSL is a better electron donor than any other molecule. Indeed, as calculated, AMSL possesses the lowest oxidation potential. It should be note that AMSL has the smallest LUMO energy value, resulting readily reduction compared with EC, EMC and DEC. As for comparison, vinylene carbonate (VC), the wellknown SEI forming additive for anodes, is calculated by the same method. Clearly, the LUMO energy of VC is comparable with that of AMSL. Therefore, AMSL may be considered as a potential additive for anodes. A low Liþ binding affinity is a prerequisite of SEI forming additive, since the desolvation process of Liþ is the major energy consuming step in Li-ion battery [33,34]. In other words, weak interaction between additive and Liþ is beneficial for the accumulation of additive on the cathode surface. The optimized structures and Liþ binding affinities of EC, EMC, DEC and AMSL with Liþ are presented in Fig. 1. The Eb value (1.883 kcal mol1) of AMSL is smaller than the EC, EMC and DEC values of 4.817, 4.146 and 4.450 kcal mol1 respectively. Since the interaction is weaker between AMSL and Liþ, it is more easily for AMSL to enrich the cathode surface. In conclusion, AMSL is a promising oxidative additive candidate that may form protective film on the surface of Table 1 Frontier molecular orbital energy (eV) and calculated oxidation potential (V vs. Li/ Liþ) of EC, EMC, DEC, AMSL and VC in solution. Material

HOMO

LUMO

Oxidation potential

EC EMC DEC AMSL VCa

8.6848 8.3691 8.3188 7.2682 7.2605

0.0531 0.1513 0.1763 0.3192 0.2887

7.06 6.73 6.71 5.37 5.68

a VC (Vinylene carbonate) is used for comparison here, which has similar HOMO and LUMO energies to AMSL.

43

LiNi0.5Mn1.5O4. The detail effects of AMSL were identified through evaluation of the electrochemical performances of cells and surface compositions of the cycled LiNi0.5Mn1.5O4 cathode. 3.2. Battery performance at room temperature The cyclic performances of Li/LiNi0.5Mn1.5O4 cells using electrolyte with and without AMSL are given in Fig. 2a. A better cycling stability of LiNi0.5Mn1.5O4 cathode is achieved in the additive containing electrolyte, delivering a capacity of 133.6 mAh g1 initially and maintaining 107.2 mAh g1 after 500 cycles. The discharge capacity retention is improved from 73.1% to 80.2% by using AMSL as an electrolyte additive. The charge capacity of the cell with AMSL is about 7 mAh g1 higher than that without AMSL in the first cycle; while the coulombic efficiency of the former (86.5%) is lower than the latter (89.7%). This lower initial efficiency and capacity fade with cycling can be explained by the oxidative reaction of AMSL and formation of SEI film on the LiNi0.5Mn1.5O4 cathode. As shown in Fig. 2b, the discharge capacity of the cell with blank electrolyte starts to decrease significantly at 1 C and is 121.4 mAh g1 after 10 cycles at 1 C. Whereas, the cell with AMSL containing electrolyte deliver capacities of 107.4 and 95.6 mAh g1 at high discharge rates of 2 C and 3 C, which are much higher than that with blank electrolyte. In addition, when the rate returned to 0.1 C, the discharge capacity of both the cells can be partly restored. It is believed that the AMSL derived SEI film not only can inhibit the continuous electrolyte decomposition, but also facilitate Liþ and charge transfer at cathode/electrolyte interface. More clear evidences for the role of AMSL on the remarkable cycling performance of the LiNi0.5Mn1.5O4 cathode are presented in Fig. 2c and d. Electrochemical impedance spectra of Li/ LiNi0.5Mn1.5O4 cells with and without AMSL were collected before cycle and after 3, 300 and 500 cycles, respectively. Each curve is consisted of a high frequency semicircle, an intermediate frequency semicircle and a low frequency slop line, except the one measured before cycle, which is only composed of a semicircle and a tail. As previously reported, the high frequency range of the semicircle is attributed to the surface film resistance (Rf), and the middle frequency range of the semicircle is corresponded to the charge transfer resistance (Rct) between electrode and electrolyte [35,36]. The low frequency region of the straight line is related to the Liþ diffusion process in the bulk of the electrode. The spectra were fitted with the equivalent circuit model insert in Fig. 2c and the detailed resistances are summarized in Table 2. Before cycling, the impedance of the cell with AMSL is much smaller. This is probably due to the weaker interaction ability between AMSL and Liþ promotes the accumulation of AMSL onto the cathode surface, and thus resulting in lower impedance. As can be seen, the impedance of the cell with blank electrolyte considerably increases to 187 U for Rf and to 264 U for Rct after 500 cycles. This implies that undesired SEI film has been formed from the interfacial side reaction between cathode and electrolyte and functioned as a resistive layer for ionic conduction and charge transfer. In contrary, both Rf and Rct in the cell with AMSL containing electrolyte slowly increase to 98 U and 232 U after 500 cycles. They are much lower than those of the cell with blank electrolyte, indicating that SEI film provided by AMSL is relatively more compact and stable. The incorporation of AMSL is an effective way to enhance the interfacial stability between cathode and electrolyte, leading to better cycle stability and high rate capability. 3.3. Impact of AMSL at elevated temperature To better understand the positive impact of AMSL, galvanostatic charge/discharge cycles of Li/LiNi0.5Mn1.5O4 cells were carried out

44

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

Fig. 1. Optimized structures together with Liþ binding affinity values (Eb, in kcal mol1) in solution.

Fig. 2. Cycling stability (a) and rate capability (b) of Li/LiNi0.5Mn1.5O4 cells with and without AMSL under a voltage range of 3.5e5.0 V at 25  C. Electrochemical impedance spectra of LiNi0.5Mn1.5O4 cathodes cycled in the electrolyte without (c) and with (d) AMSL. The insert in (a) is the equivalent circuit model for spectra fitting.

Table 2 Fitted results of EIS spectra for Li/LiNi0.5Mn1.5O4 cells with and without AMSL. Sample Before cycle After 3 cycles After 300 cycles After 500 cycles

Rf þ Rct (U) R f ( U) Rct (U) R f ( U) Rct (U) R f ( U) Rct (U)

Blank electrolyte

With 0.5% AMSL

317 76 279 119 218 187 264

54 60 187 75 148 98 232

at a high temperature of 55  C, where the unwanted interfacial side reaction between cathode and electrolyte becomes more severe. A comparison of the cycling performance of the cathode cycled in blank and AMSL containing electrolytes is shown in Fig. 3a. Obviously, cell with the addition of AMSL presents dramatically improved discharge capacity. It delivers a discharge capacity of 117.5 mAh g1 after 100 cycles and exhibits capacity retention of 92.5%. While for the cell without AMSL only remains 67.1 mAh g1 after 100 cycles, corresponding to the capacity retention of 52.4%. The substantially improved cycling performance is ascribed to the

suppression of the electrode polarization and transition metal ions dissolution. More details will be discussed later. Fig. 3b also compares the coulombic efficiency of the cells with different electrolytes. The coulombic efficiency of the cell with blank electrolyte is low and unstable during cycling, whereas that of the cell with AMSL containing electrolyte shows much more stable and reaches above 92%. It can be confirmed that substantial electrolyte decomposition and excessive SEI growth, which result in a severe electrode polarization, are heavily alleviated. Fig. 3c and d presents the evolution of charge/discharge profiles of Li/LiNi0.5Mn1.5O4 cells with and without AMSL at different cycles. The potential difference between charge and discharge plateaus for the cell cycled in blank electrolyte is much larger than that cycled in AMSL containing electrolyte, indicating severe polarization of LiNi0.5Mn1.5O4 cathode. The excessive SEI film generated by aggressive interfacial side reaction between electrode and electrolyte covers the cathode quickly and consequently increases the electrode polarization under high voltages and elevated temperature [36]. When in the presence of AMSL, the polarization is effectively decreased due to the compact and stable SEI film formed by AMSL. The thermal stability of fully charged cathodes soaked in

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

45

Fig. 3. Cycling performance of LiNi0.5Mn1.5O4 cathodes cycled in the electrolyte with and without AMSL at 0.5 C and 55  C under a voltage range of 3.5e5.0 V: discharge capacity (a), coulombic efficiency (b), and charge/discharge profiles of LiNi0.5Mn1.5O4 cathodes cycled in blank electrolyte (c) and AMSL containing electrolyte (d).

electrolyte with and without AMSL was investigated to further elucidate the difference in cycling performance. As shown in Fig. 4, the main exothermic peaks are related to the violently reaction between cathode and electrolyte [37,38]. Heat generation initiated by this interfacial reaction is clearly reduced by introducing AMSL to the blank electrolyte: DH ¼ 209 J g1 in the absence of AMSL and DH ¼ 149 J g1 in the presence of AMSL. Moreover, the onset temperature of the first exothermic peak is increased by about 16  C with AMSL addition. These results suggest that AMSL derived SEI film is robust enough to suppress the undesirable interfacial reaction even under such harsh conditions and thus promise the

enhanced thermal stability. Transition metal (Mn and Ni) ions dissolution from LiNi0.5Mn1.5O4, which is usually provoked by HF, is another crucial factor responsible for the capacity fading of cells at high temperature. The dissolved Mn2þ and Ni2þ migrates to the anode and consequently can be reduced to metallic Mn and Ni, leading to consumption of cyclable Liþ and destruction of cathode. Mn and Ni deposition amount on the Li anode after 100 cycles at 55  C was measured by ICP-AES and listed in Table 3. Fewer amounts of Mn and Ni ions was observed on the Li foil cycled in AMSL containing electrolyte, when compared to that of the one cycled in blank electrolyte. Therefore, the AMSL derived SEI film effectively protects the LiNi0.5Mn1.5O4 surface from the attack of HF, resulting in retardation of Mn and Ni dissolution. 3.4. Surface analysis of the cycled LiNi0.5Mn1.5O4 cathode The SEM images of the fresh and cycled LiNi0.5Mn1.5O4 cathodes are presented in Fig. 5. The very clean and smooth surface of the fresh cathode can be seen in Fig. 5a. After cycling at high temperature in blank electrolyte, the surface of LiNi0.5Mn1.5O4 particle is covered with a rough film (Fig. 5b), which is due to that the decomposition products of the electrolyte continuously deposit at the unprotected cathode surface. In addition, this process usually accompanies the dissolution of transition metal ions into electrolyte, leading to structure detriment of the cathode. Differently, the

Table 3 The deposition amount of Mn and Ni ions on the Li anode after 100 cycles at 55  C in the electrolyte with and without AMSL.

Fig. 4. DSC profiles of interfacial exothermic reaction between the charged LiNi0.5Mn1.5O4 cathode and the electrolyte with and without AMSL.

Sample

Mn (ppm)

Ni (ppm)

Blank electrolyte With 0.5% AMSL

7.243 1.650

1.527 0.511

46

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

Fig. 5. SEM images of fresh LiNi0.5Mn1.5O4 cathode (a) and the cathodes after 100 cycles in blank electrolyte (b) and AMSL containing electrolyte (c).

cathode cycled in AMSL containing electrolyte exhibit a comparable clean surface to the fresh cathode in Fig. 5c. The compact and stable SEI film provided by AMSL inhibits direct contact of the electrolyte with LiNi0.5Mn1.5O4 cathode and thus ensures successfully cycling. The surface chemistry of LiNi0.5Mn1.5O4 cathodes after 100 cycles at 55  C in the electrolyte with and without AMSL was analyzed by XPS, and the results are shown in Fig. 6. The C1s spectrum of the fresh cathode shows three peaks which are CeC, CeH in graphite (284.6 eV) and CeF in PVDF binder (286 and 290.9 eV). Additional peaks at 286.1 and 288.7 eV associated with C]O and CeO containing species, which are generally originated from the decomposition of electrolyte components [11], arise in cycled cathodes.

The cathode with AMSL exhibits a stronger CeO peak but a less intense C]O peak compared to that without AMSL, demonstrating that the eOCH2CH]CH2 group in AMSL may contribute to the formation of the SEI film and this resulting film can effectively restrain the oxidative decomposition of carbonate solvents. In addition, the CeF peak is still observable on the cathode cycled with AMSL. This is because the AMSL derived SEI film is thinner than the one formed by the blank electrolyte. Metal-O peak in O1s spectra supports this point of view as well. In O1s spectra, the intensity of CeO and C]O peaks are weaker for the cathode with additive which is consistent with the less decomposition reactions of carbonate solvent occur on the cathode surface.

Fig. 6. XPS spectra of fresh LiNi0.5Mn1.5O4 cathode and the cathodes after 100 cycles in the electrolyte with and without AMSL.

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

47

is the appearance of SieC peak at 102.1 eV, which produced by the oxidation of AMSL. This suggests that AMSL is helpful to create a protective SEI film. Fig. 7 shows the ATR-FTIR spectra of fresh LiNi0.5Mn1.5O4 cathode and cathodes cycled with and without AMSL. The spectra of all samples are dominated by the peaks attributed to PVDF binder at 1377, 1160, 1055, 875 and 836 cm1 [40]. A strong peak at 1730 cm1 observed with the cathode after 100 cycles at 55  C in blank electrolyte are characteristic of poly (ethylene carbonate) (PEC) due to the EC oxidation [13,41]. Incorporation of AMSL results in a dramatic decrease in the intensity of PEC on the cathode cycled at elevated temperature. To clearly illustrate the difference between the cycled cathode with and without AMSL, we increased the concentration of AMSL to 5% and only carried out three formation cycles to highlight the impact of AMSL. The cathode with 5% AMSL exhibits new peak at 1278 cm1 assigned to eSi(CH3)3 [42,43], confirmed by the more intense peaks at 875 and 836 cm1 which are overlapped with the PVDF peaks. The more intense peak at 1055 cm1 is probably due to the eCH2eOeCH2e functional group derived from the decomposition of AMSL. The infrared spectroscopic results further confirm that the AMSL additive is preferentially oxidized to form a SEI film composed of organic silicon based species and ester moiety and effectively inhibits further electrolyte decomposition. 3.5. Compatibility of AMSL toward graphite anode

Fig. 7. ATR-FTIR spectra of fresh LiNi0.5Mn1.5O4 cathode and the cycled cathodes. The spectrum of the 3 times cycled cathode with 5% AMSL is plotted to highlight the impact of AMSL.

In F1s spectra of the cycled cathodes, there are two major peaks: the one at 687.9 eV corresponds to PVDF binder and the other is attributed to LiF generated from the reaction of HF with Liþ based compounds and thermal decomposition of LiPF6. The intensity of the LiF peak on the cathode cycled in blank electrolyte is remarkably stronger than that on the cathode cycled in AMSL containing electrolyte. Since LiF is insulated for electron and Liþ, it can not only increases the SEI film resistance but also hinder the charge transfer reaction between cathode and electrolyte [39]. This result is in good agreement with the EIS results as discussed above. Two peaks appeared at 134 and 135.9 eV in P2p spectra are characteristic of LixPOyFz and LixPFy, leading to the same result with respect to the F1s spectra. A noticeable feature for the cathode cycled with AMSL

The electrochemical behavior of graphite anode in blank and AMSL containing electrolytes was evaluated using cyclic voltammetry and charge/discharge test. Fig. 8a compares the voltammograms for the Li/graphite cells in two different electrolytes. For the cell cycled in blank electrolyte, the peak appeared at 0.69 V is known to be corresponding to the reduction of EC [44], which lead to the SEI film formation on graphite surface. Subsequently, a reversible Liþ intercalation and deintercalation process between 0.01 and 0.4 V is observed. With the addition of AMSL, electrochemical behavior does not change below 1 V except a small new cathodic irreversible peak appears at 1.47 V. This peak can be assigned to the reduction of AMSL, which is involved in the buildup of SEI film on the graphite surface. Consistent with the calculation result, AMSL is reduced prior to the main solvent of the electrolyte. The cyclic performances of Li/graphite cells with and without AMSL are presented in Fig. 8b. In the first three cycles, all the cells were subjected to a formation cycle at 0.1 C and then cycled at 0.5 C. When cycling at 0.5 C, the cell without AMSL delivers an initial discharge capacity of 350.5 mAh g1 and gradually increases to 379.1 mAh g1 after 100 cycles. In contrast, relatively high and stable discharge capacity is achieved in the cell with AMSL, maintaining 383.1 mAh g1 after 100 cycles. Since a large

Fig. 8. Cyclic voltammogram (a) and cycling performance (b) of graphite anodes in blank electrolyte and AMSL containing electrolyte.

48

J. Chen et al. / Journal of Power Sources 303 (2016) 41e48

quantity of acetylene carbon black is used in the graphite electrode, it delivers larger capacity than the calculated capacity of graphite. Coulombic efficiency is maintained close to 100% over 100 cycles irrespective of the introduction of AMSL. These results indicate that AMSL has good compatibility with graphite anode during the electrochemical process. But more detail effects bought by this additive require further investigation. 4. Conclusion We demonstrated that AMSL, which was screened out through DFT calculation, is a promising additive to enhance the interfacial stability between LiNi0.5Mn1.5O4 cathode and electrolyte and lead to improved cycling stability and rate capability of LiNi0.5Mn1.5O4. By incorporating 0.5% AMSL into electrolyte, Li/LiNi0.5Mn1.5O4 cell showed an enhanced discharge capacity retention of 80.2% after 500 cycles at room temperature. At elevated temperature, dramatic improvement in cycling performance was achieved, elevating the discharge capacity retention to 92.5% from 52.4%. Moreover, the discharge capacity at high C rate of 3 C was improved from 76.8 mAh g1 to 95.6 mAh g1 by the incorporation of additive. EIS, DSC and surface composition analyses clearly revealed that less resistive and high thermal stable SEI film was formed on the surface of LiNi0.5Mn1.5O4 due to the addition of AMSL. The other notable contribution of this SEI film was effective prevention of direct contact of the LiNi0.5Mn1.5O4 surface to electrolyte, which indicated that it is essential to suppress the undesired electrolyte decomposition and the dissolution of transition metal ions. This additive also showed positive effect toward graphite anode, delivering a slightly higher discharge capacity than the one without additive. We believe that AMSL is a promising additive candidate that can be used for both cathode and anode materials in advanced Li-ion batteries. Acknowledgments This work was financially supported by Scientific and Technological Research and Development Foundation of Shenzhen City (JCYJ20140418193546111), 973 National Defense Project (613142020201) and National Natural Science Foundation (51574166). References [1] M. Armand, J.M. Tarascon, Nature 451 (2008) 652e657. [2] Y.-K. Sun, S.-T. Myung, B.-C. Park, J. Prakash, I. Belharouak, K. Amine, Nat. Mater. 8 (2009) 320e324. [3] B. Scrosati, J. Hassoun, Y.K. Sun, Energy Environ. Sci. 4 (2011) 3287e3295. [4] N.-S. Choi, Z. Chen, S.A. Freunberger, X. Ji, Y.-K. Sun, K. Amine, G. Yushin, L.F. Nazar, J. Cho, P.G. Bruce, Angew. Chem. Int. Ed. 51 (2012) 9994e10024. [5] J.B. Goodenough, Acc. Chem. Res. 46 (2013) 1053e1061. [6] M.M. Thackeray, C. Wolverton, E.D. Isaacs, Energy Environ. Sci. 5 (2012) 7854e7863. [7] J.B. Goodenough, Y. Kim, Chem. Mater. 22 (2010) 587e603. [8] D. Liu, W. Zhu, J. Trottier, C. Gagnon, F. Barray, A. Guerfi, A. Mauger, H. Groult, C.M. Julien, J.B. Goodenough, K. Zaghib, RSC Adv. 4 (2014) 154e167. [9] M. Hu, X. Pang, Z. Zhou, J. Power Sources 237 (2013) 229e242.

[10] H. Duncan, D. Duguay, Y. Abu-Lebdeh, I.J. Davidson, J. Electrochem. Soc. 158 (2011) A537eA545. [11] D. Lu, M. Xu, L. Zhou, A. Garsuch, B.L. Lucht, J. Electrochem. Soc. 160 (2013) A3138eA3143. [12] S.F. Lux, I.T. Lucas, E. Pollak, S. Passerini, M. Winter, R. Kostecki, Electrochem. Commun. 14 (2012) 47e50. [13] N.P.W. Pieczonka, L. Yang, M.P. Balogh, B.R. Powell, K. Chemelewski, A. Manthiram, S.A. Krachkovskiy, G.R. Goward, M. Liu, J.-H. Kim, J. Phys. Chem. C 117 (2013) 22603e22612. [14] J. Mun, T. Yim, K. Park, J.H. Ryu, Y.G. Kim, S.M. Oh, J. Electrochem. Soc. 158 (2011) A453eA457. [15] J. Kalhoff, D. Bresser, M. Bolloli, F. Alloin, J.Y. Sanchez, S. Passerini, ChemSusChem 7 (2014) 2939e2946. [16] L. Hu, K. Amine, Z. Zhang, Electrochem. Commun. 44 (2014) 34e37. [17] L. Hu, Z. Zhang, K. Amine, Electrochem. Commun. 35 (2013) 76e79. [18] A. Abouimrane, I. Belharouak, K. Amine, Electrochem. Commun. 11 (2009) 1073e1076. [19] H. Duncan, N. Salem, Y. Abu-Lebdeh, J. Electrochem. Soc. 160 (2013) A838eA848. [20] Z. Zhang, L. Hu, H. Wu, W. Weng, M. Koh, P.C. Redfern, L.A. Curtiss, K. Amine, Energy Environ. Sci. 6 (2013) 1806e1810. [21] Q. Zhang, J.J. Chen, X.Y. Wang, C. Yang, M.S. Zheng, Q.F. Dong, Phys. Chem. Chem. Phys. 17 (2015) 10353e10357. [22] N.-S. Choi, J.-G. Han, S.-Y. Ha, I. Park, C.-K. Back, RSC Adv. 5 (2015) 2732e2748. [23] G. Yan, X. Li, Z. Wang, H. Guo, X. Xiong, J. Power Sources 263 (2014) 231e238. [24] H. Rong, M. Xu, L. Xing, W. Li, J. Power Sources 261 (2014) 148e155. [25] Y.-M. Song, J.-G. Han, S. Park, K.T. Lee, N.-S. Choi, J. Mater. Chem. A 2 (2014) 9506e9513. [26] W. Huang, L. Xing, R. Zhang, X. Wang, W. Li, J. Power Sources 293 (2015) 71e77. [27] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision C.01, Gaussian, Inc., Wallingford CT, 2010. [28] N. Shao, X.G. Sun, S. Dai, D.E. Jiang, J. Phys. Chem. B 115 (2011) 12120e12125. [29] M.D. Halls, K. Tasaki, J. Power Sources 195 (2010) 1472e1478. [30] L.D. Xing, C.Y. Wang, M.Q. Xu, W.S. Li, Z.P. Cai, J. Power Sources 189 (2009) 689e692. [31] Y.-K. Han, Y. Moon, K. Lee, Y.S. Huh, Curr. Appl. Phys. 14 (2014) 897e900. [32] E.G. Leggesse, J.C. Jiang, J. Phys. Chem. A 116 (2012) 11025e11033. [33] Y.-K. Han, K. Lee, S.C. Jung, Y.S. Huh, Comput. Theor. Chem. 1031 (2014) 64e68. [34] K. Xu, A. von Cresce, U. Lee, Langmuir 26 (2010) 11538e11543. [35] Y.-M. Lee, K.-M. Nam, E.-H. Hwang, Y.-G. Kwon, D.-H. Kang, S.-S. Kim, S.W. Song, J. Phys. Chem. C 118 (2014) 10631e10639. [36] J. Zheng, J. Xiao, M. Gu, P. Zuo, C. Wang, J.-G. Zhang, J. Power Sources 250 (2014) 313e318. [37] Y.-S. Kim, T.-H. Kim, H. Lee, H.-K. Song, Energy Environ. Sci. 4 (2011) 4038e4045. [38] Z.D. Li, Y.C. Zhang, H.F. Xiang, X.H. Ma, Q.F. Yuan, Q.S. Wang, C.H. Chen, J. Power Sources 240 (2013) 471e475. [39] W.-K. Shin, J.-H. Yoo, D.-W. Kim, J. Power Sources 279 (2015) 737e744. [40] S. Dalavi, M. Xu, B. Knight, B.L. Lucht, Electrochem. Solid-State Lett. 15 (2011) A28eA31. [41] M. Xu, N. Tsiouvaras, A. Garsuch, H.A. Gasteiger, B.L. Lucht, J. Phys. Chem. C 118 (2014) 7363e7368. [42] M. Urbanova, J. Pola, J. Anal. Appl. Pyrolysis 62 (2002) 197e203. [43] B. Narayanan, R. Kumar, P.D. Foo, Microelectron. J. 33 (2002) 971e974. [44] H.-C. Lee, S.-D. Choi, S.-H. Choi, H.-J. Kim, Y.-S. Choi, S.-J. Yoon, J.-J. Cho, Electrochem. Commun. 9 (2007) 801e806.