Efficient MoWO3/VO2/MoS2/Si UV Schottky photodetectors; MoS2 optimization and monoclinic VO2 surface modifications

The distinctive properties of strongly correlated oxides provide a variety of possibilities for modulating the properties of 2D transition metal dichalcogenides semiconductors; which represent a new class of superior optical and optoelectronic interfacing semiconductors. We report a novel approach to scaling-up molybdenum disulfide (MoS2) by combining the techniques of chemical and physical vapor deposition (CVD and PVD) and interfacing with a thin layer of monoclinic VO2. MoWO3/VO2/MoS2 photodetectors were manufactured at different sputtering times by depositing molybdenum oxide layers using a PVD technique on p-type silicon substrates followed by a sulphurization process in the CVD chamber. The high quality and the excellent structural and absorption properties of MoWO3/VO2/MoS2/Si with MoS2 deposited for 60 s enables its use as an efficient UV photodetector. The electronically coupled monoclinic VO2 layer on MoS2/Si causes a redshift and intensive MoS2 Raman peaks. Interestingly, the incorporation of VO2 dramatically changes the ratio between A-exciton (ground state exciton) and trion photoluminescence intensities of VO2/(30 s)MoS2/Si from < 1 to > 1. By increasing the deposition time of MoS2 from 60 to 180 s, the relative intensity of the B-exciton/A-exciton increases, whereas the lowest ratio at deposition time of 60 s refers to the high quality and low defect densities of the VO2/(60 s)MoS2/Si structure. Both the VO2/(60 s)MoS2/Si trion and A-exciton peaks have higher intensities compared with (60 s) MoS2/Si structure. The MoWO3/VO2/(60 s)MoS2/Si photodetector displays the highest photocurrent gain of 1.6, 4.32 × 108 Jones detectivity, and ~ 1.0 × 1010 quantum efficiency at 365 nm. Moreover, the surface roughness and grains mapping are studied and a low semiconducting-metallic phase transition is observed at ~ 40 °C.


Scientific RepoRtS
| (2020) 10:15926 | https://doi.org/10.1038/s41598-020-72990-9 www.nature.com/scientificreports/ bandgap 33 . The nanostructured thin layer of Mo 0.2 W 0.8 O 3 was deposited on the surface of VO 2 as a protective and anti-reflection layer 34 . The optical anti-reflective layer (Mo 0.2 W 0.8 O 3 ) is used to improve the responsivity of the photodetector and strongly eliminate the optical interference to minimize its undesired effects. Numerous applications including photodiodes 35 , image sensor 36 , and semiconductor photodetectors 37 have recorded the uses of the antireflective coating. The structural and morphological characteristics are also studied here. Raman, PL, electrical, optoelectronic characterization of strongly correlated oxide (VO 2 ), and 2D VDW heterostructure (MoS 2 -Si) will be discussed.

Results and discussion
Structural and composition. The crystal structure of vanadium oxide and molybdenum oxide was calculated in detail to understand the structural analysis of the deposited films. Figure 2a shows the XRD pattern of the deposited VO 2 thin film. XRD result of VO 2 shows a monoclinic phase with a JCPDS card number of [#96-153-0871] with a space group of C12/m 1 (12). The inserted crystal structure and visualizing 3D data are obtained using VESTA crystal software. , decrease the semiconductor-metal phase transition (SMT) 39 as reported here and to narrow the bandgap in the monoclinic phase 40 . Oxygen vacancies are acting as an electron donor with n-type conductivity which can change the electron orbital occupancy, band structures, and contribute to high photocurrent gain when staking with MoS 2 41 . The preparation of molybdenum oxide was achieved using the Mo target at sputtering temperature of 400 °C. Figure 2b shows the XRD Rietveld refinement of Mo-O bonding in molybdenum oxide. The calculated R-factors were found to be R wp = 7.52, R exp = 3.26, and χ = 25.31. The Refinement analysis shows that MoO 3 and MoO 2 Raman characterization. The physics behind interfacing structures such as 2D semiconductors and correlated oxides should receive high attention. The importance of these structures can be highlighted by controlling the band alignment of the 2D materials such as MoS 2 . Moreover, controlling the carrier mobility, coupling, and strain effect (as reported in the current work) 42 . Raman spectra of the deposited multilayer structure MoWO 3 /VO 2 /MoS 2 on p-type Si substrate are depicted in Fig. 3. The full range (200-1700 cm −1 ) Raman spectra of the deposited structures are shown in Fig. 3a and the magnified ranges of the full range spectra are shown in Fig. 3b-e. The Raman peaks of VO 2 at RT are shown in Fig. 3a,b,d which confirm its monoclinic phase. Whereas, Fig. 3c shows the two characteristic peaks of MoS 2 of E 1 2g (385 cm −1 ) and A 1g (405 cm −1 ) originate from their inplane and out-of-plane phonon characteristics 43 . The peaks positions, intensities, distance between E 1 2g and A 1g position at different sputtering time of MoS 2 are summarized in Table 1 and Fig. 4.
Many authors tried to study the effect of VO 2 in contact with transition metal dichalcogenides semiconductors (TMDs) layers such as MoS 2 and WS 2 31,32 . While in our current study, we tried to study the matching behavior   44,45 and consequently increases the film crystallinity and photocurrent as reported here. It is known that the intensity of Raman peaks is referring to high crystallinity effects. In most cases, Raman scattering is sensitive to the degree of crystallinity in a sample. Typically a crystalline material yields a spectrum with very sharp, intense Raman peaks, whilst an amorphous material will show less intense Raman peaks [46][47][48] . Table 1 and Fig. 4 show the Raman intensities of the corresponding thin films. These results show high attention to the applications of enhancing the Raman signal/intensity. The difference between peak positions of E 1 2g and A 1g in the presence and absence of VO 2 layer are deposited in Table 1. It shows a decrease with increasing the MoS 2 layer thickness from 30 to 240 s for both MoS 2 /Si and VO 2 /MoS 2 /Si structures, which may contribute to www.nature.com/scientificreports/ a decrease in the film layers and enhance the band gap 49 as seen from Table 1. However, the film deposited at 180 s is out of this base. Meanwhile, the differences between the intensities of the peaks show an enhancement in the peak intensity with increasing the MoS 2 film thickness which may contribute to high crystallinity effects. In counter, the film deposited at 240 s shows a decrease in the intensity. The highest intensity was observed for VO 2 / MoS 2 /Si with 180 s, whereas the lowest intensity was attributed to 240 s film. These results are concluded that the MoS 2 sputtering time of 180 s is optimized for VO 2 and MoS 2 optical coupling. Consequently, the observed results may highlight the importance of incorporating strongly correlated oxide through 2D VDWs MoS 2 structure to control film crystallinity, surface-enhanced Raman spectroscopy (SERS) of MoS 2 for better signal detection and spatial resolution [50][51][52] , optical coupling 53,54 , Plasmonic local-field enhancement 53 and optoelectronic behavior 51 .
For more details about intensity distribution versus position, Raman mappings of E 1 2g , A 1g and Si peaks of MoS 2 /Si at different sputtering time of Mo-O (30, 60, 120, 180, and 240 s) are investigated at 385, 410, and 520 cm −1 , respectively, and shown in Fig. S3 (a, b, c, d, and e). The Raman mapping was realized at large area (1.6 mm × 1 mm) with 320 × 200 data point using Leica microscope at 5 × magnification to provide an evidence about the ability to scale up the MoS 2 thin films. The wavelength, power and integration time of the used laser were 532 nm, 3 mW and 1 s, respectively. This analysis illustrated that 30, 60, and 120 s samples show high homogeneity over the full scale (1.6 × 1 mm 2 ). Although, 180 s sample shows regular agglomerations of nanoparticles for both E 2g and A 1g positions, the 240 s sample shows irregular clusters. The scaling up of MoS 2 can therefore be demonstrated until sputtering time < 180 min.
Photoluminescence measurements (PL). The PL spectrum of MoS 2 is strongly dependent on the number of layers of 2D VDWs structures. In other words, a strong PL peak may observe in single-layer MoS 2 or WS 2 film and decreasing with increasing the number of layers 55 . Trion is defined as a quasi-particle that can potentially carry out more information and data than electrons for which make them useful towards different applications such as optoelectronics and quantum computing 56 . Trions are consisting of three charged particles bound together by very weak bonding energy that makes them quickly fall apart 57 . It is known that the dominated peak in Figs. 5 and 6 attributed to the recombination of the photogenerated electron-hole pair, whereas the observed weaker peak at lower wavelength may be attributed to the valance band splitting due to the presence of strong spin-orbit coupling of MoS 2 58 . In the literature review, the definition of trion and exciton peak is dominated by their locations and the trion peak is located at lower energy than exciton peak [59][60][61] . Figure 5a and b shows the PL spectra of MoS 2 /Si and MoWO 3 /VO 2 /MoS 2 /Si structures at RT. The MoS 2 /Si has two peaks at ~ 679 and ~ 620 nm which corresponding to components from trion and A-exciton 62 . However, trion and A-exciton positions have changed by controlling the deposition time of MoS 2 as seen in Table 2. Also, this table shows the position of the characteristic peaks (A-exciton, B-exciton (higher spin-orbit splitting state), and trion) in the presence and absence of the VO 2 layer. Generally, it is seen that the trion peak shifted to lower energy by implementing the VO 2 layer onto the surface of the MoS 2 /Si structure as seen in Table 2.
The A-exciton peak has a higher intensity in the case of the VO 2 /MoS 2 /Si structure than in the case of MoS 2 /Si structure for 30 and 60 s deposition time of MoS 2 layer, Fig. 6a and b. The trion peak is enhanced and shifted to longer wavelengths by increasing the deposition time to 60 s. The increase in PL intensity refers to an enhancement in light emission efficiency and increases the density of states of the photo carriers by modifying the band structure 44 . At 30 s, the A-exciton peak intensity is higher for the VO 2 /MoS 2 /Si structure than the MoS 2 /Si structure. i.e., the incorporation of VO 2 dramatically changes the ratio between photoluminescence intensities of A-exciton and trion from < 1 to > 1 for VO 2 /(30 s)MoS 2 /Si structure. However, the opposite case is observed for the trion. Meanwhile, by incorporating the VO 2 layer, a blue shift in the trion peak is observed, while a redshift is observed for the A-excitons. The observed peak position of PL that shifted towards lower energy (redshift) attributed to the non-radiative electron-hole recombination effect. However, by increasing the deposition time to 60 s, the A-exciton peak have higher intensity compared with the trions peak in the case of VO 2 /MoS 2 /Si www.nature.com/scientificreports/ structure than MoS 2 /Si structure as seen in Fig. 6b. By increasing the deposition time of MoS 2 from 60 to 180 s, the B-exciton/A-exciton relative intensity increases, whereas the lowest ratio at 60 s deposition time refers to the high quality and the low defects densities of VO 2 /(60 s) MoS 2 /Si structure. Moreover, a redshift was observed for the trion peak and a slight blue shift when incorporating the VO 2 layer. The increase in PL intensity refers to an enhancement in light emission efficiency and increases the density of states of the carriers by modifying the band structure and consequently enhance the radiative recombination of carriers, similar results were observed in a compressively strained trilayer MoS 2 sheet 44,63,64 . This result shows that strong coupling between VO 2 and MoS 2 at 60 s Mo-O deposition time was observed at room temperature. While the PL intensity with MoS 2 /VO 2 structure has only enhanced with increasing the film temperature 31,32,65 , while our reported results show a dramatic enhancement in the PL intensity at RT by incorporating VO 2 layer on the surface of MoS 2 /Si structure. www.nature.com/scientificreports/ On the other hand, trion peak quenching was found in Fig. 6c,d,e; with increasing the Mo-O layer from 120, 180, and 240 s when VO 2 is deposited on MoS 2 /Si structure. We thought that the quenching of PL spectra in Fig. 6c,d,e may be owing to the fact that MoS 2 is an n-type with a close Fermi level to the conduction band. However, the deposited VO 2 layer at thicker Mo-O layer (120, 180, and 240 s) may shift Fermi level to the midband gap by drawback the electron coupling of VO 2 and MoS 2 66,67 . Similar results observed using back-gating with SiO 2 /Si 66 , dopants molecule like F4-TCNQ, metal-centered Phthalocyanine molecules on the surface of monolayered TMD 55,68 . It is interesting to note that in all Fig. 6a-e, A-excitons have higher intensities when incorporated VO 2 layer. This result draws high attention for enhancing A-exciton peak intensity and raises strong spin-orbit coupling by incorporating the monoclinic VO 2 thin layer. Moreover, the B-exciton peak was observed in the MoS 2 /Si structure at 606 nm (2.04 eV), but it did not appear in the VO 2 /MoS 2 /Si structure as seen in the insets of Fig. 6b-e. It is known that the PL spectra of MoS 2 , surprisingly, increases with decreasing layer thickness 62 . However, the origin of PL spectrum in MoS 2 arises from the direct excitonic electronic transitions which shows higher radiative recombination rate than nanocrystals 69 . Therefore, the enhanced photoluminescence with increasing the deposition time of Mo-O has to be attributed to a dramatically slower electronic relaxation factor κ relax as in Eq. 1, suggesting a substantial change in electronic structure of MoS 2 when going from the short to longer deposition time of Mo-O as seen in Fig. 6.
where κ rad , κ defect , κ relax , and η Lum are representing the rates of radiative recombination, defect trapping, and electron relaxation, Luminescence quantum efficiency within the conduction and valence bands, respectively.
It is concluded that by depositing the VO 2 layer on MoS 2 /Si structure, both trion and exciton peaks get shifted as seen in Table 2. It is implemented that the presence of the VO 2 layer on the surface of the MoS 2 /Si structure results in a redshift through trion peaks, while a blue shift for A-exciton. The peak position of PL for trion is shifted towards lower energy due to the occurred non-radiative electron-hole recombination. However, 30 s sample is out of this rule with a blue and redshift in the trion and A-exciton peak on the VO 2 layer, respectively. Surface topography and grain boundary mapping. AFM has been used to investigate the surface topography, roughness, grain, and grain boundary mapping. Surface topography and parameters such as average roughness R a (nm), root mean square or standard deviation of the height value R q (nm), height different or peakto-valley (R pv ), ten-point height (R z ), skewness (R sk ) and kurtosis (R ku ) as well as fractal and grain analysis were inspected by the XEI software. Figure 7a,d,g,j,m shows a 2D surface topography, Fig. 7b,e,h,k,n shows the 3D visualization, and Fig. 7c,f,I,l,o shows the grain boundary mapping of the prepared MoWO 3 /VO 2 /MoS 2 /Si thin film with sputtering time of 30, 60, 120, 180, and 240 s, respectively. The films that deposited at short deposition time show higher uniformity, while with increasing the deposition time a small clusters of different sizes less than 100 nm have been observed. The average roughness values of the prepared thin films have summarized in Table 3 and show that 30 s and 180 s thin films have the lowest and highest R a value of 3.28 and 48.0 nm, respectively. It seems that with increasing the deposition time of Mo-O, the accumulated nanoparticles show bigger sizes, consequently higher roughness factors. The calculation of the grain and grain boundaries of interfacing thin films are important parameters that provided information about the nature of interfaces between two layers.  www.nature.com/scientificreports/  www.nature.com/scientificreports/ grain/grain boundaries mapping in Fig. 7, the scanning area of the AFM images was selected to be 5 µm × 5 µm to provide evidence about the ability to scale up the MoS 2 thin films.

Temperature-resistance measurement (T-R).
The phase transition of the prepared VO 2 thin film has been performed using a four-probe measurement system connected to a heating stage ranging from RT to 78 °C. We investigated the influence of the 50 nm VO 2 thin layer on Raman, PL, and optoelectronic measurements of a few-layers MoS 2 with different sputtering times of Mo-O layer. So, the electrical semiconductor-metal phase change of VO 2 has been tested as depicted in Fig. 8. In our case, the VO 2 phase transition temperature was calculated to be 40 °C by controlling the sputtering condition (high vacuum and long-time annealing temperature) which may affect the lattice-strain and oxygen vacancy concentrations of VO 2 [71][72][73] . The reason for the low semiconductor-metallic phase transition may owing to the high concentration of oxygen vacancies. Oxygen deficient in vanadium oxide (VO 2−δ ) has reported many times to stabilize its metallic state 10 , decrease the semiconductor-metal transition (SMT) 11 and narrowing the bandgap of the monoclinic VO 2 phase 12 . Oxygen vacancies are electron donors with n-type conductivity which can change the electron orbital occupancy and band structures and contribute to the high photocurrent generation when staking by MoS 2 13 . To check these coupling effects between theses layer, UV-optoelectronic measurements have been carried out.
Electric and optoelectronic properties. This section discusses the electric characterization of MoS 2 /Si heterostructures before and after depositing the VO 2 layer under dark and UV conditions. In order to investigate the I-V and photoresponse of the prepared devices, we measured the I-V curve under dark and upon UV light illumination by applying a sweep voltage from − 5 to + 5 V for different sputtering times of Mo-O as shown in Fig. 9. This figure shows the electrical and optoelectronic properties of MoWO 3 /VO 2 /MoS 2 /Si thin film with different thicknesses of the MoS 2 layer. Figure 9a-e shows the current-voltage (I-V) curves under dark and UV illumination at MoS 2 deposition time of 30, 60, 120, 180, and 240 s, respectively. The observed photocurrent in this figure is larger than that reported for previously proposed MoS photodetector with lateral contacts arrangement 31 . The back Al and front Pd-Au contacts may paly important rule in that because these contacts allow the vertical electron transport in the heterostructure photodetector besides the lateral electron transportation and consequently 2D conductivity measurements 31,37 . The vertical electron transport offers a high density of active edge sites 37 . Also, our optimized heterojunction photodetector does not have a high-resistance layer like the SiO 2 layer that was previously used and obstruct the vertical electron transport 31 . The back Al contact has been used for better collecting signals. Al metal makes Ohmic contact type with p-Si, which is also observed when probed on two contact pads on the same side, however, noble metals such as Ag, Au, etc. make Schottky contact with p-Si. On the other hand, the Au-Pd front contact was built in the anti-reflection Mo 0.2 W 0.8 O 3 layer, in which the formed Schottky barrier height and width could be controlled by the current passing through the metal-semiconductor contact. Under dark conditions and as predicted by Basyooni et al., a non-linear I-V curve was obtained indicating that a good double-Schottky contact behavior was formed between the front Pd-Au contact and the film surface 34 . The position of asymmetric metal contacts can provide an integrated potential gradient assigned to the work function difference of asymmetric electrodes as previously stated for various applications, such as gas sensors 34 and photodetector 74 , which leads to enhanced device performance as reported here. For instance, Casalino et al. used an asymmetric Al-Si-Cu (metal-semiconductor-metal) structure-based Si photodetector 75 . Moreover, several studies used asymmetric metal contacts for the photodetector application to control the dark current 70,74 . www.nature.com/scientificreports/ Figure S4 shows the linear and semi-logarithmic scale current-voltage characteristics of MoS 2 /Si device without VO 2 layer under the dark condition with different sputtering times of MoS 2 layer; 30, 60, 120, 180, and 240 s. The positive part shows an increase in the associated dark current with increasing the sputtering time from 30 to 60 s. While at 120 s, a jump in the forward dark current is observed due to the related folding-effects in MoS 2 . Folding effect decreases the interlayer coupling and enhances the photoluminescence emission yield of A-and B-exciton peaks as seen in Fig. 5. Whereas instability measurements were observed at the negative bias part. The large increase in the negative dark current for MoS 2 (60 s) may be attributed to the release of charges that were trapped on MoS2 's surface at the interface trap sites (oxygen sites). The highest reverse dark current, which suggests the lowest potential barrier, was observed at 60 s. This may be ascribed to the values of the optical band gaps as shown in our previous study, whereas the 60 s MoS 2 -Si thin film displayed optical band gaps of 1.75 and 2.01 eV 33 . By increasing the sputtering time the reverse dark current decreases and almost becomes identical for sputtering time ≥ 180 s, as shown in Fig. S4(a,b). Figure 10a-e shows the log-current curves under dark and UV illumination of 30, 60, 120, 180, and 240 s conditions, respectively. It is clearly seen that UV illumination shifts the logarithmic I-V curve towards the negative voltage region. This behavior may address the induced strain effects from the VO 2 layer or unidirectional charge transport mechanism from the top to bottom layer due to the different electron concentrations 76 . Interestingly, it seems that the VO 2 layer enhances the positive and negative current. Meanwhile, the dark current obtained after depositing the VO 2 layer on the surface of MoS 2 /Si is about 2-3 folds' improvement over pure MoS 2 /Si device for 30 and 60 s samples, as in Fig. 10a,b. The observed higher value of photocurrent under UV illumination is also attributed to enhancement though the band-to-band excitation in the VO 2 /MoS 2 /Si region. Moreover, carrier recombination and tunneling across the device junction may be addressed as a reason for the enhancement I-V under UV illumination. Nevertheless, fast response and recovery times, high responsivity, high reliability, and low signal-to-noise ratio are important characteristics for detector applications 77,78 , which is discussed below in detail.  Figure S5 (supplementary information). Note that the ON/OFF and OFF/ON transitions of the UV light source are repeated many times for each 5 s at a bias voltage of 1 V. The time-resolved photoresponse curves in Fig. 11 show different sputtering time dependence. The curves at 30 and 120 s show increasing on/off behaviors, whereas the curves at 60, 180, and 240 s show stable on/off behaviors. The increasing on/off behaviors in Fig. 11a,c may come from some organic trap states that accumulated during the CVD sulfurization process. Wile, the stable on/off behaviors in Fig. 11b,d,e can be attributed to the high stability, high quality, and low density of the defects.  www.nature.com/scientificreports/ The response/raise time was measured when the light source turned on, while the recovery/decay time was measured when the light turned off as shown in Fig. 12a. The response and recovery times have been estimated from the ON/OFF dynamic photoresponses at different sputtering times, Fig. 11. The combined sputtering and CVD deposition process of MoWO 3 /VO 2 /MoS 2 /Si UV photodetector device shows symmetrical response and recovery time which not exceed 0.25 s using the selected wavelength (365 nm) and bias (1 V), as shown in Fig. 12b. Consequently, our proposed photodetector is considered more efficient than the previously tested photodetector by Ang et al. 78 .

Transient response.
The fast response and recovery speed, indicating that electron-hole pairs could be effectively generated and separated in the proposed structure under UV illumination at room temperature. It is important to note that the fabricated device using MoS 2 (120 s) shows the fastest response/recovery times (0.19 s at 1 V) among the studied devices as shown in Fig. 12b. The fast response/recovery time at 120 s can indicate the fast and stable generation and separation of the electron-hole pairs. Unlike Dhyani and Das who reported rapid response for the silicon-MoS 2 photodetector@ 580 nm and 3 V bias 79 , our measurements are performed without any external series resistance. Nevertheless, it is known that higher applied bias voltage can generate more photocurrent and consequently decrease the response and recovery time. A clear high photocurrent can be observed in the ON state at 1 V which makes the gate voltage lowers the potential barriers at the contacts, resulting in highly efficient photogenerated carrier extraction and thus increased photocurrent at a low applied voltage (1 V). The reason behind this is that the gate voltage can affect the height of the Schottky barrier between the metal contact and film surface and thus shift the Fermi level 80,81 . It seems that our designed photodetector did not require high bias voltage which makes it more applicable for low power photodetector technology.

Photocurrent gain (Pg) and photoresponsivity ( R ). The induced photocurrent I ph is given by
where I ph increases with increasing the applied voltage and the light power 82 . Photocurrent gain (P g ) can be defined and determined by P g = (I photo − I dark )/I dark , where I photo and I dark are photocurrent and dark current respectively 77 .
Also, the detector responsivity ( R ) can be expressed as R = �I/(A × P) , where I is the difference between the photocurrent and dark current, A is the illuminated area, and P is the UV light power. Figure 13a shows the photocurrent and photocurrent gain of the tested samples under a 365 nm UV illumination source.  83 . The enhanced photoresponsivity by interfacing the VO 2 layer may be owing to the film strain include stresses arising from the different thermal expansion coefficients of the VO 2 and MoS 2 /Si film due to a high deposition/sulphurization temperatures of ~ 400/650 °C and growth stresses arising from crystal structure changes after deposition. Nevertheless, more efficient light absorption involving more e-h pairs generation, resulting in higher mobility and more detection capability.  Fig. 14a, where EQE varies from 6.6 × 10 8 to ~ 1.0 × 10 10 at 365 nm which considered higher than the mesoscopic multilayer MoS 2 as reported before 85 . Another important figure of merit of a photodetector is the detectable signal 82 , referred by the specific detectivity measured in Jones, which given by D * = (AB) 0.5 R i n (cmHz 1 2 W −1 ) , where A is the effective area of the d in cm 2 , B is the bandwidth, and i n is the measured noise current. If the shot noise from the dark current is the main noise source, the specific detectivity can be simplified as D * = R A 0.5 (2eI dark ) 0.5 ,Where e is the charge of an elementary electron 86  Mechanism. Now we turn to the underlying photoresponse mechanism of the VO 2 /MoS 2 /Si as a UV photodetector device. The photoresponse properties of VO 2 /MoS 2 /Si heterojunction can be understood from the  www.nature.com/scientificreports/ energy-band alignment diagram as in Fig. 1b. Due to the free dangling bonds of the surfaces of MoS 2 film, the MoS 2 /Si heterojunction can be affected by lattice matching. Clearly, the implanting of VO 2 layer-based UV photodetector was overwhelmingly play an important role in enhancing the Raman signal/intensity, PL intensity, electrical and optoelectronic performance of MoS 2 /Si device. Under the zero-bias condition, an insignificant current was observed due to the high depletion layer at the n-p (MoS 2 -Si) junction which restricts the movement of the carriers. Under VO 2 interfacing, both positive and negative current increased significantly and the photocurrent I ph of VO 2 /MoS 2 /Si film is much higher than that in MoS 2 -Si, which can be attributed to the more photon absorption on the top of MoWO 3 /VO 2 layer and larger photocurrent-gain due to higher carrier mobility 87,88 . With increasing the amount of Mo-O (deposition time of 30, 60, 120, 180, and 240 s), the induced current is enhanced due to the more electron-hole pair generation by UV light absorption and the applied voltage shifts dramatically towards negative voltage, which indicates the continuous accumulation of electrons in the vertical VO 2 -MoS 2 channel as seen from the logarithmic scale current. In the positive voltage region, the MoS 2 /Si n-p structure shows that umpteen electrons are accumulated on the MoS 2 band which shifts Fermi level near the conduction band. Because 1 V is able to decrease the depletion width and the barrier height, electrons are able to overcome the barrier height through thermionic emission, resulting in a high flux of photocurrent and more efficient photocurrent extraction. It is interesting to observe that under a shorter sputtering time of Mo-O (30, then 60 s), a significant forward photocurrent was observed which did not observe before in MoS 2 /Si structures 52,89 . Meanwhile, with increasing the Mo-O content, MoO 3 starts to get folded and the reverse photocurrent starts to get highlighted as seen in the semi-logarithmic scale I-V, Fig. 10.

Conclusion
In summary, the next generation of optoelectronic devices integrates the physics of light-matter interaction of 2D materials at nanoscale for light-harvesting applications and these optoelectronic devices can control the light that converts trions, excitons, and photons to electrical signals. Our approach is based on a high vacuum deposition of Mo-O compound at 400 °C, followed by a sulphurization process in a chemical vapor deposition tube.

Materials and methods
Device fabrication. Preparation of MoS 2 layer on p-type Si substrate has prepared through two steps in physical vapor deposition (PVD)-Radio Frequency magnetron sputtering system, followed by chemical vapor deposition (CDV) process. Si substrates were cleaned through many steps; firstly, kept in NH 4 OH-H 2 O 2 solution diluted with de-ionized (DI) water for 5 min at 75 °C, then rinsed with DI water for 5 min. After that, they left in HF (%5) solution for 5 s, then rinsed in DI water and dried with high purity N 2 . Immediately, the cleaned Si-substrates transferred to a 3 × 10 −7 Torr RF magnetron sputtering system (VAKSIS Midas 3M1T). In-situ Ar-plasma source has activated for 10 min at a power of 100 W and low pressure of 6 × 10 −3 Torr at room temperature to activate the Si surface. Mo-O thin films were grown using a 3-inch. pure molybdenum target (99.9%) utilizing Ar plasma as a carrier gas and O 2 as a reactive gas. The substrate temperature was stabilized at 400 °C for more than 30 min before the deposition process with steps of 100 °C/30 min. The O 2 and Ar flow rates were kept constant, whereas O 2 /(Ar + O 2 ) of ¼ ratio. The deposition was carried out at 5 × 10 −3 Torr and 137 W with different sputtering times of 30, 60, 120, 180, and 240 s. The system was kept to cool down normal up to the room temperature (RT), then immediately transferred to the two-zone CVD quartz chamber (MTI-OTF 1200 system) for the sulphurization process. The as-deposited molybdenum oxide (Mo-O) thin films transferred to the center of the CVD furnace and the temperature is raised to 650 °C. Sulphur powder (0.5 g) is put in a ceramic boot with 100sccm high purity Ar source. An external heating belt with a distance of 50 cm to the substrate was used to evaporate the sulphur for 22 min. Then, the system cooled down until RT with the same flow rate of Ar (100 sccm). After forming the MoS 2 layer, a thin layer of monoclinic VO 2 has grown. The same sputtering system was used with a 190 W deposition power and Ar/O 2 ratio was 41/2.2 sccm, while the deposition time was set to produce 50 nm film thickness. Then the samples were in-situ annealed at 400 °C for 2 h with 50 sccm Ar flow. A protective and anti-reflection thin layer of Mo 0.2 W 0.8 O 3 was deposited on the surface of VO 2 as optimized in our previous work 34 . High vacuum thermal evaporation system was used to deposit aluminum and gold-palladium that used as a back and front contacts, respectively. Device characterization. The crystal structures were analyzed using Grazing Incidence X-ray diffraction (XRD GNR ADP PRO 2000), with CuKα (λ = 1.5405 Å) radiation source with a step of 0.01. VO 2 layer was deposited at a high vacuum condition to ensure its high crystallinity and low semiconductor-metallic phase transition. Parameters such as space group, diffraction peaks, angles, Wyckoff position of vanadium (V) and www.nature.com/scientificreports/ oxygen (O) atoms, ratio of O:V, and oxygen vacancy concentrations were calculated from refinement analysis. The refinement calculations were done using Match and Fullprof Suite program. Moreover, the refined structures were plotted in a three-dimensional view using 3D visualization VESTA program. The surface morphology was recorded using scanning electron microscopy (SEM) TESCANMAIA3 XMU. Atomic Force Microscopy (AFM) has been used to investigate the surface topography, roughness, and grain mapping. Each sample was characterized by XE-6 AFM (Park Systems Corp., Suwon-Korea) that controlled with XEP software for data acquisition and XEI software for image analysis and processing. AFM images were obtained through a 0.5 × 0.5 μm area (x-y accessible area) at a 0.5 Hz scan rate. Measurements were taken with a non-contact mode using a PPP-NCHR silicon cantilever consisting of tip radius < 10 nm and 42 N/m force constant (Nanosensors TM, Neuchâtel-Switzerland). Raman measurements and photoluminescence (PL) spectra were carried out using Renishaw in Via Confocal Raman microscope with a 532 nm laser beam, while an incident laser power of 3 mW was chosen to acquire a single Raman spectrum. The temperature-resistance measurement of monoclinic and high crystalline nanostructure VO 2 thin film has been performed using a four-probe measurement system connected to a heating stage ranging from RT to 100 °C. The electrical and optoelectronics measurements were measured using 2450 Kethley Source -Meter and 365 nm ultraviolet (UV) light lamp for optoelectronic measurements.