Thickness-Dependent Sign Change of the Magnetoresistance in VTe 2 Thin Films

: Transition metal dichalcogenides of type VX 2 (X = S, Se, Te) have recently attracted great interest as it has been predicted that they host ferromagnetism at room temperature. Whether ferromagnetism is indeed present is an open experimental question. An in-depth study of the structural and magnetoelectric properties of VTe 2 thin ﬁlms is presented in this work. The VTe 2 thin ﬁlms were grown through molecular beam epitaxy, which allows for precise control of thicknesses, ranging from several nanometers down to monolayers. The low-temperature magnetoelectric transport studies reveal no sign of intrinsic ferromagnetism. However, a transition from positive to negative magnetoresistance is present upon decreasing


Introduction
We are living in the information age where the development of information storage technologies is progressively becoming faster and more indispensable. The continuous reduction of the size of electronic devices with an increasing density of circuits is associated with a rise in energy dissipation due to leakage currents and parasitic capacitance. These negative effects considerably limit the progress in the development of electronic devices [1].
The field of spintronics could present a solution to these limiting effects since it is based on the spin of the electron instead of its charge. It has the promise of faster switching speeds, less total energy consumption, and a higher density of circuit elements. The absence of a net charge current would potentially avoid problems arising from capacitances or Joule heating [2,3]. Recent studies propose that the next generation of spintronic devices can be based on room-temperature ferromagnetic semiconductors [2,4]. However, due to the conflicting requirements of the electronic structure, semiconductor materials with ferromagnetic properties and a high Curie temperature are not common.
Transition metal dichalcogenides (TMDs) comprise unexplored 2D materials that are currently receiving much attention, due to their novel properties and promising applications including photo-electricity [5], ferromagnetism [6][7][8], quantum Hall effect [9], and superconductivity [10]. However, most TMD family members are intrinsically non-magnetic and a magnetic order can only be introduced by applying strain, doping, hydrogenation, or creating vacancies. It has recently been predicted, based on DFT calculations and/or the Monte Carlo method, that monolayers of type VX 2 (X = S, Se, or Te) are intrinsically ferromagnetic at room temperature [11][12][13][14].
Indications for ferromagnetism at room temperature have been observed in vanadium ditelluride (VTe 2 ) nanoplates on SiO 2 /Si substrates grown by chemical vapor deposition [15]. However, Wong et al. [16] attribute the ferromagnetic signatures to either a substrate contribution or possibly from extrinsic perturbations due to the used technique of vibrating sample magnetometry. In addition, some works report the absence of ferromagnetic order in monolayers of VTe 2 through angle-resolved photoemission spectroscopy (ARPES) [17] and X-ray magnetic circular dichroism measurements [18]. A closely related Solids 2022, 3 501 compound, VSe 2 , evoked a similar debate, i.e., room temperature ferromagnetism has been recently reported in monolayers [6,19], while ARPES measurements have revealed a nonmagnetic electronic structure [20]. The contradictory results in VX 2 compounds, but also the origin of its ferromagnetism in general, are still under discussion. Meanwhile, the Kondo effect [21] and ultrahigh conductivity [22] were observed in single-crystalline 1T-VTe 2 nanosheets. Moreover, recent predictions announce strong intrinsic ferromagnetism and semiconducting properties in a new phase, called monolayer puckered pentagonal VTe 2 [23].
With this work, we intend to clarify the ferromagnetic behavior of VTe 2 through direct studies of its magnetic properties by magnetoelectric transport measurements. Our results did not show any indication of ferromagnetism. However, it was found that, as the thickness is reduced, VTe 2 exhibits different magnetoresistance behavior when compared to bulk samples. A crossover was observed from positive to negative magnetoresistance upon reducing the film thickness.

Materials and Methods
The VTe 2 thin films were grown in an ultrahigh vacuum Octoplus 300 MBE System from Dr. Eberl MBE Komponenten in Te-rich conditions at 0.07 nm/min. V (99.99 %) and Te (99.9999%) were evaporated from a custom high-temperature effusion cell and standard Knudsen cell, respectively. Al 2 O 3 (001) is an appropriate substrate for van der Waals epitaxy of TMD materials due to its atomically smooth surface without dangling bonds with a hexagonal lattice symmetry. Moreover, Al 2 O 3 is an insulating substrate suitable for electrical measurements. Prior to the deposition, the substrate was thermally treated in air for 90 min at 1050 • C to obtain atomically flat terraces. Subsequently, the substrate was degassed in situ for 60 min at 500 • C in the presence of Te vapor. The result is an appropriate substrate for van der Waals epitaxy. Finally, the substrate temperature was decreased to 225 • C, at which the growth process was performed during the rotation of the sample. The base pressure during the growth was 1 × 10 −10 mbar. The sample thickness ranges between a monolayer to a few layers. All the samples were in situ capped with 15 nm of Te (deposited at room temperature) to prevent possible oxidation.
X-ray diffraction (XRD) and X-ray reflectometry (XRR) measurements were collected using a Bruker D8 Dimension system in the Bragg-Brentano configuration. Texture analysis was performed by making pole diagrams using a Panalytical system positioning the detector at a fixed 2θ Bragg's angle corresponding to a specific crystallographic plane. The diffracted beam intensity was mapped as the sample was rotated 360 • around the azimuthal axis (ϕ) and tilted from 20 • to 80 • around the tilt axis (ψ). X-ray photoelectron spectroscopy (XPS) spectra were measured using an Omicron nanotechnology surface analysis system, equipped with a monochromatic aluminum source. The surface morphology was examined through atomic force microscopy (AFM) by means of a Bruker Dimension Icon system. AFM measurements were carried out in tapping mode to analyze the 3D surface topography. Low-temperature (4.5 K) transport measurements were carried out using an Oxford 4He cryostat equipped with a 7T superconducting magnet, connected to a lock-in amplifier measurement set up to send/measure the AC current/voltage. Hall bar devices (410 × 820 mm) were fabricated by optical lithography. For this, OiR 907-17 photoresist was spin-coated over the surface of the sample at 4000 rpm for 60 s and baked for 1 min at 100 • C. After UV exposure, the sample was developed in AZ-351 for 1 min. Argon ion milling was used to structure the film. A second step of lithography was made to define the Au contacts using RF sputter deposition.

Results and Discussion
Bulk VTe 2 presents a monoclinic distorted 1T' phase below 207 • C [24,25] but recently it has been shown that an undistorted hexagonal 1T structure appears upon a reduction of the thickness [18]. Lattice distortions related to thickness, temperature, and electrical field have led to quantum phases such as charge density waves (CDWs) phases [26,27].
The 1T structure is composed of one layer of hexagonally arranged vanadium atoms sandwiched between two layers of tellurium atoms in an octahedral configuration, stacked together by weak van der Waals forces (Figure 1a). In a trigonal layered structure, 1T-VTe 2 crystallizes in the space group P-3m1 with hexagonal lattice parameters of a = b = 3.64 Å and c = 6.51 Å [24,28,29]. has a smooth surface with a roughness of 170 pm. The terrace termination is similar to the Al2O3 substrate surface, which indicates an epitaxial growth. Figure 1c shows a typical XRD pattern corresponding to 15 nm 1T-VTe2 deposited on Al2O3 (001). The obtained material was identified as hexagonal VTe2 according to Powder Diffraction File 98-060-3582 and other reports [22,26]. The presence of only the (00l) family peaks indicate that the obtained films have the c-axis of the crystallites oriented perpendicular to the substrate surface. The thickness obtained by XRR was 15.7 nm (left inset in Figure 1c). Using Bragg's law, the estimated c-axis lattice constant was calculated to be 6.58 Å, which is in good agreement with the expected value for 1T-VTe2. The in-plane lattice constant extracted from the streak spacing obtained by the reflection high-energy diffraction (RHEED) patterns was 3.7 Å, also consistent with 1T-VTe2 (right inset in Figure 1c). The obtained lattice parameter discards the presence of the 1T' structure or phases related to the CDWs transitions.  Intensity 2θ (°) The surface morphology was studied by AFM in Figure 1b. The 5 nm 1T-VTe 2 sample has a smooth surface with a roughness of 170 pm. The terrace termination is similar to the Al 2 O 3 substrate surface, which indicates an epitaxial growth. Figure 1c shows a typical XRD pattern corresponding to 15 nm 1T-VTe 2 deposited on Al 2 O 3 (001). The obtained material was identified as hexagonal VTe 2 according to Powder Diffraction File 98-060-3582 and other reports [22,26]. The presence of only the (00l) family peaks indicate that the obtained films have the c-axis of the crystallites oriented perpendicular to the substrate surface. The thickness obtained by XRR was 15.7 nm (left inset in Figure 1c). Using Bragg's law, the estimated c-axis lattice constant was calculated to be 6.58 Å, which is in good agreement with the expected value for 1T-VTe 2 . The in-plane lattice constant extracted from the streak spacing obtained by the reflection high-energy diffraction (RHEED) patterns was 3.7 Å, also consistent with 1T-VTe 2 (right inset in Figure 1c). The obtained lattice parameter discards the presence of the 1T' structure or phases related to the CDWs transitions.
The in-plane orientation and the epitaxial relation with the substrate were studied through texture analysis. Figure 1d shows the pole diagram taken at 2θ = 39.6 • corresponding to the (102) crystalline plane of VTe 2 . The well-defined peaks around ψ = 46 • correspond with the angle between the (001) and (102) crystalline planes of VTe 2 . The presence of the six distinct peaks is in agreement with the six-fold symmetry of the 1T-VTe 2 hexagonal structure and is an indication of the in-plane orientation of the crystals. The pole diagram corresponding to the (024) crystalline plane of Al 2 O 3 was also measured. The obtained peaks are at the same φ angle as the VTe 2 peaks, which means that the VTe 2 films are growing according to a textured epitaxial growth mode, i.e., both the c-and a-axis are oriented with respect to the substrate axes [30].
In situ high-resolution XPS spectra for V-2p and Te-3d of 5 nm 1T-VTe 2 thin films are given in Figure 2. Both spectra are composed of one chemical state, without metallic or oxidized bonding. The binding energies of the distinct V-2p and Te-3d doublet peaks indicate that the VTe 2 film consists of the 1T-VTe 2 phase [24,25]. The stoichiometry of the grown film was calculated by fitting the XPS peaks, for which we performed a Shirley background subtraction from the data. As a result, we found a V:Te ratio of 1:2.09, consistent with VTe 2 . Ex situ measurements showed high levels of oxidation, which means that for an ex situ study of thin samples (<5 nm) a capping layer is required (see double peaks in the insets in Figure 2).
The in-plane orientation and the epitaxial relation with the substrate were studied through texture analysis. Figure 1d shows the pole diagram taken at 2θ = 39.6° corresponding to the (102) crystalline plane of VTe2. The well-defined peaks around ψ = 46° correspond with the angle between the (001) and (102) crystalline planes of VTe2. The presence of the six distinct peaks is in agreement with the six-fold symmetry of the 1T-VTe2 hexagonal structure and is an indication of the in-plane orientation of the crystals. The pole diagram corresponding to the (024) crystalline plane of Al2O3 was also measured. The obtained peaks are at the same φ angle as the VTe2 peaks, which means that the VTe2 films are growing according to a textured epitaxial growth mode, i.e., both the c-and aaxis are oriented with respect to the substrate axes [30].
In situ high-resolution XPS spectra for V-2p and Te-3d of 5 nm 1T-VTe2 thin films are given in Figure 2. Both spectra are composed of one chemical state, without metallic or oxidized bonding. The binding energies of the distinct V-2p and Te-3d doublet peaks indicate that the VTe2 film consists of the 1T-VTe2 phase [24,25]. The stoichiometry of the grown film was calculated by fitting the XPS peaks, for which we performed a Shirley background subtraction from the data. As a result, we found a V:Te ratio of 1:2.09, consistent with VTe2. Ex situ measurements showed high levels of oxidation, which means that for an ex situ study of thin samples (<5 nm) a capping layer is required (see double peaks in the insets in Figure 2). Low-temperature magnetoelectric transport measurements were performed on thin VTe2 films (less than 2 nm thick) using a Hall bar configuration (right inset in Figure 3a) in an out-of-plane magnetic field direction. Considering a non-linear growth rate during the initial stage of the growth and a capping with 10 nm of Te to avoid oxidation, an absolute estimate of the thickness was not made. Therefore, we quote the growth time when we compare different samples. In addition, considering previous reports on the use of Te as a capping layer in other material systems [31,32], where a slight change in the nominal value of the resistance while keeping the same magnetoelectric behavior was observed, we could safely neglect the effect of the Te capping layer on our measurements.
The temperature-dependent resistivity of samples with different thicknesses is shown in Figure 3a. For the thicker 1T-VTe2 sample (30 min), the resistivity shows a small decrease with decreasing temperature, suggesting a metallic behavior, typical of bulk VTe2 (left inset in Figure 3a) [21] with a slight increase of the resistance below 105 K. Upon decreasing the growth time (<30 min), we observed semiconducting behavior in the entire temperature range, i.e., an increase in resistivity while decreasing the temperature. The thinnest samples show the largest increase in resistivity reaching resistances in the order of MOhms at 7 K, leaving the film nearly insulating and no longer possible to be measured with our lock-in configuration. These results are in agreement with previous reports that  Low-temperature magnetoelectric transport measurements were performed on thin VTe 2 films (less than 2 nm thick) using a Hall bar configuration (right inset in Figure 3a) in an out-of-plane magnetic field direction. Considering a non-linear growth rate during the initial stage of the growth and a capping with 10 nm of Te to avoid oxidation, an absolute estimate of the thickness was not made. Therefore, we quote the growth time when we compare different samples. In addition, considering previous reports on the use of Te as a capping layer in other material systems [31,32], where a slight change in the nominal value of the resistance while keeping the same magnetoelectric behavior was observed, we could safely neglect the effect of the Te capping layer on our measurements.
The temperature-dependent resistivity of samples with different thicknesses is shown in Figure 3a. For the thicker 1T-VTe 2 sample (30 min), the resistivity shows a small decrease with decreasing temperature, suggesting a metallic behavior, typical of bulk VTe 2 (left inset in Figure 3a) [21] with a slight increase of the resistance below 105 K. Upon decreasing the growth time (<30 min), we observed semiconducting behavior in the entire temperature range, i.e., an increase in resistivity while decreasing the temperature. The thinnest samples show the largest increase in resistivity reaching resistances in the order of MOhms at 7 K, leaving the film nearly insulating and no longer possible to be measured with our lock-in configuration. These results are in agreement with previous reports that attribute the strong thickness-dependent metal to insulator transition to two different patterns of CDW phases in VTe 2 films [27,29]. However, no sudden step in resistance, characteristic of CDW phase transitions driven by temperature and/or electrical field (DC current), was observed during the cool-down and warm-up of the samples (Figure 3b,c). attribute the strong thickness-dependent metal to insulator transition to two different patterns of CDW phases in VTe2 films [27,29]. However, no sudden step in resistance, characteristic of CDW phase transitions driven by temperature and/or electrical field (DC current), was observed during the cool-down and warm-up of the samples (Figure 3b−c). No magnetic field hysteresis loop was observed, independently of the film thicknesses and temperature, which indicates that no intrinsic long-range ferromagnetic order occurs (Figure 3d). The absence of hysteresis loops was confirmed for in-plane oriented applied magnetic fields. Therefore, we can conclude that no anomalous Hall effect was observed in any of the VTe2 samples. These results are in line with the results from previous studies performed using ARPES [17] and X-ray magnetic circular dichroism measurements [18]. Extracting the Hall resistance, the sheet carrier concentration (n2D) and the carrier mobility (μ) were calculated. Upon decreasing the thickness, n2D decreased from 7.02 × 10 13 to 1.89 × 10 13 cm −2 while μ ranged from 43.8 to 1.9 cm 2 /Vs. For a temperature range of 4.5 to 30 K (inset in Figure 3d), the electrical conduction is dominated by p-type charge carriers, consistent with the hole-like bands determined by ARPES [33] and the Hall measurements in VTe2 single crystals [34]. Figure 4 reveals the magnetoresistance as a function of perpendicular magnetic field (MR = (ρxx(B)−ρxx(0))/ρxx(0) × 100%) for a thick (a) and thin (b) VTe2 sample at various temperatures and VTe2 samples of different thicknesses at 4.5 K (c) and 30 K (d). For thicker samples (Figure 4a), a positive and quadratic field dependence MR is observed independent of temperature, which is in agreement with the classical theory of ordinary MR. However, for thinner samples at low temperature (4.5 K), the MR is linear with a steep slope at low magnetic fields and a shallow slope above approximately 3 T (Figure 4b). Moreover, contrary to the thicker sample, the MR undergoes a gradual transition from positive to negative increasing the temperature. The influence of the VTe2 film thickness on the electronic properties of the material is clearer in Figure 4c−d. At 4.5 K, a positive MR is obtained for all the samples while at 30 K a transition from positive to negative MR appears by decreasing the film thickness. Negative MR has been reported in VTe2 nanoplates of approximately 87 nm deposited by CVD [21] or single crystals [34]. This behavior is attributed to the Kondo effect, which is related to the presence of localized magnetic moments even when there is no sign of intrinsic long-range ferromagnetism. This effect is characterized by negative MR at temperatures below the so-called Kondo temperature, and in our films, there is a transition from positive to negative MR with increasing temperature. Therefore, we must rule out this effect as the origin of the negative MR observed in our thin films. No magnetic field hysteresis loop was observed, independently of the film thicknesses and temperature, which indicates that no intrinsic long-range ferromagnetic order occurs (Figure 3d). The absence of hysteresis loops was confirmed for in-plane oriented applied magnetic fields. Therefore, we can conclude that no anomalous Hall effect was observed in any of the VTe 2 samples. These results are in line with the results from previous studies performed using ARPES [17] and X-ray magnetic circular dichroism measurements [18]. Extracting the Hall resistance, the sheet carrier concentration (n 2D ) and the carrier mobility (µ) were calculated. Upon decreasing the thickness, n 2D decreased from 7.02 × 10 13 to 1.89 × 10 13 cm −2 while µ ranged from 43.8 to 1.9 cm 2 /Vs. For a temperature range of 4.5 to 30 K (inset in Figure 3d), the electrical conduction is dominated by p-type charge carriers, consistent with the hole-like bands determined by ARPES [33] and the Hall measurements in VTe 2 single crystals [34]. Figure 4 reveals the magnetoresistance as a function of perpendicular magnetic field (MR = (ρ xx (B)−ρ xx (0))/ρ xx (0) × 100%) for a thick (a) and thin (b) VTe 2 sample at various temperatures and VTe 2 samples of different thicknesses at 4.5 K (c) and 30 K (d). For thicker samples (Figure 4a), a positive and quadratic field dependence MR is observed independent of temperature, which is in agreement with the classical theory of ordinary MR. However, for thinner samples at low temperature (4.5 K), the MR is linear with a steep slope at low magnetic fields and a shallow slope above approximately 3 T (Figure 4b). Moreover, contrary to the thicker sample, the MR undergoes a gradual transition from positive to negative increasing the temperature. The influence of the VTe 2 film thickness on the electronic properties of the material is clearer in Figure 4c,d. At 4.5 K, a positive MR is obtained for all the samples while at 30 K a transition from positive to negative MR appears by decreasing the film thickness. Negative MR has been reported in VTe 2 nanoplates of approximately 87 nm deposited by CVD [21] or single crystals [34]. This behavior is attributed to the Kondo effect, which is related to the presence of localized magnetic moments even when there is no sign of intrinsic long-range ferromagnetism. This effect is characterized by negative MR at temperatures below the so-called Kondo temperature, and in our films, there is a transition from positive to negative MR with increasing temperature. Therefore, we must rule out this effect as the origin of the negative MR observed in our thin films.

Conclusions
High-quality 1T phase VTe 2 thin films have been grown by MBE. Our experimental findings show no signs of intrinsic long-range ferromagnetic order and clearly demonstrate that our VTe 2 films exhibit non-trivial temperature-dependent magneto-transport properties. The MR behavior of VTe 2 films can be tuned by varying the thickness and temperature. This provides important insights to design sign-adjustable magneto-electronic devices. Further studies of the resistivity at low temperatures and under different magnetic fields and angle-dependent MR studies are still needed to understand the origin of the MR behavior of VTe 2 .