Introduction
Organic–inorganic hybrid halide perovskite-based solar cells (PSCs) received tremendous attention in recent years owing to its high absorption coefficient, small exciton-binding energy, high carrier mobility, and extensive carrier-diffusion lengths [1–9]. Single-junction PSCs have achieved a power conversion efficiency (PCE) of 22.1% [10–14]. Tandem solar cells present a promising strategy to further enhance device performance due to the Shockley-Queisser limit for single-junction devices [15,16]. With its proper small band gap (Eg) of 0.9 to 1.2 eV, mixed lead-tin (Pb-Sn) perovskite is an ideal light-harvesting material for rear cell in tandem solar cells. Recently, PSCs based on a mixed Pb-Sn perovskite-based absorber with an efficiency of 17% have been demonstrated [17].
Two device architectures are commonly used for PSCs: the conventional n-i-p structure and the inverted p-i-n structure. In both structures, a perovskite active layer is sandwiched between a hole-transporting layer (HTL) and an electron-transporting layer (ETL). For conventional n-i-p structures, the most frequently used ETLs are transition metal oxides (TiO2 and ZnO), whereas HTLs are mainly organic materials, such as 2,2′,7,7′-tetrakis [N,N-di(4-methoxyphenyl) amino]-9,9′-spirobifluorene (spiro-OMeTAD) [18–21]. Inverted devices generally employ NiOx, poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS) as HTL, and [6,6]-phenyl C61-butyric acid methyl ester (PCBM), ZnO, or SnO2 as the top ETL [22–25]. For inverted device structures, both ETL and HTL can use inorganic materials, thus facilitating better device stability. Numerous studies have attempted to optimize the contact between the active layer and the carrier-transporting layers, as well as between the carrier-transporting layers and the electrodes, to achieve quick carrier transport and less interface recombination [26–30]. Given the tedious device-fabrication process for each layer (especially the requirement of high-temperature annealing process for the fabrication of some oxide films), ETL- or HTL-free devices were developed to simplify device fabrication [26,31–37]. HTL-free inverted-structure solar cells based on both Pb and Sn perovskites have been reported with good performance and stability [34–38].
In the current work, we constructed an HTL-free planar Pb-Sn binary PSC by preparing perovskite film on bare indium tin oxide (ITO) for the first time. Homogeneous and mirror-like FAPb0.5Sn0.5I3 films were developed through anti-solvent assisted crystallization and solvent engineering. The effect of the Pb-Sn ratio as well as the concentration of SnF2 on film morphology and device performance were investigated. Based on the inverted-structure (ITO/perovskite/PCBM/BCP/Al), solar cells with a highest PCE of 7.94% under the illumination of AM 1.5G solar simulator were presented.
Experimental section
Perovskite precursor preparation
The perovskite precursor solution was obtained by mixing the solutions of FASnI3 and FAPbI3 at a specific ratio. FASnI3 and FAPbI3 solutions of 1-M†) concentration were prepared in a mixed solvent of N,N-dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) (volume ratio of 3:2). The mole ratio of FAI to SnI2/PbI2 was 1:1. Furthermore, 0.1 M SnF2 was added to the tin precursor to prevent oxidation of the Sn2+ to Sn4+ unless otherwise stated. The final solution was stirred at 70°C for 1 h and subsequently filtered with a PTFE filter (0.22 µm pore size) before use.
Perovskite film fabrication
For FAPb1−xSnxI3 film (x>0.25) fabrication, the mixed precursor prepared above was spin coated onto the substrate via a one-step anti-solvent assisted crystalline process at 1000 and 5000 rpm for 10 and 30 s, respectively. In the second process, 800 µL toluene was slowly dropped in the middle of the substrate. The perovskite film was subsequently annealed at 70°C for 30 min.
For FAPbI3 and FAPb0.75Sn0.25I3 films, the annealing temperature was increased to 170°C and 100°C, respectively. All of the operations were performed in a glovebox (O2≤1 ppm, H2O≤1 ppm).
Device fabrication
ITO-coated glasses were sequentially cleaned with Triton X-100 detergent, acetone, deionized water, and isopropanol. The cleaned ITO glasses were treated with UVO for 3 min before being transferred into the glovebox. Perovskites were fabricated on top of the ITO glasses using the procedure described above. The PCBM film was prepared via a spin-coating process (speed of 2000 rpm and time of 60 s) using chlorobenzene as the solvent (concentration of 10 mg/mL), followed by annealing at 80°C for 5 min. Subsequently, a saturated solution of BCP in anhydrous methanol was spin coated at 6000 rpm for 15 s. Aluminum (of 100 nm thickness) was thermally evaporated as the top electrode at a rate of 1 Å/s.
Device characterization
Current-voltage (J-V) curves were measured using a Keithley 2400 source unit under AM1.5G solar simulator (Enli Tech, Taiwan) at 100 mW/cm2 (1 sun). The light intensity was calibrated by a certified KG-5 Si diode. The spectral mismatch correction factor is 0.49%. The devices were measured in reverse scan (from 0.7 to 0 V, at increments of 0.01 V) and the delay time was 30 ms. J-V curves for all devices were measured by masking the active area using a metal mask with an area of 0.04 cm2. The external quantum efficiency (EQE) spectra were obtained by a commercial system (Solar Cell Scan 100, Beijing Zolix Instruments Co., Ltd). Additionally, the cells were subjected to monochromatic illumination (150 W Xe lamp passing through monochromator filters). The light intensity was calibrated by a standard photodetector (QE-B3/S1337-1010BQ, Zolix). The light beam was chopped at 180 Hz, and the response of the cell was acquired using a Stanford Research SR830 lock-in amplifier.
Results and discussion
We firstly investigated the optical properties of the mixed Pb-Sn perovskites. As the Sn is added, the absorption spectra of FAPb1−xSnxI3 (x = 0.25, 0.50, 0.75, 1.00) are all red shifted. Based on the absorption onset point, the band gaps of FAPb0.75Sn0.25I3 and FAPb0.5Sn0.5I3 are estimated to be 1.31 and 1.27 eV, respectively (Fig. 1(a)). The absorption tail mainly results from sub-band gap states [39]. Photoluminescence (PL) spectra show similar trend as the absorption spectra (Fig. 1(b)). The Eg values estimated from both absorption and PL are shown in Table 1. As the Sn content increases from 0% to 75%, the estimated Eg decreases from 1.507 to 1.282 eV, and recovers to 1.40 eV for FASnI3. The small band gaps of mixed Pb-Sn perovskites can be attributed to the synergistic effect of the relatively higher valence band maximum of Sn and lower conduction band minimum of Pb. The multi-PL peaks of FAPbI3 and FAPb0.75Sn0.25I3 films were ascribed to phase impurity.
Tab.1 Band gaps for Pb-Sn perovskite alloys calculated from absorption edge and PL peak, respectively |
| FAPb1−xSnxI3 | 0 | 0.25 | 0.50 | 0.75 | 1.00 |
|---|---|---|---|---|---|
| band gap from absorption/eV | 1.490 | 1.310 | 1.270 | − | − |
| band gap from PL/eV | 1.507 | 1.355 | 1.286 | 1.282 | 1.400 |
XRD experiment was conducted to study the structure of FAPb1−xSnxI3 films (Fig. 1(c)). All of the samples show diffraction peaks near 14°, 25°, and 28°, corresponding to the (100), (110), and (200) planes of the cubic structure of FAPb1−xSnxI3. High-temperature post-annealing was employed to convert FAPbI3 and FAPb0.75Sn0.25I3 to cubic structure. FAPbI3 and FAPb0.75Sn0.25I3 films show extra diffraction peaks corresponding to a triclinic structure (Fig. S1), possibly deriving from the high formation energy of a cubic structure, despite the high-temperature post-annealing employed to convert FAPbI3 and FAPb0.75Sn0.25I3 to a cubic structure. As the Sn content increased, the extra peaks disappeared, while a pure cubic structure was observed. We conclude that the addition of Sn reduced the formation energy of the cubic phase [40].
Given its uniform crystal structure and extended absorption range, we explored FAPb0.5Sn0.5I3 films as a light absorber for solar cells based on inverted planar architecture: ITO/perovskite/PCBM/BCP/Al (Fig. 2(a)). In this study, PCBM and BCP serve as the ETL and HTL. According to the cross-sectional scanning electron microscopy (SEM) image (Fig. 2(b)), the thickness of the FAPb0.5Sn0.5I3, PCBM, and BCP films are approximately 270, 55, and 45 nm, respectively. Figure 2(c) shows the band alignment diagram of the HTL-free devices. The conduction band and valence band position are estimated to be 4.00 and 5.25 eV, respectively [40]. The small difference between the work function of ITO and the valence band of the perovskite allows for effective hole injection. As shown in Fig. 2(d), our champion cell with 10% (molar ratio) SnF2 achieves a maximum PCE of 7.94%, with an open-circuit voltage (Voc) of 0.59 V, a short-circuit current density (Jsc) of 23.13 mA/cm2, and a fill factor (FF) of 58%.
From the EQE spectra, the solar cells based on mixed Pb-Sn perovskite exhibit high EQE in a broad absorption range of up to 1050 nm (Fig. 2(e)), and the highest EQE reaches 80.4% at 520 nm. The integrated Jsc from the EQE spectrum (22.57 mA/cm2) matches well with the Jsc obtained in J-V curves measured under solar simulator (with a difference within 3%). This high Jsc demonstrates that holes can be effectively collected by ITO without the HTL. We subsequently investigated the stability of the devices under continuous simulated AM 1.5G illumination (Fig. S2), in a N2-filled glovebox with O2<2 ppm, H2O<1 ppm. Device efficiency slightly increased initially, and then remained steady. The illumination-initiated improvement of efficiency can be ascribed to the drift of ions in perovskites driven by photovoltage-induced electric field, forming anin situ p-i-n homojunction [41].
Fig.2 Photovoltaic structure and performance of HTL-free devices based on FAPb0.5Sn0.5I3 films. (a) Schematics of the device architecture; (b) SEM cross-sectional image; (c) energy band diagram; (d) J-V characteristics; (e) EQE spectrum and integrated Jsc of the highest performance of HTL-free PSC |
To investigate the impact of the energy-level alignment between ITO and perovskite, we prepared a control sample using NiOx as an HTL (Fig. S3). The XRD patterns of perovskite films prepared on ITO and NiOx substrates exhibited similar cubic structures and good crystallinity (see Fig. 3(a)). The Jsc and FF of the PSC with NiOx as HTL, however, are significantly lower compared with the HTL-free one. This condition may result from the deep work function of NiOx, which affected the hole transfer from perovskite. To confirm this speculation, we examined the PL spectra of FAPb0.5Sn0.5I3 films on ITO and NiOx. As illustrated in Fig. 3(b), the emission intensity at 964 nm of the film on ITO is significantly weaker, implying worse hole transport between NiOx/FAPb0.5Sn0.5I3. We further compared the perovskite film morphology through SEM, as shown in Fig. 3(c). The FAPb0.5Sn0.5I3 film exhibits higher surface coverage and larger crystal domain size on bare ITO (left). More pinholes and smaller grains appear when NiOx (right) was employed, which would lead to direct contact between NiOx and ETL and severe interface charge recombination.
SnF2 is generally employed to stabilize the FASnI3 perovskites from oxidation owing to its reducing character [42,43]. The addition of excess amount of SnF2, however, will induce phase separation of the film, thus generating poor device performance [42,43]. We systematically investigated the impact of SnF2 molar concentration on film morphology and device performance. As shown in Figs. 4(a)–4(e), pure FAPb0.5Sn0.5I3 exhibits a full coverage and small grain size. Larger grain size along with pinholes appeared when 5 mol% SnF2 was added. Upon adding 10 mol% of SnF2, rock-like grains with large domain size emerged with a relatively high coverage ratio. When more SnF2 was incorporated, both grain size and coverage rate decreased. The maximum value of the average grain size was around 280 nm with SnF2 ratio of 10 mol% (Fig. 4(f)). We further compared the Sn4+ content in the FAPb0.5Sn0.5I3 films with or without SnF2 additive using X-ray photoelectron spectroscopy. To exclude the effect of surface oxidation, samples were etched before the measurement (Fig. S4 and Table S1). As shown in Table S1, the Sn4+ amount in the 10 mol% SnF2-doped FAPb0.5Sn0.5I3 film is significantly lower than that in pure FAPb0.5Sn0.5I3, implying that the addition of SnF2 prevents Sn2+ from oxidation.
The PSCs based on FAPb0.5Sn0.5I3 with various molar concentrations of SnF2 were fabricated. As shown in Fig. S5 and Table S2, the devices exhibited low Jsc and efficiency without SnF2 additives. This condition can be attributed to the high defect density resulting from the oxidation of Sn2+ to Sn4+, causing significantly reduced carrier-diffusion lengths [42–44]. With the addition of SnF2, the device performance is effectively improved by suppressing Sn oxidation. However, film morphology deteriorated when the concentration of SnF2 is increased to 20%, leading to poor performance. Therefore, both oxidation and film morphology are important to device performance.
Conclusion
In this study, we constructed an inverted planar mixed Pb-Sn perovskite solar cell on bare ITO glass without an independent HTL. The Pb-to-Sn ratio and the concentration of SnF2 were systematically optimized. Highly uniform reflective FAPb0.5Sn0.5I3 thin films on ITO glass with a band gap of 1.28 eV were achieved. Solar cells based on HTL-free ITO/FAPb0.5Sn0.5I3/PCBM/BCP/Al structure exhibit a promising PCE of 7.94% with a short-circuit current density of 23.13 mA/cm2, EQE of 80%, and extended light-harvesting range to 1050 nm. The removal of HTL allows effective carrier transfer between the perovskite and the ITO electrode. It therefore provides an alternative simple structure for low-temperature, small band gap single-junction or tandem PSCs.