Durba Roy

An improved four-site ionic liquid model.

Computer simulations provide insight into the molecular-level details responsible for the unique properties of ionic liquids. Due to the sluggish dynamics and nanostructured nature of many ionic liquids, coarse-grained models are an important complement to fully atomistic simulations because they enable simulation of much larger system sizes and much longer times, which are often of interest. This paper reports a four-site, coarse-grained model for studying ionic liquids and their solutions. It is intended to be a generic model representative of common ionic liquids currently in use, but it is parametrized to fit the properties of 1-butyl-3-methylimidazolium hexafluorophosphate, [Im(41)][PF(6)]. The present model is a variant of one introduced in J. Phys. Chem. B 114, 8410 (2010). Reduction of ion charges to ±0.78e and fine-tuning Lennard-Jones parameters from the original model leads to a remarkable improvement in the realism of the model and surprisingly good agreement between simulation and experiment for a variety of static and dynamic properties of [Im(41)][PF(6)]. This idealized model should prove valuable for studies of solute-based dynamics and other phenomena occurring on nanosecond and longer time scales, which are not feasible with all-atom simulations.

Measurements of the complete solvation response of coumarin 153 in ionic liquids and the accuracy of simple dielectric continuum predictions

The complete solvation response of coumarin 153 (C153) has been determined over the range 10(-13)-10(-8) s in a variety of ionic liquids by combining femtosecond broad-band fluorescence upconversion and picosecond time-correlated single photon counting measurements. These data are used together with recently reported dielectric data in eight ionic liquids to test the accuracy of a simple continuum model for predicting solvation dynamics. In most cases the features of the solvation response functions predicted by the dielectric continuum model are similar to the measured dynamics of C153. The predicted dynamics are, however, systematically faster than those observed, on average by a factor of 3-5. Computer simulations of a model solute/ionic liquid system also exhibit the same relationship between dielectric predictions and observed dynamics. The simulations point to spatial dispersion of the polarization response as an important contributor to the over-prediction of solvation rates in ionic liquids.

Dynamics in an idealized ionic liquid model.

An idealized four-site ionic liquid model having characteristics approximating those of 1-butyl-3-methylimidazolium hexafluorophosphate ([Im(41)][PF(6)]) is introduced as a low-cost alternative to existing all-atom models for purposes of simulating solute-based dynamics over nanosecond and longer time scales. The structural and energetic properties of the model are in reasonable agreement with those of [Im(41)][PF(6)] and similar ionic liquids, but the dynamics are unrealistically slow. A temperature shift of approximately 100 K is required to produce agreement between the viscosity and diffusion coefficients of the model and experimental values. Several aspects of the ion dynamics such as subdiffusive translational motions, non-Gaussian van Hove distributions, and jumplike displacements in both positions and orientations, are similar to behavior observed in supercooled liquids. Translational diffusion coefficients and rotational correlation times show roughly the proportionalities to viscosity expected from hydrodynamic models, and slip hydrodynamic calculations provide reasonable accuracy in some cases. But anomalously high rotational diffusion coefficients which decouple from viscosity at low temperature are also observed. These anomalies are explained in terms of the prevalence of 180 degrees rotational jumps coupled to the presence of marked heterogeneity in rotational motions, especially about one molecular axis. Comparisons between the dynamics observed in the ionic liquid (IL) model and a neutral mixture (NM) counterpart help to explain the origins of the distinctive dynamics in ionic liquids compared to conventional solvents. The requirement for balancing electrostatic interactions in the IL leads to uniform and interleaved distributions of cations and anions resembling a distorted ionic lattice, similar to the structure of molten NaCl. The resistance to reorganizing this structure is what leads to the slow dynamics of ionic liquids. The coupling among large collections of ions is presumably responsible for the similarity of ionic liquids to supercooled conventional liquids.

Temperature dependence of solvation dynamics and anisotropy decay in a protein: ANS in bovine serum albumin

Temperature dependence of solvation dynamics and fluorescence anisotropy decay of 8-anilino-1-naphthalenesulfonate (ANS) bound to a protein, bovine serum albumin (BSA), are studied. Solvation dynamics of ANS bound to BSA displays a component (300 ps) which is independent of temperature in the range of 278-318 K and a long component which decreases from 5800 ps at 278 K to 3600 ps at 318 K. The temperature independent part is ascribed to a dynamic exchange of bound to free water with a low barrier. The temperature variation of the long component of solvation dynamics corresponds to an activation energy of 2.1 kcal mol(-1). The activation energy is ascribed to local segmental motion of the protein along with the associated water molecules and polar residues. The time scale of solvation dynamics is found to be very different from the time scale of anisotropy decay. The anisotropy decays are analyzed in terms of the wobbling motion of the probe (ANS) and the overall tumbling of the protein.

Simulations of Solvation and Solvation Dynamics in an Idealized Ionic Liquid Model

Equilibrium and nonequilibrium molecular dynamics simulations of solvation and solvation dynamics of a variety of solutes have been performed in the coarse-grained ionic liquid model ILM2 (Roy, D.; Maroncelli, M. J. Phys. Chem. B 2010, 114, 12629). Some comparisons are made between ionic and dipolar solvation using parallel simulations in CH(3)CN. Despite the fact that the multipolar character of electrostatic interactions and their spatial extent differ in the two solvents, solvation energies are equal to within about 10% in ILM2 and CH(3)CN. This near equality also holds with reduced accuracy in the case of reorganization energies. Solvation energies of spherical solutes in ILM2 and its variants can be correlated as a function of solute and solvent size using a Born-type expression with an effective cavity size. Solvation time correlation functions in ILM2 exhibit a subpicosecond inertial component followed by a broadly distributed component related to solvent viscosity, comparable to what has been observed in experiment. Direct comparison of simulation to experiment using the solute coumarin 153 (C153) shows general agreement on the time scales and character of the fast and slow components, but the amplitude of the fast component is overestimated by the simulations. Solute motion can significantly increase the speed of solvation, even in the case of large solutes such as C153. Good agreement is found between linear response estimates and the nonequilibrium dynamics associated with electronic excitation of C153. In contrast, perturbations involving changes of a full electron charge in atomic solutes lead to local heating which greatly hastens solvation compared to linear response predictions. The mechanism of charge solvation in atomic solutes is examined in some detail. It is found that ion translation dominates the inertial dynamics. The rotational contribution only becomes comparable to the translation contribution in the tail of the response. Adjustments of ion positions over distances of ~30% of their diameters are all that is required to relax the solvation energy in these systems.

A femtosecond study of excitation wavelength dependence of solvation dynamics in a PEO-PPO-PEO triblock copolymer micelle

Excitation wavelength (lambdaex) dependence of solvation dynamics of coumarin 480 (C480) in the micellar core of a water soluble triblock copolymer, PEO20-PPO70-PEO20 (Pluronic P123), is studied by femtosecond and picosecond time resolved emission spectroscopies. In the P123 micelle, the width of the emission spectrum of C480 is found to be much larger than that in bulk water. This suggests that the P123 micelle is more heterogeneous than bulk water. The steady state emission maximum of C480 in P123 micelle shows a significant red edge excitation shift by 25 nm from 453 nm at lambdaex=345 nm to 478 nm at lambdaex=435 nm. The solvation dynamics in the interior of the triblock copolymer micelle is found to depend strongly on the excitation wavelength. The excitation wavelength dependence is ascribed to a wide distribution of locations of C480 molecules in the P123 micelle with two extreme environments-a bulklike peripheral region with very fast solvent response and a very slow core region. With increase in lambdaex, contribution of the bulklike region having an ultrafast component (< or =2 ps) increases from 7% at lambdaex=375 nm to 78% at lambda(ex)=425 nm while the contribution of the ultraslow component (4500 ps) decreases from 79% to 17%.

Ultrafast fluorescence resonance energy transfer in a micelle

Ultrafast fluorescence resonance energy transfer (FRET) from coumarin 153 (C153) to rhodamine 6G (R6G) is studied in a neutral PEO(20)-PPO(70)-PEO(20) triblock copolymer (P123) micelle and an anionic micelle (sodium dodecyl sulfate, SDS) using a femtosecond up-conversion setup. Time constants of FRET were determined from the rise time of the acceptor emission. It is shown that a micelle increases efficiency of FRET by holding the donor and the acceptor at a close distance (intramicellar FRET) and also by tuning the donor and acceptor energies. It is demonstrated that in the P123 micelle, intramicellar FRET (i.e., donor and acceptor in same micelle) occurs in 1.2 and 24 ps. In SDS micelle, there are two ultrafast components (0.7 and 13 ps) corresponding to intramicellar FRET. The role of diffusion is found to be minor in the ultrafast components of FRET. We also detected a much longer component (1000 ps) for intramicellar FRET in the larger P123 micelle.