See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323475152

Systematic comparison of force fields for molecular dynamic simulation of

Au(111)/Ionic liquid interfaces

Article  in  Fluid Phase Equilibria · May 2018

DOI: 10.1016/j.fluid.2018.01.024

CITATIONS READS

16 273

4 authors, including:

Runxi Wang Sheng Bi

Huazhong University of Science and Technology Huazhong University of Science and Technology
20 PUBLICATIONS   297 CITATIONS    31 PUBLICATIONS   436 CITATIONS   

SEE PROFILE SEE PROFILE

Guang Feng
Huazhong University of Science and Technology
132 PUBLICATIONS   4,399 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

CO2 EOR View project

Electrical Double Layer View project

All content following this page was uploaded by Sheng Bi on 03 December 2018.

The user has requested enhancement of the downloaded file.

 Fluid Phase Equilibria 463 (2018) 106e113

Contents lists available at ScienceDirect

Fluid Phase Equilibria

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / fl u i d

Systematic comparison of force fields for molecular dynamic

simulation of Au(111)/Ionic liquid interfaces
Runxi Wang a, b, Sheng Bi a, b, Volker Presser c, d, Guang Feng a, b, *
a
State Key Laboratory of Coal Combustion, School of Energy and Power Engineering, Huazhong University of Science and Technology (HUST), Wuhan
430074, China
b
Nano Interface Centre for Energy, School of Energy and Power Engineering, Huazhong University of Science and Technology, 430074, China
c
INM e Leibniz Institute for New Materials, 66123 Saarbrücken, Germany
d
Saarland University, 66123 Saarbrücken, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Selecting the most suitable force field is a key to meaningful molecular dynamics (MD) simulation. To
Received 9 October 2017 select the appropriate gold force field to model the Au(111)/ionic liquid interface, a systematic com-
Received in revised form parison of four different widely used force fields of gold and a typical carbon force field has been studied
2 January 2018
by MD simulations with constant potential method. We calculated the ion adsorption behavior and
Accepted 22 January 2018
Available online 21 February 2018
differential capacitance of interfaces between the gold electrode and ionic liquid 1-butyl-1-
methylpyrrolidinium bis(trifluoromethylsulfonyl)imide ([PYR][TFSI]) in comparison with the experi-
mental results and showed the effects on the solid-liquid interfaces from the van der Waals interaction,
Keywords:
Au(111)/ionic liquid interface
image force effect and cumulative ions. Based on the comparison between the results of simulations and
Force fields experiments, we recommend two types of force fields to properly model the Au(111)/ionic liquid
Ion adsorption interfaces.
Differential capacitance © 2018 Published by Elsevier B.V.
Molecular dynamics

1. Introduction multilayered structure at electrified gold/RTILs interface and ion

Room-temperature ionic liquids (RTILs) are promising electro-
lytes because of their remarkable advantages, such as low volatility,
wide electrochemical windows, and designable properties [1e3].
Beyond that, they have been used in various applications including
energy storage, electrocatalysis, electrowetting, and electronic
gating [4e7]. The interfacial structures of RTILs at solid-liquid in-
terfaces determine the performance of such applications. There-
fore, many studies have investigated fluid-solid interfaces with ILs.
Among these, gold electrodes are commonly employed due to
gold's high electrical conductivity, chemical stability, and compat-
ibility with conventional microfabrication procedures [8,9]. Rele-
vant experimental studies have been carried out to investigate
gold/RTILs interfaces in order to reveal the mechanism of the
electric double layer (EDL) formation. For example, Atkin et al. [10]
used atomic force microscopy (AFM) to reveal that there is a
* Corresponding author. State Key Laboratory of Coal Combustion, School of
Energy and Power Engineering, Huazhong University of Science and Technology
(HUST), Wuhan 430074, China.
E-mail address: gfeng@hust.edu.cn (G. Feng).
arrangements change as a function of potential, which suggested
the RTILs interface can be shaped by the electrode potential and
ions structure. Later, Mao et al. [11] obtained high quality AFM force
curves to prove the existence of charged interior layers and neutral
exterior layers. Gebbie et al. [12] employed a surface force appa-
ratus (SFA) to investigate the electric screen of ionic liquids at
charged gold and mica surfaces. Their results showed that both
Stern and diffuse layers are formed to screen the charged surfaces
and the diffuse double layer consists of dissociated cations and
anions. In addition, neutron or X-ray reflectometry and surface
tunneling microscopy are further methods found in studies of solid/
RTILs interfaces [13e17].
In addition to experimental research, there are many theoretical
works and simulations investigating gold/RTIL interfaces. Using
molecular dynamics (MD) simulations, Sha et al. [18] reported the
change of ion structure on gold electrode surface for different
applied voltages. They also calculated the interaction energy of
gold-ions and binding energy of cation-anion, which are difficult to
extract directly from experimental data. Ferreira et al. [19] used MD
simulations to investigate the orientation of ions on the solid sur-
face and found the orientations of cations are more dependent on
the interface instead of the anions. Furthermore, Nikitina et al. [20]

https://doi.org/10.1016/j.fluid.2018.01.024
0378-3812/© 2018 Published by Elsevier B.V.
 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113
combined MD simulations and quantum mechanical theory to
study the electron transfer reaction at the gold/RTIL interface. They
found a higher energy barrier that impedes the charged reactants
from approaching the gold surface, which influences the electron
transfer. In these simulations, the accuracy of the force fields plays a
crucial role in determining the amount of ions adsorption, molec-
ular orientation, energy barrier, and so on. However, so far most MD
studies used different gold force fields to investigate the interface of
gold/RTILs and the validation of the applicability of the force fields
is usually ignored before modeling. Meanwhile, there are several
kinds of force fields developed for gold, thus it is necessary to make
systematic comparisons between them to reveal the suitable one
for modeling gold/RTILs interfaces.
In the present work, we used MD simulations to compare four
different types of often-used force fields of gold with a typical
carbon force field systematically, using the ionic liquid 1-butyl-1-
methylpyrrolidinium bis(trifluoromethylsulfonyl)imide ([PYR]
[TFSI]) as the electrolyte. We investigated the electro-adsorption
behavior and differential capacitance of Au(111)/RTILs interface
and attributed the interfacial phenomena to van der Waals inter-
action, image force effects, and cumulative ions.
2. Simulation details
2.1. Simulation methods
The simulation unit cell is shown in Fig. 1. Ionic liquids were
enclosed between two electrodes with a size of 5.488
5.488 nm2.
Electrodes are modeled as Au (111) single-crystal surfaces and
graphene sheets. The distance between positive electrode and
negative electrode is 8.232 nm and the electrodes are fixed during
the simulation, which should be wide enough to ensure that the
RTILs in the central portion of the system are bulk-like (see Fig. S1).
The expectation that [PYR][TFSI] in the central portion of the sys-
tem are bulk-like is corroborated by the observation that, beyond
positions ~3.0 nm from the electrode surface, the ionic space charge
density is practically zero (i.e., both cations and anions are
distributed homogeneously) and the electrical potential profile is
flat (see Fig. S2 and Fig. S3). Detailed number of the ionic liquids
and electrode atoms are gathered in Table S1 for all studied sys-
tems. The equal potential was maintained on the electrode surfaces
and located at a plane across the center of surface atoms. This
method was able to take the electronic polarizability of the
Fig. 1. a) The schematic configuration of the simulation unit cell. b) All-atoms molecular structure of [PYR][TFSI].
107
electrode surfaces into account, for instance, a local redistribution
of charges on the electrode surface due to interaction with RTILs.
This approach has been verified by previous work [21,22], and it
was shown that this method is much more realistic comparing to
assigning a uniform and constant partial on each electrode atom
when simulating supercapacitors based on RTILs. The number of
ion pairs was tuned to guarantee the electrolyte density in the
center of channel was in accord with the bulk. Long-range Coulomb
interactions were quantified by using the particle mesh Ewald
(PME) method [23] with a correlation for slab-geometry [24]. An
FFT grid spacing of 0:1143_ nm and cubic interpolation for charge
distribution were used to compute the electrostatic interactions in
the reciprocal space. A cutoff length of 1.2 nm was used in the
calculation of electrostatic interactions in the real space. The non-
electrostatic interactions were computed by direct summation
with a cutoff length of 1.2 nm. LJ interactions between different
atoms were calculated using Lorentz-Berthelot mixing rules. Sim-
ulations were performed in the NVT ensemble using a customized
MD code Gromacs [25], with a time step of 2 fs. The ensemble
temperature was maintained at 338 K using the Nose-Hoover
thermostat. All configurations were first run by annealing from
800 K to 338 K for 20 ns, and then an 80 ns production run was
performed for analysis.
2.2. Force fields
For the gold electrode surface, we choose four highly cited gold
force fields that all used in modeling gold electrode. In this work,
the form of the force fields adopts 12-6 Lennard-Jones potential,
where rij is the distance and εij is the potential well describing the
interaction intensity between atom i and atom j.
2 !12 !6 3


4
sij sij 5
VLJ rij ¼ 4εij 
rij rij
2.2.1. Face-centered cubic metals (FCCM) force field
The face-centered cubic metals force field [26] was parameter-
ized for simulating the surface and interface of face-centered cubic
metals, such as Au, Ag, Al, and Cu. It was developed to simulate the
interface between metals and organic, inorganic, and biological
108 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113
mixtures. The Lennard-Jones parameters were obtained by fitting
the density of metal, metal-vapor surface tension, metal-water
interface tension, and elastic moduli to experimental results un-
der ambient conditions. This force field is compatible with many
universal force fields (OPLS-AA, AMBER, CHARMM, COMPASS, CVFF,
and PCFF) to mimic the interface properties with water, bio-
polymers, organic molecules and inorganic components. Most of
the existing work used this force field in metal/biopolymer inter-
face [27,28] to investigate molecular interactions, peptide adsorp-
tion and so on. Only a few studies have studied the EDL structure,
surface interaction, and capacitance when using gold as electrode
[18,19,29].
2.2.2. Force field developed by Halicioǧlu and Pound (HPFF)
Halicioǧlu and Pound developed a simple method to calculate
12-6 Lennard-Jones potential parameters for metal, using crystal-
line state physical properties (such as cohesive energy, compress-
ibility, lattice constants, heat capacity, etc.) [30]. They found that
the potential parameters are mainly dependent on DH (DH¼Hv-Hc,
Hv, and Hc denote the total enthalpies of the vapor and the crys-
talline states, respectively) and the volume of the crystal. Although
this force field is derived from experimental data of metal, several
previous works have used HPFF to investigate the diffusion of
adsorbed hydrocarbon species [31e33], and the dynamic behavior
of atomic oxygen on metal surfaces [34]. So far, HPFF has been in-
tegrated into consistent Valence force field (CVFF) [35], which is a
universal force field used by LAMPPS, Material Studio, etc.
2.2.3. Force field developed by Maranas et al. (MaFF)
The Force Field developed by Maranas et al. [36] utilizes the 12-6
Lennard-Jones potential to describe the interaction of peptides and
water with gold surface. The size parameter s was set depending on
the van der Waals radius of gold atom from the Universal Force
Field [37]. Then they tuned the value of ε to make sure that water
has a zero contact angle on gold surface. Finally, ε can be deter-
mined by comparing the adsorption energy with ab initio result.
Similarly, this force field was mainly used for proteins, amino acids,
and peptides, because most of such systems are water-based.
2.2.4. GolP force field
GolP is the abbreviation of Gold-Protein, which is designed for
the interaction of proteins and the Au(111) surface, but not limited
to simulations of proteins [38]. Different from the above mentioned
three force fields, this force field introduced two Lennard-Jones
virtual sites at the hollow site among gold atoms to drive the
adsorption of atoms to the correct site and added the image
interaction. To confirm the reliability of the force field, quantum
mechanical calculations of small molecule adsorption on gold
surface were compared with experimental data. The GolP force
field has been proved that it can model the experimental phe-
nomenon in a physiological environment, but the interaction with
ionic liquids still needs further exploration.
The force field of carbon was adapted from the work by Borodin
et al. [39], which is widely used in modeling carbon/RTILs in-
terfaces. The parameters of gold and carbon electrodes are listed in
Table 1. All studied force fields parameters are summarized in the
FF files provided in Supporting Information.
A refined force field developed by Chaban et al. [40] from the
force field of Lopes et al. [41,42] was used for [PYR][TFSI], in which
they scaled partial charges on the IL so that the total charges on
cation and anion are ± 0.8e, for the sake of refinement for more
realistic reproduction of transport properties. This force field was
verified to show that it can accurately provide the thermodynamic
properties such as density and heat of vaporization (see Table S2
and Table S3).
Table 1
Force fields parameters of electrodes.
s (nm) ε (kJ/mol)
FCCM 0.2629 22.1333
HPFF 0.2637 10.1800
MaFF 0.3293 4.3932
GolPa 0.3200 0.6500
Carbon 0.3400 0.3599
a
The LJ interaction is assigned to virtual sites in this force field. Because all
simulations are calculated by constant potential method, we do not consider the
original polarization in this force field.
3. Results and discussion
3.1. Electric double-layer structure
To observe different ion electro-adsorption phenomena at
Au(111)/[PYR][TFSI] interfaces when using four types of force fields,
we calculated the density profiles of [PYR]þ and [TFSI]- ions as a
function of distance from the electrode and the potential drop
across the EDL, shown in Figs. 2e3. The ion number density profiles
were calculated based on the location of the ion's center-of-mass as
obtained from simulations. According to the electro-adsorption
behavior, we divided the four force fields into two groups: strong
adsorption (FCCM and HPFF) and weak adsorption (MaFF and GolP).
For strong adsorption, more cations can be adsorbed in the first
layer at neutral and negative potential electrodes. The peak value of
first layer is close to 80 #/nm3 at 3 V at 0.34 nm from the electrode
surface, which is much higher than for the weak adsorption group.
At zero potential, the strong adsorption electrode also electro-
adsorbs more cations in the first layer. As potential becomes
increasingly positive when sweeping from 0 V, there is still a strong
adsorption of [PYR]þ. This observed strong adsorption of [PYR]þ at
the positive potential surface is in agreement with the experi-
mental phenomena deduced from neutron reflectometry [43]. In
line with observations for strong adsorption group, Lauw et al.
experimentally found a cation [PYR]þ rich interface at a positively
charged Au(111) electrode, which indicates a strong adsorption of
[PYR]þ on the gold electrode. The similar strong adsorption of RTIL
cations or anions on other metal surface has also been reported in
the literature [44,45]. However, the adsorption of [PYR]þ decreases
rapidly as the potential becomes more positive for the weak
adsorption group.
For clearly comparing the different adsorption behavior, we also
calculated the cation number density of [PYR][TFSI]/carbon surface,
shown in Fig. 4. The number density of [PYR]þ in the first layer near
the carbon electrode shows the same pattern as can be seen for the
weak adsorption group: the first layer has only one peak and the
peak value decreases significantly with the increase of positive
voltage. This finding implies that MaFF and GolP force fields may
not adequately reflect the strong adsorption of [PYR]þ at Au(111)
surface as revealed in experiment.
Fig. 3 shows the number density of [TFSI]- as a function of the
potential drop. The number of [TFSI]- in the first adsorbed layer
increases with the potential well (ε) of the force field. All four gold
force fields and carbon (Fig. 4b) present the same pattern, showing
that anions adsorb at positive electrode and are rejected from the
negative electrode. This clearly shows that the interaction between
[TFSI]- and the electrode is weaker than for [PYR]þ. In addition,
specific adsorption of [TFSI]- on the gold electrode is not observed
in our simulations (in contrast to [PYR]þ).
3.2. Van der Waals interaction
To understand the strong adsorption of [PYR]þ on the surface of
 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113 109

Fig. 2. The number density of [PYR]þ as a function distance from the electrode surface for a series of potential across EDL (blue and red in floor color correspond to negative and
positive potential respectively). a) FCCM (s ¼ 0.2629 nm, ε ¼ 22.1333 kJ/mol), b) HPFF (s ¼ 0.2637 nm, ε ¼ 10.1800 kJ/mol), c) MaFF (s ¼ 0.3293 nm, ε ¼ 4.3932 kJ/mol), d) GolP
(s ¼ 0.3200 nm, ε ¼ 0.6500 kJ/mol). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3. The number density of [TFSI]- as a function distance from the electrode surface for a series of potential across EDL (blue and red in floor color correspond to negative and
positive potential, respectively). a) FCCM (s ¼ 0.2629 nm, ε ¼ 22.1333 kJ/mol), b) HPFF (s ¼ 0.2637 nm, ε ¼ 10.1800 kJ/mol), c) MaFF (s ¼ 0.3293 nm, ε ¼ 4.3932 kJ/mol), d) GolP
(s ¼ 0.3200 nm, ε ¼ 0.6500 kJ/mol). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

the gold electrode, we first investigated the van der Waals inter-
action. As shown in Fig. 5a, the Lennard-Jones (LJ) potential be-
tween the electrode and the cation [PYR]þ has been calculated for
all five force fields, including the one for carbon. There are two
distinct observations: (1) the potential well of the strong adsorp-
tion group is lower than for weak adsorption group, which suggests
the strong adsorption group could give stronger interactions with
cations; (2) the location of the potential well of the strong
adsorption group is smaller than for the weak adsorption group;
therefore, the first layer of cation is in closer proximity to the sur-
face. To eliminate the effect of the applied voltage (note that we do
consider the image force effect due to the electronic polarizability
110 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113

Fig. 4. The ions number density as a function distance from the carbon electrode surface for a series of potential across EDL (blue and red in floor color correspond to negative and
positive potential, respectively). a) [PYR]þ distribution, b) [TFSI]- distribution. (For interpretation of the references to color in this figure legend, the reader is referred to the Web
version of this article.)

Fig. 5. a) The LJ potential between electrode and carbon atom of [PYR]þ. b) The number density of [PYR]þ at applied potential of 0 V. To facilitate reading, the parameters of the force
fields are as follows. FCCM (s ¼ 0.2629 nm, ε ¼ 22.1333 kJ/mol), HPFF (s ¼ 0.2637 nm, ε ¼ 10.1800 kJ/mol), MaFF (s ¼ 0.3293 nm, ε ¼ 4.3932 kJ/mol), GolP (s ¼ 0.3200 nm,
ε ¼ 0.6500 kJ/mol), Carbon (s ¼ 0.3400 nm, ε ¼ 0.3599 kJ/mol).

of the electrode surfaces), we compared the cation number density
of the five force fields (including carbon) at 0 V in Fig. 5b. Consistent
with the previous data, more cations are electro-adsorbed at the
electrode surface and the location of first peak is closer to the
electrode surface. For the same reason, the strong adsorption group
yields in the simulation for gold surfaces more electro-adsorbed
anions. Therefore, the LJ potential mainly determines the adsorp-
tion behavior of ions, where the larger value of ε could drive more
ions to be adsorbed on the gold surface; the value of s relates the
location of the first ion layer.
3.3. Image force effect
Without applied voltage, the ion adsorption behavior is not only
determined by the LJ potential but also affected by the image forces
between ions and the electrode. To show the image force effect
clearly, the 2D surface charge distribution has been illustrated in
Fig. 6 (note that in MD modeling work with constant charge
method, there would be no results like such 2D surface charge
distribution, since the constant charge method assume the charge
on surface is zero everywhere at zero applied potential). As can be
seen, the charge distribution pattern changes when using different
force fields. In Fig. 6aeb, most of the area is deep red or blue, which
suggests that the ions near the surface induced opposite charge.
Comparatively, the area and value of the induced charge are smaller
in Fig. 6c and the surface charge distribution in Fig. 6d shows that
only few charges have been induced by the ionic liquids. This
change of the surface charge distribution is mainly due to the
different adsorption of ionic liquids in the first layer. As mentioned
before, a strong adsorption electrode electro-adsorbs more ions
and attract them closer to surface; thereby, stronger opposite
charges are induced which results in lager image forces to electro-
adsorb more ions in the first ionic layer.
The specific adsorption of cations can be attributed to the
combined effect of van der Waals interaction and the image force
effect. Van der Waals interaction plays a major role in the adsorp-
tion behavior: a larger value of ε and smaller value of s cause that
more ions are electro-adsorbed closer to the electrode surface. The
strong adsorption would induce more image charge at electrode
surface, which makes the interaction between ions and electrode
even stronger. Although this adsorption phenomenon was pre-
dicted in many theoretical studies, the image force is mostly
ignored in most previous simulation studies due to the heavy
computational cost. The deep effect of image forces still requires
further investigations in the future works. Interestingly, we found
that there is no specific adsorption of anions, probably because of
the geometry of anion different from that of cation: the cation
possesses a ring structure and a long alkyl chain, which would
facilitate the cation closely to pack on the flat gold surface and then
generate a high attraction to the electrode.
 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113 111

Fig. 6. 2D charge density distribution on gold surface at zero applied potential (unit: C/m2). a) FCCM (s ¼ 0.2629 nm, ε ¼ 22.1333 kJ/mol), b) HPFF (s ¼ 0.2637 nm, ε ¼ 10.1800 kJ/
mol), c) MaFF (s ¼ 0.3293 nm, ε ¼ 4.3932 kJ/mol), d) GolP (s ¼ 0.3200 nm, ε ¼ 0.6500 kJ/mol).

3.4. Differential capacitance
For better comparison, we also show the differential capacitance
of carbon in Fig. 7a. The differential capacitance (DC) curves in
Fig. 7a were obtained from a local fit of surface charges and the EDL
potential drop (see Supporting Information for details). The dif-
ferential capacitance curves obtained by using FCCM and HPFF
force fields have a characteristic camel shape and when using MaFF,
GolP, or carbon, we see a curve with bell shape. Lauw and Lockett
et al. [43,46] measured the differential capacitance of [PYR][TFSI] at
Au(111) surface by impedance spectroscopy and confirmed for this
system a camel-like shaped curve. In the latter work, the relatively
large ion sizes and nonhomogeneous ion polarization at the inter-
face were correlated with the camel shaped differential capacitance
curve. The camel-shaped differential capacitance has been theo-
retically predicted in the literature [47,48].
Overall, the differential capacitance curves of strong adsorption
electrodes have a larger amplitude than that of weak adsorption
group and carbon; however, the peak value of weak adsorption
group and carbon are similar. As reported in the literature [49,50],
the capacitance observed for gold electrodes is larger compared to
carbon surfaces when using the same ionic liquid. This aligns with
the issue that the weak adsorption group force field could not
precisely reproduce the differences between gold and carbon
electrodes. Previous studies [51] showed that the differential
capacitance is mainly determined by the electro-adsorption of ionic
liquids; therefore, we have calculated the cumulative number of
ions for all four different gold force fields in first EDLs layer as a
function of potential drop (Fig. 7b). Because of the strong adsorp-
tion, the change of cumulative cations of FCCM force field is less
obvious compared to other force fields. The increase of net charge
(i.e., the number of cations minuses that of anions) for FCCM from
0.5 V to 2.0 V is larger than that for weak adsorption group.
Beyond 0.5 V, the variation changes slightly for FCCM. Thus, the
differential capacitance curve of FCCM has larger amplitude than
what we found for the weak adsorption group at negative potential
and become less defined at positive potential. For the HPFF force
field, the variation of cations and anions are steeper than weak
adsorption group within the voltage range. Such a tendency in the
change of ions number leads to a higher capacitance of strong
adsorption group compared with that of weak adsorption group.
The results of strong adsorption group are qualitatively consistent
with experimental findings, which indicate that these force fields
are more suitable for modeling the realistic interaction between
ionic liquids and gold electrode surface. For a meaningful quanti-
tative comparison, the simulations and experiments should be
done in a way where conditions should be identical, such as tem-
perature, electrode structure, impurities and so on.
4. Conclusion
We systematically compared the accuracy of four different gold
112 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113
Fig. 7. a) Differential capacitance as a function of the EDL potential. b) The cumulative number of cations (solid lines) and anions (dashed lines). To facilitate reading, the parameters
of the force fields are as follows. FCCM (s ¼ 0.2629 nm, ε ¼ 22.1333 kJ/mol), HPFF (s ¼ 0.2637 nm, ε ¼ 10.1800 kJ/mol), MaFF (s ¼ 0.3293 nm, ε ¼ 4.3932 kJ/mol), GolP (s ¼ 0.3200 nm,
ε ¼ 0.6500 kJ/mol), Carbon (s ¼ 0.3400 nm, ε ¼ 0.3599 kJ/mol).
force fields when modeling Au(111)/RTILs interfaces at various
electric potentials with MD simulations. The constant potential
condition was imposed at the locations of all electrode atoms by
allowing the fluctuations of charges on the electrode surface. The
specific adsorption of cations was found at the strong adsorption
group of force fields even when the potential becomes positive,
which is in agreement with previous experimental work [43]. This
specific adsorption results from stronger van der Waals interaction
and the image force effect. Moreover, MD simulations reveal that
FCCM and HPFF can result in a camel-shape of the calculated dif-
ferential capacitance pattern of Au/[PYR][TFSI] interfaces, while
MaFF and HPFF yield a bell shape. These curve shapes are mainly
due to the different adsorption behaviors of ions at electrode sur-
faces. Further observations show that the strong adsorption group
of force fields is more suitable for modeling the realistic interaction
between ionic liquids and gold electrode surface. However, the
strongest adsorption of FCCM force field may complicate the
simulation because more ions are crowed in the limited interface
space, which leads to declining diffusion of ions so that it requires
much more time to reach the equilibrium and longer trajectories
are needed for statistical analysis. Weak adsorption group of force
fields (MaFF and HPFF) are more suitable for simulating the inter-
face between biological molecules and gold in aqueous solution,
since both force fields considered extra interaction with water
during the development process. MaFF used a modified TIP3P
model [52] and GolP added extra van der Waals interaction be-
tween gold and water molecules. Since water molecules occupy
most area of gold surface, weak adsorption group force fields are
preferred for biological simulation system; however, they may not
be suitable for Au(111)/RTILs interfaces. This study provides
important information for researchers to choose suitable force
fields from existing gold force fields to study the gold/RTILs
interface.
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (51406060) and the Fundamental Research
Funds for the Central Universities (2016YXZD006). V.P. thanks Prof.
Eduard Arzt (INM) for his continued support. All the computations
were carried out at the National Supercomputing Center in Tianjin
(Tianhe-1A) and Guangzhou (Tianhe II).
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.fluid.2018.01.024.
References
[1] M.V. Fedorov, A.A. Kornyshev, Ionic liquids at electrified interfaces, Chem. Rev.
114 (2014) 2978e3036.
[2] H. Niedermeyer, J.P. Hallett, I.J. Villar-Garcia, P.A. Hunt, T. Welton, Mixtures of
ionic liquids, Chem. Soc. Rev. 41 (2012) 7780e7802.
[3] M. Galin  ski, A. Lewandowski, I. Ste˛ pniak, Ionic liquids as electrolytes, Elec-
trochim. Acta 51 (2006) 5567e5580.
[4] F. Endres, Ionic liquids: solvents for the electrodeposition of metals and
semiconductors, ChemPhysChem: Eur. J. Chem. Phys. Phys. Chem. 3 (2002)
144e154.
[5] S. Millefiorini, A.H. Tkaczyk, R. Sedev, J. Efthimiadis, J. Ralston, Electrowetting
of ionic liquids, J. Am. Chem. Soc. 128 (2006) 3098e3101.
[6] F. Chen, Q. Qing, J. Xia, J. Li, N. Tao, Electrochemical gate-controlled charge
transport in graphene in ionic liquid and aqueous solution, J. Am. Chem. Soc.
131 (2009) 9908e9909.
[7] A. Balducci, U. Bardi, S. Caporali, M. Mastragostino, F. Soavi, Ionic liquids for
hybrid supercapacitors, Electrochem. Commun. 6 (2004) 566e570.
[8] M.M. Collinson, Nanoporous gold electrodes and their applications in
analytical chemistry, ISRN Anal. Chem. 2013 (2013).
[9] R. Zhang, H. Olin, Porous gold filmsda short review on recent progress, Ma-
terials 7 (2014) 3834e3854.
[10] R. Hayes, N. Borisenko, M.K. Tam, P.C. Howlett, F. Endres, R. Atkin, Double
layer structure of ionic liquids at the Au(111) electrode interface: an atomic
force microscopy investigation, J. Phys. Chem. C 115 (2011) 6855e6863.
[11] X. Zhang, Y.X. Zhong, J.W. Yan, Y.Z. Su, M. Zhang, B.W. Mao, Probing double
layer structures of Au (111)-BMIPF6 ionic liquid interfaces from potential-
dependent AFM force curves, Chem. Commun. (Camb) 48 (2012) 582e584.
[12] M.A. Gebbie, M. Valtiner, X. Banquy, E.T. Fox, W.A. Henderson,
J.N. Israelachvili, Ionic liquids behave as dilute electrolyte solutions, Proc. Natl.
Acad. Sci. U.S.A. 110 (2013) 9674e9679.
[13] R. Steitz, V. Leiner, R. Siebrecht, R.v. Klitzing, Influence of the ionic strength on
the structure of polyelectrolyte films at the solid/liquid interface, Colloid.
Surface. Physicochem. Eng. Aspect. 163 (2000) 63e70.
[14] T. Albrecht, K. Moth-Poulsen, J.B. Christensen, J. Hjelm, T. Bjørnholm,
J. Ulstrup, Scanning tunneling spectroscopy in an ionic liquid, J. Am. Chem.
Soc. 128 (2006) 6574e6575.
[15] E. Sloutskin, B.M. Ocko, L. Tamam, I. Kuzmenko, T. Gog, M. Deutsch, Surface
layering in ionic liquids: an X-ray reflectivity study, J. Am. Chem. Soc. 127
(2005) 7796e7804.
[16] U. Doman  ska, K. Paduszyn  ski, (Solidþ liquid) and (liquidþ liquid) phase
 R. Wang et al. / Fluid Phase Equilibria 463 (2018) 106e113
equilibria measurements and correlation of the binary systems {tri-iso-butyl
(methyl) phosphonium tosylateþ alcohol, orþ hydrocarbon}, Fluid Phase
Equil. 278 (2009) 90e96.
[17] P. Paricaud, A. Galindo, G. Jackson, Recent advances in the use of the SAFT
approach in describing electrolytes, interfaces, liquid crystals and polymers,
Fluid Phase Equil. 194 (2002) 87e96.
[18] M. Sha, Q. Dou, F. Luo, G. Zhu, G. Wu, Molecular insights into the electric
double layers of ionic liquids on Au (100) electrodes, ACS Appl. Mater. In-
terfaces 6 (2014) 12556e12565.
[19] E.S. Ferreira, C.M. Pereira, M.N.l.D. Cordeiro, D.J. dos Santos, Molecular dy-
namics study of the gold/ionic liquids interface, J. Phys. Chem. B 119 (2015)
9883e9892.
[20] V.A. Nikitina, S.A. Kislenko, R.R. Nazmutdinov, M.D. Bronshtein, G.A. Tsirlina,
Ferrocene/ferrocenium redox couple at Au (111)/ionic liquid and Au (111)/
acetonitrile interfaces: a molecular-level view at the elementary act, J. Phys.
Chem. C 118 (2014) 6151e6164.
[21] C. Merlet, C. Pean, B. Rotenberg, P.A. Madden, P. Simon, M. Salanne, Simulating
supercapacitors: can we model electrodes as constant charge surfaces?
J. Phys. Chem. Lett. 4 (2013) 264e268.
[22] C. Pe an, C. Merlet, B. Rotenberg, P.A. Madden, P.-L. Taberna, B. Daffos,
M. Salanne, P. Simon, On the dynamics of charging in nanoporous carbon-
based supercapacitors, ACS Nano 8 (2014) 1576e1583.
[23] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen,
A smooth particle mesh Ewald method, J. Chem. Phys. 103 (1995) 8577e8593.
[24] I.-C. Yeh, M.L. Berkowitz, Ewald summation for systems with slab geometry,
J. Chem. Phys. 111 (1999) 3155e3162.
[25] B. Hess, C. Kutzner, D. Van Der Spoel, E. Lindahl, GROMACS 4: algorithms for
highly efficient, load-balanced, and scalable molecular simulation, J. Chem.
Theor. Comput. 4 (2008) 435e447.
[26] H. Heinz, R. Vaia, B. Farmer, R. Naik, Accurate simulation of surfaces and in-
terfaces of face-centered cubic metals using 12 6 and 9 6 Lennard-Jones
potentials, J. Phys. Chem. C 112 (2008) 17281e17290.
[27] H. Heinz, B.L. Farmer, R.B. Pandey, J.M. Slocik, S.S. Patnaik, R. Pachter, R.R. Naik,
Nature of molecular interactions of peptides with gold, palladium, and Pd Au
bimetal surfaces in aqueous solution, J. Am. Chem. Soc. 131 (2009)
9704e9714.
[28] J. Feng, R.B. Pandey, R.J. Berry, B.L. Farmer, R.R. Naik, H. Heinz, Adsorption
mechanism of single amino acid and surfactant molecules to Au {111} sur-
faces in aqueous solution: design rules for metal-binding molecules, Soft
Matter 7 (2011) 2113e2120.
[29] K.C. Jha, H. Liu, M.R. Bockstaller, H. Heinz, Facet recognition and molecular
ordering of ionic liquids on metal surfaces, J. Phys. Chem. C 117 (2013)
25969e25981.
[30] T. Halicio, G.M. Pound, Calculation of potential energy parameters form
crystalline state properties, Phys. Status Solidi 30 (1975) 619e623.
[31] H. Tamura, M. Yoshida, K. Kusakabe, C. Youngmo, R. Miura, M. Kubo,
K. Teraishi, A. Abhijit Chatterjee, A. Miyamoto, Molecular dynamics simulation
of friction of hydrocarbon thin films, Langmuir: ACS J. surface. colloids 15
(1999) 7816e7821.
[32] D. Huang, Y. Chen, K.A. Fichthorn, A molecular-dynamics simulation study of
the adsorption and diffusion dynamics of short n-alkanes on Pt(111), Chem-
Inform 26 (1995) 11021e11030.
[33] K.A. Fichthorn, P.G. Balan, Y. Chen, Simulation and analysis of the motion of n-
butane on Pt(111), Surf. Sci. 317 (1994) 37e44.
View publication stats
113
[34] S.M. Levine, S.H. Garofalini, Molecular dynamics simulation of atomic O on
Pt(111), Surf. Sci. 167 (2015) 198e206.
[35] P. Dauberosguthorpe, V.A. Roberts, D.J. Osguthorpe, J. Wolff, M. Genest,
A.T. Hagler, Structure and energetics of ligand binding to proteins: Escherichia
coli dihydrofolate reductase-trimethoprim, a drug-receptor system, Proteins
Struct. Funct. Bioinf. 4 (1988) 31e47.
[36] A.V. Verde, J.M. Acres, J.K. Maranas, Investigating the specificity of peptide
adsorption on gold using molecular dynamics simulations, Bio-
macromolecules 10 (2009) 2118e2128.
[37] A.K. Rappe , C.J. Casewit, K. Colwell, W. Goddard Iii, W. Skiff, UFF, a full periodic
table force field for molecular mechanics and molecular dynamics simula-
tions, J. Am. Chem. Soc. 114 (1992) 10024e10035.
[38] F. Iori, R. Di Felice, E. Molinari, S. Corni, GolP: an atomistic force-field to
describe the interaction of proteins with Au (111) surfaces in water,
J. Comput. Chem. 30 (2009) 1465e1476.
[39] O. Borodin, Polarizable force field development and molecular dynamics
simulations of ionic liquids, J. Phys. Chem. B 113 (2009) 11463e11478.
[40] V.V. Chaban, I.V. Voroshylova, Systematic refinement of canongia
LopeseP adua force field for pyrrolidinium-based ionic liquids, J. Phys. Chem. B
119 (2015) 6242e6249.
[41] J.N. Canongia Lopes, J. Deschamps, A.A.H. P adua, Modeling ionic liquids using a
systematic all-atom force field, J. Phys. Chem. B 108 (2004), 11250e11250.
[42] J.N. Canongia Lopes, A.A.H. P adua, Molecular force field for ionic liquids
composed of triflate or bistriflylimide anions, J. Phys. Chem. B 108 (2004)
16893e16898.
[43] Y. Lauw, M.D. Horne, T. Rodopoulos, V. Lockett, B. Akgun, W.A. Hamilton,
A.R. Nelson, Structure of [C4mpyr][NTf2] room-temperature ionic liquid at
charged gold interfaces, Langmuir: ACS J. surface. colloids 28 (2012)
7374e7381.
[44] C. Aliaga, S. Baldelli, Sum frequency generation spectroscopy and double-layer
capacitance studies of the 1-butyl-3-methylimidazolium Dicyanamide
platinum interface, J. Phys. Chem. B 110 (2006) 18481e18491.
[45] M.T. Alam, M.M. Islam, T. Okajima, T. Ohsaka, Measurements of differential
capacitance at mercury/room-temperature ionic liquids interfaces, J. Phys.
Chem. C 111 (2007) 18326e18333.
[46] V. Lockett, M. Horne, R. Sedev, T. Rodopoulos, J. Ralston, Differential capaci-
tance of the double layer at the electrode/ionic liquids interface, Phys. Chem.
Chem. Phys. 12 (2010) 12499e12512.
[47] A.A. Kornyshev, Double-layer in ionic liquids: paradigm change? J. Phys.
Chem. B 111 (2007) 5545e5557.
[48] Y. Lauw, M. Horne, T. Rodopoulos, F. Leermakers, Room-temperature ionic
liquids: excluded volume and ion polarizability effects in the electrical
double-layer structure and capacitance, Phys. Rev. Lett. 103 (2009) 117801.
[49] M.T. Alam, M.M. Islam, T. Okajima, T. Ohsaka, Capacitance measurements in a
series of room-temperature ionic liquids at glassy carbon and gold electrode
interfaces, J. Phys. Chem. C 112 (2008) 16600e16608.
[50] V. Lockett, M. Horne, R. Sedev, T. Rodopoulos, J. Ralston, Differential capaci-
tance of the double layer at the electrode/ionic liquids interface, Phys. Chem.
Chem. Phys.: PCCP 12 (2010) 12499e12512.
[51] G. Feng, J. Zhang, R. Qiao, Microstructure and capacitance of the electrical
double layers at the interface of ionic liquids and planar electrodes, J. Phys.
Chem. C 113 (2009) 4549e4559.
[52] E. Neria, S. Fischer, M. Karplus, Simulation of activation free energies in mo-
lecular systems, J. Chem. Phys. 105 (1996) 1902e1921.