Figure 1 shows typical Raman spectra of [TeX
6](Y
2) at room temperature. The positions of the band maxima are indicated only for the stretching vibrations of Cl
2 (spectrum 1), Br
2 (2) and I
2 (3). The remaining modes relate to vibrations of [TeX
6]
2− octahedra, the study of which is beyond the scope of the present work and are not considered here. In some cases, the spectra also contain vibrational lines of counterions, which are also not considered.
Table 1 lists the vibration wavenumber of Y
2 and the length of the X–I and I–I bonds. As can be seen from the table, in [TeX
6](I
2) compounds, which represent a complete series of halogen type X, the mode frequency ν(I
2) depends significantly on X, and the interpretation of this effect is one of the goals of this work.
Figure 2 shows vibration wavenumber of Cl
2, Br
2 and I
2 molecular halogens vs temperature for three different [TeX
6](Y
2) compounds. Vibration frequencies of ν(Cl
2) (
Figure 2a) and ν(Br
2) (
Figure 2b) decrease upon cooling down. This behavior is anomalous since, as is well known, the vibration frequency of a conventional anharmonic oscillator should increase with decreasing temperature (see, for example, Ref. [
29,
30]). An interpretation of the observed anomaly is another goal of the work. At the same time, the frequency of the ν(I
2) mode in [TeI
6](I
2) (
Figure 2c) shows the expected positive shift with decreasing temperature, and this difference in the temperature behavior of the vibration frequencies of the molecular halogens also needs to be interpreted.
The outer electronic shell of halogen atoms is
s2p5. All molecular orbitals in the Y
2 molecule are occupied except for the σ*(2
px) term (LUMO), which is entirely unoccupied (
Figure 4). On the other hand, halogen X in the [TeX
6]
2− complex has a saturated outer electron shell (HOMO). Its charge state is X
− due to bonding with the complexing metal Te(IV) and charge migration from the counterion in the crystal lattice. An excess negative charge of the halogen atom in the [TeX
6]
2− complex provokes the spreading of electron density to antibonding σ*(2
px) orbital of Y
2. It will results in a weakening of the intramolecular Y−Y bond and a decrease in Y
2 vibrational frequency.
Charge spreading to the unoccupied antibonding orbital σ*(2
px) of the I
2 dimer in [TeX
6](I
2) compounds is limited by the energy ratio of the dimer LUMO and the [TeX
6]
2− HOMO. The less the energy of the electronic term X
− in [TeX
6]
2− compare to the energy of σ*(2
px) orbital of Y
2, the smaller the charge shift to the Y
2. The outer electronic shell of Cl atoms is 3
s23
p5. On the energy scale, these states are located significantly lower than the state of the antibonding orbital σ*(2
px) of the I
2 dimer, whose electronic configuration is 5
s25
p5. For this reason, the charge transfer from Cl
− to I
2 in [TeCl
6](I
2) is negligibly small. The same, but to a lesser extent, applies to [TeBr
6](I
2). However, the HOMO energy of the I
− ion is quite comparable to the LUMO energy of the I
2 dimer, and the charge spreading from I
− to I
2 in [TeI
6](I
2) will be more significant than in the two previous cases. It agrees with the observed dependence of ν(I
2) on the type of halogen X in the [TeX
6](I
2) series (
Table 1,
Figure 3).
Thus, the assumption of charge transfer from the surrounding cavity and from the halogen atom X
− in [TeX
6](I
2) compounds to the unoccupied orbital of the Y
2 dimer is the only correct one for the interpretation of the experimental data given in
Table 1 and
Figure 3. Another reliable experiment confirming the proposed mechanism of interaction of halogen atoms in the contact X
−−Y
2 is the temperature dependence of the Y
2 vibrational frequency. It is presented in the next paragraph
Temperature Dependence of ν(Y2) in [TeX6](Y2)
The temperature of the crystal determines the population of the vibrational states of Y
2 molecules. As the temperature decreases, the vibrations freeze out, i.e., decrease in vibrational quantum number. In an anharmonic oscillator, which is described by the 6-12 potential of Lennard-Jones, it should result in a shortening of bond lengths, i.e., Te−X, X−Y and Y−Y bond lengths in the case, and an increase in their vibration frequency (see Ref. [
29,
30] for details). On the other hand, the vibration frequency of Y
2 should fall at the X−Y bond lengths shortening due to an increase in the population of the σ*(2
px) orbital of the Y
2. The compromise between these two processes determines the temperature dependence of the vibration frequency ν(Y
2). It is the temperature dependence that makes it possible to determine which of the two processes, anharmonicity of vibrations or a change in the population of the antibonding σ*(2
px) orbital, prevails in the temperature dependence of the ν(Y
2).
The vibration frequency of Cl
2 in the lattice is about 540 cm
-1 (
Table 1). The population of the first exited vibrational state is negligible (about 0.08 at room temperature). In other words, in the temperature range 5 K – 300 K, only a transition from the zero vibrational state to the first excited state is observed in the vibrational spectrum and the anharmonic contribution to the frequency of the Cl
2 stretching vibration does not change with temperature. At the same time, the frequency of translational modes of the Cl
2 dimer is much lower, about 220 cm
-1 (
Figure 5), and the population of the first excited state of this mode is 0.5 – 0.6 at room temperature. Consequently, its freezing will affect the length of Cl
−−Cl
2 and the population of the σ*(2
px) orbital of Cl
2. For this reason, the ν(Cl
2) stretching mode shows a strong decrease in the range of 5 K – 300 K compared to ν(Br
2) and ν(I
2) (
Figure 2).
Figure 5 shows the spectra of the compound [TeCl
6](Cl
2) at 5 K and 290 K in the frequency range of stretching (around 500 cm
-1) and translational (around 200 cm
-1) vibrations of Cl
2. The splitting of the band of stretching and translational vibrations of Cl
2 into three components is associated with the isotopic composition of chlorine and makes their assignment reliable. The ratio of the natural content of the heavy isotope
37Cl and the light isotope
35Cl can be taken with sufficient accuracy as 0.25:0.75. In this case, the number of pairs with heavy, mixed composition and light isotopes in the crystal should be in a ratio of
. The observed ratio of the integral intensities of the three translation and three stretching modes corresponds with the calculated values with great accuracy. To indicate the frequency of Cl
2 vibration we will use the peak position of the mid mode.
It can be seen that if the wavenumber of the intramolecular mode of Cl2 decreases from 516 cm-1 at 290 K to 513 cm-1 at 5 K, then the wavenumber of the translational mode of Cl2 increases from 214 cm-1 to 217 cm-1 at the same temperature range, which confirms the above assignment of the temperature dependence of vibration frequencies. In other words, these results highlight that strengthening of intermolecular interactions [TeX6]2− − Y2 provokes to weakening of intramolecular bonds.
However, in all [TeI
6](I
2) species, the ν(I
2) increases at cooling down (
Figure 2 c). This means that the contribution of charge transfer from [TeI
6]
2− to the antibonding orbital of the I
2 dimer either changes very little or even decreases with decreasing temperature in contrary to a similar process in [TeCl
6](Cl
2) and [TeBr
6](Br
2). This effect should not be confused with the composition dependence (
Figure 3) obtained at room temperature only. This somewhat unusual result can be understood if we take into account the parameters of the halogen atoms. The empirical atomic radius of iodine (1.4 Å) is significantly larger than that of Br (1.15 Å) and Cl (1.0 Å) atoms. It means probably that bromine and chlorine atoms can track the anharmonic shortening of the Te-X bond lengths in the [TeX
6]
2− octahedron at cooling down. Still, iodine atoms cannot do it because of their large size and an emergence of steric hindrance.
Of the six Te−X bonds, the maximum effect from steric restrictions will be obtained by the two weakest Te−X bonds, the halogen atoms of which have additional bonds with neighboring Y
2 dimers. (
Figure 6, which shows a fragment of the crystal structure).
It is these two bonds of the complex that will not be able to respond to a decrease in temperature, and the charge acquired by the corresponding iodine atoms during interaction with Te(IV) may either not change or even decrease with decreasing temperature, just like the charge transferred from these two I
− to the antibonding orbital I
2. At the same time, the population of the I
2 vibrational mode drops from 0.8 at room temperature to 0.0 at 5 K, suggesting a noticeable anharmonic effect. This explains the difference in the temperature behavior of the ν(Y
2) in [TeI
6](I
2) compared to [TeCl
6](Cl
2) and [TeBr
6](Br
2) (
Figure 2).
The assumption of steric hindrance in the [TeI
6]
2− octahedron is confirmed experimentally.
Figure 7 shows the spectra of [TeI
6](I
2) at various temperatures. In the low-temperature spectrum, two packets of vibrational modes are observed, i.e., in the region of 110 cm
-1 and 160 cm
-1. The first of them relates to Raman-active (symmetric), and the second to IR-active (asymmetric) vibration modes of the [TeI
6]
2− octahedron [
31]. Asymmetric modes arise in the spectrum at T < 200 K. The arising of vibrations forbidden in the Raman spectrum means a distortion of the octahedron, in which various (short and long) Te–I bonds appear in the structure. This phenomenon is not observed in [TeCl
6](Cl
2) and [TeBr
6](Br
2).
Halogen Crystals Cl2, Br2, I2
The crystal structures of the halogens Cl
2, Br
2, and I
2, as well as chalcogens, for example, O
2, are very similar between them.
Figure 8 shows the layout of I
2 molecules in the plane of the I
2 crystal and the value of the electron localization function calculated in [
7].
As mentioned above, the LUMO of the I
2 molecule is σ*(2
px), and HOMO is π*(2
py, 2
pz). I
2 molecules are located in the crystal relative to each other in such a way that the occupied π*(2
py) orbital of each molecule can interact with the unoccupied σ*(2
px) orbital of the neighboring molecule (shown by arrows in
Figure 8). This interaction is more substantial the lower the temperature of the crystal. For this reason, the frequency of the I
2 stretching vibration decreases slightly with decreasing temperature (
Figure 3 in Ref. [
7]) even despite the decrease in the anharmonic contribution.
Intermolecular interaction in halogen crystals is much weaker than intramolecular ones. In other words, the spring that characterizes intermolecular interaction is much weaker than the spring that determines intramolecular bonding. In an experiment with external pressure, at low pressure, the weak spring compresses first, and then, after a certain threshold (3–5 GPA), when the resource for shortening the intermolecular distance ends, the firm spring of the intramolecular bond begins to compress. But the first of the two processes is a decrease in frequency ν(I
2), and the second one is its increase. Such a nonmonotonic dependence of ν(I
2) in I
2 crystals was obtained experimentally and in calculations [
7]. A similar effect was observed also in solid oxygen O
2 (see
Figure 3 in Ref. [
32]). This is unsurprising, since the crystal structure of halogens and chalcogens as well as population of their molecular orbitals are very similar. Thus, the observed effects in both types of crystals, I
2 and O
2, with changes in temperature or pressure, are in good agreement with the model of intermolecular interaction proposed in this work.