摘要:
The effect of lithium triflate and lithium bromide on the vibrational frequencies of DMEDA Shawna S. York, Scott E. Boesch, Ralph A. Wheeler* and Roger Frech* Methodology Experimental Department of Chemistry and Biochemistry, University of Oklahoma, Norman, OK 73019, USA Received 1st May 2002, Accepted 20th June 2002 Published on the Web 4th July 2002 An experimental and computational investigation of the structures and vibrational frequencies of N,N’- dimethylethylenediamine–lithium triflate (DMEDA–LiTf; Tf2 ~ CF3SO32) has been done using a combination of hybrid Hartree–Fock/density functional calculations and Raman and IR spectroscopy. Band assignments for DMEDA–LiBr were made by comparing the experimental Raman and IR spectra of a 5 : 1 DMEDA–LiBr sample with the calculated vibrational frequencies of the DMEDA–Li1 and DMEDA–LiBr complexes.Band assignments for DMEDA–LiTf were made by comparing the experimental spectra of samples over a composition range of 20 : 1 to 1.5 : 1 with calculations done on the DMEDA–LiTf complexes. The effect of the lithium cation and lithium triflate on the geometries and vibrational frequencies of the DMEDA will be examined. The combined experimental data and computational results clearly show that the intramolecular hydrogen bonding in pure DMEDA is broken upon addition of lithium bromide or lithium triflate. Introduction 2- The potential importance of polymer electrolytes in a variety of applications, including rechargeable batteries, has stimulated intense interest in these systems.A wide variety of studies have focused on developing a molecular level understanding of the factors controlling ionic conductivity in poly(ethylene oxide)– salt systems.1–9 Fundamental investigations into the nature of interactions in polymer electrolyte systems can be complemented by considering a polymer having a backbone structure similar to that of poly(ethylene oxide), or PEO, but with a nonoxygen heteroatom. Linear poly(ethylenimine), or LPEI, [CH CH2NH]n, is structurally analogous to PEO, with anN–Hgroup in place of the oxygen.10 Several groups have studied polymer electrolytes based on LPEI with dissolved lithium salts.11–18 In order to better understand the interactions within LPEI– salt systems, it is useful to do theoretical and experimental studies on small-molecule model compounds. The simplest model compound for LPEI is N,N’-dimethylethylenediamine.We have previously conducted a complete vibrational analysis of DMEDA, comparing the experimental Raman and IR spectra of DMEDA to the calculated vibrational frequencies.19 We now present calculations done on the DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf complexes using hybrid Hartree–Fock/density functional methods and compare these to the experimental IR and Raman spectra of DMEDA–LiBr and DMEDA–LiTf systems. The DMEDA–LiTf complex is shown in Fig. 1. Lithium triflate is used because the vibrations for the triflate anion are well studied and ionic association of the triflate anion i.e.the presence of discrete, ionically associated species, may be determined spectroscopically.20–25 In DMEDA, several triflate anion vibrational modes are coincident with DMEDA modes, complicating the interpretation of the spectra; therefore we have also studied the DMEDA–LiBr system. N,N’-dimethylethylenediamine (DMEDA) 99% was used as received from Aldrich. Lithium triflate, LiCF3SO3, (LiTf) and DOI: 10.1039/b204103k lithium bromide obtained from Aldrich were dried under vacuum at y120 uC for 24 h. All materials were stored in a nitrogen-atmosphere glovebox with moisture less than 1 ppm. Desired ratios of DMEDA and LiTf or LiBr were mixed in the glovebox and stirred for 24 h. The DMEDA–LiTf samples were made with compositions ranging from a 1.5 : 1 to a 20 : 1 DMEDA to LiTf ratio.The DMEDA–LiBr sample was made at a 5 : 1 composition. Infrared spectra were recorded with a Bruker IFS66V FT-IR over a range of 4000 to 500 cm21 at a resolution of 1 cm21. The DMEDA–salt samples were placed between ZnSe plates in a sealed sample holder in the glovebox. The IR spectra were taken under dry air purge. Raman spectra were collected using a Jobin-Yvon T64000 Raman spectrometer with a CCD detector. The 514.5 nm line of an argon ion laser at a power of 300 mW at the laser head was used for excitation. Spectra were collected in a 90u scattering geometry. Raman samples were sealed in cuvettes inside the glovebox. Curve fitting analysis was accomplished using a commercial program (Galactic Grams version 5.05).Spectra were curve-fit to a straight base Fig. 1 DMEDA–LiTf complex in the TGT conformation. Click image or here to access a 3D representation. PhysChemComm, 2002, 5(16), 99–111 This journal is # The Royal Society of Chemistry 2002 Paper 99line and one Gaussian–Lorenzian product function for each band using a non-linear least squares method. X X X C C XC C Computational The B3LYP hybrid Hartree–Fock/density functional(HF/DF) method26,27 was used to perform complete geometry optimizations and vibrational frequency calculations on the DMEDA– Li1, DMEDA–LiBr, and DMEDA–LiTf complexes. The three-parameter HF/DF method employs a weighted sum of Hartree–Fock (EHF X ), local DF, and gradient corrected DF expressions for the exchange and correlation energies as in the following equation: E ~ aESlater 1 (1 2 a)EHF 1 bEBecke 1 cELYP 1 (1 2 c)EVWN X where (ESlater) is Slater’s local spin density functional for exchange,28 EBeche is Becke’s gradient corrected exchange functional,29 EVWN is the local density correlation functional of Vosko, Wilk and Nusair,30 and ELYP is the gradient-corrected correlation functional of Lee, Yang and Parr.31 Coefficients giving the relative weights of various approximations for the exchange and correlation energies in this method were optimized to reproduce thermochemical data for a variety of small molecules using a slightly different correlation functional.32 Unless otherwise noted, all calculations reported here were performed using the 6–31G(d) split-valence plus polarization basis set.33 This basis set was chosen because it accurately reproduces the structures and vibrational spectra of medium-sized organic molecules and is small enough for rapid calculations.The quantum chemistry programs GAUSSIAN9426 and GAUSSIAN9834 were used for all calculations. Berny’s optimization algorithm35 was used to perform full geometry optimizations in C1 symmetry. Harmonic frequency calculations were performed at the optimized geometries without correcting for anharmonicity. It has become customary to scale calculated frequencies to facilitate comparisons with experiment and we chose to use the multiplicative scaling factors of Scott and Radom.36 All frequencies less than 1000 cm21 are multiplied by 1.0013 and all frequencies greater than 1000 cm21 are multiplied by 0.9614. Vibrational mode assignments were performed by animating each mode using the program XMOL37 and comparing the modes of one geometry to another using the program ViPA, an acronym for Vibrational Projection Analysis.38–40 The ViPA program exploits the vector properties of vibrational normal modes to assess the similarity between modes of an object molecule and a structurally similar basis molecule.The program first aligns the two molecules and calculates each molecule’s normal modes and vibrational frequencies. For each molecule, each of the normal vibrational modes is a column vector, which is orthonormal to all other normal modes of the same molecule.The vector projection operation is done by sequentially projecting each normal mode of the object molecule on the modes of the basis molecule. The similarity of any mode of the object molecule to any mode of the basis molecule can then be expressed as a percentage by calculating the sum of squares of the matrix elements and multiplying by 100. Vibrational projection analysis has been used to compare normal modes modified by isotopic or chemical substitution, oxidation/reduction, non-covalent contacts, and conformation.19,38–41 The conformations of DMEDA–Li, DMEDA–LiBr, and DMEDA–LiTf can be identified by the dihedral angles of the DMEDA. For example, a TGT conformation signifies that the first dihedral angle (C–N–C–C) is trans, T, corresponding to an angle of 180 ¡ 60u, the second dihedral angle (N–C–C–N) is PhysChemComm, 2002, 5(16), 99–111 100 gauche, G, corresponding to an angle 60 ¡ 60u, and the third dihedral angle (C–C–N–C) is trans.Relative energy 0a 0.5 1.0 27.3 29.1 30.2 0a 0.3 00.7 a 0.3 0.9 aResults Calculated geometries Geometry optimizations were done on a complex with DMEDA and a lithium cation. The lowest energy structure has the DMEDA in a TGT conformation, TGT–Li1. Henceforth in this paper we will refer to the calculated DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf complexes by the three-letter conformation of the DMEDA followed by Li1, LiBr, or LiTf. In TGT–Li1 the lithium cation is coordinated to both nitrogens of the DMEDA.Several other calculations were done to determine if other low energy structures of the DMEDA–Li1 complex exist. In addition to the TGT–Li1, geometry optimizations were done on five other possible conformations. In order of increasing energy, they are TGG–Li1, GGG–Li1, TTT–Li1, TTG–Li1, and GTG–Li1. Table 1 shows the relative energies of the different conformations. As in the TGT–Li1 conformation, the TGG–Li1 and GGG–Li1 systems have the lithium cation coordinated to both nitrogens and these conformations are only slightly higher in energy than the TGT–Li1 complex. In contrast, the TTT–Li1, TTG–Li1, and GTG–Li1 conformations have the lithium cation coordinated to single nitrogen and are significantly higher in energy than the TGT–Li1, TGG–Li1, and GGG–Li1 conformations. In order to investigate the effect of the bromine or triflate anion on the structure and vibrational frequencies of DMEDA, we did calculations on complexes of DMEDA and lithium bromide (DMEDA–LiBr) and DMEDA and lithium triflate (DMEDA–LiTf).For both of these complexes, the calculations were done with the DMEDA in three different conformations (TGT, TGG, and GGG). We chose these conformations because of results found for the TGT–Li1 complexes. In each of the calculations with lithium triflate, it was found that the lithium was coordinated to the two nitrogen atoms of the DMEDA and two oxygen atoms of the triflate anion. Previous calculations have been done on DMEDA and it was found that DMEDA has three low-energy TGT conformations with zero, one, and two intramolecular bonds (referred to as DMEDA(0), DMEDA(1), and DMEDA(2) respectively), with DMEDA(1) and DMEDA(0) providing the closest frequency correlations to the experimental spectrum of DMEDA.19 Table 2 compares the calculated bond distances, angles, and dihedral angles of the low-energy structures of DMEDA and the DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf complexes. The most striking effect of Li1 coordination by Table 1 Comparison of the relative energies (kcal mol21) of the various DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf conformations Conformation TGT–Li1 TGG–Li1 GGG–Li1 TTT–Li1 TTG–Li1 GTG–Li1 TGT–LiBr TGG–LiBr GGG–LiBr TGT–LiTf TGG–LiTf GGG–LiTf The actual calculated energy for TGT–Li1 is 2276.5345960 hartrees, TGT–LiBr is 22848.5129096 hartrees, and TGT–LiTf is 21238.2244363 hartrees.Table 2 A comparison of the bond distances (a°ngstro�ms), bond angles (degrees), and torsional angles (degrees) of DMEDA(0), DMEDA(1), DMEDA–TGT–Li1, DMEDA–TGT–LiBr, and DMEDA–TGT–LiTf- calculated using the B3LYP hybrid Hartree–Fock/density functional method with a 6–31G(d) basis C2–N5 N5–C6 C6–C9 C9–N12 N12–C13 N5–H17 N12–H18 N5–Li19 N12–Li19 C2–N5–C6 N5–C6–C9 C6–C9–N12 N9–C12–C13 C2–N5–C6–C9 N5–C6–C9–N12 C6–C9–N12–C13 H17–N5–C6–C9 H18–N12–C9–C6 Li19–Br Li19–O20 Li19–O21 O20–S22 O21–S22 O23–S22 S22–C24 C24–F26 C24–F27 C24–F25 DMEDA is the elimination of intramolecular hydrogen bonds.The addition of the lithium cation also slightly increases the C–N bond distances, followed by a small decrease when the triflate anion is added. The C–C distance remains essentially unchanged when the lithium or lithium triflate is added and there is little change in the bond angles. However, a signifi- cant geometrical change occurs in the dihedral angles. In DMEDA(0), the central N–C–C–N dihedral angle is 46.7u. This is increased to 60.4u in the TGT–Li1 complex and to 58.8u in the TGT–LiTf complex. This is an interesting contrast to the PEO oligomer, monoglyme (CH3OCH2CH2OCH3), in which the O–C–C-O torsional angle is 72.3u in pure monoglyme and decreases to 48.6u in monoglyme–Li1 and 53.6u in monoglyme– LiTf.42 Notice that the DMEDA–Li1 and DMEDA–LiTf complexes have no intramolecular hydrogen bond, unlike the lowest energy TGT conformation in pure DMEDA. Vibrational spectra Calculated frequencies Tables 3, 4, 5, and 6 list the calculated harmonic vibrational frequencies, IR intensities, Raman activities, and mode assignments for the TGT–Li1, TGT–LiBr, TGT–LiTf, and TGG– LiTf complexes, respectively.We will concentrate on a detailed comparison of the calculated frequencies and those bands of significant intensities observed in the IR and Raman spectra. The lowest frequency modes calculated for the TGT–Li1 complex involve methyl wag or twist (78–257 cm21). There is C–N–Li/C–N–C bending at 275 cm21, and at 284 cm21 there is methyl twist mixed with C–N–Li/N–Li–N bend.The out-ofphase C–N–C bends are at 347 and 568 cm21, while the inphase C–N–C bend is at 349 cm21. The Li–N stretches are at 499 cm21 (symmetric) and 500 cm21 (antisymmetric). The vibrational modes of the TGT–Li1 and the TGT–LiBr complexes are very similar to each other. The lowest frequency modes for the TGT–LiBr complex are torsions, methyl twists, and methyl wags. The mode calculated at 290 cm21 has been assigned to C–N–Li bend, while the mode at 297 cm21 is a DMEDA(1) DMEDA(0) 1.459 1.464 1.527 1.455 1.454 1.020 1.019 1.455 1.456 1.530 1.456 1.455 1.017 1.017 113.4 110.4 110.4 113.4 189.4 62.9 185.8 64.5 -48.2 113.3 112.6 112.6 113.3 166.0 46.7 166.0 41.8 41.8 TGT–Li1 1.485 1.490 1.529 1.491 1.485 1.023 1.023 2.01 2.01 112.4 110.8 110.8 112.4 188.8 60.4 189.0 71.8 71.8 ————————— mixture of methyl twist and wag.The out-of-phase C–N–C bends are at 353 and 568 cm21, while the in-phase C–N–C bend is at 365 cm21. The Li–N symmetric stretch is highly mixed with Li–Br stretch and is calculated at 641 cm21. The Li–N antisymmetric stretch is calculated at 401 cm21. The DMEDA–LiTf complexes have low frequency modes that involve movement of the entire triflate anion with respect to the DMEDA. The five lowest frequency modes for the TGT– LiTf involve these motions. The next ten modes involve methyl wag or twist.At 320 cm21, there is a C–N–C bend. There is a triflate twist at 327 cm21, and the mode at 342 cm21 has contributions from triflate wag (a mixture of SO3 and CF3 wag), C–N–C bend, and methyl wag. The modes at 356 and 567 cm21 are C–N–C bend. The Li–N stretches occur at 370 and 421 cm21. The seven lowest frequency modes of the TGG– LiTf have been assned to torsions of the DMEDA–lithium triflate complex. Methyl wags and twists occur in eight modes (153–290 cm21). The C–S stretch of the triflate ion is computed to be at 302 cm21. Two modes assigned to triflate ion wag are at 327 and 340 cm21. The C–N–C bends are calculated at 361 and 432 cm21. The Li–N stretch is calculated to occur at 600 cm21. It is interesting to compare the Li–N stretches in the DMEDA–Li1 and –LiTf complexes to the corresponding calculated Li–O stretches monoglyme systems.In DMEDA– Li1, the Li–N stretches are at 499 and 500 cm21, and in monoglyme–Li1, the Li–O stretches are at 505 and 541 cm21. In the TGT conformation of DMEDA–LiTf, the Li–N stretches are at 370 and 421 cm21, while in monoglyme–LiTf the Li–O stretches are at 294 and 311 cm21.42 Comparison of calculated and experimental frequencies The IR and Raman spectra of 5 : 1 DMEDA–LiBr were compared to calculated frequencies for DMEDA and the TGT–Li1, TGG–Li1, GGG–Li1, and TGT–LiBr complexes. It was found that the frequencies and intensities calculated for the TGG and GGG complexes were not as consistent with the PhysChemComm, 2002, 5(16), 99–111TGT–LiTf TGT–LiBr 1.469 1.471 1.530 1.476 1.474 1.022 1.021 2.092 2.087 113.8 110.4 110.6 112.9 189.1 58.8 187.6 67.6 67.4 1.472 1.473 1.530 1.473 1.472 1.021 1.021 2.093 2.093 114.1 110.0 110.0 114.1 190.9 58.1 190.9 70.1 70.1 2.227 1.955 2.019 1.504 1.503 1.460 1.863 1.335 1.336 1.349101Table 3 Scaled calculated vibrational frequencies (cm21), mode assignments, IR intensities, and Raman activities for the DMEDA–TGT–Li1 complex Mode assignment Methyl wag; N–H perpendicular bend Methyl wag; N–C–C–N torsion Methyl twist Methyl twist Methyl wag C–N–Li/C–N–C bend; methyl wag Methyl twist; C–N–Li/N–Li–N bend C–N–C bend (out-of-phase) C–N–C bend (in-phase) N–Li–N bend Li–N symmetric stretch Li–N antisymmetric stretch C–N–C bend (out-of-phase) CH2 twist CH2 rock; N–H perpendicular bend CH2 wag; methyl wag N–H perpendicular bend; CH2 rock N–H perpendicular bend; methyl wag C–C stretch N–H perpendicular bend; methyl wag CH2 rock; N–H parallel bend; methyl wag C–N antisymmetric stretch C–N symmetric stretch; C–C stretch CH2 twist; N–H perpendicular bend; methyl wag Methyl wag Methyl wag CH2 twist; N–H perpendicular bend; methyl wag CH2 twist CH2 twist CH2 wag CH2 wag N–H parallel bend; methyl antisymmetric deformation Methyl symmetric deformation; N–H parallel bend Methyl symmetric deformation N–H parallel bend; methyl antisymmetric deformation Methyl antisymmetric deformation Methyl symmetric deformation CH2 scissors; methyl antisymmetric deformation CH2 scissors Methyl antisymmetric deformation CH2 scissors; methyl antisymmetric deformation C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch N–H stretch N–H stretch experimental spectra as those calculated for the TGT conformation.For example, in the IR spectrum of DMEDA–LiBr, we assign the observed band at 840 cm21 to a mixture of CH2 rock and CH2 twist. This assignment is comparable to the calculated frequency of 831 cm21 in the TGT–Li1 conformation and 832 cm21 in the TGT–LiBr conformation. However, the calculated frequency for this mode in the TGG–Li1 conformation is 813 cm21 and for GGG–Li1 conformation is 808 cm21, which are both significantly lower than the experimental frequency.Assignments for the modes in the experimental spectra of DMEDA–LiBr are given in Table 7. We have compared the calculated frequencies and intensities of DMEDA and the DMEDA–LiTf complexes to the IR and Raman spectra of DMEDA–LiTf over a composition range of 1.5 : 1 to 20 : 1 and list the assignments for the experimentally observed modes in Table 8. It was found that the TGT and TGG conformations of the DMEDA–LiTf complex consistently provided the best agreement with the experimental frequencies. Some of the modes in Table 8 involving the triflate anion vibrations have 102 PhysChemComm, 2002, 5(16), 99–111 IR intensity Frequency 78 140 198 223 257 275 284 347 349 499 500 568 831 853 969 964 983 991 1000 1042 1065 1075 1129 1157 1215 1272 1273 1343 1367 1415 1426 1427 1428 1463 1464 1470 1474 1481 1484 2930 2932 2938 2938 2967 2979 3008 3008 3034 3034 3312 3312 123456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 80260 31 15 111 70 479 13 19 84 100 480 987 10 19249 14 110870 228 18 1706 266 12 454 15 20 13 14 12 101 10 been previously assigned.22,23,43 In the 1.5 : 1 composition the majority of the bands arise from the DMEDA–LiTf complex; however, in the more dilute compositions, there are also bands that arise from uncomplexed DMEDA.Specific features in the experimental IR and Raman spectra of DMEDA–LiTf and DMEDA–LiBr are discussed below. The spectral region from 1075 to 500 cm21 Fig. 2a shows the DMEDA–LiTf composition-dependent IR spectra in the region from 500 to 1075 cm21. There are several interesting spectral features in this region. Sharp bands at 519 and 639 cm21 arise from the das(CF3) and ds(SO3) modes, respectively, which are calculated at 523 and 634 cm21 in DMEDA(TGT)–LiTf and at 524 and 636 cm21 in DMEDA- (TGG)–LiTf. There are notable changes in the DMEDA IR bands in the region from 700 to 900 cm21 as lithium triflate composition changes.The bands in this region are assigned to a mixture of CH2 rock and N–H bend. The broad band around 770 cm21 Raman activity 0000001016204 11032211 10122764245981 11 195 151 28 21 77 11 279 91 20 55 54 67 64 979Table 4 Scaled calculated vibrational frequencies (cm21), mode assignments, IR intensities, and Raman activities for the DMEDA–TGT–LiBr complex Mode description Torsion Torsion Torsion NCCN torsion and methyl wag NCCN torsion and methyl wag Methyl twist Methyl twist Methyl wag Methyl wag; C–N–Li bend Methyl twist/wag C–N–C bend (out of phase) C–N–C bend (in phase) Li–N antisymmetric stretch C–N–C bend (out of phase) Li–N symmetric stretch; Li–Br stretch CH2 rock; N–H perpendicular bend CH2 wag; N–H perpendicular bend; methyl wag N–H perpendicular bend; CH2 rock N–H perpendicular bend; methyl wag CH2 rock; methyl wag C–C stretch CH2 rock; methyl wag; N–H parallel bend CH2 twist; methyl wag; N–H bend C–N stretch; methyl wag Methyl wag; C–N stretch; C–C stretch Methyl wag Methyl wag Methyl wag; N–H perpendicular bend CH2 twist CH2 twist CH2 wag CH2 wag Methyl symmetric deformation; N–H parallel bend Methyl symmetric deformation N–H parallel bend; methyl antisymmetric deformation N–H parallel bend Methyl antisymmetric deformation Methyl antisymmetric deformation CH2 scissors; methyl antisymmetric deformation Methyl antisymmetric deformation CH2 scissors CH2 scissors; methyl antisymmetric deformation C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch N–H stretch N–H stretch decreases drastically in intensity as the LiTf composition increases, disappearing in the 1.5 : 1 sample except for a shoulder at 808 cm21.This corresponds to the calculated N–H bends at 788 and 825 cm21 in DMEDA(1). A band appears at 836 cm21 in the 20 : 1 sample, shifting to 839 cm21 in the 1.5 : 1 composition. This is closest to calculated bands at 834 and 858 cm21 in TGT–LiTf and at 817 and 859 cm21 in TGG–LiTf, which are assigned to a mixture of CH2 rock and N–H bend. As the LiTf composition increases, the DMEDA band at 880 cm21 shifts from 881 cm21 in the 20 : 1 composition to 898 cm21 in the 1.5 : 1 composition.The mode calculated at 904 cm21 for DMEDA(1) and assigned primarily to CH2 rock moves to calculated frequencies of 917 and 903 cm21 in the TGT–LiTf and TGG–LiTf complexes, respectively. It is important to note that these modes are assigned to a mixture of CH2 rock and N–H bend. These calculated changes very nicely underscore the point that the interaction of the cation with the host DMEDA Frequency 35 39 98 105 156 195 238 261 290 297 353 365 401 568 641 832 857 913 958 985 986 1005 1073 1079 1093 1132 1154 1213 1260 1267 1344 1364 1425 1426 1248 1441 1463 1464 1472 1480 1482 1485 2885 2896 2899 2901 2938 2956 2989 2989 3032 3032 3330 3331 123456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 molecule can significantly affect mode mixing and hence the nature of the normal modes.In the Raman spectra in this region (Fig. 2b) one can see the DMEDA band at 880 cm21, assigned to primarily N–H bend, broaden and disappear as the composition of lithium triflate increases.A band assigned to a mixture of N–H bend and CH2 rock grows in at 836 cm21 and steadily increases in intensity, just as in the IR spectra in this region. The comparison of the experimental spectra with the calculated frequencies indicates that these spectral changes reflect the structural changes seen in the geometry optimization upon addition of LiTf, i.e., that the central dihedral angle of the DMEDA is changed (as indicated in Table 1) and that an intramolecular hydrogen bond is broken. The N–H bending contribution to the modes in this region is probably most influenced by the change in hydrogen bonding. The dramatic changes are observed in the CH2 rocking and N–H bending modes upon complexation of DMEDA with LiTf are not Raman activity IR intensity 41330200 11270 261 99 26 64 75 41 20614 744 23022 10 142412 167 21 1650 156 38 116 847 40 21 22 13714 000010001113113473151110 12424 1232577557 223 283 15 35 47 57 364 898 50 65 59 56 992 103 PhysChemComm, 2002, 5(16), 99–111Table 5 Scaled calculated vibrational frequencies (cm21), mode assignments, IR intensities, and Raman activities for the DMEDA–TGT–LiTf complex Mode assignment Torsion Torsion Torsion Torsion Torsion Methyl wag Methyl wag Methyl wag Methyl twist Methyl twist Methyl twist; triflate wag Methyl twist Methyl twist Methyl wag Methyl twist Methyl twist; CH2 rock; C–S stretch C–N–C bend; Triflate twist Triflate wag; C–N–C bend; methyl wag C–N–C bend Li–N stretch Li–N stretch SO3 antisymmetric deformation CF3 antisymmetric deformation CF3 antisymmetric deformation SO3 antisymmetric deformation CF3 antisymmetric deformation SO3 antisymmetric deformation C–N–C bend CF3/SO3 antisymmetric deformation Li–O; Li–N stretch SO3 symmetric deformation CF3 symmetric deformation 2 rock 3 symmetric stretch; N–H perpendicular bend 2 twist; methyl wag N–H perpendicular bend; CH CH2 wag; CH2 twist N–H perpendicular bend N–H perpendicular bend SO Methyl wag; C–N stretch C–C/C–N stretch; CH2 wag CH2 rock; methyl wag; N–H parallel bend N–H parallel bend; C–N stretch; CH2 twist CH SO3 antisymmetric stretch; N–H perpendicular bend C–N stretch; methyl wag Methyl wag C–S stretch; CF3 antisymmetric stretch Methyl wag CF3 antisymmetric stretch; C–S stretch CF3 antisymmetric stretch 3 antisymmetric stretch 2 scissors N–H perpendicular bend; methyl wag; CH2 rock SO CH2 twist CH2 twist; N–H perpendicular bend CH2 wag CH2 wag Methyl symmetric deformation; N–H parallel bend Methyl symmetric deformation N–H parallel bend; methyl symmetric deformation N–H parallel bend Methyl antisymmetric deformation Methyl antisymmetric deformation Methyl antisymmetric deformation; CH2 scissors CH2 scissors Methyl antisymmetric deformation Methyl antisymmetric deformation; CH C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch N–H stretch N–H stretch 104 PhysChemComm, 2002, 5(16), 99–111 Frequency 18 29 42 48 64 80 122 155 180 200 205 208 252 260 298 299 320 327 342 356 370 421 491 12345678910 11 12 13 14 15 16 17 18 19 20 21 22 23 523 24 555 25 26 567 577 595 634 759 834 858 917 959 979 986 989 1007 1071 1080 1092 1094 1133 1153 1157 1180 1202 1212 1255 1261 1267 1345 1366 1425 1426 1430 1449 1465 1465 1471 1480 1481 1487 2881 2893 2897 2900 2938 2966 2980 3000 3025 3037 3331 3333 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 Raman activity IR intensity 00000000000000231231211 11014410104110216109923 21 0 8 4 13 131110 566417 4312159333204410 635857525 12 4327 17 36 69 153 226 81 19 64 60 59 50 51 58 023 224 99 226 69 121 83 129 13 8123 54 177 22 25 164 8118 270 1343 422 17 150516 15 11 10 1911 755 108 77 16 27 25 19 15 10 51Table 6 Scaled calculated vibrational frequencies (cm21), mode assignments, and IR intensities for the DMEDA–TGG–LiTf complex Mode assignment Torsion Torsion Torsion Torsion Torsion Torsion Torsion Methyl wag Methyl twist Methyl twist Methyl twist Methyl twist Methyl twist Methyl twist Methyl twist C–S stretch Triflate wag Triflate wag C–N–C bend N–Li–O bend Li–N stretch C–N–C bend SO3 antisymmetric deformation CF3 antisymmetric deformation CF3 antisymmetric deformation SO3 antisymmetric deformation CF3 antisymmetric deformation SO3 antisymmetric deformation CH2 twist; N–H perpendicular bend CF3 antisymmetric deformation SO3 antisymmetric deformation Li–N stretch Li–N stretch; SO3 symmetric deformation 3 symmetric deformation 2 rock; N–H perpendicular bend 2 wag; CH2 twist CF CH CH N–H perpendicular bend N–H perpendicular bend SO3 symmetric stretch 2 twist 2 rock 2 twist; methyl wag CH CH CH2 rock; methyl wag; N–H parallel bend N–H perpendicular bend; CH C–N stretch SO3 antisymmetric stretch 3 antisymmetric stretch; C–S stretch 2 rock 3 antisymmetric stretch CH2 wag; N–H perpendicular bend; methyl wag Methyl wag CF Methyl wag CF3 antisymmetric stretch Methyl wag; CH CF SO3 antisymmetric stretch 2 twist 2 wag 2 wag 2 scissors; methyl antisymmetric deformation 2 scissors 2 scissors; methyl antisymmetric deformation CH CH2 twist CH CH Methyl symmetric deformation; N–H parallel bend Methyl symmetric deformation N–H parallel bend N–H parallel bend; methyl antisymmetric deformation CH Methyl antisymmetric deformation Methyl antisymmetric deformation; CH CH Methyl antisymmetric deformation; N–H parallel bend Methyl antisymmetric deformation; CH2 scissors C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch C–H stretch N–H stretch N–H stretch Frequency 27 31 41 60 66 86 130 153 171 189 204 207 240 260 290 302 327 340 361 365 412 432 494 12345678910 11 12 13 14 15 16 17 18 19 20 21 22 23 524 24 557 25 561 578 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 600 636 758 817 859 903 928 978 1001 982 1008 1060 1083 1094 1114 1128 1144 1149 1181 1199 1202 1257 1259 1271 1346 1363 1421 1427 1435 1440 1459 1466 1470 1476 1478 1484 2884 2895 2898 2900 2949 2973 2995 3006 3031 3047 3335 3373 Raman activity IR intensity 00000000000000042320101 0012123012210030001817 819 0 9 4 5 13 330 259 79 3849 98 31 177 14 719 16 31 190 17 18 179 3118 9264 345 18 714 10 4010 3013 18 811 11 10 47 86 96 12 29 18 20 11 714 1110 382319 331544523221039834892619 17 82532 40 53 57 336 89 11 58 47 52 56 60 72 105 PhysChemComm, 2002, 5(16), 99–111Table 7 Comparison of experimental IR frequencies (cm21) of DMEDA–LiBr 5 : 1 with the calculated frequencies (cm21) for DMEDA–TGT–LiBr, DMEDA–Li1 and DMEDA(1) with mode assignments DMEDA–LiBr 566 775 811 840 898 983 1006 1034 1105 1121 1146 1226 1251 1281 1349 1410 1449 1472 2686 2790 2844 2889 2934 3210 3270 unexpected.In a related system, diglyme (CH3OCH2- CH2OCH2CH2OCH3)–LiTf, changes in the frequency and intensity of the modes assigned to mixed CH2 rocking have been shown to indicate changes in the central dihedral angleupon interaction with the lithium ion.2,9,44 The CF3 symmetric deformation can be seen in the IR spectra as a small peak around 757 cm21 growing in on top of 10 : 1 5 : 1 1 : 1 Table 8 Comparison of experimental IR frequencies (cm21) of DMEDA–LiTf of varying compositions with the calculated frequencies (cm21) for DMEDA–TGT–LiTf, DMEDA–TGG–LiTF, and DMEDA(1) with mode assignments ds(CF3) peak is unobscured by DMEDA bands and the peak center is seen to shift from 757 cm21 in the 20 : 1 sample to 760 cm21 in the 5 : 1 sample, and then to 763 cm21 in the 1.5 : 1 sample.The calculated CF3 symmetric deformation is at 759 cm21 in TGT–LiTf and 758 cm21 in TGG–LiTf. The symmetric SO DMEDA–TGG– LiTf 518 518 518 558 574 640 759 573 639 757 573 639 757 823 839 898 1006 839 897 1006 838 896 1006 1033 1033 1038 1092 1100 1104 1122 1159 1104 1122 1160 1166 1226 1259 1352 1226 1259 1352 1228 1257 1353 1370 1458 1450 1450 1474 1474 1475 2689 2795 2847 2887 2937 3295 3323 2689 2794 2847 2888 2937 3295 3323 2692 2811 2864 2902 2954 3301 3332 106 PhysChemComm, 2002, 5(16), 99–111 DMEDA–LiBr DMEDA–Li1 568 568 831 832 913 986 1005 1079 983 1000 1042 1129 1157 1215 1132 1154 1213 1260 1267 1344 1425 1272 1343 1415 1470 2938 2967 2979 3008 3034 3312 3312 1472 2938 2956 2989 2989 3032 3330 3331 20 : 1 DMEDA–TGT– LiTf 519 492 523 556 578 634 759 573 639 758 770 880 834 or 859 916 1008 1033 1092 1094 1106 1121 1154 1226 1257 1347 1447 1473 2679 2789 2843 2887 2935 3287 1133 1153 1180 1202 1256 1345 1366 1449 1464 1472 1486 2892 2898 2900 2936 2963 3332 3334 DMEDA(1) 788 825 904 1115 1248 1343 1425 1451 1476 2771 2822 2831 2842 2944 3338 3363 the CH2 rocking band at 770 cm21.In the Raman spectra the 3 stretches appear at 1039 and 1032 cm21 in the DMEDA(1) 494 523 557 578 636 758 788 825 904 (CH2 rock) 817 or 859 902 1009 994 n(C–C) 1022 (CH2 rock) 1094 1083 1115 1134 1129 1144 1180 1202 1257 1346 1363 1435 1466 1479 1483 2883 2896 2899 2900 2950 3335 3373 1342 1362 1444 1451 1476 1482 2770 2822 2831 2842 2944 3338 3363 Assignment C–N–C bend N–H bend CH2 rock CH2 rock; CH2 twist CH2 rock C–C stretch CH2 rock, N–H bend, methyl wag C–N stretch Methyl wag Methyl wag CH2 twist; N–H bend CH2 twist CH2 wag N–H bend; methyl deformation Methyl deformation; CH2 scissors C–H stretch N–H stretch Assignment SO3 antisymmetric deformation; CF3 antisymmetric deformation CF3 antisymmetric deformation CF3/SO3 antisymmetric deformation SO3 symmetric deformation CF3 symmetric deformation N–H bend N–H bend CH2 rock; N–H bend N–H bend CH2 rock; methyl wag; N–H bend SO3 antisymmetric stretch C–N stretch, methyl wag Methyl wag C–N stretch CF3 antisymmetric stretch CF3 antisymmetric stretch SO3 antisymmetric stretch CH2 wag CH2 wag N–H bend; methyl deformation CH2 scissors; methyl deformation C–H stretch N–H stretch N–H stretchFig. 2 Composition-dependent IR (a) and Raman (b) spectra of DMEDA–LiTf in the region from 500 to 1075 cm21.Click image or here to access enhanced versions. IR spectra, with the intensity of the 1039 cm21 band increasing as LiTf composition increases. The bands are calculated in the TGT–LiTf and the TGG–LiTf conformations at 1092 and 1094 cm21, respectively. The SO3 symmetric stretches appear at 1034 and 1040 cm21 in the 20 : 1 Raman spectrum, shifting slightly to 1036 and 1044 cm21 in the 1.5 : 1 sample, with the relative intensity of the higher-frequency band increasing with increasing LiTf composition. Curvefitting analysis has been done on the IR and Raman data in the ds(CF3) and ns(SO3) regions. These modes are very sensitive to ionic speciation, i.e., as the triflate anion becomes more associated, the frequencies Fig.3 Curvefitting analysis of the DMEDA–LiTf spectra for a composition range of 20 : 1 to 1.5 : 1 in the (a) Raman ds(CF3) region; (b) IR ds(CF3) daggregate appears to be only 16% as determined from the relative integrated intensities in the ns(SO3) region. In the 20 : 1, 10 : 1 and 5 : 1 samples, the contact ion pair seems to dominate, with the ‘‘free’’ ion band making up less than 15% of the total area in the curvefitting of the Raman ds(CF3) region, while in the ns(SO3) region the intensities of the ‘‘free’’ ion and ion pair components for these three compositions are nearly equal. Discrepancies in the ionic speciation as determined from the ds(CF3) and ns(SO3) regions have been observed by Ferry et al. in a PEO oligomer45 and in poly(propylene glycol).46 discrepancies by using tetrabutylammonium triflate (TbaTf). Cation–anion interactions are minimized by using Tba1, a bulky cation with a well-protected charge.47 We have examined region; (c) Raman ns(SO3) region and (d) IR ns(SO3) region.of these modes increase.23 These bands are non-degenerate, and the presence of multiple bands arises from the triflate anion vibrating in different local environments. Fig. 3 shows the results of the curvefitting. The symmetric CF3 deformation can be deconvoluted into three bands centered around 755, 758– 760 and 763 cm21. According to the literature,23 bands at 752– 753 cm21 are attributed to ‘‘free’’ ions, bands at 756–757 cm21 are attributed to contact ion pairs, and bands at 763 cm21 are attributed to the [Li2Tf]1 aggregate.The symmetric SO3 stretching region can be deconvoluted into bands centered around 1032–1036, 1038–1044, and 1050–1054 cm21. According to the literature,8 ‘‘free’’ ions are at 1032–1033 cm21, ion pairs are at 1042 cm21, and aggregate at 1050–1052 cm21. Based on a comparison of the ds(CF3) and ns(SO3) regions, it appears that in the DMEDA–LiTf system the ‘‘free’’ ion and ion pair components of the ds(CF3) band are at slightly higher frequencies in the DMEDA–LiTf system than in ethylene oxide systems.8,9,21,22,24 The curvefitting of the ds(CF3) in the IR and the ns(SO3) in the Raman is complicated by the presence of DMEDA bands in these spectral regions, therefore the relative intensities obtained by curvefitting may have a larger error than in other regions of these spectra.There are qualitative similarities between the ionic speciation as determined from these spectral regions: in all cases, for the 20 : 1, 10 : 1 and 5 : 1 composition, the LiTf appears to exist as a mixture of ‘‘free’’ ions and/or ion pairs, while in the 1.5 : 1 sample a significant amount of aggregate is present. There are, however, notable differences in the quantitative comparison of the species as determined from the CF3 symmetric deformation or the SO3 symmetric stretch. The amount of aggregate in the 1.5 : 1 sample is 100% as determined from the Raman spectra in the s(CF3) region, while in the same spectrum the amount of We have attempted to determine the cause of these 107 PhysChemComm, 2002, 5(16), 99–111the IR spectra of DMEDA–TbaTf in the ds(CF3) and ns(SO3) regions over a composition range of 44 : 1 to 7 : 1.We found neither a complex band structure or a shift in frequency in these two modes with increasing salt composition, indicating no change in ionic association, as expected. The ns(SO3) mode exhibits a single band at 1031 cm21, consistent with a ‘‘free’’ ion, although at a slightly lower frequency than is usual in ethylene oxide based systems. In the ds(CF3) region, a single band is seen at 755 cm21, which is higher that the ‘‘free’’ triflate frequency in LiTf–ethylene oxide systems.21,23,24 In a previous study,47 it was noted that the frequency of the ds(CF3) mode of TbaTf is solvent-dependent, while the ns(SO3) mode appears to be solvent-independent.However, recent work suggests that significant hydrogenbonding interaction between the triflate anion and a protic solvent may slightly lower the ns(SO3) frequency from the usual values expected for a ‘‘free’’ ion in an aprotic environment. The frequency of 755 cm21 for the ds(CF3) mode observed in the DMEDA system with both LiTf and TbaTf, may therefore arise from a ‘‘free’’ ion, with the unusually high frequency being due to anion–solvent interactions rather than cation–ion interactions. Furthermore, the probability of the frequency of the CF3 symmetric stretching mode being affected by solvent interactions suggests that the SO3 symmetric stretch may be a better indicator of ionic association in the DMEDA system.However this conclusion must be very carefully qualified. There are clearly significant interactions of the triflate ion with the solvent, presumably a hydrogen-bonding interaction between the triflate oxygen atoms and the amine hydrogen atom of the DMEDA. Therefore, any conclusions about the relative composition of ion species based on comparisons of relative integrated intensities from either the ds(CF3) or the ns(SO3) spectral region should be regarded as highly problematic in systems that have hydrogen bonding. The IR and Raman spectra of DMEDA, DMEDA–LiBr (5 : 1) and DMEDA–LiTf (5 : 1) in this region are shown in Fig. 4a and 4b. The bands centered around 780 and 880 cm21 in the IR spectrum of DMEDA(4a) broaden and shift to higher wavenumbers when LiBr or LiTf is added.A comparison of the IR spectra of DMEDA–LiBr with DMEDA in the region around 1000 cm21 shows that the bands at 983 and 1006 cm21 in the DMEDA are unchanged by LiBr addition. These bands are consistent with the C–C stretch at 999 cm21 and the CH2 rock at 1022 cm21 in the DMEDA(1) calculation, with the C–C stretch at 983 cm21 and CH2 rock mixed with N–H bend and methyl wag at 1000 cm21 in the DMEDA–Li1 calculation and with the methyl wag at 1005 cm21 in the DMEDA–LiBr calculation. The very weak, broad band appearing around 800 cm21 in the pure DMEDA Raman spectrum (4b) is suppressed with the addition of either LiBr or LiTf. This band has been attributed to N–H bend from the DMEDA(1) calculation.19 A large band at 836 cm21 appears in the Raman spectra with the addition of LiBr or LiTf.This band is closest to the mode calculated at 831 cm21 in the DMEDA–Li1 complex and 832 cm21 in the DMEDA–LiBr complex, which we assign to a mixture of CH2 rock and N–H perpendicular bend. The band at 880 cm21 in the Raman spectrum of pure DMEDA decreases in intensity when LiBr or LiTf is added. These changes in the CH2 rocking and N–H bending regions are consistent with what is seen in the IR spectra. The same type of change in the dihedral angle of DMEDA and the breaking of intramolecular hydrogen bonds that is indicated in the LiTf system also appears to occur in the LiBr system. This is expected, given that those changes were seen in the geometry optimization when a lithium cation was added to DMEDA.The spectral region from 1600 to 1075 cm21 Fig. 5 shows the IR spectra of pure DMEDA and DMEDA– LiTf from the 20 : 1 to the 1.5 : 1 compositions in the 1600 to 108 PhysChemComm, 2002, 5(16), 99–111 Fig. 4 (a) IR spectra of pure DMEDA, DMEDA–LiTf, and DMEDA– LiBr in the 500 to 1075 cm21 region. (b) Raman spectra of pure DMEDA, DMEDA–LiTf, and DMEDA–LiBr in the 700 to 1075 cm21 region. Click image or here to access enhanced versions. Fig. 5 Composition-dependent IR spectra of DMEDA–LiTf in the 1075 to 1600 cm21 region. Click image or here to access enhanced version. 1075 cm21 region. There are several significant changes in this region of the IR spectra of DMEDA when salt is added.The band at 1106 cm21 gradually shifts to 1100 cm21 as LiTf composition increases, and a band grows in at 1092 cm21. These bands are closest to calculated frequencies of 1094 or 1083 cm21 in the LiTf complexes, which are assigned to a mixture of C–N stretch and methyl wag. The analogous mode in DMEDA(1) is calculated at 1134 cm21 and assigned to C–N stretch. The band at 1123 cm21 decreases in intensity as LiTf composition increases and appears to vanish in the 1.5 : 1 composition. This band is assigned to a methyl wag, which is calculated at 1115 cm21 in DMEDA(1) and at 1133 or 1128 cm21 for the TGT– or TGG–LiTf complexes. The DMEDA band at 1150 cm21, assigned to C–N stretching, isreplaced by new bands at 1146 and 1165 cm21 in DMEDA– LiTf, with the intensity of the 1165 band increasing relative to the 1146 cm21 as LiTf composition increases.These bands are closest in frequency to calculated bands at 1153 and 1144 cm21 in the TGT–LiTf and TGG–LiTf and are assigned to the CF3 antisymmetric stretch, with some contribution from the C–S stretch. The CF3 antisymmetric stretch appears at 1226 cm21 in the IR and 1227 cm21 in the Raman, but is calculated at 1202 cm21 in both conformations. There is a band calculated at 1251 cm21 in DMEDA(1), which appears in the pure DMEDA spectrum but which is experimentally indistinguishable under the broad, intense bands due to the SO3 stretches at 1296, 1277, and 1257 cm21 in the DMEDA–LiTf spectra.The SO3 antisymmetric stretch is doubly degenerate in an isolated triflate ion. Upon interaction with a cation, the degeneracy is broken. We assign the band at 1277 cm21 to an unperturbed or ‘‘free’’ triflate ion, based on previous studies.24,43,48 The two bands at 1296 and 1257 cm21 are then attributed to the two components of the SO3 symmetric stretch resulting from the breaking of the two-fold degeneracy in some fraction of the triflate ions present. The SO3 antisymmetric stretch is calculated at 1255 or 1257 cm21. The bands from 1500 to 1400 cm21 broaden as the LiTf composition increases. Modes involving N–H parallel bend, methyl deformation and CH2 scissors are calculated from 1425 to 1487 cm21 in the TGT– LiTf complex and from 1421 to 1484 cm21 in the TGG–LiTf complex.The IR spectrum of DMEDA–LiBr 5 : 1 has similar features to the DMEDA–LiTf spectra in the region from 1575 to 1075 cm21. In the DMEDA–LiBr spectrum, a shoulder grows in at 1094 cm21 on the 1105 cm21 band and the band at 1121 cm21 decreases in intensity. The modes calculated at 1042 cm21 in the DMEDA–Li1 complex and 1079 in the DMEDA–LiBr complex are assigned to a mixture of C–N stretch and methyl wag. The band at 1151 cm21 shifts to slightly lower wavenumbers. In the Raman spectra, the C–N stretch at 1114 cm21 (calculated at 1115 cm21 in DMEDA(1)) and the methyl wags at 1126 and 1151 cm21 (calculated at 1134 cm21 in DMEDA(1), 1129 and 1157 cm21 in DMEDA– Li1, and 1132 and 1154 cm21 in DMEDA–LiBr) are relatively unchanged upon the addition of LiBr.The Raman band at 1257 cm21 in DMEDA, attributed to CH2 twist, decreases in intensity with addition of LiBr. Bands between 1400 and 1500 cm21 are slightly broadened with addition of either LiBr or LiTf in both the IR and Raman spectra. These modes, involving N–H bend, methyl deformation, or CH2 scissors, are calculated to occur from 1415 to 1484 cm21 in the DMEDA– Li1 complex and 1425 to 1485 cm21 in the DMEDA–LiBr complex. The spectral region from 4000 to 1600 cm21 There are several bands in the IR spectra of both DMEDA– LiTf 5 : 1 and DMEDA–LiBr 5 : 1 from 2685 to 2965 cm21. These are assigned to C–H stretching vibrations, although the bands lower in frequency than 2800 cm21 may originate from a Fermi resonance interaction between CH2 bending overtones and the higher frequency C–H stretching fundamentals.These bands are calculated between 2881 and 3037 cm21 in the TGT– LiTf complex, between 2884 and 3047 cm21 in the TGG–LiTf complex, between 2885 and 3032 in the TGT–LiBr complex, and between 2930 and 3034 cm21 in the TGT–Li1 complex. It is well-known that density-functional methods give X–H stretching modes too high by ~100–200 cm21 and this appears to be the case here. Fig. 6 shows the IR spectra for DMEDA, DMEDA–LiBr 5 : 1 and DMEDA–LiTf 1.5 : 1 to 20 : 1 in the N–H stretching region. The calculated N–H stretches are at 3338 and 3363 cm21 for DMEDA(1), at 3335 and 3373 cm21 for TGG– LiTf, at 3331 and 3333 cm21 for TGT–LiTf, at 3330 and Fig.6 IR spectra of the N–H stretching region for DMEDA, DMEDA–LiBr and DMEDA–LiTf. Click image or here to access enhanced version. 3331 cm21 for TGT–LiBr, and two modes both calculated at 3312 cm21 for TGT–Li1. In pure DMEDA there are two bands, at 3280 and 3320 cm21, in both the IR and Raman spectra. In DMEDA–LiBr there are bands at 3210 and 3270 cm21 in the IR spectrum and at 3220, 3280 and 3320 cm21 in the Raman spectrum. In DMEDA–LiTf there are distinct bands at 3295 and 3323 cm21 in the IR spectrum and at 3301 and 3327 cm21 in the Raman spectrum. In the composition-dependent IR spectra of the DMEDA–LiTf, the frequencies of the N–H stretches increase as the LiTf composition increases.An interpretation of the N–H stretching region is complicated by the variety of possible hydrogen bonding interactions. In the pure DMEDA, there is a possibility of both intramolecular hydrogen bonds and hydrogen bonds between neighboring DMEDA molecules.19 In the DMEDA–salt systems, there may still be hydrogen bonding between DMEDA molecules, as well as between the amine hydrogens and the anions. Our calculations coupled with observations from other regions of the vibrational spectra indicate that intramolecular hydrogen bonds are broken when salt is added to DMEDA; however at low salt compositions some intramolecular hydrogen bonds may remain. Conclusions In order to determine the effects of adding LiBr and LiTf to DMEDA, hybrid Hartree–Fock/density functional calculations have been performed on DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf complexes.We found that changes occur in the central dihedral angle and in the intramolecular hydrogen bonding when DMEDA complexes lithium. In a previous study,19 we found that pure DMEDA exists in a mixture of low energy conformations, predominantly TGT with intramolecular hydrogen bonding. The calculated geometries for both the DMEDA–Li1, DMEDA–LiBr, and DMEDA–LiTf show that this intramolecular hydrogen bond in the TGT conformation is broken when lithium is added. The differences in the experimentally observed band structure of DMEDA–LiTf or DMEDA–LiBr compared to that of pure DMEDA are also indicative of conformational changes in the DMEDA.The N–H bending and CH2 rocking regions of the spectra are particularly sensitive to conformational changes of the DMEDA, and the N–H stretching region is strongly affected by changes in the hydrogen bonding. From the comparison of the experimentally observed frequency shifts and changes in intensity in the 700–900 cm21 region to the calculated frequencies for the CH2 rocking and NH bending modes for the different conformations in the pure and 109 PhysChemComm, 2002, 5(16), 99–111complexed DMEDA, a change in the dihedral angle of the DMEDA–LiBr and DMEDA–LiTf complexes compared with pure DMEDA is inferred. Comparison of the experimental IR and Raman spectra of DMEDA–LiBr to the calculated DMEDA–Li1 and DMEDA–LiBr frequencies for the three low energy conformations leads us to conclude that the TGT conformation is favored in the DMEDA–LiBr complex.The calculated frequencies of the DMEDA–LiTf complex in both the TGT and TGG conformations are very similar and are in good agreement with the frequencies observed in the DMEDA–LiTf spectra. However, the frequencies and intensities of the TGT conformation provide a closer match with the IR and Raman spectra in the important N–H bending, CH2 rocking and N–H stretching regions. As the composition of LiTf in the DMEDA system increases, the degree of ionic association of the triflate anion appears to increase, changing from a mixture of ‘‘free’’ ions and ion pairs to a mixture that includes a more highly associated aggregate.One triflate aggregate is the triple cation, [Li2Tf]1, in which each triflate ion is coordinated to two lithium cations. The [Li2Tf]1 aggregate may exist as an independent species, chargebalanced by ‘‘free’’ triflate anions, or as part of a more complex structure such as is seen in the diglyme–LiTf system.9 In that system, a combination of vibrational spectroscopy and X-ray diffraction of a diglyme–LiTf crystal shows that this spectroscopic signature (763 cm21 for the ds(CF3) mode) arises from a dimer structure, diglyme2Li2Tf2, in which each lithium is coordinated to three oxygen atoms of a diglyme molecule and one oxygen from each triflate.9 Each triflate then vibrates as an Li2Tf1 although it is part of an Li2Tf2 dimer.It is not unreasonable to hypothesize a similar bonding geometry could exist in the DMEDA–LiTf system. Differences in the frequency shifts of the ds(CF3) and ns(SO3) modes show that the ds(CF3) mode and perhaps the ns(SO3) mode as well is sensitive to anion–solvent interactions, probably through hydrogen-bonding interactions. The addition of lithium salt has a profound effect on the hydrogen bonding in the DMEDA system. It is well known that the N–H stretch is sensitive to hydrogen bonding and that lower frequencies indicate a greater degree of hydrogen bonding.49 There may be both intramolecular and intermolecular hydrogen bonding in DMEDA.19,49 The increase in the experimentally observed frequencies of the N–H stretches upon the addition of LiTf to DMEDA would seem to indicate that hydrogen bonding interactions are decreased. This is consistent with the breaking of the intramolecular hydrogen bond seen in the geometry optimization.The measured frequencies shift to lower wavenumbers in the DMEDA–LiBr system, suggesting that there are stronger hydrogen-bonding interactions in this system than in the pure DMEDA. One possible explanation for the stronger hydrogen bonding in the LiBr system is that there are intermolecular hydrogen bonds between the DMEDA and the Br anion. Because the triflate anion is coordinated to two lithium cations in the DMEDA–LiTf systems at high salt compositions, there are fewer available triflate oxygen atoms to hydrogen-bond to the DMEDA.The presence of more than one N–H stretching band in these systems may indicate that the N–H moieties exist in several different environments, with different types of hydrogen bonding interactions present. Acknowledgement The research described in this publication was made possible by a grant to R.F. from the National Science Foundation (DMR- 0072544). We would like to thank the National Computational Science Alliance for allocations of computer time at the University of Illinois and the University of Kentucky through award number MCA96–N019. 110 PhysChemComm, 2002, 5(16), 99–111 References 1 C. Berthier, W. Gorecki, M. Minier, M. B. Armand, J. M. Chabagno and P. Rigaud, Solid State Ionics, 1983, 11, 91. 2 R. Frech and W. Huang, Macromolecules, 1995, 28, 1246.3 R. Frech, S. Chintapalli, P. G. Bruce and C. A. Vincent, Chem. Commun., 1997, 157. 4 M. Gauthier, A. Belanger, B. Kapfer, G. Vassort and M. Armand, Polym. Electrolyte Rev., ed. J. R. MacCallum and C. A. Vincent, 1989, 1. 5 S. P. Gejji, J. Tegenfeldt and J. Lindgren, Chem. Phys. Lett., 1994, 226, 427. 6 P. Johansson, J. Tegenfeldt and J. Lindgren, Polymer, 1999, 40, 4399. 7 B. L. Papke, M. A. Ratner and D. F. Shriver, J. Electrochem. Soc., 1982, 129, 1434. 8 G. Petersen, P. Jacobsson and L. M. Torell, Electrochim. Acta, 1992, 37, 1495. 9 C. P. Rhodes and R. Frech, Macromolecules, 2001, 34, 2660. 10 Y. Chatani, T. Kobatake, H. Tadokoro and R. Tanaka, Macromolecules, 1982, 15, 170. 11 C. K. Chiang, G.T. Davis, C. A. Harding and T. Takahashi, Macromolecules, 1985, 18, 825. 12 C. K. Chiang, G. T. Davis, C. A. Harding and T. Takahashi, Solid State Ionics, 1986, 18&19, 300. 13 C. S. Harris, D. F. Shriver and M. A. Ratner, Macromolecules, 1986, 19, 987. 14 C. S. Harris, M. A. Ratner and D. F. Shriver, Macromolecules, 1987, 20, 1778. 15 R. Tanaka, T. Fujita, N. Nishibayashi and S. Saito, Solid State Ionics, 1993, 60, 119. 16 S. York, R. Frech, A. Snow and D. Glatzhofer, Electrochim. Acta, 2001, 46, 1553. 17 H. Dong, J.-K. Hyun, C. Durham and R. A. Wheeler, Polymer, 2001, 42, 7809. 18 P. Johansson, Polymer, 2001, 42, 4367. 19 S. E. Boesch, S. S. York, R. Frech and R. A. Wheeler, PhysChemComm, 2001, 1. 20 S. Chintapalli and R. Frech, Electrochim.Acta, 1995, 40, 2093. 21 R. Frech and W. Huang, Solid State Ionics, 1994, 72, 103. 22 R. Frech, W. Huang and M. A. K. L. Dissanayake, Mater. Res. Soc. Symp. Proc., 1995, 369, 523. 23 W. Huang, R. Frech and R. A. Wheeler, J. Phys. Chem., 1994, 98, 100. 24 W. Huang, R. Frech, P. Johansson and J. Lindgren, Electrochim. Acta, 1995, 40, 2147. 25 P. K. Muhuri, B. Das and D. K. Hazra, J. Phys. Chem., 1997, 101, 3329. 26 M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y.Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon, C. Gonzalez and J. A. Pople, GAUSSIAN 94, rev. B.2, Gaussian, Inc., Pittsburgh, PA, 1995. 27 P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. Frisch, J. Phys. Chem., 1994, 98, 11623. 28 J. C. Slater, Quantum Theory of Molecules and Solids, McGraw- Hill, New York, 1974, vol. 4. 29 A. D. Becke, Phys. Rev. A, 1988, 38, 3098. 30 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200. 31 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B, 1988, 58, 785. 32 A. D. Becke, J. Chem. Phys., 1993, 98, 1372. 33 W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab Initio Molecular Orbital Theory, John Wiley & Sons, New York, 1986. 34 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Aala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres,C. Gonzalez, M. Head-Gordon, E. S. Replogle and J. A. Pople, GAUSSIAN 98, rev. A.6, Gaussian, Inc., Pittsburgh, PA, 1998. 35 H. B. Schlegel, J. Comput. Chem., 1986, 3, 214. 36 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502. 37 C. Wasikowski and S. Klemm, XMOL, Research Equipment, Inc. d.b.a. Minnesota Supercomputer Center, Inc., Minneapolis, MN, 42 S. E. Boesch, C. P. Rhodes, R. Frech and R. A. Wheeler, work in progress. 43 W. Huang, R. A. Wheeler and R. Frech, Spectrochim. Acta, 1994, 50A, 985. 44 H. Matsuura, K. Fukuhara and H. Tamaoki, J. Mol. Struct., 1987, 156, 293. 45 A. Ferry, G. Ora�dd and P. Jacobson, J. Chem. Phys., 1998, 108, 7426. 46 A. Ferry, J. Phys. Chem. B, 1997, 101, 150. 47 R. Frech and W. Huang, J. Solution Chem., 1994, 23, 469. 48 R. Frech, S. Chintapalli, P. G. Bruce and C. A. Vincent, Macromolecules, 1999, 32, 808. 49 P. J. Krueger, Can. J. Chem., 1967, 45, 2143. 1993. 38 A. K. Grafton and R. A. Wheeler, J. Comput. Chem., 1998, 19, 1663. 39 A. K. Grafton and R. A. Wheeler, Comput. Phys. Commun., 1998, 113, 78. 40 K. E. Wise, B. Pate and R. A. Wheeler, J. Phys. Chem. B, 1999, 103, 4764. 41 M. R. Razeghifard, S. Kim, J. S. Patzlaff, R. S. Hutchison, T. Krick, I. Ayala, J. J. Steenhuis, S. E. Boesch, R. A. Wheeler and B. A. Barry, J. Phys. Chem. B, 1999, 103, 9790. 111 PhysChemComm, 2002, 5(16), 99–1
ISSN:1460-2733
DOI:10.1039/b204103k
出版商:RSC
年代:2002
数据来源: RSC