Chemical Bath Deposition of NiOx‐NiSO4 Heterostructured Hole Transport Layer for Perovskite Solar Cells

Herein, the scalable chemical bath deposited NiOx‐NiSO4 heterostructured films are reported as the efficient hole transport layers (HTLs) in perovskite solar cells. The NiOx‐NiSO4 films show excellent hole extraction ability and reduce interfacial charge recombination in solar cell devices. By using NiOx‐NiSO4 HTLs, a high power conversion efficiency of 20.55% is obtained, which is about 12.23% greater than that of the pure NiOx transport layer. This study provides a simple solution‐processing route toward the large‐area production and fabrication of full inorganic transport layers for perovskite photovoltaics.


Introduction
Lead halide hybrid perovskites are strong competitors for next-generation solar cells, benefiting from their broadband light absorption, [1] high absorption coefficient, [2] long carrier lifetime, [3] and feasible solution deposition. [4] Perovskite solar cells (PSCs) have advanced the efficiency apace, which from 3.8% in 2009 to 25.7% within the last few years. [5] The typical PSCs compose of perovskite absorbers, charge transport materials, and metallic contacts, in which the charge transport materials selectively extract the charge carriers and block the counter charges. [6] Recent progress in power conversion efficiency (PCE) heavily depended on the advancement of electron transport layers, such as the synthesis of n-type SnO 2 [7] or TiO 2 . [8] In contrast, the inorganic hole transport layers (HTLs) have made very limited success, whose efficiencies were generally lower than that of the organic ones.
Here, we report a NiO x -NiSO 4 heterostructured layer by a chemical bath deposition (CBD) as efficient HTLs in perovskite photovoltaics. Such method provides a low-cost solution and low-temperature process for the preparation of NiO x -based HTLs, which also gain scalable and compact films without the inherent weakness in spin-coating. The NiO x -NiSO 4 HTLs showed excellent hole extraction ability and reduced interfacial charge recombination in PSCs. As a result, perovskite devices based on NiO x -NiSO 4 HTLs achieved a high PCE of 20.55%, which was about 1.2 times higher than the solar cells based on pure NiO x HTLs.

Results and Discussion
The NiO x -NiSO 4 heterostructured layer was deposited by gradual hydrolysis of nickel chloride with the presence of thiourea and ammonia water at 75°C. After thermal annealing, the product shows uniform morphologies that comprises nanosized particles as depicted by the images of scanning electron microscopy (SEM) in Figure 1a. The grain size of NiO x -NiSO 4 film is estimated to be 10-30 nm. By prolonging the reaction time, small particles join together and form large aggregates at 60 min ( Figure S1 Table 1). The short-circuit photocurrent density ( J SC ) values of both devices also match well with the integrated photocurrent from the spectra of external quantum efficiency (EQE) (Figure 2c). The devices employing NiO x and NiO x -NiSO 4 HTLs obtain a stabilized photocurrent (21.12 and 19.09 mA cm À2 ) under simulated one sun illumination at the maximum power point (MPP), which correspond to the stable PCEs of 19.85% and 17.31%, respectively ( Figure 2d). The increased J SC can be associated with the improvement in light absorption and interfacial carrier extraction. The light absorption spectra of NiO x and NiO x -NiSO 4 films are very close ( Figure S3a, Supporting Information). However, there is an evident promotion in light absorption from MAPbI 3 film, which is deposited on NiO x -NiSO 4 HTLs. This can be interpreted as the enhanced quality of perovskite on NiO x -NiSO 4 HTLs ( Figure S3b, Supporting Information). Therefore, such HTL not only improved the quality of perovskite but also optimized the interface between NiO x -NiSO 4 HTLs and perovskite with optimized charge collection.
To scrutinize the carrier transfer at perovskite/HTLs interface, we carried out the measurements of photoluminescence (PL) and time-resolved photoluminescence (TRPL). As shown in Figure 3a, the PL intensity of MAPbI 3 film deposited on NiO x -NiSO 4 film is mildly stronger than that on NiO x film, which indicates the NiSO 4 can neutralize the interfacial defects and www.advancedsciencenews.com www.advenergysustres.com reduce the nonradiative recombination velocity. [27] In fact, the sulfate anions can coordinate with surface cationic Pb centers of perovskite to passivate corresponding trap states. [1] We also operated thermal admittance spectroscopy (TAS) measurement to prove this. The PSC with NiO x -NiSO 4 HTL has much lower trap density of state (tDOS) than the NiO x one at the range from 0.33 to 0.46 eV ( Figure S4, Supporting Information). TRPL curves of different HTLs in Figure 3b can be fit by a biexponential function, [22] which can obtain the slow and fast delay lifetimes. The fast decay lifetimes of NiO x -NiSO 4 HTLs samples (2.4 ns) were shorter than the NiO x ones (4.9 ns), revealing a shorter decay time and a faster decay rate for the perovskite films on NiO x -NiSO 4 HTLs. We also calculate the average carrier lifetimes, which indicate its good ability of carrier extraction. The average carrier lifetimes of NiO x -NiSO 4 and NiO x HTLs were 40.1 and 43.2 ns, respectively. To explain why the charge extraction was accelerated, we prepared lateral devices with structures of FTO/HTLs/Au to obtain the conductivity data of HTLs ( Figure S5, Supporting Information). We found that the bilayer of NiO x and NiSO 4 has a poor electrical conductivity than pure NiO x HTL. However, the current density of NiO x -NiSO 4 HTLs shows a rising tendency at the same forward bias, revealing higher conductivity than others. This phenomenon can be explained by the formation of NiO x -NiSO 4 heterostructured   www.advancedsciencenews.com www.advenergysustres.com HTLs. In this structure, NiSO 4 species are likely to play an important role in confining the charge carriers within NiO x network and shortening the transmission path of charge carriers. [27] The increased conductivity resulted in an improved charge extraction rate and a suppressed nonradiative recombination.
To further explore the behavior of carrier recombination in the as-fabricated PSCs, we then collected the data of dependence between incident light intensity and V OC . By fitting log-scaled light intensity versus V OC values linearly (Figure 4a), a slope of 1.837 k B T/q for the device with NiO x and a flatter one of 1.382 k B T/q for the device with NiO x -NiSO 4 are obtained, where k B is defined as the Boltzmann constant, T denotes the absolute temperature, and q is elementary charge. The slope (ideality factor, n id ) of 1 normally denotes a radiative recombination of free holes and electrons in solar cells, whereas n id ¼ 2 indicates a trap-assisted Shockley-Read-Hall recombination process. [27] The results here pinpoint the dominating role of bimolecular recombination channel in NiO x -NiSO 4 HTLs based devices. Figure 4b shows the electrochemical impedance spectra (EIS) spectra of as-fabricated devices. We found that PSCs with NiO x -NiSO 4 HTLs at the low-frequency region exhibited a larger semicircle than NiO x based devices, which is a signature of large recombination resistance (R rec ). The dramatically suppressed charge recombination was also revealed by transient photovoltage decay (TPV) measurements. As shown in Figure 4c, the average carrier lifetime is fitted to be 2.8 and 4.8 μs, respectively. The device with NiO x -NiSO 4 HTLs exhibited large recombination resistance and long carrier lifetime, which can be attributed to the sulfate anions in NiO x -NiSO 4 HTLs. The effect of defect passivation was achieved by the coordination between sulfate anions and Pb cations. Carrier lifetime was improved and carrier recombination was suppressed, which resulted in of the improved V OC .

Conclusion
In summary, a CBD method for facile, large-area deposition of efficient NiO x -NiSO 4 heterostructured HTLs has been developed. The PSCs device based on NiO x -NiSO 4 HTLs obtained a high PCE of 20.55%, which was much higher than the device based on pure NiO x HTLs (18.31%). Benefiting from the CBD process, the prepared devices exhibited enhanced stability, repeatability and should be compatible with large-area production. Our preliminary results have verified the compatibility of these HTLs with large-area solar cells ( Figure S6, Supporting Information). The device may be further increased by optimizing the morphologies of the heterostructures, such as the phase distribution, crystalline grain size, and aggregate structure. Other nickel-based oxysalts are also promising in the application of transport layers. Our study provides a new option for the large area and reliable production of inorganic HTLs for perovskitebased photovoltaics.
Preparation of HTL Films: Nickel chloride (0.714 g) and thiourea (0.516 g) were dissolved in deionized (DI, 30 mL) water. Then, ammonia (2 mL) water was added, in order to adjust the value of pH (>7) after the powder complete dissolving. The cleaned FTO substrates (cleaned by the solution of detergent, acetone, and ethanol) were immersed into the mixed solution and subsequently placed on the hot plate at 70°C with different processing time (20,40, and 60 min). The treated FTO substrates were rinsed with DI water and ethanol. Finally, the substrates were heated at different temperature (150, 300, and 500°C) for 60 min. NiO x films were prepared on the basis of the reported method. [28] Device Fabrication: PbI 2 powders (1.1986 g) were added into DMSO (185 μL) and DMF (2 mL). CH 3 NH 3 I (80 mg MAI synthesized as the reported method [29] ) powders were dissolved in IPA (2 mL). Before using, they are all stirred at 45°C for 12 h. The precursor of PbI 2 was coated on the NiO x or NiO x -NiSO 4 films (3000 rpm, 30 s), then spinning coating MAI solution on the PbI 2 films (5000 rpm, 30 s), after that the films annealing at 115°C for 10 min. The charge-transport layer (ETL) was prepared by spin-coating PC 61 BM solution (20 mg mL À1 , dissolved in CB) at 2000 rpm for 45 s and BCP solution (0.5 mg mL À1 , dissolved in ethanol) at 4000 rpm for 45 s, then heated at 70°C for 15 min. Finally, a metal electrode www.advancedsciencenews.com www.advenergysustres.com (Ag, 100 nm) was used in the method of thermal evaporation to deposit on the ETL. Characterization: Field-emission SEM characterized the morphology of hole transfer layer films and perovskite samples using Hitachi S4800. Powder XRD patterns were used to analyze the crystallographic information of hole transfer layer films by Bruker Advance D8 X-Ray diffractometer (Cu Kα radiation, 40 kV). XPS was conducted to explore the elemental valence states and composition of the hole transfer layer films using PHI5300 (Mg anode, 250 W, 14 kV). Cary 500 UV-Vis-NIR spectrophotometer was equipped to collect the absorption spectra of HTL films and MAPbI 3 films. The PL measurement was acquired at room temperature using the spectrophotometer (Ocean optics QE pro) with a light source (365 nm). TRPL spectra were obtained by the high-resolution photoluminescence spectrofluorometer.
J-V measurements were conducted under AM 1.5G irradiation (calibrate by a standard Si cell) by a Keithley 2400 digital source meter with a scan speed of 0.15 V s À1 . The active areas (0.0625 and 1 cm 2 ) of the devices were defined using the metal masked. Newport-74 125 system was used to measure the EQE. The steady-state photocurrent output was measured at MPP for 300 s. The electrochemical workstation (Parstat 2273, Princeton) was used to measure EIS under different positive bias voltages from 100 000 to 1 Hz at dark conditions. An attenuated laser pulse was used to obtain the data of TPV. I-V measurements for NiO x and NiO x -NiSO 4 films were based on the FTO/HTLs/Au structures. TAS was performed to detect the tDOS.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.