Improved Photovoltaic Performance of CsPbI2Br Perovskite Films via Bivalent Metal Chloride Doping

Cesium-based all-inorganic perovskite absorbers have attracted increasing attention due to their superior thermal stability, compared to their organic–inorganic counterparts. Up to now, it is a challenge to fabricate high-efficiency all-inorganic perovskite solar cells (PSCs) with low defect densities. Herein, we used bivalent metal chloride salts (SrCl2 and NiCl2) to optimize CsPbI2Br films. The experimental results indicate that this method could deliver high-quality films with improved electronic property. As a result, the champion device based on the 0.01 M SrCl2–doped CsPbI2Br film achieved a power conversion efficiency (PCE) of 16.07% with a high open voltage (V OC) of 1,322 mV, which is about 18% higher than that of the pristine device.

Partial substitution of iodine for bromine anions can effectively stabilize the perovskite structure by tailoring the tolerant factor (Rehman et al., 2017;Fu et al., 2019;Dong et al., 2020). A paradigm is the CsPbI 2 Br perovskite with a bandgap of ∼1.9 eV and a stable cubic phase at room temperature (Sutton et al., 2016;Yan et al., 2018). However, the PCEs of CsPbI 2 Br solar cells were still much lower than those of CsPbI 3 devices. Controlling the deposition process of perovskite thin films could optimize film coverage at a long-range scale, as well as reducing the atomic crystal defects Duan et al., 2020;Liu et al., 2020). Processing parameters, such as coating speed and annealing temperature, have been initially studied in previous reports (Yu and Gao, 2017;Zhang et al., 2020a). Besides, additives can selectively adsorb on the perovskite surface and regulate perovskite crystallization . A number of organic additives, such as dimethylammonium iodide, choline iodine, and dithiocarbamate, have been used to achieve high-quality perovskite films (Yang et al., 2019c;Fu et al., 2020;He et al., 2020). However, inorganic salts have been less investigated in this field, particularly for optimizing the film morphology and electronic property of all-inorganic perovskites (Wang et al., 2019c;Zhang et al., 2019). Bivalent metal salts are commonly used as inorganic additives to improve films' quality, which is suitable for different organic-inorganic perovskite formulations (Aydin et al., 2019;Wang et al., 2019a). ZnI 2 (Shai et al., 2018) and SnI 2 (Eperon and Ginger, 2017) have been incorporated into organic-inorganic perovskites, which could enhance PCEs of the corresponding PSCs attributed to increased grain size and crystallinity. In addition, the Cl − ion could passivate grain boundaries and trap defects of perovskite polycrystalline films (Aydin et al., 2019).
Here, we studied the effect of bivalent metal chloride salts (SrCl 2 and NiCl 2 ) on the film quality of the CsPbI 2 Br perovskite. We revealed that certain amount of metal chloride salts is substantially beneficial for the formation of the perovskite with less trap sites, while excess metal chloride salts are detrimental to the perovskite films. By utilizing SrCl 2 dopants, the best planar heterojunction CsPbI 2 Br perovskite solar cell achieved a high PCE of 16.07%, with ∼18% increment than the pristine one.

Device Fabrication
Fluorine-doped tin oxide (FTO) glasses (Nippon Sheet Glass, 8 Ω/ square) were cleaned with soap, acetone, ethanol, and deionized water sequentially for 15 min, respectively. Then, the cleaned FTO glass substrates were treated by a plasma cleaner with O 2 gas for 3 min at middle power radio frequency (10.2 W). Then, the substrates were immersed in a 25 mM TiCl 4 aqueous solution for 60 min at 70°C and washed with distilled water and ethanol, followed by annealing at 500°C for 60 min in a muffle oven to form a compact TiO 2 blocking layer. The preparation of CsPbI 2 Br films was based on a modified method according to the reported article . P3HT solution (15 mg/ml in chlorobenzene) as a hole transport layer (HTL) was spin-coated on perovskite films at 2,500 rpm for 30 s, followed by annealing at 120°C for 10 min. Finally, 100 nm-thick Ag was thermally evaporated onto the HTL as a metal electrode. All steps were carried out in a nitrogen glove box (H 2 O < 0.01 ppm, O 2 < 0.01 ppm).

Characterization
Field emission scanning electron microscopy (FESEM, HITACHI S4800) was used to characterize the morphology of films. X-ray diffraction (XRD) patterns were acquired by powder X-ray diffraction (PXRD, Bruker Advance D8 X-ray diffractometer Cu Kα radiation, 40 kV). Ultraviolet and visible (UV-Vis) absorption spectrometry was collected using a Cary 500 UV-Vis-NIR spectrophotometer. Photoluminescence (PL) spectra were characterized at room temperature by exciting the sample deposited on the FTO substrate. The excitation wavelength of the Fluorolog-3-p spectrophotometer is 380 nm. The solar cells were illuminated by a solar light simulator (Solar IV-150A, Zolix), and light intensity was calibrated by a standard Newport calibrated KG5-filtered Si reference cell. Current density-voltage (J-V) curves of solar cells were measured by a Keithley 2400 digital source meter under simulated AM 1.5G illumination of 100 mW cm −2 with a reverse scan rate of 0.15 V s −1 . Solar cells were masked with a metal aperture to define the active area of 0.0625 cm 2 . External quantum efficiency (EQE) spectra of solar cells were measured by an SCS10-X150-HBSD system. The electrochemical impedance spectra (EIS) were acquired under an applied voltage of 0-1 V with frequency ranging from 1 Hz to 1 MHz in dark condition. In transient photovoltage (TPV) decay tests, the cells were connected directly in series with an oscilloscope (DSOX3104T), and the input impedance of DSO was 10 6 Ω. An attenuated green laser pulse (DSP-532-A Laser) was used as a small perturbation to the background illumination on the device. The laser-pulse-induced photovoltage variation and the V OC are produced by the background illumination. The wavelength of the laser was 532 nm, the repeating frequency was about 10 Hz, and the pulse width was less than 10 ns.

RESULTS AND DISCUSSION
The deposition of CsPbI 2 Br perovskites is based on the spincoating of perovskite precursors dissolved in DMSO solution. We firstly included SrCl 2 into perovskite precursor solution in the concentrations of 0.005, 0.01, and 0.05 M, respectively. The morphological characteristics of CsPbI 2 Br films were firstly visualized by SEM characterizations. As shown in Figure 1A, the pristine film presents a non-flat surface with a high coverage. By the addition of 0.005-0.01 M SrCl 2 , the CsPbI 2 Br films retained the fully covered morphology (Figures 1B,C). For the film with 0.05 M SrCl 2 ( Figure 1D), a rough surface with many pinholes was observed, implying the adverse effect of excess SrCl 2 on film formation.
We also noted that the 0.05 M SrCl 2 -doped films degraded rapidly in ambient condition. XRD patterns in Figure 2A disclose the nonperovskite phase of the doped film with 0.05 M SrCl 2 . Although the ionic radius of Sr 2+ (118 pm) is similar to that of Pb 2+ (119 pm) (Shannon, 1976), we still infer that excess metal dopants would favor the formation of non-perovskite phases (Phung et al., 2020). An intriguing phenomenon is that the full width at half maximum (FWHM) of the (100) diffraction peak reduces from 0.161 of pristine to 0.139 of film with 0.01 M SrCl 2 . The narrowed XRD diffraction peak signifies the enhanced crystallinity with less crystal defects . Therefore, we conclude that the SrCl 2 additive is a double-edged sword for the perovskite: low concentration of SrCl 2 can promote the formation, while high concentration can generate the Frontiers in Energy Research | www.frontiersin.org May 2021 | Volume 9 | Article 692059 non-perovskite structure. The identical trend was also observed in the UV-Vis spectra in Figure 2B.
To further examine the effect of bivalent metal chlorides on CsPbI 2 Br perovskite films, we used NiCl 2 as an additive into perovskite films. The concentration of NiCl 2 is 0.01 M for all samples, which corresponds to 0.83% of the molar ratio with Pb ions. As expected, the NiCl 2 -doped sample exhibits smooth and uniform morphology (Figure 3). XRD patterns in Figure 4A further ascertain the cubic perovskite phase of all samples. Absorbance offsets of films were determined to be ∼650 nm by UV-Vis spectra in Figure 4B. Steady-state PL spectra were operated to assess the photoelectrical property of perovskite films deposited on glass substrates. As shown in Figure 4C, the PL intensities of SrCl 2 -and NiCl 2 -doped perovskite films are 4.46 and 2.72 times higher than that of the pristine film, respectively, indicating mitigated non-radiative recombination in bivalent metal chloride-doped films (Han et al., 2016;You et al., 2018).
We then measured the J-V characteristics of as-prepared solar cells. Solar cells were fabricated with a planar n-i-p configuration of FTO/compact-TiO 2 (c-TiO 2 )/CsPbI 2 Br/P3HT/Au ( Figure 5A). All J-V characteristics were measured under simulated AM 1.5G illumination with a reverse scan rate of 0.15 V s −1 . As shown in Figure 5B and Table 1 Figure 5C and Table 2). Notably, negligible hysteresis was observed for this champion device. In contrast, the best pristine sample exhibited a J SC of 15.69 mA cm −2 , a V OC of 1,217 mV, a FF of 0.73, and a PCE of 13.66% under reverse scan condition (Supplementary Figure S1 and Supplementary Table S1). We then held these devices under maximum power point (MPP) to measure the stabilized power output, as shown in Figure 5D and Supplementary Figure S2. The current density at the MPP was 14.39 mA cm −2 at a bias voltage of 1,100 mV, corresponding to an efficiency of 15.83%. The current density at the MPP was 13.47 mA cm −2 at a bias voltage of 960 mV, corresponding to an efficiency of 12.93%. Supplementary Figure S3 exhibits the external quantum efficiency (EQE) spectra of PSCs. The integration of the EQE spectrum yields the photocurrent of 15.37 mA cm −2 for the 0.01 M SrCl 2 -doped device and 15.24 mA cm −2 for the pristine device, which are consistent with the corresponding J-V measurements shown in Figure 5C and Supplementary Figure S1.
A statistical analysis of 20 individual devices reveals the good productivity of SrCl 2 -doped devices such that the mean PCEs of pristine and SrCl 2 -doped devices are 14.66 and 12.96%, respectively ( Figure 5E). Moreover, inclusion of NiCl 2 also improved the device performance significantly ( Figure 5F and Supplementary Table S2). The V OC of the NiCl 2 -doped device increased to 1,245 mV, yielding a PCE of 13.83%. The results presented herein manifest our strategy as versatile and universal with a large variety of possible metal salts for perovskite devices.
We then studied the optical and electrical properties of the doped perovskite devices to probe the origin for performance enhancement. EIS spectra were operated to characterize the carrier behaviors of as-fabricated CsPbI 2 Br devices, and the equivalent circuit model Zhang et al., 2020b) is displayed in Figure 6A. The charge recombination resistance (R rec ) of the SrCl 2 -doped device is much higher than that of the pristine device at a bias of 0.8 V ( Figure 6A). The R rec values of the SrCl 2 -doped device are about one time larger than R rec values of the pristine one under the bias range of 0-1 V, which confirms the longer carrier lifetime of the SrCl 2 -doped device ( Figure 6B) (Quilettes et al., 2015). We further investigated charge-transport properties of PSCs with and without SrCl 2 by using TPV measurements. The TPV curve revealed that incorporation of SrCl 2 increased the charge-carrier lifetime ( Figure 6C), indicating a decrease in the undesired charge-carrier recombination (Seitkhan et al., 2020). As shown in Supplementary Figure S4, the ideality factor (m) was measured to evaluate the recombination process in PSCs. The value of m for SrCl 2 -doped devices (1.31 kT/e) is smaller than 1.74 kT/e of the pristine device, illustrating the reduced trap-assisted charge recombination, where K is the Boltzmann constant, T is an absolute temperature of 300 K, and e is the elementary charge (Aydin et al., 2019;Dong et al., 2021).

CONCLUSION
In summary, we report a simple strategy to modulate the CsPbI 2 Br perovskite by using bivalent metal chloride additives, which facilitates the formation of high-quality perovskite films. A high PCE of 16.07% was achieved on solar cells by SrCl 2 doping, which could eliminate defective trap states and extend carrier lifetime. Our strategy would provide new insights into controllably fabricating high-efficiency allinorganic perovskite solar cells.