Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields
Isothermal-isobaric molecular dynamics simulations are used to examine the microscopic structure and some properties of water-methanol liquid mixture. The TIP4P/2005 and SPC/E water models are combined with the united atom TraPPE and the all-atom force field model for methanol. Our principal focus...
Збережено в:
| Дата: | 2019 |
|---|---|
| Автори: | , , |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Інститут фізики конденсованих систем НАН України
2019
|
| Назва видання: | Condensed Matter Physics |
| Онлайн доступ: | http://dspace.nbuv.gov.ua/handle/123456789/157476 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields / M. Cruz Sanchez, H. Dominguez, O. Pizio // Condensed Matter Physics. — 2019. — Т. 22, № 1. — С. 13602: 1–14. — Бібліогр.: 52 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-157476 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1574762019-06-21T01:29:09Z Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields Cruz Sanchez, M Dominguez, H. Pizio, O. Isothermal-isobaric molecular dynamics simulations are used to examine the microscopic structure and some properties of water-methanol liquid mixture. The TIP4P/2005 and SPC/E water models are combined with the united atom TraPPE and the all-atom force field model for methanol. Our principal focus is to evaluate the quality of predictions of different combinations of model force fields concerning the composition dependence of basic properties of this system. Specifically, we explored the composition effects on density, excess molar volume and excess entropy, as well as on the surface tension and static dielectric constant. In addition, the structural properties are described in terms of the coordination numbers and the average number of hydrogen bonds between molecules of constituent species. Finally, the composition dependence of self-diffusion coefficients of the species is evaluated. All theoretical predictions are tested with respect to experimental data. Моделювання методом молекулярної динамiки в iзотермiчно-iзобаричному ансамблi застосовано до дослiдження мiкроскопiчної структури та деяких властивостей рiдкої сумiшi вода-метанол. Моделi води TIP4P/2005 i SPC/E поєднано з моделлю об’єднананих атомiв TraPPE i моделлю силових полiв всiх атомiв для метанолу. Основною метою даної роботи є дати якiсну оцiнку передбаченням рiзних комбiнацiй модельних силових полiв стосовно концентрацiйних залежностей основних властивостей системи. Зокрема, ми дослiдили вплив концентрацiї на густину, надлишковий молярний об’єм i надлишкову ентропiю, а також на поверхневий натяг i статичну дiелектричну сталу. Крiм цього, описано структурнi властивостi на мовi координацiйних чисел i середнього числа водневих зв’язкiв мiж молекулами компонентiв сумiшi. Нарештi, здiйснено оцiнку концентрацiйної залежностi коефiцiєнтiв самодифузiї компонентiв. Усi теоретичнi передбачення перевiрено по вiдношенню до експериментальних даних. 2019 Article Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields / M. Cruz Sanchez, H. Dominguez, O. Pizio // Condensed Matter Physics. — 2019. — Т. 22, № 1. — С. 13602: 1–14. — Бібліогр.: 52 назв. — англ. 1607-324X PACS: 61.20.-p, 61.20-Gy, 61.20.Ja DOI:10.5488/CMP.22.13602 arXiv:1903.11479 http://dspace.nbuv.gov.ua/handle/123456789/157476 en Condensed Matter Physics Інститут фізики конденсованих систем НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
Isothermal-isobaric molecular dynamics simulations are used to examine the microscopic structure and some
properties of water-methanol liquid mixture. The TIP4P/2005 and SPC/E water models are combined with the
united atom TraPPE and the all-atom force field model for methanol. Our principal focus is to evaluate the
quality of predictions of different combinations of model force fields concerning the composition dependence
of basic properties of this system. Specifically, we explored the composition effects on density, excess molar
volume and excess entropy, as well as on the surface tension and static dielectric constant. In addition, the
structural properties are described in terms of the coordination numbers and the average number of hydrogen
bonds between molecules of constituent species. Finally, the composition dependence of self-diffusion coefficients of the species is evaluated. All theoretical predictions are tested with respect to experimental data. |
| format |
Article |
| author |
Cruz Sanchez, M Dominguez, H. Pizio, O. |
| spellingShingle |
Cruz Sanchez, M Dominguez, H. Pizio, O. Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields Condensed Matter Physics |
| author_facet |
Cruz Sanchez, M Dominguez, H. Pizio, O. |
| author_sort |
Cruz Sanchez, M |
| title |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields |
| title_short |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields |
| title_full |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields |
| title_fullStr |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields |
| title_full_unstemmed |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields |
| title_sort |
molecular dynamics simulations of the properties of water-methanol mixtures. effects of force fields |
| publisher |
Інститут фізики конденсованих систем НАН України |
| publishDate |
2019 |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/157476 |
| citation_txt |
Molecular dynamics simulations of the properties of water-methanol mixtures. Effects of force fields / M. Cruz Sanchez, H. Dominguez, O. Pizio // Condensed Matter Physics. — 2019. — Т. 22, № 1. — С. 13602: 1–14. — Бібліогр.: 52 назв. — англ. |
| series |
Condensed Matter Physics |
| work_keys_str_mv |
AT cruzsanchezm moleculardynamicssimulationsofthepropertiesofwatermethanolmixtureseffectsofforcefields AT dominguezh moleculardynamicssimulationsofthepropertiesofwatermethanolmixtureseffectsofforcefields AT pizioo moleculardynamicssimulationsofthepropertiesofwatermethanolmixtureseffectsofforcefields |
| first_indexed |
2025-07-14T09:54:08Z |
| last_indexed |
2025-07-14T09:54:08Z |
| _version_ |
1837615650260910080 |
| fulltext |
Condensed Matter Physics, 2019, Vol. 22, No 1, 13602: 1–14
DOI: 10.5488/CMP.22.13602
http://www.icmp.lviv.ua/journal
Molecular dynamics simulations of the properties of
water-methanol mixtures. Effects of force fields
M. Cruz Sanchez1, H. Dominguez2, O. Pizio1 ∗
1 Instituto de Química, Universidad Nacional Autónoma de México, Circuito Exterior,
04510 Cd. de México, México
2 Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de México, Circuito Exterior,
04510 Cd. de México, México
Received February 22, 2019
Isothermal-isobaric molecular dynamics simulations are used to examine the microscopic structure and some
properties of water-methanol liquid mixture. The TIP4P/2005 and SPC/E water models are combined with the
united atom TraPPE and the all-atom force field model for methanol. Our principal focus is to evaluate the
quality of predictions of different combinations of model force fields concerning the composition dependence
of basic properties of this system. Specifically, we explored the composition effects on density, excess molar
volume and excess entropy, as well as on the surface tension and static dielectric constant. In addition, the
structural properties are described in terms of the coordination numbers and the average number of hydrogen
bonds between molecules of constituent species. Finally, the composition dependence of self-diffusion coeffi-
cients of the species is evaluated. All theoretical predictions are tested with respect to experimental data.
Key words: water-methanol mixtures, mixing properties, surface tension, molecular dynamics simulations
PACS: 61.20.-p, 61.20-Gy, 61.20.Ja
1. Introduction
This manuscript is the first part of our two-stage project that involves water-methanol mixtures.
Namely, in this first part, we present a set of results coming from the isothermal-isobaric (NPT) com-
puter simulation concerning composition changes of density, excess mixing volume and entropy, first
coordination numbers of species and average number of hydrogen bonds. Moreover, we explore the
behavior of the surface tension and the dielectric constant on composition, as well as the self-diffusion
coefficients of species.
The forthcoming, second part of the project, is devoted to the exploration of changes of all the
properties mentioned above, brought by addition of NaCl salt to water-methanol solvent. Again, the NPT
computer simulation technique will be applied. One of the principal issues we would like to address is to
evaluate the validity and quality of theoretical results coming from different combinations of force fields
with respect to experimental data.
For specific purposes of our project, it is worth mentioning that molecular dynamics computer
simulations have been widely applied to mixtures of water with various solvents. Most frequently studied
seem to be the mixtures of water with alcohols, specifically the water-methanol mixtures, see, e.g., [1–
11] and references therein for a rather comprehensive account of the previously applied modelling. Our
present report has been inspired by most recent contributions concerning the system in question [6, 8, 9].
On the other hand, important experimental observations concerning this system and used in the present
study have been discussed in [12–17].
∗Corresponding author, E-mail: oapizio@gmail.com
This work is licensed under a Creative Commons Attribution 4.0 International License . Further distribution
of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI.
13602-1
https://doi.org/10.5488/CMP.22.13602
http://www.icmp.lviv.ua/journal
http://creativecommons.org/licenses/by/4.0/
M. Cruz Sanchez, H. Dominguez, O. Pizio
The majority of computer simulation studies cited above have been focused on the application of
a single combination of force fields describing each constituent species, water and methanol with the
exception of [9]. This latter work involved two methanol force fields, namely the TraPPE model [18]
and OPLS-all atom model [19]. As concerns water, the SPC/E and TIP4P models were chosen, see,
e.g., [20] and [21], respectively. Formally, we apply similar strategy. However, in contrast to [9], we
apply the TIP4P/2005 water model [22] rather than the TIP4P version [21]. It is known that the the
TIP4P/2005 version provides a better performance of the microscopic structure of water [23] and yields
even better description of various properties compared to the TIP4P, see, e.g., table 2 of reference [24].
Moreover, we are specifically interested in using the TIP4P/2005, because it has been quite recently
applied to parametrize the description of properties of NaCl aqueous solution, see [25, 26]. It opens up
the possibility to study NaCl solutions with water-methanol solvent in subsequent work. On the other
hand, the gained knowledge would permit to explore complex solutes in mixed solvents tuning with
confidence their solubility on solvent composition. Some specific molecules of interest in medicinal
chemistry, for example curcumin, are marginally soluble in water, but dissolve well in, e.g., alcohols
or dimethylsulfoxide [27–29]. Additional comments concerning the methodological difference of our
procedure and technical details in comparison with reference [9] are given in the body of the manuscript
below.
To summarize, the principal objective of the present report is to investigate a set of properties of water-
methanol mixtures in the entire interval of composition. All the properties are validated by comparison
with experimental results.
2. Models and simulation details
Preliminarily discussing the simulation methodology, for the sake of convenience for the reader, in
table 1 we list some of the models used in previous studies of water-methanol mixtures and related to the
present work. Within the united atom model for methanol, like OPLS/UA and TraPPE (see table 1), the
CH3 group is considered as a single site. On the other hand, the all atom models explicitly involve all the
hydrogens of the methanol molecule.
Table 1.Models of water-methanol liquid mixtures and the combination rules (according to GROMACS
nomenclature).
water methanol combination rules
Galicia-Andres et al. [3] TIP4P/Ew [30] OPLS/UA [31, 32] CR2 (L-B)
SPC/E [20]
Galicia-Andres et al.. [2] TIP4P/Ew [30] OPLS/AA [19] CR3
SPC/E [20]
Wensink et al. [1] TIP4P [21] OPLS/AA [19] CR3
Guevara-Carrion et al. [33] SPC/E [20] UA-own design [34] CR2 (L-B)
TIP4P/2005 [22]
Kohns et al.. [35] SPC/E [20] UA-own design [34] CR2 (L-B)
Požar et al. [36] SPC/E [20] TraPPE [18] CR2 (L-B)
Obeidat et al. [9] TIP4P [21] TraPPE [18] CR2 (L-B)
SPC/E [20] OPLS/AA [19]
2.1. Technical details
In this work, we explore the SPC/E model [20] and the TIP4P/2005 model [22] for water. For
methanol, we used two models, namely the united atom model-TraPPE [18], and all atom model [19],
denominated as MET/TraPPE, and MET/AA, respectively. The Lorentz-Berthelot combination rules are
used to determine the cross parameters.
13602-2
Water-methanol mixtures
The long-range electrostatic interactions were handled by the particle mesh Ewald method imple-
mented in the GROMACS software package (fourth order, Fourier spacing equal to 0.12) with a precision
of 10−5. The nonbonded interactions were cut off at 1.4 nm. The van der Waals tail correction terms
to the energy and pressure were taken into account. In order to maintain the geometry of the water and
methanol molecules, the LINCS algorithm was used.
Our calculations were performed in the isothermal-isobaric (NPT) ensemble at 1 bar, and at a
temperature of 298.15 K. We used the GROMACS software package [37], version 5.1.2. Concerning
the procedure, a periodic cubic simulation box was set up for each system. The GROMACS genbox tool
was employed to randomly place all particles in the simulation box. To remove the possible overlaps of
particles introduced by the procedure of preparation of the initial configuration, each system underwent
energy minimization using the steepest descent algorithm implemented in the GROMACS package.
Minimization was followed by a 50 ps NPT equilibration run at 298.15 K and 1 bar using a timepstep of
0.25 fs. We used the Berendsen thermostat and barostat with τT = 1 ps and τP = 1 ps during equilibration.
Constant value of 4.5 · 10−5 bar−1 for the compressibility of the mixtures was employed. In the case of
pure methanol solvent, the compressibility was taken to be 1.2 · 10−4 bar−1. The V-rescale thermostat
and Parrinello-Rahman barostat with τT = 0.5 ps and τP = 2.0 ps and the time step 2 fs were used during
production runs. Statistics for each mole solvent composition and various ions concentration for any of
the properties were collected over several 10 ns NPT runs, each starting from the last configuration of
the preceding run. The time extension for each series of calculations will be mentioned below in the
appropriate place but not less than 70 ns.
While exploring the composition changes, the total number of molecules was kept fixed at 3000. The
composition is described by the mole fraction of methanol molecules Xm, Xm = Nm/(Nm + Nw).
3. Results and Discussion
In order to make things clear from the very beginning, we study four models of water-methanol liquid
mixtures described by the following force fields: TIP4P/2005-MET/TraPPE, TIP4P/2005-MET/AA,
SPC/E-MET/TraPPE and SPC/E-MET/AA. In all cases, the Lorentz-Berthelot combination rules are
used to obtain cross interaction parameters. The behavior of mixture density on composition, Xm, is given
in figure 1. Our NPT computer simulation results are supplemented by the experimental data [14, 38].
It can be seen that the density is perfectly well described by the TIP4P/2005-MET/TraPPE model.
The SPC/E-MET/TraPPE combination of force fields yields only a bit worse predictions in the interval
of intermediate compositions. Two models, the TIP4P/2005-MET/AA and SPC/E-MET/AA, are less
accurate regarding the density dependence on composition, apparently because the density of pure liquid
methanol is not described accurately within the MET/AA model. These trends of behavior have been
mentioned recently in [2, 3] exploring TIP4P/Ew water model [30] combined with methanol models with
a different degree of sophistication. On the other hand, the density dependence on composition from
the NPT simulations by Soetens and Bopp also agrees with the experimental data perfectly well [8]. It
is worth mentioning that the authors used well established BJH [39] and PHH [40] flexible models for
water and methanol, respectively. The results presented in figures 2 and 3 of [9] are slightly off all the
observations mentioned above concerning the density dependence on methanol molar fractions. Namely,
in figure 2 of that work, the liquid density of pure methanol from TraPPE and OPLS-AAmodels at 300 K
looks too low compared to the values reported in TraPPE database and by other authors. Moreover, the
discrepancy between the simulation predictions and experimental data given in figure 3 of [9] (the y-scale
of this figure seems to be not very appropriate to appreciate the total density) is too big compared to
other works. Apparently, trends of behavior of density from this work [9] are affected by the number of
particles, system size, method of calculations using a rather small liquid slab surrounded by vacuum and
by simulation time. To summarize the discussion of our figure 1 and the observations of other authors,
it seems that the dependence of density on composition is one of the important properties, but not too
demanding computationally, if the system size and time of simulations are properly chosen.
Next, in the spirit of previous works from our laboratory [2, 3] and of the development by Soetens
and Bopp [8], we turn our attention to the mixing properties. The excess molar volume, ∆Vmix, is defined
13602-3
M. Cruz Sanchez, H. Dominguez, O. Pizio
0 0.2 0.4 0.6 0.8 1
X
m
800
850
900
950
1000
ρ
(k
g
/m
3
)
water: SPC/E - triangles (black)
water: TIP4P/2005 - circles (red)
squares - experimental data [15,52]
dashed line - MET-AA
solid line - MET-TraPPE
solid line - MET-TraPPE
dashed line - MET/AA
Figure 1. (Colour online) Composition dependence of water-methanol mixture density from the present
NPT simulations of the TIP4P/2005-MET/TraPPE, TIP4P/2005-MET/AA, SPC/E-MET/TraPPE, and
SPC/E-MET/AA models, together with experimental data [14, 38]. The nomenclature of lines and
symbols is given in the figure.
as,
∆Vmix = Vmix − (1 − Xm)Vw − XmVm , (1)
whereVmix is the volume of the mixture at a certain composition,Vw andVm refer to the molar volumes of
pure water and pure methanol, respectively. Our simulation results for ∆Vmix are shown in figure 2 (a). As
in the case of density, cf. figure 1, here we observe again that the TIP4P/2005-MET/TraPPE model yields
the best agreementwith experimental results [16]. The entire set of experimental points iswell reproduced,
slight overestimation of the magnitude of ∆Vmix is observed in the interval of intermediate compositions
and when methanol content is higher than of water. Nevertheless, the experimentally observed minimum
at Xm = 0.5 is reproduced by simulations. If methanol is described by all-atom model (MET/AA),
the agreement between simulation results and experimental points deteriorates, the minimum shifts to
Xm ≈ 0.6. Quite similar dependence of ∆Vmix on composition (to the TIP4P/2005-MET/AA force field)
has been obtained from simulations of flexible BJH-PHH model, cf. figure 2 of reference [8]. If one
combines the SPC/E water with each of two methanol models in question, the values for ∆Vmix in almost
entire range of composition are underestimated, figure 2 (a). This kind of inaccuracy has been observed
previously and was discussed in detail in references [3, 41]. Energetic aspects of mixing are given by the
excess enthalpy of mixing, ∆Hmix. It is defined similarly to equation 1,
∆Hmix = Hmix − (1 − Xm)Hw − XmXm , (2)
where Hmix refers to the mixture entalphy whereas Hw and Hm describe the entalphy of each species
in pure state at the same temperature and pressure. From the comparison of the simulation results in
figure 2 (b) with experimental data [16], we conclude that none of the employed force field combinations
reproduce the experimental trends perfectly well. Concerning the magnitude of ∆Hmix, the TIP4P/2005-
MET/TraPPE and SPC/E-MET/AA models provide a better description than the TIP4P/2005-MET/AA
and SPC/E-MET/TraPPE. However, all the models of this study predict the minimum of ∆Hmix at Xm in
the interval between 0.5 and 0.6 rather than the experiment that shows the corresponding minimum at
Xm = 0.3. Similar kind of inaccuracy has been observed and discussed for a set of previously studied
models [2, 3]. Unfortunately, the excess entalphy of mixing from NPT simulations has not been presented
in reference [8] for the flexible BJH-PHH model to provide a critical evaluation of the present modelling.
Three data points previously reported for the excess potential energy from the NVE simulations of the
13602-4
Water-methanol mixtures
0 0.2 0.4 0.6 0.8 1
X
m
-1
-0.5
0
∆
V
m
ix
TIP4P/2005-MET/AA
TIP4P/2005-MET/TraPPE
SPC/E-MET/TraPPE
SPC/E-MET/AA
experimental data [16]
a
0 0.2 0.4 0.6 0.8 1
X
m
-1.5
-1
-0.5
0
∆
H
m
ix
experimental data [17]
SPC/E-MET/AA
SPC/E-MET/TraPPE
TIP4P/2005-MET/AA
TIP4P/2005-MET/TraPPE
b
Figure 2. (Colour online) Panel (a): Excess mixing volume of water-methanol mixtures on methanol
molar fraction for different combinations of models for each species together with the experimental data
from reference [15] (blue squares). Panel (b): A comparison of computer simulation data for mixing
entalphy with experimental data from [16] (blue squares). The nomenclature of lines and symbols is like
in figure 1.
combination of flexible models, see table 1 of reference [42], do not provide a definite answer in this
respect, unfortunately.
The microscopic structure of water-methanol mixtures has been discussed in terms of evolution of the
pair distribution functions on composition in many occasions. However, quantitative insights about the
changes of structure on Xm are usually described in terms of the first coordination numbers of the species
and of the number of hydrogen bonds. Here, in order to avoid unnecessary repetition, we adopt a similar
point of view. The first coordination numbers of the species are defined from the running coordination
numbers as common,
ni j(r) = 4πρj
r∫
0
gi j(R)R2dR , (3)
where the upper limit of integration in this equation is taken to correspond to the first minimum rmin of
13602-5
M. Cruz Sanchez, H. Dominguez, O. Pizio
the pair distribution function gi j(r) of the species i and j. A detailed description of the behavior of this
property, for a set of models related to the present study, was given in [2, 3, 43]. Very recent reports also
concern the evolution of the first coordination numbers, see, e.g., figure 7 of reference [8] and figures 24
and 25 of reference [9].
In order to evaluate the average number of H-bonds between molecules, the corresponding utility of
GROMACS software is applied with default distance — angle criterion. However, for distance cutoff we
used the first minimum of the corresponding pair distribution function, see also discussion of this issue
in references [2, 3, 43]. Averaging is performed over a piece of or the entire simulated trajectory.
Changes of ni j = ni j(rmin) on Xm for oxygens that belong to water or methanol, OW-OW and OW-
OM, and resulting from our simulations, are shown in figure 3(a). The coordination number nOW-OW
0 0.2 0.4 0.6 0.8
X
m
0
1
2
3
4
5
n
O
W
-O
W
,
n
O
W
-O
M
,
<
n
H
B
>
(
w
w
,
w
m
)
w - w
w - m
CN
HB
HB
CN
TIP4P/2005 (circles), SPC/E (triangles)
a
MET/TraPPE
0 0.2 0.4 0.6 0.8 1
X
m
0
0.5
1
1.5
2
2.5
3
n
O
M
O
M
,
n
O
M
-O
W
,
<
n
H
B
>
(m
m
,
m
w
) water (TIP4P-2005)
m - m
m - w
CN
HB
HB
CN
MET/TraPPE (red, circles)
MET/AA (blue, triangles)
b
Figure 3. (Colour online) The dependence of the first coordination number of oxygens for each species
(CN) and of average number of H-bonds (HB) on composition for models as in figure 1 and figure 2.
The coordination numbers are given by solid lines and symbols, whereas the dotted lines refer to the
numbers of hydrogen bonds. Normalization of the average number of H-bonds is performed per number
of molecules of the first species of the notation: w-w, w-m and m-m and m-w. Panel (a) is for w-w and
w-m where the panel (b) os for m-m and m-w. The nomenclature of lines and symbols is given in the
figure.
13602-6
Water-methanol mixtures
monotonously, almost linearly, decreases with an increasing methanol mole fraction starting from the
value ≈ 4.5 for pure water. On the other hand, nOW-OM increases up to ≈ 3 while approaching Xm = 1.
Two combinations of force fields, TIP4P-2005-MET/TraPPE and SPC/E-MET/TraPPE, yield almost the
same results. The crossing point between nOW-OW(Xm) and nOW-OM(Xm) describing “inversion” in the type
of predominant neighbors around water molecule occurs at Xm ≈ 0.65. The crossing point at the same
composition characterizes the BJH-PHH combination of flexible models for two species, figure 7 of [8].
On the other hand, evolution of the average number of hydrogen bonds 〈nHB〉 per water molecule on the
methanol mole fraction from our simulations is given in figure 3 (a) as well. Actually the dependences of
〈nHB〉 on Xm follow the trends of behavior of the first coordination numbers. The crossing point describing
the predominant number of H-bonds between two kinds of molecular species is observed at Xm ≈ 0.65.
Again, two water models (TIP4P/2005 and SPC/E), if combined with MET/TraPPE, yield very similar
predictions concerning the trends of behavior of 〈nHB〉 on Xm. The “inversion” or crossover point in
nOW-OW(Xm) and nOW-OM(Xm) or 〈nHB〉 is located close to the minimum of the excess mixing volume,
cf. figure 2 (a), as it was discussed in [3]. Finally, we would like to mention that the approximation of
average hydrogen bond numbers by the corresponding coordination numbers is not very appropriate in
the case OW-OW, see figure 3 (a), in contrast to what was claimed in [9].
Concerning the behavior of methanol coordination number and the average number of hydrogen
bonds in a mixture with varying composition, figure 3 (b), we would like to mention the following
trends. We explored two methanol models, the MET/TraPPE and MET/AA, combined with TIP4P-
2005 water model. The models yield a similar behavior. In pure methanol, the coordination number is
around three and the average number of hydrogen bonds between methanol molecules is only slightly
less. The methanol coordination smoothly decreases with a decreasing Xm. On the other hand, the cross
coordination numberOM-OW increaseswith a decreasing Xm, the corresponding hydrogen bonds number
follows this behavior. The “inversion” of the composition of the surrounding of a methanol molecule
on average occurs at Xm ≈ 0.65. In the region of low methanol mole fractions, methanol molecule
incorporates the hydrogen bonded structure of water and on average forms a bit more than two hydrogen
bonds with water molecules.
An excellent detailed analysis of the coordination numbers and hydrogen bonds network topology
was performed in reference [44], though the energetic definition of hydrogen bonds was used in that work.
Nevertheless, all trends of behavior deduced from geometric criterion for hydrogen bonding qualitatively
agree with the predictions described by using energetic definition.
One of the properties representing a quite demanding test of the employed force field is the surface
tension that describes the changes of surface free energy upon changing the surface area. Therefore, we
have undertaken additional calculations focused on the evaluation of surface tension for a set of models
under study. Relevant experimental results are available [17]. Moreover, this property has been explored
in detail in several computer simulation studies of water-methanol system, see, e.g., [9–11].
The simulations aiming at surface tension calculations at each point of composition axis in terms
of Xm, have been performed by using the final configuration of particles in the box from the NPT run.
Next, the box edge along z-axis was elongated by a factor of 3, generating a box with liquid slab and
two liquid-mixture – vacuum interfaces in the x − y plane, in close similarity to the procedure used
in reference [45]. The total number of particles, 3 × 103, seem to be reasonable as it provides the area
of the x − y face of the liquid slab large enough to avoid size effects. The elongation of the liquid
slab along z-axis is sufficient as well. To be more specific, we would like just to mention that the box
dimensions were 4.63 × 4.63 × 13.89 nm3 for Xm = 0.1 up to 5.75 × 5.75 × 17.25 nm3 for Xm = 0.9, for
the TIP4P/2005-MET/TraPPE model calculations. These numbers favourably compare with the cut-off
distance for non-bonded interactions 1.4 nm. The executable molecular dynamics file was modified by
deleting fixed pressure condition, just the V-rescale thermostatting with the same parameters as in the
NPT runs has been preserved. Other corrections have not been used.
The values for the surface tension, γ, follow from the combination of the time averages for the
components of the pressure tensor,
γ = Lz
〈[
Pzz −
1
2
(
Pxx + Pyy
) ]〉/
2 , (4)
where Pi j (i, j = x, y, z) are the components of the pressure tensor, and 〈...〉 denotes the time average.
13602-7
M. Cruz Sanchez, H. Dominguez, O. Pizio
We performed a set of NVT runs, not less than 5–6, each with the time duration of 10 ns, and obtained
the results for γ making the block average.
Our results for the surface tension and the excess mixing surface tension are given in two panels of
figure 4. It is known that the TIP4P/2005 model leads to a better prediction for the surface tension of
pure water in comparison with SPC/E [24]. This is also documented in figure 4 (a). Our result for pure
methanol within MET/TraPPE model agrees perfectly well with reference [46]. On the other hand, the
application of the same procedure with MET/AA model yields a slightly higher value for pure methanol,
very close to the experimental result, figure 4 (a). Small discrepancy of our result with the value given in
the supplementary material to [45] can be attributed to slightly different technical details of simulations,
e.g., cut-off, corrections to the electrostatics for a slab system, etc.
Overall trends of behavior of γ(Xm) in the entire composition interval are qualitatively correctly
reproduced by all four combinations of the force fields. A better description at low values of Xm is
reached if the TIP4P/2005 model is invloved. For high values of Xm, the application of the MET/AA
model leads to a better agreement with the experimental data. The values calculated by us are lower
0 0.2 0.4 0.6 0.8 1
X
m
20
30
40
50
60
70
γ
(m
N
/m
)
Experiment Ref.[18]
TIP4P/2005
SPC/E
MET/AA
MET/TraPPE
a
0 0.2 0.4 0.6 0.8 1
X
m
-25
-20
-15
-10
-5
0
∆
γ
(
m
N
/m
)
b
Figure 4. (Colour online) Panel (a): Surface tension of water-methanol mixtures on methanol molar
fraction as compared with experimental data from reference [17] (blue squares). Panel (b): Excess
mixing surface tension on composition. Both panels refer to the same models as in previous figures.
13602-8
Water-methanol mixtures
than the experimental data [17] at all Xm, due to the defficiency of performance of models for each
component, if the TIP4P/2005 and MET/TraPPE are used. With the MET/AA model of methanol, the
surface tension values are higher than the experimental points at intermediate and high methanol mole
fractions. Interestingly, the surface tension substantially decreases even if a small amount of methanol is
added to water. This behavior can be attributed to the tendency for methanol to be located close to the
interface with vacuum. Our findings are in agreement with the results reported in [9] for the same system
but with a smaller number of particles.
If one focuses on the deviation of the surface tension from ideal mixing behavior, figure 4 (b), the
TIP4P/2005-MET/TraPPE model is the closest to the experimental predictions. The present satisfactory
modelling can be seemingly used with confidence to other more complex systems. It is worth mentioning
that experimental results predict a maximum absolute value of deviation from ideality at Xm ≈ 0.3 and
computer simulation results reproduce this behavior. This particular composition actually coincides with
the maximum deviation of mixing entalphy, cf. figure 2 (b), that has not been predicted by the models in
question. Seemingly, the local composition fluctuations with specific orientations of molecules (yielding
nonideality of surface tension) are missing in the bulk phase to provide a correct behavior of nonideality
of entalphy. At present, it is difficult to offer a recipe how to improve the performance of nonpolarizable
rigid models for surface tension and for nonideality of entalphy, though.
In order to explore how the particles of different species are distributed in the box when Xm changes,
we have plotted the number density profiles of water and methanol oxygens in the entire slab plus vacuum
system in figure 5. At a low value of Xm, see Xm = 0.1 as an example, water is uniformly distributed in the
slab. Methanol species are distributed uniformly in the inner part of the slab, though we observe a high
maximum for methanol oxygens distribution at each liquid slab — vacuum interface. At intermediate
composition, Xm = 0.5, both species are almost uniformly distributed inside the liquid slab, but the
maxima of Om observed for Xm = 0.1 disappear, in expence of the growing amount of methanol oxygens
in the part of each interface exposed to vacuum. The two profiles describing the intermediate composition,
Xm = 0.5, are not perfectly symmetric with respect to the box center. This situation occurs very rarely
but can have consequences, if one pretends to obtain the precise density of species from this kind of
procedure. Specifically, in reference [9] the authors failed to estimate the liquid density both for methanol
and water at Xm = 0.3. Finally, at Xm = 0.9 in figure 5, it can be seen that the interface predominantly
contains methanol molecules whereas water molecules “hide” in the inner part of the liquid slab. Only a
small fraction of water molecules penetrates the interfacial region.
-4 -2 0 2 4
Relative position from box center (nm)
0
5
10
15
20
25
30
N
u
m
b
er
d
en
si
ty
p
ro
fi
le
s
(n
m
-3
)
X
m
=0.1 - solid
X
m
=0.5 - dashed
X
m
=0.9 - dash-dot
OW - red
OM - black
TIP4P/2005
MET/TraPPE
(circles)
(triangles)
Figure 5. (Colour online) Examples of the number density profiles of water and methanol oxygens for
mixtures (TIP4P/2005-MET/TraPPE model) at different composition.
13602-9
M. Cruz Sanchez, H. Dominguez, O. Pizio
The final part of the manuscript is concerned with the description of the self-diffusion coefficients
of species and changes of the dielectric constant with composition of water-methanol mixtures. The
self-diffusion coefficients of water and methanol were calculated from the mean-square displacement
(MSD) of a particle via Einstein relation,
Di =
1
6
lim
t→∞
d
dt
〈|ri(τ + t) − ri(τ)|2〉 , (5)
where i refers to water or methanol and τ denotes the time origin. Default settings of GROMACS were
used for the separation of the time origins.Moreover, the fitting interval (from 10% to 50% of the analyzed
trajectory) has been used to calculate Dm and Dw. Moreover, a special care has been taken to fitting for
the cases with a small number of particles on the extremes along the Xm axis. A set of our results is given
in figure 6.
The results for Dw, are qualitatively correct, the TIP4P/2005 model slightly underestimates Dw, it
yields 2.1 for purewater, whereas the SPC/E overestimates Dw, it yields≈ 2.6, see figure 6 (a). Concerning
the self-diffusion coefficient of pure methanol, it can be seen that MET/TraPPE underestimates Dm, it
0 0.2 0.4 0.6 0.8 1
X
m
1
1.5
2
2.5
D
w
x
1
0
-5
(c
m
2
/s
)
Experiment
SPC/E
TIP4P/2005
MET/AA
MET/TraPPE
MET/TraPPE
MET/AA
a
0 0.2 0.4 0.6 0.8 1
X
m
1
1.5
2
2.5
3
D
m
x
1
0
-5
(c
m
2
/s
)
SPC-E
MET/AA
MET/TraPPE
MET/AA
MET/TraPPE
TIP4P-2005
Experiment
b
Figure 6. (Color online) Self-diffusion coefficients of water, Dw, and of methanol, Dm, in water-methanol
mixture on composition [panels (a) and (b), respectively] and experimental data [47]. The nomenclature
of lines and symbols is given in the figure.
13602-10
Water-methanol mixtures
gives ≈ 2.25, whereas the all atom model, MET/AA, substantially overestimates Dm, figure 6 (b). This
tendency has been already discussed in reference [43]. These inaccuracies prohibit quantitatively correct
predictions for the composition changes of the self-diffusion coefficients of species.
Namely, the dependence Dw(Xm) is qualitatively correct. The TIP4P/2005, if combined with
MET/TraPPE or with MET/AA methanol model, leads to a very good prediction of Dw up to Xm ≈ 0.3,
for higher values of Xm simulation data deviate from experimental results [47]. The miminum value for
Dw from simulations is in the interval Xm between 0.4 and 0.5 whereas the experiment predicts this
minimum at 0.3. The minimum value of Dw from simulations along the composition axis coincides with
the minimum of excess mixing volume at Xm = 0.5. At high values of Xm, the growth of Dw is similar in
simulations and in experiment [47]. If water is descibed in the framework of the SPC/E model, the resuts
for Dw(Xm) substantially overestimate this self-diffusion coefficient in a wide range of composition.
Solely at high values of Xm, the agreement with experimental values becomes more acceptable. It is
worth mentioning that we were unable to get a better shape of Dw(Xm) within the SPC/E-MET/TraPPE
model. Apparently, there are two minima, one at Xm = 0.4 and anoter at Xm = 0.7. It is difficult to
establish (without additional exploration of various properties) if this behavior is related to the clustering
of species at local scale as it has been discussed in the experimental study [12]. Here, we would like just
to mention that similar evolution of Dw on Xm was reported in our recent work [3] in the study of SPC/E
model combined by OPLS/UA model of methanol [31].
Concerning the trends of behavior of the self-diffusion of methanol species in mixtures of different
composition Dm(Xm), we would like to emphasize the following. The best combination of force fields
is provided by the TIP4P/2005-MET/TraPPE model. It describes the changes of the function Dm(Xm)
pretty well in the entire composition range, figure 6 (b). The minimum value of Dm is described at a
correct place, Xm ≈ 0.3. All other combinations of the force fields exhibit defficiencies either due to the
water model like SPC/E or due to all-atom modelling of methanol at MET/AA level. An overall most
satisfactory picture of the dependence of the self-diffusion coefficients of two species, therefore, results if
the TIP4P/2005-MET/TraPPEmodel is used. We believe that alternative calculations of the self-diffusion
coefficients by applying velocity autocorrelation functions should lead to similar conclusions.
Our final concern is the evolution of the dielectric constant with composition. The long-range, asymp-
totic behavior of correlations between molecules possessing a permanent dipole moment is determined
by the dielectric constant, ε. Usually, long molecular dynamics runs are necessary to obtain reasonable
values for ε, because it is calculated from the time-average of the fluctuations of the total dipole moment
of the system [48],
ε = 1 +
4π
3kBTV
(
〈M2〉 − 〈M〉2
)
, (6)
where kB is the Boltzmann constant and V is the volume of the simulation box.
The lines from our NPT simulations of ε are shown in figure 7 (a). An overall behaviour of ε(Xm) is
that it decreases starting from a high value for pure water to a lower value corresponding to pure methanol
with an increasing Xm. As it follows from the comparison of the simulation results and experimental
data [49], all four combinations of the force fields substantially underestimate the values for ε in the entire
composition range. The reason is that the dielectric constant for two constituents, water and methanol,
is essentially underestimated. It is highly probable to improve the dependence of the static dielectric
constant on composition by applying the model for water specifically parametrized to reproduce the
dielectric constant for water [50]. Simultaneously, it would require parametrization of the force field for
methanol, see, e.g., [51]. However, common experience is that parametrization of a single property leads
to worse predictions for other properties. Therefore, this issue requires additional computation efforts.
On the other hand, a sensitive test is provided by comparison of the excess dielectric constant,
∆εmix = εm−[Xmεm+ (1− Xm)εw], with the experimental predictions [49]. Experimental points indicate
a negative deviation from ideality in the entire composition range, figure 6 (b). Maximal (negative)
deviation from the ideal type of behaviour reported from the experimental measurements is at Xm ≈ 0.45.
The simulations results reproduce the position of a minimum approximately; namely, the minimum is in
the interval 0.15–0.4, dependent on the combination of the force fields. These trends are in accordance
with observations concerning the deviations from ideality of thermodynamic properties. The magnitude
of the excess static dielectric constant is overestimated if the SPC/E model for water is used and is
13602-11
M. Cruz Sanchez, H. Dominguez, O. Pizio
0 0.2 0.4 0.6 0.8 1
X
m
20
30
40
50
60
70
80
d
ie
le
ct
ri
c
co
n
st
an
t
Experiment
TIP4P/2005
MET/TraPPE
MET/AA
SPC/E - MET/TraPPE
a
SPC/E - MET/AA
0 0.2 0.4 0.6 0.8 1
X
m
-8
-6
-4
-2
0
∆
ε
TIP4P-2005
MET/AA
MET/TraPPE
b
SPC/E - MET/TraPPE
SPC/E - MET/AA Experiment
Figure 7. (Colour online) Panel (a): Dielectric constant of water-methanol mixtures on methanol molar
fraction and experimental data from reference [49] (blue squares). Panel (b): Excess dielectric constant
on composition. The experimental data are from [49] (blue squares). The nomenclature of lines and
symbols is given in the figure.
underestimated if TIP4P/2005 is involved. The TIP4P/2005-MET/TraPPE model is the most close to the
experimental data combination of models. The excess dielectric constant curve as function of chemical
composition of the mixture can be related to the excess refractive index measurements, see, e.g., [52].
Consequently, a complementary comparisons with experiments would be desirable in future work.
4. Summary and conclusions
This work has been principally inspired by the necessity to evaluate various properties of water-
methanol mixtures using a specific set of models for each component. The choice of the force fields of
this study permits extensions necessary to describe ionic solutions with such combined solvents in the
spirit of very recent contributions from the research laboratory of C. Vega [25, 26] concerning aqueous
NaCl solutions with a novel force field for ions.
13602-12
Water-methanol mixtures
In the present work, extensive NPT molecular dynamics simulations were conducted to study ther-
modynamic, dynamic and structural properties of water-methanol mixtures. Two different water and two
methanol nonpolarizable models were combined and simulated with the purpose of testing their predic-
tions for an ample set of properties in the entire range of compositions. Comparisons with the available
experimental data were performed. Considering the scope of the models, the predictions obtained for
the mixtures appear to be qualitatively correct, particular properties were a bit better described at low
methanol compositions, i.e., for water-rich compositions. However, as a general trend, it is observed
that the best predictions are given with the water (TIP4P/2005)-methanol(TraPPE) mixtures since these
models reasonably well predict several properties of the pure components. The results indicate that a good
agreement with laboratory experiments could be obtained when both force fields, of the two components
in the mixture, are good. In fact, the dielectric constant is not well predicted by any of the simulated
mixtures since none of the selected models (water and methanol) predicts correctly that property.
Acknowledgements
M. Cruz and O.P. are grateful to M. Aguilar for technical support of this work at the Institute of
Chemistry of the UNAM.M. Cruz acknowledges support of CONACyT of Mexico for Ph.D. scholarship.
References
1. Wensink E.J.W., Hoffmann A.C., van Maaren P.J., van der Spoel D., J. Chem. Phys., 2003, 119, 7308,
doi:10.1063/1.1607918.
2. Galicia-Andrés E., Pusztai L., Temleitner L., Pizio O., J. Mol. Liq., 2015, 209, 586,
doi:10.1016/j.molliq.2015.06.045.
3. Galicia-Andrés E., Dominguez H., Pusztai L., Pizio O., Condens. Matter Phys., 2015, 18, 43602,
doi:10.5488/CMP.18.43602.
4. Perera A., Sokolić F., Almásy L., Koga Y., J. Chem. Phys., 2006, 124, 124515, doi:10.1063/1.2178787.
5. Perera A., Zoranić L., Sokolić F., Mazighi R., J. Mol. Liq., 2011, 150, 52, doi:10.1016/j.molliq.2010.05.006.
6. Bakó I., Pusztai L., Temleitner L., Sci. Rep., 2017, 7, 1073, doi:10.1038/s41598-017-01095-7.
7. Palinkás G., Bakó I., Heinzinger K., Bopp P., Mol. Phys., 1991, 73, 897, doi:10.1080/00268979100101641.
8. Soetens J.-C., Bopp P.A., J. Phys. Chem. B, 2015, 119, 8593, doi:10.1021/acs.jpcb.5b03344.
9. Obeidat A., Abu-Ghazleh H., AIP Adv., 2018, 8, 065203, doi:10.1063/1.5025575.
10. Matsumoto M., Takaoka Y., Kataoka Y., J. Chem. Phys., 1993, 98, 1464, doi:10.1063/1.464310.
11. Chang T.-M., Dang L.X., J. Phys. Chem. B, 2005, 109, 5759, doi:10.1021/jp045649v.
12. Takamuku T., Yamaguchi T., Asato M., Matsumoto M., Nishi N., Z. Naturforsch., A: Phys. Sci., 2000, 55, 513,
doi:10.1515/zna-2000-0507.
13. Wakisaka A., Ohki T., Faraday Discuss., 2005, 129, 231, doi:10.1039/B405391E.
14. Mikhail S.Z., Kimel W.R., J. Chem. Eng. Data, 1961, 6, 533, doi:10.1021/je60011a015.
15. McGlashan M.L., Williamson A.G., J. Chem. Eng. Data, 1976, 21, 196, doi:10.1021/je60069a019.
16. Lama R.F., Lu B.C.-Y., J. Chem. Eng. Data, 1965, 10, 216, doi:10.1021/je60026a003.
17. Vazquez G., Alvarez E., Navaza J.M., J. Chem. Eng. Data, 1995, 40, 611, doi:10.1021/je00019a016.
18. Chen B., Potoff J.J., Siepmann J.I., J. Phys. Chem. B, 2001, 105, 3093, doi:10.1021/jp003882x.
19. Jorgensen W.L., Maxwell D.S., Tirado-Rives J., J. Am. Chem. Soc., 1996, 118, 11225, doi:10.1021/ja9621760.
20. Berendsen H.J.C., Grigera J.R., Straatsma T.P., J. Phys. Chem., 1987, 91, 6269, doi:10.1021/j100308a038.
21. Jorgensen W.L., Chandrasekhar J., Madura J.D., Impey R.W., Klein M.L., J. Chem. Phys., 1983, 79, 926,
doi:10.1063/1.445869.
22. Abascal J.L.F., Vega C., J. Chem. Phys., 2005, 123, 234505, doi:10.1063/1.2121687.
23. Pusztai L., Pizio O., Sokolowski S., J. Chem. Phys., 2008, 129, 184103, doi:10.1063/1.2976578.
24. Vega C., Abascal J.L.F., Phys. Chem. Chem. Phys., 2011, 13, 19663, doi:10.1039/C1CP22168J.
25. Benavides A.L., Aragones J.L., Vega C., J. Chem. Phys., 2016, 144, 124504, doi:10.1063/1.4943780.
26. Benavides A.L., Portillo M.A., Chamorro V.C., Espinosa J.R., Abascal J.L.F., Vega C., J. Chem. Phys., 2017,
147, 104501, doi:10.1063/1.5001190.
27. Hazra M.K., Roy S., Bagchi B., J. Chem. Phys., 2014, 141, 18C501, doi:10.1063/1.4895539.
28. Samanta S., Roccatano D., J. Phys. Chem. B, 2013, 117, 3250, doi:10.1021/jp309476u.
29. Patsahan T., Ilnytskyi J.M., Pizio O., Condens. Matter Phys., 2017, 20, 23003, doi:10.5488/CMP.20.23003.
30. Horn H.W., SwopeW.C., Pitera J.W., Madura J.D., Dick T.J., Hura G.L., Head-Gordon T., J. Chem. Phys., 2004,
120, 9665, doi:10.1063/1.1683075.
13602-13
https://doi.org/10.1063/1.1607918
https://doi.org/10.1016/j.molliq.2015.06.045
https://doi.org/10.5488/CMP.18.43602
https://doi.org/10.1063/1.2178787
https://doi.org/10.1016/j.molliq.2010.05.006
https://doi.org/10.1038/s41598-017-01095-7
https://doi.org/10.1080/00268979100101641
https://doi.org/10.1021/acs.jpcb.5b03344
https://doi.org/10.1063/1.5025575
https://doi.org/10.1063/1.464310
https://doi.org/10.1021/jp045649v
https://doi.org/10.1515/zna-2000-0507
https://doi.org/10.1039/B405391E
https://doi.org/10.1021/je60011a015
https://doi.org/10.1021/je60069a019
https://doi.org/10.1021/je60026a003
https://doi.org/10.1021/je00019a016
https://doi.org/10.1021/jp003882x
https://doi.org/10.1021/ja9621760
https://doi.org/10.1021/j100308a038
https://doi.org/10.1063/1.445869
https://doi.org/10.1063/1.2121687
https://doi.org/10.1063/1.2976578
https://doi.org/10.1039/C1CP22168J
https://doi.org/10.1063/1.4943780
https://doi.org/10.1063/1.5001190
https://doi.org/10.1063/1.4895539
https://doi.org/10.1021/jp309476u
https://doi.org/10.5488/CMP.20.23003
https://doi.org/10.1063/1.1683075
M. Cruz Sanchez, H. Dominguez, O. Pizio
31. Jorgensen W.L., J. Phys. Chem., 1986, 90, 1276, doi:10.1021/j100398a015.
32. Haughney M., Ferrario M., McDonald I.R., J. Phys. Chem., 1987, 91, 4934, doi:10.1021/j100303a011.
33. Guevara-Carrion G., Nieto-Draghi C., Vrabec J., Hasse H., J. Phys. Chem. B, 2008, 112, 16664,
doi:10.1021/jp805584d.
34. Schnabel T., Srivastava A., Vrabec J., Hasse H., J. Phys. Chem. B, 2007, 111, 9871, doi:10.1021/jp0720338.
35. Kohns M., Horsch M., Hasse H., Fluid Phase Equilib., 2018, 458, 30, doi:10.1016/j.fluid.2017.10.034.
36. Požar M., Kerasidou A., Lovrinčević B., Zoranić L., Mijaković M., Primorac T., Sokolić F., Teboul V., Perera A.,
J. Chem. Phys., 2016, 145, 144502, doi:10.1063/1.4964487.
37. Van der Spoel D., Lindahl E., Hess B., Groenhof B., Mark A.E., Berendsen H.J.C., J. Comput. Chem., 2005,
26, 1701, doi:10.1002/jcc.20291.
38. Washbrun E.W. (Ed.), International Critical Tables of Numerical Data, Physics, Chemistry and Technology,
Knovel, New York, 2003.
39. Bopp P., Jancsó G., Heinzinger K., Chem. Phys. Lett., 1983, 98, 129, doi:10.1016/0009-2614(83)87112-7.
40. Palinkas G., Hawlicka E., Heinzinger K., J. Phys. Chem., 1987, 91, 4334, doi:10.1021/j100300a026.
41. González-Salgado D., Nezbeda I., Fluid Phase Equilib., 2006, 240, 161, doi:10.1016/j.fluid.2005.12.007.
42. Pálinkás G., Bakó I., Z. Naturforsch., A: Phys. Sci., 1991, 46, 95, doi:10.1515/zna-1991-1-215.
43. Galicia-Andrés E., Dominguez H., Pusztai L., Pizio O., J. Mol. Liq., 2015, 212, 70,
doi:10.1016/j.molliq.2015.08.061.
44. Bakó I., Megyes T., Bálint S., Grósz T., Chihaia V., Phys. Chem. Chem. Phys., 2008, 10, 5004,
doi:10.1039/b808326f.
45. Fischer N.M., van Maaren P.J., Ditz J.C., Yildrim A., van der Spoel D., J. Chem. Theory Comput., 2015, 11,
2938, doi:10.1021/acs.jctc.5b00190.
46. Biscay F., Ghoufi A., Malfreyt P., J. Chem. Phys., 2011, 134, 044709, doi:10.1063/1.3544926.
47. Derlacki Z.J., Easteal A.J., Edge A.V.J., Woolf L.A., Roksandic Z., J. Phys. Chem., 1985, 89, 5318,
doi:10.1021/j100270a039.
48. Neumann M., Mol. Phys., 1983, 50, 841, doi:10.1080/00268978300102721.
49. Albright P.S., Gosting L.J., J. Am. Chem. Soc., 1946, 68, 1061, doi:10.1021/ja01210a043.
50. Alejandre J., Chapela G.A., Saint-Martin H., Mendoza N., Phys. Chem. Chem. Phys., 2011, 13, 19728,
doi:10.1039/C1CP20858F.
51. Salas F.J., Méndez-Maldonado G.A., Núñez-Rojas E., Aguilar-Pineda G.E., Domínguez H., Alejandre J.,
J. Chem. Theory Comput., 2015, 11, 683, doi:10.1021/ct500853q.
52. Gofurov Sh., Ismailova O., Makhmanov U., Kokhkharov A., Int. J. Chem. Mol. Nucl. Mater. Metall. Eng., 2017,
11, 330.
Моделювання методом молекулярної динамiки
властивостей сумiшей вода - метанол. Вплив силових полiв
М. Круз Санчес1, Г. Домiнгес2, О. Пiзiо1
1 Iнститут хiмiї, Нацiональний автономний унiверситет м.Мехiко,Мехiко,Мексика
2 Iнститут матерiалознавства, Нацiональний автономний унiверситет м.Мехiко,Мехiко,Мексика
Моделювання методом молекулярної динамiки в iзотермiчно-iзобаричному ансамблi застосовано до до-
слiдження мiкроскопiчної структури та деяких властивостей рiдкої сумiшi вода-метанол. Моделi води
TIP4P/2005 i SPC/E поєднано з моделлю об’єднананих атомiв TraPPE i моделлю силових полiв всiх атомiв
для метанолу. Основною метою даної роботи є дати якiсну оцiнку передбаченням рiзних комбiнацiй мо-
дельних силових полiв стосовно концентрацiйних залежностей основних властивостей системи. Зокре-
ма, ми дослiдили вплив концентрацiї на густину, надлишковий молярний об’єм i надлишкову ентропiю,
а також на поверхневий натяг i статичну дiелектричну сталу. Крiм цього, описано структурнi властивостi
на мовi координацiйних чисел i середнього числа водневих зв’язкiв мiж молекулами компонентiв сумiшi.
Нарештi, здiйснено оцiнку концентрацiйної залежностi коефiцiєнтiв самодифузiї компонентiв. Усi теоре-
тичнi передбачення перевiрено по вiдношенню до експериментальних даних.
Ключовi слова: сумiшi вода-метанол, властивостi змiшування, поверхневий натяг, моделювання
методом молекулярної динамiки
13602-14
https://doi.org/10.1021/j100398a015
https://doi.org/10.1021/j100303a011
https://doi.org/10.1021/jp805584d
https://doi.org/10.1021/jp0720338
https://doi.org/10.1016/j.fluid.2017.10.034
https://doi.org/10.1063/1.4964487
https://doi.org/10.1002/jcc.20291
https://doi.org/10.1016/0009-2614(83)87112-7
https://doi.org/10.1021/j100300a026
https://doi.org/10.1016/j.fluid.2005.12.007
https://doi.org/10.1515/zna-1991-1-215
https://doi.org/10.1016/j.molliq.2015.08.061
https://doi.org/10.1039/b808326f
https://doi.org/10.1021/acs.jctc.5b00190
https://doi.org/10.1063/1.3544926
https://doi.org/10.1021/j100270a039
https://doi.org/10.1080/00268978300102721
https://doi.org/10.1021/ja01210a043
https://doi.org/10.1039/C1CP20858F
https://doi.org/10.1021/ct500853q
Introduction
Models and simulation details
Technical details
Results and Discussion
Summary and conclusions
|