Enhanced Photoresponse and Wavelength Selectivity by SILAR-Coated Quantum Dots on Two-Dimensional WSe2 Crystals.

High-performance photodetectors play crucial roles as an essential tool in many fields of science and technology, such as photonics, imaging, spectroscopy, and data communications. Demands for desired efficiency and low-cost new photodetectors through facile manufacturing methods have become a long-standing challenge. We used a simple successive ionic layer adsorption and reaction (SILAR) method to synthesize CdS, CdSe, and PbS nanoparticles directly grown on WSe2 crystalline flakes. In addition to the excellent wavelength selectivity for (30 nm) CdS, (30 nm) CdSe, and (6 nm) PbS/WSe2 heterostructures, the hybrid devices presented an efficient photodetector with a photoresponsivity of 48.72 A/W, a quantum efficiency of 71%, and a response time of 2.5-3.5 ms. Considering the energy band bending structure and numerical simulation data, the electric field distribution at interfaces and photocarrier generation/recombination rates have been studied. The introduced fabrication strategy is fully compatible with the semiconductor industry process, and it can be used as a novel method for fabricating wavelength-tunable and high-performance photodetectors toward innovative optoelectronic applications.

Introduction

Photodetector devices convert light irradiance to electrical signals, and as a vital optoelectronic elements have a wide range of applications in optical imaging,[1] communication,[2] remote sensing,[3] and environmental monitoring.[4] Common and core materials such as silicon,[5] germanium,[6] and GaAs[7] are the primary materials that are used for the fabrication of commercial photodetectors. Apart from the bulk nature of these materials, operating in a wide wavelength range is one of the drawbacks because the maximum photoresponse cannot be tuned at the desired wavelength for specific applications.[8]

Transition-metal dichalcogenides (TMDs) are emergent materials for photodetection that are broadly studied due to their capability of energy band gap engineering, layered structure, and chemical and physical doping possibilities.[9] Tungsten diselenide (WSe2) is one of the TMDs that is less explored compared to its counterparts, such as WS2 and MoS2. The WSe2 flakes have unique properties such as high carrier mobility (250 cm2/V s),[10] strong optical absorption (2.13 cm–1),[11] large luminescence intensity,[12] and high photoconversion efficiency[13] that make them a potent choice among 2D materials for electronic and optoelectronic applications. WSe2 has a band gap in the range of ∼1.3 eV (as bulk form) to ∼1.8 eV (monolayer) that causes excellent absorption to the visible and near-infrared (NIR) regions, rather than the UV region.[14] Stronger optical absorption can be achieved by fabricating heterostructures using TMDs and active semiconducting materials with designed energy band gap matching in which the responsivity can be enhanced by injecting more photocarriers and less recombination.[15,16]

It is known that most quantum dots as artificial atoms have tunable band gaps, strong optical absorption to a broad range of wavelengths from UV to IR, and high quantum efficiencies that make them interesting materials for heterostructure photodetectors.[17] For instance, CdS, CdSe, and PbS nanoparticles (NPs) are n-type semiconducting materials with band gaps in the range of ∼1.5 to ∼2.5 eV, which depends on the synthesis parameters.[18] They exhibit potential applications in the fields of photodetectors,[19] QLEDs,[20] solar cells,[21] and bioimaging.[22]

Recently, low-dimension (0D/2D) hybrid photodetectors have attracted massive attention due to heterojunction formation between the NPs and the layered materials, which causes van der Waals contact at the interfaces.[23] This contact creates a built-in electric field in the junction that leads to an efficient photoinduced charge transport in the 2D nanoflake.[24] These heterostructures display unique properties such as high light absorption,[25] band tunability,[24] and scalability,[26] giving rise to a high potential for developing high-performance photodetectors and overcoming existing challenges.

Although there are many studies on 0D/2D photodetectors, there is still a necessity to build high-precision, adjustable, and affordable photodetectors using feasible approaches for commercialization.[27] To reduce the device preparation cost, synthesis methods of the materials are an important factor. One of the simple, inexpensive, highly efficient, and fast methods for synthesizing different semiconducting NPs is successive ionic layer adsorption and reaction (SILAR).[28]

In this study, we have presented a low-cost, accurate, and a rather simple method based on a novel NP/WSe2 heterostructure on SiO2/Si substrates by the SILAR method. The heterostructure formation between the WSe2 flakes and the assembly of CdS, CdSe, and PbS NPs resulted in the enhanced absorption of light, responsivity, external quantum efficiency, and wavelength-tunable photodetectors. Moreover, using numerical simulation and scanning tunneling spectroscopy, the proposed mechanisms have been investigated.

Results and Discussion

The X-ray diffraction (XRD) pattern of the grown material showed the crystalline phase of WSe2 (Figure a). The diffraction peaks at 31.92°, 37.82°, 41.54°, 47.34°, and 56.42° are assigned to the (100), (103), (006), (105), and (008) crystal planes, respectively (38-1388, PDF 2 database). The confocal Raman spectrum (Figure b) from the transferred WSe2 flake exhibits a peak centered at 247 cm–1 attributed to the A1g mode of WSe2.[31] The mean thickness of the used WSe2 flakes was measured to be about 40 nm using an atomic force microscope (AFM) (Figure c). Step-like edges of the flake verify the layered structure of the chemical vapor transport (CVT)-grown crystal.

Figure 1

Characterization of the WSe2 flake. (a) XRD pattern of the CVT-grown WSe2 crystal. (b) Raman spectrum of the WSe2 layer. (c) AFM image and the corresponding height profile from the edge of a typical flake.

The schematic in Figure a shows the design of the device based on the WSe2 flake and typical CdS NPs deposited using the SILAR method. The optical image of the WSe2 flake is shown in Figure b, which is connected on both sides with Cr/Au electrodes. Figure c shows the field-emission scanning electron microscopy (FESEM) image of the edge of the WSe2 flake before (the inset image in Figure c) and after CdS NP deposition. More detailed FESEM images indicate the successful synthesis of scattered NPs on both surfaces of the WSe2 flake and the SiO2 substrate (Figures S3 and S4). Using the analysis performed by ImageJ software, the particle size is estimated to be between 40 ± 15 nm.

Figure 2

Characterization of NP/WSe2. (a) Schematic illustration of the cross-sectional view of the CdS/WSe2 NP device. (b) Optical image of the device based on the WSe2 flake (before SILAR synthesis of the CdS NPs). (c) Scanning electron microscopy (SEM) image of the edge of the WSe2 flake (before and after SILAR synthesis of CdS). (d) Raman spectrum of CdS, CdSe, and PbS NPs.

Figure d shows the Raman spectra of the CdS, CdSe, and PbS NPs, measured by a 532 nm laser excitation source at room temperature. For the CdS NPs coated on the WSe2 sample, the spectrum exhibits intense peaks at 294 and 515 cm–1 that can be assigned to the longitudinal optical (LO) phonon mode and its overtone (2LO), respectively.[32] In the case of CdSe NPs coated on the WSe2 sample, the peaks occur at 200 cm–1 (1LO) and 406 cm–1 (2LO).[33] In the PbS NP-coated sample, the spectrum exhibits peaks at 150 and 283 cm–1, corresponding to the 1LO and 2LO phonon modes, respectively.[34]

Although deposition of the WSe2 flake with NPs increases the absorption of light, which is an advantage for photodetectors, it can also increase the electrical resistance due to carrier scattering and prevent photons from interacting with the interface. For this reason, it is necessary to obtain an optimal condition for SILAR synthesis and deposition of the NPs on the WSe2 flake in terms of thickness and morphology. Various parameters such as dip-coating time (T) and the number of cycles (C) varied in five sample types (named as CdS(T-C)) have been investigated to obtain optimal conditions, where T is the dip-coating time in minutes (1.5, 2, 5, 10, and 20) and C presents the number of cycles (1, 3, 5, 10, and 20). For all samples, 0.05 M of the precursor concentration was utilized. At low T and C values, sparse decoration of the NPs was observed on the surface of WSe2, while for SILAR deposition at high T and C conditions (e.g., CdS (10–10)), the substrate was almost entirely covered by NPs (Figure S4).

Figure a shows the photocurrent density results for bare WSe2 and CdS NP-deposited samples using the SILAR method. Under the irradiation of different wavelengths, all samples demonstrate a small photoresponse peak at 730 nm from WSe2 and a significant peak at ∼520 nm corresponding to the CdS NPs (an average size of ∼30 nm and an energy band gap of 2.4 eV).[35] The photoresponse peak for the CdS (1.5–1)/WSe2 sample shifts to 460 nm due to the smaller size of NPs. Since photocurrent for CdS (10–10)/WSe2 and CdS (20–20)/WSe2 samples is in the same order, the optimal condition for deposition of the CdS NPs on the WSe2 flake was selected as 10 min of dip-coating time for 10 cycles. In Figure b, the photocurrent density is plotted as source-drain voltage (Vsd) versus the irradiated wavelengths that produce the maximum photocurrent. The photocurrent density is very low (<3 A/m2) for the bare WSe2 and CdS samples and increases significantly for the CdS/WSe2 heterostructure device (the dark current for the CdS/WSe2 structure is 0.16 A/m2, shown in Figure a as a dotted line). It can be observed that the CdS (20–20)/WSe2 sample presented less photocurrent in comparison with CdS (10–10)/WSe2 because more NPs accumulating on the surface increases the electrical resistance of the sample (due to carrier scattering and covering/shadowing of the WSe2 surface by NPs).

Figure 3

(a) Photocurrent density as a function of light-emitting diode (LED) irradiance at different wavelengths for the various conditions of SILAR synthesis. (b) Photocurrent density as a function of Vsd in the irradiated wavelengths that produce maximum photocurrent (a wavelength of 515 nm for CdS and CdS/WSe2 and 730 nm for WSe2 samples).

According to the optimal conditions of the SILAR process, we apply similar T (10 min) and C (10 cycles) on different precursor solutions with concentrations of 0.05 M for CdSe and 0.02 M for PbS NP deposition on WSe2 flakes. In Figure a, the photocurrent density is plotted at different wavelengths for the three samples, namely, CdS/WSe2, CdSe/WSe2, and PbS/WSe2. The maximum photocurrent values are observed at wavelengths of 520 nm (CdS/WSe2), 730 nm (CdSe/WSe2), and 840 nm (PbS/WSe2). In Figure b, the photocurrent density is plotted as a function of Vsd for different samples at the irradiation wavelengths of the photocurrent peak. The hybrid structures ((CdS, CdSe, or PbS)/WSe2) exhibited more than six times increase in photocurrent at Vsd of 2 V compared to the bare WSe2, CdS, CdSe, and PbS structures, demonstrating the enhanced photoabsorption and synergic effect of the NPs on WSe2 flakes to provide effective interfaces/junctions.

Figure 4

(a) Photocurrent as a function of wavelengths for CdS/WSe2, CdSe/WSe2, and PbS/WSe2 samples. (b) Photocurrent density as a function of Vsd at the irradiated wavelengths that produce the maximum photocurrent (wavelength of 515 nm for CdS/WSe2, 730 nm for CdSe/WSe2, and 840 nm for PbS/WSe2 samples).

To study the energy band structures of the NP/WSe2 interface, scanning tunneling spectroscopy (STS) was utilized as a powerful tool to estimate the band gap of materials. According to the STS results shown in Figure a, all the materials represent an n-type behavior. For WSe2, the band gap is 1.6 eV, and the electron affinity is 3.53 eV. Also, CdS, CdSe, and PbS showed band gaps (electron affinity) of 2.55 (3.9) eV, 1.7 (4.2) eV, and 1.5 (3.12) eV, respectively (Table ).[36−38] Regarding these measured values from the STS analysis, their band structure diagrams are illustrated in Figure b.

Figure 5

(a) STS spectra of the WSe2 flake and CdS, CdSe, and PbS NPs on the Si substrate. The valence band and conduction band of each one are shown with dashed lines. (b) Band diagrams. (c) Energy band diagram at an unbiased mode for CdS/WSe2, (d) CdSe/WSe2, and (e) PbS/WSe2.

Table 1

Comparison of CdS-, CdSe-, and PbS-Measured Band Gaps with Reported Values for Different NP Sizes

structures	reported NP band gap in eV [NP size in nm]	measured band gap [eV], this work		references
		optical	STS
CdS	2.4 [>6], 2.8 [3], 3.7 [2], 4.6 [1.2]	2.4 ± 0.2	2.55 ± 0.2	(35−39)
CdSe	1.8 [>5], 2.3 [3], 2.8 [2], 3.8 [0.8]	1.7 ± 0.2	1.7 ± 0.2	(39)
PbS	0.5 [>13], 0.75 [7], 0.9 [5], 1.4 [3], 2.3 [2]	1.5 ± 0.2	1.5 ± 0.2	(40)

According to the average particle sizes measured by the SEM images (Figures S3 and S4), the diameters were estimated to be ∼40 ± 15 nm for CdS and CdSe and ∼3 ± 1.5 nm for PbS. Since the purpose of this research was to fabricate photodetectors in different wavelength regions, SILAR synthesis for PbS was performed under the condition that the band gap of this material was located near the IR region. These results are in agreement with the previously reported energy band gap data (Table ).

To understand the carrier injection and transport at interfaces under light irradiation, band alignments of the samples are demonstrated in Figure c–e. These diagrams show rather small ΔCB (the difference in the conduction band of the two connected materials shown in Figure c) as shown in Table , resulting in more electron transfer to WSe2 than pristine materials. In the case of CdS/WSe2, the ΔVB (the difference in the valence band of the two connected materials shown in Figure c) is higher than others (1.13 eV), which facilitates photocarrier (holes) transfer at the interface. In addition, band bending formation and the potential barrier (especially in CdSe/WSe2 and PbS/WSe2) at the interfaces hinder photoinduced electrons and holes from being recombined. Hence, the photocurrent in these structures increases significantly compared to the pristine WSe2, CdS, CdSe, and PbS (Figure b).

Table 2

Comparison of ΔCB and ΔVB in CdS/WSe2, CdSe/WSe2, and PbS/WSe2 Heterostructures

structures	ΔCB (eV)	ΔVB (eV)
CdS/WSe2	0.37	1.13
CdSe/WSe2	0.27	0.33
PbS/WSe2	0.47	0.37

To understand the nonlinear behavior of photocurrent and mechanism of charge carrier generation, the CdS/WSe2 heterostructure was numerically simulated by technology computer-aided design (TCAD) software known as Silvaco Atlas. The ray tracing model was used to describe light propagation. The simulation of the optoelectronic device was divided into two specific models that are calculated concurrently at each DC bias point (transient time step):

1. Optical ray trace calculates the optical intensity at each grid point by using the real part of refractive index.

2. The absorption or photogeneration model uses the imaginary refractive index component to calculate a new carrier concentration at each grid point.

Figure a shows a schematic of the simulated structure. The dashed area indicates the WSe2 and CdS interface, which is zoomed out in Figure b and used to represent the local electric field distribution. The relatively high electric field at the interface is due to the transfer of electrons from n+(CdS) to n–(WSe2), as predicted by the STS analysis data (Figure b). It is known that in the presence of applied Vsd, photoinduced charge carriers drift into the electrodes and decrease their recombination rate. Therefore, more charge carriers can participate in the conductance, as observed in Figure b. The simulated total current density (electrons and holes) in the CdS/WSe2 heterostructure is shown in Figure a, which is as high as 7.5 A/cm2.

Figure 6

(a) Schematic of the CdS/WSe2 structure ( and represent two directions). (b) Simulated electric field of the dashed area in panel (a) in log scale and Vsd = +2 V.

Figure 7

(a) Total current density in log scale for the CdS/WSe2 structure and (b) impact generation rate in log scale for the CdS/WSe2 structure at Vsd = +2 V. (c) Recombination rate in log scale for the CdS/WSe2 structure at Vsd = 0 V and (d) Vsd = +2 V.

As shown in Figure a, moving away from the interface (in the direction), the electric field and current density decrease. Since the carrier mobility in WSe2 (145–173 cm2/Vs) is higher than that of CdS (100 cm2/Vs),[41,42] the total current density in the CdS layer is much lower than that in the WSe2 region.

On the other hand, the high current density is due to the considerable impact generation rate, a three-particle production phenomenon, that is, high energy charge carriers in the built-in electric field experience scattering with bonded electrons in the valence band and excite them into the conduction band creating a new electron–hole pair. Due to the high energy of the secondary electron–hole pairs, an avalanche effect may be triggered.

Figure b shows the impact generation rate of the CdS/WSe2 structure at Vsd = +2 V. The impact generation rate increases dramatically upon applying lateral voltage, where carriers gain enough energy for electron–hole production. However, the current density in the interface area is high (Figure a). Therefore, a high scattering rate results in a lower impact generation rate at the interface while it increases in both directions.

Figure c,d show the recombination rate in the CdS/WSe2 heterostructure for Vsd = 0 and Vsd = 2 V, respectively. The uniform rather high recombination rate at Vsd = 0 condition ( direction) indicates less chance for majority photoinduced charge carriers to be transferred to the contacts. By applying voltage in the direction, tending carrier distribution to the left-hand side led to having less chance for recombination as is shown in the right side of Figure d[43] (22.3 cm–3 s–1 at Vsd = 0 V and 28 cm–3 s–1 at Vsd = 2 V).

To investigate the optical sensing properties of the samples, various parameters have been studied: photoresponsivity (R), EQE, NEP, D*, and photoconductivity gain (G) of the samples. Figure shows the measured optoelectronic data under irradiation of LEDs at different wavelengths for (CdS, CdSe, and PbS)/WSe2 devices.

Figure 8

Photodetection parameters of the device. (a) Responsivity as a function of LED irradiance at different wavelengths. (b) External quantum efficiency (EQE) as a function of LED irradiance at different wavelengths. (c) Spectral dependence of the noise equivalent power (NEP) of the device. (d) Specific detectivity (D*) of the device. (e) Response time of the CdS/WSe2 device to a pulsed light (515 nm) source (the turn on and turn off times are the same, at 10 ms.) (f) Rising and falling times of the CdS/WSe2 device to a pulsed light (515 nm) source (the turn on and turn off times are the same, at 10 ms).

The photoresponsivity (R) is defined as the ratio of photocurrent generated to the incident light power and can be calculated as:where I and Id are the photocurrents under light illumination and dark mode, respectively, and Pin indicates the optical power density of the incident light. The photoresponsivity of the device is shown in Figure a, for the wavelengths ranging from 395 to 970 nm. The maximum responsivity of the samples was measured to be 8.78 A/W at 515 nm, 15.4 A/W at 735 nm, and 48.72 A/W at 850 nm for CdS/WSe2, CdSe/WSe2, and PbS/WSe2, respectively.

The EQE parameter that is a ratio of the collected charge carriers (NC) to the incident number of photons (NI) is defined as:where h is the Planck constant, c indicates the speed of light, e is the electron charge, and λ is the wavelength of the incident light. Its direct relation to R keeps the peak positions in the same wavelengths as shown in Figure b. The EQE for PbS/WSe2 is about 71%, which is a significant increase compared to bare WSe2 (Figure S5), and for CdS/WSe2 and CdSe/WSe2, the EQE values are 21 and 26%, respectively.

As an essential parameter for a photodetector, the NEP parameter is evaluated which defines the required optical input power to achieve a signal-to-noise ratio (SNR) of one within a bandwidth of 1 Hz. The NEP is expressed as:where R denotes responsivity and Id is the dark current. The NEP values shown in Figure c for PbS/WSe2 vary from 3.8 × 10–15 to 3.3 × 10–16, while for other samples, the NEP values are in the range of 7.1 × 10–14 to 1.4 × 10–15.

The sensitivity of photodetectors relies on parameters such as bandwidth, the geometry of the structure, and the detector’s active area and can be expressed as D*. This parameter is calculated as:where A is the active area of the photodetector. Figure d shows the detectivity of the structures. Detectivity of CdS/WSe2 and CdSe/WSe2 varies from 7.8 × 1010 to 4.0 × 1012 Jones, and for PbS/WSe2, it is in the range of 5.2 × 1011 to 6.1 × 1012 Jones.

The photoconductivity gain (G) that indicates the ratio of the detected charge carriers per single incident photon is represented as:

In this relation, τtransit is defined as τtransit = L2/μVsd, where L is the length of the channel and μ is the carrier mobility. The τlife is approximated by the falling time of the transient Iph during the on/off cycles of illumination. Since, for all structures, the channel length, mobility, and Vsd are approximately similar together, therefore, the difference in gain can be assigned to their τlife (which is measured by the falling time). The measured photoconductivity gains of CdS/WSe2, CdSe/WSe2, and PbS/WSe2 are 1.27 × 105, 6.25 × 104, and 1 × 105, respectively.

Figure e shows the typical response of CdS/WSe2 to the pulsed light (515 and 10 ms duration). The calculated rising and falling times are 2.5 ms and 3.5 ms, respectively (Figure f). The rising (falling) time is defined as the time interval from 10% (90%) to 90% (10%) of the maximum (minimum) value, which is presented by green (orange) color. The long-term stability diagram, which shows the stability of the photodetector over time, is also shown in Figure S6. The response time diagram for CdSe/WSe2 and PbS/WSe2 photodetectors is shown in Figure S7.

Table shows the comparison of the photodetector parameters (selectivity, responsivity, and response time) and fabrication methods presented in this research with other similar reported devices.

Table 3

Comparison of Photodetector Parameters and Fabrication Methods of NP/WSe2 with Similar Reported Devices

structures	wavelength (nm)	responsivity (A/W)	response time (s)	fabrication method	ref
WSe2 (25 layers)	405–980	6.5 × 10–3	6.5 ms	exfoliation	(44)
MoS2/rGO NPs (monolayer)a	visible	2.1	18 ms	CVD/transferring	(45)
MoS2/MoTe2 (eight layers)	visible	0.62	10 μs	exfoliation/transferring	(46)
CdS nanowire	visible	360	10 ms	CVD	(47)
CdS/CdSO4	400–600	0.3	14.4 ms	chemical methods	(48)
Sm:CdS	visible	0.2	120 ms	spraying	(49)
CdSSe nanobelt (60 layers)	550–650	10.4	4.7 ms	CVD	(50)
Sn2+:PbS	660–980	15	∼ 30 μs	chemical bath deposition	(51)
CdS/WSe2 (50 layers)	450–650	8.8	2.5 ms	exfoliation/SILAR	this work
CdSe/WSe2 (50 layers)	600–850	15.4	2.5 ms	exfoliation/SILAR	this work
PbS/WSe2 (50 layers)	600–950	48.7	2.5 ms	exfoliation/SILAR	this work

rGO: reduced graphene oxide.

Wavelength selectivity is one of the most important challenging and vital parameters in photodetectors, and our sample shows better selectivity compared to other reported devices. Photodetectors based on TMDs and their heterostructure[44−46] have lower responsivity in the visible region compared to the samples of the current study. Although those photodetectors based on CdS obtained through a more complicated fabrication procedure presented higher photoresponsivity, the response time is still higher than our introduced SILAR-coated devices. Hence, the simple SILAR method, a rather high responsivity, and low response time in our samples make them as promising candidates for the photodetector industry.

Conclusions

In summary, we fabricated photodetectors based on NP (CdS, CdSe, and PbS)/WSe2 heterostructures to improve the selectivity, responsivity, and response time compared with those made of WSe2 and NPs. NPs create a built-in field resulting in a higher impact generation rate, good photoresponsivity (48.72 A/W), high EQE (71%), and an appropriate response time (2.5–3.5 ms). In addition, high selectivity (for wavelengths of 520, 735, and 850 nm) was achieved. The introduced structure is an innovative approach to build high-selectivity photodetectors that can open new windows for low-cost, mass-produced photodetectors.

Methods

The crystal growth of WSe2 was performed using the CVT method as follows: tungsten and selenium precursors were placed in a vacuum seal ampoule with a stoichiometric ratio of 1:2. The ampoule with a dimension of 20 cm length was placed in a furnace at 850 °C for 2 weeks (Figure S1). The CVT-grown WSe2 crystal was mechanically exfoliated into a few layers of flakes and was transferred onto SiO2/Si substrates using scotch tape. Then, Cr/Au electrodes were deposited on both sides of different selected flakes by physical vapor deposition and using a shadow mask. To deposit NPs, the SILAR method was carried out as given in the following text

Deposition of CdS

Cd2+ ions are adsorbed on the surface from a methanolic 0.05 M solution of Cd (NO3)2 and a 0.05 M solution of Na2S in methanol/water (50/50 v/v) utilized as a S2– source. A single SILAR cycle for CdS deposition was used as successive cycles of 1 min dip-coating each for Cd2+ precursor and then S2– solution. After each precursor bath, the sample was rinsed by the corresponding solvent to remove the chemical residuals from the surface and then dried with an N2 gun.[29]

Deposition of CdSe

Cd (NO3)2 (0.05 M) in ethanol solution was used as the Cd2+ precursor, and a 0.03 M selenide solution was prepared by dissolving 0.33 g of SeO2 and 0.45 g of NaBH4 in 60 mL of ethanol under a N2 atmosphere and used as the Se2– precursor.

Deposition of PbS

Pb(NO3)2 (0.02 M) in ethanol/water (50/50 v/v) was used as the Pb2+ precursor, and a 0.02 M solution of Na2S in ethanol/water (50/50 v/v) was employed as the S2– precursor source.[30]

To explore the crystal structure of the CVT-grown WSe2 flake, XRD patterns were measured using a PANalytical diffractometer (Cu Kα, λ = 0.15418 nm), operating at 40 kV and equipped with a copper Kα radiation source. The thickness of the WSe2 flake and the morphology of NP/WSe2 were investigated using an AFM (Park Scientific CP-Research, Veeco) and an FESEM (MIRA3, T-Scan), respectively. Raman scattering measurements (XploRA, confocal Raman microscope, Horiba) were carried out at room temperature using a 532 nm laser as the excitation source. Electrical characterization of the devices was recorded using a Keithley 6487 picoammeter voltage source instrument. In the UV-to-NIR region, 13 LEDs were employed to illuminate the samples by wavelengths of 395, 415, 435, 445, 470, 510, 520, 595, 625, 650, 735, 850, and 970 nm. To achieve similar power density for different wavelengths illuminated on the sample, we adjusted the output power of LED sources using an optical power meter. To evaluate the time response of the device, a pulsed light was used and the photodetector response was controlled using a GW Instek GDS-1052-U oscilloscope along with the current-to-voltage converter circuit.