Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
The results of Molecular Dynamics simulations of supercritical ammonia are reported for the first time. Qualitative agreement is found with neutron diffraction studies, the only experimental evidence on the structure of supercritical ammonia, so far. Based on the experiences with aqueous electrolyt...
Збережено в:
| Дата: | 2000 |
|---|---|
| Автори: | , , |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Інститут фізики конденсованих систем НАН України
2000
|
| Назва видання: | Condensed Matter Physics |
| Онлайн доступ: | https://nasplib.isofts.kiev.ua/handle/123456789/121029 |
| 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 supercritical ammonia and metal-ammonia solutions / S. Hannongbua, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2000. — Т. 3, № 2(22). — С. 381-392. — Бібліогр.: 26 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-121029 |
|---|---|
| record_format |
dspace |
| spelling |
nasplib_isofts_kiev_ua-123456789-1210292025-02-09T21:33:12Z Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions Моделювання методом молекулярної динаміки надкритичного аміаку та розчинів метал-аміак Hannongbua, S. Kiselev, M. Heinzinger, K. The results of Molecular Dynamics simulations of supercritical ammonia are reported for the first time. Qualitative agreement is found with neutron diffraction studies, the only experimental evidence on the structure of supercritical ammonia, so far. Based on the experiences with aqueous electrolyte solutions, it is discussed how the potentials describing ion-ammonia interactions in dilute and concentrated metal-ammonia solutions at subcritical conditions have to be modified for supercritical simulations. Вперше представлені результати моделювання методом молекулярної динаміки аміаку в надкритичному стані. Встановлено якісне узгодження з дослідженнями за дифракцією нейтронів, які є єдиним експериментальним свідченням про структуру аміаку у надкритичному стані. На основі відомих даних про водні розчини електролітів обговорюється питання про те, в який спосіб необхідно модифікувати потенціали, що описують взаємодії іон-аміак у розведених та концентрованих розчинах металів у аміаку, з метою їх застосування у моделюванні аміаку в надкритичному стані. Financial support by the Russian Fund for Basic Research through Grant No. RFBR–99–03–32064 (M.K.) and by INTAS Project No. 96–1989 (M.K. and K.H.) is gratefully acknowledged. It is a pleasure to also thank the National Electronic and Computer Technology Center in Bangkok for a generous supply of computer time. 2000 Article Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions / S. Hannongbua, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2000. — Т. 3, № 2(22). — С. 381-392. — Бібліогр.: 26 назв. — англ. 1607-324X DOI:10.5488/CMP.3.2.381 PACS: 61.20.J, 61.20.Qg, 71.15.Pd https://nasplib.isofts.kiev.ua/handle/123456789/121029 en Condensed Matter Physics application/pdf Інститут фізики конденсованих систем НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
The results of Molecular Dynamics simulations of supercritical ammonia
are reported for the first time. Qualitative agreement is found with neutron
diffraction studies, the only experimental evidence on the structure of supercritical ammonia, so far. Based on the experiences with aqueous electrolyte solutions, it is discussed how the potentials describing ion-ammonia
interactions in dilute and concentrated metal-ammonia solutions at subcritical conditions have to be modified for supercritical simulations. |
| format |
Article |
| author |
Hannongbua, S. Kiselev, M. Heinzinger, K. |
| spellingShingle |
Hannongbua, S. Kiselev, M. Heinzinger, K. Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions Condensed Matter Physics |
| author_facet |
Hannongbua, S. Kiselev, M. Heinzinger, K. |
| author_sort |
Hannongbua, S. |
| title |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| title_short |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| title_full |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| title_fullStr |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| title_full_unstemmed |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| title_sort |
molecular dynamics simulations of supercritical ammonia and metal-ammonia solutions |
| publisher |
Інститут фізики конденсованих систем НАН України |
| publishDate |
2000 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/121029 |
| citation_txt |
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions / S. Hannongbua, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2000. — Т. 3, № 2(22). — С. 381-392. — Бібліогр.: 26 назв. — англ. |
| series |
Condensed Matter Physics |
| work_keys_str_mv |
AT hannongbuas moleculardynamicssimulationsofsupercriticalammoniaandmetalammoniasolutions AT kiselevm moleculardynamicssimulationsofsupercriticalammoniaandmetalammoniasolutions AT heinzingerk moleculardynamicssimulationsofsupercriticalammoniaandmetalammoniasolutions AT hannongbuas modelûvannâmetodommolekulârnoídinamíkinadkritičnogoamíakutarozčinívmetalamíak AT kiselevm modelûvannâmetodommolekulârnoídinamíkinadkritičnogoamíakutarozčinívmetalamíak AT heinzingerk modelûvannâmetodommolekulârnoídinamíkinadkritičnogoamíakutarozčinívmetalamíak |
| first_indexed |
2025-12-01T00:24:19Z |
| last_indexed |
2025-12-01T00:24:19Z |
| _version_ |
1850263378540363776 |
| fulltext |
Condensed Matter Physics, 2000, Vol. 3, No. 2(22), pp. 381–392
Molecular Dynamics simulations of
supercritical ammonia and
metal-ammonia solutions
S.Hannongbua 1 , M.Kiselev 2 , K.Heinzinger 3
1 Department of Chemistry, Faculty of Science, Chulalongkorn University,
Bangkok 10330, Thailand
2 Institute of Solution Chemistry, Russian Academy of Sciences,
153045 Ivanovo, Russia
3 Max-Planck-Institut für Chemie (Otto-Hahn-Institut),
D-55020 Mainz, Germany
Received March 3, 2000
The results of Molecular Dynamics simulations of supercritical ammonia
are reported for the first time. Qualitative agreement is found with neutron
diffraction studies, the only experimental evidence on the structure of su-
percritical ammonia, so far. Based on the experiences with aqueous elec-
trolyte solutions, it is discussed how the potentials describing ion-ammonia
interactions in dilute and concentrated metal-ammonia solutions at subcrit-
ical conditions have to be modified for supercritical simulations.
Key words: Molecular Dynamics simulation, aqueous electrolyte
solutions, supercritical ammonia, metal-ammonia solutions
PACS: 61.20.J, 61.20.Qg, 71.15.Pd
Dedicated to Professor Igor Yukhnovskii on the occasion of his 75th birthday.
1. Introduction
There is an increasing interest in the investigation of supercritical electrolyte
solutions. The state of art of the simulation of water and aqueous solutions under
supercritical conditions can be found in the recent review by Chialvo and Cum-
mings [1]. Ammonia is also of interest as a solvent. The first, and so far the only,
neutron diffraction study of supercritical ammonia was reported by Bausenwein et
al. [2]. They measured pure ammonia at 449 K and four different densities. They an-
alyzed the experimental data by the Reverse Monte Carlo method (RMC) and com-
pared the results for the three radial distribution functions (RDF) with those from
c© S.Hannongbua, M.Kiselev, K.Heinzinger 381
S.Hannongbua, M.Kiselev, K.Heinzinger
calculations with the site-site-Ornstein-Zernicke (SSOZ) equation with the HNC
closure by employing various potential models available in the literature.
It is expected that MD simulations lead to more reliable results than the SSOZ
calculations. Therefore, we have performed simulations with the NH3 model em-
ployed for the simulation of Li+-liquid ammonia solutions some time ago [3]. The
same model was also used in the SSOZ calculations mentioned above and has led to
a reasonable agreement with the experimental results, at least at the higher densi-
ties. It will be interesting to see how much the MD simulations differ from those of
the SSOZ calculations under supercritical conditions.
In the next chapter the details of the first simulations of supercritical pure ammo-
nia will be given and preliminary results will be presented. In the following chapter
the solvation shell structures of various alkali and alkaline earth ions in ammonia will
be compared with those in aqueous solutions, all derived from simulations with a sin-
gle ion under subcritical conditions. From this comparison together with the results
for supercritical aqueous solutions, the applicability of the ion-ammonia potentials,
employed so far, for simulations of supercritical solutions will be discussed.
From simulations of concentrated Li+-ammonia solutions, cluster formation be-
tween solvated Li+ has been deduced [4]. In these simulations the pseudopotential
theory has been employed. In this way the electrons do not have to be treated explic-
itly and classical simulations can be performed [5]. The applicability of this theory
to supercritical concentrated metal-ammonia solutions will be discussed.
2. Simulations of supercritical pure ammonia
2.1. Details of the simulations
The three simulations, the results of which are reported here, were performed for
an NVT ensemble consisting of 215 flexible ammonia molecules. The ammonia model
employed has been taken from [3]. For the subcritical simulation a temperature of
220 K was chosen. With a density of 0.71 g cm−3, a pressure of 20 MPa results. For
both simulations at supercritical conditions, the temperature was set to 430 K, with
densities of 0.53 and 0.73 g cm−3 pressures of 20 and 400 MPa resulted, respectively.
The sidelength of the basic cube was about 20 Å. The time step length was chosen
to be 0.125 fs. The simulations extended over 10 ps each. For further details the
reader is referred to [6].
2.2. Preliminary results and comparison with experimental data
The N-N and N-H RDFs from two simulations at supercritical conditions are
compared in figure 1 with one of liquid ammonia at 220 K and a density of 0.71 g cm−3
[6]. The subcritical simulation agrees favourably with a former simulation with the
same ammonia model at 235 K and a density of 0.690 g cm−3 [3].
It can be seen from figure 1 that at the higher temperature, the height of the
first peak is lower and broadened for all RDFs. Also the other maxima and minima
382
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
r / Å
N−H
2
1
0
0 2 10864
2
1
0
g (r)
N−N
Figure 1. Nitrogen-nitrogen and nitrogen-hydrogen radial distribution functions
from simulations of supercritical ammonia. The full, dashed, and dotted lines
refer to temperatures of 220, 430, and 430 K and densities of 0.71, 0.73, and
0.53 g cm−3, respectively.
are less pronounced. The effect is more significant for the N-N when compared with
the N-H RDFs as the latter ones are already very broad even at a low temperature.
Furthermore, the positions of the first maxima remain unchanged with the in-
crease in temperature at a constant density. If the temperature is kept constant,
but the density decreased, then the positions of the first maxima shift to higher dis-
tances while the positions of all other maxima and minima seem to be unchanged.
The shoulder at the long distance side of the first peak in the N-N RDF – discussed
before in all simulations of pure ammonia – can be found even under supercritical
conditions. All RDFs extend at the high temperature to shorter distances, more
pronounced at the lower than at the higher density.
To the best of our knowledge there is no other simulation and only one experi-
mental investigation of supercritical ammonia for comparison. Bausenwein et al. [2]
performed neutron diffraction measurements at 449 K and four different densities
ranging from 0.318–0.700 g cm−3. The authors calculated from the intermolecular
total RDF the partial ones by RMC simulations and compared them with the results
from SSOZ calculations by employing various ammonia models from the literature,
one of which was the model used in this work, too.
There are some differences in the positions and the heights of the first maxima
in the RDFs between the simulation data presented in figure 1 and the SSOZ calcu-
lations. This is not surprising as it is well known that the employment of the SSOZ
method for hydrogen bonded liquids is connected with some difficulties.
There are also some small differences between the positions and the heights of the
first maxima of all RDFs between simulation results and neutron diffraction data.
383
S.Hannongbua, M.Kiselev, K.Heinzinger
But the tendencies of the changes in going from sub- to supercritical conditions are
the same: At a constant density the heights of the first peaks decrease and broaden
with the increasing temperature while the positions remain unchanged. At a constant
temperature and a decreasing density, the positions shift to larger distances and the
heights decrease for the N-N RDFs but increase for the N-H RDFs. This means that
even under supercritical conditions the hydrogen bonding increases with a decreasing
density.
A quantitative comparison between simulated and experimental results is not
possible because the partial RDFs derived from the total RDF by RMC show very
broad and noisy first peaks. In addition, it has to be kept in mind that there exists
a problem with the uniqueness of the RMC method.
The qualitative agreement between simulation and experiment indicates – at
least at this stage – that the ammonia model employed in the simulations does not
lead to serious discrepancies with the experiment. Therefore, it is justified to use this
model for further simulations of supercritical ammonia as well as metal-ammonia
solutions for the time being. The potentials necessary to describe the ion-ammonia
interactions are discussed in the following chapter.
3. Supercritical aqueous electrolyte and metal-ammonia
solutions
With the beginning of this decade the simulation of metal-liquid ammonia solu-
tions developed in three different directions:
1) The extension of single ion simulations to various alkali and alkaline earth
metals by neglecting the electron. The investigation of many-body effects in the
ion-ammonia potentials.
2) Quantum simulations with electron concentrations of up to 9 MPM (mole per-
cent of metal) where the positive ions are either treated as neutralizing background
or, at a high concentration, the Li+ are considered explicitly.
3) The simulation of concentrated Li+-ammonia solutions with concentrations
up to 20 MPM through the application of the pseudopotential theory. Renormalized
effective interatomic potentials are derived in this way, so that the electrons can
be excluded from explicit consideration and classical simulation methods can be
employed.
In the preceding report these three approaches towards the understanding of the
structural and dynamical properties of metal-liquid ammonia solutions have been
discussed in some detail [7]. Now it has to be checked whether the experience with
the potentials used in these simulations justifies their employment also in simulations
of supercritical metal-ammonia solutions.
3.1. Single ion simulations
It is difficult to settle the question of the need of many-body potentials for
the correct description of the structure of ionic solvation shells in metal-ammonia
384
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
Table 1. Ion-nitrogen and ion-oxygen first neighbour distances and solvation num-
bers as deduced from simulations for five cations. The temperatures for the am-
monia and the aqueous solutions were about 240 K and 300 K, respectively.
Ion rINmax n(rmin) Reference rIOmax n(rmin) Reference
Li 2.20 6;4 [9];[10] 2.13 6.1 [11]
Na 2.42 7 [12] 2.36 6.5 [13]
K 2.88 7 [14] 2.80 7.8 [15]
Mg 2.18 6 [16] 2.00 6 [17]
Ca 2.53 9 [18] 2.39 9.2 [19]
solutions because of a lack of experimental data which in turn is a result of the
temperature range of interest. The simulation results (mainly positions of the first
maxima of the ion-nitrogen RDFs and the coordination numbers) have already been
collected in [7] and will not be discussed here again. An indication of the reliability
of the results can be found by comparison with the hydration shells of those ions in
aqueous solutions where a large amount of experimental evidence is available, mainly
from X-ray and neutron diffraction measurements. In table I the first neighbour
ion-nitrogen and ion-oxygen distances and the coordination numbers, defined by the
integration up to the first minimum of the corresponding RDFs, are collected for five
cations. All results presented are from simulations. In the case of aqueous solutions
the simulation results are found to be in good agreement with experimental data [8].
The ion-nitrogen first neighbour distances are found to be slightly larger than
the ion-oxygen ones, while the solvation numbers seem to be quite similar, except
for the lithium ion, where some simulations lead to a value of four. The comparison
of the solvation number of Li+ in ammonia with that in water also strongly favours
a value of six. All pair potential simulations of various lithium halide solutions lead
to a hydration number of six in agreement with neutron diffraction studies with
isotopic substitution [20], which are expected to provide the most reliable data on
RDFs and, with it, on coordination numbers. It is very difficult to imagine why there
should be only four ammonia but six water molecules in the first solvation shell of
Li+ although the Li+-N distance is by about 0.1 Å larger than the Li+-O one. A
final answer cannot be derived from the simulations at this stage.
After it has been demonstrated in the preceding chapter that neat supercritical
ammonia can be simulated reliably, the question arises about supercritical metal-
ammonia solutions. So far no reports can be found in the literature. But there are
reports on simulations of supercritical aqueous solutions. So far only three cations
have been investigated. Driesner et al. [21] simulated a Na+ in 2177 water molecules,
while Lee and Cummings [22] reported on Li+, Na+, and Cs+.
The effect of the supercriticality on the structure of the hydration shell of the
cations in water can be described in short: The first neighbour ion-oxygen distances
do not change significantly with temperature or density. The hydration numbers
385
S.Hannongbua, M.Kiselev, K.Heinzinger
also do not change with temperature but decrease almost linearly with density.
These results have been achieved with the SPC/E model for water and ion-water
potentials which were taken from earlier simulations of aqueous electrolyte solutions
under ambient conditions, where simulations have led to the results in agreement
with the experimental data and also a simulation of a LiI solution at a constant
density and temperatures up to 500 K [23].
Unfortunately, there is no direct experimental evidence that the transfer of ion-
water potentials from ambient to supercritical conditions is justified for the ions
given in table 1. But the tendency that there is only a slight change in the first
neighbour distances with temperature and a decrease in the hydration number with
a decreasing density has been confirmed by neutron diffraction and EXAFS mea-
surements for Ni2+, Ag+, Rb+, and Cl− [21]. Another indirect way to learn about the
reliability of the potentials is the comparison of the simulated molar conductances for
very dilute aqueous solutions with experimental data for supercritical alkali halide
solutions. The molar conductances from the single ion simulations have to be added
up for a comparison with the experimental data from the salt solutions. Possible ion
pair formations might result in errors; but they are expected to be small because
of the very dilute solutions in the experiment, too. Good agreement has been found
for NaBr and CsBr, while there is a significant discrepancy in the case of the LiCl
solution which has to be attributed to the unclear conductance of Li+ [22]. From
the general agreement in the conductances it has been concluded that the hydra-
tion shell structures are also correct, at least qualitatively. The discrepancy for Li+
has been explained by the disagreement with the hydration number of Li+. From
the supercritical simulations, an almost density independent value of four follows
while from neutron diffraction studies with isotopic substitution clearly a hydration
number for Li+ of six results at ambient conditions.
In spite of this inconsistency one may conclude that in the case of aqueous
electrolyte solutions the use of pair potentials, which have been successfully employed
in the simulations at ambient conditions, do not lead to a significant disagreement
with experiments in the case of supercritical solutions. Based on this qualitative
agreement, it seems to be justified to use the pair potentials employed in single
ion metal-ammonia simulations also for supercritical solutions, at least for the time
being. In the next paragraph the problem of the potentials for the use in concentrated
supercritical metal-ammonia solutions will be investigated.
3.2. Concentrated metal-ammonia solutions
Scientifically more interesting than single ion studies are the concentrated metal-
ammonia solutions. It has been demonstrated that classical simulations can be
performed without taking into account the electrons explicitly by employing the
pseudopotential theory [5]. Then, the total potential describing effective site-site
interactions is given by
V ij
tot(R) = V ij
dir(R) + V ij
ind(R), (1)
386
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
0
25
75
50
−150
−50
0
0
50
100
150
−100
0 4 8 12 R / Å
N − N
N − H
H − H Li − Li
Li − N
Li − H
0
100
150
50
0
50
75
0
−150
−450
−300
−600
25
0 4 8 12 R / Å
V / kJ.mol−1
Figure 2. Site-site potentials for the six different interactions in the 11.64 MPM
lithium-liquid ammonia solution at a temperature of 240 K. The full and dashed
lines denote the direct and the total potential, respectively.
where V ij
dir(R) is the potential for the direct interactions, the same as used in the
single ion simulations (preceding chapter), and V ij
ind(R) is caused by free electrons
and derived within the pseudopotential theory [5], [24]. R denotes the distance
between two sites i and j belonging to different molecules or between a metal ion
and an ammonia molecule site. In principle the pseudopotential theory works well
if electron solvation does not occur and the metal-ammonia solution behaves like a
liquid metal. As it is to be expected that in a certain range of metal concentrations
the electrons are only weakly solvated [25], this situation is approximately realized
and the pseudopotential theory can be employed for the metal-ammonia solutions.
With this approach, a 11.64 MPM lithium-ammonia solution was simulated. In
this MC calculation the basic cube contained 135 Li+ and 1025 NH3 molecules. With
an experimental density of 0.554 g/cm3 a sidelength of the periodic cube of 35.89 Å
resulted. For the direct contributions to the total potential a rigid version of the
ammonia model described in [3] was employed while new ab initio calculations were
performed for the Li+-Li+ and the Li+-NH3 interactions [4].
In figure 2 the total potentials for the six site-site interactions in the Li+-NH3
solution as well as the direct contributions are depicted. The indirect contribu-
tions calculated from the pseudopotential theory result in a strong screening of the
Coulomb interactions in all site-site potentials. It is expected from these now much
387
S.Hannongbua, M.Kiselev, K.Heinzinger
weaker site-site interactions that the polarizibility effect becomes much less impor-
tant which means, of course, that problems with the many-body interactions, as
discussed in the single ion case [7], are much less relevant and need not be con-
sidered in first approximation in the concentrated metal-ammonia solutions. It is
interesting to note that the consequences of the screening of both, Coulombic at-
traction and repulsion, leaves the overall ammonia-ammonia and lithium-ammonia
potentials almost unchanged as can be seen from figure 3.
−50
0 4 8 12 R / Å
0
−100
−150
0
−4
4
−8
−12
V / kJ.mol−1
Figure 3. Ammonia-ammonia (top)
and lithium-ammonia (bottom) pair
potentials as a function of nitrogen-
nitrogen and lithium-nitrogen dis-
tances for orientations as shown in the
insertions. The full and dashed lines
refer to the direct and the total po-
tential, respectively.
A more detailed analysis of the con-
figurations resulting from the MC calcu-
lation shows that the partial RDFs can
only be understood by the formation of
clusters [4], where 33% of all NH3 belong
to the bulk. From the remaining 67% of
the molecules 55% are coordinated to one
Li+ and 12% simultaneously to two Li+.
These data are consistent with the vari-
ous partial RDFs and their running inte-
gration numbers only if we assume that
clusters are formed as shown in figure 4.
The following distribution of the clusters
results: about 35% of the solvated Li+ are
not connected, 10% have one (a), 15% have
two (b), and 40% have three ammonia
molecules in common [4].
Different from the single ion case there
is no possibility to compare with aqueous
electrolyte solutions in going to the su-
percritical solution because the electrons
cannot be neglected any longer. In addi-
tion, there are only results reported for a
1 molal aqueous NaCl solution [21] while
the concentrations of interest here are
of the order of 10 MPM. The result is
very similar to the single ion case. The
first neighbour Na-O distance remains in
the limits of statistical uncertainty un-
changed, while the hydration number de-
creases slightly with increasing tempera-
ture but decreases strongly with decreasing
density. The small hydration number of Na+ at ambient conditions of 4.8 compared
with 6 for Li+ is explained by a significant amount of ion pairing. But there is no ex-
perimental evidence for ion pairing at room temperature for a 1 molal NaCl solution.
There is no indication that the pseudopotential theory cannot be also employed
for the simulations under supercritical conditions. As the densities of free electrons
388
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
a
b
c
Figure 4. Solvated lithium ions with one (a), two (b), and three (c) ammonia
molecules in common, calculated from a 11.64 MPM lithium-ammonia solution
at 240 K.
are abnormally low even in the concentrated metal-ammonia solutions [5], [24], treat-
ing the electron-ion interaction within the local field function approach faces some
difficulties. The latter could be more pronounced under supercritical conditions.
Therefore, the deviation of the indirect potential in equation (1) should be modi-
fied. Work on this subject is in progress [26].
4. Summary and conclusions
Molecular Dynamics simulations of supercritical ammonia are reported for the
first time. The results are compared with those of neutron diffraction studies under
similar conditions which provide the only experimental information on the structure
of supercritical ammonia, so far. The agreement found justifies the use of the ammo-
nia model employed not only for further simulations of supercritical pure ammonia
but also of supercritical metal-ammonia solutions, at least for the time being.
In the simulation of metal-ammonia solutions, different treatments of the ion-
ammonia potentials are necessary. If we are concerned with the solvation shells
of an ion only, then single ion simulations can be employed where the electron is
389
S.Hannongbua, M.Kiselev, K.Heinzinger
simply neglected. By comparison of the changes occurring in aqueous electrolyte
solutions when going from sub- to supercritical conditions it can be concluded that
the same ion-ammonia potentials can be also employed here in both cases. It has
been demonstrated before that for concentrated solutions at subcritical conditions
by introduction of the pseudopotential theory the electrons can be excluded from
explicit consideration and classical simulations can be performed. The same proce-
dure should be also possible under supercritical conditions, but new potentials have
to be derived and calculated.
Acknowledgements
Financial support by the Russian Fund for Basic Research through Grant
No. RFBR–99–03–32064 (M.K.) and by INTAS Project No. 96–1989 (M.K. and
K.H.) is gratefully acknowledged. It is a pleasure to also thank the National Elec-
tronic and Computer Technology Center in Bangkok for a generous supply of com-
puter time.
References
1. Chialvo A.A., Cummings P.T. Molecular-based modelling of water and aqueous solu-
tions at supercritical conditions. // Adv. Chem. Phys., 1999, vol. 109, p. 115–205.
2. Bausenwein T., Bertagnolli H., David A., Goller K., Zweier H., Tödheide K., Chieux P.
Structure and intermolecular interactions in fluid ammonia: An investigation by neu-
tron diffraction at high pressure, statistical-mechanical calculations, and computer
simulations. // J. Chem. Phys., 1994, vol. 101, p. 672–682.
3. Hannongbua S., Ishida T., Spohr E., Heinzinger K. Molecular Dynamics study of a
lithium ion in ammonia. // Z. Naturforsch., 1988, vol. 43a, p. 572–582.
4. Hannongbua S., Kokpol S., Gurskii Z., Heinzinger K. Cluster formation in a con-
centrated lithium-liquid ammonia solution. A Monte Carlo study. // Z. Naturforsch.,
1997, vol. 52a, p. 828–834.
5. Gurskii Z., Hannongbua S., Heinzinger K. On the structure of concentrated lithium-
liquid ammonia solutions. A Molecular Dynamics study. // Mol. Phys., 1993, vol. 78,
p. 461–474.
6. Kiselev M., Kerdcharoen T., Hannongbua S., Heinzinger K. Structural properties of
sub- and supercritical ammonia as studied by Molecular Dynamics simulations. // (to
be published).
7. Heinzinger K. Computer simulations of metal-liquid ammonia solutions. // J. Molec.
Liquids (in press).
8. Ohtaki H., Radnai T. Structure and dynamics of hydrated ions. // Chem. Rev., 1993,
vol. 93, p. 1157–1204.
9. Hannongbua S. On the solvation of lithium ions in liquid ammonia: Monte Carlo
simulations with a three-body potential. // Chem. Phys. Lett., 1998, vol. 288, p. 663–
668.
10. Kerdcharoen T., Liedl K.R., Rode B.M. A QM/MM simulation method applied to the
solvation of Li+ in ammonia. // Chem. Phys., 1996, vol. 211, p. 313–323.
390
Molecular Dynamics simulations of supercritical ammonia and metal-ammonia solutions
11. Szász Gy.I., Heinzinger K., Pálinkás G. The structure of the hydration shell of the
lithium ion. // Chem. Phys. Lett., 1981, vol. 78, p. 194–196.
12. Hannongbua S. Molecular Dynamics simulation of the structure of a sodium ion in
liquid ammonia. // Aust. J. Chem., 1991, vol. 44., p. 447–456.
13. Heinje G., Luck W.A.P., Heinzinger K. Molecular Dynamics simulation of an aqueous
NaClO4 solution. // J. Phys. Chem., 1987, vol. 91, p. 331–338.
14. Tongraar A., Hannongbua S., Rode B.M. Molecular Dynamics simulations of a potas-
sium ion and an iodide ion in liquid ammonia. // Chem. Phys., 1997, vol. 219, p. 279–
290.
15. Migliore M., Fornili S. L., Spohr E., Pálinkás G., Heinzinger K. A Molecular Dynamics
study of the structure of an aqueous KCl solution. // Z. Naturforsch., 1986, vol. 41a,
p. 826–834.
16. Hannongbua S. The role of nonadditive effects in the first solvation shell of Na+ and
Mg2+ in liquid ammonia: Monte Carlo studies including three-body corrections. //
J. Chem. Phys., 1997, vol. 106, p. 6076–6081.
17. Dietz W., Riede W.O., Heinzinger K. Molecular Dynamics simulation of an aqueous
MgCl2 solution. // Z. Naturforsch., 1982, vol. 37a, p. 1038–1048.
18. Sidhisoradej W., Hannongbua S., Ruffolo D. Three-body effects in calcium(II)-
ammonia solutions: Molecular Dynamics simulations. // Z. Naturforsch., 1998,
vol. 53a, p. 208–216.
19. Probst M.M., Radnai T., Heinzinger K., Bopp P., Rode B.M. Molecular Dynamics and
X-ray investigation of an aqueous CaCl2 solution. // J. Phys. Chem., 1985, vol. 89,
p. 753–759.
20. Enderby J.E., Cummings S., Herdman G.J., Neilson G.W., Salomon P.S., Skipper N.
Diffraction and the study of aqua ions. // J. Phys. Chem., 1987, vol. 91, p. 5851–5858.
21. Driesner T., Seward T.M., Tironi I.G. Molecular Dynamics simulation study of ionic
hydration and ion association in dilute and 1 molal aqueous sodium chloride solutions
from ambient to supercritical conditions. // Geochim. Cosmochim. Acta, 1998, vol. 62,
p. 3095–3107.
22. Lee S.H., Cummings P.T. Molecular Dynamics simulation of limiting conductances
for LiCl, NaBr, and CsBr in supercritical water. // J. Chem. Phys., 2000, vol. 112,
p. 864–869.
23. Szász Gy.I., Heinzinger K. Hydration shell structures in a LiI solution at elevated
temperature and pressure: A Molecular Dynamics study. // Earth Planet. Sci. Lett.,
1983, vol. 64, p. 163–167.
24. Gurskii Z., Kushaba V., Hannongbua S., Heinzinger K. On unusually strong temper-
ature dependences of interatomic potentials in metal-ammonia solutions. // Metal
Phys. Adv. Techn., 1995, vol. 17, p. 62–67.
25. Deng Z., Martyna G.J., Klein M.L. Quantum simulation studies of metal-ammonia
solutions. // J. Chem. Phys., 1994, vol. 100, p. 7590–7601.
26. Gurskii Z., Kiselev M., Hannongbua S., Heinzinger K. Molecular Dynamics simulations
of supercritical concentrated metal-ammonia solutions. // (in preparation).
391
S.Hannongbua, M.Kiselev, K.Heinzinger
Моделювання методом молекулярної динаміки
надкритичного аміаку та розчинів метал-аміак
С.Ганнонґбуа 1 , М.Кісєльов 2 , К.Гайнцінгер 3
1 Відділення хімії, факультет природничих наук,
університет м.Чулалонгкорн,
Бангкок 10330, Тайланд
2 Інститут хімії розчинів Російської академії наук,
153045 Іваново, Росія
3 Інститут хімії Макса Планка (Інститут Отто Гана),
D-55020 Майнц, Німеччина
Отримано 3 березня 2000 р.
Вперше представлені результати моделювання методом молеку-
лярної динаміки аміаку в надкритичному стані. Встановлено якісне
узгодження з дослідженнями за дифракцією нейтронів, які є єдиним
експериментальним свідченням про структуру аміаку у надкритич-
ному стані. На основі відомих даних про водні розчини електролітів
обговорюється питання про те, в який спосіб необхідно модифікува-
ти потенціали, що описують взаємодії іон-аміак у розведених та кон-
центрованих розчинах металів у аміаку, з метою їх застосування у мо-
делюванні аміаку в надкритичному стані.
Ключові слова: метод молекулярної динаміки, водні розчини
електролітів, аміак в надкритичному стані, розчини металів у аміаку
PACS: 61.20.J, 61.20.Qg, 71.15.Pd
392
|