The demand for implementing photodetection in the infrared range (>760 nm) continues to increase for numerous applications such as optical fiber communications, environmental surveillance and metrology applying in agriculture, pharmaceuticals and biology^[1−5]. Up to now, several materials have been successfully applied in fabricating infrared photodetectors, i.e., germanium, mercury cadmium telluride, and indium gallium arsenide^[6−8]. Nevertheless, the abovementioned materials usually require costly epitaxial techniques and complicated processes to be integrated into the target infrared photodetectors, which inevitably result in high cost and limit their large-scale residential application^[9]. Compared to the traditional materials, the recently emerging alternatives such as organic bulk-heterojunction, organic−inorganic hybrid perovskite and infrared-response quantum dot (QD) gradually attract widespread attention in infrared photodetectors owing to their low-cost solution processing engineering, high absorption coefficient, and tunable optical bandgap^[10−15].

Colloidal QD is an attracting nanobuilding block that shows great potential in application of light emmitting diodes, solar cells, and imaging^[16−18]. Among the foregoing materials, PbS QD displays great potential for large-scale commercial application in infrared photodetectors for their facile synthesis, sufficient stability and outstanding photoresponse range (100−3100 nm)^[19−21]. With the great efforts on surface ligand exchange, QD size control, and surface defect passivation, the photodetection of PbS QD infrared photodetectors is largely improved^[22−25]. However, due to the point-to-point hopping process of charge carriers in PbS QD system, there is severe non-radiative recombination of charge carriers in PbS QD-only photodetectors, which highly impedes the further improvement of their photodetection performance^{[26, 27]}. Chen and coworkers incorporated two-dimensional graphene into PbS QD system to construct bulk-heterojunction photodetector. Due to the quasi-continuous charge transport pathways as well as the improved carrier mobility, the target photodetectors obtained superior responsivity (R) of 145 mA/W at 808 nm than that of control devices (73 mA/W)^[28]. Besides, Gong et al. introduced thieno(3,4−b)diathiazole-based conjugated polymers into PbS QD system to establish bulk-heterojunction photodetector. Benefiting from the continuous charge transport channels and the photoinduced charge carrier transfer behavior, the target photodetectors deliver R of 73 mA/W at 808 nm under bias of −0.1 V. Moreover, the R of the abovementioned photodetectors could be further improved to 545 mA/W at 808 nm under bias of −0.5 V. However, the reported graphene and conjugated polymers exhibited complex preparation procedures and purification steps with large batch-to-batch variation, which could increase production costs and affect the reliability of applied photodetectors^[29]. Besides, the performance of those reported PbS QD photodetectors is far from satisfaction versus that of the traditional counterparts. Therefore, seeking for another material with commercial yet cost-effective availability to construct efficient PbS QD photodetectors is highly imperative for their large-scale application.

Based on the abovementioned consideration, we select the commercially available yet inexpensive CuSCN to assist PbS QD in establishing bulk-heterojunction self-powered photodetector, a photovoltaic type detector that works without external bias voltage^{[30, 31]}. It is noteworthy that built-in potential would form between CuSCN being p-type semiconductor and PbS QD as n-type semiconductor, promoting the instant splitting of photo-generated excitons and reducing non-radiative recombination losses of charge carrier. Meanwhile, the quasi-three dimensional charge carrier transport channels provided by CuSCN lead to the improvement of charge transport for target photodetector. Benefiting from the ameliorative charge carrier dynamics, the PbS QD/CuSCN bulk-heterojunction photodetector delivers higher specific detectivity (D*) of 4.38 × 10¹² Jones and R of 782 mA/W at 808 nm than that of control devices (4.66 × 10¹¹ Jones and 338 mA/W). Moreover, higher linear dynamic range (LDR) of 75.87 dB and signal-to-noise ratio (SNR) of 3.10 × 10³ are found for the PbS QD/CuSCN bulk-heterojunction photodetector than that of control devices (53.98 dB and 1.02 × 10³), which verifies the effectiveness of the former.

The preparation process schematic of PbI₂−PbS QD-only film (control film) and PbI₂−PbS QD/CuSCN bulk-heterojunction film (target film) is shown in Fig. 1(a). Generally, the primary OA−PbS QD capped with long oleic acid (OA) ligand is readily dissolved in octane (non-polar solvent, the upper part in Fig. 1(a_Ⅰ))^[32]. When successfully undergoing ligand exchange from OA to PbI₂, the obtained PbI₂−PbS QD would facilely dissolve into DMF (polar solvent, the lower part in Fig. 1(a_Ⅱ)). Meanwhile, the PbI₂−PbS QD features less surface defects and stronger electronic coupling than OA−PbS QD, which is beneficial for charge transport to some extent (Figs. 1(b) and 1(c)). However, the PbI₂−PbS QD photodetector is still composed of isolated QDs, which would result in severe charge carrier recombination and insufficient charge collection (Fig. 1(c))^{[33, 34]}. As exhibited in Fig. 1(d), when incorporating CuSCN into PbI₂−PbS QD system to construct bulk-heterojunction photodetector, the quasi-three dimensional carrier transport channels provided by CuSCN can significantly improve the charge transport. Moreover, the built-in potential formed between CuSCN as p-type material and PbI₂−PbS QD being n-type material can promote the prompt splitting of photo-generated excitons and decrease non-radiative recombination losses of charge carriers of the derived system^{[35, 36]}. Therefore, better photodetection performance of PbI₂−PbS QD/CuSCN bulk-heterojunction photodetectors would be anticipated for its overall ameliorated charge dynamics.

As displayed in Fig. 2(a), the PbI₂−PbS QDs are observed to be embedded within the CuSCN matrix for the obvious contrast, which is confirmed by the different type of crystal lattice distance in the corresponding HRTEM image (Fig. 2(b)). As shown, the crystal lattice distances of 0.297 and 0.210 nm are attributed to the (200) crystal plane of PbI₂−PbS QD and the (321) crystal plane of CuSCN, respectively^{[37, 38]}. Furthermore, the Cu element from CuSCN and the Pb element from PbI₂−PbS QDs are both uniformly distributed over the whole sample region (Figs. 2(c) and 2(d)), indicating that CuSCN could assist the construction of quasi-three dimensional carrier transport channels for PbI₂−PbS QDs, which is highly conducive to the charge transport and the derived collection in the corresponding photodetectors^[39]. As shown in Fig. 2(e), the maximum absorption peak of the target film in the long-wavelength presents a slight bathochromic-shift compared to that of the control film. The abovementioned phenomenon originates from trace amounts of Cu atoms incorporate into the crystal lattice of PbI₂−PbS QDs, which leads to higher density of states near the valence band edge of PbI₂−PbS QDs and smaller Moss−Burstein effect, eventually resulting in a redshift of absorption spectrum via band tailing^[40]. The target film featuring quasi-three dimensional carrier transport channels which display higher hole mobility of 3.62 × 10⁻³ cm²∙V⁻¹∙s⁻¹ than that of the control film (1.75 × 10⁻³ cm²∙V⁻¹∙s⁻¹, Figs. 2(f) and Fig. S1). The improved charge mobility and enhanced infrared light response range characteristics would be beneficial for constructing efficient infrared photodetectors.

As depicted in Fig. 3(a), the valence band maximum (V_max) of PbI₂−PbS QD film could be obtained via the equation, V_max = 21.22 eV − (E_cutoff − E_onset), where 21.22 eV, E_cutoff, and E_onset denote the width value of He I in UPS spectra, secondary electron cut-off, and Fermi onset, respectively^[41]. Therefore, a deep V_max of −5.34 eV could be obtained for PbI₂−PbS QD film. Then the conduction band minimum (C_min) can be calculated based on equation, C_min = V_max + E_opt, in which E_opt represents the optical bandgap of PbI₂−PbS QD film (Fig. S2). And an appropriate C_min of −4.06 eV is obtained for PbI₂−PbS QD film. As exhibited in Fig. 3(b), all compositions in target photodetector could form cascaded energy level alignment, which is not only profitable for the charge transfer and transport, but also conducive to suppressing charge recombination. The favorable charge carrier transfer behavior is corroborated by the PL intensity quenching efficiency (22.6%) of target film compared to the control one (Fig. 3(c)). Moreover, the improved conductivity of target film (2.56 × 10⁻² mS/cm) versus that of control film (1.25 × 10⁻² mS/cm, Fig. 3(d)) benefits for the charge carrier transport, which is anticipated to realize high-performance photodetection.

To investigate the potential of PbI₂−PbS QDs/CuSCN bulk-heterojunction strategy in infrared photodetectors, the self-powered photodetectors based on sandwiched planar ITO/SnO₂/active layer/p−PbS/Carbon structure were fabricated. Meanwhile, the control devices based on PbI₂−PbS QDs-only were also constructed for comparison. Note that the concentration of CuSCN is elaborately screened and the derived optimal concentration is determined to be 0.025 mg/mL for the target photodetector fabrication (Fig. S3). The effective area of the photodetector was determined by the area of carbon electrode that was patterned by plastic tapes.

As shown in Fig. 4(a), the target devices present higher photoresponse than control devices in the wavelength from 400 to 1100 nm. The enhancement in wide response range reflects that the addition of CuSCN helps on charge transport and collection that significantly contributes to the EQE; the wide bandgap of 3.5 eV for the CuSCN takes no effect on the light absorption of the device in this visible−infrared range. The type-Ⅱ energy level alignment between the CuSCN layer and the quantum dot (QD) layer facilitate the efficient transfer of these photogenerated carriers to the QD layer (Fig. 3(b)), enhancing the charge separation and transport processes while suppressing charge carrier recombination, which is also confirmed by the lower dark current of the former (Fig. 4(b)). Furthermore, the current density−voltage curves of target devices and control devices were also measured under AM 1.5 G illumination (Fig. S4). The results show that the photoelectric conversion performance of target devices is significantly improved versus that of control ones, in which the open-circuit voltage (V_oc), short-circuit current density (J_sc) and fill factor are all increased. This result indicates that applying CuSCN to assist to construct PbI₂−PbS QD/CuSCN bulk-heterojunction indeed promote carrier extraction and transport effectively.

Furthermore, when illuminated via 808 nm laser, the target devices showed higher light current than control ones in the broad power range from 10 mW/cm² to 2.3 μW/cm² (Fig. 4(c)). More importantly, the target devices still present appropriate light current in low power intensity of 0.7 μW/cm² (6.50 × 10⁻⁸ A) and 0.1 μW/cm² (1.84 × 10⁻⁸ A), while control devices are almost out of work under the same condition (the inset in Fig. 4(c)). The superior photodetection performance of target devices should be associated well with their improved charge carrier dynamics such as enhanced charge carrier transport, suppressed charge carrier recombination, and increased charge carrier transfer and collection. The R and D* as the important parameters for photodetectors can be described as: R = (I_light − I_dark)/PA and D* = R/(2eJ_dark)^1/2, respectively, where I_light is the photocurrent at each wavelength, I_dark is the dark current, P is the intensity of incident light, A is the area of active region of the device, e is the elementary charge and J_dark is the dark current density^{[42, 43]}. Noted that R and D* of target photodetectors and control ones would both decrease along with the increasing of P_light because of the enhanced charge carrier recombination at higher P_light^[44]. However, the performance of target photodetectors is still superior than that of control ones. The ameliorated charge carrier dynamics of target devices should be resulted from the construction of quasi-three dimensional charge carrier transport pathways based on PbI₂−PbS QD/CuSCN bulk-heterojunction, leading to higher light current (Fig. 4(d)), larger responsivity (Fig. 4(e)), and superior detectivity (Fig. 4(f)) for target photodetectors under different wavelength.

To further compare the photodetection performance of target devices and control devices under self-powered mode (0 V bias voltage), the R and D* for derived devices at different P_light when illuminated by 808 nm laser were performed. As depicted in Figs. 5(a) and 5(b), the target devices present higher R and D* values than control ones across the whole range of P_light. Especially, the target photodetectors deliver remarkable R of 782 mA/W and D* of 4.38 × 10¹² Jones than control ones (338 A/W and 4.66 × 10¹¹ Jones) at P_light of 0.7 μW/cm² (Table 1).

To provide insight into the superior photodetection performance of target devices, the charge carrier dynamics of devices were systematically studied. Firstly, the P_light dependence of J_sc and V_oc of devices were characterized. Generally, the relationship between P_light and J_sc can be described as power law equation, J_sc ∝ P_light^α, in which the α represents the degree of bimolecular charge recombination^{[45, 46]}. As shown in Fig. 5(c), the target devices display larger α of 0.89 than control ones (0.79), demonstrating the bimolecular charge recombination is relatively suppressed in the former^[47]. On the other hand, the ideality factor (n) obtained from the equation of V_oc ∝ nkT/e are 1.63 and 1.91 for target devices and control ones, respectively (Fig. 5(d))^[48]. Obviously, the smaller n of target devices signifies their retarded trap-assisted charge recombination^[49]. Fig. 5(e) displays the Nyquist plots of electrochemical impedence spectrum (EIS) of the two devices an open circuit bias (−0.6 V) under dark conditions (Fig. S4−Fig. S6 and Table S1). The EIS reponse is mainly located at middle and low frequency range (100−1 K Hz) that corresponds to charge transfer recombination in the active layer^[50]. The radius of the arc in the middle frequency represents transfer recombination resistance that is relatively large for the target device compared to the control, which further indicates a depressed recombination processes in the CuSCN device.

Besides, the target devices show faster photocurrent decay trend than control ones. Meanwhile, more efficient charge carrier collection is strongly confirmed by the smaller charge carrier lifetime of 1.63 µs (Fig. 5(f)) for target devices compared to control ones (3.46 µs). Overall, the improved charge carrier dynamics resulted from the quasi-three dimensional charge carrier transport pathways based on PbI₂−PbS QD/CuSCN bulk-heterojunction leads to the superior photodetection performance for target devices.

The signal-to-noise ratio (SNR) and linear dynamic range (LDR) were also calculated for investigating the sensitivity of photodetectors. Generally, SNR represents the proportion between useful signal and background noise that includes thermal noise, scattering noise, generation−recombination (G−R) noise, and 1/f noise. And the LDR refers to the photodetector's ability to accurately detect or measure the minimum intensity of light signals^{[51, 52]}. LDR and SNR can be calculated based on the equations, LDR = 20 × log10(P_max/P_min) and SNR = (I_light − I_dark)/I_dark, where I_light,P_max, and P_min are the photocurrent at 10⁻³ W/cm², the maximum and minimum incident light intensities when the photocurrent deviates from linearity, respectively^[53]. As shown in Fig. 6, the target devices exhibit higher LDR of 75.87 dB and SNR of 3.10 × 10³ than that of control devices (53.98 dB and 1.02 × 10³, respectively), which also verifies the effectiveness of the CuSCN matrix in constructing robust and efficient self-powered photodetects.

In summary, we adopt the p-type semiconducting CuSCN matrix to assist PbI₂−PbS QDs to establish bulk-heterojunction photodetectors. It is found that the quasi-three dimensional carrier transport channels provided by CuSCN matrix can improve the charge transport in the derived photodetector. Meanwhile, the built-in potential would form between CuSCN channel and PbI₂−PbS QDs, promoting the prompt splitting of photo-generated excitons and reducing non-radiative recombination losses of charge carriers. Benefiting from the ameliorative charge carrier dynamics, the PbI₂−PbS QD/CuSCN bulk-heterojunction photodetector delivers higher D* of 4.38 × 10¹² Jones and R of 782 mA/W at 808 nm than that of control devices (4.66 × 10¹¹ Jones and 338 mA/W). Besides, higher LDR of 75.87 dB and SNR of 3.10 × 10³ are found for the PbI₂−PbS QD/CuSCN bulk-heterojunction photodetector than that of control devices (53.98 dB and 1.02 × 10³). The whole result indicates that the proposed strategy in this work should be a simple, efficient, and novel method to construct infrared photodetector featuring desirable specific detectivity and responsivity.