Efficient Wide-Bandgap Mixed-Cation and Mixed-Halide Perovskite Solar Cells by Vacuum Deposition

Vacuum deposition methods are increasingly applied to the preparation of perovskite films and devices, in view of the possibility to prepare multilayer structures at low temperature. Vacuum-deposited, wide-bandgap solar cells based on mixed-cation and mixed-anion perovskites have been scarcely reported, due to the challenges associated with the multiple-source processing of perovskite thin films. In this work, we describe a four-source vacuum deposition process to prepare wide-bandgap perovskites of the type FA1–nCsnPb(I1–xBrx)3 with a tunable bandgap and controlled morphology, using FAI, CsI, PbI2, and PbBr2 as the precursors. The simultaneous sublimation of PbI2 and PbBr2 allows the relative Br/Cs content to be decoupled and controlled, resulting in homogeneous perovskite films with a bandgap in the 1.7–1.8 eV range and no detectable halide segregation. Solar cells based on 1.75 eV bandgap perovskites show efficiency up to 16.8% and promising stability, maintaining 90% of the initial efficiency after 2 weeks of operation.

were sublimed at varying deposition rates, in the range of 0.25-0.45 Å/s and 0.07-0.22 Å/s, respectively. During the perovskite deposition, the pressure of the chamber was mantained at 8·10 -6 mbar and the substrates were kept at room temperature. Typical sublimation temperatures for the precursors were 150 ºC for FAI, 310 ºC for PbI 2 , 280-300 ºC for PbBr 2 and 490 -520ºC for CsI.
Characterization. Absorption spectra were collected using fiber optics based Avantes Avaspec2048 Spectrometer. The photoluminescence spectra were measured with an Avantes Avaspec2048 spectrometer and films were illuminated with a diode laser of integrated optics, emitting at 522 nm. All the spectra were collected with an integration time of 1 s. The crystalline structure of the powder and film samples was studied by X-ray diffraction (XRD). The patterns were collected in Bragg-Brentano geometry on an Empyrean PANalytical powder diffractometer with a copper anode operated at 45 kV and 40 mA. Further analysis including Le Bail fits were performed with Fullprof software. Scanning Electron Microscopy (SEM) images were performed on a Hitachi S-4800 microscope operating at an accelerating voltage of 2 kV over platinummetallized samples.
High resolution x-ray photoemission spectroscopy (XPS, Sceinta Omicron SES-100, nonmonochromatic Al−Kα = 1486.6 eV) measurements were performed to quantify the chemical composition of the top surface of the perovskite film. The peak fitting and atomic percentage calculations were performed using CasaXPS 2.3.16 software. Shirley background lines and gaussian-lorentzian shape lines were used for fitting the components. The binding energy (BE) for XPS was calibrated by measuring Fermi edge (E F = 0 eV) and Au-4f 7/2 (84.0 eV) on a clean Au surface. The x-ray gun was operated at 250W. The estimated energy resolution of XPS is 0.7 eV. XPS measurements were performed in an ultra-high vacuum (UHV) chamber with pressure of about 1× 10 -10 mbar. The acquisition was performed with pass energy of 20 eV and dwell time of 0.1 sec. X-ray-induced sample damage was monitored by taking several consecutive spectra and comparing those spectra. In case of no damage observed, all the collected spectra for each core level were averaged into the final high resolution spectrum. Time acquisition for each scan varied from 40 to 70 sec depending on the core level region. To minimize any possible unnecessary x-ray exposure time, special care was taken while optimizing and acquiring XPS measurements. No X-ray-induced damage was observed on the films. All the calculated atomic For the sensitive EQE measurements, the cell was illuminated by a Quartz-Tungsten-Halogen lamp (Newport Apex 2-QTH) through a monochromator (Newport CS130-USB-3-MC), a chopper at 279 Hz and a focusing lens. The device current was measured as a function of energy from 2.1 eV to 1.2 eV in 0.02 eV steps using a lock-in amplifier (Stanford Research Systems SR830). The system was calibrated and the solar spectrum mismatch was corrected using a calibrated Silicon reference cell. Solar cell stability measurements were recorded using a maximum power point tracker (mppt) system, with a white LED light source under 1 sun equivalent, developed by Candlelight. During the mppt measurements, a flow of N 2 gas was used and temperature was kept at 300 K using a water-circulating cooling system. In order to confirm that perovskite films obtained with CsI deposition rates > 0.3 Å/s are photostable, we have fit the PL spectrum in Figure S1c. The generalized Planck's law describes the radiation of a non-black body under a potential difference (quasi Fermi-level splitting, ) ∆ and reads in the Wien-approximation where is the photon's energy, is Planck's constant, is the speed of light, is the ℎ ( ) absorptance of the semiconductor and is the thermal energy. From this, one can see that the slope of the PL is determined by the term in the low-wavelength region while it is To not be limited by the measurement noise starting at 750 nm, we extended the Urbach slope of the EQE down to 900 nm. The resulting theoretical spectrum assuming a temperature of 300 K is essentially identical to the measured PL. The small red-shift of 5 nm is likely due to a small discrepancy in the wavelength-calibration of the two measurement setups or to batch-to-batch variations in the perovskite deposition. Importantly, the quasi Fermi-level splitting does not influence the shape of the spectrum but just its intensity. We can hence conclude that we cannot see any additional phase from the PL spectra for perovskites obtained with CsI deposition rate > 0.3 Å/s.          (1) The slope of this decay is called the Urbach energy, , and indicates the structural and thermic disorder in the absorber. 2,3 Here, we assumed that the EQE is proportional to the absorption coefficient and obtained from a fit of the exponential decay of the EQE.
Furthermore, one can find an effective bandgap by looking at the derivative of the EQE. For the ideal semiconductor as in the Shockler-Queisser theory (SQ), this should be just a Dirac delta function at the bandgap, i.e. an infinitely high and infinitely narrow peak at the bandgap.
Following Rau et al., 4 can be interpreted as a distribution of SQ-bandgaps for a real semiconductor. It often looks similar to a Gaussian peak, so we extract the bandgap as the central energy of a Gaussian fit.
The radiative limit of the V oc can be calculated with the measured J sc and with the radiative limit of the dark saturation current, (2) This value includes the voltage losses because of difference between the measured and the theoretical possible value in the SQ theory as well as the radiative losses originating from the finite absorption tail. 4 To calculate the latter (corresponding to ), we can use again the 0, absorption tail of the EQE as a product with the black-body radiation at 300 K to calculate the luminescence spectrum at equilibrium (reciprocity relation) and integrate this: In the SQ theory is calculated by replacing the EQE spectrum with an step-function which 0, yields lower values for and thus a higher Voc. 0, Figure S12. a) External quantum efficiency (EQE) spectrum (red) as in Fig. 4a Figure S13. PL spectra of a perovskite film on glass, with and without the charge transport layers used in the solar cells. The relative PLQY for each variation is calculated by integration of the spectra and normalized to the one of the bare perovskite (PLQY =1). The corresponding QFLS difference calculated for each layout is also reported.