Low-Temperature Chemical Vapor Deposition Growth of MoS2 Nanodots and Their Raman and Photoluminescence Profiles

We report on the growth of an ordered array of MoS2 nanodots (lateral sizes in the range of ∼100–250 nm) by a thermal chemical vapor deposition (CVD) method directly onto SiO2 substrates at a relatively low substrate temperature (510–560°C). The temperature-dependent growth and evolution of MoS2 nanodots and the local environment of sulfur-induced structural defects and impurities were systematically investigated by field emission scanning electron microscopy, micro-Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS) techniques. At the substrate temperature of 560°C, we observed mostly few-layer MoS2, and at 510°C, multilayer MoS2 growth, as confirmed from the Raman line shape analysis. With reduced substrate temperature, the density of MoS2 nanodots decreases, and layer thickness increases. Raman studies show characteristic Raman modes of the crystalline MoS2 layer, along with two new Raman modes centered at ∼346 and ∼361 cm−1, which are associated with MoO2 and MoO3 phases, respectively. Room temperature photoluminescence (PL) studies revealed strong visible PL from MoS2 layers, which is strongly blue-shifted from the bulk MoS2 flakes. The strong visible emission centered at ∼ 658 nm signifies a free excitonic transition in the direct gap of single-layer MoS2. Position-dependent PL profiles show excellent uniformity of the MoS2 layers for samples grown at 540 and 560°C. These results are significant for the low-temperature CVD growth of a few-layer MoS2 dots with direct bandgap photoluminescence on a flexible substrate.


INTRODUCTION
Since the discovery of graphene, two-dimensional materials have attracted much interest due to their interesting optical, electrical, and mechanical characteristics. Molybdenum disulfide (MoS 2 ), a transition metal dichalcogenide (TMD), has a layered structure of hexagonally arranged covalently bonded Mo and S atoms. The vast difference in the electronic structure of the bulk in comparison with the monolayer MoS 2 offers a great opportunity for diverse applications. It has an excellent on/off ratio of 10 8 (Radisavljevic et al., 2011) and a high carrier mobility up to 200 cm 2 / Vs at room temperature (Radisavljevic et al., 2011), making it a promising candidate for optoelectronic applications, such as photodetectors  , (Lee et al., 2012)photovoltaics, (Feng et al., 2012;Bernardi et al., 2013;Fontana et al., 2013), light emitters (Ye et al., 2014), and photo (electro)catalysis. (Biroju et al., 2017a;Biroju et al., 2017b). Its high mechanical strength and flexibility open up new scope for future electronic and optoelectronic applications. Different methods, such as hydrothermal synthesis (Rao and Nag, 2010), different types of exfoliation (Zeng et al., 2011;Zhou et al., 2011;Li et al., 2012a), and physical vapor deposition (Gong et al., 2013;Muratore et al., 2014) techniques, have been used for the growth of MoS 2 . However, the lateral size of MoS 2 domains synthesized by these methods is often restricted to several micrometers, making the synthesis of a large-area MoS 2 thin layer challenging. Among the different methods, chemical vapor deposition (CVD) has been one of the most promising methods of producing large-area and high-quality MoS 2 thin films Jeon et al., 2015;Liu et al., 2015;Yu et al., 2015).
However, a certain amount of defects and impurities is usually present in single-crystalline nanomaterials and even with small imperfections. Particularly, in the case of CVD-grown graphene and related 2D-layered semiconductors, plenty of in-plane point defects and extended defects are often introduced during the CVD growth process. These defects play an important role in tuning the electrical and optical properties of MoS 2 . The change in the nature of bandgap in MoS 2 from indirect to direct type, as the layer number is thinned down to a monolayer (Mak et al., 2010), leads to an extraordinary visible photoluminescence (PL) and a robust visible light response in monolayer MoS 2 . However, quite often, the CVD-grown monolayer and few-layer MoS 2 have lagged behind the expectation for a high-quality direct bandgap semiconductor, mainly because of the intrinsic defects such as sulfur-induced vacancies and Mo-and S-induced complex/ligand impurities and heavy intrinsic n-type doping in monolayer MoS 2 (Mak et al., 2013). Further, the CVD growth of the MoS 2 layer at a relatively low temperature compatible with the microelectronic industry is quite challenging, including the question as to how this study improves the existing knowledge in this area about the MoS 2 dots and quantum dots.
Here, we demonstrate the CVD growth of highly crystalline MoS 2 dots on SiO 2 at a relatively low growth temperature (510-560°C). CVD growth of TMDs on various substrates such as SiO 2 , quartz, and sapphire is usually carried out at high temperatures ranging from 650 to 850°C using atmospheric CVD to grow single crystals or low-pressure CVD to grow films. (Zhang, 2015;Ganatra and Zhang, 2014;Lee et al., 2320). Note that there are very few reports on the direct CVD growth of TMDs on various kinds of substrates. (Mun et al., 2016a;Gong et al., 2017). For flexible and stretchable devices, MoS 2 is one of the suitable materials among the TMDs. Flexible substrates usually have a low melting temperature. So direct growth on these substrates poses a considerable challenge; usually, the MoS 2 grown at high temperature is transferred onto these substrates to fabricate the devices. When a comparison is made between the transfer and the direct growth method, the former method may have some structure deformation which may create some defects that may affect its semiconducting properties. Thus, a lowtemperature direct growth method is always preferable for the fabrication of nanoelectronic devices (Mun et al., 2016b). We have investigated the sulfur (S)-induced surface defects on as-grown MoS 2 dots using position-dependent micro-Raman spectroscopy and X-ray photoelectron spectroscopy (XPS) analyses. Some new Raman bands at 346 and 365 cm −1 were identified, besides the characteristic Raman signatures of MoS 2 (E 2g and A 1g ), which are associated with the presence of MoO 2 and MoO 3 . The Raman features of the as-grown MoS 2 dots were fully consistent with the XPS analysis. The XPS spectral line shape analysis of Mo-3d and S-2p core level peaks reveals a high amount of S vacancies in the asgrown MoS 2 dots.

EXPERIMENTAL
Few layered MoS 2 dots were directly grown on the SiO 2 substrates using an in-house developed CVD system. Commercially procured MoO 3 and S powder were used as precursors for the CVD growth of MoS 2 in a quartz tube-based horizontal muffle furnace. High-purity argon (Ar) gas was used as the carrier gas. At first, the S precursor was kept near the entrance of the quartz tube, away from the central heating zone of the furnace, while the MoO 3 powder was placed at the center of the furnace, and a few pieces of bare SiO 2 substrates were placed near the source powder at the rear end side of the quartz tube. Note that both S and MoO 3 precursors were separately loaded in two mini quartz tubes (inner diameter∼10 mm) and kept at a distance of 53 and 33 cm away from the substrates (see Supplementary Figure S1). This arrangement provides stable evaporation of S and MoO 3 sources at suitable temperatures and avoids any crosscontaminations during the MoS 2 growth. The quartz chamber was pumped down to a base pressure of 1.9 × 10 -2 mbar, and the central zone temperature was raised to 760°C at the rate of 17°C/ min for 45 min and was allowed to stabilize at this temperature for 15 min with Ar as the carrier gas with a flow rate of 100 sccm (standard cubic centimeters per minute), keeping the pressure at 2.2 mbar. Then, it was ramped up slowly to 800°C at the rate of 4°C/ min and then, maintained at this temperature for 30 min without changing the gas flow. After the growth process, the system was allowed to cool down to room temperature with an Ar gas flow of 300 sccm. The experimental setup and the complete timeline profile of the CVD-grown MoS 2 dots are shown in Supplementary Figure S1. The substrate temperature is controlled by the proper positioning of the substrate inside the furnace. The temperature at each position corresponding to the substrate was calibrated using an external thermocouple prior to the growth of MoS 2 nanodots. The MoS 2 growth was monitored at four different substrate temperatures. The samples grown at substrate temperatures of 560, 540, 510, and 310°C are named as MS-560, MS-540, MS-510, and MS-310, respectively.

CHARACTERIZATION
Micro-Raman and PL measurements were performed with a high-resolution spectrometer (Horiba 171 LabRam HR), with excitation wavelengths (λ ex ) of 488 and 514.5 nm (Ar + ion laser). The excitation source was focused with a 100× objective lens, a Frontiers in Nanotechnology | www.frontiersin.org November 2021 | Volume 3 | Article 775732 spot size of 2 μm, and a laser power of 1.5 mW, and the signal was collected by a CCD in a back scattering geometry sent through a multimode fiber grating of 1800 grooves/ mm. X-ray photoelectron spectroscopy (XPS) measurements were performed with a fully automated XPS microprobe (PHI-Xtool, Ulvac-Phi) using an Al Kα X-ray beam (1486.7 186 eV). The standard carbon 1s spectrum was used to calibrate the XPS spectra recorded for various samples with the spectral resolution of 0.2 eV. Surface morphology and microstructure of the MoS 2 dots were examined by using a field emission scanning electron microscope (FESEM), FEI-Quanta 400 with an acceleration voltage of 20 kV with spatial resolution: 1.2 nm at 30 kV in SE, 2 nm at 30 kV in BSE (gold/carbon), detectors: SE, BSE. Spectral line shapes of the micro-Raman features were fitted by Lorentzian line shapes. The XPS line shape profile for all the samples was performed by Shirley peak fitting using Peakfit software with peak parameters as free variables. The best fit was obtained by a large number of iterations.

RESULTS AND DISCUSSIONS FESEM Studies
In order to understand the morphology of the as-grown samples, FESEM imaging was carried out. A detailed description of the sample growth conditions and FESEM features is provided in Supplementary Table S1. Figures  The uniformity of these dots is almost maintained throughout the surface of the sample, estimated as 1 cm × 1 cm. We observed that the MS-560 has a doughnut-shaped structure of MoS 2 dots, which are denser and smaller than that of MS-510. The shape of the MS-510 dots is mostly circular, and a regular array type structure is formed. The observed regular array of MoS 2 dots is likely to be due to the direction of the outgoing carrier gas flow. Frontiers in Nanotechnology | www.frontiersin.org November 2021 | Volume 3 | Article 775732

AFM Studies
show the AFM images of the as-grown samples grown at different substrate temperatures. For the same growth duration, but at different growth temperatures, we notice an evolution of the morphology of the MoS 2 nanodots. In the experiment, we observed that as the substrate is at higher temperature and close to the center of the quartz tube, higher density of the MoS 2 nanodots are formed and the corresponding sizes are smaller. In Figures 2A,B for the MS-560 sample, the MoS 2 nanodots show some kind of a ring-like structure. These rings are evenly distributed throughout the sample. The average size is estimated to be ∼200 nm. We can also see from the AFM height profile ( Figure 2C) that bilayer and a few-layer MoS 2 were grown at this temperature (560°C). In Figures 2D,E, for the MS-540 sample, we notice that the MoS 2 ring structures are less in density than that of MS-560, and there is also presence of larger particle sizes ranging from ∼200 to 500 nm. These MoS 2 layers are found to be thicker, and from the height profile ( Figure 2F), we can see that they are mutilayered. Few larger circular ring structures are also present with diameter ∼1 µm. In Figures  2G,H, for MS-510 sample, the nanodots are seen to be less dense than that of MS-560 and MS-540 cases. There is also presence of larger circular rings with sizes ranging from ∼800 nm to 2 µm.

Micro-Raman Studies
Micro-Raman measurements were performed on the as-grown MoS 2 dot samples, and it was compared with that of the bulk MoS 2 flakes. Figure 3A represents a comparison of the Raman spectra for different samples, which exhibit two important characteristic Raman modes, and this is a signature of the few-layer MoS 2 , associated with the E 2g and A 1g phonon modes at 379 and 405 cm −1 , respectively. In general, the E 2g mode is an in-plane vibration of two S atoms with respect to the Mo atoms, while the A 1g mode is associated with the out-of-plane vibration of only S atoms in opposite directions with respect to the Mo atoms. These two Raman modes are very sensitive to the local in-plane, edge defects, and impurities present on the atomic layers of MoS 2 . The separation between these two modes is a measure of the number of layers in the MoS 2 . It is clear from Figure 3A that a few-layer MoS 2 has been grown for samples MS-560, MS-540, and MS-510 as evidenced by the presence of E 2g and A 1g Raman bands, which is also consistent with the FESEM images ( Figures 1A-D). On the other hand, in the case of sample MS-310, there was no signature of the MoS 2 layer, as evidenced by the absence of the A 1g peak in the Raman spectrum. The Raman spectrum of MS-310 shows a strong Raman peak at 338 cm −1 , which signifies the presence of MoO 3 and negligible presence of S. We also observed a phonon mode stiffening of the characteristic Raman modes, reflected by a blue shift of both the peaks by 3-5 cm −1 for all the as-grown samples. The blue shift is expected to be closely related to the compressive strain induced by the adlayer-substrate lattice mismatch (Hui et al., 2013). The line shape parameters of Raman peaks are shown in Table 1, which depicts the peak position and frequency difference (Δk) between the A 1g and E 2g . The observed Δk values signify the growth of bilayer and trilayer MoS 2 . The Lorentzian line shape was fitted for the E 2g and A 1g bands, as shown in Figure 3B. There are two new Raman bands fitted at the lower and higher frequency side at 380 and 415 cm −1 , respectively. An earlier report suggests that the splitting of the A 1g peak into two peaks is due to the appearance of a Raman inactive mode (B 1u ) at 415 cm −1 , which is not prominent in the case of bulk MoS 2 . However, it becomes significant in the case of monolayer MoS 2 , which makes the A 1g peak (406.9 cm −1 ) asymmetric towards a higher wavenumber. (Li et al., 2012b). The observed splitting in the E 2g Raman mode at 380 cm −1 can be attributed to the mechanical strain in the MoS 2 dots. (Conley et al., 2013). Note that the symbols represent the experimental data, and blue solid lines signify the fitted data in Figures 3B,E. Figures 3C,D show the position-dependent Raman spectra of the as-grown MoS 2 dot array in MS-560 and the corresponding OM image showing the scanned area for each spectrum marked with the respective serial number. We observed that the Raman spectra for different locations of a sample show quite similar features. In spots 1 and 2, we found that the frequency difference Δk between the Raman bands is ∼21.5 cm −1 and 21.9 cm −1 , respectively, which implies the presence of bilayer MoS 2 . However, at spots 3 and 4, the frequency differences Δk are found to be 20.2 and 19.9 cm −1 , respectively, which implies the coverage of few layers to monolayer MoS 2 . We notice a shift of the E 2g mode without any significant shift of the A 1g mode for different spots of the sample MS-560. The summary of the peak parameters is shown in Table 2. There are some additional Raman bands at 346.3, 351.6, and 365.3 cm −1 , which are assigned to the stretching modes of the doubly coordinated oxygen (Mo−O−Mo). The observed Raman bands of MoO 3 and MoO 2 compare well with the reported literature values. (Dieterle and Mestl, 2002). Note that MoO 2 and MoO 3 are present in the sample due to lowtemperature growth.
Position-dependent Raman spectra of as-grown MoS 2 dots in MS-540 show similar uniformity and growth of monolayer and bilayer MoS 2 dots, as shown in Supplementary Figure S2; Supplementary Table S2. There is also the presence of an additional band corresponding to the MoO 2 and MoO 3 phases (see Supplementary Figure S2; Supplementary  Table S2). In the case of MS-510, we observed mostly trilayer and multi-layered MoS 2 growth (Supplementary Figure S2; Supplementary Table S3, ). In this case, the E 2g mode is downshifted substantially. Thus, growth at 540-560°C is suitable for the growth of monolayer and bilayer MoS 2 . However, there are recent research works that have been progressing on low temperature CVD growth in which a new root is being proposed using the atmospheric CVD technique to grow monolayers of MoS 2 at 550°C by using sodium (Na) as the catalyst. (Safeer et al., 2021). Authors have produced single and polycrystalline films by controlling the NaNo 3 /MoO 3 catalyst/precursor ratio and growth time. Based on the first-principle calculations, authors have determined that Na was the nucleation site for the growth of MoS 2 , and the precursor's ratio is crucial to decrease the formation energy and growth temperature. In the present case, the additional Raman bands associated with the E 2g mode; all the samples are attributed to the nucleation sites of MoO 3 / MoO 2 that may be more favorable for the sulfurization process in order to form MoS 2 nanodots at low-temperature regimes i.e., 510, 540, and 560°C.  Figure 4 shows the core level XPS spectra of the Mo-3d and S-2p bands for the as-grown MoS 2 dots grown at different substrate temperatures. In Figures 4A,B, several Mo 3d 5/2 peaks could be fitted for each Mo (3d) envelope, indicating that more than one Mo species was present. In most of the spectra, the main Mo component (Mo 4+ ) was MoS 2 . Besides MoS 2 , all the samples contain additional Mo (3d) peaks corresponding to the oxides of the Mo metal (Mo 6+ ). Note that the peak positions and local environment of S induced surface defects, and impurities were estimated from the Gaussian line shape of the Mo (3d) and S (2p) core level peaks, and the fitting parameters along with the peak  identity are summarized in Table 3. Note that the as-grown MS-560 and MS-510 show quite similar spectra. However, in the case of MS-310 ( Figure 4C), we notice that the peak intensity corresponding to MoO 3 is much higher than that corresponding to MoS 2, and the sulfur (2s) peak is absent. This may be due to the higher content of the physisorbed oxygen. Note that MS-310 is grown at a very low temperature, and due to the low vapor pressure of S, the formation of the MoO 3 phase dominates. Figures 4D-F shows the core level XPS S-2p spectra for the asgrown MoS 2 dots. The band that is observed in S-2p is the 2p 3/2 , which is significant for the crystalline S bonded with Mo. The S-2p spectra for all the as-grown samples show the presence of a number of S components. The content of S in all the samples is found to be very low. For MS-310, the amount of S is found to be much less than that of MS-560 and MS-510. This may be due to the lack of formation of MoS 2 , which is consistent with the Raman spectra discussed earlier. Additionally, we found some new XPS bands in the range 160-180 eV that may be due to the S-related defects. It is expected from the CVD-grown samples that the additional bands from the 2p 3/2 may be due to the S or Mo complexes. Due to low substrate temperature, the formation of polysulphides and ionized S (due to the weak bonding with Mo) is quite expected. The higher binding energy of S-2p peaks may correspond to the polysulphide or thiomolybdate (Spevack and McIntyre, 1993) (which has a core structure of Mo x S y where x ≠ y < 1) species. There may also be a formation of oxysulphide species, where S-O bonding is present, and this would produce the intense XPS bands in the range 167-170 eV. We also observed some additional peaks in the range 170-180 eV, which may be due to the presence of some impurities. Figures 4G-I show the core level O-1s XPS spectra for the asgrown MoS 2 dots. The O-1s spectrum has been fitted by two peaks (components) for all the samples. The peak position and the intensity ratio are summarized in Table 4. The peaks P1 and P2 are associated with the physisorbed oxygen and SiO 2 , respectively. We see that the ratio of the intensities of peaks P1/P2 for MS-560 is 0.11 and MS-510 is 0.17, while for MS-310 it is 6.44. This clearly reveals the presence of a higher content of physisorbed oxygen in the case of MS-310. This is consistent with the Mo (3d) XPS spectra, where we observed the presence of a higher integrated intensity of the Mo species associated with the MoO 3 (see Figure 4, peaks at 235.7 and 237 1eV). This is also consistent with the Raman analysis that showed the presence of the characteristic peak of MoO 3 (Figure 2A).

Photoluminescence Studies
The as-grown MoS 2 dots show strong visible PL at room temperature. Figure 5A shows the PL spectra of the MoS 2 dots grown at different substrate temperatures (MS-560, MS-540, and MS-510) along with that of the commercial MoS 2 flakes (MS). We observed a clear blue shift in the PL peak positions for the MS-560, MS-540, and MS-510 with respect to that of bulk MoS 2 . The blue shift is maximum for the MS-560 sample, which is due to the direct band gap transition in the monolayer MoS 2 , consistent with the previous reports. (Splendiani et al., 2010;Nan et al., 2014). Furthermore, the intensity of PL in MoS 2 dots is about one order of a magnitude higher than that of the bulk MoS 2 flakes. The strongly enhanced PL is direct evidence for the direct bandgap nature of the MoS 2 dots, as compared to the indirect gap of the bulk MoS 2 . In order to understand the origin of the PL band, we have deconvoluted each spectrum of the MoS 2 dot sample by using four Gaussian peaks (Figures 5B-D). The summary of the peak positions in each sample is listed in Table 5. Two characteristic PL peaks, P1 and P2, centered at ∼ 621-631 nm and ∼660-667 nm are ascribed to B exciton and A exciton, respectively. (Splendiani et al., 2010). An exciton peak is derived from the direct bandgap of MoS 2 , and the B exciton peak arises from the direct gap transition between the minima of the conduction band and the lower-level valence band maxima that is created by strong valence-band spinorbit splitting at the K point (Splendiani et al., 2010). We also  observed the presence of the A − exciton (P 3 ) peak, which is understood as arising from the trions. The trion emission is caused by light n-type doping from the substrate or from charged impurities in the CVD-grown MoS 2 dots. (Tongay et al., 2013;Dhakal et al., 2014). We also observed the presence of peak P4 in all the samples at ∼ 696-700 nm, and this peak is attributed to the radiative recombination of bound excitons. (Tongay et al., 2013). This defect peak may be due to the presence of S vacancies in the samples. Note that the relative intensity of P4 is higher in the case of MS-510, than the other two samples. This is because of the lower temperature growth. All four peaks have been observed only in MS-560, MS-540, and MS-510. On the other hand, MS-310 did not show any PL, consistent with the lack of formation of MoS 2 as evidenced from the Raman and XPS spectra discussed earlier. We notice a small change in peak positions with the growth temperatures (see Table 5), which could be due to the non-uniformity in layer numbers for different samples. Figures 6A,B show the position-dependent PL spectra of the MS-560 sample and the corresponding optical micrograph image showing the scanned area for each spectrum marked with a serial number. We observed that there is a no substantial shift in the PL peak position for different locations of the same sample, indicating fairly uniform growth of the MoS 2 dots. A marginal shift in the peak position at location 3 may be due to a small variation in the thickness of the MoS 2 dots. Previous reports suggest that the P 2 peak is blue-shifted, and the PL intensity is increased when decreasing the thickness from multilayer to monolayer MoS 2 films. (Mohapatra et al., 2016). Note that at positions 1 and 2, the integrated PL intensity ratio of the P2/P1 peak is much higher than that at the positions 3 and 4. We also observe a slight broadening of the PL peak in positions 3 and 4, and this may be due to the high trion emission, as the integrated intensity ratio of the P2/P1 peak is lowest in region 4. The P2 peak decreases in intensity due to the enhanced nonradiative decay  when MoS 2 is doped with electrons. Since the growth is carried out at a relatively low temperature, the presence of structural defects and strain cannot be ignored. MS-540 and MS-510 samples show PL similar to that of MS-560 (see Supplementary Figures S3, S4). This visible PL at room temperature is a direct consequence of the direct gap nature of bilayer MoS 2 . Thus, our results demonstrate that the CVD growth of an array of monolayer MoS 2 dots is possible at a temperature as low as 560°C. Note that monolayer CVD growth of MoS 2 is typically achieved at ∼1000°C, though the recent report shows that it can be grown at 680°C. However, the present report shows that monolayer CVD growth is achieved at 560°C.

Growth Mechanism
The growth mechanism of MoS 2 nanodots on the SiO 2 substrate is shown in Figure 7. Here, MoO 3 and sulfur are the precursors used and are placed at a particular distance away from the substrate as mentioned earlier. As the temperature of the single-zone furnace is ramped up to 800°C, first, the MoO 3 vapor is partially transferred by the Ar carrier gas and the MoO 3 molecules get attached onto the substrates. Next, as the sulfur vaporization takes place, the S atoms appear later near the substrates and react with the MoO 3 atoms adhered onto the substrate, resulting in the formation of MoS 2 (2MoO 3 + 7S 2 → MoS 2 + 3SO 2 ). However, as the substrates are placed at the region of lower temperatures (510-560°C), due to the short diffusion length on the surface, the MoS 2 nanodots grown on the SiO 2 substrates are of non-uniform thickness, with a few tens of layers at the center. (Mun et al., 2016c). Additionally, the high flow rate of the Ar carrier gas throughout the experiment has also resulted in the formation of a uniform array of the MoS 2 nanodots.

CONCLUSION
Growth of highly luminescent bilayer and few-layer MoS 2 dots with controlled size and shape by a low-temperature (510-560°C) CVD technique was reported. The intense visible PL at room temperature with free excitonic emission indicated the direct gap nature of the crystalline MoS 2 dots array. Sulfur-induced surface defects are found and detected from the micro-Raman and XPS studies and correlated with the bound excitonic emission in the PL spectrum. Position-dependent Raman and PL profiles showed  Frontiers in Nanotechnology | www.frontiersin.org November 2021 | Volume 3 | Article 775732 uniformity in the MoS 2 layer for the growth at 540 and 560°C, while some nonuniformity in the MoS 2 layer was apparent for growth at 510°C. We believe the low-temperature growth of the luminescence MoS 2 dots array will stimulate further studies on growth on flexible substrates and its diverse range of applications in optoelectronics, bioimaging, and related areas.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding authors.