Since 2004, significant attention has been directed towards two-dimensional (2D) materials such as graphene (GP), black phosphorus (BP), and molybdenum disulfide (MoS₂), due to their extraordinary atomic-scale thickness, remarkable mechanical strength, and robust light−matter interaction^[1−3]. As a nascent member of 2D materials, tin diselenide (SnSe₂) is an instinct n-type Ⅳ−Ⅵ semiconductor with a Se−Sn−Se sandwich-layered structure and a bandgap of ~1 eV^[4]. Furthermore, with its exceptional qualities of high carrier mobility (462 cm²/(V∙ s))^[5], absorption coefficient (7.25 × 10⁵ cm⁻¹)^[6], environmental friendliness^[7], and chemical stability^[8], SnSe₂ is a superior material for high-sensitivity and low-consumption photodetectors, which can be widely used in image acquisition^[9], information communication^[10], and photo-electronic memory^[11]. For example, Chen et al. fabricated a SnSe₂-based photodetector via low-temperature molecular beam epitaxy, which demonstrated a normalized detectivity of 6.09 × 10⁹ cm∙Hz^1/2∙W⁻¹ under 532 nm illumination^[12]. Zhai et al. fabricated a SnSe₂-based photodetector using chemical vapor deposition, which exhibited a normalized detectivity of 1.01 × 10¹⁰ cm∙Hz^1/2∙W⁻¹ and a response time of 14.5 ms under 530 nm illumination^[13].

However, SnSe₂-based photodetectors suffer from drawbacks such as large dark current, slow response time, and low normalized detectivity, which limits their potential for commercial application. To surmount these challenges and enhance the photodetection performance of SnSe₂-based photodetectors, stacking van der Waals heterostructures has been established as an effective strategy. This is due to the unique properties of van der Waals heterostructures, including the absence of dangling surfaces, strong interlayer coupling, and ultrafast charge transfer^{[14, 15]}. For example, Zhou et al. successfully achieved a photodetector based on a SnSe₂/MoS₂ heterostructure. The photodetector exhibited a normalized detectivity of 9.3 × 10¹⁰ cm∙Hz^1/2∙W⁻¹ under 500 nm illumination, which was one order of magnitude higher than that of the photodetector based on the individual SnSe₂ and MoS₂ components^[16]. Zheng et al. reported a photodetector based on an In₂Se₃/SnSe₂ heterostructure that outperformed the device based on the individual components by one order of magnitude, achieving a high normalized detectivity of 8.8 × 10¹¹ cm∙Hz^1/2∙W⁻¹ under 477 nm illumination^[17]. However, more efforts should be devoted to elucidating the underlying mechanisms of SnSe₂-based heterostructures.

As members of the Ⅳ−Ⅵ compounds, tin mono-sulfide (SnS) and tin mono-selenide (SnSe) exhibit inherent p−type semiconductivity with an orthorhombic Sn−S (Se) layered structure^{[18, 19]}. Given their common metallic atoms, it is both feasible and convenient to realize p−n heterostructures of SnS/SnSe₂ and SnSe/SnSe₂. In this work, photodetectors based on SnS/SnSe₂ and SnSe/SnSe₂ p−n heterostructures (denoted as the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors) were constructed by a polydimethylsiloxane (PDMS)-assisted dry-transfer method, and then investigated. Our results demonstrate that these photodetectors manifested superior photodetection performance, encompassing enhanced responsivity, normalized detectivity, and response speed in comparison to photodetectors based on individual SnSe₂, SnS, and SnSe components. This work expands the research scope of SnSe₂-based photodetectors and holds great promise for future optoelectronic applications.

Single crystals of SnSe₂, SnS, and SnSe were prepared by a chemical vapor transport (CVT) technique. The purity of precursors (Se, S, and Sn powders) was 99.99%, and they were purchased from Aladdin (Shanghai, CN). For SnSe₂, Sn and Se powders were hermetically vacuum-sealed in a quartz tube under a vacuum of 10⁻⁵ Torr with a stoichiometric ratio of 1∶2. The precursors were heated by gradient heating: initially, the temperature was heated to 610 ℃, and held for 24 h. In the next step, the temperature was decreased to 600 ℃ and held for 120 h. Finally, they were cooled to room temperature. As to SnS, Sn and S powders were also vacuum-sealed in a quartz tube under a vacuum of 10⁻⁵ Torr with a stoichiometric ratio of 1∶1. The precursors were subjected to gradient heating: initially, the temperature was raised to 950 ℃ and held for 24 h. In the subsequent step, the temperature was lowered to 850 ℃ and held for 120 h. Finally, they were cooled to room temperature. For SnSe, Sn and Se powders were hermetically sealed with a stoichiometric ratio of 1∶1. The precursors were subjected to gradient heating: initially, the temperature was raised to 900 ℃ and maintained for 24 h. In the subsequent step, the temperature was cooled to 800 ℃ and held for 120 h. Finally, the precursors were cooled down to room temperature.

The composition and phase of the crystals were probed using an X−ray diffractometer (XRD, Smartlab, Rigaku) with Cu Kα radiation of λ = 0.15406 nm. The elemental composition and valence of the samples were assessed using an X−ray photoelectron spectroscope with an excitation source of 532 nm (XPS, Escalab 250 Xi, Thermo Fisher). The Raman spectrometer (InVia, Renishaw) was utilized to characterize the vibrational mode and material identification under the excitation source of 532 nm. The surface morphology was evaluated using an atomic force microscope (AFM, Ntegra Spectra, NT−MDT) in tapping mode and an optical microscope (OM, Olympus BX51M).

Photodetectors based on SnS/SnSe₂ and SnSe/SnSe₂ heterostructures were stacked by a polydimethylsiloxane (PDMS)-assisted dry-transfer method, and Au electrodes were fabricated on substrates via means of lithography and electron beam evaporation. In the case of SnS/SnSe₂, SnSe₂ and SnS thin films were exfoliated from their single crystals and transferred onto the substrate using PDMS films. Subsequently, SnSe₂ thin films were stacked onto the SnS thin films with the assistance of an optical microscope. For SnSe/SnSe₂, SnSe₂ and SnSe thin films were exfoliated from their single crystals, and SnSe thin films were transferred onto the substrate using PDMS films. Subsequently, SnSe₂ thin films were stacked onto the SnSe thin films with the assistance of an optical microscope. The heterostructures were then transferred onto the substrate.

The photodetection performance of the photodetectors was assessed by a self-assembled photoelectric measurement system, which comprised a probe station (PW−600, Advanced), a semiconductor device analyzer (B1500A, Keysight), a waveform generator (33500B, Keysight), and four semiconductor lasers (with a wavelength of 405, 532, 650, and 850 nm and a spot diameter of 200 μm). The power density of the lasers was ascertained by a handheld laser power and energy meter (Nova Ⅱ, Ophir).

Crystals of SnSe₂, SnS, and SnSe were synthesized via CVT, and their phase structures were probed by XRD. Fig. 1(a) shows XRD patterns of the as-grown crystals. The diffraction peaks are consistent with the standard cards of hexagonal-SnSe₂ (JCPDS No.23-0602), orthorhombic-SnS (JCPDS No.39-0354), and orthorhombic-SnSe (JCPDS No.48-1224)^[20−22]. No phases related to other impurities like SnO and SnO₂ were detected. The strongest peaks corresponding to the (001), (003), (004), and (005) planes of SnSe₂ indicate the c-axis preferred orientation. The (111) plane preferred orientation of SnS and SnSe, on the other hand, is indicated by the strongest peaks at 31.97^° and 30.89^°, respectively. The inset of Fig. 1(a) shows the corresponding optical microscopy images and all the crystals showed gray-black metallic lusters with a lateral size of >4 mm.

In order to verify their chemical compositions and chemical states, XPS measurements were conducted. Fig. 1(b) illustrates the XPS spectra of the SnSe₂, SnS, and SnSe crystals. For SnSe₂, four peaks located at 52.5, 53.5, 485.3, and 494.6 eV are ascribed to the Se3d_5/2, Se3d_3/2, Sn3d_5/2, and Sn3d_3/2 orbitals of SnSe₂, indicating the presence of the Se²⁻ and Sn⁴⁺ states^[23]. The atomic ratio of Sn to Se was determined to be 1 : 2.03 (SnSe_2.03). For SnS and SnSe, two peaks located at 493.3 and 484.91 eV are attributed to the Sn3d_5/2 and Sn3d_3/2 orbitals of the Sn²⁺ state. Four peaks located at 161.3, 160.4, 54.7, and 53.1 eV correspond to the S2p_1/2, S2p_3/2, Se3d_5/2, and Se3d_3/2 orbitals of SnS and SnSe, respectively^{[24, 25]}. The atomic ratios of Sn to S and Sn to Se were determined to be 1 : 1.08 (SnS_1.08) and 1 : 1.02 (SnSe_1.02), respectively.

Following the fabrication of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors through a polydimethylsiloxane (PDMS)−assisted dry−transfer technique, the two types of heterostructures were characterized in−situ (Fig. S1 (a)). The surface morphologies of the heterostructures were detected by AFM. As illustrated in Fig. 2(a), the SnS/SnSe₂ and SnSe/SnSe₂ heterostructures possess clean surfaces devoid of any impurities, undulations, or bubbles. The height profiles of the SnS/SnSe₂ heterostructure reveal that the individual SnS and SnSe₂ thin layers are 8 nm (~15 layers) and 9 nm (~12 layers) in thickness, respectively^{[4, 18]}. Meanwhile, the thicknesses of the SnSe and SnSe₂ in the SnSe/SnSe₂ heterostructure were determined to be 9 nm (~15 layers) and 8 nm (~12 layers), respectively^[19].

Fig. 2(b) presents the Raman spectra of the SnS/SnSe₂ and SnSe/SnSe₂ heterostructures. And the Raman spectrum of each heterostructure can be regarded as the combination of the individual Raman spectra of SnS, SnSe, and SnSe₂, thereby confirming the successful formation of the heterostructures. In the Raman spectrum of the SnS/SnSe₂ heterostructure, five peaks were observed (position A1). Two of these peaks located at 109.7 and 184.6 cm⁻¹ were attributed to the E_g and A_1g vibration modes of SnSe₂ (position C1), while the remaining three peaks at 161.2, 188.4, and 216.1 cm⁻¹ were attributed to the B_3g and A_g vibration modes of SnS (position B1)^{[26, 27]}. Similarly, peaks corresponding to the E_g and A_1g vibration modes of SnSe₂ (position C2), as well as the Ag1, B_g, and Ag2 vibration modes of SnSe (position B2) can be found in the Raman spectrum of the SnSe/SnSe₂ heterostructure (position A2)^[19].

The electrical characteristics of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors were probed by employing Au layers (50 nm) as the source and drain electrodes, while a heavily p−doped Si layer was utilized as the back gate. The output characteristics of the photodetectors were acquired by sweeping the source−drain voltage from −5 to 5 V and varying the gate−source voltage V_gs from −50 to 50 V, as illuminated in Figs. 3(a) and 3(b). The p−n junction demonstrated remarkable gate-tunable rectification behaviors, with rectification ratios of 55 and 110 observed at V_gs = 25 V for the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors, respectively^[28]. Figs. 3(c) and 3(d) depict the transfer curves of the photodetectors, which exhibited bipolar behavior attributed to the p-type SnS (SnSe) and n-type SnSe₂. Notably, for the SnS/SnSe₂ photodetector, the charge transport was dominated by holes when V_gs < −42.55 V, whereas electrons governed the charge transport when V_gs > −42.55 V. The mobility can be calculated using the following formula^[13]:

1μ=dIdsdVgs×LW×Cox×Vds,

where L and W represent the length and width of the channel (L: 5 μm and W: 4 μm), C_ox is the capacitance of SiO₂ with a value of 11.5 nF cm⁻². The electron mobility (μ_e) and hole mobility (μ_h) of the SnS/SnSe₂ photodetector were determined to be 45.3 and 41.2 cm²/(V·s), respectively, while the μ_h and μ_e values of the SnSe/SnSe₂ photodetector were determined to be 21.5 and 18.3 cm²/(V·s), respectively. These values were found to be considerably higher than those of the individual SnSe₂ (10 cm²/(V·s)), SnS (5.29 cm²/(V·s)), and SnSe (6.84 cm²/(V·s)) thin layers as shown in Fig. S2.

To evaluate the photodetection performance of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors under 405 nm illumination, the semi-logarithmic current−voltage (I_ds−V_ds) curves were measured at various power densities (17.8, 35.7, 75.3 μW·mm⁻²), as depicted in Figs. 4(a) and 4(b). The rectification behavior remained consistent as the light power density increased, with a noticeable increase in the forward current. The dark current (I_dark) of both heterostructure photodetectors (0.5 μA) was observed to be one order of magnitude lower than that of SnSe₂-based photodetectors (5.1 μA, Fig. S3).

The current−time (I_ds−T) curves of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors were obtained at V_ds = 5 V and V_gs = 0 V with a modulation frequency of 50 mHz, as presented in Figs. 4(c) and 4(d), respectively. An immediate response to the 405 nm light was observed, with the current increasing to saturation as soon as the laser was activated, and swiftly decaying to the initial state when the laser was turned off. After storage for three weeks, the photodetectors still exhibited stable and repeatable switching cycles, which can be attributed to the films possessing inert properties^[29].

The response time, including the rise time (τ_r) and fall time (τ_f), was defined as the time of the photocurrent I_ph (I_ph= I_ds − I_dark) increased from 10% to 90% and decayed from 90% to 10%, respectively^[30]. The rise/fall time of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors was estimated to be 3.13/4.54 ms and 4.74/5.18 ms, respectively, which is one order of magnitude less than that of the photodetector based on the SnSe₂ thin layers (Fig. S3), as shown in Fig. 4(e). The scatter plots of photocurrent−light power intensity are displayed in Fig. 4(f), and can be fitted using the formula^[16]:

2Iph=aPα,

where a is a constant, P is the light power intensity, and α is the fitted parameter, reflecting the efficiency of photocurrent conversion. The α values were estimated to be 0.69 and 0.60 for the SnS/SnSe₂ and SnSe/SnSe₂ photodetector, respectively, which are larger than that of the photodetector based on the SnSe₂ thin layers (0.32, Fig. S4), indicating an increased photocurrent conversion efficiency. The observed α values were all less than 1, a behavior that has been widely reported in photodetectors based on 2D materials, owing to the presence of trap centers within the channel material^[16].

Both the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors exhibit broad-spectrum photodetection characteristics, as illustrated in Figs. 5(a) and 5(b), as well as in Fig. S5 and S6, upon illumination with wavelengths ranging from 405 to 850 nm. While the photocurrents of the two photodetectors vary under different wavelengths of illumination, their exceptional broad-spectrum photodetection abilities are evident across the near-ultraviolet to near-infrared spectrum. The photoresponsivity (R) and external quantum efficiency (EQE), two fundamental parameters for assessing a photodetector, can be derived using the following formulas^{[31, 32]}:

3R=IphPS,

4EQE=hcReλ,

where λ is the incident wavelength, S is the effective area, and the size of the channel is smaller than the laser spot, therefore the effective area is 20 μm² (length × width: 4 × 5 μm²)^[33]. Typically, noise in photodetectors comprises flicker noise, thermal noise, and shot noise^[34]. The noise current (I_N) of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors can be derived from the fast Fourier transform of the dark current (see Fig. S7)^[35]. At frequencies lower than 14.7 Hz (12.5 Hz), flicker noise is the dominant noise source and demonstrates a linear relationship with frequency. At higher frequencies exceeding 14.7 Hz (12.5 Hz), the noise is composed of thermal noise (IT=4KBTRsh) and shot noise (Is=2eIdark). As a result, the normalized detectivity (D*), which is indicative of the photodetector's ability to detect weak signals, is calculated using the following formula^[36]:

5D*=RsIN=RS2eIdark+4kBTRsh,

where k_B is the Boltzmann constant, and R_sh is the resistance. Table S1 and Fig. 5(b) present the computed R and D^* values, which reveal a declining trend as the incident wavelength increases. When exposed to 405 nm light, the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors exhibit maximal R values of 4.99 × 10³ and 5.91 × 10³ A∙W⁻¹, respectively. These values are notably greater than those of photodetectors based on individual SnSe₂ (2.46 × 10³ A∙W⁻¹, Fig. S8), SnS (1.19 × 10³ A∙W⁻¹), and SnSe (1.37 × 10³ A∙W⁻¹). The corresponding EQE values are 1.53 × 10⁴% and 1.81 × 10⁴% (Fig. S9). The D^* values of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors are calculated to be 5.78 × 10¹² and 7.03 × 10¹² cm∙Hz^1/2∙W⁻¹, respectively, which are one order of magnitude greater than those of photodetectors based on individual SnSe₂ (8.79 × 10¹¹ cm∙Hz^1/2∙W⁻¹, Fig. S8), SnS (5.69 × 10¹¹ cm∙Hz^1/2∙W⁻¹), and SnSe (9.82 × 10¹¹ cm∙Hz^1/2∙W⁻¹). It is worth noting that the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors respond well to 850 nm light, demonstrating considerable photoresponsivities of 9.39 × 10² and 1.10 × 10³ A∙W⁻¹, as well as substantial normalized detectivities of 1.02 × 10¹² and 8.38 × 10¹¹ cm∙Hz^1/2∙W⁻¹.

The noise equivalent power (NEP) is an important evaluation parameter in commercial applications, which is defined as the incident light power that produces a signal equivalent to the noise level in a 1 Hz bandwidth. NEP can be calculated using the formula^[37]:

6NEP=SD*.

The NEP values of the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors were estimated to be 8.02 × 10⁻⁵ and 7.29 × 10⁻⁵ pW∙Hz^1/2, respectively. These values are notably lower by five orders of magnitude than those of conventional photodetectors (12 pW∙Hz^1/2, PDB230A, Thorlabs) as depicted in Fig. S9. Hence, considering the photoresponsivity, normalized detectivity, noise equivalent power, and response time, the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors demonstrate superior performance compared to photodetectors based on other heterostructures, such as WSe₂/MoS₂^[38], SnTe/Ge^[39], and others^{[16, 40−47]}, as demonstrated in Fig. 5(c). Furthermore, the photodetectors are comparable to, and in some cases even surpass, commercial photodetectors.

Furthermore, first-principles calculations based on density functional theory (DFT) were performed to provide additional insights into the properties of the heterostructures^[48]. The detailed setting information is shown in the supplementary information. Figs. 6(a) and 6(b) depict planar-averaged and three-dimensional isosurfaces of the charge density differences for the SnS/SnSe₂ and SnSe/SnSe₂ heterostructures. The blue part signifies electron accumulation, while the yellow part represents electron depletion. The redistribution of charge density was observed at the interface of the two heterostructures, where electrons were accumulated in SnSe₂ but depleted in SnS and SnSe, demonstrating for the presence of the built-in electrical field. The DFT-computed band gaps of SnSe₂, SnS, and SnSe were determined to be 0.88, 1.39, and 1.07 eV, respectively, as illustrated in Fig. S10. The band alignments of the SnS/SnSe₂ and SnSe/SnSe₂ heterostructures were displayed in Figs. 6(c) and 6(d). Because the conduction band minimum and the valence band maximum of SnSe₂ were lower than those of SnS and SnSe, both the SnS/SnSe₂ and SnSe/SnSe₂ heterostructures were categorized as type−Ⅱ heterostructures.

The excellent photodetection performance of heterostructures can be attributed to the following reasons. Firstly, the built-in electric field effectively separated the photo-generated carriers, reducing their recombination rate^[49]. As a result, both SnS/SnSe₂ and SnSe/SnSe₂ photodetectors demonstrated increased photocurrent and photoresponsivity. And the higher photoresponsivity of the SnSe/SnSe₂ photodetector than that of the SnS/SnSe₂ photodetector can be attributed to the higher work function difference or potential barrier. Moreover, the barrier region at the heterostructure interface prevented the majority carriers, hence decreasing the dark current^[50]. Because the normalized detectivity is proportional to photoresponsivity and inversely related to the dark current, the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors displayed at least an order of magnitude greater normalized detectivity than their individual components. Finally, the fast response time can be attributed to two reasons: (Ⅰ) the rapid movement of photo-generated carriers toward the electrodes, enabled by the heterostructure's band offset, and (Ⅱ) the presence of clean and atomically sharp interfaces because of the interlayer van der Waals forces, which facilitated the transportation of photo-generated carriers^{[51, 52]}.

In this work, photodetectors based on SnS/SnSe₂ and SnSe/SnSe₂ heterostructures were successfully constructed using a polydimethylsiloxane (PDMS)-assisted dry-transfer method, which exhibited remarkable performance in detecting a broad range of wavelengths from 405 to 850 nm. Compared with their individual components, the SnS/SnSe₂ and SnSe/SnSe₂ photodetectors demonstrated superior characteristics such as high photoresponsivities of 4.99 × 10³ and 5.91 × 10³ A∙W⁻¹, elevated normalized detectivities of 5.80 × 10¹² and 7.03 × 10¹² cm∙Hz^1/2∙W⁻¹, and faster response times of 3.13 and 4.74 ms, respectively. The SnS/SnSe₂ and SnSe/SnSe₂ heterostructures provide new platforms for the development of SnSe₂-based photodetectors, and are considered promising candidates for future high-performance optoelectronic devices.