Lithium solvation in bis(trifluoromethanesulfonyl)imide

View Article Online / Journal Homepage / Table of Contents for this issue

COMMUNICATION

www.rsc.org/pccp | Physical Chemistry Chemical Physics

Lithium solvation in bis(trifluoromethanesulfonyl)imide-based ionic liquids Jean-Claude Lasse`gues,* Joseph Grondin and David Talaga

Downloaded by Arizona State University on 09 February 2013 Published on 16 November 2006 on http://pubs.rsc.org | doi:10.1039/B615127B

Received 19th October 2006, Accepted 10th November 2006 First published as an Advance Article on the web 16th November 2006 DOI: 10.1039/b615127b

The lithium solvation in (1 x)(EMI-TFSI), xLiTFSI ionic liquids where EMI+ is the 1-ethyl-3-methylimidazolium cation and TFSI the bis(trifluoromethanesulfonyl)imide anion, is shown by Raman spectroscopy to involve essentially [Li(TFSI)2] anionic clusters for 0 o x o 0.4, but addition of stoichiometric amounts of solvents S such as oligoethers changes the lithium solvation into [Li(S)m]+ cationic clusters; the lithium transference number in TFSI-based ionic liquid electrolytes for lithium batteries should thus be strongly improved. Lithium conducting electrolytes based on ionic liquids such as (EMI+)TFSI , where EMI+ is the 1-ethyl-3-methylimidazolium cation and TFSI the bis(trifluoromethanesulfonyl) imide anion, can be prepared by adding a lithium salt such as LiTFSI.1–4 At room temperature, the (1 x)(EMI-TFSI), xLiTFSI mixtures are liquid in the 0 r x r 0.4 concentration range and their conductivity decreases from B10 to 2 mS cm 1 when the LiTFSI mole fraction x increases from 0 to 0.35.1,2 A room temperature conductivity of a few mS cm 1 is quite acceptable. Much more severe limitations come from the electrochemical stability and the lithium transference number. Because of its acidic imidazolium protons, the cation introduces a cathodic limit of B1 V versus Li+/Li0, which renders the (1 x)(EMI-TFSI), xLiTFSI electrolytes unsuitable for batteries with Li-metal or graphite anodes. Addition of solvents such as vinylene carbonate,3 chemical modifications of the cation,4 of the anion,5 or plastification of polymers,6,7 have been proposed. Even if the electrochemical stability is improved, a fundamental problem remains: which is the lithium transference number, TLi, in ionic liquids where bulky cations such as EMI+, Li+ cations and TFSI anions coexist, with or without an added solvent? To answer this question, it is important to understand the lithium solvation mechanisms. Molecular dynamics simulations are promising,8 but direct measurements of diffusion coefficients by NMR or impedance spectroscopy, and of local interactions by vibrational spectroscopy are complementary.1,7,9 In spite of its limited electrochemical performances, we have chosen to study the (1 x)(EMITFSI), xLiTFSI system by Raman spectroscopy because it is representative of a large family of lithium-doped ionic liquids. Furthermore, the vibrational spectra of the EMI+ and TFSI

Laboratoire de Physico-Chimie Molećulaire, UMR 5803, CNRS, Universite´ Bordeaux I, 351 Cours de la Libe´ration, 33405 Talence Cedex, France. E-mail: [email protected]; Fax: 33 540008402; Tel: 33 540006355

This journal is

c

the Owner Societies 2006

ions in various environments have already been thoroughly investigated by this technique10–14 and many studies have also been devoted to the interactions between Li+ and TFSI .11 Thus, it has been established that the more intense Raman line of TFSI , which corresponds to a whole anion expansion and contraction,10 occurs at 739–742 cm 1 for ‘‘free’’ anions and solvent-separated ion-pairs and is shifted to 746–750 cm 1 when contact ion-pairs or aggregates with Li+ are formed. This is illustrated in Fig. 1 for the (1 x)(EMI-TFSI), xLiTFSI system studied at room temperature as a function of x. The band of the ‘‘free’’ anions is observed at 742 cm 1 for x = 0. ‘‘Free’’ anions means here TFSI anions fully solvated by weakly interacting EMI+ cations. When x increases, a new band grows at about 748 cm 1. It is due to TFSI anions in direct interaction with Li+ cations to form ion-pairs or aggregates. The intensity of the 742 cm 1 band is quantitatively transferred into the 748 cm 1 band. This means that the scattering cross section of the considered vibration is very similar for free anions and ion-paired anions. It is then possible to evaluate the population of the two types of anions from the integrated intensity of the corresponding bands. The

Fig. 1 Raman spectra of (1 x)(EMI-TFSI), xLiTFSI ionic liquids for x = 0, 0.1, 0.15, 0.2, 0.3 and 0.4 at room temperature. For a convenient comparison, all the spectra have been normalized with respect to their total intensity integrated between 717 and 775 cm 1. The population of ‘‘free’’ TFSI anions, i.e. the fraction of intensity contained in the 742 cm 1 band, is given in the insert. The experimental conditions are those described in ref. 11 and 22.

Phys. Chem. Chem. Phys., 2006, 8, 5629–5632 | 5629


View Article Online

Fig. 2 Comparison of the Raman spectra at room temperature of (1 x)(EMI-TFSI), xLiTFSI ionic liquids for x = 0 (dotted line) and x = 0.4 (solid line). Characteristic bands of the ‘‘free’’ cisoid (C1) and transoid (C2) conformers of TFSI are indicated.

two components have been fitted with Voigt profiles with a proportion of Lorentzian imposed to 60%. The results, reported in the insert of Fig. 1, indicate that the population of free anions decreases with a slope close to 2, i.e. that ionpairs of the [Li(TFSI)2] type are formed. Raman spectroscopy can also give some indication on the conformation adopted by the anion in these ion-pairs. We have previously shown that the 260–360 cm 1 spectral range contains specific bands of the cisoid and transoid conformers of TFSI of C1 and C2 symmetry, respectively.11,12 A comparison of the Raman spectra of the x = 0 and x = 0.4 compositions (Fig. 2) indicates that the intensity of the 341 and 297 cm 1 bands of the transoid conformers decreases when x increases. Therefore, the cisoid conformational state of TFSI would be favoured in the [Li(TFSI)2] cluster, although it is of higher energy than the transoid state in the gas state,10,14 in solution,11,12 and even in the (EMI+)TFSI ionic liquid.14 It must be pointed out that this conclusion remains very qualitative because it is based only on the knowledge of the spectra of the ‘‘free’’ conformers at x = 0. The spectra of the ion-paired anion conformers in this region are unknown and one can simply suppose that they are not too different from those of the ‘‘free’’ anions. Finally, we propose a [Li(TFSI)2] structure in which the lithium is, for example, tetrahedrally coordinated to four oxygens of two anions, with the latter in the cisoid (C1) conformation. A recent crystallographic study of the x = 0.66 composition indicates that bidentate Li+ coordinations with two oxygens bound to different sulfur atoms of the same anion and cisoid (C1) anion conformations are indeed predominant.15 It is also interesting to know the binding energy of the lithium ion within [Li(TFSI)2] . The x = 0.2 composition has been studied as a function of the temperature in the 248–408 K temperature range. Some Raman spectra are reported in Fig. 3 and an Arrhenius plot has been built using the integrated intensities of the bands of the ‘‘free’’ ions and of the ion-pairs. 5630 | Phys. Chem. Chem. Phys., 2006, 8, 5629–5632

Fig. 3 Raman spectra of the 0.8(EMI-TFSI), 0.2LiTFSI ionic liquid at different temperatures. The Arrhenius plot in the insert corresponds to ln(I741/I747) = f (1000/T) where I741 and I747 are the integrated intensities of the 741 cm 1 band (free ions) and of the 747 cm 1 band (ion-pairs).

The DH = 1.2 0.2 kJ mol 1 value obtained from the slope of this plot indicates that Li+ is not very strongly bonded to the two anions. Nevertheless, nearly all the available Li+ cations are involved in a [Li(TFSI)2] cluster at room temperature, as shown in the insert of Fig. 1 by a quantity of free anions very close to the theoretical value expected for a solvation of all Li+ cations in [Li(TFSI)2] . At x = 0.2, the population of free anions is 64% instead of the theoretical 60% ones, which means that only 2% Li+ are not involved in a [Li(TFSI)2] cluster. The fact that a large majority of Li+ is transported in an anionic species is detrimental for a lithium-metal or a lithium-ion battery. How can this situation be changed? There have been in the past many attempts to improve the lithium transference number, for example by grafting anions on polymer chains, by complexing Li+ within crown-ether molecules to avoid ion-pairing effects, etc.16 One can also remark in the literature that small amounts of additives S such as organic carbonates, ethers or nitriles, are often used.1,3 When they improve the cyclability, they are generally claimed to favour the formation of a protective layer at the interfaces. We suggest that they may have another very positive action on the present system by transforming the [Li(TFSI)2] anionic cluster into a [Li(S)m]+ cationic cluster. Indeed, the recent and numerous results obtained by several groups,17–19 including ours,20,21 on the complexing of Li+ by glymes Gn of formula CH3(OCH2CH2)nOCH3, indicate, for example, that solventseparated ion-pairs (SSIP) of the [Li(Gn)m]+ family are easily formed. With diglyme G2 (n = 2) and tetraglyme G4 (n = 4), [Li(G2)2]+ and [Li2(G4)2]2+ complexes are obtained with a number of anions. In [Li(G2)2]+, the lithium interacts with six oxygens of two diglymes, whereas in [Li2(G4)2]2+, the two ligands adopt a double-helix structure to establish a sixcoordination structure with two Li+. This journal is

c



View Article Online

Fig. 4 Raman spectra at room temperature of the 0.8(EMI-TFSI), 0.2LiTFSI ionic liquid: (a, dotted line), doped by y = 0.2: (b) or 0.4: (c) moles of G2.

It is easy to show that these glyme complexes are more stable than the [Li(TFSI)2] cluster. Indeed, when y = 0.2 or 0.4 G2 moles are added to the 0.8(EMI-TFSI), 0.2LiTFSI ionic liquid, the spectra of Fig. 4 are obtained. From the fitted areas of the bands of the ‘‘free’’ anions at 742 cm 1 and of the ion-paired anions at 748 cm 1, it is found that the population of free anions increases from 64% at y = 0, to 83% at y = 0.2 and to 94% at y = 0.4. Exactly the same fractions of free anions, 83 and 94%, is found when y = 0.1 and 0.2 moles of G4 are added, respectively. These results indicate that G2 and G4 transform nearly completely the [Li(TFSI)2] anionic clusters into [Li(G2)2]+ and [Li2(G4)2]2+ cationic clusters, respectively. A further proof of this transformation is provided by the Raman spectra recorded in the region of the so-called ‘‘ring-breathing’’ modes of the complexed glyme (Fig. 5). The band near 880 cm 1, accompanied by the weaker bands at 844 and 832 cm 1 are quite characteristic of the coupled C–O stretching/CH2 rocking modes occurring in various [Li(G2)2]+ X complexes (X = ClO4, BF4, PF6, AsF6, SbF6) of a known structure.17–21 The spectrum of [Li(G2)2]+ PF6 (Fig. 5d) is given for comparison. The experiments developed here for the TFSI anion can easily be extended to the bis(perfluoroethanesulfonyl)imide anion (BETI ) which presents very similar conformational properties and coordination ability for Li+.22 Understanding the lithium solvation mechanisms is the first step. It would now be crucial to measure directly the lithium transference numbers with and without added glymes. Indeed, this lithium transference number may depend on the respective times during which the lithium remains solvated within [Li(TFSI)2] or [Li(S)m]+ complexes and diffuses as anionic or cationic clusters, compared to the characteristic exchange times of Li+ between different solvates.8 We believe however that the transformation of [Li(TFSI)2] into [Li(S)m]+ should be highly beneficial for lithium conducting electrolytes based This journal is

c


Fig. 5 Same as Fig. 4, but on a 16 times smaller ordinate scale and with spectrum (d) of [Li(G2)2]+PF6 added for comparison. The asterisk indicates a band of EMI+.

on ionic liquids involving the TFSI or BETI anions. Addition of a solvent such as diglyme should also improve the conductivity. Note added in proof A paper just published by Hardwick et al.23 describes similar Raman experiments on the same (1 x)(EMI-TFSI), xLiTFSI system, where addition of ethylene carbonate (EC) destroys the existing ion-pairs between Li+ and TFSI to produce [Li(EC)4]+ clusters.

References 1 M. Galin´ski, A. Lewandowski and I. St˛epniak, Electrochim. Acta, 2006, 51, 5567. 2 B. Garcia, S. Lavalleé, G. Perron, C. Michot and M. Armand, Electrochim. Acta, 2004, 49, 4583. 3 M. Holzapfel, C. Jost and P. Nova´k, Chem. Commun., 2004, 2098. 4 S. Seki, Y. Kobayashi, H. Miyashiro, Y. Ohno, A. Usami, Y. Mita, N. Kihira, M. Watanabe and N. Terada, J. Phys. Chem. B, 2006, 110, 10228. 5 M. Ishikawa, T. Sugimoto, M. Kikuta, E. Ishiko and M. Kono, J. Power Sources, 2006, 162, 658. 6 J.-H. Shin, W. A. Henderson and S. Passerini, Electrochem. Commun., 2003, 5, 1026. 7 D. Bansal, F. Cassel, F. Croce, M. Hendrickson, E. Plichta and M. Salomon, J. Phys. Chem. B, 2005, 109, 4492. 8 O. Borodin, G. D. Smith and W. A. Henderson, J. Phys. Chem. B, 2006, 110, 16879. 9 A. Noda, K. Hayamizu and M. Watanabe, J. Phys. Chem. B, 2001, 105, 4603. 10 I. Rey, P. Johansson, J. Lindgren, J. C. Lasse`gues, J. Grondin and L. Servant, J. Phys. Chem. A, 1998, 102, 3249. 11 M. Herstedt, M. Smirnov, P. Johansson, M. Chami, J. Grondin, L. Servant and J. C. Lasse`gues, J. Raman Spectrosc., 2005, 36, 762. 12 M. Herstedt, W. A. Henderson, M. Smirnov, L. Ducasse, L. Servant, D. Talaga and J. C. Lasse`gues, J. Mol. Struct., 2006, 783, 145. 13 Y. Umebayashi, T. Fujimori, T. Sukizaki, M. Asada, K. Fujii, R. Kanzaki and S.-i. Ishiguro, J. Phys. Chem. A, 2005, 109, 8976. 14 K. Fujii, T. Fujimori, T. Takamuku, R. Kanzaki, Y. Umebayashi and S.-i. Ishiguro, J. Phys. Chem. B, 2006, 110, 8179.

Phys. Chem. Chem. Phys., 2006, 8, 5629–5632 | 5631

View Article Online 20 J. Grondin, L. Ducasse, J.-L. Bruneel, L. Servant and J.-C. Lasse`gues, Solid State Ionics, 2004, 166, 441. 21 J. Grondin, J. C. Lasse`gues, M. Chami, L. Servant, D. Talaga and W. A. Henderson, Phys. Chem. Chem. Phys., 2004, 6, 4260. 22 J. Grondin, D. Talaga, J. C. Lasse`gues, P. Johansson and W. A. Henderson, J. Raman Spectrosc., 2006, DOI: 10.1002/jrs. 1573. 23 L. J. Hardwick, M. Holzapfel, A. Wokaun and P. Nova´k, J. Raman Spectrosc., 2006, DOI: 10.1002/jrs.1632.


15 K. Matsumoto, R. Hagiwara and O. Tamada, Solid State Sci., 2006, 8, 1103. 16 F. M. Gray, Polymer Electrolytes, RSC Materials Monographs, The Royal Society of Chemistry, Cambridge, 1997. 17 W. A. Henderson, N. R. Brooks, W. W. Brennessel and V. G. Young, Jr, J. Phys. Chem. A, 2004, 108, 225. 18 Y. G. Andreev, V. Senevitrane, M. Khan, W. A. Henderson, R. E. Frech and P. G. Bruce, Chem. Mater., 2005, 17, 767. 19 W. A. Henderson, F. McKenna, M. A. Khan, N. R. Brooks, V. G. Young, Jr and R. Frech, Chem. Mater., 2005, 17, 2284.

5632 | Phys. Chem. Chem. Phys., 2006, 8, 5629–5632

This journal is

c
