Compressibility Anisotropy and Electronic Properties of Oxyanionic

Aug 14, 2017 - The crystal structure of lithium perchlorate trihydrate (LPT) has been determined both from X-ray and neutron diffraction data.(7, 8) A...
0 downloads 12 Views 2MB Size
Subscriber access provided by AUSTRALIAN NATIONAL UNIV

Article

Compressibility Anisotropy and Electronic Properties of Oxyanionic Hydrates Dmitry Vasilievich Korabel'nikov, and Yuriy N. Zhuravlev J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.7b04776 • Publication Date (Web): 14 Aug 2017 Downloaded from http://pubs.acs.org on August 15, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Compressibility Anisotropy and Electronic Properties of Oxyanionic Hydrates Dmitry V. Korabel’nikov*1, Yuriy N. Zhuravlev Institute of Fundamental Sciences, Kemerovo State University, Krasnaya 6, 650043, Kemerovo, Russia ABSTRACT

The structural and electronic properties of oxyanionic crystalline hydrates, LiNO3·3H2O, LiClO4·3H2O and NaClO4·H2O, have been studied using density functional theory including van der Waals interactions. It is established that the linear compressibility of lithium perchlorate trihydrate is anisotropic (a < c) and positive, while lithium nitrate trihydrate and sodium perchlorate monohydrate demonstrate negative linear compressibility along the b and c- axes, respectively. Deformation of Ow-H···O hydrogen bonding motifs is correlated with the negative linear compressibility. The band gaps of lithium nitrate and lithium perchlorate trihydrates decrease with pressure, whereas band gap of sodium perchlorate monohydrate increases. Introduction Water is an exclusive agent that has many abnormal properties. Most of these abnormal properties are due to the fact that water molecules form an extended three dimensional network of hydrogen bonds. Aqueous salt hydrates have become attractive model systems for confined water molecules characterized by a well-defined geometrical arrangement. A well-defined geometrical arrangement and an accordingly well-defined strength of hydrogen bonds is typical for crystal water in hydrated salts. In recent years the most attention has been drawn to the hydrogen bond of lithium nitrate hydrate.1-3 Infrared (IR) spectroscopy is a powerful tool to study interactions between water molecules as the OH stretch vibrations greatly depends on the strength of hydrogen-bond (HB): the stronger the hydrogen bond, the lower its frequency.4 In ref. 1 it was described that the superposition of the stretching bands of the hydroxyl groups involved in strong, weak, and bifurcated HBs caused the OH stretching region of the IR spectrum of lithium nitrate trihydrate. Lithium nitrate trihydrate (LNT) is an inorganic salt, its lattice forces confine water molecules into a characteristic crystalline arrangement. The crystal structure of LNT has been earlier determined from X-ray and neutron diffraction data.5,6 The crystal structure refers to the orthorhombic group Cmcm with two various types of water molecules. The number of formula units is Z=4. The crystal bc- plane, which is indicated as (b) in Fig. 1, contains the nitrate anion, lithium cation, and one of the two nonequivalent water molecules. Bifurcated hydrogen bonds (O4-H3...O1 and O4-H3...O2) to two nonequivalent oxygen atoms of a nitrate anion are formed by this water molecule. Type (b) adjacent planes are interconnected transversely by water molecules of different type. They form the ab planes which are shown in Fig. 1 as (a). A relatively strong hydrogen bond (O3-H2...O4) to the water molecule in plane (b) is formed by one *

Corresponding author. Tel./fax: +73842 583195. E-mail address: [email protected] (D.V. Korabel’nikov) 1

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 22

of the OH groups. At the same time a relatively weak hydrogen bond (O3-H1...O2) to the nitrate anion is formed by the other OH group.

Figure 1. Crystal structures of LNT ((a)- ab plane, (b) - bc plane), LPT (c), SPM (d).

The crystal structure of lithium perchlorate trihydrate (LPT) has been determined both from X-ray and neutron diffraction data.7,8 At ambient conditions LPT crystallizes into hexagonal structure P63mc (Fig. 1(c)) with two formula units. Formula units consist of a lithium cation, a perchlorate anion and three water molecules. Perchlorate anion contains two non-equivalent oxygen atoms (O1, O2). Unlike LNT, lithium perchlorate trihydrate in ab plane has one type of water molecules linked by O3-H...O2 hydrogen bonds to perchlorate anions. As in case of LNT and LPT, the crystal structure of sodium perchlorate monohydrate (SPM) has been determined both from X-ray and neutron diffraction data.9,10 At ambient conditions SPM crystallizes into monoclinic structure C2/c (Fig. 1(d)) with eight formula units. The unit cell contains two nonequivalent sodium cations, perchlorate anions with four nonequivalent oxygen atoms (O1,O2,O3,O4) and water molecules with two nonequivalent hydrogen atoms (H1,H2). The water molecules are linked via O5-H1...O3 and O5-H2...O1 hydrogen bonds to perchlorate anions. One of the OH groups of water molecule donates a relatively strong hydrogen bond (O5-H2...O1), while the other OH group donates a relatively weak hydrogen bond (O5-H1...O3). 2

ACS Paragon Plus Environment

Page 3 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The study of materials behavior under extreme conditions is of special interest.11 A thorough knowledge of structure and properties under pressure is necessary. Only a small number of materials exhibit negative linear compressibility (NLC), so these materials are very much desirable for applications, such as ultrasensitive pressure detectors, shock resistance materials, artificial muscles, body armor, pressure switches.12-16 NLC effects were observed in some hydrates. Gagnon et al. measured the lattice parameters of hydrated ZAG-4 (Zn(HO3PC4H8PO3H)·2H2O) under pressure with single crystal X-ray diffraction and observed NLC.17 They used the 'wine-rack' motif to explain the NLC of hydrated ZAG-4 after the proton transition. However, recently, Hui Wang et al. using first-principles calculations established that this NLC is explained via the deformation of H3O+ tetrahedron.18 At the same time, calculations for dehydrated ZAG-4 show no NLC. This means that water molecules are essential to the NLC of ZAG-4. Also, relatively recently, NLC was discovered for methanol monohydrate in which the C-D···O hydrogen bonds acted as hinges to form the “wine-rack” motif.19 Recently, using experimental measurements and theoretical calculations Bo Zou et al. demonstrated that the NH···O hydrogen bonding “wine-rack” motifs result in the NLC in oxyanionic ammonium oxalate monohydrate ((NH4)2C2O4·H2O).20 However, no study of structure for oxyanionic hydrates LiNO3·3H2O, LiClO4·3H2O, NaClO4·H2O under pressure was reported. Electronic structure is also important, which is intimately related to the fundamental physical and chemical properties of crystals. However, the electronic properties of LNT, LPT, SPM have not been studied so far. At the same time the structure and electronic properties at ambient conditions and under pressure have been studied for anhydrous nitrates and perchlorates.21-28 It was noted that the band structure is characterized by practically flat valence bands and unoccupied bands of localized anionic excitations. The molecular anionic bands and the subsequent unoccupied bands are separated by an energy space. It is established from the first principles calculations that to interpret the optical properties, intramolecular excitations are very important. Anisotropy compressibility and band gap increase with pressure have also been established for lithium nitrate, lithium and sodium perchlorate.27, 28 At the same time NLC effects for LiNO3, LiClO4 and NaClO4 have not been observed. It is interesting to study the influence of crystal water on the structural and electronic properties of the oxyanionic crystals. The determination of the exact location of hydrogen atoms, especially by X-ray diffraction, is a well-known problem.29,30 Neutron-diffraction measurements31,32 or first-principle calculations33 contribute a more detailed atomic structure of hydrogen-containing compounds. However, even neutron-diffraction measurements are not enough to receive reliable structural information because of the thermal libration effect (Ow-H shortening) produced by the large librational freedom of the water molecule.10,34 The crystal structures of LNT, LPT, SPM at ambient conditions and under pressure, using ab initio calculations, are still not identified. Understanding of the pressure effects on the hydrogen bonds is important. The H-bond transformations data is also necessary for designing new functional materials,35 such as ferroelectrics and relaxors36-38 for electronic applications. The study of the microscopic properties of energetic materials with complex chemical behavior presents certain difficulties. Theoretical calculations are an effective way to model the physical and chemical properties of solids on the atomic level. Furthermore, systematic computational procedures are more and more often used to predict or design new materials11, including NLC materials.16,18,20,39-43 Density functional theory (DFT) proved its efficiency in the simulation and prediction of the physical and chemical properties of a wide range of materials.44 Quite reasonable agreement with the experimental data is generally observed in the results obtained via DFT calculations. In case when crystals contain molecular complexes taking into account weak intermolecular van der Waals interaction is important. After correction of exchange correlation functional to describe weak intermolecular interactions, DFT can predict the properties for energetic materials, even for alkali-metal nitrates, chlorates and pechlorates 3

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 22

anhydrates, that will have good agreement with the experiment.27, 28, 45-56 Over recent years, the study of structures and properties of various materials under hydrostatic compression has been successfully conducted using density functional theory (DFT) method with dispersion correction.27, 28, 45-53, 57-59 In this paper we present the compressibility and electronic properties of hydrated oxyanionic crystals, LiNO3·3H2O, LiClO4·3H2O, NaClO4·H2O, based on DFT with dispersion correction.

Computational method In this paper, theoretical methods based on the density functional theory were used to study geometry and electronic structure for LNT, LPT and SPM. Calculations were carried out with the CRYSTAL code60 both at ambient pressure and under hydrostatic compression. The advantage of standard software packages and methods is that they have been thoroughly tested. It should be noted that CRYSTAL code was also used to study even elastic properties of NLC materials.39, 42, 43, 50 The basis set of linear combinations of atomic orbitals (LCAO) is used in the code. We used LCAO basis sets61-66 and generalized-gradient approximation functional of Perdew, Burke, and Ernzerhof (PBE-GGA)67 in our calculations. To include the long range intermolecular interactions, semiempirical dispersion correction schemes, integrated in the standard DFT description, have been used. Well-known Grimme scheme,68 employed via the functional, was applied in the present study. The functional is an empirical correction to DFT, which takes into account the dispersive interactions based on damped atomic pairwise potentials. In concordance with the semi-empirical dispersion correction of the DFT-D method, the system total energy can be expressed as EDFT-D= EDFT+ Edisp, where EDFT is the usual self-consistent Kohn–Sham energy, obtained from the chosen functional, while Edisp is an empirical dispersion correction.68 Minimization of the total energy, interatomic forces and displacements were used in ab initio determination of the crystal structure. As consistent with the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm,69 all atoms and cell parameters were allowed to relax. The Mulliken scheme was used in the calculations of the populations of the atoms electron shells and the overlap (the bond populations). Convergence was everywhere better than 10-7 eV.

Results and Discussion Although DFT-D method, which takes into account the dispersive vdW interactions, is not new, however, there are relatively few works regarding DFT-D calculations for ionic-molecular crystals.27,28,56,59 We have performed DFT-D calculations for inorganic oxyanionic hydrates in the present work. Table 1 exhibits the equilibrium structural parameters for LNT, LPT, SPM hydrates at ambient pressure, calculated by DFT and DFT-D methods with the PBE functional.67 The existing experimental6,8,10 data have also been delivered for comparison. Compared to the experimental data, the hydrates lattice parameters calculated by DFT method are overestimated. As regards the DFT-D method, with account of the dispersion correction68 the hydrates lattice parameters decrease (especially for the axes perpendicular nitrate-anions and water molecules), and in some cases are less than experimental values as the calculations do not take into consideration the thermal expansion and respond to T = 0 K. Thus, it is seen from Table 1 that Van-der-Waals interaction is important for LNT, LPT, SPM hydrates and taking into account the dispersion correction more reasonable results are obtained.

4

ACS Paragon Plus Environment

Page 5 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Table 1. Computed and experimental lattice parameters for hydrates at ambient pressure. Hydrates LNT

LPT

SPM

Method DFT DFT-D Expt.6 DFT DFT-D Expt.8 DFT DFT-D Expt.10

a(Å) 6.897 6.616 6.713 7.783 7.676 7.719 15.970 15.124 15.542

b(Å) 12.574 12.489 12.669 7.783 7.676 7.719 5.684 5.466 5.540

c(Å) 6.112 6.075 5.968 5.678 5.416 5.453 11.232 11.155 11.046

β(O) 90 90 90 120 120 120 111.26 110.28 110.67

V(Å3) 530.05 501.96 507.56 297.87 276.36 281.38 950.18 864.99 889.87

The shortest and strongest calculated LNT hydrogen bond is O3-H2...O4 bond between water molecules (H2-O4 length is 1.77 Å). Relatively weak hydrogen bond is O3-H1...O2 bond between water molecule and nitrate anion (H1-O2 length is 1.80 Å). The longer bonds correspond to bifurcated hydrogen bonds O4-H3...O1 and O4-H3...O2 with difference in H-O lengths ~ 0.3 Å. This HBs types are in accordance with IR measurements.1 Because H2 atoms are involved in stronger hydrogen bonds than H1, the O3–H2 bond length (1.000 Å) is longer than O3–H1 (0.987 Å). Also for LNT the N–O2 (1.287 Å) bond length is longer than N–O1 (1.268 Å) because the hydrogen bond is stronger for O2 than that for O1. As to LPT, all hydrogen atoms involved in equivalent hydrogen bonds O3-H...O2 with H-O2 lengths equal to 1.965 Å. So, LPT hydrogen bonds are weaker than LNT ones. Because O2 atoms are involved in hydrogen bonds, the Cl–O2 bond length is longer than Cl–O1 by value of 0.016 Å. In SPM the water molecule is linked to perchlorate anions via HBs, one of which (O5H2...O1) is relatively strong (1.856 Å), while the other (O5-H1...O3) is relatively weak (2.039 Å). Therefore, the O5–H2 bond length (0.992 Å) is longer than O5–H1 (0.985 Å) and the Cl–O1 bond length is longer than Cl–O3 by value of 0.014 Å. The relative standard (root mean square) deviation between calculated and X-ray5,7,9 hydrogen bond lengths H-O at ambient pressure is 13.9, 14.4 and 23.9 % for LNT, LPT and SPM, respectively. These differences are due to the well-known problem, which is the X-ray diffraction determination of the exact location of hydrogen atoms.29, 30 Corresponding root mean square deviation for calculated and neutron diffraction6,8,10 HB lengths are smaller values ~ 5.8, 5.3 and 14.0 %. Root mean square deviations reduced from 29.5, 36.5, 58.2 % (X-ray diffraction) down to 3.1, 3.6, 7.6 % (neutron diffraction) for Ow-H covalent bond lengths of LNT, LPT and SPM, respectively. As it was pointed out above, some differences between the calculated and neutron diffraction data are due to the thermal libration effect produced by the large librational freedom of the water molecule.10, 34 The calculated pressure dependencies of relative lattice parameters (a/a0, b/b0, c/c0) and relative volumes (V/V0) for LNT, LPT, SPM can be seen in Fig. 2. The calculated bulk moduli for LNT, LPT, SPM are 22.6, 23.0, 17.9 GPa, respectively. Thus, the bulk modulus of hydrated perchlorates decreases with cation radius increase. It should be noted that bulk moduli for hydrates under consideration are ~ 1.5-2 times less than for the corresponding anhydrates.27, 28 As it can be seen from Fig. 2, the linear compressibility of lithium perchlorate trihydrate along the crystallographic axes is anisotropic. For c-axis the compressibility is maximum and for baxis is minimum. It should be noted that Fig. 2 does not show the pressure dependency for a/a0 of lithium perchlorate trihydrate, because a=b. Compressibility anisotropy is associated with bond anisotropy. It indicates a strong intralayer bond (Ow-H covalent, hydrogen) in the ab plane of water molecules and a weaker interlayer bond (van der Waals, covalent Cl–O1 bonds). Let us 5

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 22

remark here that in ref. 27, 50, 52 strong anisotropy for nitrates and TATB is also proven. It can be due to the fact that short interlayer covalent bonds are less compressible than long intralayer bonds. The compressibility of LiClO4·3H2O can be indirectly estimated by the elastic constants measured by Haussuhl.70 We should point out that LPT compressibility anisotropy has agreement with the experimental measurements of elastic constants,70 where c11 = c22 > c33. At the same time from LPT elastic constants it follows70, 71 that linear compressibilities are positive which is in agreement with our calculations.

Figure 2. Pressure dependences of relative unit cell dimensions a/a0 (dashed lines), b/b0 (bold lines), c/c0 (solid lines), V/V0 (bold lines), hydrogen bond distance H-O (bold lines) for LNT, LPT, SPM from DFT-D.

Compressibility anisotropy for LNT and SPM is also observed, it is maximum for a-axis (Fig. 2). Moreover, there is quite interesting peculiarity of LNT and SPM, that is abnormal increase of b and c parameters, respectively, with pressure (NLC). The linear expansion under pressure for LNT and SPM at 2 GPa is ~ 0.5 and 0.8 %, respectively. As negative (dV/dP) is necessary due to the thermodynamics requirements,12, 72 NLC is connected with large positive compressibility along at least one direction. Thus, positive compressibility along the a- axis has a compensating effect for LNT and SPM (~ 6% decrease at 2 GPa). It should be pointed out, that the first 6

ACS Paragon Plus Environment

Page 7 of 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

principle calculations also allow discovering NLC in other crystals,18,20,42,50,57 including hydrates, which agrees with the experimental results. So, Cai and Katrusiak have experimentally discovered strong NLC for crystal Ag(en)N,73 which then was confirmed from theoretical DFTD calculations.50 It is especially remarkable, that Ortiz, et al.42 predicted, based on DFT-D calculations, the NLC behavior of soft porous crystal MIL-53(Al), which was experimentally proven two years later by Gascon, et al.74 Negative linear compressibility (NLC) is not a common phenomenon for hydrates. NLC has been so far discovered only for three hydrates, which are metal-organic hydrated ZAG-4, organic methanol monohydrate and ammonium oxalate monohydrate.17,19,20 In the present work we have shown that inorganic hydrates LiNO3·3H2O, NaClO4·H2O also demonstrate NLC. Pressure-induced transformations in hydrogen bonds are particularly interesting, the interest has been instigated by the search of substitutes for ceramic ferroelectrics, ferroelectric relaxors and other materials for electronic and optoelectronic applications. These materials can be more practical in terms of their cost and environmental protection and useful compared to traditional perovskite ceramics impregnated with lead. Fig. 2 shows pressure dependencies of hydrogen bond lengths H-O for LNT, LPT and SPM. The atoms numeration is in accordance with Fig. 1. It can be seen that all H-O lengths of hydrates decreased with pressure, therefore HBs becomes stronger. At the same time, for LNT, SPM the differences between H1-O2 and H2-O4, H1-O3 and H2-O1 lengths become, respectively, smaller. Moreover, we considered pressure behavior of Ow-H···O hydrogen bonding motifs in LNT, LPT and SPM hydrates. Fig. 3 shows pressure dependences of interatomic distances for HB motifs in LNT, LPT and SPM.

Figure 3. Pressure dependences of interatomic distances for HB motifs in LNT, LPT and SPM. It can be seen that all LPT interatomic distances (Ow1-O3, O1-O2, Ow2-Ow3) decreased and HB "honey-comb" motif contracted in all directions. So, LPT shows no NLC behavior under pressure. In contrast to LPT, LNT interatomic distance in a direction (O1-Ow2) decreased, whereas in b direction (Ow1-Ow3 distance) increased. Thus, LNT hydrogen bonding motif pressure behavior is similar to "wine-rack" one and correlates with NLC along b axis. Similar NLC mechanism was reported only for organic ammonium oxalate monohydrate.20 This NLC mechanism has not been previously reported for inorganic crystals. Baughman et al.13 identified that structures, containing a “wine-rack” topology, were among those which, as predicted, 7

ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 22

exhibited negative linear compressibility, because in wine-rack one diagonal decreased whereas another increased. SPM hydrate has zigzag chains linked by water molecules. It can be seen (Fig. 3), L2 distance decreased, whereas L1 distance increased under pressure that lead to expansion of SPM hydrogen bonding motif in c direction and contraction in b direction. Therefore, SPM hydrogen bonding motif pressure behavior is similar to "zigzag" chain motif (expand in one direction and contract in another) and correlates with NLC in c direction. It should be noted that hydrogen bonding NLC motifs differ from other NLC motifs hinged by strong covalent bonds. The novelty of the presented work is to discover negative linear compressibility of inorganic oxyanionic hydrates LiNO3·3H2O, NaClO4·H2O (Fig. 2) and its relation with hydrogen bonding motifs (Fig. 3). The character of the chemical bond can be determined by studying the electron density distribution. Fig. 4 shows the total electron density distributions for HB motifs (marked by lines) in LNT (a), LPT (b), SPM (c). The interval between isodensity curves is 0.01 e/bohr3.

Figure 4. Total electron density distributions for HB motifs in LNT (a), LPT (b) and SPM (c). It is seen that inside the nitrate anions (between nitrogen and oxygen), perchlorate anions (between chlorine and oxygen) and water molecules (between hydrogen and oxygen) the chemical bond is significantly covalent. The atoms have a large number of common closed contours which is characteristic of the covalent bond. An interesting feature here is that the water hydrogen atoms are involved in the hydrogen bond with oxygen atoms of the nitrate (perchlorate) anions and other water molecules. In HB motifs water molecules and anions (or different water molecules) have common contours, which give evidence of the covalent component of the Ow-H···O hydrogen bonds. In particular, it is related to LNT water molecules having relatively strong O3-H2···O4 bonds and practically four common contours. Corresponding hydrogen bond angle is 177o which also indicates covalent component of HB because bond directionality is a characteristic feature of the covalent bond. It should be noted that the hydrogen bonds with small lengths (