Highly efficient self-powered perovskite photodiode with an electron-blocking hole-transport NiOx layer

Hybrid organic–inorganic perovskite materials provide noteworthy compact systems that could offer ground-breaking architectures for dynamic operations and advanced engineering in high-performance energy-harvesting optoelectronic devices. Here, we demonstrate a highly effective self-powered perovskite-based photodiode with an electron-blocking hole-transport layer (NiOx). A high value of responsivity (R = 360 mA W−1) with good detectivity (D = 2.1 × 1011 Jones) and external quantum efficiency (EQE = 76.5%) is achieved due to the excellent interface quality and suppression of the dark current at zero bias voltage owing to the NiOx layer, providing outcomes one order of magnitude higher than values currently in the literature. Meanwhile, the value of R is progressively increased to 428 mA W−1 with D = 3.6 × 1011 Jones and EQE = 77% at a bias voltage of − 1.0 V. With a diode model, we also attained a high value of the built-in potential with the NiOx layer, which is a direct signature of the improvement of the charge-selecting characteristics of the NiOx layer. We also observed fast rise and decay times of approximately 0.9 and 1.8 ms, respectively, at zero bias voltage. Hence, these astonishing results based on the perovskite active layer together with the charge-selective NiOx layer provide a platform on which to realise high-performance self-powered photodiode as well as energy-harvesting devices in the field of optoelectronics.

strategies reported thus far will inevitably lead to grain boundary issues and large variations in the morphologies of the resultant perovskite active layers in the devices. Further, the relatively small values of the responsivity R and detectivity D of perovskite photodiodes and their comparatively large dark current densities are the major hurdles to overcome to realise high-performance devices. Moreover, suitable materials capable of forming a good interface with the perovskite active layer remain elusive in the field of perovskite-based photodiode during the effort to achieve high R and D values as well as fast response times. Thus, the device performance capabilities of recent perovskite-based photodiode remain insufficient.
Meanwhile, researchers remain motivated to realise an innovative photodiode exhibiting astonishing performance and a high R value using other functional layers as hole transport layers (HTLs) and electron transport layers (ETLs) together with hybrid heterostructures exhibiting tailored, novel, and improved characteristics [14][15][16] . In order to achieve excellent device performance with regard to photoconductivity, the proper selection of the HTLs is crucial, not only to decrease the dark current density but also to enhance and promote perfect light absorption of the perovskite active layer in the wide visible region [17][18][19] . Metal-oxide (MO) or ternary MO nanoparticles (NPs) synthesised by a solution process are favourable for unlocking their potential in solar cells as HTLs given their low cost, good stability, and promising optical characteristics. However, exotic organic ligands adopted for the purpose of ensuring a small size and a nano-dispersion are associated with poor conductivity, which thus impedes their use in electrical applications 20 . Well-dispersed NiCo 2 O 4 ternary MO NPs synthesised by a unique method without exotic ligands have been used successfully as a HTL in PVSCs. The pinhole-free films of NiCo 2 O 4 NPs facilitated the formation of large grains of perovskite films. As a result, the power conversion efficiency (PCE) was enhanced to 18%, with promising stability 21 . Other ternary MO NPs of In-doped CuCrO 2 have also been proposed as an efficient HTL material system. Interestingly, the PCE of the PVSCs with the In:CuCrO 2 HTL was increased to 20.5% with good repeatability and photostability 22 . Thus, new approaches are mandatory for the realisation of excellent perovskite photodiode using heterostructures of organic and inorganic materials combined with perovskite materials.
Recently, among inorganic semiconductor thin films, nickel oxide (NiO x ) has been considered as a crucial building block for modern optoelectronic devices owing to its versatile physical and chemical features along with its appropriate band structure. Specifically, NiO x has been used in organic and perovskite PV cells as a HTL owing to its efficient chemical and thermal stability, outstanding hole-transport characteristics, large band gap (> ~ 3.7 eV), the ability to control the valence-band energy level effectively, and its deep valence band position (> 5.2 eV) [23][24][25] . Chen et al. used a NiO x HTL synthesised by a sputtering method at a low temperature to devise inverted-type perovskite PV cells, achieving a PCE of 11.6% 26 . Zhu et al. introduced a solution-processed NiO x :PbI 2 nanocomposite structure which operated at room-temperature to assist with compact and crystalline MAPbI 3 film growth, which is critical during the fabrication of effectual photodetectors. This nanocomposite served as an electron-blocking, hole-extracting, and passivation layer and ultimately suppressed the dark current, resulting in improved photo characteristics, such as better detectivity and faster response times 27 . More recently, a thin interlayer of NiO x , synthesised by a comparatively simple solution-processed method demonstrated tunable work functions (5.0 -5.6 eV) and efficient hole conductivity outcomes 28,29 . A thin layer of NiO x as a HTL in perovskite PV cells was also shown to hinder the penetration of water and oxygen into a device effectively 30 . Zin et al. proposed a simple method by which to synthesise a NiO x layer and used it as a HTL in inverted perovskite PV cells. Their device showed greatly improved values of PCE (~ 20.2%) 24,31 . Particularly, a thin NiO x film showed enhanced charge carrier density and conductivity through improvements of the interfacial charge extraction, thus reducing the interface trap density and providing better energy level alignment with the perovskite layer in perovskite solar cells 32 .
Inspired by these remarkable characteristics of NiO x layers, a heterostructure approach which involves the assembly of a halide perovskite with the inorganic semiconducting material of NiO x may represent an alternative means by which to design photodiode that are anticipated to have enhanced photosensitivity. In this work, a high-performance and unique perovskite photodiode with a HTL of NiO x exhibiting a high responsivity value is reported. We designed a self-powered hybrid perovskite-based photodiode structure that suppresses the dark leakage current and enhances the performance of the perovskite photodiode. The effects of the HTL of NiO x in the perovskite photodiode on the device performance and limitation factors are investigated by evaluating the optoelectronic characteristics of the photodiode. We measured the optoelectronic characteristics of the photodiode device under different power levels and wavelengths of incident laser light to obtain the R, D, and EQE values at a given bias voltage. Further, we took temporal photoresponse measurements to determine the response time (rise and decay time) characteristics. We also verified the EQE values through incident-photonto-electron conversion efficiency (IPCE) measurements and confirmed the degree of self-consistency in the results. Moreover, we measured the characteristics of reference devices with poly (3, 4-ethylene dioxythiophene) polystyrene sulfonate (PEDOT:PSS) as a HTL in order to compare and support our results, showing that NiO x as a HTL exhibits much improved photodiode performance as compared to the conventional HTL of PEDOT:PSS.

Results
Initially, in order to understand the surface properties of the fabricated sample NiO x and reference PEDOT:PSS layers, atomic force microscopy (AFM), Kelvin probe force microscopy (KPFM), and contact angle measurements were utilised to characterise the layers, especially with regard to their surface roughness and surface potential levels. Figure 1a, 33 , finding a value of approximately 5.1 eV, slightly higher than that (5.0 eV) of the PEDOT:PSS layer. We also measured the water contact angles of the NiO x layer and compared these outcomes with those of the reference PEDOT:PSS layer, as shown in Fig. 1c. The observed water contact angles were 16.1° ± 2.8° for the NiO x layer and 13.4° ± 1.4° for the PEDOT:PSS layer, indicating that the NiO x surface is more hydrophobic than the PEDOT:PSS layer. Thus, the relatively non-wetting NiO x surface may improve the film formation of the active perovskite layer and facilitate uniform and continuous perovskite film morphologies with large crystalline domains 34,35 .
Next, we examined the effects of the NiO x and PEDOT:PSS layers on the formation of a light harvester, in this case a perovskite layer (CH 3 NH 3 PbI 3 , MAPbI 3 ). Here, the perovskite layers were prepared using a toluene-assisted rapid-crystalline technique 36 , to attain a uniform, continuous, flat, and full-coverage film, as mentioned in the "Experimental section". The thickness of the perovskite layer was fixed at 250 nm in all layers and devices in this study, as the device performance of a perovskite photodiode can be influenced by the perovskite film thickness 37 . To compare the film quality levels of the fabricated perovskite layers on the NiO x and PEDOT:PSS underlying layers as HTLs, the surface morphologies of the perovskite films on the NiO x and PEDOT:PSS HTLs were characterised by scanning electron microscopy (SEM). Figure 2a,b show surface SEM images of the CH 3 NH 3 PbI 3 perovskite layers on the NiO x and PEDOT:PSS HTLs, respectively. Notably, the grain size (≈ 254 nm) of the perovskite film on the NiO x HTL greatly exceeds that (≈ 166 nm) on the PEDOT:PSS HTL. This result verifies that the relatively non-wetting NiO x surface suppresses perovskite nucleation sites, which facilitates the growth of larger grains 38 . The large grain size of the perovskite film can reduce the recombination rate of the charge carriers due to the decreased charge trap density 34 .
As depicted in Fig. 2c, the perovskite films on the NiO x and PEDOT:PSS HTLs were also characterised by X-ray diffractometry (XRD). Robust and major Bragg peaks were observed at 14.3°, 28.6°, and 32.1°, corresponding to the (110), (220), and (310) planes of the MAPbI 3 perovskite film given its orthorhombic crystal structure 34 . On the NiO x HTL, we observed a slight improvement in the XRD peak intensity levels of the perovskite film www.nature.com/scientificreports/ as compared to those on the PEDOT:PSS HTL, also implying the larger grains with high crystallinity of the perovskite film. Note that there was no significant change in the XRD peak ratios, implying that the different HTLs did not cause any substantial change in the crystal orientations of the perovskite films. Figure 2d shows the optical absorption spectra of the perovskite films on the NiO x and PEDOT:PSS HTLs. We observed that the absorption spectra of the perovskite layers in both cases on the HTLs were nearly identical to each other 34 . It was also noted that the optical transmittance levels of the NiO x layer are comparable but slightly lower in the visible range compared to those of the PEDOT:PSS layer (Fig. S1). Next, in order to examine the effects of the NiO x and PEDOT:PSS HTLs on the surface roughness and energy level of the perovskite layer, AFM and KPFM were used to characterise the perovskite layers on the NiO x and PEDOT:PSS HTLs. Figures 3a and b show the results of observations of the AFM topologies and the KPFM surface potential maps of the perovskite film on the NiO x HTL, respectively. For comparison, AFM images and KPFM surface potential maps of the perovskite film on the PEDOT:PSS HTL are shown in Fig. S2. The roughness values of the perovskite films on the NiO x and PEDOT:PSS HTLs were approximately 8.5 and 9.7 nm, respectively. These results indicate that the active MAPbI 3 perovskite layer formed on the NiO x HTL is more continuous, flat, and uniform than that on the PEDOT:PSS HTL. We also determined the Fermi levels of the perovskite layers on the HTLs using the obtained surface potentials 33 , finding that they were approximately 5.1 eV. It is noted that the Fermi level of the NiO x HTL is closer to that of the perovskite film as compared to that of the PEDOT:PSS HTL. Thus, we considered that the interface quality and affinity between the perovskite layer and the NiO x HTL are better than those between the perovskite layer and the PEDOT:PSS HTL due to the good crystallinity with large grains, low surface roughness, and a small difference in the Fermi level of the perovskite layer on the NiO x layer, all of which may promote the performance of perovskite photodiodes 39,40 .
Subsequently, we investigated the device performance of a self-powered perovskite photodiode. A schematic illustration of the self-powered perovskite photodiode is presented in Fig. 4a. In the structure, ITO is used as the anode material, NiO x serves as the HTL, the perovskite layer (CH 3 NH 3 PbI 3 ) functions as the active layer, phenyl-C61-butyric acid methyl ester (PCBM 60 ) and ZnO NPs form the ETLs, bathocuproine (BCP) is used as a buffer layer for a reduction of the nonradiative recombination of excitons, and Al is utilised as the cathode (for details, see the "Experimental section").  41,42 . Thus, the NiO x HTL provides a more favourable alignment of the energy level with respect to the VBM of the perovskite active layer. The high value of the VBM of the NiO x HTL would decrease the energy barrier, which may in turn increase the build-in potential (V bi ) and facilitate improved hole transportation, consequently leading to a larger value of the open-circuit voltage (V oc ) 24,34,41 . In order to estimate the PV characteristics of the fabricated self-powered perovskite photodiodes, the photocurrent density-voltage (J-V) characteristics of the photodiodes were measured under the AM1.5G illumination source. Figure 4c shows the observed J-V characteristics of the sample perovskite photodiode with the NiO x HTL (hereafter, the sample device) and of the reference photodiode with the PEDOT:PSS HTL (hereafter, the reference device). The estimated PCE value from the J-V curve was 13% in the sample device with an open-circuit voltage V OC of 1.03 V, a short-circuit current density (J SC ) of 21 mA cm −2 , and a fill factor (FF) of 61%; the PCE in this case is much higher than the PCE value of 8% in the reference device with V oc = 0.95 V, J sc = 19 mA cm −2 , and FF = 45%. These PV results are consistent with previous results for perovskite PV cells 26,43,44 . Thus, it is obvious that the perovskite photodiodes based on the NiO x HTL exhibit improved PV performance. Note that the increased V oc value of the sample device is mainly due to the high VBM of the NiO x HTL, which implies an increase in the built-in potential 29 . We also took measurements via forward and backward scans of the photodiode devices studied here (Fig. S3) to augment our discussion of the hysteresis in the J-V scans of the perovskite devices [45][46][47] . We calculated the hysteresis index (HI) using the relationship of HI = J RS (0.8Voc)−JFS(0.8Voc)) , where J RS (0.8V OC ) and J FS (0.8V OC ) represent the photocurrent densities at a bias voltage of 0.8 V OC for the reverse and forward scans, respectively 47 . The estimated values of HI for the photodiodes with the NiO x and PEDOT:PSS HTLs are 0.06 and 0.60, respectively, which indicates that the sample photodiode with the NiO x HTL shows negligible hysteresis. Thus, the much smaller value of HI for the photodiode with the NiO x HTL than for that with the PEDOT:PSS HTL is mainly owing to the high crystallinity of the perovskite layer and rapid charge extraction by the NiO x HTL [45][46][47] .
At this point, we focus our attention on the photodiode performance of the sample device fabricated with NiO x . Figure 4d shows the current density as a function of the bias voltage (J-V) at different input power levels (0 -349 mW cm −2 ) of incident laser light having a wavelength of = 532nm . We swept the voltage from − 1.0 to + 2.0 V and measured the current density J. Both the forward and reverse current density levels increase gradually as the input power of the incident light increases.  where c, P, and θ are a proportional constant, the input power intensity of the incident laser light, and the powerlaw index, respectively. With best-fit parameters, the obtained value of θ for the sample device with NiO x is 0.89, which is close to 1.0 for an ideal photodiode with a low trap state junction and which is noticeably higher than that ( θ = 0.69) for the reference device with PEDOT:PSS (Fig. S5). Such a relatively large θ value for the sample device indicates that a small number of trap states exist in the perovskite layer on the NiO x HTL, which is valuable for the realisation of high photo-sensing ability through the efficient collection of a large number of photo-excited charge carriers. Next, to evaluate R and D of the perovskite photodiode at zero bias voltage (V bias = 0 V, self-powered condition), we used the following relationship 49 , where J PH is the net photocurrent density J PH = J light − J dark and A denotes the illuminating junction area of the photodiode for the incident laser light. Figure 5a shows the R value, estimated from the current density data (shown in Fig. 4d), as a function of the input power of incident laser light (532 nm) for the sample device. The highest estimated R value of the sample device is 340 mA W −1 at zero bias voltage (self-powered), which is much higher as compared to previously reported values, mainly due to the efficient suppression of the dark leakage current. Moreover, the value of R increases to 440 mA W −1 when the bias voltage V bias = − 1.0 V. We also estimated the detectivity D value of the sample device using the following relationship 50 , www.nature.com/scientificreports/ where q denotes the charge of the electron. Figure 5b shows the D values obtained for several input power levels of incident laser light (532 nm). The highest estimated D value of the sample device with NiO x is approximately 1.9 × 10 11 Jones, which is also much higher than earlier values reported in the literature. Additionally, we measured the R and D values of the sample device with NiO x with incident laser light at wavelengths of 594 and 633 nm to quantify the wavelength selectivity of the photodiode, as shown in Fig. 5c,d, respectively. In this case, we also obtained high corresponding R values of 360 and 330 mA W −1 with improved D values of 2.08 × 10 11 and 1.8 × 10 11 Jones under incident laser light illumination with wavelengths of 594 and 633 nm, respectively, at zero bias voltage. These results clearly demonstrate the highly efficient self-powered operation of the perovskite photodiode studied here. Figure 6a shows comparisons of the R values obtained from the sample and reference devices in three incident light with different wavelengths. Clear differences were noted in R values, and the R value is enhanced from 240 to 360 mA W −1 at 594 nm when NiO x is introduced to replace the conventional PEDOT:PSS as the HTL. This outcome provides clear evidence that the sample device based on the NiO x HTL exhibits superior responsivity R relative to the reference device based on the conventional PEDOT:PSS HTL. We also measured the stability of the photodiodes studied here with the NiO x and PEDOT:PSS HTLs (Fig. S6). As shown in the figure, the device with NiO x is much more stable as compared to that with PEDOT:PSS 25,51 . Thus, it is clear that our self-powered perovskite-based photodiode with NiO x is promising for highly sensitive photodiodes requiring low energy consumption levels.

Discussion
In order to understand the cause of such high responsivity of the sample device with NiO x , the dark and photo J-V characteristics were compared and analysed, as shown in Fig. 6b. In this figure, we also compare the J-V curves of the reference device with PEDOT:PSS. In a dark condition, we observed clear diode behaviour with a www.nature.com/scientificreports/ high rectification ratio of 2.0 × 10 4 for the sample device. This rectification ratio is nearly 27 times higher than that ( 7.5 × 10 2 ) for the reference device. Moreover, the sample device with NiO x clearly shows relatively low leakage current compared to the reference device with PEDOT:PSS. The dark current density is suppressed in the sample device to 7.84 × 10 −6 mA cm −2 , a much smaller value as compared to that ( 2.84 × 10 −5 mA cm −2 ) of the reference device at zero bias voltage. This reduction in the dark current density confirms the low leakage current in the perovskite photodiode with NiO x 27 . Thus, it is clear that the perovskite photodiode based on the NiO x HTL exhibits much better diode characteristics than the reference device with the conventional PEDOT:PSS HTL. Further, the sample device shows higher photocurrent density levels under incident laser light in both the forward and reverse regions as compared to the results of the reference device. The photocurrent density of the sample device increases to 7.5 mA cm −2 under incident laser light (532 nm) with an input power of 349 mW cm −2 at zero bias voltage. Therefore, we considered that the relatively low leakage current in the dark and the increased photocurrent outcomes of the sample device with NiO x may promote the improved device performance through the realisation of a leakage-free photodiode.
To clarify the remarkable effect of the NiO x HTL on the high responsivity of the perovskite photodiode, the dark current flows of the devices (Fig. 6b) are analysed using the following relationship for the Shockley diode in a single-junction device 52 , where J 0 , n, K, and T represent the saturation current density, ideality factor, Boltzmann's constant, and temperature, respectively. For an ideal diode (p-n junction), the ideality factor n is expected to be equal to 1.0 without charge carrier trapping 53 . Thus, the ideality factor n can be a key parameter to estimate the suppression of the recombination rate in perovskite photodiodes. From the analyses of the dark currents, the estimated values of ns are 1.1 and 1.7 for the sample and reference devices, respectively. The smaller value of n for the sample device as compared to that for the reference device stems from the reduced number of monomolecular recombinations 53 . To verify this reduction of recombinations, we also estimated the recombination resistance R rec for the charge carriers using the following relationship 54 , where V e , and R s represent the effective voltage, the minimum voltage used to extract n from the J-V curves, and the series resistance, respectively. Based on the above relationship, we obtained a higher value of R rec (≈ 105�) for the sample device with NiO x as compared to that (≈ 79�) for the reference device with PEDOT:PSS. Thus, the higher R rec for the sample device verifies higher resistivity to carrier recombinations in the photodiode with the NiO x HTL 54,55 .
To clarify and examine the defect states and trap recombination at the HTL/perovskite interface, the spacecharge limited current model is employed. The trap-state density (n trap ) can be calculated using the following relationship 56 , where ε, ε 0 , V TFL , e, and L correspondingly represent the relative dielectric constant of perovskite, the dielectric constant of a vacuum, the trap-filling limited voltage, the elementary charge, and the thickness of the perovskite film. The value of V TFL can be derived from the J-V characteristics of reference and sample photodiode devices in a dark condition (Fig. S7). The V TFL values of the devices with the NiO x and PEDOT:PSS HTLs are 0.42 and 0.74 V, respectively. Thus, the estimated n trap values of the devices are 1.83 × 10 15 cm −3 for the reference device with the PEDOT:PSS HTL and 1.12 × 10 15 cm −3 for the sample device with the NiO x HTL, clearly demonstrating fewer trap recombinations at the NiO x /perovskite interface 20,57 .
Next, the reduction of charge recombination of the sample device was analysed further on the basis of the open-circuit voltage V oc because all of the photo-excited charge carriers will recombine within the device at the end at the open-circuit condition 36,58 . By linear least squares fitting of the V oc data (Fig. 4e), slopes of 1.2K B T/q and 1.4K B T/q were attained for the sample and reference devices, respectively. In principle, when the slope is equal to 1.0K B T/q , the device operates without trapping charge carriers or is governed by bimolecular recombinations, and the active layer is considered as having recombinations free of electrons and holes; when the slope exceeds 1.0K B T/q , monomolecular Shockley-Read-Hall (SRH) recombinations are involved 53 . Thus, our NiO x -based sample device exhibits a smaller slope of 1.2K B T/q than that ( 1.4K B T/q ) of the reference device, verifying that the NiO x -based device can efficiently reduce monomolecular recombinations with less charge carrier trapping, contributing to the improvement of the device performance 59 .
Further, we estimated V bi of the perovskite photodiode using the relationship ( V bi = −nK B T q (ln J 0 )) of the diode model 60 . The estimated value of V bi is ≈ 0.52V for the sample device with NiO x , while the value of V bi is ≈ 0.36 V for the reference device with PEDOT:PSS. Such a high V bi value of the sample device is a direct signature of the improved charge-selecting properties of the NiO x HTL. Thus, it is clear that the hole-selecting capability of NiO x is much greater than that of PEDOT:PSS, causing the high responsivity of the self-powered perovskite photodiode to be greater with NiO x 29 . Next, to gain a deeper understanding of the charge transfer in the perovskite photodiodes based on distinct HTLs, impedance spectroscopy (IS) measurements were taken for the photodiodes in the dark. Figure 6c shows www.nature.com/scientificreports/ Nyquist plots from IS data for the photodiodes with NiO x and PEDOT:PSS at zero bias voltage. We observed clear semicircles that distinguish the intermediate-frequency (f) regions for both photodiodes. These are linked to the charge transfer at the HTL/perovskite/ETL interfaces, primarily owing to recombinations. From the fitting of the IS data, the capacitance (C 1 ), parallel resistance (R 1 ), and series resistance (R 2 ) values of the photodiodes with NiO x and PEDOT:PSS were deduced. The equivalent circuit diagram is shown in the inset figure. The obtained values of C 1 , R 1, and R 2 are correspondingly 3.87 nF, 87.04 Ω and 34.94 Ω for the sample device with NiO x and are 3.90 nF, 95.48 Ω and 35.14 Ω for the reference device with PEDOT:PSS. The C 1 and R 2 values of the sample device are slightly lower than those of the reference device, indicating fewer surface trap state charges with less charge accumulation at the interface between the perovskite layer and the NiO x HTL compared to the PEDOT:PSS HTL. Moreover, the R 1 value of the sample device is clearly lower than that of the reference device. This low R 1 of the NiO x -based perovskite photodiode can be attributed to the low interfacial charge transfer resistance at the perovskite/NiO x interface. The smaller value of R 1 also implies faster hole transport at the NiO x /perovskite interface compared to that at the PEDOT:PSS/perovskite interface. Next, we also measured the temporal responses of the sample device with NiO x as the HTL at various input power levels (P = 3, 69, 139, 209, 279, and 349 mW cm −2 ) while turning the incident laser light (532 nm) on and off (Fig. 6d). The rapid rise ( τ r ) and decay ( τ d ) response times of the sample device were found to be τ r ≈ 0.9 and τ d ≈ 1.8 ms, respectively, at zero bias voltage. For comparison, we also measured the photoresponse of the reference device (Fig. S8). From the comparison, it was found that the photoresponse of the sample device is much faster than that (τ r ≈ 3, τ d ≈ 10 ms) of the reference device. Thus, it is clearly shown that the NiO x HTL is a key component required to realise a fast self-powered perovskite-based photodiode.
Next, the EQEs of the perovskite photodiodes were also estimated from their R values using the following relationship 61 , where h and c correspondingly denote the Planck constant 6.62 × 10 −34 J s and the speed of light 3.8 × 10 8 ms −1 . The estimated EQE values are 76.5% and 61% for the sample device with NiO x and the reference device with PEDOT:PSS, respectively, at a wavelength of 532 nm. As expected, the EQE value of the sample device with NiO x is much higher than that of the reference device with PEDOT:PSS. This high EQE value for the sample device is the result of the improved charge-selecting properties as well as the improved exciton dissociation at the perovskite/HTL interface with the help of the high built-in electric field in the heterostructure. In order to confirm the estimated EQE values, we also measured the EQE spectra using an incident photon-to-current collection efficiency (IPCE) system, as shown in Fig. 7a. The measured EQE values from IPCE measurements were 76% and 60% for the sample and reference devices, respectively, at the wavelength of 532 nm, which are nearly identical to those estimated from the R values. For further comparison with other wavelengths, Fig. 7b shows the high EQE values estimated from both R and IPCE measurements of the sample device with NiO x , verifying the close correspondence of these values and supporting the high device performance of the perovskite photodiode with the NiO x HTL.
We also compared D with other reported values in Table S1, as shown in Fig. 7d 50,65,68,[71][72][73][74][75][76][77] . It is clear from these figures that R and D of our photodiode with NiO x in this study are much higher compared to those reported previously. These findings overall show that the interface quality and recombination activities in perovskite photodiodes are mainly influenced by the electron-blocking NiO x HTL, which will surely provide a platform for further improvements in the performance of photodiodes coupled with excellent charge-selecting layers.

Conclusions
In summary, we designed a self-powered hybrid organic-inorganic perovskite photodiode with an effective and capable electron-blocking hole-transport NiO x layer. We used NiO x as the HTL in the fabrication of the perovskite photodiode to improve the interface quality by suppressing the dark leakage current reaching to 7.84 × 10 −6 mA cm −2 . The perovskite photodiode fabricated with NiO x exhibited better PV performance with a PCE of 13% as compared to the photodiode with conventional PEDOT:PSS. A remarkably high responsivity R value of 360 mA W −1 with detectivity D = 2.08 × 10 11 Jones and EQE = 76.5% for the self-powered perovskite photodiode with the NiO x HTL was noted under incident laser light with a wavelength of 594 nm at zero bias voltage. Further, the performance capabilities of the perovskite photodiode were estimated at different bias voltages; the value of R increased gradually to 428 mA W −1 with D = 3.6 × 10 11 Jones and EQE = 77% at a bias voltage of V bias = − 1 V. Based on the diode model, we deduced an ideality factor of 1.1 and a high built-in potential value of V bi ≈ 0.52V for the photodiode with the NiO x HTL, thus providing direct evidence of the improvement of the charge-selecting characteristics of the NiO x layer. Furthermore, we observed fast rise and decay times of approximately 0.9 to 1.8 ms, respectively, for the perovskite photodiode with the NiO x HTL at zero bias voltage, values which are much faster than those of the reference photodiode with the conventional PEDOT:PSS HTL. Therefore, the self-powered perovskite photodiode studied here opens up an opportunity for applications of hybrid perovskite heterostructures with the electron-blocking NiO x HTL in highly sensitive light-detecting optoelectronic devices that consume low amounts of energy, such as optical sensors, waveguideintegrated photodiodes, and/or nano-photodetectors. To form the HTLs of NiO x (40 nm) and PEDOT:PSS (40 nm), the prepared solutions were spin-coated onto ITO substrates at 3000 rpm for 40 s and at 4000 rpm for 35 s, respectively. For NiO x , the coated precursor layer was annealed at 70 °C for 1 min and then at 300 °C for 1 h, while for PEDOT:PSS, the coated layer was annealed at 120 °C for 30 min. To make the perovskite precursor solution, a mixture of MAI and PbI 2 at a 1:1 molar ratio was dissolved in a mixed solvent of GBL and DMSO at a 7:3 volume ratio. The precursor solution was stirred overnight at 60 °C. The perovskite precursor solution (50 µl) was spin-coated at 1000 rpm for 10 s and then at 3000 rpm for 25 s to obtain the desired thickness of 250 nm onto the NiO x and PEDOT:PSS layers. During the spinning process, an anti-solvent (toluene ~ 150 µl) was poured onto the precursor layer at 17 s. The perovskite precursor layer was then annealed at 100 °C for 20 min. For the ETL, a 20 mg/ml of PCBM 60 solution, dissolved www.nature.com/scientificreports/ in CB, was spin-coated at 1500 rpm for 60 s onto the perovskite layer. A dispersion of ZnO NPs mixed with IPA (7:3) was then spin-coated at 1000 rpm for 10 s and at 4000 rpm for 40 s onto the PCBM 60 layer. Finally, BCP (12 nm) and Al (70 nm) were deposited onto the ZnO NP layer using a thermal evaporation system. Thus, the device structure of the sample photodiode was [ITO/NiO x /CH 3 NH 3 PbI 3 /PCBM 60 /ZnO NPs/BCP/Al]. The active area of the fabricated devices was 6 mm 2 in size.
Film and device characterisation. The water contact angles of the fabricated NiO x and PEDOT:PSS layers were measured using a contact angle meter (Phoenix 300 Touch, Surface Electro Optics). The fabricated perovskite film was characterised using a field emission scanning electron microscope (SEM, Model JSM-6700F, JEOL Co.) to analyse the surface morphology. To investigate the surface roughness and surface potential of the organic and inorganic functional layers, atomic force microscopy (AFM) and Kelvin probe force microscopy (KPFM, FlexAFM, Nanosurf AG) were used. To measure the surface potential, a Pt/Ir-coated silicon tip (resonance frequency = 87 kHz and a force constant = 3.9 Nm −1 , NanoWorld, Inc.) was used while applying AC voltage of 1 V at a frequency of 18 kHz. An X-ray diffractometer (XRD-Rigaku D/Max 2200, λ = 0.154 nm) was used to check the crystallinity of the perovskite layers on the NiO x and PEDOT:PSS layers. To estimate the optical absorption spectra of the perovskite layers, a UV-visible spectroscopy system (8453, Agilent) was employed. The photocurrent versus bias voltage (J-V) characteristics were measured using a source meter (2400, Keithley) under illumination of incident laser light with several different wavelengths and were calibrated using a reference commercial silicon photodiode (THORLABS-PDA10A2) (see Fig. S9). The PV performance of the fabricated photodiode was measured under an illumination intensity of 100 mW cm −2 generated by an AM1.5 light source (Newport, 96,000 Solar Simulator) and calibrated using a reference cell (Bunkoh-keiki, BS-520). Impedance spectroscopy (IS) measurement of the photodiode was performed in the dark using an impedance analyser (HP 4192A) with an AC oscillating amplitude of 100 mV (RMS) to maintain the linearity of the response. The EQE spectra were also measured using an incident photon-to-current collection efficiency (IPCE) measurement system (IQE-200 EQE/IQE, Newport).