Reactive Dynamics in Micelles: Auramine O in ... - ACS Publications

Sep 22, 2010 - The reaction rate at the regular micelle−water interface is slower than in bulk water but significantly faster than in nanoconfined w...
0 downloads 0 Views 1MB Size
J. Phys. Chem. B 2010, 114, 12859–12865

12859

Reactive Dynamics in Micelles: Auramine O in Solution and Adsorbed on Regular Micelles Minako Kondo, Ismael A. Heisler, and Stephen R. Meech* School of Chemistry, UniVersity of East Anglia, Norwich NR4 7TJ, United Kingdom ReceiVed: June 25, 2010; ReVised Manuscript ReceiVed: August 11, 2010

The role of confinement in the suppression of the excited-state reaction of the dye molecule auramine O in nanoscale water droplets is investigated by contrasting the behavior of the dye in solution and on regular and reverse micelles. Auramine O photophysics are studied in bulk water, at the interface between regular micelles and bulk water, and in the aerosol OT (AOT)-stabilized aqueous nanodroplet. It is shown that the reaction of auramine O in bulk water is to a first approximation determined by aqueous solvation dynamics rather than solvent viscosity. This is in contrast to the result in more viscous and slowly relaxing solvents, where the solvent viscosity controls the rate. This result suggests the possibility of multiple reaction pathways on the excited-state surface. The reaction rate at the regular micelle-water interface is slower than in bulk water but significantly faster than in nanoconfined water, indicating a distinct effect of confinement on the reaction rate. It is suggested that the degree of perturbation of the water structure at the interface is the factor controlling the rate of the reaction. Specifically, the water structure is strongly perturbed at the AOT-confined water interface, suppressing the ability of fast collective solvent reorientation to promote the auramine O excitedstate reaction. This effect is less marked at the interface between the micelle and bulk water. The contrast between these results indicates a route whereby confinement modifies ultrafast reaction dynamics in micelles. Introduction The effect of nanoscale spatial confinement on the properties of liquid water has been the topic of intense study.1-6 Much of the interest in nanoconfined water (NCW) derives from its importance in naturally occurring systems. The crowded environment of the living cell comprises about 70% water, much of which is necessarily close to an interface. Confined water layers, typically one to a few molecules thick, are important for protein and membrane stability,7-11 while water-filled cavities containing perhaps only tens of molecules exist in a number of proteins.12 Consequently there are already a number of important studies of the structure and dynamics of such interfacial water. While understanding the nature of water in the living cell is certainly of great importance, such natural systems provide only limited opportunities for modifying or manipulating the structure of the confined water. A system which offers the possibility of controlling both the spatial extent of confinement of NCW and the nature of the interface is the reverse micelle.1,13 Spherical NCW droplets can be stabilized by the surfactant aerosol OT (AOT) in the size range zero to tens of nanometres simply by varying the molar ratio of water to surfactant, w ) [H2O]/[AOT], and using the relationship between w and the droplet radius, rw ) 0.18w.14,15 Thus the AOT system has been widely adopted in studies of NCW.1 Here we test the role of confinement and the interface on reaction dynamics by comparing the excited-state decay of the reactive dye auramine O (AuO) adsorbed at the interface between water and regular micelles (unconfined, negative curvature) with previously reported studies in AOT-stabilized NCW (positive curvature).16-19 Infrared (IR) spectroscopy is a convenient experimental means of studying the structure and dynamics of the AOT-stabilized NCW.20 A number of steady state and transient IR studies * To whom correspondence should be addressed. E-mail: s.meech@ uea.ac.uk.

broadly supported a “core-shell” model of NCW.4,21 However, some measurements suggested a more complex picture.22 The recent transient IR experiments by Moilanen et al23 reported a crossover at rw ≈ 2 nm from collective dynamics to core-shell type behavior in larger droplets, suggesting that the interface can extend its influence on water dynamics up to a distance of 2 nm. These direct studies of dynamics in NCW can be compared with the results of MD simulations. Faeder and Ladanyi found that dynamics associated with water molecules in the interface with AOT were dramatically slowed compared to the bulk, consistent with a core-shell model.24-26 Berkowitz and coworkers found in their simulations that for a charged surfactant the water orientational dynamics were dramatically slowed in the first aqueous solvation layer but that the influence of the interface on the dynamics had almost disappeared at a distance of 1 nm.27,28 Very recently Pieniazek et al.29 combined MD simulations with calculations of IR spectra and found that in smaller AOT-stabilized micelles the effect of the interface was longer range than in larger micelles, in good agreement with the recent experiments of Moilanen et al.23 The properties of NCW have also been investigated through the inclusion of fluorescent probe molecules.30 A large number of studies have been made of solvation dynamics about the increased dipole moment following electronic excitation of a solute in AOT-stabilized NCW.31-35 Measurements have been reported on the femto- to picosecond and pico- to nanosecond time scales; the field has recently been surveyed by Levinger.1 In bulk water, the dominant solvation dynamics are subpicosecond with a minor component on the picosecond time scale.36 In contrast measurements in NCW reveal components of the solvation dynamics on the time scale of tens of picoseconds or even nanoseconds.1 It has been suggested that the nanosecond dynamics may reflect probe interconversion between sites in the reverse micelle.3 The rw dependence of the solvation dynamics is significant; they are essentially frozen (on

10.1021/jp105878p  2010 American Chemical Society Published on Web 09/22/2010

12860

J. Phys. Chem. B, Vol. 114, No. 40, 2010

Figure 1. Structure of AuO and the surfactants sodium dodecylsulfate (SDS) and cetyltrimethylammonium bromide (CTAB).

the fluorescence time scale) in the smallest micelles, while in larger micelles they range from ultrafast (bulk-like) to hundreds of picoseconds.37 These measurements suggest the existence of slow dynamics in AOT-stabilized NCW, which are not observed in IR studies. Similar conclusions can be drawn from studies of orientational motion of probe molecules in micelles. Dutt38,39 observed that probe orientational relaxation became faster with increasing rw but the rate of increase was a function of the charge state of the probe. All of these data are consistent with slow dynamics in NCW but also suggest that the probe location is important in determining the dynamics observed. In many cases the dominant slow dynamics appear to arise from probe molecules preferentially located in the interfacial or shell region. Ladanyi and co-workers extended their MD calculations on NCW to simulate solvation dynamics about a charged probe molecule.40 The positively charged probe was located close to the AOT/NCW interface and exhibited slower solvation dynamics than the core localized probe. In addition to studying the dynamics of NCW there is great interest in understanding the effects of NCW on chemical reactivity. Such information will be critical in, for example, understanding reactivity at protein and membrane interfaces. The two classes of reaction which have been most widely studied are excited-state proton transfer41-45 and reactions involving large scale structural reorganization;16-18,46,47 the latter class are the focus here. The reactive probe molecule which has been most studied in NCW is the cationic dye AuO (Figure 1). Following electronic excitation in solution the phenyl rings undergo a barrierless reorientation to form a dark intermediate state of charge transfer (CT) character.48-51 The large-scale reorientation is opposed by medium friction, so the excitedstate reaction rate is sensitive to the viscosity of the environment.52 In addition the stabilization of the dark CT state means that the rate may also be influenced by the solvent reorganization in the same way as for solvation dynamics. Hirose and co-workers first studied AuO in AOT-stabilized NCW by transient absorption.47 They found that the excited-state decay and the ground-state recovery were suppressed by more than an order of magnitude compared to bulk water, and both were a strong function of rw. We investigated the same reaction by ultrafast fluorescence up-conversion spectroscopy.16-18 The fluorescence decay rate was observed to be sensitive to rw for small micelles (up to 5 nm) but independent of rw in larger micelles, although the fastest decay rate was still five times slower than in bulk water, suggesting that the probe is not localized in a core (bulk-like) region of the micelle. The very fast decay of AuO in bulk water compared even to very large micelles may suggest the operation of some specific quenching mechanism in bulk water.47 The AuO fluorescence decay in NCW is a strong function of emission wavelength, indicating a time dependent fluorescence spectrum. The spectra were

Kondo et al. reconstructed and modeled with a generalized Smoluchowski equation,51 which revealed the population dynamics on the bright to dark excited-state potential energy surface.16 Clearly in terms of both interpreting the results reported by solvation probes of NCW and in understanding the influence of NCW on reaction dynamics it is necessary to separate the effect of confinement from the effect of localization of the probe at the surfactant-NCW interface. In addition, for the specific case of the reactive AuO probe a better understanding of the origin of the ultrafast decay in bulk water is required before the origin of the observed suppression of the reaction can be firmly assigned. To achieve this we undertook measurements of AuO excited-state decay in aqueous solutions and in the presence of different (regular) micelles. In this way the role of the interface in the absence of confinement can be probed. Further we study AuO decay as a function of ionic strength and in a range of solvents. We conclude that there is a water specific channel in the decay of AuO arising from ultrafast solvation dynamics and, critically, that confinement itself leads to a suppression of the excited-state reaction. Some possible mechanisms are discussed. Experimental Section Time resolved fluorescence was measured with an upconversion spectrometer described in detail elsewhere.16 By minimizing dispersive optical elements and carefully balancing group velocity dispersion a time resolution of better than 50 fs is possible.16 In the present experiments additional glass filters were employed to remove scattered light which degraded the resolution to 70 fs. For steady state electronic spectroscopy the AuO concentration was kept at 5 × 10-6 mol dm-3, which gave an optical density (OD) of 0.1-0.2 with 1 cm cuvettes at the 415 nm excitation wavelength. For the fluorescence up-conversion measurements the concentration was 1 × 10-4 mol dm-3 giving 0.2-0.3 OD in 1 mm cuvettes at 415 nm. The NCW samples were prepared from dried AOT as described previously, avoiding multiple occupancy.16 The salt solution was prepared by dissolving sodium sulfate (Na2SO3, Sigma-Aldrich SigmaUltra) in distilled water (analytical reagent grade, Fisher Scientific). The regular (non-inverse) micelle solutions were formed using two surfactants (see Figure 1), SDS (Sigma-Aldrich g99.0%) and CTAB (Sigma 99%). SDS is an anionic surfactant that has an SO4- headgroup and Na+ counterion, similar to AOT. The critical micelle concentration (cmc) of SDS is 8.3 × 10-3 mol dm-3 in water.53 CTAB has a positively charged ammonium headgroup, with a Br- counterion, and a cmc of 0.9 × 10-3 mol dm-3 in water.54 The concentration of SDS was changed from 1 × 10-3 to 0.1 mol dm-3 and CTAB from 0.1 × 10-3 to 0.1 mol dm-3, and therefore, the measurements were performed for AO in solution both above and below the cmc. The fluorescence up-conversion data were fitted to a multiexponential function convoluted with the instrumental response function, which was obtained by up-conversion of the Raman scatter from pure heptane. A sum of three exponentials function, F(t) ) ∑i 3) 1Ai exp(-t/τi), fit the data around the maximum emission wavelength. On the red and blue edge an additional damped sine function [i.e., F(t) ) ∑3i ) 1Ai exp(-t/τi) + A4 sin(ωt + φ) exp(-t/τV)] was required to account for the contribution of a coherently excited mode in the excited state, as described previously.19 Finally the mean fluorescence lifetime is calculated from the three exponential decay terms according to the equation 〈τ〉 g ∑Aiτi/∑Ai.

Auramine O in Solution and Adsorbed on Regular Micelles

J. Phys. Chem. B, Vol. 114, No. 40, 2010 12861

Figure 2. The mean fluorescence lifetime, 〈τ〉, of AuO in several solvents plotted as a function of the viscosity (filled circles, lower axis) and mean solvation time (crosses, upper axis). The arrow highlights the anomalous result for the viscosity dependence in liquid water.

Results and Discussion AuO in Solution. The excited-state reaction of AuO in solution is well characterized.48-52 In a series of alcohol solvents or as a function of temperature the reaction was shown to be sensitive to medium viscosity.50,51 The reaction coordinate in the excited state involves the barrierless conversion from an emissive locally excited state to a dark CT state through rotation of the phenyl rings.52 The solvent friction resists the phenyl group rotation conferring the viscosity dependence. The CT state subsequently decays back to the ground state in a second step. Our measurements (Figure 2) of the mean fluorescence decay time of AuO in a series of fluid solvents, measured at the peak of the fluorescence intensity, are broadly consistent with this picture. The important exception is the aqueous solution of AuO, which has a much faster decay than expected on the basis of the viscosity of water. The same anomaly was noted by Hirose and co-workers in their transient absorption study of AuO.47 They also noted that the fast decay time was close to the solvation time in water as measured using the Stokes shift method.36 Since we are interested here in understanding the excited-state reaction of AuO in NCW this anomaly in bulk water deserves further investigation. To examine the role of solvation dynamics in the excitedstate reaction of AuO we measured its time dependent fluorescence spectra in aqueous solution. In Figure 3a the wavelengthresolved fluorescence decays are shown. The decay on the blue edge is ultrafast and highly nonexponential. The mean decay time increases as the observation wavelength increases. The decay becomes approximately exponential on the red edge of the emission, with no discernible rise time-resolved, a feature which distinguishes these observations from conventional solvation dynamics associated with a non reactive solute. These data were combined with the steady state spectrum to recreate the time dependent emission spectra,55,56 which were fit to a log-normal function (Figure 3b). The dynamics of AuO on the excited-state potential energy surface have previously been modeled by a one-dimensional generalized Smoluchowski equation,51 which yields the time dependent population density, F(z,t), along the reaction coordinate (z)

(

)

∂ ∂ ∂ 1 ∂ F(z, t) ) D(t) + S (z) F(z, t) - κΓ(z)F(z, t) ∂t ∂z ∂z kBT ∂z r

Figure 3. (a) The wavelength resolved fluorescence up-conversion data and (b) the time-resolved fluorescence spectra of AuO in water.

in which D(t) is the (in general) time dependent diffusion coefficient. The excited-state potential energy surface Sr(z) is obtained from coupling a harmonic emissive state with a harmonic dark state, assumed to be of CT character.51 For simplicity ground, emissive and dark states are assumed to have the same force constant, although different energies, and the coupling constant was fixed for all calculations at 800 cm-1. In the second term on the right-hand side, κ is the rate coefficient associated with relaxation from the CT (dark) state to the ground state, which is assumed to also depend upon z in a fashion modeled by Γ(z), a Gaussian “sink” function.57 When this analysis was applied to model the temperature dependent excited-state dynamics of AuO in alcohol solvents a static time independent diffusion coefficient was recovered, which scaled with temperature divided by viscosity.51 This is entirely consistent with diffusive dynamics on the excited-state potential energy surfaces and thus the reported viscosity dependence of the AuO reaction. When the same model was applied to the decay of AuO in NCW a good fit was only recovered if a time dependent diffusion coefficient was employed,16 in which58-60

D(t) ) -〈(δz)2〉

˙ (t) ∆ ∆(t)

and ∆(t) is a normalized reaction coordinate time correlation function with δz ) z - zeq. For a reaction that is driven by solvent reorientation to stabilize a new charge distribution (as

12862

J. Phys. Chem. B, Vol. 114, No. 40, 2010

Figure 4. (a) The time-dependent excited-state population distribution, F(z,t), on the excited-state potential energy surface of AuO in water. The dashed line represents the potential energy surface of AuO in decanol. (b) The first moment of AuO in water with the data fitted by the Smoluchowski analysis utilizing a time dependent diffusion coefficient.

may be the case for a charge separation reaction such as occurs in AuO49) the ∆(t) may be approximated by the experimentally determined solvation time correlation function, C(t). This approach was only partially successful in NCW, especially when applied to AOT with several different counterions.19 Here we fit the time dependent spectra for AuO in bulk water assuming a biexponential solvation correlation function; this extends our previous analysis where only a static (though anomalously large) diffusion coefficient was assumed.16 The D(t) so calculated yields the F(z,t) which are in turn used to calculate the time dependent fluorescence spectra, Ifl

Ifl ∝

∫ dzg(ν0(z), ν(z) - ν0(z))M2(z)F(z, t)ν3

in which M(z) is a normalized transition moment scaling between 1 and 0 as the population moves along the reaction coordinate from bright to dark (CT) state and g(ν0(z),ν(z) ν0(z))is a log-normal line shape function accounting for the Franck-Condon factor. The calculated spectra are then compared with those reconstructed from the decay data. A good quality fit was obtained over the time where signal-to-noise is high with C(t) ) 0.7 exp(-t/0.2 ps) + 0.3 exp(-t/0.8 ps) which yields the F(z,t) of Figure 4a and the fluorescence spectra in 4b, where the calculated and experimental data are compared. This result may be compared with earlier measurement of C(t) for nonreactive dyes, which yielded a two- or threecomponent solvation response with a mean solvation time of

Kondo et al. between 250 and 350 fs.36,61 This is in good agreement with the mean C(t) recovered here (380 fs), which is thus consistent with aqueous solvation dynamics being a major factor determining the excited-state reaction of AuO in water, rather than the solvent friction. Thus the ultrafast collective water reorientation which yields the ultrafast solvation response62 is also the driving force for the nuclear and electronic structure change as AuO evolves from the locally excited to CT state in aqueous solution. We note that to obtain the fit in Figure 4b it was also necessary to lower the energy of the directly excited state slightly and the CT state significantly compared to that which had been used to fit alcohol solutions (the reactive potential energy surface used to fit the data recovered for the relatively nonpolar decanol solution16 are shown in Figure 4a for comparison). These modifications are consistent with the greater ability of a highly polar solvent like water to stabilize a CT state. This result raises the question why does viscosity apparently control the AuO excited-state reaction in alcohol and other solvents, while solvation dynamics is the determining factor in bulk aqueous solution? The viscosity dependence is illustrated by the data in Figure 2 in which three solvents (acetonitrile, methanol, and ethanol) show a linear relationship between decay time and viscosity. However, for these fluid solvents an equally good linear dependence of lifetime on the mean solvation time is observed (0.26, 5, and 16 ps, respectively63); critically water also fits on this line. Thus, we conclude that in water (and possibly other solvent with ultrafast solvation times, Figure 2) solvation dynamics plays a key role in determining the reaction rate, while in low temperature or very slowly relaxing solvents macroscopic viscosity is the controlling factor.51,52 A plausible explanation for this result is that the one-dimensional reaction coordinate implicit in the Smoluchowski model is too simple to describe the AuO reaction. Specifically, following ideas developed by Hynes,64 we suggest that the reaction coordinate in general involves both solvation and nuclear reorganization coordinates. If the solvation dynamics are sufficiently fast they can promote a facile barrierless formation of the CT state. On the other hand slower solvation dynamics (or less polar media) require diffusive motion from the initial locally excited state to the final CT state, possibly over a small barrier. In terms of this picture alcohol solvents at low temperature or with long alkyl chains lead to reaction by the viscosity dependent mechanism, while polar solvents exhibiting fast solvation, such as water, promote a faster solvation driven reaction. We note that the very fast solvation dynamics in liquid water are not the result of reorientation of isolated solvent molecules but rather a collective reorganization of the H-bonded network.62 To conclude the present data suggest that both solvation dynamics and intramolecular viscosity dependent reorganization contribute to the excited-state reaction rate, and which dominates depends on the medium. AuO in Aqueous Surfactant Solution. The AuO absorption spectrum is unchanged by the addition of the anionic SDS until the surfactant concentration reaches 5 mM, when a red shift is observed. The spectrum shifts further with addition of more SDS up to 10 mM, above which no further change occurs. The corresponding emission intensity is shown in Figure 5. Between 0 and 5 mM SDS the fluorescence intensity increases steadily. We assign this increase in yield to the preferential solvation of AuO by SDS molecules or premicellar aggregates, leading to the disruption of the water solvation layer around AuO. As described above, this may lead to relatively less efficient formation of the dark state by slowing the aqueous solvation dynamics. The rate of increase in fluorescence with concentra-

Auramine O in Solution and Adsorbed on Regular Micelles

Figure 5. Fluorescence intensity (crosses) and mean fluorescence lifetime (filled circles) of AuO as a function of SDS concentration (arbitrarily normalized at 50 mM SDS for comparison).

tion decreases sharply at ca. 10 mM SDS (Figure 5). Fitting straight lines to the regions above and below 10 mM reveals a crossing point at 7.3 mM, which is close to the literature value reported for the cmc of SDS in water (8.3 mM).53 Thus AuO is sensitive to the cmc of SDS, and the subsequent small increase in intensity with SDS concentration above the cmc can be ascribed to adsorption of further AuO molecules as the number density of micelles increases. A similar set of measurements was made for the cationic surfactant CTAB. In this case there is no shift in the absorption spectrum and the fluorescence intensity does not show any variation as the concentration is increased from zero to the cmc (0.9 mM). This lack of interaction between CTAB and AuO can reasonably be ascribed to Coulombic repulsion between these positively charged species. As the concentration of CTAB increased further a small increase in AuO fluorescence intensity was observed, which may indicate that at high concentration the solvation structure around AuO is disrupted by CTAB, or may simply indicate the onset of a viscosity effect on the AuO decay, since the viscosity of a 0.1 M CTAB solution is 20 to 25% higher than water. These results strongly suggest that the cationic dye AuO is adsorbed at the water micelle interface of the anionic surfactant SDS. In addition it is apparent that the excited-state reaction leading to quenching of AuO fluorescence is suppressed at the interface relative to bulk water. In contrast AuO does not adsorb at the interface of a CTAB micelle, or if it does the excitedstate reaction is unaffected. Adsorption, Medium, and Confinement Effects on the AuO Reaction. Fluorescence up-conversion data for AuO in several media are shown in Figure 6, measured at the maximum of the fluorescence intensity. The corresponding fitting data are presented in Table 1. In these data we include as reference points the fluorescence decay in neat water and in a large (rw ) 10 nm) AOT-stabilized reverse micelle. In the latter case this corresponds to ca. 50 water molecules per surfactant and is the largest NCW droplet studied; all micelles with a smaller rw have a longer relaxation time.16 The AuO decay in NCW and the micellar media should not be compared directly with an aqueous solution of AuO, since the surfactant necessarily modifies the ionic strength of the solution. Any ionic strength effect will be most significant for AOT-stabilized NCW, where there is one sulfate headgroup (two ions) for 50 water molecules. In Figure 6 and Table 1, 1 M Na2SO4 dissolved in an aqueous solution of AuO is shown to have a negligible effect on the fluorescence decay time. Thus

J. Phys. Chem. B, Vol. 114, No. 40, 2010 12863

Figure 6. Fluorescence up-conversion data of AuO in water (black), Na2SO4 (blue), 3 mM SDS (green), and 3 mM CTAB (red).

TABLE 1: Fitting Components of the Fluorescence Up-Conversion Data for AuO in Several Solvents at the Maximum Emission Wavelength sample water AOT reversed micelle rw ) 10 Na2SO4 (mol dm-3) 0.25 0.5 1 NaCl 5 mol dm-3 SDS (× 10-3 mol dm-3) 3 4 5 8 10 20 50 100 CTAB 3 × 10-3 mol dm-3

τ1/ps

A1

τ2/ps

A2

τ3/ps

A3

〈τ〉/ps

0.14 0.78 0.42 0.22 0.33 0.58 1.64 0.34 7.33 0.08

0.20 1.36

0.22 0.20 0.20 0.10

0.93 0.89 0.87 0.90

1.01 0.70 0.76 0.84

0.07 0.11 0.13 0.10

0.28 0.26 0.27 0.18

0.12 0.13 0.12 0.18 0.15 0.14 0.17 0.19 0.13

0.65 0.64 0.57 0.56 0.55 0.48 0.52 0.56 0.87

0.67 0.83 0.65 0.84 0.73 0.63 0.77 0.91 0.47

0.30 0.31 0.35 0.36 0.35 0.40 0.37 0.34 0.13

3.17 3.65 3.17 3.94 3.22 3.27 3.74 4.45

0.05 0.04 0.07 0.07 0.10 0.12 0.11 0.10

0.44 0.50 0.54 0.69 0.67 0.70 0.79 0.84 0.18

we can exclude simple ionic strength effects from being the origin of the slow excited state decay for AuO in NCW. The effect of increasing surfactant concentration on the mean fluorescence decay time is consistent with the steady state data. The AuO decay time is unaffected by CTAB but becomes longer in the presence of SDS (Figure 6). The mean decay time is sensitive to the cmc of SDS and then increases by less than 10% as the SDS concentration increases up to 100 mM, the maximum concentration studied. The increase in mean lifetime follows the increase in fluorescence yield, as expected (Figure 5). The most significant result of Figure 6 is that the excitedstate reaction of AuO at the SDS-water interface is significantly faster than that of AuO at the AOT-NCW interface, but slower than in bulk water (Figure 6, Table 1). The principal objective here is to investigate the effect of confinement on the excited-state reaction of AuO. The comparison (above) of the fluorescence decay of AuO at the SDS and AOT-stabilized interfaces confirms the significant effect of confinement on the reaction rate (which is also consistent with studies using nonionic surfactants18). Further analysis requires a microscopic picture of the structure of the “internal” (AOT-NCW) and “external” (SDS-water) interfaces. Important information can be obtained from the MD simulations of

12864

J. Phys. Chem. B, Vol. 114, No. 40, 2010

surfactant-water systems that have appeared in the past decade. Internal interfaces (AOT and other surfactants) have been most thoroughly investigated.24-26,40,65,66 The external interface has received rather less attention, but Bruce and co-workers presented detailed simulations of structure and dynamics at the SDS-water interface.67,68 They found a strong perturbation of the orientational dynamics of water molecules in the first solvation shell, with a much smaller effect on translational diffusion. For the AOT-NCW interface Faeder et al. identified a population of essentially trapped water molecules, which had orientational diffusion suppressed in a manner similar to that observed at the SDS-water interface, but they found that the translational diffusion was also strongly suppressed.25 The structures of the two interfaces were also studied in these simulations. Both agree that there are effectively two shells of water molecules, one highly restricted by direct interaction with the surfactant headgroup and a second molecular layer bound to the first which exhibits substantial deviations from bulk-like dynamics. Beyond 1-nm bulklike water dynamics dominate. In addition there is a shell of sodium counterions associated with the sulfate headgroup for both interfaces. This shell is quite diffuse for SDS-water but rather well-defined for small AOT-NCW micelles; however, as rw increases this counterion shell becomes more diffuse, and some sodium ions are solvated in the center of the NCW.65 A significant difference between the local structures of the two interfaces was found in these simulations. For the smallest NCW droplets the interface is calculated to have a pseudo lattice structure in which the sodium ions are hexacoordinated both by water molecules and the AOT sulfate oxygen atoms. As rw increases this structure is loosened somewhat with intervening water molecules.65 The less dense arrangement of the SDS headgroups in the external interface results in a less ordered structure.68 This degree of order has important effects on the water H-bonding structure at the interfaces. It is difficult to compare quantitatively the results of the two calculations, which used different definitions of the H-bond, but it can safely be concluded that the mean number of water-water H-bonds in the first solvation shell is larger for the SDS-water than for the AOT-NCW interface, and the bulk water value (3.7 H-bonds per water) is attained in a shorter distance from the SDS-water interface.65,68 The study of AuO in aqueous solution showed that the anomalously fast reaction dynamics were accurately modeled assuming a time dependent diffusion coefficient close to that expected from the solvation dynamics of bulk water. Thus, the fastest reaction probably comprises a coupling of the AuO excited-state reorganization with aqueous solvation dynamics, which are themselves driven by collective reorganization of the aqueous environment.62 The collective MD in liquid water (and thus NCW) requires a reorganization of the H-bond network. At the interface this network is disrupted by the formation of water-surfactant H-bonds and solvation of the counterions. The MD calculations described above suggest that this effect may be more extensive at the AOT-NCW interface than at the SDS-water interface, consistent with the observed slower dynamics in the AOT-NCW. However, the actual range of this disruption is small (