Theoretical calculation of the physico-chemical properties of 1-butyl-4-
methylpyridinium based ionic liquids

Artur Giełdoń, Maciej Bobrowski, Aleksandra Bielicka-Giełdoń, Cezary

Czaplewski

PII: S0167-7322(16)33119-1
DOI: doi: 10.1016/j.molliq.2016.11.087
Reference: MOLLIQ 6634

To appear in: Journal of Molecular Liquids

Received date: 17 October 2016

Accepted date: 25 November 2016

Please cite this article as: Artur Gieldoń, Maciej Bobrowski, Aleksandra Bielicka-Gieldoń,
Cezary Czaplewski, Theoretical calculation of the physico-chemical properties of 1-
butyl-4-methylpyridinium based ionic liquids, Journal of Molecular Liquids (2016), doi:
10.1016/j.molliq.2016.11.087

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Theoretical calculation of the physico-chemical properties of 1-butyl-4-methylpyridinium

based ionic liquids.

Artur Giełdoń1,2, Maciej Bobrowski2, Aleksandra Bielicka-Giełdoń1, Cezary Czaplewski1,2,*

T
IP
1
Faculty of Chemistry, University of Gdańsk, Wita Stwosza 63, 80-308 Gdańsk, Poland,
2
Faculty of Technical Physics and Applied Mathematics, Technical University of Gdańsk,

R
Naturowicza 11/12, 80-233 Gdańsk, Poland

SC
*Corresponding author, e-mail [email protected]

Abstract
NU
MA
Ionic liquids (ILs) have attracted much attention for their unique physicochemical
properties, which can be designed as needed by altering the ion combinations. Besides
experimental work, numerous computational studies have been concerned with prediction of
D

physical properties of ILs. The results of molecular dynamics simulations of ILs depend
TE

strongly on the proper force field parameterization. Classical force fields and the force fields
designed specifically for ILs have been found to yield diffusion coefficients, that are too low
P

for the liquid state. This has been attributed to the lack of electronic polarizability in the most
CE

force fields. One simple solution to this problem has been uniform by scaling down of the
partial charges of all ions. In this work we investigated the influence of the charge scaling, the
size of the simulated system and the temperature factor on calculated density, radial
AC

distribution function and the diffusion coefficients of the cation and anion of two
methylpyridinium based ionic liquids: 1-butyl-4-methylpyridinium tetrafluoroborate
([b4mpy][BF4]) and 1-butyl-4-methylpyridinium chloride ([b4mpy][Cl]). We show that the
parameterization is the key for a proper reproduction of the diffusion coefficient and as a
consequence the melting temperature and ionic conductivity. We were also able to get some
structural informations on the cation-anion relationships in the investigated ILs.

Keywords:
Molecular Dynamics, Ionic Liquids, 1-butyl-4-methylpyridinium, ([b4mpy][BF4]),
([b4mpy][Cl]), charge distribution, density calculation, melting point calculation,

1
 ACCEPTED MANUSCRIPT

Introduction
Ionic liquids (IL) are novel organic salts generally composed of an organic cation and
an anion, which is usually inorganic. It is known that ionic liquids exhibit extremely low
vapor pressure, thermal stability up to 400 ºC and relatively low flammability. The unique

T
physicochemical properties of ionic liquids make them a good replacement of standard

IP
organic solvents in a great number of different chemical reactions. Due to the chemical
stability and possibility of the recycled usage, the ionic liquids are in the field of interest of

R
the industry [1,2]. The properties of IL can be modified by simple changes in the nature of the

SC
cation or anion. The knowledge of the way how to change physicochemical properties of ionic
liquids is crucial for their optimized design [3,4]. According to Katrizky et al., there are

NU
approximately 1018 anion-cation combinations [5], therefore all methods which can be used
for prediction of some IL properties can be extremely useful.
MA
Theoretical chemistry offers three ways of the prediction of ionic liquid properties.
Quantitative structure – property relationship (QSPR) methods are based on the extrapolation
of already known or calculated properties of the training set of molecules. The newly built
D

model of relationship can be used for a prediction of the specific properties for the new
TE

molecules [6,7,8,9]. Semi-empirical and ab-initio methods can be used to calculate the
electron density (charge distribution, polarizability), molecular geometry and volume of the
P

ion pairs [10,11]. The high correlation between molecular volume and basic physicochemical
CE

properties for the ionic liquids was shown by Slattery et al [12] and Wilenska et al [13].
Empirical force field and molecular dynamics methods (MD) can be used for the prediction
almost all physicochemical properties of ionic liquids, however as it is shown in this study the
AC

obtained results are highly dependent on the simulation parameters.

Molecular dynamics simulations of ionic liquids are preformed for almost 20 years
[14,15]. Lee and co workers studied the ionic conductivity for 1-n-butyl-3-methylimidazolium
ionic liquids with 1 ns molecular dynamics and show, that the self-diffusion coefficient of
ionic liquids is an important factor in determination of the ionic conductivity [16]. Prado and
Freitas aimed to investigate relationships between structure and thermodynamical properties
of the 1-butyl-3-methylimidazolium tetrafluoroborate by using 5 ns MD simulation and
OPLS-AA force field and show the difference between the diffusion of the cation and anion
[17]. Kodderman and co workers calculated viscosities for imidazolium-based ionic liquids by
using Green-Kubo relation and 10 ns MD and show that the parameterization of the force field
is crucial in the investigation of the ILs properties [18]. Picalek and co workers investigated
the surface and IL-water interface of aqueous solutions of 1-butyl-3-methylimidazolium

2
 ACCEPTED MANUSCRIPT

tetrafluoroborate with 4 ns MD and MACSIMUS package and show that the surface tension
can be reproduced only by the polarizable force fields [19]. Zhang and Maginn compared
several methods for calculation of the melting point temperature using molecular dynamics
simulations for 1-n-butyl-3-methylimidazolium chloride [bmim][Cl] [20]. They postulate, that

T
the pseudo-supercritical path (PSCP) method is reliable for prediction of the melting point

IP
temperature. The effects of linear scaling of the atomic point-charges were investigated by
Youngs and Hardacre for the imidazolium based ionic liquid [21]. They show correlation

R
between the charge scaling factor and the diffusion coefficient. Kowsari and co workers

SC
calculated dynamics and diffusion coefficient [22] and transport coefficients [23] of twelve 1-
alkyl-3-methylimidazolium-based ionic liquids by using OPLS and AMBER forcefield. They

NU
show that the Green-Kubo relation gives lower values for electrical conductivity than the
Nerst-Einstein relation. Kislenko and co workers studied the electrochemical interface
MA
between a graphite surface and the 1-butyl-3-methylimidazolium with 4 ns MD and AMBER
forcefield [24]. They observed, confirmed experimentally, formation of the interfacial layers.
It has been demonstrated, that enzymes are stabilized in several ILs well beyond room
D

temperature. Klahn and co workers investigated the stability of the lipase enzyme in several
TE

ionic liquids by using GROMACS package [25].

Herein, we report the results of our theoretical studies on two ILs: 1-butyl-4-
P

methylpyridinium tetrafluoroborate ([b4mpy][BF4]) (CAS Number 343952-33-0) and 1-butyl-

CE

4-methylpyridinium chloride ([b4mpy][Cl]) (CAS Number 112400-86-9) by molecular

dynamics (MD) simulation. The most available literature data on ([b4mpy][BF4]) describes
viscosity [26,27,28,29,30,31,32] and density [28,29,31,32,33] as measured in a function of the
AC

temperature. There are also a substantial amount of literature data according physicochemical
properties of binary [26,34,35,36,37] and ternary [38] systems of ([b4mpy][BF4]). It is known,
that ([b4mpy][BF4]) improves enantioselectivity of lipases [39], can be used as a solvent for
separation of n-alkanes [40,41] and has an inhibition effect of the copper corrosion [42].
There is also information on ([b4mpy][BF4]) toxicity [2,43,44]. The melting temperature of
([b4mpy][BF4]) is 303 K [45]. If we compare it with 301 K of ([b2mpy][BF4]) [46] we can
conclude, that isomerization of the 1-butyl-4-methylpyridinium cation seems to have a slight
effect on the melting temperature.
With our best knowledge there is only one theoretical work on ([b4mpy][BF4]) [47].
The authors used COSMO-RS method to reproduce the excess enthalpy and solubility of the
binary mixtures containing ([bXmpy][BF4]) and B2(CH2)υBr (X=2-4; υ=2-6). The anion BF4-
can be considered as a hydrogen bond acceptor and the hydrogen atoms of the pyridine ring

3
 ACCEPTED MANUSCRIPT

can act as donors. The alkyl group of the cation promotes interactions with apolar compounds
[47]. [b4mpy][Cl] can exists in two forms, the first is solid with a melting temperature of
about 332 – 334 K, the second comes from the crystallization from dry acetone and form of a
white highly hygroscopic needles with the melting temperature ranges between 430 and 433K

T
[48]. If we compare the melting temperature of the first polymorphic form with 368 K [49],

IP
384 K [50] and 370 K [51] of [b3mpy][Cl]) we can make the same conclusion as in the case
of ([b4mpy][BF4]) that isomerization of the 1-butyl-4-methylpyridinium seems to have a

R
small effect on the melting temperature. Ellis and coworkers investigated the usage of

SC
([b4mpy][Cl]) with AlCl3 as a solvent for the linear dimerisation of but-1-ene in biphasic
mode [52].

NU
In this work we investigated the influence of the charge distribution, the size of the
simulated system and the temperature factor on calculated density, radial distribution function
MA
and the mobility of the cation and anion. Additionally we investigated the influence of the
anion and verified, if the molecular dynamics is able to predict the state of the simulated
chemical specimen (liquid or solid at given temperature).
D
TE

Methods
1-butyl-4-methylpyridinium tetrafluoroborate ([b4mpy][BF4]) and 1-butyl-4-
P

methylpyridinium chloride ([b4mpy][Cl]) were parameterized by using the antechamber

CE

program from AMBER v.12 [53,54] package. The partial charges of cation and anion were
fitted to ±1 (Fig 1.). Three different molecular systems for both ionic liquids were built by
placing 6x6x6, 8x8x8 and 10x10x10 unit boxes in the x, y and z dimensions and with a
AC

volume of each unit box of about 2000 Å3. We assume that every pair of cation and anion can
occupy one unit box. In the next step we performed 2 ns molecular dynamics simulation with
a temperature set to 300 K and NTP ensemble. The initial density of each simulated molecular
system was about 0.2 g/cm3 and the equilibrium density was reached after 1 ns in all
simulations. This procedure let us to obtain the quasi random IL structure which was used for
a further investigation. Following, in all of the newly obtained IL molecular systems (three of
([b4mpy][BF4]) and three of ([b4mpy][Cl]) with a size of 63, 83and 103 cation - anion pairs),
the partial charges were scaled by a factor of 1.0, 0.96, 0.92 and 0.88 and simulated in time of
10 ns in five different temperatures: 270, 300, 330, 360 and 390 K by using MD as
implemented in AMBER v.12 package [53]. The General AMBER force field (GAFF) was
already used with a success for prediction of some thermodynamic properties of the ionic
liquids [24,55,56]. All simulations were done with the following conditions: 1 fs time step,

4
 ACCEPTED MANUSCRIPT

constant pressure (1 atm) and periodic boundary conditions. Long-range electrostatic

interactions were simulated using PME (Particle Mesh Ewald) method [57]. Non-bonding
interactions were updated every 25 steps and data were collected every 1 ps. All plots were
created and fitted using GNUPLOT [58], Origin [59] and MS EXCEL program. Spatial

T
Distribution Analysis was performed by using Travis [60] and VMD program [61].

IP
Results and discussion

R
In the case of ionic liquid the description of the electrostatic interactions is the key to

SC
reproduce experimental results by using MD simulations with empirical force field. However,
the scaling factor which mimics polarizability has to be used with extreme caution, due to the

NU
fact, that the details in the radial distribution function (RDF) are lost, as the charges on the
ions decrease [21,62]. This task seems to be difficult, because the linear change of the charge
MA
distribution is causing exponential changes in the mobility as described by the mean-squared
displacement (MSD) curves. In the investigated range of the charge scaling factor (0.88 to
1.0) we did not observed any significant change in the shape of the RDFs as shown in Fig 2.
D

With the temperature incensement the value of the first maximum responsible for the
TE

description of the most probable cation-anion distance, was decreased and the peak was
flattened. We observed this tendency with the charge scaling factor of 0.92 and 0.88, after
P

crossing the IL melting temperature. This effect was caused by an increased mobility of the
CE

ions for temperatures higher than melting temperature. With the scaling factor of 0.96 and
1.00, before and after crossing the experimentally measured melting point temperature, the
RDF plots are almost identical. This effect is visible in the case of all analyzed interactions
AC

(cation-anion, anion-anion and cation-cation). To analyze in detail the cation-anion

relationships, we performed Spatial Distribution Function (SDF) analysis as shown on Fig. 3.
In the case of both analyzed ionic liquids the most probable cation-anion interaction place is
located with the hydrogen atom in the vicinity of the nitrogen from the pyridine ring. This is
due to the highest difference in the partial charge. Due to the presence of the highly mobile
alkyl chain, the interaction area is rotated against the pyridine ring. There is the second
interaction area with the second hydrogen atom from the pyridine ring. In this case we also
observed the rotation, however it is caused by the cation-cation interaction. Additionally, we
observed a π-π interaction type as available on both sides of the pyridine ring. After crossing
the melting temperature, the first area of the cation-cation interactions is gone, due to the
increased mobility of the anions. All blue spots visible on Fig. 3 corresponds to the second
interaction area around 15 Å as visible on the RDF plots.

5
 ACCEPTED MANUSCRIPT

To observe in detail the change of the mobility of the ions as a function of the temperature for
different charge scaling factors, we have calculated mean square displacement (MSD) curves.
The MSD allows for easy calculation of self-diffusion coefficient. The first step to do so, is
the linear regression of the MSD plots, to obtain time related MSD. In our case it is nano

T
second time scale [Å2/ns], and in the manuscript will be called MSD-ns. This value can be

IP
placed directly to the Einstein formula for calculation of the self-diffusion coefficient Di.
d 1
Di  lim  [ri (t )  ri (0)]2 

R
dt 6

SC
Other physicochemical parameters such as viscosity, ionic conductivity or thermal
conductivity depend mostly on diffusion. Therefore the calculation of the mobility is the key

NU
for a proper theoretical prediction of the physical properties of ILs. The direct correlation
between diffusion and viscosity is defined by the Stokes-Einstein equation, in which those
values are inverse proportional [63,64,65].
MA
MSD plots calculated from MD simulations of [b4mpy][BF4] with different box sizes are
presented on Fig. 4. For the small system with only 216 ion pairs, the deviation from the
D

linear lines are quite large and can be associated with local ordering and faster/slower
movement of some ions. For larger systems with 512 or 1000 ions pairs MSD plots do not
TE

deviate from linear lines. With the large enough molecular system, the local deviations are not
statistically important and does not have a big influence for the obtained MSD-ns. The
P

summary of the obtained MSD-ns values are shown on Fig. 5. Our result show, that for MD
CE

simulation with 1.0 and 0.96 scaling factor, MSD-ns values are very small for all
temperatures. The results with scaling factor 0.88 are in the best agreement with experimental
AC

data, since the MSD plots (see Fig. 5) fits best to the viscosity plot [27]. However, in this
case, the estimation error is up to 50%. With the charge scaling factor of 0.7, the estimation
error can be even up to 80% as shown by Sprenger et al. [56].
The most adequate way for calculation of the melting temperature for ionic liquids seems to
be the Pseudo-Supercritical Path (PSCP) method postulated by Margin and co-workers
[20,66], however, its application is rather problematic. According to the Lindemann criterion,
the melting occurs, when the thermal motion of the atoms of the crystal reaches MSD of about
one tenth of the interatomic distance [67,68]. It can be briefly defined as (MSD/d2) ~ 0.1,
where “d” is the average intermolecular distance. In the case of the liquids, such universal
criterion does not exist, but have to be found individually, as it was shown by Niss et al. [69].
The authors named it as the intrinistic Lindemann coefficient. In the crystals MSD comes
mostly from the vibrations, in the case of the liquids, the major contribution comes from

6
 ACCEPTED MANUSCRIPT

displacements. With the temperature increase in the crystals, the ordered state melts at Tm,
whereas in the amorphous state the disordered structure softens at the glass transition
temperature Tg, above which the flow connected with the viscosity, occurs [70]. The empirical
law Tg≈ 2/3 Tm suggest that those two phenomena have a common basis [71]. Many ILs

T
slowly form glasses at low temperatures, because they have negligible vapor pressures. In

IP
general, they do not evaporate or boil at high temperatures [72]. In our case the melting
temperatures for ([b4mpy][BF4]) and ([b4mpy][Cl]) are about 303 and 333 K, respectively. It

R
suggest the following glass transition temperatures: 202 and 222 K for ([b4mpy][BF4]) and

SC
([b4mpy][Cl]), respectively. This prediction holds well, since the glass transition temperature
Tg for ([b4mpy][BF4]) is 203 K [73]. In such a low temperature it is very difficult to catch a

NU
proper MSD value to identify the glass transition temperature. Therefore the direct
identification of the melting temperature, seems to be a better solution. The straightforward
MA
calculation, from the MD simulation with the temperature restraints of the melting point
temperature, can be overestimated or underestimated, due to the existence of the superheating
/ supercooling phenomenon [20,74]. To avoid that problem we decided to perform a set of 10
D

ns long MD simulations in a wide range of temperatures. As a consequence the estimation

TE

error was within temperature step. In the case of ([b4mpy][BF4]) the experimentally measured
melting temperature is 303 K [45]. The mean square displacements (MSD-ns) in 300K were
3.06 and 2.34 and in 310 K were 5.07 and 3.30 [Å2/ns] for cation and anion, respectively (see
P

Fig 5). The average cation – anion distances at 300 and 310 K were 4,7 and 4,8 Å as identified
CE

with the RDF plots (see Fig. 2 and supplementary data). The experimentally measured
melting temperature of ([b4mpy][Cl]), is about 333K. According to our calculation the
AC

average distance between cation and anion is about 4.5 Å in 340 and 350K. The mean square
displacements (MSD-ns) are 2.44 and 2.36 in 340 K and 8.24 and 6.40 in 350 K for cation and
anion, respectively (see Fig 5). The summary of the calculation of the so called by Niss et al.
[69] intrinistic Lindemann criterion are summarized in Table 1.

7
 ACCEPTED MANUSCRIPT

Table 1. The summary of the calculation of the intrinistic Lindemann coefficient for both
simulated ionic liquids. The units of the diffusion coefficients are [Å2/ns], the average
distances are in [Å].
([b4mpy][BF4]) ([b4mpy][Cl])

T
300K 310K 340K 350K

IP
cation anion cation anion cation anion cation anion
3.06 2.34 5.07 3.30 2.44 2.36 8.24 6.40
 0.138  0.106  0.220  0.143  0.120  0.116  0.407  0.316

R
4.7 2 4.7 2 4.8 2 4.8 2 4.5 2 4.5 2 4.5 2 4.5 2

SC
According to Larini et al. [70] the ratio of MSD/d2 is close to 0.129 for the melting of a hard-
sphere face-centred cubic solid. In the case of ionic liquids, this value seems to be a little bit

NU
higher. According to our results, in the melting temperature, the ratio of ([MSD-ns] / d2 ) is
about 0.20 and 0.14 for cation and anion, respectively. The differences are caused by the fact,
MA
that the diffusion coefficient for the anion is smaller, than for the cation [17]. In general we
can postulate, that in the MD simulations the melting temperature will be crossed, when the
MSD will be grater than the average distance between cation and anion.
D

The self-diffusion coefficient of ionic liquid is the most important factor in

TE

determination of the ionic conductivity. In this work we calculated the ionic conductivity of
[b4mpy][BF4] by using the nerst-Einstein equation, and compared it with available
P

experimental data [73] (see Fig. 6).

CE

Ne 2 ( Dself ( cat )  Dself ( an) )

c 
Vk BT
AC

[σc – ionic conductivity; N – number of ions in the simulation cell; T – temperature; e –

electron charge; kB – Boltzmann’s constant; Dself – self-diffusion coefficient, as obtained with
Einstein formula.]

With the charge scaling factor of 0.88, the obtained ionic conductivity is systematically lower,
with an error similar as in the prediction of the viscosity (about 50%). Picalek and Kolafa [75]
are explaining this phenomenon by a short-time clustering which is caused by the correlation
in the ion motion. With the decreasement of the charge scaling factor, the interactions in the
clusters are weaker and the ion motion is higher.
Density of IL is a static parameter which is less sensitive to the charges used in
parameterization of empirical force field used in MD simulation then a dynamic parameter as
diffusion coefficient. Densities calculated from MD simulations of [b4mpy][BF4] at

8
 ACCEPTED MANUSCRIPT

temperatures from 270 to 390 K are compared to experimental data on Fig. 7. In contrast to
MSD, the results from the density calculation, with charge scaling factor 0.88 are further from
experimental data than results with scaling factor 1.0. However, both sets of simulations
properly describe linear dependency of the density on temperature.

T
IP
Conclusions
We performed 60 simulations for 1-butyl-4-methylpyridinium tetrafluoroborate

R
([b4mpy][BF4]) and for 1-butyl-4-methylpyridinium chloride ([b4mpy][Cl]) testing different

SC
sizes of the simulation system in different temperatures and with different charge scaling
factors. The minimal size of the simulated system have to be at least 500 cation-anion pairs.

NU
The size of the simulated system affects calculation of mean square displacement and only for
large enough molecular system, MSD curves are linear. In the case of the small molecular
MA
system, the local disturbances have a large influence on the obtained result. Scaling down all
partial charges by a factor of 0.88 is a simple solution to mimic polarization effect in IL and
gives adequate description of ionic mobility and its changes with temperature. The charge
D

scaling factor of 1.00 and 0.96 does not reproduce the diffusion coefficient and as a
TE

consequence viscosity and the other diffusion based properties. Molecular dynamics
simulation is able to predict viscosity and ionic conductivity, however, our results show, that
P

the error is highly dependent on the charge scaling factor. While the charge scaling factor
CE

decreases, the estimation error of the viscosity and ionic conductivity decreases as well. In our
case the estimation error was up to 50%. MD is a good tool for the structural description,
however we are losing some structural information with the decreasement of the charge
AC

scaling factor. According to our results, the best compromise is 0.88. Using intrinistic
Lindemann criterion of about 0.20 and 0.14 for cation and anion, respectively, we are able to
identify the investigated IL melting point with an error less than 5%. However, to extend this
statement for the other IL’s, further investigation is needed. We were able to predict density
and radial distribution functions which are less sensitive to the charge parameterization used
in the simulations. In the case of the density, MD simulation was able to reproduce properly
linear dependency on temperature with the estimation error less than 5%.

9
 ACCEPTED MANUSCRIPT

Acknowledgements
This work was supported by Switzerland through the Swiss Contribution to the enlarged
European Union [grant number PSPB-051/2010]. Calculations were carried out at the
supercomputers of Informatics Center of the Metropolitan Academic Network (IC MAN) at

T
the Gdansk University of Technology in Poland

R IP
SC
NU
MA
D
P TE
CE
AC

10
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D

Fig. 1. Partial charges calculated for ([b4mpy][BF4]) and ([b4mpy][Cl]) ions using
TE

antechamber AM1-BCC method.

P
CE
AC

11
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D
P TE
CE
AC

Fig 2. Radial distribution function (RDF) for ([b4mpy][BF4]) and ([b4mpy][Cl]) ionic liquid calculated
from MD simulations with 0.88 and 1.0 charge scaling factors, before and after crossing melting
temperature. All plots correspond to the system size of 10x10x10 (10_10) cation-anion pairs.

12
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D
P TE
CE
AC

Fig 3. Spatial distribution function (SDF) for ([b4mpy][BF4]) (top, A – 300K and B – 360K)
and ([b4mpy][Cl]) (bottom, C – 330K and D – 390K). The red area corresponds to the
location of the anion around the cation, the blue area corresponds to the location of the cation.
All plots are taken in the distance of 0.4 Å in the case of the anion and 1.4 Å in the case of the
cation from the first maximum of interactions as visible on RDF plots. Fig. A corresponds to
51 and 55 % of population in the case of the cation and anion, respectively. Fig. B – 50 and 49
%, Fig. C and D – 48 and 55 % of population. One have to take the account, that after
crossing the melting temperature, the first sphere of the cation-cation interactions is gone and
the blue spots belong to the second peak on the RDF plots. Both ionic liquids are with the
0.88 charge scaling factor.

13
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D
P TE
CE
AC

Fig 4. Mean square displacement (MSD) [Å2] plots for [b4mpy][BF4] with different box size: 1000
(10_10) ion pairs (top), 512 (8_8) ion pairs (middle), 216 (6_6) ion pairs (bottom). The linear
approximation are given on the top right on the plots. The slope coefficient is called in the manuscript,
diffusion coefficient [Å2/ns].

14
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA
D
P TE
CE
AC

Fig 5. Mean square displacement as a function of temperature and charge scaling factor used
in the simulations. Viscosity data, as shown on the bottom-left plot (light-blue) are taken from
Bandres at al [27].

15
 ACCEPTED MANUSCRIPT

T
R IP
SC
NU
MA

Fig 6. The electrical conductivity, σ, in a function on the inverse of the absolute temperature
for ([b4mpy][BF4]). VTF – Vogel-Fulcher-Tammann equation. Experimental data are taken
D

from Bandres at al. [73].

P TE
CE
AC

Fig 7. Density calculated from MD simulations of ([b4mpy][BF4]) as a function of the

absolute temperature, compared to experimental data (LIT 1: Nonthanasin et al. [33], LIT 2:
Sanchez et al. [28].)

16
 ACCEPTED MANUSCRIPT

References:

T
IP
R
SC
NU
Graphical abstract MA
D
P TE
CE
AC

17
 ACCEPTED MANUSCRIPT

Highlights
Molecular dynamics simulation is able to predict density, viscosity and ionic conductivity of
ILs.

T
Scaling down partial charges can mimic polarization and is necessary to reproduce physico-

IP
chemical properties of ILs in MD.

R
Intrinistic Lindemann criterion can identify the ILs melting point, the increased mobility is

SC
shown on SDF plots.

NU
Insufficient size of the simulated system can effect in large local disturbances.

1. M. Palacio, B. Bhushan, A Review of Ionic Liquids for Green Molecular Lubrication in

MA
Nanotechnology, Tribology Letters. 40 (2010) 247–268. doi:10.1007/s11249-010-9671-8.
2. S. Keskin, D. Kayrak-Talay, U. Akman, Ö. Hortaçsu, A review of ionic liquids towards
supercritical fluid applications, The Journal of Supercritical Fluids. 43 (2007) 150–180.
doi:10.1016/j.supflu.2007.05.013.
D

3. M. Freemantle, DESIGNER SOLVENTS: Ionic liquids may boost clean technology

development, Chemical & Engineering News. 76 (1998) 32–37. doi:10.1021/cen-
TE

v076n013.p032.
4. C. Chiappe, D. Pieraccini, Ionic liquids: solvent properties and organic reactivity, Journal
of Physical Organic Chemistry. 18 (2005) 275–297. doi:10.1002/poc.863.
P

5. A.R. Katritzky, R. Jain, A. Lomaka, R. Petrukhin, M. Karelson, A.E. Visser, R.D. Rogers,
CE

Correlation of the Melting Points of Potential Ionic Liquids (Imidazolium Bromides and
Benzimidazolium Bromides) Using the CODESSA Program, Journal of Chemical Information
and Computer Sciences. 42 (2002) 225–231. doi:10.1021/ci0100494.
6. D.M. Eike, J.F. Brennecke, E.J. Maginn, Predicting melting points of quaternary
AC

ammonium ionic liquidsElectronic supplementary information (ESI) available: training sets B

and C. See http://www.rsc.org/suppdata/gc/b3/b301217d/, Green Chemistry. 5 (2003) 323.
doi:10.1039/b301217d.
7. R. Bini, M. Malvaldi, W.R. Pitner, C. Chiappe, QSPR correlation for conductivities and
viscosities of low-temperature melting ionic liquids, Journal of Physical Organic Chemistry.
21 (2008) 622–629. doi:10.1002/poc.1337.
8. H. Matsuda, H. Yamamoto, K. Kurihara, K. Tochigi, Computer-aided reverse design for
ionic liquids by QSPR using descriptors of group contribution type for ionic conductivities
and viscosities, Fluid Phase Equilibria. 261 (2007) 434–443. doi:10.1016/j.fluid.2007.07.018.
9. I. Billard, G. Marcou, A. Ouadi, A. Varnek, In Silico Design of New Ionic Liquids Based
on Quantitative Structure−Property Relationship Models of Ionic Liquid Viscosity, The
Journal of Physical Chemistry B. 115 (2011) 93–98. doi:10.1021/jp107868w.
10. J. Forsman, C.E. Woodward, M. Trulsson, A Classical Density Functional Theory of Ionic
Liquids, The Journal of Physical Chemistry B. 115 (2011) 4606–4612.
doi:10.1021/jp111747w.
11. C.-W. Cho, U. Preiss, C. Jungnickel, S. Stolte, J. Arning, J. Ranke, A. Klamt, I. Krossing,
J. Thöming, Ionic Liquids: Predictions of Physicochemical Properties with Experimental

18
 ACCEPTED MANUSCRIPT

and/or DFT-Calculated LFER Parameters To Understand Molecular Interactions in Solution,

The Journal of Physical Chemistry B. 115 (2011) 6040–6050. doi:10.1021/jp200042f.
12. J.M. Slattery, C. Daguenet, P.J. Dyson, T.J.S. Schubert, I. Krossing, How to Predict the
Physical Properties of Ionic Liquids: A Volume-Based Approach, Angewandte Chemie
International Edition. 46 (2007) 5384–5388. doi:10.1002/anie.200700941.

T
13. D. Wileńska, I. Anusiewicz, S. Freza, M. Bobrowski, E. Laux, S. Uhl, H. Keppner, P.
Skurski, Predicting the viscosity and electrical conductivity of ionic liquids on the basis of

IP
theoretically calculated ionic volumes, Molecular Physics. 113 (2015) 630–639.
doi:10.1080/00268976.2014.964344.

R
14. E. Hawlicka, T. Dlugoborski, Molecular dynamics simulations of the aqueous solution of
tetramethylammonium chloride, Chemical Physics Letters. 268 (1997) 325–330.

SC
doi:10.1016/S0009-2614(97)00229-7.
15. E.J. Maginn, Molecular simulation of ionic liquids: current status and future opportunities,
Journal of Physics: Condensed Matter. 21 (2009) 373101. doi:10.1088/0953-

NU
8984/21/37/373101.
16. S.U. Lee, J. Jung, Y.-K. Han, Molecular dynamics study of the ionic conductivity of 1-n-
butyl-3-methylimidazolium salts as ionic liquids, Chemical Physics Letters. 406 (2005) 332–
MA
340. doi:10.1016/j.cplett.2005.02.109.
17. C.E. Resende Prado, L.C. Gomide Freitas, Molecular dynamics simulation of the room-
temperature ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate, Journal of Molecular
Structure: THEOCHEM. 847 (2007) 93–100. doi:10.1016/j.theochem.2007.09.009.
18. T. Köddermann, D. Paschek, R. Ludwig, Molecular Dynamic Simulations of Ionic
D

Liquids: A Reliable Description of Structure, Thermodynamics and Dynamics,

TE

ChemPhysChem. 8 (2007) 2464–2470. doi:10.1002/cphc.200700552.

19. J. Picálek, B. Minofar, J. Kolafa, P. Jungwirth, Aqueous solutions of ionic liquids: study
of the solution/vapor interface using molecular dynamics simulations, Physical Chemistry
P

Chemical Physics. 10 (2008) 5765. doi:10.1039/b806205f.

20. Y. Zhang, E.J. Maginn, A comparison of methods for melting point calculation using
CE

molecular dynamics simulations, The Journal of Chemical Physics. 136 (2012) 144116.
doi:10.1063/1.3702587.
21. T.G.A. Youngs, C. Hardacre, Application of Static Charge Transfer within an Ionic-
AC

Liquid Force Field and Its Effect on Structure and Dynamics, ChemPhysChem. 9 (2008)
1548–1558. doi:10.1002/cphc.200800200.
22. M.H. Kowsari, S. Alavi, M. Ashrafizaadeh, B. Najafi, Molecular dynamics simulation of
imidazolium-based ionic liquids. I. Dynamics and diffusion coefficient, The Journal of
Chemical Physics. 129 (2008) 224508. doi:10.1063/1.3035978.
23. M.H. Kowsari, S. Alavi, M. Ashrafizaadeh, B. Najafi, Molecular dynamics simulation of
imidazolium-based ionic liquids. II. Transport coefficients, The Journal of Chemical Physics.
130 (2009) 014703. doi:10.1063/1.3042279.
24. S.A. Kislenko, I.S. Samoylov, R.H. Amirov, Molecular dynamics simulation of the
electrochemical interface between a graphite surface and the ionic liquid [BMIM][PF6],
Physical Chemistry Chemical Physics. 11 (2009) 5584. doi:10.1039/b823189c.
25. M. Klähn, G.S. Lim, P. Wu, How ion properties determine the stability of a lipase enzyme
in ionic liquids: A molecular dynamics study, Physical Chemistry Chemical Physics. 13
(2011) 18647. doi:10.1039/c1cp22056j.
26. A. Heintz, J.K. Lehmann, C. Wertz, Thermodynamic Properties of Mixtures Containing
Ionic Liquids. 3. Liquid−Liquid Equilibria of Binary Mixtures of 1-Ethyl-3-
methylimidazolium Bis(trifluoromethylsulfonyl)imide with Propan-1-ol, Butan-1-ol, and

19
 ACCEPTED MANUSCRIPT

Pentan-1-ol †, Journal of Chemical & Engineering Data. 48 (2003) 472–474.

doi:10.1021/je0201931.
27. I. Bandrés, B. Giner, H. Artigas, F.M. Royo, C. Lafuente, Thermophysic Comparative
Study of Two Isomeric Pyridinium-Based Ionic Liquids, The Journal of Physical Chemistry
B. 112 (2008) 3077–3084. doi:10.1021/jp077259p.

T
28. L.G. Sánchez, J.R. Espel, F. Onink, G.W. Meindersma, A.B. de Haan, Density, Viscosity,
and Surface Tension of Synthesis Grade Imidazolium, Pyridinium, and Pyrrolidinium Based

IP
Room Temperature Ionic Liquids, Journal of Chemical & Engineering Data. 54 (2009) 2803–
2812. doi:10.1021/je800710p.

R
29. J. Safarov, I. Kul, W.A. El-Awady, J. Nocke, A. Shahverdiyev, E. Hassel,
Thermophysical properties of 1-butyl-4-methylpyridinium tetrafluoroborate, The Journal of

SC
Chemical Thermodynamics. 51 (2012) 82–87. doi:10.1016/j.jct.2012.02.018.
30. M. García-Mardones, A. Barrós, I. Bandrés, H. Artigas, C. Lafuente, Thermodynamic
properties of binary mixtures combining two pyridinium-based ionic liquids and two alkanols,

NU
The Journal of Chemical Thermodynamics. 51 (2012) 17–24. doi:10.1016/j.jct.2012.02.023.
31. H. Guerrero, M. García-Mardones, P. Cea, C. Lafuente, I. Bandrés, Correlation of the
volumetric behaviour of pyridinium-based ionic liquids with two different equations,
MA
Thermochimica Acta. 531 (2012) 21–27. doi:10.1016/j.tca.2011.12.020.
32. M. García-Mardones, I. Gascón, M.C. López, F.M. Royo, C. Lafuente, Viscosimetric
Study of Binary Mixtures Containing Pyridinium-Based Ionic Liquids and Alkanols, Journal
of Chemical & Engineering Data. 57 (2012) 3549–3556. doi:10.1021/je300557g.
33. T. Nonthanasin, A. Henni, C. Saiwan, Densities and low pressure solubilities of carbon
D

dioxide in five promising ionic liquids, RSC Advances. 4 (2014) 7566.

TE

doi:10.1039/c3ra46339g.
34. J. Ortega, R. Vreekamp, E. Penco, E. Marrero, Mixing thermodynamic properties of 1-
butyl-4-methylpyridinium tetrafluoroborate [b4mpy][BF4] with water and with an alkan-1ol
P

(methanol to pentanol), The Journal of Chemical Thermodynamics. 40 (2008) 1087–1094.

doi:10.1016/j.jct.2008.02.019.
CE

35. M.B. Shiflett, B.A. Elliott, A. Yokozeki, Phase behavior of vinyl fluoride in room-
temperature ionic liquids [emim][Tf2N], [bmim][N(CN)2], [bmpy][BF4], [bmim][HFPS] and
[omim][TFES], Fluid Phase Equilibria. 316 (2012) 147–155. doi:10.1016/j.fluid.2011.11.030.
AC

36. M. García-Mardones, H.M. Osorio, C. Lafuente, I. Gascón, Ionic Conductivities of Binary

Mixtures Containing Pyridinium-Based Ionic Liquids and Alkanols, Journal of Chemical &
Engineering Data. 58 (2013) 1613–1620. doi:10.1021/je301347v.
37. Y. Li, M. Zhang, Q. Liu, H. Su, Phase behaviour for aqueous two-phase systems
containing the ionic liquid 1-butylpyridinium tetrafluoroborate/1-butyl-4-methylpyridinium
tetrafluoroborate and organic salts (sodium tartrate/ammonium citrate/trisodium citrate) at
different temperatures, The Journal of Chemical Thermodynamics. 66 (2013) 80–87.
doi:10.1016/j.jct.2013.06.011.
38. S.I. Abu-Eishah, A.M. Dowaidar, Liquid−Liquid Equilibrium of Ternary Systems of
Cyclohexane + (Benzene, + Toluene, + Ethylbenzene, or + o -Xylene) + 4-Methyl- N -butyl
Pyridinium Tetrafluoroborate Ionic Liquid at 303.15 K, Journal of Chemical & Engineering
Data. 53 (2008) 1708–1712. doi:10.1021/je7007268.
39. S.H. Schöfer, N. Kaftzik, U. Kragl, P. Wasserscheid, Enzyme catalysis in ionic liquids:
lipase catalysed kinetic resolution of 1-phenylethanol with improved enantioselectivity,
Chemical Communications. (2001) 425–426. doi:10.1039/b009389k.
40. G. Inoue, Y. Iwai, M. Yasutake, K. Honda, Y. Arai, Measurement of infinite dilution
activity coefficients of n-alkanes in 4-methyl-n-butylpyridinium tetrafluoroborate using gas–

20
 ACCEPTED MANUSCRIPT

liquid chromatography, Fluid Phase Equilibria. 251 (2007) 17–23.

doi:10.1016/j.fluid.2006.10.016.
41. K.H.A.E. Alkhaldi, M.S. Al-Tuwaim, M.S. Fandary, A.S. Al-Jimaz, Separation of
propylbenzene and n-alkanes from their mixtures using 4-methyl-N-butylpyridinium
tetrafluoroborate as an ionic solvent at several temperatures, Fluid Phase Equilibria. 309

T
(2011) 102–107. doi:10.1016/j.fluid.2011.06.036.
42. M. Scendo, J. Uznanska, Inhibition Effect of 1-Butyl-4-Methylpyridinium

IP
Tetrafluoroborate on the Corrosion of Copper in Phosphate Solutions, International Journal of
Corrosion. 2011 (2011) 1–12. doi:10.1155/2011/761418.

R
43. A. Latała, M. Nędzi, P. Stepnowski, Toxicity of imidazolium and pyridinium based ionic
liquids towards algae. Chlorella vulgaris, Oocystis submarina (green algae) and Cyclotella

SC
meneghiniana, Skeletonema marinoi (diatoms), Green Chemistry. 11 (2009) 580.
doi:10.1039/b821140j.
44. N. Papaiconomou, J. Estager, Y. Traore, P. Bauduin, C. Bas, S. Legeai, S. Viboud, M.

NU
Draye, Synthesis, Physicochemical Properties, and Toxicity Data of New Hydrophobic Ionic
Liquids Containing Dimethylpyridinium and Trimethylpyridinium Cations †, Journal of
Chemical & Engineering Data. 55 (2010) 1971–1979. doi:10.1021/je9009283.
MA
45. K. Sakizadeh, L.P. Olson, P.J. Cowdery-Corvan, T. Ishida, D.R. Whitcomb, Thermographic
materials containing ionic liquids. (2007) US Patent No.: US 7,163,786 B1.
46. I. Bandrés, G. Pera, S. Martín, M. Castro, C. Lafuente, Thermophysical Study of 1-Butyl-
2-Methylpyridinium Tetrafluoroborate Ionic Liquid, The Journal of Physical Chemistry B.
113 (2009) 11936–11942. doi:10.1021/jp906133t.
D

47. A. Navas, J. Ortega, J. Palomar, C. Díaz, R. Vreekamp, COSMO-RS analysis on mixing

TE

properties obtained for the systems 1-butyl-X-methylpyridinium tetrafluoroborate [X = 2,3,4]

and 1,ω-dibromoalkanes [ω = 1–6], Physical Chemistry Chemical Physics. 13 (2011) 7751.
doi:10.1039/c0cp02169e.
P

48. S. Ardizzone, C.L. Bianchi, P. Drago, D. Mozzanica, P. Quagliotto, P. Savarino, G.

Viscardi, Adsorption of 1-alkyl-4-methylpyridinium salts at solid-liquid and water-air
CE

interfaces, Colloids and Surfaces A: Physicochemical and Engineering Aspects. 113 (1996)
135–144. doi:10.1016/0927-7757(96)03632-1.
49. T. Heinze, K. Schwikal, S. Barthel, Ionic Liquids as Reaction Medium in Cellulose
AC

Functionalization, Macromolecular Bioscience. 5 (2005) 520–525.

doi:10.1002/mabi.200500039.
50. A.B. Pereiro, A. Rodriguez, M. Blesic, K. Shimizu, J.N. Canongia Lopes, L.P.N. Rebelo,
Mixtures of Pyridine and Nicotine with Pyridinium-Based Ionic Liquids, Journal of Chemical
& Engineering Data. 56 (2011) 4356–4363. doi:10.1021/je2001446.
51. E.S. Sashina, D.A. Kashirskii, G. Janowska, M. Zaborski, Thermal properties of 1-alkyl-
3-methylpyridinium halide-based ionic liquids, Thermochimica Acta. 568 (2013) 185–188.
doi:10.1016/j.tca.2013.06.022.
52. B. Ellis, W. Keim, P. Wasserscheid, Linear dimerisation of but-1-ene in biphasic mode
using buffered chloroaluminate ionic liquid solvents, Chemical Communications. (1999) 337–
338. doi:10.1039/a808346k.
53. D.A. Pearlman, D.A. Case, J.W. Caldwell, W.S. Ross, T.E. Cheatham, S. DeBolt, D.
Ferguson, G. Seibel, P. Kollman, AMBER, a package of computer programs for applying
molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations
to simulate the structural and energetic properties of molecules, Computer Physics
Communications. 91 (1995) 1–41. doi:10.1016/0010-4655(95)00041-D.

21
 ACCEPTED MANUSCRIPT

54. D.A. Case, T.E. Cheatham, T. Darden, H. Gohlke, R. Luo, K.M. Merz, A. Onufriev, C.
Simmerling, B. Wang, R.J. Woods, The Amber biomolecular simulation programs, Journal of
Computational Chemistry. 26 (2005) 1668–1688. doi:10.1002/jcc.20290.
55. X. Wu, Z. Liu, S. Huang, W. Wang, Molecular dynamics simulation of room-temperature
ionic liquid mixture of [bmim][BF4] and acetonitrile by a refined force field, Physical

T
Chemistry Chemical Physics. 7 (2005) 2771. doi:10.1039/b504681p.
56. K.G. Sprenger, V.W. Jaeger, J. Pfaendtner, The General AMBER Force Field (GAFF)

IP
Can Accurately Predict Thermodynamic and Transport Properties of Many Ionic Liquids, The
Journal of Physical Chemistry B. 119 (2015) 5882–5895. doi:10.1021/acs.jpcb.5b00689.

R
57. U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, A smooth
particle mesh Ewald method, The Journal of Chemical Physics. 103 (1995) 8577.

SC
doi:10.1063/1.470117.
58. Williams, T.; Kelley, C. Gnuplot 4.5: an interactive plotting program. (2011), URL
http://gnuplot.info.

NU
59. Origin (OriginLab, Northampton, MA), URL http://www.originlab.com
60. M. Brehm, B. Kirchner, TRAVIS - A Free Analyzer and Visualizer for Monte Carlo and
Molecular Dynamics Trajectories, Journal of Chemical Information and Modeling. 51 (2011)
MA
2007–2023. doi:10.1021/ci200217w.
61. W. Humphrey, A. Dalke, K. Schulten, VMD: Visual molecular dynamics, Journal of
Molecular Graphics. 14 (1996) 33–38. doi:10.1016/0263-7855(96)00018-5.
62. M.G. Del Pópolo, R.M. Lynden-Bell, J. Kohanoff, Ab Initio Molecular Dynamics
Simulation of a Room Temperature Ionic Liquid, The Journal of Physical Chemistry B. 109
D

(2005) 5895–5902. doi:10.1021/jp044414g.

TE

63. A. Einstein, Über die von der molekularkinetischen Theorie der Wärme geforderte
Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen. Annalen der Physik 17
(1905), 549–560. doi: 10.1002/andp.19053220806
P

64. M.-M. Huang, S. Bulut, I. Krossing, H. Weingärtner, Communication: Are hydrodynamic

models suited for describing the reorientational dynamics of ions in ionic liquids? A case
CE

study of methylimidazolium tetra(hexafluoroisopropoxy)aluminates, The Journal of Chemical

Physics. 133 (2010) 101101. doi:10.1063/1.3478221.
65. A.W. Taylor, P. Licence, A.P. Abbott, Non-classical diffusion in ionic liquids, Physical
AC

Chemistry Chemical Physics. 13 (2011) 10147. doi:10.1039/c1cp20373h.

66. D.M. Eike, E.J. Maginn, Atomistic simulation of solid-liquid coexistence for molecular
systems: Application to triazole and benzene, The Journal of Chemical Physics. 124 (2006)
164503. doi:10.1063/1.2188400.
67. Lindemann, F.A. The calculation of molecular vibration frequencies. Physik Z. 11 (1910),
609-612.
68. U. Buchenau, R. Zorn, M.A. Ramos, Probing cooperative liquid dynamics with the mean
square displacement, Physical Review E. 90 (2014). doi:10.1103/PhysRevE.90.042312.
69. K. Niss, C. Dalle-Ferrier, B. Frick, D. Russo, J. Dyre, C. Alba-Simionesco, Connection
between slow and fast dynamics of molecular liquids around the glass transition, Physical
Review E. 82 (2010). doi:10.1103/PhysRevE.82.021508.
70. L. Larini, A. Ottochian, C. De Michele, D. Leporini, Universal scaling between structural
relaxation and vibrational dynamics in glass-forming liquids and polymers, Nature Physics. 4
(2008) 42–45. doi:10.1038/nphys788.
71. A. Hunt, A simple connection between the melting temperature and the glass temperature
in a kinetic theory of the glass transition, Journal of Physics: Condensed Matter. 4 (1992)
L429–L431. doi:10.1088/0953-8984/4/32/001.

22
 ACCEPTED MANUSCRIPT

72. M. Freemantle, An Introduction to ionic liquids, RSC Pub, Cambridge, UK, 2010, (4), 31-
40. ISBN: 978-1-84755-161-0.
73. I. Bandrés, D.F. Montaño, I. Gascón, P. Cea, C. Lafuente, Study of the conductivity
behavior of pyridinium-based ionic liquids, Electrochimica Acta. 55 (2010) 2252–2257.
doi:10.1016/j.electacta.2009.11.073.

T
74. P.M. Agrawal, B.M. Rice, D.L. Thompson, Molecular dynamics study of the melting of
nitromethane, The Journal of Chemical Physics. 119 (2003) 9617. doi:10.1063/1.1612915.

IP
75. J. Picálek, J. Kolafa, Molecular dynamics study of conductivity of ionic liquids: The
Kohlrausch law, Journal of Molecular Liquids. 134 (2007) 29–33.

R
doi:10.1016/j.molliq.2006.12.015.

SC
NU
MA
D
P TE
CE
AC

23