Vladimir
Poborchii
*a and
Dmitrij
Rappoport
b
aNational Institute of Advanced Industrial Science and Technology, Tsukuba 305-8565, Japan. E-mail: Vladimir.poborchii@gmail.com
bDepartment of Chemistry, 1102 Natural Sciences 2, University of California, Irvine, CA 92697-2025, USA
First published on 13th April 2024
The Te8 ring molecule (cluster) is poorly investigated due to the lack of experimental data. Here, we report an experimental and theoretical study of a regular array of oriented Te8 rings formed in the ∼1.14 nm diameter cavities of zeolite LTA, which are arranged in a cubic lattice with a spacing of ∼1.2 nm. Single crystals of LTA with encapsulated tellurium (LTA-Te) were studied using Raman spectroscopy (RS) and optical absorption spectroscopy (OAS). The experimental LTA-Te spectra were found to be in agreement with those calculated using density functional theory (PBE0 hybrid functional and def2-TZVP basis sets) for the crown-shaped Te8 ring molecule with D4d symmetry. Using polarization–orientation RS, we show that the Te8 rings are oriented by their major axes along the 4-fold axes of cubic LTA. We also show that the site symmetry of Te8 in LTA-Te is lower than D4d. Te8 bond-bending modes are well described in the harmonic approximation, while bond-stretching modes are mixed due to the reduced ring symmetry and, probably, anharmonicity. Importantly, OAS data of LTA-Te display dependence on the Te8 concentration, implying the interaction of the rings from neighbouring LTA cavities with the generation of the valence and conduction electron bands of such a cluster crystal.
Zeolites provide a unique opportunity to form and accommodate uniform guest species in their cavities/channels where the species are oriented due to the crystalline nature of zeolites.7–13 For example, only zeolites allowed obtaining polarized Raman spectra (RS) and optical absorption spectra (OAS) of isolated Te chains and rings.7–11 Moreover, fabrication of such regular high-density Te species arrays can be considered as an important direction in the development of new functional materials, so-called cluster crystals.14–16
Here, we study Te8 rings regularly arranged in the large cavities of zeolite LTA. Importantly, we are using the polarization–orientation RS examination of zeolite single crystals. On one hand, this method can be considered as an alternative to X-ray diffraction (XRD) for the structural study of zeolite-confined clusters. On the other hand, it provides very useful information about the cluster properties and their interaction with zeolites. We also study the OAS spectra of Te8 and demonstrate the effect of the interaction between Te8 rings in neighboring cavities, which is important for the fabrication of real cluster crystals.
The idea of Te8 ring formation in LTA cavities was proposed in ref. 17 where the only observed RS feature at ∼168 cm−1 was assigned to the ring with an argument that its frequency is different from the dominant trigonal tellurium (t-Te) band frequency at ∼120.5 cm−1.18,19 However, this observation was not confirmed later. In ref. 20, the RS of LTA-Te microcrystalline powder was re-examined. A strong band at ∼182 cm−1 and a weaker one at ∼45 cm−1 were observed and attributed rather to Te12 than to Te8 since Te8 requires a relatively strong symmetric bond-bending mode band at around 60 cm−1, which was not clearly detected due to laser-induced band broadening, as was found later. The LTA-Te band at ∼62 cm−1 accompanied by ∼182 cm−1 and ∼45 cm−1 bands was observed in ref. 21. This was a rather strong argument in favor of Te8 ring formation in LTA cavities. The observed bands at ∼45, ∼62 and ∼182 cm−1 look similar to the bending E2 and A1 mode bands and stretching A1 mode bands of S8 (∼153, ∼222, ∼480 cm−1) and Se8 (∼76, 111, 267 cm−1) rings.21 The RS of another zeolite AFI with encapsulated tellurium (AFI-Te) showed similar bands at 41–47 cm−1, 65 cm−1 and 184 cm−1.7 The bands were attributed to the Te8 ring. However, no polarized RS study of LTA-Te single crystals has been performed yet, which could provide crucial evidence for Te8 ring formation. Here, we present this evidence. Importantly, no other species but Te8 was found in LTA-Te. This is beneficial for realizing the idea of cluster crystal formation.
Dehydration of zeolites in a vacuum was performed in Pyrex ampoules for several hours at t ∼ 550 °C. The design of the ampoules made it possible to add crystalline Te powder to the zeolite after dehydration without exposing the dehydrated samples to air. The subsequent adsorption of the Te vapour was carried out at t ∼ 550 °C for a few days until the brownish colour of the sample was saturated.
RS spectra of the LTA-Te single crystals were studied using a Renishaw micro-Raman spectrometer equipped with Semrock edge-filters and an Olympus micro-objective lens allowing an ∼1 μm focused laser probe size. A 633 nm wavelength line of the He–Ne laser and a 785 nm line of the light-emitting diode laser were used for RS excitation. Weak absorption of Te8 at these wavelengths is beneficial for Te8 compared to the non-absorbing LTA matrix for obtaining the cluster Raman signal with nearly no contribution of LTA. An additional measurement with 561 nm laser diode line excitation was performed using a Nanofinder-30 Raman/AFM system (Tokyo Instruments Inc.) specially designed for the maximal RS detection sensitivity around this wavelength, which allowed LTA-Te Raman measurement at relatively high Te8 absorption and very low laser excitation power in order not to strongly heat or destroy the clusters. Moreover, a Nanofinder-30 (Tokyo Instruments Inc.) was equipped with a set of three volume Bragg grating notch filters (OptiGrate Co), allowing the detection of the Raman signal with frequencies down to ∼6 cm−1. UV-visible OAS was performed using a Carl Zeiss micro-optical spectrometer with a light probe size of ∼5 μm. The samples were intentionally broken to minimize their optical density in the ultraviolet (UV) spectral range and placed in glycerol between two cover glasses to minimize surface light scattering.
The Te8 RS, infrared (IR) spectra and OAS were simulated using density functional theory (DFT).23–26 The structure and vibrational frequencies of Te8 were computed using the hybrid PBE0 exchange–correlation functional26 and the resolution-of-identity approximation for the Coulomb part (RI-J).27 The def2-TZVP basis sets28 together with the corresponding auxiliary basis sets29 and small-core effective core potentials (ECPs)30 were employed for Te. RS intensities were obtained in the double harmonic approximation using analytical derivatives of static electronic polarizabilities31 and analytical energy Hessians.32,33 Calculations of static polarizability derivatives used the PBE0 functional and augmented def2-TZVPD basis sets.34 IR spectra were computed in the double harmonic approximation from the analytical dipole and energy Hessians.32,33 The vibrational frequencies were not scaled in this work. OAS was performed with time-dependent DFT (TDDFT)35 using the PBE0 functional with the RI-J approximation36 and def2-TZVPD basis sets. All calculations were performed with the Turbomole program package, version 7.5.23,37
Theoretical point group analysis for the vibration modes of Te8 is similar to that made for the S8 molecule with a similar crown shape.43 The Te8 vibration irreducible representations (symmetries) with calculated frequencies and IR/Raman activities are shown in Table 1. The E1 symmetry corresponds to the IR activity for the light polarized in the ring plane (E//X or E//Y), while the B2 symmetry corresponds to the E//Z IR activity. Calculated Te8 Raman intensities for different polarization configurations are equal to their squared Raman tensor components (XX, ZZ, XY, XZ) for each mode, which are determined by the corresponding atomic displacements and related molecular polarizability derivatives. The corresponding values in Table 1 are given in atomic units of 10−2e2a0/Eh, where e, a0 and Eh are the electron charge, Bohr radius and Hartree energy, respectively.
Symmetry | Frequency (cm−1) | IR intensity (km mol−1) | Raman activity XX | Raman activity ZZ | Raman activity XY | Raman activity XZ |
---|---|---|---|---|---|---|
E2 | 21.7 | 0 | −0.49 | 0 | −2.08 | 0 |
21.7 | 0 | −2.08 | 0 | 0.49 | 0 | |
E2 | 43.8 | 0 | −4.38 | 0 | −3.36 | 0 |
43.8 | 0 | 3.36 | 0 | −4.38 | 0 | |
E1 | 56.9 | 0.524 | 0 | 0 | 0 | 0 |
56.9 | 0.524 | 0 | 0 | 0 | 0 | |
A1 | 65.4 | 0 | −7.04 | 0.56 | 0 | 0 |
B2 | 68.6 | 0.128 | 0 | 0 | 0 | 0 |
E3 | 72.7 | 0 | 0 | 0 | 0 | −1.86 |
72.7 | 0 | 0 | 0 | 0 | −0.86 | |
B1 | 179.9 | 0 | 0 | 0 | 0 | 0 |
E3 | 185.3 | 0 | 0 | 0 | 0 | −3.54 |
185.3 | 0 | 0 | 0 | 0 | −0.17 | |
A1 | 186.9 | 0 | −8.48 | −6.95 | 0 | 0 |
E1 | 188.1 | 1.29 | 0 | 0 | 0 | 0 |
188.1 | 1.29 | 0 | 0 | 0 | 0 | |
E2 | 192.1 | 0 | 1.21 | 0 | −1.05 | 0 |
192.1 | 0 | −1.05 | 0 | −1.21 | 0 |
The calculated RS spectra of Te8 for different polarization configurations, corresponding to the Raman tensor components from Table 1, are shown in Fig. 2(b). Raman bands are represented as Lorentz curves with the amplitudes corresponding to the calculated Raman activities and half-widths of 5 cm−1 for all bands.
The dominant band in the calculated RS spectra of Te8, displaying comparable XX and ZZ activities, corresponds to the A1 symmetric bond-stretching mode at a frequency of 186.9 cm−1. In contrast, the A1 bond-bending mode at 65.4 cm−1 displays high XX activity and nearly negligible ZZ activity. Fig. 2(c) shows atomic displacements and Raman tensors of these two A1 modes, which explain the difference in their ZZ activities. One can consider the A1 bond-stretching as a breathing mode since both in-plane and off-plane Te8 dimensions expand and contract the in-phase, while the A1 bond bending mode is an important instrument for Te8 ring orientation determination.
The E2 bond-bending mode at 43.8 cm−1 is rather active in the XX and XY configurations, while inactive in the ZZ configuration. Similar but weaker activity is displayed by the E2 torsion and stretching modes at 21.7 cm−1 and 192.1 cm−1, respectively. The E3 bond-bending mode at 72.7 cm−1 and the bond-stretching mode at 185.3 cm−1 dominate in the XZ configuration spectrum.
In addition to the Te8 Raman-active modes, there are several inactive modes: E1 bond-bending at 56.9 cm−1, B2 bond-bending at 68.6 cm−1, B1 bond stretching at 179.9 cm−1 and E1 bond stretching at 188.1 cm−1, which can become active when the ring D4d symmetry is reduced. As we mentioned above, the E1 and B2 modes are IR-active.
The calculated partial molecular orbital (MO) diagram of Te8 including the highest occupied MOs and the lowest unoccupied MOs is presented in Fig. 2(d). These MOs consist of combinations of Te 5p orbitals, as determined by Mulliken population analysis. The occupied orbitals correspond to Te lone pairs, while the unoccupied orbitals have antibonding character with respect to Te–Te bonds. The largest MO contributions to the allowed electronic transitions with photon energies <4 eV and the largest oscillator strengths are shown in red. It should be noted that the electron transition energies take into account excitonic effects. Therefore, the transition energies are less than the orbital energy differences of the dominant MO contributions. All these four transitions have E1 symmetry, implying the polarization of light E//X.
There are also E//Z-allowed transitions of B2 symmetry. However, their oscillator strengths appeared to be much lower than those of the strongest E//X transitions (Table 2). The calculated OAS spectra of Te8 for polarizations of light E//X and E//Z are shown in Fig. 2(e). Allowed electronic transitions (E1 for E//X and B2 for E//Z) with photon energies of up to 5.5 eV were computed. Due to the difference in the oscillator strengths, the E//X absorption is stronger than the E//Z one. The E//X absorption is mainly represented by transitions at photon energies of ∼2.9, ∼3.1, ∼4.2, ∼4.6 and ∼5.4 eV, while the E//Z one is determined by transitions at photon energies of ∼3, ∼3.6, ∼4 and ∼5.2 eV. The corresponding absorption bands are approximated by the Lorentz curves with half-widths of ∼0.6 eV.
Symmetry | Photon energy (eV) | Oscillator strength |
---|---|---|
E1 | 2.90 | 0.104 |
B2 | 3.03 | 0.016 |
E1 | 3.09 | 0.053 |
B2 | 3.45 | 0.003 |
E1 | 3.53 | 0.001 |
B2 | 3.61 | 0.003 |
E1 | 3.89 | 0.012 |
B2 | 3.96 | 0.004 |
E1 | 4.11 | 0.006 |
E1 | 4.21 | 0.357 |
E1 | 4.35 | 0.002 |
E1 | 4.57 | 0.525 |
B2 | 5.21 | 0.071 |
E1 | 5.42 | 0.333 |
B2 | 5.77 | 0.008 |
E1 | 5.82 | 0.229 |
B2 | 5.94 | 0.002 |
RS spectra of LTA-Te with a loading density of ∼8 at. per cav. at an excitation wavelength of 785 nm are shown in Fig. 3(a). We performed measurements in four different polarization configurations: aa, cc, ab and cd (see the inset in Fig. 3(a)) similar to the polarization–orientation Raman study of LTA with sulphur LTA-S12 and LTA with selenium LTA-Se.12,44,45 Experimentally, we rotated the LTA-Te crystals in the ab plane with incident and scattered light polarizations (1) parallel and (2) perpendicular to each other. The procedure was described in detail earlier.12 Theoretical RS spectra of LTA-Te for the aa, cc, ab and cd polarization configurations were obtained via summation of the Raman responses of three Te8 rings in their three possible orientations in LTA crystals [Fig. 3(b)].
The experimental LTA-Te RS spectra display rather good correspondence with the calculated ones. Indeed, the bands at 162 cm−1 and 183 cm−1 can be assigned to the Te8 E3 and A1 bond-stretching modes similar to the 442 cm−1 and 480 cm−1 bands of S8 in LTA-S [Fig. 3(c)]. The correspondence between the experiment and theory looks especially good for the A1 mode frequency (experimental: ∼183 cm−1vs. theoretical: 186.9 cm−1) and intensity (the strongest band in the spectrum). At first glance, the polarization dependence of its intensity also looks reasonable: the band is strong in aa and cc polarization configurations, while it is weak in ab and cd configurations as expected for the A1 modes. Actually, as we have shown below, the situation is more complicated for the bond stretching modes due to a possible mode mixing effect but, in a rough approximation, the assignment of the 183 cm−1 band to the A1 bond-stretching mode looks reasonable.
Polarization dependencies of the LTA-Te bands in the Te8 bond-bending mode region appear to be qualitatively similar to those of the bond-bending modes of S812 [Fig. 3(c)] and Se812,44,45 rings in LTA. Indeed, the 46–50 cm−1 band polarization–orientation dependence is similar to that of the E2 band at 74–76 cm−1 or 77–79 cm−1 of Se845 and the E2 band at 153–155 cm−1 of S8 [Fig. 3(c)]. The calculated Te8 E2 bond-bending mode frequency of 43.8 cm−1 is in good agreement with the observed 46–50 cm−1 band frequencies of LTA-Te. The splitting of the degenerate E2 mode indicates structural distortion of Te8 in LTA and therefore a reduced symmetry compared to the initial D4d symmetry of the ring. For example, a slight reduction of the Te8 symmetry from D4d to C4v would split the E2 mode into B1 and B2 modes, active in XX and XY polarization configurations, respectively. This is exactly what we experimentally observed. Interestingly, the splitting of the Te8 E2 mode in LTA is very strong, ∼8.3%, in contrast to the weaker splitting of the corresponding E2 modes of S8, ∼ 1.3%, and Se8, ∼ 2.6%.
The LTA-Te 64 cm−1 band displays polarization dependence similar to the A1 symmetric bond-bending mode band of S8 at ∼221 cm−1 [Fig. 3(c)] and the analogous band of Se8 at ∼112 cm−1.12,44,45 The frequency of 64 cm−1 is very close to the calculated frequency of 65.4 cm−1 of the Te8 A1 bond-bending mode [Fig. 2(b); 3(b)]. The experimental 64 cm−1 band is rather strong in the aa and cc polarization configurations and it is still not weak in the cd configuration, while the band completely disappears in the ab configuration. This is a consequence of the 2-dimensional character of the Raman tensor of this mode with only two strong diagonal tensor components in contrast to the rather 3-dimensional tensor of the A1 bond-stretching mode [Fig. 2(c)]. Correspondingly, the single free-standing Te8 A1 bond-bending mode band is strong in the XX polarization configuration and negligibly weak in the ZZ configuration with the zero XY and XZ activities [Fig. 2(b)]. Theoretical calculation of this mode activity in the RS spectra of LTA-Te [Fig. 3(b)] is also in very good agreement with the experiment, namely high aa activity and no ab activity. This is a clear indication of the orientation of the Te8 4-fold axis along the 4-fold axis of LTA.
The experimental LTA-Te ∼ 83 cm−1 band looks similar to the 248 cm−1 E3 bond-bending mode band of S8 [Fig. 3(c)]. It can be attributed to the E3 bond-bending mode of Te8 with a calculated frequency of 72.7 cm−1. The ∼29 cm−1 band of LTA-Te can be attributed to the E2 torsional mode of Te8 with a calculated frequency of 21.7 cm−1. A broad ∼25 cm−1 band, probably, originates from the superposition of the second component of the E2 torsional mode and the ring librations in the LTA cavity.
As we mentioned above, the bands of LTA-Te in the bond-stretching mode region require more detailed consideration. Fig. 3(d) shows LTA-Te RS in this region taken at temperature T ∼ 77 K and an excitation wavelength of 633 nm with better resolved bands due to their temperature-induced narrowing. Fig. 3(e) shows room temperature LTA-Te RS at the same excitation wavelength in a wider spectral range. Four bands at ∼163 cm−1, ∼173 cm−1, and ∼183 cm−1 and its shoulder at ∼186 cm−1 can be observed in Fig. 3(d). The shoulder is, probably, associated with the Te8 E2 bond-stretching mode (theoretical frequency of ∼192 cm−1). Taking into account our preliminary assignment of the 182–183 cm−1 and 162–163 cm−1 bands to the Te8 A1 and E3 modes, respectively, we have to attribute the 171–173 cm−1 band to the forbidden B1 (theoretical frequency of ∼180 cm−1) mode which can be activated due to the reduced symmetry of Te8 in the LTA cavity, and we found that the bands in the bond-stretching mode region can be mixed.
The band at 182–183 cm−1 shows slightly stronger activity in the cc polarization configuration than that in the aa configuration, which is clearly seen in the room temperature RS of LTA-Te obtained by excitation with 633 nm light, as shown in Fig. 3(e). This can be associated with the mode mixing due to the reduced symmetry of the Te8 rings. Interestingly, the 2nd order Raman band of Te8 can be recognized in the cc configuration spectrum at ∼360 cm−1 [Fig. 3(e)]. An even stronger 2nd order Raman band can be clearly seen at an excitation wavelength of 561 nm [Fig. 3(f), green curve] due to the resonant Raman enhancement. The band is very broad covering the spectral range from ∼300 cm−1 to ∼375 cm−1. Importantly, LTA bands do not contribute to the 2nd order Raman band of Te8 since even the strongest LTA band at ∼490 cm−1 [Fig. 3(f), black curve] does not contribute to LTA-Te RS obtained by excitation with 561 nm light.
Summarizing this section, the polarization dependence of the A1 bond-bending mode Raman band of Te8 indicates the orientation of the ring by its 4-fold axis along the LTA 4-fold axis, while that of the E2 bond-bending mode suggests a noticeable symmetry reduction of the ring from the ideal D4d point group symmetry. Te8 bond-stretching modes are mixed due to the ring symmetry reduction. The Raman bands at 162–163 cm−1 and 171–173 cm−1, probably, originate from the E3 mode and formerly inactive B1 mode while the Raman bands at 182–183 cm−1 and 186 cm−1 originate from the A1 and E2 modes of Te8. The originally inactive E1 bond-stretching mode may also contribute to the observed Raman bands.
First, the observed experimental OAS spectra of LTA-Te [Fig. 4(a and b)] appeared to be in nearly perfect agreement with the theoretical OAS spectra of Te8 [Fig. 2(e)]. The agreement with theoretical OAS for a single Te8 ring is especially good for LTA-Te(3) with approximately two rings per five cavities. We have to note that surface light scattering, which increases with the photon energy, contributes to the experimental spectra. In contrast, theoretical spectra do not include this effect and do not include high-energy transitions (>5.5 eV in Fig. 2(e)). This may produce an impression that the experimental and theoretical spectra are different. However, actually, they are very similar.
Secondly, differences between the OAS of LTA-Te(3) and LTA-Te(8) indicate a noticeable interaction between the Te8 rings in neighbouring LTA cavities, in particular, for LTA-Te(8) with one ring per cavity. In Fig. 4(b), we show the spectral positions of the Raman excitation laser lines, clearly indicating rather strong absorption of LTA-Te(8) at a wavelength of 561 nm corresponding to the resonant Raman effect [Fig. 3(f)]. To avoid laser-induced heating/destruction of Te8, we used a very low 561 nm laser power of ∼0.07 mW. At weaker LTA-Te absorption corresponding to wavelengths of 785 nm and 633 nm [Fig. 4(b)], we were able to use higher laser powers of ∼1 mW and ∼0.3 mW, respectively [Fig. 3(f)].
In order to examine a possible effect of the ring interaction on the OAS spectrum of LTA-Te, we theoretically considered three possible Te8 ring dimers with a separation of ∼1.2 nm between the ring centres [Fig. 4(c)]. The corresponding calculated dimer OAS spectra are shown in Fig. 4(d). The effect of the ring interaction is indicated by the absorption enhancement with an increase in the ring interaction, which is the weakest in the “face-to-face” configuration and the strongest in the “side-to-side” configuration. Due to the inter-ring interaction, the unoccupied MOs of the Te8 rings organize conduction bands of the LTA-Te cluster crystals, while the highest occupied MOs organize its valence band. The conduction and valence bands should be rather narrow. Therefore, a variety of non-linear electric conductance effects predicted for narrow semiconductor superlattice mini-bands46 can be expected in the LTA-Te electric properties. A negative differential electric resistance of LTA-Te observed in ref. 47 confirms this.
The experimentally observed splitting of the E2 bond-bending mode [Fig. 3(a, d and e)] is, definitely, a sign of D4d symmetry reduction. It is important to verify whether the observed E2 mode splitting fits the ring structure distortion shown in Fig. 5(a) or not. Due to the Te8 symmetry reduction from D4d to C4v, the E2 bond-bending mode splits into B1 and B2 modes, displaying XX and XY Raman activities, respectively. We performed calculations of the B1 and B2 mode frequencies and intensities as a function of the distortion Δϕ. The results are shown in Fig. 5(b). The B1 mode frequency increases while the B2 mode frequency decreases with Δϕ. This corresponds well to our experimental observation for the E2 bond-bending mode with the LTA-Te aa-active component frequency higher than that of the ab-active component. Thus, we conclude that the Te8 structural distortion shown in Fig. 5(a) reasonably describes the experimentally observed splitting of the E2 bond-bending mode of Te8. However, even stronger symmetry reduction of Te8 is possible.
As we mentioned in the Introduction, Raman polarization–orientation spectroscopy in application to the zeolite-confined cluster structure study can be considered in some sense as an alternative to XRD. Indeed, using Raman polarization–orientation spectroscopy, we can determine a type of cluster (molecule) and its orientation in a zeolite. Then we can suggest its possible location in the zeolite when we know its size. However, we can only guess what sort of interaction occurs between the Te atoms and the zeolite. For understanding this point, the XRD data would be very useful. However, a structural analysis of LTA-Te by XRD is rather difficult because Te is distributed statistically inside the cavities. The resulting scattering power of Te sites with partial occupancy is comparable to those of Na+ cations and water molecules filling spaces not occupied by Te.
The ∼20 micron size of LTA-Te is insufficient for single-crystal XRD, while powder XRD of zeolites with clusters is, definitely, not effective for the examination of the cluster structure. It shows that the zeolite filling with clusters or molecules causes just changes in the relative intensities of the XRD peaks like, for example, in ref. 48 and its ESI, where a number of different types of zeolites, including LTA, with/without sulphur clusters were examined using powder XRD. In contrast, XRD analysis of ∼65 micron size LTA single crystals containing sulphur clearly showed double S8 ring cluster formation in each large cavity,49 which is in agreement with the published RS results12,50,51 and this work [Fig. 3(c)].
An XRD study of ∼60 micron size LTA-Te single crystals was performed in ref. 52. The XRD analysis suggested the interaction of Te atoms with the zeolite oxygen atoms and cations, while it showed low occupancy and high temperature factors for Te atomic sites. Therefore, the authors could not make any univocal conclusion about the Te cluster structure and they proposed several options including Te8 rings.
Using the RS of LTA-Te, the interaction between the Te8 ring and the zeolite can be probed via observation of the Te ring librations vs. temperature. Taking into account the S8 libration frequency of ∼37 cm−1 [Fig. 3(c)] and assuming approximately the same van der Waals interaction force with LTA for both S8 and Te8, we can roughly estimate the expected Te8 room temperature libration frequency as ∼18.5 cm−1 using the square root of the Te and S mass ratio, which is ∼2.
In the LTA-Te RS spectra obtained by excitation with 785 nm light, Te8 librations, probably, contribute to the broad ∼22–25 cm−1 band [Fig. 3(a)]. Instrumental limitations did not allow RS detection at frequencies <20 cm−1 at this excitation wavelength. In contrast, such measurements were available at 561 nm light excitation. A weak point of this wavelength is that the corresponding Te8 absorption and light-induced heating are not negligible.
Fig. 5(c and d) show the Stokes and anti-Stokes low-frequency RS of LTA-Te in ab and cd polarization configurations at ∼25 °C and ∼50 °C temperatures of the heat-controlling table. We roughly estimated the laser-induced heating from the Stokes/anti-Stokes intensity ratio as ∼20–30 °C, the value that should be added to the heating table temperature. In the RS taken at ∼25 °C table temperature, one can see a ∼25 cm−1 band in combination with a ∼14 cm−1 band for both configurations. The last band, definitely, originates from the Te8 librations. With the increase of the table temperature to ∼50 °C, it shifts to a lower frequency of ∼11 cm−1 and displays significant enhancement. This corresponds to the intensification of librations and the reduction of the ring interaction with LTA. The position of the ∼25 cm−1 feature remains unchanged, confirming its origin from the internal torsional E2 mode of Te8.
Thus, the Te8 libration amplitudes strongly increase with a moderate increase in the LTA temperature. This corresponds to a rather weak van der Waals interaction of Te8 with LTA. At room temperature, the libration amplitudes are, probably, quite strong as well. This explains high temperature factors observed for Te sites,52 which complicates the XRD analysis of LTA-Te. In contrast, this effect does not cause any trouble for polarization–orientation Raman spectroscopy, which appears to be a very fruitful method of zeolite-confined cluster study providing information on the cluster structure with distortions, orientation and interaction with zeolites.
This journal is © The Royal Society of Chemistry 2024 |