Revealing Long-Range Density Fluctuations in Dialkylimidazolium

2 Aug 2007 - Process optimization and performance evaluation on sequential ionic liquid dissolution–solid acid saccharification of sago waste. Kiat ...
0 downloads 0 Views 101KB Size
Download PDF
9858

J. Phys. Chem. B 2007, 111, 9858-9863

Revealing Long-Range Density Fluctuations in Dialkylimidazolium Chloride Ionic Liquids by Dynamic Light Scattering Qinglin Kuang,†,‡ Jun Zhang,† and Zhigang Wang*,† CAS Key Laboratory of Engineering Plastics, Joint Laboratory of Polymer Science and Materials, Beijing National Laboratory for Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing, 100080 P. R. China, and Graduate School of Chinese Academy of Sciences, Beijing, 100049 P. R. China ReceiVed: March 2, 2007; In Final Form: June 15, 2007

In this study, the structures and dynamics of ionic liquids of 1-allyl-3-methylimidazolium chloride ([AMIM][Cl]) and 1-n-butyl-3-methylimidazolium chloride ([BMIM][Cl]) were studied by dynamic light scattering with polarized and depolarized geometries in the temperature range from 300 to 400 K. The temperature range covered supercooled and liquid states for [BMIM][Cl] and covered the liquid state for [AMIM][Cl]. The results show that for these ionic liquids at all chosen temperatures only one ultraslow relaxation is observed in the polarized component of dynamic light scattering, however, the ultraslow relaxation is not observed in the depolarized component. The ultraslow relaxation exhibited several typical features of the “cluster” mode generally found in glass-forming liquids and polymer melts, such as diffusive, strongly scattering-vectordependent, and nearly exponential characters, which thus corresponded to long-range density fluctuations. The physical origin for long-range density fluctuations was the existence of heterogeneities with large characteristic length scales in these ionic liquids. It was further considered that molecules of these ionic liquids not only tended to aggregate to form dynamic clusters but also possibly formed dynamic networks in the supercooled state and the heterogeneities could exist even at temperatures higher than the melting points.

Introduction The increasing number and relevance of applications for room-temperature ionic liquids are related to their possible exploitation as green solvents or reaction media because of their negligible vapor pressure and dissolvability for a wide range of polar or nonpolar, organic or inorganic molecules,1 which stimulate a deep understanding of the structures of ionic liquids in bulk liquid at the molecular scale. One of the most widely studied ionic liquid families is imidazolium-based halide. Experiments2-6 and simulations7-12 show that there exists strong hydrogen bonding between anion and cation in dimethylimidazolium-based halide and a considerable degree of order presents in the local structure of the liquid phase, closely resembling that found in the crystalline phase. Moreover, an extended structure, namely, network, can be formed through hydrogen bonding for dialkylimidazolium chloride12 or through Coulombic force for dialkylimidazolium iodide.3 Dupont13 presents a reasonable concept in one of his reviews that “pure” imidazolium ionic liquids can be described as polymeric hydrogen-bonded supermolecules, which are highly ordered hydrogen-bonded materials. Recently, several studies14-17 showed that there exists spatial heterogeneity in the local environments for some imidazoliumbased ionic liquids in the liquid state. In the study of timedependent fluorescence of dipolar molecules in 1-butyl-3methylimidazolium hexafluorophosphate ([BMIM][PF6]), Samanta and co-workers14-15 observed the “red-edge effect” * To whom correspondence should be addressed. Tel: 011-86-1062558172; Fax: 011-86-10-62558172; E-mail: [email protected]. † Chinese Academy of Sciences. ‡ Graduate School of Chinese Academy of Sciences.

phenomenon, which is a characteristic of low-temperature glasses, polymers, and organized assemblies such as micelles. This suggests that imidazolium-based ionic liquids not only are dynamically heterogeneous but also display locally heterogeneous environments. Hu et al.16 also predicted the existence of heterogeneity in [BMIM][PF6]. Simulations17 with multi-scale coarse-graining models showed that, for 1-n-butyl-3-methylimidazolium nitrate ([BMIM][NO3]), neutral tail groups of the cations aggregate to form spatially heterogeneous domains of the tails, while the charged head groups and anions distribute as uniformly as possible due to strong electrostatic interactions. However, direct experimental evidence still lacks to prove this simulation result. Dynamic light scattering (DLS) is a noninvasive technique with a wide dynamic range (at least 10-6-103 s) that suits the study of the structure and dynamics of an amorphous system. In the literature, a few papers have demonstrated structured heterogeneities of low molar mass18-23 and polymeric24-27 glassforming liquids by DLS. In these studies, it has been discovered that, besides R-relaxation and β-relaxation, an additional ultraslow relaxation with correlation length scales of up to 100 nm and with a characteristic relaxation time of many orders of magnitude longer than the R-relaxation time can be observed in the polarized component of DLS, which can be explained by long-range density fluctuations, the so-called “clusters”. This type of ultraslow relaxation shows diffusive, strongly scatteringvector-dependent, and nearly exponential characters. Dialkylimidazolium halides are popular ionic liquids. Recently, they were used as new types of solvents for natural polymers, such as cellulose28-29 and silk.30 To understand the dissolving mechanisms for natural polymers, it is necessary to

10.1021/jp071733+ CCC: $37.00 © 2007 American Chemical Society Published on Web 08/02/2007

Dialkylimidazolium Chloride Ionic Liquids understand the structures and dynamics of these ionic liquids in the liquid state first. In the literature, primary works have been done to explore the local structures of dialkylimidazolium halides by using neutron scattering and NMR techniques, but few works have been performed by using DLS. In this study, ionic liquids of dialkylimidazolium chlorides, 1-allyl-3-methylimidazolium chloride ([AMIM][Cl]) and 1-n-butyl-3-methylimidazolium chloride ([BMIM][Cl]), were chosen to study by using DLS. DLS measurements with polarized and depolarized geometries, respectively, were performed at various scattering angles and various temperatures. The DLS results show that one ultraslow relaxation always exists in both ionic liquids at all chosen temperatures. This ultraslow relaxation exhibits typical features of the “cluster” mode found in glass-forming liquids, such as diffusive, strongly scattering-vector-dependent, and nearly exponential characters. Interpretations to this ultraslow relaxation associated with long-range density fluctuations and the structures of the ionic liquids will be reported in this article. Experimental Section Samples. Ionic liquids of dialkylimidazolium chlorides, 1-allyl-3-methylimidazolium chloride ([AMIM][Cl]) and 1-nbutyl-3-methylimidazolium chloride ([BMIM][Cl]), were synthesized for the present study according to previous works.31 Synthesized colorless ionic liquids were vacuum distilled at 60 °C for 24 h, then dried at 60 °C for 12 h and at 100 °C for 6 h with dry nitrogen purge to remove unreacted reagents. Finally, the purified ionic liquids were dried in vacuum oven at 60 °C for 48 h. The vacuum oven contained phosphorus pentoxide (P2O5) for removing water and impurities from the samples. The obtained colorless ionic liquids had purities of >99 wt % measured by using NMR spectroscopy. Water content was 0.11 wt % for [AMIM][Cl] and 0.12 wt % for [BMIM][Cl], which were measured by using the Karl Fischer titration method. The refractive indexes were 1.5465 for [AMIM][Cl] and 1.5510 for [BMIM][Cl], which were measured at 298 K by using Abbe’s refractometer. The glass transition temperature (Tg) and melting point (Tm) were 222 and 290 K, respectively, for [AMIM][Cl] and 230 and 339 K, respectively, for [BMIM][Cl], which were obtained from differential scanning calorimetry (DSC) measurements. DLS Measurements. To prepare dust-free samples, purified colorless [AMIM][Cl] and [BMIM][Cl] were filtered through 1 µm pore-size nylon membrane filters into dust-free light scattering cells at 70 °C in a glove box. Note that these ionic liquids cannot dissolve nylon. Pore sizes of the nylon membrane filters should be large enough to avoid affecting the structures of ionic liquids. It was found that 1 µm pore-size nylon membrane filters were sufficient to use to clean the samples without any obvious side effects on DLS measurements. Then, samples were sealed and allowed to cool to room temperature. Samples were stored in a desiccator containing P2O5 for a week before DLS measurements. Note that the structures and properties of ionic liquids could be remarkably affected by water absorption,32-36 thus it was necessary for samples to avoid water absorption in air. A commercial light scattering spectrometer (ALV/DLS/SLS5022F) equipped with a multi-τ digital time correlator (ALV5000) and a cylindrical 22 mW UNIPHASE He-Ne laser (wavelength λ0 of 632.8 nm) was used for the DLS measurements. The ALV/ DLS/SLS-5022F system allowed measurements of the autocorrelation functions of the scattered light, which was detected by a single photon detector. Samples were measured over a wide

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9859 temperature range from 300 to 400 K. This temperature range covered supercooled and liquid states for [BMIM][Cl] and covered the liquid state for [AMIM][Cl]. The experimental temperatures less than 400 K were chosen due to the following reason: It is known that ionic liquids have only a limited thermal stability. Even though the thermal gravimetric analysis (TGA) results showed that the onset temperatures of thermal decomposition were above 523 K for both ionic liquids used in this study, temperatures less than 400 K for DLS measurements were specifically chosen mainly considering the possible duration effect on thermal decomposition at high temperature. In fact, NMR experiments showed that, for both ionic liquids sealed in the light scattering cells and held at 400 K for 24 h, only [AMIM][CL] showed a decomposition of about 5 wt %. Therefore, when the DLS experiment was performed at a temperature higher than 373 K for [AMIM][CL], the DLS experimental duration was less than 2 h to avoid any thermal decomposition. The sample was allowed to equilibrate at each experimental temperature for 1 h before the DLS experiment was commenced. Sample cell assembly was thermostatic to maintain temperature fluctuations within 0.1 °C from the set value. In DLS, the normalized time-dependent intensity correlation function g(2)(t,q) in the self-beating mode was measured (eq 1), where t is the time and q the modulus of the scattering vector (q ) (4πn/λ0)sin(θ/2), where n is the refractive index of the medium, λ0 the wavelength, and θ the scattering angle)37

g(2)(t,q) )

〈I(0,q)I(t,q)〉 〈I(0,q)〉2

(1)

It should be noted that the scattered light intensities were detected and processed in an autocorrelation scheme using one avalanche photodetector. Equation 1 could be expressed in terms of the normalized first-order electric field time correlation function g(1)(t,q) through the Siegert relation as38

g(2)(t,q) ) 1 + f |g(1)(t,q)|2

(2)

where f is the instrumental coherence factor, which depends on the coherence areas seen by the detector and on the finite speed of signal processing and averaging. In DLS experiments, the spectrometer had a high coherence factor of f ∼ 0.95 because of the novel single-mode fiber optics coupled with an efficient avalanche photodiode. In DLS measurements, the scattering angles were set larger than 20° to exclude the effect of heterodyne contribution. In a realistic experimental situation, there is a possible contribution from stray light, for example, from dusts on the cell windows or from the cell windows themselves. Therefore, the experiments had to be designed in such a way that only the scattered light from the sample itself was collected by the detector. By setting the scattering angle larger than 20° and making the incidence light intensity large enough, the possible interference from a heterodyne contribution to the correlation function could be excluded. The g(1)(t) values at different temperatures were fitted to the stretched-exponential Kohlrausch-Williams-Watts (KWW) function as39

g(1)(t) ) A(e-Γt)β, (0 < β e 1)

(3)

where A, Γ, and β are the fitting parameters and A is the amplitude of g(1)(t,q), Γ the relaxation rate, and β the stretched exponent. The value of β depended on the type of relaxation. For R-relaxation, the value of β was about 0.8, while for the “cluster” mode, the value of β was about 0.98.

9860 J. Phys. Chem. B, Vol. 111, No. 33, 2007

Kuang et al.

Figure 2. Normalized time-dependent intensity correlation functions for [AMIM][Cl] measured at θ ) 20, 30, 45, 60, 90, 120, and 150° at 338 K.

Figure 1. (a) Normalized time-dependent intensity correlation functions for [AMIM][Cl] and [BMIM][Cl] at θ ) 90° at 338 K. The solid lines are the fitting curves by using single KWW functions (eq 3). (b) Residuals from the curve fittings.

Results Figure 1a shows typical plots of normalized time-dependent intensity correlation functions for the ionic liquids of [AMIM][Cl] and [BMIM][Cl] measured at a scattering angle θ of 90° and a temperature of 338 K. This temperature is above or near their respective melting points. The solid lines in Figure 1a represent the fitting curves by using single KWW functions (eq 3). The residuals of the fittings are shown in Figure 1b. The small residuals demonstrate that the fitting curves by using single KWW functions were in good agreement with the correlation functions for both ionic liquids. The normalized time-dependent intensity correlation functions consist of one relaxation mode. The relaxation time τ is 0.11 s for [AMIM][Cl] and 1.20 s for [BMIM][Cl]. The exponent β is 0.96 for [AMIM][Cl] and 1.00 for [BMIM][Cl]. The above values suggest that the relaxation mode is close to the single-exponential type. It can be also seen from Figure 1a that the amplitudes of g(1)(t,q), A, are different for these ionic liquids, possibly indicating the different degrees of heterogeneity induced by different molecular structures in the ionic liquids. According to the order of relaxation time, the relaxation should be an ultraslow one, which is similar to the “cluster” mode or the Fischer mode.22 The nature of the ultraslow relaxation will be the main issue in this article. Figure 2 shows typical normalized time-dependent intensity correlation functions for [AMIM][Cl] measured at 338 K at different scattering angles. It can be seen that the relaxation exhibits strong scattering-vector-dependence according to the obvious changes of amplitude and relaxation time. The relaxation time changes about 2 orders of magnitude when the scattering angle changes from 20° to 150°. A similar phenomenon was observed for [BMIM][Cl]. The relaxation rates, Γ, against q2 are plotted in Figure 3, noting that the experimental temperature is 338 K for [AMIM][Cl] and 377 K for [BMIM][Cl]. Relaxation rates exhibit nearly linear dependences on q2 and the linear regression lines pass through the origin, which demonstrate that the ultraslow relaxation mode has diffusive character. Hence, it is possible to assign a “cluster” diffusion coefficient, Dcl, to the ultraslow relaxation process according

Figure 3. Plots of relaxation rate Γ against q2 for [AMIM][Cl] at 338 K and [BMIM][Cl] at 377 K. Note that only the relaxation rates measured at low temperatures of 303 and 318 K have relatively large errors of less than 5%.

TABLE 1: Dynamic Viscosity of Ionic Liquids at Different Temperatures Measured with a Fixed Steady Shear Rate of 1 s-1 temperature (K)

viscosity of [AMIM][Cl] (Pa‚s)

viscosity of [BMIM][Cl] (Pa‚s)

303 318 338 357 377 394

1.15 0.31 0.092 0.036 0.017 0.015

3.95 1.05 0.28 0.10 0.042 0.027

to the slope. If dynamic viscosity is available, then it is possible to determine the hydrodynamic size or correlation length, ξ, through the Kawasaki-Stokes-Einstein equation

ξ ) kT/6πηDcl

(4)

where k is the Boltzmann constant and η the dynamic viscosity at temperature T. Dynamic viscosities of the ionic liquids at different temperatures were measured in steady shear with a fixed shear rate of 1 s-1 during the temperature ramping step from 303 to 400 K by using the stress-controlled rheometer (TA AR2000, TA company, USA). Table 1 lists the dynamic viscosities of these ionic liquids at different temperatures. Thus, a correlation length of 120 ( 16 nm at 338 K was obtained for [AMIM][Cl] and a correlation length of 205 ( 26 nm at 377 K was obtained for [BMIM][Cl]. When compared at the same temperature of 338 K, the correlation length was 120 ( 16 nm for [AMIM][Cl] and 421 ( 12 nm for [BMIM][Cl]. Note that the errors for correlation length ζ were obtained by considering the errors for η and Dcl. Viscosity η was directly obtained from rheological measurement with an instrument precision of 0.01 mPa‚s. The error of viscosity η was less than 5%. The error for Dcl was associated with the errors for the scattering vector q and relaxation rate Γ. In fact, the scattering

Dialkylimidazolium Chloride Ionic Liquids

Figure 4. Normalized time-dependent intensity correlation functions for [BMIM][Cl] at θ ) 90° measured at temperatures of 303, 318, 338, 357, 377, 377 (annealed for 24 h), and 394 K (arranged from left to right). Inset shows the temperature dependence of stretched exponent β for [AMIM][Cl] and [BMIM][Cl].

Figure 5. Arrhenius plots of relaxation time in the temperature range from 303 to 394 K for [AMIM][Cl] and [BMIM][Cl]. Note that only the relaxation times measured at low temperatures of 303 and 318 K have relatively large errors of less than 5%.

vector q was obtained from 4πn/λ0 sin(θ/2) with a precision of 0.0001 nm-1 and the relaxation rate Γ was obtained from the average of the data evaluation procedure for 3 times. In addition, the magnitude of correlation lengths for [AMIM][Cl] and [BMIM][Cl] is 100 nm, which is similar to that observed in other glass-forming liquids19 and polymer melts.24-27 Figure 4 shows the temperature effects on normalized timedependent intensity correlation functions measured at a scattering angle of 90° for [BMIM][Cl]. It can be found that the shapes of the correlation functions for various temperatures look similar, however, the relaxation time of the ionic liquid decreases dramatically with increasing temperature, that is, from 14.7 s at 303 K to 32 ms at 394 K. Note that, even in the supercooled state (T < 338 K), only one ultraslow relaxation mode for [BMIM][Cl] still exists. Moreover, even after [BMIM][Cl] had been annealed at 377 K for 24 h, the correlation function did not show any obvious changes. It is evinced that the entropy in the liquid phase of the ionic liquid is stable in a broad temperature range, and this entropy stabilization may contribute largely to degrading the melting point. The results of [AMIM][Cl] are similar to that of [BMIM][Cl]. The inset in Figure 4 shows the changes of stretched exponent β with temperature. The stretched exponents are close to 1 for both ionic liquids, indicating that the temperature effect on the stretched exponent is quite small. Figure 5 shows the Arrhenius plots of relaxation time of [AMIM][Cl] and [BMIM][Cl] in the same temperature range. Two features can easily be found in Figure 5. The first feature is that both relaxations show good Arrhenius relations with temperature. The second feature is that the slope for [AMIM][Cl] is lower than that for [BMIM][Cl], suggesting less fragility for the former. The activation energy of the ultraslow relaxation

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9861

Figure 6. Temperature dependences of correlation length for [AMIM][Cl] and [BMIM][Cl].

can be determined from the slope, which is 53.6 kJ/mol for [AMIM][Cl] and 67.4 kJ/mol for [BMIM][Cl]. In terms of a network model, the activation energy is associated with the thermally activated creation and subsequent relaxation of excursions of the network from its equilibrium configurations. The activation energies of the ionic liquids are much larger than the activation energy of about 29 kJ/mol for the cluster diffusions in melts of poly(n-laurylmethacrylate) (PLMA)27 and poly(ethylene oxide) (PEO).40 Obviously, the interactions between clusters due to hydrogen bonding and Coulombic force in the ionic liquids are much stronger than those in polymer melts. The temperature dependences of correlation length obtained from the Kawasaki-Stokes-Einstein equation for [AMIM][Cl] and [BMIM][Cl] are shown in Figure 6. It can be found that the correlation lengths decrease dramatically when temperature just passes through the melting points for both ionic liquids. For [BMIM][Cl], the correlation length at above the melting point decreases exponentially with increasing temperature. For [AMIM][Cl], correlation length at above the melting point also decreases exponentially with increasing temperature. However, the decrease of correlation length is less significant for [AMIM][Cl] than for [BMIM][Cl]. It is worth mentioning that to understand the origin of this ultraslow relaxation mode, comparisons were made between the correlation functions from polarized and depolarized geometries for both [AMIM][Cl] and [BMIM][Cl]. The results showed that the ultraslow relaxation mode was evidently measurable in the polarized component but was not measurable in the depolarized component in the time window of DLS for [AMIM][Cl] and [BMIM][Cl]. The above results indicate the absence of depolarization from the molecular orientation or shear waves in the depolarized component of DLS and the existence of long-range density fluctuations associated with the ultraslow relaxation mode in the polarized component of DLS. Discussion A comprehensive DLS study on imidazolium chloride ionic liquids within a wide temperature range from 300 to 400 K has been performed. The experimental time window covers eight decades (10-6-102 s). Correlation functions for [AMIM][Cl] and [BMIM][Cl] at all chosen temperatures exhibit only one ultraslow relaxation mode in the polarized component, with correlation length in the range of 90-470 nm; however, correlation functions are flat in the depolarized component. Although the physical origin of the ultraslow relaxation mode has not been confirmed by other methods, it could be attributed to the “cluster” mode or long-range density fluctuations, which are similar to those observed in low molar mass and polymeric glass-forming liquids. Among the low molar mass glass-forming

9862 J. Phys. Chem. B, Vol. 111, No. 33, 2007 liquids, o-terphenyl has been widely studied in the literature.19 Besides the two-step relaxations (namely, R- and β-relaxations) for common glass-forming liquids at near Tg, an additional slow relaxation appears, which is several orders of magnitude slower than the usual density fluctuations (R-relaxation). Fischer accounted this slow relaxation to the “cluster” mode,22 because the samples without “clusters” behave like simple liquids at above Tg. In the case of these ionic liquids, only one ultraslow relaxation exists in the correlation functions of DLS at above Tm. Recall that the structured heterogeneity inherent in ionic liquids, which may be referred to interpret the ultraslow relaxation mode, has been studied by using many other techniques. Neutron diffraction measurements demonstrate that ionic liquids have intermediate-range charge ordering and the local structures resemble those found in solid states.2 The molecular packing and interactions in the first two or three coordination shells are similar in both the crystal and liquid states, although the absolute distances can be altered due to melting. The short-range order in 1 nm is local and microscopic and is not sufficient to prove the existence of the heterophase in all ionic liquids, for example, 1-n-butyl-3-methylimidazolium iodide ([BMIM][I]) ionic liquid is isotropic and transparent.3 However, a few recent studies show that special spatial heterogeneity exists in some ionic liquids. Observation of the “red-edge effect” phenomenon in the fluorescence study of 1-butyl-3-methylimidazolium hexafluorophosphate ([BMIM][PF6]) suggests that imidazolium-based ionic liquids not only are dynamically heterogeneous but also display local heterogeneous environments. Hu et al.16 theoretically predicted the existence of heterogeneity in [BMIM][PF6]. Simulations17 on 1-butyl-3-methylimidazolium nitrate ([BMIM][NO3]) show that neutral tail groups of the cations aggregate to form spatially heterogeneous domains, while the charged head groups and anions distribute as uniformly as possible due to strong electrostatic interactions. Moreover, an extended structure, namely, network, can be formed by hydrogen bonds in dialkylimidazolium chloride. A recent study on hydrogen bonds in dialkylimidazolium halide by ab initio and density functional theory (DFT) calculations12 shows that in the local structure of 1-ethyl-3-methylimidazolium chloride ([EMIM][Cl]), each [Cl] anion is surrounded by three [EMIM] cations and the case is similar for the [EMIM] cation. Further study by the periodic boundary conditions shows that the local structures can be periodically extended in three dimensions and the hydrogenbonded network exists in dialkylimidazolium halide. Dupont13 pointed out that pure 1,3-dialkylimidazolium ionic liquids could be well described as hydrogen-bonded polymeric supermolecules of the type {[(DAI)x(X)x-n]n+[(DAI)x-n(X)x]n-}n, where DAI is the 1,3-dialkylimidazolium cation and X the anion. For the ionic liquids studied in this article, the heterogeneity is considered to exist. Due to strong hydrogen bonding between anions and cations, short-range order forms. Because both the alkyl tails of the cations are long enough, they may aggregate to form “dynamic clusters”. Anions and some minority of water molecules surrounding the clusters may form polar regimes. Due to electrostatic repulsive interactions, “cavities” between the clusters also possibly exist. The dynamic clusters and cavities can result in the spatial heterogeneities in the ionic liquids. Furthermore, a transient physical network may exist due to the extended hydrogen bonding between the clusters, which explains that entropy in the liquid phase of dialkylimidazolium halide ionic liquids can be stable in a broad temperature range. With tie points continually forming and disintegrating, network rearrangements, namely, long-range density fluctuations, can

Kuang et al. give rise to ultraslow relaxation. However, fast relaxations with the “cluster” mode like in other glass-forming liquids cannot be found in these ionic liquids in our DLS measurements. It has been found that glassy [BMIM][Cl] shows a fast β-relaxation above Tg in the quasielastic neutron scattering measurements;41 therefore, a possible explanation as to the absence of fast relaxations in DLS measurements for the ionic liquids is that the fast relaxations may shift out the time window of DLS at the chosen temperatures. In addition, relaxation time, correlation length, and activation energy of the ultraslow relaxation depend on the length of alkyl substituent on imidazolium cation and the different anion. Unsaturated alkyl substituent in [AMIM][Cl] can greatly increase electronic cloud density and decrease polarity of the molecule. Butyl tails of the cations in [BMIM][Cl] can aggregate more easily with each other than allyl tails of the cations in [AMIM][Cl]. These different molecular structures cause relatively weaker interactions in [AMIM][Cl] than in [BMIM][Cl]. Therefore, at the same temperature, slower relaxation, longer correlation length, and higher activation energy can be found in [BMIM][Cl] than in [AMIM][Cl]. The physical meaning of the lower contrast observed in the correlation function for [AMIM][Cl] in comparison with [BMIM][Cl] in Figure 1a might be due to the less degree of heterogeneity in [AMIM][Cl] than in [BMIM][Cl], which is also induced by the above-mentioned different molecular structures of the ionic liquids. Conclusions In this work, dialkylimidazolium chloride ionic liquids were studied by DLS over a wide temperature range from 300 to 400 K, including supercooled and liquid states. Only one ultraslow relaxation was found in the polarized component of DLS at all chosen temperatures for both ionic liquids, even after annealing for a long time at high temperature, but the ultraslow relaxation could not be found in the depolarized component. The ultraslow relaxation mode showed diffusive, strongly scattering-vector-dependent, nearly exponential characters, which were usually found in low molar mass glass-forming liquids and polymer melts. The relaxation could be attributed to the “cluster” mode or long-range density fluctuations. Due to strong hydrogen bonding between anions and cations and aggregation of alkyl tails of the cations, dynamic cluster structures, which could result in long-range density fluctuations, were considered to form in both ionic liquids. It was further considered that dynamic networks possibly existed in the supercooled state and that the heterogeneities could exist even at temperatures higher than the melting points. Acknowledgment. Z.W. acknowledges financial support from the “One Hundred Talents” Program of the Chinese Academy of Science, the National Science Foundation of China with Grant No. 10590355 for the Key Project on Evolution of Structure and Morphology during Polymer Processing and the National Science Foundation of China with Grant No. 20674092. References and Notes (1) Rogers, R.; Seddon, K. R. Ionic Liquids IIIB: Fundamentals, Progress, Challenges and Opportunities; ACS Symposium Series 902; American Chemical Society: Washington, DC, 2005. (2) Hardacre, C.; Holbrey, J. D.; McMath, S. E. J.; Bowron, D. T.; Soper, A. K. J. Chem. Phys. 2003, 118, 273. (3) Katayanagi, H.; Hayashi, S.; Hamaguchi, H. O.; Nishikawa, K. Chem. Phys. Lett. 2004, 392, 460. (4) Abdulsada, A. K.; Greenway, A. M.; Hitchcock, P. B.; Mohammed, T. J.; Seddon, K. R.; Zora, J. A. J. Chem. Soc., Chem. Commun. 1986, 24, 1753.

Dialkylimidazolium Chloride Ionic Liquids (5) Elaiwi, A.; Hitchcock, P. B.; Seddon, K. R.; Srinivasan, N.; Tan, Y. M.; Welton, T.; Zora, J. A. J. Chem. Soc., Dalton Trans. 1995, 21, 3467. (6) Huang, J. F.; Chen, P. Y.; Sun, W.; Wang, S. P. Spectrosc. Lett. 2001, 34, 591. (7) Urahata, S. M.; Ribeiroa, M. C. C. J. Chem. Phys. 2003, 120, 1855. (8) Wang, Y.; Li, H. R.; Han, S. J. J. Chem. Phys. 2005, 123, 174501. (9) Wang, Y.; Li, H. R.; Han, S. J. J. Chem. Phys. 2006, 124, 044504. (10) Shim, Y.; Choi, M. Y.; Kim, H. J. J. Chem. Phys. 2005, 122, 044510. (11) Shim, Y.; Choi, M. Y.; Kim, H. J. J. Chem. Phys. 2005, 122, 044511. (12) Dong, K.; Zhang, S. J.; Wang, D. X.; Yao, X. Q. J. Phys. Chem. A 2006, 110, 9775. (13) Dupont, J. J. Braz. Chem. Soc. 2004, 15, 341. (14) Mandal, P.; Sarkar, M.; Samanta, A. J. Phys. Chem. A 2004, 108, 9048. (15) Paul, A.; Mandal, P. K.; Samanta, A. J. Phys. Chem. B 2005, 109, 9148. (16) Hu, Z.; Margulis, C. J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 831. (17) Wang, Y.; Voth, G. A. J. Am. Chem. Soc. 2005, 127, 12192. (18) Fischer, E. W.; Becker, C.; Hagenah, J. U.; Meier, G. Prog. Coll. Polym. Sci. 1989, 80, 198. (19) Fischer, E. W. Physica A 1993, 201, 183. (20) Kanaya, T.; Patkowski, A.; Fischer, E. W. Acta Polym. 1994, 45, 137. (21) Patkowski, A.; Fischer, E. W.; Steffen, W.; Glaeser, H.; Baumann, M.; Ruths, T.; Meier, G. Phys. ReV. E 2001, 63, 061503. (22) Fischer, E. W.; Bakai, A.; Patkowski, A.; Steffen, W.; Reinhardt, L. J. Non-Cryst. Solids 2002, 307, 584. (23) Bakai, A. S.; Fischer, E. W. J. Chem. Phys. 2004, 120, 5235. (24) Jian, T.; Anastasiadis, S. H.; Semenov, A. N.; Fytas, G.; Adachi, K.; Kotaka, T. Macromolecules 1994, 27, 4762.

J. Phys. Chem. B, Vol. 111, No. 33, 2007 9863 (25) Kanaya, T.; Patkowsk, A.; Fischer, E. W.; Seils, J.; Glaser, H.; Kaji, K. Macromolecules 1995, 28, 7831. (26) Stepanek, P.; Lodge, T. P. Macromolecules 1996, 29, 1244. (27) Brown, W.; Stepanek, P. J. Polym. Sci., Part B: Polym. Phys. 1997, 35, 1013. (28) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am. Chem. Soc. 2002, 124, 4974. (29) Zhang, H.; Wu, J.; Zhang, J.; He, J. S. Macromolecules 2005, 38, 8272. (30) Phillips, D. M.; Drummy, L. F.; Conrady, D. G.; Fox, D. M.; Naik, R. R.; Stone, M. O.; Trulove, P. C.; De Long, H. C.; Mantz, R. A. J. Am. Chem. Soc. 2004, 126, 14350. (31) Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis; Wiley: Weinheim, Germany, 2003. (32) Anthony, J. L.; Maginn, E. J.; Brennecke, J. F. J. Phys. Chem. B 2001, 105, 10942. (33) Hanke, C. G.; Lynden-Bell, R. M. J. Phys. Chem. B 2003, 107, 10873. (34) Pandey, S.; Fletcher, K. A.; Baker, S. N.; Baker, G. A. Analyst 2004, 129, 569. (35) Xu, H. T.; Zhao, D. C.; Xu, P.; Liu, F. Q.; Gao, G. J. Chem. Eng. Data 2005, 50, 133. (36) Widegren, J. A.; Saurer, E. M.; Marsh, K. N.; Magee, J. W. J. Chem. Thermodyn. 2005, 37, 569. (37) Cummins, H. Z.; Pike, E. R. Photon Correlation and Light Beating Spectroscopy; Plenum Press: New York, 1974. (38) Berne, B. J.; Pecora, R. Dynamic Light Scattering; Plenum Press: NewYork, 1976. (39) Williams, G.; Watts, D. C. Trans. Faraday Soc. 1970, 66, 80. (40) Walkenhorst, R.; Selser, J. C.; Piet, G. J. Chem. Phys. 1998, 109, 11043. (41) Inamuraa, Y.; Yamamuroa, O.; Hayashib, S.; Hamaguchib. H. Physica B-Condens. Matter 2006, 385, 732.