High-Performance Inverted Perovskite Solar Cells with Mesoporous NiO x Hole Transport Layer by Electrochemical Deposition.

Perovskite solar cells (PSCs) based on a NiO x hole transport layer (HTL) with an inverted p-i-n configuration have yielded highly efficient and relatively stable devices. Here, we develop a simple electrochemical deposition method for quickly and evenly preparing a mesoporous NiO x film. It is demonstrated that the increasing thickness and decreasing surface roughness of the NiO x film are beneficial for light transmission. The optimal condition for preparing NiO x films is achieved by adjusting the deposition time at a certain applied current density, which exhibits excellent optical transmittance and suitable thickness and band gap, thus reducing optical loss and enhancing hole extraction at the interface between HTL and the perovskite layer and therefore improving photovoltaic performances. The finite-difference time-domain simulation confirms the optimal thickness of the NiO x layer and coincides with our experiment results. An optimal power conversion efficiency (PCE) of 17.77% with an active area of 0.25 cm2 is achieved. The prepared device shows negligible hysteresis, high reproducibility, and high uniformity with a PCE difference of 2% for measuring the different sites from edge to center. This simple fabrication process paves a novel way to the evolution of PSCs based on NiO x and rapid commercialization.

Introduction

The organometal halide perovskite solar cells (PSCs) have attracted much attention due to their optimal band gap (≈1.5 eV),[1] large charge carrier mobility (33 and 115 cm2 V–1 s–1 for microcrystalline thin films and single crystal, respectively),[2−4] long electron–hole diffusion length (>1 μm for microcrystalline films and >175 μm for single crystal),[5,6] long carrier lifetimes (>250 ns),[7] and low exciton-binding energy (≈2 meV).[8] Besides, the simple fabrication process and low cost make it suitable for commercial applications competing with silicon-based solar cells.[9,10] Although its initial power conversion efficiency (PCE) was very low (4%) in 2009,[11] the rapid increase astonished researchers after only a decade’s development, with the highest PCE incredibly reaching 23.3% in 2018.[12] The inverted planar heterojunction PSCs with a p–i–n configuration have been demonstrated to show excellent performances, such as simple fabrication process, low hysteresis, and high stability.[13−16]

Recently, inorganic metal oxides performing p-type semiconductor property, such as Cu2O,[17] CoO,[18] NiO,[19] and MoO3,[20] have been studied as a hole transport layer (HTL) to replace poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) and its modifications because these organic materials are always expensive and show high instability to air, moisture, and temperature.[21,22] Among them, nickel oxide (NiO) exhibits great potential due to its wide band gap for light penetration, appropriate valence band edge exactly aligning with that of perovskite to allow for efficient hole transport, and high carrier mobility for rapid extraction and transport of carriers,[21,23] thus leading to high open circuit voltage (VOC). Besides, NiO is a low-cost material with superior thermal and chemical stabilities due to its inorganic nature, which brings about a longer device lifetime.[24−26] To date, research on PSCs using NiO as HTL have made great progress, and NiO can be prepared by various methods (Table S1). Spin-coating is the most commonly used method because of the simple operation approach.[27] For example, Yang et al.[23] fabricated NiO-based PSCs and achieved uncertified maximum PCE value of 16.1% by synthesizing NiO precursor solution and then spin-coating and postannealing. However, the solution-derived spin-coating method is time consuming,[28,29] relatively complicated for preparing the NiO precursor, and needs very high temperature (>400 °C) for annealing after spin-coating.[30] Furthermore, the spin-coating process results in great waste of precursor solution, and it is not easy to uniformly cover the transparent conductive substrate unless the NiO layer is thick enough.[31] In this case, the high resistance between the perovskite layer and the transparent electrode could block the transportation of carriers, negatively affecting the performance of PSCs.[32] Other methods, such as magnetron sputtering,[33] pulse laser deposition,[34] and atomic-layer deposition (ALD),[35] are wasteful of raw materials, time consuming, costly to process, and not suitable for large-scale fabrication.

Electrochemical deposition (ECD) is always used to fabricate the covering layer for macroscopic components and micro devices. Compared with the spin-coating method, the thickness of the NiO film prepared by the ECD method can be easily controlled by regulating the deposition time and applied current density. Besides, a uniform film with a large area can be obtained, thus significantly reducing the possibility of direct contact between the perovskite layer and the transparent conductive layer. In addition, this method is environment-friendly and suitable for mass production. This treatment does not lead to the production of waste and significantly reduces fabrication costs. Kim et al.[36] showed that using the ECD method for preparing NiO films resulted in much improved uniformity for devices and realized PCE value of 17.0% for large area (∼1.084 cm2) and 19.2% for small area (∼0.1 cm2) after introducing an extremely thin polyethylenimine ethoxylated layer. However, several fundamental questions remain unanswered. For example, it is not clear whether the optical transmittance will be influenced by morphology of the NiO film after long-time deposition, which will limit the light absorption for the perovskite layer. Moreover, there has been no investigation on the effect of thickness and surface roughness of the electrochemically deposited NiO film on the role of grain size, grain boundary, and the interface quality of perovskite thin films grown on top of NiO HTL, which are critical for both performance and stability.[37]

In this paper, we focus on the morphology effect of the mesoporous NiO film on the performances of PSCs with a structure of fluorine-doped tin oxide (FTO)/NiO/MAPbI3/phenyl-C61-butyric acid methyl ester (PCBM)/bathocuproine (BCP)/Ag, in which the NiO film is prepared by using the ECD method, and its thickness was accurately controlled by adjusting deposition time at constant current density. We report that the optical transmittance of the NiO film abnormally increases in the visible range with increasing time during the ECD process. A detailed research on the effect of the thickness and surface roughness of the NiO film on photovoltaic performance of fabricated PSCs is conducted. By optimizing the ECD process and the interfaces between each layer, a promising PCE of 17.77% at an active area of 0.25 cm2 without obvious hysteresis is achieved, which is comparable to the best-in-class PSCs using NiO HTL prepared by spin-coating,[38] ALD,[35] or sputtering.[39] The fabricated PSCs exhibit high uniformity with a small difference of 2% when measuring three different positions of a typical device. Further simulation based on finite-difference time-domain (FDTD) analysis gives reasonable results and verifies that the thickness of the NiO film has effect on PSC performances, which conforms to our experimental conclusion. This work provides a new angle of view and a meaningful reference to fabricate high-efficiency inverted PSCs based on electrochemically deposited NiO films.

Results and Discussion

The p-type NiO layer was deposited on the cleaned FTO/glass substrate via applying constant current density of 0.1 mA cm–2 and subsequent annealing at 300 °C for 2 h, as illustrated in Figure . In step 1, the Ni(OH)2 film was first deposited on the FTO substrate; the electrochemical reactions on the FTO surface could be expressed by eqs and 2. In step 2, the device was transferred into a muffle furnace for thermal annealing, and this process could be explained by eq , and finally the mesoporous NiO film was obtained. The real sample pictures and the corresponding scanning electron microscopy (SEM) images can also be seen from Figure , which indicates the formation of the NiO film from an initially semitransparent surface to a finally transparent surface through the thermal annealing process.The morphology of the prepared NiO film was examined by SEM and atomic force microscopy (AFM). As can be seen from the top-view SEM images of Figure , the FTO nanograins are totally covered by the mesoporous NiO film, and they become much denser with increasing deposition time. As shown in the insets of Figure a–d, the measured average thicknesses of the NiO film according to the cross-sectional SEM images are 18, 54, 71, and 87 nm for different deposition times of 60, 90, 120, and 150 s, respectively. Note that the morphology of the NiO film is affected by the FTO underneath due to its rough pristine surface, and the deposited NiO film seems smoother through longer deposition time according to the cross-sectional SEM images. The NiO film prepared through the ECD method exhibits very high uniformity; the SEM images are shown in Figure S2. The mesoporous morphology of the NiO film is beneficial for filling of perovskite into the nanoporous pore and conformal surface coverage during its deposition process,[40] thus preventing direct contact between the electron transport layer (PCBM) and HTL (NiO). Besides, the thin overlayer capping of perovskite on the mesoporous NiO film is advantageous for acquiring higher VOC.

Figure 1

Schematic illustration for preparing the NiO film on FTO/glass substrate.

Figure 2

Top-view SEM images of FTO/NiO layers obtained by electrochemical deposition at 0.1 mA cm–2 with (a) 60 s, (b) 90 s, (c) 120 s, and (d) 150 s. The insets are the corresponding cross-sectional images.

Figure a–e illustrates three-dimensional (3D) AFM images of the pristine FTO/glass substrate and NiO films. Note that the surface of pristine FTO is quite rugged with a root-mean-square (RMS) roughness of 14.3 nm (Figure a). Interestingly, as shown in Figure b–e, the RMS roughness decreases after electrochemically depositing the NiO film on the FTO/glass substrate; more remarkable, it decreases almost linearly with deposition time. This phenomenon is consistent with the result mentioned above in Figure . Figure f summarizes the dependence of NiO thin film’s thickness and roughness on deposition time according to Figures a–d and 3b–e. The thickness of the deposited NiO film almost linearly increases with deposition time, whereas the surface roughness unusually decreases. This could be ascribed to the filling effect, where the pits on the surface of pristine FTO are gradually filled by the deposited thin NiO film.

Figure 3

AFM 3D images of (a) pristine FTO substrate and NiO films electrochemically deposited on the FTO substrate by applying constant current density at 0.1 mA cm–2 for (b) 60 s, (c) 90 s, (d) 120 s, and (e) 150 s. The RMS roughnesses are 14.3, 13.8, 12.9, 10.9, and 9.7 nm, respectively. (f) Dependence of the NiO thin film’s thickness and roughness on deposition time.

The optical properties of the deposited NiO films with different deposition times were determined with UV–vis absorption spectra, as shown in Figure a. As a reference, the optical transmittance of the bare FTO/glass substrate was also measured. Both bare FTO and NiO films show high transmittance (>80%) over the entire visible wavelength range. Further, the transmittance of the FTO/glass substrate with the deposited NiO film (labeled as FTO/NiO) is larger than that of the bare FTO/glass substrate (labeled as FTO), and the transmittance dramatically increases with increasing deposition time. This result seems totally different from the previous reports showing negligible effect on the optical properties of the NiO layer spin-coated on the transparent conductive substrate.[30,35,41] This can be explained by the fact that the NiO films reported in previous studies were relatively compact because of the preparation method such as ALD, magnetic sputtering, spin-coating, and so on; therefore, the light was hindered when passing through the thicker NiO film. In our experiment, the thin NiO film prepared by the ECD method is mesoporous, and the incident light cannot be effectively reflected; meanwhile, the thicker mesoporous NiO film with increasing deposition time reduces surface roughness as mentioned above, thus decreasing the possibility of incident light being trapped in the pits of the substrate. As a result, the transmittance gradually enhances with the increasing thickness of NiO HTL. This is highly desirable for photovoltaic devices as it generally leads to enhanced light-harvesting capability of the perovskite layer.

Figure 4

(a) Transmission spectra and (b) X-ray diffraction (XRD) patterns of pristine FTO/glass substrate and electrochemically deposited NiO films as hole transport layers. (c) X-ray photoelectron spectroscopy (XPS) spectra of the Ni 2p3/2 core level of the prepared NiO film on the FTO/glass substrate. (d) Tauc plot for the NiO film electrochemically deposited on FTO/glass substrate, showing that the band gap of NiO is 3.66 eV.

Figure b shows XRD patterns of the pristine FTO/glass substrate and electrochemically deposited NiO film (50 nm in thickness) on the substrate. As displayed, the XRD pattern of the NiO film exhibits weak peaks compared with the pristine FTO, whereas the small diffraction peaks at 37.2 and 43.3° match very well with that of the cubic Fm3m crystal structure of NiO (PDF# 47-1049), which can be assigned to the (111) and (200) planes. Figure c displays XPS spectra of the Ni 2p3/2 core level, and they are consistent with the previous reports.[35,38] Three distinct peaks can be separated from the XPS spectrum: the peak at 853.8 eV is ascribed to Ni2+, indicating the existence of NiO in the prepared film; the peak at 855.5 eV should be ascribed to Ni3+, indicating the existence of Ni2O3; and the broad peak at 860.9 eV is ascribed to the shakeup process of the NiO structure according to the literature.[23,42] Therefore, the prepared film in this paper is noted as NiO rather than barely NiO or Ni2O3. Furthermore, we have derived the optical band gap of the samples from the Tauc plots in Figure d, in which (αhν)2 is plotted as a function of hν from the absorption spectra, with α, h, and ν representing the absorption coefficient, Planck constant, and light frequency, respectively.[43] The calculated optical band gap of the NiO film is 3.66 eV, which corresponds closely to the value reported in previous literature.[32,35,44]

Figure a–e exhibits the top view of SEM images of the prepared CH3NH3PbI3 (MAPbI3) films deposited on NiO mesoporous films fabricated by the ECD method with different deposition times from 0 to 150 s, indicating that the compact perovskite completely covers the mesoporous NiO film underneath without visible pinholes. The deposition time of 0 s means that the perovskite layer was directly coated on the FTO/glass substrate without NiO HTL. Even though the difference of average grain size is not significant for different deposition times from the statistic distribution diagram of perovskite grain size, illustrated in Figure f, it is evident that the largest average grain size of about 320 nm is obtained at the deposition time of 90 s for the prepared NiO film. The improved crystallization of the perovskite layer based on NiO at optimal deposition condition helps reduce carrier recombination and improve the photocurrent of PSCs.[45] The schematic diagram of the device structure is shown in Figure g, which is in accordance with the cross-sectional SEM image of a completed solar cell shown in Figure h. It is clear that the light-harvester layer is composed of compact perovskite grains with a uniform stacking structure, and the perovskite layer closely contacts the hole and electron transport layers. The energy band diagram for the device structure of the prepared inverted planar PSCs based on NiO films is described in Figure i. Here PCBM and BCP work as the electron transport layer and the hole isolation layer, respectively.

Figure 5

Top-view SEM images of the perovskite film deposited on the NiO HTL prepared by the ECD method with deposition times of (a) 0 s, (b) 60 s, (c) 90 s, (d) 120 s, and (e) 150 s. (f) Statistic diagram of perovskite grain size based on the NiO film with different deposition times. (g) Structural diagram, (h) cross-sectional view SEM image, and (i) the corresponding energy band diagram of a typical device based on NiO HTL.

The photovoltaic performance of the prepared PSCs based on NiO HTL with different deposition times is illustrated in Figure a and listed in Table . The J–V curves were characterized under AM 1.5G simulated sunlight (100 mW cm–2), and the scan rate was set to 100 mV s–1. It is clearly seen that the control device without the electrochemically depositing NiO layer exhibits a PCE of only 3.64%, with a VOC of 0.705 V, a short-circuit current density (JSC) of 14.31 mA cm–2, and a fill factor (FF) of 36.07%. However, employing NiO HTL can significantly improve the performance of the PSC device, resulting in an optimal PCE of 16.18, 17.77, 16.34, and 14.68% for the NiO layer with different deposition times of 60, 90, 120, and 150 s, respectively. To examine the photocurrent difference of the fabricated solar cell based on NiO HTL and bare FTO, the incident photon-to-current efficiency (IPCE) spectra were measured, as shown in Figure b. The IPCE in the whole visible region is significantly enhanced after applying NiO HTL. The integrated current densities calculated from IPCE spectra are well consistent with the measured JSC illustrated in Figure a. For better understanding the electrical performance affected by NiO HTL, the distributions of PCE, FF, JSC, and VOC are summarized in Figure c. The statistical results are obtained from 40 devices and show relatively small error bars, indicating high uniformity of the fabricated PSCs. The average VOC of the prepared PSCs slightly increases from ∼1.00 to ∼1.01 V and then remains stable with the increasing deposition time of NiO HTL. As shown in Figure f, the surface roughness of the deposited NiO film decreases with increasing deposition time, which leads to higher shunt resistance (Rsh) and stronger ability of blocking electron transport, and eventually increases VOC.[23]Rsh of the PSCs calculated from the J–V curves actually increases from 7.5 to 13.4 kΩ cm2 with the deposition time increasing from 60 to 150 s. However, the increased thickness of NiO HTL will lead to the increase of series resistance Rs, which can largely hamper JSC and FF, for example, on increasing the deposition time from 90 to 150 s, JSC and FF reduce from 21.55 mA cm–2 and 81.34% to 20.24 mA cm–2 and 71.59%, respectively. This is probably due to the increased recombination of dissociated carriers and a higher resultant series resistance because of holes traveling a longer distance to reach the FTO electrode.[32]

Figure 6

(a) J–V curves of PSCs based on NiO HTL with different deposition times. (b) IPCE spectra of the champion device and the control device without NiO HTL. (c) Photovoltaic parameters of the perovskite devices based on hole contacts with different times for electrochemically depositing NiO HTL.

Table 1

Photovoltaic Parameters of the PSCs Based on NiO HTL at Various Deposition Times

deposition time (s)	VOC (V)	JSC (mA cm–2)	FF (%)	PCE (%)
0	0.705	14.31	36.07	3.64
60	1.001	20.53	78.77	16.18
90	1.014	21.55	81.34	17.77
120	1.005	20.42	79.63	16.34
150	1.013	20.24	71.59	14.68

The optimal PSC was prepared by electrodepositing NiO HTL for 90 s, and its electrical performance is shown in Figure a. The measurement was implemented at room temperature and humidity of 60% with scan rate of 100 mV s–1. It demonstrates a high conversion efficiency of 17.47% in forward scan with a JSC of 21.36 mA cm–2, a VOC of 1.011 V, and an FF of 80.93%; and a PCE of 17.77% in reversed scan with a JSC of 21.55 mA cm–2, a VOC of 1.014 V, and an FF of 81.34%. The PCE difference between the forward and reverse scans is much smaller than 0.1% in absolute PCE values, exhibiting negligible hysteresis. Although several possible mechanisms have been proposed to explain the phenomenon of hysteresis, including ferroelectric effect,[46] ionic displacement,[47] charge accumulation at interfaces,[48] defects in the perovskite,[49] and so on, the physical origin of hysteresis has not been clearly pointed out. In our case, the efficient hole extraction ability of NiO HTL and high quality of the perovskite film should be the two most important reasons for the negligible hysteresis behaviors of our devices due to suitable thickness of NiO and good interface contact, which are probably ascribed to better control capacity of the ECD method and the mesoporous morphology of NiO.[50] Some studies reported that the electrical property of the PSCs could be affected by changing the scan rate during the measurement process,[51] whereas in this paper, as shown in Figure S2, the J–V curves almost overlap in the whole range of scan voltage with different scan rates at 50, 100, 200, and 300 mV s–1. Output uniformity of the photovoltaic performance is essential for achieving a highly efficient large-area device. Figure b shows the J–V curves tested at different positions corresponding to the three electrodes in the lower inset. The curves almost overlap each other with a negligible discrepancy, and the detailed parameters are listed in the upper inset. The three positions exhibit very even PCE performances (17.54% for A, 17.77% for B, and 17.41% for C) with a small difference of 2%, demonstrating excellent uniformity of the prepared PSCs. Figure c illustrates the PCE distribution based on NiO HTL electrochemically deposited for 90 s with applying constant current density of 0.1 mA cm–2. The majority of PCE values are located in the range of ∼16.7%, which is consistent with the above result shown in Figure c and Table .

Figure 7

(a) Dual-direction scan of the optimal PSC device. (b) J–V responses at different positions of a typical PSC device. (c) PCE distribution histogram of NiO-based PSCs. (d) Photovoltaic parameters’ stability of the prepared device.

Stability of PSCs has become a pivotal issue for commercial application. In this regard, we further examined the long-term stability of our NiO-based PSCs. The unencapsulated PSC devices were measured at room temperature with humidity of 60% and stored in a glovebox filled with nitrogen after each measurement and characterization. The degradation trends of normalized photovoltaic parameters (VOC, JSC, FF, and PCE) are shown in Figure d. The fabricated devices demonstrate good stability over a period of 40 days and maintain 80.5% of their initial efficiency. It is noteworthy that two other parameters, JSC and VOC, do not decline too much; the maximum degradation of PCE comes from the reduction of FF due to interfacial degradation.[23,44] We also notice that the measuring pins of the test fixture stab on the perovskite absorber through pierced silver electrodes after multiple repeated measurements, which also finally causes degradation in FF and PCE.

To further validate the impact of prepared NiO thickness on the performance of the inverted p–i–n configuration perovskite solar cells theoretically, we used the computer simulation package of Lumerical’s optical simulation software of finite-difference time-domain (FDTD) solutions 2017a and electrical simulation solver of DEVICE 2015a to simulate detailed photovoltaic parameters (VOC, JSC, FF, and PCE). The simulation was based on four basic types of recombination: radiative, Auger, Shockley–Read–Hall, and surface recombination. A two-dimensional (2D) modular numerical simulation was carried out by two steps due to the device structure and software configuration features. The first step was to obtain the electron–hole pair generation rate inside the considered device based on 2D Maxwell’s equations. The standard AM 1.5G spectrum was introduced as the incident light source, and the refractive index (n, k) of each layer of the PSC structure was referenced from previous literature.[52−56]

Furthermore, the photogenerated carrier rate extracted from the optical calculation was employed for the exact electrical characteristic calculations based on drift-diffusion models, in which five layers of BCP, n-type PCBM, MAPbI3, p-type NiO, and FTO were considered. The basic parameters of materials, such as thickness, shunt, and series resistance, come from experimental data, and other electrical parameters, for instance, band gap, carrier mobility, carrier lifetime, and so on, are referenced from the literature,[54,55,57] which are all listed in Table S2. Here, Eg is band-gap energy; εr is relative permittivity; χ is electron affinity; NA and ND denote acceptor and donor densities, respectively; NC and NV are effective state density of conduction and valence bands, respectively; and μn and μp are mobilities of electron and hole, respectively. τn and τp are the lifetimes of electron and hole for trap-assisted recombination, respectively.

The simulated J–V curves shown in Figure a indicate that the performance of NiO-based PSCs is highly affected by the thickness of NiO HTL. The optimal thickness of the NiO layer is around 60 nm, as shown in Figure b; this matched well with our experimental result. The photovoltaic parameters, especially VOC and PCE from the simulation results, are much higher than those in our experiment. This is mainly because the simulation process is executed only in terms of the electrical parameters listed in Table S2, but ignoring the interface defects, material impurity, humidity, temperature, and so on. The light intensity in the interior of the prepared PSC device with a 60 nm-thick NiO layer is shown in Figure c, demonstrating that incident light is efficiently coupled into the perovskite layer by optical resonance and light scattering provided by the mesoporous NiO film. The simulation results indicate that there is much more room for improvement of NiO-based PSC performance. We believe that the PCE of NiO-based PSC can be very probably enhanced over 20% by further optimizing the preparation process of the NiO layer and reducing the contact resistance and defects at interfaces between each layer.

Figure 8

Simulation results. (a) Simulated J–V curves and (b) photovoltaic parameters of PSCs based on NiO HTL with different thicknesses. (c) Internal light intensity of the PSC device. E represents electric field intensity.

Conclusions

In summary, mesoporous NiO films with different thicknesses and surface roughness have been successfully deposited on FTO/glass substrates by the ECD process. The as-prepared NiO films exhibit high optical transmittance in the visible range with a band gap of 3.66 eV and show higher transmittance with the increasing deposition time at constant current density. The thickness and surface roughness of the NiO layer deposited with different duration times are found to affect the JSC and FF without significantly attenuating the VOC. As a result, an optimized efficient inverted planar heterojunction PSC with a structure of FTO/NiO/CH3NH3PbI3/PCBM/BCP/Ag is fabricated, exhibiting an optimal PCE of 17.77% with an active area of 0.25 cm2. The devices show high uniformity and stability without obvious hysteresis. The simulation results coincide with experimental results and prove that incident light is efficiently coupled into the light-harvester layer through the mesoporous NiO layer. This study opens a new way for preparing high-quality NiO HTL, and further investigation is necessary to improve JSC and VOC for high-performance PSCs.

Experimental Section

Materials

The patterned fluorine-doped tin oxide (FTO)-coated glass substrate was purchased from Shanghai MaterWin New Materials Co., Ltd, China. Nickel foil (300 μm in thickness, 99.99%) was purchased from Shengshida metal materials Co., Ltd, China. Nickel nitrate (Ni(NO3)2·6H2O, 99%) was acquired from InnoChem, China. Methylammonium iodide (CH3NH3I, 99.5%), bathocuproine (BCP), and phenyl-C61-butyric acid methyl ester (PCBM, 99.5%) were acquired from Xi’an Polymer Light Technology Corp. Lead iodide (PbI2, 99.9%), dimethyl sulfoxide (DMSO, 99.8%), and chlorobenzene (99.5%) were acquired from Sigma-Aldrich, and γ-butyrolactone (GBL, 99.8%) was supplied from Aladdin.

Preparation of the NiO Film

NiO film was prepared by the electrochemical deposition method and then a thermal annealing process. In brief, Ni(NO3)2·6H2O (2 mMol) was dissolved in 100 mL of deionized water with magnetic stirring until a light-green solution was obtained. The cleaned FTO/glass substrate was connected to the cathode of the electrochemical workstation, whereas the nickel foil with the same shape was connected to the anode. A thin layer of Ni(OH)2 was first obtained by applying a stable current density (0.1 mA cm–2) between the working electrode (FTO/glass) and the counter electrode (nickel foil). The thickness of Ni(OH)2 was precisely controlled by adjusting the electrodeposition time. The NiO film was finally obtained by annealing the prepared sample in a muffle furnace (Thermolyne, Thermo Scientific) at 300 °C for 2 h.

Fabrication of PSCs

The patterned FTO/glass substrate was cleaned with detergent mixed in deionized water, isopropanol, and ethanol sequentially by sonication for 15 min. The cleaned FTO/glass substrate was dried by blowing nitrogen and treated by ultraviolet-ozone for 20 min. An approximate 50 nm-thick NiO film was deposited on the FTO/glass substrate by the previously described method. Thereafter, the substrate was transferred into a glovebox filled with nitrogen for coating the perovskite layer by a one-step method. The perovskite precursor solution was prepared by dissolving 1 mMol PbI2 and 1 mMol CH3NH3I in DMSO (0.3 mL) and GBL (0.7 mL). The spin-coating process for the perovskite precursor solution was set to run at 500 rpm for 12 s and then 4000 rpm for 30 s. After spinning for 20 s of the second spinning process, 100 μL of anhydrous chlorobenzene was rapidly dropped on top of the substrate. The substrates were heated in a muffle furnace at 100 °C for 10 min. Subsequently, PCBM solution (dissolved in chlorobenzene of 20 mg mL–1) and BCP solution (0.5 mg mL–1 in isopropanol) were spun onto the perovskite layer at 2000 rpm for 30 s and 4000 rpm for 30 s, respectively. Each prepared solution was filtered through poly(tetrafluoroethylene) filters (0.45 μm) before the spin-coating process. Finally, a 120 nm-thick silver layer was deposited using thermal evaporation (PECVD350, Shenyang Xinlantian vacuum technology Co., Ltd, China) and a shadow mask. The active area of the prepared solar cell device is 0.25 cm2.

Film and PSC Characterization

Scanning electron microscopy (SEM, Carl Zeiss, Germany) and atomic force microscopy (AFM, Nanoscope IIIa Multimode) were employed to characterize the morphology of the deposited NiO film and prepared perovskite solar cells. Further, the structure of NiO HTL was confirmed via X-ray diffraction (XRD, D8 ADVANCE, Germany). X-ray photoelectron spectroscopy (XPS, AXIS UltraDLD, Japan) was used to analyze the elemental composition of the NiO film. Optical transmission spectra of the samples were collected by a UV–vis–NIR spectrophotometer (LAMBDA750, PerkinElmer). Photocurrent density–voltage (J–V) curves were measured by using Keithley 2400 SourceMeter under a standard 1 sun AM 1.5G solar simulator (Newport, 2612A), which was calibrated by a silicon reference cell under a light intensity of 100 mW cm–2. J–V curves were measured by means of reverse (from 1.2 to −0.2 V) or forward (from −0.2 to 1.2 V) scanning at room temperature with scan rate of 100 mV s–1 and relative humidity of 60%. The incident photon to current efficiency (IPCE) of the fabricated PSCs was characterized using a quantum efficiency measurement system (QEX10, PV measurements).