Experimental and Theoretical Studies of the Surface Oxidation Process of Rare‐Earth Tritellurides

Recent studies have established Van der Waals (vdW) layered and 2D rare‐earth tritellurides (RTe3) as superconductors and near room‐temperature charge density wave (CDW) materials. Their environmental stability raises natural concern owing to aging/stability effects observed in other tellurium‐based layered crystals. Here, the results establish the stability and environmental aging characteristics of these RTe3 systems involving a variety of metals such as La, Nd, Sm, Gd, Dy, and Ho. The atomic force microscopy (AFM) and scanning electron microscopy (SEM) results show that all the RTe3 sheets oxidize to form thin TeOx layers that are primarily confined to the surface, edges, and grain boundaries. Time‐resolved in situ Raman spectroscopy measurements are used to understand the kinetics of the oxidization process for different lanthanide metal cations and establish their relative stability/resilience to oxidization. Overall results indicate that the vdW layers show higher air stability as the 4f electron number decreases going from Ho to La, resulting in the most stable LaTe3 compared to the least stable HoTe3. Comprehensive quantum mechanical simulations reveal that environmental degradation originates from a strong oxidizing reaction with O2 molecules, while humidity (H2O) plays a negligible role unless Te vacancies are present. Moreover, the simulations explain the effects of 4f electrons on the work function and Te vacancies formation, which directly impact the aging characteristics of RTe3 layers. Interestingly, optical and electrical measurements show that the CDW response is still observed in aged RTe3 layers owing to the presence of underlying pristine/nonoxidized RTe3 layers, except CDW transition temperatures increase due to the thickness effect. Overall results offer the first in‐depth environmental aging studies on these materials, which can be applied to engineer and design their chemical stability, surface properties, and overall CDW characteristics.


Introduction
Van der Waals layered crystals that exhibit charge density waves (CDW) formation belong to a broad class of material systems with unique physical properties. [1][2][3][4][5][6] Superconductivity state, [1,7,8] temperatureinduced Peierls dimerization, [9,10] and other exotic quantum phenomena [5] have been widely studied for these unique class of materials. One of the examples of this class of vdW CDW crystals is the lanthanide tritellurides series with the chemical formula of rare-earth tritellurides (RTe 3 ), where R stands for the rare-earth elements from La to Tm. [11] Previous studies have shown the emergence of temperature-driven CDW phase in RTe 3 based on Raman spectroscopy, [12,13] electrical resistivity measurements, [4,14] angle-resolved photoelectron spectroscopy (ARPES), [15,16] as well as electron diffraction. [17,18] The presence of the CDW state was attributed to the Fermi surface nesting or strong electron-phonon interaction and can be accompanied by the Kohn anomaly. [13,19,20] Overall, RTe 3 materials exhibit CDW phase transition temperature (T CDW ) ranging from below room temperature (220 K for TmTe 3 ) to well-above room temperature (550 K for LaTe 3 ) across the entire lanthanide series. [11,14] Recent studies have established Van der Waals (vdW) layered and 2D rareearth tritellurides (RTe 3 ) as superconductors and near room-temperature charge density wave (CDW) materials. Their environmental stability raises natural concern owing to aging/stability effects observed in other tellurium-based layered crystals. Here, the results establish the stability and environmental aging characteristics of these RTe 3 systems involving a variety of metals such as La, Nd, Sm, Gd, Dy, and Ho. The atomic force microscopy (AFM) and scanning electron microscopy (SEM) results show that all the RTe 3 sheets oxidize to form thin TeO x layers that are primarily confined to the surface, edges, and grain boundaries. Time-resolved in situ Raman spectroscopy measurements are used to understand the kinetics of the oxidization process for different lanthanide metal cations and establish their relative stability/resilience to oxidization. Overall results indicate that the vdW layers show higher air stability as the 4f electron number decreases going from Ho to La, resulting in the most stable LaTe 3 compared to the least stable HoTe 3 . Comprehensive quantum mechanical simulations reveal that environmental degradation originates from a strong oxidizing reaction with O 2 molecules, while humidity (H 2 O) plays a negligible role unless Te vacancies are present. Moreover, the simulations explain the effects of 4f electrons on the work function and Te vacancies formation, which directly impact the aging characteristics of RTe 3 layers. Interestingly, optical and electrical measurements show that the CDW response is still observed in aged RTe 3 layers owing to the presence of underlying pristine/ nonoxidized RTe 3 layers, except CDW transition temperatures increase due to the thickness effect. Overall results offer the first in-depth environmental aging studies on these materials, which can be applied to engineer and design their chemical stability, surface properties, and overall CDW characteristics.
For practical applications of RTe 3 materials, however, their air stability plays a critical role that ultimately dictates the reliability and lifetime of the device. Because of their large surfaceto-volume ratio, 2D materials and even vdW layered crystals raise concerns regarding their stability. [21][22][23] This is particularly true for tellurium-based material systems, including but not limited to tellurene, GaTe, Bi 2 Te 3 , InTe, and others. [22,[24][25][26][27] While the physical properties of RTe 3 layers were extensively studied, their environmental stability remains at its seminal stages. Prior work has shown that high carrier mobility and antiferromagnetic GdTe 3 with T CDW above room temperature (377 K) [28,29] should be more air-sensitive compared to early RTe 3 (toward LaTe 3 ), [30] which was claimed based on the visual observation of the change in crystal color. In general, more studies are needed to establish the environmental stability of RTe 3 materials using in situ techniques and understand the aging mechanism behind these degradation effects.
Here, we present comprehensive environmental degradation studies across the lanthanide tritellurides series from early to late RTe 3 compounds, namely, LaTe 3 , NdTe 3 , SmTe 3 , GdTe 3 , DyTe 3 , and HoTe 3 . Using optical and microscopy techniques, our studies establish the aging dynamics (kinetics) and the resilience to degradation when different RTe 3 materials are subjected to the same environmental conditions. In situ Raman spectroscopy studies show that there is a correlation between environmental aging time scales, metal cation atomic number, and b out-of-plane lattice constant (chemical pressure). At the same time, SEM and atomic force microscopy (AFM) techniques shed light on aging characteristics on the surface. Experimental findings are explained within the density functional theory (DFT) studies to reveal the chemical origin of degradation as well as the correlation between atomic numbers and the kinetics of the aging process.

Material Growth and Characteristics
In our studies, RTe 3 crystals were grown using the chemical vapor transport (CVT) or self-flux (flux) technique. NdTe 3 , SmTe 3 , GdTe 3 , and DyTe 3 crystals were realized using the CVT technique from elemental precursors of metal lanthanide and tellurium powders (99.999%, Alfa Aesar). [31] The typical growth procedure involved sealing these precursors with 20 mg of iodine (I 2 ) as a transporting agent in a 2 mm thick quartz ampoule and subsequently pumping it down to pressure better than 10 −5 Torr. Crystal growth was realized at 830 °C with ΔT = 10 °C thermal gradient for 1-2 weeks to produce millimeter to centimeter-sized crystals. LaTe 3 and HoTe 3 crystals were grown using a flux technique wherein the molar ratio R:Te = 3:97 mixture was kept in an alumina crucible under a vacuum. The quartz ampoule was heated to 700 °C within 8 h to create a mixture solute, and the ampoule was slowly cooled down to 515 °C at a rate of 2 °C min −1 to reject metal cations and to form the desired RTe 3 crystals. LaTe 3 and HoTe 3 were centrifuged to separate vdW crystal from Te liquid to collect crystals.
Typical growth procedure provided millimeter to centimeter-sized crystals, as shown in Figure 1a, which exhibit clear edges and a van der Waals nature. The energy-dispersive X-ray (EDS) results show that lanthanide-based metals and tellurium are distributed uniformly across van der Walls sheets without any phase separation, as shown in Figure 1c. Additionally, X-ray diffraction (XRD -obtained by the Malvern PANalytical Aeris system with CuKα radiation) confirmed the high crystallinity and layered nature of the studied crystals, as shown in Figure 1b. As the R lanthanide metal cation atomic number increases, the XRD reflection shifts to higher values, corresponding to lattice constant reduction (from 26.27 Å for LaTe 3 to 25.36 Å for HoTe 3 ). This behavior is related to the so-called lanthanide contraction effect, where the atomic radius decreases due to poorer shielding of 4f orbital from the increased charge of atomic nuclei. [32] SEM (Zeiss Auriga FIB-SEM) for one of the freshly exfoliated lanthanide tritellurides material is shown in Figure 1d. The high crystallinity of the material can be evidenced by the 135° angles between the edges of the crystal sheet, which is related to its symmetry.

Microscopy Studies to Understand the Aging Effects
Tellurium-based layered vdWs material systems are widely studied as they exhibit different properties, such as singlephoton emission, [33] the nonlinear effect, [34] magnetic order, [35] superconductivity, [36] and CDW state. [6] Previously it was shown that they could suffer from poor environmental stability, [27] and studies performed for GaTe provided information on the resilience of that material to different gases exposure. [22] That issue raises a question about the stability of the RTe 3 series, which hosts the CDW state within the Te atoms sheet enclosed by the RTe slab. That CDW state is formed along the c-axis, and for selected compounds, namely TmTe 3 , ErTe 3 , HoTe 3 , and DyTe 3 , below a specific temperature (in the range from 52 to 180 K), a second CDW along the a-axis is displayed. [14] AFM and Kelvin probe force microscopy (KPFM) measurements were performed on RTe 3 crystal immediately after exfoliation (before oxidization Figure 2a,b) and after subjecting the same sheets (thickness = 20 nm) to air for 24 h in ambient conditions (after oxidization Figure 2d,e). Here, we note that the results in Figures 2,3 primarily focus on R = La owing to their higher air stability, allowing us to demonstrate aging effects across longer periods of time, which will be discussed in greater detail in later sections. AFM studies show that the air-exposed samples produce rough surfaces (Figure 2b) while the work function of the LaTe 3 increases from 4.700 ± 0.025 to 4.735 ± 0.025 eV as per KPFM surface analysis (the approach of work function estimation is described in Section S2.1, Supporting Information). Despite being small, the increase of the work function is uniform across the measured crystal which can be related to the formation of the oxide layer on its surface. To provide further information on the surface topography changes, vdW LaTe 3 was imaged under in situ SEM, as shown in Figure 2c,f. Before aging, vdW crystals show clear edges and terraces, as depicted in Figure 2c. Once the aging takes place, nanoparticle-like features appear on the surface, at the crystal edges, as well as at the terraces edge sites, which are visible in SEM images as well as oxygen atoms in EDS maps (inset Figure 2f). This behavior can be attributed to the higher chemical reactivity at the edges and terraces due to unpassivated/open chemical bonds lowering the chemical barrier for www.advelectronicmat.de oxidization. While crystal surfaces show more resilience to oxidation than the edges, the surface still undergoes substantial environmental degradation after 1 day of exposure. Moreover, the formation of the oxide layer was confirmed unambiguously by XPS measurements performed for pristine and aged DyTe 3 crystals. These results, described in Section S2.2 (Supporting Information), show the emergence of new peaks related to Te oxide complexes.

In Situ Raman Spectroscopy and Kinetics of the Reaction
Time-dependent (in situ) Raman spectroscopy measurements were performed within a ≈2 months period in the backscattering configuration (with the use of CW 632 nm laser line) using 50× Mitutoyo objective with ≈5 µm spatial resolution. While collecting spectra, to avoid laser induce degradation in the presence of water molecules, [22] the crystals were kept under a vacuum (10 −3 Torr) in a Linkam chamber. Here in Figure 3a, freshly exfoliated LaTe 3 sheets exhibit three prominent optical modes located at 88, 97, and 105 cm −1 , whereas the low-frequency Raman peak (amplitude mode at 70 cm −1 ) is related to the CDW phase since LaTe 3 forms that phase at room temperature. [12,31] In situ Raman measurements show that 10 days after preparing fresh thin crystal by exfoliation, new peaks start to emerge at around 128 and 145 cm −1 . At the same time fundamental optical modes and CDW amplitude mode gradually reduce in Raman intensity across ≈2 months. A 2D contour plot constructed from the raw data ( Figure 3a) is shown in Figure 3b to illustrate better the environmental degradation, which shows a clear transition from fundamental Raman modes to new emergent modes related to oxidation of the studied material. These emergent peaks exhibit much larger full-width-at-half-max (FWHM) than fundamental modes, suggesting that these oxidebased regions are disordered/amorphous in nature. Previous studies have shown that these modes at 130 and 145 cm −1 correspond to tellurium oxide complexes TeO 2-x arising primarily from the interaction between H 2 O (g) and tellurium. [22] To understand the kinetics of the oxidization reaction qualitatively, we have analyzed the ratio of integrated intensity between Figure 1. Characterization of a crystal structure for studied in this work rare-earth tritellurides (RTe 3 ) material systems. a) Exemplary optical image of millimeter size LaTe 3 crystal together with schematics of its unit cell. b) (080) X-ray diffraction (XRD) reflection for all studied here materials. c) Energy-dispersive X-ray (EDS) maps showing the spatial distribution of La and Te elements across that sample (1:3 stochiometric ratio is confirmed), and d) scanning electron microscopy (SEM) image for thin LaTe 3 crystal.

www.advelectronicmat.de
LaTe 3 (88 cm −1 ) and TeO x (128 cm −1 ) Raman peaks. Since the intensity of the individual Raman modes is linearly proportional to the material quantity, the calculated time-dependent intensity ratio of LaTe 3 /TeO x denotes the ability of a material to resist oxidization. The Raman intensity ratio of LaTe 3 /TeO x (Figure 3c) shows that this integrated intensity ratio decreases with time, suggesting that the TeO x peaks start to form from the LaTe 3 material. The observed response can be fitted to exponential decay. Such behavior is understood within the Deal-Grove model as the thickness of the oxide layer will form linearly but slow down in the diffusion-limited formation regime. [37] As such, the significant portion of the initial drop in the LaTe 3 /TeO x ratio is ultimately followed by slow saturation as the formed TeO x layer acts as a diffusion barrier, much similar to the oxidization process in silicon or other oxidizing material systems. Here we find that the half-life time equals 9 days, meaning the material-related Raman signal loses its original intensity by half.

In Situ Measurements in Other RTe3 Systems
Using a similar approach, we have extended our studies across the lanthanide tritellurides series, including NdTe 3 , SmTe 3 , GdTe 3 , DyTe 3 , and HoTe 3 (mentioned results are provided in Section S2.3, Supporting Information). These measurements were repeated on a large number of samples (+20) for a better statistical representation of the oxidization characteristics. The results (Figure 4) reveal that DyTe 3 and HoTe 3 have very small resilience to oxidization, meaning they readily oxidize when exposed to ambient conditions, as obtained by the halftime oxidization analysis presented for LaTe 3 in Figure 3c. In contrast, NdTe 3 , SmTe 3 , and GdTe 3 exhibit moderate resilience to oxidization with half-time oxidization characteristics ranging from 2 to 6 days for GdTe 3 and NdTe 3 , respectively. Whereas, LaTe 3 has the highest durability to aging transformation (for that reason, our previous Figures show results obtained for LaTe 3 material, which allows detailed analysis of the kinetics of the reaction). Overall findings summarized in Figure 4 clearly indicate a strong relationship between oxidization and the 4f electron number. Since the out-of-plane lattice constant (b) is smaller (also in-plane (a) and (c) lattice constants reduced) for increasing 4f electron number (see Figure 1b), similar conclusions can be drawn for oxidation tendency and lattice constant (b) (Figure 4 inset). A detailed explanation of the nature of the oxidation phenomenon is provided in the following section.

Theoretical Insights into Oxidation Reactions and Environmental Stability
DFT simulations were carried out to unravel the origin of the observed oxidation resilience trend and provide insight into the oxidation mechanisms of RTe 3 materials. To explain the origin of the resilience trend, we also performed DFT calculations of the work function (Φ) for pristine RTe 3 crystal. In general, a higher Φ should refer to a higher energy barrier that the oxidant needs to overcome in order to capture electrons from the surface of the material to form a chemical bond, which should translate into a higher oxidation resilience. In Figure 4b, we show a plot of Φ values calculated for monolayers of the different RTe 3 materials considered in this study. The plot indeed exhibits a trend of Φ with 4f electrons similar to the oxidation resilience trend. This behavior of Φ can be attributed to the poor shielding effect of 4f electrons, which leads to a greater effective attraction of electrons by the nucleus as the number of 4f electrons increases. [38,39] This results in a decrease in the atomic radius which is manifested by smaller lattice parameters ( Figure 4a, Figure S2, Supporting Information), as well as an increase in the electronegativity of the lanthanide atom with the increase in the number of 4f electrons. In turn, a smaller electronegativity difference between R and Te atoms results in a weaker chemical bonding, making it easier for the oxidant to capture electrons from the surface of RTe 3 . This weaker

www.advelectronicmat.de
bonding is evinced by a decrease in the cohesive energy of RTe 3 materials with the increase in the number of 4f electrons ( Figure S3a, Supporting Information).
Furthermore, to determine the types of gas molecules responsible for the strong surface interaction, DFT simulations were performed to investigate the reactivity of both H 2 O and O 2 on top of clean LaTe 3 and HoTe 3 sheets. The adsorption of both molecules is examined on top of 3 × 1 × 3 supercells at three high symmetry sites (Te, R, and hollow sites) with two possible orientations (parallel and vertical) with respect to the monolayer as illustrated in Figure S1 (Supporting Information). Considering a pristine monolayer of RTe 3 , independent from the potential, three binding sites and how close H 2 O molecules interact with the RTe 3 surfaces, H 2 O molecules did not dissociate instead only remained in their physisorption state (see Section (d), Supporting Information). This implies a substantially high dissociation barrier and excludes H 2 O (humidity) as one of the aging catalysts for pristine RTe 3 . On the contrary, oxygen molecules exhibited strong dissociative chemisorption, as shown in Figure 5. The upper row of Figure 5 depicts four distinct equilibrium configurations that resulted from exothermic O 2 chemical adsorption on pristine LaTe 3 and HoTe 3 . Table 1 outlines the binding energies, bond lengths, and charge transfer for these configurations. From Table 1, we can first observe that for (b), (c), and (d) configurations, the binding energies are always stronger for HoTe 3 , which originates from the shorter average bond lengths between oxygen atoms and their nearest neighbor Ho atoms caused by the poor shielding of 4f electrons. Second, the charge transfer to oxygen atoms is approximately identical for both LaTe 3 and HoTe 3 in all configurations due to the fact that 4f electrons have limited radial extension and thus are not contributing to charge transfer and bond formation. For configuration (a), where oxygen atoms bind to the upper side of the Te sheet, we can generally observe www.advelectronicmat.de smaller binding energies ascribed to the smaller electronegativity difference between O and Te compared to O and R. In addition, the binding energies of oxygen atoms in configuration (a) are equal and independent of their average distance from the nearest neighbor lanthanide atoms, which demonstrates the short-sightedness of the 4f electrons, meaning that the influence of 4f electrons on the binding characteristics can be considered negligible unless oxygen atoms bind directly to R atoms. This behavior agrees with the observed very small difference in physisorption energies (≈0.015 eV lower for HoTe 3 ) (Table S3, Supporting Information), as well as the nearly equal energies (≈0.01 eV difference) of dissociation transition states of O 2 above LaTe 3 and HoTe 3 (see Section (f), Supporting Information). From these observations, we can infer that O 2 on top of HoTe 3 would exhibit (1) slightly more stable physisorption states, leading to a higher number of trial attempts for crossing the dissociation barrier and getting chemisorbed, (2) relatively more stable chemisorption states, and thus a lower rate of oxygen desorption from the surface, leading to faster aging. This generally indicates the higher reactivity of RTe 3 materials to oxygen as the number of 4f electrons increases, as observed from the experimental results and predicted as well by the work function calculations. Sections (e)-(f) (Supporting Information) include detail of our simulations of O 2 adsorption on RTe 3 .
From the pristine configurations in Figure 5, we can notice the higher tendency of oxygen atoms to diffuse under the Te sheets to bind to lanthanide atoms. Moreover, for cases where oxygen binds to the upper side of the Te sheet, such as in configuration (a), we observe identical binding energies (Table 1), and accordingly, one would expect similar aging behavior of different RTe 3 , which is contradictory to the experimental observations. This, in general, promotes the idea that the observed resilience trend originates from a dominant direct binding of the oxidant atoms to lanthanide atoms. Furthermore, the fast oxidation of RTe 3 materials observed in water (shown in Figure S20, Supporting Information), along with the DFT results showing relatively high dissociation barriers of H 2 O above pristine RTe 3 monolayers, suggest that H 2 O molecules could interact with RTe 3 if they have the opportunity to diffuse more easily under the Te sheets and directly interact with lanthanide atoms. This can be related to the influence of Te vacancy defects (V Te ) on reactivity. To investigate this, we performed DFT simulations of defected LaTe 3 and HoTe 3 monolayers where the upper tellurium sheet is completely removed. Although a full removal of the upper Te sheet might appear as a severe assumption, it helps to investigate the behavior of surface spots that have multiple V Te where O 2 or H 2 O molecules get the opportunity to interact directly with the inner R atoms. The lower row of Figure 5 depicts the distinct configurations with exothermic adsorption reactions, while Table 2 outlines their characteristics. From Table 2, we can observe consistent behaviors of roughly equal charge transfers, along with shorter bonds and stronger binding energies for HoTe 3 . Second, the oxygen binding energies are noticed to be three to five times stronger compared to the pristine RTe 3 model.
This can indicate the role played by the upper Te sheet as a protective barrier against environmental aging. The obtained stronger binding energies further agree with the reductions observed in the work function after removing the upper Te sheet (Figure 4b). Interestingly, we can now observe H 2 O dissociative chemisorption (configuration (h) in Figure 5), which indicates that the dissociation barrier for H 2 O becomes finite after the removal of the upper Te sheet, promoting its interaction with RTe 3 materials. These results suggest that V Te might be playing a vital role in determining the overall environmental stability characteristics of these material systems. In this retrospect, we calculated the Te vacancy formation energies of all RTe 3 materials considered in this work. Figure 4b (open black circle) shows that the formation energy of V Te decreases with the number of 4f electrons, which is an expected behavior considering the similar behavior of RTe 3 cohesive energies ( Figure S3, Supporting Information). This point establishes V Te formation as an extra factor contributing to the observed resilience trend, in addition to the role played by the work function. As the number of 4f electrons increases, the RTe 3 material has a higher probability of the V Te formation due to its lower formation energy, which results in much faster oxidation due to the higher binding energies of O 2 and H 2 O to the inner R-Te slab. Overall, although 4f electrons do not directly contribute to chemical bonding, their effects on the work function and the vacancy formation energy can still strongly affect the binding characteristics and the overall environmental stability of the material. More details about the DFT simulations, including binding energies, dissociation barriers, charge transfer, and structure optimization schemes for physisorption and chemisorption situations for both pristine and defected structures, are outlined in the Supporting Information file.

Effect of Oxidization on the CDW Characteristics
After discussing the origin of the aging effect, additional temperature-dependent Raman spectroscopy studies were carried out to understand how the environmental aging effect influences CDW formation. Here, DyTe 3 layers are ideal material testbed systems owing to their relatively fast oxidation characteristics as well as above liquid nitrogen CDW temperature allowing temperature-dependent Raman measurements to capture clean CDW transitions. Freshly exfoliated DyTe 3 sheets were subjected to continuous air exposure for 12 h, and temperature-dependent Raman measurements were performed on pristine (as-exfoliated) as well as aged DyTe 3 sheets, to assess the CDW behavior, as shown in Figure 6a,b, respectively. Previously, the temperature variation of Raman modes was obtained by several authors to track the anti-crossing interaction between coupled phonon and amplitude mode and to determine the T CDW . [12,31,40] As shown in Figure 6c, such anticrossing behavior was observed for both pristine and aged DyTe 3 samples, where the frequency of CDW amplitude (at 68 cm −1 ) and phonon (at 55 cm −1 ) mode reduces as temperature increases starting from 80 K. Subsequently, at around 140 K **nature of both branches changes, and the top one is more phononic, whereas the bottom resembles amplitude mode. It can be seen that amplitude mode softens much quicker compared to phonon mode since the temperature diminishes the CDW order abruptly. At the same time, the temperature does not significantly affect the frequency of phonon mode as mainly related to the thermal expansion of the lattice. Comparison between oxidized and nonoxidized samples shows that this CDW amplitude and phonon cross-over regime remains similar. Moreover, based on the Ginzburg-Landau model, the extracted T CDW values remain close for both samples (308 ± 5 K before and 312 ± 7 K after oxidation). These results suggest that the oxidation process is primarily a surface-limited reaction and oxidized surface ultimately prevent further oxidization from taking place. The formation on the surface of the oxide layer was shown by AFM and SEM/EDS measurements in Figure 2, while it can be seen in Figure 3c, that initial quick drop in the ratio of LaTe 3 /TeO x mode intensity is followed by slow saturation. Such behavior supports the idea of oxidation happening from the surface. Moreover, the surface oxidization leaves an oxidized surface in conjunction with the nonoxidized regions within the exfoliated flakes. While the oxidized amorphous surface offers smaller Raman signals, the underlying RTe 3 layers remain intact and produce CDW Raman behavior, as observed in Figure 6b,c.
This was further confirmed by electrical resistivity measurements shown in Figure 6c inset. Pristine and aged DyTe 3 sheets still exhibit signature CDW behavior (Figure 6c inset), except the CDW transition temperature increases by about ≈20 ± 5 K as the RTe 3 layers get oxidized or the active pristine RTe 3 thickness is reduced. Here, an aging-induced increase in T CDW can Figure 6. Studies of T CDW after oxidation of DyTe 3 material, i.e., rare-earth tritellurides (RTe 3 ) compound for which T CDW transition is around room temperature. a,b) Temperature-dependent Raman spectra before and after oxidation of the studied sample. c) Raman shift in the function of the temperature for charge density wave (CDW) amplitude and phonon mode (label of which is in red in panels (a) and (b)), inset graph shows normalized resistivity spectra obtained before/after oxidation. d) 2D contour plot of Raman spectra obtained at 238 K with time; shows how insensitive to oxidation is CDW amplitude mode (at ≈35 cm −1 ). After 22 min, the laser power was increased to accelerate the oxidation process.
www.advelectronicmat.de be attributed to the reduced thickness of nonoxidized RTe 3 as demonstrated for the GdTe 3 [40] and TiSe 2 [9] in which the T CDW was found to be increasing for thinner CDW materials. Our results show full agreement with these studies and suggest that electrical current mainly probes the nonoxidized electrically conductive CDW portion of the DyTe 3 sheets while bypassing the oxidized TeO x portion of the material or contributing to increased resistivity as observed in our samples (Figure 6c inset red-solid line). It is also noteworthy to mention that optical Raman spectroscopy measurements also show a slight increase in T CDW values by ≈4K (Figure 6a,b), this effect falls within the error bar in our measurements due to the spectral resolution of the CDW amplitude modes (0.5 cm −1 ) and inherently large FWHM values and low intensity count of the CDW amplitude modes presenting increased difficulties in assigning T CDW values.

Conclusion
Comprehensive in situ Raman studies were carried out on van der Waals charge density wave RTe 3 lanthanide tritelluride materials to establish their oxidization characteristics, stability, and elucidate the origin of oxidization effects. Results show that the material stability increases as the 4f electrons are depleted or going from R = Ho → La. Computational studies show that O 2 molecules are the primary catalysts for the oxidization while humidity (H 2 O) plays a negligible role, unless Te vacancies are present, in the environmental surface transformation process. Moreover, we propose an explanation based on the electronegativity of the R lanthanide cation element to provide a simple scheme of the observed experimental dependency of resilience to oxidation versus 4f electron number. Due to the higher electronegativity of the late rareearth lanthanides (i.e., Ho, Dy) compared to early ones (e.g., La, Nd), the electron cloud is pulled from the Te anion toward the R cation. That shift of the electron cloud away from the Te atom leads to weaker attraction with tellurium and makes the transfer of electrons from tellurium to oxygen easier. Detailed microscopy and spectroscopy measurements show that all the RTe 3 systems eventually oxidize to form thin TeO x layers at the edges, terraces, and on the surface. However, a significant portion of the pristine material still resists oxidization due to the diffusion barrier created by the thin amorphous TeO x surface layers. Aged materials still exhibit characteristics of CDW response except for their CDW transition temperatures which increase potentially due to reduced CDW material thickness after aging induced oxidization effect. The results on the oxidization kinetics and aging-induced changes in RTe 3 CDW behavior offer the first environmental aging insights into these material systems, which can be applied to engineer and design their chemical stability, surface properties, and overall CDW characteristics.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.