Photodetectors are essential devices that can detect photons and convert them into electrical signals making them widely used in various fields such as Artificial Silicon Retina (Retinal Prosthesis–MEMS), telecommunications, image sensors, biomedicine, military surveillance, and security equipment (Chenais et al., 2021; Yu et al., 2016; García-Hemme et al., 2014; Dong et al., 2014; Sun et al., 2017; Rogalski, 2002; Yan et al., 2017a). Photodetectors are characterized by several stringent requirements for optimum functioning, including spectral response, photosensitivity, low dark current, low noise equivalent power, high quantum efficiency, and fast timing response (rise time and fall time) (Supplementary Figure S1) at their working wavelength range (Gupta and Kumar, 2022; Shelke et al., 2020). The distinct properties of silicon, such as a large absorption coefficient at near-infrared, low cost, and excellent compatibility, have attracted the research community to develop Si-based photodetectors (WX₂/Si, MoX₂/Si where X = S, Se) (Sun et al., 2018; Hasani et al., 2019). Rhenium-based transition metal dichalcogenides (TMDCs) have also generated interest due to their unique in-plane anisotropic optoelectronic properties (Liu et al., 2015; Jariwala et al., 2016a; Hart et al., 2016; Hart et al., 2016; Arora et al., 2017). ReSe₂ has a stable distorted 1T phase with layer-independent indirect band structure in the range of 0.9–1.3 eV from bulk to monolayer (Arora et al., 2017; Jiang et al., 2018; Li et al., 2020). Therefore, the ReSe₂/Si system is an ideal candidate for the fabrication of high-performance broadband photodetectors. However, the responsivity of ReSe₂ (Zhu et al., 2021) and Si (Haynes and Hornbeck, 1953) semiconductors suffer from intrinsic defect states in the forbidden zone at higher illumination intensities. We report a highly sensitive and chemical vapor deposition (CVD)-grown ReSe₂/Si photodetector with trap-assisted photoresponsivity at lower illumination intensities. The high ideality factor (4.8) and power coefficient (α = 0.5) reveal defect and trap states. The heterojunction photodiode has conductive gain at −3 V under NIR radiation (1,060 nm) with good reproducibility. The responsivity, specific detectivity, and EQE are evaluated to be 465 A/W, 1.4 × 10¹³ Jones, and 54368.2%, respectively. The heterojunction photodetector exhibits broadband spectral selectivity with a peak response at 1,060 nm. The generalization of the above result proves that this ReSe₂/Si sensor will have potential application in future ultra-sensitive photodetectors.

The evaluation of photo response to different light intensities is a primary experiment to determine the photodetection performance of the device. Figure 2A shows a schematic of our heterojunction device and 2b shows current versus voltage (I–V) characteristics of the ReSe₂/Si photodetector in the dark at different temperatures. The dark current decreases with decreasing temperature due to a decrease in the generation of thermal electrons. The barrier height (0.39 eV) of the ReSe₂/Si heterojunction is estimated from the thermionic emission theory of charge carriers. A Richardson plot is drawn using the equation: ln I 0 T 2 = ln A A * − q ϕ B k T where A^* is the effective Richardson constant, A is the junction area (0.00005 cm²) of the device, k is Boltzmann’s constant, ϕ_B is the barrier height of the heterojunction, and T is the temperature, as shown in the inset of Figure 2C. This calculation was carried out in the temperature range of 295–473 K. Figure 2C shows the current versus voltage (I-V) curves of the ReSe₂/Si heterojunction for dark and photocurrent at various wavelengths. To verify the enhanced photo-current from ReSe₂/Si heterojunction, we conducted separate experiments to compare the photoresponses between the ReSe₂/Si heterojunction and Si. Supplementary Figure S3 shows the current changes and ON/OFF ratio for each case and it clearly shows. The high photocurrent was generated in the ReSe₂/Si heterojunction, indicating very low photocurrent generated in the Si substrate.

Under reverse bias, the photocurrent increases according to the various wavelengths of incident lasers and the corresponding ON/OFF ratios increase significantly, respectively as shown in Figure 3A. As a result of better diode characteristics, under 1,060 nm light illumination (636.6 mW/cm²), the ON/OFF ratio is 5.8 × 10⁴ at open circuit voltage as shown in the inset of Figure 3B. The forward-bias characteristics of the n-p heterojunction diode under dark and light illumination were fitted with the standard diode equation: I = I 0 exp q V n k T − 1 where I₀, V, q, n, k, and T indicate reverse bias saturation current, voltage, electron charge, ideality factor, Boltzmann constant, and absolute temperature of the heterojunction, respectively. The ideality factor, n provides (inset: Figure 2B) a high value of 4.8 under the dark conditions in the voltage range of 0.25–0.55 V, which is shown with a dotted line in Figure 2B. The high ideality factor indicates the existence of high-density interface defects and trap states in the ReSe₂/Si heterojunction. These defects and trap states play an important role in the conduction of photo-generated charge carriers in the device. The photo response of the device is affected by defect trap states at lower illumination intensities.

The photovoltaic characteristics of the ReSe₂/Si heterojunction were investigated in the intensity range from 127.3 to 636.6 mW/cm² of 1,060 nm laser illumination, as shown in Figures 3A–D. It is worth noting that the heterojunction exhibits photovoltaic properties under laser illumination without external bias. The significant photovoltaic effect was observed at zero voltage as shown in Figure 3A. The zero-voltage photocurrent highly increased from 1.8 × 10⁻⁵ A at 405 nm to 8.1 × 10⁻⁵ A at 808 nm, indicating that the ReSe₂/Si heterojunction device can function as a self-powered photodetector. We investigated the photoresponsivity of this detector as a function of applied reverse bias voltages (V = 0, −0.5, −1, −2, −3 V), as shown in Figure 3B. Under the same incident intensity of 1,060 nm laser illumination, the ReSe₂/Si heterojunction exhibited a larger photocurrent at the higher reverse bias. Since the effective barrier height at the ReSe₂-Si junction increases with the application of a larger reverse bias, charge separation of the heterojunction is increased by heightened in-built potential. The inset: shows the ON/OFF ratio (5.8 × 10⁴) at V_oc. Interestingly, the highest detectivity of the device (4.8 × 10¹³ Jones) is at zero bias due to the lowest dark current when compared to reverse bias voltages, as shown in Figure 3C.

The external quantum efficiency (EQE = hcR/eλ) and responsivity (R = I_ph/pA) of the device at different intensities under 1,060 nm illumination are shown in Figure 3D. Here, A is the active area (∼0.00005 cm²) of the device, I_ph is the photocurrent, p is the optical power, h is Planck’s constant, e is the electron charge, and c is the speed of light. The EQE and responsivity decrease with increasing light intensity due to interfacial defects and volume trap states in the ReSe₂/Si heterojunction. The heterojunction device exhibits a photoresponsivity of R = 144 A/W with a corresponding EQE of 16858% at 0 V under the wavelength of 1,060 nm. These findings reveal that the ReSe₂/Si device is useful for the development of advanced photovoltaic devices, such as solar cell applications.

The photocurrent of the ReSe₂/Si heterojunction device increases with increasing reverse bias voltage, as shown in Figure 4A. This can be explained by the increased electric field of the ReSe₂/Si space charge region with reverse bias, which leads to a decrease in photocarrier transit time and a decrease in recombination of photo-generated charge carriers. Additionally, the photocurrent depends on the intensity of light and follows a power law of I_ph = AP^α, where α = 0.5, 0.7, 0.81, 0.83, and 0.84 at V = 0, −0.5, −1, −2, and −3, respectively. Before electron recombination, it passes through many trap states, and at higher α values (0.84 at −3 V) electrons are separated before recombination by the higher electric field. Therefore, trap-assisted electron excitation is lower compared with lower α values (0.5 at 0 V). Similar behavior of the photocurrent with optical power depending on voltage has been reported for other similar structures (Liu et al., 2016; Yan et al., 2017b; Wei et al., 2017; Lv et al., 2018; Wei et al., 2019).

The responsivity and detectivity values under different illumination intensities were evaluated and plotted in Figure 4D. The maximum values of responsivity and detectivity were 465 A/W and 1.5 × 10¹³ Jones, respectively, under 1,060 nm at −3 V. The responsivity and detectivity both decreased with increasing light intensity, indicating the presence of localized trap states in the forbidden gap of ReSe₂ (Zhu et al., 2021) and Si (Haynes and Hornbeck, 1953). To understand the spreading of the trap states in the forbidden gap of the ReSe₂/Si photodetector, optical power dependent I_ph was measured at different voltages as shown in Figure 4A. The dependence of I_ph on the laser power P (127.3 mW < p < 636.6 mW) followed the power law I_ph ∼ P^α with 0.5 < α < 0.84 at different voltages. Lower α values indicate the presence of localized trap states between the conduction and valence bands, as shown in Figure 4B (Rose, 1955; Rose, 1963). These trap states, which exist above the steady state Fermi level for electrons (SSFL)_n and below the steady state Fermi level for holes (SSFL)_h are called shallow traps. The falling electrons into shallow traps are promptly re-excited into the conduction band by thermal excitations without participating in the recombination process. However, the states that exist between the SSFL_n and SSFL_h are called deep traps, which are responsible for the recombination of electrons or holes. At higher intensities, the number of photo-excited charge carriers increases, and the two SSFL_n and SSFL_h move away from each other as shown in Figure 4B, resulting in an increase in the number of recombination centers (Kao, 2004). The increase in the number of recombination centers reduces the lifetime of the free carriers, thereby decreasing the responsivity and detectivity with increasing intensity. At lower intensities, the photocarriers dominate the number of recombination centers, leading to higher responsivity through the re-excitation of trapped electrons into the conduction band (Shi et al., 2013). Figure 4C shows the positive dependence of the ON/OFF ratio on the illumination intensity for different voltages. The maximum ON/OFF ratio of the heterojunction was about 2303 at −1 V reverse bias. The ON/OFF ratio decreased with increasing negative bias voltage at constant optical power density because the dark current decreased with an increase in negative voltage, while photocurrent increased due to the wide depletion region. It also showed a higher on-off ratio, up to 57920 at 0 V bias (shown in Supplementary Figure S5).

The spectral responsivity of the ReSe₂/Si heterojunction is shown in Figure 5A. Remarkably, the heterojunction exhibits a broad spectral response that covers the entire visible and infrared wavelength range due to its unique band alignment, which facilitates the direct transition of valence electrons of Si to the conduction band of ReSe₂. The peak responsivity of the heterojunction is observed to be 465 A/W at 1,060 nm, as shown in Figure 5A. The detectivity exhibits a similar trend to the responsivity with a maximum value of 1.5 × 10¹³ Jones at −3 V as shown in Figure 5A. The spectral behavior of EQE is also presented in Figure 5B, where the maximum EQE value of 54368% is achieved at 1,060 nm. The broadband response from 400 nm to 1,100 nm wavelength under LED light source is shown in Supplementary Figure S4 and the highest responsivity of 1056 A/W was measured at 1,000 nm.

The band alignment of Si and ReSe₂ before contact is shown in Figure 6A. The electron affinities of ReSe₂ and Si are 3.9 and 4.05 eV, respectively. ReSe₂ exhibits an n-type behavior with a layer-independent indirect band gap (1.31 eV) (Jariwala et al., 2016b; Hart et al., 2017) and Si also has an indirect band gap of 1.12 eV. The Fermi level (E_f) of ReSe₂ is located close to the conduction band (E_c), while the Fermi level of p-type Si is close to the valence band (E_v). Upon contact, a p-n heterojunction is formed, and the diffusion of electrons and holes leads to the local band bending at the interface between ReSe₂ and Si. Furthermore, the equilibrium of carrier diffusion at the heterojunction interface causes the E_f of both semiconductors to reach the same level. The band alignment of the heterojunction after contact is presented in Figure 6B, where electrons transfer from ReSe₂ to Si, and holes in Si transfer in the opposite direction due to the difference between Fermi levels (ϕ_B = 0.39 eV). Figure 6C shows the band offset of the heterojunction under reverse bias with light illumination. Photoexcited electrons are generated in ReSe₂ and Si under 405–1,060 nm laser light illumination. Under the external reverse bias, the direction of the electric field of the barrier potential in the ReSe₂/Si heterojunction aligns with the applied electric field. Additionally, the staggered gaps (Type II) band offset causes photoelectrons in Si to flow into ReSe₂, while the holes flow in the opposite direction. Finally, under laser illumination, the photo-generated carriers are collected by the electrodes.

We have fabricated a heterojunction photodiode by depositing a ReSe₂ monolayer on p-type silicon and applying Ti/Au contacts. We have investigated the photovoltaic and photo response properties of the device illumination with different wavelengths and intensities. Our results show that interface and subgap states strongly affect the device’s photocurrent, with these traps re-exciting trapped electrons into the conduction band at lower intensities. The ReSe₂/Si heterojunction exhibits excellent spectral selectivity, long-term stability, and good reproducibility. We have measured an ON/OFF ratio of 5.8 × 10⁴, a responsivity of 465 A/W, and a detectivity of 4.8 × 10¹³, indicating the photovoltaic behavior of the device. The device exhibits a low dark current of 1.4 × 10⁻⁹ A and a high external quantum efficiency of 54368.2% at −3 V under 1,060 nm laser light, representing a photoconductive gain. The maximum photoresponsivity (R = 465 A/W) is achieved at reverse bias of −3 V under 1,060 nm illumination.