Heterogeneous Solute Dynamics in Room Temperature Ionic Liquids

Nov 9, 2007 - ... Tuhin Pradhan , Didier Touraud , Werner Kunz and Sekh Mahiuddin ..... Deboleena Sarkar , Bhaswati Bhattacharya , Nitin Chattopadhyay...
0 downloads 0 Views 186KB Size
13473

2007, 111, 13473-13478 Published on Web 11/09/2007

Heterogeneous Solute Dynamics in Room Temperature Ionic Liquids Hui Jin, Xiang Li, and Mark Maroncelli* Department of Chemistry, The PennsylVania State UniVersity, 104 Chemistry Building, UniVersity Park PennsylVania 16802 ReceiVed: September 8, 2007; In Final Form: October 10, 2007

The excitation wavelength dependence of the emission kinetics of several solutes is used to demonstrate the presence of dynamic heterogeneity in two representative room temperature ionic liquids, dimethyl-isopropylpropyl-ammonium bis(trifluoromethylsulfonyl)imide [Nip311+][Tf2N-] and N-propyl-N-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide [Pr31+][Tf2N-]. The solute kinetics examined here include rotation and solvation of coumarin 153, isomerization of two malononitriles, and intramolecular charge transfer in crystal violet lactone. The rates of most of these processes vary significantly with excitation wavelength, especially for excitation on the red edges of the solute absorption bands, indicating that energetically selected subpopulations relax at distinct rates. The results presented here suggest more generally that dynamical processes taking place on the subnanosecond time scale in typical ionic liquids near room temperature are likely to be heterogeneous in character.

A number of recent studies of both solvent1,2 and solute dynamics3-6 in room temperature ionic liquids (ILs) have pointed to the dispersive nature of kinetics in these systems. For example, we have observed that the time-dependent response to a solute electronic perturbation in several imidazolium ionic liquids extends over 3 or more decades in time, exhibiting a time evolution that can be reasonably represented by a stretched exponential function with a stretching parameter β ∼ 0.4.3 Such kinetics resemble those typically found in fragile liquids in the supercooled regime, despite the fact that most ionic liquids near room temperature are above (albeit close to7) their melting points and roughly 100 K away from their glass transition temperatures.8 In the case of conventional supercooled liquids, it is generally agreed that dispersive kinetics are the result of “dynamic heterogeneity”: the fact that molecules in different local environments relax at significantly different rates.9,10 It is reasonable to suppose that dynamic heterogeneity also underlies the dispersive kinetics observed in room temperature ionic liquids. An obvious prerequisite for such dynamics is variability of the instantaneous local structures and energetics encountered by a molecule, which we term “static heterogeneity”. The presence of static heterogeneity in ionic liquids has been clearly demonstrated both experimentally11-14 and in computational studies.15-18 However, whereas computer simulations have shown how such static heterogeneity translates into dynamic heterogeneity of solvent17 and solute19 motions, experimental evidence for dynamic heterogeneity in room temperature ionic liquids is lacking. The purpose of the present Letter is to provide such evidence based on the kinetics of several probe solutes. We do so by observing differences in the time-dependent emission of a solute produced by the site (energy) selection afforded by red-edge excitation.20,21 Because of the asymmetric nature of vibronic structure, excitation near the peak or on the * Corresponding author. E-mail: [email protected].

10.1021/jp077226+ CCC: $37.00

SCHEME 1

blue side of an electronic absorption tends to excite a broad distribution of energetically different solute-solvent environments. In contrast, excitation on the red edge of an absorption band preferentially selects a narrowed and red-shifted subset of the overall distribution. For polar solvatochromic probes of the sort examined here, red-edge excitation is capable of selecting subpopulations that are energetically stabilized by ∼2000 cm-1 (10 kBT at room temperature) as compared to the average solvation state.22 Samanta and co-workers13 have already demonstrated that site selection is possible in room temperature ionic liquids by monitoring the shifts in steady-state emission of short-lived solutes caused by red-edge excitation.23 Here, we extend this approach to explore the extent to which red-edge excitation leads to kinetics that are different from those observed upon nonselective excitation. We will focus mainly on dynamics in the ionic liquid dimethyl- isopropyl-propyl-ammonium bis(trifluoromethylsulfonyl)imide [Nip311+][Tf2N-] (Scheme 1) at 25 °C. [Nip311+][Tf2N-] was chosen for the present study because it exhibits the least impurity emission out of all of the ILs in our collection, an important consideration for edge excitation experiments. Beyond its optical purity, this IL is in no way distinctive from the other short chain, nonfunctionalized ionic liquids we have studied.4 The features observed with this particular IL should therefore be representative of most of the ILs in common use today. For comparison purposes, some physical properties of [Nip311+][Tf2N-] (25 °C) are the following: glass and fusion temperatures, Tg ) 189 K, Tfus ) 290 K; density, d ) 1.40 g cm-3; refractive index, nD ) 1.42; viscosity η ) 113 cP.24 In one example, we instead use the closely related ionic liquid N-propyl-N-methylpyrroli© 2007 American Chemical Society

13474 J. Phys. Chem. B, Vol. 111, No. 48, 2007

Figure 1. Excitation frequency dependence of the emission frequencies, rotation times, and solvation times of C153 in [Nip311+][Tf2N-] at 25 °C. The smooth curve in each panel shows the normalized absorption spectrum. Points display various quantities relative to their values observed when excitation is at the absorption maximum. Emission frequencies measured as the first moment of the emission band are represented by 2∆ν/Γ0 ) 2(ν - ν0)/Γ0, where ν0 and Γ0 are the frequency and full width at half-maximum of the emission excited at the absorption maximum (Γ0 ) 3440 cm-1). Rotation and solvation times are represented by ∆τ/τ0 ) (τ - τ0)/τ0 where τ0 is the time observed when exciting at the absorption maximum. Rotation times are integral times derived from stretched exponential fits of the anisotropy data (τ0 ) 6.0 ns). Solvation times shown as the open triangles are integral times of the observed ν(t) data (τ0 ) 0.52 ns), and those shown as filled circles are the 1/e times of the spectral response functions (τ0 ) 0.12 ns) after approximately accounting for the (∼50%) fast component not observed in these experiments (see ref 4). For clarity, the two sets of points in the bottom panel are horizontally offset by a small amount.

dinium bis(trifluoromethylsulfonyl)imide [Pr31+][Tf2N-] (Tfus ) 283 K; d ) 1.40 g cm-3; nD ) 1.42; η ) 54 cP24). The experiments reported here employ either steady-state absorption and emission measurements or time-correlated single-photon counting (TCSPC) using the instruments and methods detailed in ref 4. We note here only that the TCSPC instrument used has a response time of 25 ps fwhm, which means we poorly sample or entirely miss dynamics taking place at times earlier than ∼10 ps. The first solute to be considered is the well-known solvatochromic probe coumarin 15325 (C153; Figure 1). The top panel of Figure 1 shows the excitation frequency dependence of the steady-state emission (filled symbols; “SS”) of C153. Here and in later figures, we represent various observables by normalized differences from the values observed when excited at the absorption maximum. Actual values are tabulated in the Supporting Information. In the case of frequencies, we plot 2(ν - ν0)/Γ0, where ν is the first moment of the emission band, and ν0 and Γ0 are the frequency and full width of the emission band excited at the absorption maximum. As previously noted by Mandal et al.,13 the steady-state emission of C153 is relatively insensitive to excitation frequency. Over the range of excitation wavelengths employed, the shift of the steady-state spectrum is a mere 4% (140 cm-1) of its width, Γ0 3440 cm-1. This insensitivity is expected,13,20,21 given that the emission lifetime of C153 is 6 ns, whereas the majority of the spectral dynamics of C153 in [Nip311+][Tf2N-] take place at subnanosecond times.4 Also shown in the top panel of Figure 1 (labeled “t ) 0”) are the relative frequencies expected in the absence of solvent

Letters relaxation, that is, when the solvent is immobile on the emission time scale, which is identical to the spectrum expected immediately after excitation. These “time-zero” frequencies are calculated using estimates of the polar solvent contribution to the inhomogeneous broadening of the absorption spectrum as described in refs 22 and 26. Such calculations for C153 accurately reproduce the red-edge excitation shifts observed at low temperatures22 and the earliest spectra in dynamic Stokes shift experiments when performed with high time resolution.25 In contrast to the small steady-state shifts, the time-zero frequencies estimated in Figure 1 shift by 30% of the spectral width (1000 cm-1). This sizable shift demonstrates that significant site (energy) selection is provided by the range of excitation frequencies accessible in these experiments. We note that if C153 had a much shorter lifetime, as do several of the other solutes examined later, we would observe much larger shifts of the steady-state spectra than we do. But observation of an excitation-dependent shift of the emission spectrum only indicates the presence of a static heterogeneity, which is a necessary but not a sufficient condition for the dynamic heterogeneity of interest here. The first dynamical property we examine is the rotation of C153. In previous work, it was noted that the rotational correlation functions of C153 in most ionic liquids such as [Nip311+][Tf2N-] are nonexponential.4,6,27 Although there is some disagreement on the best description of such data, we find that stretched exponential functions of time with stretching exponents β in the range 0.6-0.8 provide reasonable fits.4 One might suppose that this nonexponentiality reflects heterogeneity of the rotation dynamics, and if so, one might expect to observe an excitation wavelength dependence to the rotation times. In the middle panel of Figure 1, we plot the rotational correlation times of C153 in [Nip311+][Tf2N-] measured from fits to emission anisotropy data collected at eight excitation frequencies. At these frequencies, we find anisotropy decays that are the same to within uncertainties, with 〈τ〉 ) 6.0 ( 0.1 ns and β ) 0.72 ( 0.01 ((1σ). This lack of excitation dependence could be interpreted to mean that rotations are homogeneous, that is, that all solutes undergo the same nonexponential rotational motion, independent of their particular solvent environment. Given that much of the solvation response in [Nip311+][Tf2N-] (discussed below) is 10-fold faster than these rotation times, such rotational homogeneity might be expected.28 As discussed in ref 4, this is our current interpretation, that rotation of C153 is not significantly heterogeneous in most ILs. We suggest that the nonexponential rotational correlation functions observed instead reflect the non-Markovian nature of friction in these slowly relaxing solvents.29 However, another explanation should also be considered. The primary friction on rotation of C153 results from short-range repulsive interactions with surrounding solvent molecules,29 whereas the electronic spectral shifts used for site selection are dominated by electrostatic interactions. Given that there might be little or no correlation between the magnitudes of these two types of interactions, the present experiments might not select molecules in a manner relevant to rotational motion. The distinction between what is selected for via edge excitation versus what controls the dynamics of interest must always be kept in mind when using such experiments to explore dynamic heterogeneity. Finally, the bottom panel of Figure 1 contains data on the excitation dependence of the solvation response of C153 as monitored by its dynamic Stokes shift. Representative timeresolved spectra, from which these results were derived, are shown in Figure 2a. In contrast to the situation with rotation,

Letters

Figure 2. (a) Time-resolved emission spectra of C153 in [Nip311+][Tf2N-] at 25 °C excited at 380 nm ) 26 300 cm-1. Points show actual data at 0 (circles) and 10 ns (triangles). The smooth solid curves are log-normal fits to the data at 0, 10, 20, 50, 100, 200, and 500 ps and 1, 2, and 10 ns (later times correspond to lower frequencies). The dashed curves show the estimated time-zero spectra.26 (b) Peak frequencies versus time derived from log-normal fits to such spectra.

what is being observed here is the relaxation of the emission frequency and, assuming the absorbing and emitting transitions are identical,30 the selection process is performed directly on the coordinate of interest. It is expected that changing the excitation frequency will change the initial position of the relaxing spectrum and, therefore, the total extent of the Stokes shift will also change.21,22,26 As illustrated in the peak frequency versus time data in Figure 2b, the expected variations are, indeed, observed in the experimental spectra. But, as shown by the estimated time-zero spectrum in Figure 2a, due to the limited time resolution of the experiment, we fail to observe a significant fraction of the initial relaxation. This shortcoming complicates the interpretation of the kinetics slightly. If we examine the times associated with the peak (Figure 2a) or first moment frequencies of the spectra (not shown), we find times that decrease with increasing νexc. The open symbols in the bottom panel of Figure 1 are the integral times associated with these observed ν(t) data.31 There is a systematic variation with excitation wavelength that amounts to 15-20% of the central value τ0 ) 0.52 ns. The trend is such that ν(t) decays more slowly as νexc is decreased and the magnitude of the Stokes shift decreases. Although these variations appear significant as compared to the uncertainties in the data, they are small, and one could question whether they might not be an artifact of the insufficient time resolution employed here. (For example, we have previously found that the larger the Stokes shift, the easier it is to detect fast relaxation components and, therefore, the faster the observed response in TCSPC experiments.3) To ascertain whether such an artifact might be responsible for the variations observed, we also calculated normalized response functions, S(t) ) {ν(t) - ν(∞)}/{ν(0) - ν(∞)}, using values of ν(0) obtained from estimates of the time-zero spectra26 and a common value of ν(∞) for all νexc. The 1/e times of these S(t) functions are shown as the filled symbols in the bottom panel of Figure 1. Despite their large uncertainties, these times exhibit a clear and much larger dependence on νexc. The variation is (35% about the peak value, τ0 ) 0.12 ns. Thus, although the present

J. Phys. Chem. B, Vol. 111, No. 48, 2007 13475 experiments do not resolve all of the dynamics, they provide clear evidence that the solvation response in [Nip311+][Tf2N-] is site- or solvation-energy-dependent, that is, that the solvation dynamics is heterogeneous. We note that whereas the excitation dependence of the dynamic Stokes shift has only rarely been studied in simple solvents,21 it has recently been finding increasing use in studies of organized assemblies, such as micelles.32 In simple solvents, heterogeneity in solvation dynamics has been inferred mainly from the presence of a maximum in the dynamic spectral width that can sometimes be detected in experiment10,33 and simulation.34,35 Although data such as those shown in Figure 2a do not reveal any pronounced width variations, examination of a large collection of such data, recorded with C153 and with the alternative solvation probe 4-dimethylamino-4′-cyanostilbene, suggests that the nonmonatonic variation expected for heterogeneous dynamics36 is present in ionic liquids but is difficult to observe given the spectral and time resolution typically available. It should also be noted that heterogeneity of solvation dynamics in ionic liquids is to be expected on the basis of the molecular descriptions of solvation dynamics provided by recent computer simulations.37-40 As discussed in ref 3, simulations show that the solvation energy relaxes primarily via collective small-amplitude translational motions of ions near to the solute. The nature of this motion is such that the majority of the solvation energy is relaxed before any significant reorganization of the solvation structure can occur. Assuming the solvation time to depend on initial solvation structure, the disparity between solvation structure relaxation times and solvation energy relaxation times leads to heterogeneous solvation dynamics. Another example of dynamic heterogeneity is found in the excited-state decay of the so-called “molecular rotors”, p(dimethylamino)benzylidene malononitrile (DMN) and julolidine malononitrile (JDMN) (see Figure 3). The fluorescence quantum yields of these molecules are sensitive to solvent fluidity, and for this reason, they have long been used as “microvisosity” probes in conventional solvents,41 polymers,42 and biological systems43 and recently in ionic liquids.44,45 Femtosecond experiments and electronic structure calculations show that the S1 lifetimes of these molecules are on the order of a few picoseconds in conventional solvents at room temperature due to internal conversion via a barrierless isomerization about the double bond.46 There is also a large (∼10 D47) increase in dipole moment between S0 and S1, which renders these molecules highly solvatochromic; however, no clear connection between solvent polarity and isomerization rates has yet been discerned. In ionic liquids such as [Nip311+][Tf2N-], emission decays measured with TCSPC are nonexponential, consisting of components having time constants in the 10-50 ps range, but a significant portion of the decay is too rapid to observe with such experiments. We therefore estimate S1 lifetimes and thus isomerization rates indirectly, using measured quantum yields (2-6 × 10-3) and radiative rates (∼2.8 × 108 s-1 48,49).50 The results are summarized in Figure 3. The lifetimes of DMN and JDMN upon peak excitation are 8 and 17 ps, respectively. These times are much shorter than the solvation response of [Nip311+][Tf2N-] so that, unlike C153, both solutes exhibit pronounced red-edge excitation shifts of their emission spectra. The observed shifts are more than one-half of the widths of the respective emission spectra, or roughly 700 cm-1 in DMN and 1000 cm-1 in JDMN. These shifts are a sizable fraction of what is predicted should be observed in frozen solvents. More importantly, the estimated isomerization times depend strongly

13476 J. Phys. Chem. B, Vol. 111, No. 48, 2007

Figure 3. Excitation wavelength dependence of the emission frequencies and lifetimes of DMN and JDMN in [Nip311+][Tf2N-] at 25 °C. Frequencies, ν, are the average (first moment) frequencies of the emission bands, and the lifetimes, τ, are estimated from the quantum yields as described in the text. The values used for normalizing the DMN data are Γ0 ) 2825 cm-1 and τ0 ) 8.0 ps, and for JDMN, the values are Γ0 ) 2335 cm-1 and τ0 ) 16.9 ps.

on excitation wavelength. These times are ∼50% (JDMN) and ∼70% (DMN) slower when excited on the red edge, as compared to excitation at the absorption peak. We note that comparable excitation dependence is also observed in the timeresolved decays of these two solutes measured with TCSPC, but the times are uniformly longer due to the presence of components too fast to observe with this method. These data provide unequivocal evidence for the presence of dynamic heterogeneity on the tens of picoseconds time scale in [Nip311+][Tf2N-]. Given the slow dynamics in ionic liquids, finding such rapid reactions to be heterogeneous is perhaps not surprising, but it should be noted again that even for such fast reactions, one would not observe an excitation dependence unless the environmental factors affecting the reaction rate and the absorption shift were correlated. The fact that we do observe a clear excitation effect implies that the electrical interactions selected for by the spectroscopy influence the effective potential on which the isomerization occurs, via either energetic or frictional means. We have also observed signs of dynamic heterogeneity in several ultrafast, intramolecular electron-transfer reactions in ionic liquid solvents. As a last example, we present results on the dual fluorescent solute crystal violet lactone (CVL; Figure 4). Karpiuk51 has shown that excitation into the S1 state of CVL produces a polar excited state (“LE”; µ ∼ 11 D) localized on the aminophthalide ring, which subsequently transforms into a more polar charge-transfer state (“CT”; µ ∼ 25 D) in which an electron is transferred from one of the dimethylaniline donor groups to the aminophthalide acceptor. We have recently begun investigating the LE f CT process of CVL and its connection to solvation dynamics in conventional dipolar solvents. We have found that in strongly polar and highly fluid solvents, such as acetonitrile, the emission dynamics observed are those expected for a reversible 2-state interconversion. Equilibrium constants are in the range 50-100, and reaction times are 10-30 ps, similar to what is observed in alkylaminobenzonitriles.52 In ionic liquids, the interconversion is much slower (>100 ps) and highly nonexponential. Available data indicates that reaction times

Letters

Figure 4. Top panel: Steady-state emission spectra of CVL in [Pr31+][Tf2N-] at 25 °C at excitation frequencies of 25.0, 25.6, 26.3, 27.0, 27.8, and 28.6 × 103 cm-1. Points show actual data for the two extreme values of νexc, and the smoothed curves are fits to these data. The dashed curves are the decomposition into LE and CT components for νexc ) 25 000 cm-1. These spectra have been normalized to equal areas (of fitted spectra) to best display the relative intensities of the LE and CT bands. Bottom panel: Normalized values of the peak LE emission frequency (ν, filled circles, ν0 ) 22 940 cm-1, Γ0 ) 3780 cm-1) and the ratio of the LE to the CT band areas (r, open triangles, r0 ) 0.90) determined from these spectra. The absorption spectrum is also shown for reference.

roughly track the viscosity of the ionic liquid, suggesting that solvation dynamics or large-amplitude intramolecular motions (or both) play an important role in the charge transfer. Because the time scales of solvation, reaction, and excited-state decay of CVL in ionic liquids are all comparable, analysis of the observed kinetics is complicated. For this reason, we simply use the steady-state spectra shown in Figure 4 to provide a qualitative demonstration of dynamic heterogeneity in this reaction. We show data collected in [Pr31+][Tf2N-] at 25 °C because the variations in the steady-state spectra are much more pronounced in this liquid than in [Nip311+][Tf2N-]. The same qualitative features are observed for CVL in [Nip311+][Tf2N-], but the 2-fold smaller viscosity of [Pr31+][Tf2N-] modifies the relative rates of reaction and excited-state decay so as to greatly enhance the edge excitation effects on the steady-state spectra. As shown in the bottom panel of Figure 4, there is a significant shift (∼350 cm-1) of the LE band with excitation frequency. Much more striking is the dramatic variation in the relative intensities of the LE and CT bands, measured by the ratio r ) ICT/ILE. As shown in the bottom panel of Figure 4, r increases by nearly a factor of 2.5 (∆r/r0 ) 1.5) between peak excitation and the reddest excitation employed. Given that the equilibrium lies far to the CT side of the reaction, this variation in band area reflects a large increase in the net LE f CT interconversion during the ∼1 ns lifetime of the excited state and, thus, in the reaction rate with excitation frequency. Quantitative estimates of ∆τ/τ0 of the sort made for the malononitriles will require a more detailed analysis, but it is not unreasonable to assume factors of 2 or more will be found on the basis of these steadystate results. Thus, for this charge-transfer process in which there is a small but significant barrier to reaction, the effects of environmental heterogeneity appear to be quite large. To summarize, we have used the site selection provided by red-edge excitation of solutes to look for experimental evidence of dynamic heterogeneity in the representative ionic liquids

Letters [Nip311+][Tf2N-] and [Pr31+][Tf2N-]. Such heterogeneity is anticipated for processes that are faster than the lifetime of a typical solvation environment. The solvation response, as measured by the dynamic Stokes shifts of a solute such as C153, provides one estimate of this environment lifetime.28 In [Nip311+][Tf2N-] at 25 °C, solvation occurs over a broad time window: roughly 50% of the solvation response is complete in 10 ps, 75% in 300 ps, 90% in 1 ns, and >98% in 10 ns. Given this characterization, it is not surprising that we do not observe any excitation dependence to the rotation of C153, which is a relatively slow process, ) 6 ns. But the rates of all of the other processes investigated here, such as the solvation of C153 itself, the barrierless isomerization of the fluidity probes DMN and JDMN (7-25 ps), and the low-barrier LE f CT intramolecular charge transfer in CVL (100-500 ps) do all vary significantly as functions of excitation wavelength. Although these examples provide the first clear experimental evidence for dynamic heterogeneity in room temperature ionic liquids, we do not consider such behavior to be unusual. Given the sluggish and broadly distributed solvation response of these solvents, we anticipate that heterogeneous kinetics will be the norm rather than the exception for processes that occur in the subnanosecond time regime. It will be important to recognize this heterogeneity when attempting to model fast kinetics in ionic liquids. Finally, some comment on the nature of the static heterogeneity underlying the heterogeneous dynamics observed here may be useful. Simulations show that dynamic heterogeneity can be produced by subtle differences in local packing, even in model liquids consisting of spherical molecules.53 In the case of ILs, the ions comprising the liquid of are themselves “heterogeneous” in that they are often conformationally flexible and amphiphilic in character, which can lead to other sorts of structural heterogeneity and even to local polar/nonpolar domain segregation.11,15,16 Although these latter sources of static heterogeneity cannot be excluded in the present systems, we note that solvatochromic measurements of the solutes examined here, as well as others,54,55 indicate that the extent of the energetic heterogeneity in these liquids is not significantly different from that found in simple dipolar solvents having comparable solvation energies. On this basis, we conjecture that dynamic heterogeneity of the sort measured here is readily observable in room temperature ionic liquids primarily because of their much slower relaxation as compared to conventional solvents,4 rather than to any special structural features they might also possess. Further experiments and computer simulations of solute spectra and dynamics18,39 will be needed to define the origins of dynamic heterogeneity in ionic liquids and explore its effect on chemical processes carried out in them. Acknowledgment. The authors thank Gary Baker for preparing the ionic liquids used here and Sergei Arzhantsev for help with many aspects of these experiments. This research was sponsored by the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Science, U.S. Department of Energy. Supporting Information Available: Six tables containing the unscaled data presented in this paper. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Weinga¨rtner, H.; Sasisanker, P.; Daguenet, C.; Dyson, P. J.; Krossing, I.; Slattery, J. M.; Schubert, T. J. Phys. Chem. B 2007, 111, 4775.

J. Phys. Chem. B, Vol. 111, No. 48, 2007 13477 (2) Li, J.; Wang, I.; Fruchey, K.; Fayer, M. D. J. Phys. Chem. A 2006, 110, 10384. (3) Arzhantsev, S.; Jin, H.; Baker, G. A.; Maroncelli, M. J. Phys. Chem. B 2007, 111, 4978. (4) Jin, H.; Baker, G. A.; Arzhantsev, S.; Dong, J.; Maroncelli, M. J. Phys. Chem. B 2007, 111, 7291. (5) Ito, N.; Richert, R. J. Phys. Chem. B 2007, 111, 5016. (6) Funston, A. M.; Fadeeva, T. A.; Wishart, J. F.; Castner, E. W. J. Phys. Chem. B 2007, 111, 4963. (7) We note that Hamaguchi and co-workers,59 have suggested the existence of solidlike regions within ionic liquids, and if this suggestion is correct, proximity to the melting point may have some relevance to the liquid dynamics. (8) Wasserscheid, P., Welton, T., Eds. Ionic Liquids in Synthesis; Wiley-VCH: Weinheim, 2003. (9) Ediger, M. D. Annu. ReV. Phys. Chem. 2000, 51, 99. (10) Richert, R. J. Phys.: Condens. Matter 2002, 14, R703. (11) Triolo, A.; Russina, O.; Bleif, H. J.; DiCola, E. J. Phys. Chem. B 2007, 111, 4641. (12) Mandal, P. K.; Paul, A.; Samanta, A. J. Photochem. Photobiol. A 2006, 182, 113. (13) Mandal, P. K.; Sarkar, N.; Samanta, A. J. Phys. Chem. A 2004, 108, 9048. (14) Shigeto, S.; Hamaguchi, H. Chem. Phys. Lett. 2006, 427, 329. (15) Lopes, J. N. C.; Padua, A. A. H. J. Phys. Chem. B 2006, 110, 3330. (16) Wang, Y.; Voth, G. A. J. Phys. Chem. B 2006, 110, 18601. (17) Hu, Z.; Margulis, C. J. Proc. Natl. Acad. Sci. 2006, 103, 831. (18) Hu, Z.; Margulis, C. J. J. Phys. Chem. B 2006, 110, 11025. (19) Shim, Y.; Jeong, D.; Choi, M. Y.; Kim, H. J. J. Chem. Phys. 2006, 125, 061102. (20) Demchenko, A. P. Luminescence 2002, 17, 19. (21) Nemkovich, N. A.; Rubinov, A. N.; Tomin, V. I. In Topics in Fluorescence Spectroscopy; Lakowicz, J. R., Ed.; Plenum Press: New York, 1991; Vol. 2, pp 367. (22) Fee, R. S.; Milsom, J. A.; Maroncelli, M. J. Phys. Chem. 1991, 95, 5170. (23) Pronounced red-edge excitation shifts in the emission of neat ionic liquids have also been well documented by Samanta and coworkers.12,56-58 (24) Jin, H.; O’Hare, B.; Dong, J.; Arzhantsev, S.; Baker, G. A.; Wishart, J. F.; Benesi, A.; Maroncelli, M. J. Phys. Chem. B, in press. (25) Horng, M. L.; Gardecki, J. A.; Papazyan, A.; Maroncelli, M. J. Phys. Chem. 1995, 99, 17311. (26) Fee, R. S.; Maroncelli, M. Chem. Phys. 1994, 183, 235. (27) Chakrabarty, D.; Hazra, P.; Chakraborty, A.; Seth, D.; Sarkar, N. Chem. Phys. Lett. 2003, 381, 697. (28) Wang, L. M.; Richert, R. J. Chem. Phys. 2004, 120, 11082. (29) Horng, M.-L.; Gardecki, J.; Maroncelli, M. J. Phys. Chem. 1997, 101, 1030. (30) Lewis, J. E.; Maroncelli, M. Chem. Phys. Lett. 1998, 282, 197. (31) The data labeled ν(t) in Figure 1 are weighted averages of the integral correlation times obtained from stretched exponential fits of the peak and first moment frequencies of the spectra. No account of the unobserved portion of the spectra is made in these fits. (32) Ghosh, S.; Mandal, U.; Adhikari, A.; Dey, S.; Bhattacharyya, K. Int. ReV. Phys. Chem. 2007, 26, 421. (33) Maroncelli, M. J. Mol. Liq. 1993, 57, 1. (34) Maroncelli, M. J. Chem. Phys. 1991, 94, 2084. (35) Carter, A. E.; Hynes, J. T. J. Chem. Phys. 1991, 94, 5961. (36) Richert, R. J. Chem. Phys. 2001, 114, 7471. (37) Shim, Y.; Jeong, D.; Manjari, S.; Choi, M. Y.; Kim, H. J. Acc. Chem. Res., ACS ASAP. (38) Hu, Z.; Margulis, C. J. Acc. Chem. Res. 2007, ASAP. (39) Kobrak, M. N. J. Chem. Phys. 2006, 125. (40) Patel, N.; Conte, S.; Maroncelli, M. 2007, unpublished results. (41) Law, K. Y. Chem. Phys. Lett. 1980, 75, 545. (42) Hooker, J. C.; Torkelson, J. M. Macromolecules 1995, 28, 7683. (43) Haidekker, M. A.; Theodorakis, E. A. Org. Biomol. Chem. 2007, 5, 1669. (44) Gutkowski, K. I.; Japas, M. L.; Aramendia, P. F. Chem. Phys. Lett. 2006, 426, 329. (45) Lu, J.; Liotta, C. L.; Eckert, C. A. J. Phys. Chem. A 2003, 107, 3995. (46) Jin, H.; Arzhantsev, S.; Swalina, C.; Maroncelli, M., unpublished results. (47) Blanchard-Desce, M.; Wortmann, R.; Lebus, S.; Lehn, J. M.; Kramer, P. Chem. Phys. Lett. 1995, 243, 526. (48) Loutfy, R. O.; Law, K. Y. J. Phys. Chem. 1980, 84, 2803. (49) Mqadmi, S.; Pollet, A. J. Photochem. Photobiol. A 1990, 53, 275. (50) Accurate measurements of quantum yields for these molecules require that one account for the fact that their emission in most solvents is not depolarized. This means that the polarization properties of the quantum yield reference and the polarization sensitivity of the detection system must be considered. Here, all emission spectra and emission responsivity

13478 J. Phys. Chem. B, Vol. 111, No. 48, 2007 correction factors were measured using magic angle detection. Radiative rates were assumed to vary as ν3 (i.e., emission transition moments were pk 3 assumed constant), and lifetimes were calculated as τ ) (φ/kref rad)(νem/νem) where νpk em and νem refer to the first moment frequencies excited at the peak and at an arbitrary excitation frequency, and kref rad is the reference radiative rate taken to be 2.8 × 108 s-1 for both DMN and JDMN.48,49 (51) Karpiuk, J. J. Phys. Chem. A 2004, 108, 11183. (52) Dahl, K.; Biswas, R.; Ito, N.; Maroncelli, M. J. Phys. Chem. B 2005, 109, 1563. (53) Glotzer, S. C. J. Non-Cryst. Solids 2000, 274, 342.

Letters (54) Fletcher, K. A.; Pandey, S.; Storey, I. K.; Hendricks, A. E.; Pandey, S. Anal. Chim. Acta 2002, 453, 89. (55) Ito, N.; Arzhantsev, S.; Maroncelli, M. Chem. Phys. Lett. 2004, 396, 83. (56) Samanta, A. J. Phys. Chem. B 2006, 110, 13704. (57) Paul, A.; Mandal, A. A.; Samanta, A. Chem. Phys. Lett. 2005, 402, 375. (58) Paul, A.; Mandal, A. A.; Samanta, A. J. Phys. Chem. B 2005, 109, 9148. (59) Hamaguchi, H.; Ozawa, R. AdV. Chem. Phys. 2005, 131, 85.