Ultrafast and stable planar photodetector based on SnS2 nanosheets/perovskite structure

Two-dimensional (2D) transition metal dichalcogenides are promising candidates of photodetectors where they are commonly grown parallel to the substrate due to their 2D characteristics in micrometer scales from exfoliation of bulk crystals or through high temperature chemical vapor deposition (CVD) methods. In this study, semi-hexagonal vertical nanosheets of SnS2 layered have been fabricated on FTO substrate without using Sn source through CVD method at relatively low temperature (500 °C). Due to exceptional band alignment of triple cation lead perovskite (TCLP) with semi-hexagonal SnS2 nanosheets, an improved photodetector has been fabricated. This type of photodetectors fabricated through lithography-free and electrodes metallization free approach with remarkable fast response (20.7 µs/31.4 µs as rising /falling times), showed high photoresponsivity, external quantum efficiency and detectivity of 1.84 AW−1, 513% and 1.69 × 1011, respectively under illumination of incident light with wavelength of 445 nm. The stability of the photodetectors has been studied utilizing a protective PMMA layer on the perovskite layer in 100% humidity. The introduced growth and fabrication process of the planar photodetector, including one/two dimensional interface through the edges/basal planes of layered materials with perovskite film, paves a way for the large scale, cost-effective and high-performance optoelectronic devices.

www.nature.com/scientificreports/ with adding a little amount of Cesium (Cs) leads to better stability and performance of perovskite material 24,25 . The lower trap density as 1.75 × 10 13 cm −3 for three cationic lead iodide based perovskite on formamidinium and suitable electron mobility in the range of 54.6-157.1 cm 2 s −1 V −1 makes this type of perovskite the strong light absorbing layer in the photodetector devices 26 .
Although solar-cell like multilayers are capable to be planar photodetector, these structures often suffer from low sensitivity and huge electrical hysteresis 27 . Applying the interface perovskite layers with other functional materials such as TMDs 28 , carbon based material 29 , and metal oxide 30 , is powerful approach to improve the requirements of photodetectors. In this research, the vertical hexagonal SnS 2 nanosheets on FTO substrate has been achieved using a modified CVD method without using extra tin (Sn) source at a relatively low temperature of 500 °C, which inherent Sn source in the FTO substrate plays a significant role to prepare these structures. To avoide the complexity of device fabrication based on TMDs, such as mechanical/chemical exfoliations, flakes transferring and lithography steps along with high vacuum deposition for electrode preparation, the planar TMD photodetector design based on patterned FTO has been presented here. Also, to improve photodetection, designing based on triple cation perovskite of Cs 0.0 (FA 0.83 MA 0.17 ) 0.95 Pb(I 0.83 Br 0.17 ) 3 chemical deposition on vertically SnS 2 nanosheets has been considered; this engineered photodetector exhibits significant performance in the wide spectrum range as well as fast photoresponse. The stability of the device has been investigated at 100% humidity and room temperature conditions, through a water resistance poly (methyl methacrylate) (PMMA) protective layer.

Result and discussion
Physical and chemical characterizations. Here, we show that the vertical growth orientation of hexagonal with 2D SnS 2 nanosheets through a standard CVD system at low temperature (500 °C) on FTO glass. Although, some reports shows vertically SnS 2 nanosheets grown on the substrate at low temperature, here the FTO plays both as the substrate and tin source for SnS 2 growth in our CVD setup (Fig. 1a). The proposed method is capable of producing SnS 2 nanosheets in a large scale as shown in Fig. 1b. The FESEM analysis of the prepared samples depicts the vertically orientated morphology with clear semi-hexagonal shape including three or five crystal edges. The Fig. 1c shows a thickness of about 40 nm from an isolated SnS 2 flake on Si/SiO 2 substrate. The Fig. 1d illustrates the vertically SnS 2 nanosheets on FTO as well as the thickness of about 5 µm, by applying 500 mg of the Sn source. Figure 1e-g demonstrate elemental mapping images of S, Sn and Sn:S atoms, respectively, using energy dispersive X-ray spectroscopy (EDS). The related homogenous elemental distribution reveals the proportional elemental ratio for S to Sn in the nanosheets (Fig. S1). Further analysis has been performed for as-grown SnS 2 nanosheets using HRTEM (Fig. 1h). The interplanar lattice spacing can be extracted as 3.15 Å corresponding to (100) and (010) planes and the lattice spacing of 5.9 Å is corresponding to the basal plane of (001).
The selected area electron diffraction (SAED) pattern of the sample clarifies a single set of the perfect hexagonal pattern, indicating the crystal nature of the SnS 2 nanosheet with the hexagonal crystal structure (Fig. 1i). The crystal planes in SAED are (100) and (010) planes with the angles of ~ 120° degree; the main exposed surfaces of SnS 2 nanosheets are (001) basal planes. These results indicate good agreement with the previously reported lattice parameters of SnS 2 crystals grown by CVD methods 15,31,32 .
The crystallographic structure of SnS 2 nanosheets extracted from the diffraction pattern of XRD spectra, well-matched with a hexagonal structure of 2H-SnS 2 nanosheets (JCPDS no 01-83-1705). Figure 2a shows diffraction peaks corresponding to (001), (002), (101), (102), (003), (110), (111) and (110) facets. It's clear that (001) is the preferential orientation of crystal growth, which agrees with SAED pattern of SnS 2 nanosheets. Moreover, based on Bragg's equation, average spacing in the basal plane of SnS 2 is about 5.8 Å which is consistent with the HRTEM analysis and disclose the (001) direction as the main basal planes of grown SnS 2 nanosheets with lattice spacing distance of 5.9 Å. XPS analysis of the prepared SnS 2 sample indicates the presence of the Sn and S elements, without any other additives (Fig. 2b). The high-resolution S 2p core level analysis at the binding energy of 161.2 eV and 162.3 eV shown in Fig. 2c, attributes to S 2p 3/2 and S 2p 1/2 , respectively. The strong peaks around 486.4 eV and 494.8 eV (Fig. 2d) assign to Sn 3d 5/2 and Sn 3d 3/2 , respectively is corresponding to chemical states of Sn 4+ and S 2− of the SnS 2 33 . Raman analysis (Fig. 2e) displays two peaks located at 320 cm −1 and 208 cm −1 appointed to A 1g (out of the plane) and E g (in-plane) modes of the SnS 2 , presenting an explicit fingerprinting of 2H phase 34 of the SnS 2 crystals in this work. These results confirm the growth of pure SnS 2 nanosheets with high crystal and chemical quality. The diffusion transmission and reflectance spectroscopy of SnS 2 on FTO/glass substrate was utilized to calculate the absorption coefficient spectra, using the following equation 35 where α , T , R, d, h, ν , A and E g are absorption coefficient, transmission and reflectance, film thickness, Planck's constant, frequency of light and bandgap respectively. In Eq. (2), n = 4 represents an indirect bandgap semiconductor such as the SnS 2 bulk sample 36 . As shown in Fig. 2f, the estimated optical band gap for the film containing SnS 2 nanosheet arrays-grown by applying 500 mg of the tin source was obtained as 2.11 eV by extrapolating of the linear part of the ( αhν) 1/2 versus photon energy plot. Moreover, Mott-Schottky (MS) analysis was utilized  where C sc is the capacitance of the space charge layer, e 0 is the electron charge; ɛ r = 17.7 is the dielectric constant of the SnS 2 38 , ɛ 0 is the vacuum permittivity, V is the applied potential versus Ag/AgCl, T is the absolute temperature, and K is the Boltzmann constant. The positive slope of capacitance variation as the function of applied voltage (Fig. 2g), indicates the fabricated SnS 2 nanosheet is an n-type semiconductor. The flat band potentials obtained by linear extrapolation of the MS plot is about − 0.44 V versus Ag/AgCl which stands for E f as − 4.25 eV versus vacuum level energy. Carrier density of the SnS 2 nanosheets is about 3.97 × 10 17 cm −3 ; the conduction band energy of the n-type SnS 2 versus absolute vacuum energy state (AVS) can be extracted through the following equation 38 The N CB is the effective density of the states close to the conduction band which can be evaluated as 7.32 × 10 18 cm −3 within the electron mass of 0.43m 0 39 . Using Eq. (4), for a bandgap value of 2.11 eV, the conduction and the valance band of the growth SnS 2 nanosheets are positioned at 4.20 eV and 6.31 eV, respectively versus absolute vacuum standard level (Fig. 2h). These results are also in good agreement with other reports 40,41 . Crystal growth mechanisms. The influence of inter-distances between substrates and sulfur source position 42 , temperature and time parameters 40 on the vertical growth of the SnS 2 nanosheets in CVD method has been reported. Here, through a modified CVD process and using the FTO/glass as the substrate with no extra Sn source, the effect of sulfur precursor on the SnS 2 nanosheet growth has been surveyed by FESEM analysis. In the case of S = 200 mg, as shown in Fig. 3a, a quite low density of SnS 2 nanosheets have been distributed on FTO substrate; the average lateral size and thickness of the SnS 2 with semi hexagonal structure was about, 5.6 µm and 135 nm, respectively (Figs. S2a and S3a). By increasing the sulfur source to 300 mg, the quantity of the SnS 2 nanoflakes on FTO has been incredibly increased (Fig. 3b); the average lateral size of SnS 2 nanoflakes increased to 9.5 µm while the thickness sharply decreased to 40 nm (Figs. S2b and S3b). However, by using further sulfur Therefore, the increment of the sulfur precursor leads to more SnS 2 nanosheets with larger lateral sizes and sharper edges. A graph as an inset of Fig. 5d with 2Y axis summarizes these trends. The nanosheets showed relatively similar aspect ratios of semi hexagonal shapes due to high oriented growth in almost [100] and [010] rather than [001] direction 43 .
Crystal growth of hexagonal SnS 2 nanosheets can be discussed from two aspects: Bravais rule and surface energy theory. According to the Bravais rule, the most morphologically favourable crystal forms would be those planes having the highest reticular densities 44 . Based on XRD and HRTEM characterizations, SnS 2 has a hexagonal layered structure with a unit cell of a = b = 3.65 Å and c = 5.9 Å. Similar to other TMDs structure, the Sn and S atoms are covalently bonded together in the layers, while the adjacent layers with the highest lattice distance of c , are stacked through weak Wan der Waals forces 45 . The anisotropic nature of atomic bonding in SnS 2 (Fig. 4a)  Since the lattice space of (001) is 5.9 Å, which is larger than the {010} planes, the highest reticular densities are related to {100}. On the other hands, regarding surface energy theory, the growth of a crystal occurs through the plane which has the lowest surface energy based on the Gibbs principle in thermodynamic 46 . Hence, for the stable crystal nuclei with the constant free energy per unit volume, the surface energy should be minimized in a crystal that is bounded by n faces as in the following equation where a i is the area of the i th the face of the crystal and g i is the surface energy per unit area of the i th face. As the {001} plane of SnS 2 may own the lowest surface energy (g 100 = 0.034 eV/Å 2 and g 001 = 0.0065 eV/Å 2 ) 47 , the growth of crystal should occur in (001) plane. As illustrated in Fig. 4b, after forming the SnS 2 nucleus on the FTO, the new species favourably consolidate themselves into crystal lattice along the {100} planes rather than {001} planes due to much higher surface energy along [100] and [010] than the [001] 42,47 . Therefore, rapid transverse growth and slow vertical growth makes the (001) plane as the most exposed basal plane.
The lattice mismatch of nucleation A on the substrate affected the growth of the final structure (Fig. 4c). When the substrate and the film keep their different bulk lattice constants, the growth is stress-free, or incoherent. The growth is pseudomorphic or coherently strained when the film is stretched or compressed so that the film and substrate of the in-plane lattice constant are matched 48 . In the latter, the elastic energy of mismatch is released by the formation of islands or dislocations, partially or totally. Moreover, in this study, as the growth process has been performed at a relatively low temperature (500 °C), overcoming the energy barrier by SnS 2 species is almost impossible and the lateral diffusion on the surface of the substrate is practically dismissed 42 . So, it seems that the stored elastic energy is released by the formation of 3D islands or misfit-dislocations. In the 3D islands mode, the Gibbs free energy is g 3D which is related to the formation n density of 3D island. By regarding the square-based pyramidal island on the wetted substrates which is similar for the typical SnS 2 nanosheet growth in this study (Fig. 4d), this parameter is equal to the following equation 44 where H and S are height and average lateral size of an island A on the substrate B; θ is the intersection angle between pyramid edge and substrate, and g i is the surface energy of lateral face i of the polyhedral island. By adding the sulfur amount, the probability of the reaction and formation of nuclease increases, so the surface energy consumed per unite area increases. Based on the above equation, by increasing the g 3D in the same height of nuclease, it is expected that the lateral size of pyramided nuclease increases. We observed this phenomenon for the SnS 2 nanosheets grown on the FTO substrate. Subsequently, based on these equations, it is predictable that the thickness of the SnS 2 nanosheet decreases by adding the sulfur amount through the introduced CVD method.
As the last point, the Wulff 's theory could be taken into account to understand the morphology of SnS 2 nanosheet in this research. In this regard, the equilibrium shape of free crystal A can be described by introducing the constant ratio of = g j h j between the surface energy g j of each j facet bounding to the crystal of the given volume and distance h j of that facet from a common point inside the crystal so-called Wulff 's point 44,46 .
It follows that for a free crystal growing near the equilibrium condition, the crystal shape for different λ (different sizes) is self-similar around the Wulf 's point (Fig. S4a). Hence, the constancy of the shape ratio can be expressed as For a supported crystal "A" on substrate "B", the equilibrium shapes were determined by minimization of surface and interface energies. In the absence of misfit between A and B, the self-similarity around the common point S on the substrate has been sustained (Fig. S4b) which is a result of the size independence of shapes ratio (Wulff-Kaishev theorem) 44,46 .
where H = h A + h AB is the emerging height of crystal; h AB is the common distance of the interface surface from all the pyramid apexes (Wulff 's point) taken positive or negative according to whether the Wulff 's point is outside or inside the substrate, respectively; β AB is the specific adhesion free energy of face i = 1 of A on B and g A is the corresponding free energy of this face.
For the case of a strained polyhedral crystal, A supported on substrate B, similar to SnS 2 nanosheet growth in this study, elastic bulk energy plays an important role. Assuming coherence growth, the elastic strain accumulates both in crystal and in the nearby region of the substrate during growth (Fig. S4c), which is reduced by relaxation-energy factor R. In this case, by minimization of the free energy at the constant volume, the equilibrium shape of the crystal can be gained, which contains the elastic energy. The shape ratios for crystal "A" with volume V A having free facets of area S i and a contact area S AB with substrate "B" of area S B , based on generalized Wulf-Kaishew theorem, obeys the following equation 44 (5) n i a i g i = min www.nature.com/scientificreports/ where C 1 and C 2 are the constant elements. Thus, the shape ration depends on the size of the crystal and there is no more self-similarity of equilibrium shape, consequently. In our catalyst-free growth, as no kinks exist on the substrate, the nucleation can occur at any site on it 43 . The strained polyhedral SnS 2 nucleus crystal deposited on lattice mismatches of FTO substrate and asymmetric SnS 2 nanosheets growth happened to release its stored elastic energy. These processes occur for nucleation-growth as layer by layer for SnS 2 flakes 15,42,43,47 .
The vertically grown crystal is strongly suitable for planar electronic devices because it provides a higher density of SnS 2 nanoflakes with enormously exposed surface/edges per unit area 12 . Due to this fact, we have used this fascinating structure as a primary material for photodetectors.
Photodetector operation. Pristine SnS 2 based photodetector. As vertically species provides enormously exposed surface/edges per unit area 12 , we have used this fascinating SnS 2 structure as a primary material for low-cost planar photodetector. This photodetector been developed by using the FTO/glass with 25 µm etched strip, in which two sides of the etched area act as the planar electrical contacts thanks to the highly conductive properties of FTO film. In this study, by using the introduced CVD method, grown SnS 2 nanosheets helps to bridge between two separated parts on the substrate, as shown in Fig. 5a,b. So, some traditional steps such as TMD's flake transferring, lithography, oxygen plasma process, and metallization process for electrical contacts such as Cr/Au or Pt deposition, have been avoided which offer photodetector fabrication with no need for high vacuum condition process. The higher magnification of the vertically SnS 2 nanosheets has been shown in Fig. 5c. The interconnection of these structure is more clear in the Fig. 5d.
The schematic of the planar photodetector is illustrated in Fig. S5a, where the light is illuminated from the backside. The current evolution vs. different light wavelengths at 3 V (Fig. 5f) presents the highest photocurrent about 27.3 × 10 -8 A at 445 nm of illumination. Figure 5g, displays the dependence of photocurrent on the light intensity at 445 nm wavelengths, which can be described by power-law as I ∼ P β , where P is the light intensity and β is the parameter related to the trap states on the photovoltaic materials. The trap states originating from sulfur vacancies and/or charged impurities in the SnS 2 .More trap sites could be filled with photogenerated charge carriers as the light intensity rises 49 . By fitting the measured data by the power-law equation, the value of β is 0.61 which is close to the other reports 15,49 .
Next, the photoresponse time as one of the critical parameters of photodetector was tested at wavelengths of 445 nm with 5mW/cm 2 light intensity at 3 V applying bias. Figure S5c,d display the time-resolved photoresponse curve under illumination/dark exposure cycles for a long time at 1 Hz and 1 kHz, respectively. It is obvious that the photodetector has reasonable stability and steady response toward the incident light. The one cycle of current vs. time of SnS 2 based photodetector at 1 kHz (Fig. 5h), describes a fast rise time (τ r ) about 0.36 ms and decay time ( τ d ) of 0.47 ms. This record of the rise and decay times that are both above tenth of milliseconds reflects an enhancement by at least two orders of magnitude relative to other parallel and vertical SnS 2 nanosheet based photodetectors 34,49 . SnS 2 /TCLP photodetector. The Cs 0.05 (FA 0.83 MA 0.17 ) 0.95 Pb(I 0.83 Br 0.17 ) 3 , abbreviated as Cs 0.05 M, where M stands for mixed perovskite-with appreciated electrical properties and remarkable stability besides its high light absorption coefficient, is a promising candidate for optoelectronic devices 25 . The XRD analysis of this material (Fig. S6a) confirms that crystalline planes of Cs 0.05 M perovskite layer have been formed properly. The estimated optical bandgap is about 1.63 eV which is in agreement with the clear peak of photoluminescence spectra in 765 nm (Fig. S6b). Therefore, the Cs 0.05 M should be a potential candidate to be utilized as an absorbing layer with SnS 2 nanosheets due to the well-matching of energy bandgaps 25 . In this research, facile, low cost and large scalable planar design of SnS 2 /TCLP photodetector has been fabricated through deposition of the TCLP on the SnS 2 nanosheets. The schematic of this photodetector is shown in Fig. 6a. The inset Fig. 6a indicates the SnS 2 /TCLP layer in its atomic structure. The top-view FESEM analysis of SnS 2 /TCLP (Fig. 6b) displays full embedding of the SnS 2 nanosheets grown on patterned FTO/glass by TCLP film. While the thickness of pure SnS 2 (5 µm) layer and SnS 2 /TCLP (5.5 µm) (Fig. S7a) layers are almost the same, the SnS 2 /TCLP indicates more light absorption in the range of 320-740 nm, in comparison with pristine SnS 2 and perovskite layers (Fig. S7b). It can be expected that the SnS 2 /TCPL layers presents enhancement in the photocurrent rather than pure SnS 2 and TCLP films.
Figures 6c and S7c show the current evolution of the photodetector under the illumination of different wavelengths under the same intensity of 5mW/cm 2 . As Fig. 6d displays by applying the 3 V and variation the irradiation wavelength from 395 to 450 nm, the photocurrent raised and reached a maximum value at 445 nm. By increasing the wavelengths of illumination up to 520 nm, the current slightly decreases and at 595 nm, the current falling takes place for longer wavelengths. A gentle decrease of the photocurrent by wavelength variation for the device can be attributed to the characteristics of light absorption of the perovskite layer in the visible region. The current variation vs. light intensity evolution fitted by the power-law equation gives β = 0.7 (Fig. 6e). The increment of this parameter compared to that of pristine SnS 2 based device indicates the lower carrier recombination process in SnS 2 /TCLP device. The photocurrent evolution versus time with on/off modes of (9)   Fig. 6g before contact together. After the physical connection of these two layers, the interface of the SnS 2 /TCLP, band bending occurs with band offset at the valence and conduction bands (Fig. 6h) which facilities charge transferring in one preferred direction. Different Fermi levels of these two materials also lead to the formation of depletion regions and provide a built-in electric field to drive the separated electron-hole pairs in the TCLP film. When the light is illuminated, electron-holes pairs are created in the TCLP layer (Fig. 6i). Due to the existence of the built-in electric field, electrons are transferred to the SnS 2 nanosheets, while holes are restricted in the perovskite layer. The accumulated charge carriers in the SnS 2 nanosheets increase (decrease) the Fermi level in the SnS 2 (TCLP), respectively 50 . This engineered band bending results in more photo-charges creation/transferring at the interface toward metallic contacts. The photocurrent generated per unit power of incident light on the effective area-photoresponsivity (R)sensitivity of the photodetector-specific detectivity (D*)-and the number of electrons detected per incident photon-external quantum efficiency (EQE)-are the key parameters of photodetectors which calculate through following equations: where I ph is I p (light current)-I d (dark current), P 0 is the light intensity; S is the effective illuminated area which here is about 13.5 × 10 -9 m 2 ; e stands out the elementary charge; h is Plank's constant; c represents the light velocity; R is the responsivity; and λ is the incident light wavelength 49,50 . Figure 6j, h and l demonstrate the evolution of the R (AW −1 ), EQE % and, D* (Jones) as the function of incident light intensity at 445 nm for the photodetectors based on pristine SnS 2 and SnS 2 /TCLP. These graphs show maximum values of responsivity of 140 mAW −1 , 41% as external quantum efficiency, and detectivity of 1.33 × 10 11 Jones for pristine SnS 2 -based device, while these parameters enhanced to 1.84AW −1 , 513% and 1.69 × 10 11 Jones, respectively for SnS 2 /TCLP photodetectors. Moreover, the photodetector's parameters for three SnS 2 , TCLP and SnS 2 /TCLP based photodetectors has been evaluated through different light wavelength illumination at 5 mW/cm 2 (see Fig. S7g, h and i). These improvements are assigned to the enhancement of charge carrier production and better separation due to the well-engineered band alignment of SnS 2 and three cationic perovskite. Stability test. As the stability of the perovskite-based optoelectronic device in humidity has always been a challenge, at the next step, we tested the stability of the SnS 2 /TCLP encapsulated device. Although, the literature shows that conventional conducting polymer or small molecules as the protecting layer 54 , using the stable, ecofriendly and inexpensive material for protecting the perovskite layer through a cost-effective method, is still under debating 55 . In this study, the PMMA layer with low water absorptivity (0.3%) 56 , has been used as a protecting layer on the SnS 2 /TCLP device. The PMMA layer due to its interring cross-linked network can delay the degradation of the TCLP layer from oxygen and moisture of environmental conditions 55 . Since the light beam is illuminated from the backside of the photodetector, the protective layer should not be optically perturbed due to incident light.
The current-voltage graphs of the protected and unprotected photodetectors under the wavelength of 445 nm with 5 mW/cm 2 were measured for different storing time in the 100% humidity and room temperature, as shown in Fig. 7a. The inset graph shows the I-V characteristics of the unprotected photodetector. The trend of current degradation percentage ( I I 0 * 100 ) at a bias voltage of 3 V, illustrates in Fig. 7b, which indicates that for 10, 16 and 22 min, the current degradation of the protected photodetector are 70%, 64.5% and 31.5% while these corresponding values decreased to 4%, 3.5 and 1.5%, respectively. Figure S8 shows a picture of photodetectors with and without PMMA layer, after storing in 100% humidity and room temperature condition for 10 min.
The clear change in the perovskite layer from dark brown to yellowish (Fig. S8) indicates the decomposition of Pb-I in the TCLP layer, as reported for the decomposition of MAPbI 3 57 , which confirms the existence of moisture in the unprotected device. Therefore, the 20% PMMA, with low water absorptivity, is applicable as the protecting layer for photodetector or other optoelectronic devices through a simple solution method.

Conclusion
In summary, the vertically SnS 2 nanosheets have been grown directly through a low-temperature CVD process on FTO/glass with no extra Sn source. The mechanism study showed the growth of the polyhedral structure of the SnS 2 crystals on the FTO that could be related to elastic energy and the mismatching between SnS 2 nanosheets crystals and substrate. As the vertically TMDs have high densities of exposed surface area and edges per unit area, it should be beneficial for optoelectronic devices. In this research, the cost-effective photodetector based on vertically grown SnS 2 nanosheets has been presented using laser scribed FTO substrate and CVD method. To enhance the device performance, in terms of light absorbance and carrier separation, three cationic perovskite layer of Cs 0.05 (FA 0.83 MA 0.17 ) 0.95 Pb(I 0.83 Br 0.17 ) 3 with proper band alignment has been deposited on SnS 2 nanosheets. the photodetector shows an impressive enhancement in parameters including high responsivity of 1.84 AW −1 , EQE = 513% and detectivity = 1.69 × 10 11 Jones under the incident light of 445 nm with 5 mW/cm 2 , with rising and falling time of ~ 20.7 µs and 31.4 µs, respectively. Besides, a protected (with PMMA layer) and unprotected SnS 2 /TCLP photodetectors have been investigated for different storing times at 100% humidity conditions in which the protected device showed much more stability than unprotected device.  Growth of vertically aligned SnS 2 hexagonal sheets. The cleaned and pattered FTO/glass substrates were sulfurized in a quartz tube furnace under argon atmosphere. The sulfur contents were varied to study the SnS 2 nanosheets growth. In this case, elemental S powder was taken into a ceramic crucible for several amounts of 200, 300, 400 and 500 mg, in each separated experiments, while the substrate was kept 20 cm away from downstream. Before sulfurization process, the trace of oxygen in the quartz tube was removed by purging the pure argon gas for 30 min; then the sample sulfurized at 500 °C for 60 min, eventually. The sulfur vapor specie (S) was incorporated by the tin source in the FTO substrate and then naturally was cooled to room temperature.
Perovskite precursor solution. The organic compound, methylammonium bromide (MABr) and Foramidinium iodide (FAI), Cesium iodide (CsI) were purchased from Sigma Aldrich, and the other solvents were acquired by the Merck Company.
The perovskite precursor solution were prepared by (1. Device fabrication. To fabricate the planar photodetector, engraving process of the FTO crystalline layer performed on FTO/glass substrate. In this regard, a ~ 25 µm gap on the FTO substrate created by physical etching using laser Nd:YAG laser, QCW fiber, wavelength of 1064 nm and 10 Watt laser cutter (Connect laser technology Co. LTD). The laser-engraved FTO/glass substrate was used in CVD process along with 500 mg of sulfur element 2 (Fig. S9a-c)as described in the "Experimental method" section. After CVD procedure, the SnS 2 nanosheets growth on the substrate and filled the laser groove (Fig. S9d) To fabricate the SnS 2 /TCLP PD, the TCLP deposition was done through two-step spin-coating process at 1000 rpm and 4000 rpm for 10 s and 30 s, respectively, while the chlorobenzene anti-solvent was mildly casted at the final 5 s of the second step (Fig. S9e). The film was then baked at 100ºC for 1 h (Fig. S9f.). To make electrical contact, in all PDs (pure SnS 2 , pure TCLP and SnS 2 / TCLP,), the grown SnS 2 nanosheets on FTO, were mechanically removed from both sides of the laser patterned gap. And then, to provide electrical contacts, the Ag paste was placed onto both side of the laser etched areas. Also, a mask with the effective area about 1.35E −6 cm 2 has been regarded and finally, the incident light illuminated from backside of the fabricated PD as shown in Fig. S9g. Stability test. The prepared 20% wt PMMA in chlorobenzene as the protective layer was spin-coated on the perovskite/SnS 2 device at a rotation speed of 6000 rpm for 90 s and then dried for 30 min at room temperature. The non-protected and protected photodetector were stored under 100% humidity at room temperature for 10 to 22 min. Next, the I-V measurement of both photodetectors was performed for the same storing time under bias voltage range of − 3 V to 3 V under illumination (wavelength of 445 nm at 5mW/cm 2 by LED source).
Characterization. The surface morphologies of the isolated SnS 2 sheet were studied by atomic force microscopy (AFM) (Park Scientific CP-Research, VEECO) in tapping mode. Field emission scanning electron microscope (FESEM, TESCAN, MIRA3), energy dispersive X-ray analysis (Oxford instrument) and high-resolution transmission electron microscopy (HRTEM, JEOL, Model JEM-2010F UHR) were used to study morphology, elemental analysis and lattice distance of the SnS 2 and SnS 2 /perovskite heterojunction, respectively. X-ray diffraction (X' Pert Pro, PANalytical) was performed to study of crystallization and structural characteristics of samples. X-ray photoelectron spectroscopy (XPS) was utilized to investigate chemical states of the flakes while the data acquisition was made by using a hemispherical analyzer with an Al Ka X-ray source (hν = 1486.6 eV) operating at a vacuum better than 10 -7 Pa. Optical properties were explored by diffuse reflectance/transmittance spectroscopy (Avantas-Avaspec-2048-TEC). Raman spectrum analysis was performed using a 532 nm laser (Explora, Horiba). The mott-Schottky analysis was carried out by a three-electrode setup with Ag/AgCl in 3 M saturated KCl and Pt wire as the reference and counter electrode, respectively and the electrolyte was an aqueous 1 M Na 2 SO 4 . The potential was swept in the range of stability diagram with an AC signal of 10 mV of amplitude superposed on the DC component using a Potentiostatic-galvanostatic system (Auto-lab system, PGSTAT30).
The opto-electrical/electrical measurements of the devices were carried out using KEITHLEY 6487 picoammeter voltage source instrument. Light sources were arranged as an array of several LEDs from an extended range of UV to Vis wavelengths (395-740 nm). To measure the rising and falling time of the photodetector, an operational amplifier circuit and an oscilloscope have been utilized. A pulsed light has been used controlled by GWINSTEK GDS-1052-U oscilloscope along with a current to voltage converter circuit (Fig. S10). The amount of the current in the dark and light has been calculated via the amount the out-put voltage and the resistance in the circuit.

Data availability
Derived data supporting the findings of this study are available from the corresponding author on request.