High-Performance Photoresistors Based on Perovskite Thin Film with a High PbI2 Doping Level

We prepared high-performance photoresistors based on CH3NH3PbI3 films with a high PbI2 doping level. The role of PbI2 in CH3NH3PbI3 perovskite thin film was systematically investigated using scanning electron microscopy, X-ray diffraction, time-resolved photoluminescence spectroscopy, and photoconductive atomic force microscope. Laterally-structured photodetectors have been fabricated based on CH3NH3PbI3 perovskite thin films deposited using precursor solution with various CH3NH3I:PbI2 ratios. Remarkably, the introduction of a suitable amount of PbI2 can significantly improve the performance and stability of perovskite-based photoresistors, optoelectronic devices with ultrahigh photo-sensitivity, high current on/off ratio, fast photo response speed, and retarded decay. Specifically, a highest responsivity of 7.8 A/W and a specific detectivity of 2.1 × 1013 Jones with a rise time of 0.86 ms and a decay time of 1.5 ms have been achieved. In addition, the local dependence of photocurrent generation in perovskite thin films was revealed by photoconductive atomic force microscopy, which provides direct evidence that the presence of PbI2 can effectively passivate the grain boundaries of CH3NH3PbI3 and assist the photocurrent transport more effectively.

Photodetectors are the critical component for converting incident light to electrical signal and are broadly utilized in optical communication, environmental monitoring, medical analysis, and astronomy [27,28]. Compared with inorganic semiconductors (such as ZnO, GaN, and Si), organic-inorganic perovskite-based photodetectors possess many advantages in terms of high photoresponsivity, high sensitivity, and fast response speed, in addition to their simple and low-cost processing requirements for large-scale fabrication [29][30][31][32][33].
For a certain perovskite material of AMX 3 stoichiometry, chemical composition is a crucial factor affecting device performance. CH 3 NH 3 PbI 3 (MAPbI 3 ) is one of the most widely studied photoactive layer materials for solar cells and photodetectors [34][35][36][37]. Previous reports have investigated that a slight excess of PbI 2 in the MAPbI 3 film is beneficial for solar cell performance. Kim et al. reported the presence of PbI 2 in MAPbI 3 significantly reduces the hysteresis of IV characteristics and ion defect migration and results in a high efficiency of 19.75% [38]. Carmona et al. reported that an excess of PbI 2 can improve the crystallinity of the MAPbI 3 film and the electron transfer to the TiO 2 layer, leading to a high average cell efficiency of 18% [39]. Yang et al. showed that PbI 2 filling grain boundaries of the perovskite not only reduced the number of trap sites but also generated significant band bending [40]. The doping level of PbI 2 is critical, which plays an important role in the chemical stability and optoelectronic performance of perovskites. Zhang et al. reported that the PbI 2 formed at grain boundaries of the perovskite can effectively passivate perovskite films and improve the charge separation efficiency with a longer PL lifetime and enhanced open-circuit voltage [41]. Besides, it has also been reported that PbI 2 may passivate the TiO 2 interface and reduce the charge recombination [42]. Similarly, an excessive amount of PbI 2 can also act as an electron blocking layer and facilitate hole transport between the interface of the perovskite and hole transport layer [43]. However, most of the current studies on the role of PbI 2 are limited to solar cell applications. With increasing research interest in the photo-detection study of perovskites, there is a demand to systematically investigate the role of PbI 2 in photodetectors.
In this study, we present a high-performance photoresistor photodetector based on MAPbI 3 with a high PbI 2 content. Different from previous reports in which only a slight excess of PbI 2 was proved to be beneficial, the optimized content of PbI 2 in our study was much higher. Our results demonstrate that a PbI 2 -rich perovskite precursor (MAI:PbI 2 = 0.5:1) is beneficial to photodetector performance in terms of photocurrent, responsivity, and photocurrent on/off ratio. Furthermore, the stability of the device can be significantly improved with the incorporation of PbI 2 .

Materials and Preparation of Perovskite Precursor Solution
CH 3 NH 3 I (MAI) and PbI 2 were purchased from Xi'an Polymer Light Technology Corp. Dimethyl formamide (DMF, 99.9%) was purchased from Aladdin. All chemicals were directly used without further purification. For the perovskite precursor solution preparation, MAI and PbI 2 with various mole ratios (0.25, 0.5, 0.75, 1.0) were dissolved in DMF and stirred overnight at room temperature. PbI 2 is fixed at 1 mol/L in all precursor solutions. All prepared solutions were filtered with a polytetrafluoroethylene filter (pore size~0.2 µm) before film deposition by spin-coating method.

Fabrication of the Devices
Indium tin oxide (ITO) glass with a sheet resistance of~6 Ω sq −1 was used as a conductive substrate for device fabrication. The ITO was laser-etched to form electrode-patterns and subsequently ultrasonicated in acetone, isopropanol, ethanol, and deionized water, and blown dry with N 2 . Then the substrates were treated with air plasma for 20 min. After that, perovskite thin films were spin-coated at 4000 rpm for 25 s and annealed on a hot plate at 100 • C for 30 min. The photoresistors prepared with different precursor solutions (MAI-to-PbI 2 ratio of 0.25:1, 0.5:1, 0.75:1, and 1:1) were denoted as S1, S2, S3, and S4, respectively. All procedures were conducted in a glove box filled with argon. The active area for all photoresistors is 0.02 cm × 0.7 cm (0.014 cm 2 ).

Characterization of Perovskite Thin Films and Device Performance
The surface and cross-section morphology of the perovskite thin film was examined by scanning electron microscopy (FEG SEM SU-70, Hitachi, Tokyo, Japan). The steady-state photoluminescence (PL) spectra measurements were conducted by the WITec's Raman microscope alpha300 R using 532 nm as excitation source. The time-resolved photoluminescence (TRPL) spectroscopy was measured by FS5 Fluorescence Spectrometer (Edinburgh Instruments, Edinburgh, England). The X-ray diffraction (XRD) measurements were carried out by a Rigaku D/MAX-2500 diffractometer with Cu K α X-ray source. The photoelectronic performance of the devices was measured in air. A fiber-coupled LED (MCWHF2, Thorlabs) was used to provide white light, and different band pass filters were installed on the LED lamp to achieve monochromatic light. The IV curves under dark and white light conditions were measured by an electrochemical station (CHI 760e, CH Instruments Ins., Shanghai, China) in a dark box. The light intensities of white light and monochromatic light were recorded by TES-132 solar power meter and a Thorlabs power meter (PM100D with a S121C standard photodiode power sensor), respectively, and light intensity was adjusted by regulating the LED lamp's working current. We set the LED lamp in a pulse mode to achieve pulse illumination, which allowed the recording of the time-dependent photocurrent response (I-t) by the electrochemical station.
Photo conductive atomic force microscopy (PC-AFM) was conducted using a Bruker Dimension ICON AFM. A Pt-Ir coated silicon probe with a force constant of 3 N/m was used for PC-AFM measurement with peak-force TUNA mode. The light source was a 532 nm laser (Changchun New Industries Optoelectronics Technology Co. Ltd., Changchun, China). The laser was introduced by an inverted optical microscope with an objective lens and vertically shined onto the sample. Topography and current images were achieved simultaneously and compared directly. All of the AFM measurements were conducted in the dark and under 532 nm laser illumination. The NanoScope-Analysis software package was used to analyze the AFM images.

Results and Discussion
The compositions of the as-synthesized perovskite thin films were characterized by XRD (as shown in Figure 1a). For samples S1 to S3, an additional diffraction peak at 12.56 • was observed which originated from the reflection of (001) lattice plane of PbI 2 [34], while the (001) diffraction peak from PbI 2 became nearly absent in the sample with a MAI:PbI 2 ratio of 1:1 (sample S4). Three typical peaks of perovskite located at 14.09 • , 28.44 • , and 31.85 • can be observed, corresponding to the reflections from (110), (220), and (310) lattice plane [9,30,39], respectively. As the MAI:PbI 2 ratio varied from 0.25:1 to 1:1, typical perovskite peaks got enhanced gradually. The relatively strong intensities were probably caused by the preferred crystal orientation of (110). Figure 1b-e shows the SEM morphology of PbI 2 -rich perovskite thin films (S1-S3) and the pure-phase perovskite (S4). The grain size of the thin films is ranging from~10 nm to~100 nm, and becomes larger as the MAI:PbI 2 ratio varies from 0.25:1 to 1:1. Figure 2a shows the absorption spectra of all the perovskite films with a thickness of~300 nm. Similar absorption band-edge positions (~770 nm) were observed for all films. Furthermore, the absorption spectra show an obvious increase with the increasing amount of MAPbI 3 in the range between 450 nm and 850 nm, which can be ascribed to a higher absorption coefficient of MAPbI 3 than PbI 2 . Furthermore, the peak shoulder observed at 500 nm in the absorption spectra is attributable to the band-edge absorption of PbI 2 [38]. Steady-state PL spectra of the thin films are shown in Figure 2b and the inset shows the change of PL emission peak positions with various MAI:PbI 2 ratios. It can be clearly seen that the peak position shifts towards 776 nm as the MAI:PbI 2 ratio varies from 0.25:1 to 1:1. It is reported that the interplay between inorganic and organic interactions has an important effect on the lattice strain and the band gap of semiconductors [39]. A head-to-tail configuration of organic cations would cause a reduced lattice strain and larger grain size of perovskite materials. This reduced lattice strain leads to lattice expansion and a red shift in PL emission peak position. Here, the SEM results (Figure 1b-e) show that the grain size of perovskite increased with varying MAI:PbI 2 ratios (from 0.25:1 to 1:1), and a red shift on PL emission peak position (Figure 2b) is observed for these samples. In a recent work, Jones et al. observed that lattice strain is directly related to enhanced defect concentrations and non-radiative recombination on the micro-scale [44]. Here, the PL intensities are stronger for the PbI 2 -rich perovskites compared with the pure-phase perovskite, which is consistent with previous reports [38,39,45]. The presence of PbI 2 can passivate the grain boundaries of the perovskite films and this passivation effect on PL intensity is dominant compared with lattice strain. The sample with a MAI:PbI 2 ratio of 0.5:1 (sample S2) shows the strongest PL emission, which suggests an effective suppression of non-radiative recombination.  Figure 2a shows the absorption spectra of all the perovskite films with a thickness of ~300 nm. Similar absorption band-edge positions (~770 nm) were observed for all films. Furthermore, the absorption spectra show an obvious increase with the increasing amount of MAPbI3 in the range between 450 nm and 850 nm, which can be ascribed to a higher absorption coefficient of MAPbI3 than PbI2. Furthermore, the peak shoulder observed at 500 nm in the absorption spectra is attributable to the band-edge absorption of PbI2 [38]. Steady-state PL spectra of the thin films are shown in Figure  2b and the inset shows the change of PL emission peak positions with various MAI:PbI2 ratios. It can be clearly seen that the peak position shifts towards 776 nm as the MAI:PbI2 ratio varies from 0.25:1 to 1:1. It is reported that the interplay between inorganic and organic interactions has an important effect on the lattice strain and the band gap of semiconductors [39]. A head-to-tail configuration of organic cations would cause a reduced lattice strain and larger grain size of perovskite materials. This reduced lattice strain leads to lattice expansion and a red shift in PL emission peak position. Here, the SEM results (Figure 1b-e) show that the grain size of perovskite increased with varying MAI:PbI2 ratios (from 0.25:1 to 1:1), and a red shift on PL emission peak position (Figure 2b) is observed for these samples. In a recent work, Jones et al. observed that lattice strain is directly related to enhanced defect concentrations and non-radiative recombination on the micro-scale [44]. Here, the PL intensities are stronger for the PbI2-rich perovskites compared with the pure-phase perovskite, which is consistent with previous reports [38,39,45]. The presence of PbI2 can passivate the grain boundaries of the perovskite films and this passivation effect on PL intensity is dominant compared with lattice strain. The sample with a MAI:PbI2 ratio of 0.5:1 (sample S2) shows the strongest PL emission, which suggests an effective suppression of non-radiative recombination. To investigate the charge carrier dynamics, TRPL spectroscopy was used to compare the carrier lifetime of the two typical perovskite samples, S2 (MAI:PbI2 ratio 0.5:1) and S4 (MAI:PbI2 ratio 1:1), as shown in Figure 2c,d. The TRPL curves can be fitted by the bi-exponential function To investigate the charge carrier dynamics, TRPL spectroscopy was used to compare the carrier lifetime of the two typical perovskite samples, S2 (MAI:PbI 2 ratio 0.5:1) and S4 (MAI:PbI 2 ratio 1:1), as shown in Figure 2c,d. The TRPL curves can be fitted by the bi-exponential function where τ 1 and τ 2 are the corresponding fast and slow time constant. A 1 and A 2 are the amplitude of the fast and slow decay component, respectively. We further calculate the contribution from the fast (A 1 /(A 1 + A 2 )) and slow (A 2 /(A 1 + A 2 )) process, and all the parameters are listed in Table 1 [46]. The defect states or shallow trapping levels in the grain boundaries of the perovskite will result in PL emission quenching and therefore behave as non-radiative recombination centers [47]. The observed longer PL lifetime in sample S2 suggests that the grain boundaries of MAPbI 3 have been effectively passivated by PbI 2 [41]. The defect states or shallow trapping levels in the grain boundaries of the S2 sample are passivated by PbI 2 in the film; hence, the S2 sample shows a high PL intensity, which is consistent with the steady-state PL results in Figure 2b. The spatially resolved photo response study of perovskite thin films by PC-AFM provides reliable and location-dependent electrical analysis of photosensitive materials [48]. The PC-AFM measurements are conducted using a Pt-Ir-coated conductive AFM tip. Bias voltages are applied on the conductive ITO substrate and the Pt-Ir tip is held at ground [49]. For the PC-AFM technique, topography and photocurrent of MAPbI 3 can be recorded simultaneously. Therefore, it is possible to investigate and compare the charge transport process from different regions under exactly the same experimental conditions. The photocurrent measurement combined with topographic images provides a much more accurate way to investigate local defects (especially at the edges) and spatial fluctuation on local charge distributions. The topographies of the two samples are exhibited in Figure 3a,c. Consistent with observations from SEM images, the PbI 2 -rich perovskite thin film had a smaller grain size and smoother surface with an average roughness of~4.93 nm, while that of the pure-phase perovskite film is~10.5 nm. The scanning current images as shown in Figure 3b,d reflect the local optoelectronic and electrical properties at a nanoscale spatial resolution. To directly compare the photocurrent generation in the pure MAPbI 3 and MAPbI 3 -PbI 2 , we measured the current value change with and without light illumination as shown in Figure 3b,d, which displays the corresponding current mapping in the dark and under 532 nm laser illumination for samples S4 (MAI:PbI 2 ratio 1:1) and S2 (MAI:PbI 2 ratio 0.5:1). For the two samples, the current intensities under laser illumination are higher than those in dark. By contrast, the photocurrent mapping for S2 (MAI:PbI 2 ratio 0.5:1) is larger than that of S4 (MAI:PbI 2 ratio 1:1), which could be due to the grain boundary passivation effect of PbI 2 . When the grain boundary is passivated by PbI 2 , the defects at grain boundaries can be reduced, resulting in more photogenerated charge carriers transferring along the grain boundaries. Therefore, the photocurrent at the grain boundary becomes further enhanced compared with the center of single-crystalline perovskite grain (as shown in Figure 3d). In addition, the current photo response at 1 V is extracted as shown in Figure 3e, which clearly demonstrates a higher current density in the PbI 2 -rich perovskite thin film. Nanomaterials 2019, 9, x FOR PEER REVIEW 7 of 13  To demonstrate the effect of the chemical composition on the performance of photodetectors, perovskite-based photodetectors with a lateral device structure of ITO/perovskite/ITO were fabricated. Figure 4a schematically shows the stepwise procedures for device fabrication. The IV curves of the devices measured under the white light with an illumination intensity of 4.5 mW/cm 2 are shown in Figure 4b. It can be clearly observed that all the devices exhibit a linear IV curve, which suggests good Ohmic contact between the ITO and perovskite thin films. Among all devices, the photodetector fabricated from the precursor solution with a MAI:PbI 2 ratio of 0.5:1 (sample S2) exhibited the highest photocurrent. The dark IV curves of all devices are compared as shown in Figure S1. The addition of PbI 2 increased the resistance of the perovskite films and reduced the dark current of the devices. The pure-phase perovskite device showed the highest dark current. The time-dependent photocurrent curves of the devices are shown in Figure 4c. For all the devices, upon photo illumination, the photocurrents increased rapidly. The on/off period was set to 2.5 s and all devices kept a repeatable photocurrent level and response speed after several on/off switching cycles, indicating the excellent reproducibility and stability of the devices. The device fabricated from precursor solution with a MAI:PbI 2 ratio of 0.5:1 shows the highest photocurrent, which is consistent with the results of the IV measurements.  Responsivity (R) is an important parameter to assess the performance of a photodetector. R can be calculated using the following equation: where I p , I d , P, and A are the photocurrent, dark current, incident light intensity, and active area of the device, respectively [49]. The responsivities of all devices at a 5 V bias voltage are summarized in Table 2. The S2 device displays the highest responsivity of 64 mA/W, much higher than the rest of devices. The excellent R value suggests that S2 (MAI:PbI 2 ratio 0.5:1) possesses a high photon-to-electron/hole conversion efficiency. The electrical measurements indicate that an appropriate excessive amount of PbI 2 in MAPbI 3 is beneficial for improving the photoelectrical performance. Response time is another important parameter for a photodetector. The rise edges and decay edges in the time-dependent photo response curve were measured under white light illumination (4.5 mW/cm 2 ) at 5 V and enlarged as shown in Figure S2. The rise times (τ r , defined as the time for the device increasing to 90% of the photocurrent maximum) and decay times (τ d , defined as the time for the device decaying to 10% of the photocurrent) are summarized in Table 2. It can be clearly seen that devices fabricated from perovskite-PbI 2 thin films show a shorter response time than the device fabricated from pure-phase perovskite (sample S4). The S2 (MAI:PbI 2 ratio 0.5:1) device exhibited an ultrafast response speed with a rise time of 0.86 ms and a decay time of 1.5 ms, which are outstanding values among the reported lateral-structured photodetectors and even heterostructure devices [11]. The spectral selectivity of the S2 (MAI:PbI 2 ratio 0.5:1) photoresistor was also studied. where h is Plank's constant, c is the velocity of the incident light, R λ is the responsivity, e is the electronic charge, and λ is the wavelength of the incident light [50]. The maximum EQE of 400% is achieved under 405 nm illumination. The photoconduction results from the electron-hole pairs excited by the illumination light with photon energy larger than the semiconductor band gap. The 405 nm light with larger energy can excite more electron-hole pairs in perovskite thin films, leading to a higher photocurrent.   Figure 5b shows the log-scale dark and light IV curves of the S2 (MAI:PbI2 ratio 0.5:1) photoresistor under white light with illumination intensities varying from 0.14 to 4.46 mW/cm 2 . The photocurrent increased with ascending light intensity. Specific detectivity (D*) is another important parameter to assess the performance of a photodetector and it can be calculated using Equation (4): ( ) Moreover, the responsivities and specific detectivities of the S2 device under different light intensities are calculated and plotted in Figure 5c. It can be seen that both the responsivity and specific detectivity decrease with increasing light intensity. Remarkably, the device shows a highest  Figure 5b shows the log-scale dark and light IV curves of the S2 (MAI:PbI 2 ratio 0.5:1) photoresistor under white light with illumination intensities varying from 0.14 to 4.46 mW/cm 2 . The photocurrent increased with ascending light intensity. Specific detectivity (D*) is another important parameter to assess the performance of a photodetector and it can be calculated using Equation (4): Moreover, the responsivities and specific detectivities of the S2 device under different light intensities are calculated and plotted in Figure 5c. It can be seen that both the responsivity and specific detectivity decrease with increasing light intensity. Remarkably, the device shows a highest responsivity value of 7.8 A/W and specific detectivity of 2.1 × 10 13 Jones under 405 nm light with an intensity of 5 µW/cm 2 , which is higher than the commercial Si-based photodetector (4.2 × 10 12 Jones) [16]. These excellent photo response performance values are some of the best results compared with previously reported MAPbI 3 -based photodetectors [11,[51][52][53][54][55], as shown in Table 3. It is noticeable that the responsivity and specific detectivity decreased rapidly when light intensity was below 0.5 mW/cm 2 , while a further increase in intensity would result in a very slow decrease until saturation is reached. This saturation is observed in many other studies [50] and has been attributed to the filling mechanism of the sensitizing centers of the perovskite thin film at a high light intensity. The saturation value is close to Liang's report (0.6 mW/cm 2 ) [56]. Figure 5d gives the EQE of the S2 device at 5 V bias with different incident light intensity. The EQE reaches the highest value of 2.37 × 10 3 % under 5 µW/cm 2 405 nm light illumination. The EQE displays nearly the same trend as responsivity and specific detectivity. In the end, we measured the I-t curves of the pristine S2 device under a strong intensity of 224 mW/cm 2 to study its stability, as shown in Figure S3. The device shows a good repeatability after several on/off cycles. Then, the I-t curve of the same device was measured again after one month, as shown in Figure S3. A high-level photocurrent remained with a~15% decay compared with the pristine device. In contrast, the pure perovskite devices all decomposed. This result demonstrates that the incorporation of PbI 2 in perovskite film is beneficial to improve the stability of the device to some extent.

Conclusions
In summary, a high-performance lateral-structured photodetector, photoresistor, based on perovskite film with a high PbI 2 doping level was successfully prepared by simple spin coating method. The photodetector fabricated from precursor solution with the MAI:PbI 2 ratio of 0.5:1 shows the best device performance, which is higher than the device based on pure perovskite thin film. The device displays a fast response speed with a rise time of 0.86 ms and a decay time of 1.5 ms. Under 405 nm light illumination, a high responsivity of 7.8 A/W, specific detectivity of 2.1 × 10 13 Jones, and an EQE of 2.5 × 10 3 % were realized. The analysis of local electrical properties demonstrated that the presence of the excess PbI 2 in the perovskite film improved its photo-electrical properties. The presence of PbI 2 can effectively passivate the grain boundaries of MAPbI 3 , resulting in a longer PL lifetime and a higher carriers transport efficiency. Furthermore, the device based on PbI 2 -rich perovskite also showed better stability compared with pure-phase perovskite devices. These results indicate that non-stoichiometric perovskites could be beneficial to device performance and chemical stability, which could also be extended to the exploration of lead-free perovskites for the application in high-performance photodetection.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/9/4/505/s1, Figure S1: The dark I-V curves of all the devices fabricated from perovskite films with different MAI: PbI 2 ratio; Figure S2: The enlarged rise edges and decay edges of all the devices under white LED light illumination of 4.5 mW/cm 2 at 5 V bias; Figure S3: the I-t curves of the device measured as pristine and after a month under 100 mW/cm 2 light illumination.