Low-Frequency Spectra of Metallocenium Ionic Liquids Studied by

Feb 4, 2011 - Takashi Inagaki,. ‡. Motohiro Banno,. †. Tomoyuki Mochida,. ‡ and Keisuke Tominaga*. ,†,‡. †. Molecular Photoscience Researc...
0 downloads 0 Views 888KB Size
ARTICLE pubs.acs.org/JPCA

Low-Frequency Spectra of Metallocenium Ionic Liquids Studied by Terahertz Time-Domain Spectroscopy Anjan Chakraborty,† Takashi Inagaki,‡ Motohiro Banno,† Tomoyuki Mochida,‡ and Keisuke Tominaga*,†,‡ † ‡

Molecular Photoscience Research Center, Kobe University, Kobe, 657-8501 Japan Department of Chemistry, Graduate School of Science, Kobe University, Nada, Kobe, 657-8501 Japan

bS Supporting Information ABSTRACT: Terahertz (THz) time-domain spectroscopic measurements have been done on five novel metallocenium ionic liquids based on the electro-optic sampling method. The study covered the spectral range from 10 to 85 cm-1. The complex dielectric spectra were broad and dispersive in nature, and the imaginary part of the dielectric constant consisting of part of the dielectric constant was simulated with different combinations of model functions to unravel the intermolecular dynamics. We compared our results with the previous results on the other ionic liquid. It was revealed that the librational motion of the cations as well as the interion vibration between the cations and the anions are responsible for observed dynamics in THz region. No intramolecular vibrational mode has been found in the low-frequency region.

’ INTRODUCTION Ionic liquids (ILs) have been extensively used as green substitutes for volatile solvents for the past few years.1 ILs are organic salts composed of anions and cations and remain in the liquid state at ambient conditions. Because of various unique chemical and physical properties such as thermal stability, low melting temperature, and low vapor pressure, ILs have been widely used for organic synthesis, electrochemical applications, solar batteries, biopolymers, and many more applications.2-15 However, because of their relative novelty, studies of their physical properties, particularly intermolecular dynamics, are still lagging behind. Femtosecond optical heterodyne detected Raman induced Kerr effect spectroscopy (OHD-RIKES) and terahertz timedomain spectroscopy (THz-TDS) are such kinds of techniques which enable us to investigate different intermolecular interactions in the liquid state.16-29 In recent years, studies have been extensively carried out on several liquids to reveal their intermolecular interactions. Castner and Chang studied the solvation dynamics and low-frequency Raman modes using the OHDRIKES technique in conventional liquids.21 Lotshaw and coworkers have undergone detailed comparative studies between OHD-RIKES22 and THz-TDS17 studies of several conventional liquids. Giraud and Wynne reported the vibrational studies of different conventional solvents studied by IR and OHDRIKES.23 In most of these studies the spectral line shapes obtained by Fourier transformation of the time-domain results were analyzed with model functions such as Lorentzian and Gaussian functions. The fitting results revealed different intraand intermolecular modes in the solvents. Schmuttenmaer and co-workers studied the dielectric properties of several conventional liquids using THz-TDS.24-26 The spectra were fitted with a multi-Debye relaxation model. Recently, Dutta and Tominaga r 2011 American Chemical Society

performed THz-TDS studies on solutions of a polar solute in a nonpolar solvent such as nitrobenzene in alkanes27,28 and acetone in cyclohexane.29 Since ILs are completely different in many properties from the conventional solvents, it is of great interest to investigate the intermolecular interactions in ILs by spectroscopic methods on the low-frequency region. Recently, several groups reported OHD-RIKES and THz-TDS studies on different ILs.30-49 Quitevis and co-workers reported OHD-RIKES studies of a series of alkyl methyl imidazolium ILs as Raman active components.30 The imaginary part of the dielectric susceptibility (Im D(ω)) was monitored as a function of alkyl chain length and fitted with linear combinations of antisymmetrized Gaussian functions and a function corresponding to so-called collisioninduced component. The authors observed different intermolecular modes which depend on the chain length of imidazolium cations. The picosecond slow dynamics was directly dependent on viscosity. Wynne and co-workers reported the OHD-RIKES studies on a series of imidazolium salts using both the cations and the anions.31 The imaginary part of dielectric susceptibility was fitted with five Lorentzian functions and found that the librational modes of the imidazolium rings have characteristic frequencies at around 30 cm-1, 65 cm-1, and 100 cm-1, which arise due to local configuration of the anions with respect to the cations. Turton et al. reported the intermolecular dynamics 1,3-dialkylimidazolium based IL with different anions using OHDRIKES, and dielectric relaxation study.32 It was found that the librational band is overlapped with the intermolecular vibrational Received: August 26, 2010 Revised: December 6, 2010 Published: February 4, 2011 1313

dx.doi.org/10.1021/jp108097q | J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A band and this contributes to the higher frequency component and the low frequency relaxation is in agreement with the simulation. Recently, Castner's group and Shirota's group reported extensive OHD-RIKES studies on different kinds of ILs.33-39 Intra- and intermolecular vibrational line shapes were interpreted by the multimode Brownian oscillator model for pyrrolidinium ILs, and Ohmic and antisymmetrized functions were used to fit the data of silyl- and siloxy substituted imidazolium ILs.33,34 These authors also studied the alkyl substituted imidazolium ILs and compared the properties with its silyl analogue.35 It was found in their study that the frequency obtained by fitting of Im D(ω) is consistent with weaker ion-ion interaction in the silyl derivative. They also studied the IL 1-methoxyethylpyridinium dicyanoamide and compared the results with those of the analogous isoelectronic binary solution, comprised of equal parts of 1-methoxyethylbenzene and dicyanomethane.36 Recently, Shirota et al. reported vibrational dynamics of a 1-butyl3-methyl imidazolium cation based IL with anions [PF6]-, [AsF6]-, and [SbF6]-.37 They found that heavy atom substitution leads to the weaker force constant, and hence a lowering in frequency is observed.37 Fujisawa et al. studied the intermolecular vibrational dynamics in ILs and concentrated electrolyte solution and compared the result between the IL and that of electrolyte solution.38 Weingartner et al. studied the dielectric response of imidazolioum based ILs and observed the effect of cation variation on the frequency-dependent dielectric permittivity.40 Fayer and co-workers reported studies of imidazolium ILs.41,42 They have studied dynamics beyond one picosecond at various temperatures in the supercooled region. According to their study the slowest component can be welldescribed by the hydrodynamic Stokes-Einstein Debye relation, while in the first few hundred picoseconds the relaxation was better described by a dual power law decay function. Quitevis and co-workers also studied the intermolecular dynamics of imidazolium ILs at different temperatures by OHD-RIKES.43 Xiao et al. reported the nanostructural organization and the effect of anions and temperature on the ILs and binary ionic liquid mixture.44-46 Recently, Quitevis and co-workers studied the OKE spectrum of CS2 in the imidazolium IL to understand tail aggregation in ILs.47 Very recently, Xiao et al. studied the effect of “cation symmetry” on the intermolecular dynamics of imidazolium IL.48,49 While Raman active low-frequency modes of ILs have been often studied by OHD-RIKES, THz-TDS has been rarely used to investigate the intermolecular dynamics in ILs. Recently, Koeberg et al. reported a THz-TDS study in the mixtures of imidazolium ILs and water.50 The frequency-dependent real and imaginary parts of the dielectric constant were fitted with a doubleDebye model. Very recently, Yamamoto et al. published the experimental results of THz-TDS measurements on ILs having an imidazolium cation and different anions.51 They concluded that intermolecular dynamics are more dominant than intramolecular dynamics in the THz region. However, this study did not reveal any intermolecular mode conclusively. It should be noted that most of the studies involve imidazolium, pyrrolidinium, or such kinds of moieties in their cationic part.31-39 Therefore, to clarify the intermolecular dynamics, it is inevitable to use a new kind of ILs which is different from ILs studied so far in the OHDRIKES and THz-TDS experiments. The study of these new ILs will enable us to have a better insight regarding the intermolecular dynamics. Inagaki and Mochida have recently developed unconventional ILs containing metallocenium cations.52 These materials can be

ARTICLE

Scheme 1. Structure of the Ionic Liquids

regarded as functional liquids containing metal ions, which exhibit unusual physical properties. The metallocenium ILs are characterized by their bulky molecular shape and charge localization inside the cation, which make a sharp contrast to imidazoliumbased conventional ILs. Therefore, the molecular motions as well as the intermolecular interactions in these liquids are of great interest. Since the low-frequency dynamics are governed by both the intramolecular and collective intermolecular motions, by varying the cation part with uitable metals and aromatic rings and by a varying anion, we aim to understand the exact dynamics playing a crucial role in the low-frequency region. The structures of the ILs are shown in Scheme 1.

’ EXPERIMENTAL SECTION Our THz-TDS apparatus is based on the electro-optic sampling (EO) method.53 In brief, a femtosecond pulse from a Ti: Sapphire laser was used to generate and detect the pulsed THz radiation. Most of the pulses are focused on to a (100) surface of an InAs wafer. The wafer is placed in a magnetic field with a flux of 1.6 T to enhance the intensity of the radiation. The rest of the pulses are used for detecting the electric field of the pulsed THz radiation with the EO sampling method by using a ZnTe crystal. The sample solutions were placed in a cell with two Si plates with a thickness of 3 mm. A path length was 0.5-0.2 mm. The reference signal was obtained from measurement on an empty cell. We first obtained frequency-dependent refractive index n and extinction coefficient κ of the sample from the analysis of the experimental results. Then, dielectric permittivity (ε = ε0 þ iε) is calculated from the relations ε0 = n2 - κ2 and ε00 = 2nκ. All of the measurements were done at room temperature (20 °C). It was ensured that the system covered the frequency range region from 10 to 85 cm-1. Beyond the 85 cm-1 frequency range, the signal-tonoise ratio became worse; hence, we do not accept data in this region. All of the ILs investigated in this work were synthesized by the procedure reported elsewhere.52 The ILs were dried under vacuum for 1 day at 80-90 °C to remove a trace of water which might be left in the liquids. During the experiment the sample does not absorb any moisture and was not exposed to air. ’ SPECTRAL ANALYSIS In OHD-RIKES and THz-TDS the low-frequency spectra have been often analyzed in terms of different types of functions 1314

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A

ARTICLE

such as Lorentzian and Gaussian functions, and so forth. Different linear combinations of these functions have been used to represent the experimental data. In the present case we used minimal number of functions for the fitting criteria. We modeled the low-frequency spectra of the ILs according to the procedure described by Lotshaw and co-workers.17 The linear combination of Lorentzian functions which correspond to the solution for the equation of a damped driven oscillator is used to fit the imaginary part of the dielectric constant. In the time domain, the response function is as17 follows RðtÞ ¼ Gj exp½ - ðγj - βj Þt½1 - expð - 2βj tÞ

ð1Þ

where γi is the damping coefficient, βj = (γj2 - ωj2)1/2 is the angular velocity of the jth damped oscillator whose undamped frequency of the oscillator is ωj. In the case of overdamped oscillators, γj > ωj. The Laplace-Fourier transform of the above equation turns into the following equation where the imaginary part of the complex permittivity is given by, ε00 ðωÞ ¼

2Gj ωγj ðγj 2 þ ω2 Þ - 2βj 2 ðγj 2 - ω2 Þ þ βj 4

ð2Þ

The Lorenzian functions are used to define the homogeneously broadened vibrational band in the Brownian oscillator analysis of the vibrational spectra. The overdamped Lorentzian function is used to represent the diffusive reorientational coordinate in the OHD-RIKES study. The other function that we used to fit the imaginary part of the dielectric constant is a linear combination of the generalized Ohmic function or Bucaro-Litoviz (BL) function and the antisymmetrized Gaussian function. The generalized Ohmic function has been used to describe the low-frequency dynamics in light scattering experiment of nonpolar liquids and expressed as follows, ε00 BL ðωÞ ¼ ABL ωR expð - ω=ωBL Þ

ð3Þ

Though in most cases R is either equal to one or greater than one, there are cases where R is less than one in the low-to-intermediate frequency range.17 The antisymmetrized Gaussian function is used to describe the higher frequency region and is expressed as follows: ε00 G ðωÞ ¼ AG fexp½ - ðω - ωG Þ2 =σ2  - exp½ - ðωþωG Þ2 =σ2 g

ð4Þ

where σ is related to a full width at the half-maximum. In the present study the Gaussian function is used to fit the high-frequency term to obtain the best quality of the fit. Generally, the Gaussian function is used to represent librational motion of the molecular liquid at higher-frequency region. In the case of aromatic molecular liquids the interionic dynamics which have characteristic components between 0 and 50 cm-1 originate from the interaction-induced motion, and this is coupled with the librational motion which is like a hindered motion. The librational motion takes place at the intermediate to the higher frequency region.54-57

’ RESULTS AND DISCUSSION The experimental results of the imaginary part of the dielectric constant of the metallocenium ILs are shown as red points in

Figure 1. One interesting observation is that the complex dielectric spectra of all of the ILs are broadly structured, suggesting contributions of several bands. The observation implies a high degree of association in the ILs. It was found in the work of Yamamoto and co-workers that for imidazolium ILs the complex dielectric spectra were very much broad and dispersive.51 The ε0 (ν) spectra were found to have structured features in shape which was composed of several components of dielectric response. This feature of imidazolium ILs is consistent with the present THz study on mettalocenium ILs. The structured dielectric spectra in the imidazolium ILs as well as those in the present case imply that there may be some local structures in the ILs. We first tried to fit the imaginary part of the dielectric constant with a multi-Debye model. It was found that for all of the ILs up to 50 cm-1 the imaginary part of the dielectric constant could be fitted with a double-Debye or triple-Debye model function. The time constants obtained from the fitting are around 2 ps, 0.2 ps, and 0.080 ps (data are not shown). The uncertainty of the fitting is relatively large, especially for the ultrafast component shorter than 100 fs. However, spectral data beyond 50 cm-1 could not be fitted with a multi-Debye function. The impossibility to fit data at the high-frequency region suggests inhomogeneity in the IL structure. The present result also implies that local structure may exist in ILs which leads to inhomogeneity. The same result was obtained in the previous study on ILs with an imidazolium cation, and the presence of some crystal structure in the ILs was mentioned there.51 Recently there are reports where the nanostructures in ILs have been invoked by tail aggregation of the cations.44-47 The difficulties in fitting the data to a multi-Debye model arise because of the band structured in the ILs that appeared at 40-50 cm-1. It was found in the case of imidazolium ILs that the ε0 (ν) part is very much similar to that of a short-chain alcohol but the imaginary part of dielectric spectra, that is, ε00 (ν), was strikingly different.25 The complex dielectric spectra of short chain alcohol can be fitted with a multi-Debye relaxation function. However, in the case of imidazolium ILs as well as in the present case the ε00 (ν) part of dielectric spectra cannot be reproduced at the high-frequency region. This is because the band structure appears in the ε00 (ν) spectra above 50 cm-1. This fact suggests that the ILs are inhomogeneous in structure. Since the polarizability measurement in the OHD-RIKES study (χ00 (ν)) directly corresponds to the dipole dynamics (ε00 (ν)) in THz-TDS, we may compare the present THz results with that obtained from the OHD-RIKES studies on different ILs. In the Spectral Analysis section we briefly described the fitting of the spectral data with a linear combination of Lorentzian functions; let us first discuss results obtained from the fitting with a sum of Lorentzian functions. We found that at least four Lorentzian functions are needed to reproduce the experimentally obtained spectrum from 10 to 85 cm-1 satisfactorily well. ε00 ð~ν Þ ¼

4 X

2Gj ~ν γj 2 ν Þ - 2βj 2 ðγj 2 j ¼ 1 ðγj þ ~ 2

- ~ν 2 Þ þ βj 4

ð5Þ

The frequency-domain data fitted with a sum of four Lorentzian functions are shown in Figure 1. The red points are the experimental data, and the black curve is the fitted graph. The obtained parameters are summarized in Table 1. We fitted the imaginary part of the dielectric constant with a sum of three antisymmetrized Gaussian functions and one BucaroLitoviz (BL) or collision-induced function. The results are 1315

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Fitting of an imaginary dielectric constant part with four Lorenztian functions. In all of the cases the red curve at low frequency corresponds to an over-damped oscillator, while the others correspond to under-damped oscillators.

summarized in the Supporting Information. It is revealed that in all of the ILs the antisymmetrized Gaussian functions and the BL function yield bands at four distinct regions. The BL function yields frequency in the region 15-20 cm-1 while the antisymmetrized Gaussian bands appear from the intermediate region to the high-frequency region. It is seen that the fitting with antisymmetrized Gaussian functions and the BL function qualitatively produces the similar results to that obtained by a model with four Lorentzian functions. Therefore, we discuss about the results by Lorentzian functions in detail. It is revealed from Figure 1 that fitting by Lorentzian functions yields four distinct bands in all of the ILs. In the four Lorentzian functions, one is for an overdamped oscillator, which is called band I. The others are for the underdamped oscillator, which are

denoted as band II, III, and IV from the low-frequency sides to high-frequency sides. The overdamped and underdamped oscillator model is often used in the study of OHD-RIKES for conventional solvents.17, 20-22 In the present study, the structures of ILs show that, except for IL1, the cation part is symmetrical in all of the ILs. Thus, it is expected that neutral species of the cations of these ILs, that is, symmetrical neutral metallocene molecules, will not have any permanent dipole moment. The study of pure metallocene compounds such as ferrocene [bis(cyclopentadienyl)iron] and ruthenocene [bis(cyclopentadienyl)ruthenium] reveals no characteristic vibrational mode in the region 0-90 cm-1.54 However, [bis(pentamethylcyclopentadienyl)iron] shows a weak band at around 54 cm-1. Unlike the intramolecular bands, this band is weak and appears at the lower 1316

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A

ARTICLE

Table 1. Parameters Obtained by Fitting with the Four Lorentzian Functionsa Gj %

~νj (cm-1)

γj (cm-1)

band I

26.1

20.7

26.1

band II band III

37.5 26.9

44.2 60.7

25.8 16.0

ILs IL1

band IV IL2

IL3

IL4

IL5

9.0

73.1

9.9

band I

16.6

17.1

22.6

band II

30.3

38.4

26.4

band III

30.0

54.5

20.9

band IV

22.9

68.1

16.2

band I

21.0

18.3

25.6

band II band III

34.0 26.4

42.0 56.3

30.1 21.5

band IV

18.6

68.7

15.6

band I

14.5

16.9

20.1

band II

26.6

35.2

24.3

band III

32.1

52.9

21.3

band IV

26.9

67.2

18.2

band I

14.9

17.6

24.1

band II band III

17.9 28.2

32.6 47.9

21.1 18.5

band IV

39.0

64.4

18.3

a

The error for band I is around (10% and for bands II, III, and IV is around 5%.

frequency side. This mode was suggested to be an intermolecular vibrational mode by Nishizawa and co-workers.54 The Raman spectral measurement shows that in metallocene compounds the symmetric stretching, antisymmetric stretching, and antisymmetric ring tilt modes have frequencies at 169 cm-1, 451 cm-1, and 515 cm-1, respectively.54 Again, the ring-metal-ring bending mode and out-of-plane CH3 oscillation take place at 140 and 200 cm-1, respectively. Now if we look at Table 1 and Figure 1, it is seen that, for IL1, IL2, IL3, IL4, and IL5, the bands III appear at 60.7 cm-1, 54.5 cm-1, 56.3 cm-1, 52.9 cm-1, and 47.9 cm-1, respectively. These modes are close to the intermolecular mode in [bis(pentamethylcyclopentadienyl)iron] and may come from the intermolecular vibration between the cations and the anions. Therefore, it is clear that bands II, III, and IV are not solely due to the cationic part. The intermolecular vibrations between the cations and the anions are also important. It is to be noted that the intermolecular interaction in ILs is stronger than those in any ordinary liquids. Thus, more numbers of functions are required to fit the spectral data in the low-frequency region, which indicates the inhomogeneous nature of the IL. In the study of ILs with an imidazolium cation the imaginary part of the dielectric constant was fitted with five Lorentzian functions.31 The lowest-frequency component corresponds to diffusive motion, and the other modes correspond to the intermolecular librational mode. In our case the damping coefficients for the underdampled oscillators are the same in all of the ILs. As the damping coefficient represents how a particular motion is coupled to the bath, similar damping constants in all of the ILs suggest that they undergo a similar type of motions. As we have already mentioned that intramolecular or interionic vibration takes place at a much higher frequency, therefore, in the present case these motions are the librational motions of the cations which involve anions also and collective interionic

motions between the cations and the anions. It should be mentioned here that the OHD-RIKES data of imidazolium ILs varying by chain length and anions were fitted with five Lorentzian functions.31 In this study it was assumed that the second, third, and fourth Lorentzian bands (30 cm-1, 65 cm-1, and 100 cm-1) have the origin of librational motions of imidazolium cations. In all of the imidazolium ILs it was found that the amplitude of the band increases from the low- to high-frequency sides, like 15%, 25%, and 60% for the second, third, and fourth Lorentzian bands, respectively. The authors assumed that these amplitudes of the second, third, and fourth Lorenztian bands correspond to the anion population which are preferentially located at three regions around the cation.31 Similarly, in the present case the THz-TDS data were fitted with four Lorentzian functions. The band I is the overdamped oscillator, and the remaining three are coming either from the librational motion of the cations which depend on the preferential location of the anions as the case of imidazolium ILs and/or intermolecular vibration between the cations and the anions. Let us now compare the results of different ILs studied in the present case. Both IL1 and IL2 possess the same anion ((CF3)2SO2N-), but the cations are structurally different. In IL1, a cation consists of an Fe atom which is attached to an ethyl cyclopentadienyl ring and an ethyl phenyl ring; on the other hand in IL2 the Fe atom is attached to two ethyl cyclopentadienyl rings. Therefore, any difference between the THz spectra of IL1 and IL2 may be assumed to be coming from the cation part and the arrangement of anions around the cations. It is revealed from Table 1 that for IL1 fitting of the imaginary part of the dielectric constant yields four bands at 20.7 cm-1, 44.2 cm-1, 60.7 cm-1, and 73.1 cm-1, and the corresponding amplitudes are 26.1%, 37.5%, 26.9%, and 9.0%, respectively. For IL2 the peaks appear at 17.1 cm-1, 38.4 cm-1, 54.5 cm-1, and 68.1 cm-1, and the amplitudes are 16.6%, 30.3%, 30.0%, and 22.9%, respectively. We already mentioned earlier that [bis(pentamethylcyclopentadienyl)iron] has a weak vibrational band at around 54 cm-1 and this mode was suggested to be an intermolecular vibration.54 In the present case we see that, for IL1, the bands II and the bands III have frequencies at 44.2 cm-1 and 60.7 cm-1, respectively, and for IL2 the bands II and III are located at 38.4 cm-1 and 54.5 cm-1, respectively. We, therefore, assume that these two bands are due to intermolecular vibrations which involve both the cations and the associated anions attached to the cations. As already mentioned, the intramolecular modes take place at a much higher frequency for the cation of these ILs. Therefore, the present mode may be considered as an intermolecular mode arising from the interionic vibration between cations and anions. Now the question arises about the origin of the band IV that appears at 73.1 cm-1 and 68.1 cm-1 for IL1 and IL2, respectively. In the present work, the cationic part of all of the ILs consists of a metal atom which is attached to either two cyclopentadienyl rings (IL2-IL5) or one cyclopentadienyl and one phenyl ring (IL1). During the librational motion of the cation these aromatic moieties may remain in-plane or out-of-plane. Now for pure aromatic moieties like benzene and toluene an OHD-RIKES study detects only out-of-plane librational motion, while in THzTDS both in-plane and out-of-plane motion is observed for aromatic molecules like benzene or toluene.55 Though the inplane motion contributes little to the overall dynamics, the alkyl chains of the aromatic rings are likely to affect the interionic dynamics when the in-plane motion takes place. Therefore, the THz spectra originating from the librational motion of the 1317

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A cations of ILs in the present case should be highly affected by the in-plane or out-of-plane of aromatic rings. This is also evident from the fact that the present imaginary part of the dielectric spectra has a great resemblance to that of toluene and benzene.55 The THz studies reveal that the dynamics in the THz region for benzene and toluene is mainly governed by both in-plane and out-of-plane librational motions. Therefore, it is clear that the band IV in all ILs is coming from the librational motion of the cation and depends on the spatial arrangement of anions. From the OHD-RIKES studies it has been established that in case of the aromatic ILs the component at the higher frequency range is attributed to the out-of-plane librational motion of the aromatic cation.31 Moreover, aromatic liquids such as benzene, toluene, ethyl benzene, and pyridines exhibit librational bands in the OHD-RIKES study.58,59 The molecular dynamics simulation has shown that the high frequency part in the intermolecular vibrational spectra of benzene and biphenyl arises from the tumbling motion of the aromatic rings.56,57 Therefore, it is not unworthy to consider that the fourth band originates from the librational motion of the cation in the present case. From Table 1, it can be seen that the damping coefficients for the four Lorentzian components of IL1 are close to the corresponding damping coefficients of IL2. This suggests that these two ILs undergo a similar kind of motion. Another interesting observation is that the amplitudes for the band III for both the IL1 and the IL2 are the same (26.9% and 30.0%), while of IL1 and IL2 the band IV has an amplitude of 9.0% and 22.9%, respectively. A large increase in the amplitude for the band IV on going from IL1 to IL2 also supports that the band IV has different dynamics from the band II and III. From Table 1 it is seen that if the Fe atom is replaced by the Co atom, the frequency of the Lorentzian modes is shifted to the higher frequencies. In IL2, the Lorentzian bands appear at 17.0 cm-1 (16.6%), 38.4 cm-1 (30.3%), 54.5 cm-1 (30.0%), and 68.1 cm-1 (22.9%), while in IL3 the same appear at 18.3 cm-1 (21.0%), 42.0 cm-1 (34.0%), 56.3 cm-1 (26.4%), and 68.7 cm-1 (18.6%). As the Fe atom is replaced by the Co atom, the ring-metal distance decreases. Hence, the total volume of the cation decreases. Therefore, intermolecular interaction between the cation and the associated anions becomes stronger in IL3 which brings the vibrational frequency to the higher wavenumber. From Table 1 it is revealed that, for IL2, the bands III and IV have amplitudes of 30.0% and 22.9%, while for IL3 the amplitudes of the bands III and IV are 26.4% and 18.5%, respectively. The higher amplitude for IL2 in the high frequency region implies that IL2 is more ordered in structure than IL3. The structure of IL2 and IL3 are almost identical in structure except in metal. Thus, it is not expected that the arrangement of the anions around these two cations may be very different. However, in the case of IL1 one cyclopentadienyl ring and one phenyl ring are attached to the Fe atom; due to the presence of the different rings the external environment for this cation, that is, the arrangement of the anions around this cation, the intermolecular motion may be different from IL2 or IL3. Now we compare the results between IL2 and IL4. The structural difference between the two ILs is that IL2 has an ethyl chain in cationic part while IL4 possesses a butyl chain in the cationic part. The Lorentzian bands for IL4 appear at 16.9 cm-1, 35.2 cm-1, 52.9 cm-1, and 67.2 cm-1, respectively, with the amplitude of 14.5%, 26.6%, 32.1%, and 26.9%, respectively. It is revealed from Table 1 that the Lorentzian bands for IL4 appear at the lower-frequency side compared to IL2. This observation is different from the OHD-RIKES study of alkyl imidazolium ILs

ARTICLE

where it was found that with an increase in the chain length the Lorentzian modes shifted to the higher wavenumbers.30 We assume that the arrangement of the anions around the cations may be a little different for IL2 and IL4. As IL4 is larger in size than IL2, the interionic interaction may be weaker for IL4, and thus the vibrational band is shifted at the lower frequency region. However, torsional motion of the alkyl chain length perhaps does not play any role in governing dynamics in this region. In IL2 the amplitudes of the band III (54.5 cm-1) and band IV (68.1 cm-1) are 30.0% and 22.9%, respectively, while in IL4 the amplitudes corresponding to these modes are 32.1% and 26.9%, respectively. The little increase in the percentage of the amplitude for IL4 compared to IL2 indicates that IL4 is highly ordered in structure compared to IL2. This further confirms our conjecture that the THz spectra appear from the librational motion of the cations and intermolecular motion between the cations and the anions. If we compare results between IL2 and IL5, several interesting results can be found. Both IL2 and IL5 possess the same cation but the different anions. For IL2 the anion is (CF3)2SO2N-, and for IL5 the anion is (C3F7)2SO2N-. The Lorentzian band frequency appears for IL5 at 17.6 cm-1, 32.6 cm-1, 47.9 cm-1, and 64.4 cm-1, respectively, with the amplitudes of 14.9%, 17.9%, 28.2%, and 39.0%, respectively. It is seen from Table 1 that, by replacing the (CF3)2SO2N- anion in IL2 with (C3F7)2SO2N-, the underdamped Lorentzian bands shifted to the lower frequency. This is not unexpected because of the fact that, as the mass and the volume of the anion increase, it will lead to a weaker force constant leading to lower frequency. This result is consistent with the observation by Shirota et al.37 Shirota et al. observed that with heavier atom substitution in anion the libration mode goes to a lower frequency. Another observation is that, in IL2, the amplitudes of band II to band IV decrease from lower- to higher-frequency sides. The amplitudes corresponding to the band II (38.4 cm-1), the band III (54.5 cm-1), and the band IV (68.1 cm-1) are 30.3%, 30.0%, and 22.9%. However, for IL5 the amplitudes of the Lorentzian band increases from the lower frequency to higher frequency. The amplitudes corresponding to band II (32.6 cm-1), the band III (47.9 cm-1), and band IV (64.4 cm-1) are 17.9%, 28.2%, and 39.0%, respectively. This indicates that IL5 is more ordered in structure than IL2. This also confirms us that the different Lorentzian bands are associated with the librational modes of cations and the intermolecular motion between the cations and the anions. In the case of aromatic molecular liquid, the low-frequency band below 50 cm-1 is ascribed to interaction-induced motion and the higher-frequency due to the librational motions. This assumption may have some merits for our case.

’ CONCLUSION The low-frequency study of some novel metallocenium ILs has been conducted by terahertz time-domain spectroscopy. The complex dielectric spectra were found to be broad and dispersive. The imaginary spectra were found to be consisting of band structures and cannot be produced by a multi-Debye model of dielectric relaxation. The imaginary part of the dielectric constant was fitted with combinations of four Lorentzian functions. It was found that the dynamics in the THz region is governed by interion dynamics as well as librational motion of the cations. We see that the second and third Lorentzian bands correspond to the interion vibration between the cations and the anions, while the fourth Lorentzian band corresponds to the librational motion. 1318

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319

The Journal of Physical Chemistry A

’ ASSOCIATED CONTENT

bS

Supporting Information. Parameters obtained by fitting using three ant-symmetrized Gaussian functions and one BL function (Table S1). Fitting of the imaginary part of the dielectric constant using three anti-symmetrized Gaussian functions and one BL function (Figure S1). This material is available free of charge via the Internet at http://pubs.acs.org.

’ ACKNOWLEDGMENT This work was supported by a Grant-in-Aid for Creative Scientific Research (No. 17GS0209) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. ’ REFERENCES (1) (a) Seddon, K. R. Nat. Mater. 2003, 2, 363.(b) Rodgers, R; Seddon, K. Ionic Liquids, Industrial Applications For Green Chemistry; ACS Symposium Series 818; American Chemical Society: Washington DC, 2002. (2) Welton, T. Chem. Rev. 1999, 99, 2071. (3) Dupont, J.; de Souza, R. F.; Suarez, P. A. Z. Chem. Rev. 2002, 102, 3667. (4) Electrochemical Aspects of Ionic Liquids; Ohno, H., Eds.; Wiley Interscience: New York, 2005. (5) Ionic Liquids IIIB Fundamental Progress, Challenges and opportunities; ACS Symposium Series 901; Rogers, R. D., Seddon, K. R., Eds.; American Chemical Society: Washington DC, 2005. (6) Ionic Liquids IIIB Fundamental Progress, Challenges and opportunities properties, and structure; ACS Symposium Series 902; Rogers, R. D., Seddon, K. R., Eds.; American Chemical Society: Washington DC, 2005. (7) Wassercheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 3773. (8) Wang, P.; Zakeeruddin, S. M.; Comte, P.; Exnar, L.; Gratzel, M. J. Am. Chem. Soc. 2003, 125, 1166. (9) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am. Chem. Soc. 2002, 124, 4974. (10) Wilkes, J. S. Green Chem. 2002, 4, 73. (11) Song, C. E.; Roh, E. J. Chem. Commun. 2000, 837. (12) Hu, A.; Ngo, H. L.; Lin, W. Angew. Chem., Int. Ed. 2004, 43, 2501. (13) Cave, G. W. V.; Raston, C. L.; Scott, J. L. Chem. Commun. 2001, 2159. (14) Holbrey, J. D.; Seddon, K. R. Clean Products and Processes; Matsunaga, T., Ed.; Springer Verlag: Berlin, 1999; Vol. 1, pp 223-236. (15) Earle, M. J.; Seddon, K. R. Pure Appl. Chem. 2000, 72, 398. (16) Palese, S.; Mukamel, S.; Miller, R. J. D.; Lotshaw, W. T. J. Phys. Chem. 1996, 100, 10380. (17) Beard, M. C.; Lotshaw, W. T.; Korter, T. M.; Heilweil, E. J.; McMorrow, D. J. Phys. Chem. A 2004, 108, 9348. (18) Keiding, S. R. J. Phys. Chem. A 1997, 101, 5250. (19) (a) Smith, N. A.; Meech, S. A. J. Phys. Chem. A 2000, 104, 4223.(b) Hunt, N. T.; Jaye, A. A.; Meech, S. R. Phys. Chem. Chem. Phys. 2007, 9, 2167. (20) Lotshaw, W. T.; McMorrow, D.; Thantu, N.; Melinger, J. S.; Kitchenham, R. J. Raman Spectrosc. 1995, 26, 571. (21) Chang, Y. J.; Castner, E. W., Jr. J. Chem. Phys. 1993, 99, 7289. (22) McMorrow, D.; Thantu, N.; Kleiman, V.; Melinger, J. S.; Lotshaw, W. T. J. Phys. Chem. A 2001, 105, 7960. (23) Giraud, G.; Wynne, K. J. Chem. Phys. 2003, 119, 11753. (24) Kindt, J. T.; Schmuttenmaer, C. A. J. Phys. Chem. 1996, 100, 10373. (25) Schmuttenmaer, C. A. Chem. Rev. 2004, 104, 1759 and references therein. (26) Venables, D. S.; Chiu, A.; Schmuttenmaer, C. A. J. Chem. Phys. 2000, 113, 3243. (27) Dutta, P.; Tominaga, K. J. Mol. Liq. 2009, 147, 45.

ARTICLE

(28) Dutta, P.; Tominaga, K. Mol. Phys. 2009, 107, 1845. (29) Dutta, P.; Tominaga, K. J. Phys. Chem. A 2009, 113, 8235. (30) Hyun, B. R.; Dzyuba, S. V.; Bartsch, R. A.; Quitevis, E. L. J. Phys. Chem. A 2002, 106, 7579. (31) Giraud, G.; Gordon, C. M.; Dunkin, I. R.; Wynne, K. J. Chem. Phys. 2003, 119, 464. (32) Turton, D. A.; Hunger, J.; Stoppa, A.; Hefter, G.; Thoman, A.; Walther, M.; Buchner, R.; Wynne, K. J. Am. Chem. Soc. 2009, 131, 11140. (33) Shirota, H.; Funston, A. M.; Wishart, J. F.; Castner, E. W., Jr. J. Chem. Phys. 2005, 122, 184512. (34) Shirota, H.; Wishart, J.; Castner, E., Jr. J. Phys. Chem. B 2007, 111, 4819. (35) Shirota, H.; Castner, E., Jr. J. Phys. Chem. B 2005, 109, 21576. (36) Shirota, H.; Castner, E. W., Jr. J. Phys. Chem. A 2005, 109, 9388. (37) Shirota, H.; Nishikawa, K.; Ishida, T. J. Phys. Chem. B 2009, 113, 9831. (38) Fujisawa, T.; Nishikawa, K.; Shirota, H. J. Chem. Phys. 2009, 131, 244519. (39) Castner, E. W., Jr.; Wishart, J.; Shirota, H. Acc. Chem. Res. 2007, 40, 1217. (40) Weing€artner, H.; Sasisanker, P.; Daguenet, C.; Dyson, P. J.; Krossing, I.; Slattery, J. M.; Schubert, T. J. Phys. Chem. B 2007, 111, 4775. (41) Cang, H.; Li, J.; Fayer, M. D. J. Chem. Phys. 2003, 119, 13017. (42) Li, J.; Wang, I.; Fruchey, K.; Fayer, M. D. J. Phys. Chem. A 2006, 110, 10384. (43) Rajian, J. R.; Li, S. F.; Bartsch, R. A.; Quitevis, E. L. Chem. Phys. Lett. 2004, 393, 372. (44) Xiao, D.; Rajian, J. R.; Li, S. F.; Bartsch, R. A.; Quitevis, E. L. J. Phys. Chem. B 2006, 110, 16174. (45) Xiao, D.; Rajian, J. R.; Hines, L. G., Jr.; Li, S. F.; Bartsch, R. A.; Quitevis, E. L. J. Phys. Chem. B 2008, 112, 13316. (46) Xiao, D.; Rajian, J. R.; Li, S. F.; Bartsch, R. A.; Quitevis, E. L. J. Phys. Chem. B 2007, 111, 4669. (47) Xiao, D.; Hines, L. G., Jr.; Bartsch, R. A.; Quitevis, E. L. J. Phys. Chem. B 2009, 113, 4544. (48) Xiao, D.; Hines, L. G.; Bartsch, R. A.; Quitevis, E. L.; Russina, O.; Triolo, A. J. Phys. Chem. B 2009, 113, 6426. (49) Xiao, D.; Hines, L. G., Jr.; Holtz, M. W.; Song, K.; Bartsch, R. A.; Quitevis, E. L. Chem. Phys. Lett. 2010, 497, 37. (50) Koeberg, M.; Wu, C.; Kim, D.; Bonn, M. Chem. Phys. Lett. 2007, 439, 60. (51) Yamamoto, K.; Tani, M.; Hangyo, M. J. Phys. Chem. B 2007, 111, 4854. (52) Inagaki, T.; Mochida, T. Chem. Lett. 2010, 39, 572. (53) Yamamoto, K.; Kabir, H. M.; Tominaga, K. J. Opt. Soc. Am. B 2005, 22, 2417. (54) Nishizawa, J.; Tanno, T.; Suto, K. J. Organomet. Chem. 2006, 691, 2479. (55) Ronne, C.; Jensby, K.; Loughnane, B. J.; Fourkas, J.; Faurskov Nielsen, O.; Keiding, S. R. J. Chem. Phys. 2001, 114, 5246. (56) Tao, G.; Stratt, R. M. J. Phys. Chem. B 2006, 110, 976. (57) Ryu, S.; Stratt, R. M. J. Phys. Chem. B 2004, 108, 6782. (58) Shirota, H. J. Chem. Phys. 2005, 122, 44514. (59) Zhong, Q.; Fourkas, T. J. J. Phys. Chem. B 2008, 112, 15342.

1319

dx.doi.org/10.1021/jp108097q |J. Phys. Chem. A 2011, 115, 1313–1319