The structure of concentrated Li-ammonia solutions as derived from MD simulations
The results of Molecular Dynamics simulations of lithium-ammonia solutions over the whole concentration range from 0.5 to 19.6 MPM at 240 K are reported. The pseudopotential theory is employed at the higher concentrations and the direct contribution to the total potential which has been derived...
Gespeichert in:
| Datum: | 2003 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Інститут фізики конденсованих систем НАН України
2003
|
| Schriftenreihe: | Condensed Matter Physics |
| Online Zugang: | http://dspace.nbuv.gov.ua/handle/123456789/120755 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | The structure of concentrated Li-ammonia solutions as derived from MD simulations / S. Hannongbua, T. Remsungnen, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2003. — Т. 6, № 3(35). — С. 459-470. — Бібліогр.: 20 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-120755 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1207552017-06-14T18:24:26Z The structure of concentrated Li-ammonia solutions as derived from MD simulations Hannongbua, S. Remsungnen, T. Kiselev, M. Heinzinger, K. The results of Molecular Dynamics simulations of lithium-ammonia solutions over the whole concentration range from 0.5 to 19.6 MPM at 240 K are reported. The pseudopotential theory is employed at the higher concentrations and the direct contribution to the total potential which has been derived from ab initio calculations has been supplemented by three-body terms. The resulting partial structure and radial distribution functions are compared in detail with recent X-ray and neutron diffraction studies with isotopic substitution. There is an overall good agreement between simulation and experiment. Differences are discussed. The solvation number of the lithium ion is found to be six for the dilute solution and four for the higher concentrations in analogy to the hydration of Li⁺ found for various aqueous solutions. В даній роботі представлені результати, отримані методом молекулярної динаміки для розчинів літій-аміакy для усього діапазону концентрацій від 0.5 до 19.6 MPM при температурі 240 К. Використовується теорія псевдопотенціалів при вищих концентраціях. Прямий внесок загального потенціалу, який отримується з ab initio розрахунків, доповнюється тричастинковим доданком. Результуюча частинна структура і радіальні функції розподілу детально порівнюються з рентгенівським і нейтронним розсіюванням з ізотопним заміщенням. Результати експерименту і моделювання добре узгоджуються. Знайдено сольватаційні числа іонів літію для розведених розчинів (шість), а також для розчинів з вищою концентрацією (чотири) по аналогії з гідратацією іонів літію у різних водних розчинах. 2003 Article The structure of concentrated Li-ammonia solutions as derived from MD simulations / S. Hannongbua, T. Remsungnen, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2003. — Т. 6, № 3(35). — С. 459-470. — Бібліогр.: 20 назв. — англ. 1607-324X PACS: 61.20.J, 61.20.Qg, 71.15.Pd DOI:10.5488/CMP.6.3.459 http://dspace.nbuv.gov.ua/handle/123456789/120755 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 |
The results of Molecular Dynamics simulations of lithium-ammonia solutions
over the whole concentration range from 0.5 to 19.6 MPM at 240 K
are reported. The pseudopotential theory is employed at the higher concentrations
and the direct contribution to the total potential which has been
derived from ab initio calculations has been supplemented by three-body
terms. The resulting partial structure and radial distribution functions are
compared in detail with recent X-ray and neutron diffraction studies with
isotopic substitution. There is an overall good agreement between simulation
and experiment. Differences are discussed. The solvation number of
the lithium ion is found to be six for the dilute solution and four for the higher
concentrations in analogy to the hydration of Li⁺ found for various aqueous
solutions. |
| format |
Article |
| author |
Hannongbua, S. Remsungnen, T. Kiselev, M. Heinzinger, K. |
| spellingShingle |
Hannongbua, S. Remsungnen, T. Kiselev, M. Heinzinger, K. The structure of concentrated Li-ammonia solutions as derived from MD simulations Condensed Matter Physics |
| author_facet |
Hannongbua, S. Remsungnen, T. Kiselev, M. Heinzinger, K. |
| author_sort |
Hannongbua, S. |
| title |
The structure of concentrated Li-ammonia solutions as derived from MD simulations |
| title_short |
The structure of concentrated Li-ammonia solutions as derived from MD simulations |
| title_full |
The structure of concentrated Li-ammonia solutions as derived from MD simulations |
| title_fullStr |
The structure of concentrated Li-ammonia solutions as derived from MD simulations |
| title_full_unstemmed |
The structure of concentrated Li-ammonia solutions as derived from MD simulations |
| title_sort |
structure of concentrated li-ammonia solutions as derived from md simulations |
| publisher |
Інститут фізики конденсованих систем НАН України |
| publishDate |
2003 |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/120755 |
| citation_txt |
The structure of concentrated
Li-ammonia solutions as derived from
MD simulations / S. Hannongbua, T. Remsungnen, M. Kiselev, K. Heinzinger // Condensed Matter Physics. — 2003. — Т. 6, № 3(35). — С. 459-470. — Бібліогр.: 20 назв. — англ. |
| series |
Condensed Matter Physics |
| work_keys_str_mv |
AT hannongbuas thestructureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT remsungnent thestructureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT kiselevm thestructureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT heinzingerk thestructureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT hannongbuas structureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT remsungnent structureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT kiselevm structureofconcentratedliammoniasolutionsasderivedfrommdsimulations AT heinzingerk structureofconcentratedliammoniasolutionsasderivedfrommdsimulations |
| first_indexed |
2025-07-08T18:31:31Z |
| last_indexed |
2025-07-08T18:31:31Z |
| _version_ |
1837104620679200768 |
| fulltext |
Condensed Matter Physics, 2003, Vol. 6, No. 3(35), pp. 459–470
The structure of concentrated
Li-ammonia solutions as derived from
MD simulations∗
S.Hannongbua 1 , T.Remsungnen 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),
POB 3060, 55020 Mainz, Germany
Received May 16, 2003
The results of Molecular Dynamics simulations of lithium-ammonia solu-
tions over the whole concentration range from 0.5 to 19.6 MPM at 240 K
are reported. The pseudopotential theory is employed at the higher con-
centrations and the direct contribution to the total potential which has been
derived from ab initio calculations has been supplemented by three-body
terms. The resulting partial structure and radial distribution functions are
compared in detail with recent X-ray and neutron diffraction studies with
isotopic substitution. There is an overall good agreement between simula-
tion and experiment. Differences are discussed. The solvation number of
the lithium ion is found to be six for the dilute solution and four for the higher
concentrations in analogy to the hydration of Li+ found for various aqueous
solutions.
Key words: Molecular Dynamics simulation, lithium-ammonia solutions,
pseudopotential theory
PACS: 61.20.J, 61.20.Qg, 71.15.Pd
1. Introduction
Recent investigations of various metal-ammonia solutions by Molecular Dynam-
ics (MD) simulations demonstrated the importance of three-body contributions for
the structure of the solvation shells of the ions [1–4]. In addition, the comparison with
neutron and X-ray diffraction studies showed some discrepancies, especially with the
∗Dedicated to Professor Myroslav Holovko on the occasion of his 60th birthday.
c© S.Hannongbua, T.Remsungnen, M.Kiselev, K.Heinzinger 459
S.Hannongbua et al.
latest neutron diffraction studies with isotopic substitution (NDIS) at high concen-
trations [5–9]. Although earlier simulations of very dilute Li-ammonia solutions did
not show any effect of three-body contributions to the lithium ion solvation [10], we
decided to simulate again Li-ammonia solutions over the whole concentration range
where the pair potentials had been supplemented by three-body terms.
By applying the pseudopotential theory, effective renormalized interatomic po-
tentials were derived through which electrons can be excluded from an explicit con-
sideration and classical computer simulation methods can be used to calculate the
structural and dynamical properties of concentrated metal-ammonia solutions [11].
Thus, the total potential which describes the effective interactions between the sites
of i and j in the metal-ammonia system containing free unlocalized electrons is given
by
V
ij
tot(r) = V
ij
dir
(r) + V
ij
ind
(r), (1)
where r is the distance between the two sites belonging to different molecules, be-
tween two metal ions, or between a metal ion and an ammonia molecule site. The
direct contribution is based on ab initio calculations while the indirect contribution
can be derived solely from the pseudopotential theory provided that the metal con-
centration is sufficiently high for the electrons not to be localized. In this way, the
structure of a 11.64 MPM Li-ammonia solution (about half saturated) has been de-
rived from an MC simulation with a rigid ammonia model [12]. The reliability of the
thus achieved results could not be checked because of a lack of detailed experimental
information for this concentration.
In this work, we report on the simulations of Li-ammonia solutions over the
whole concentration range. In the next chapter we discuss the potentials and the
details of the simulations followed by a chapter where the structure of such solutions
and its dependence on the lithium concentration is discussed in comparison with the
results from neutron and X-ray diffraction studies.
2. Potentials and details of the simulations
A flexible four site model was employed to describe the ammonia-ammonia in-
teractions. The experimental gas phase geometry, with an N-H distance of 1.0124 Å
and an HNH angle of 106.67◦, was taken from reference [13]. The partial charges of
0.2674|e| on the hydrogen and −0.8022|e| on the nitrogen atom are taken from SCF
calculations of the ammonia molecule [14].
All of the direct contributions to the potentials have been derived from ab initio
calculations and are given for the ammonia-ammonia interactions by [15]:
VNN(r) =
894
r
+
33.554 · 105
r12
−
819
r6
,
VNH(r) = −
298
r
+ 0.6275 {exp[−4.6(r − 2.4)] − 2 exp[−2.3(r − 2.4)]} ,
VHH(r) =
99.4
r
+ 2929 exp(−3.7r) (2)
460
Structure of concentrated Li-ammonia solutions
and for the Li-Li and Li-ammonia interactions by [12]:
VLiLi(r) =
1390
r
− 18.71 exp(−1.62r),
VLiN(r) = −
1115
r
−
3516
r4
+ 120417 exp(−3.31r),
VLiH(r) =
372
r
+
49.31
r4
+ 2241 exp(−2.61r), (3)
where the energies are given in kJ/mol with r in Å. For further details of the po-
tentials and especially the indirect contributions [11], the reader is referred to the
references cited.
The results of an MD simulation of a saturated Li-ammonia solution with these
potentials has led to some discrepancies with the experimental findings. As a con-
sequence of these discrepancies we have modified the potentials by introducing a
3-body term for the Li-ammonia interactions as described in detail in reference [10],
although the simulation of a single lithium ion in 201 ammonia molecules did not
significantly change the Li-N radial distribution function (RDF). The 3-body term
employed is given by [10]:
∆E3bd = a exp[−b(r1 + r2) − cr3], (4)
where r1 and r2 are the distances between Li+ and each of the two nitrogen atoms
and r3 is the distance between the two nitrogen atoms. The fitting parameters are
found to be: a = 656.92 kJ/mol, b = 0.2261 Å−1, and c = 0.9645 Å−1. There was no
need to introduce a 3-body term for the ammonia-ammonia interactions as discussed
in reference [16].
With these potentials four different MD simulations have been performed, the
details of which are collected in table 1. Due to the strong screening of the Coulombic
interactions by the indirect contributions to the total potential (see reference [12])
there was no need to use the Ewald summation. Instead, the shifted force method
with a cut-off length of half the box size has been employed together with periodic
boundary conditions. The time step length was chosen to be 0.125 fs.
Table 1. Details of the simulations.
Conc. No. No. Density Temperature Sim. Time
[MPM] Li+ NH3 [g/cm3] [K] [ps]
19.6 50 205 0.498 241 10
11.8 27 202 0.544 241 10
6.4 14 205 0.618 241 10
0.5 1 201 0.690 240 10
461
S.Hannongbua et al.
3. Results and discussion
3.1. Radial distribution and structure functions
The site-site radial distribution functions (RDFs) together with the running
integration numbers are shown in figure 1 for four different lithium ion concentrations
according to table 1. This report presents exclusively the results for the simulations
including 3-body interactions. The characteristic values of the RDFs are presented
in table 2.
1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
1 2 3 4 5 6 7 8 9 10
0.00
0.25
0.50
0.75
1.00
r/Å
n(r)g(r)
Li-Li
0
1
2
3
4
5
6
7
8
0
10
20
30
40
50
n(r)g(r) Li-H
0
2
4
6
8
10
12
14
0
2
4
6
8
10
n(r)g(r) Li-N
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0
10
20
30
40
50
n(r)g(r) N-H
1 2 3 4 5 6 7 8 9 10
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
r/Å
g(r) n(r)H-H
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0
5
10
15
20
n(r)g(r) N-N
Figure 1. Site-site radial distribution functions, g(r), and running integration
numbers, n(r), from MD simulations of lithium-ammonia solutions for four dif-
ferent concentrations: 0.46 MPM (dotted), 6.4 MPM (dash-dotted), 11.8 MPM
(dashed), and 19.6 MPM (full).
462
Structure of concentrated Li-ammonia solutions
Table 2. Characteristic values of the radial distribution functions gij(r) for Li-
ammonia solutions at four different concentrations with the Li mole fractions ci:
0.0046, 0.064, 0.118, and 0.1958 denoted by 1, 2, 3, and 4, respectively. Rl, RMl,
and Rml are the distances where for the lth time gij(r) is unity, has a maximum,
or a minimum, respectively.
ij c R1 RM1 gij(RM1) R2 Rm1 gij(Rm1) RM2 gij(RM2) nij(R2) nij(Rm1)
NN 1 3.22 3.70 2.1 4.52 5.24 0.7 6.96 1.2 9.1 12.8
2 2.90 3.21 3.6 3.96 4.62 0.7 6.01 1.3 6.9 9.1
3 2.98 3.30 2.5 4.07 4.71 0.8 6.27 1.2 5.9 8.3
4 3.13 3.49 2.0 4.20 4.91 0.8 6.99 1.1 4.7 7.3
NH 1 3.18 3.74 1.4 4.74 5.38 0.8 7.06 1.1 29.1 40.9
2 2.36 2.77; 3.891 1.4; 1.5 4.55 4.73 0.9 5.98 1.2 26.8 29.6
3 2.77 2.92; 3.961 1.1; 1.4 4.62 4.88 0.9 6.14 1.1 24.2 28.2
4 3.39 4.07 1.4 4.73 5.21 0.9 7.37 1.0 20.2 26.8
HH 1 3.14 3.96 1.2 4.82 5.42 0.9 7.20 1.1 29.5 41.7
2 2.18 2.57; 3.411 1.1; 1.4 – – – – – – –
3 3.13 2.49; 3.741 0.8; 1.2 5.48 5.54 1.0 – – – 41.4
4 3.59 4.38 1.2 5.19 5.72 0.9 7.35 1.03 26.6 35.5
LiN 1 1.98 2.20 12.4 2.58 3.15 0.0 4.55 2.0 4.4 6.0
2 1.92 2.13 10.5 2.51 3.17 0.0 – – 3.8 4.0
3 1.92 2.13 10.6 2.51 3.10 0.1 – – 3.7 4.0
4 1.92 2.16 11.8 2.54 3.21 0.1 – – 3.7 4.0
LiH 1 2.43 2.75 6.3 3.14 3.55 0.1 5.0 1.6 17.2 18.0
2 2.21 2.57 3.5 3.06 3.72 0.2 – – 10.8 12.7
3 2.22 2.60 3.6 3.07 3.61 0.2 – – 10.9 12.6
4 2.21 2.64 4.1 3.10 3.54 0.3 – – 10.9 12.5
LiLi 1 – – – – – – – – – –
2 2.10 2.53 9.2 3.20 3.8 0.1 7.2 1.8 0.6 0.7
3 2.38 2.60 3.0 3.04 4.2 0.0 7.2 1.8 0.3 0.4
4 2.42 2.66 1.9 2.97 4.2 0.0 7.2 1.5 0.4 0.5
Peak splitting (see figure 1).
It can be seen from the Li-N RDFs that the first peak is most pronounced for
the single ion simulation where not only the first but also the second solvation
shell can be formed undisturbed by other lithium ions. The results for this dilute
case reproduce the findings presented already in reference [15] and proved to be
unchanged when 3-body interactions were included [10]. In accordance with these
findings the solvation number remained six while for the higher concentrations the
addition of the 3-body term reduced the solvation number to four. This change leads
to an agreement with the NDIS measurements of a saturated Li-ammonia solution
[6]. An experimental confirmation of a solvation number of six for the dilute solution
does not exist. The hydration number of the lithium ion is discussed controversially
463
S.Hannongbua et al.
as well. There are, similarly, indications for 6 at low concentrations and for 4 at
higher ones (see e.g. [17]).
It should be mentioned here that the NDIS measurements show a prepeak at
1 Å−1 which is only indicated in the simulation data at medium concentrations and
reduces to a shoulder at the saturated solution. This prepeak is expected to reflect
a medium range structure in real space but such structures can neither be seen
in the RDFs from the experiment nor in those from the simulation. There exists
only a well defined second solvation shell in the very dilute solution (see figure 1 and
table 2). There is – in the limits of uncertainty – no difference between simulation and
experiment in the positions of the first peaks in the Li-N and Li-H RDFs. It should
be noted that the first maximum in the Li-N RDF is higher for the saturated solution
than that for the 6.4 and 11.8 MPM ones, indicating a more pronounced solvation
shell, which might indicate the cluster formation with the increasing concentration.
Similar conclusions can be drawn from the Li-H RDFs. Here, as well as in the Li-N
case the first neighbor distances are slightly larger for the dilute versus concentrated
solutions (see table 2). While for the 6.4 and 11.8 MPM solutions both the Li-N and
Li-H RDFs very slowly but monotonously approach the uniform distribution, there
is an indication of a slightly further ranging structure for the saturated solution.
With the screening of the Coulomb interactions in the pseudopotential theory,
obviously the 3-body term in the Li-N and Li-H interactions becomes significantly
more important than in the single ion case where the electron is simply neglected.
This is very different from the simulation of the beryllium ion in water, where the
3-body term immediately reduced the hydration number from six to four even in
the single ion case [18,19].
The relatively small number of Li+ leads to a strong noise in the Li-Li RDF
but this limitation does not hide the obvious effect that the ion-ion structure is
significantly more pronounced for the 6.4 MPM solution than for the other two
concentrations. This special feature is also reflected in all three solvent RDFs (see
below). Consequently, in spite of the much smaller average ion density, the number
of the nearest neighbors is about 0.7 while it is only about 0.5 for the 11.8 and
19.6 MPM solutions. The first neighbor Li-Li distance is about the same for all
concentrations. The meaning of the Li-Li RDF for the solution structure is not yet
clear. It deserves a further analysis to be derived from the simulation data.
The most detailed experimental confirmation on the change of the solvent struc-
ture with the lithium ion concentration has been derived from the neutron diffraction
study with hydrogen isotope substitution [9]. The measured partial structure func-
tions (pstfs) and the RDFs for N-N, N-H, and H-H will be compared here with those
derived from simulations with lithium ion concentrations of 6.4, 11.8, and 19.6 MPM.
They are presented in figure 2. The comparison between simulation and experiment
shows that the main peaks in all three pstfs and all concentrations are in the range
1.9-2.2 Å−1 where for the highest concentration the experimental ones are slightly
shifted by about 0.1 Å−1 to larger k. There is an agreement between simulation
and experiment as far as the changes of position and height of the main peak with
concentration are concerned. The positions shift in all cases to lower k-values with
464
Structure of concentrated Li-ammonia solutions
the increasing lithium concentration while its height decreases for the N-N pstf and
increases for N-H and H-H with concentration.
Figure 2. Partial structure functions for
the N-N, N-H, and H-H interactions de-
rived from MD simulations of lithium-
ammonia solutions for three differ-
ent concentrations: 6.4 MPM (dotted),
11.8 MPM (dashed), and 19.6 MPM
(solid).
Simulation and experiment also show
agreement in respect to the shoulder at
about 2.5 Å−1 in the N-N pstf for the
saturated solution. In addition, the sec-
ond and the third small peaks at about
4 and 6 Å appear at similar positions.
The pstfs N-H and H-H from the simu-
lations practically disappear for k-values
larger than 6 Å−1, slightly different from
the experimental results. The only sig-
nificant discrepancy between simulation
and experiment concerns the prepeak at
about 1 Å−1. It is most obvious and well
pronounced in the total structure func-
tion for the saturated solution [9]. In the
simulation data, all three pstfs and at all
concentrations also show a prepeak, but
much less pronounced, which additional-
ly reduces to a shoulder when the concen-
tration is increased. It remains unclear
why this discrepancy has been found.
According to the overall agreement
even for the pstfs, a similar agreement
between simulation and experiment for
the partial RDFs is expected. The broad
peaks resulting from the NDIS measure-
ments [9] make it difficult to determine
the positions and heights of the peaks in
the RDFs and thereby the running in-
tegration numbers. For a better analysis,
Gaussian functions have been fitted [8,9].
The problems connected with this kind
of analysis become obvious when the po-
sitions and the areas under the various
peaks in the N-N RDF from the X-ray
measurement [8] and the NDIS experi-
ment [9] are compared. The same infor-
mation is easily deduced from the simu-
lation. The positions of the peaks and the
number of the nearest neighbors (run-
ning integration up to the first minimum) for all partial RDFs and the different
concentrations are given in table 2.
465
S.Hannongbua et al.
The position of the main peak in the N-N RDF is shifted to shorter distances
by about 0.15 Å and its height decreases in going from the dilute to the saturated
solution. The integration of the first peak up to about 5 Å (the first minimum)
leads to a change in the number of the nearest neighbors from about 12 to 7.5
(figure 1 and table 2). This means a very good agreement with the experimental
results, considering the difficulties in the analysis of the measured pstfs mentioned
in the preceding paragraph [8,9]. The N-N RDFs show a remarkable feature. The
first and the second neighbor shells are most pronounced for the 6.4 MPM solution.
It indicates – together with similar features in the N-H and H-H RDFs – a special
solvent structure in the range of medium concentration which is not reflected in
the experimental results and has yet to be analyzed more in detail. The medium
distance range order in real space expected from the prepeak at about 1 Å−1 in
the experimental N-N pstf is not reflected in the corresponding RDF [9]. It might
follow again from the difficulties in the analysis of experimental pstfs mentioned
above. It remains unclear whether the shallow maxima and minima beyond 6 Å in
the simulated N-N RDF lead to the weak prepeak found also in the simulated N-N
pstf at 1 Å−1 (figure 2).
The characteristic data for the N-H RDF show – again in the limits of uncertainty
– good agreement between simulation and experiment [9] for the dilute and the
saturated solution. This is true not only for the position of the main peak but also for
the number of the nearest neighbors. In both investigations a pronounced shoulder
at about 2.5 Å is obvious which reduces with the increasing concentration. The
formation of the shoulder has been discussed in reference [15] for the dilute solution.
It is a result of the existence of two kinds of hydrogen atoms. The ones at the shorter
distances are those which form bonds with the neighboring molecule. This is very
similar to the water case where the integral over the first part of the double peak
has often been employed as a measure of the number of hydrogen bonds formed.
But it has to be kept in mind that a distance criterion alone cannot be a correct
measure of hydrogen bonding. The relative orientation of the two molecules has to
be included in the definition. But having employed this estimation here would mean
that the largest number of hydrogen bonds between solvent molecules is formed in
the 6.4 MPM solution (see table 2 and reference 20) in accordance with the result
for the N-N RDF mentioned above. The positions of the main peaks of the H-H
RDFs again agree between simulation and experiment but the change of the RDF
with concentration is much more detailed in the simulation when compared with
the experiment. This is again especially true for the solution with the concentration
of 6.4 MPM.
3.2. Structure of the solvation shells
The geometrical arrangement of the ammonia molecules in the solvation shells
of the lithium ions can be deduced from the simulation by the calculation of the
distribution of cos ϑ, where ϑ is defined as the N-Li-N angle. The result is depicted
in figure 3, where only the ammonia molecules in the first solvation shell are included
in the distribution.
466
Structure of concentrated Li-ammonia solutions
Figure 3. Distribution of cos ϑ –
where theta is defined as the nitrogen-
lithium-nitrogen angle (see insertion) –
calculated for the ammonia molecules
in the first solvation shell of the lithi-
um ion for four different concentra-
tions: 0.46 MPM (dotted), 6.4 MPM
(dash-dotted), 11.8 MPM (dashed),
and 19.6 MPM (full).
Figure 4. Normalized distributions of
cos θ for the ammonia molecules in
the first solvation shell of the lithi-
um ion for four different concentra-
tion: 0.46 MPM (dotted), 6.4 MPM
(dash-dotted), 11.8 MPM (dashed),
and 19.6 MPM (full). θ is defined in
the insertion.
It is obvious from figure 3 that the six ammonia in the dilute solution are arranged
octahedrally around the lithium ions. Integration over the distribution shows that
there is another solvation shell molecule at about 180◦ and four more at about
90◦. For all three higher concentrations the solvation number is four (table 2). The
distributions of cos ϑ for these concentrations are very similar. There exists only one
broad maximum around 102◦ indicating not very well pronounced nearly tetrahedral
arrangement of the ammonia molecules in the first solvation shells.
The orientation of the ammonia molecules in the solvation shells of the lithium
ions can be described by the distribution of cos θ, where θ is defined as the angle
between the dipole moment vector of the ammonia molecule and the vector pointing
from the nitrogen atom towards the ion. The normalized distributions are depicted
in figure 4.
For all four concentrations there is a preference for the dipole moment of the
ammonia molecule pointing away from the ion. The distributions are quite similar for
all three concentrated solutions but much broader than for the single ion case where
solely the interaction between the solvation shell molecule and the ion determines
the orientation because of the relatively weak interactions between the solvation
shell molecules and those in the bulk. In the concentrated solutions, the solvation
shells of different ions partly overlap and a competition between the two ions for the
energetically most favorable orientation leads to a broad distribution.
467
S.Hannongbua et al.
4. Conclusions
The most important change caused by the introduction of three-body contribu-
tions is the reduction of the solvation number of Li+ from 6 to 4 for the concentrated
solutions. It remained 6 for the dilute one. This change of the solvation number with
concentration in the ammonia solution is similar to what has been found for this ion
in various aqueous solutions (see e.g. [17]). Very recent neutron diffraction studies
with hydrogen isotope substitution [9] provide the possibility for a detailed compar-
ison between simulation and experiment as far as the change of the solvent structure
with the lithium ion concentration is concerned. There is an overall good agreement
for the N-N, N-H, and H-H partial structure functions except for some differences
in the prepeak near 1 Å−1. But this discrepancy is not reflected in the partial radial
distribution functions.
The N-N, N-H, and H-H RDFs from the simulations indicate the enhancement of
the solvent structure in the medium concentration range, especially for the 6.4 MPM
solution. This surprising feature is not obvious in the partial structure functions.
Considering the overall good agreement between simulation and experiment, it is
not to be expected that this special solvent structure is an artifact resulting from the
simulation. Therefore, a further detailed analysis of the data and, most probably,
additional simulations seem to be appropriate to clarify this interesting feature.
Acknowledgements
Financial support by the Russian Fund for Basic Research through Grant
No. RFBR-99-03-32064, the INTAS Project No. 96-1989, the Thailand Research
Fund, and the Chulalongkorn University is gratefully acknowledged.
References
1. Hannongbua S., Kerdcharoen T. // J. Chem. Phys., 1992, vol. 96, p. 6945.
2. Hannongbua S. // J. Chem. Phys., 1997, vol. 106, p. 6076.
3. Sidhisordej W., Hannongbua S., Ruffolo D. // Z. Naturforsch, 1998, vol. 53a, p. 208.
4. Tongraar A., Hannongbua S., Rode B.M. // Chem. Phys., 1997, vol. 219, p. 279.
5. Chieux P., Bertagnolli H. // J. Phys. Chem., 1984, vol. 88, p. 3726.
6. Wasse J.C., Hayama S., Skipper N.T., Fischer H.E. // Phys. Rev. B, 2000, vol. 61,
p. 11993.
7. Wasse J.C., Hayama S., Skipper N.T., Benmore C.J., Soper A.K. // J. Chem. Phys.,
2000, vol. 112, p. 7147.
8. Hayama S., Skipper N.T., Wasse J.C., Thompson H. // J. Chem. Phys., 2002, vol. 116,
p. 2991.
9. Thompson H., Wasse J.C., Skipper N.T., Hayama S., Bowron D.T., Soper A.K. //
JACS, 2003, vol. 125, p. 2572.
10. Hannongbua S. // Chem. Phys. Lett., 1998, vol. 288, p. 663.
11. Gurskii Z., Hannongbua S., Heinzinger K. // Mol. Phys., 1993, vol. 78, p. 461.
468
Structure of concentrated Li-ammonia solutions
12. Hannongbua S., Kokpol S., Gurskii Z., Heinzinger K. // Z. Naturforsch., 1997, vol. 52a,
p. 828.
13. Benedict W.S., Plyler E.K. // Can. J. Phys., 1985, vol. 35, p. 890.
14. Stillinger F.H. // Israel J. Chem., 1975, vol. 14, p. 130.
15. Hannongbua S., Ishida T., Spohr E., Heinzinger K. // Z. Naturforsch., 1988, vol. 43a,
p. 572.
16. Hannongbua S. // J. Chem. Phys., 2000, vol. 113, p. 4707.
17. Egorov A.V., Komolkin A.V., Chizhik V.I., Yushmanov P.V., Lyubartsev A.P., Laak-
sonen A. // J. Phys. Chem. B, 2003, vol. 107, p. 3234.
18. Probst M.M., Spohr E., Heinzinger K. // Chem. Phys. Lett., 1989, vol. 161, p. 405.
19. Marx D., Sprik M., Parinello M. // Chem. Phys. Lett., 1997, vol. 273, p. 360.
20. Kiselev M., Kerdcharoen T., Hannongbua S., Heinzinger K. // Chem. Phys. Lett.,
2000, vol. 327, p. 425.
469
S.Hannongbua et al.
Структура концентрованих розчинів Li-аміак.
Молекулярна динаміка
С.Ганоннґбуа 1 , Т.Ремшугнен 1 , М.Кісєльов 2 ,
К.Гайнцінгер 3
1 Відділення хімії, природничий факультет,
Університет Чулалонгкорн,
Бангкок 10330, Таїланд
2 Інститут хімії розчинів, Російська Академія Наук,
153045 Іваново, Росія
3 Інститут хімії Макса Планка (Інститут Отто Гана),
55020 Майнц, Німеччина
Отримано 16 травня 2003 р.
В даній роботі представлені результати, отримані методом молеку-
лярної динаміки для розчинів літій-аміакy для усього діапазону кон-
центрацій від 0.5 до 19.6 MPM при температурі 240 К. Використову-
ється теорія псевдопотенціалів при вищих концентраціях. Прямий
внесок загального потенціалу, який отримується з ab initio розрахун-
ків, доповнюється тричастинковим доданком. Результуюча частин-
на структура і радіальні функції розподілу детально порівнюються
з рентгенівським і нейтронним розсіюванням з ізотопним заміщен-
ням. Результати експерименту і моделювання добре узгоджуються.
Знайдено сольватаційні числа іонів літію для розведених розчинів
(шість), а також для розчинів з вищою концентрацією (чотири) по
аналогії з гідратацією іонів літію у різних водних розчинах.
Ключові слова: метод молекулярної динаміки, розчини літій-аміак,
теорія псевдопотенціалів
PACS: 61.20.J, 61.20.Qg, 71.15.Pd
470
|