Article pubs.acs.org/JPCB
Molecular Topology and Local Dynamics Govern the Viscosity of Imidazolium-Based Ionic Liquids Yong Zhang,† Lianjie Xue,‡ Fardin Khabaz,§ Rose Doerfler,† Edward L. Quitevis,*,‡ Rajesh Khare,*,§ and Edward J. Maginn*,† †
Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, Indiana 46556, United States Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, Texas 79409, United States § Department of Chemical Engineering, Texas Tech University, Lubbock, Texas 79409, United States ‡
ABSTRACT: A series of branched ionic liquids (ILs) based on the 1-(iso-alkyl)-3-methylimidazolium cation from 1-(1methylethyl)-3-methylimidazolium bistriflimide to 1-(5-methylhexyl)-3-methylimidazolium bistriflimide and linear ILs based on the 1-(n-alkyl)-3-methylimidazolium cation from 1propyl-3-methylimidazolium bistriflimide to 1-heptyl-3-methylimidazolum bistriflimide were recently synthesized and their physicochemical properties characterized. For the ILs with the same number of carbons in the alkyl chain, the branched IL was found to have the same density but higher viscosity than the linear one. In addition, the branched IL 1-(2methylpropyl)-3-methylimidazolium bistriflimide ([2mC3C1Im][NTf 2]) was found to have an abnormally high viscosity. Motivated by these experimental observations, the same ILs were studied using molecular dynamics (MD) simulations in the current work. The viscosities of each IL were calculated using the equilibrium MD method at 400 K and the nonequilibrium MD method at 298 K. The results agree with the experimental trend. The ion pair (IP) lifetime, spatial distribution function, and associated potential of mean force, cation size and shape, and interaction energy components were calculated from MD simulations. A quantitative correlation between the liquid structure and the viscosity was observed. Analysis shows that the higher viscosities in the branched ILs are due to the relatively more stable packing between the cations and anions indicated by the lower minima in the potential of mean force (PMF) surface. The abnormal viscosity of [2mC3C1Im][NTf 2] was found to be the result of the specific side chain length and molecular structure.
1. INTRODUCTION Ionic liquids (ILs) are defined as salts that melt at or below 373 K. A typical IL is composed of a bulky organic cation and an inorganic or organic anion. As a group, many ILs tend to have low volatility, reasonable conductivity, and a wide electrochemical window and tend to be miscible with many organic solvents. These special properties make ILs attractive alternatives in synthesis,1 chemical separations,2−4 electrochemistry,5,6 catalysis,7 and solar cells.8 Another unique characteristic of ILs is that their physicochemical properties can be easily tuned by changing the combination or the structure of their component ions. Thus, ILs are also called designer solvents or task-specific solvents.9 Thus, understanding the correlation between molecular structure and macroscopic properties is necessary for the rational design of ILs for certain specific applications, and there has been extensive research carried out toward this objective. Watanabe and co-workers10 systematically studied the correlation between alkyl chain length and physicochemical properties of imidazolium-based ILs, and concluded that the change in the alkyl chain length causes a change in the interaction forces, and the properties of the ILs are determined by the cumulative effect of the electrostatic interaction between © 2015 American Chemical Society
the ionic species and the induction interactions between the ions, aggregates, and clusters. Zhang and Maginn11 studied the melting points of a series of 1-alkyl-3-methylimidazolium hexafluorophosphate ILs ([CnC1Im][PF6]) with n = 2, 4, 10, 12, and 14 using molecular dynamics (MD) simulations. It was found that, for the ILs with long alkyl chains, melting points increase with the alkyl chain length as a result of the increased fusion enthalpies due to the enhanced van der Waals interaction brought about by the long nonpolar alkyl chains. For the ILs with short alkyl chains, fusion entropy plays a dominant role so that [C4C1Im][PF6] has a lower melting temperature than [C2C1Im][PF6]. Entropy also plays an essential role in deciding dynamics in certain ILs. For example, the higher viscosities of the imidazolium ILs with the substitution of a methyl group for a hydrogen at the C2 position of the cation ring relative to the C2 protonated cations are due to the lower entropy in the former.12,13 The effect of cation alkyl chain length and cation symmetry on morphology and physicochemical properties was also Received: August 24, 2015 Revised: October 16, 2015 Published: October 27, 2015 14934
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B Table 1. Structures of the Linear [CnC1Im][NTf 2] and Branched [(n − 2)mCn−1C1Im][NTf 2] ILs
systematically studied by Quitevis,14 Triolo,15,16 and Maroncelli17 and co-workers. Schubert18 and Husson19 studied the effect of branch structure on the physicochemical properties of imidazolium-based ILs. Their results showed that branched ILs have higher viscosities and glass transition temperatures than their linear analogues. Recently, Xue et al.20 synthesized a series of branched imidazolium-based ILs (1-(iso-alkyl)-3-methylimidazolium bistriflimide, [(n − 2)mCn−1C1Im][NTf 2]) (see Table 1). The viscosities of each IL were measured and compared with a series of linear ILs (1-(n-alkyl)-3-methylimidazolium bistriflimide, [CnC1Im][NTf 2]). They found that the viscosity of each branched IL was higher than that of its linear analogue. For both branched and linear ILs, except for [2mC3C1Im][NTf 2], the viscosities increased with increasing alkyl chain length. On the other hand, [2mC3C1Im][NTf 2] showed an abnormally high viscosity, even higher than that of
[3mC4C1Im][NTf 2]. These experimental observations motivated the current simulation study.
2. SIMULATION PROCEDURES Both equilibrium (EMD) and nonequilibrium molecular dynamics (NEMD) methods were used in the current study. A common simulation setup is described below followed by specific simulation details for EMD and NEMD methods, respectively. 2.1. Force Field and Common Simulation Setup. Classical MD simulations were carried out using the package LAMMPS.21 Periodic boundary conditions were applied, and a time step of 1 fs was used. The long-range electrostatic interactions were calculated using the particle−particle particle−mesh (PPPM) method22 with a real space cutoff of 12 Å. The Nosé−Hoover thermostat23 and the extended 14935
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B
every 5 fs for each trajectory. The first 1 ns was considered equilibration and discarded, and the rest of the trajectory was used to calculate viscosity using eq 1. The parameter tcut was determined by taking the time that the calculated standard deviation reached the value of 40% of the estimated viscosity of each system (the value of the flat region of the averaged running integral).32 2.3. NEMD Simulations. To prepare the model structure for NEMD simulations of the IL, initially, 1728 ion pairs were placed in a cubic simulation box at a low density. The density of the system was then increased by subjecting it to an MD simulation at a pressure of 150 atm. The simulation box was subsequently relaxed for a period of 2 ns at temperature T = 298 K and P = 1 atm. Five replicas were built for each IL. The viscosity calculations were performed using the NEMD method34 at constant NVT conditions. The NEMD method has previously been used to compare the viscosity of linear and branched alkanes.35 Six different shear rates in the range 107− 109 s−1 were studied.
Lagrangian approach24 were applied to control the temperature and pressure. The general Amber force field (GAFF)25 was used to describe the interactions in the systems. In order to derive atomic charges, electronic structure calculations were carried out on an isolated ion to optimize the structures at the B3LYP/ 6-311++g(d,p) level using the Gaussian 09 package.26 The atomic charges were then derived on the basis of the optimized structure by fitting the electrostatic potential surface obtained from these calculations using the RESP method.27 Such charges have a total value of ±1 e on the cation and anion, respectively. To include the effect of charge transfer and polarizability in the bulk phase, the partial charges were scaled uniformly by 0.8. The 0.8-scaled charges have been found to be reliable for the study of dynamic properties of similar ILs.28 2.2. EMD Simulations. EMD simulation details closely followed the procedure described in our previous publications.13,29 The simulation box was built up for each IL by putting 300 ion pairs randomly in a cubic box using the package Packmol.30,31 The systems were then equilibrated for 2 ns in the isothermal−isobaric (NPT) ensemble followed by 20 ns production runs in the canonical (NVT) ensemble. The densities were calculated on the basis of the NPT ensemble simulation, and the last 18 ns of the NVT trajectories were used for analysis. The pressure was fixed at 1 atm in all constant pressure simulations with isotropic volume fluctuations. The shear viscosity of each IL was calculated using the recently developed time-decomposition method.32 In this method, the shear viscosity was calculated from the integral over time of the stress tensor autocorrelation function following the Green−Kubo relation η=
V kBT
∫0
3. RESULTS AND DISCUSSION 3.1. Density Is Independent of Chain Structure. The calculated densities of the ILs as a function of temperature are shown in Figure 1. The experimentally measured densities20 are
∞
⟨ταβ(t ) ·ταβ(0)⟩ dt
(1)
where V is the system volume, kB is the Boltzmann constant, and ταβ denotes the αβ-component of the stress tensor. For better statistics, the shear viscosity was calculated by averaging over six independent terms of the stress tensor, namely, τxy, τyz, τzx, 0.5(τxx − τyy), 0.5(τyy − τzz), and 0.5(τxx − τzz).33 Multiple short independent trajectories were generated, and the running integral in eq 1 was calculated for each trajectory. The average of the running integrals was then calculated. The corresponding time-dependent standard deviation of the running integrals was also calculated as σ (t ) =
1 N−1
Figure 1. Liquid phase density as a function of temperature from experiments20 and MD simulations. Calculated results were fit by a linear function and extrapolated to experimental temperatures (shown as dashed lines). The largest deviation in the calculated density from experiments is less than 0.6%. A direct comparison of calculated and experimental densities at 298 K is provided in Table 2.
N
also provided for comparison. In the experimental results, for a given chain length n, the density of the branched IL almost equals that of the linear IL. The largest difference was found for the n = 3 pair, in which the density of the branched IL is slightly higher than that of the linear one by 0.13%. From the figure, we can see that the densities of the branched and linear ILs vary linearly in the temperature range. For faster dynamics and better sampling, MD simulations were carried out at higher temperatures between 350 and 500 K with 50 K intervals. Consistent with experimental results, for a given n, the densities of the branched IL and the linear IL are very close to each other. Except for n = 3, the branched IL has a slightly (160 times the bulk density, which is much higher than that of ∼40 times bulk density observed in [C3C1Im][NTf 2]. This suggests that the preferred distribution of the CT atoms in this region in [2mC3C1Im][NTf 2] will block the approach of the anions from the side of the imidazolium ring and results in the SDF shown in Figure 5. This steric effect reduces anion mobility and is the cause of the abnormally high viscosity in this IL. When the side chains become even longer, for both linear and branched ILs, the CT distribution becomes less localized due to the flexibility of the alkyl chain (results not shown). The sizes of the cations were characterized using the radius of gyration Rg, whereas the asphericity (b) and acylindricity (c) parameters were calculated to characterize the shape of the cations. Recently, these quantities were used to study the relationship between the viscosity and the chain shape and size in dilute polymer solutions.43 These quantities were calculated using the following expressions44 S̲ =
N ∑i = 1 (R̲ i − R̲ com)2
R g 2 = λ1 + λ 2 + λ3
b = λ1 − 0.5(λ 2 + λ3)
c = λ 2 − λ3
(7)
N λ1 > λ 2 > λ3
(8) (9) (10)
where S̲ is the gyration tensor, R̲ i is the position vector of each atom, R̲ com is the center of mass of a cation, N is the total number of atoms per cation, and λi’s are the eigenvalues of S̲ . If a cation has a perfectly spherical shape, then the parameter b will have a value of zero, and if a cation has a perfectly cylindrical shape, then the parameter c will have a value of zero. Figure 10 (top panel) shows the mean squared radius of gyration ⟨Rg2⟩ of cations of linear ILs and branched ILs plotted as a function of carbon number. As can be seen in the figure, the ⟨Rg2⟩ values of branched IL cations are systematically smaller than those of the corresponding linear cations. The extent of asphericity (middle panel in Figure 10) and acylindricity (bottom panel in Figure 10) of the cations were plotted as a function of carbon number. As expected, the degree of asphericity (deviation from the spherical shape) increases monotonically with the number of carbon atoms for both linear and branched ILs. Furthermore, the cations of the branched ILs are more spherical in shape than the cations of the corresponding linear ILs. Due to the relative rigidity of [1mC 2 C 1 Im] + and [2mC 3 C 1 Im] + compared to other branched-chain cations, the differences in b/Rg2 between linear and branched ILs are larger for n = 3 and 4. The value of the acylindricity (deviation from the cylindrical shape) of the cations of the linear ILs and branched ILs, in general, decreases monotonically with the number of carbon atoms; i.e., cation shape becomes more cylindrical as chain length is increased.
Figure 10. Top: Mean squared radius of gyration of cations of linear and branched ILs as a function of carbon number n. Middle: Normalized asphericity of cations for linear and branched ILs as a function of carbon number n. Bottom: Normalized acylindricity of cations for linear and branched ILs as a function of carbon number n.
Two exceptions to this behavior, however, are the acylindricity of the cation of [1mC2C1Im][NTf 2] and [2mC3C1Im][NTf 2]. Due to its specific size and structure (see Figure 8), [2mC3C1Im]+ exhibits an even larger deviation from the cylindrical shape than [1mC2C1Im]+. These results are consistent with the observation presented earlier in this section that the folding back of the methyl groups in [2mC3C1Im][NTf 2] prevents approach of anions to the side of the imidazolium ring, thus, in turn, leading to higher viscosity for this branched IL. 3.6. Viscosity Is Not Sensitive to Charge Localization. To investigate the effect of intermolecular interactions on the liquid phase structure and therefore the dynamics such as viscosity, the total nonbonded energy (i.e., sum of van der Waals and electrostatic energy) per atom of the linear and branched systems was determined and plotted as a function of the number of carbon atoms (Figure 11, top panel). The total 14941
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B
high attractive energy per atom. To determine the origin of this observation, the partial charges on the imidazolium ring of the ILs are also plotted in the same figure (right axis). Except for n = 3, the partial charges in the imidazolium ring are larger in the branched cations compared to those of the linear cations with the same carbon number; i.e., the partial charges are more localized on the imidazolium ring in the branched cations. These graphs clearly show that the electrostatic energy per atom is directly correlated with the magnitude of the localized partial charge on the imidazolium ring. It is also interesting to note that [2mC3C1Im][NTf 2] has the most attractive nonbonded interactions in the ILs studied in the current work, consistent with its most localized partial charges on the imidazolium ring (see Figure 11). At the outset, these results suggest that the origin of the abnormally high viscosity of the [2mC3C1Im][NTf 2] branched IL is related to the higher localized partial charge on the imidazolium ring of this molecule. To further test this effect of charge localization on the molecular interaction and viscosity, a model system was set up in which the partial charges of [C5C1Im]+ were applied to [3mC4C1Im]+. All of the other parameters in the force field for [3mC4C1Im]+ and [NTf 2]− were kept the same. The viscosity of the modified [3mC4C1Im][NTf 2] was computed using both EMD and NEMD methods. Using the EMD method, the viscosity was found to be 8.25 mPa·s at 400 K, which is about halfway between the viscosities of the original [C5C1Im][NTf 2] (7.79 mPa·s) and [3mC4C1Im][NTf 2] (8.75 mPa·s). On the other hand, using NEMD at a temperature of 298 K, the viscosity of the charge modified [3mC4C1Im][NTf 2] was found to be the same as that of the original [3mC4C1Im][NTf 2] molecule within statistical uncertainties. These results suggest that the origin of the abnormally high viscosity of the branched IL with n = 4 is not in the charge localization; the effect is primarily governed by the molecular structure as discussed in the previous section.
4. CONCLUDING REMARKS A series of ILs with a branched alkyl chain ([(n − 2) mCn−1C1Im][NTf 2], n = 3−7) and a linear alkyl chain ([CnC1Im][NTf 2]) were studied using classical MD simulations. The calculated densities showed excellent agreement with the experimental values. The experimental trend in viscosity was also successfully captured using both EMD and NEMD methods. On the basis of MD simulations, the trend in viscosities is connected to the microscopic IP lifetimes, consistent with a previous study.29 The IP lifetimes were found to be linearly correlated with the depth of minima on the PMF surface derived from spatial distribution functions of the ions in the liquid phase. The higher viscosities of the ILs with branched side chains relative to the ones with a linear side chain are likely caused by the relatively more stable packing between the cations and anions in the liquid phase of the branched ILs compared to the linear ILs, as indicated by the lower free energy minima in the branched IL systems. The abnormally high viscosity of [2mC3C1Im][NTf 2] was found to be caused by the specific shape and length of its alkyl side chain, which blocks the approach of anions to the side of the imidazolium ring. Our results suggest a direct quantitative connection between the transport properties of the ILs and their liquid phase structure. It is known that transport properties such as viscosity are relatively hard to calculate and the results usually involve a
Figure 11. Calculated total (sum of van der Waals and electrostatic, top panel), van der Waals (middle), and electrostatic (bottom) energies per atom in linear and branched ILs at 400 K. Partial charges localized on the imidazolium ring of the cations are also provided (bottom panel, right axis).
nonbonded energy/atom of the linear ILs shows a weak increase with an increase in the number of carbon atoms. On the other hand, the nonbonded energy/atom exhibits a much stronger variation for the branched ILs. It is interesting to note that, except for n = 3, the total nonbonded energy/atom is lower in the branched IL compared to that of the linear analogue with the same carbon number and, furthermore, the cation size dependence is not monotonic. To further investigate this, the two energy components, van der Waals and electrostatic interactions, were studied separately. As shown in Figure 11 (middle panel), the van der Waals energy values per atom for the linear and branched ILs with the same carbon numbers are very close to each other. For both linear and branched ILs, a monotonic increase is observed with an increase in carbon number. As seen in the bottom panel of Figure 11, the origin of the nonmonotonic behavior of the total nonbonded energy is in the electrostatic energy contribution per atom. Furthermore, the branched IL with n = 4 (the one that shows an abnormally high viscosity) exhibits an abnormally 14942
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B large uncertainty,32 whereas it is relatively easy to get accurate results of structural properties such as radial distribution functions (RDFs) and SDFs. The quantitative correlation between these properties demonstrated here suggests that the liquid structural properties can be used to predict transport properties of ILs, which, in turn, could significantly speed up the design of new ILs for a given application.
■
(12) Hunt, P. A. Why Does a Reduction in Hydrogen Bonding Lead to an Increase in Viscosity for the 1-Butyl-2,3-dimethyl-imidazoliumbased Ionic Liquids? J. Phys. Chem. B 2007, 111, 4844−4853. (13) Zhang, Y.; Maginn, E. J. The Effect of C2 Substitution on Melting Point and Liquid Phase Dynamics of Imidazolium Based-ionic Liquids: Insights from Molecular Dynamics Simulations. Phys. Chem. Chem. Phys. 2012, 14, 12157−12164. (14) Zheng, W.; Mohammed, A.; Hines, L. G., Jr.; Xiao, D.; Martinez, O. J.; Bartsch, R. A.; Simon, S. L.; Russina, O.; Triolo, A.; Quitevis, E. L. Effect of Cation Symmetry on the Morphology and Physicochemical Properties of Imidazolium Ionic Liquids. J. Phys. Chem. B 2011, 115, 6572−6584. (15) Russina, O.; Triolo, A.; Gontrani, L.; Caminiti, R.; Xiao, D.; Hines, L. G., Jr.; Bartsch, R. A.; Quitevis, E. L.; Plechkova, N.; Seddon, K. R. Morphology and Intermolecular Dynamics of 1-Alkyl-3methylimidazolium Bis((trifluoromethane)sulfonyl)amide Ionic Liquids: Structural and Dynamic Evidence of Nanoscale Segregation. J. Phys.: Condens. Matter 2009, 21, 424121. (16) Rocha, M. A. A.; Neves, C. M. S. S.; Greire, M.; Russina, O.; Triolo, A.; Coutinho, J. A. P.; Santos, L. M. N. B. F. Alkylimidazoilium Based Ionic Liquids: Impact of Cation Symmetry on Their Nanoscale Structural Organization. J. Phys. Chem. B 2013, 117, 10889−10897. (17) Jin, H.; O’Hare, B.; Dong, J.; Arzhantsev, S.; Baker, G. A.; Wishart, J. F.; Benesi, A. J.; Maroncelli, M. Physical Properties of Ionic Liquids Consisting of the 1-Butyl-3-Methylimidazolium Cation with Various Anions and the Bis(trifluoromethylsulfonyl)imide Anion with Various Cations. J. Phys. Chem. B 2008, 112, 81−92. (18) Erdmenger, T.; Vitz, J.; Wiesbrock, F.; Schubert, U. S. Influence of Different Branched Alkyl Side Chains on the Properties of Imidazolium-based Ionic Liquids. J. Mater. Chem. 2008, 18, 5267− 5273. (19) Andresova, A.; Storch, J.; Traikia, M.; Wagner, Z.; Bendova, M.; Husson, P. Branched and Cyclic Alkyl Groups in Imidazolium-based Ionic Liquids: Molecular Organization and Physico-Chemical Properties. Fluid Phase Equilib. 2014, 371, 41−49. (20) Xue, L.; Tamas, G.; Koh, Y. P.; Shadeck, M.; Gurung, E.; Simon, S. L.; Maroncelli, M.; Quitevis, E. L. Effect of Alkyl Branching on Physicochemical Properties of Imidazolium-based Ionic Liquids. J. Chem. Eng. Data, submitted for publication, 2015. (21) Plimpton, S. Fast Parallel Algorithms for Short-range Molecular Dynamics. J. Comput. Phys. 1995, 117, 1−19. (22) Hockney, R.; Eastwood, J. Computer Simulation Using Particles; Adam Hilger: New York, 1989. (23) Hoover, W. G. Canonical Dynamics: Equilibrium Phase-space Distributions. Phys. Rev. A: At., Mol., Opt. Phys. 1985, 31, 1695−1697. (24) Shinoda, W.; Shiga, M.; Mikami, M. Rapid Estimation of Elastic Constants by Molecular Dynamics Simulation under Constant Stress. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 69, 134103. (25) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004, 25, 1157−1174. (26) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (27) Bayly, C.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A Wellbehaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: the RESP Model. J. Phys. Chem. 1993, 97, 10269−10280. (28) Zhang, Y.; Maginn, E. J. A Simple AIMD Approach to Derive Atomic Charges for Condensed Phase Simulation of Ionic Liquids. J. Phys. Chem. B 2012, 116, 10036−10048. (29) Zhang, Y.; Maginn, E. J. Direct Correlation Between Ionic Liquid Transport Properties and Ion Pair Lifetimes: A Molecular Dynamics Study. J. Phys. Chem. Lett. 2015, 6, 700−705. (30) Martinez, J. M.; Martinez, L. Packing Optimization for Automated Generation of Complex System’s Initial Configurations for Molecular Dynamics and Docking. J. Comput. Chem. 2003, 24, 819−825.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Y.Z. and E.J.M. are supported by the U.S. Department of Energy, Basic Energy Science, Joint Center for Energy Storage Research under Contract No. DE-AC02-06CH11357. Computational resources were provided by the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, and the Center for Research Computing (CRC) at the University of Notre Dame. L.X. and E.L.Q. are supported by the National Science Foundation (NSF) under Grant No. CHE-1153077.
■
REFERENCES
(1) Iranpoor, N.; Firouzabadi, H.; Azadi, R. A New Diphenylphosphinite Ionic Liquid (IL-OPPh2) as Reagent and Solvent for Highly Selective Bromination, Thiocyanation or Isothiocyanation of Alcohols and Trimethylsilyl and Tetrahydropyranyl Ethers. Tetrahedron Lett. 2006, 47, 5531−5534. (2) Tian, K.; Qi, S.; Cheng, Y.; Chen, X.; Hu, Z. Separation and Determination of Lignans from Seeds of Schisandra Species by Micellar Electrokinetic Capillary Chromatography. J. Chromatogr. A 2005, 1078, 181−187. (3) Llovell, F.; Oliveira, M. B.; Coutinho, J. A. P.; Vega, L. F. Solubility of Greenhouse and Acid Gases on the [C4mim][MeSO4] Ionic Liquid for Gas Separation and CO2 Conversion. Catal. Today 2015, 255, 87−96. (4) Zhang, Z.; Hu, A.; Zhang, T.; Zhang, Q.; Sun, M.; Sun, D. Separation of Methyl Acetate + Methanol Azeotropic Mixture Using Ionic Liquids as Entrainers. Fluid Phase Equilib. 2015, 401, 1−8. (5) Hapiot, P.; Lagrost, C. Electrochemical Reactivity in Roomtemperature Ionic Liquids. Chem. Rev. 2008, 108, 2238−2264. (6) MacFarlane, D. R.; Forsyth, M.; Howlett, P. C.; Pringle, J. M.; Sun, J.; Annat, G.; Neil, W.; Izgorodina, E. I. Ionic Liquids in Electrochemical Devices and Processes: Managing Interfacial Electrochemistry. Acc. Chem. Res. 2007, 40, 1165−1173. (7) Ranu, B. C.; Banerjee, S.; Jana, R. Ionic Liquid as Catalyst and Solvent: The Remarkable Effect of a Basic Ionic Liquid, [Bmim]OH on Michael Addition and Alkylation of Active Methylene Compounds. Tetrahedron 2007, 63, 776−782. (8) Tsuda, T.; Hussey, C. L. Electrochemical Applications of Roomtemperature Ionic Liquids. Electrochem. Soc. Interface 2007, 42−49. (9) Earle, M. J.; Seddon, K. R. Ionic Liquids. Green Solvents for the Future. Pure Appl. Chem. 2000, 72, 1391−1398. (10) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. Physicochemical Properties and Structures of Room Temperature Ionic Liquids. 2. Variation of Alkyl Chain Length in Imidazolium Cation. J. Phys. Chem. B 2005, 109, 6103−6110. (11) Zhang, Y.; Maginn, E. J. Molecular Dynamics Study of the Effect of Alkyl Chain Length on Melting Points of [CnMIM][PF6] Ionic Liquids. Phys. Chem. Chem. Phys. 2014, 16, 13489−13499. 14943
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B (31) Martinez, L.; Andrade, R.; Martinez, J. M. Software News and Update Packmol: A Package for Building Initial Configurations for Molecular Dynamics Simulations. J. Comput. Chem. 2009, 30, 2157− 2164. (32) Zhang, Y.; Otani, A.; Maginn, E. J. Reliable Viscosity Calculation from Equilibrium Molecular Dynamics Simulations: A Time Decomposition Method. J. Chem. Theory Comput. 2015, 11, 3537− 3546. (33) Rey-Castro, C.; Vega, L. F. Transport Properties of Ionic Liquid 1-Ethyl-3-methylimidazolium Chloride from Equilibrium Molecular Dynamics Simulation. The Effect of Temperature. J. Phys. Chem. B 2006, 110, 14426−14435. (34) Evans, D.; Morriss, G. Statistical Mechanics of Nonequilibrium Liquids; Cambridge University Press: Cambridge, 2008. (35) Khare, R.; de Pablo, J.; Yethiraj, A. Rheological, Thermodynamic, and Structural Studies of Linear and Branched Alkanes under Shear. J. Chem. Phys. 1997, 107, 6956−6964. (36) Yan, T.; Burnham, C.; Del Popolo, M.; Voth, G. Molecular Dynamics Simulation of Ionic Liquids: The Effect of Electronic Polarizability. J. Phys. Chem. B 2004, 108, 11877−11881. (37) Borodin, O. Polarizable Force Field Development and Molecular Dynamics Simulations of Ionic Liquids. J. Phys. Chem. B 2009, 113, 11463−11478. (38) Schröder, C. Comparing Reduced Partial Charge Models with Polariable Simulations of Ionic Liquids. Phys. Chem. Chem. Phys. 2012, 14, 3089. (39) Zhao, W.; Leroy, F.; Heggen, B.; Zahn, S.; Kirchner, B.; Balasubramanian, S.; Muller-Plathe, F. Are There Stable Ion-pairs in Room-temperature Ionic Liquids? Molecular Dynamics Simulations of 1-n-butyl-3-methylimidazolium Hexafluorophosphate. J. Am. Chem. Soc. 2009, 131, 15825−15833. (40) Kohagen, M.; Brehm, M.; Thar, J.; Zhao, W.; Muller-Plathe, F.; Kirchner, B. Performance of Quantum Chemically Derived Charges and Persistence of Ion Cages in Ionic Liquids. A Molecular Dynamics Simulations Study of 1-n-butyl-3-methylimidazolium Bromide. J. Phys. Chem. B 2011, 115, 693−702. (41) E, W.; Ren, W.; Vanden-Eijnden, E. String Method for the Study of Rare Events. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 66, 052301. (42) Maragliano, L.; Fischer, A.; Vanden-Eijnden, E.; Ciccotti, G. String Method in Collective Variables: Minimum Free Energy Paths and Isocommittor Surfaces. J. Chem. Phys. 2006, 125, 024106. (43) Khabaz, F.; Khare, R. Effect of Chain Architecture on the Size, Shape, and Intrinsic Viscosity of Chains in Polymer Solutions: A Molecular Simulation Study. J. Chem. Phys. 2014, 141, 214904. (44) Theodorou, D. N.; Suter, U. Shape of Unperturbed Linear Polymers: Polypropylene. Macromolecules 1985, 18, 1206−1214.
14944
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Molecular Topology and Local Dynamics Govern the Viscosity of Imidazolium-Based Ionic Liquids Yong Zhang,† Lianjie Xue,‡ Fardin Khabaz,§ Rose Doerfler,† Edward L. Quitevis,*,‡ Rajesh Khare,*,§ and Edward J. Maginn*,† †
Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, Indiana 46556, United States Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, Texas 79409, United States § Department of Chemical Engineering, Texas Tech University, Lubbock, Texas 79409, United States ‡
ABSTRACT: A series of branched ionic liquids (ILs) based on the 1-(iso-alkyl)-3-methylimidazolium cation from 1-(1methylethyl)-3-methylimidazolium bistriflimide to 1-(5-methylhexyl)-3-methylimidazolium bistriflimide and linear ILs based on the 1-(n-alkyl)-3-methylimidazolium cation from 1propyl-3-methylimidazolium bistriflimide to 1-heptyl-3-methylimidazolum bistriflimide were recently synthesized and their physicochemical properties characterized. For the ILs with the same number of carbons in the alkyl chain, the branched IL was found to have the same density but higher viscosity than the linear one. In addition, the branched IL 1-(2methylpropyl)-3-methylimidazolium bistriflimide ([2mC3C1Im][NTf 2]) was found to have an abnormally high viscosity. Motivated by these experimental observations, the same ILs were studied using molecular dynamics (MD) simulations in the current work. The viscosities of each IL were calculated using the equilibrium MD method at 400 K and the nonequilibrium MD method at 298 K. The results agree with the experimental trend. The ion pair (IP) lifetime, spatial distribution function, and associated potential of mean force, cation size and shape, and interaction energy components were calculated from MD simulations. A quantitative correlation between the liquid structure and the viscosity was observed. Analysis shows that the higher viscosities in the branched ILs are due to the relatively more stable packing between the cations and anions indicated by the lower minima in the potential of mean force (PMF) surface. The abnormal viscosity of [2mC3C1Im][NTf 2] was found to be the result of the specific side chain length and molecular structure.
1. INTRODUCTION Ionic liquids (ILs) are defined as salts that melt at or below 373 K. A typical IL is composed of a bulky organic cation and an inorganic or organic anion. As a group, many ILs tend to have low volatility, reasonable conductivity, and a wide electrochemical window and tend to be miscible with many organic solvents. These special properties make ILs attractive alternatives in synthesis,1 chemical separations,2−4 electrochemistry,5,6 catalysis,7 and solar cells.8 Another unique characteristic of ILs is that their physicochemical properties can be easily tuned by changing the combination or the structure of their component ions. Thus, ILs are also called designer solvents or task-specific solvents.9 Thus, understanding the correlation between molecular structure and macroscopic properties is necessary for the rational design of ILs for certain specific applications, and there has been extensive research carried out toward this objective. Watanabe and co-workers10 systematically studied the correlation between alkyl chain length and physicochemical properties of imidazolium-based ILs, and concluded that the change in the alkyl chain length causes a change in the interaction forces, and the properties of the ILs are determined by the cumulative effect of the electrostatic interaction between © 2015 American Chemical Society
the ionic species and the induction interactions between the ions, aggregates, and clusters. Zhang and Maginn11 studied the melting points of a series of 1-alkyl-3-methylimidazolium hexafluorophosphate ILs ([CnC1Im][PF6]) with n = 2, 4, 10, 12, and 14 using molecular dynamics (MD) simulations. It was found that, for the ILs with long alkyl chains, melting points increase with the alkyl chain length as a result of the increased fusion enthalpies due to the enhanced van der Waals interaction brought about by the long nonpolar alkyl chains. For the ILs with short alkyl chains, fusion entropy plays a dominant role so that [C4C1Im][PF6] has a lower melting temperature than [C2C1Im][PF6]. Entropy also plays an essential role in deciding dynamics in certain ILs. For example, the higher viscosities of the imidazolium ILs with the substitution of a methyl group for a hydrogen at the C2 position of the cation ring relative to the C2 protonated cations are due to the lower entropy in the former.12,13 The effect of cation alkyl chain length and cation symmetry on morphology and physicochemical properties was also Received: August 24, 2015 Revised: October 16, 2015 Published: October 27, 2015 14934
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B Table 1. Structures of the Linear [CnC1Im][NTf 2] and Branched [(n − 2)mCn−1C1Im][NTf 2] ILs
systematically studied by Quitevis,14 Triolo,15,16 and Maroncelli17 and co-workers. Schubert18 and Husson19 studied the effect of branch structure on the physicochemical properties of imidazolium-based ILs. Their results showed that branched ILs have higher viscosities and glass transition temperatures than their linear analogues. Recently, Xue et al.20 synthesized a series of branched imidazolium-based ILs (1-(iso-alkyl)-3-methylimidazolium bistriflimide, [(n − 2)mCn−1C1Im][NTf 2]) (see Table 1). The viscosities of each IL were measured and compared with a series of linear ILs (1-(n-alkyl)-3-methylimidazolium bistriflimide, [CnC1Im][NTf 2]). They found that the viscosity of each branched IL was higher than that of its linear analogue. For both branched and linear ILs, except for [2mC3C1Im][NTf 2], the viscosities increased with increasing alkyl chain length. On the other hand, [2mC3C1Im][NTf 2] showed an abnormally high viscosity, even higher than that of
[3mC4C1Im][NTf 2]. These experimental observations motivated the current simulation study.
2. SIMULATION PROCEDURES Both equilibrium (EMD) and nonequilibrium molecular dynamics (NEMD) methods were used in the current study. A common simulation setup is described below followed by specific simulation details for EMD and NEMD methods, respectively. 2.1. Force Field and Common Simulation Setup. Classical MD simulations were carried out using the package LAMMPS.21 Periodic boundary conditions were applied, and a time step of 1 fs was used. The long-range electrostatic interactions were calculated using the particle−particle particle−mesh (PPPM) method22 with a real space cutoff of 12 Å. The Nosé−Hoover thermostat23 and the extended 14935
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B
every 5 fs for each trajectory. The first 1 ns was considered equilibration and discarded, and the rest of the trajectory was used to calculate viscosity using eq 1. The parameter tcut was determined by taking the time that the calculated standard deviation reached the value of 40% of the estimated viscosity of each system (the value of the flat region of the averaged running integral).32 2.3. NEMD Simulations. To prepare the model structure for NEMD simulations of the IL, initially, 1728 ion pairs were placed in a cubic simulation box at a low density. The density of the system was then increased by subjecting it to an MD simulation at a pressure of 150 atm. The simulation box was subsequently relaxed for a period of 2 ns at temperature T = 298 K and P = 1 atm. Five replicas were built for each IL. The viscosity calculations were performed using the NEMD method34 at constant NVT conditions. The NEMD method has previously been used to compare the viscosity of linear and branched alkanes.35 Six different shear rates in the range 107− 109 s−1 were studied.
Lagrangian approach24 were applied to control the temperature and pressure. The general Amber force field (GAFF)25 was used to describe the interactions in the systems. In order to derive atomic charges, electronic structure calculations were carried out on an isolated ion to optimize the structures at the B3LYP/ 6-311++g(d,p) level using the Gaussian 09 package.26 The atomic charges were then derived on the basis of the optimized structure by fitting the electrostatic potential surface obtained from these calculations using the RESP method.27 Such charges have a total value of ±1 e on the cation and anion, respectively. To include the effect of charge transfer and polarizability in the bulk phase, the partial charges were scaled uniformly by 0.8. The 0.8-scaled charges have been found to be reliable for the study of dynamic properties of similar ILs.28 2.2. EMD Simulations. EMD simulation details closely followed the procedure described in our previous publications.13,29 The simulation box was built up for each IL by putting 300 ion pairs randomly in a cubic box using the package Packmol.30,31 The systems were then equilibrated for 2 ns in the isothermal−isobaric (NPT) ensemble followed by 20 ns production runs in the canonical (NVT) ensemble. The densities were calculated on the basis of the NPT ensemble simulation, and the last 18 ns of the NVT trajectories were used for analysis. The pressure was fixed at 1 atm in all constant pressure simulations with isotropic volume fluctuations. The shear viscosity of each IL was calculated using the recently developed time-decomposition method.32 In this method, the shear viscosity was calculated from the integral over time of the stress tensor autocorrelation function following the Green−Kubo relation η=
V kBT
∫0
3. RESULTS AND DISCUSSION 3.1. Density Is Independent of Chain Structure. The calculated densities of the ILs as a function of temperature are shown in Figure 1. The experimentally measured densities20 are
∞
⟨ταβ(t ) ·ταβ(0)⟩ dt
(1)
where V is the system volume, kB is the Boltzmann constant, and ταβ denotes the αβ-component of the stress tensor. For better statistics, the shear viscosity was calculated by averaging over six independent terms of the stress tensor, namely, τxy, τyz, τzx, 0.5(τxx − τyy), 0.5(τyy − τzz), and 0.5(τxx − τzz).33 Multiple short independent trajectories were generated, and the running integral in eq 1 was calculated for each trajectory. The average of the running integrals was then calculated. The corresponding time-dependent standard deviation of the running integrals was also calculated as σ (t ) =
1 N−1
Figure 1. Liquid phase density as a function of temperature from experiments20 and MD simulations. Calculated results were fit by a linear function and extrapolated to experimental temperatures (shown as dashed lines). The largest deviation in the calculated density from experiments is less than 0.6%. A direct comparison of calculated and experimental densities at 298 K is provided in Table 2.
N
also provided for comparison. In the experimental results, for a given chain length n, the density of the branched IL almost equals that of the linear IL. The largest difference was found for the n = 3 pair, in which the density of the branched IL is slightly higher than that of the linear one by 0.13%. From the figure, we can see that the densities of the branched and linear ILs vary linearly in the temperature range. For faster dynamics and better sampling, MD simulations were carried out at higher temperatures between 350 and 500 K with 50 K intervals. Consistent with experimental results, for a given n, the densities of the branched IL and the linear IL are very close to each other. Except for n = 3, the branched IL has a slightly (160 times the bulk density, which is much higher than that of ∼40 times bulk density observed in [C3C1Im][NTf 2]. This suggests that the preferred distribution of the CT atoms in this region in [2mC3C1Im][NTf 2] will block the approach of the anions from the side of the imidazolium ring and results in the SDF shown in Figure 5. This steric effect reduces anion mobility and is the cause of the abnormally high viscosity in this IL. When the side chains become even longer, for both linear and branched ILs, the CT distribution becomes less localized due to the flexibility of the alkyl chain (results not shown). The sizes of the cations were characterized using the radius of gyration Rg, whereas the asphericity (b) and acylindricity (c) parameters were calculated to characterize the shape of the cations. Recently, these quantities were used to study the relationship between the viscosity and the chain shape and size in dilute polymer solutions.43 These quantities were calculated using the following expressions44 S̲ =
N ∑i = 1 (R̲ i − R̲ com)2
R g 2 = λ1 + λ 2 + λ3
b = λ1 − 0.5(λ 2 + λ3)
c = λ 2 − λ3
(7)
N λ1 > λ 2 > λ3
(8) (9) (10)
where S̲ is the gyration tensor, R̲ i is the position vector of each atom, R̲ com is the center of mass of a cation, N is the total number of atoms per cation, and λi’s are the eigenvalues of S̲ . If a cation has a perfectly spherical shape, then the parameter b will have a value of zero, and if a cation has a perfectly cylindrical shape, then the parameter c will have a value of zero. Figure 10 (top panel) shows the mean squared radius of gyration ⟨Rg2⟩ of cations of linear ILs and branched ILs plotted as a function of carbon number. As can be seen in the figure, the ⟨Rg2⟩ values of branched IL cations are systematically smaller than those of the corresponding linear cations. The extent of asphericity (middle panel in Figure 10) and acylindricity (bottom panel in Figure 10) of the cations were plotted as a function of carbon number. As expected, the degree of asphericity (deviation from the spherical shape) increases monotonically with the number of carbon atoms for both linear and branched ILs. Furthermore, the cations of the branched ILs are more spherical in shape than the cations of the corresponding linear ILs. Due to the relative rigidity of [1mC 2 C 1 Im] + and [2mC 3 C 1 Im] + compared to other branched-chain cations, the differences in b/Rg2 between linear and branched ILs are larger for n = 3 and 4. The value of the acylindricity (deviation from the cylindrical shape) of the cations of the linear ILs and branched ILs, in general, decreases monotonically with the number of carbon atoms; i.e., cation shape becomes more cylindrical as chain length is increased.
Figure 10. Top: Mean squared radius of gyration of cations of linear and branched ILs as a function of carbon number n. Middle: Normalized asphericity of cations for linear and branched ILs as a function of carbon number n. Bottom: Normalized acylindricity of cations for linear and branched ILs as a function of carbon number n.
Two exceptions to this behavior, however, are the acylindricity of the cation of [1mC2C1Im][NTf 2] and [2mC3C1Im][NTf 2]. Due to its specific size and structure (see Figure 8), [2mC3C1Im]+ exhibits an even larger deviation from the cylindrical shape than [1mC2C1Im]+. These results are consistent with the observation presented earlier in this section that the folding back of the methyl groups in [2mC3C1Im][NTf 2] prevents approach of anions to the side of the imidazolium ring, thus, in turn, leading to higher viscosity for this branched IL. 3.6. Viscosity Is Not Sensitive to Charge Localization. To investigate the effect of intermolecular interactions on the liquid phase structure and therefore the dynamics such as viscosity, the total nonbonded energy (i.e., sum of van der Waals and electrostatic energy) per atom of the linear and branched systems was determined and plotted as a function of the number of carbon atoms (Figure 11, top panel). The total 14941
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B
high attractive energy per atom. To determine the origin of this observation, the partial charges on the imidazolium ring of the ILs are also plotted in the same figure (right axis). Except for n = 3, the partial charges in the imidazolium ring are larger in the branched cations compared to those of the linear cations with the same carbon number; i.e., the partial charges are more localized on the imidazolium ring in the branched cations. These graphs clearly show that the electrostatic energy per atom is directly correlated with the magnitude of the localized partial charge on the imidazolium ring. It is also interesting to note that [2mC3C1Im][NTf 2] has the most attractive nonbonded interactions in the ILs studied in the current work, consistent with its most localized partial charges on the imidazolium ring (see Figure 11). At the outset, these results suggest that the origin of the abnormally high viscosity of the [2mC3C1Im][NTf 2] branched IL is related to the higher localized partial charge on the imidazolium ring of this molecule. To further test this effect of charge localization on the molecular interaction and viscosity, a model system was set up in which the partial charges of [C5C1Im]+ were applied to [3mC4C1Im]+. All of the other parameters in the force field for [3mC4C1Im]+ and [NTf 2]− were kept the same. The viscosity of the modified [3mC4C1Im][NTf 2] was computed using both EMD and NEMD methods. Using the EMD method, the viscosity was found to be 8.25 mPa·s at 400 K, which is about halfway between the viscosities of the original [C5C1Im][NTf 2] (7.79 mPa·s) and [3mC4C1Im][NTf 2] (8.75 mPa·s). On the other hand, using NEMD at a temperature of 298 K, the viscosity of the charge modified [3mC4C1Im][NTf 2] was found to be the same as that of the original [3mC4C1Im][NTf 2] molecule within statistical uncertainties. These results suggest that the origin of the abnormally high viscosity of the branched IL with n = 4 is not in the charge localization; the effect is primarily governed by the molecular structure as discussed in the previous section.
4. CONCLUDING REMARKS A series of ILs with a branched alkyl chain ([(n − 2) mCn−1C1Im][NTf 2], n = 3−7) and a linear alkyl chain ([CnC1Im][NTf 2]) were studied using classical MD simulations. The calculated densities showed excellent agreement with the experimental values. The experimental trend in viscosity was also successfully captured using both EMD and NEMD methods. On the basis of MD simulations, the trend in viscosities is connected to the microscopic IP lifetimes, consistent with a previous study.29 The IP lifetimes were found to be linearly correlated with the depth of minima on the PMF surface derived from spatial distribution functions of the ions in the liquid phase. The higher viscosities of the ILs with branched side chains relative to the ones with a linear side chain are likely caused by the relatively more stable packing between the cations and anions in the liquid phase of the branched ILs compared to the linear ILs, as indicated by the lower free energy minima in the branched IL systems. The abnormally high viscosity of [2mC3C1Im][NTf 2] was found to be caused by the specific shape and length of its alkyl side chain, which blocks the approach of anions to the side of the imidazolium ring. Our results suggest a direct quantitative connection between the transport properties of the ILs and their liquid phase structure. It is known that transport properties such as viscosity are relatively hard to calculate and the results usually involve a
Figure 11. Calculated total (sum of van der Waals and electrostatic, top panel), van der Waals (middle), and electrostatic (bottom) energies per atom in linear and branched ILs at 400 K. Partial charges localized on the imidazolium ring of the cations are also provided (bottom panel, right axis).
nonbonded energy/atom of the linear ILs shows a weak increase with an increase in the number of carbon atoms. On the other hand, the nonbonded energy/atom exhibits a much stronger variation for the branched ILs. It is interesting to note that, except for n = 3, the total nonbonded energy/atom is lower in the branched IL compared to that of the linear analogue with the same carbon number and, furthermore, the cation size dependence is not monotonic. To further investigate this, the two energy components, van der Waals and electrostatic interactions, were studied separately. As shown in Figure 11 (middle panel), the van der Waals energy values per atom for the linear and branched ILs with the same carbon numbers are very close to each other. For both linear and branched ILs, a monotonic increase is observed with an increase in carbon number. As seen in the bottom panel of Figure 11, the origin of the nonmonotonic behavior of the total nonbonded energy is in the electrostatic energy contribution per atom. Furthermore, the branched IL with n = 4 (the one that shows an abnormally high viscosity) exhibits an abnormally 14942
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B large uncertainty,32 whereas it is relatively easy to get accurate results of structural properties such as radial distribution functions (RDFs) and SDFs. The quantitative correlation between these properties demonstrated here suggests that the liquid structural properties can be used to predict transport properties of ILs, which, in turn, could significantly speed up the design of new ILs for a given application.
■
(12) Hunt, P. A. Why Does a Reduction in Hydrogen Bonding Lead to an Increase in Viscosity for the 1-Butyl-2,3-dimethyl-imidazoliumbased Ionic Liquids? J. Phys. Chem. B 2007, 111, 4844−4853. (13) Zhang, Y.; Maginn, E. J. The Effect of C2 Substitution on Melting Point and Liquid Phase Dynamics of Imidazolium Based-ionic Liquids: Insights from Molecular Dynamics Simulations. Phys. Chem. Chem. Phys. 2012, 14, 12157−12164. (14) Zheng, W.; Mohammed, A.; Hines, L. G., Jr.; Xiao, D.; Martinez, O. J.; Bartsch, R. A.; Simon, S. L.; Russina, O.; Triolo, A.; Quitevis, E. L. Effect of Cation Symmetry on the Morphology and Physicochemical Properties of Imidazolium Ionic Liquids. J. Phys. Chem. B 2011, 115, 6572−6584. (15) Russina, O.; Triolo, A.; Gontrani, L.; Caminiti, R.; Xiao, D.; Hines, L. G., Jr.; Bartsch, R. A.; Quitevis, E. L.; Plechkova, N.; Seddon, K. R. Morphology and Intermolecular Dynamics of 1-Alkyl-3methylimidazolium Bis((trifluoromethane)sulfonyl)amide Ionic Liquids: Structural and Dynamic Evidence of Nanoscale Segregation. J. Phys.: Condens. Matter 2009, 21, 424121. (16) Rocha, M. A. A.; Neves, C. M. S. S.; Greire, M.; Russina, O.; Triolo, A.; Coutinho, J. A. P.; Santos, L. M. N. B. F. Alkylimidazoilium Based Ionic Liquids: Impact of Cation Symmetry on Their Nanoscale Structural Organization. J. Phys. Chem. B 2013, 117, 10889−10897. (17) Jin, H.; O’Hare, B.; Dong, J.; Arzhantsev, S.; Baker, G. A.; Wishart, J. F.; Benesi, A. J.; Maroncelli, M. Physical Properties of Ionic Liquids Consisting of the 1-Butyl-3-Methylimidazolium Cation with Various Anions and the Bis(trifluoromethylsulfonyl)imide Anion with Various Cations. J. Phys. Chem. B 2008, 112, 81−92. (18) Erdmenger, T.; Vitz, J.; Wiesbrock, F.; Schubert, U. S. Influence of Different Branched Alkyl Side Chains on the Properties of Imidazolium-based Ionic Liquids. J. Mater. Chem. 2008, 18, 5267− 5273. (19) Andresova, A.; Storch, J.; Traikia, M.; Wagner, Z.; Bendova, M.; Husson, P. Branched and Cyclic Alkyl Groups in Imidazolium-based Ionic Liquids: Molecular Organization and Physico-Chemical Properties. Fluid Phase Equilib. 2014, 371, 41−49. (20) Xue, L.; Tamas, G.; Koh, Y. P.; Shadeck, M.; Gurung, E.; Simon, S. L.; Maroncelli, M.; Quitevis, E. L. Effect of Alkyl Branching on Physicochemical Properties of Imidazolium-based Ionic Liquids. J. Chem. Eng. Data, submitted for publication, 2015. (21) Plimpton, S. Fast Parallel Algorithms for Short-range Molecular Dynamics. J. Comput. Phys. 1995, 117, 1−19. (22) Hockney, R.; Eastwood, J. Computer Simulation Using Particles; Adam Hilger: New York, 1989. (23) Hoover, W. G. Canonical Dynamics: Equilibrium Phase-space Distributions. Phys. Rev. A: At., Mol., Opt. Phys. 1985, 31, 1695−1697. (24) Shinoda, W.; Shiga, M.; Mikami, M. Rapid Estimation of Elastic Constants by Molecular Dynamics Simulation under Constant Stress. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 69, 134103. (25) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004, 25, 1157−1174. (26) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, revision A.1; Gaussian, Inc.: Wallingford, CT, 2009. (27) Bayly, C.; Cieplak, P.; Cornell, W. D.; Kollman, P. A. A Wellbehaved Electrostatic Potential Based Method Using Charge Restraints for Deriving Atomic Charges: the RESP Model. J. Phys. Chem. 1993, 97, 10269−10280. (28) Zhang, Y.; Maginn, E. J. A Simple AIMD Approach to Derive Atomic Charges for Condensed Phase Simulation of Ionic Liquids. J. Phys. Chem. B 2012, 116, 10036−10048. (29) Zhang, Y.; Maginn, E. J. Direct Correlation Between Ionic Liquid Transport Properties and Ion Pair Lifetimes: A Molecular Dynamics Study. J. Phys. Chem. Lett. 2015, 6, 700−705. (30) Martinez, J. M.; Martinez, L. Packing Optimization for Automated Generation of Complex System’s Initial Configurations for Molecular Dynamics and Docking. J. Comput. Chem. 2003, 24, 819−825.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Y.Z. and E.J.M. are supported by the U.S. Department of Energy, Basic Energy Science, Joint Center for Energy Storage Research under Contract No. DE-AC02-06CH11357. Computational resources were provided by the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231, and the Center for Research Computing (CRC) at the University of Notre Dame. L.X. and E.L.Q. are supported by the National Science Foundation (NSF) under Grant No. CHE-1153077.
■
REFERENCES
(1) Iranpoor, N.; Firouzabadi, H.; Azadi, R. A New Diphenylphosphinite Ionic Liquid (IL-OPPh2) as Reagent and Solvent for Highly Selective Bromination, Thiocyanation or Isothiocyanation of Alcohols and Trimethylsilyl and Tetrahydropyranyl Ethers. Tetrahedron Lett. 2006, 47, 5531−5534. (2) Tian, K.; Qi, S.; Cheng, Y.; Chen, X.; Hu, Z. Separation and Determination of Lignans from Seeds of Schisandra Species by Micellar Electrokinetic Capillary Chromatography. J. Chromatogr. A 2005, 1078, 181−187. (3) Llovell, F.; Oliveira, M. B.; Coutinho, J. A. P.; Vega, L. F. Solubility of Greenhouse and Acid Gases on the [C4mim][MeSO4] Ionic Liquid for Gas Separation and CO2 Conversion. Catal. Today 2015, 255, 87−96. (4) Zhang, Z.; Hu, A.; Zhang, T.; Zhang, Q.; Sun, M.; Sun, D. Separation of Methyl Acetate + Methanol Azeotropic Mixture Using Ionic Liquids as Entrainers. Fluid Phase Equilib. 2015, 401, 1−8. (5) Hapiot, P.; Lagrost, C. Electrochemical Reactivity in Roomtemperature Ionic Liquids. Chem. Rev. 2008, 108, 2238−2264. (6) MacFarlane, D. R.; Forsyth, M.; Howlett, P. C.; Pringle, J. M.; Sun, J.; Annat, G.; Neil, W.; Izgorodina, E. I. Ionic Liquids in Electrochemical Devices and Processes: Managing Interfacial Electrochemistry. Acc. Chem. Res. 2007, 40, 1165−1173. (7) Ranu, B. C.; Banerjee, S.; Jana, R. Ionic Liquid as Catalyst and Solvent: The Remarkable Effect of a Basic Ionic Liquid, [Bmim]OH on Michael Addition and Alkylation of Active Methylene Compounds. Tetrahedron 2007, 63, 776−782. (8) Tsuda, T.; Hussey, C. L. Electrochemical Applications of Roomtemperature Ionic Liquids. Electrochem. Soc. Interface 2007, 42−49. (9) Earle, M. J.; Seddon, K. R. Ionic Liquids. Green Solvents for the Future. Pure Appl. Chem. 2000, 72, 1391−1398. (10) Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. Physicochemical Properties and Structures of Room Temperature Ionic Liquids. 2. Variation of Alkyl Chain Length in Imidazolium Cation. J. Phys. Chem. B 2005, 109, 6103−6110. (11) Zhang, Y.; Maginn, E. J. Molecular Dynamics Study of the Effect of Alkyl Chain Length on Melting Points of [CnMIM][PF6] Ionic Liquids. Phys. Chem. Chem. Phys. 2014, 16, 13489−13499. 14943
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
Article
The Journal of Physical Chemistry B (31) Martinez, L.; Andrade, R.; Martinez, J. M. Software News and Update Packmol: A Package for Building Initial Configurations for Molecular Dynamics Simulations. J. Comput. Chem. 2009, 30, 2157− 2164. (32) Zhang, Y.; Otani, A.; Maginn, E. J. Reliable Viscosity Calculation from Equilibrium Molecular Dynamics Simulations: A Time Decomposition Method. J. Chem. Theory Comput. 2015, 11, 3537− 3546. (33) Rey-Castro, C.; Vega, L. F. Transport Properties of Ionic Liquid 1-Ethyl-3-methylimidazolium Chloride from Equilibrium Molecular Dynamics Simulation. The Effect of Temperature. J. Phys. Chem. B 2006, 110, 14426−14435. (34) Evans, D.; Morriss, G. Statistical Mechanics of Nonequilibrium Liquids; Cambridge University Press: Cambridge, 2008. (35) Khare, R.; de Pablo, J.; Yethiraj, A. Rheological, Thermodynamic, and Structural Studies of Linear and Branched Alkanes under Shear. J. Chem. Phys. 1997, 107, 6956−6964. (36) Yan, T.; Burnham, C.; Del Popolo, M.; Voth, G. Molecular Dynamics Simulation of Ionic Liquids: The Effect of Electronic Polarizability. J. Phys. Chem. B 2004, 108, 11877−11881. (37) Borodin, O. Polarizable Force Field Development and Molecular Dynamics Simulations of Ionic Liquids. J. Phys. Chem. B 2009, 113, 11463−11478. (38) Schröder, C. Comparing Reduced Partial Charge Models with Polariable Simulations of Ionic Liquids. Phys. Chem. Chem. Phys. 2012, 14, 3089. (39) Zhao, W.; Leroy, F.; Heggen, B.; Zahn, S.; Kirchner, B.; Balasubramanian, S.; Muller-Plathe, F. Are There Stable Ion-pairs in Room-temperature Ionic Liquids? Molecular Dynamics Simulations of 1-n-butyl-3-methylimidazolium Hexafluorophosphate. J. Am. Chem. Soc. 2009, 131, 15825−15833. (40) Kohagen, M.; Brehm, M.; Thar, J.; Zhao, W.; Muller-Plathe, F.; Kirchner, B. Performance of Quantum Chemically Derived Charges and Persistence of Ion Cages in Ionic Liquids. A Molecular Dynamics Simulations Study of 1-n-butyl-3-methylimidazolium Bromide. J. Phys. Chem. B 2011, 115, 693−702. (41) E, W.; Ren, W.; Vanden-Eijnden, E. String Method for the Study of Rare Events. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 66, 052301. (42) Maragliano, L.; Fischer, A.; Vanden-Eijnden, E.; Ciccotti, G. String Method in Collective Variables: Minimum Free Energy Paths and Isocommittor Surfaces. J. Chem. Phys. 2006, 125, 024106. (43) Khabaz, F.; Khare, R. Effect of Chain Architecture on the Size, Shape, and Intrinsic Viscosity of Chains in Polymer Solutions: A Molecular Simulation Study. J. Chem. Phys. 2014, 141, 214904. (44) Theodorou, D. N.; Suter, U. Shape of Unperturbed Linear Polymers: Polypropylene. Macromolecules 1985, 18, 1206−1214.
14944
DOI: 10.1021/acs.jpcb.5b08245 J. Phys. Chem. B 2015, 119, 14934−14944
