CHEMPHYSCHEM ARTICLES DOI: 10.1002/cphc.201402205
Ion Pairing in Protic Ionic Liquids Probed by Far-Infrared Spectroscopy: Effects of Solvent Polarity and Temperature Koichi Fumino,[a] Verlaine Fossog,[c] Peter Stange,[a] Kai Wittler,[a] Wigbert Polet,[a] Rolf Hempelmann,[c] and Ralf Ludwig*[a, b] The cation–anion and cation–solvent interactions in solutions of the protic ionic liquid (PIL) [Et3NH][I] dissolved in solvents of different polarities are studied by means of far infrared vibrational (FIR) spectroscopy and density functional theory (DFT) calculations. The dissociation of contact ion pairs (CIPs) and the resulting formation of solvent-separated ion pairs (SIPs) can be observed and analyzed as a function of solvent concentration, solvent polarity, and temperature. In apolar environments, the CIPs dominate for all solvent concentrations and
temperatures. At high concentrations of polar solvents, SIPs are favored over CIPs. For these PIL/solvent mixtures, CIPs are reformed by increasing the temperature due to the reduced polarity of the solvent. Overall, this approach provides equilibrium constants, free energies, enthalpies, and entropies for ion-pair formation in trialkylammonium-containing PILs. These results have important implications for the understanding of solvation chemistry and the reactivity of ionic liquids.
1. Introduction Ionic liquids (ILs) are liquids that consist entirely of ions. The potential applications of these Coulomb fluids depend on the properties of the particular liquid material.[1–4] To a large extent, the structure and properties of ILs are determined by the intermolecular interactions between anions and cations.[5, 6] In particular, the subtle balance between Coulomb forces, hydrogen bonds, and dispersion forces is of great importance in the understanding of ILs.[7, 8] This mlange of interactions of varying type, strength, and directionality leads to the formation of charged and neutral clusters, and ion pairs.[9] Moreover, ion pairing is crucial in the determination of the properties of Coulomb fluids. Two types of ILs are of increasing interest in science and technology: aprotic and protic ILs.[1–4, 10–16] It is known that the addition of hydrophobic organic solvents to aprotic ILs leads to electrical conductance minima and liquid–liquid miscibility gaps.[17, 18] In highly diluted systems, neutral aggregates and contact ion pairs (CIPs) are formed. This situation is [a] Dr. K. Fumino, P. Stange, K. Wittler, W. Polet, Prof. Dr. R. Ludwig Universitt Rostock Institut fr Chemie Abteilung fr Physikalische Chemie Dr.-Lorenz-Weg 1 18059, Rostock (Germany) E-mail: [email protected] [b] Prof. Dr. R. Ludwig Leibniz-Institut fr Katalyse an der Universitt Rostock e.V. Albert-Einstein-Str. 29a 18059 Rostock (Germany) [c] V. Fossog, Prof. Dr. R. Hempelmann Universitt des Saarlandes Physikalische Chemie 66123 Saarbrcken (Germany) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402205.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
similar to that of simple electrolytes in dilute solutions.[18] For protic ILs (PILs), the initial position can be the opposite. Possible proton transfer from the Brønsted acid to the Brønsted base leads to the formation of distinct proton acceptor and donor sites and results in strong hydrogen bonding.[10–16] This holds in particular for PILs that include the triethylammonium ion [Et3NH] + as the cation. Beside the Coulomb interaction, a strong and directional hydrogen bond between the NH group of the cation and a polar atom of the anion provides strongly bound CIPs in the pure Coulomb fluid.[19–21] Although each ion is surrounded by counterions, the tight cation–anion interaction leads to the existence of quasi CIPs in the neat PIL. We recently showed that this specific cation–anion interaction gives rise to distinct vibrational bands in the low frequency range between 100 and 180 cm1 that depend on the interaction potential of the involved ions.[14, 19] Thus, far-infrared (FIR) spectroscopy is expected to be a suitable method for the study of ion pairing in this type of PIL. Usually, ion pairing proceeds in stages as originally suggested by Eigen and Tamm.[22] In the first stage, free solvated ions form a doubly solvent-separated ion pair (2SIP), in which both ions essentially retain their first solvation shells. In the second stage, partial desolvation of 2SIP occurs to form solvent-shared ion pairs (SIP). In the final stage, the solvent molecules between the ions are eliminated to form CIPs. All species are in chemical equilibrium, which is described by ion association and solvent elimination constants.[23–25] The process for PILs described in this work is the reverse of that process. We start with CIPs that are characterized by long-range Coulomb interactions, which are strongly enhanced by short-range, spatially directed hydrogen-bonding interactions between the Lewis acid and the Lewis base. We then study the dissociation of CIPs into SIPs under different solvent concentrations and polarChemPhysChem 2014, 15, 2604 – 2609
2604
CHEMPHYSCHEM ARTICLES ities expressed by the dielectric constant, e. Thus, the dissociation constant can be regarded as the equilibrium constant that describes the equilibrium between CIPs and SIPs (see Scheme 1). The solvent concentrations used here do not allow further transfer from SIPs to 2SIPs. The maximum 9:1 ratio between solvent molecules and CIPs does not provide sufficient
Scheme 1. Ion-pair concept as defined for the PIL [Et3NH][I]. Here, dimethyl sulfoxide (DMSO) is used as the solvent molecule.
www.chemphyschem.org ture, they were measured in pellets of polyethylene. Both spectra share a common vibrational feature around 70 cm1 that is usually attributed to cage rattling of interacting ions and bending modes of directional intermolecular interaction.[19, 20] The additional vibrational band at 106 cm1 for TEAI can be clearly assigned to the + NH···I interaction, which is strongly enhanced by hydrogen bonding.[14] Consequently, this feature is missing in the spectrum of [Et4N][I]. This finding is supported by DFT calculated frequencies of aggregates [Et3NH][I]n with n = 1–3. The calculated average value for all species of about 103 cm1 is in perfect agreement with the measured frequency.[27–33] If we include Grimme’s DFT–D3 method for calculating the non-covalent interactions, we obtain a value of 114 cm1, which is still reasonable considering that no correction for the harmonic approximation is applied (see the Supporting Infor-
solvent molecules to form full solvation shells for each ion. It is expected that the purely CIP contributions in the neat PIL will change with solvent polarity and temperature, which will result in new vibrational bands in the FIR spectra that indicate the presence of SIP configurations. The purpose of the present study is manifold. In the literature it is claimed that in general and with some exceptions, conventional spectroscopy detects only CIP species.[23–26] For example, Anderson et al. stated in Figure 1. FIR spectra of a) [Et4N][I] and [Et3NH][I] (TEAI) in polyethylene pellets, b) TEAI dissolved in mixtures of 60, in 80 mol % CDCl3 as a function of a recent article about the direct 80, and 90 mol % deuterated chloroform, CDCl3, at 298 K, and c) TEAI dissolved temperature between 298 and 323 K. The vibrational band at about 106 cm1 could be assigned to the cation– observation of CIP formation in anion interaction along the + NH···I bond, and indicates the CIP configurations. The shape and position of this aqueous solution: “Spectroscop- vibrational mode does not change with concentration or temperature, which indicates that all CIP configurations ic methods are the exception, are preserved. but these are prone to insufficient resolution, and are unable mation).[34] At this point we can conclude that the distinct vito detect solvent-separated ion pairing.”[26] Here, we show that FIR spectroscopy in the terahertz regime can be successfully brational band at 106 cm1 is well isolated and can certainly be assigned to the + NH···I hydrogen bond, thus indicating used to probe ion-pair species beyond CIPs. We then discuss the presence of exclusively CIP species in the pure PIL. In the conditions under which CIPs will dissociate and SIPs will be a second step, we dissolved TEAI in CDCl3 and obtained mixformed. Which parameters are needed to overcome the specific interaction between cation and anion? For which solvent tures containing 60, 80, and 90 mol % solvent concentrations. polarity and for which solvent concentrations can we expect No further PIL could be dissolved. The deuterated solvent was dissociation? Finally, in what way does temperature affect the chosen for better discrimination between the vibrational equilibrium between CIPs and SIPs? All these questions are adbands of the negative solvent and the positive TEAI and TEAI– dressed by FIR spectroscopy and supported by density funcsolvent contributions. In Figure 1 b the FIR spectra of the TEAI/ tional theory (DFT) calculations. CDCl3 mixtures are shown. The positive contributions result from the remaining TEAI and possibly TEAI/solvent interactions. The absorbance decreases with increasing solvent con2. Results and Discussion centration but the vibrational band at about 106 cm1, which describes the anion–cation interaction along the + NH···I First, we measured the FIR spectra of tetraethylammonium iodide, [Et4N][I], and triethylammonium iodide, [Et3NH][I] (TEAI), bond, remains even at the highest solvent concentration. All in the low frequency range between 10 and 200 cm1 as CIPs are essentially intact. This is also true for the spectra measured as a function of temperature. The spectra in Figure 1 c shown in Figure 1 a. As these salts are solid at room tempera 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2604 – 2609
2605
CHEMPHYSCHEM ARTICLES show neither changes in the CIP vibrational features nor the appearance of new bands. The + NH···I bond is so tight that the low polarity chloroform (e = 4.6) is unable to dissociate the CIPs at any concentrations or temperatures. The situation changes fundamentally on using deuterated dimethylsulfoxide (C2D6SO, [D6]DMSO) as solvent. Due to the significantly higher polarity of DMSO, which is expressed by the relatively high dielectric constant (e = 46), only mixtures containing 80 and 90 mol % solvent could be prepared. In Figure 2 a,b, the FIR spectra for both TEAI/[D6]DMSO mixtures are shown. This time, we observe strong negative contributions that result from the subtracted background spectra of [D6]DMSO. However, the vibrational mode at 106 cm1, which indicates the remaining CIP contributions, is still observed. Additionally we now detect a new vibrational band at 150 cm1. In recent work we could assign this vibrational band to the + NH···OS(CH3)2 interaction between the cation and solvent molecule in [Et3NH][CF3SO3]/DMSO mixtures.[21] The polarity of [D6]DMSO is clearly required to overcome the strong cation– anion interaction and to transfer CIPs to SIPs on increasing the solvent concentration. Considering the negative contributions by [D6]DMSO, all spectra could be properly deconvoluted (see the Supporting Information). An example is shown in Figure 3 for the mixture with 80 mol % [D6]DMSO at 353 K. The deconvoluted vibrational bands below 75 cm1 are probably related to the bending modes and unspecific cage rattling of both ion-pair species and solvent molecules but are not discussed in detail here. The assignment of all vibrational modes could be confirmed by DFT calculated frequencies for clusters including one TEAI ion pair and n = 1–9 solvent molecules. In particular the measured shift of about 50 cm1 between CIP and SIP vibrational modes is supported by the calculated anion–cation and cation–solvent frequencies (see the Supporting Information). If the CIP dissociates and a solvent molecule penetrates between the two ions, the SIP configuration is characterized by cation–solvent molecule vibrational modes at about 150 cm1 as shown in Figures 2 and 3. In principle, we are now able to determine the equilibrium constants from the integrated intensities of the CIP and SIP vibrational bands given in Table 1. However, the absorption coefficients for the vibrational bands of the cation–anion and the cation–solvent molecule interactions may be different. This was checked carefully for the calculated intensities of the related frequencies in the clusters given above that include one ion pair and n = 1–9 solvent molecules. The average CIP to SIP intensity ratios are 1.18 and 1.29 for the calculations with and without including dispersion forces, respectively. As the intensity ratio from the calculations that take dispersion forces into account should be more reliable, the value of 1.18 was used to calculate the equilibrium constants by K = ICIP*/ISIP (see Table 1). The calculated uncertainty for the intensity ratios given above is indicated by the error bars in Figures 4 and 5. The standard free energies DGq can be now derived from the logarithm of the equilibrium constant using DGq = RT ln (K). The free energies for the reactions from CIPs to SIPs are given in Table 1 as a function of [D6]DMSO concentration 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 2. FIR spectra of TEAI dissolved in mixtures of a) 80 mol % and b) 90 mol % deuterated dimethylsulfoxide [D6]DMSO as a function of temperature between 298 and 353 K. Beside the vibrational band at about 106 cm1, which indicates the cation–anion interaction along the + NH···I bond, a new band occurs at about 150 cm1 and represents the + N H···OS(CH3)2 interaction. For both mixtures it is observed that CIPs are reformed on increasing the temperature.
Figure 3. Deconvoluted FIR spectrum of TEAI dissolved in 80 mol % [D6]DMSO at 353 K. The contributions for the subtracted solvent (dashed lines) were taken from the background spectra of [D6]DMSO and multiplied by the solvent concentration of the corresponding mixture (here 0.8). As a result, we obtained the vibrational band for the + NH···I interaction at 106 cm1 (CIPs) and a new vibrational band at 150 cm1, which can be assigned to the cation–[D6]DMSO interaction, + NH···OS(CD3)2 (SIPs). The deconvoluted spectra for all solvent concentrations and temperatures are given in the Supporting Information.
and temperature. If the CIPs dominate, DGq is negative, and if the SIPs take over, DGq becomes positive. This is only the case for the 80 mol % mixture above 333 K. Interestingly, the vibrational band at 106 cm1 increases at the expense of the band at around 150 cm1 with increasing temperature, and this indicates the transfer from SIPs to CIPs. Assuming that the intensities for the CIP and SIP vibrational ChemPhysChem 2014, 15, 2604 – 2609
2606
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 1. Integrated intensities ICIP and ISIP from deconvoluted FIR spectra of TEAI/[D6]DMSO mixtures as a function of [D6]DMSO concentration and temperature.[a, b] T [K]
ICIP
ICIP*
ISIP
DGq [kJ mol1]
K
80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 298 0.8956 – 0.7589 – 1.7792 1.1694 0.426 – 1.53 – 303 0.8503 – 0.7206 – 1.6751 1.1049 0.430 – 1.54 – 313 1.0590 0.1511 0.8975 0.1281 1.5392 1.0329 0.583 0.124 0.81 4.83 323 1.1131 0.2682 0.9433 0.2273 1.3922 1.0855 0.678 0.209 0.43 3.59 333 1.3105 0.2819 1.1105 0.2389 1.2565 1.0279 0.884 0.232 0.05 3.42 343 1.8221 0.4061 1.5442 0.3442 1.2938 0.8833 1.194 0.390 1.14 2.05 353 2.6326 0.6327 2.2310 0.5362 1.4765 0.9052 1.511 0.592 1.86 0.88 [a] The ICIP* intensities are corrected for the different absorption coefficients for CIP and SIP vibrational modes as obtained from DFT calculations (see the Supporting Information). [b] The equilibrium constants and free energies could be derived for the ionic species present in the TEAI/[D6]DMSO mixtures.
Figure 4. Van‘t Hoff plots for the equilibrium constants, K, of the CIP/SIP ratios in TEAI/[D6]DMSO mixtures with 80 mol % (*) and 90 mol % (&) [D6]DMSO. For both mixtures the equilibrium shifts towards higher CIP concentration with increasing temperature.
bands change in the same way with temperature, we could also evaluate the equilibrium constants as a function of temperature for both mixtures. The data are listed in Table 1 for the 80 and 90 mol % mixtures, respectively. We could then plot the equilibrium constants versus the reciprocal temperature, which resulted in the Van ‘t Hoff plots expressed by Equation (1): lnðKÞ ¼
DHq DSq þ RT R
ð1Þ
The standard molar enthalpy changes DHq for the back reactions from SIPs to CIPs can be taken from the slopes of the curves. We find DHq values of 20.6 kJ mol1 for the 80 mol % mixture and 34.4 kJ mol1 for the 90 mol % mixtures, respectively (Figure 4). It is not surprising that the enthalpy changes for the conversion from SIPs to CIPs is higher for the 90 mol % mixture. Due to the higher DMSO concentration more energy is required for the back-formation from SIPs to CIPs. Such tem 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 5. Linear relationship between ln (K) for the TEAI mixtures with 80 mol % [D6]DMSO and the dielectric constant, e, of DMSO for all temperatures between 298 K and 353 K.[36]
perature-dependent behavior has been reported for aqueous salt solutions by using molecular dynamics simulations.[35] As apparent from Equation (1) we also have access to the standard molar entropy changes DSq, which are given in Table 2. The entropies, DSq, are 62.7 J mol1 K1 for the 80 mol % mixture, and 93.9 J mol1 K1 for the 90 mol % mixture, respectively. The positive values of DSq signify that several solvent molecules are released into the bulk solvent from the SIPs on formation of the CIPs. This release is stronger for the 90 mol % mixture, as expected. In recent studies for mixtures of [Et3NH][CH3SO3] and water we could show that the back-formation of CIPs is probably
Table 2. Enthalpies and entropies for the equilibrium between CIPs and SIPs in the 80 and 90 mol % mixtures of TEAI/[D6]DMSO. [D6]DMSO
DHq [kJ mol1]
DSq [J mol1 K1]
80 mol % 90 mol %
+ 20.1 + 34.4
+ 62.7 + 93.9
ChemPhysChem 2014, 15, 2604 – 2609
2607
CHEMPHYSCHEM ARTICLES due to entropic reasons. In the related SIP configurations, the water molecules were tightly bound between cation and anion due to strong cooperative effects.[21] The significantly reduced motional degrees of freedom resulted in an entropic penalty as indicated by the calculated Gibbs free energies for both ionpair configurations. For the system under investigation we did not find such behavior. For the spherical iodide and the solvent DMSO, the favorable cooperative interaction that was observed in the previous system, + NH···OH···OS(CH3)2, is not possible. Instead we summarize the variation of ion-pair dissociation constants with change in solvent polarity as a linear plot of ln (K) against the DMSO permittivity, e, for the 80 mol % mixture TEAI/[D6]DMSO at each temperature (see Figure 5). As shown in Figure 6, the dielectric constant of DMSO decreases from 46 to 40.6 over the temperature range between 298 K and 353 K.[36] It is predominantly an experimental finding
www.chemphyschem.org additional dependencies on the distance between cation and anion or ion size parameters had to be taken into account to fit the experimental data properly,[37–40] but this is not required here. In the PIL under investigation the cation–anion distance is more or less constant within the CIP configurations.
3. Conclusions Here, we show that FIR spectroscopy is sufficiently sensitive for the detection of ion-pair species beyond CIPs. Supported by DFT calculated frequencies and intensities, we could derive dissociation constants that describe the equilibrium between CIPs and SIPs. The equilibrium was studied as a function of solvent concentration, solvent polarity, and temperature. The reformation of CIPs on increasing temperature could be ascribed to the decreasing dielectric constant rather than entropic effects, as reported in other cases. Our measurements also provide important thermodynamic information. Free energy, enthalpy, and entropy changes characterize the dissociation and backformation of CIPs in PILs. However, such a quantitative analysis is only possible if either ion pair does not fall below the ratio of 1:10. Of course, the interconversion reaction from CIPs to SIPs that we have studied here involves a concerted sequence of local solvent molecule reorganization coupled to the interionic separation coordinate. This will be studied in detail by molecular dynamics (MD) simulations which are currently being performed in our laboratory.
Experimental Section All ILs were dried in a vacuum (p = 8 103 mbar) for approximately 36 h. The water content was then determined by Karl Fischer titration. Further purification was not carried out.
Figure 6. The dielectric constant, e, of pure DMSO decreases linearly with increasing temperature.[36]
that the equilibrium constants can be related to the temperature dependent dielectric constant. At the moment we have no molecular interpretation for this behavior. Experimentally, we can summarize that high polarity of DMSO is necessary to disrupt the CIPs, and that these CIPs are reformed with increasing temperature, and this is related to decreasing dielectric constants. However, another interpretation is also reasonable. Compared to CHCl3, the S=O of the DMSO molecule provides a clear proton acceptor ability. The possible formation of a local and directional hydrogen bond may be energetically and entropically favored over the cation–anion interaction. The dominant factor, the dielectric constant or the hydrogen-bonding ability, could be determined experimentally by using solvent molecules with low dielectric constant but a clear proton acceptor function or vice versa. In earlier studies of the relations between the association constant and the inversed permittivity by Fuoss and others, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The FTIR measurements were performed with a Bruker Vertex 70 FTIR spectrometer equipped with an extension for measurements in the FIR region that consists of a multilayer mylar beam splitter and a room temperature DLATGS detector with preamplifier. Polyethylene (PE) windows with an internal optical path of 0.1 mm were used. Further improvement could be achieved by using a high pressure mercury lamp and a silica beam splitter. The accessible spectral region for this configuration now lies between 10 and 680 cm1 (0.3 and 20.3 THz). The spectra were deconvoluted simultaneously as well as separately into a number of Voigt profiles (convolution of Lorentzian and Gaussian functions) following the Levenberg–Marquardt procedure. The Voigt profile has four parameters: the intensity, the frequency, the half-width of the Lorentzian, and the half-width of the Gaussian. The deconvolution procedure is described in detail in the Supporting Information. The geometries and frequencies of ion-pair aggregates TEAI and of one TEAI ion pair dissolved in an environment of DMSO molecules (n = 1–9) were calculated at the DFT B3LYP level, using the def2– SVPD basis set developed by Rappoport and Furche[32] and the ECP for iodide (see the Supporting Information) proposed by Peterson et al. as implemented in the Turbomole program.[27–33] Different conformers were calculated for all structures but only the best in energy were considered. Grimme’s DFT–D3 method was applied for the calculation of dispersion forces.[34] Different conformers were calculated for all structures but only the best in energy were ChemPhysChem 2014, 15, 2604 – 2609
2608
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
considered. The structures and geometries of all CIP and SIP clusters are given in the Supporting Information.
Acknowledgements This work has been supported by the DFG Collaborative Research Center SFB 652 “Strong correlations and collective effects in radiation fields: Coulomb systems, clusters and particles” R. L. gratefully acknowledges the support of the Leibniz-Institute for Catalysis. Keywords: density functional calculations spectroscopy · ion pairs · ionic liquids · polarity
·
FTIR
[1] P. Wasserscheid, T. Welton, Ionic Liquids in Synthesis, 2nd ed., Wiley-VCH, Weinheim, 2008. [2] F. Endres, S. Z. El Abedin, Phys. Chem. Chem. Phys. 2006, 8, 2101 – 2116. [3] H. Weingrtner, Angew. Chem. 2008, 120, 664 – 682; Angew. Chem. Int. Ed. 2008, 47, 654 – 670. [4] N. V. Plechkova, K. R. Seddon, Chem. Soc. Rev. 2008, 37, 123 – 150. [5] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2008, 120, 3890 – 3894; Angew. Chem. Int. Ed. 2008, 47, 3830 – 3834. [6] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2008, 120, 8859 – 8862; Angew. Chem. Int. Ed. 2008, 47, 8731 – 8734. [7] K. Fumino, T. Peppel, M. Geppert-Rybczynska, D. H. Zaitsau, J. K. Lehmann, S. P. Verevkin, M. Kçckerling, R. Ludwig, Phys. Chem. Chem. Phys. 2011, 13, 14064 – 14075. [8] A. Wulf, K. Fumino, R. Ludwig, Angew. Chem. 2010, 122, 459 – 463; Angew. Chem. Int. Ed. 2010, 49, 449 – 453. [9] J. Dupont, J. Braz. Chem. Soc. 2004, 15, 341 – 350. [10] T. L. Greaves, C. J. Drummond, Chem. Rev. 2008, 108, 206 – 237. [11] C. A. Angell, N. Byrne, J.-P. Belieres, Acc. Chem. Res. 2007, 40, 1228 – 1236. [12] M. Yoshizawa, W. Xu, C. A. Angell, D. R. MacFarlane, Phys. Chem. Chem. Phys. 2010, 12, 10341 – 10347. [13] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2009, 121, 3230 – 3233; Angew. Chem. Int. Ed. 2009, 48, 3184 – 3186. [14] K. Fumino, E. Reichert, K. Wittler, R. Hempelmann, R. Ludwig, Angew. Chem. 2012, 124, 6340 – 6344; Angew. Chem. Int. Ed. 2012, 51, 6236 – 6240.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[15] R. Hayes, S. Imberti, G. G. Warr, R. Atkin, Angew. Chem. 2012, 124, 7586 – 7589; Angew. Chem. Int. Ed. 2012, 51, 7468 – 7471. [16] Y. Jiang, H. Nadolny, S. Kshammer, S. Weibels, W. Schrçer, H. Weingrtner, Faraday Discuss. 2012, 154, 391 – 407. [17] T. Kçddermann, S. Klembt, D. Klasen, D. Paschek, U. Kragl, R. Ludwig, ChemPhysChem 2012, 13, 1748 – 1752. [18] J. Hunger, A. Stoppa, R. Buchner, G. Hefter, J. Phys. Chem. B 2008, 112, 12913 – 12919. [19] K. Fumino, V. Fossog, K. Wittler, R. Hempelmann, R. Ludwig, Angew. Chem. 2013, 125, 2425 – 2429; Angew. Chem. Int. Ed. 2013, 52, 2368 – 2372. [20] P. Stange, K. Fumino, R. Ludwig, Angew. Chem. 2013, 125, 3064 – 3068; Angew. Chem. Int. Ed. 2013, 52, 2990 – 2994. [21] K. Fumino, P. Stange, V. Fossog, R. Hempelmann, R. Ludwig, Angew. Chem. 2013, 125, 12667 – 12670; Angew. Chem. Int. Ed. 2013, 52, 12439 – 12442. [22] a) M. Eigen, K. Tamm, Z. Elektrochem. 1962, 66, 93 – 107; b) M. Eigen, K. Tamm, Z. Elektrochem. 1962, 66, 107 – 121. [23] Y. Marcus, G. Hefter, Chem. Rev. 2006, 106, 4585 – 4651. [24] G. Hefter, Pure Appl. Chem. 2006, 78, 1571 – 1586. [25] R. Buchner, Pure Appl. Chem. 2008, 80, 1239 – 1252. [26] M. R. Antonio, M. Nyman, T. M. Anderson, Angew. Chem. 2009, 121, 6252 – 6256; Angew. Chem. Int. Ed. 2009, 48, 6136 – 6140. [27] O. Treutler, R. Ahlrichs, J. Chem. Phys. 1995, 102, 346 – 354. [28] M. Von Arnim, R. Ahlrichs, J. Comput. Chem. 1998, 19, 1746 – 1757. [29] P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2004, 384, 103 – 107. [30] P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2002, 362, 511 – 518. [31] P. Deglmann, F. Furche, J. Chem. Phys. 2002, 117, 9535 – 9538. [32] D. Rappoport, F. Furche, J. Chem. Phys. 2010, 133, 134105. [33] K. A. Peterson, D. Figgen, E. Goll, H. Stoll, M. Dolg, J. Chem. Phys. 2003, 119, 11113 – 11123. [34] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104. [35] L. X. Dang, J. Chem. Phys. 1992, 97, 1919 – 1921. [36] J. F. Casteel, P. G. Sears, J. Chem. Eng. Data 1974, 19, 196 – 200. [37] J. V. Walther, J. Schott, Nature 1988, 332, 635 – 638. [38] M. J. Blandamer, Nature 1967, 214, 380 – 381. [39] H. Sadek, R. Fuoss, J. Am. Chem. Soc. 1954, 76, 5897 – 5901. [40] E. Grunwald, Anal. Chem. 1954, 26, 1696 – 1701. Received: April 4, 2014 Published online on June 12, 2014
ChemPhysChem 2014, 15, 2604 – 2609
2609
Ion Pairing in Protic Ionic Liquids Probed by Far-Infrared Spectroscopy: Effects of Solvent Polarity and Temperature Koichi Fumino,[a] Verlaine Fossog,[c] Peter Stange,[a] Kai Wittler,[a] Wigbert Polet,[a] Rolf Hempelmann,[c] and Ralf Ludwig*[a, b] The cation–anion and cation–solvent interactions in solutions of the protic ionic liquid (PIL) [Et3NH][I] dissolved in solvents of different polarities are studied by means of far infrared vibrational (FIR) spectroscopy and density functional theory (DFT) calculations. The dissociation of contact ion pairs (CIPs) and the resulting formation of solvent-separated ion pairs (SIPs) can be observed and analyzed as a function of solvent concentration, solvent polarity, and temperature. In apolar environments, the CIPs dominate for all solvent concentrations and
temperatures. At high concentrations of polar solvents, SIPs are favored over CIPs. For these PIL/solvent mixtures, CIPs are reformed by increasing the temperature due to the reduced polarity of the solvent. Overall, this approach provides equilibrium constants, free energies, enthalpies, and entropies for ion-pair formation in trialkylammonium-containing PILs. These results have important implications for the understanding of solvation chemistry and the reactivity of ionic liquids.
1. Introduction Ionic liquids (ILs) are liquids that consist entirely of ions. The potential applications of these Coulomb fluids depend on the properties of the particular liquid material.[1–4] To a large extent, the structure and properties of ILs are determined by the intermolecular interactions between anions and cations.[5, 6] In particular, the subtle balance between Coulomb forces, hydrogen bonds, and dispersion forces is of great importance in the understanding of ILs.[7, 8] This mlange of interactions of varying type, strength, and directionality leads to the formation of charged and neutral clusters, and ion pairs.[9] Moreover, ion pairing is crucial in the determination of the properties of Coulomb fluids. Two types of ILs are of increasing interest in science and technology: aprotic and protic ILs.[1–4, 10–16] It is known that the addition of hydrophobic organic solvents to aprotic ILs leads to electrical conductance minima and liquid–liquid miscibility gaps.[17, 18] In highly diluted systems, neutral aggregates and contact ion pairs (CIPs) are formed. This situation is [a] Dr. K. Fumino, P. Stange, K. Wittler, W. Polet, Prof. Dr. R. Ludwig Universitt Rostock Institut fr Chemie Abteilung fr Physikalische Chemie Dr.-Lorenz-Weg 1 18059, Rostock (Germany) E-mail: [email protected] [b] Prof. Dr. R. Ludwig Leibniz-Institut fr Katalyse an der Universitt Rostock e.V. Albert-Einstein-Str. 29a 18059 Rostock (Germany) [c] V. Fossog, Prof. Dr. R. Hempelmann Universitt des Saarlandes Physikalische Chemie 66123 Saarbrcken (Germany) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402205.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
similar to that of simple electrolytes in dilute solutions.[18] For protic ILs (PILs), the initial position can be the opposite. Possible proton transfer from the Brønsted acid to the Brønsted base leads to the formation of distinct proton acceptor and donor sites and results in strong hydrogen bonding.[10–16] This holds in particular for PILs that include the triethylammonium ion [Et3NH] + as the cation. Beside the Coulomb interaction, a strong and directional hydrogen bond between the NH group of the cation and a polar atom of the anion provides strongly bound CIPs in the pure Coulomb fluid.[19–21] Although each ion is surrounded by counterions, the tight cation–anion interaction leads to the existence of quasi CIPs in the neat PIL. We recently showed that this specific cation–anion interaction gives rise to distinct vibrational bands in the low frequency range between 100 and 180 cm1 that depend on the interaction potential of the involved ions.[14, 19] Thus, far-infrared (FIR) spectroscopy is expected to be a suitable method for the study of ion pairing in this type of PIL. Usually, ion pairing proceeds in stages as originally suggested by Eigen and Tamm.[22] In the first stage, free solvated ions form a doubly solvent-separated ion pair (2SIP), in which both ions essentially retain their first solvation shells. In the second stage, partial desolvation of 2SIP occurs to form solvent-shared ion pairs (SIP). In the final stage, the solvent molecules between the ions are eliminated to form CIPs. All species are in chemical equilibrium, which is described by ion association and solvent elimination constants.[23–25] The process for PILs described in this work is the reverse of that process. We start with CIPs that are characterized by long-range Coulomb interactions, which are strongly enhanced by short-range, spatially directed hydrogen-bonding interactions between the Lewis acid and the Lewis base. We then study the dissociation of CIPs into SIPs under different solvent concentrations and polarChemPhysChem 2014, 15, 2604 – 2609
2604
CHEMPHYSCHEM ARTICLES ities expressed by the dielectric constant, e. Thus, the dissociation constant can be regarded as the equilibrium constant that describes the equilibrium between CIPs and SIPs (see Scheme 1). The solvent concentrations used here do not allow further transfer from SIPs to 2SIPs. The maximum 9:1 ratio between solvent molecules and CIPs does not provide sufficient
Scheme 1. Ion-pair concept as defined for the PIL [Et3NH][I]. Here, dimethyl sulfoxide (DMSO) is used as the solvent molecule.
www.chemphyschem.org ture, they were measured in pellets of polyethylene. Both spectra share a common vibrational feature around 70 cm1 that is usually attributed to cage rattling of interacting ions and bending modes of directional intermolecular interaction.[19, 20] The additional vibrational band at 106 cm1 for TEAI can be clearly assigned to the + NH···I interaction, which is strongly enhanced by hydrogen bonding.[14] Consequently, this feature is missing in the spectrum of [Et4N][I]. This finding is supported by DFT calculated frequencies of aggregates [Et3NH][I]n with n = 1–3. The calculated average value for all species of about 103 cm1 is in perfect agreement with the measured frequency.[27–33] If we include Grimme’s DFT–D3 method for calculating the non-covalent interactions, we obtain a value of 114 cm1, which is still reasonable considering that no correction for the harmonic approximation is applied (see the Supporting Infor-
solvent molecules to form full solvation shells for each ion. It is expected that the purely CIP contributions in the neat PIL will change with solvent polarity and temperature, which will result in new vibrational bands in the FIR spectra that indicate the presence of SIP configurations. The purpose of the present study is manifold. In the literature it is claimed that in general and with some exceptions, conventional spectroscopy detects only CIP species.[23–26] For example, Anderson et al. stated in Figure 1. FIR spectra of a) [Et4N][I] and [Et3NH][I] (TEAI) in polyethylene pellets, b) TEAI dissolved in mixtures of 60, in 80 mol % CDCl3 as a function of a recent article about the direct 80, and 90 mol % deuterated chloroform, CDCl3, at 298 K, and c) TEAI dissolved temperature between 298 and 323 K. The vibrational band at about 106 cm1 could be assigned to the cation– observation of CIP formation in anion interaction along the + NH···I bond, and indicates the CIP configurations. The shape and position of this aqueous solution: “Spectroscop- vibrational mode does not change with concentration or temperature, which indicates that all CIP configurations ic methods are the exception, are preserved. but these are prone to insufficient resolution, and are unable mation).[34] At this point we can conclude that the distinct vito detect solvent-separated ion pairing.”[26] Here, we show that FIR spectroscopy in the terahertz regime can be successfully brational band at 106 cm1 is well isolated and can certainly be assigned to the + NH···I hydrogen bond, thus indicating used to probe ion-pair species beyond CIPs. We then discuss the presence of exclusively CIP species in the pure PIL. In the conditions under which CIPs will dissociate and SIPs will be a second step, we dissolved TEAI in CDCl3 and obtained mixformed. Which parameters are needed to overcome the specific interaction between cation and anion? For which solvent tures containing 60, 80, and 90 mol % solvent concentrations. polarity and for which solvent concentrations can we expect No further PIL could be dissolved. The deuterated solvent was dissociation? Finally, in what way does temperature affect the chosen for better discrimination between the vibrational equilibrium between CIPs and SIPs? All these questions are adbands of the negative solvent and the positive TEAI and TEAI– dressed by FIR spectroscopy and supported by density funcsolvent contributions. In Figure 1 b the FIR spectra of the TEAI/ tional theory (DFT) calculations. CDCl3 mixtures are shown. The positive contributions result from the remaining TEAI and possibly TEAI/solvent interactions. The absorbance decreases with increasing solvent con2. Results and Discussion centration but the vibrational band at about 106 cm1, which describes the anion–cation interaction along the + NH···I First, we measured the FIR spectra of tetraethylammonium iodide, [Et4N][I], and triethylammonium iodide, [Et3NH][I] (TEAI), bond, remains even at the highest solvent concentration. All in the low frequency range between 10 and 200 cm1 as CIPs are essentially intact. This is also true for the spectra measured as a function of temperature. The spectra in Figure 1 c shown in Figure 1 a. As these salts are solid at room tempera 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2604 – 2609
2605
CHEMPHYSCHEM ARTICLES show neither changes in the CIP vibrational features nor the appearance of new bands. The + NH···I bond is so tight that the low polarity chloroform (e = 4.6) is unable to dissociate the CIPs at any concentrations or temperatures. The situation changes fundamentally on using deuterated dimethylsulfoxide (C2D6SO, [D6]DMSO) as solvent. Due to the significantly higher polarity of DMSO, which is expressed by the relatively high dielectric constant (e = 46), only mixtures containing 80 and 90 mol % solvent could be prepared. In Figure 2 a,b, the FIR spectra for both TEAI/[D6]DMSO mixtures are shown. This time, we observe strong negative contributions that result from the subtracted background spectra of [D6]DMSO. However, the vibrational mode at 106 cm1, which indicates the remaining CIP contributions, is still observed. Additionally we now detect a new vibrational band at 150 cm1. In recent work we could assign this vibrational band to the + NH···OS(CH3)2 interaction between the cation and solvent molecule in [Et3NH][CF3SO3]/DMSO mixtures.[21] The polarity of [D6]DMSO is clearly required to overcome the strong cation– anion interaction and to transfer CIPs to SIPs on increasing the solvent concentration. Considering the negative contributions by [D6]DMSO, all spectra could be properly deconvoluted (see the Supporting Information). An example is shown in Figure 3 for the mixture with 80 mol % [D6]DMSO at 353 K. The deconvoluted vibrational bands below 75 cm1 are probably related to the bending modes and unspecific cage rattling of both ion-pair species and solvent molecules but are not discussed in detail here. The assignment of all vibrational modes could be confirmed by DFT calculated frequencies for clusters including one TEAI ion pair and n = 1–9 solvent molecules. In particular the measured shift of about 50 cm1 between CIP and SIP vibrational modes is supported by the calculated anion–cation and cation–solvent frequencies (see the Supporting Information). If the CIP dissociates and a solvent molecule penetrates between the two ions, the SIP configuration is characterized by cation–solvent molecule vibrational modes at about 150 cm1 as shown in Figures 2 and 3. In principle, we are now able to determine the equilibrium constants from the integrated intensities of the CIP and SIP vibrational bands given in Table 1. However, the absorption coefficients for the vibrational bands of the cation–anion and the cation–solvent molecule interactions may be different. This was checked carefully for the calculated intensities of the related frequencies in the clusters given above that include one ion pair and n = 1–9 solvent molecules. The average CIP to SIP intensity ratios are 1.18 and 1.29 for the calculations with and without including dispersion forces, respectively. As the intensity ratio from the calculations that take dispersion forces into account should be more reliable, the value of 1.18 was used to calculate the equilibrium constants by K = ICIP*/ISIP (see Table 1). The calculated uncertainty for the intensity ratios given above is indicated by the error bars in Figures 4 and 5. The standard free energies DGq can be now derived from the logarithm of the equilibrium constant using DGq = RT ln (K). The free energies for the reactions from CIPs to SIPs are given in Table 1 as a function of [D6]DMSO concentration 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 2. FIR spectra of TEAI dissolved in mixtures of a) 80 mol % and b) 90 mol % deuterated dimethylsulfoxide [D6]DMSO as a function of temperature between 298 and 353 K. Beside the vibrational band at about 106 cm1, which indicates the cation–anion interaction along the + NH···I bond, a new band occurs at about 150 cm1 and represents the + N H···OS(CH3)2 interaction. For both mixtures it is observed that CIPs are reformed on increasing the temperature.
Figure 3. Deconvoluted FIR spectrum of TEAI dissolved in 80 mol % [D6]DMSO at 353 K. The contributions for the subtracted solvent (dashed lines) were taken from the background spectra of [D6]DMSO and multiplied by the solvent concentration of the corresponding mixture (here 0.8). As a result, we obtained the vibrational band for the + NH···I interaction at 106 cm1 (CIPs) and a new vibrational band at 150 cm1, which can be assigned to the cation–[D6]DMSO interaction, + NH···OS(CD3)2 (SIPs). The deconvoluted spectra for all solvent concentrations and temperatures are given in the Supporting Information.
and temperature. If the CIPs dominate, DGq is negative, and if the SIPs take over, DGq becomes positive. This is only the case for the 80 mol % mixture above 333 K. Interestingly, the vibrational band at 106 cm1 increases at the expense of the band at around 150 cm1 with increasing temperature, and this indicates the transfer from SIPs to CIPs. Assuming that the intensities for the CIP and SIP vibrational ChemPhysChem 2014, 15, 2604 – 2609
2606
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 1. Integrated intensities ICIP and ISIP from deconvoluted FIR spectra of TEAI/[D6]DMSO mixtures as a function of [D6]DMSO concentration and temperature.[a, b] T [K]
ICIP
ICIP*
ISIP
DGq [kJ mol1]
K
80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 80 mol % 90 mol % 298 0.8956 – 0.7589 – 1.7792 1.1694 0.426 – 1.53 – 303 0.8503 – 0.7206 – 1.6751 1.1049 0.430 – 1.54 – 313 1.0590 0.1511 0.8975 0.1281 1.5392 1.0329 0.583 0.124 0.81 4.83 323 1.1131 0.2682 0.9433 0.2273 1.3922 1.0855 0.678 0.209 0.43 3.59 333 1.3105 0.2819 1.1105 0.2389 1.2565 1.0279 0.884 0.232 0.05 3.42 343 1.8221 0.4061 1.5442 0.3442 1.2938 0.8833 1.194 0.390 1.14 2.05 353 2.6326 0.6327 2.2310 0.5362 1.4765 0.9052 1.511 0.592 1.86 0.88 [a] The ICIP* intensities are corrected for the different absorption coefficients for CIP and SIP vibrational modes as obtained from DFT calculations (see the Supporting Information). [b] The equilibrium constants and free energies could be derived for the ionic species present in the TEAI/[D6]DMSO mixtures.
Figure 4. Van‘t Hoff plots for the equilibrium constants, K, of the CIP/SIP ratios in TEAI/[D6]DMSO mixtures with 80 mol % (*) and 90 mol % (&) [D6]DMSO. For both mixtures the equilibrium shifts towards higher CIP concentration with increasing temperature.
bands change in the same way with temperature, we could also evaluate the equilibrium constants as a function of temperature for both mixtures. The data are listed in Table 1 for the 80 and 90 mol % mixtures, respectively. We could then plot the equilibrium constants versus the reciprocal temperature, which resulted in the Van ‘t Hoff plots expressed by Equation (1): lnðKÞ ¼
DHq DSq þ RT R
ð1Þ
The standard molar enthalpy changes DHq for the back reactions from SIPs to CIPs can be taken from the slopes of the curves. We find DHq values of 20.6 kJ mol1 for the 80 mol % mixture and 34.4 kJ mol1 for the 90 mol % mixtures, respectively (Figure 4). It is not surprising that the enthalpy changes for the conversion from SIPs to CIPs is higher for the 90 mol % mixture. Due to the higher DMSO concentration more energy is required for the back-formation from SIPs to CIPs. Such tem 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 5. Linear relationship between ln (K) for the TEAI mixtures with 80 mol % [D6]DMSO and the dielectric constant, e, of DMSO for all temperatures between 298 K and 353 K.[36]
perature-dependent behavior has been reported for aqueous salt solutions by using molecular dynamics simulations.[35] As apparent from Equation (1) we also have access to the standard molar entropy changes DSq, which are given in Table 2. The entropies, DSq, are 62.7 J mol1 K1 for the 80 mol % mixture, and 93.9 J mol1 K1 for the 90 mol % mixture, respectively. The positive values of DSq signify that several solvent molecules are released into the bulk solvent from the SIPs on formation of the CIPs. This release is stronger for the 90 mol % mixture, as expected. In recent studies for mixtures of [Et3NH][CH3SO3] and water we could show that the back-formation of CIPs is probably
Table 2. Enthalpies and entropies for the equilibrium between CIPs and SIPs in the 80 and 90 mol % mixtures of TEAI/[D6]DMSO. [D6]DMSO
DHq [kJ mol1]
DSq [J mol1 K1]
80 mol % 90 mol %
+ 20.1 + 34.4
+ 62.7 + 93.9
ChemPhysChem 2014, 15, 2604 – 2609
2607
CHEMPHYSCHEM ARTICLES due to entropic reasons. In the related SIP configurations, the water molecules were tightly bound between cation and anion due to strong cooperative effects.[21] The significantly reduced motional degrees of freedom resulted in an entropic penalty as indicated by the calculated Gibbs free energies for both ionpair configurations. For the system under investigation we did not find such behavior. For the spherical iodide and the solvent DMSO, the favorable cooperative interaction that was observed in the previous system, + NH···OH···OS(CH3)2, is not possible. Instead we summarize the variation of ion-pair dissociation constants with change in solvent polarity as a linear plot of ln (K) against the DMSO permittivity, e, for the 80 mol % mixture TEAI/[D6]DMSO at each temperature (see Figure 5). As shown in Figure 6, the dielectric constant of DMSO decreases from 46 to 40.6 over the temperature range between 298 K and 353 K.[36] It is predominantly an experimental finding
www.chemphyschem.org additional dependencies on the distance between cation and anion or ion size parameters had to be taken into account to fit the experimental data properly,[37–40] but this is not required here. In the PIL under investigation the cation–anion distance is more or less constant within the CIP configurations.
3. Conclusions Here, we show that FIR spectroscopy is sufficiently sensitive for the detection of ion-pair species beyond CIPs. Supported by DFT calculated frequencies and intensities, we could derive dissociation constants that describe the equilibrium between CIPs and SIPs. The equilibrium was studied as a function of solvent concentration, solvent polarity, and temperature. The reformation of CIPs on increasing temperature could be ascribed to the decreasing dielectric constant rather than entropic effects, as reported in other cases. Our measurements also provide important thermodynamic information. Free energy, enthalpy, and entropy changes characterize the dissociation and backformation of CIPs in PILs. However, such a quantitative analysis is only possible if either ion pair does not fall below the ratio of 1:10. Of course, the interconversion reaction from CIPs to SIPs that we have studied here involves a concerted sequence of local solvent molecule reorganization coupled to the interionic separation coordinate. This will be studied in detail by molecular dynamics (MD) simulations which are currently being performed in our laboratory.
Experimental Section All ILs were dried in a vacuum (p = 8 103 mbar) for approximately 36 h. The water content was then determined by Karl Fischer titration. Further purification was not carried out.
Figure 6. The dielectric constant, e, of pure DMSO decreases linearly with increasing temperature.[36]
that the equilibrium constants can be related to the temperature dependent dielectric constant. At the moment we have no molecular interpretation for this behavior. Experimentally, we can summarize that high polarity of DMSO is necessary to disrupt the CIPs, and that these CIPs are reformed with increasing temperature, and this is related to decreasing dielectric constants. However, another interpretation is also reasonable. Compared to CHCl3, the S=O of the DMSO molecule provides a clear proton acceptor ability. The possible formation of a local and directional hydrogen bond may be energetically and entropically favored over the cation–anion interaction. The dominant factor, the dielectric constant or the hydrogen-bonding ability, could be determined experimentally by using solvent molecules with low dielectric constant but a clear proton acceptor function or vice versa. In earlier studies of the relations between the association constant and the inversed permittivity by Fuoss and others, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The FTIR measurements were performed with a Bruker Vertex 70 FTIR spectrometer equipped with an extension for measurements in the FIR region that consists of a multilayer mylar beam splitter and a room temperature DLATGS detector with preamplifier. Polyethylene (PE) windows with an internal optical path of 0.1 mm were used. Further improvement could be achieved by using a high pressure mercury lamp and a silica beam splitter. The accessible spectral region for this configuration now lies between 10 and 680 cm1 (0.3 and 20.3 THz). The spectra were deconvoluted simultaneously as well as separately into a number of Voigt profiles (convolution of Lorentzian and Gaussian functions) following the Levenberg–Marquardt procedure. The Voigt profile has four parameters: the intensity, the frequency, the half-width of the Lorentzian, and the half-width of the Gaussian. The deconvolution procedure is described in detail in the Supporting Information. The geometries and frequencies of ion-pair aggregates TEAI and of one TEAI ion pair dissolved in an environment of DMSO molecules (n = 1–9) were calculated at the DFT B3LYP level, using the def2– SVPD basis set developed by Rappoport and Furche[32] and the ECP for iodide (see the Supporting Information) proposed by Peterson et al. as implemented in the Turbomole program.[27–33] Different conformers were calculated for all structures but only the best in energy were considered. Grimme’s DFT–D3 method was applied for the calculation of dispersion forces.[34] Different conformers were calculated for all structures but only the best in energy were ChemPhysChem 2014, 15, 2604 – 2609
2608
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
considered. The structures and geometries of all CIP and SIP clusters are given in the Supporting Information.
Acknowledgements This work has been supported by the DFG Collaborative Research Center SFB 652 “Strong correlations and collective effects in radiation fields: Coulomb systems, clusters and particles” R. L. gratefully acknowledges the support of the Leibniz-Institute for Catalysis. Keywords: density functional calculations spectroscopy · ion pairs · ionic liquids · polarity
·
FTIR
[1] P. Wasserscheid, T. Welton, Ionic Liquids in Synthesis, 2nd ed., Wiley-VCH, Weinheim, 2008. [2] F. Endres, S. Z. El Abedin, Phys. Chem. Chem. Phys. 2006, 8, 2101 – 2116. [3] H. Weingrtner, Angew. Chem. 2008, 120, 664 – 682; Angew. Chem. Int. Ed. 2008, 47, 654 – 670. [4] N. V. Plechkova, K. R. Seddon, Chem. Soc. Rev. 2008, 37, 123 – 150. [5] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2008, 120, 3890 – 3894; Angew. Chem. Int. Ed. 2008, 47, 3830 – 3834. [6] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2008, 120, 8859 – 8862; Angew. Chem. Int. Ed. 2008, 47, 8731 – 8734. [7] K. Fumino, T. Peppel, M. Geppert-Rybczynska, D. H. Zaitsau, J. K. Lehmann, S. P. Verevkin, M. Kçckerling, R. Ludwig, Phys. Chem. Chem. Phys. 2011, 13, 14064 – 14075. [8] A. Wulf, K. Fumino, R. Ludwig, Angew. Chem. 2010, 122, 459 – 463; Angew. Chem. Int. Ed. 2010, 49, 449 – 453. [9] J. Dupont, J. Braz. Chem. Soc. 2004, 15, 341 – 350. [10] T. L. Greaves, C. J. Drummond, Chem. Rev. 2008, 108, 206 – 237. [11] C. A. Angell, N. Byrne, J.-P. Belieres, Acc. Chem. Res. 2007, 40, 1228 – 1236. [12] M. Yoshizawa, W. Xu, C. A. Angell, D. R. MacFarlane, Phys. Chem. Chem. Phys. 2010, 12, 10341 – 10347. [13] K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. 2009, 121, 3230 – 3233; Angew. Chem. Int. Ed. 2009, 48, 3184 – 3186. [14] K. Fumino, E. Reichert, K. Wittler, R. Hempelmann, R. Ludwig, Angew. Chem. 2012, 124, 6340 – 6344; Angew. Chem. Int. Ed. 2012, 51, 6236 – 6240.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[15] R. Hayes, S. Imberti, G. G. Warr, R. Atkin, Angew. Chem. 2012, 124, 7586 – 7589; Angew. Chem. Int. Ed. 2012, 51, 7468 – 7471. [16] Y. Jiang, H. Nadolny, S. Kshammer, S. Weibels, W. Schrçer, H. Weingrtner, Faraday Discuss. 2012, 154, 391 – 407. [17] T. Kçddermann, S. Klembt, D. Klasen, D. Paschek, U. Kragl, R. Ludwig, ChemPhysChem 2012, 13, 1748 – 1752. [18] J. Hunger, A. Stoppa, R. Buchner, G. Hefter, J. Phys. Chem. B 2008, 112, 12913 – 12919. [19] K. Fumino, V. Fossog, K. Wittler, R. Hempelmann, R. Ludwig, Angew. Chem. 2013, 125, 2425 – 2429; Angew. Chem. Int. Ed. 2013, 52, 2368 – 2372. [20] P. Stange, K. Fumino, R. Ludwig, Angew. Chem. 2013, 125, 3064 – 3068; Angew. Chem. Int. Ed. 2013, 52, 2990 – 2994. [21] K. Fumino, P. Stange, V. Fossog, R. Hempelmann, R. Ludwig, Angew. Chem. 2013, 125, 12667 – 12670; Angew. Chem. Int. Ed. 2013, 52, 12439 – 12442. [22] a) M. Eigen, K. Tamm, Z. Elektrochem. 1962, 66, 93 – 107; b) M. Eigen, K. Tamm, Z. Elektrochem. 1962, 66, 107 – 121. [23] Y. Marcus, G. Hefter, Chem. Rev. 2006, 106, 4585 – 4651. [24] G. Hefter, Pure Appl. Chem. 2006, 78, 1571 – 1586. [25] R. Buchner, Pure Appl. Chem. 2008, 80, 1239 – 1252. [26] M. R. Antonio, M. Nyman, T. M. Anderson, Angew. Chem. 2009, 121, 6252 – 6256; Angew. Chem. Int. Ed. 2009, 48, 6136 – 6140. [27] O. Treutler, R. Ahlrichs, J. Chem. Phys. 1995, 102, 346 – 354. [28] M. Von Arnim, R. Ahlrichs, J. Comput. Chem. 1998, 19, 1746 – 1757. [29] P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2004, 384, 103 – 107. [30] P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 2002, 362, 511 – 518. [31] P. Deglmann, F. Furche, J. Chem. Phys. 2002, 117, 9535 – 9538. [32] D. Rappoport, F. Furche, J. Chem. Phys. 2010, 133, 134105. [33] K. A. Peterson, D. Figgen, E. Goll, H. Stoll, M. Dolg, J. Chem. Phys. 2003, 119, 11113 – 11123. [34] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132, 154104. [35] L. X. Dang, J. Chem. Phys. 1992, 97, 1919 – 1921. [36] J. F. Casteel, P. G. Sears, J. Chem. Eng. Data 1974, 19, 196 – 200. [37] J. V. Walther, J. Schott, Nature 1988, 332, 635 – 638. [38] M. J. Blandamer, Nature 1967, 214, 380 – 381. [39] H. Sadek, R. Fuoss, J. Am. Chem. Soc. 1954, 76, 5897 – 5901. [40] E. Grunwald, Anal. Chem. 1954, 26, 1696 – 1701. Received: April 4, 2014 Published online on June 12, 2014
ChemPhysChem 2014, 15, 2604 – 2609
2609
