Fourier-transform infrared (FTIR) and time-resolved IR spectroscopies have been used to study vibrational band positions, vibrational energy relaxation (VER) rates, and reorientation times of anions in several ionic liquid (IL) solutions. The ILs primarily investigated are based on the 1-butyl-2,3-dimethylimidazolium ([BM(2)IM]) cation with thiocyanate (NCS-), dicyanamide (N(CN)2-), and tetrafluoroborate (BF4-) anions. Spectroscopic studies are carried out near 2000 cm-1 for the C[Triple Bond]N stretching bands of NCS- and N(CN)2- as the IL anion as well as for NCS-, N(CN)2-, and azide (N3-) anions dissolved in [BM2IM][BF4]. The VER studies of N(CN)2- are reported for the first time. VER of N3-, NCS-, and N(CN)2- is measured in normal solvents, such as N-methylformamide, to compare with the IL solutions. The spectral shifts and VER rates of the anions in IL solution are quite similar to those in polar aprotic, conventional organic solvents, i.e., dimethylsulfoxide, and significantly different than those in methanol, in which there is hydrogen bonding. Similar studies were also carried out for the anions in another IL, 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF4]), in which the C2 hydrogen is present. The results for the anions are similar to those in the [BM2IM] containing ILs, in which the C2 hydrogen is methyl substituted. This suggests that substituting this hydrogen has, at most, a minor effect on the degree of hydrogen bonding in the anion-IL solvation interaction based on the infrared spectra and dynamics.

Ultrafast infrared spectroscopy has been used to measure vibrational energy relaxation (VER) and reorientation (Tr) times for the high frequency vibrational bands of potassium ferrocyanide and ferricyanide (CN stretches), and sodium nitroprusside (SNP, CN, and NO stretches) in water and several other solvents. Relatively short VER times (4-43 ps) are determined for the hexacyano species and for the NO band of SNP, but the CN band of SNP relaxes much more slowly (55-365 ps). The solvent dependence of the VER times is similar for all the solutes and resembles what has been previously observed for triatomic molecular ions [Li et al., J. Chem. Phys. 98, 5499 (1993)]. Anisotropy decay times are also measured from the polarization dependence of the transient absorptions. The Tr times determined for SNP are different for the different vibrational bands; for the nondegenerate NO mode of nitroprusside (SNP) they are much longer (>15 ps), correlate with solvent viscosity, and are attributed to overall molecular rotation. The short Tr (<10 ps) times for the CN band in SNP and for the hexacyanoferrates are due to dipole orientational relaxation in which the transition moment rapidly redistributes among the degenerate modes. There is no evidence of intramolecular vibrational relaxation (IVR) to other high frequency modes. VER times measured for hexacarbonyls and SNP in methanol are similar, which suggests that the generally faster VER for the latter is in part because they are soluble in more strongly interacting polar solvents. The results are compared to those for small ions and metal carbonyls and are discussed in terms of the importance of solute charge and symmetry on VER.

Ultrafast infrared spectroscopy has been used to measure vibrational energy relaxation (VER) and reorientation times (T r) for the antisymmetric stretching band of azide ion (N3 -) in several reverse micelle (RM) systems using cationic, anionic, and nonionic surfactants. RMs were formed using H2O for all surfactants and D2O for the anionic and cationic surfactants. The vibrational dynamics depended on the RM charge. The charge dependence is attributed to differences in ion location in the RM because of Coulombic interactions. The VER times in anionic (AOT, sodium bis(2-ethylhexyl) sulfosuccinate) RMs are indistinguishable from those in bulk solution and T r times are longer only for the smallest RMs studied. In cationic (CTAB, cetyltrimethylammonium bromide) RMs, VER and T r times are longer and weakly depend on the RM water content and water pool size. The results are attributed to the azide anion being attracted toward the RM wall when it is cationic and repelled into the bulklike center when it is anionic. VER times are also longer in small nonionic RMs (NP and Brij-30) but, unlike cationic RMs, approach bulk behavior as the RM size is increased. Comparative studies are performed using mixtures of water and tri(ethylene glycol) monomethyl ether (TGE) in which the latter resembles the hydrophilic portion of the nonionic surfactants. The similar results for nonionic RMs and TGE/water mixtures provide evidence that water penetrates into and hydrates the poly-oxo chains in the nonionic RMs before a water pool is formed. The interfacial region in nonionic RMs include the poly-oxo chains, which appear to be hydrated, so that the boundary between the interface and the water core is less clearly defined than for the ionic RMs.

We present an analysis of the Duschinsky effect and its application to real molecules. We discuss the many subtle aspects of applying the theory to calculations and give examples of a nonlinear normal coordinate transformation. We show how to judge if nonlinear effects are small enough to be neglected through use of the zero-order axis-switching approximation, which allows calculation of Franck-Condon factors (FCF). However, even with the zero-order axis-switching approximation, nonorthogonality can occur in the Duschinsky matrix, and this must be corrected to allow proper FCF calculations. We have calculated the Duschinsky effect for two systems that form the anion in an electron-transfer ion pair, V(CO) 6and Co(CO) 4 -. The formation of the D 3d neutral vanadium species is accompanied by a small geometric distortion and small Duschinsky effect, despite the change in point group from O h . We discuss how to perform the calculations to properly represent degenerate vibrations and how to test if the linear approximation is adequate. The tetrahedral cobalt anion undergoes a much larger geometrical distortion, which results in a larger Duschinsky effect, upon formation of the nearly C 3V neutral species. The analysis of the cobalt system, with a C 1 symmetry for the neutral, demonstrates the methods required when there is no simplification from symmetry. These two examples show the validity of the zero-order axis-switching approximation. The cobalt complex has much larger reorganization energy and a much greater dependence of reorganization energy on the choice of reference state, as expected when the Duschinsky effect is larger. We briefly outline the method of applying these computations to electron-transfer rate calculations.

We systematically test how the Duschinsky mixing of normal coordinate vibrations affects transition rates for electron transfer (ET). We find that ET rates in the inverted region can increase many orders of magnitude from Duschinsky mixing, and both totally symmetric and nontotally symmetric vibrations are very important. The Duschinsky effect arises when two electronic states have vibrational normal mode coordinate systems that are rotated and translated relative to each other. We use a conventional quantum rate model for ET, and the examples include 6-8 vibrations, where two vibrational modes are mixed with different amounts of coordinate rotation. The multidimensional Franck-Condon factors (FCF) are computed with standard algorithms and recently developed recursion relations. When displaced, totally symmetric modes are involved, rates with Duschinsky mixing can increase several orders of magnitude for inverted electron transfer reactions and modest mixing. The peak location in a rate vs energy gap plot can depend on the degree of Duschinsky mixing, and therefore it corresponds to a sum of solvent and an effective vibrational reorganization energy that is not predictable by simple models that exclude mixing. In addition, for some examples of inverted region ET we observe significant flattening of the usual parabolic curve at large degrees of mixing. We demonstrate that large rate effects can occur with very little change in either the calculated absorption or emission spectra, depending on the details of the Duschinsky mixing. The origin of the rate effect is the increased FCF between the initial vibrational state and the higher lying final vibrational states when the Duschinsky effect is taken into account. The rate effect of totally symmetric modes is greater than nontotally symmetric modes, but since there are many nontotally symmetric modes, in real molecules these modes can make a large total contribution to ET rates.