J. Phys. Chem. A 2010, 114, 687–688
687
Reply to the “Comment on ’New Interpretation of the CH Stretching Vibrations in Imidazolium-Based Ionic Liquids’” Jean-Claude Lasse`gues,*,† Joseph Grondin,† Dominique Cavagnat,† and Patrik Johansson‡ ISM, UMR 5255, CNRS, UniVersite´ Bordeaux I, 351 Cours de la Libe´ration, 33405 Talence Cedex, France, and Department of Applied Physics, Chalmers UniVersity of Technology, SE-41296, Go¨teborg, Sweden ReceiVed: October 12, 2009 In their Comment, Wulf et al. raise doubts about the general applicability of our new assignment. However, we clearly stated that the proposed new interpretation of the CH stretching vibrations in imidazolium-based ionic liquids (ILs) applies mainly to weakly coordinating anions such as bis(trifluoromethanesulfonyl)imide (TFSI-).1 Other examples are the widely used AlCl4-, PF6-, BF4-, and B(CN)4- anions.2-5 It is well-known from the literature that more basic, and thus better coordinating, anions produce marked spectral changes in the IR spectra as illustrated in Figure 1 by a series of 1-ethyl-3methylimidazolium (EMIM) derivatives. The ILs with weakly coordinating BF4- and TFSI- anions are characterized by similar profiles, but stronger proton acceptors such as CF3SO3-, N(CN)2-, and Br- do produce clearly visible stronger, broader, and red-shifted absorptions that have been generally interpreted in terms of hydrogen bond interactions of increasing strength. Wulf et al. present about the same kind of experimental evidence with another series of anions and, using the complementary information provided by DFT calculations of clusters and NMR proton chemical shifts, they conclude that the C(4,5)H stretching vibrations are situated in the 3200-3120 cm-1 range, whereas those of C(2)H are red-shifted to 3120-3020 cm-1. In other words, with the weakly coordinating TFSI-, BF4-, or B(CN)4- anions, the C(4,5)H and C(2)H stretching vibrations would produce the two main absorptions at ∼3160 and ∼3120 cm-1, respectively (Figure 1). As the 3120 cm-1 absorption often involves two components, also two kinds of hydrogen bond associations for C(2)H have been necessary to invoke.6,7 This kind of assignment seems natural when considering the two above series of results alone. However, it is not corroborated by the selective deuteration of the imidazolium CH groups.1 If two families of νC(2)H were involved at 3125 and 3105 cm-1 for the fully hydrogenated cations,6 two νC(2)D bands should be resolved when the C(2)H group is deuterated. In all the systems investigated until now, only one νC(2)D absorption band is observed. It is situated at ∼2350 cm-1 for imidazoliumd1 derivatives associated with the weakly coordinating anions AlCl4-,2 BF4-,3 or TFSI-,1 and shifted to 2300 and 2280 with I- 4 and Cl-,2 respectively, in good agreement with the increasing basicitiy of the halides. At the same time, three components are systematically observed in the νCH region where only the νipC(4,5)H and νopC(4,5)H vibrations are expected. As far as we know, IR spectra of a selectively deuterated C(4)D and C(5)D imidazolium derivatives have only be recorded for [EMIM-d2][TFSI].1 One νC(2)H component is expected, but * Corresponding author. E-mail:
[email protected]. † ISM. ‡ Chalmers University of Technology.
Figure 1. IR spectra of EMIM derivatives at room temperature, except [EMIM]Br melted at 353 K. The intensity of the absorption band at ∼1575 cm-1 due to the R1 and R2 in-plane ring modes (not shown) has been normalized for a more convenient comparison.
two bands are observed at 3160 and 3096 cm-1. This is a very clear example of Fermi resonance between the νC(2)H fundamental level and the 2R1 overtone. It corresponds to the highest value of the calculated coupling constants (Table SI2 of ref 1). All other νC(2)D vibrations, including those of the fully deuterated-d3 derivative, are predicted by the anharmonic gas phase model. Wulf et al. do not at all comment on these results although some of them on -d1 derivatives have been published by other groups.2-4 We believe, however, that isotopic effects constitute the more severe test of any vibrational assignment. As far as we know, anharmonic calculations of alkyl-substituted and selectively deuterated imidazolium cations were performed for the first time and they reveal Fermi resonances never taken into account until now.1 Wulf et al. seem to admit that Fermi resonances do contribute to the νCH spectra. On the other hand, we are aware that a gas phase model is just a first approximation used to support anharmonic calculations. Ideally, these anharmonic calculations have now to be applied to more realistic models, especially when more basic anions are involved. Even in cases where hydrogen bonding plays a minor role, medium effects should be taken into account, for example, within the framework of a polarized continuum model (PCM). This latter approach has proved to be essential also to predict the NMR properties of imidazolium-based ILs.8,9 Interestingly, however, the νC(2)H vibrations for the 1,3-dimethylimidazolium (MMIM+) cation in the gas phase and for the [MMIM][PF6] ion-pair solvated in a PCM benzene medium have both been calculated at 3298 cm-1.8 Coming back to the IR spectra of the isotopic derivatives, the more striking observation is made for the [EMIM-d2][TFSI] derivative where the interacting νC(2)H and 2R1 levels share about the same intensity in the observed bands at 3160 and 3096 cm-1. We did not emphasize enough in our Letter that both these bands have a νC(2)H character. The unperturbed νC(2)H level must be situated in between, at ∼3128 cm-1, giving an
10.1021/jp909770s 2010 American Chemical Society Published on Web 12/16/2009
688
J. Phys. Chem. A, Vol. 114, No. 1, 2010
isotopic ratio of about 1.33 with the νC(2)D at 2350 cm-1. With more basic anions, the R1 and R2 vibrations are little perturbed, but the νC(2)H level is certainly red-shifted and the Fermi resonance interaction is going to change in a way that is presently unpredictable. Therefore, we fully agree with the last sentence of Wulf et al.’s Comment saying that it is highly desirable to separate the vibrational contribution in the νCH region stemming either from hydrogen bonding or from Fermi resonances. This is the subject of a forthcoming publication.10 A fundamental question raised by this discussion is to know when, where, and according to which criteria a C-H · · · anion interaction deserves the name of hydrogen bond. Wulf et al. cite experimental and theoretical studies on imidazolium-based ILs where hydrogen bonding is shown to play a crucial role. However, and again in the case of weakly coordinating anions, which is the main topic and cornerstone of the new assignment, there are as many studies leading to the opposite conclusion. Tsuzuki et al. claim, for example, that the nature of the C(2)H · · · anion interaction is completely different from that of a conventional hydrogen bond and remains negligibly small compared to the charge-charge interactions.11 Very recent NMR experiments combined with DFT calculations confirm this view for [EMIM][BF4] and its solutions in dichloromethane.9 In the IL, the anion can take several positions and orientations with respect to the cation with very small energy costs; in particular, the orientation dependence of the interaction energy is found to be weak and thus in contrast with directional hydrogen bonds. Dilution of [BMIM][BF4] in water,4 or of [EMIM][BF4] in dichloromethane,12 does not appreciably change the νCH IR profile. More generally, the solvent or temperature effects presented until now in the literature do not produce spectral changes that can clearly be interpreted in terms of an equilibrium between associated and dissociated species. Other fundamental physicochemical properties of the ILs are difficult to understand. For example, substitution of a methyl group on the C(2) carbon of the [BMIM][TFSI] and [BMIM][BF4] ILs leads to an unexpected increase of the viscosity and melting point.13 This counterintuitive behavior is explained by the fact that the loss of hydrogen bonding is outweighed by a loss in entropy coming from the elimination of ion pairs and an increase in the rotational barrier of the butyl chain.13 However, the same methyl substitution on the C(2) carbon of [MMIM]]TFSI] also produces an increase in the melting point from 26 to 108 °C,14 whereas conformational isomerism is excluded for both trimethyl as well as dimethyl derivatives. This unexpected increase of the melting
Comments point is even observed when TFSI is substituted by the stronger proton acceptor chloride anion.15 It has been shown that the total interaction energy of a [EMIM][BF4] ion-pair is only 4% larger than that of the same ion pair where the C(2)H bond of the cation has been methylated.16 Therefore, the role played by hydrogen bond interactions is far from being easy to evaluate in a general description of the ILs energy landscape.15 For vibrational spectroscopy to contribute to the understanding of ionic interactions in ILs, well-established vibrational assignments are a priority. We believe that our new interpretation of the νCH spectra of imidazolium-based ILs associated with weakly coordinating anions is not an overstatement. It rather points out the importance of Fermi resonance interactions in these systems, the necessity to revisit many literature studies based on temperature, pressure or solvent effects in the νCH region, and the possibility to use the much less complicated νCD vibrations. References and Notes (1) Lasse`gues, J. C.; Grondin, J.; Cavagnat, D.; Johansson, P. J. Phys. Chem. A 2009, 113, 6419. (2) Dieter, K. M.; Dymek, C. J., Jr.; Heimer, N. E.; Rovang, J. W.; Wilkes, J. S. J. Am. Chem. Soc. 1998, 110, 2722. (3) Katsyuba, S. A.; Zvereva, E. E.; A. Vidisˇ, A.; Dyson, P. J. J. Phys. Chem. A 2007, 111, 352. (4) Jeon, Y.; Sung, J.; Seo, C.; Lim, H.; Cheong, H.; Kang, M.; Moon, B.; Ouchi, Y.; Kim, D. J. Phys. Chem. B 2008, 112, 4735. (5) Scheers, J.; Johansson, P.; Jacobsson, P. J. Electrochem. Soc. 2008, 155, A628. (6) Ko¨ddermann, T.; Wertz, C.; Heintz, A.; Ludwig, R. ChemPhysChem 2006, 7, 1944. (7) Fumino, K.; Wulf, A.; Ludwig, R. Angew. Chem., Int. Ed. 2008, 47, 3830. and 8731. (8) Palomar, J.; Ferro, V. R.; Gilarranz, M. A.; Rodriguez, J. J. J. Phys. Chem. B 2007, 111, 168. (9) Katsyuba, S. A.; Griaznova, T. P.; Vidisˇ, A.; Dyson, P. J. J. Phys. Chem. B 2009, 113, 5046. (10) Lasse`gues, J. C.; Grondin, J.; Cavagnat, D.; Johansson, P.; Holomb, R. Manuscript in preparation. (11) Tsuzuki, S.; Tokuda, H.; Mikami, M. Phys. Chem. Chem. Phys. 2007, 9, 4780. (12) Katsyuba, S. A.; Dyson, P. J.; Vandyukova, E. E.; Chernova, A. V.; Vidisˇ, A. HelV. Chim. Acta 2004, 87, 2556. (13) Hunt, P. A. J. Phys. Chem. B 2007, 111, 4844. (14) Henderson, W. A. Private communication. (15) Zahn, S.; Bruns, G.; Thar, J.; Kirchner, B. Phys. Chem. Chem. Phys. 2008, 10, 6921. (16) Tsuzuki, S.; Tokuda, H.; Hayamizu, K.; Watanabe, M. J. Phys. Chem. B 2005, 109, 16474.
JP909770S