PDF A Solution-Processed Dopant-Free Tin Phthalocyanine (SnPc) Hole ... - NJU

A Solution-Processed Dopant-Free Tin Phthalocyanine (SnPc) Hole Transport Layer for E ffi cient and Stable Carbon-Based CsPbI

₂

Br Planar Perovskite Solar Cells Prepared by a Low-Temperature Process

Xiang Zhang, Naitao Gao, Yuzhu Li, Lai Xie, Xiang Yu, Xubing Lu, Xingsen Gao, Jinwei Gao,*

Lingling Shui, Sujuan Wu,* and Jun-Ming Liu

Cite This:ACS Appl. Energy Mater.2020, 3, 7832−7843 Read Online

ACCESS

Metrics & More Article Recommendations

*

^sı Supporting Information

ABSTRACT: Carbon-based inorganic CsPbX₃(X = I, Br, Cl) perovskite solar cells (PSCs) are attracting great attention in the photovoltaicfield because of their low cost, simple process, and superior thermal stability. However, the large difference in energy band between the CsPbX₃ and carbon electrode leads to the poor hole extraction and unfavorable charge recombination, thus deteriorating device’s efficiency. In this work, a solution-processed dopant-free tin phthalocyanine (SnPc) film is used as a hole-transport layer (HTL) to fabricate carbon-based CsPbI₂Br PSCs by a low-temperature process. At the optimal process, the device achieves a maximum efficiency of 11.39% with less hysteresis, which is much higher than 9.22% of reference device without the SnPc HTL. Moreover, the unencapsulated device exhibits the improved stability and remains about 90% of its initial efficiency after 30 days in ambient air (20−25°C) with 25−35% relative humidity (RH). Thisfinding reveals that the solution-processed dopant-free SnPc HTL can significantly promote charge

transport and suppress charge recombination at the CsPbI₂Br/carbon interface. This work provides a low-temperature and solution- processed method to fabricate the efficient and stable carbon-based inorganic PSCs.

KEYWORDS: low-temperature process, CsPbI₂Br perovskite solar cells, solution-processed dopant-free SnPc, carbon electrode, photovoltaic performance

■

INTRODUCTION

In the past few years, organic−inorganic hybrid perovskite solar cells (PSCs) have drawn extensive attention in view of their excellent photoelectric properties.¹⁻³ Until now, the efficiency of PSCs has skyrocketed rapidly from 3.8% to a certified value of 25.2%.^4,5 Nevertheless, they always suffer from a structural instability under thermal/moisture conditions due to the volatilization of organic components such as methylammonium (MA) and formamidinium (FA).⁶⁻¹³ To address this issue, replacing the organic ions with Cs⁺to form inorganic CsPbX₃ (X = I, Br, Cl) perovskites should be an effective strategy.^7,8,10,11Among them, CsPbI₃perovskite has a ideal band gap (E_g) of 1.73 eV, and the corresponding PSCs have achieved an efficiency of over 19%.^7,10,14,15However, its cubic α-phase is prone to degradation in an ambient atmosphere because the Cs⁺ with too small radius cannot hold [PbX₆]⁴⁻ octahedral.¹⁶ Although CsPbBr₃ perovskite shows better thermal/moisture stability, it has a largeE_gof 2.3 eV and can only absorb the light below 540 nm, resulting in the limited device efficiency.^15,17Fortunately, the mixed-halide CsPbI₂Br perovskite owns reasonable E_g of 1.91 eV and considerable phase stability in air, demonstrating the great

potential in photovoltaic applications.^6,15 Recently, the efficiency of CsPbI2Br PSCs has exceeded 16% through the crystal morphology control and interface engineering.^7,10,18,19 Despite the rapid progresses, most of the efficient CsPbI₂Br PSCs still employ the doped organic hole-transport layer (HTL) and noble metal electrode.^7,10,18,19 This considerably limits the stability of CsPbI₂Br PSCs. On the one hand, the hygroscopic bis(trifluoromethane)sulfonimide lithium salt (Li- TFSI) must be used to dope the 2,2′,7,7′-tetrakis(N,N-di-p- methoxyphenylamine)-9,9′-spirobifluorene (Spiro-MeOTAD) to increase the conductivity of HTL, which will cause the moisture-induced phase transition of CsPbI₂Br perovskite.⁸ Moreover, it is difficult to control the oxidation of Spiro- MeOTAD, resulting in the instability of CsPbI₂Br PSCs.²⁰On

Received: May 26, 2020 Accepted: July 20, 2020 Published: July 20, 2020 Downloaded via NANJING UNIV on October 10, 2020 at 11:07:05 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

the other hand, a metal ion (Au and Ag) from the electrode diffuse across the HTL into the perovskite layer can produce shunts across the device and create deep defects within the semiconductor, which accelerates the degradation of CsPbI₂Br PSCs.²¹Delightedly, the commercialized carbon electrode has drawn increasing attention because of its earth abundant, low- cost, excellent chemical stability and suitable work func- tion.²²⁻²⁴Furthermore, the carbon electrode (such as single- walled carbon nanotubes, etc.) have demonstrated their feasibility as a promising alternative to a conventional electrode.²⁴

The carbon-based CsPbI₂Br PSC with a classical con- struction of glass/FTO/compact TiO₂/CsPbI₂Br/carbon has achieved an efficiency of 10%first reported by Han’s group.²³ Subsequently, Wang et al. and co-workers obtained an efficiency of 10.21% with the same structure by antisolvent- assisted multistep deposition of the CsPbI₂Brfilm.²⁵However, the CsPbI₂Brfilms in the above two work have been fabricated by a high-temperature annealing (>280°C) process to form cubic α-phase because of the high formation energy of CsPbI₂Br perovskite.^15,26 This is not only energy- and cost- consuming but also undesirable for multijunction tandem and flexible solar cells. Zhu et al. have achieved an efficiency of 6.55% by controlling the precursor solution aging time to reduce the crystallization temperature of CsPbIBr₂film to 100

°C.²⁷ Wang et al. have tried to reduce the crystallization temperature of CsPbI₂Br film to 100 °C by incorporating nonvolatile additive, leading to an efficiency of 8.44%.²⁸ Regrettably, the efficiency of low-temperature carbon-based CsPbX₃PSCs is still much lower than that of the PSCs with organic HTL and noble metal electrode, which mainly originates from the poor hole extraction and serious charge recombination at the CsPbX₃/carbon interface resulting from the large energy level difference.²⁹ An effective strategy is to introduce HTLs at the CsPbX₃/carbon interface to suppress the charge recombination and promote the hole extrac- tion.^8,29−31In this regard, a wide number of HTLs including modified PEDOT:PSS, organic molecules, and conjugated polymers have been synthesized and investigated.³²⁻³⁵ Recently, the small molecule metal phthalocyanines (MPc) with excellent hydrophobic properties and high hole mobility (10⁻³−10⁻⁵cm²V⁻¹s⁻¹) have been extensively used as p-type semiconductors in light-emitting diodes and solar cells.^29,31−43 Generally, the MPcfilms need to be deposited through thermal evaporation method.^31,37−43 But the thermal evaporation process is highly energy-consuming, resulting in the additional expenditure. Furthermore, the MPc derivatives as HTLs still need to be p-doped to get higher device efficiency, which will reduce the long-term stability of PSCs.⁴⁴ In this case, a solution-processed dopant-free SnPc is more desirable to realize good performance of PSCs.

Based on these considerations, the solution-processed dopant-free tin phthalocyanine (SnPc) is used as a HTL.

Carbon-based CsPbI₂Br PSCs with the simple structure of glass/ITO/SnO₂/CsPbI₂Br/SnPc/carbon were fabricated by a low-temperature process. In our work, two types offilms/PSCs are discussed: the reference-film/PSC corresponds to the CsPbI₂Br film/PSC without SnPc HTL, while the SnPc-film/

PSC refers to the CsPbI₂Brfilm/PSC with SnPc HTL. At the best SnPc solution concentration, the SnPc-PSC yields a champion efficiency of 11.39% and a steady-state efficiency of 11.23% with less hysteresis and improved device stability.

Systematic characterizations and analysis have been performed

to reveal the underlying mechanism for the enhanced photovoltaic performance. The results suggest that the SnPc HTL with the intermediate energy level can effectively suppress the charge recombination at the CsPbI₂Br/carbon interface, resulting in the enhanced hole extraction efficiency.

This work paves a new route to prepare the better-performing carbon-based inorganic PSCs by a low-temperature process.

■

EXPERIMENTAL SECTION

Materials. Lead(II) iodide (PbI2, 99.99%), poly(3,4-ethylene- dioxythiophene):poly(styrenesulfonate) (PEDOT:PSS), and lead(II) bromide (PbBr₂, 99.99%) were supplied by Xi’an Polymer Light Technology Corp. Cesium iodide (CsI, 99.999%), tin(IV) oxide (SnO₂) solution (15 wt % in H₂O colloidal dispersion), SnPc, and chlorobenzene (CB, 99.9%) were purchased from Alfa Aesar.

Isopropanol (IPA, ≥99.9%) and N,N-dimethylformamide (DMF,

≥99.9%) were provided by Aladdin. Indium-doped tin oxide glass substrate ((ITO, 15 Ω/sq), dimethyl sulfoxide (DMSO, ≥99.9%), and conductive carbon paste (printing ink) (CH-8 (MOD2)) were obtained from Nippon Sheet Glass (NSG) Company Limited of Japan, Sigma-Aldrich, and Jujo Printing Supplies & Technology (Pinghu) Co., Ltd., respectively. The related properties of carbon paste are as follows: adhesion 100/100; specific gravity 1.15; volume resistivity 1.0×10⁻²Ω·cm; surface resistivity 10.0Ω/cm²; moisture resistance 90% RH/240 h, 40 °C (changing within 10%); heat resistance 150°C/72 h (changing within 10%).

Synthesis of PbI₂(DMSO) and PbBr₂(DMSO) Adducts.1.0 M PbI₂and PbBr₂were dissolved in DMSO and stirred for 24 h at 60

°C. Subsequently, a 50 mL IPA was slowly dropped into the PbI2or PbBr2solution with continuous stirring, and then a white precipitate was formed. Finally, the obtained precipitate wasfiltered and dried in a vacuum oven at 60°C for 48 h to get the white PbI₂(DMSO) and PbBr₂ (DMSO) adducts. Figure S1 shows the XRD patterns and photo images of PbI₂, PbI₂ (DMSO), PbBr₂, and PbBr₂ (DMSO) powders.

Device Fabrication. The ITO glass substrate was chemically etched by zinc (Zn) powders together with 2 M hydrochloric acid (HCl). Then they were successively ultrasonic cleaned with detergent, deionized water, acetone, IPA, and ethanol for 15 min. After being dried by a nitrogen (N₂) stream, the ITO glass substrate was treated by UV-ozone for 5 min. A SnO₂solution (3 wt %) was spin-coated on the ITO glass substrate at 4000 rpm for 30 s and annealed at 150°C for 30 min in ambient air. After cooling to room temperature, the glass/ITO/SnO2substrate was treated under UV-ozone for 10 min, and then the CsPbI2Br precursor solution with 234 mg of CsI, 243 mg of PbI₂(DMSO), and 200 mg of PbBr₂(DMSO) in the mixed solvent of DMF and DMSO (4:1, v/v) at 0.9 M was spin-coated on the substrates at 1000 rpm for 10 s and subsequently at 3500 rpm for 40 s.

At about 20 s in the last spin-coating step, 150μL of CB solvent was dropped on them, followed by annealing at 35°C for 3 min and 120

°C for 10 min. The SnPc solutions were prepared by dissolving the SnPc in CB with the concentrations of 1, 2, 3, 4, and 5 mg mL⁻¹ under ultrasonic overnight. The SnPc HTL was deposited on CsPbI₂Br surface by spin-coating the SnPc solution at 3000 rpm for 30 s and annealed at 120 °C for 10 min. Finally, the conductive carbon paste was doctor-bladed on the sample surface and annealed at 120°C for 20 min. The size of the devices was 0.09 cm².

Characterization.The current density−voltage (J−V) curves of PSCs were performed by a Keithley 2420 source meter under standard AM 1.5G light illumination at 100 mW cm⁻² (Newport 91160). The light intensity was calibrated with a standard silicon solar cell (certified by NREL) as the reference. The scan voltage, scan speed, and delay time were −0.2−1.5 V, 100 mV s⁻¹, and 50 ms, respectively. The surface morphology, cross-sectional images, and energy-dispersive X-ray spectroscopy (EDS) elemental mapping were observed with a scanning electron microscope (SEM) (ZEISS Gemini 500). The contact angle was investigated by video-based optical contact angle measurement instrument (Dataphysics OCA Pro 15, Germany). The atomic force microscope (AFM) (Asylum Research,

Cypher) was employed to investigate the root-mean-square (RMS) roughness and average contact potential difference (CPD). The X-ray diffractometer (XRD) measurement was carried on the PANalytical X’Pert PRO with a Cu Kαradiation source. The ultraviolet−visible (UV−vis) absorption spectra were collected on the Shimadzu UV- 2550 spectrophotometer. The external quantum efficiency (EQE) was detected by a standard EQE system (Newport 66902). The darkJ−V curves were obtained by a Keithley 2420 source meter with a scan rate of 100 mV s⁻¹. The steady-state photoluminescence (PL) spectra were characterized on the fluorescence spectrophotometer (HITA- CHIF-5000), and the laser excitation wavelength is 322 nm. The time-resolved PL (TRPL) decay spectra were recorded by a fluorescence spectrophotometer (Fluorolog-3, Horiba) equipped with a time-correlated single-photon counting method excited by a 473 nm laser beam. The electrochemical impedance spectroscopy (EIS) was measured by an electrochemical workstation (Zahner Zennium, Germany) under the 10 mW cm⁻² white LED light,

applying a 10 mV AC signal with the frequency ranging from 1 Hz to 1 MHz. The capacitance−voltage (C−V), transient photocurrent (TPC), and transient photovoltage (TPV) measurements were recorded by the above electrochemical workstation. All of measure- ments were performed in ambient air (20−25°C) with a relative humidity (RH) of 25−35%.

■

RESULTS AND DISCUSSION

Figures 1a and1b present the schematic diagram of SnPc-PSC and SnPc molecular structure, respectively. As seen from the cross-sectional SEM image in Figure 1c and the partial magnification SEM image inFigure 1d, the SnPc-PSC exhibits a better layer-by-layer structure in different components. The average thickness for SnO₂, CsPbI₂Br, and carbon electrode is 30 nm, 310 nm, and 22.5μm, respectively. To more clearly Figure 1.(a) Schematic diagram of SnPc-PSC. (b) Molecular structure of SnPc. (c) Cross-sectional and (d) partial magnification SEM images of SnPc-PSC. (e) Energy level diagram of corresponding materials used in SnPc-PSC.

Figure 2.Top-view SEM images with different scale bars and contact angles: (a, b) reference-film; (c, d) SnPc-film. (e) Top-view SEM image and (f−j) EDS elemental mapping images of SnPc-film. The scale bars are 5μm. (k) XRD patterns of reference-film and SnPc-film deposited on the glass/ITO/SnO₂substrate. (l) UV−vis absorption spectra of reference-film and SnPc-film deposited on the glass/ITO/SnO₂substrate. Inset: the E_gvalues.

distinguish the thickness of SnPc HTL, the SnPc solution with a concentration of 3 mg mL⁻¹was spin-coated on a smooth silicon substrate. The thickness of SnPc HTL is measured by an ellipsometer.¹For this best condition, the thickness of SnPc HTL is 11.24 nm, as illustrated inFigure S2.Figure 1e depicts the energy level diagram of corresponding materials in SnPc- PSC. The HOMO energy level (≈ −5.24 eV) of SnPc is between the valence band maximum (VBM≈ −5.92 eV) of CsPbI₂Br and the work function (≈ −5.0 eV) of a carbon electrode, which will reduce the energy level difference and facilitate the hole extraction and transport.8,11,41,42,45

To optimize the process, the effect of SnPc solution concentrations on the J−V performance of SnPc-PSC has been investigated. Figure S3 shows the J−V curves and the detailed photovoltaic parameters of SnPc-PSCs as a function of SnPc solution concentrations. The detailed photovoltaic parameters are summarized in Table S1. It is noted that all photovoltaic parameters of SnPc-PSCs increase with the SnPc solution concentration at the beginning and then began to decrease when the concentration is greater than 3 mg mL⁻¹. Therefore, the optimal concentration of SnPc solution is 3 mg mL⁻¹, which is the condition for the SnPc-film and SnPc-PSC in this work.

The top-view SEM images of the reference-film and SnPc- film are presented inFigure 2a−d. Compared to the reference- film, the SnPc-film exhibits fewer surface trap states, which

indicates that the solution-processed SnPc fills the cracks/

pinholes at the grain boundary of the CsPbI₂Brfilm. The water contact angles for the reference-film and SnPc-film are inserted in the Figure 2a,c. They are 44.9° and 71.1°, respectively.

Compared to the reference-film, the increased contact angle in the SnPc-film will benefit to enhance the hydrophobicity and prevent moisture intrusion of CsPbI₂Br film, which will be further discussed in the following. To investigate the coverage of SnPc on the CsPbI₂Br film, the EDS elemental mapping images and spectra of the SnPc-film are shown inFigure 2e−j andFigure S4, respectively. The quantified atomic ratio of the SnPc-film is shown inTable S2. The Sn element from SnPc is observed in EDS elemental mapping image and spectra, which confirms the coverage of SnPc on the CsPbI₂Br surface.Figure 2k shows the XRD patterns of the reference-film and SnPc- film. The diffraction peaks at 14.6 and 29.5°can be assigned to the (100) and (200) crystal planes of CsPbI₂Br perovskite, respectively.^6,30 The intensity and position of the diffraction peaks in the SnPc-film do not change, indicating that the SnPc does not change the crystal structure of CsPbI₂Br perovskite.

The diffraction peaks of SnPc are almost obscured by CsPbI₂Br perovskite, and only two new peaks corresponding to SnPc at 7.5° and 12.6° (Figure S5) can be observed, which further confirm the existence of SnPc on the CsPbI₂Br surface.⁴⁶

Figure 2l depicts the UV−vis absorption spectra of the reference-film and SnPc-film. Similar absorption spectra Figure 3.(a)J−Vcurves of reference-PSC and SnPc-PSC measured under the RS direction. (b) EQE and integrated current density curves of reference-PSC and SnPc-PSC. (c)J−Vcurves of reference-PSC and SnPc-PSC under the RS and FS directions. (d) Steady-state current density and PCE as a function of time for reference-PSC and SnPc-PSC measured at their maximum power output point. (e) PCE statistics histograms of reference-PSC and SnPc-PSC. (f) Evolution of PCE over time measured in ambient air (20−25°C) at a RH of 25−35%.

confirmed that the SnPc does not change the optical absorption of the CsPbI₂Br film. Moreover, the E_g of the reference-film and SnPc-film can be calculated by the Tauc plots transformed from the UV−vis absorption spectra viaeq 1:11,23,25,43,47

h k h E

( ) ( _g)^1/2

α υ = υ− ₍₁₎

whereα,hυ, andkare the absorption coefficient, photo energy, and arbitrary constant, respectively. As illustrated inFigure 2l, theE_g value for both the reference-film and SnPc-film is 1.91 eV, confirming that the incorporation of SnPc does not change the absorption edge of the CsPbI₂Br film.⁴⁵

In theory, the open-circuit voltage (V_oc) of PSCs is determined by the difference between E_fn and E_fp: V_oc = E_fn

−E_fp, in which theE_fnandE_fprefer to the electron quasi-Fermi energy level and hole quasi-Fermi energy level, respec- tively.^30,48,49Figure S6provides the schematic of energy level diagram of reference-PSC and SnPc-PSC. For the reference- PSC, E_fn depends on the interface between CsPbI₂Br and SnO₂, andE_fpis determined by the interaction of CsPbI₂Br and carbon electrode (Figure S6a).^48,49In the SnPc-PSC, the E_fp should be determined by the interaction between the CsPbI₂Br and the SnPc.^48,49The incorporation of SnPc does not change E_fn but decreases the level value of E_fp (Figure S6b). The reducedE_fpis mainly attributed to the lower HOMO energy level of SnPc than the Femi level of carbon electrode.

Therefore, theV_ocof SnPc-PSC is supposed to be higher than that of the reference-PSC.⁴⁸To confirm the above result, we fabricated the reference-PSC and SnPc-PSC at the same time.

Figure 3a shows theJ−V curves of reference-PSC and SnPc- PSC measured in the reverse scan (RS) direction. The detailed photovoltaic parameters are presented in Table 1. It can be seen that the reference-PSC and SnPc-PSC achieve the champion power conversion efficiency (PCE) of 9.22% and 11.39%, respectively. The enhanced PCE in SnPc-PSC is attributed to the simultaneous improvement of V_oc, short- circuit current density (J_sc), and fill factor (FF). In particular, the SnPc-PSC shows a significantly high V_ocof 1.244 V. The improvedV_ocis consistent with the above theory. In addition, Table S3andFigure S7summarize and compare the annealing temperature,V_oc, and PCE values of carbon-based CsPbI₂Br solar cells reported in recent years. It is noted that theV_ocin our work have achieved the record values of low-temperature carbon-based CsPbI₂Br PSCs reported so far. The value of FF is correlated to the ratio of the shunt resistance (R_sh) to series resistance (R_s) (R_sh/R_s).⁵⁰The higher FF in SnPc-PSC can be attributed to the larger R_sh/R_s.⁵⁰ The increasedJ_scfor SnPc- PSC is further confirmed by the EQE measurements.Figure 3b shows the EQE spectra and integrated current density curves of reference-PSC and SnPc-PSC. The integrated current density values for reference-PSC and SnPc-PSC are 12.78 and 13.37 mA cm⁻², respectively, which are very close to the J−Vresults (errors less than 3%).

It is reported that theJ−Vhysteresis is inevitable in mixed- halide inorganic PSCs due to the iodide and bromide phase

segregation under light illumination.⁵⁰To investigate the effect of SnPc on the hysteresis behavior, theJ−Vcurves have been measured in the RS and forward scan (FS) directions, as shown in Figure 3c. The corresponding photovoltaic parameters are summarized in Table S4. Generally, the hysteresis index (HI) can be used to describe the J−V hysteresis of PSCs, which is defined asformula 2.^11,30,50

HI PCE PCE

PCE

RS FS

RS

= −

(2) Here the HI decreases from 0.113 (reference-PSC) to 0.062 (SnPc-PSC), indicating that theJ−Vhysteresis in SnPc-PSC is suppressed compared to the reference-PSC. In addition, considering that such photocurrent hysteresis is related to charge trapping, the reference-PSC with severeJ−Vhysteresis can be deteriorated by trapped-charge-driven degradation of CsPbI₂Br perovskite.⁵¹ The J−V hysteresis is in accordance with the steady-state current density as a function of time, as shown inFigure S8.^38,50The current density for SnPc-PSC is quickly stabilized within a few seconds, whereas the current density for reference-PSC does not seem to be stabilized even in 60 s. Much faster stabilization of current density can partly explain the reduced HI in SnPc-PSC.^39,51The detailed reasons will be further discussed in the following.

Given that there is theJ−Vhysteresis in PSCs, it is necessary to measure the steady-state current density and PCE output under their respective optimal bias voltage (V_opt) at the maximum power point to confirm the reliability of J−V measurements, as seen inFigure 3d. Here the values ofV_optfor reference-PSC and SnPc-PSC are 0.847 and 0.968 V, respectively. Under standard AM 1.5G continuous light illumination for 110 s, the values of PCE for reference-PSC and SnPc-PSC stabilize at 8.90% and 11.23% with current density of 11.69 and 12.17 mA cm⁻², respectively. These values are close to the values measured from the J−V curves. This confirms the reliability of ourJ−Vresults. To demonstrate the reproducibility of reference-PSC and SnPc-PSC, the statistical distributions of PCE,V_oc,J_sc, FF, andR_sh/R_sfrom 50 individual devices are shown inFigure 3e andFigure S9. All parameters obtained from SnPc-PSCs exhibit a narrower distribution than those of the reference-PSCs, which implies the better reproducibility in SnPc-PSC.²³

Figure 3f demonstrates the normalized PCE of reference- PSC and SnPc-PSC as a function of storage time. After exposed ambient air (20−25 °C) with 25−35% RH for 30 days, the values of PCE for the unsealed reference-PSC and SnPc-PSC remain at about 76% and 90% of its initial value, respectively. Significantly, the SnPc-PSC demonstrates a better stability. The improved stability for SnPc-PSC is mainly attributed to the fact that the hydrophobic SnPc HTL can block water molecules from infiltrating into CsPbI₂Br film, which is consistent with the above-mentioned increased water contact angle.

To gain insight into the improvement of the V_oc and the reduced HI in SnPc-PSC, we fabricated hole-only devices with the structure of glass/ITO/PEDOT:PSS/CsPbI₂Br/with or Table 1. Photovoltaic Parameters of Reference-PSC and SnPc-PSC in RS Direction

devices V_oc(V) J_sc^c(mA cm⁻²) J_sc^d(mA cm⁻²) FF PCE (%)

reference-PSC 1.142 13.11 12.78 0.616 9.22^a(8.73)^b

SnPc-PSC 1.244 13.69 13.37 0.669 11.39^a(10.86)^b

aBest PCE.^bAverage PCE from 50 devices.^cMeasuredJ_scvalues from theJ−Vcurves.^dIntegratedJ_scvalues from the EQE curves.

without SnPc/carbon to estimate the hole mobility (μh) in the full device. The μh values of these two devices can be calculated from thefitting log(J)−log(V) curves inFigure 4by

the space charge limited current (SCLC) method according to Mott−Gurney’s square law:¹⁰

JL V 8

h 9

3

0

μ 2

= εε

(3) where ε, ε0, and L are the relative permittivity (ε = 8.5),⁴⁵ vacuum permittivity (ε0 = 8.85 × 10⁻¹⁴ F cm⁻¹), and the thickness of the CsPbI₂Br film (L ≈ 310 nm, as shown in Figure 1d andFigure S10), respectively. The device with SnPc HTL exhibits theμhvalue of 7.02×10⁻⁴cm²V⁻¹s⁻¹, which is higher than that of the device without SnPc HTL (5.63×10⁻⁴ cm² V⁻¹ s⁻¹). This result indicates that the SnPc HTL can effectively accelerate the hole transport and collection at the CsPbI₂Br/carbon interface, which is the further evidence for the improvement of device performance.

To provide more evidence of the improved performance in SnPc-PSC, some microstructure and electrical measurements were carried out. The micrograph and RMS roughness of reference-film and SnPc-film were measured by AFM, as shown inFigure 5a−c. It can be seen that the values of RMS

roughness for the reference-film and SnPc-film are 30.17 and 26.03 nm. As mentioned before, the thickness of the SnPc HTL is only 11.24 nm under the ideal conditions. Therefore, the SnPc HTL deposited on the rough CsPbI₂Br surface by a solution process cannot be conformal (Figure S11a), and it should be discontinuous for hole extraction and transport (Figure S11b).¹According to previous reports, the change of the obtained average CPD from Kelvin probe force microscopy (KPFM) at the nanometer scale is consistent with the bulk measurement, which is related to the device performance as well.^50,52 To investigate the effect of SnPc on the local electrical properties of the CsPbI₂Brfilm at the nanoscale level, we measured the average CPD of SnPc films with different SnPc solution concentrations by KPFM, as shown in Figure 5d−f andFigure S12. The average CPD values for SnPcfilms are summarized inTable S5. It can be seen that the SnPc-film exhibits a highest average CPD. According to previous reports, the increased CPD value may be attributed to the reduced surface defects of CsPbI₂Br film, or the WF of SnPc HTL treated with 3 mg mL⁻¹ SnPc solution is the highest, thus promoting the charge separation and achieving higher V_ocof SnPc-PSC.^32,33,52 This result is in agreement with the J−V measurements.

Figure 6a compares the darkJ−V curves of reference-PSC and SnPc-PSC. As seen in Figure 6a, the leakage current is lower than that of the reference-PSC. This suggests that the charge recombination at the CsPbI₂Br/carbon interface is suppressed in SnPc-PSC, which is beneficial for the improve- ment of J_sc and FF.^41−43,53 Figure 6b shows the linear-scale darkJ−Vcurves. It is known that theV_occan be evaluated from the linear-scale dark J−V curves, according to the intercept between the linear portion of darkJ−Vcurves and the voltage axis.^42,43,53 The evaluated V_oc from dark J−V curves of reference-PSC and SnPc-PSC is 1.135 and 1.233 V, respectively. This implies that the SnPc-PSC has a higher intrinsicV_oc.17,42,43,53

This is in good accordance with theJ−V measurements.

To further investigate the effect of the inserted SnPc HTL on the efficiency of exciton dissociation, charge collection, and Figure 4.The log(J)−log(V) curves of hole-only devices with the

inserted architecture.

Figure 5.AFM images: (a) the reference-film; (b) the SnPc-film. (c) Histograms of RMS roughness obtained from the AFM measurement. KPFM images: (d) the reference-film; (e) the SnPc-film. (f) Histograms of average CPD obtained from the KPFM measurement. The scale bar is 2μm.

recombination, the photocurrent density (J_ph) versus effective voltage (V_eff) curves (J_ph−V_eff) of reference-PSC and SnPc- PSC were investigated, as shown inFigure 6c.^17,54−59Here the J_phis defined asJ_ph=J_light−J_dark, in whichJ_lightandJ_darkare the current density under light illumination and in the dark, respectively.17,54,56−58TheV_eff is defined as V_eff=V₀−V_a, in which V₀ is the voltage at J_ph = 0 and V_a is the applied voltage.17,54,57−59EachJ_phincreases linearly with the increase ofV_effandfinally becomes a saturation current density (J_sat) at highV_eff(>1 V) region, where all photogenerated excitons are dissociated into free carriers and adequately collected by electrodes.¹⁷ This means that the charge recombination in PSCs is at a minimum.¹⁷ Generally, the J_satis related to the exciton dissociation efficiency (J_ph*/J_sat), charge transport and collection efficiency (J_ph^#/J_sat), and maximum exciton gen- eration rate (G_max).^54,57,58HereJ_ph*andJ_ph^#correspond to a J_phvalue at short circuit and maximal power output conditions, respectively.⁵⁴G_maxcan be calculated byeq 4:^56,57,59

J_sat =qG_maxL ₍₄₎

whereqis the elementary charge. The correspondingJ_ph*,J_ph^#, J_sat, J_ph*/J_sat, J_ph^#/J_sat ,and G_max values of reference-PSC and SnPc-PSC are summarized in Table S6. It is noted that the values of J_sat and J_ph*/J_sat at the short circuit condition for SnPc-PSC are higher than that of the reference-PSC. These suggest that the inserted SnPc HTL promotes the exciton dissociation in SnPc-PSCs.^54,57,58At the maximal power output condition, theJ_ph^#/J_satvalues for SnPc-PSC are also larger that of the reference-PSC, confirming that the more efficient charge transport and collection in SnPc-PSC.^54,57,58According toeq 4, the SnPc-PSC exhibits the higher G_maxvalue compared to that of the reference-PSC, which reveals that the inserted SnPc HTL suppresses the charge recombination at the CsPbI₂Br/

carbon interface.⁵⁸Therefore, better photovoltaic performance can be achieved in SnPc-PSC.

To compare the charge recombination and extraction process in reference-PSC and SnPc-PSC, the PL measurements were carried out.Figure 6d displays the steady-state PL spectra

of glass/CsPbI₂Br, glass/CsPbI₂Br/carbon and glass/

CsPbI₂Br/SnPc/carbon samples. All samples show a PL emission peak at about 650 nm, which is consistent with previous reports.^23,50 It is noted that the glass/CsPbI₂Br/

SnPc/carbon sample presents the lowest PL intensity, even lower than that of the glass/CsPbI₂Br/carbon sample, indicating that the inserted SnPc HTL can promote the hole extraction.29,30,37,39

Figure 6e presents the TRPL decay spectra which can be fitted by the following biexponential decay function:^11,51

y t( )=A₁e^−t/τ1+A₂e^−t/τ2+y_{0 (5)} whereA₁andA₂are the relative amplitudes,τ1andτ2are the fast decay component and slow decay component, respectively, and y₀ is a constant for the baseline offset.⁵⁰ The relevant parameters are summarized inTable S7. The average carrier lifetimes (τave) can be estimated via formula 6:^50,55

A

A A

A

A A

ave

1 1

1 1 2 2

1

2 2

1 1 2 2

τ τ 2

τ τ τ τ

τ τ τ

= + +

+ ₍₆₎

Theτaveof glass/CsPbI₂Br, glass/CsPbI₂Br/carbon, and glass/

CsPbI₂Br/SnPc/carbon samples is 32.57, 28.16, and 23.66 ns, respectively. According to the literature, the reduced τave

indicates the effective hole extraction and suppressed charge recombination, which contribute to the better photovoltaic performance of SnPc-PSC.^31,37

The C−V measurements were then employed to estimate the built-in potential (V_bi) and understand the origin of the enhancedV_ocin SnPc-PSC.^30,60,61Figure 6f shows theC⁻²−V curves for reference-PSC and SnPc-PSC, from which theV_bi can be obtained through the Mott−Schottky equa- tion:18,23,30,60

C

V V

A q N

1 2( )

2

bi 2

εε0 A

= −

(7) Figure 6.DarkJ−Vcurves of reference-PSC and SnPc-PSC plotted on a (a) semilog scale and (b) linear scale. (c) TheJ_ph−V_effcurves of reference- PSC and SnPc-PSC. (d) Steady-state PL spectra and (e) TRPL decay spectra of glass/CsPbI₂Br, glass/CsPbI₂Br/carbon, and glass/CsPbI₂Br/

SnPc/carbon samples. (f)C⁻²−Vcharacteristics for reference-PSC and SnPc-PSC.

where A and N_A are the device active area and carrier concentration, respectively. The V_bi of SnPc-PSC is 1.15 V, which is much higher than 1.07 V of reference-PSC. It was reported that theV_ocvalue depends on the value ofV_bi.^30,60,61 The higher V_bi value contributes to the larger V_oc in SnPc- PSC.^30,60The increasedV_bi will not only increase the driving force for carrier separation but also favor to form an extended depletion region to efficiently suppress charge recombina- tion.^61,62 Hence, the SnPc-PSCs demonstrate the higher V_oc compared to the reference-PSCs.^61,62

To further study the relationship between charge carrier dynamics and device performance, various characterization methods including TPC, TPV, EIS, and light intensity (P_light)- dependent J−V characteristics were carried out.¹⁹ Generally, the TPC decay measurement can probe the charge extraction and transport features in PSCs.^63,64As seen inFigure 7a, under short-circuit conditions, the SnPc-PSC exhibits a faster photocurrent decay time (4.59 μs) than that of the reference-PSC (5.55 μs). This indicates that the charge extraction and transport in SnPc-PSC are more efficient than that of the reference-PSC, which benefits to achieve the higher J_scand FF.17,45,63,64

The TPV decay measurement is related to the charge recombination process in PSCs.⁶³⁻⁶⁵ Figure 7b shows a longer decay lifetime of 0.19 ms for SnPc-PSC compared to the reference-PSC (0.09 ms). This result confirms

that the SnPc can efficiently decrease the charge recombination rate at the CsPbI₂Br/carbon interface, thereby leading to the enhanced V_oc in SnPc-PSC.17,22,63−65 Figure 7c shows the Nyquist plots of reference-PSC and SnPc-PSC measured at a bias of 1.1 V under light illumination, where the inserted equivalent circuit model is used to fit the experimental data.

According to reports, the R_s corresponds to the value of the starting point at the real part (Z′) in thefitted Nyquist plots, which is related to the sheet resistance of ITO glass substrate and the contact resistance of PSCs.⁶⁵The transport resistance (R_tra) corresponds to the high-frequency arc, which is associated with the charge extraction and transport at the SnO₂/CsPbI₂Br/carbon interfaces.^41−43,50 The recombination resistance (R_rec) corresponds to the low-frequency arc, which is attributed to the charge recombination in CsPbI₂Brfilm and at above interfaces.^41−43,50In view of the spatial inhomogeneity induced by the defects and impurities at interfaces, the constant phase angle element (CPE_traand CPE_rec) are used to replace the ideal capacitance.⁵⁰ Table S8 provides all of the fitted parameters. It is seen that the SnPc-PSC exhibits a smaller R_s than the reference-PSC, which is beneficial for achieving higher FF.^42,43,66 Given that the SnO₂/CsPbI₂Br interface is identical, R_tra and R_rec are determined from the charge transport and recombination process at the CsPbI₂Br/

carbon interface. As shown in Figure 7d, the SnPc-PSC Figure 7.(a) Normalized TPV and (b) TPC decay curves of reference-PSC and SnPc-PSC. (c) Nyquist plots of reference-PSC and SnPc-PSC measured under light illumination. Solid symbol: experimental data. Solid line:fitting results. Inset: equivalent circuit model. (d) Histograms ofR_tra andR_recfrom thefitting results of Nyquist plots. (e)R_traand (f)R_recas a function of applied bias for reference-PSC and SnPc-PSC extracted from the Nyquist plot measured under light illumination.

exhibits a smallerR_tra and largerR_reccompared to that of the reference-PSC. Meanwhile,Figures 7e and7f show theR_traand R_rec fitted from the Nyquist plots measured under light illumination at different applied bias, respectively. Compared with the reference-PSC, the SnPc-PSC exhibits the lowerR_tra and higherR_recat the same applied bias condition. This result confirms that the inserted SnPc HTL can promote hole extraction and suppress charge recombination at the CsPbI₂Br/carbon interface, which is responsible for the increased photovoltaic parameters in SnPc-PSC.30,42,43,50,63

In addition, the V_oc and J_sc as a function of P_light for reference-PSC and SnPc-PSC were measured. The linear relationship of V_oc and the natural logarithm of P_light can be plotted according to the following relationship:^29,50

V nkT

q ln(P ) constant

oc= light +

(8) wherenis the ideal factor,kis the Boltzmann constant, andT is the absolute temperature. In principle, the slope ofV_oc vs P_lightwill be equal to kT/q when no trap-assisted recombina- tion or bimolecular recombination determines the performance of PSCs.18,50,54,67

If theV_ochas a stronger dependence onP_light, the slope would be greater thankT/qdue to the involvement of the trap-assisted recombination.^18,50,67As depicted inFigure 8a, the reference-PSC and SnPc-PSC show a slope of 1.44kT/q and 1.19kT/q, respectively. This result means restrained charge recombination in SnPc-PSC.^50,67 The dependence of J_sc on P_lightcould be described by the power law formula19,29,50,54,57

J_sc∝ P_light^S (S≤1) ₍₉₎

where S is an exponential factor. According to the previous reports, the value of Sis 0.75 for PSCs limited by the space charge due to an interfacial barrier or a charge carrier imbalance.⁶⁸ The S value is close to 1 for PSCs which have not been affected by space charge.^19,68As displayed inFigure 8b, the value ofSfor SnPc-PSC is 0.992, which appears to be not limited by space charge, whereas the reference-PSC has a smaller S value of 0.946. The higher S value for SnPc-PSC indicates that the inserted SnPc HTL significantly mitigates the bimolecular recombination at the CsPbI₂Br/carbon inter- face.^57,69,70 All of these characterizations indicate that the SnPc modification at the CsPbI2Br/carbon interface can form a favorable energy level alignment and prevent moisture penetration into the CsPbI₂Br film, resulting in the reduced charge recombination and improved device performance.

■

CONCLUSIONS

In summary, we have successfully demonstrated a simple strategy to improve the photovoltaic performance of low- temperature carbon-based CsPbI₂Br planar PSCs by introduc- ing the solution-processed and dopant-free SnPc HTL. At the optimal SnPc solution concentration, the SnPc-PSC achieves a champion PCE of 11.39% and a steady-state PCE of 11.23% as well as less hysteresis, better reproducibility, and improved device stability in ambient air. The results confirm that the inserted SnPc HTL can form a favorable energy level alignment at the CsPbI₂Br/carbon interface, resulting in the promoted hole extraction and suppressed charge recombina- tion. Our work shows that the inserted SnPc HTL at the perovskite/carbon interface benefits to further improve the photovoltaic performance of low-temperature inorganic PSCs, which may make carbon-based PSCs closer to the market.

■

ASSOCIATED CONTENT

*^sı Supporting Information

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsaem.0c01184.

Photovoltaic parameters corresponding to Table S1, Table S4, Figure S3, and Figure S9; quantified atomic ratio for the SnPc-film as obtained from EDS measure- ment; a summary of annealing temperature, V_oc, and PCE values measured under RS direction for carbon- based CsPbI₂Br PSCs reported in recent years; the values ofJ_ph*,J_ph^#,J_sat,J_ph*/J_sat,J_ph^#/J_sat, andG_max;fitting parameters of TRPL decay spectra; fitting values of different electronic parameters obtained from the Nyquist plots; XRD patterns and photo images of PbI₂, PbI₂(DMSO), PbBr₂, and PbBr₂(DMSO) pow- ders; thickness of SnPc HTL measured by an ellipsometer; J−V curves of SnPc-PSCs with different SnPc solution concentrations; EDS spectra of the SnPc- film; XRD patterns of SnPc powder and SnPc film annealed at 120 °C; the schematic of energy level diagram; the steady-state current density as a function of time measured at the maximum power output point;

spectrum for the thickness of CsPbI₂Br layer; two possible growth models while depositing solution- processed dopant-free SnPc on the CsPbI₂Br surface;

KPFM images and CPD values of SnPcfilms based on different SnPc solution concentrations (PDF)

Figure 8.(a)V_ocand (b)J_scas a function ofP_lightfor reference-PSC and SnPc-PSC.

■

AUTHOR INFORMATION Corresponding Authors

Sujuan Wu−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China; orcid.org/

0000-0002-8269-7811; Email:[email protected] Jinwei Gao−Institute for Advanced Materials, South China

Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China; Email:gaojw@

scnu.edu.cn Authors

Xiang Zhang−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China

Naitao Gao− Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China

Yuzhu Li− Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China

Lai Xie−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China

Xiang Yu−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China

Xubing Lu−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China; orcid.org/

0000-0002-2552-9571

Xingsen Gao−Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China; orcid.org/

0000-0002-2725-0785

Lingling Shui−Guangdong Provincial Key Laboratory of Optical Information Materials and Technology, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China.; orcid.org/

0000-0001-8517-1535

Jun-Ming Liu− Institute for Advanced Materials, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China; orcid.org/

0000-0001-8988-8429

Complete contact information is available at:

https://pubs.acs.org/10.1021/acsaem.0c01184

Notes

The authors declare no competingfinancial interest.

■

ACKNOWLEDGMENTS

We acknowledge thefinancial support of the National Key R &

D Program of China (2016YFB0401502), Science and Technology Program of Guangzhou (No. 2019050001), NSFC-Guangdong Joint Fund (No. U1801256), the Natural S c i e n c e F o u n d a t i o n o f G u a n g d o n g P r o v i n c e (2020A1515010731), Guangdong Provincial Key Laboratory of Optical Information Materials and Technology (2017B030301007), and the MOE International Laboratory for Optical Information Technologies.

■

(1) Jiang, Q.; Zhao, Y.; Zhang, X.; Yang, X.; Chen, Y.; Chu, Z.; Ye,^REFERENCES Q.; Li, X.; Yin, Z.; You, J. Surface Passivation of Perovskite Film for Efficient Solar Cells.Nat. Photonics2019,13, 460−466.

(2) Tsai, H.; Asadpour, R.; Blancon, J.-C.; Stoumpos, C. C.; Durand, O.; Strzalka, J. W.; Chen, B.; Verduzco, R.; Ajayan, P. M.; Tretiak, S.;

et al. Light-Induced Lattice Expansion Leads to High-Efficiency Perovskite Solar Cells.Science2018,360, 67−70.

(3) Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith, H. J. Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites.Science2012,338, 643−647.

(4) Kojima, A.; Teshima, K.; Shirai, Y.; Miyasaka, T. Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells.

J. Am. Chem. Soc.2009,131, 6050−6051.

(5) Chart, N. https://www.nrel.gov/pv/cell-efficiency.html (ac- cessed February 2020).

(6) Zhao, H.; Yang, S.; Han, Y.; Yuan, S.; Jiang, H.; Duan, C.; Liu, Z.; Liu, S. A High Mobility Conjugated Polymer Enables Air and Thermally Stable CsPbI2Br Perovskite Solar Cells with an Efficiency Exceeding 15%.Adv. Mater. Technol.2019,4, 1900311.

(7) Zhang, Y.; Wu, C.; Wang, D.; Zhang, Z.; Qi, X.; Zhu, N.; Liu, G.;

Li, X.; Hu, H.; Chen, Z.; Xiao, L.; Qu, B. High Efficiency (16.37%) of Cesium BromidePassivated All-Inorganic CsPbI₂Br Perovskite Solar Cells.Sol. RRL2019,3, 1900254.

(8) Wang, G.; Liu, J.; Chen, K.; Pathak, R.; Gurung, A.; Qiao, Q.

High-Performance Carbon Electrode-Based CsPbI₂Br Inorganic Perovskite Solar Cell Based on Poly (3-Hexylthiophene)-Carbon Nanotubes Composite Hole-Transporting Layer. J. Colloid Interface Sci.2019,555, 180−186.

(9) Fan, Y.; Fang, J.; Chang, X.; Tang, M.-C.; Barrit, D.; Xu, Z.;

Jiang, Z.; Wen, J.; Zhao, H.; Niu, T.; et al. Scalable Ambient Fabrication of High-Performance CsPbI2Br Solar Cells.Joule2019,3, 1−18.

(10) Wang, K.-L.; Wang, R.; Wang, Z.-K.; Li, M.; Zhang, Y.; Ma, H.;

Liao, L.-S.; Yang, Y. Tailored Phase Transformation of CsPbI₂Br Films by Copper (II) Bromide for High-Performance All-Inorganic Perovskite Solar Cells.Nano Lett.2019,19, 5176−5184.

(11) Guo, Z.; Zhao, S.; Liu, A.; Kamata, Y.; Teo, S.; Yang, S.; Xu, Z.;

Hayase, S.; Ma, T. Niobium Incorporation into CsPbI₂Br for Stable and Efficient All-Inorganic Perovskite Solar Cells.ACS Appl. Mater.

Interfaces2019,11, 19994−20003.

(12) Liu, D.; Yang, C.; Bates, M.; Lunt, R. R. Room Temperature Processing of Inorganic Perovskite Films to Enable Flexible Solar Cells.iScience2018,6, 272−279.

(13) Xiang, W.; Wang, Z.; Kubicki, D. J.; Tress, W.; Luo, J.;

Prochowicz, D.; Akin, S.; Emsley, L.; Zhou, J.; Dietler, G.; Gratzel, M.; Hagfeldt, A. Europium-Doped CsPbI₂Br for Stable and Highly Efficient Inorganic Perovskite Solar Cells.Joule2019,3, 205−214.

(14) Wang, Y.; Liu, X.; Zhang, T.; Wang, X.; Kan, M.; Shi, J.; Zhao, Y. The Role of Dimethylammonium Iodine in CsPbI3 Perovskite Fabrication: Additive or Dopant? Angew. Chem., Int. Ed. 2019, 58, 16691−16696.

(15) Zai, H.; Zhang, D.; Li, L.; Zhu, C.; Ma, S.; Zhao, Y.; Zhao, Z.;

Chen, C.; Zhou, H.; Li, Y.; Chen, Q. Low-Temperature-Processed Inorganic Perovskite Solar Cells via Solvent Engineering with Enhanced Mass Transport.J. Mater. Chem. A2018,6, 23602−23609.

(16) Zeng, Z.; Zhang, J.; Gan, X.; Sun, H.; Shang, M.; Hou, D.; Lu, C.; Chen, R.; Zhu, Y.; Han, L. In Situ Grain Boundary Functionalization for Stable and Efficient Inorganic CsPbI2Br Perovskite Solar Cells.Adv. Energy Mater.2018,8, 1801050.

(17) Zhu, W.; Zhang, Q.; Chen, D.; Zhang, Z.; Lin, Z.; Chang, J.;

Zhang, J.; Zhang, C.; Hao, Y. Intermolecular Exchange Boosts Efficiency of Air-Stable, Carbon-Based All-Inorganic Planar CsPbIBr2

Perovskite Solar Cells to Over 9%. Adv. Energy Mater. 2018, 8, 1802080.

(18) Chen, W.; Chen, H.; Xu, G.; Xue, R.; Wang, S.; Li, Y.; Li, Y.

Precise Control of Crystal Growth for Highly Efficient CsPbI2Br Perovskite Solar Cells.Joule2019,3, 191−204.