12204
J. Phys. Chem. B 2007, 111, 12204-12210
Imidazolium-Based Ionic Liquids Formed with Dicyanamide Anion: Influence of Cationic Structure on Ionic Conductivity Yukihiro Yoshida,*,† Osamu Baba,† Carlos Larriba,‡ and Gunzi Saito*,†,§ DiVision of Chemistry, Graduate School of Science, Kyoto UniVersity, Sakyo-ku, Kyoto 606-8502, Japan, Research Center for Low Temperature and Materials Sciences, Kyoto UniVersity, Sakyo-ku, Kyoto 606-8502, Japan, and Mechanical Engineering Department, Yale UniVersity, New HaVen, Connecticut 06520-8286 ReceiVed: June 12, 2007; In Final Form: August 7, 2007
Ionic liquids composed of dicyanamide anion and various imidazolium-based cations were prepared, and the influence of structural variations such as substituting a hydrogen at 2-position and changing the sort of alkyl group at 1-position of imidazolium cations on their thermal behavior, density, solvatochromic effects, viscosity, ionic conductivity, and surface tension was characterized. The substitution of the 2-hydrogen for methyl group or N-methylimidazole decreases the fluidity and ionic conductivity, mainly caused by the increased cohesive energy associated with the increasing cation size. Chain branching at 1-position also gives rise to the pronounced depression of the fluidity and ionic conductivity, presumably as a consequence of the increased π-π interactions between imidazolium rings. We found that the surface tension of the present ionic liquids is in inverse proportion to the molar concentration, which can be originally rationalized on the basis of the hole theory.
Introduction Room-temperature (RT) ionic liquids have attractive liquid properties such as negligible vapor pressure, diverse miscibility with solvents, tolerance to strong acids, excellent thermal and electrochemical stabilities, and high ionic conductivity,1 which render them especially suitable for electrochemical applications including lithium batteries, fuel cells, electric double layer capacitors, and solar cells.2 Of particular importance is that these properties can be tailored by the selection of component cations and anions.3 Among the anions, dicyanamide (N(CN)2) anion has afforded highly fluid and conductive ionic liquids made by combining hetero cations such as 1,3-dimethylimidazolium (DMI, 3.6 × 10-2 S cm-1 at 25 °C),4 1-ethyl-3-methylimidazolium (EMI, 2.8 × 10-2 S cm-1 at 25 °C)4,5 and diethylmethylsulfonium (2.68 × 10-2 S cm-1 at 20 °C),6 and thereby has been utilized as components of electrolyte for high performance dye-sensitized solar cell.7 Since the ionic liquids are entirely composed of ions as the name implies, their physical properties must be closely connected with the interionic interactions, mainly comprised of cation‚‚‚anion Coulombic attractive and cation‚‚‚cation van der Waals (vdW) attractive forces. Very recently, we demonstrated that the ion diffusion and ion-association propensities of the N(CN)2-based ionic liquids are strongly affected by the interionic vdW attractions, which were manifested in the influence of the alkyl chain length of 1-(n-alkyl)-3-methylimidazolium cations and the ring structure of heterocyclic cations.4 We now have extended our research to the influence of the sort, not the length, of substituent group in the imidazolium-based cations on the liquid behaviors. * To whom correspondence should be addressed. E-mail: yoshiday@ kuchem.kyoto-u.ac.jp (Y.Y.);
[email protected] (G.S.). † Division of Chemistry, Graduate School of Science, Kyoto University. ‡ Yale University. § Research Center for Low Temperature and Materials Sciences, Kyoto University.
These include the 1-(n-butyl)-3-methylimidazolium (BMI), and its branched isomers 1-(i-butyl)-3-methylimidazolium (iBMI) and 1-(t-butyl)-3-methylimidazolium (tBMI) cations, as drawn in Scheme 1. Taking advantage of newly characterized N(CN)2based ionic liquids, involving methyl group (BM2I) or Nmethylimidazole (BM2BI) at 2-position of the BMI cation that has an acidic hydrogen enough to form a significant hydrogen bond to the anions,8 the substituent effect at 2-position on the liquid behaviors are also investigated. Here we report the preparation, characterization, and properties such as thermal behaviors, hydrogen-bond donor (HBD) ability, polarity, viscosity, ionic conductivity, ion association, and surface tension of the N(CN)2-based ionic liquids composed of such cations. Results and Discussion Thermal Behavior and Density. Table 1 summarizes glass transition (Tg), crystallization (Tc), melting (Tm), and decomposition (Td) temperatures of N(CN)2 salts in the present study, as determined by the analysis of DSC traces. All the salts, except for [tBMI][N(CN)2], having a melting point of 59 °C, are in liquid form at RT, and exhibit either a low melting point or an extremely low glass transition without melting (Figure 1). [tBMI][N(CN)2] crystallizes at 33 °C even on rapid cooling (-30 °C min-1) and exhibits no glass transition down to -110 °C, presumably as a consequence of both high symmetry and limited orientational freedom of the t-butyl group. Such effects were manifested for the difference in melting points between t-butanol (25 °C) and n-butanol (-90 °C). For three cations including the 1-(n-butyl)-3-methylimidazolium moiety, namely BMI, BM2I, and BM2BI, the glass transition temperature steadily increases with the increase of cation size in going from BMI to BM2BI. The high Tg values driven by the 1,1′,3-trialkyl2,2′-biimidazolium cations were reported for the ionic liquids based on bis(trifluoromethanesulfonyl)imide (Tf2N) and PF6 anions by Xiao and Shreeve,9 and the influence of molecular
10.1021/jp0745236 CCC: $37.00 © 2007 American Chemical Society Published on Web 10/03/2007
Influence of Cationic Structure on Ionic Conductivity
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12205
SCHEME 1: Molecular Structures of Heterocyclic Cations in Text and Syntheses of Their Halide Saltsa
a BMI, 1-(n-butyl)-3-methylimidazolium; BM I, 1-(n-butyl)-2,3-dimethylimidazolium; iBMI, 1-(i-butyl)-3-methylimidazolium; tBMI, 1-(t-butyl)2 3-methylimidazolium; BM2BI, 1-(n-butyl)-1,3′-dimethyl-2,2′-biimidazolium.
TABLE 1: Thermal Properties of N(CN)2 Saltsa salt
Tg/°C
Tc/°C
Tm/°C
Td/°C
[BMI][N(CN)2]b [BM2I][N(CN)2] [BM2BI][N(CN)2] [iBMI][N(CN)2] [tBMI][N(CN)2]
-94 -81 -42 -80
-35 -37
-10 26
ca. 220 ca. 300 ca. 270 ca. 230 ca. 260
59
a
Tg, glass transition temperature; Tc, crystallization temperature; Tm, melting temperature; Td, decomposition temperature. All the values were determined by DSC on heating process. b Reference 4.
Figure 1. Glass transition (Tg, b) and melting (Tm, O) temperatures of N(CN)2 salts in the present study.
symmetry of the 1,1′,3-trialkyl-2,2′-biimidazolium cations on thermal behaviors of their salts will be dealt with in a separate paper.10 The order for Tg (BM2BI > BM2I > BMI) strongly suggests that the glass forming is more under the control of the van der Waals (vdW) interactions than of the Coulombic interactions, since the latter must decrease with increasing ion size. On heating, the N(CN)2 salts decompose at high temperatures (220-300 °C), providing a wide liquid region over 200 °C. As seen in Table 2, the molar concentration (C) decreases steadily in going from BMI to BM2BI, i.e., increasing cation
size. The density (d) of [BM2I][N(CN)2] is slightly larger than that reported previously (1.055 g cm-3 at 25 °C).11 For branched isomers having the same molecular weights, both d and C values of [tBMI][N(CN)2] are significantly large in comparison with those of [BMI][N(CN)2] and [iBMI][N(CN)2], which must be a reflection of the high cohesive energy of the tBMI cation, in relation to the high Tm. The lower density and molar concentration of [BMI][N(CN)2] as compared to those of [iBMI][N(CN)2], which are combined with the lower activation energies for viscosity and ionic conduction (vide infra), remind us of the case of the N(CN)2-based ionic liquids composed of 1-(n-alkyl)3-methylimidazolium cations with different alkyl chain lengths.4 In this system, the cation with the longer n-alkyl chain leads to higher cohesive ability mainly associated with the vdW attraction between the alkyl chains. Since the chain branching could exert an effect in ways that lead to the depressed orientational degree of freedom and steric hindrance relative to the case of the n-alkyl group, the source of larger density of [iBMI][N(CN)2] must lie in the increased cohesive energy driven by the π-π attractions between imidazolium rings. Also, this propensity could be a part of the reason for its higher Tg value than that of [BMI][N(CN)2]. Hydrogen-Bond Donor Ability and Polarity. A great variety of solvatochromic probes have been used to estimate the polarity of ionic liquids, by comparison to well-established empirical solvent polarity scales based on molecular liquids.12 Among them, the ET(30) value of Dimroth and Reichardt13 is a good general scale of the solvating ability of a liquid, and is defined as ET(30) (in kcal mol-1) ) 28592/λReichardt (in nm), where λReichardt is the wavelength maximum of the lowest energy band of a Reichardt’s betaine dye, 2,6-diphenyl-4-(2,4,6triphenyl-N-pyridino)phenolate. Because of its zwitterionic structure, the solvatochromic property is strongly affected by the dipolarity/polarizability (π*), and hydrogen-bond acidity (R) of the solvent (or its component). The dipolarity/polarizability
12206 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Yoshida et al.
TABLE 2: Physical Properties of N(CN)2 Saltsa salt b
[BMI][N(CN)2] [BM2I][N(CN)2] [BM2BI][N(CN)2] [iBMI][N(CN)2] [tBMI][N(CN)2]
d/g cm-3
C/mol cm-3
1.06 1.09c 1.17 1.08 1.12d
5.16 × 10 4.99 × 10-3 4.11 × 10-3 5.26 × 10-3 5.44 × 10-3 d
η/cP (Ea(η)/kcal mol-1)
σ/S cm-1 (Ea(σ)/kcal mol-1)
Λ/S cm2 mol-1
γ/dyn cm-1
29.3 (6.7) 67.2 (7.7) 977 (13.7) 61.1 (7.6)
1.1 × 10 (4.5) 4.9 × 10-3 (6.2) 4.1 × 10-4 (10.9) 5.5 × 10-3 (5.6) 4.6 × 10-5 d
2.1 0.98 0.10 1.1 8.5 × 10-3 d
46.5 48.9 35.8 45.7
-3
-2
a d: density at 20 °C. C: molar concentration at 20 °C. η: viscosity at 25 °C. Ea(η): activation energies for viscous flow. σ: ionic conductivity at 25 °C. Ea(σ): activation energies for ionic conduction. Λ: molar conductivity at 25 °C, where the density at 20 °C is used to estimate the value. γ: surface tension at 23 °C. b Reference 4. c Literature value: d ) 1.055 g cm-1 (25 °C).11 d In solid state.
TABLE 3: Solvatochromic Parameters of N(CN)2 Saltsa Reichardt’s dye salt b
[BMI][N(CN)2] [BM2I][N(CN)2] [BM2BI][N(CN)2] [iBMI][N(CN)2]
Kamlet-Taft
ET(30)/kcal mol-1
ETN
π*
R
51.4 48.8 49.1 51.6
0.639 0.558 0.567 0.646
1.05 1.04 1.02 1.06
0.533 0.373 0.405 0.544
a Concentration is on the order of 10-3 M. ET(30), ETN, π*, and R: see text. b Reference 4.
characteristics in a liquid can be readily estimated on the basis of the intramolecular charge transfer (CT) transition energy of aromatic molecules with both electron withdrawing and donating groups. The dimensionless π* value, which was initially defined by Kamlet, Abboud, and Taft,14 is normalized by taking dimethyl sulfoxide (π* ) 1.00) and cyclohexane (π* ) 0.00), and can be correlated to the wavelength maximum of the lowest energy band of N,N-dimethyl-p-nitroaniline (λDMPNA) as π* ) (28.18 - 104/λDMPNA (in nm))/3.52.15 On the other hand, a Kamlet-Taft R value is more diagnostic for the HBD ability of cations in ionic liquids, and is related to the ET(30) and π* values through ET(30) (in kcal mol-1) ) 31.2 + 11.5π* + 15.2R.16 Table 3 presents the ET(30), π*, and R values of the present salts, together with dimensionless normalized scales of ET(30), ETN, by making use of water (ETN ) 1.00) and TMS (ETN ) 0.00) as reference solvents. The order for ETN among the present salts is consistent with that for R, presumably indicating that the HBD ability of cations is the controlling influence on the ETN value. It is apparent that the ETN values of [BM2I][N(CN)2] and [BM2BI][N(CN)2] are comparable to each other and are significantly low in comparison with that of [BMI][N(CN)2]. This result is strongly indicative of the high HBD ability of 2-hydrogen of the imidazolium cations, since the solvent‚‚‚solute interaction with the dyes in 1,3-dialkylimidazolium salts is preferentially formed through the 2-hydrogen in a diluted condition. Similar tendency was found for Tf2N-based ionic liquids (ETN ) 0.642-0.644 for [BMI][Tf2N] and 0.541-0.552 for [BM2I][Tf2N]).12c,17 Figure 2 illustrates the relationship between π* and molar concentration (C) for the N(CN)2-based ionic liquids, including data from an earlier work on 1-alkyl-3-methylimidazolium cations with different alkyl chain lengths.4 It is apparent from the figure that the π* values roughly show the linear variation with C, as seen in the previous report.12g Since the π* values of most ionic liquids including the present salts are significantly high (close to unity),17,18 this trend indicates that the Coulombic interactions between component ions and solutes, that relate to the charge density, are a factor governing the π* value.12g Viscosity and Ionic Conductivity: Variation of Substituents at the 2-Position of Imidazolium Cation. For three N(CN)2-based ionic liquids including 1-(n-butyl)-3-methylimidazolium moiety (BMI, BM2I, and BM2BI), there is a pronounced increase in viscosity (η) as the cation size increases.
Figure 2. Plot of π* against molar concentration for N(CN)2 salts. A solid line is the least-squares fit to the data. Data for DMI, EMI, and C6MI are taken from ref 4.
Figure 3. Temperature dependence of viscosity for N(CN)2 salts with BMI (O), BM2I (∆), and BM2BI (0) cations.
BM2BI cation gives an exceptionally viscous ionic liquid, even by the combination with the N(CN)2 anion, which was found to provide the most fluid of a group of ionic liquids.4-6 Their temperature dependencies follow the Arrhenius equation η ) η0exp(Ea(η)/kBT) as depicted in Figure 3, where η0 is a constant and Ea(η) is the activation energy for viscous flow. As indicated above, the Ea(η) values are in the order BM2BI > BM2I > BMI, in accord with the order of Tg. The Arrhenius plots especially for [BM2BI][N(CN)2] show a slight upward curvature, which might be rationalized with the Vogel-Fulcher-Tamman (VFT) equation on account of their glass-forming behavior.19 According to the Stokes-Einstein equation (eq 1) for the self-diffusivity (D) in an ionic medium with viscosity η and average ion radius r
D ) RT/cπNAηr
(1)
where c is a constant (4 e c e 6), R is the universal gas constant, and NA is the Avogadro’s number, it is immediately apparent that the self-diffusivity decreases as the cation size increases (both η and r increase), i.e., the order BMI > BM2I > BM2BI. The higher viscosity of the BM2I salt than that of the BMI salt
Influence of Cationic Structure on Ionic Conductivity
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12207
Figure 4. Temperature dependence of ionic conductivity for N(CN)2 salts with BMI (O), BM2I (∆), and BM2BI (0) cations.
is consistent with several ionic liquids with Tf2N20 and BF420b anions. The increase of cation size induces the interionic vdW attractive force as mentioned above, whereas such modifications also cause the steric hindrance for the cation‚‚‚anion Coulombic attractions that must decrease the ion diffusion. Since the substitution at the 2-position could exert an effect in ways that lead to diminished hydrogen-bonding interactions with anions8 as verified above, it is possible that the increased interionic vdW interactions by the increasing cation size, not the substituting at 2-position, are responsible for the increased viscosity in going from BMI to BM2BI. As expected from the viscosity data, the ionic conductivity (σ) is strongly affected by the variation of substituents at 2-position of imidazolium cations; i.e., the σ value of the present salts decreases steadily with increasing cation size. Since the ion diffusion decreases in going from BMI to BM2BI as indicated above, this trend in ionic conduction is straightforwardly explained on the basis of the Nernst-Einstein equation (eq 2) for the relationship between the self-diffusivity (D) of the liquid and its molar conductivity (Λ ) σ/C)
D ) RTΛ/z2F2
(2)
where z is the charge of ions and F is the Faraday constant. The temperature dependences can be fit well to the Arrhenius equation σ ) σ0 exp(-Ea(σ)/kBT) over the measured temperature range (25-70 °C, Figure 4), where σ0 is a constant and Ea(σ) is the activation energy for ionic conduction. The σ value of [BM2BI][N(CN)2] is as small as 4.6 × 10-3 S cm-1 even at 70 °C, which is comparable to the value of [BM2I][N(CN)2] at 25 °C (4.9 × 10-3 S cm-1). It appears that each Ea(σ) value is lower than the corresponding Ea(η) value, as is the case in most ionic liquids.21,22 It is noticed, from literature, that the marked difference in activation energy is a measure of the degree of decoupling motion of more diffusive ions on relative to resting ions. Thus, the ratio Ea(σ)/Ea(η) is in inverse proportion to the logarithm of the decoupling index.22 However, there is no obvious correlation between the ratio and cation size (0.67 for BMI, 0.81 for BM2I, and 0.80 for BM2BI), although all the values are less than unity and are close to the “universal” value for the molten salts.23 Viscosity and Ionic Conductivity: Variation of Sort of Alkyl Chain at 1-Position of Imidazolium Cation. Table 2 also presents some physical parameters of the three N(CN)2based salts composed of 1-alkyl-3-methylimidazolium cations with a branched alkyl group at 1-position, namely i-butyl (iBMI) and t-butyl (tBMI), as well as n-butyl (BMI). We replot the temperature dependences of viscosity and ionic conductivity for
Figure 5. Temperature dependence of viscosity for N(CN)2 salts with BMI (O), iBMI (∆), and tBMI (0) cations.
Figure 6. Temperature dependence of ionic conductivity for N(CN)2 salts with BMI (O), iBMI (∆), and tBMI (0) cations. An arrow indicates a melting temperature of [tBMI][N(CN)2] (Tm ) 59 °C).
[BMI][N(CN)2] in Figures 5 and 6, respectively, and compare them to those of the two salts in question. Since crystalline [tBMI][N(CN)2] has a melting event at 59 °C, the viscosity measurement using a rotational viscometer is available above the temperature. It appears from Figure 5 that the temperature dependences of viscosity follow well the Arrhenius equation and the viscosity at 70 °C is in the order tBMI (11.9 cP) > iBMI (11.2 cP) > BMI (6.9 cP). For ionic conduction, the value of [tBMI][N(CN)2] steadily increases up to its Tm, at which temperature it shows an abrupt jump and eventually attains a value of about 4 × 102 times the RT conductivity. The temperature at which the differential dσ/dT shows a maximum (ca. 63 °C) is consistent with the off-set temperature of the endothermic DSC peak on the heating process. The ionic conductivity at 70 °C falls in the order BMI (2.9 × 10-2 S cm-1) > tBMI (2.0 × 10-2 S cm-1) > iBMI (1.9 × 10-2 S cm-1), which does not coincide with the reciprocal order of viscosity. The higher ionic conductivity of [tBMI][N(CN)2], despite having the high Tm, than that of [iBMI][N(CN)2] might be a reflection of the smaller ion radius (r) of the tBMI cation, since the self-diffusivity is originally anticipated to vary as the inverse first power of r for a given viscosity, according to eq 1. The lower fluidity and conductivity of [iBMI][N(CN)2] and [tBMI][N(CN)2] than those of [BMI][N(CN)2] provide us with further confirmation of the higher cohesive ability of such branched isomers. The order of fluidity and ionic conductivity BMI > iBMI is consistent with those of the ionic liquids with Tf2N anion.12a The Walden Plot in Relation to Ion Association. The empirical Walden rule based on a wide range of aqueous
12208 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Yoshida et al.
rh ) (8/5π)(kBT/γ)1/2
Figure 7. Plot of the molar conductivity against the reciprocal of viscosity for N(CN)2 salts in the present study. The diagonal dotted line indicates the ideal Walden line (see text). Inset is the product of molar conductivity and viscosity of the N(CN)2 salts. The horizontal dotted line indicates the ideal Walden product (see text).
solution systems,24 namely that the molar conductivity (Λ) is inversely proportional to the viscosity (η) of the conducting medium, is understood in terms of the same manner as the combination of the Stokes-Einstein (eq 1) and Nernst-Einstein (eq 2) equations, provided that the effect of ion radius r on ion diffusion is not pronounced and no ion pairs exist. Indeed, the Walden plot on a log-log scale, depicted in Figure 7, demonstrates that the present N(CN)2-based ionic liquids with structurally related cations roughly follow the relation. However, we should note that all the salts reside on or below the line evaluated from the data of KCl aqueous solution, in which the ions are completely dissociated. A departure from the ideal Walden line is strongly indicative of a significant degree of ion association, which leads the diminished conductivity at a given viscosity.22 Therefore, it appears that the Walden product Λη, which is expected to decrease as the fraction of the ion association increases, is more diagnostic for the contribution of the ion-association propensity to the ion diffusivity. As seen in the inset of Figure 7, the present salts except for [BM2BI][N(CN)2] have similar Λη values, and are apparently less than that of KCl aqueous solution (100 S cm2 cP mol-1). [BM2I][N(CN)2] (66 S cm2 cP mol-1) having a methyl group at 2-position has a similar Λη value to that of [BMI][N(CN)2] (62 S cm2 cP mol-1), presumably indicating that the hydrogen bond to anions given above for the solvatochromic effect is not large enough to affect the ion-association propensity. The higher Λη value of [BM2BI][N(CN)2] (98 S cm2 cP mol-1) does not come from the larger cation size, because the increased r value could give the Walden product in the negative direction, according to eqs 1 and 2. It is more likely that the possible difference in ion diffusion between cations and anions is responsible for the high Λη value, since such a difference could result in the high ionic conductivity by more diffusive ions in a quasi-lattice composed of resting ions with a given viscosity. Surface Tension and Molar Concentration Relation. There have been limited reports on systematic study on surface tensions of ionic liquids21f,25 although surface properties of such conducting medium have attracted much attention in recent years, associated with the interface phenomena on the electrodes for electrochemical processes including electrodeposition, batteries, and solar cells.2 According to the hole theory, which was developed by Fu¨rth,26 molten salts contain empty spaces that move and change their size by thermal fluctuations of component ions in local spaces, and the holes with random size and distribution are closely connected with the surface tension (γ) by the expression27
(3)
where rh is the hole radius. Although it is difficult to assess precisely the radii of the nonspherical ions with support of both experimental and theoretical approaches, the effective sum of the ion radii of components is well represented by the molar volume that is reciprocal of the molar concentration (C). Figure 8 shows the correlation between the surface tension and molar concentration for the N(CN)2-based ionic liquids including the present salts.28 Assuming that the effect of ion radius r on the concentration is not pronounced, the linear correlation is simply explained in terms of the smaller holes in the denser liquid. It is also apparent that the γ value of the 1-alkyl-3-methylimidazolium salts decreases as the alkyl chain of the cations elongates in going from DMI (54.9 dyn cm-1) to C6MI (40.6 dyn cm-1). Such behaviors were earlier seen for the imidazolium-based ionic liquids composed of BF4, PF6, and Tf2N anions.25a-c,i On the basis of eq 3, the rh values were estimated to be 1.52 Å for BMI, 1.54 Å for iBMI, 1.49 Å for BM2I, and 1.74 Å for BM2BI salts, which are significantly small in comparison with the vdW radii of the smallest BMI cation (3.03 Å)21d and N(CN)2 anion (2.70 Å).4 As far as the hole theory goes, therefore, the holes in the present ionic liquids are not large enough to allow the ions to jump in the quasi-lattice, namely impeding the motion of ions. Concluding Remarks In this paper, our research interest is centered on the N(CN)2based ionic liquids composed of imidazolium cations. The way of structural modifications is enormously broad, and we have only touched on substituting a hydrogen at the 2-position and chain branching of the alkyl chain at the 1-position of imidazolium cations. Although both modifications result in the fluidity and ionic conductivity in the negative direction, we found that the n-butyl groups of BMI cations are not correlated in ways that lead to diminished conductivity through vdW attractions between the alkyl chains but give rise to the steric hindrance for the π-π attractions between imidazolium rings. It also appears that the hydrogen bond between cations and anions has little effect on the Walden product, that relates to the ionassociation propensity. Notably, dipolarity/polarizability (π*) and surface tension of the present salts are roughly proportional to molar concentration. One can therefore envisage the opening of many opportunities for designing an ionic liquid with such desired properties, since the way to assess the molar concentration of various ionic liquids was recently supposed on the basis of molecular structures of the component ions.29 Experimental Section Synthesis. Solvents (water, methanol, acetonitrile, and ethyl acetate) were distilled prior to use. N-(t-Butyl)imidazole was synthesized according to the literature procedures,30 and purified by the distillation twice. 1,1′-Dimethyl-2,2′-biimidazole was also synthesized according to the literature procedures,9,31 and purified by column chromatography on alumina (elution: hexane/ethyl acetate (1:1)) followed by sublimation under reduced pressure. White crystalline solids [BMI]Br and [BM2I]Br were synthesized by Menschutkin reaction of appropriate alkylimidazoles with n-butyl bromide, and purified by reprecipitation from acetonitrile/ethyl acetate. Ag[N(CN)2] was prepared by the metathesis of recrystallized Na[N(CN)2] with AgNO3 (Aldrich, 99.9999%) in the dark in water. Preparation of [BMI][N(CN)2] was reported in detail in our previous
Influence of Cationic Structure on Ionic Conductivity
Figure 8. Plot of surface tension against molar concentration for N(CN)2 salts. A solid line is the least-squares fit to the data.
publication.4 2,6-Diphenyl-4-(2,4,6-triphenyl-N-pyridino)phenolate (Reichardt’s betaine dye) and N,N-dimethyl-p-nitroaniline were obtained from Fluka (99.0%) and Tokyo Kasei (98%), respectively, and used without purification. [iBMI]Br. To N-methylimidazole (9.1 g, 110 mmol) was added a slight excess of 1-bromo-2-methylpropane (19.7 g, 144 mmol) at RT. The reaction mixture was stirred at 110 °C for 16 h. Pale yellow viscous liquid was decolorized by activated charcoal powder in methanol solution and dried in vacuo at RT for 2 days (21.4 g, 97.6 mmol, 89% yield). Tg: -49 °C. IR (KBr) νmax: 3146m, 3087m, 2964m (νC-H) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 1.00 (d, J ) 6.8 Hz, 6H, CH2CH(CH3)2), 2.32 (sept, J ) 6.8 Hz, 1H, CH2CH(CH3)2), 4.16 (s, 3H, CH3), 4.34 (d, 2H, J ) 6.8 Hz, CH2CH(CH3)2), 7.91 (s, 1H, CH), 7.96 (s, 1H, CH), 10.27 (s, 1H, NCHN) ppm. Anal. Calcd for C8H15N2Br: C, 43.85; H, 6.90; N, 12.78; Br, 36.47%. Found: C, 42.95; H, 6.89; N, 12.70; Br, 36.35%. [tBMI]I. To a solution of N-(t-butyl)imidazole (4.6 g, 37 mmol) in methanol (50 cm3) was added a slight excess methyl iodide (6.4 g, 45 mmol) at RT, and the reaction mixture was refluxed for 18 h. Evaporation of the solvent afforded a pale yellow crystalline solid. A white crystalline solid was yielded by recrystallization from acetonitrile, washed by ethyl acetate, and dried in vacuo (8.0 g, 30 mmol, 81% yield). Tm: 146 °C. IR (KBr) νmax: 3140s, 3111m, 3071vs, 2993vw, 2968s, 2948w, 2922w (νC-H) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 1.79 (s, 9H, C(CH3)3), 4.13 (s, 3H, CH3), 7.83 (s, 1H, CH), 8.02 (s, 1H, CH), 9.78 (s, NCHN) ppm. Anal. Calcd for C8H15N2I: C, 36.11; H, 5.68; N, 10.53; I, 47.69%. Found: C, 35.96; H, 5.43; N, 10.64; I, 47.49%. [BM2BI]Br. To 1,1′-dimethyl-2,2′-biimidazole (3.9 g, 24 mmol) was added an equimolar amount of n-butyl bromide (3.3 g, 24 mmol) at RT. The reaction mixture was stirred at 100 °C for 1 day. A white crystalline solid was yielded by reprecipitation twice from acetonitrile/ethyl acetate, washed by ethyl acetate, and dried in vacuo (5.6 g, 19 mmol, 77% yield). Tm: 185 °C. IR (KBr) νmax: 3153vw, 3110m, 2962m, 2936w (νC-H) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 0.87 (t, 3H, J ) 7.6 Hz, CH2CH2CH2CH3), 1.31 (m, 2H, CH2CH2CH2CH3), 1.86 (m, 2H, CH2CH2CH2CH3), 3.96 (s, 3H, CH3), 4.02 (s, 3H, CH3), 4.32 (m, 2H, CH2CH2CH2CH3), 7.39 (s, 1H, CH), 7.70 (s, 1H, CH), 8.27 (s, 1H, CH), 8.34 (s, 1H, CH) ppm. Anal. Calcd for C12H19N4Br: C, 48.17; H, 6.40; N, 18.72; Br, 26.71%. Found: C, 47.62; H, 6.43; N, 18.60; Br, 26.57%. [iBMI][N(CN)2]. A slight excess of Ag[N(CN)2] (5.74 g, 33.0 mmol) and [iBMI]Br (6.57 g, 30.0 mmol) were dissolved in distilled water (50 cm3). The resulting suspension was stirred for 1 day in the dark at RT, and filtered to remove any trace of AgBr and unreacted Ag[N(CN)2]. Evaporation of the filtrate
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12209 was performed under vacuum at 40 °C to give transparent liquid (5.64 g, 27.5 mmol, 92% yield). IR (KBr) νmax: 3151m, 3109m, 3025vw, 2965m (νC-H), 2240s, 2199s, 2142vs (νCN) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 1.01 (d, J ) 6.8 Hz, 6H, CH2CH(CH3)2), 2.28 (sept, J ) 6.8 Hz, 1H, CH2CH(CH3)2), 4.12 (s, 3H, CH3), 4.25 (d, 2H, J ) 7.6 Hz, CH2CH(CH3)2), 7.80 (s, 1H, CH), 7.82 (s, 1H, CH), 9.17 (s, 1H, NCHN) ppm. Anal. Calcd for C10H15N5: C, 58.52; H, 7.37; N, 34.12; Br, 0.00%. Found: C, 58.06; H, 7.31; N, 34.08; Br, 0.00%. [tBMI][N(CN)2]. Transparent liquid was prepared by the procedure described above for [iBMI][N(CN)2] except that [tBMI]I was used instead of [iBMI]Br. Recrystallization from acetonitrile/ethyl acetate gave white needle crystals. Yield: 84%. IR (KBr) νmax: 3152w, 3101w, 2985w (νC-H), 2240s, 2200s, 2140vs (νCN) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 1.79 (s, 9H, C(CH3)3), 4.11 (s, 3H, CH3), 7.81 (s, 1H, CH), 8.02 (s, 1H, CH), 9.26 (s, NCHN) ppm. Anal. Calcd for C10H15N5: C, 58.52; H, 7.37; N, 34.12; I, 0.00%. Found: C, 58.43; H, 7.18; N, 34.38; I, 0.00% [BM2I][N(CN)2]. Transparent liquid was prepared by the procedure described above for [iBMI][N(CN)2] except that [BM2I]Br was used instead of [iBMI]Br. Yield: 97%. IR (KBr) νmax: 3182vw, 3141w, 3025vw, 2962m, 2937w (νC-H), 2238s, 2198s, 2138vs (νCN) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 0.99 (t, 3H, J ) 7.4 Hz, CH2CH2CH2CH3), 1.44 (m, 2H, CH2CH2CH2CH3), 1.89 (m, 2H, CH2CH2CH2CH3), 2.82 (s, 3H, CCH3), 3.99 (s, 3H, NCH3), 4.33 (t, 2H, J ) 7.6 Hz, CH2CH2CH2CH3), 7.62 (s, 1H, CH), 7.69 (s, 1H, CH) ppm. Anal. Calcd for C11H17N5: C, 60.26; H, 7.82; N, 31.93; Br, 0.00%. Found: C, 59.83; H, 7.85; N, 31.76; Br, 0.00%. [BM2BI][N(CN)2]. Transparent viscous liquid was prepared by the procedure described above for [iBMI][N(CN)2] except that [BM2BI]Br was used instead of [iBMI]Br. Yield: 86%. IR (KBr) νmax: 3153w, 3139w, 3119m, 3026w, 2964m, 2937w (νC-H), 2239s, 2198s, 2139vs (νCN) cm-1. 1H NMR (400 MHz, acetone-d6) δ: 0.88 (t, 3H, J ) 7.6 Hz, CH2CH2CH2CH3), 1.31 (m, 2H, CH2CH2CH2CH3), 1.87 (m, 2H, CH2CH2CH2CH3), 3.92 (s, 3H, CH3), 3.99 (s, 3H, CH3), 4.27 (m, 2H, CH2CH2CH2CH3), 7.40 (s, 1H, CH), 7.68 (s, 1H, CH), 8.05 (s, 1H, CH), 8.11 (s, 1H, CH) ppm. Anal. Calcd for C14H19N7: C, 58.94; H, 6.71; N, 34.35; Br, 0.00%. Found: C, 59.23; H, 6.76; N, 34.64; Br, 0.00%. Measurements. 1H NMR measurements were conducted on a JEOL JNM-FX400 spectrometer operating at 400 MHz, and acetone-d6 was used as solvent. FT-IR spectra were taken in dispersed KBr pellets on a Perkin-Elmer 1000 series spectrophotometer (400-4000 cm-1). Glass transition (Tg), crystallization (Tc), melting (Tm), and decomposition (Td) temperatures were determined from differential scanning calorimetry (DSC) thermograms during the heating scans (10 °C min-1) on a Shimadzu DSC-60 instrument equipped with nitrogen cryostatic cooling. The samples were sealed in aluminum pans under an inert atmosphere of helium gas in a glovebox (H2O, O2 < 1 ppm), and the temperature was calibrated by water and indium. Density values, except for [tBMI][N(CN)2], were obtained by measuring the weight of the sample in a 1 cm3 pycnometer, whereas that of crystalline [tBMI][N(CN)2] was obtained using a floatation method in a mixed solution of tetrachloromethane and cyclohexane, in the glovebox. Viscosities were measured using a cone-type Tokyo Keiki RE-80L rotational viscometer. Conductance measurements were performed in a two platinum electrode conductivity cell (cell constant is 38.3 cm-1), which was set in the glovebox. The conductance of the samples was determined from the first real axis touchdown point in the Cole-
12210 J. Phys. Chem. B, Vol. 111, No. 42, 2007 Cole plot of the impedance data, using an Agilent Technologies impedance analyzer 4294A over a frequency range 40 Hz to 110 MHz. UV-vis absorption spectra of Reichardt’s betaine dye and DMPNA dissolved in ionic liquids were taken in a quartz cell with light path length of 1 mm on a Shimadzu UV3100 spectrophotometer (300-800 nm). Concentration is on the order of 10-3 M, and the samples were sealed into the cell in the glovebox. Surface tension values were obtained using a capillary rise method, which was described in more detail elsewhere.25g All surface tension measurements were done at 23 °C with the same batch of capillaries. The radius of one of the capillaries was measured at different points to make sure there was no variation in radius through out the length. Elemental analyses (C, H, N, and halogen) were carried out by the Center for Organic Elemental Microanalysis of Kyoto University. Acknowledgment. The authors thank Masatoshi Kondo for obtaining 1H NMR spectra. This work was in part supported by 21st Century COE program on Kyoto University Alliance for Chemistry and Grant-in-Aid for Scientific Research (15205019) from the Ministry of Education, Culture, Sports, Science and Technology, Japan. Y.Y. acknowledges the financial support of Grants-in-Aid for Scientific Research (17750126) from Japan Society for the Promotion of Science (JSPS). C.L. has been supported by the U.S. AFOSR through Grants F-49620-01-1-0416 and FA9550-06-1-0104, and AFOSR subcontracts from the companies Busek and Connecticut Analytical Corporation. Supporting Information Available: Numerical data of ionic conductivity and viscosity as a function of temperature for the present ionic liquids. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) For selected reviews, see: (a) Welton, T. Chem. ReV. 1999, 99, 2071-2083. (b) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 3772-3789. (2) For selected reviews, see: (a) Electrochemical Aspects of Ionic Liquids; Ohno, H., Ed.; John Wiley and Sons: New York, 2005; Chapters 14-17. (b) Buzzeo, M. C.; Evans, R. G.; Compton, R. G. ChemPhysChem 2004, 5, 1106-1120. (3) For selected reviews, see: (a) Hagiwara, R.; Ito, Y. J. Fluorine Chem. 2000, 105, 221-227. (b) Wilkes, J. S. J. Mol. Catal. A: Chem. 2004, 214, 11-17 and references therein. (c) Chiappe, C.; Pieraccini, D. J. Phys. Org. Chem. 2005, 18, 275-297. (d) Poole, C. F. J. Chromatogr., A 2004, 1037, 49-82. (4) Yoshida, Y.; Baba, O.; Saito, G. J. Phys. Chem. B 2007, 111, 47424749. (5) (a) Yoshida, Y.; Muroi, K.; Otsuka, A.; Saito, G.; Takahashi, M.; Yoko, T. Inorg. Chem. 2004, 43, 1458-1462. (b) Yoshida, Y.; Fujii, J.; Muroi, K.; Otsuka, A.; Saito, G.; Takahashi, M.; Yoko, T. Synth. Met. 2005, 153, 421-424. (c) MacFarland, D. R.; Golding, J.; Forsyth, S.; Forsyth, M.; Deacon, G. B. Chem. Commun. 2001, 1430-1431. (6) Gerhard, D.; Alpaslan, S. C.; Gores, H. J.; Uerdingen, M.; Wasserscheid, P. Chem. Commun. 2005, 5080-5082. (7) (a) Wang, P.; Zakeeruddin, S. M.; Moser, J.-E.; Gra¨tzel, M. J. Phys. Chem. B 2003, 107, 13280-13285. (b) Kawano, R.; Matsui, H.; Matsuyama, C.; Sato, A.; Susan, M. A. B. H.; Tanabe, N.; Watanabe, M. J. Photochem. Photobiol., A 2004, 164, 87-92. (8) (a) Tait, S.; Osteryoung, R. A. Inorg. Chem. 1984, 23, 4352-4360. (b) Dieter, K. M.; Dymek, C. J.; Heimer, N. E.; Rovang, J. W.; Wilkes, J. S. J. Am. Chem. Soc. 1988, 110, 2722-2726. (c) Elaiwi, A.; Hitchcock, P. B.; Seddon, K. R.; Srinivasan, N.; Tan, Y.-M.; Welton, T.; Zora, J. A. J. Chem. Soc., Dalton Trans. 1995, 3467-3472. (9) Xiao, J.-C.; Shreeve, J. M. J. Org. Chem. 2005, 70, 3073-3078. (10) Yoshida, Y.; Baba, O.; Saito, G. In preparation. (11) Andre, M.; Loidl, J.; Schottenberger, H.; Bentivoglio, G.; Wurst, K.; Ongania, K.-H. Anal. Chem. 2005, 77, 702-705. (12) (a) Bonhoˆte, P.; Dias, A.-P.; Armand, M.; Papageorgiou, N.; Kalyanasundaram, K.; Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168-1178. (b) Carmichael, A. J.; Seddon, K. R. J. Phys. Org. Chem. 2000, 13, 591-595. (c) Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. J. Chem. Soc., Perkin
Yoshida et al. Trans. 2 2001, 433-435. (d) Aki, S. N.; Brennecke, J. F.; Samanta, A. Chem. Commun. 2001, 413-414. (e) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc. 2002, 124, 14247-14254. (f) Kawai, A.; Hidemori, T.; Shibuya, K. Chem. Lett. 2004, 33, 1464-1465. (g) Fujisawa, T.; Fukuda, M.; Terazima, M.; Kimura, Y. J. Phys. Chem. A 2006, 110, 6164-6172. (13) (a) Dimroth, K.; Reichardt, C.; Siepmann, T.; Bohlmann, F. Justus Liebigs Ann. Chem. 1963, 661, 1-37. (b) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358. (14) (a) Kamlet, M. J.; Taft, R. W. J. Am. Chem. Soc. 1976, 98, 377383. (b) Taft, R. W.; Kamlet, M. J. J. Am. Chem. Soc. 1976, 98, 28862894. (c) Yokoyama, T.; Taft, R. W.; Kamlet, M. J. J. Am. Chem. Soc. 1976, 98, 3233-3237. (d) Kamlet, M. J.; Abboud, J. L.; Taft, R. W. J. Am. Chem. Soc. 1977, 99, 6027-6038. (15) Laurence, C.; Nicolet, P.; Dalati, M. T.; Abboud, J.-L. M.; Notario, R. J. Phys. Chem. 1994, 98, 5807-5816. (16) Marcus, Y. Chem. Soc. ReV. 1993, 22, 409-416. (17) Crowhurst, L.; Mawdsley, P. R.; Pe´rez-Arlandis, J. M.; Salter, P. A.; Welton, T. Phys. Chem. Chem. Phys. 2003, 5, 2790-2794. (18) (a) Crowhurst, L.; Lancaster, N. L.; Pe´rez-Arlandis, J. M.; Welton, T. J. Am. Chem. Soc. 2004, 126, 11549-11555. (b) Chiappe, C.; Pieraccini, D. J. Phys. Chem. A 2006, 110, 4937-4941. (19) (a) Vogel, H. Phys. Z. 1921, 22, 645-646. (b) Fulcher, G. S. J. Am. Ceram. Soc. 1923, 8, 339-355. (c) Tamman, G.; Hesse, W. Z. Anorg. Allg. Chem. 1926, 156, 245-257. (20) (a) McLean, A. J.; Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. Chem. Commun. 2002, 1880-1881. (b) Okoturo, O. O.; VanderNoot, T. J. J. Electroanal. Chem. 2004, 568, 167-181. (21) For examples: (a) McEwen, A. B.; Ngo, H. L.; LeCompte, K.; Goldman, J. L. J. Electrochem. Soc. 1999, 146, 1687-1695. (b) Hagiwara, R.; Matsumoto, K.; Nakamori, Y.; Tsuda, T.; Ito, Y.; Matsumoto, H.; Momota, K. J. Electrochem. Soc. 2003, 150, D195-D199. (c) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. J. Phys. Chem. B 2004, 108, 16593-16600. (d) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. J. Phys. Chem. B 2005, 109, 6103-6110. (e) Tokuda, H.; Ishii, K.; Susan, M. A. B. H.; Tsuzuki, S.; Hayamizu, K.; Watanabe, M. J. Phys. Chem. B 2006, 110, 2833-2839. (f) Zhou, Z.-B.; Matsumoto, H.; Tatsumi, K. ChemPhysChem 2005, 6, 1324-1332. (g) Yoshida, Y.; Saito, G. J. Mater. Chem. 2006, 16, 1254-1262. (22) (a) Xu, W.; Cooper, E. I.; Angell, C. A. J. Phys. Chem. B 2003, 107, 6170-6178. (b) Xu, W.; Wang, L.-M.; Nieman, R. A.; Angell, C. A. J. Phys. Chem. B 2003, 107, 11749-11756. (23) (a) Voronel, A.; Veliyulin, E.; Machavariani, V. Sh.; Kisliuk, A.; Quitmann, D. Phys. ReV. Lett. 1998, 80, 2630-2633. (b) Veliyulin, E.; Shasha, E.; Voronel, A.; Machavariani, V. Sh.; Seifer, Sh.; Rosenberg, Yu.; Shumsky, M. G. J. Phys.: Condens. Matter 1999, 11, 8773-8784. (24) Walden, P. Z. Phys. Chem. 1906, 55, 207-249. (25) (a) Law, G.; Watson, P. R. Langmuir 2001, 17, 6138-6141. (b) Huddleston, J. G.; Visser, A. E.; Reichert, W. M.; Willauer, H. D.; Broker, G. A.; Rogers, R. D. Green Chem. 2001, 3, 156-164. (c) Dzyuba, S. V.; Bartsch, R. A. ChemPhysChem 2002, 3, 161-166. (d) Abbott, A. P. ChemPhysChem 2004, 5, 1242-1246. (e) Bagno, A.; Butts, C.; Chiappe, C.; D’Amico, F.; Lord, J. C. D.; Pieraccini, D.; Rastrelli, F. Org. Biomol. Chem. 2005, 3, 1624-1630. (f) Rebelo, L. P. N.; Lopes, J. N. C.; Esperanc¸ a, J. M. S. S.; Filipe, E. J. Phys. Chem. B 2005, 109, 6040-6043. (g) Martino, W.; Fernandez de la Mora, J.; Yoshida, Y.; Saito, G.; Wilkes, J. Green Chem. 2006, 8, 390-397. (h) Tong, J.; Hong, M.; Guan, W.; Li, J.-B.; Yang, J.-Z. J. Chem. Thermodyn. 2006, 38, 1416-1421. (i) Zaitsau, D. H.; Kabo, G. J.; Strechan, A. A.; Paulechka, Y. U.; Tschersich, A.; Verevkin, S. P.; Heintz, A. J. Phys. Chem. A 2006, 110, 7303-7306. (j) Greaves, T. L.; Weerawardena, A.; Fong, C.; Drummond, C. J. J. Phys. Chem. B 2007, 111, 4082-4088. (26) (a) Fu¨rth, R. Proc. Cambridge Philos. Soc. 1941, 37, 252-275. (b) Fu¨rth, R. Proc. Cambridge Philos. Soc. 1941, 37, 281-290. (27) Bockris, J. O’M.; Reddy, A. K. N. Modern Electrochemistry; Plenum Press: New York, 1970; Vol. 1, Chapter 6. (28) Surface tensions of the DMI, EMI, BMI, and C6MI salts were tabulated in our recent paper (ref 25g). However, the values of the DMI (54.9 dyn cm-1) and EMI (49.1 dyn cm-1) salts are slightly different from the paper, presumably due to the insufficient drying prior to measurements. The data are available in our recent papers: (a) Larriba, C.; Garoz, D.; Bueno, C.; Romero-Sanz, I.; Castro, S.; Fernandez de la Mora, J.; Yoshida, Y.; Saito, G.; Hagiwara, R.; Matsumoto, K.; Wilkes, J. In Ionic Liquids: Not Just SolVents Anymore; Brennecke, J., Rogers, R. D., Seddon, K. R., Eds.; ACS Symposium Series; in press. (b) Garoz, D.; Bueno, C.; Larriba, C.; Castro, S.; Romero-Sanz, I.; Fernandez de la Mora, J.; Yoshida, Y.; Saito, G. J. Appl. Phys., in press. (29) Ye, C.; Shreeve, J. M. J. Phys. Chem. A 2007, 111, 1456-1461. (30) (a) Arduengo, A. J.; Gentry, F. P.; Taverkere, P. K.; Simmons, H. E. U.S. Patent 6,177,575, 2001. (b) Liu, J.; Chen, J.; Zhao, J.; Zhao, Y.; Li, L.; Zhang, H. Synthesis 2003, 2661-2666. (31) Xiao, J.-C.; Twamley, B.; Shreeve, J. M. Org. Lett. 2004, 6, 38453847.