Ionic Vapor Composition in Pyridinium-Based Ionic LiquidsVitaly V. Chaban† and Oleg V. Prezhdo*,‡

†Instituto de Ciencia e Tecnologia, Universidade Federal de Sao Paulo, 12231-280, Sao Jose dos Campos, SP, Brazil‡Department of Chemistry, University of Southern California, Los Angeles, California 90089, United States

ABSTRACT: Strong electrostatic interactions in ionic compounds makevaporization a complex process. The gas phase can contain a broad range ofionic clusters, and the cluster composition can differ greatly from that inthe liquid phase. Room-temperature ionic liquids (RTILs) constitute acomplicated case due to their ionic nature, asymmetric structure, and ahuge versatility of ions and ionic clusters. This work reports vapor−liquidequilibria and vapor compositions of butylpyridinium (BPY) RTILs formedwith hexafluorophosphate (PF6), trifluoromethanesulfonate (TF), andbis(trifluoromethanesulfonyl)imide (TFSI) anions. Unlike inorganiccrystals, the pyridinium-based RTILs contain significant percentages ofcharged clusters in the vapor phase. Ion triplets and ion quadruplets each constitute up to 10% of the vapor phase composition.Triples prevail over quadruples in [BPY][PF6] due to the size difference of the cation and the anion. The percentage of chargedionic clusters in the gas phase is in inverse proportion to the mass of the anion. The largest identified vaporized ionic clustercomprises eight ions, with a formation probability below 1%. Higher temperature fosters formation of larger clusters due to anincrease of the saturated vapor density.

■ INTRODUCTION

Room-temperature ionic liquids (RTILs) are salts with lowtemperature melting points. They are composed entirely ofions, and many of them are liquid at room conditions.1−8 In thebeginning of the 21st century, RTILs rapidly became the mostactively investigated group of solvents ever. Nowadays, RTILsare actively explored in physics, chemistry, and molecularbiology as universal solvents, reaction media, and electrolytesolutions, for separation applications, and in chemical synthesis.Low volatility and flammability, wide liquid range, amphiphi-licity, chemical stability (in most known cases), tunability, andionic conductivity are among the practically importantcharacteristics of the pursued RTILs.The alkylpyridinium-based RTILs are of great value for the

field,9−16 since precisely these compounds launched the newera of ionic liquids when aluminum(III) chloride was mixedwith N-alkylpyridinium.17 Representatives of this establishedclass of RTILs have already been successfully applied in asignificant number of processes. Examples include, while notlimited to, conductive liquids in electrochemistry, photo-induced electron transfer,18−21 organocatalysis for well-knownreactions,22,23 and desulfurization of fuels.24−27 Furthermore,pyridinium-based RTILs exhibit an outstanding thermalstability,28 which depends primarily on the chain length ofthe cation and the nature of the anion. This property appearspractically important, because certain chemical reactionsrequire elevated temperatures for initiation.Using a thermogravimetric analyzer, Bittner and co-work-

ers28 reported physical chemical properties and thermalstabilities for a number of pyridinium-based RTILs composedof the pyridinium cations with different substituents and thefollowing anions: bis(trifluoromethanesulfonyl)imide [TFSI],

trifluoromethanesulfonate [TF], tetrafluoroborate [BF4], andtris(pentafluoroethyl)trifluorophosphate [FAP]. The authorssystematically compared the ethylpyridinium [EPY], butylpyr-idinium [BPY], and hexylpyridinium [HPY] cations. An effectof the grafted methyl group in the “2” [B2MPY], “3” [B3MPY],or “4” [B4MPY] positions of the pyridine ring was alsocarefully investigated. It was found that all RTILs maintain aperfect stability up to 500 K, whereas a complete thermaldecomposition takes place above 900 K. A longer aliphaticchain results in a higher decomposition temperature, defined asthe temperature of the maximum decomposition rate: 706, 724,and 784 K for [EPY], [BPY], and [HPY], respectively. Theanion determines the thermal decomposition mechanism andthe corresponding temperature. Compare, [B3MPY][TF] and[B3MPY][TFSI] are totally decomposed above 850 K. In turn,the decomposition temperature of both [B3MPY][FAP] and[B3MPY][BF4] exceeds 1000 K. While critical points for thepyridinium-based RTILs were not identified, it is clear from oursimulations that they are significantly higher than thedecomposition temperatures. Thus, critical temperatures ofthese RTILs are outside of experimental reach.While earlier RTILs were considered to be nonvolatile

liquids, now it is recognized that distillation of certain RTILswithout degradation is indeed possible.29,30 At relatively lowtemperatures, RTILs exhibit extremely low but still nonzerovapor pressures,31 which increase continuously upon furtherheating. The recent findings highlight an importance ofunderstanding the vapor−liquid equilibria (VLE) of these

Received: March 27, 2016Revised: May 3, 2016Published: May 10, 2016

Article

pubs.acs.org/JPCB

© 2016 American Chemical Society 4661 DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

Dow

nloa

ded

via

UN

IV O

F SO

UT

HE

RN

CA

LIF

OR

NIA

on

Nov

embe

r 7,

201

9 at

23:

02:0

4 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

compounds. VLE provide valuable information, from afundamental standpoint, regarding the nature of interionicinteractions simultaneously in the liquid phase and the vaporphase. The VLE data are necessary to assess the potentialperformance of different RTILs as lubricants, coatings, andseparating agents. As discussed above, the stability of thepyridinium-based RTILs allows one to measure their saturatedvapor pressure, prior to decomposition, and investigateexperimentally vapor composition.The aggregation state of ions in the vapor phase of RTIL is

still debated actively.30,32−35 Much of the uncertainty arisesfrom experimental hurdles, since the work has to be performedat high temperatures and low vapor pressures of RTILs. Traceimpurities can introduce significant biases to the results.Computer simulations,36−45 such as molecular dynamics(MD), allow one to avoid the enumerated technical problemsand provide an atomically precise description of the vaporphase. However, reliable interaction potentials are required,which have to work equally well over the entire temperaturerange of interest.An insightful simulation study was published by Chen and

co-workers.46 Vaporization enthalpy and vapor-phase specieswere obtained from MD simulations and ab initio calculationsof the 1,1,3,3-tetramethylguanidinium-based RTILs. Accordingto the authors, both electrostatic and van der Waals interionicinteractions contribute critically to the vaporization behavior.Additionally, changes of ionic configurations upon vaporizationalso appear important and should be accounted for.Chaban and Prezhdo40 reported MD simulations to observe

ionic vapor in real time. They concluded that, while bothneutral and charged species were present in the vapor phase of1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonate)-imide [EMIM][TFSI] and N-butylpyridinium tetrafluoroborate[BPY][BF4], the neutral ion pair dominated in both RTILs.Accurate MD simulations of 1-butyl-3-methylimidazoliumhexafluorophosphate with an emphasis on critical parameterswere contributed by Weiss47 very recently.Mass-spectrometry measurements and density functional

theory calculations of Eberlin and co-workers48 agree with theabove-mentioned theoretical results, suggesting that the vaporphase of [EMIM][BF4] contains not only a single ion pair butalso dimers and trimers. Additional interesting investigations ofthe ionic vapor species employ analytical spectrum meth-ods.49,50

The present work characterizes, for the first time, the vapor−liquid interfaces for the three pyridinium-based ILs, [BPY]-[PF6], [BPY][TF], and [BPY][TFSI], at relatively hightemperatures, 700−850 K, using atomically precise classicalMD simulations. The interfaces are explicitly implemented inthe simulated MD cells (Figure 1), and the exchange ofmolecules between the liquid phase and the vapor phase isobserved in real time. The RTILs chosen for this studyrepresent different anion structures, responsible for varyingphysicochemical properties of RTILs.

■ METHODOLOGYClassical MD simulations in the constant temperature, constantvolume ensemble were used to derive molecular trajectoriescorresponding to VLEs of [BPY][PF6], [BPY][TF], and[BPY][TFSI]. The constant temperature was maintained bythe velocity rescaling thermostat (relaxation constant 200 fs),which generates a correct velocity distribution for a simulatedensemble of particles.51 The integration time-step was set to

1.0−2.0 fs depending on the simulated temperature. Accord-ingly, the equations of motion were propagated for 50−100 ns.The first 5 ns of each MD simulation were regarded asequilibration and excluded from subsequent statistical analysis.Atomic coordinates were saved every 5 ps. Energy-relatedcomponents were recorded every 2 ps.The Coulomb interactions were computed directly if the

separation between the interaction centers did not exceed 1.2nm and by the reaction-field-zero scheme for largerseparations.52 The Lennard-Jones (12,6) interactions weremodified to smoothly approach zero in the range 1.1−1.2 nm.The long-range dispersion corrections were used to computeenergy and pressure during MD. The list of neighbors for agiven particle was updated every 10 time-steps, irrespective ofthe chosen integration time-step, within the radius of 1.5 nm.The quickly oscillating carbon−hydrogen covalent bondlengths were constrained by the LINCS algorithm.53 Theparticle decomposition scheme was used to organize parallelcomputations. Four to 16 cores were used per MD systemdepending on the available resources.The force field (FF), developed and thoroughly validated

versus experimental data previously, was used to represent thesimulated pyridinium-based RTILs.54,55 The employed FF isknown to provide highly accurate dynamic properties at roomand elevated temperatures. One can argue that realistictransport properties correspond to a realistic phase behavior,since both phenomena require proper balance betweenpotential and kinetic energies. Not only potential energyminima but also energy barriers should be reproducedcorrectly. In comparison, a structural characterization requireslocations of the minima, while energy barriers remain largelyirrelevant. Even quantitative differences between the minimaare not particularly important, as long as the energy order ispreserved. Therefore, a model producing accurate structuralproperties may err for transport properties, while a model with

Figure 1. MD configuration of the [BPY][PF6] system obtained afterequilibration at 800 K. Cations are red, and anions are blue. Allfrequent ionic formations (single ion, neutral pair, charged triplet,neutral quadruplet) existing in the saturated vapor can be observed inthis MD snapshot. Noteworthy, an ion triplet consists of one cation(BPY) and two anions (PF6), and therefore, it is negatively charged.The preference is rationalized by the size difference between the ions.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4662

accurate transport properties can be expected to predict thecorrect phase behavior.Further, it was recently shown that electronic polarization

only insignificantly decays with temperature. The models ofRTILs derived for room-temperature simulations are, accord-ingly, suitable for higher-temperature simulations. Each ioniccomposition was simulated at 700, 750, 800, and 850 K (Table1). These temperatures were identified through the preliminaryannealing simulations, in which the vapor−liquid interfaceswere heated from 400 to 1000 K with a speed of 0.01 K ps−1,and the vapor pressure was computed on the fly. The box size(Table 1) was chosen to allow for a sufficiently larger numberof vaporized ions to properly sample different ionic clusters inthe vapor phase. The vapor density is lower at smaller T.Consequently, the volume allocated to the vapor was increased.Gromacs 4 and supplementary utilities52,56,57 of this program

package were used to generate and analyze the atomistictrajectories. VMD,58 Gabedit,59 and Packmol60 were used tobuild and visualize the MD systems.

■ RESULTS AND DISCUSSION

MD simulations allow one to observe all properties of thevapor−liquid interface in its equilibrium state. Three conditionsshould be rigorously met. First, the simulated temperatureshould be high enough to generate an observable vaporpressure. Second, the MD cell should be sufficiently large toaccommodate a significant number of vapor particles forstatistics. Ideally, the ratio of the liquid phase volume to thevapor phase volume should be commensurate with the ratio ofthe liquid phase density to the vapor phase density. Third, afteran equilibrium is established, the number of particles in theliquid phase should not be smaller than the thermodynamiclimit, and the linear size of the liquid phase should be at leasttwice larger than the interaction energy cutoffs. Otherwise, thesimulated liquid would likely be unstable, and the correspond-ing saturated vapor pressure might be assessed incorrectly.Important methodological considerations were provided by Liuand co-workers while simulating decomposition of nitro-methane.61

Preliminary MD calculations can be used to address theenumerated concerns and to identify suitable methodologicalconditions, particularly temperature and MD box size. We

simulated continuous heating of [BPY][PF6], [BPY][TF], and[BPY][TFSI] from 400 to 1000 K with a temperature elevationrate of 0.01 K ps−1. Upon heating, ions of the liquid phasespontaneously joined the vapor, creating a certain vaporpressure. As long as the liquid phase and the vapor phasecoexist, the recorded pressure in the MD system is denoted as asaturated vapor pressure.It was found that vapor pressures of the considered RTILs

are of the same order of magnitude, with [BPY][PF6] beingsystematically less volatile (Figure 2). This observation is in linewith a stronger cation−anion coupling in this RTIL revealedpreviously.54 At 850 K, the vapor pressure of [BPY][PF6]reaches 1 bar, which is sufficient to investigate the compositionof the vapor phase of the RTIL. The vapor pressures generatedby [BPY][TF] and [BPY][TFSI] are several times larger, andthus, the sampling of the corresponding phase spaces wasbetter. A greater number of evaporated ions both improves thesampling rate and allows a larger diversity of species. Therefore,less simulation time is required to properly sample the vaporphase. Logarithmic dependence of vapor pressure on temper-ature reveals that the pressure calculations were sufficientlyaccurate in all cases. We chose 700, 750, 800, and 850 K for adetailed investigation of the vapor phase composition. Althoughincreasing the simulated temperatures to a certain extent,especially for [BPY][PF6], would further improve the sampling,higher temperatures may extend beyond the decompositiontemperature that is undesirable for possible comparisons withexperiments in the future.To identify an optimal MD cell size for every simulated

system, the vapor density profiles were computed (Figure 3).The ionic vapor is uniformly distributed around the liquidphase throughout the MD cell. Since [BPY][PF6] is the leastvolatile liquid, the corresponding box side was set to 45 nm,and box volume to 91 125 nm3, for enhancement of sampling.The volatilities of [BPY][TF] and [BPY][TFSI] are higher(hence, the vapor density is higher); consequently, the box sizeswere set to smaller values, 30 nm. Note that the vapor volumeshould not be too large to prevent complete evaporation of theliquid phase. The observed volatility of [BPY][TF] at 700 K issomewhat larger than that of [BPY][TFSI], due to a smallerand lighter anion. Table 1 lists cell sizes, which were selectedfor production runs.

Table 1. Simulated Systems, Their Compositions, and Basic Simulation Detailsa

# RTIL # ions # atoms time-step (fs) trajectory length (ns) temperature (K) MD box side (nm)

1 [BPY][PF6] 680 21 080 2.0 100 700 452 [BPY][PF6] 680 21 080 1.5 100 750 453 [BPY][PF6] 680 21 080 1.0 70 800 204 [BPY][PF6] 680 21 080 1.0 60 850 205 [BPY][TF] 648 20 736 2.0 80 700 306 [BPY][TF] 648 20 736 1.5 70 750 257 [BPY][TF] 648 20 736 1.0 60 800 208 [BPY][TF] 648 20 736 1.0 50 850 209 [BPY][TFSI] 540 21 060 2.0 80 700 3010 [BPY][TFSI] 540 21 060 1.5 60 750 2511 [BPY][TFSI] 540 21 060 1.0 60 800 2012 [BPY][TFSI] 540 21 060 1.0 50 850 20

aThe simulated temperatures were chosen on the basis of the preliminary annealing simulations, as discussed below. The MD systems weresimulated until the difference between the time-block averaged vapor compositions became acceptably small. Therefore, the simulated trajectorylengths are unique for each MD system. The time-step was decreased as the temperature was increased to provide a better total energy conservation.The sides of the cubic MD cells were longer at smaller temperatures to allow for more particles to join the vapor phase, thus improving statistics. Inparticular, the statistics is improved due to an ability of many ions to form a variety of possible ionic structures faster.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4663

The largest fraction of ions constituting vapor of each RTILexists as ion pairs. This is a rather expected phenomenon in allsystems at all temperatures (Figures 4−6), since the electro-static cation−anion interactions are strong. 10−15% of single

ions are observed in all RTIL vapors. The percentage is higherin [BPY][PF6] and lower in [BPY][TFSI]. Since the [PF6]

−

anion is spherical and relatively light, Mr = 145.0 amu, its speedis higher in the vapor phase. Thus, the percentage of single ionsis also higher in this RTIL. Vice versa, the [TFSI]− anion is thebulkiest and heaviest particle, Mr = 280.1 amu. As a result, ionpairs dominate significantly over freely moving ions in[BPY][TFSI]. Interestingly, ion triplets are somewhat moreabundant in [BPY][PF6], 10−15%, than neutral quadruplets.This happens likely due to the different sizes of [BPY]+ and[PF6]

−, leading to the geometry in which the first anion getsattached above the pyridine ring and the second anion getsattached below the ring. The [BPY][PF6]2

− structures can beclearly seen in the VLE system snapshot (Figure 1). In turn,quadruplets are more stable in [BPY][TF] and [BPY][TFSI].Temperature variation, 700−850 K, does not qualitatively

alter percentages of ion clusters, although it slightly affectsprobabilities of larger clusters. Higher temperatures result inhigher vapor densities. Therefore, it is reasonable to expectformation of larger ionic clusters. These events are infrequent,and the larger clusters are short-lived. It would be interesting to

Figure 2. Pressure generated by the vapor phases of [BPY][PF6](circles), [BPY][TF] (squares), and [BPY][TFSI] (triangles) as afunction of temperature. Derived initially from the preliminaryannealing simulations, the depicted pressures were then averagedover time for each simulated RTIL and temperature based on theequilibrium MD simulations (12 statistically independent points). Thediscrepancy of the pressures obtained from the annealing simulationsand from the equilibrium (production) simulations was insignificantdue to a sufficiently slow heating rate and inherent pressurefluctuations.

Figure 3. Vapor density of the pyridinium-based RTILs: [BPY][PF6](circles), [BPY][TF] (squares), and [BPY][TFSI] (triangles), derivedfrom the 10 ns long MD simulations at 700 K.

Figure 4. Probability of formation of different ionic clusters in thevapor phase of [BPY][PF6] at 700, 750, 800, and 850 K. The samescale was used in all tabs for easier comparison.

Figure 5. Probability of formation of different ionic clusters in thevapor phase of [BPY][TF] at 700, 750, 800, and 850 K. The samescale was used in all tabs for easier comparison.

Figure 6. Probability of formation of different ionic clusters in thevapor phase of [BPY][TFSI] at 700, 750, 800, and 850 K. The samescale was used in all tabs for easier comparison.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4664

attempt observing such large clusters experimentally, forinstance, employing high-precision variations of mass spec-trometry. Interestingly, the largest observed cluster is neutraland comprises eight ions, [BPY]4[anion]4. It is observed only athigher temperatures and is more probable with the [TF]− and[TFSI]− anions than with the spherical [PF6]

− anion. At 700 K,the probability of [BPY]4[anion]4 is either zero ([BPY][PF6]and [BPY][TFSI]) or marginal ([BPY][TF]).The fraction of ion pairs (Figure 7) in the ionic vapors clearly

decreases with increasing temperature in all RTILs. The effect

is strongest in [BPY][TFSI]. As mentioned above, this leads toemergence of large clusters, whereas the fraction of single ionsis altered marginally in the considered temperature range.Although ion pairs dominate in all RTILs and at alltemperatures, the portion of other ionic compositions issignificant. The quantity of independent vapor particlesdetermines vapor pressure.Neutral ionic clusters are more thermodynamically favorable

than charged ones below a critical temperature. Destruction ofan ion pair or an aggregate of multiple ion pairs requires asignificant additional energy to overcome the cation−anionelectrostatic attraction. Higher temperatures favor chargedclusters (Figure 8), both positive and negative. In [BPY][TFSI]

at 700 K, only 12% of all evaporated ions form charged clusters,whereas their number reaches 30% at 850 K. The majority ofthe charged clusters are single ions and ion triples (Figures4−6). The number of charged clusters is in direct proportion tothe mobility of the anions: [TFSI]− > [TF]− > [PF6]

−.The equilibration process of the [BPY][TF] vapor−liquid

interface is depicted in Figure 9. The process is qualitatively thesame for all investigated RTILs. At time 0, all ions are in the

liquid phase. Once vacuum is added to the MD cell, surfaceions evaporate very quickly, within 100 ps. The maximumpopulation of the vapor phase is observed at 150 ps, 20% of allions, after which certain condensation takes place. The vaporphase stabilizes with 10% of ions. These processes take lessthan 500 ps. Therefore, after the first 500 ps of MD simulation,a vapor composition can already be analyzed.

■ CONCLUSIONS

The VLEs of three pure pyridinium-based ionic liquids wereinvestigated by classical MD simulations at 700−850 K. Thecompositions and relative abundances of the ionic clustersexisting in the vapor phase were identified and correlated to thecation−anion interaction energy. It was found that mostevaporated ions exist as ion pairs (more than 40% in anyRTIL). The vapor phase of [BPY][TFSI] at 700 K containsmost ion pairs, 71%. This percentage decreases down to 44%when temperature increases up to 850 K. Due to the vapordensity increase at higher temperatures, ion pairs tend to fuse,giving rise to larger ionic aggregates, up to eight ions. However,the fraction of such large aggregates is very small, usually below1%. It is yet unclear whether they can be detected by modernexperimental techniques. Studies of the lifetimes of the six toeight ion aggregates in all RTILs may constitute an interestingextension of our present work.Interestingly, the amount of ion triplets and neutral ion

quadruplets exceeds the amount of single ions in most systems.Thus, single ions in the vapor phaseboth cation andanionsare relatively unstable. The size of the ionic aggregatesis generally proportional to the density of the vapor, which is, inturn, proportional to temperature. The reported results are ingeneral agreement with the earlier reports40,48 on the vapor ofthe imidazolium-based RTILs, showing that ion pairs constitutethe dominant species.The reported data are important from the fundamental

standpoint, providing information on the vapor composition ofionic liquids. Vapors of molecular substances contain individualmolecules, and vapors of inorganic melts contain mostly ionpairs (with modest admixtures of neutral ionic quadruples andneutral ionic sextets).42−45,62−65 In comparison, vapors ofRTILs exhibit a more complicated structure that needs to beaccounted for while building analytic theories. For instance, thesaturated pressure at given conditions depends on the densityof evaporated particles, rather than on their size and mass.From the practical point of view, information regarding VLE isimportant for lubrication, coating, and separation applications.

Figure 7. Percentage of ion pairs in the vapor phase of RTIL:[BPY][PF6] (circles), [BPY][TF] (squares), and [BPY][TFSI](triangles). Compare dependences on the anion and temperature.

Figure 8. Percentage of charged clusters in the vapor of [BPY][PF6](circles), [BPY][TF] (squares), and [BPY][TFSI] (triangles) versustemperature.

Figure 9. Fraction of total ions belonging to the vapor phase versussimulation time in the [BPY][TF] RTIL at 850 K.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4665

■ AUTHOR INFORMATION

Corresponding Author*E-mail: prezhdo@usc.edu. Phone: +1 (213) 821-3116.

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

Fruitful discussions with Andrew Sifain (University of SouthernCalifornia) are gratefully acknowledged. The research wasfunded in part by the US Department of Energy, Grant No.DE-SC0014607.

■ REFERENCES(1) Chaban, V. V.; Prezhdo, O. V. Ionic and Molecular Liquids:Working Together for Robust Engineering. J. Phys. Chem. Lett. 2013,4, 1423−1431.(2) Lei, Z. G.; Dai, C. N.; Zhu, J. Q.; Chen, B. H. ExtractiveDistillation with Ionic Liquids: A Review. AIChE J. 2014, 60, 3312−3329.(3) Karadas, F.; Atilhan, M.; Aparicio, S. Review on the Use of IonicLiquids (Ils) as Alternative Fluids for CO2 Capture and Natural GasSweetening. Energy Fuels 2010, 24, 5817−5828.(4) Aparicio, S.; Atilhan, M.; Karadas, F. Thermophysical Propertiesof Pure Ionic Liquids: Review of Present Situation. Ind. Eng. Chem. Res.2010, 49, 9580−9595.(5) Esperanca, J. M. S. S.; Lopes, J. N. C.; Tariq, M.; Santos, L. M. N.B. F.; Magee, J. W.; Rebelo, L. P. N. Volatility of Aprotic Ionic Liquids- a Review. J. Chem. Eng. Data 2010, 55, 3−12.(6) Blasucci, V. M.; Hart, R.; Pollet, P.; Liotta, C. L.; Eckert, C. A.Reversible Ionic Liquids Designed for Facile Separations. Fluid PhaseEquilib. 2010, 294, 1−6.(7) Araque, J. C.; Yadav, S. K.; Shadeck, M.; Maroncelli, M.;Margulis, C. J. How Is Diffusion of Neutral and Charged TracersRelated to the Structure and Dynamics of a Room-Temperature IonicLiquid? Large Deviations from Stokes-Einstein Behavior Explained. J.Phys. Chem. B 2015, 119, 7015−7029.(8) Araque, J. C.; Hettige, J. J.; Margulis, C. J. Ionic Liquids-Conventional Solvent Mixtures, Structurally Different but DynamicallySimilar. J. Chem. Phys. 2015, 143, 134505.(9) Vilas, M.; Rocha, M. A. A.; Fernandes, A. M.; Tojo, E.; Santos, L.M. N. B. F. Novel 2-Alkyl-1-Ethylpyridinium Ionic Liquids: Synthesis,Dissociation Energies and Volatility. Phys. Chem. Chem. Phys. 2015, 17,2560−2572.(10) Kurnia, K. A.; Sintra, T. E.; Neves, C. M. S. S.; Shimizu, K.;Lopes, J. N. C.; Goncalves, F.; Ventura, S. P. M.; Freire, M. G.; Santos,L. M. N. B. F.; Coutinho, J. A. P. The Effect of the Cation Alkyl ChainBranching on Mutual Solubilities with Water and Toxicities. Phys.Chem. Chem. Phys. 2014, 16, 19952−19963.(11) Harroun, S. G.; Abraham, T. J.; Prudhoe, C.; Zhang, Y. T.;Scammells, P. J.; Brosseau, C. L.; Pye, C. C.; Singer, R. D.Electrochemical Surface-Enhanced Raman Spectroscopy (E-Sers) ofNovel Biodegradable Ionic Liquids. Phys. Chem. Chem. Phys. 2013, 15,19205−19212.(12) Zeng, S. J.; He, H. Y.; Gao, H. S.; Zhang, X. P.; Wang, J.; Huang,Y.; Zhang, S. J. Improving So2 Capture by Tuning Functional Groupson the Cation of Pyridinium- Based Ionic Liquids. RSC Adv. 2015, 5,2470−2478.(13) Deyko, A.; Hessey, S. G.; Licence, P.; Chernikova, E. A.;Krasovskiy, V. G.; Kustov, L. M.; Jones, R. G. The Enthalpies ofVaporisation of Ionic Liquids: New Measurements and Predictions.Phys. Chem. Chem. Phys. 2012, 14, 3181−3193.(14) Cao, Y. Y.; Chen, Y.; Sun, X. F.; Zhang, Z. M.; Mu, T. C. WaterSorption in Ionic Liquids: Kinetics, Mechanisms and Hydrophilicity.Phys. Chem. Chem. Phys. 2012, 14, 12252−12262.(15) Ab Rani, M. A.; Brant, A.; Crowhurst, L.; Dolan, A.; Lui, M.;Hassan, N. H.; Hallett, J. P.; Hunt, P. A.; Niedermeyer, H.; Perez-

Arlandis, J. M.; et al. Understanding the Polarity of Ionic Liquids. Phys.Chem. Chem. Phys. 2011, 13, 16831−16840.(16) Deyko, A.; Lovelock, K. R. J.; Corfield, J. A.; Taylor, A. W.;Gooden, P. N.; Villar-Garcia, I. J.; Licence, P.; Jones, R. G.; Krasovskiy,V. G.; Chernikova, E. A.; et al. Measuring and Predicting Delta H-Vap(298) Values of Ionic Liquids. Phys. Chem. Chem. Phys. 2009, 11,8544−8555.(17) Chum, H. L.; Koch, V. R.; Miller, L. L.; Osteryoung, R. A.Electrochemical Scrutiny of Organometallic Iron Complexes andHexamethylbenzene in a Room-Temperature Molten-Salt. J. Am.Chem. Soc. 1975, 97, 3264−3265.(18) Wu, B. N.; Liang, M.; Maroncelli, M.; Castner, E. W.Photoinduced Bimolecular Electron Transfer from Cyano Anions inIonic Liquids. J. Phys. Chem. B 2015, 119, 14790−14799.(19) Koch, M.; Rosspeintner, A.; Angulo, G.; Vauthey, E. BimolecularPhotoinduced Electron Transfer in Imidazolium-Based Room-Temperature Ionic Liquids Is Not Faster Than in ConventionalSolvents. J. Am. Chem. Soc. 2012, 134, 3729−3736.(20) Vieira, R. C.; Falvey, D. E. Photoinduced Electron-TransferReactions in Two Room-Temperature Ionic Liquids: 1-Butyl-3-Methylimidazolium Hexafluorophosphate and 1-Octyl-3-Methylimida-zolium Hexafluorophosphate. J. Phys. Chem. B 2007, 111, 5023−5029.(21) Marquis, S.; Ferrer, B.; Alvaro, M.; Garcia, H.; Roth, H. D.Photoinduced Electron Transfer in Ionic Liquids: Use of 2,4,6-Triphenylthiapyrylium as a Photosensitizer Probe. J. Phys. Chem. B2006, 110, 14956−14960.(22) Ni, B. K.; Zhang, Q. Y.; Headley, A. D. Pyrrolidine-Based ChiralPyridinium Ionic Liquids (Ils) as Recyclable and Highly EfficientOrganocatalysts for the Asymmetric Michael Addition Reactions.Tetrahedron Lett. 2008, 49, 1249−1252.(23) Xiao, Y.; Malhotra, S. V. Friedel-Crafts Acylation Reactions inPyridinium Based Ionic Liquids. J. Organomet. Chem. 2005, 690,3609−3613.(24) Zhang, C.; Pan, X. Y.; Wang, F.; Liu, X. Q. Extraction-OxidationDesulfurization by Pyridinium-Based Task-Specific Ionic Liquids. Fuel2012, 102, 580−584.(25) Verdia, P.; Gonzalez, E. J.; Rodriguez-Cabo, B.; Tojo, E.Synthesis and Characterization of New Polysubstituted Pyridinium-Based Ionic Liquids: Application as Solvents on Desulfurization ofFuel Oils. Green Chem. 2011, 13, 2768−2776.(26) Zhao, D. S.; Wang, Y. A.; Duan, E. H.; Zhang, J. A. OxidationDesulfurization of Fuel Using Pyridinium-Based Ionic Liquids asPhase-Transfer Catalysts. Fuel Process. Technol. 2010, 91, 1803−1806.(27) Gao, H. S.; Zeng, S. J.; Liu, X. M.; Nie, Y.; Zhang, X. P.; Zhang,S. J. Extractive Desulfurization of Fuel Using N-Butylpyridinium-BasedIonic Liquids. RSC Adv. 2015, 5, 30234−30238.(28) Bittner, B.; Wrobel, R. J.; Milchert, E. Physical Properties ofPyridinium Ionic Liquids. J. Chem. Thermodyn. 2012, 55, 159−165.(29) Rebelo, L. P. N.; Lopes, J. N. C.; Esperanca, J. M. S. S.; Filipe, E.On the Critical Temperature, Normal Boiling Point, and VaporPressure of Ionic Liquids. J. Phys. Chem. B 2005, 109, 6040−6043.(30) Earle, M. J.; Esperanca, J. M. S. S.; Gilea, M. A.; Lopes, J. N. C.;Rebelo, L. P. N.; Magee, J. W.; Seddon, K. R.; Widegren, J. A. TheDistillation and Volatility of Ionic Liquids. Nature 2006, 439, 831−834.(31) Paulechka, Y. U.; Zaitsau, D. H.; Kabo, G. J.; Strechan, A. A.Vapor Pressure and Thermal Stability of Ionic Liquid 1-Butyl-3-Methylimidazolium Bis(Trifluoromethylsulfonyl)Amide. Thermochim.Acta 2005, 439, 158−160.(32) Ballone, P.; Pinilla, C.; Kohanoff, J.; Del Popolo, M. G. Neutraland Charged 1-Butyl-3-Methylimidazolium Triflate Clusters: Equili-brium Concentration in the Vapor Phase and Thermal Properties ofNanometric Droplets. J. Phys. Chem. B 2007, 111, 4938−4950.(33) Lovelock, K. R. J.; Armstrong, J. P.; Licence, P.; Jones, R. G.Vaporisation and Thermal Decomposition of DialkylimidazoliumHalide Ion Ionic Liquids. Phys. Chem. Chem. Phys. 2014, 16, 1339−1353.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4666

(34) Armstrong, J. P.; Hurst, C.; Jones, R. G.; Licence, P.; Lovelock,K. R. J.; Satterley, C. J.; Villar-Garcia, I. J. Vapourisation of IonicLiquids. Phys. Chem. Chem. Phys. 2007, 9, 982−990.(35) Brunetti, B.; Ciccioli, A.; Gigli, G.; Lapi, A.; Misceo, N.; Tanzi,L.; Ciprioti, S. V. Vaporization of the Prototypical Ionic LiquidBmimntf(2) under Equilibrium Conditions: A Multitechnique Study.Phys. Chem. Chem. Phys. 2014, 16, 15653−15661.(36) Iwahashi, T.; Ishiyama, T.; Sakai, Y.; Morita, A.; Kim, D.; Ouchi,Y. Liquid/Liquid Interface Layering of 1-Butanol and [Bmim]Pf6Ionic Liquid: A Nonlinear Vibrational Spectroscopy and MolecularDynamics Simulation Study. Phys. Chem. Chem. Phys. 2015, 17,24587−24597.(37) Palchowdhury, S.; Bhargava, B. L. Effect of Spacer Chain Lengthon the Liquid Structure of Aqueous Dicationic Ionic Liquid Solutions:Molecular Dynamics Studies. Phys. Chem. Chem. Phys. 2015, 17,11627−11637.(38) Paredes, X.; Fernandez, J.; Padua, A. A. H.; Malfreyt, P.;Malberg, F.; Kirchner, B.; Pensado, A. S. Bulk and Liquid-VaporInterface of Pyrrolidinium-Based Ionic Liquids: A MolecularSimulation Study. J. Phys. Chem. B 2014, 118, 731−742.(39) Paredes, X.; Fernandez, J.; Padua, A. A. H.; Malfreyt, P.;Malberg, F.; Kirchner, B.; Pensado, A. S. Using Molecular Simulationto Understand the Structure of [C(2)C(1)Im](+)-Alkylsulfate IonicLiquids: Bulk and Liquid-Vapor Interfaces. J. Phys. Chem. B 2012, 116,14159−14170.(40) Chaban, V. V.; Prezhdo, O. V. Ionic Vapor: What Does ItConsist Of? J. Phys. Chem. Lett. 2012, 3, 1657−1662.(41) Galamba, N.; de Castro, C. A. N.; Ely, J. F. ThermalConductivity of Molten Alkali Halides from Equilibrium MolecularDynamics Simulations. J. Chem. Phys. 2004, 120, 8676−8682.(42) Galamba, N.; Cabral, B. J. C. First Principles MolecularDynamics of Molten NaCl. J. Chem. Phys. 2007, 126, 124502.(43) Galamba, N.; de Castro, C. A. N.; Ely, J. F. Equilibrium andNonequilibrium Molecular Dynamics Simulations of the ThermalConductivity of Molten Alkali Halides. J. Chem. Phys. 2007, 126,204511.(44) Galamba, N. Molecular Dynamics Study of the Vaporization ofan Ionic Drop. J. Chem. Phys. 2010, 133, 124510.(45) Galamba, N.; Cabral, B. J. C. First Principles MolecularDynamics of Molten NaI: Structure, Self-Diffusion, PolarizationEffects, and Charge Transfer. J. Chem. Phys. 2007, 127, 094506.(46) Yu, G. R.; Chen, X. C.; Asumana, C.; Zhang, S. J.; Liu, X. M.;Zhou, G. H. Vaporization Enthalpy and Cluster Species in Gas Phaseof 1,1,3,3-Tetramethylguanidinium-Based Ionic Liquids from Com-puter Simulations. AIChE J. 2011, 57, 507−516.(47) Weiss, V. C. Liquid-Vapor Equilibrium and Critical Parametersof the Ionic Liquid 1-Butyl-3-Methylimidazolium Hexafluorophos-phate from Molecular Dynamics Simulations. J. Mol. Liq. 2015, 209,745−752.(48) Neto, B. A. D.; Meurer, E. C.; Galaverna, R.; Bythell, B. J.;Dupont, J.; Cooks, R. G.; Eberlin, M. N. Vapors from Ionic Liquids:Reconciling Simulations with Mass Spectrometric Data. J. Phys. Chem.Lett. 2012, 3, 3435−3441.(49) AbdulSada, A. K.; Elaiwi, A. E.; Greenway, A. M.; Seddon, K. R.Evidence for the Clustering of Substituted Imidazolium Salts ViaHydrogen Bonding under the Conditions of Fast Atom BombardmentMass Spectrometry. Eur. Mass Spectrom. 1997, 3, 245−247.(50) Bini, R.; Bortolini, O.; Chiappe, C.; Pieraccini, D.; Siciliano, T.Development of Cation/Anion ″Interaction″ Scales for Ionic Liquidsthrough ESI-MS Measurements. J. Phys. Chem. B 2007, 111, 598−604.(51) Bussi, G.; Donadio, D.; Parrinello, M. Canonical Samplingthrough Velocity Rescaling. J. Chem. Phys. 2007, 126, 014101−014108.(52) Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. Gromacs 4:Algorithms for Highly Efficient, Load-Balanced, and ScalableMolecular Simulation. J. Chem. Theory Comput. 2008, 4, 435−447.(53) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M.Lincs: A Linear Constraint Solver for Molecular Simulations. J.Comput. Chem. 1997, 18, 1463−1472.

(54) Voroshylova, I. V.; Chaban, V. V. Atomistic Force Field forPyridinium-Based Ionic Liquids: Reliable Transport Properties. J. Phys.Chem. B 2014, 118, 10716−10724.(55) Canongia Lopes, J. N.; Padua, A. A. H. Molecular Force Fieldfor Ionic Liquids III: Imidazolium, Pyridinium, and PhosphoniumCations; Chloride, Bromide, and Dicyanamide Anions. J. Phys. Chem. B2006, 110, 19586−19592.(56) Berendsen, H. J. C.; van der Spoel, D.; van Drunen, R. Gromacs- a Message-Passing Parallel Molecular-Dynamics Implementation.Comput. Phys. Commun. 1995, 91, 43−56.(57) Lindahl, E.; Hess, B.; van der Spoel, D. Gromacs 3.0: A Packagefor Molecular Simulation and Trajectory Analysis. J. Mol. Model. 2001,7, 306−317.(58) Humphrey, W.; Dalke, A.; Schulten, K. Vmd: Visual MolecularDynamics. J. Mol. Graphics 1996, 14, 33−38.(59) Allouche, A. R. Gabedit-a Graphical User Interface forComputational Chemistry Softwares. J. Comput. Chem. 2011, 32,174−182.(60) Martinez, L.; Andrade, R.; Birgin, E. G.; Martinez, J. M.Packmol: A Package for Building Initial Configurations for MolecularDynamics Simulations. J. Comput. Chem. 2009, 30, 2157−2164.(61) Liu, L. M.; Car, R.; Selloni, A.; Dabbs, D. M.; Aksay, I. A.;Yetter, R. A. Enhanced Thermal Decomposition of Nitromethane onFunctionalized Graphene Sheets: Ab Initio Molecular DynamicsSimulations. J. Am. Chem. Soc. 2012, 134, 19011−19016.(62) Pitzer, K. S. Critical-Point and Vapor-Pressure of Ionic FluidsIncluding NaCl and KCl. Chem. Phys. Lett. 1984, 105, 484−489.(63) Pitzer, K. S. Ionic Fluids - near-Critical and Related Properties. J.Phys. Chem. 1995, 99, 13070−13077.(64) Gillan, M. J. Liquid-Vapor-Equilibrium in the RestrictedPrimitive Model for Ionic Liquids. Mol. Phys. 1983, 49, 421−442.(65) Weiss, V. C. Predicting Critical Temperatures of Ionic and Non-Ionic Fluids from Thermophysical Data Obtained near the MeltingPoint. J. Chem. Phys. 2015, 143, 144503.

The Journal of Physical Chemistry B Article

DOI: 10.1021/acs.jpcb.6b03130J. Phys. Chem. B 2016, 120, 4661−4667

4667