12420
J. Phys. Chem. B 2008, 112, 12420–12430
Density, Refractive Index, Interfacial Tension, and Viscosity of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] in Dependence on Temperature at Atmospheric Pressure Andreas P. Fro¨ba,*,† Heiko Kremer,‡ and Alfred Leipertz‡ Erlangen Graduate School in AdVanced Optical Technologies, UniVersity of Erlangen-Nuremberg, Paul-Gordan-Strasse 6, D-91052 Erlangen, Germany, and Institute of Chemical and Bioengineering, Department of Engineering Thermodynamics, UniVersity of Erlangen-Nuremberg, Am Weichselgarten 8, D-91058 Erlangen, Germany ReceiVed: May 15, 2008; ReVised Manuscript ReceiVed: July 26, 2008
The density, refractive index, interfacial tension, and viscosity of ionic liquids (ILs) [EMIM][EtSO4] (1ethyl-3-methylimidazolium ethylsulfate), [EMIM][NTf2] (1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide), [EMIM][N(CN)2] (1-ethyl-3-methylimidazolium dicyanimide), and [OMA][NTf2] (trioctylmethylammonium bis(trifluoromethylsulfonyl)imide) were studied in dependence on temperature at atmospheric pressure both by conventional techniques and by surface light scattering (SLS). A vibrating tube densimeter was used for the measurement of density at temperatures from (273.15 to 363.15) K and the results have an expanded uncertainty (k ) 2) of (0.02%. Using an Abbe refractometer, the refractive index was measured for temperatures between (283.15 and 313.15) K with an expanded uncertainty (k ) 2) of about (0.0005. The interfacial tension was obtained from the pendant drop technique at a temperature of 293.15 K with an expanded uncertainty (k ) 2) of (1%. For higher and lower temperatures, the interfacial tension was estimated by an adequate prediction scheme based on the datum at 293.15 K and the temperature dependence of density. For the ILs studied within this work, at a first order approximation, the quantity directly accessible by the SLS technique was the ratio of surface tension to dynamic viscosity. By combining the experimental results of the SLS technique with density and interfacial tension from conventional techniques, the dynamic viscosity could be obtained for temperatures between (273.15 and 333.15) K with an estimated expanded uncertainty (k ) 2) of less than (3%. The measured density, refractive index, and viscosity are represented by interpolating expressions with differences between the experimental and calculated values that are comparable with but always smaller than the expanded uncertainties (k ) 2). Besides a comparison with the literature, the influence of structural variations on the thermophysical properties of the ILs is discussed in detail. The viscosities mostly agree with values reported in the literature within the combined estimated expanded uncertainties (k ) 2) of the measurements while our density and interfacial tension data differ by more than (1% and (5%. Introduction Ionic liquids (ILs) are molten salts with a melting point below the boiling point of water at atmospheric pressure. Most ILs are composed of an organic cation and an inorganic, polyatomic anion. ILs have attracted a rapid increasing interest during the last ten years. They represent an emerging class of versatile materials with a long history. Formerly used for specialized electrochemical applications, nowadays ILs are being considered in diverse areas of technical interest.1 These extend from energy2,3 over process engineering,4-7 biotechnology,8,9 material engineering10-13 to sensor technology.14 ILs offer many possibilities as solvents for catalytic reactions, chemical synthesis, and separation technology or as electrolytes in batteries, capacitors, fuel cells, solar cells, and chemical sensors. The potential use of ILs in a wide range of applications is mainly founded by their unique properties. They are liquid over a wide * Author to whom correspondence should be addressed. Telephone: +499131-85-29789. Fax: +49-9131-85-29901. E-mail:
[email protected]. † Erlangen Graduate School in Advanced Optical Technologies, University of Erlangen-Nuremberg. ‡ Institute of Chemical and Bioengineering, Department of Engineering Thermodynamics, University of Erlangen-Nuremberg.
temperature range, nearly nonvolatile, nonflammable, relatively thermally and electrically stable and they possess an excellent dissolving power for a wide range of inorganic and organic materials. Due to an almost unlimited number of potential combinations of cations and anions, ILs can be tailored to a specific application. Yet, for identifying their usefulness, e.g., as solvents, knowledge about their physical and chemical properties is required. Of course, the properties of every conceivable IL cannot be obtained by carrying out appropriate measurements, since this would represent a substantial investment.1 An alternative approach, where the composition of an IL exhibiting a given set of properties can be predicted, would be very useful. For this, however, knowledge about the origins of the fundamental properties of ILs is necessary, resulting in a better understanding of structure-property relationships. Until now, for the design of new ILs with particular characteristics, trial-and-error methods are used. For overcoming this limit, quantitative prediction methods with reasonable certainty must be developed, see, e.g. refs.15-17 Yet, all prediction methods can only be as accurate as the experimental data used for the evaluation of their performance.
10.1021/jp804319a CCC: $40.75 2008 American Chemical Society Published on Web 09/04/2008
Properties of Ionic Liquids At present, one impediment for the progress in the field of ILs is the lack of reliable data.18-20 For many physicochemical properties of ILs only a single datum near room temperature, usually (293 or 298) K, and at atmospheric pressure is reported in the literature. Most of these studies are dedicated to characterize an IL after its synthesis and are of limited reliability. It is fairly common to find differences between published data sets, which are more than one order of magnitude larger than the combined uncertainties of the measurements. This situation holds both for equilibrium data and for transport properties, whereas it is more pronounced for the latter. Meanwhile it is well-known that impurities, e.g., arising from the sample preparation or the absorption of atmospheric moisture, can significantly affect the thermophysical properties of ionic liquids.21-25 Beside sample handling and preparation procedures, the current insufficient data situation for pure ionic liquids as well as their mixtures with cosolvents may also be caused by the use of routine laboratory analytic methods, e.g., rotational viscometers,22,26,27 which are often overestimated regarding the achievable accuracy over a wide viscosity range. Moreover, taking into account a strong dependence of the viscosity of ionic liquids on temperature, even problems regarding its control and determination could be responsible for the disagreement between published data sets.27 The current situation, however, is hindering both a more fundamental understanding of the structureproperty relationships of ILs as well as their successful and large-scale application. Here, a substantial contribution of scientists who are engaged in the field of thermophysical properties would be helpful. The major aim of the present work was to provide accurate and reliable data for the dynamic viscosity of a selected set of pure ILs by surface light scattering (SLS). In comparison to conventional viscometers, the main advantage of the SLS technique relies on the possibility of determining the viscosity in macroscopic thermodynamic equilibrium, and this in an absolute way, without the need of any calibration procedure involving a fluid of known viscosity.28-30 The fluids investigated were the newly developed [OMA][NTf2] (trioctylmethylammonium bis(trifluoromethylsulfonyl)imide), as well as the imidazolium-based ILs [EMIM][EtSO4] (1-ethyl-3-methylimidazolium ethylsulfate), [EMIM][NTf2] (1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide), and [EMIM][N(CN)2] (1ethyl-3-methylimidazolium dicyanamide). The latter were selected because they are most promising for technical application. Moreover, the tetraalkylammonium-based IL was included as a representative of another class of ILs for which only few data are available. For high viscosity fluids, as it is relevant for the ILs mentioned above, the evaluation of the viscosity form SLS requires interfacial tension and density data. Beside refractive index, these were measured as a function of temperature by conventional techniques. The investigations, however, should point out the influence of the anion on the thermophysical properties of [EMIM]-based ILs and the effect of the cation structure for ILs with the [NTf2]-anion. In the following experimental section, some information about the used conventional techniques and SLS method is given. The sample preparation procedure, experimental conditions and the uncertainties are described and stated in detail. Subsequently, with the results for density, refractive index, interfacial tension, and dynamic viscosity, first the influence of structural variations on the thermophysical properties of the ILs is discussed. Finally, the data of the present work are compared with available literature data originating from other techniques.
J. Phys. Chem. B, Vol. 112, No. 39, 2008 12421 TABLE 1: Molecular Weight M, Concentration of Water by Mass w, and Nominal Purity of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] Studied in This Work [EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] M/g · mol-1 w/ppm purity/%
236.29 1051 >99
391.34 509 >99
177.22 1886 >98
648.86 523 >98
Experimental Section Materials and Sample Preparation Procedure. [EMIM][EtSO4] was obtained by the synthesis procedure described by Maier and co-workers.31 [EMIM][NTf2] was synthesized by reacting equimolar amounts of [EMIM][EtSO4] and [Li][NTf2] in aqueous solution at 60 °C under vigorous stirring. The upper aqueous phase was decanted and the lower phase was separated, washed with distilled water, and dried under reduced pressure and elevated temperature. The purity of [EMIM][EtSO4] and [EMIM][NTf2] was proved by 1H NMR analysis (JEOL, ECX +400 spectrometer), where dimethylsulfoxide-d6 (DMSO-d6) was used as solvent. The total peak integral in the 1H NMR spectrum was found to correspond to a nominal purity higher than 99%. [EMIM][N(CN)2] and [OMA][NTf2] were purchased from Solvent Innovation GmbH, Cologne, Germany with a nominal purity higher than 98%. Before use, the ILs were dried at about 333.15 K for a time period of 3 h on a vacuum line (0.5 mbar) with an oil-sealed vacuum pump and a liquid nitrogen trap. For the dried ILs the concentration of water was proved by Karl Fischer coulometric titration (Metrohm, 756 KF Coulometer). The expanded uncertainty (k ) 2) of the water content determinations performed within this work is estimated to be less than (20%. The molecular weight, nominal purity and water content of the ILs studied in this work are summarized in Table 1. All parts of the measuring devices, which were in contact with the sample, as well as all glassware used for sample handling were cleaned, rinsed with double-distilled water, and oven-dried. Furthermore, all transfer of samples was performed under an argon atmosphere. Vibrating Tube MethodsDensity. Density measurements at atmospheric pressure are based on the vibrating tube method. For the density meter (Anton Paar, DMA 5000) used here, longterm drift is eliminated by a reference oscillator built into the measuring cell and only one adjustment at 293.15 K is sufficient to reach a high accuracy for the whole measuring temperature range. The DMA 5000 allows a full-range viscosity correction, whereby all viscosity related errors inherent to all known types of oscillating U-tube density meters are automatically eliminated. The temperature of the U-tube is controlled within (1 mK and measured by a high-precision platinum resistance probe with an uncertainty of (10 mK. For the density meter calibration, standard water and air were used. The calibration procedure was checked by measuring the liquid density of toluene at atmospheric pressure for temperatures between (278.15 and 343.15) K, in intervals of 5 K. Here, the difference between the density values determined by our densimeter and those calculated by the equation of state by Lemmon and Span32 are smaller than 0.01%. For toluene the uncertainty in the equation of state by Lemmon and Span32 for the saturated liquid density approaches 0.01% around 300 K. However, the expanded uncertainty (k ) 2) of the present density measurements for ILs is estimated to be less than (0.02%. For this, the calibration error of the apparatus of 0.01% and the error associated with the following measurement procedure for ILs have been taken into account. The precision or repeatability of the instrument was better than (0.001%.
12422 J. Phys. Chem. B, Vol. 112, No. 39, 2008 Abbe RefractometersRefractive Index. The refractive index nD at the sodium line (λD ) 589.3 nm) and the refractive index difference nF - nC for the Fraunhofer lines F (λF ) 486.1 nm) and C (λC ) 656.3 nm) were measured with an Abbe refractometer (Leo Kuebler, R 6000 G). The temperature of the samples was controlled with a laboratory thermostat within (0.1 K and measured by a mercury thermometer with an uncertainty of (0.5 K. The refractometer was calibrated with water. The expanded uncertainty (k ) 2) in the measurement of the refractive index is estimated to be less than (0.0005. Pendant Drop TechniquesInterfacial Tension. For the evaluation of the dynamic viscosity from SLS, interfacial tension data are needed in the case of fluids of high viscosity and/or low interfacial tension. For an accurate determination of the dynamic viscosity of high viscosity fluids, however, one has to ensure that the liquid surface under investigation corresponds to the interfacial tension values used for data evaluation. This was ensured by investigating identical samples of ILs by SLS and the pendant drop method for measuring the interfacial tension. Here, a universal surface analyzer (OEG, SURFTENS universal) was used, where the geometrical profile of a pendant drop is compared with the theoretical drop profile obtained from the Laplace equation. The measurements were performed inside an optical glass cell for photometry (Hellma, 402.000) at well defined conditions for a temperature of (293.15 ( 0.1) K. For the interfacial tension data of all ILs, the expanded uncertainty (k ) 2) was estimated to be less than (1%. Surface Light Scattering (SLS)sViscosity. SLS analyzes the dynamics, i.e., frequency ωq and/or damping Γ or mean lifetime τC ) 1/Γ, of surface fluctuations at a given wavenumber q present at the phase boundary of the fluid system under investigation.33,34 At low viscosity and/or large interfacial tension, surface fluctuations show an oscillatory behavior and both properties are determined simultaneously by SLS.28 For high viscosity, as it is relevant for the investigations within this work, surface fluctuations are overdamped and do not propagate (ωq ) 0). In this case, analyzing the mean lifetime of surface fluctuations, τC ≈ 2 η/(σ q), at a first approximation SLS gives only access to the ratio of dynamic viscosity η to interfacial tension σ.29,30 Even though in the case of high viscosity fluids interfacial tension data are needed for the evaluation of the viscosity, the application of the SLS method is of particular interest. In contrast to conventional viscometers, in which the fluid is subjected to shear stress resulting in a gradient in the fluid velocity, SLS allows the determination of viscosity in macroscopic thermodynamic equilibrium. Furthermore, while almost all conventional methods determine the viscosity in a relative manner, for the SLS technique no calibration procedure using a fluid of known viscosity is needed. For details about the fundamentals and methodological principles of the SLS technique, the reader is referred to the specialized literature.28-30,33-36 The experimental setup used here for the investigation of ILs is the same as that employed in our former SLS investigations for numerous pure refrigerants,37 refrigerant mixtures,38 and reference fluids.28,29 The temperature of the SLS measuring cell was measured with two calibrated 100 Ω platinum resistance probes integrated into the main body of the vessel, with a resolution of 0.25 mK, using an ac bridge (Anton Paar, MKT 100). The uncertainty of the absolute temperature measurement was less than (15 mK. The temperature stability during an experimental run was better than (1 mK. For each temperature, at least six measurements at different wave numbers q of surface fluctuations were performed.
Fro¨ba et al. TABLE 2: Density G of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] from T ) (273.15 to 363.15) K at Atmospheric Pressure T/K 273.15 278.15 283.15 288.15 293.15 298.15 303.15 308.15 313.15 318.15 323.15 328.15 333.15 338.15 343.15 348.15 353.15 358.15 363.15
[EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] F/kg · m-3 F/kg · m-3 F/kg · m-3 F/kg · m-3 1254.00 1250.46 1246.91 1243.48 1240.09 1236.70 1233.32 1229.94 1226.56 1223.20 1219.85 1216.51 1213.17 1209.86 1206.55 1203.25 1199.96 1196.69 1193.43
1543.92 1538.82 1533.71 1528.61 1523.53 1518.45 1513.41 1508.39 1503.38 1498.39 1493.41 1488.46 1483.52 1478.61 1473.70 1468.81 1463.94 1459.08 1454.24
1125.97 1122.53 1119.10 1115.68 1112.28 1108.91 1105.55 1102.21 1098.89 1095.58 1092.29 1089.01 1085.75 1082.51 1079.29 1076.08 1072.89 1069.71 1066.55
1116.60 1112.48 1108.59 1104.80 1101.01 1097.22 1093.43 1089.64 1085.96 1082.28 1078.58 1074.89 1071.20 1067.51 1063.83 1060.16 1056.49 1052.83
In the present work, however, data for the dynamic viscosity η of ILs were obtained by an exact description of the dynamics of surface fluctuations in the case of a free liquid surface.33 For this, data obtained for the dynamics of surface waves, i.e. the mean lifetime τC at a defined wave vector q, were combined with reference data for the interfacial tension σ and density F. For the latter, even approximate or less accurate values allow a successful determination of the dynamic viscosity.29,30 For [EMIM][EtSO4] at a temperature of 313.15 K, for example, an uncertainty in the density of (10% would result in an uncertainty of about (0.1% for the dynamic viscosity. Although a first order approximation, τC ≈ 2 η/(σ q), for data evaluation does not allow the determination of the dynamic viscosity with high accuracy, the approximation can be applied to get a good estimate for the uncertainty of our SLS results in an analytical manner by combining in quadrature the errors for the interfacial tension σ, mean lifetime τC, and wave vector q. For the relative uncertainty of the interfacial tension σ between (273.15 and 333.15) K a value of (2.0% has been estimated. Taking into account relative uncertainties of clearly less than ((0.5 and (0.15)% for the mean lifetime τC and wave vector q, respectively, the expanded uncertainty (k ) 2) of our values for the dynamic viscosity η of ILs investigated in this work is estimated to be less than (3.0%. Results and Discussion In the following, our results for different ILs are discussed sequentially for the density, refractive index, surface tension, and dynamic viscosity. In the subsections, first the properties are interpreted with regard to the structure of the ions and the nature of their interactions. Then our results are discussed in detail in comparison with literature. While for [EMIM][EtSO4] and [EMIM][NTf2] numerous data are available in literature, for [EMIM][N(CN)2] and [OMA][NTf2] only a limited number of investigations is reported. Density. Our experimental density data obtained at atmospheric pressure are summarized in Table 2 and in Figure 1. The densities for all investigated ILs can be represented by a polynomial of second order,
F ) F0 + F1T + F2T2
(1)
where all data have been taken into account with the same statistical weight. In eq 1, T is the temperature in K and F0, F1,
Properties of Ionic Liquids
J. Phys. Chem. B, Vol. 112, No. 39, 2008 12423
Figure 1. Liquid density F of ionic liquids at atmospheric pressure as a function of temperature: O, [EMIM][EtSO4]; 0, [EMIM][NTf2]; ∆, [EMIM][N(CN)2]; ], [OMA][NTf2].
TABLE 3: Coefficients of Equation 1 for the Density of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] listed in Table 2
Figure 3. Comparison between the densities F of different authors for [EMIM][NTf2] at atmospheric pressure for temperatures between (273.15 and 363.15) K (s, eq 1 using coefficients from Table 3): O, Fro¨ba et al., this work; ∆, Jacquemin et al.;27 3, Krummen et al.;46 2, Heintz et al.;51 [, Fredlake et al.;39 · · · , Tokuda et al.;40 9, Tokuda et al.;55 ], Noda et al.;52 0, Bonhoˆte et al.;49 b, Gardas et al.;50 x, Lopes et al.;53 1, Hong et al.;56 !, Dzyuba and Bartsch.57
F0/kg · m-3 F1/kg · m-3 · K-1 F2/kg · m-3 · K-2 rmsa [EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] a
1461.271 1850.643 1339.847 1354.662
-0.82497 -1.21781 -0.87546 -0.94094
0.24083 × 10-3 0.34773 × 10-3 0.33846 × 10-3 0.30299 × 10-3
0.0029 0.0005 0.0003 0.0070
Standard percentage deviation of F to the fit.
Figure 2. Comparison between the densities F of different authors for [EMIM][EtSO4] at atmospheric pressure for temperatures between (273.15 and 363.15) K (s, eq 1 using coefficients from Table 3): n, Fro¨ba et al., this work; -, Yang et al.;42,43 ∆, Jacquemin et al.;27 [, Arce et al.;44 +, Rodrı´guez and Brennecke;26 1, Blanchard et al.;41 3, Krummen et al.;46 x, Vila et al.;45 0, Go´mez et al.47
and F2 are the fit parameters given in Table 3. For all ILs the deviations of the experimental density data from their correlation according to eq 1 lie clearly within the measurement uncertainty of (0.02%. In general, the densities of the ILs containing the [EMIM]cation increase with increasing molecular weight of the associated anion. Fredlake and co-workers39 have already described this dependency for imidazolium-based ILs. They found this behavior for anions that are small enough to easily occupy a position close to the relatively large cation. The difference in density between ILs containing the [NTf2]-anion can be explained by the different size of the associated cation. The IL with the larger [OMA]-cation exhibits lower densities than that with the smaller [EMIM]-cation. This is in agreement with investigations of ILs containing 1-alkyl-3-methylimidazoliumcations, where decreasing densities with increasing alkyl chain length could be observed.40 Figure 2 shows for [EMIM][EtSO4] the deviations of our experimental density data from eq 1 as well as for that available in the literature. With the exception of the data of Blanchard et al.41 and Yang et al.,42,43 who measured the density using a
pycnometer and a Westphal balance with an uncertainty of (0.007 g · cm-3 and (0.0001 g · cm-3, all other data sets are based on the vibrating U-tube method and instruments of the same manufacturer. Moreover, Arce et al.44 and Vila et al.45 as well as Krummen et al.46 and Rodrı´guez and Brennecke26 used even identical vibrating U-tube instruments. While Arce et al.,44 Vila et al.,45 and Rodrı´guez and Brennecke26 specify the uncertainty of their measurements to be (0.00001 g · cm-3, (0.0001 g · cm-3, and (0.00005 g · cm-3, respectively, corresponding information can not be found for the data of Krummen et al.46 The density measurements reported by Jacquemin et al.27 and Go´mez et al.47 are stated with uncertainties of (0.001 g · cm-3 and (0.00001 g · cm-3. Taking into account both the error connected with the calibration and that of the subsequent measuring procedure, some of the above stated uncertainties for the vibrating U-tube method seem to be underestimated, i.e. below the expected values. As it can be seen from Figure 2, with the exception of the data by Blanchard et al.,41 the deviations between our data and those from literature are generally larger than the combined estimated expanded uncertainties of the measurements. The same statement holds also for the comparison of the literature data among themselves. The differences, however, can not be explained by the differences in the water content, which varied within the different studies between 23 and 1051 ppm. Such a difference would result in a maximum difference for the density of [EMIM][EtSO4] of about 0.02% as revealed in our work for IL cosolvent mixtures.48 The large deviations between the different data sets of up to about 1% may be explained by different and/or undefined impurities, which can affect the thermophysical properties of ILs to this extent. In Figure 3 the deviations between the correlation, eq 1, developed from our experimental results and literature data are shown for [EMIM][NTf2]. Besides data from the already mentioned refs 27 and 46, see discussion above for [EMIM][EtSO4], values for the density included in Figure 3 are obtained with the vibrating U-tube method with the exception of data from Fredlake et al.,39 who utilized a pycnometer with an uncertainty between ((0.0015 and 0.0023) g · cm-3. Bonhoˆte et al.49 did not specify the applied technique and its uncertainty at all. In contrast to [EMIM][EtSO4], for [EMIM][NTf2] it seems that the authors do not underestimate the uncertainty achievable by the vibrating U-tube method. In detail, Gardas et al.,50 Heintz et al.,51 and Noda et al.52 claim uncertainties of less than (0.001 g · cm-3, (0.0002 g · cm-3, and (0.001 g · cm-3, respectively.
12424 J. Phys. Chem. B, Vol. 112, No. 39, 2008
Fro¨ba et al.
TABLE 4: Refractive Index nD and Refractive Index Difference nF - nC of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] from T ) (283.15 to 313.15) K at Atmospheric Pressure [EMIM][EtSO4]
[EMIM][NTf2]
[EMIM][N(CN)2]
[OMA][NTf2]
T/K
nD
nF - nC
nD
nF - nC
nD
nF - nC
nD
nF - nC
283.15 293.15 303.15 313.15
1.4830 1.4800 1.4774 1.4749
0.0096 0.0094 0.0094 0.0092
1.4272 1.4244 1.4216 1.4186
0.0098 0.0082 0.0080 0.0080
1.5140 1.5110 1.5092
0.0140 0.0140 0.0139
1.4413 1.4388 1.4359 1.4328
0.0066 0.0073 0.0074 0.0072
Lopes et al.53 state for their density measurements for [EMIM][NTf2] at atmospheric pressure a precision of 0.001% and estimate the accuracy to be one or two orders of magnitude lower taking into account the inconsistencies associated with the calibrating fluids’ data and sample’s impurities. Krummen et al.,46 Bonhoˆte et al.,49 Tokuda et al.,54,55 Hong et al.,56 and Dzyuba and Bartsch57 give no information about the experimental uncertainty. In addition, a correlation for the density of [EMIM][NTf2] by Tokuda et al.40 is depicted in Figure 3, based on measurements with a vibrating U-tube densimeter, whereas again no uncertainty is specified. As it can be seen from Figure 3, with the exception of the datum by Heintz et al.,51 the data set by Fredlake et al.,39 and the two data points by Gardas et al.50 at the highest and lowest temperature, the maximum deviations between the different data sets are generally not larger than the estimated combined uncertainties of the measurements. The datum by Heintz et al.51 deviates about 0.45% with respect to our data correlation. The data by Krummen et al.,46 Bonhoˆte et al.,49 Tokuda et al.,54,55 Hong et al.,56 and Dzyuba and Bartsch,57 however, show relatively good agreement with our data with maximum deviations smaller than 0.05%, 0.10%, 0.17%, 0.01%, and 0.04%, respectively. For the data correlation published by Tokuda et al.40 in 2005 for [EMIM][NTf2], the maximum deviation from this work is 0.23%. It should be noted that the maximum deviation between the different data sets of Tokuda et al.40,54,55 is about 0.2%. In general, the scatter of the different data sets is somewhat smaller than that found for [EMIM][EtSO4]. For the density of [EMIM][N(CN)2] only two values can be found in literature, namely one datum by Yoshida et al.58 and another by MacFarlane et al.59,60 Yoshida et al.58 used a pycnometer with an unspecified uncertainty. Vice versa MacFarlane et al.59,60 claim an uncertainty of (5%, whereas the method is not stated. However, the values of Yoshida et al.58 and MacFarlane et al.59,60 show deviations of -2.9% and -4.4% from our data correlation, eq 1. Also for the density of [OMA][NTf2] a scarce data situation can be found. At present, only one experimental study is available in literature by Kilaru et al.,61 who report values for
Figure 4. Refractive index nD of ionic liquids at atmospheric pressure as a function of temperature: O, [EMIM][EtSO4]; 0, [EMIM][NTf2]; 4, [EMIM][N(CN)2]; ); [OMA][NTf2].
TABLE 5: Coefficients of Equation 2 for the Refractive Index n of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] listed in Table 4 [EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] a
n0
n1/K-1
(∆n/∆λ)/m-1
rmsa
1.55903 1.50822 1.58176 1.52187
-0.000269 -0.000286 -0.000240 -0.000284
-55 400 -50 000 -81 900 -42 000
0.009 0.004 0.018 0.011
Standard percentage deviation of nD to the fit.
the interfacial tension and density of [OMA][NTf2]. Their data obtained between 298.15 and 333.15 K by a gravity bottle with an uncertainty of (0.0005 g · cm-3 deviate between 1.7% and 1.2% from our data correlation, eq 1. Of course, these deviations are clearly outside the estimated combined uncertainties of the measurements. Refractive Index. The data obtained for the refractive index nD and refractive index difference nF - nC of ILs [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] in the temperature range from 283.15 to 313.15 K at atmospheric pressure are summarized in Table 4 and shown in Figure 4. The refractive index n in dependence on temperature and wavelength at atmospheric pressure can well be represented by a linear equation,
n ) n0 + n1T +
∆n (λ - λD) ∆λ
(2)
Here, T is the temperature in K, λ the wavelength in m, and λD ()589.3 × 10-9 m) the wavelength of the sodium vapor line. The fit parameters of eq 2, n0 and n1, as well as the mean dispersion ∆n/∆λ are given in Table 5. The mean dispersion, which was assumed to be independent of temperature, represents the mean value of the measured refractive index differences nFnC listed in Table 4. For all ILs the deviations of the experimental refractive index data nD from their correlation according to eq 2 are clearly within the measurement uncertainty of (0.0005. The data of this work are in good agreement with the behavior of the refractive index of fluids in general as described by Brocos et al.62 and studied especially for ILs by Deetlefs et al.17 They report that the larger the reduced molar volume, i.e., the unoccupied part of the molar volume of a substance, the smaller the refractive index. For all [EMIM]-based ILs investigated within this work an increasing refractive index with decreasing molar volume is found. Furthermore, when taking the refractive indices of both ILs containing the [NTf2]-anion into consideration, again an increasing refractive index with decreasing molar volume is assessed. For a comparison of the refractive index of the ILs, only for [EMIM][EtSO4] and [EMIM][NTf2] data are available in literature. Figure 5 includes measurements of the refractive index of [EMIM][EtSO4] by Go´mez et al.,47 who studied the temperature dependence between (288.15 and 343.15) K, and by Arce et al.44 at a single temperature of 298.15 K. Both Go´mez et al.47 and Arce et al.44 measured the refractive index of
Properties of Ionic Liquids
J. Phys. Chem. B, Vol. 112, No. 39, 2008 12425
Figure 5. Comparison between the refractive indices nD of different authors for [EMIM][EtSO4] at atmospheric pressure for temperatures between (283.15 and 313.15) K (s, eq 2 using coefficients from Table 5): O, Fro¨ba et al., this work; [, Arce et al.;44 0, Go´mez et al.47
TABLE 6: Surface Tension σ20°C of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] at T ) 293.15 K and Atmospheric Pressure [EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] σ20°C/mN · m-1
47.31
35.59
44.14
27.82
[EMIM][EtSO4] with an Abbe refractometer with an uncertainty of (0.00004. Within combined estimated uncertainties of the measurements, good agreement of our data can be found with those by Arce et al.44 and Go´mez et al.47 It is worth mentioning that the data by Arce et al.44 and Go´mez et al.47 are clearly outside their combined uncertainty. This may be explained not only by the sample purity but also by an overestimation of experimental accuracy. For the refractive index of [EMIM][NTf2], a single datum of 1.4231 at a temperature of 293.15 K is reported by Bonhoˆte et al.49 The authors give no information about the used measuring technique and its uncertainty. Their value, however, deviates about -0.09% from our datum, which is clearly outside the expanded uncertainty (k ) 2) of about (0.0005 of this work. Interfacial Tension. Interfacial tension values obtained within this work at a single temperature of 293.15 K at atmospheric pressure are summarized in Table 6. From these values and the temperature dependency of density, the interfacial tension was predicted for temperatures between (273.15 and 333.15) K for evaluating the dynamic viscosity from the SLS experiment, see experimental section. The prediction method is based on the MacLeod-Sudgen correlation given in ref 63. MacLeod64 suggested a relation between the interfacial tension σ, the liquid density Fl, and vapor density FV, σ1/4 ) C (Fl-FV), where C is a constant. Sugden65 modified MacLeod’s expression and introduced the temperature-independent parameter, the parachor, P ) C M, which can be estimated from the structure of the molecule. Taking into account a negligible vapor pressure or density for ILs, for the prediction of the temperature dependency of interfacial tension only a single experimental value for the interfacial tension at an arbitrary temperature and the knowledge of the temperature dependency of the density is needed. Within this work, however, the interfacial tension σ at any temperature was predicted by
σ ) σ20°C(F ⁄ F20°C)4
(3)
where σ20 °C and F20 °C are the measured interfacial tension and density at the reference temperature of 20 °C. The proposed prediction scheme represents the interfacial tension of high viscosity fluids typically with an uncertainty of less than 2%, which was tested for several reference fluids, see, e.g., ref 29.
Figure 6. Comparison between the interfacial tension σ of different authors for [EMIM][EtSO4] at atmospheric pressure for temperatures between (273.15 and 333.15) K (s, eq 3 using coefficients from Table 6 and density data according to eq 1): O, Fro¨ba et al., this work; -, Yang et al.;42 b, Martino et al.;67 0, Go´mez et al.,47 – –, Knotts et al., QSPR.68
Until now there is only a limited number of studies about the interfacial tension of ILs available in the literature. Law and Watson66 were one of the first to investigate the interfacial tension of several ILs. They found that ILs containing the same anion show decreasing interfacial tension values with increasing alkyl-chain length in a homologous series of cations. Furthermore, they revealed that the interfacial tension of ILs containing the same cation increases with increasing molecular size of the anion, which were highly symmetrical molecules in their study. Our results comply well with their first finding. We found a lower surface tension for [OMA][NTf2], having a larger molar volume in comparison to [EMIM][NTf2]. Yet, our results for the [EMIM]-based ILs with symmetric anions reveal that the IL containing the small [N(CN)2]-anion shows a higher interfacial tension than that with the clearly larger [NTf2]-anion. Thus, the interfacial tension of ILs containing the same cation should be strongly influenced by the strength of electrostatic interactions between anion and cation. Tokuda et al.55 explain that the interactions between different anions and the same cation are governed by the anionic donor ability (Lewis basicity) of the anions. Furthermore, they found the anionic donor ability to be high for anions having locally large negative charges with an asymmetric distribution and to be low for anions having symmetrically distributed low negative charges. In this context we found the highest interfacial tension for the IL with the [EtSO4]-anion, which is characterized by the lowest charge delocalization of all anions in this study. Figure 6 shows the deviations of available interfacial tension data for [EMIM][EtSO4] from eq 3. Included are experimental data by Go´mez et al.,47 who measured the interfacial tension between (288.15 and 313.15) K with the pendant drop technique with an accuracy of 0.1 mN · m-1. Yang et al.42 performed interfacial tension measurements with the forced bubble method between (278.15 and 328.15) K with a stated accuracy of 0.01 mN · m-1. Yet, a closer look at the results of Yang et al.42 reveals that they have mistakenly reported all their interfacial tension values one order of magnitude too low. Here, their interfacial tension values and also their experimental uncertainty has been hence assumed to be one order of magnitude larger. Martino et al.67 applied the capillary rise method to measure the interfacial tension of [EMIM][EtSO4] at a single temperature of 296.15 K without specifying explicitly the experimental accuracy. Within combined uncertainties Figure 6 shows agreement between our data and that of Go´mez et al.47 In contrast to this, the data by Yang et al.42 deviate with respect to eq 3 as well as to the data of Go´mez et al.47 between (3 and 7)%, which is clearly outside the estimated combined uncertainties. The same statement holds
12426 J. Phys. Chem. B, Vol. 112, No. 39, 2008
Fro¨ba et al. TABLE 7: Dynamic Viscosity η of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] from T ) (273.15 to 333.15) K at Atmospheric Pressurea T/K
[EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] η/mPa · s η/mPa · s η/mPa · s η/mPa · s
273.15 278.15 283.15 293.15 303.15 313.15 323.15 333.15
Figure 7. Comparison between the interfacial tension σ of different authors for [EMIM][NTf2] at atmospheric pressure for temperatures between (273.15 and 333.15) K (s, eq 3 using coefficients from Table 6 and density data according to eq 1): O, Fro¨ba et al., this work; 3, Kilaru et al.;61 b, Martino et al.;67 !, Dzyuba and Bartsch;57 •, Rebelo et al.;69 – –, Knotts et al., QSPR.68
for the datum by Martino et al.,67 which deviates from eq 3 by -9.5%. Martino et al.67 attribute the discrepancies between their results and that of Yang et al.42 to a higher water content in the sample of the latter. Yet, our previous work for IL cosolvent mixtures48 points up that the interfacial tension of [EMIM][EtSO4] shows no significant change for a concentration of water up to about 11% by mass. Thus, the discrepancies between the different data sets cannot be explained by insufficient drying precautions during sample handling as mentioned by Martino et al.67 Similar to the findings for the density of [EMIM][EtSO4], the main reasons for the discrepancies may be found in the presence of other impurities other than water or in an overestimation of experimental accuracy. Finally, in Figure 6 calculated values for the interfacial tension of [EMIM][EtSO4] are included, which are based on a quantitative structure-property relationship (QSPR) using existing parachor contribution data for neutral compounds by Knotts et al.68 Of course, the electrostatic interactions lacking in that method can be found for systems where ions participate. The calculation of parachors of a range of ionic liquids, however, using neutral parachor contribution values was first performed by Deetlefs et al.17 They showed that despite parachor contribution data for neutral organics being employed in the calculations, the differences between the corresponding calculated and experimental values for the parachor are remarkably small. In detail, they found that for numerous ILs based on 1-alkyl-3methylimidazolium cations, the deviations between the predicted and experimental interfacial tensions are smaller than 5%. For evaluating the interfacial tension of the ILs studied within this work from parachor contribution data by Knotts et al.,68 information about the molecular weight M and density F is required, which were adopted from Table 1 and eq 1. With respect to eq 3, the interfacial tension values based on the parachor contribution data by Knotts et al.68 show for the whole temperature range a deviation of about +13%. This relatively large difference may be explained by the fact that the parachors calculated for the ILs using neutral contribution data do not account for electrostatic interactions present in the ILs. The constant offset between our predicted values, eq 3, and the corresponding data by Knotts et al.68 is attributable to the consistent density values, eq 1, used for both predictions. In Figure 7 the deviations between eq 3 and literature data are shown for the interfacial tension of [EMIM][NTf2]. Beside the two values by Martino et al.67 at 296.65 K also the single datum by Dzyuba and Bartsch57 at 298.15 K is based on the capillary rise method. Dzyuba and Bartsch57 do not specify their
458.3 315.9 224.0 125.4 77.8 50.6 35.6 25.2
99.2 75.7 60.1 38.6 27.1 19.4 14.9 11.8
52.1 33.8 22.8 15.4 11.4 -
3729.6 2324.7 1473.7 676.3 359.9 200.7 121.4 77.7
a Directly measured values of the mean decay time τC at a defined wave vector q of surface fluctuations were combined with interfacial tension data σ, eq. 3, and density data F, eq 1 using coefficients from Table 3, to derive η by an exact numerical solution of the dispersion relation.
Figure 8. Arrhenius plots of dynamic viscosity η for ionic liquids: O, [EMIM][EtSO4]; 0, [EMIM][NTf2]; ∆, [EMIM][N(CN)2]; ], [OMA][NTf2].
TABLE 8: Coefficients of Equation 4 for the Viscosity of Ionic Liquids [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] listed in Table 7 [EMIM][EtSO4] [EMIM][NTf2] [EMIM][N(CN)2] [OMA][NTf2] η0/mPa · s B/K C/K rmsa a
0.22348 751.942 174.477 0.84
0.28421 612.09 168.635 0.66
0.00624 1766.797 77.541 0.96
0.07478 1167.259 165.226 0.92
Standard percentage deviation of η to the fit.
experimental uncertainty as already mentioned for the work of Martino et al.67 The measurements by Kilaru et al.61 were performed with a du Nou¨y tensiometer, a method for which it is questionable if low uncertainties can be achieved. Although Kilaru et al.61 report to have checked the tensiometer with water at 293 K, where the deviations from a value of 72.8 mN · m-1 were smaller than 0.1%, the uncertainty of their measurements for ILs is not explicitly stated. They state only that the standard deviation for all their measurements on a single substance was smaller than 0.05 mN · m-1. Neither the method nor its experimental uncertainty can be found in the work of Rebelo et al.,69 who report an interfacial tension of [EMIM][NTf2] in a range between (34.7 and 32.8) mN · m-1 for temperatures between (302 and 338) K. Finally, in Figure 7 we compare interfacial tension data for [EMIM][NTf2] based on the parachor contribution data by Knotts et al.68 As can be seen from Figure 7, for the datum given by Dzyuba and Bartsch57 at 298.15 K a positive deviation of 11% from our values can be found. The data by Kilaru et al.61 show with respect to our data systematic positive deviations of up to 24%, which are clearly outside the estimated combined uncertainties. In contrast, for the interfacial tension of [EMIM][NTf2] our
Properties of Ionic Liquids
J. Phys. Chem. B, Vol. 112, No. 39, 2008 12427
values are in good agreement with the data given by Martino et al.67 within their experimental uncertainty. For temperatures up to 320 K, our values are also in agreement with the range given for the interfacial tension of [EMIM][NTf2] by Rebelo et al.69 With increasing temperatures, the difference between their and our data becomes larger, yet the deviations are less than 4% for the highest temperatures studied in this work. In comparison to [EMIM][EtSO4], for [EMIM][NTf2] a significantly smaller deviation of less than 3% can be found between our data and those based on the parachor contribution data by Knotts et al.68 Taking into account that the [NTf2]-anion in comparison to the [EtSO4]-anion exhibits a more pronounced charge delocalization, the behavior may indicate that the parachor contribution data for neutral compounds by Knotts et al.68 show a better performance for salts with smaller charge contribution of the anion. At present, only one experimental data set is available in the literature for the interfacial tension of [EMIM][N(CN)2] and [OMA][NTf2]. These are measurements of Kilaru et al.,61 which were performed by a du Nou¨y tensiometer, and of Martino et al.,67 which were obtained by the capillary rise method. The methods and their respective accuracies have already been discussed in context with [EMIM][EtSO4] and [EMIM][NTf2]; see above. While the deviation of the datum by Martino et al.67 from our data for [EMIM][N(CN)2] at a temperature of about 297 K is -2.7%, the corresponding deviations of the data set by Kilaru et al.61 for [OMA][NTf2] at temperatures from (299.46 and 336.79) K are between +(6.3 and 8.2)% and hence clearly exceed the combined estimated uncertainties. As observed for [EMIM][NTf2], also for [OMA][NTf2] interfacial tension data based on the parachor contribution data by Knotts et al.68 show good agreement with our data with a deviation of less than 2.5%. On the contrary, for [EMIM][N(CN)2] a deviation of 37% can be found between our data and that based on the parachor contribution data by Knotts et al.68 Dynamic Viscosity. The dynamic viscosity of ILs [EMIM][EtSO4], [EMIM][NTf2], [EMIM][N(CN)2], and [OMA][NTf2] in the temperature range (273.15 to 333.15) K at atmospheric pressure from SLS are summarized in Table 7 and Figure 8. Here, each data point represents the average value of at least six independent measurements for different moduli of the wave vector q, cf. Experimental Section. The drawn lines in Figure 8 represent the correlation of the viscosity data in the form of a Vogel equation,
η ) η0exp[B ⁄ (T - C)]
(4)
where T is the temperature in K and η0, B, and C are the fit parameters given in Table 8. Here, also the standard deviation (root-mean-square, rms) of our data relative to those calculated by eq 4 is listed. For the data correlation, the statistical weight of each data point has been assumed to be the same. The residuals of the viscosity data from the fit, eq 4, are clearly smaller than the expanded uncertainty (k ) 2) of less than (3%. It should be noted that for our viscosity data at 293.15 K, due to the lower uncertainty of the surface tension data used for data evaluation of the SLS experiment, even a lower expanded uncertainty (k ) 2) of less than (2% can be stated. In general, the viscosity of ILs is strongly governed by Coulomb forces like van der Waals interactions and hydrogen bonding. Bonhoˆte et al.49 revealed in their work about the viscosity of ILs that a stronger charge delocalization within the anion leads to the weakening of molecular interactions based on hydrogen bonding. If this effect is not overcompensated by increasing van der Waals interactions, the viscosity of the IL is
Figure 9. Comparison between the dynamic viscosity η of different authors for [EMIM][EtSO4] at atmospheric pressure for temperatures between (273.15 and 333.15) K (-, eq 4 using coefficients from Table 8): O, Fro¨ba et al., this work; +, Rodrı´guez and Brennecke;26 [, Arce et al.;44 ∆, Jacquemin et al.;27 0, Go´mez et al.47
Figure 10. Comparison between the dynamic viscosity η of different authors for [EMIM][NTf2] at atmospheric pressure for temperatures between (273.15 and 333.15) K (s, eq 4 using coefficients from Table 8): O, Fro¨ba et al., this work; ∆, Jacquemin et al.;27 +, Widegren et al.;21 ], Crosthwaite et al.;72 · · · · · · · · · , Tokuda et al.;40 9, Tokuda et al.;55 – –, Noda et al.;52 x, McEwen et al.;73 0, Bonhoˆte et al.;49 !, Dzyuba and Bartsch;57 b, Pringle et al.74
lowered. For the [EMIM]-based ILs investigated in this work, a decrease in viscosity from the [EtSO4]- over the [NTf2]- to the [N(CN)2]-anion was found. The highest viscosity of the IL containing the [EtSO4]-anion can be explained by strong molecular interactions due to hydrogen bonding. The lower viscosity of the [NTf2] salt can be explained by the charge delocalization in the [NTf2]-anion combined with a moderate molecular weight. The lower viscosity found for the [N(CN)2]in comparison to the [NTf2]-salt can be explained by the distinct lower molecular weight and size of the first one. The highest viscosity found for [OMA][NTf2] having the highest molecular weight of all ILs studied within this work is in agreement with the findings of Bonhoˆte et al.49 They report low viscosities for ILs where the cation has a sufficient side chain mobility and a low molecular weight. In the following, the results for the dynamic viscosity of ILs at atmospheric pressure from SLS are discussed in comparison to measurements originating from viscometers of different types and operating principles. For the comparison shown in Figures 9 and 10 for [EMIM][EtSO4] and [EMIM][NTf2] the correlation of our data, eq 4, serves as a basis. Data for the viscosity of [EMIM][EtSO4] included in Figure 9 comprise measurements of Arce et al.44 and Go´mez et al.,47 which were both performed by capillary viscometers. Arce et al.44 state for their investigated IL sample a water content of 98 ppm, yet no information about the experimental uncertainty can be found. In contrast, Go´mez et al.47 specify the water content and experimental uncertainty to be lower than 300 ppm and (0.01 mPa · s. With regard to the latter
12428 J. Phys. Chem. B, Vol. 112, No. 39, 2008 value, it is worth mentioning that all data sets from conventional techniques are based on a calibration of the viscometer with an adequate reference fluid of known viscosity, which in some cases70 is the primary viscosity reference, i.e., water. If the “true” value of viscosity is required, one should be prepared to add an uncertainty of up to (0.3% to the uncertainty of viscosity measurements owing to the water reference point.71 This additional uncertainty, however, seems not to be taken into account in the work of Go´mez et al.,47 according to the experimental uncertainty specified by them. Also shown in Figure 9 are measurements by Rodrı´guez and Brennecke,26 who have studied the viscosity of [EMIM][EtSO4] for temperatures between 278.15 and 348.15 K in intervals of 10 K with a rotational viscometer. Rodrı´guez and Brennecke26 used two samples of [EMIM][EtSO4] with a nominal purity greater than 98% and a water content of (23 and 45) ppm. They estimate the uncertainty in the measurements to be approximately (2% for high viscosity samples but as much as (2 mPa · s for medium viscosity samples (between 50 and 100 mPa · s) and (1 mPa · s for low viscosity samples (below 50 mPa · s). Finally, viscosity values for [EMIM][EtSO4] for a concentration of water of 100 ppm by mass as reported by Jacquemin and co-workers27 have been included. Their data are also based on a rotational viscometer with an expected overall uncertainty lower than (1%. Within the temperature range from (273.15 to 333.15) K, our viscosity data for [EMIM][EtSO4] agree with the literature data within the combined estimated uncertainties of the measurements. The discrepancies between the literature data sets among themselves are partly outside their combined uncertainties and may be caused not only by varying concentrations of impurities, e.g., water, but also by an overestimation of experimental accuracy. In Figure 10, data available in literature for the dynamic viscosity of [EMIM][NTf2], which are based on two different methods, are compared with the correlation, eq 4, of our data from SLS. The data by Jacquemin et al.,27 Tokuda et al.,40 Bonhoˆte et al.,49 Noda et al.,52 Tokuda et al.,55 Crosthwaite et al.,72 and McEwen et al.73 are based on rotational viscometers. The uncertainty of the data by Jacquemin et al.27 has already been discussed in context with [EMIM][EtSO4]; see above. They state for their investigated [EMIM][NTf2] sample a concentration of water of 50 ppm by mass. Bonhoˆte et al.49 and Crosthwaite et al.72 performed measurements on samples of [EMIM][NTf2] with a concentration of water of about 25 ppm by mass and of 204 ppm by mass with uncertainties of 5% and 2%. The measurements by Tokuda et al.55 as well as their respective correlation by Tokuda et al.40 were obtained for a sample with a concentration of water of less than 10 ppm by mass. Both in the work of Tokuda et al.55 from 2006 and in that of Tokuda et al.40 from 2005 information regarding the measurement uncertainty is missing. The same statement holds also for the depicted datum by McEwen et al.73 and the data correlation by Noda et al.,52 for which, moreover, the concentration of water is not specified. The measurement data by Widegren et al.,21 Dzyuba and Bartsch,57 and Pringle et al.74 are based on the capillary method. From these authors only Widegren et al.,21 who investigated the viscosity of [EMIM][NTf2] in dependence on the concentration of water at a temperature of 293.15 K, estimated their expanded uncertainty (k ) 2) in the kinematic viscosity measurements to be (1%. Further-
Fro¨ba et al. more, neither Dzyuba and Bartsch57 nor Pringle et al.74 specify the concentration of water of their samples. As can be seen from Figure 10, the scatter of literature data available for the dynamic viscosity of [EMIM][NTf2] is in the range of +8 to -22%, where our data from surface light scattering seem to form the center. Within the combined uncertainties of the measurements, the results of this work are in agreement with the data given by Widegren et al.21 From their work four values are included in total, for which the concentration of water is ranging from (10 to 1010) ppm. It is noteworthy that the smallest deviation of 0.06% of their data from eq 4 can be found for the datum corresponding to a water content of 480 ppm, which is close to the water content of 510 ppm of our sample. Agreement can be found also between our data correlation and that published by Tokuda et al.40 in 2005, whereby the absolute average deviation is less than 1.1%. A corresponding value of 2.3% indicates that the agreement is not this good for the experimental data published by Tokuda et al.55 in 2006, which deviate from our correlation up to +6.9% at a temperature of 283.15 K. Deviations of the data by Jacquemin et al.27 from eq 4 of up to +7.9% cannot be attributed to the concentration of water alone, although it was for their sample about 460 ppm by mass lower than for our sample. However, such a difference in the concentration of water would result for a temperature of 293.15 K in a change for the viscosity of [EMIM][NTf2] of about 1.5%, which can be derived from the work of Widegren and co-workers.21 In the temperature range between (298 and 343) K, agreement can be found between our data from SLS and the data by Crosthwaite et al.72 within the combined uncertainties of the measurements. In contrast, at a temperature of 283 and 293 K, their data deviate systematically from eq 4 by -13.3% and -7.7%. This disagreement becomes more evident, if the concentration of water in both cases is considered, which is about 300 ppm by mass lower in the sample investigated by Crosthwaite et al.72 For the data of Bonhoˆte et al.,49 Dzyuba and Bartsch,57 McEwen et al.,73 and Pringle et al.,74 who reported the viscosity of [EMIM][NTf2] at a single temperature of 298.15 K, 293.15 K, 299.15 K, and 293.15 K, the deviations with respect to our data correlation, eq 4, are -12%, -22%, -9%, and -12%, respectively. The latter studies are mainly dedicated to investigating the influence of structural variations on the physical properties of ILs to characterize these after their synthesis and are most likely of limited reliability. For the viscosity of [EMIM][N(CN)2], only two values can be found in the literature, namely one datum by Yoshida et al.58 for a temperature of 295.15 K and another by MacFarlane et al.59,60 for a temperature of 298.15 K. Yoshida et al.58 used a rotational viscometer and MacFarlane et al.60 determined the viscosity from the time required for a fixed volume to flow through a narrow orifice in a calibrated glass viscometer. Yoshida et al.58 and MacFarlane et al.59,60 did not specify the uncertainties of their viscosity measurements as well as the concentration of water of their investigated samples. However, the values of Yoshida et al.58 and MacFarlane et al.59,60 show with respect to our data correlation, eq 4, deviations of -18.9% and +11.9%. Due to the lack of literature data, a data comparison for the viscosity of [OMA][NTf2] could not be drawn. Conclusions This work presents an extensive experimental study of several thermophysical properties of three pure imidazolium-
Properties of Ionic Liquids based ILs, [EMIM][EtSO4], [EMIM][NTf2], and [EMIM][N(CN)2], as well as of one tetraalkylammonium-based IL, [OMA][NTf2], in dependence on temperature at atmospheric pressure. Based on the provision of reliable experimental data for density, refractive index, interfacial tension, and viscosity, the influence of structural variations on these thermophysical properties was discussed qualitatively. Finally, a critical evaluation and comprehensive comparison of data available in literature was performed. The disagreement between different data sources may be attributed to an inconsistent sample purity, yet it seems not to be the only reason. Here, the discrepancies for the physicochemical properties, including equilibrium properties, originating from different sources were also found to be influenced by inconsistent experimental techniques as well as an inadequate estimation of their uncertainty. In this context, for the existing research activities for the physiochemical properties of ILs, some parallels may be drawn to previous research activities for hydrofluorocarbons HFCs used today in refrigeration. First investigations of the thermophysical properties of HFCs revealed the demand for a continuous improvement of experimental methods and their proper application. The data situation for HFCs, however, could be considered satisfactory only long time after their adoption as alternatives for the former used chlorofluorocarbons CFCs. Acknowledgment. We thank Cristina Botero and Benjamin Hasse for their valuable assistance in carrying out many of the SLS experiments. We are grateful to Dirk Gerhard and Peter Wasserscheid from the Department of Chemical Reaction Engineering of the University of Erlangen-Nuremberg for the sample preparation, analysis of the water content, and 1H NMR spectra determination. This work was supported by the Max-Buchner-Forschungsstiftung and by the German National Science Foundation (Deutsche Forschungsgemeinschaft, DFG). References and Notes (1) Wilkes, J. S. In Ionic Liquid Synthesis; Wasserscheid, P.; Welton, T. , Eds.; Wiley-VCH Verlag: Weinheim, Germany, 2003; pp 1-6. (2) Balducci, A.; Soavi, F.; Mastragostino, M. Appl. Phys. A: Mater. Sci. Process. 2006, 82, 627–632. (3) Van Valkenburg, M. E.; Vaughn, R. L.; Williams, M.; Wilkes, J. S. Thermochim. Acta 2005, 425, 181–188. (4) Welton, T. Chem. ReV. 1999, 99, 2071–2083. (5) Dupont, J.; de Souza, R. F.; Suarez, P. A. Z. Chem. ReV. 2002, 102, 3667–3692. (6) Wilkes, J. S. In Ionic Liquids: Industrial Applications for Green Chemistry; Rogers, D., Seddon, K. R., Eds; ACS Symposium Series 818; American Chemical Society: Washington, DC, 2002; pp 214-229. (7) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 3773– 3789. (8) Scho¨fer, S. H.; Kaftzik, N.; Kragl, U.; Wasserscheid, P. Chem. Commun. 2001, 425–426. (9) Schroer, K.; Tacha, E.; Lutz, S. Organic Process Research & DeVelopment 2007, 11, 836–841. (10) Farag, H. K.; Endres, F. J. Mater. Chem. 2008, 18, 442–449. (11) Birbilis, N.; Howlett, P. C.; MacFarlane, D. R.; Forsyth, M. Surf. Coat. Technol. 2007, 201, 4496–4504. (12) Fukushima, T.; Aida, T. Chem. Eur. J. 2007, 13, 5048–5058. (13) Wang, H.; Lu, Q.; Ye, C.; Liu, W.; Cui, Z. Wear 2004, 256, 44– 48. (14) Wei, D.; Ivaska, A. Anal. Chim. Acta 2008, 607, 126–135. (15) Tochigi, K.; Yamamoto, H. J. Phys. Chem. C 2007, 111, 15989– 15994. (16) Zhang, S.; Sun, N.; He, X.; Lu, X.; Zhang, X. J. Phys. Chem. Ref. Data 2006, 35, 1475–1517. (17) Deetlefs, M.; Seddon, K. R.; Shara, M. Phys. Chem. Chem. Phys. 2006, 8, 642–649. (18) Marsh, K. N.; Boxall, J. A.; Lichtenthaler, R. Fluid Phase Equilib. 2004, 219, 93–98. (19) Wilkes, J. S. J. Mol. Catal. A: Chem. 2004, 214, 11–17.
J. Phys. Chem. B, Vol. 112, No. 39, 2008 12429 (20) Heintz, A. J. Chem. Thermodynamics 2005, 37, 525–535. (21) Widegren, J. A.; Laesecke, A.; Magee, J. W. Chem. Commun. 2005, 1610–1612. (22) Seddon, K. R.; Stark, A.; Torres, M. J. Pure Appl. Chem. 2000, 72, 2275–2287. (23) Perry, R. L.; Jones, K. M.; Scott, W. D.; Liao, Q.; Hussey, C. L. J. Chem. Eng. Data 1995, 40, 615–619. (24) Huddleston, J. G.; Visser, A. E.; Reichert, W. M.; Willauer, G. A.; Rogers, R. D. Green Chem. 2001, 3, 156–164. (25) Poole, C. F.; Kersten, B. R.; Ho, S. S. J.; Coddens, M. E.; Furton, K. G. J. Chromatogr. 1986, 352, 407–425. (26) Rodrı´guez, H.; Brennecke, J. F. J. Chem. Eng. Data 2006, 51, 2145– 2155. (27) Jacquemin, J.; Husson, P.; Padua, A. A. H.; Majer, V. Green Chem. 2006, 8, 172–180. (28) Fro¨ba, A. P.; Leipertz, A. Int. J. Thermophys. 2003, 24, 895–921. (29) Fro¨ba, A. P.; Leipertz, A. J. Chem. Eng. Data 2007, 52, 1803– 1810. (30) Fro¨ba, A. P.; Leipertz, A. To be submitted for publication in Int. J. Thermophys. (31) Maier, F.; Gottfried, J. M.; Rossa, J.; Gerhard, D.; Schulz, P. S.; Schwieger, W.; Wasserscheid, P.; Steinru¨ck, H.-P. Angew. Chem., Int. Ed. 2006, 45, 7778–7780. (32) Lemmon, E. W.; Span, R. J. Chem. Eng. Data 2006, 51, 785–850. (33) Levich, V. G. Physicochemical Hydrodynamics; Prentice Hall: Englewood Cliffs, NJ, 1962. (34) Lucassen-Reynders, E. H.; Lucassen, J. AdVan. Colloid Interface Sci. 1969, 2, 347–395. (35) Langevin, D. Light Scattering by Liquid Surfaces and Complementary Techniques; Marcel Dekker: New York, 1992. (36) Fro¨ba, A. P. Simultane Bestimmung von Viskosita¨t und Oberfla¨chenspannung transparenter Fluide mittels Oberfla¨chenlichtstreuung, Dr.Ing. thesis, Friedrich-Alexander-Universita¨t Erlangen-Nu¨rnberg, Germany, 2002. (37) Fröba, A. P.; Will, S.; Leipertz, A. Int. J. Thermophys. 2000, 21, 1225–1253. (38) Fro¨ba, A. P.; Leipertz, A. Int. J. Thermophys. 2003, 24, 1185– 1206. (39) Fredlake, C. P.; Crosthwaite, J. M.; Hert, D. G.; Aki, S. N. V. K.; Brennecke, J. F. J. Chem. Eng. Data 2004, 49, 954–964. (40) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. J. Phys. Chem. B 2005, 109, 6103–6110. (41) Blanchard, L. A.; Gu, Z.; Brennecke, J. F. J. Phys. Chem. B 2001, 105, 2437–2444. (42) Yang, J.-Z.; Lu, X.-M.; Gui, J.-S.; Xu, W.-G. Green Chem. 2004, 6, 541–543. (43) Yang, J.-Z.; Lu, X.-M.; Gui, J.-S.; Xu, W.-G.; Li, H.-W. J. Chem. Thermodyn. 2005, 37, 1250–1255. (44) Arce, A.; Rodil, E.; Soto, A. J. Chem. Eng. Data 2006, 51, 1453– 1457. (45) Vila, J.; Gines, P.; Rilo, E.; Cabeza, O.; Varela, L. M. Fluid Phase Equilib. 2006, 247, 32–39. (46) Krummen, M.; Wasserscheid, P.; Gmehling, J. J. Chem. Eng. Data 2002, 47, 1411–1417. (47) Go´mez, E.; Gonza´lez, B.; Calvar, N.; Tojo, E.; Domı´nguez, A. J. Chem. Eng. Data 2006, 51, 2096–2102. (48) Fro¨ba, A. P.; Wasserscheid, P.; Gerhard, D.; Kremer, H.; Leipertz, A. J. Phys. Chem. B 2007, 111, 12817–12822. (49) Bonhoˆte, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.; Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168–1178. (50) Gardas, R. L.; Freire, M. G.; Carvalho, P. J.; Marrucho, I. M.; Fonseca, I. M. A.; Ferreira, A. G. M.; Coutinho, J. A. P. J. Chem. Eng. Data 2007, 52, 1881–1888. (51) Heintz, A.; Kulikov, D. V.; Verevkin, S. P. J. Chem. Eng. Data 2002, 47, 894–899. (52) Noda, A.; Hayamizu, K.; Watanabe, M. J. Phys. Chem. B 2001, 105, 4603–4610. (53) Lopes, J. N. C.; Cordeiro, T. C.; Esperanc¸a, J. M. S. S.; Guedes, H. J. R.; Huq, S.; Rebelo, L. P. N.; Seddon, K. R. J. Phys. Chem. B 2005, 109, 3519–3525. (54) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. J. Phys. Chem. B 2004, 108, 16593–16600. (55) Tokuda, H.; Tsuzuki, S.; Susan, M. A. B. H.; Hayamizu, K.; Watanabe, M. J. Phys. Chem. B 2006, 110, 19593–19600. (56) Hong, G.; Jacquemin, J.; Husson, P.; Gomes, M. F. C.; Deetlefs, M.; Nieuwenhuyzen, M.; Sheppard, O.; Hardacre, C. Ind. Eng. Chem. Res. 2006, 45, 8180–8188. (57) Dzyuba, S. V.; Bartsch, R. A. ChemPhysChem 2002, 3, 161–166. (58) Yoshida, Y.; Muroi, K.; Otsuka, A.; Saito, G.; Takahashi, M.; Yoko, T. Inorg. Chem. 2004, 43, 1458–1462. (59) MacFarlane, D. R.; Golding, J.; Forsyth, S.; Forsyth, M.; Deacon, G. B. Chem. Commun. 2001, 1430–1431.
12430 J. Phys. Chem. B, Vol. 112, No. 39, 2008 (60) MacFarlane, D. R.; Forsyth, S. A.; Golding, J.; Deacon, G. B. Green Chem. 2002, 4, 444–448. (61) Kilaru, P.; Baker, G. A.; Scovazzo, P. J. Chem. Eng. Data 2007, 52, 2306–2314. ´ .; Bravo, R.; Amigo, A. Phys. Chem. Chem. (62) Brocos, P.; Pin˜eiro, A Phys. 2003, 5, 550–557. (63) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of Gases and Liquids; McGraw-Hill: New York, 1987. (64) MacLeod, D. B. Trans. Faraday Soc. 1923, 19, 38–41. (65) Sugden, S. J. Chem. Soc., Trans. 1924, 125, 32–41. (66) Law, G.; Watson, P. R. Langmuir 2001, 17, 6138–6141. (67) Martino, W.; de la Mora, J. F.; Yoshida, Y.; Saito, G.; Wilkes, J. Green Chem. 2006, 8, 390–397.
Fro¨ba et al. (68) Knotts, T. A.; Wilding, W. V.; Oscarson, J. L.; Rowley, R. L. J. Chem. Eng. Data 2001, 46, 1007–1012. (69) Rebelo, L. P. N.; Lopes, J. N. C.; Esperanc¸a, J. M. S. S.; Filipe, E. J. Phys. Chem. B 2005, 109, 6040–6043. (70) Caetano, F. J. P.; Fareleira, J. M. N. A.; Oliveira, C. M. B. P.; Wakeham, W. A. J. Chem. Eng. Data 2005, 50, 1875–1878. (71) Marsh, K. N. Pure Appl. Chem. 2000, 72, 1809–1818. (72) Crosthwaite, J. M.; Muldoon, M. J.; Dixon, J. K.; Anderson, J. L.; Brennecke, J. F. J. Chem. Thermodynamics 2005, 37, 559–568. (73) McEwen, A. B.; Ngo, H. L.; LeCompte, K.; Goldman, J. L. J. Electrochem. Soc. 1999, 146, 1687–1695. (74) Pringle, J. M.; Golding, J.; Baranyai, K.; Forsyth, C. M.; Deacon, G. B.; Scott, J. L.; MacFarlane, D. R. New J. Chem. 2003, 27, 1504–1510.
JP804319A
