Journal of Power Sources 328 (2016) 397e404
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Novel choline-based ionic liquids as safe electrolytes for high-voltage lithium-ion batteries Tianqiao Yong a, b, Lingzhi Zhang a, *, Jinglun Wang a, Yongjin Mai a, Xiaodan Yan a, Xinyue Zhao a a
CAS Key Laboratory of Renewable Energy, Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou 510640, China State Key Laboratory of Applied Microbiology Southern China and Guangdong Provincial Key Laboratory of Microbial Culture Collection and Application, Guangdong Institute of Microbiology, Guangzhou 510663, China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Novel choline-based ionic liquids synthesized as safe electrolytes for Li-ion batteries. Delivered a stable capacity of 152 mAh g1 over 90 cycles at a cutoff voltage of 4.4 V. Displayed a lower propagation rate than the commercial carbonate electrolyte.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 31 March 2016 Received in revised form 13 June 2016 Accepted 9 August 2016
Three choline-based ionic liquids functionalized with trimethylsilyl, allyl, and cynoethyl groups are synthesized in an inexpensive route as safe electrolytes for high-voltage lithium-ion batteries. The thermal stabilities, viscosities, conductivities, and electrochemical windows of these ILs are reported. Hybrid electrolytes were formulated by doping with 0.6 M LiPF6/0.4 M lithium oxalydiﬂuoroborate (LiODFB) as salts and dimethyl carbonate (DMC) as co-solvent. By using 0.6 M LiPF6/0.4 M LiODFB trimethylsilylated choline-based IL (SN1IL-TFSI)/DMC as electrolyte, LiCoO2/graphite full cell showed excellent cycling performance with a capacity of 152 mAh g-1 and 99% capacity retention over 90 cycles at a cut-off voltage of 4.4 V. The propagation rate of SN1IL-TFSI)/DMC electrolyte is only one quarter of the commercial electrolyte (1 M LiPF6 EC/DEC/DMC, v/v/v ¼ 1/1/1), suggesting a better safety feature. © 2016 Elsevier B.V. All rights reserved.
Keywords: Ionic liquid Choline High-voltage Electrolyte Lithium-ion battery
1. Introduction Ionic liquids (ILs) have a unique suite of properties such as low volatility, and high electrochemical and thermal stabilities, which
* Corresponding author. CAS Key Laboratory of Renewable Energy, Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou 510640, China. E-mail address: [email protected]
(L. Zhang). http://dx.doi.org/10.1016/j.jpowsour.2016.08.044 0378-7753/© 2016 Elsevier B.V. All rights reserved.
makes them as important candidates for a number of energy storage and conversion related applications, such as super-capacitors, dye sensitised solar cells and batteries. Advantages that they offer for advanced lithium-ion batteries (LIBs) to replace the anodic limited and high ﬂammable organic-carbonate electrolytes are their low volatility and non-ﬂammability, making it possible to formulate electrolytes with enhanced safety without compromise of cyclic stability . Based on the types of cations, ILs could be primarily categorised into sulfonium, imidozonium, phosphonium
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
and ammonium [ 2]. Amongst them, quaternary ammonium ILs are widely investigated as electrolytes for LIBs because of their high cathodic stablility. For instance, cyclic pyrrolidinium and piperidinium bis(triﬂuoromethylsulfonyl)imide (TFSI) ILs are typically less viscous and cathodically stable than ILs with aliphatic cations, thus show highly efﬁcient lithium cycling [3,4]. While the lack of the ability of forming a stable solid electrolyte interface (SEI) on graphitic anode before lithium intercalation can be effectively overcame by the incorporation of ethylene carbonate (EC)  or other well-known additives such as vinyl carbonate [6,7], ﬂuoroethylene carbonate , lithium bis(oxalato) borate  and lithium oxalyldiﬂuoroborate (LiODFB) , the design and synthesis of novel ILs with low viscosity, advanced electrochemical performance and especially low cost is still a challenge. Choline chloride is an ammonium based ionic liquid produced in industry scale with competitive price, and is environmentally benign and biocompatible [11,12]. Thus choline chloride and its derivatives are vastly used in foodstuff [13,14], biodegradation , catalysis , electrochemical synthesis  and surfactants . However, choline chloride ionic liquid cannot be directly used as electrolyte in lithium-ion batteries, due to its poor compatibility with highly reductive anodes and oxidative cathodes because of its hydroxyl group and chloride ion. Additional modiﬁcation, therefore, is necessary to use it as electrolyte in LIBs. Fortunately, the electrochemical compatibility of ILs could be tuned by integrating various groups [19,20], e.g. organosilicon, double bond and cyano groups. Besides that, organosilicon offers good biocompatibility , double bond SEI ﬁlm forming capability  and cyano group protective effect for aluminum collector . Meanwhile, different anions  have been used to functionalize ILs, such as BF 4 , B(CN)-4 , PF 6 , ODFB , bis(ﬂuorosulfonyl)imide  and bis(triﬂuoromethylsulfonyl)imide (TFSI) . It is demonstrated that TFSI anion exhibits superior stability against oxidation and reduction, which offers wide electrochemical window . In this paper, we report the synthesis of three new cholinebased ionic liquids prepared by substituting hydroxyl group with trimethylsilyl, allyl or cyanoethyl groups and exchanging chlorine with TFSI anions, aiming to combine the inexpensive features of choline ILs and the high-voltage feature of these functional moieties (Scheme 1). The thermal stabilities, viscosities, ionic conductivities and electrochemical windows are characterized. Using hybrid electrolytes formulated by doping the ionic liquids with LiPF6/LiODFB as salts and taking dimethyl carbonate (DMC) as cosolvent, high-voltage LiCoO2 (LCO)/graphite full cells are tested with an upper cut off voltage of 4.4 V. And the effect of this IL on electrochemical properties is analysed in detail by electrochemical impedance spectroscopy (EIS), Fourier transform infrared (FTIR) and scanning electron micrograph (SEM). The ﬂammability of these
electrolytes is characterized by measuring the ﬂame propagation rate and compared with a commercial carbonate electrolyte. 2. Experimental 2.1. Materials Choline chloride, hydroxylethyldimethylamine, methyl iodide, acrylonitrile, acetonitrile, allyl bromide, sodium hydroxide, Nmethyl-2-pyrrolidone (NMP) and hexamethylsilazane were purchased from Aladdin Reagent Co. (China). Lithium bis(triﬂuoromethylsulfonyl)imide (LiTFSI) was bought from Acros Organics Co. (USA). DMC, LiPF6, LiODFB and a conventional electrolyte of 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) were donated by Dongguan Shanshan Battery Materials Co. (China) and used as received. Artiﬁcial graphite powder and high-voltage LCO powder were offered by Amperex Technology Co. (China). Celgard 2400 (Celgard, USA) microporous polypropylene membrane was used as separators. 2.2. Apparatus 1 H NMR, 13C NMR and 29Si NMR spectra were taken on a Bruker avance 400 spectrophotometer. The water content of the synthesized ionic liquids was less than 10 ppm, determined by Karl-Fisher coulometric moisture titrator (831 KF, Metrohm Co., Sweden). Viscosity (h) measurements were performed on SPb-2 Viscometer (Nirun Intelligent Technology Co., China). Thermal gravimetric analysis (TGA) measurements were conducted on a STA409C/PCPFEIEFFER VACUUMTGA-7 analyzer (NERZSCH-Gertebau GmbH, Germany) in an Ar atmosphere with a ﬂow rate of 30 mL min1 from 40 C to 450 C at a heating rate of 10 C/min. Variable temperature conductivity measurements of the prepared ionic liquids were conducted using a conductivity meter (model DDS-310, Dapu Instru. Co., China), utilizing an oil bath to control temperatures between 0 C and 90 C. Linear scanning voltammetry (LSV) experiments using an electrochemical workstation (BAS-ZAHNER IM6, Germany) were performed in a custom-made three-electrode cell with a platinum as working electrode, a lithium metal as reference and a lithium metal counter electrode at a scan rate of 5 mV s1 between 0.5e6.7 V. Electrochemical impedance spectroscopy (EIS) measurements were conducted on a electrochemical workstation (BAS-ZAHNER IM6, Germany) by applying an alternative voltage of 5 mV over the frequency ranging from 0.01 to 105 Hz. Fourier transform infrared spectroscopy (FTIR) was recorded on a TENSOR 27 spectrometer (Bruker, Germany) from 4000 to 400 cm1 at a resolution of 4 cm1. The morphology of the electrodes was observed by a scanning electron microscopy (Hitachi S-
Scheme 1. Synthesis routes of the functionalized choline ILs: (a) SN1IL-TFSI, (b) AN1IL-TFSI and (c) CEN1IL-TFSI.
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
4800, Japan). The ﬂame propagation rate tests were conducted based on a modiﬁed condition of UL Flame RatingdUL94 . The coin cells (CR2025) were assembled to test electrochemical performance of the prepared electrolytes and the conventional base electrolyte for comparison. The cathodes were prepared by mixing the LiCoO2 powder with carbon black and polyvinylidene ﬂuoride (PVdF) at a weight ratio of 90:5:5 in NMP to form homogeneous slurries. The anodes were prepared by slurrying the graphite powder with carbon black and PVdF at a weight ratio of 90:5:5 in an NMP solution. The slurries were coated onto an aluminum foil for cathodes or copper foil for anodes, followed by drying at 75 C under vacuum for 12 h. The load of active material is about 2 and 5 mg cm2 for graphite electrode and LCO electrode, respectively. All the cells were assembled in a dry argon ﬁlled glove-box (Mikrouna, H2O and O2 < 1 ppm). The cycling performance tests were conducted on a multi-channel battery test system (NEWARE BTS-610, China). The LiCoO2/graphite full cells were galvanostatically charged and discharged with the current density of C/5 (1C ¼ 170 mA/g) in a voltage range of 2.7e4.4 V after three formation cycles (typically between 2.7 and 4.4 V at C/10).
neutralized by hydrochloric acid. The pure product was obtained by fractional distillation. Yield: 89%. 1H NMR (300 MHz, CDCl3): s 2.25 (s, 6H, -N(CH3)2), 2.48 (m, 2H, -NCH2CH2O), 2.59 (m, 2H, -NCH2CH2O), 3.57 (m, 2H, CH2eCH2eOCH2), 3.66 (m, 2H, -OCH2CH2CN); 13C NMR (75 MHz, CDCl3): s 18.77, 45.80, 58.66, 66.66, 69.44, 117.85. Synthesis of (2-cyanoethoxyethyl)trimethylammonium bis(triﬂuoromethanesulfon-yl)imide (CEN1IL-TFSI): To a suspension of CEN1 (26.76 g, 0.188 mol) in 200 mL ethyl ether was added methane iodide (29.37 g, 0.207 mol) dropwise. The resulting white solid (CEN1IL-I) was washed with ethyl ether repeatedly and then dried under vacuum. The metathesis of I by TFSI for the white solid obtained was performed in water. The product was extracted by dichloromethane. The solvent was removed in vacuo at 60 C for 24 h. Yield: 84%. 1H NMR (300 MHz, CDCl3): s 3.23 (m, 9H, þN(CH3)3), 2.74 (m, 2H, -NCH2CH2O), 3.60 (m, 2H, -NCH2CH2O), 3.75 (m, 2H, CH2eCH2eOCH2), 3.96 (m, 2H, -OCH2CH2CN); 13C NMR (75 MHz, CDCl3): s 18.18, 54.37, 58.46, 64.44, 65.73, 118.20, 121.34.
3.1. Synthesis and characterization
Synthesis of (2-hydroxyethyl)trimethylammonium bis(triﬂuoromethanesulfonyl)im-ide (N1IL-TFSI): To a solution of choline chloride (20 g, 143 mmol) in 50 mL deionized water was added LiTFSI (41 g, 143 mmol). The mixture was stirred for 4 h. The product was extracted by dichloromethane. Yield: 83%. 1H NMR (300 MHz, CDCl3): d 3.16 (s, 9H, þN(CH3)3), 3.40 (s, 1H, CH2OH), 3.45 (s, 2H, CH2OH), 4.03 (s, 2H, CH2Nþ); 13C NMR (75 MHz, CDCl3): 54.06, 56.21, 67.66, 119.75. Synthesis of (2-trimethylsililoxyethyl)trimethylammonium bis(triﬂuoromethanesulf-onyl)imide (SN1IL-TFSI): To N1IL-TFSI (40 g, 103.5 mmol) was added hexamethyldisilazane (16.7 g, 103.5 mmol) dropwise at 0 C. The mixture was heated to reﬂux for 16 h. After removing the residual hexamethyldisilazane through distillation, the pure product was obtained as colourless liquid. Yield: 99%. 1H NMR (300 MHz, CDCl3): d 0.16 (s, 9H, Si(CH3)3), 3.22 (s, 9H, þN(CH3)3), 3.50 (s, 2H, CH2O), 4.00 (s, 2H, CH2Nþ); 13C NMR (75 MHz, CDCl3): 1.04, 54.55, 56.81, 67.86, 119.87. Synthesis of (2-allyloxyethyl)trimethylammonium chloride (AN1IL-Cl): To a suspension of choline chloride (56.4 g, 403 mmol), sodium hydroxide (32.31 g, 806 mmol) in 100 mL acetonitrile was added allyl bromide (78 g, 645 mmol) dropwise at 0 C. The mixture was heated to reﬂux for 16 h. The solid salt was removed by ﬁltration. After removing solvent, the pure product was obtained by recrystallization in acetonitrile/ethyl ether. Yield: 87%. 1H NMR (300 MHz, CDCl3): s 3.47 (m, 9H, þN(CH3)3), 3.90, 3.94 (dd, 4H, OCH2CH2O), 4.02 (m, 2H, CH2]CHeCH2eO), 5.23 (ddq, 2H, CH2] CHeCH2eO), 5.84 (ddt, 1H, CH2]CHeCH2eO); 13C NMR (75 MHz, CDCl3): s 54.61, 63.98, 65.68, 72.21, 118.43, 133.27. Synthesis of (2-allyloxyethyl)trimethylammonium bis(triﬂuoromethanesulfonyl)imi-de (AN1IL-TFSI): To AN1IL-Cl (23.28 g, 123 mmol) in 50 mL deionized water was added LiTFSI (35.85 g, 123 mmol). The mixture was stirred for 4 h. The product was extracted by dichloromethane. The solvent was removed in vacuo at 60 C for 24 h. Yield: 87%. 1H NMR (300 MHz, CDCl3): s 3.19 (m, 9H, þN(CH3)3), 3.58, 3.86 (m, 4H, OCH2CH2O), 4.05 (m, 2H, CH2]CHeCH2eO), 5.28 (ddq, 2H, CH2]CHeCH2eO), 5.85 (m, 1H, CH2]CHeCH2eO); 13C NMR (75 MHz, CDCl3): s 54.65, 63.50, 66.20, 72.31, 118.70, 132.97. Synthesis of cyanoethoxyethyldimethylamine (CEN1): To a suspension of hydroxylethyldimethylamine (89.14 g, 1 mol) in NaOH solution (0.4 g NaOH, 1 mL water) was added acrylonitrile (58.3 g, 1.1 mol) dropwise. The mixture was stirred for 16 h and then
The choline-based ILs were prepared via a straightforward twostep method consisting of an anion exchange reaction of chlorine ion with LiTFSI and an functionalization reaction by reaction with hexamethylsilazane, allyl bromid, and acrylonitrile (Scheme 1). For SN1IL-TFSI, anion exchange of Cl with TFSI was carried out before trimethylsilylation in order to avoid the hydrolysis of SieO bond . In the case of AN1IL-TFSI, it was synthesized through nucleophilic substitution reaction of allyl bromide with choline chloride in the presence of sodium hydroxide and a subsequent anion exchange reaction of AN1IL-Cl with LiTFSI due to the formed CeO bond for AN1IL-Cl is stable and resistant to hydrolysis. Earlier, it is reported that Michael addition of alcohol compounds with acrylonitrile is usually performed in the presence of strong bases directly . In our case for CN1IL-TFSI, however, we found that it is difﬁcult to separate the obtained nitrile group functionalized choline chloride from the reaction system because of partial polymerization of acrylonitrile . Therefore, CN1IL-TFSI was synthesized through a Michael addition of hydroxylethyldimethylamine with acrylonitrile to afford intermediate CEN1, followed by a quarterization with iodomethane and an anion exchange reaction of CEN1IL-I with LiTFSI. The chemical structures of the synthesized ILs have been characterized by 1H, 13C, and 29Si NMR.
3. Results and discussion
3.2. Physiochemical properties As described in the literature, the viscosity of ILs has been found to have considerable inﬂuence on their performances as electrochemical electrolytes; for instance it partly deﬁnes the conductivity of the ILs. Therefore, we measured the viscosities of all the synthesized ILs (Table 1). Generally, viscosity is affected dominantly by cation-anion interaction, ion size and ﬂexibility of ions . Among the synthesized ILs, CEN1IL-TFSI has the highest viscosity due to the strong interaction of nitrile-containing cation with TFSI anion [1,2]. SN1IL-TFSI has a bulky trimethylsiloxy group and therefore it has a large viscosity too. Surprisingly, the viscosity of AN1IL-TFSI is 36 cp, signiﬁcantly lower than the other two with the similar chemical structures, possibly resulted from the small size and stereo hindrance of allyloxy group. The values of anodic and cathodic limit of the synthesized ILs were determined by linear scanning voltammetry (LSV) at room temperature (Fig. 1). The electrochemical windows calculated from
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
Table 1 Physical data, conductivity and VTF parameters for functionalized choline ILs. Ionic liquids
SN1IL-TFSI AN1IL-TFSI CEN1IL-TFSI P(2o1)2-TFSIb
1.35 4.29 1.10 2.29
125.4 36.0 140.5 55.0
336 325 330 379
0.00 0.65 1.30 0.40
5.30 5.40 5.78 5.40
5.30 4.75 4.48 5.00
251.2 279.2 332.7 146
84.66 100.4 78.32 201
4.76 4.37 4.90 3.36
Conductivities were measured with pure functionalized choline ILs at 25 C; EW: electrochemical window. Ref. .
them are listed in Table 1. All three ILs oxidatively decompose at potential higher than 5.3 V (vs. Li/Liþ), making them potentially suitable for high-voltage lithium-ion batteries. Compared with the oxidative limit of 5.3 V for SN1IL-TFSI and 5.4 V for AN1IL-TFSI, the CN substitution increases the oxidative limit up to 5.8 V for CEN1ILTFSI. SN1IL-TFSI and AN1IL-TFSI degrade reductively at 0 V and 0.65 V, respectively. The reductive potential for CEN1IL-TFSI shifted to a much higher potential of 1.3 V, which can be attributed to the introduction of reductive -CN group . The ionic conductivities of the synthesized ILs at room temperature are listed in Table 1. Neat ILs have conductivities in the range of 1.10e4.29 mS/cm. Generally, the ionic conductivity of the ionic liquid is usually affected by factors such as viscosity, molecular weight, density and ion size. But it seems that the viscosity dominates the conductivities of these ILs. CEN1IL-TFSI shows the lowest ionic conductivity of 1.10 mS/cm while SN1IL-TFSI the modest of 1.35 mS/cm. Surprisingly, AN1IL-TFSI has the highest of 4.29 mS/cm which is higher than those ammonium-TFSI ILs reported in literature. The temperature dependence of conductivity for the neat ILs is plotted in Fig. 2a over the temperature range from 0 to 85 C. For each IL, the conductivity increases with increasing temperature, due to the fast ion motion with decreased viscosity at the elevated temperatures. The conductivity of AN1IL-TFSI shows a more pronounced increase with temperature than that for both SN1IL-TFSI and CEN1IL-TFSI. The conductivity of the functionalized choline ILs can be described by the empirical VTF equation :
sT1=2 ¼ s0 e½B=kðTT0 Þ
where s0 (S cm1) is a constant, B (K) is the pseudo-activation energy, and T0 (K) is the vanishing ion mobility temperature. The parameters s0, T0, and Ea (B ¼ Ea/R, ideal gas constant R ¼ 8.31 J K1 mol1) are then obtained (Table 1). According to the VTF equation, ln(sT1/2) should have a linear relationship with (TT0)1. Using the T0 values in Table 1, the ln(sT1/2) versus (T-T0)1 can
Fig. 1. Linear sweep voltammograms of the functionalized choline ILs: (a) SN1IL-TFSI, (b) AN1IL-TFSI and (c) CEN1IL-TFSI.
be established and ﬁt well linearly (Fig. 2b). Also, the value of Ea, reﬂecting the ease of ion motion in ILs, follows the order of AN1ILTFSI > SN1IL-TFSI > CEN1-TFSI, agreeing with the trend of viscosity for the ILs. The prefactor s0 relates to the number of mobile charge carriers in the ILs. CEN1IL-TFSI has the highest s0 of 332.7 mS/cm due to its enhanced ion dissociation by the polar -CN group; while SN1IL-TFSI has the lowest s0 of 251.2 mS/cm due to that the trimethylsilyl group reduces the ion complexing ability of the neighbouring oxygen atom. The thermal properties of the ILs were studied by TGA (Fig. S1). The values of the thermal decomposing temperature (Tde) are summarized in Table 1. The prepared ILs are stable before 300 C and start to decompose at 325e336 C which is slightly lower than that for cyclic ammonium-TFSI ILs (e.g. 375 C for P(2o1)2-TFSI ). This suggests that the aliphatic ammonium ILs are less thermally stable than the cyclic ammonium ionic liquids. Tde follows the order of AN1IL-TFSI (325 C) < CEN1IL-TFSI (330 C) < SN1IL-TFSI (336 C) which is consistent with the previous report in literature [38,39]. 3.3. Cell performance of SN1IL as co-solvent The above measurements show that our ILs have high oxidative potential and thermal stability, but still their viscosities (36e140 cp) are relatively higher than that of the conventional organic carbonate solvents (e.g. 2.5 cp for propylene carbonate). This would lead to poor wetting problem and poor rate capability if the neat ILs are directly used for electrolyte. Fortunately, previous reports demonstrate that a viable strategy is to mix ILs with organic carbonate solvents to form hybrid electrolyte. Such hybrid electrolytes show good safety without compromising electrochemical performance . Therefore, we selected DMC as dilute solvent for study. On the other hand, LCO will be a more attractive cathode if it is charged to 4.4 V because it will increase approximately 30% extra energy capacity than that of 4.2 V . But the high oxidative surface of delithiated LCO initiates side reactions of electrolytes at interface, resulting in poor cyclic stability . Therefore, exploring electrolytes capable of suppressing such side reactions for 4.4 V LCO is signiﬁcant and industrial promising. Given that SN1IL-TFSI is a compromise of stable cathodic and anodic stabilities, we chose it to mix with DMC (v/v ¼ 1:1). LiODFB was chosen to partially replace LiPF6 because it provides an effective passivation ﬁlm on carbon anode. Thus, hybrid electrolytes (denoted as SN1IL-DMC electrolyte) formulated with 0.6 M LiPF6 and 0.4 M LiODFB are applied to 4.4 V LCO/graphite systems. Fig. 3a compares the cycling performance of LCO/graphite full cell using SN1IL-DMC electrolyte and 1 M LiPF6 EC/DEC/DMC (v/v/ v ¼ 1:1:1) at 0.2 C, respectively. The cell using SN1IL-DMC electrolyte delivers a capacity of 152 mAh g1 and shows excellent cycling performance with capacity retention of 99% over 90 cycles; while the cell with the conventional electrolyte, 1 M LiPF6 EC/DEC/ DEC, fades rapidly with capacity retention of only 25% after 90 cycles. This excellent cyclic stability at high-voltage for SN1IL-DMC electrolyte may be attributed to the high anodic limit and anti-
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
Fig. 2. (a) Ionic conductivity at variable temperature and (b) VTF curves of neat functionalized choline ILs.
Fig. 3. (a) Cycling performance of LCO/graphite cells with SN1IL-DMC electrolyte and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte, respectively; (b) rate performance of LCO/ graphite cell with SN1IL-DMC electrolyte.
oxidation capability of SN1IL-TFSI. Moreover, the cell with SN1ILDMC electrolyte also exhibits good rate performance, retaining 72% capacity of 0.2C at 2 C (Fig. 3b). And when the cell cycles at 0.2 C after 2 C test, it fully recovered its capacity to 0.2 C level of 152 mAh g1. The propagation rate of SN1IL-DMC electrolytes with various ratios of SN1IL-TFSI/DMC was plotted and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) in Fig. 4. The conventional electrolyte shows the highest propagation rate of 32 mm S1. For the SN1IL based hybrid electrolytes, the effect of adding IL is signiﬁcant even only with 25% IL addition. All IL-containing electrolytes show reduced ﬂame propagation rates. Speciﬁcally, the propagation rate is 25, 8 and 6 mm S1 when the content ratio of SN1IL-TFSI is 25%, 50% and 75% respectively. It is worth noticing that the propagation rate of SN1ILDMC (v/v ¼ 1/1) electrolyte is only one quarter of that of the
Fig. 4. Propagation rates from the ﬂammability test of 1 M LiPF6 EC/DEC/DMC (v/v/ v ¼ 1/1/1) electrolyte and SN1IL based hybrid electrolytes with various ratios of SN1ILTFSI and DMC.
conventional electrolyte (1 M LiPF6 EC/DEC/DMC, v/v/v ¼ 1/1/1). Fig. 5 displays the voltage-capacity curves between different cycles of the cells using SN1IL-DMC electrolyte and the reference electrolyte of 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1). It was found that all curves show the typical charge-discharge plateau of LiCoO2/ graphite cell at about 3.7/4.4 V. The cell with 1 M LiPF6 EC/DEC/DMC electrolyte shows a decreased capacity and increased voltage polarization from 5th to 10th cycle (Fig. 5a). By contrast, similar capacity change and voltage polarization growth are not observed for SN1IL-DMC electrolyte even after 90 cycles (Fig. 5b). The mean voltage value (the voltage at half time of the cells' running) reﬂects how long the cell can run under testing voltage. For 1 M LiPF6 EC/ DEC/DMC electrolyte, the mean voltage value increases gradually over cycling and end at about 4.1 V after 90 cycles. For SN1IL-DMC electrolyte; however, the mean voltage value remains a stable value of about 3.8 V over 90 cycles (Fig. S2). The above analysis about voltage-capacity curves and mean voltage changes between the two electrolytes suggests that the excellence of the cell with SN1ILDMC electrolyte may be resulted from the suppressed internal impedance growth over cycling, comparing with the reference electrolyte. This is well supported by the results of EIS, which compares the Nyquist plots of the cells with different electrolytes before cycling and after 90 cycles (full discharge) (Fig. 6). The semicircles at the high-frequency, medium-frequency region and the inclined line at an approximate 45 to the real axis in the Nyquist plots can be attributed to the SEI ﬁlm resistance, chargetransfer impedance and lithium diffusion process within electrode respectively. It is noted that the cells with the both electrolytes show almost the same semicircle diameter before cycling, suggesting they have almost the same cell impedance. For the 1 M LiPF6 EC/DEC/DMC electrolyte, the semicircle diameter increases four times after 90 cycles. In the case of SN1IL-DMC electrolyte, although the semicircle diameter also increases due to the formation of SEI, the impedance increase is much lower than that for the
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
Fig. 5. Voltage proﬁles of (a) 5th, 10th and 90th cycles of LCO/graphite cell using SN1IL- DMC electrolyte; (b) 5th and 10th cycles of LCO/graphite cell with 1 M LiPF6 EC/DEC/DMC (v/ v/v ¼ 1/1/1) electrolyte.
Fig. 6. Nyquist plots of LCO/graphite cells with SN1IL-DMC and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte before cycling (a) and after 90 cycles (b).
reference electrolyte. This indicates that the cell with SN1IL-DMC has smaller SEI ﬁlm resistance and charge-transfer impedance than that for 1 M LiPF6 EC/DEC/DMC. The morphologies of graphite before/after cycling in different electrolytes were observed by SEM (Fig. 7aec). While the pristine graphite shows clean and smooth surface before cycling (Fig. 7a), both the cycled graphite are covered by SEI ﬁlms on surface. But the SEI ﬁlm is more compact and smoother for the graphite anode with SN1IL-DMC when compared with that with the reference electrolyte (1 M LiPF6 EC/DEC/DMC); and the later SEI layer is ﬂuffy with tiny holes (Fig. 7bec). These observations suggest that SN1IL-DMC tend to form a more stable SEI ﬁlms on graphite. The morphologies
of LCO electrodes before/after cycling in different electrolytes were also observed by SEM (Fig. 7def). After cycling, LCO cathode with SN1IL-DMC electrolyte displays a surface as fresh and smooth as the pristine LCO (Fig. 7dee). As a comparison, LCO with the reference electrolyte displays a rough and coarse surface due to the decomposition of the electrolyte on high oxidative surface of the charged LCO at 4.4 V cut-off voltage as expected (Fig. 7f). FT-IR spectroscopy was used to characterize the species on the surface of both electrodes before or after LCO/graphite full cells cycled in different electrolytes. Fig. 8a presents the FT-IR spectra of the pristine graphite anode and the cycled with SN1IL-DMC and 1 M LiPF6 EC/DEC/DMC electrolyte. The characteristic peaks of
Fig. 7. SEM images of the pristine graphite anode (a), the cycled graphite anode after 90 cycles with SN1IL- DMC (b) and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte (c); the pristine LCO cathode (d), the cycled LCO cathode after 90 cycles with SN1IL- DMC (e) and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte (f).
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
Fig. 8. FTIR spectra of the pristine graphite and the cycled graphite with SN1IL- DMC and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte (a) and the pristine LCO and cycled LCO with SN1IL- DMC and 1 M LiPF6 EC/DEC/DMC (v/v/v ¼ 1/1/1) electrolyte (b).
PVdF, the stretching vibration of -CH3 at 2928 cm1, -CH2 at 2853 cm1 and -C-F at 1200 cm1, disappeared for the graphite anode cycled in 1 M LiPF6 EC/DEC/DMC electrolyte. Furthermore, the characteristic peaks from the decomposed electrolyte components, e.g. 1511 cm1 for carbonates, carboxylates and 515 cm1 for LiO2 , are clearly observed for graphite anode with 1 M LiPF6 EC/ DEC/DMC electrolyte. This demonstrates that the graphite anode is covered by the decomposed products of the reference electrolyte, thus the absorption peaks of PVdF binder disappeared. For the graphite anode with SN1IL-DMC, the characteristic peaks of PVdF and LiODFB (B-O at 596 cm1 and F-B-F at 1096 cm1) are clearly observed. Besides, the vibration of C-H in -Si(CH3)3 at 850 cm1 and the symmetric stretching vibration of Si-C in -Si(CH3)3 at 776 cm1 resulted from SN1IL-TFSI are also observed. Therefore, it can be concluded that both LiODFB and SN1IL-TFSI contribute to the formation of SEI layer on the surface of graphite anode. Fig. 8b presents the FT-IR spectra of the pristine LCO cathode and the cycled with SN1IL-DMC and 1 M LiPF6 EC/DEC/DMC electrolyte. The typical peaks of PVdF are not observed for the cycled LCO cathode with 1 M LiPF6 EC/DEC/DMC electrolyte. Furthermore, the characteristic peaks from the decomposed electrolyte components, e.g. 1547 and 1160 cm1 for carbonates and carboxylates , are clearly observed for LCO cathode with 1 M LiPF6 EC/DEC/DMC electrolyte. This demonstrates that the LCO cathode is covered by the decomposed products of the reference electrolyte, thus the absorption peaks of PVdF binder disappeared. For the LCO cathode with SN1IL-DMC, the vibration of -N-C at 867 cm1 and the vibration of C-H in -Si(CH3)3 at 837 cm1 resulted from SN1IL-TFSI are observed . This result suggests that SN1IL-TFSI can form a thin passivation ﬁlm on LCO cathode which can explain the better cycling stability for the LCO/graphite cell with SN1IL-DMC electrolyte cycled at 4.4 V.
4. Conclusions In summary, three choline-based ILs with trimethylsilyl, allyl, and cynoethyl group are designed and synthesized as safe electrolytes for high-voltage lithium-ion batteries through a straightforward two-step method consisting of an anion exchange reaction of chlorine ion with LiTFSI and an functionalization reaction by reaction with hexamethylsilazane, allyl bromid, and acrylonitrile. The IL with allyl group, AN1IL-TFSI, shows a high conductivity of 4.29 mS cm1; the IL with cynoethyl group, CEN1IL-TFSI, exhibits a high oxidative potential of 5.78 V versus Li/Liþ. LiCoO2/graphite full cell using SN1IL/DMC (v/v ¼ 1/1) doped with 0.6 M LiPF6/0.4 M LiODFB salts showed excellent cycling performance, delivering a capacity of 152 mAh g1 over 90 cycles at a cut-off voltage of 4.4 V,
and good rate capability, retaining 72% capacity of 0.2C at 2 C rate. The propagation rate of the electrolyte is only one quarter of the commercial reference electrolyte (1 M LiPF6 EC/DEC/DMC, v/v/ v ¼ 1/1/1), suggesting a better safety feature. EIS, FTIR and SEM results revealed that SN1IL-TFSI can form a passivation ﬁlm on LCO cathode which can explain the better cycling stability for the LCO/ graphite cell with SN1IL-DMC electrolyte cycled at 4.4 V. Acknowledgments This work was supported by National Natural Science Foundation of China (21573239), Guangdong Provincial Project for Science & Technology (2014TX 01N014/2014A050503050/ 2015B010135008), Guangzhou Municipal Project for Science & Technology (2014Y2-00219/201509010018), Dongguan Municipal Project for Science & Technology (2013509104210). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2016.08.044. References  D.R. MacFarlane, N. Tachikawa, M. Forsyth, J.M. Pringle, P.C. Howlett, G.D. Elliott, J.H. Davis, M. Watanabe, P. Simonf, C.A. Angellg, Energy Environ. Sci. 7 (2014) 232.  A. Lewandowski, A. Swiderska-Mocek, J. Power Sources 194 (2009) 601.  H. Li, J. Pang, Y. Yin, W. Zhuang, H. Wang, C. Zhai, S. Lu, RSC Adv. 3 (2013) 13907.  H. Sakaebe, H. Matsumoto, Electrochem. Commun. 5 (2003) 594.  L. Lombardo, S. Brutti, M.A. Navarra, S. Panero, P. Reale, J. Power Sources 227 (2013) 8.  Y. An, P. Zuo, C. Du, Y. Ma, X. Cheng, J. Lin, G. Yin, RSC Adv. 2 (2012) 4097.  M. Holzapfel, C. Jost, P. Novaka, Chem. Commun. 0 (2004) 2098.  I.A. Profatilova, N.-S. Choi, S.W. Roh, S.S. Kim, J. Power Sources 192 (2009) 636.  Y. An, P. Zuo, X. Cheng, L. Liao, G. Yin, Electrochim. Acta 56 (2011) 4841.  J. Xiang, F. Wu, R. Chen, L. Li, H. Yu, J. Power Sources 233 (2013) 115.  W.H. Meck, C.L. Williams, Devel. Brain Res. 118 (1999) 51.  S.H. Zeisel, Annu. Rev. Nutr. 26 (2006) 229.  F. Ilgen, D. Ott, D. Kralisch, C. Reil, A. Palmbergera, B. Konig, Green Chem. 11 (2009) 1948.  E.A. Drylie, D.S. Wragg, E.R. Parnham, P.S. Wheatley, A.M.Z. Slawin, J.E. Warren, R.E. Morris, Angew. Chem. Int. Ed. 46 (2007) 7839.  S. Sekar, M. Surianarayanan, V. Ranganathan, D.R. MacFarlane, A.B. Mandal, Environ. Sci. Technol. 46 (2012) 4902.  M. Avalos, R. Babiano, P. Cintas, J.L. Jimnez, J.C. Palacios, Angew. Chem. Int. Ed. 45 (2006) 3904.  C. Gu, J. Tu, Langmuir 27 (2011) 10132.  S. Song, P. Fu, Y. Wang, Z. Wang, Y. Qian, Colloids Surfaces A Physicochem. Eng. Asp. 399 (2012) 100.  T. Tsuda, K. Kondo, T. Tomioka, Y. Takahashi, H. Matsumoto, S. Kuwabata, C.L. Hussey, Angew. Chem. Int. Ed. 50 (2011) 1310.  A.J.R. Rennie, N. Sanchez-Ramirez, R.M. Torresi, P.J. Hall, J. Phys. Chem. Lett. 4 (2013) 2970.
T. Yong et al. / Journal of Power Sources 328 (2016) 397e404
 W. Weng, Z. Zhang, J. Lu, K. Amine, Chem. Commun. 47 (2011) 11969.  J. Reiter, M. Nadherna, Electrochim. Acta 71 (2012) 22.  M. Egashira, M.T. Nakagawa, I. Watanabe, S. Okada, J. Yamaki, J. Power Sources 160 (2006) 1387.  H. Nakagawa, S. Izuchi, K. Kuwana, T. Nukuda, Y. Aihara, J. Electrochem. Soc. 150 (2003) A695.  H. Yoon, G.H. Lane, Y. Shekibi, P.C. Howlett, M. Forsyth, A.S. Best, D.R. MacFarlane, Energy Environ. Sci. 6 (2013) 979.  S.-Y. Lee, H.H. Yong, Y.J. Lee, S.K. Kim, S. Ahn, J. Phys. Chem. B 109 (2005) 13663.  J.L. Allen, D.W. McOwen, S.A. Delp, E.T. Fox, J.S. Dickmann, S.-D. Han, Z.B. Zhou, T.R. Jow, W.A. Henderson, J. Power Sources 237 (2013) 104.  M. Ishikawa, T. Sugimoto, M. Kikuta, E. Ishiko, M. Kono, J. Power Sources 162 (2006) 658.  H. Sakaebe, H. Matsumoto, Electrochem. Commun. 5 (2003) 594.  K. Xu, Chem. Rev. 104 (2004) 4303.  L. Zhang, L. Lyons, J. Newhouse, Z. Zhang, M. Straughan, Z. Chen, K. Amine,
R.J. Hamers, R. West, J. Mater. Chem. 20 (2010) 8224.  L. Zhang, Z. Zhang, S. Harring, M. Straughan, R. Butorac, Z. Chen, L. Lyons, K. Amine, R. West, J. Mater. Chem. 18 (2008) 3713.  X. Li, S.H. Goh, Y.H. Lai, S.-M. Deng, J. Appl. Polym. Sci. 73 (1999) 2771.  O. Torre, I. Alfonsoa, V. Gotor, Chem. Commun. 0 (2004) 1724.  Y. Abu-Lebdeh, I. Davidson, J. Power Sources 189 (2009) 576.  M.L. Williams, R.F. Landel, J.D. Ferry, J. Am. Chem. Soc. 77 (1955) 3701.  S. Fang, Z. Zhang, Y. Jin, L. Yang, S. Hirano, K. Tachibana, S. Katayamad, J. Power Sources 196 (2011) 5637.  Z.-B. Zhou, H. Matsumoto, K. Tatsumi, Chem. Eur. J. 11 (2005) 752.  Y.-S. Kim, T.-H. Kim, H. Lee, H.-K. Song, Energy Environ. Sci. 4 (2011) 4038. , A. Vijh, K. Zaghib,  A. Guerﬁ, M. Dontigny, P. Charest, M. Petitclerc, M. Lagace J. Power Sources 195 (2010) 845.  J. Cho, Y.J. Kim, B. Park, Chem. Mater. 12 (2000) 3788.  Z. Chen, J.R. Dahn, Electrochim. Acta 49 (2004) 1079.  P. Verm, P. Maire, P. Nov ak, Electrochim. Acta 55 (2010) 6332.  M. Francisco, A. Bruinhorst, M.C. Kroon, Green Chem. 14 (2012) 2153.