Ultrafast Dynamics in the Dispersed Phase of Oil-In-Water

Time-resolved optical Kerr effect spectroscopy has been used to probe the molecular environment afforded by the hydrophobic core of oil-in-water ...
0 downloads 0 Views 164KB Size
Download PDF
1238

Langmuir 2005, 21, 1238-1243

Ultrafast Dynamics in the Dispersed Phase of Oil-In-Water Microemulsions: Monosubstituted Benzenes Incorporated into Dodecyltrimethylammonium Bromide (DTAB) Aqueous Micelles Andrew A. Jaye, Neil T. Hunt, and Stephen R. Meech* School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich NR4 7TJ, United Kingdom Received September 28, 2004. In Final Form: November 8, 2004 Time-resolved optical Kerr effect spectroscopy has been used to probe the molecular environment afforded by the hydrophobic core of oil-in-water microemulsions. This was achieved by measuring the ultrafast dynamics of a series of benzene derivatives (benzonitrile, nitrobenzene, fluorobenzene, styrene, and toluene) incorporated as the oil phase within oil-in-water microemulsions and comparing them to the dynamics in neat liquid and the liquid diluted in nonpolar solvent. Polar and strongly interacting liquids (benzonitrile and nitrobenzene) showed dynamics in the microemulsion that are similar to those in the solution phase, while weakly interacting and mildly polar liquids (fluorobenzene, styrene and toluene) reveal dynamics more similar to those of the neat liquid. This suggests stabilization of the polar dispersed phase in polar regions of the micelle.

Introduction Microemulsions are thermodynamically stable singlephase complex fluids composed of an oil phase and an aqueous phase, made miscible by an amphiphilic surfactant molecule.1 The system consists of nanometer-sized oil droplets dispersed in water (or vice versa) and their structure (droplet diameter, interparticle distance, etc.) can be varied by changing the ratio of oil to water and by choosing a suitable surfactant molecule.2 Microemulsions have proved valuable in many areas of chemistry, such as nanolatex synthesis, reactions in confined geometries, drug delivery, and oil recovery.1 Their properties have been studied extensively by experimental techniques ranging from conductivity3,4 and viscosity2,4 to light and neutron scattering.5,2 Despite large amounts of published work relating to microemulsion structure, there are only a few studies6-14 that focus on the liquid dynamics of the dispersed phase. Most of these have studied the dynamics * Corresponding author. E-mail: [email protected]. (1) Paul, B. K.; Moulik, S. P. Curr. Sci. 2001, 80, 990. (2) Cazabat, M. Fisica Degli Anfifili: Micelle, Vesicole e Microemulsioni; Degiorgio, V., Ed.; North-Holland: Amsterdam, 1983; pp 723753. (3) Forgiarini, A.; Esquena, J.; Gonzalez, C.; Solans, C. Langmuir 2001, 17, 2076-2083. (4) Hunt, N. T.; Jaye, A. A.; Meech, S. R. J. Phys. Chem. B 2003, 107, 3405. (5) Handbook of Surface and Colloid Chemistry; Birdi, K., Ed.; CRC Press: Boca Raton, FL, 1997. (6) Boyd, J. E.; Briskman, A.; Sayes, C. M.; Mittleman, D.; Colvin, V. J. Phys. Chem. B 2002, 106, 6346. (7) Boyd, J. E.; Briskman, A.; Colvin, V.; Mittleman, D. M. Phys. Rev. Lett. 2001, 87, 147401. (8) Venables, D. S.; Huang, K.; Schumuttenmaer, C. A. J. Phys. Chem. B 2001, 105, 9132. (9) Willard, D. M.; Riter, R. E.; Levinger, N. E. J. Am. Chem. Soc. 1998, 120, 4151. (10) Riter, R. E.; Undiks, E. P.; Levinger, N. E. J. Am. Chem. Soc. 1998, 120, 6062. (11) Riter, R. E.; Undiks, E. P.; Kimmel, J. R.; Pant, D. D.; Levinger, N. E. J. Phys. Chem. B 1998, 102, 7931. (12) Shirota, H.; Horie, K. J. Phys. Chem. B 1999, 103, 1437. (13) Hunt, N. T.; Jaye, A. A.; Hellman, A.; Meech, S. R. J. Phys. Chem. B 2004, 108, 100-108. (14) Hunt, N. T.; Jaye, A. A.; Meech, S. R. Chem. Phys. Lett. 2003, 371, 301-310.

of the aqueous phase of water-in-oil microemulsions, either directly, using techniques such as terahertz and infrared spectroscopy, or indirectly through observation of the dynamics of a fluorescence probe.6-12 In contrast, our recent work4,13,14 exploits the optically heterodyne detected optical Kerr effect (OHD-OKE) technique as a means of directly recovering the dynamics of the dispersed oleic phase of oil-in-water microemulsions. Here we apply the OHD-OKE method to make the first systematic investigation of the dynamics of the dispersed oil phase as a function of the properties of the dispersed molecules. The OHD-OKE technique is among the most experimentally straightforward routes to observing the ultrafast dynamics of liquids. The technique produces high quality high time resolution data that are readily analyzed through established Fourier transform deconvolution procedures.15 The method and its application to the dynamics of pure and mixed liquids have been reviewed extensively.15-17 More recently, complex fluids such as liquid crystals18 and polymer solutions19 have been studied. In the work most relevant to the present study, we reported the ultrafast dynamics of the dispersed oil phase of two oil-in-water microemulsions, CS2-dodecyltrimethylammonium bromide (DTAB)-water and styrene (STY)DTAB-water.4,13 In both cases, careful selection of the dispersed, continuous, and surfactant components of the microemulsions ensured that the signal from the dispersed phase dominated the OHD-OKE signal. It was established that the ultrafast dynamics of CS2 in a microemulsion are intermediate between those for solvated and neat CS2. Measurement of diffusional reorientational relaxation in the oil phase also indicated a decoupling between the macroscopic and microscopic viscosity in the system.4 In (15) Lotshaw, W. T.; McMorrow, D.; Thantu, N.; Melinger, J. S.; Kitchenham, R. J. Raman. Spectrosc. 1995, 26, 571-583. (16) Kinoshita, S.; Kai, Y.; Ariyoshi, T.; Shimada, Y. Int. J. Mod. Phys. B 1996, 10, 1229-1272. (17) Smith, N. A.; Meech, S. R. Int. Rev. Phys. Chem. 2002, 21, 75. (18) Cong, H.; Novikov, V. N.; Fayer, M. D. J. Chem. Phys. 2003, 118, 2800. (19) Shirota, H.; Castner, E. W. J. J. Am. Chem. Soc. 2001, 123, 12877.

10.1021/la047599r CCC: $30.25 © 2005 American Chemical Society Published on Web 01/20/2005

Dynamics in Oil-in-Water Microemulsions

Langmuir, Vol. 21, No. 4, 2005 1239

contrast, the ultrafast dynamics of styrene in the dispersed phase of a microemulsion exhibited more bulklike character than was observed for CS2.13 These differences in the behavior for dispersed CS2 and STY suggest that the balance of the intermolecular interactions between molecules in the oil phase and between them and their heterogeneous environment plays an important role in determining the ultrafast dynamics within a microemulsion droplet. To further investigate these phenomena, we have studied a series of monosubstituted benzenes incorporated as the oil phase within DTAB microemulsions. These were chosen as the oil phase because of the wide variation in intermolecular interactions for a similar molecular size and shape and because the dynamics of the pure liquids are rather well characterized.13,20-22 The study of this series of microemulsions leads to further understanding of the interactions that determine the dynamics and structure of the dispersed phase. Experimental Section DTAB (Lancaster, 97%), benzonitrile (BN), nitrobenzene (NB), fluorobenzene (FB), STY (all Aldrich, 99%), toluene (TOL) (Sigma, 99%), and n-dodecane (Lancaster, 99%) were all used without further purification. Purified water was obtained via a Millipore filtration unit. Microemulsion samples were characterized in terms of weight percentage of each component (as opposed to the ratio of oil to surfactant used in some studies). The oil was titrated into a 30 wt % solution of DTAB in water, except in the case of STY, where the data are taken from a recent publication13 in which a 20 wt % solution was used. It has been shown elsewhere that variation in the concentration of DTAB in this range does not affect the ultrafast dynamics recovered from the dispersed phase.4 In terms of their phase boundary, all microemulsions were found to behave similarly to the styrene-DTAB-water systems measured previously, which allowed up to 10 wt % fraction of the oil phase.4,13 To avoid any complications arising from the system being near to the Windsor IV/Windsor I phase boundary, samples for OHD-OKE measurements were prepared with dispersed phase components 1 wt % lower than the maximum uptake. All the benzenoids dissolved readily in dodecane for analysis of the solution phase dynamics. Dodecane was chosen as a suitable solvent because it is an analogue of the dodecyl chains of the DTAB molecule. Dilution to 25 wt % was a compromise between retaining strong signal and ensuring that the aromatic/dodecane molar ratio in the “diluted” solution was always less than the aromatic/DTAB ratio. Each sample was filtered (to remove dust particles) through a 0.22 µm syringe filter into a 1 mm path length quartz cuvette prior to measurement and allowed to equilibrate at the laboratory temperature of 295 ( 1 K. The OHD-OKE spectrometer used in this present study has been described in detail elsewhere17,4 and utilizes both differential detection and intensity normalization. The ultrafast light source was a Kerr lens mode-locked titanium sapphire laser (Clark MXR), pumped with a cw neodymium vanadate laser operating at 4 W. This combination produced near transform limited pulses of 45 fs duration (as measured by second-order autocorrelation at the sample position) at a repetition rate of 100 MHz. The operating wavelength was 800 nm with a bandwidth of 23 nm. Time domain traces were recorded with 6.6 fs/point resolution, up to pump-probe delay times of 1.5 ps, then with an intermediate resolution of 33 fs/point, and the remaining long time data at 100 fs/point resolution, out to pump-probe delays of 60 ps. The data presented are the average of nine scans.

Results and Discussion Figure 1 shows the short time dynamics (up to 1 ps) of the benzenoid liquids when neat, in the microemulsion, (20) Smith, N. A.; Meech, S. R. J. Phys. Chem. A 2000, 104, 42234235. (21) McMorrow, D. Opt. Commun. 1991, 86, 236-244. (22) Chang, Y. J.; Castner, E. W. J. J. Phys. Chem. 1996, 100, 3330.

Figure 1. Time domain OHD-OKE data up to 1 ps for five monosubstituted benzenes, each in three different molecular environments: neat liquid (black line), 25 wt % in dodecane (dashed line), and DTAB microemulsion (dotted line).

and solvated in dodecane. All time domain traces show an intense peak at time zero. This is assigned to the response arising from the instantaneous molecular electronic hyperpolarizability. At positive time delays, a distinct shoulder between 100 and 300 fs is seen, arising from nuclear dynamics. The intensity of the shoulder relative to the electronic response varies proportionately with the number density of the monosubstituted benzenes. The empty DTAB micelles show only the instantaneous electronic response and a very weak ultrafast relaxation.4 A shift of the shoulder due to the nuclear response to longer time delays is discernible upon dissolution of the monosubstituted benzenes in dodecane. Also evident in Figure 1 are underdamped oscillations due to impulsive excitation of intramolecular vibrational modes (see below) which persist beyond 1 ps. As expected, these are more prominent at higher concentrations of the oil phase but can still be seen in microemulsions containing as little as 4 wt % of the oil. Beyond the 1 ps time window shown, the data relax as a sum of two exponentials. This slow relaxation has previously been assigned to diffusive orientational relaxation. Further analysis of the ultrafast response is best carried out in the frequency domain. The frequency domain representation of the nuclear dynamics is obtained from the imaginary part of the ratio of the Fourier transforms of the OHD-OKE data and the laser pulse autocorrelation function. This procedure removes the effect of the finite laser pulse width and distortions due to the instantaneous electronic response.15-17 It produces the spectral densities in the frequency domain shown in Figure 2. These spectral densities are analogous to frequency domain low-frequency Raman or dynamic light scattering measurements, but free from complications due to thermal population effects, which tend to obscure the data at the lowest frequency.15-17 At the lowest frequencies, there is an approximately Lorentzian peak that is assigned to the diffusive reorientational component of the molecular relaxation (corresponding to the slow exponential relaxation in the time domain). To permit a focus on the ultrafast response, this component is subtracted from the data in the time domain,

1240

Langmuir, Vol. 21, No. 4, 2005

Figure 2. Spectral densities of neat liquid (solid line), dodecane solvated (dashed line), and microemulsion (dotted line).

following procedures described in detail elsewhere.15 The result of the Fourier transform procedure is then the nondiffusive (or ultrafast) spectral density, which is shown in Figure 3. The subtraction procedure is valid so long as the picosecond diffusional response is slow enough to allow it to be separated from the ultrafast response. The diffusive orientational relaxation times of the liquids studied here are >10 ps (see below), while the ultrafast dynamics are subpicosecond. The broad feature in the ultrafast spectral densities, between 0 and 150 cm-1 (Figure 3), has no counterpart in gas-phase spectra and is assigned to molecular librational motion and interaction-induced (dipole-induced dipole and collisional) effects. Similar assignments have been made in OKE and light scattering studies of these and a number of other aromatic liquids.21-28 The nondiffusive spectral densities of the neat liquids are remarkably similar: all the peaks have a mean frequency, ν˜ av, in the range of 56 ( 3 cm-1 and a full width at half-maximum (fwhm) of 88 ( 4 cm-1 (Table 1, Figure 3). All spectra display a shoulder or broadening on the lowfrequency side (ca. 20 cm-1). In addition, the spectra show (23) McMorrow, D.; Lotshaw, W. T. Chem. Phys. Lett. 1993, 201, 369-376. (24) McMorrow, D.; Lotshaw, W. T.; Kenney-Wallace, G. A. IEEE J. Quantum Electron. 1988, 24, 443. (25) Lotshaw, W. T.; McMorrow, D.; Kalpouzos, C.; Kenney-Wallace, G. A. Chem. Phys. Lett. 1987, 136, 323. (26) Vohringer, P.; Scherer, N. F. J. Phys. Chem. 1995, 99, 2684. (27) Kinoshita, S.; Kai, Y.; Yamaguchi, M.; Yagi, T. Chem. Phys. Lett. 1995, 236, 259. (28) Cong, P.; Deuel, H. P.; Simon, J. D. Chem. Phys. Lett. 1995, 240, 72.

Jaye et al.

Figure 3. Ultrafast spectral densities: neat liquid (solid line), dodecane solvated (dashed line), and microemulsion (dotted line). Table 1. Characteristic Frequencies of the Ultrafast Spectral Density in Figure 3: the First Moment, ν˜ av ) ∫0∞ν˜ D(ν˜ )/∫0∞D(ν˜ ) dν˜ , the fwhm of the Low-Frequency Feature, the Intramolecular Peak Frequencies, and Comparable Values From Literature ν˜ av/ fwhm/ intramolecular derivative environment cm-1 cm-1 frequency/cm-1 BN NB FB STY TOL

e

neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 12% ME neat 25% in C12 8% ME

55.6 47.5 53.9 56.2 48.9 42.2 54.1 48.6 54.7 58.6 50.8 55.9 56.6 50.3 56.8

90.7 75.0 83.4 88.8 67.9 74.8 82.6 70.0 83.2 88.8 73.2 86.5 88.4 76.0 90.2

173 170 170 179 175 177 241 239 241 238 & 212 240 & 215 240 & 212 216 219 217

a Reference 36. b Reference 20. c Reference 37. Reference 22.

d

literature/ cm-1 170a,b 180b 240c 238 & 212d 220e

Reference 13.

the intramolecular bending or “wagging” modes of the ring substituent between 150 and 300 cm-1 (corresponding to the underdamped response in the time domain, Figure 1). The agreement between the intramolecular frequencies recovered from OKE data and those in the literature is good (Table 1). When the aromatics are dissolved in dodecane, the ultrafast spectral densities shift to the red, and the peaks narrow (Table 1, Figure 3). Similar results

Dynamics in Oil-in-Water Microemulsions

Langmuir, Vol. 21, No. 4, 2005 1241 Table 2. Results of the Frequency Domain Fitting Procedure of the Ultrafast Spectral Densities to Equations 1 and 2 ωG/ ωBL/ ∆ω/ derivative environment cm-1 cm-1 cm-1 a/(a + b) BN NB FB

Figure 4. Fit of eq 1 + eq 2 to the ultrafast spectral density of BN. Individual contributions are also shown.

were observed for BN, NB,20 CS2,4 and STY13 in dodecane solution. In contrast to the rather uniform behavior in nonpolar solvent, the microemulsion environment yields ultrafast spectral densities which are dependent on the particular benzenoid. One feature common to all spectra in the microemulsion is that the low-frequency edge of the spectral density approaches that of the dodecane-solvated peak of each benzenoid. For BN and NB in the microemulsion environment, the higher frequency edge of the spectral density also shows a shift to lower frequency, such that the spectrum narrows and the spectral densities are intermediate between the neat and solvated phase spectra (Figure 3). In contrast, FB, STY, and TOL have a high-frequency edge that reflects more closely that of the neat liquid spectrum (Figure 3), and hence a somewhat broader ultrafast spectral density (Table 1). To characterize the frequency domain behavior further, the ultrafast spectral densities were fit to an empirical line shape model, allowing a more quantitative representation of the results. Recent analyses of neat and dilute liquids22,29-32 show that the frequency domain response can be well represented by the sum of an antisymmetrized Gaussian (eq 1) and an expression related to the BucaroLitovitz, or generalized Ohmic, line shape (eq 2):

[ {

IG(ω) ) a exp

} {

-(ω - ωG)2 2σ2

- exp

( )

IBL(ω) ) bωR exp

}]

-(ω + ωG)2

-ω ωBL

2σ2

(1) (2)

in which σ ) ∆ω/(2 ln 2)1/2. The fitting parameters are the Gaussian frequency, ωG, and half-width, ∆ω; the exponent R; the Bucaro-Litovitz frequency, ωBL; and the amplitudes, a and b. An example of the quality of fit is shown in Figure 4 for BN along with the individual contributions arising from eqs 1 and 2. The results of applying this fitting procedure to all data sets are shown in Table 2. The quality of the fit is very good. Indeed, the principle advantage of fitting these characteristically broad asymmetric flat top line shapes to eqs 1 and 2 is that a good description of the line shape is obtained with a minimum of variable parameters. In the present case, the Gaussian component dominates the ultrafast spectral density (Table 2). This Gaussian component has previously been assigned (29) Chang, Y. J.; Castner, E. W. J. J. Chem. Phys. 1993, 99, 7289. (30) Smith, N. A.; Lin, L.; Meech, S. R. J. Phys. Chem. 1997, 101, 9578-9586. (31) Neeklakandan, M.; Pant, D. D.; Quitevis, E. L. J. Phys. Chem. A 1997, 101, 2936. (32) Smith, N. A.; Lin, L.; Meech, S. R.; Yoshihara, K. J. Phys. Chem. A 1997, 101, 3641.

STY TOL

neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 12% ME neat 25% in C12 8% ME

61.3 51.9 49.4 43.6 35.8 34.6 60.4 46.7 56.4 42.3 30.5 47.7 54.6 44.5 54.9

12.7 13.8 9.8 9.3 19.8 10.1 14.0 14.3 14.8 8.5 10.5 8.8 10.4 11.5 12.5

37.3 34.6 43.0 49.7 45.2 47.7 37.9 41.4 42.6 52.9 50.0 47.8 43.9 42.3 44.6

0.92 0.92 0.95 0.95 0.92 0.95 0.93 0.94 0.97 0.98 0.97 0.96 0.95 0.95 0.95

R 1.176 1.161 1.377 1.297 1.117 1.330 1.259 1.373 1.563 1.630 1.499 1.589 1.349 1.318 1.238

to molecular librational motion, on the basis of its dependence on concentration in nonpolar solvent and molecular moment-of-inertia.20-22 These librational dynamics arise from the relaxation of coherently excited harmonic orientational motion of molecules in a potential well defined by their nearest neighbors. It is the reduced curvature of the potential that arises from replacing the nearest neighbors with weakly interacting aliphatic molecules which results in a red shift of the spectral density in solution (Table 2). Recent molecular dynamics simulations and instantaneous normal-mode analysis of the OKE spectra of liquid benzene have confirmed the importance of librational dynamics. However they also show that the underlying librational response has a much more complex line shape than can properly be represented by eq 1.33 The assignment of the minor low-frequency component fit by eq 2 is more ambiguous. The original BL function was derived for collision-induced line shapes in atomic liquids. Its suitability for molecular liquids is much less apparent. Possible assignments of the low-frequency response that have been discussed include intermolecular (interaction-induced) dynamics, a second librational mode, modes associated with an intermolecular complex and motional narrowing. Again, recent simulations of liquid benzene dynamics provide useful insights and suggest that both interaction-induced and librational orientational dynamics contribute in this frequency range.33 The simulations also suggest that the empirical analysis using eqs 1 and 2 underestimates the contribution of interactioninduced components at high frequency, and that of librational components at low frequency. Bearing in mind these limitations, the fitting parameters (Table 2) nevertheless reveal a number of trends in the ultrafast spectral densities. The ωG shows the greatest dependence on the environment. In contrast, ωBL remains almost constant at 12 ( 2 cm-1 with R in the range of 1.1-1.6 in all the samples studied. The ωG values show an average shift of 10 ( 2 cm-1 to lower frequencies for all benzene derivatives dissolved in dodecane. This is consistent with librational motion in the potential well defined by the neighboring molecules, as discussed above. In the microemulsion, there is a greater variety of values of ωG. For dispersed BN and NB, the values are very similar to the dodecane-solvated ωG. In contrast, the ωG values for microemulsions of FB, STY, and TOL more closely reflect those of the corresponding neat liquids. Next we consider the data recorded on the 1-60 ps time scale, assigned on the basis of detailed studies of the neat (33) Ryu, S.; Stratt, R. M. J. Phys. Chem. B 2004, 108, 6782-6795.

1242

Langmuir, Vol. 21, No. 4, 2005

Jaye et al.

Table 3. Biexponential Parameters Resulting from a Fit over the Range of 1.5-40 ps and Some Molecular and Physical Properties of the Neat Liquid derivative

molecular environment

% a1

τ1a/ps

τ2a/ps

dipole momentb/D

polarizabilityc/Å3

∆Hvapb/kJ mol-1

ηb/cP

BN

neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 10% ME neat 25% in C12 12% ME neat 25% in C12 8% ME

0.56 0.75 0.70 0.58 0.73 0.64 0.56 0.85 0.73 0.51 0.71 0.57 0.56 0.82 0.63

1.4 2.0 2.0 1.6 2.1 1.8 1.3 2.0 1.3 1.3 1.9 2.5 1.1 2.0 1.5

18.0 21.8 34.9 28.5 25.2 20.2 4.6 17.3 8.2 9.7 16.1 10.7 5.2 17.7 11.0

4.2

12.5

47.5

1.24d

4.2

13.0

50.9

2.03e

1.6

10.3

33.4

0.60e

0.1

14.4

40.3

0.70d

0.4

12.3

39.2

0.60e

NB FB STY TOL

a

Fitting error of 10%. b Reference 38. c Polarizability calculated using the Clausius-Mossotti equation (ref 39).

d

At 25 °C. e At 20 °C.

liquids20-22 to the diffusive orientational relaxation. The data are shown in Figure 5, normalized at 1.5 ps. The data after 1.5 ps (i.e., ignoring the ultrafast librational and interaction-induced components) were fit to the following function:

rdiffusive(t) ) [a1 exp(-t/τ1) + a2 exp(-t/τ2)][1 exp(-2ωavt)] (3) a rapid rise followed by a biexponential decay, in which ωav is the mean frequency of the spectral density which is kept constant at 15 ps-1, and τ1 and τ2 are relaxation times. The results of fitting the data to eq 3 are shown in Table 3. The quality of the fits to the data, as judged by residuals and χ2 values, was found to improve for some solutions when a third exponential relaxation term was included. However, in this case the uncertainties in the recovered time constants were large. Most decays were described as well or better by two exponentials, so the triexponential results are not presented. However, the picosecond dynamics in the microemulsion (and in dodecane solution) are probably multiexponential or dispersive in character, and the biexponential fit is only an approximation. This is in contrast to the neat liquids which are accurately described by a dual exponential decay function. This behavior is evident in Figure 5. The time constants retrieved from the fitting have been discussed in detail for the neat liquids elsewhere; τ1 is too fast to be attributed to diffusive molecular reorientation. Its value scales with that of τ2, and it has been assigned (in simpler liquids) to motional narrowing.34 In the present study, τ1 falls in the range of 1-2 ps. The slower relaxation time, τ2, for neat NB and BN has been assigned to diffusive rotational motion due to the good agreement with the predictions of the Stokes-Einstein-Debye equation.20 Comparison of τ2 with viscosity (Table 3) for all the neat liquids studied here suggests that this assignment is also appropriate for STY, TOL, and FB. The reorientational decay profiles (Figure 5) along with the biexponential fitting results (Table 3) show that the monosubstituted benzenes in dodecane solution exhibit longer orientational relaxation times than the neat liquids, with the exception of NB. For solutions other than BN, this is in the direction expected from the Stokes-EinsteinDebye equation, using η ) 1 cP for dodecane. The unexpected increase in relaxation time for BN and the appearance of a multiexponential form even in the slow (34) Loughlane, B. J.; Scoudinu, A.; Farrer, R. A.; Fourkas, J. T.; Mohanty, U. J. J. Chem. Phys. 1999, 111, 2686.

Figure 5. Long time OHD-OKE data (1.5-25 ps) for the monosubstituted benzene series. Neat liquid (black), dodecane solutions (red), microemulsion (blue).

component (Figure 5) indicate that other factors may also be important, including clustering of the solute molecules in the solvent, as discussed previously.4,13,35 Incorporation within a microemulsion produces orientational decay times that contrast differently with the neat and solvated times depending upon the particular monosubstituted benzene. For FB, STY, and TOL, the reorientation time is closer to that of the neat liquid than the dodecane solution. For NB and BN, no clear trend between the three environments is discernible. In all cases, we note that the diffusive reorientation of the oil phase is certainly decoupled from the macroscopic viscosity of the microemulsion, which is large at these large weight fractions of DTAB.4 These picosecond and sub-picosecond dynamics can be discussed in terms of the distribution of the “oil” phase within the micelle. The low-frequency spectral densities of the neat liquids (Figure 3) are in good agreement with those observed previously.13,20-22,36 For all benzenoids (35) McMorrow, D.; Thantu, N.; Melinger, J. S.; Kim, S. K.; Lotshaw, W. T. J. Phys. Chem. 1996, 100, 10389-10399. (36) Wiewior, P.; Radzewicz, C. Opt. Appl. 2000, 30, 103-120.

Dynamics in Oil-in-Water Microemulsions

dissolved in dodecane, the librational peak narrows and shifts to lower frequencies (Figure 3 and Table 2). This phenomenon has already been noted in studies of polarizable molecules dissolved in alkane solvents4,13,35 and is consistent with a dominant contribution to the higher frequency part of the spectra from molecular librational dynamics. Upon incorporation within a microemulsion, the spectra for BN and NB reveal spectral densities that are intermediate between those of neat liquid and solvated samples (Figure 3). This is also evident from the ωG and ∆ω values recovered for BN and NB in each environment (Table 2). In contrast, the ultrafast spectral densities of FB, STY, and TOL constrained within a microemulsion are more similar to those of the neat liquids (Figure 3, Table 2). The low-frequency edge of the librational peak mimics the dodecane-solvated spectrum (as is also the case for BN and NB), but the high-frequency edge closely matches that of the neat sample. Thus, all the aromatics studied show evidence of a degree of spectral broadening when incorporated into a microemulsion environment. However, the three weakly polar liquids (FB, STY, and TOL) reveal a more bulklike response in the microemulsion than do the two polar liquids NB and BN. In some respects, this is a surprising result, as the tendency for “like to dissolve like” would lead one to expect that in the nonpolar core of a DTAB micelle the polar molecules would tend to associate, while the less polar ones disperse. However, a micelle does not provide a continuous nonpolar medium but is in fact a highly heterogeneous environment. We propose that the nonpolar and slightly polar FB, STY, and TOL associate in the nonpolar core of the micelle, while the highly polar BN and NB are pushed out of the core and are stabilized by interactions with the polar outer layer of the micelle. This rationalization may be testable by neutron scattering experiments. (37) Nielsen, O. F.; Fajolles, C.; Lund, P. A.; Praestgaard, E. J. Mol. Liq. 1989, 43, 13. (38) CRC Handbook of Chemistry and Physics, 67th ed.; Weast, R. C., Ed.; CRC Press: Boca Raton, FL, 1986. (39) Atkins, P.; De Paula, J. In Physical Chemistry, 7th ed.; Oxford University Press: New York, 2002; p 693.

Langmuir, Vol. 21, No. 4, 2005 1243

The picosecond orientational dynamics for the neat liquids (Figure 5 and Table 3) are in good agreement with those observed previously (where available).13,20,22,36 The slower reorientational component (τ2) shows a strong correlation with viscosity, indicative of Stokes-EinsteinDebye type diffusive reorientation. On dilution with dodecane, it appears that the dynamics are still largely diffusive, although the nonexponential form of the slow relaxation suggests that other factors may be important. When the monosubstituted benzenes are incorporated within a microemulsion, there is a tendency for the diffusive orientational relaxation kinetics to become dispersive or multiple exponential in character, possibly an indication of an inhomogeneous distribution of environments, consistent with the spectral broadening noted above. For FB, STY, and TOL, the mean orientational relaxation time is faster than in dodecane, suggesting that the interior of the micelle is moderately fluid. The relaxation time tends toward but is never as fast as that in the neat liquids. Conclusion We have reported the first systematic investigation of the ultrafast dynamics of the oil component of oil-in-water microemulsions. The dynamics in the dispersed phase were discussed in terms of those measured in the neat oil and its solutions in a nonpolar solvent. The three oils which had weaker intermolecular interactions revealed dynamics in the dispersed phase which were similar to those of the neat liquid. In contrast, oil phases comprising more strongly interacting polar liquids revealed dynamics more similar to those of the solvated state. We suggest that this arises because of the stabilization of the polar molecules close to the interfacial region of the micelle. This causes the polar liquid to be dispersed throughout the micelle. In contrast, the less polar but polarizable molecules cluster in the nonpolar micelle core. In all cases, the data suggest an inhomogeneous distribution of sites for the dispersed liquids. Acknowledgment. We are grateful to the EPSRC for financial support of this work. A.A.J. is grateful to EPSRC for a studentship. LA047599R