Solid State Ionics 175 (2004) 267 – 272 www.elsevier.com/locate/ssi
New lithium salts for rechargeable battery electrolytes Braja Mandala,*, Thanasat Sooksimuanga,1, Brian Griffinb, Akshaya Padhib, Robert Fillerb a
Department of Biological, Chemical and Physical Sciences, Illinois Institute of Technology, Chicago, IL 60616, United States b TechDrive, Inc., 3255 S. Dearborn Street, Chicago, IL 60616, United States Received in revised form 10 November 2003; accepted 21 November 2003
Abstract The facile syntheses of new, low-cost, non-fluorinated, sulfonyl-substituted imide and methide lithium salts are described. These salts, prepared for potential application in lithium ion rechargeable battery electrolytes, exhibit very good electrochemical and thermal behavior. While the salts are very soluble in DMSO and sulfolane, their solubilities in standard carbonate solvents is less than adequate for battery operations. Molecular modifications to improve solubility are in progress. D 2004 Elsevier B.V. All rights reserved. Keywords: Non-fluorinated lithium salts; Lithium rechargeable batteries; Liquid electrolytes; Electric and hybrid electric vehicles
1. Introduction Small lithium-ion (Li-ion) batteries, which possess high energy density compared to other secondary batteries, are commercially available (with a capacity of 1300 to 1900 mA h) to power portable electronic devices such as cellular phones, camcorders, computers, and cameras [1–3]. Fullsize batteries are now under consideration for use in electric and hybrid electric vehicles (EV/HEV) to provide a longer driving range, higher acceleration, long lifetime and a reduction in environmental pollution. One of the major obstacles that prevents the mass scale production of these batteries is the high cost of lithium salts. Current state of the art Li-ion batteries use expensive low lattice energy lithium salts such as LiPF6, LiN(SO2CF3)2 (Fluorad HQ-115) and LiN(SO2CF2CF3)2 (BETI). Although the systems derived from these salts provide desired characteristics for small Li-ion batteries, they are unsuitable for EV/HEV batteries (where large volumes of electrolytes
* Corresponding author. Tel.: +1 312 567 3446; fax: +1 312 567 3436. E-mail address: [email protected]
(B. Mandal). 1 Visiting Research Fellow. Permanent address: National Metal and Materials Technology Center, NSTDA Building, 73/1 Rama VI rd. Rajdhevee, Bangkok 10400, Thailand. 0167-2738/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2003.11.037
are required) because they are too costly, and especially, in the case of widely used LiPF6, is unattractive because of poor thermal stability and potential hydrolytic decomposition to form HF, which destroys the cell cathode [4,5]. Highly fluorinated compounds also present environmental concerns involving biodegradability. In contrast, cheaper non-fluorine containing salts, such as lithium methanesulfonate (CH3SO3Li), possess high interionic attractive forces (high lattice energy) and their use results in low ionic conductivities. Therefore, there is a strong demand for low lattice energy lithium salts that are not only inexpensive, but also environmentally benign. The availability of such salts will be a major advance in rechargeable battery technology.
2. Experimental 2.1. Lithium bis(methyl)sulfonimide (NLS-011) In a 250-mL three-neck flask fitted with an argon inlet, a thermometer and a reflux condenser, 70 mL of 1.4 M MeLi was added dropwise to a solution of 7 g (32.7 mmol) of IMC  in 50 mL of dry benzene at 40 8C. After stirring for 2 h at room temperature, the product was filtered and washed with 50 mL each of benzene and ether. The product was then stirred with 2-propanol (200 mL) and the white
B. Mandal et al. / Solid State Ionics 175 (2004) 267–272
product filtered. Yield: 98%, mp. N300 8C. FT-IR (KBr, cm 1): 2932, 1142, 949. 1H-NMR (DMSO-d6, ppm): 4.20 (s, 6H). 2.2. Lithium bis(sec-butyl)sulfonimide (NLS-012) Following the above procedure and using 7 g (32.7 mmol) of IMC and 98 mL of 1.0 M sBuLi, the title compound was obtained in 96% yield. Mp. 280 8C (dec). FT-IR (KBr, cm 1): 2930, 1160, 944. 2.3. Lithium (methanesulfonyl)cyanamide (NLS-013) In a 250-mL three-neck flask fitted with an argon inlet, a thermometer and a reflux condenser, 5.0 g (119 mmol) of cyanamide in 50 mL THF was added dropwise to a solution of 2.86 g (119 mmol) of NaH in 50 mL of dry THF while maintaining the temperature between 30 and 40 8C. After refluxing for 2 h, the solution was cooled to 0 8C and 13.63 g (119 mmol) of methanesulfonyl chloride was added dropwise. The solution was then brought to room temperature and stirred overnight. The solvent was evaporated and the product was washed with cold water. The yield of this nitrogen acid was 68%. The neutralization was carried out by using 1 eq. of LiOH. The evaporation of water provided an almost quantitative yield of NLS-013. Mp. N300 8C. FT-IR (KBr, cm 1): 3015, 2932, 2229, 1287, 1218, 1069, 983. 2.4. Bis(methanesulfonyl)methane In a 500-mL round bottom flask, fitted with a condenser, 180 mL (7 eq.) of 30% hydrogen peroxide was added dropwise to a solution containing 24.66 g (225.6 mmol) of bis(methylthio)methane (Aldrich) in 120 mL of acetic acid. The temperature of the reaction was maintained between 45 and 55 8C. After addition, the temperature was then raised to 100 8C and heated for 6 h. On cooling, white needles formed, which were filtered and dried under vacuum. Yield: 92%. Mp. 144–145 8C.
2.5. Bis(ethanesulfonyl)methane (two steps in one-pot) In a 500-mL round bottom flask, was added 4.5 g (49.5 mmol) of 1,3,5-trioxane, 450 mg of p-toluenesulfonic acid and 30 mL (393 mmol) of ethanethiol. The mixture was heated at 40 8C for 6 h. Excess ethanethiol was then removed by distillation. The residue was dissolved in 150 mL of acetic acid. 100 mL of 30% hydrogen peroxide was added dropwise while maintaining the temperature between 45 and 55 8C. After addition, the temperature was raised to 100 8C and heated for 6 h. On cooling, white needles formed, which were filtered and dried under vacuum. Yield: 88%, Mp. 102–104 8C. 2.6. Tris(methanesulfonyl)methane In a three-neck flask, fitted with a reflux condenser, a thermometer and an addition funnel, was added 13.5 mL (174.6 mmol) of methanesulfonyl chloride in 250 mL of dry ethanol. 34.0 g (174.4 mmol) of sodium bis(methanesulfonyl)methide, prepared by reacting bis(methanesulfonyl)methane and sodium ethoxide, was then added in portions during 0.5 h. The mixture was then refluxed for 1 h and cooled to room temperature. The white product was filtered, washed with cold water until the silver nitrate test for chloride ion was negative. Yield: 85%, Mp. N300 8C. 2.7. Bis(methanesulfonyl)(ethanesulfonyl)methane Following the above procedure and using 10.33 g (80.3 mmol) of ethanesulfonyl chloride and 15.50 g (79.9 mmol) of sodium bis(methanesulfonyl)methide, the title compound was obtained in 76% yield. 2.8. Bis(ethanesulfonyl)(methanesulfonyl)methane Following the above procedure and using 8.72 g (76.1mmol) of methanesulfonyl chloride and 16.80 g (75.7 mmol) of sodium bis(ethanesulfonyl)methide, the title compound was obtained in 72% yield.
B. Mandal et al. / Solid State Ionics 175 (2004) 267–272
Fig. 1. Synthesis of non-fluorinated lithium imides.
2.9. Lithium tris(methanesulfonyl)methide (NLS-021)
3. Results and discussion
An aqueous suspension containing 23.23 g of tris(methanesulfonyl)methane was titrated with 0.2 M LiOH until an end point was detected using a phenolphthalein indicator. The solution was then evaporated to dryness and washed with THF. Finally, the product was dried under vacuum at 150 8C for 6 h, during which time traces of bis(methanesulfonyl)methane impurities sublimed out of the salt. Yield: 85%, Mp. N300 8C. FT-IR (KBr, cm 1): 3017, 2936, 1281, 1115, 1043, 970. 1H-NMR (DMSO-d6, ppm): 2.92(s, 9H). Analysis: Calc. C, 18.75; H, 3.54. Found. C, 18.88; H, 3.43.
3.1. Synthesis of lithium salts
2.10. Lithium bis(methanesulfonyl)(ethanesulfonyl)methide (NLS-022) Following the above procedure and using 11.0 g (41.7 mmol) of bis(methanesulfonyl)(ethanesulfonyl)methane, the title compound was obtained in 82% yield, Mp. N300 8C. FT-IR (KBr, cm 1): 3038, 2930, 1293, 1113, 1029, 948. 1HNMR (DMSO-d6, ppm): 1.13(t, 3H), 2.49(s, 6H), 3.00(q, 2H). Analysis: Calc. C, 22.22; H, 4.10. Found. C, 22.05; H, 3.81. 2.11. Lithium bis(ethanesulfonyl)(methanesulfonyl)methide (NLS-023) Following the above procedure and using 3.9 g (14.0 mmol) of bis(ethanesulfonyl)(methanesulfonyl)methane, the title compound was obtained in 84% yield, Mp. N300 8C. FT-IR (KBr, cm 1): 2986, 2946, 1281, 1115. 1 H-NMR (DMSO-d6, ppm): 1.14(t, 6H), 2.91(s, 3H). Analysis: Calc. C, 25.35; H, 4.61. Found. C, 25.69; H, 4.24.
3.1.1. Lithium imides The primary goal of our study was to identify new, stable, low-cost, and environmentally friendly lithium salts which exhibit good electrochemical and thermal properties, for potential use in Li-ion rechargeable batteries. We have designed, synthesized on a small scale, and characterized more than a dozen novel, inexpensive non-fluorinated lithium salts (NLS) which can be classified into two categories, imide and methide. In this study, we systematically investigated the effect on stability of salts substituted with various electron-withdrawing groups (EWG), including, CN, SO2, NO2 and PO(OEt)2. We found that the sulfonyl (SO2) and cyano (CN) groups are clearly superior to other EWG in terms of hydrolytic stability, thermal stability, electrochemical stability and ease of synthesis. For instance, the salts with a NO2 group decompose, as indicated by color change and those of PO(OEt)2 are seemingly very hygroscopic. By contrast, the cyano group is very stable if not contaminated by water. The sulfonyl group also possesses a special, unique ability to further stabilize an adjacent negative charge, owing to sulfur d-orbital participation . As a consequence, compounds such as (RSO2)3CH are exceptionally strong carbon acids with pK a’s ranging from b1 to 3, which results in unusually stable salts. Thus, (C2H5SO2)3CH exhibits a titration curve almost identical to that of HCl [8,9]. This compound and several closely related analogs have been reported previously , but their lithium salts have not been described prior to our study. The SO2 group also leads to a significant increase in the nitrogen acidity and stabilization of the resulting imide ion (Scheme 1).
Fig. 2. Synthesis of a lithium cyanamide.
B. Mandal et al. / Solid State Ionics 175 (2004) 267–272
Fig. 3. Synthesis of non-fluorinated lithium methides.
NLS-011 and NLS-012 were prepared in two steps, as shown in Fig. 1. The first step is the preparation of imidodisulfuric acid chloride (IMC) , which was obtained in greater than 80% yield. The second step, lithiation/ alkylation, was quantitative. The attributes of this methodology are inexpensive starting materials and very high yield. However, we were disappointed that NLS-011 was only sparingly soluble in standard solvent (1:1 EC/DMC), though very soluble in dimethyl sulfoxide (DMSO) and sulfolane. In contrast, NLS-012, was moderately soluble in standard solvent. Another challenge of this methodology is the difficulty in freeing the salt from traces of LiCl, a byproduct of step 2. Isopropanol was used to remove most of the LiCl from the imide salt. NLS-013, a lithium cyanamide, is perhaps the most promising candidate among the nitrogen salts. The pK a of its conjugate acid is probably near 1 (pK a: CH3CONH2=14, NCNH2=7 and CH3CO(CN)NH=3.8) . This salt, prepared in two steps in good yield (Fig. 2), is very soluble in DMSO and acetonitrile. 3.1.2. Lithium methides The salts of this class are even more promising than the imides, due to the presence of three EWG. The synthetic approaches also involve inexpensive starting materials. First, we demonstrated the feasibility of the methodology by preparing the symmetrical methide, NLS-021, starting with commercially available bis(thiomethyl)methane (Fig. 3). The oxidation step proceeded almost quantitatively by using H2O2 in acetic acid. In step 2, the free carbon acid was
prepared by reaction with methanesulfonyl chloride. Finally, the salt was obtained by neutralization with LiOH. This methide salt was partially soluble in carbonates (standard solvent), but very soluble in DMSO and sulfolane. Solubility has been improved traditionally by removing structural symmetry and/or attaching branched aliphatic chains. In order to address the solubility issue, we prepared an unsymmetrical methide, NLS-022, by using ethanesulfonyl chloride in step 2 (Fig. 3). The solubility of this salt was much improved, but not to the point acceptable for electrolyte applications. With these encouraging results on solubility, we then designed and synthesized another unsymmetrical methide, NLS-023, which contains two ethyl groups and one methyl group (Fig. 4). The synthesis involved inexpensive materials, such as trioxane, and avoided relatively expensive bis(thiomethyl)methane. The desired bis(ethanesulfonyl)methane was prepared in a one-pot reaction in high yield and subsequently, NLS-023. This methide exhibited very good solubility (~0.4 M) in standard solvent. A detailed cost analysis of the new salts indicates that they can be produced at a cost 50–60% less than for LiPF6 and Fluorad HQ-115. 3.2. Properties of NLS 3.2.1. Solubility While we accomplished our objective of efficiently synthesizing low-cost NLS, we observed that their solubilities are low (0.1–0.4M) in standard solvents, but very soluble in DMSO or sulfolane (0.3–4 M). Although the solubilities of NLS-022 and NLS-023 in standard solvents Table 1 Solubilities of selected non-fluorinated lithium salts
Fig. 4. Synthesis of fairly soluble methide salt.
NLS-011 NLS-012 NLS-013 NLS-021 NLS-022 NLS-023
Poor (b0.1 M) Fair (~0.3 M) Very good (N1 M) Good (~0.5 M) Excellent (N4 M) Excellent (N4 M)
Poor (b0.1 M) Poor (b0.1 M) Good (~0.5 M) Fair (~0.3 M) Very good (N1 M) Very good (N1 M)
Insoluble Poor (b0.1 M) Poor (b0.1 M) Poor (b0.1 M) Fair (~0.3 M) Good (~0.5 M)
B. Mandal et al. / Solid State Ionics 175 (2004) 267–272 Table 2 Electrochemical and thermal properties of NLS Compound
Thermal decomposition peaksc (8C)
LiPF6 NLS-011 NLS-012 NLS-013 NLS-021 NLS-022 NLS-023
9.8 3.2 3.2 7.1 4.3 4.3 4.3
4.9 –e –e 4.35 –e 4.57 4.58
108, 196, 241 – – 239 none none none
92, 203 302 285 394 388 385 390
a b c d e
0.5 M solution in DMSO at 25 8C. CV was performed with 0.1 M NLS+0.5 M LiPF6 in standard solvent. Exothermic peaks below 250 8C (from DSC trace). Product decomposition temperature (PDT) was obtained by TGA. Not performed due to insolubility in standard solvent.
have been enhanced, further structural modifications are required for potential applications. We are confident that the solubility of these salts can be improved by substituting appropriate solubilizing groups (Table 1). 3.2.2. Electrochemical and thermal stability Since redox reactions can be observed even at low concentration of salts, we performed CV of one representative salt from each category in the conducting medium of carbonates containing LiPF6. All the salts showed good electrochemical stability (Table 2).
Fig. 5. Nyquist plots: (a) 0.5 M LiPF6 in DMSO and (b) 0.5 M NLS-013 in DMSO.
3.2.3. Conductivity The conductivity of the salts and LiPF6 were measured in 0.5 M DMSO for a fair comparison. All the salts, especially NLS-013, exhibit very good conductivity with respect to LiPF6 (Table 2). The data are very favorable for a low dielectric constant solvent such as DMSO. Our study indicates that these salts would be capable of offering conductivity near 10 mS/cm, if they were soluble in high concentration (~1M) in carbonate solution or other high dielectric constant solvent systems. Fig. 5 shows the Nyquist plots of NLS-013 and LiPF6 in DMSO. 3.2.4. Thermal stability The product decomposition temperature (PDT), at which rapid decomposition begins, of each NLS was determined by thermogravimetric analysis (TGA) (Fig. 6). All NLS showed excellent thermal stability compared to that of LiPF6 (Table 2). The thermal stability of the salts was also determined by differential scanning calorimetry (DSC) of neat sample. No exothermic or endothermic peaks were detected below 150 8C. Fig. 6 shows representative TGA and DSC traces of selected salts.
Fig. 6. Thermal stability studies of selected salts under nitrogen: (a) TGA (heating rate 50 8C/min, 3 M test specification)  and (b) DSC (heating rate 20 8C/min).
B. Mandal et al. / Solid State Ionics 175 (2004) 267–272
Facile synthesis of new low-cost non-fluorinated lithium imides and methides for potential use in lithium-ion battery electrolytes are described. These salts exhibit very good electrochemical and thermal properties. While the solubilities of the salts in standard carbonate solvents are somewhat less than required for battery operations, they are soluble in DMSO and sulfolane. Structural tailoring to improve the solubilities of these salts is in progress.
 D. Linden, Handbook of Batteries, 2nd ed., McGraw-Hill, New York, 1995.  B. Scrosati, Applications of Electroactive Polymers, Chapman & Hall, London, 1993.  G. Pistoia, Lithium Batteries: New Materials, Developments and Perspectives, vol. 5, Elsevier, Amsterdam, 1994.  K. Xu, S. Zhang, T.R. Jow, W. Xu, C.A. Angell, Electrochem. SolidState Lett. 5 (2002) A26 – A29.  J. Cho, Chem. Mater. 13 (2001) 4537.  M. Becke-Goehring, E. Fluck, Inorg. Synth. 8 (1966) 105.  S. Oae, W. Tagaki, A. Ohno, Tetrahedron 20 (1964) 427.  W. von E. Doering, L.K. Levy, J. Am. Chem. Soc. 77 (1955) 509.  E. Same´n, Arkiv Kemi, Miner. Geol. 24B (1947) 1.  H. Bfhme, R. Marx, Berita 74B (1941) 1667.  J.S. Dewar, The Electronic Theory of Organic Chemistry, Oxford University Press, Oxford, 1949, p. 92.
Acknowledgment The authors express their thanks and appreciation for the financial resources provided by the U.S. Department of Energy (DE-FG02-99ER82905) in support of this study.