Temperature Dependence of the Primary ... - ACS Publications

Jun 1, 2009 - Olga Russina, Mario Beiner, Catherine Pappas, Margarita Russina, Valeria Arrighi, Tobias Unruh, Claire L. Mullan, Christopher Hardacre a...
0 downloads 0 Views 1MB Size
Download PDF
J. Phys. Chem. B 2009, 113, 8469–8474

8469

ARTICLES Temperature Dependence of the Primary Relaxation in 1-Hexyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl}imide Olga Russina,† Mario Beiner,‡ Catherine Pappas,§ Margarita Russina,§ Valeria Arrighi,| Tobias Unruh,⊥ Claire L. Mullan,# Christopher Hardacre,# and Alessandro Triolo*,†,§ Istituto per i Processi Chimico-Fisici, Consiglio Nazionale delle Ricerche, Salita Sperone, C. da Papardo, 98158 Faro Superiore, Messina, Italy, Department of Physics, Martin-Luther-UniVersity Halle-Wittenberg, D-06099 Halle, Germany, Helmholtz Zentrum Berlin fu¨r Materialen und Energien, SF1-BENSC, Glienicker Str. 100, D-14109 Berlin, Germany, School of Engineering and Physical Sciences, Chemistry, Heriot-Watt UniVersity, EH14 4AS Edinburgh, U.K., Technische UniVersitaet Muenchen, Forschungsneutronenquelle Heinz Maier-Leibnitz FRM II and Physik Department E13, Lichtenbergstr. 1, 85747 Garching, Germany, and QUILL Center, School of Chemistry & Chemical Engineering, Queen’s UniVersity, Belfast BT9 5AG, U.K. ReceiVed: January 5, 2009; ReVised Manuscript ReceiVed: May 4, 2009

We present results from complementary characterizations of the primary relaxation rate of a room temperature ionic liquid (RTIL), 1-hexyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl} imide, [C6mim][Tf2N], over a wide temperature range. This extensive data set is successfully merged with existing literature data for conductivity, viscosity, and NMR diffusion coefficients thus providing, for the case of RTILs, a unique description of the primary process relaxation map over more than 12 decades in relaxation rate and between 185 and 430 K. This unique data set allows a detailed characterization of the VTF parameters for the primary process, that are: B ) 890 K, T0 ) 155.2 K, leading to a fragility index m ) 71, corresponding to an intermediate fragility. For the first time neutron spin echo data from a fully deuteriated sample of RTIL at the two main interference peaks, Q ) 0.76 and 1.4 Å-1 are presented. At high temperature (T > 250 K), the collective structural relaxation rate follows the viscosity behavior; however at lower temperatures it deviates from the viscosity behavior, indicating the existence of a faster process. Introduction Room temperature ionic liquids (RTILs) are attracting great attention due to a number of important physicochemical properties such as negligible vapor pressure and tunability of bulk properties. RTILs are typically built up by an asymmetric, bulky organic cation and an inorganic anion. Such a chemical architecture generally leads to enhanced stability of the liquid state as compared with the solid state, and accordingly, the melting points and glass transition points of RTILs are lower than room temperature. Detailed knowledge of the morphology and dynamics at microscopic scales is of fundamental importance for fully exploiting the smart properties of these materials. Recently, a great deal of research effort has been undertaken in this direction and a wide range of experimental techniques have been employed.1,2 Here we focus on the structural and dynamical properties of 1-hexyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl}imide ([C6mim][Tf2N]), a representative RTIL (see Scheme 1). This material has been characterized by electrical conductivity, NMR diffusion coefficient, viscosity, and density measurements.3a-c * To whom correspondence should be addressed. E-mail: [email protected]. † Consiglio Nazionale delle Ricerche. ‡ Martin-Luther-University Halle-Wittenberg. § Helmholtz Zentrum Berlin fu¨r Materialen und Energien. | Heriot-Watt University. ⊥ Technische Universitaet Muenchen. # Queen’s University.

SCHEME 1: Schematic representation of 1-hexyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl} imide, [C6mim][Tf2N]

Furthermore, its thermodynamic properties were unravelled by adiabatic calorimetry.4,5 These studies indicate the presence of a glass transition at Tg ∼183-185 K and showed that the material is easily supercooled into the glassy state. However, it is also possible to observe up to three crystalline phases (in ref 5 one crystalline phase was observed) by controlling the crystallization conditions. From the structural perspective, the liquid state morphology of [Cnmim][Tf2N] salts has only been studied for the cases of n ) 1 (1,3-dimethyl-imidazolium) and n ) 2 (1-ethyl-3-methylimidazolium). Hardacre et al. investigated molten [C1mim][Tf2N] by neutron diffraction6 and highlighted the significantly lower degree of charge ordering as compared to analogous

10.1021/jp900142m CCC: $40.75  2009 American Chemical Society Published on Web 06/01/2009

8470

J. Phys. Chem. B, Vol. 113, No. 25, 2009

chloride and hexafluorophosphate salts.7,8 They attributed this to the asymmetry, flexibility and complex behavior of the [Tf2N]- anion. More recently, Takamuku and co-workers used X-ray diffraction to study the structure of liquid [C2mim][Tf2N].9 A number of studies have been undertaken to understand the dynamics and the diffusive processes of these salts. Watanabe and co-workers3a analyzed the density, electrical conductivity, and NMR-derived diffusion coefficients at temperatures above 263 K for [Cnmim][Tf2N] with n ) 1, 2, 4, 6, 8. More recently viscosity, density, speed of sound, and electric conductivity measurements were reported for [C6mim][Tf2N] by the groups of Magee3b and Goodwin.3c The electric conductivity has also been studied in detail by Leys and co-workers.10 The purpose of this contribution is to merge the information on the dynamical behavior of [C6mim][Tf2N] obtained from different experimental techniques. In particular we compare existing experimental data (viscosity, conductivity, NMR diffusion coefficient) with our results from shear measurements, dielectric spectroscopy, and neutron spin echo. These techniques probe the primary relaxation process (also called R-process) that is characteristic for the transition from the liquid to the glasslike behavior. The primary process relaxation rates found by these different techniques quite often follow the same VogelFulcher-Tammann (VFT)11 trend. Here we succeed in obtaining universal curves by scaling the rates with factors which account for the differences in the observables monitored by the different experimental techniques. We thus may describe the primary process in [C6mim][Tf2N] over more than 12 decades in time between 185 and 430 K covering a much wider temperature and relaxation rate range than previously accessed. The results give information on the characteristic parameters for the VFT trend in this glass-former and a reliable estimate for its fragility index. Moreover the NSE technique with its unique momentum transfer dependence provides additional information on the spatial aspects of the relaxation. Experimental Section. Materials and Synthesis. [C6mim][Tf2N], used for dielectric and mechanical measurements, was purchased from IOLITEC; its declared purity is 99% and was used without further purification. Before measurements the sample was kept for several days under dynamic vacuum at 50 °C, to reduce the moisture content. The glass transition of the sample was estimated from conventional DSC measurements using a Perkin-Elmer Pyris I, with a heating rate of 10 K/min the glass transition was estimated (Tg ) 187 K) as the value of the inflection point. A fully deuterated 1-hexyl(d13)-3-methyl(d3)-imidazolium(d1) bis{(trifluoromethyl)sulfonyl}imide sample for the neutron scattering experiments was synthesized starting from 1-methyld3-imidazole prepared from imidazole and methanol-d4 according to literature methods.12 Then, 1-hexyl(d13)-3-methyl(d3)imidazolium bromide was synthesized by the usual alkylation methods and 1-hexyl(d13)-3-methyl(d3)-imidazoliumbis{(trifluoromethyl)sulfonyl}imide was obtained by metathesis of the bromide, then stirred in D2O to obtain 10 g of 1-hexyl(d13)-3-methyl(d3)imidazolium(d1) bis{(trifluoromethyl)sulfonyl}imide. The different steps are described in detail below. 1-Methyl-d3-imidazole. Anhydrous ruthenium chloride (0.6 g, 2.4 mmol), tri(n-butyl)phosphate (2.3 g, 9.2 mmol), and imidazole (10.8 g, 0.16 mol) were dissolved in 1,4-dioxane (300 cm-3) in a stirred 500 cm-3 autoclave. After addition of methanol-d4 (20 g, 0.55 mol), the mixture was heated at 180 °C for 24 h under 38 bar pressure of nitrogen and then cooled to room temperature and the liquid decanted from the spent catalyst as well as the solvent and excess methanol were

Russina et al. removed under reduced pressure. The resulting pale yellow liquid was vacuum distilled at 80 °C to give 1-methyl-d3imidazole as a colorless liquid (6.0 g, ∼44% yield) and analyzed by GC-MS. 1-Hexyl(d13)-3-methyl(d3)-imidazolium bromide. 1-Methyld3-imidazole (1.9 g, 0.02 mol) and bromohexane-d13 (4.0 g, 0.02 mol) were combined in acetonitrile and heated with stirring at 70 °C in a round-bottomed flask sealed with a Safe-Lab pressure seal. After heating for 24 h, the solvent was removed and the product was dried under reduced pressure. 1-Hexyl(d13)-3-methyl(d3)-imidazolium(d1) bis[(trifluoromethyl)sulfonyl]imide. The bis{(trifluoromethyl)sulfonyl}imide ionic liquid was prepared from 1-hexyl(d13)-3-methyl(d3)imidazolium bromide (6.0 g, 0.02 mol) and lithium bis{(trifluoromethyl)sulfonyl}imide (6.5 g, 0.02 mol), in D2O. The mixture cleared to two transparent liquid phases on heating and the lower phase, the ionic liquid, was collected. D2O was added and the mixture left to stir for a few days and then dried under reduced pressure and finally in vacuo with heating to 80 °C to yield 1-hexyl(d13)-3-methyl(d3)-imidazolium(d1) bis{(trifluoromethyl)sulfonyl}imide. 1 H NMR shows no residual hydrogen peaks and the 13C NMR was consistent with the product obtained. Mass spectroscopy gave a RMM of 184. Analysis by 1H and 2H NMR, and elemental analysis showed >97% isotopic exchange. Water content was determined by Karl Fischer analysis and was found to be 2.2 mol % water. Steady Shear and Dynamic Mechanical Measurements. Different types of shear experiments were performed using a Rheometrics RDA II instrument. (i) The temperature dependent viscosity η(T) was determined from isothermal steady shear measurements at shear rates dγ/dt between 10 and 100 s-1 and temperatures between 263 and 203 K. For these measurements parallel plates with 25 mm diameter were used. The sample thickness was about 1 mm. (ii) Dynamic mechanical experiments at 0.3, 1, 3, 10, 30, and 100 rad s-1 were performed at lower temperatures. Isochrones are constructed from measurements between 193 and 123 K with a temperature step of 2.5 K. Data were collected after 60 s of isothermal annealing at each temperature. (iii) Isothermal frequency sweeps in the range from 0.1 to 100 rad s-1 with 5 points per frequency decade were performed after 600 s of isothermal annealing at nine temperatures between 193 and 173 K. In cases (ii) and (iii) the strain was 0.1-0.2% and a fiber-like sample geometry was used. The samples were drawn between two 8 mm plates at 193 K from a highly viscous sample with an original thickness of about 2 mm extended to ∼10 mm in length which leads to concave fibers. A sample specific correction factor was calculated from two shear measurements before and after drawing the fiber at 193 K. This correction factor was used to calculate the shear modulus from raw data measured at all lower temperatures. This method results in relatively large uncertainties of up to 50% for the absolute values of the shear modulus |G*|. However, this has no consequences for the relaxation times and shift factors determined from these data. Dielectric Spectroscopy Measurements. Dielectric spectroscopy data were collected between 183-223 K in the range 2 × 10-2 to 3 × 106 Hz, applying the gain-phase analysis technique with a Solartron SI-1260 analyzer and a Novo-Control BDS 4000 spectrometer. The sample was sandwiched between two gold plated flat electrodes, with diameter of 20 mm and with a sample thickness of 50 µm. The temperature was controlled by a Quatro cryosystem.

Primary Relaxation in [C6mim][Tf2N]

Figure 1. Frequency dependence of the imaginary part of the dielectric modulus, M′′ of 1-hexyl-3-methylimidazolium bis{(trifluoromethyl)sulfonyl}imide, [C6mim][Tf2N], at selected temperatures in the range between 183 and 227 K (in steps of 1 K for 183 e T (K) e 193 and in steps of 2 K for 195 e T (K) e 227). In the inset the frequency/ temperature superposition of the dielectric modulus data is presented, leading to a master curve.

Neutron Scattering Measurements. Neutron spin echo (NSE) measurements were performed at the wide angle NSE spectrometer SPAN of the Helmholtz Zentrum Berlin. The sample was fully deuterated in order to enhance the coherent scattering and reduce as much as possible the incoherent contribution of hydrogen. The sample was in a aluminum annular cylindrical cell, with a sample thickness of 0.15 cm. The incident neutron wavelength was 6.5 Å and the NSE spectra were recorded at a momentum transfer of 1.4 Å-1 in the range 5-300 K with an Orange Cryostat. The diffraction pattern of the sample was also obtained, by integrating the S(2θ, time-of-flight) data at 4 K over all of the time-of-flight channels, t, at a given detector at an angle θ. Measurements of the dynamic structure factor S(2θ, t) were collected at the neutron time-of-flight spectrometer TOFTOF at the Forschungsneutronenquelle (FRM II) Heinz Maier-Leibnitz, Munich, with a wavelength of 6.0 Å.13 The TOF data were corrected for detector efficiency, empty cell contribution and were normalized against vanadium. The momentum transfer dependence was obtained considering that Q ) 4π/λ sin(θ) for elastic scattering, thus leading to the diffraction pattern, I(Q). Results and Discussion Figure 1 shows the frequency dependence of the imaginary part of the dielectric modulus, M′′, for selected temperatures over the range 183-223 K, where the main peak is related to a characteristic rate for the charge transport process. The characteristic time is proportional to the inverse of the frequency, where the peak is centered.14 For the sake of comparison with other spectroscopic techniques that access the primary R relaxation process, these frequency values are brought on a master curve after normalization by the so-called timetemperature shift factor, aM′′, The master curve for log M” was constructed from the individual isotherms by shifting them horizontally until they superimpose perfectly. The chosen reference temperature was 193 K. The resulting master curve is shown in the inset of Figure 1. In general, the shift factors from different experimental techniques follow the same temperature dependence although the R relaxation process may appear under isothermal conditions at slightly different frequency positions for different susceptibilities.15 Figure 2 displays

J. Phys. Chem. B, Vol. 113, No. 25, 2009 8471

Figure 2. Primary process relaxation map for 1-hexyl-3-methylimidazolium bis{(trifluoromethyl) sulfonyl}imide, [C6mim][Tf2N]. Viscosity data sets from Watanabe [red open square; ref 3a], Widegreen [black filled square; ref 3b], Goodwin [black open square; ref 3c] and our present mechanical spectroscopy measurements [blue filled circle]. Conductivity data sets from Watanabe [blue filled triangle; ref 3a], Goodwin [black open triangle; ref 3c], Leys [s; ref 10] and our present measurements. NMR self-diffusion coefficients from Watanabe [green filled triangle; ref 3a]. The characteristic times of the Arrhenius-like activated relaxation process observed with dielectric spectroscopy are also plotted and the (extrapolated) linear temperature dependence is indicated with the blue line to highlight a possible rationalization of the observed deviations of the neutron spin echo relaxation times [red open circle] from the viscosity trend. The other continuous lines refer to the Vogel-Tammann-Fulcher (VTF) fit of the Leys data set (black) and to the VTF fit of the whole data set obtained by merging the literature data together with our present data sets (red), thus covering an extended temperature range. In the inset a magnification of the high temperature portion of the whole data sets is presented.

the temperature dependence of aM′′, which is in excellent agreement with the electrical conductivity measurements by Leys and co-workers10 as well as the electrical conductivity, viscosity and NMR diffusion coefficient results from the reports from the groups of Watanabe,3a Goodwin,3b and Widegren.3c Our data expand the existing range of dynamic information on this RTIL.3a-c,10 Upon decreasing the temperature, another relaxation process enters into the probed dynamic window, as indicated by the existence of a second peak in the M′′ plot (Figure 1). A characteristic time for this process has been evaluated as given by the inverse of the peak position. The temperature dependence of such a characteristic time for the secondary process is Arrhenius-like as seen by the linear trend highlighted in Figure 2. This indicates the occurrence of more localized motions (β) as often found in glass-forming liquids. Our dynamic mechanical and viscosity measurements characterize the softening behavior of [C6mim][Tf2N] more in detail and over a wider temperature range than previous studies. Figure 3 shows isochronal data for the shear loss modulus G′′(T) for a wide temperature range. Apart from the pronounced primary (R) relaxation peak at high temperatures, a very broad additional peak is seen at significantly lower temperatures in the range 173-123 K. This finding supports the existence of more localized motions contributing to a secondary relaxation in [C6mim][Tf2N]. As expected the peak maximum shifts to higher temperatures for both the processes (R and β) when the measurement frequency increases. Their temperature dependence, however, is significantly different. The temperature dependence of the R process is clearly non-Arrhenius-like while the β process shows an Arrhenius-like behavior. Additional information on the temperature dependence of the R relaxation is obtained by applying the time temperature

8472

J. Phys. Chem. B, Vol. 113, No. 25, 2009

Figure 3. Isochronal shear loss modulus, G′′(T), between 125 and 193 K for 1-hexyl-3-methylimidazolium bis{(trifluoromethyl) sulfonyl}imide, [C6mim][Tf2N] (open circles, 10 rad/s; open squares, 100 rad/s). In the inset, the master curves for log G′ and log G′′ constructed from nine isothermal frequency sweeps measured at temperatures ranging from 173 to 193 K is presented. Reference temperature for the master curve is 193 K.

superposition principle to dynamic mechanical data. Nine individual isotherms for log G′ and log G′′ were horizontally shifted against each other until they superimpose optimally using the software package RSI Orchestror 6.5. The resulting master curve is shown in the inset of Figure 3. The chosen reference temperature is 193 K. The master curve is that of a typical glassforming liquid with a transition from the elastic behavior in the glass plateau at high frequencies with a storage modulus approaching 1 GPa to a liquid-like behavior at low frequencies with typical slopes d log G′′/d log ω and d log G′′/d log ω of 2 and 1, respectively. The additional wing in the G′′ curve at high frequencies indicates the existence of a pronounced secondary relaxation process. The obtained shift factors fit quite well with temperature-dependent R relaxation times τR (T) determined based on the isochrones as shown in Figure 3, with η(T) values obtained from additional viscosity measurements as well as η(T) and conductivity σ(T) data for [C6mim][Tf2N] reported in the literature (Figure 2). As a further step we made the very first NSE measurements on an ionic liquid. In order to maximize the coherent contribution and minimize incoherent scattering, we used a fully deuterated sample of 1-hexyl-3-methylimidazolium bis{(trifluoromethyl) sulfonyl}imide. These measurements directly access the collective relaxation at selected momentum transfer values. The NSE spectra and were fitted by a stretched exponential (Kohlrausch-Williams-Watts (KWW)16) function: φ(t) ) A exp(-t/τR)βKWW, where A reflects the Debye-Waller factor, τR is the characteristic time for the observed primary relaxation and βKWW (0 < βKWW e 1) is a stretching factor, accounting for deviations from a purely Debye relaxation (βKWW ) 1). The diffraction pattern, I(Q), as measured at the TOFTOF spectrometer for the fully deuteriated [C6mim] [Tf2N], is plotted in the inset of Figure 4. We note that I(Q) shows three peaks. The two high Q peaks at 0.76 and 1.4 Å-1 are related to intermolecular first neighbor spatial correlations. A third peak appears at Q ) 0.4 Å-1, where generally molecular fluids show featureless diffraction patterns. In the present ionic liquid, I(Q) bears the signature of structural heterogeneities at the nanoscale range which we detected in similar RTILs.17a-d Figure 4 shows the NSE spectra at Q ) 1.4 Å-1, which is the main interference peak and corresponds to first neighbor

Russina et al.

Figure 4. Dynamic structure factor at the main peak position (Q ) 1.4Å-1)forfullydeuterated1-hexyl-3-methylimidazoliumbis{(trifluoromethyl) sulfonyl}imide, [C6mim][Tf2N], at selected temperatures between 220 and 300 K (every 10 K). The continuous lines represent the fit of the experimental data in terms of the Kohlrausch-Williams-Watts (KWW) law. For the case T ) 270 K, representative error bars are indicated. In the inset the neutron diffraction pattern of fully deuterated 1-hexyl3-methylimidazoliumbis{(trifluoromethyl)sulfonyl}imide,[C6mim][Tf2N], at 4 K is presented.

Figure 5. Dynamic structure factor at the low Q peak position (Q ) 0.76 Å-1) for fully deuterated 1-hexyl-3-methylimidazolium bis{(trifluoromethyl) sulfonyl}imide, [C6mim][Tf2N], at selected temperatures between 270 and 330 K. The continuous lines represent the fit of the experimental data in terms of the Kohlrausch-Williams-Watts (KWW) law. Representative error bars are shown for data at 320 K.

spatial correlations. The spectra were collected every 10 K between 220 and 300 K and were fitted with βKWW fixed to 0.52 for all the temperatures (after preliminary fittings that led to values for this parameter statistically fluctuating around the chosen value). The corresponding characteristic times τR are included in the form ωR ) 1/τR in Figure 2. In Figure 5, we also show the NSE spectra collected at Q ) 0.76 Å-1, the lower peak of S(Q), collected at selected temperatures in the range between 270 and 320 K. These data were fitted by a KWW law with βKWW ) 0.5 (after preliminary fittings that led to values for this parameter statistically fluctuating around the chosen value) and relaxation times that, apart from a scaling factor, follow the same temperature dependence observed at Q ) 1.4 Å-1. The data reported in Figure 2 represent temperature shift factors; such a scaling procedure is standard, when dealing with characteristic rates of the same relaxation processes probed through different experimental techniques. Upon appropriate scaling, the whole set of data show an excellent superposition on a temperature trend that reflects the relaxation map for the

Primary Relaxation in [C6mim][Tf2N]

J. Phys. Chem. B, Vol. 113, No. 25, 2009 8473

primary process in [C6mim][Tf2N] over more than twelve decades in time and between 185 and 430 K. The conductivity data reported previously by Leys et al. (the data set that covers the largest temperature window among the ones available in the literature) accessed temperatures as low as 200 K and were fitted to a Vogel-Fulcher-Tammann (VFT) function:

τ ) τ0 exp[B/(T - T0)]

Conclusions

(1)

where τ0 is the high temperature limit of the R relaxation time, B the curvature and T0 is the Vogel temperature. From these parameters, Leys et al. determined the fragility parameter m according to

m ) B/2.303Tg /(Tg-T0)2

usually to an independent dynamics of the methylene units. As proposed by Rivera et al.20 after comparing with our previous QENS data on a similar system,23 the observed Arrhenius-like relaxation process might reflect conformational rearrangements of the highly flexible alkyl groups in the alkylimidazolium cation which are still observable in the glassy material below Tg.

(2)

with m ) 57, which is substantially different from the values found for other similar salts based on [BF4]- anions.10 The present combined data set covers an extended temperature range and the fit to a VFT function leads to more reliable parameters: B ) 890 K, T0 ) 155.2 K. We deduce m ) 71, which is well in the range of values reported by Leys et al. for other salts and is a intermediate value in the classification of glass strength as defined by Angell.18 We note that only in the high temperature regime (250 < T (K) < 300) the characteristic relaxation times for the collective structural relaxation obtained from NSE (both for Q ) 0.76 and 1.4 Å-1) follow the temperature dependence of the primary process seen by other methods. Similar behavior is often found in glass forming materials, both inorganic19a and polymeric.19b,c Below 250 K, the NSE characteristic times deviate systematically from the primary relaxation master curve. In this temperature range the NSE relaxation rate is significantly faster than the one of the primary relaxation process (Figure 2). A comparison of the NSE spectra at T < 250 K with model curves obtained considering relaxation rates following the primary process master curve shows differences of several orders of magnitude. We attribute this discrepancy to the appearance of an additional relaxation process which is faster than the primary one and appears in the dynamic window accessible by NSE. This hypothesis is strongly supported by the existence of an additional Arrhenius-like activated process at low temperature that has been observed in the isochronal G′′ data (Figure 3) as well as in isothermal M′′ data dielectric spectroscopy data (Figure 1). Note, that a similar relaxation process has also been observed in 1-butyl,3-methyl imidazolium bis{(trifluoromethyl) sulfonyl}imide ([C4mim][Tf2N]) by Rivera et al.20 in the same frequency temperature range. These dielectric data are shown for comparison in Figure 2 and fit quite well the temperature dependence of our NSE data. Obviously, the relaxation rates probed by NSE nicely extrapolate this Arrhenius-like process toward high temperatures, i.e. this process is not only dielectrically active but also detectable by NSE. It is known from dielectric measurements under pressure that the Arrhenius-like process in ([C4mim][Tf2N]) does not show any relevant pressure dependence.21 Although the nature of the Arrhenius-like relaxation process in ([C4mim][Tf2N]) needs further investigation there seem to be good reasons to assume that this process can be understood as a consequence of structural peculiarities on the nanoscale appearing in this and related RTILs [unpublished data] as well as in other systems containing long alkyl groups22 which lead

We reported on complementary experimental characterizations of the primary relaxation rate for a representative room temperature ionic liquid, [C6mim][Tf2N]. Literature data for electrical conductivity, viscosity and NMR diffusion coefficients were merged with our electrical conductivity, viscosity, shear and NSE data, providing a detailed description of the primary relaxation process over more than twelve decades in time and for temperatures between 185 and 430 K. This large data set allows a detailed characterization of the VTF parameters, B ) 890 K and T0 ) 155.2 K, which corresponds to a steepness index of m ) 71, indicating intermediate fragility. The collective structural relaxation rate deduced from the NSE spectra of a fully deuteriated sample at Q ) 0.76 and 1.4 Å-1 nicely follows the primary relaxation behavior for high temperatures (T > 250 K), similarly to the behavior of other glass forming liquids. Toward lower temperatures but still above the glass transition at Tg ) 187 K, the relaxation rate deviates significantly from the curve for the primary relaxation process. This highlights the existence of an additional, faster relaxation process, which is also seen in our shear and dielectric measurements. Further investigations are in progress to understand the nature of this fast Arrhenius-like relaxation process and to better access the momentum transfer dependence of the collective structural relaxation. Acknowledgment. This research project has been supported by the European Commission under the sixth Framework Programme through the Key Action: Strengthening the European Research Area, Research Infrastructures. Contract no.: RII3CT-2003-505925 (NMI3), A.T. acknowledges the hospitality of the Helmholtz Zentrum Berlin, during the period of this study. M.B. thanks Polymenakos Panagiotis for shear measurements performed in the framework of a student lab project and acknowledges financial support by the German Science Foundation (SFB 418). QUB gratefully acknowledges the EPSRC under the Portfolio Partnership Scheme, Grant EP/D029538/1 for funding. References and Notes (1) Rogers, R. D.; Voth, G. A. , Acc. Chem. Res. 2007, 40, 1077, (and references therein: special issue dedicated to RTILs). (2) Wishart, J. F.; Castner, E. W. J. Phys. Chem. B 2007, 111, 4639, and references therein: special issue dedicated to RTILs. (3) (a) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, A. B. H.; Watanabe, M. J. Phys. Chem. B 2005, 109, 6103. (b) Widegren, J. A.; Magee, J. W. J. Chem. Eng. Data 2007, 52, 2331. (c) Kandil, M. E.; Marsh, K. N.; Goodwin, A. R. H. J. Chem. Eng. Data 2007, 52, 23825. (4) Blokhin, A. V.; Paulechka, Y. U.; Kabo, G. J. J. Chem. Eng. Data 2006, 51, 1377. (5) Shimizu, Y.; Ohte, Y.; Yamamura, Y.; Saito, K.; Atake, T. J. Phys. Chem. B 2006, 110, 13970. (6) Deetlefs, M.; Hardacre, C.; Nieuwenhuyzen, M.; Padua, A. A. H.; Sheppard, O.; Soper, A. K. J. Phys. Chem. B 2006, 110, 120555. (7) Hardacre, C.; Holbrey, J. D.; McMath, S. E. J.; Bowron, D. T.; Soper, A. K. J. Chem. Phys. 2003, 118, 2735. (8) Hardacre, C.; McMath, S. E. J.; Nieuwenhuyzen, M.; Bowron, D. T.; Soper, A. K. J. Phys. Cond. Matter 2003, 15, S1595. (9) Fujii, K.; Soejima, Y.; Kyoshoin, Y.; Fukuda, S.; Kanzaki, R.; Umebayashi, Y.; Yamaguchi, T.; Ishiguro, S. I.; Takamuku, T. J. Phys. Chem. B 2008, 112, 43295.

8474

J. Phys. Chem. B, Vol. 113, No. 25, 2009

(10) Leys, J.; Wubbenhorst, M.; Menon, C. P.; Rajesh, R.; Thoen, J.; Glorieux, C.; Nockemann, P.; Thijs, B.; Binnemans, K.; Longuemart, S. J. Chem. Phys. 2008, 128, 64509. (11) (a) Vogel, H. Phys. Z. 1921, 22, 645. (b) Fulcher, G. S. J. Am. Chem. Soc. 1925, 8, 339. (c) Tamman, G.; Hesse, W. Z. Anorg. Allg. Chem. 1926, 156, 245. (12) Hardacre, C.; Holbrey, J. D.; McMath, S. E. J. Chem. Commun. 2001, 3675. (13) Unruh, T.; Neuhaus, J.; Petry, W. Neutron Instr. Methods A 2007, 580, 14145. (14) Macedo, P. B.; Moynihan, C. T.; Bose, R. Phys. Chem. Glasses 1972, 13, 1715. (15) Donth, E.; Beiner, M.; Reissig, S.; Korus, J.; Garwe, F.; Vieweg, S.; Kahle, S.; Hempel, E.; Schro¨ter, K. Macromolecules 1996, 29, 6589. (16) Kohlrausch, R. Ann. Phys., Lpz 1847, 12, 393. (b) Williams, G.; Watts, D. C. Trans. Faraday Soc 1970, 66, 80. (17) (a) Triolo, A.; Russina, O.; Bleif, H.-J.; Di Cola, E. J. Phys. Chem. B 2007, 111, 4641. (b) Triolo, A.; Russina, O.; Fazio, B.; Triolo, R.; Di

Russina et al. Cola, E. Chem. Phys. Lett. 2008, 457, 362. (c) Xiao, D.; Hines, L. G.; Li, S.; Bartsch, R. A.; Quitevis, E. L.; Russina, O.; Triolo, A. J. Phys. Chem. B 2009, 113, 6426. (d) Triolo, A.; Russina, O.; Fazio, B.; Appetecchi, G. B.; Carewska, M.; Passerini, S. J. Chem. Phys. 2009, 130, 164521. (18) Angell, C. A. Science 1995, 267, 19245. (19) (a) Mezei, F.; Knaak, W.; Farago, B. Phys. ReV. Lett. 1987, 58, 571. (b) Richter, D.; Arbe, A.; Colmenero, J.; Monkenbusch, M.; Farago, B.; Faust, R. Macromolecules 1998, 31, 1133. (c) Arrighi, V.; Pappas, C.; Triolo, A.; Pouget, S. Physica B 2001, 301, 1575. (20) Rivera, A.; Rossler, E. A. Phys. ReV. B 2006, 73, 2122015. (21) Rivera-Calzada, A.; Kaminski, K.; Leon, C.; Paluch, M. J. Phys. Chem. B 2008, 112, 31105. (22) Beiner, M.; Huth, H. Nat. Mater. 2003, 2, 595. (23) Triolo, A.; Mandanici, A.; Russina, O.; Rodriguez-Mora, V.; Cutroni, M.; Hardacre, C.; Nieuwenhuyzen, M.; Bleif, H. J.; Keller, L.; Ramos, M. A. J. Phys. Chem. B 2006, 110, 21357.

JP900142M