a review on the transport properties of ionic liquids
TRANSCRIPT
Journal of Molecular Liquids 193 (2014) 262–266
Contents lists available at ScienceDirect
Journal of Molecular Liquids
j ourna l homepage: www.e lsev ie r .com/ locate /mol l iq
Review
A review on the transport properties of ionic liquids
Xiaojing Wang a, Yanling Chi b, Tiancheng Mu c,⁎a School of Science, Beijing University of Civil Engineering and Architecture, Beijing 100044, Chinab Materials Science and Engineering College, Northeast Forestry University, Harbin 415000, Chinac Department of Chemistry, Renmin University of China, Beijing 100872, China
⁎ Corresponding author. Tel.: +86 10 62514925; fax: +E-mail address: [email protected] (T. Mu).
http://dx.doi.org/10.1016/j.molliq.2014.03.0110167-7322/© 2014 Elsevier B.V. All rights reserved.
a b s t r a c t
a r t i c l e i n f oArticle history:Received 5 November 2013Received in revised form 27 December 2013Accepted 7 March 2014Available online 20 March 2014
Keywords:Transport propertyIonic liquidCorrelationMolecular dynamics simulation
Transport properties include the thermal conductivity for heat transfer, the viscosity for themomentum transfer,and the diffusion coefficient formass transfer. The transport properties are very important for chemical engineer-ing design. The ionic liquids are liquids with good conductivity thus may have promising applications in batteryindustry. However, the high viscosity hinders the applications of them. In this paper, we firstly reviewed thevalues of transport properties of pure ionic liquids, then these properties of the mixtures of ionic liquids withother compounds were discussed, and lastly we reviewed the correlation researches on the transport propertiesof systems including ionic liquids and the molecular dynamics simulation investigations on these systems.
© 2014 Elsevier B.V. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2622. Transport properties of pure ionic liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2633. Transport properties of the mixtures including ionic liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2634. Correlations and molecular dynamics simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
4.1. Correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2644.2. Molecular dynamics simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2654.3. Comparison of the predictive methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
1. Introduction
Ionic liquids (ILs) have attracted much attention in the scientificcommunity recently due to their novel and highly customizableproperties [1,2]. They have been used in various fields such as extractionand separation [3,4], catalyst, materials fabrication [5,6], analytical tech-nology, and energy [7]. The transport properties include the diffusionrate coefficients such as the thermal conductivity (for heat transfer)[8–11], the viscosity (for momentum transfer), and diffusion coefficient(for mass transfer) [12–15]. Thus, the transport properties play crucialroles in chemical and engineering processes. The relatively high viscosityof ILs poses serious limitations to the use of ILs. Considered the high
86 10 62516444.
viscosity of most ILs, the transport properties of ILs are more important.Since the ILs are composed of cations and anions, the electric conductiv-ity of ILs is widely concerned [16–19]. According to the previous reports,the transport properties of ILs are mainly determined by hydrogenbonding, van der Waals forces, molecular weight and mobility [16,20].There have been several reviews published on the transport properties,and different models (correlation, group contribution [21], quantitativestructure property relationship (QSPR) [22], computer-aided moleculardesign [23], etc.) for the prediction of the transport properties of ILswere compared.
Hence, in this mini-review, we mainly discuss the viscosities of ILs.Both the experimental data and prediction methods were discussed.The transport properties of pure ILs, the mixtures of ILs with molecularsolvents, or the mixtures of two or more ILs were discussed. The defini-tion andnomenclature ofmixtures of ILswere given in a previous report
Table 2VFT parameters of ionic conductivity data for [BMIM]-based [27], [CnMIM] [Tf2N] [28] and[Tf2N]-based [26] ILs.
IL σ0/10−1 S cm−1 B/102 K T0/K
[BMIM][BETI] 7.1 ± 0.5 7.96 ± 0.22 169 ± 2[BMIM][Tf2N] 4.3 ± 0.2 5.65 ± 0.14 178 ± 2[BMIM][OTF] 9.8 ± 0.8 7.93 ± 0.26 162 ± 3[BMIM][PF6] 14.7 ± 1.0 8.55 ± 0.20 174 ± 2[BMIM][TFA] 9.2 ± 0.4 7.19 ± 0.13 172 ± 1[BMIM][BF4] 13.8 ± 0.6 7.41 ± 0.13 174 ± 1[MMIM][Tf2N] 6.6 ± 0.4 5.62 ± 0.18 168 ± 3[EMIM][Tf2N] 5.8 ± 0.2 5.54 ± 0.13 165 ± 2[HMIM][Tf2N] 6.1 ± 0.3 7.31 ± 0.13 168 ± 1[OMIM][Tf2N] 6.1 ± 0.2 8.11 ± 0.12 166 ± 1[BPy][Tf2N] 5.5 ± 0.5 6.33 ± 0.24 175 ± 3[BMPyrr][Tf2N] 5.6 ± 0.2 6.75 ± 0.14 171 ± 1[(n-C4H9)(CH3)3N][Tf2N] 6.1 ± 0.1 6.89 ± 0.04 177 ± 0
263X. Wang et al. / Journal of Molecular Liquids 193 (2014) 262–266
[24]; in that report, they also gave a comprehensive review of theproperties of mixtures of ionic liquids. For prediction methods, Dasand Roy have given a comprehensive review on the QSPR models [22].Therefore we will only discuss the correlation methods based ongroup contribution concept because it is simple and easy to use, andmolecular dynamics because it is molecular based.
2. Transport properties of pure ionic liquids
The transport properties of ILs are determined by the nature of itscation and anion and conditions such as temperature and pressure.Tokuda et al. investigated the influence of cation, anion, the alkylchain length of cation, and temperature on the transport properties ofIL [25–28].
The transport properties (diffusion coefficient, viscosity, ionicconductivity) of ILs: (1) with a variety of fluorinated anions combinedwith [BMIM] [27]; (2) with different cationic structures [BMIM], [BPy],[BMPyrr], and [(n-C4H9)(CH3)3N] combined with an anion, [Tf2N][26]; (3) with imidazolium based cation but different alkyl chain length([CnMIM][Tf2N], n= 1, 2, 4, 6, and 8) [28]; weremeasured and correlat-ed with Vogel–Fulcher–Tamman (VFT) equation (Eqs. (1)–(3)),the viscosity and ionic conductivity results are presented in Tables 1and 2, respectively.
D ¼ D0 exp −B1
.T−T0;1
� �h ið1Þ
η ¼ η0 exp B2
.T−T0;2
� �h ið2Þ
σ ¼ σ0 exp −B3
.T−T0;3
� �h ið3Þ
The anionic effect on the self-diffusion coefficients of the cation andanion, viscosity, and ionic conductivity of ILs [BMIM] with a variety offluorinated anions at 30 °C lead to the following conclusions: (1) TheVFT equation is suitable to describe the temperature dependencies ofthe transport properties such as self-diffusion coefficient, viscosity,ionic conductivity, and molar conductivity; (2) the self-diffusion coeffi-cients exhibit higher values for the cation compared with the anion,even if its radius is larger than that of the anionic radii. The cationicand anionic diffusion coefficients for the ILs follow the order: [BMIM][Tf2N] N [BMIM][TFA] N [BMIM][OTF] N [BMIM][BF4] N [BMIM][BETI] N[BMIM][PF6]. The order of the diffusion coefficients greatly contrasts tothe viscosity data; (3) the ratio of the experimental molar conductivityto that of the calculated data follows the order: [BMIM][PF6] N [BMIM][BF4] N [BMIM][BETI] N [BMIM][Tf2N] N [BMIM][OTF] N [BMIM][TFA][27].
Table 1VFT parameters of viscosity data for [BMIM]-based [27], [CnMIM][Tf2N] [28] and[Tf2N]-based [26] ILs.
IL η0/10−1 mPa s B/102 K T0/K
[BMIM][BETI] 1.7 ± 0.2 7.63 ± 0.38 180 ± 3[BMIM][Tf2N] 2.5 ± 0.2 6.25 ± 0.22 180 ± 2[BMIM][OTF] 3.7 ± 0.6 5.70 ± 0.37 193 ± 4[BMIM][PF6] 3.6 ± 0.5 6.39 ± 0.25 201 ± 2[BMIM][TFA] 1.1 ± 0.2 7.88 ± 0.53 177 ± 4[BMIM][BF4] 2.1 ± 0.1 6.97 ± 0.69 185 ± 6[MMIM][Tf2N] 2.9 ± 0.6 5.87 ± 0.57 178 ± 7[EMIM][Tf2N] 4.0 ± 1.3 5.09 ± 0.81 182 ± 10[HMIM][Tf2N] 1.6 ± 0.2 7.57 ± 0.39 173 ± 3[OMIM][Tf2N] 1.5 ± 0.2 8.02 ± 0.30 173 ± 2[BPy][Tf2N] 2.1 ± 0.4 6.71 ± 0.52 179 ± 5[BMPyrr][Tf2N] 2.9 ± 0.4 6.51 ± 0.31 181 ± 3[(n-C4H9)(CH3)3N][Tf2N] 4.5 ± 0.6 5.34 ± 0.27 199 ± 3
To the ILs with different type of cations, the magnitude of cationicand anionic diffusion coefficients for the ILs follows the order:[BMIM][Tf2N] N [BPy][Tf2N] N [BMPyrr][Tf2N] N [(n-C4H9)(CH3)3N][Tf2N]; which coincides with the reverse order to the viscosity data. Theratio of the experimental to that of the calculated molar conductivitydata from the ionic diffusivity using the Nernst–Einstein equationfollows the order: [BMPyrr] [Tf2N] N [(n-C4H9)(CH3)3N] [Tf2N] N [BPy][Tf2N] N[BMIM][Tf2N] at 30 °C, which also provides quantitative infor-mation on the active ions contribution to ionic conduction [26].
The investigation of the ILs with imidazolium based cation butdifferent alkyl chain length led to the following conclusions. Thesummation of the cationic and anionic diffusion coefficients followsthe order: [EMIM][Tf2N] N [MMIM][Tf2N] N [BMIM][Tf2N] N [HMIM][Tf2N] N [OMIM][Tf2N]. The ratio of the calculated and experimentaldata of molar conductivity decreases with increasing number of carbonatoms in the alkyl chain [28]. The effect of position and length of alkylsubstituent in pyridinium based ILs has been studied too. It was foundthat the position of the methyl substituent had significant influence ofon the transport properties. Walden plots indicate that these ILshave high ionicity, closer to the “ideal” KCl line, suggesting less ionassociation compared to their 1-butyl-3,5-dimethylpyridinium homo-logues [29]. The temperature dependence of transport properties ofammonium-based ILs N-alkyl-triethylammonium [Tf2N] with differentalkyl chain length on the cation (from 293.15 to 363.15) K [30] anddynamic viscosity for three phosphonium-based ionic liquids: tributylmethyl phosphonium methylsulfate (P4441C1SO4), tributyl ethylphosphonium diethylphosphate (P4442(C2)2PO4) and tributyl octylphosphonium chloride (P4448Cl) (from 293.15 to 343.15 K) can bedescribed by VFT equations.
The high-pressure (up to 126 MPa) viscosity of [EMIM], [HMIM],and [DMIM] with [Tf2N], and [HMIM] with [Tf2N], [PF6], and [BF4]was measured. The increase of the alkyl-chain length induced theviscosity increase at elevated pressures. [HMIM][PF6] and [BMIM][PF6] demonstrated a nonlinear pressure dependence even at rela-tively moderate pressures (to 30 MPa) [31]. Harris et al. carried outa series of research on the temperature and pressure dependenceof viscosity of ILs [32–37]. The transport properties of pyridiniumbased ILs [29], pyrrolidinium based ILs [38], phosphonium-basedILs [39,40] could be well described with the VFT equation.
3. Transport properties of the mixtures including ionic liquids
A considerable amount ofwork on the transport properties of pure ILhas been published. At the same time, the properties of bothmixtures ofIL with molecular solvent and the mixtures of two or more ILs areimportant because themixtures might enormously enlarge the applica-tions of the ILs.
Seddon et al. discovered that the viscosity of ionic liquid mixtureswith impurities or additiveswas dependentmainly on themole fraction
264 X. Wang et al. / Journal of Molecular Liquids 193 (2014) 262–266
of added molecular solvents and only to a lesser extent upon theiridentity. The addition of molecular solvents decreases the viscosityand density, while the addition of chloride impurities, arising from thepreparation of the ILs, increases viscosity dramatically [41]. Most ofthe ILs are hygroscopic in air [42–44], the viscosities of ILs are stronglydependent on the amount of dissolved water [45]. Recently, Khupseand Kumar reviewed the effect of molecular solvents, including wateron the viscosity of ILs [46]. They summed the experimental data withthe empirical equations. Rheological properties of a homologous seriesof ammonium-based [Tf2N] ILs were measured for both pure andwater-saturated ionic liquids [30]. It appears that the presence ofwater dramatically decreases the viscosity of ILs by up to three times[30]. There are also other researches on the viscosity and surface tensionof ILs plus water with composition [47–49].
Navis et al. investigated the viscosities of IL binary mixtures with acommon ion ([C6mim] + [C2mim]) [BF4], ([C6mim] + [C4mim]) [BF4],[C4mim] ([BF4] + [MeSO4]) and [C4mim] ([PF6] + [BF4]) within thetemperature range (298.15–308.15) K [50]. The viscosity of a binarymixture does not obey Eq. (4), which means that the interactions ofthese systems are complicated [51].
η ¼ xη1 þ 1−xð Þlog10η2 ð4Þ
It should be noted here that the transport properties reported bydifferent authors sometimes show significant discrepancy, this may becaused by the impurities, especiallywater. Therefore special care shouldbe taken to compare the transport properties of ILs from differentsources.
The temperature dependence of viscosity for mixtures containing[BMIM][SCN] and 1-alcohol (methanol, ethanol, and 1-propanol) in atemperature range (298.15 to 348.15) K and ambient pressure ofthese systems can be described by the VFT equation [52]. Similar tothe pure ionic liquid systems, the viscosity of [BMIM][SCN] + 1-alcohol binary system decreases with the temperature. As the increaseof alcohol mole fraction, the system viscosity decreases dramatically[53].
The experimental investigation on viscosity, density, and conductiv-ity ofmixture of PEG (Mw=200, 300, 400) and PEI (Mw=423)with ILs[BMIM][BF4] and [BMIM][PF6] indicates that the viscosity and conduc-tivity of the ionic liquid/solvent mixture decrease significantly as theconcentration of PEG increases [54]. Liu et al. investigated the VLE andviscosity of CO2/[BMIM][PF6]/methanol ternary mixture at pressuresup to 12.5 MPa. It demonstrated that the viscosity of the IL-rich phasedecreases significantly with increasing pressure of CO2, and indicatedthat compressed CO2 may be used to reduce the viscosity of ILs inmany applications [55].
The transport properties of ILs containing metal ions and H2O areimportant to evaluate the suitability of ILs for applications in batteries.The 1-ethyl-3-methylimidazolium dicyanamide [EMIM][DCA] contain-ing 30 mol% Zn(DCA)2 electrolyte displayed higher conductivities(15 mS.cm−1 at 25 °C) than the same IL containing 30 mol% ZnCl2(12 mS.cm−1 at 25 °C). The different solvation environments of theZn2+ in the two IL/zinc salt mixtures results in a significant improve-ment in transport properties of [EMIM][DCA] containing Zn(DCA)2over ZnCl2 [56]. Forsyth et al. investigated on the effect of organic addi-tives [57], anion mixing [58] on the transport properties of ILs/Li[Tf2N]mixtures [59]. To investigate the effect of anion mixing on the lithiumion speciation, conductivity and PFG-NMR diffusion measurements ofbinary IL system N-methyl-N-propylpyrrolidinium, and either [Tf2N])or bis(fluorosulfonyl)imide [FSI] as the anion as well as the Li-[Tf2N]-containing ternary system were investigated. The addition of thelithium salt to the IL system resulted in a decrease in conductivity. Theconductivity and ion diffusion have linear behavior as a function of theanion ratio for a fixed lithium salt composition. The addition of the FSIto the [Tf2N] IL results in a considerable increase in lithium ion diffusiv-ity without additional complex ion behavior [58]. Organic solvent
additives (toluene and THF) effect on IL/Li[Tf2N] over a wide tempera-ture range indicated that both additives enhanced the conductivitymost at low temperatures. Both the anion and lithium self-diffusivityare enhanced in the same order by the additives while that of thepyrrolidinium cation is marginally enhanced [57]. Besides that, theionic conductivity of mixtures of N-alkyl-N-methylpyrrolidinium[Tf2N] ILs with Li[Tf2N] was investigated [59]. The ionic conductivityvariability correlates well with the phase behavior [59].
4. Correlations and molecular dynamics simulations
4.1. Correlations
Since experimental data of transport properties of ILs are scarce andoften contradictory, the development of correlation and predictivemodels are quiet necessary. For transport properties, most of theproposed predictive methods are QSPR and group contributionmethods. Coutinho and his coworkers did a lot of researches on thistopic. Recently, they had given a review on the predictive methods forthe estimation of thermophysical properties of ILs [2]. In the review,they introduced some correlation works, which were complicated. Thecomplex of themodelswill decrease both the applicant and the theoret-ically base, which should not be encouraged. In this paper, we onlyintroduce the simple correlation methods and molecular dynamicssimulations.
An Orrick–Erbar-type approach [60] was employed to estimate theviscosity of ILs by Gardas and Coutinho [20]. A group contributiontechnique was employed to estimate the A and B parameters in Eq. (5):
lnηρM
¼ Aþ BT
ð5Þ
Where η and ρ are the viscosity and density, respectively. M is themolecular weight and T is the absolute temperature. For around 500data points of 29 ILs studied, in ranges of temperature, 293–393 K andviscosity, 4–21,000 cP, a MPD of 7.7% with a maximum deviation small-er than 28% was observed. The Orrick–Erbar method requires densitydata for the prediction of viscosity, in case the density is unavailable,the authors suggested using Eq. (3) to estimate it. Gardas and Coutinhoproposed a further new correlation [61] based on the VTF equation(Eq. (6)) to estimate the viscosity, electrical conductivity, thermalconductivity, refractive index, isobaric expansivity, and isothermalcompressibility, of various families of ILs. The new model is also agroup contribution method based on experimental data.
ln η ¼ Aη þBη
T−T0η
� � ð6Þ
where η is viscosity, T is temperature in K, and Aη, Bη, and T0η are adjust-able parameters. T0η is similar for all the ILs. Aη and Bη can be obtainedby a group contribution method according to:
Aη ¼Xki¼1
niai;ηBη ¼Xki¼1
nibi;η ð7Þ
where ni is the number of groups of type i, k is the total number ofdifferent groups in the IL. Similar models were proposed to estimateother transport properties; the estimated data are consistent with theexperimental data. The correlation of viscosities of [BMIM][DCA] and[BMIM][C(CN)3], for mole fractions up to 0.6, at atmospheric pressurewas consistent to the experimental viscosity data [62]. The electricalconductivity of ILs can also be predicted by the above group methods,the correlation coefficient between the experimental data (around
265X. Wang et al. / Journal of Molecular Liquids 193 (2014) 262–266
300 data points of 15 ILs) and this group contribution predicted that themethod reaches 0.9985 [62].
Ghatee [63,64] proposed a linear equation (Eq. (8)) to fit the viscosityof the ILs. The results are quite in agreement with the experiments, witha maximum deviation of 3.00%.
1η
� �ϕ¼ aþ bT ð8Þ
where a and b are substance dependent constants and ϕ is a character-istic exponent. Here we should point out again that the values oftransport properties are significantly influenced by the small amountsof impurities, which may result in the inaccuracy of the experimentaldata, therefore influence the group contribution parameters regressedfrom experimental data.
The viscosity of mixtures with ILs are very sensitive to the structureof constitute. Thus, although there aremodels to correlation of transportproperties of mixtures including ILs, most of them are based on theknowledge of viscosity of pure ILs. Usually these models can be onlyused to ideal systems. Redlich–Kister polynomial was used to fit theviscosity and molar refraction, the excess molar volume, and deviationin isentropic compressibility of [BMIM][PF6] in tetrahydrofuran,dimethylsulfoxide, methanol, and acetonitrile [65]. A model with oneparameter was proposed to calculate viscosities of binary mixturesfrom pure component physical properties to overcome the difficultythat the numbers of parameters for Redlich–Kister polynomialequations are large [66,67]. Geng et al. proposed a model based on anequation of state for estimating the viscosity of [BMIM][PF6] and DMFbinary mixtures by coupling with the excess Gibbs free energy modelof viscosity. The model gives a deviation of 8.29% for the viscosity [68].Carvalho extended the one-constant Grunberg andNissan equation [69]to describe viscosity of [BMIM][DCA] + water and [BMIM][C(CN)3] +water [62]. Equation derived data are consistent with the experimentaldata.
ln ηm ¼X
xi ln ηi þ12
Xni¼1
Xmj¼1
xix jGij ð9Þ
The x is the liquid mole fraction of the component, and Gij is atemperature dependent interaction parameter [70].
The viscosities of the IL binarymixtureswith a common ion ([C6mim]+[C2mim])[BF4], ([C6mim] + [C4mim])[BF4], [C4mim]([BF4] + [MeSO4])and [C4mim]([PF6]+ [BF4]) can be predicted by the following expression
log10 ηVð Þ ¼ x log10 η1V1ð Þ þ 1−xð Þlog10 η2V2ð Þ þ ΔgE
RTð10Þ
whereV is themolar volume andΔgE denotes the excessmolarGibbs en-ergy of activation for flow [50]. If amixture does not deviate significantlyfrom the ideal mixture model, ΔgE = 0, then the viscosity for a mixturecan be obtained from its molar volumes and from pure product molarvolumes and viscosities, which is the “ideal” Katti and Chaudhri mixinglaw [71,72].
log10 ηð Þ ¼ x log10 η1ð Þ þ 1−xð Þlog10 η2ð Þ ð11Þ
It is obvious that the key point for developing predictive methodsbased on group contribution is to have enough reliable experimentaldata. Then comprehensive investigations should be carried out on therelationship between the structure and properties of ILs. Besides that,a criteria should be established to evaluate the reliability of the experi-mental data of transport properties to avoid the use of the inaccuracydata in the regression of the group contribution parameters, just likewhat was used in selecting the activity coefficient data from DDB [73].Hence, special care should be given to the inconsistent data because
sometimes it is caused by the inaccuracy of experimental data, whilethe predicated data is more reliable.
4.2. Molecular dynamics simulations
To elucidate the physical aspects behind transport properties of ILs,molecular dynamics (MD) simulations were also used to study thetransport properties of the pure ILs [74,75] and their mixtures withconventional solvents.
Borodin et al. reported that the transport properties such as ion self-diffusion coefficients, conductivity, and viscosity of ILs obtained by MDsimulationswere in good agreementwith experimental measurements.The systems that they investigated include [MPPy][Tf2N] [76], [EMIM][Tf2N] [77,78], and 30 ILs containing [CnMIM] cations and [BF4],[CF3BF3], [CH3BF3], [OTF], [PF6], [DCA], [C(CN)3], [B(CN)4], [Tf2N], [FSI]and nitrate anions [77]. Other MD simulation research gave similarresults [79].
The investigation on the transport coefficients of ILs of [P4444] cationwith six amino acid [AA] anions indicates that the self-diffusion coeffi-cients of the [P4444] cations are smaller than those of amino acid anions.The diffusion coefficient and electrical conductivity of these [P4444][AA]ILs decrease because of the polar functional groups on the side chain of[AA] anion [74]. The dependence of the properties of the ILs on simula-tion parameters, e.g., system-size effects or the choice of the interactionpotential, is analyzed in detail [80]. Chaban et al. proposed a new, non-polarizable force fieldmodel to ILs as well as condensedmatter systemsin which the ionic interaction requires an account of polarization effects[75]. The influence of the temperature and the cation's alkyl chainlength of [CnMIM][Tf2N] (n = 4, 8, 12) ILs on the transport propertieswas thoroughly investigated by MD simulations from 298.15 to 498.15K [81]. Density functional theory is incorporated in the charge distribu-tion to calculate the partial charges. Results show that the calculation oftransport properties is very complicated, especially at low temperatures[81]. MD simulations on the transport properties of [BMIM][Tf2N],[BMIM][Tf2N], [BMMIM][Tf2N], [BMPyr][Tf2N], [N4111][Tf2N], and[N4111][COOHTf2N] show that several of these ILs are quite favorableto act as high temperature heat transfer fluids. The MD simulation re-sults of transport properties qualitatively agree with the experimentaldata. The simulated thermal conductivity of the [Tf2N]-based ionic liq-uids has good correlation with their density and molecular weight[82]. The performance of MD simulations on the transport propertiesof [PYR] [Tf2N] mixed with [Li][Tf2N] salt has been conducted by usinga polarizable force field. Mixture simulations with lithium salt molefractions between 0% and 33% at 363 and 423 K yield the simulationresults on ion self-diffusion coefficients and ionic conductivities arein very good agreement with available experimental data. The ionicconductivity of the electrolytes decreases with increasing saltconcentration [83].
4.3. Comparison of the predictive methods
The prediction methods for transport properties are either based onthe experimental data (such as correlations or group contributionmethods) or based on the a priorimethods (such asmolecular dynamicssimulations). The advantage of a priori methods is the less or nodependent on the experimental data, especially for new compoundsor molecules that have new groups. However, the predicted resultsbased only on molecular structure instead of experimental data areusually less accurate than that based on experimental data. And sincethe transport properties are seriously affected by small amount of impu-rities, a comprehensive comparison on the performance of differentmethods should be carried out to give chemists and engineering estima-tion on the error of the predicted data, just as we have done on thecomparison of the performance of COSMO-RS and UNIFAC (Do) forthe prediction of the activity coefficient at infinite dilution and vaporliquid equilibrium [84–87]. Such comparison should be carried out
266 X. Wang et al. / Journal of Molecular Liquids 193 (2014) 262–266
based on a large amount of experimental data obtained from differentauthors and exclude the unreliable data. Before such comparison ispublished, we recommendwith caution that it is better to use predicteddata based on experimental results than that from a priori methods.Even very few experimental data can benefit the prediction.
5. Conclusion
The transport properties of ILs are of importance for the applicationsof ILs. A lot of studies have been carried out on this topic. The resultsshow that temperature has significant influences on most of thetransport properties. The temperature effect on diffusion coefficient,viscosity, and ionic conductivity of ILs can be correlated with VFT equa-tion. The transport properties of ILswithmolecular solvents or inorganicsalts are also important especially for industrial applications of ILs suchas usage as electrolyte of battery. The accurate experimental data areprerequisite for industrial processes design as well as the developmentof predictive models.
Acknowledgment
Thisworkwas supported by the Plan Project of Science and Technol-ogy of BeijingMunicipal Education Committee (KM201210016007) andthe National Natural Science Foundation of China (21173267).
References
[1] T. Mu, B. Han, Struct. Bond., 151, Springer-Verlag Berlin and Heidelberg GmbH & Co.KG, 2014, pp. 107–139.
[2] J.A.P. Coutinho, P.J. Carvalho, N.M.C. Oliveira, RSC Adv. 2 (2012) 7322–7346.[3] Z. Xue, Z. Zhang, J. Han, Y. Chen, T. Mu, Int. J. Greenhouse Gas Control 5 (2011)
628–633.[4] X. Cheng, G. Yang, T. Mu, X. Guo, X. Wang, Clean-Soil Air Water 37 (2009) 245–248.[5] Z.H. Li, Z. Jia, Y.X. Luan, T.C. Mu, Curr. Opin. Solid State Mater. Sci. 12 (2009) 1–8.[6] Z. Li, R. Li, T. Mu, Y. Luan, Chem. Eur. J. 19 (2013) 6005–6013.[7] Y. Chen, J. Han, T. Wang, T. Mu, Energy Fuel 25 (2011) 5810–5815.[8] A.P. Fröba, M.H. Rausch, K. Krzeminski, D. Assenbaum, P. Wasserscheid, A. Leipertz,
Int. J. Thermophys. 31 (2010) 2059–2077.[9] D. Tomida, S. Kenmochi, T. Tsukada, K. Qiao, C. Yokoyama, Int. J. Thermophys. 31
(2010) 1888–1895.[10] R. Ge, C. Hardacre, P. Nancarrow, D.W. Rooney, J. Chem. Eng. Data 52 (2007)
1819–1823.[11] C. Frez, G.J. Diebold, C.D. Tran, S. Yu, J. Chem. Eng. Data 51 (2006) 1250–1255.[12] A. Bagno, F. D'Amico, G. Saielli, J. Mol. Liq. 131 (2007) 17–23.[13] R. Brookes, A. Davies, G. Ketwaroo, P.A. Madden, J. Phys. Chem. B 109 (2005)
6485–6490.[14] A. Heintz, J.K. Lehmann, E. Schmidt, A. Wandschneider, J. Solut. Chem. 38 (2009)
1079–1083.[15] S. Sarraute, M.F.C. Gomes, A.A.H. Padua, J. Chem. Eng. Data 54 (2009) 2389–2394.[16] P. Bonhote, A.P. Dias, N. Papageorgiou, K. Kalyanasundaram, M. Gratzel, Inorg. Chem.
35 (1996) 1168–1178.[17] A.B. McEwen, H.L. Ngo, K. LeCompte, J.L. Goldman, J. Electrochem. Soc. 146 (1999)
1687–1695.[18] H. Tokuda, S. Tsuzuki, M.A.B.H. Susan, K. Hayamizu, M. Watanabe, J. Phys. Chem. B
110 (2006) 19593–19600.[19] A.P. Fröba, H. Kremer, A. Leipertz, J. Phys. Chem. B 112 (2008) 12420–12430.[20] R.L. Gardas, J.A.P. Coutinho, Fluid Phase Equilib. 266 (2008) 195–201.[21] Y. Xiong, J. Ding, D. Yu, C. Peng, H. Liu,熊焰,丁靖,虞大红,彭昌军,刘洪来, J. Chem. Ind.
Eng. (China) 63 (2012) 667–676.[22] R.N. Das, K. Roy, Mol. Divers. 17 (2013) 151–196.[23] D.C. Weis, D.R. MacFarlane, Aust. J. Chem. 65 (2012) 1478–1486.[24] H. Niedermeyer, J.P. Hallett, I.J. Villar-Garcia, P.A. Hunt, T. Welton, Chem. Soc. Rev. 41
(2012) 7780–7802.[25] J. Jacquemin, P. Husson, A.A.H. Padua, V. Majer, Green Chem. 8 (2006) 172–180.[26] H. Tokuda, K. Ishii, M.A.B.H. Susan, S. Tsuzuki, K. Hayamizu, M. Watanabe, J. Phys.
Chem. B 110 (2006) 2833–2839.[27] H. Tokuda, K. Hayamizu, K. Ishii, M.A.B.H. Susan, M. Watanabe, J. Phys. Chem. B 108
(2004) 16593–16600.[28] H. Tokuda, K. Hayamizu, K. Ishii, M.A.B.H. Susan, M. Watanabe, J. Phys. Chem. B 109
(2005) 6103–6110.[29] N. Papaiconomou, O. Zech, P. Bauduin, J.-M. Leveque, W. Kunz, Electrochim. Acta 70
(2012) 124–130.[30] K. Machanova, A. Boisset, Z. Sedlakova, M. Anouti, M. Bendova, J. Jacquemin, J. Chem.
Eng. Data 57 (2012) 2227–2235.
[31] A. Ahosseini, A.M. Scurto, Int. J. Thermophys. 29 (2008) 1222–1243.[32] K.R. Harris, M. Kanakubo, L.A. Woolf, J. Chem. Eng. Data 51 (2006) 1161–1167.[33] K.R. Harris, M. Kanakubo, L.A. Woolf, J. Chem. Eng. Data 52 (2007) 2425–2430.[34] K.R. Harris, M. Kanakubo, L.A. Woolf, J. Chem. Eng. Data 52 (2007) 1080–1085.[35] K.R. Harris, M. Kanakubo, L.A. Woolf, J. Chem. Eng. Data 53 (2008) 1230.[36] K.R. Harris, L.A. Woolf, M. Kanakubo, J. Chem. Eng. Data 50 (2005) 1777–1782.[37] K.R. Harris, M. Kanakubo, Faraday Discuss. 154 (2012) 425–438.[38] T.-Y. Wu, I.W. Sun, S.-T. Gung, M.-W. Lin, B.-K. Chen, H.P. Wang, S.-G. Su, J. Taiwan
Inst. Chem. Eng. 42 (2011) 513–522.[39] F.J. Deive, M.A. Rivas, A. Rodriguez, J. Chem. Thermodyn. 62 (2013) 98–103.[40] M.D. Green, C. Schreiner, T.E. Long, J. Phys. Chem. A 115 (2011) 13829–13835.[41] K.R. Seddon, A. Stark, M.J. Torres, Pure Appl. Chem. 72 (2000) 2275–2287.[42] Y. Cao, Y. Chen, X. Sun, Z. Zhang, T. Mu, Phys. Chem. Chem. Phys. 14 (2012)
12252–12262.[43] Y. Cao, Y. Chen, L. Lu, Z. Xue, T. Mu, Ind. Eng. Chem. Res. 52 (2013) 2073–2083.[44] Y. Chen, Y. Cao, X. Lu, C. Zhao, C. Yan, T. Mu, New J. Chem. 37 (2013) 1959–1967.[45] J.A. Widegren, A. Laesecke, J.W. Magee, Chem. Commun. (2005) 1610–1612.[46] N.D. Khupse, A. Kumar, Proc. Nat. Acad. Sci. India Sect. A Phys. Sci. 80A (2010) 1–12.[47] J.A. Widegren, J.W. Magee, J. Chem. Eng. Data 52 (2007) 2331–2338.[48] H. Shekaari, Y. Mansoori, R. Sadeghi, J. Chem. Thermodyn. 40 (2008) 852–859.[49] A.P. Fröba, P. Wasserscheid, D. Gerhard, H. Kremer, A. Leipertz, J. Phys. Chem. B 111
(2007) 12817–12822.[50] P. Navia, J. Troncoso, L. Romani, J. Solut. Chem. 37 (2008) 677–688.[51] E. Gomez, B. Gonzalez, A. Dominguez, E. Tojo, J. Tojo, J. Chem. Eng. Data 51 (2006)
696–701.[52] U. Domanska, M. Laskowska, J. Solut. Chem. 38 (2009) 779–799.[53] U. Domanska, M. Laskowska, J. Chem. Eng. Data 54 (2009) 2113–2119.[54] T.Y. Wu, H.C. Wang, S.G. Su, S.T. Gung, M.W. Lin, C.B. Lin, J. Taiwan Inst. Chem. Eng.
41 (2010) 315–325.[55] Z.M. Liu, W.Z. Wu, B.X. Han, Z.X. Dong, G.Y. Zhao, J.Q. Wang, T. Jiang, G.Y. Yang,
Chem. Eur. J. 9 (2003) 3897–3903.[56] T.J. Simons, P.C. Howlett, A.A.J. Torriero, D.R. MacFarlane, M. Forsyth, J. Phys. Chem. C
117 (2013) 2662–2669.[57] P.M. Bayley, A.S. Best, D.R. MacFarlane, M. Forsyth, Phys. Chem. Chem. Phys. 13
(2011) 4632–4640.[58] P.M. Bayley, A.S. Best, D.R. MacFarlane, M. Forsyth, ChemPhysChem 12 (2011)
823–827.[59] Q. Zhou, P.D. Boyle, L. Malpezzi, A. Mele, J.-H. Shin, S. Passerini, W.A. Henderson,
Chem. Mater. 23 (2011) 4331–4337.[60] R.C. Reid, J.M. Prausnitz, T.K. Sherwood, McGraw-Hill, New York, (1987).[61] R.L. Gardas, J.A.P. Coutinho, AIChE J. 55 (2009) 1274–1290.[62] P.J. Carvalho, T. Regueira, L. Santos, J. Fernandez, J.A.P. Coutinho, J. Chem. Eng. Data
55 (2010) 645–652.[63] M.H. Ghatee, M. Zare, F. Moosavi, A.R. Zolghadr, J. Chem. Eng. Data 55 (2010)
3084–3088.[64] M.H. Ghatee, M. Zare, A.R. Zolghadr, F. Moosavi, Fluid Phase Equilib. 291 (2010)
188–194.[65] M.T. Zafarani-Moattar, R. Majdan-Cegincara, J. Chem. Eng. Data 52 (2007)
2359–2364.[66] B. Mokhtarani, M.M. Mojtahedi, H.R. Mortaheb, M. Mafi, F. Yazdani, F. Sadeghian, J.
Chem. Eng. Data 53 (2008) 677–682.[67] B. Mokhtarani, A. Sharifi, H.R. Mortaheb, M. Mirzaei, M. Mafi, F. Sadeghian, J. Chem.
Thermodyn. 41 (2009) 323–329.[68] Y.F. Geng, T.F. Wang, D.H. Yu, C.J. Peng, H.L. Liu, Y. Hu, Chin. J. Chem. Eng. 16 (2008)
256–262.[69] L. Grunberg, A.H. Nissan, Nature (1949) 799–800.[70] J.D. Isdale, Natl. Eng. Lab. East Kilbride, Glasgow, Scotland (1979).[71] P.K. Katti, M.M. Chandhri, J. Chem. Eng. Data 9 (1964) 442–443.[72] L. Grunberg, A.H. Nissan, Nature 164 (1949) 799–800.[73] Dortmund Data Bank and DDB Software Package, DDBST GmbH, Oldenburg,
Germany, 2013. (www.ddbst.de ).[74] M.H. Kowsari, S. Alavi, B. Najafi, K. Gholizadeh, E. Dehghanpisheh, F. Ranjbard, Phys.
Chem. Chem. Phys. 13 (2011) 8826–8837.[75] V.V. Chaban, I.V. Voroshylova, O.N. Kalugin, Phys. Chem. Chem. Phys. 13 (2011)
7910–7920.[76] O. Borodin, G.D. Smith, J. Phys. Chem. B 110 (2006) 11481–11490.[77] O. Borodin, J. Phys. Chem. B 113 (2009) 11463–11478.[78] O. Borodin, G.D. Smith, H. Kim, J. Phys. Chem. B 113 (2009) 4771–4774.[79] B.L. Bhargava, S. Balasubramanian, J. Chem. Phys. 123 (2005).[80] O. Andreussi, N. Marzari, J. Chem. Phys. 137 (2012).[81] E. Androulaki, N. Vergadou, J. Ramos, I.G. Economou, Mol. Phys. 110 (2012)
1139–1152.[82] H. Liu, E. Maginn, A.E. Visser, N.J. Bridges, E.B. Fox, Ind. Eng. Chem. Res. 51 (2012)
7242–7254.[83] Z. Li, G.D. Smith, D. Bedrov, J. Phys. Chem. B 116 (2012) 12801–12809.[84] Z. Xue, T. Mu, J. Gmehling, Ind. Eng. Chem. Res. 51 (2012) 11809–11817.[85] T.C. Mu, J. Rarey, J. Gmehling, Ind. Eng. Chem. Res. 46 (2007) 6612–6629.[86] T.C. Mu, J. Rarey, J. Gmehling, AIChE J. 53 (2007) 3231–3240.[87] T.C. Mu, J. Rarey, J. Gmehling, AIChE J. 55 (2009) 3298–3300