Multiscale Investigation of the Structural, Electrical and Photoluminescence Properties of MoS2 Obtained by MoO3 Sulfurization

In this paper, we report a multiscale investigation of the compositional, morphological, structural, electrical, and optical emission properties of 2H-MoS2 obtained by sulfurization at 800 °C of very thin MoO3 films (with thickness ranging from ~2.8 nm to ~4.2 nm) on a SiO2/Si substrate. XPS analyses confirmed that the sulfurization was very effective in the reduction of the oxide to MoS2, with only a small percentage of residual MoO3 present in the final film. High-resolution TEM/STEM analyses revealed the formation of few (i.e., 2–3 layers) of MoS2 nearly aligned with the SiO2 surface in the case of the thinnest (~2.8 nm) MoO3 film, whereas multilayers of MoS2 partially standing up with respect to the substrate were observed for the ~4.2 nm one. Such different configurations indicate the prevalence of different mechanisms (i.e., vapour-solid surface reaction or S diffusion within the film) as a function of the thickness. The uniform thickness distribution of the few-layer and multilayer MoS2 was confirmed by Raman mapping. Furthermore, the correlative plot of the characteristic A1g-E2g Raman modes revealed a compressive strain (ε ≈ −0.78 ± 0.18%) and the coexistence of n- and p-type doped areas in the few-layer MoS2 on SiO2, where the p-type doping is probably due to the presence of residual MoO3. Nanoscale resolution current mapping by C-AFM showed local inhomogeneities in the conductivity of the few-layer MoS2, which are well correlated to the lateral changes in the strain detected by Raman. Finally, characteristic spectroscopic signatures of the defects/disorder in MoS2 films produced by sulfurization were identified by a comparative analysis of Raman and photoluminescence (PL) spectra with CVD grown MoS2 flakes.


Introduction
Transition metal dichalcogenides (TMDs) are a wide family of layered van der Waals (vdW) materials with the general chemical formula MX 2 , M being a transition metal (Ti, Zr, Hf, V, Nb, Ta, Mo, W, Re, Pd, or Pt) and X a chalcogen atom (S, Se, or Te) [1]. Most of them exhibit metallic or semiconducting phases. In particular, semiconducting TMDs have been the object of increasing scientific interest in the last decade, due to their huge and diffusion of Mo atoms on the substrate surface. This latter phenomenon strongly depends on the adhesion energy and surface diffusivity of Mo atoms on the substrate.
The main disadvantage of the continuous MoS 2 films produced by the sulfurization approach is their nanocrystalline structure (with 20-30 nm grain-size) [44], typically resulting in poorer carrier mobility, if compared to the large and isolated monocrystalline MoS 2 flakes obtained by the CVD approach. However, the high uniformity and its good compatibility with the fabrication methods used in the semiconductor industry makes this approach appealing for some applications, e.g., MoS 2 /semiconductor heterojunctions [52] or hydrogen evolution applications [53]. Hence, a detailed characterisation of structural/compositional, vibrational, optical, and electrical properties of MoS 2 films produced by Mo sulfurization remains highly desirable.
In this paper, few or multilayer MoS 2 on a SiO 2 /Si substrate have been produced by sulfurization at 800 • C of very thin MoO 3 films, from~2.8 nm to~4.2 nm (i.e., the critical range for the transition from horizontally to vertically aligned layers). The compositional, morphological, structural, electrical, and optical emission properties of the grown films have been extensively investigated by the combination of several characterisation techniques with macro to nanoscale spatial resolution. This correlative analysis provides deep insight into the potentialities and limitations of this material system for applications.

Materials and Methods
The thin molybdenum-oxide films on SiO 2 (900 nm)/Si substrates were obtained by DC magnetron sputtering from a Mo-target (using a Quorum Q300-TD system), followed by natural oxidation in air. The sulfurization process, schematically illustrated in Figure 1, was carried out in a two-heating zones furnace (TSH12/38/500, Elite Thermal Systems Ltd., Market Harborough, UK), with the first zone (at a temperature of 150 • C) hosting a crucible with 300 mg sulphur (purity 99.9%, product 28260.234, VWR Chemicals, Radnor, PA, USA), and the second zone (at a temperature of 800 • C) hosting the MoO 3 /SiO 2 /Si sample. Starting from a base pressure of 4 × 10 −6 bar, the Ar carrier gas (purity 5.0, Messer, Budapest, Hungary) with a flux of 100 sccm transported the S vapours from the first to the second zone. The duration of the sulfurization process was 60 min. play an important role in MoS2 formation during sulfurization of pre-deposited MoO3. In fact, while a higher temperature may enhance the sulfurization degree, on the other hand, it can also result in increased MoO3 evaporation and diffusion of Mo atoms on the substrate surface. This latter phenomenon strongly depends on the adhesion energy and surface diffusivity of Mo atoms on the substrate. The main disadvantage of the continuous MoS2 films produced by the sulfurization approach is their nanocrystalline structure (with 20-30 nm grain-size) [44], typically resulting in poorer carrier mobility, if compared to the large and isolated monocrystalline MoS2 flakes obtained by the CVD approach. However, the high uniformity and its good compatibility with the fabrication methods used in the semiconductor industry makes this approach appealing for some applications, e.g., MoS2/semiconductor heterojunctions [52] or hydrogen evolution applications [53]. Hence, a detailed characterisation of structural/compositional, vibrational, optical, and electrical properties of MoS2 films produced by Mo sulfurization remains highly desirable.
In this paper, few or multilayer MoS2 on a SiO2/Si substrate have been produced by sulfurization at 800 °C of very thin MoO3 films, from ~2.8 nm to ~4.2 nm (i.e., the critical range for the transition from horizontally to vertically aligned layers). The compositional, morphological, structural, electrical, and optical emission properties of the grown films have been extensively investigated by the combination of several characterisation techniques with macro to nanoscale spatial resolution. This correlative analysis provides deep insight into the potentialities and limitations of this material system for applications.

Materials and Methods
The thin molybdenum-oxide films on SiO2 (900 nm)/Si substrates were obtained by DC magnetron sputtering from a Mo-target (using a Quorum Q300-TD system), followed by natural oxidation in air. The sulfurization process, schematically illustrated in Figure  1, was carried out in a two-heating zones furnace (TSH12/38/500, Elite Thermal Systems Ltd., Market Harborough, UK), with the first zone (at a temperature of 150 °C) hosting a crucible with 300 mg sulphur (purity 99.9%, product 28260.234, VWR Chemicals, Radnor, PA, USA), and the second zone (at a temperature of 800 °C) hosting the MoO3/SiO2/Si sample. Starting from a base pressure of 4 × 10 −6 bar, the Ar carrier gas (purity 5.0, Messer, Budapest, Hungary) with a flux of 100 sccm transported the S vapours from the first to the second zone. The duration of the sulfurization process was 60 min. Morphological analyses on the as-deposited MoO3 films and after the sulfurization process were carried out by Tapping mode Atomic Force Microscopy using a DI3100 system by Bruker (Santa Barbara, CA, USA) with Nanoscope V electronics. The compositional properties of the as-deposited metal films and MoS2 formation after the Morphological analyses on the as-deposited MoO 3 films and after the sulfurization process were carried out by Tapping mode Atomic Force Microscopy using a DI3100 system by Bruker (Santa Barbara, CA, USA) with Nanoscope V electronics. The compositional properties of the as-deposited metal films and MoS 2 formation after the sulfurization process were evaluated by X-ray photoelectron spectroscopy (XPS) using Escalab Xi+ equipment by Thermo Fisher (Waltham, MA, USA), with a monochromatic Al Kα X-ray source (energy = 1486.6 eV). The spectra were collected at a take-off angle of 90 • relative to the sample surface and pass energy of 20 eV. The instrument resolution was 0.45 eV (FWHM of the Ag 3d 5/2 peak). The spectra were aligned using C1s (285 eV) as reference.
High-resolution transmission electron microscopy (HR-TEM), high angle annular darkfield scanning transmission electron microscopy (HAADF-STEM), and energy dispersion spectroscopy (EDS) analyses of the MoS 2 thin films were carried out with an aberrationcorrected Titan Themis 200 microscope by Thermo Fisher (Waltham, MA USA). To this aim, cross-sectioned samples were prepared by a focused ion beam (FIB). Raman spectroscopy and mapping of MoS 2 vibrational peaks were carried out by WiTec Alpha equipment by WiTec (Ulm, Germany), using laser excitation at 532 nm, 1.5 mW power, and 100× objective. Photoluminescence spectra (PL) were collected using a Horiba (Palaiseau, France) system with a laser source of 532 nm. To confirm the uniformity of the MoS 2 thin layer across the substrate, the Raman and PL analyses have been performed at different positions on the sample. Finally, nanoscale resolution current mapping of MoS 2 on SiO 2 was performed by conductive Atomic Force Microscopy (C-AFM) with a DI3100 system by Bruker (Santa Barbara, CA, USA), using Pt-coated Si tips with~5 nm curvature radius. sulfurization process were evaluated by X-ray photoelectron spectroscopy (XPS) using Escalab Xi+ equipment by Thermo Fisher (Waltham, MA, USA), with a monochromatic Al Kα X-ray source (energy = 1486.6 eV). The spectra were collected at a take-off angle of 90° relative to the sample surface and pass energy of 20 eV. The instrument resolution was 0.45 eV (FWHM of the Ag 3d5/2 peak). The spectra were aligned using C1s (285 eV) as reference. High-resolution transmission electron microscopy (HR-TEM), high angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), and energy dispersion spectroscopy (EDS) analyses of the MoS2 thin films were carried out with an aberration-corrected Titan Themis 200 microscope by Thermo Fisher (Waltham, MA USA). To this aim, cross-sectioned samples were prepared by a focused ion beam (FIB). Raman spectroscopy and mapping of MoS2 vibrational peaks were carried out by WiTec Alpha equipment by WiTec (Ulm, Germany), using laser excitation at 532 nm, 1.5 mW power, and 100× objective. Photoluminescence spectra (PL) were collected using a Horiba (Palaiseau, France) system with a laser source of 532 nm. To confirm the uniformity of the MoS2 thin layer across the substrate, the Raman and PL analyses have been performed at different positions on the sample. Finally, nanoscale resolution current mapping of MoS2 on SiO2 was performed by conductive Atomic Force Microscopy (C-AFM) with a DI3100 system by Bruker (Santa Barbara, CA, USA), using Pt-coated Si tips with ~5 nm curvature radius.  XPS compositional analyses performed on the thinnest deposited films revealed that they are predominantly composed of MoO3, with a small (<1%) MoO2 contribution. Recently, Vangelista et al. [44] also reported the complete oxidation (ascribed to air exposure after the deposition) of evaporated Mo films with similar thickness, used for subsequent MoS2 growth by sulfurization. The same authors [44] explained the conversion of MoO3 to MoS2 upon exposure to sulphur according to the following chemical reaction: XPS compositional analyses performed on the thinnest deposited films revealed that they are predominantly composed of MoO 3 , with a small (<1%) MoO 2 contribution. Recently, Vangelista et al. [44] also reported the complete oxidation (ascribed to air exposure after the deposition) of evaporated Mo films with similar thickness, used for subsequent MoS 2 growth by sulfurization. The same authors [44] explained the conversion of MoO 3 to MoS 2 upon exposure to sulphur according to the following chemical reaction:

Results and Discussion
which is the result of two intermediate steps: i.e., the S-induced reduction of the MoO 3 to a sub-stoichiometric oxide MoO 3−x (2), followed by its conversion to MoS 2 (3), with the formation of gaseous SO 2 as a by-product. After the sulfurization process at 800 • C, XPS analyses were performed to evaluate the successful conversion of MoO 3 to MoS 2 . Figure 3a reports an overview spectrum, allowing the quantification of the percentage of elemental concentrations on the sample surface. In particular, molybdenum and sulphur percentages of 3.26% and 6.82%, respectively, were evaluated (besides the large Si and O background), which were close to the stoichiometric [Mo]/[S] ratio for MoS 2. More detailed information on the Mo and S bonding was deduced from the Mo3d 3/2 , Mo3d 5/2 , and S2s core levels in Figure 3b, and the S2p 1/2 and S2p 3/2 core levels in Figure 3c. Two doublets were found in the Mo 3d spectrum, and both doublets were fitted with a peak separation of 3.1 eV [44,54,55]. In particular, the deconvolution of the Mo3d peaks shows the predominance of the Mo 4+ component, associated with 2H-MoS 2, accompanied by a smaller Mo 6+ contribution, associated with the presence of residual MoO 3 . The two S2p 1/2 and S2p 3/2 peaks [44,54,55] in Figure 3c confirm that sulphur is mainly in the form of sulphide, with a small S-O component.
which is the result of two intermediate steps: i.e., the S-induced reduction of the MoO3 to a sub-stoichiometric oxide MoO3−x (2), followed by its conversion to MoS2 (3), with the formation of gaseous SO2 as a by-product.
After the sulfurization process at 800 °C, XPS analyses were performed to evaluate the successful conversion of MoO3 to MoS2. Figure 3a reports an overview spectrum, allowing the quantification of the percentage of elemental concentrations on the sample surface. In particular, molybdenum and sulphur percentages of 3.26% and 6.82%, respectively, were evaluated (besides the large Si and O background), which were close to the stoichiometric [Mo]/[S] ratio for MoS2. More detailed information on the Mo and S bonding was deduced from the Mo3d3/2, Mo3d5/2, and S2s core levels in Figure 3b, and the S2p1/2 and S2p3/2 core levels in Figure 3c. Two doublets were found in the Mo 3d spectrum, and both doublets were fitted with a peak separation of 3.1 eV [44,54,55]. In particular, the deconvolution of the Mo3d peaks shows the predominance of the Mo 4+ component, associated with 2H-MoS2, accompanied by a smaller Mo 6+ contribution, associated with the presence of residual MoO3. The two S2p1/2 and S2p3/2 peaks [44,54,55] in Figure 3c confirm that sulphur is mainly in the form of sulphide, with a small S-O component.  The structural properties of the MoS2 films were also investigated at nanoscale by transmission electron microscopy on cross-sectioned samples. Figure 4a,b show representative HR-TEM and HAADF-STEM analyses on the few-layers MoS2 sample obtained by sulfurization of the ~2.8 nm MoO3 film. The diffraction contrast in the HR-TEM image Figure 4a demonstrates the presence of two or three crystalline layers embedded between the amorphous SiO2 substrate and amorphous carbon (a-c) protective film. These layers are predominantly oriented parallel to the substrate, with nanometric scale corrugations. Furthermore, an interlayer spacing of ~0.6 nm is directly evaluated from the HRTEM image of a 3L-MoS2 reported in the insert of Figure 4a. The number of MoS2 layers and their nearly parallel orientation with respect to the substrate is confirmed by the HAADF-STEM image in Figure 4b collected on the same sample. On the other hand, a more irregular configuration of the layers can be observed from the HRTEM (Figure 4c) and HAADF-STEM (Figure 4d) analyses performed on the MoS2 multilayer produced by sulfurization of ~4.2 nm film. In fact, in the analysed specimen volume, horizontally oriented MoS2 layers co-exist with layers standing up with respect to the SiO2 surface. This observation is fully consistent with previous reports showing a transition from horizontal to vertically oriented growth for film thickness larger than 3 nm [47].  The layers number uniformity of the grown MoS 2 films was also investigated on micrometer scale areas and with high statistics by Raman spectroscopy. Figure 5 shows two typical Raman spectra of the few-layers (i.e., 2 L-3 L) MoS 2 (black line) and of the multilayer MoS 2 (red line) grown on SiO 2 by the sulfurization process. The two characteristic in-plane (E 2g ) and out-of-plane (A 1g ) vibrational modes of MoS 2 are clearly identified, and the typical redshift of the E 2g peak and blue shift of the A 1g with increasing the number of layers [19] is observed. In particular, the difference ∆ω = ω A1g − ω E2g between the wavenumbers of these two main modes is commonly taken as a way to evaluate the number of MoS 2 layers, with larger ∆ω values generally associated with a thicker MoS 2 .
The layers number uniformity of the grown MoS2 films was also investigated on micro-meter scale areas and with high statistics by Raman spectroscopy. Figure 5 shows two typical Raman spectra of the few-layers (i.e., 2 L-3 L) MoS2 (black line) and of the multilayer MoS2 (red line) grown on SiO2 by the sulfurization process. The two characteristic in-plane (E2g) and out-of-plane (A1g) vibrational modes of MoS2 are clearly identified, and the typical redshift of the E2g peak and blue shift of the A1g with increasing the number of layers [19] is observed. In particular, the difference Δω = ω − ω between the wavenumbers of these two main modes is commonly taken as a way to evaluate the number of MoS2 layers, with larger Δω values generally associated with a thicker MoS2.  The colour maps in Figure 5b,c illustrate the spatial distribution of the ∆ω values obtained from arrays of 50 × 50 Raman spectra collected on 10 µm × 10 µm scan areas. Figure 5d,e show the histograms of the ∆ω values reported in the two maps, with the indication of the corresponding number of MoS 2 layers according to the calibration reported in Ref. [19]. The two distributions are quite uniform and exhibit a ω ≈ 21.8 ± 0.6 cm −1 for the few-layer MoS 2 sample and ω ≈ 24.8 ± 0.4 cm −1 for the multilayer MoS 2 sample. These ∆ω values are associated with a 2 L-3 L MoS 2 thickness for the first sample, in very good agreement with TEM analyses in Figure 4, and to >4 L MoS 2 for the second one.
In the following, we will concentrate our attention on the 2 L-3 L MoS 2 sample, since the horizontal configuration of the layers makes it more suitable for electronic applications, similarly to 2H-MoS 2 samples produced by CVD or by exfoliation from bulk molybdenite.
The doping type and the biaxial strain (ε) of the thin MoS 2 film were also evaluated from the Raman maps by a correlative plot of A 1g versus E 2g peaks positions, as recently discussed in Ref. [23]. Figure 6a shows as blue circles the ω A1g and ω E2g values extracted from all the Raman spectra in the array of Figure 5. The red line in Figure 6a represents the ideal ω A1g vs. ω E2g dependence (i.e., the strain line) for a purely strained 3L-MoS 2 film. This relation is obtained from the combination of the following two expressions: Here, γ E 2g = 0.39 and γ A 1g = 0.09 are the Grüneisen parameters for the two vibrational modes of 3L-MoS 2 , estimated from the literature values of the peaks shift rates as a function of strain percentage (−3 cm −1 /% and −0.7 cm −1 /% for the E 2g and A 1g peaks, respectively) [56]. ω E 0 2g and ω A 0 1g represent the E 2g and A 1g frequencies for an ideally unstrained and undoped 3L-MoS 2 . Here, the literature values for a suspended 3L-MoS 2 membrane (ω 0 E 2g = 382.9 cm −1 and ω 0 A 1g = 406.4 cm −1 ) [56], not affected by the interaction with the substrate, were taken as the best approximation for these ideal values. This reference point is reported as a red square in Figure 6a, while the two arrows with opposite directions along the strain line indicate the tensile (red-shift) and compressive strain (blue-shift), respectively. Furthermore, the black dashed lines serve as guides to estimate the strain values. The distribution of the experimental points (blue circles) in the plot of Figure 6a clearly indicates that the thin MoS 2 film on SiO 2 is compressively strained. Figure 6b shows the 2D map of the compressive strain, calculated from the map of ω E2g values by applying Equation (4). Furthermore, the corresponding histogram of the ε values is reported in Figure 6c, from which an average strain value ε ≈ −0.78% ± 0.18% can be deduced.
The strain line separates the n-type and p-type doping regions in the ω A1g − ω E2g diagram in Figure 6a. Noteworthy, the experimental points in Figure 6a are partially located in the n-type region and partially in the p-type one. Unintentional n-type doping is typically reported for MoS 2 films produced by different synthesis methods (such as mechanical exfoliation or CVD) and it is commonly ascribed to native defects present in the material [57][58][59][60]. Here, the observed p-type doping in some regions of the MoS 2 film produced by sulfurization can be associated with the presence of residual MoO 3 , as deduced by XPS. In fact, several studies demonstrated how intentionally introducing MoO 3 in pristine (n-type) MoS 2 , e.g., by O 2 plasma treatments, results in p-type doping of the material [61,62]. The MoS2 thin layers produced by MoO3 thin films sulfurization exhibit large resistivity values in the range of 10-100 Ω•cm [63]. This can be ascribed, in part, to the nanocrystalline structure of the films, i.e., the large density of grain boundaries, which are known to introduce resistive contributions in the current path [64]. On the other hand, the local changes in the compressive strain distribution, as well in the carrier density, deduced by Raman mapping is expected to have an effect on the electrical properties of the fewlayers of MoS2. To get direct information on the homogeneity of conductivity in this film, local current mapping has been carried out by C-AFM, as schematically depicted in Figure  7a. In this configuration, the current locally injected from the AFM metal tip flows in the MoS2 film and is finally collected from the macroscopic front contact. Due to the nanoscale size of the tip contact, the dominant contributions to the measured resistance are represented by the local tip/MoS2 contact resistance and the spreading resistance in the MoS2 region underneath the tip. Figure 7b shows the contact-mode morphological image on the sample surface, from which an RMS roughness ≈ 0.5 nm slightly higher than the one of the as-deposited MoO3 film (Figure 2a) was deduced. Figure 7c,d report the corresponding C-AFM current map and the histogram of the measured current values. The current map clearly shows submicrometer lateral variations of the conductivity, which are only partially correlated to the morphology, while the histogram shows a Gaussian The MoS 2 thin layers produced by MoO 3 thin films sulfurization exhibit large resistivity values in the range of 10-100 Ω·cm [63]. This can be ascribed, in part, to the nanocrystalline structure of the films, i.e., the large density of grain boundaries, which are known to introduce resistive contributions in the current path [64]. On the other hand, the local changes in the compressive strain distribution, as well in the carrier density, deduced by Raman mapping is expected to have an effect on the electrical properties of the few-layers of MoS 2 . To get direct information on the homogeneity of conductivity in this film, local current mapping has been carried out by C-AFM, as schematically depicted in Figure 7a. In this configuration, the current locally injected from the AFM metal tip flows in the MoS 2 film and is finally collected from the macroscopic front contact. Due to the nanoscale size of the tip contact, the dominant contributions to the measured resistance are represented by the local tip/MoS 2 contact resistance and the spreading resistance in the MoS 2 region underneath the tip. Figure 7b shows the contact-mode morphological image on the sample surface, from which an RMS roughness ≈ 0.5 nm slightly higher than the one of the as-deposited MoO 3 film (Figure 2a) was deduced. Figure 7c,d report the corresponding C-AFM current map and the histogram of the measured current values. The current map clearly shows submicrometer lateral variations of the conductivity, which are only partially correlated to the morphology, while the histogram shows a Gaussian distribution of these values, resembling the shape of the strain distribution in Figure 7d. From this comparison, we can speculate that these mesoscopic-scale inhomogeneities can be partially ascribed to the lateral changes in the strain and carrier density detected by Raman.
Nanomaterials 2022, 12, x FOR PEER REVIEW 10 of 16 distribution of these values, resembling the shape of the strain distribution in Figure 7d. From this comparison, we can speculate that these mesoscopic-scale inhomogeneities can be partially ascribed to the lateral changes in the strain and carrier density detected by Raman.  In the last section of this paper, Raman and photoluminescence spectra acquired on the few-layers MoS 2 samples produced by sulfurization have been compared with reference spectra acquired on CVD-grown MoS 2 samples with a similar thickness. Figure 8 shows a typical Raman spectrum of 3L-MoS 2 on SiO 2 produced by MoO 3 sulfurization, compared with a spectrum of a 3L-MoS 2 sample grown by CVD on SiO 2 [65], reported as reference. Some remarkable differences can be clearly observed between MoS 2 layers prepared using the two different approaches. In fact, besides a lower E 2g /A 1g intensity ratio, the two vibrational peaks exhibit a more pronounced asymmetric shape in the 3L-MoS 2 produced by sulfurization as compared to the CVD-grown one. The deconvolution analysis of the Raman spectra with four Gaussian contributions, associated with the main E 2g and A 1g modes and the disorder activated LO(M) and ZO(M) modes [66], is also presented in Figure 8. These LO(M) and ZO(M) components are very small in the Raman spectra of CVD 3L-MoS 2 , whereas their weight is higher in the 3L-MoS 2 produced by sulfurization. In this latter case, they can be ascribed both to the nanocrystalline nature of the film, as well as to the presence of residual MoO 3 , as deduced from the XPS analyses. ence spectra acquired on CVD-grown MoS2 samples with a similar thickness. Figure 8 shows a typical Raman spectrum of 3L-MoS2 on SiO2 produced by MoO3 sulfurization, compared with a spectrum of a 3L-MoS2 sample grown by CVD on SiO2 [65], reported as reference. Some remarkable differences can be clearly observed between MoS2 layers prepared using the two different approaches. In fact, besides a lower E2g/A1g intensity ratio, the two vibrational peaks exhibit a more pronounced asymmetric shape in the 3L-MoS2 produced by sulfurization as compared to the CVD-grown one. The deconvolution analysis of the Raman spectra with four Gaussian contributions, associated with the main E2g and A1g modes and the disorder activated LO(M) and ZO(M) modes [66], is also presented in Figure 8. These LO(M) and ZO(M) components are very small in the Raman spectra of CVD 3L-MoS2, whereas their weight is higher in the 3L-MoS2 produced by sulfurization. In this latter case, they can be ascribed both to the nanocrystalline nature of the film, as well as to the presence of residual MoO3, as deduced from the XPS analyses. Figure 8. Raman spectrum for 3L-MoS2 produced by sulfurization (red), compared with a reference spectrum for CVD grown 3L-MoS2 (blue). Data for CVD 3L-MoS2 were adapted with permission from [65], copyright Elsevier 2020. Figure 9 shows the comparison between a PL spectrum measured on the 3L-MoS2 produced by sulfurization with a reference spectrum for CVD grown 3L-MoS2, taken from Ref. [65]. For both spectra, acquired using a 532 nm wavelength laser source, the main emission peak at an energy of 1.86 eV can be observed. However, significant differences in spectral features can be clearly identified from a detailed deconvolution analysis. Raman spectrum for 3L-MoS 2 produced by sulfurization (red), compared with a reference spectrum for CVD grown 3L-MoS 2 (blue). Data for CVD 3L-MoS 2 were adapted with permission from [65], copyright Elsevier 2020. Figure 9 shows the comparison between a PL spectrum measured on the 3L-MoS 2 produced by sulfurization with a reference spectrum for CVD grown 3L-MoS 2 , taken from Ref. [65]. For both spectra, acquired using a 532 nm wavelength laser source, the main emission peak at an energy of 1.86 eV can be observed. However, significant differences in spectral features can be clearly identified from a detailed deconvolution analysis.
The PL spectrum of CVD MoS 2 can be fitted by three Gaussian peaks, associated with the two exciton contributions (A 0 at 1.86 ± 0.01 eV and B at 1.99 ± 0.01 eV, due to the spinorbit splitting of the valence band) and the trionic contribution (X T at 1.78 ± 0.01 eV) [65]. On the other hand, the deconvolution analysis of the spectrum for the sulfurization grown sample allowed us to identify a fourth component X D at 1.75 ± 0.01 eV, besides the trion (X T at 1.78 ± 0.01 eV) and exciton peaks (A 0 at 1.86 ± 0.01 eV and B at 1.95 ± 0.01 eV). Noteworthy, the presence of this X D contribution is accompanied by a strong decrease in the spectral weight of the exciton peak B, as compared to the case of the CVD sample, as well as its FWHM reduction. The occurrence of a similar feature X D, associated with point defects in the MoS 2 lattice, has been recently reported by Chow et al. [67] for the PL spectra of MoS 2 flakes subjected to soft Ar-plasma irradiation, and it was also accompanied by a decrease in the exciton peak B with respect to unirradiated flakes. Hence, the observed X D contribution for our samples produced by sulfurization was ascribed to a higher density of point defects with respect to CVD grown samples.  . Photoluminescence (PL) spectra for 3L-MoS2 produced by sulfurization, compared with a reference spectrum for CVD grown 3L-MoS2. The deconvolution analysis indicated the presence of the excitonic contributions A 0 , B, and of the trionic contribution XT (grey lines) for the CVD grown sample. In addition, the defect-related peak XD (orange line) is identified in the sulfurization grown sample. Data for CVD 3L-MoS2 were adapted with permission from [65], copyright Elsevier 2020.
The PL spectrum of CVD MoS2 can be fitted by three Gaussian peaks, associated with the two exciton contributions (A 0 at 1.86 ± 0.01 eV and B at 1.99 ± 0.01 eV, due to the spinorbit splitting of the valence band) and the trionic contribution (XT at 1.78 ± 0.01 eV) [65]. On the other hand, the deconvolution analysis of the spectrum for the sulfurization grown sample allowed us to identify a fourth component XD at 1.75 ± 0.01 eV, besides the trion (XT at 1.78 ± 0.01 eV) and exciton peaks (A 0 at 1.86 ± 0.01 eV and B at 1.95 ± 0.01 eV). Noteworthy, the presence of this XD contribution is accompanied by a strong decrease in the spectral weight of the exciton peak B, as compared to the case of the CVD sample, as well as its FWHM reduction. The occurrence of a similar feature XD, associated with point defects in the MoS2 lattice, has been recently reported by Chow et al. [67] for the PL spectra of MoS2 flakes subjected to soft Ar-plasma irradiation, and it was also accompanied by a decrease in the exciton peak B with respect to unirradiated flakes. Hence, the observed XD contribution for our samples produced by sulfurization was ascribed to a higher density of point defects with respect to CVD grown samples. . Photoluminescence (PL) spectra for 3L-MoS 2 produced by sulfurization, compared with a reference spectrum for CVD grown 3L-MoS 2 . The deconvolution analysis indicated the presence of the excitonic contributions A 0 , B, and of the trionic contribution X T (grey lines) for the CVD grown sample. In addition, the defect-related peak X D (orange line) is identified in the sulfurization grown sample. Data for CVD 3L-MoS 2 were adapted with permission from [65], copyright Elsevier 2020.

Conclusions
In conclusion, we reported a detailed analysis of the compositional, morphological, structural, electrical, and optical emission properties of few or multilayer MoS 2 on a SiO 2 /Si substrate produced by sulfurization of very thin MoO 3 films at 800 • C. Both Raman mapping and TEM/STEM analyses showed the formation of 2-3 layers of MoS 2 nearly aligned with the SiO 2 surface after sulfurization of the thinnest MoO 3 film, whereas multilayers of MoS 2 (partially standing up) were observed for the thicker MoO 3 film. The strain distribution in the few-layer MoS 2 on SiO 2 was evaluated by the correlative plot of the characteristic A 1g -E 2g Raman modes, showing the occurrence of a compressive strain ε ≈ −0.78 ± 0.18%. Furthermore, the co-existence of submicrometer areas with n-and p-type doping is detected, with the p-type doping probably due to the presence of residual MoO 3 , as revealed by XPS analyses. Nanoscale resolution current mapping by C-AFM showed conductivity inhomogeneities in the few-layer MoS 2 , which are well correlated to the lateral changes in the strain detected by Raman. Finally, the characteristics spectroscopic signatures of the defects/disorder were identified by comparing Raman and PL spectra of sulfurization grown MoS 2 with reference analyses of CVD-grown single crystalline MoS 2 .
The demonstrated MoS 2 growth method is quite versatile and can be extended to different substrates, besides SiO 2 . In particular, the adoption of crystalline substrates (such as sapphire, GaN, and 4H-SiC) with the hexagonal basal plane and good lattice matching with MoS 2 is expected to enhance the domain size and electronic quality of the grown films. Furthermore, the homogeneous large area few-layer MoS 2 can be transferred to arbitrary substrates (including flexible ones) [68] and find applications in different fields of microelectronics, flexible electronics, and sensing.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.