Template‐Assisted Formation of High‐Quality α‐Phase HC(NH2)2PbI3 Perovskite Solar Cells

Abstract Formamidinium (FA) lead halide (α‐FAPbI3) perovskites are promising materials for photovoltaic applications because of their excellent light harvesting capability (absorption edge 840 nm) and long carrier diffusion length. However, it is extremely difficult to prepare a pure α‐FAPbI3 phase because of its easy transformation into a nondesirable δ‐FAPbI3 phase. In the present study, a “perovskite” template (MAPbI3‐FAI‐PbI2‐DMSO) structure is used to avoid and suppress the formation of δ‐FAPbI3 phases. The perovskite structure is formed via postdeposition involving the treatment of colloidal MAI‐PbI2‐DMSO film with FAI before annealing. In situ X‐ray diffraction in vacuum shows no detectable δ‐FAPbI3 phase during the whole synthesis process when the sample is annealed from 100 to 180 °C. This method is found to reduce defects at grain boundaries and enhance the film quality as determined by means of photoluminescence mapping and Kelvin probe force microscopy. The perovskite solar cells (PSCs) fabricated by this method demonstrate a much‐enhanced short‐circuit current density (  J sc) of 24.99 mA cm−2 and a power conversion efficiency (PCE) of 21.24%, which is the highest efficiency reported for pure FAPbI3, with great stability under 800 h of thermal ageing and 500 h of light soaking in nitrogen.


Supporting Information
. In situ X-ray diffraction of the C-FAPbI 3 (fabricated via conventional method) films in vacuum. The temperature increases at 5 °C/min, and the soaking time of each scan is 10 min with a characterization speed of 10 °/min.            All of the area is 10 × 10 µm 2 each. The TRPL spectrum demonstrated a substantial carrier-hole recombination in the perovskite bulks ( Figure S14a), and the SCLC showed a much higher cut-off voltage (0.78 V, Figure S14b), which revealed massive trap-states at the interface. Meanwhile, the EIS showed a lower recombination resistance that stated weak resisting ability for the recombination losses. The spectrum of DOS in Figure S14d presented a higher magnitude order of C-FAPbI 3 . In summary, for C-FAPbI 3 , it was hard to reduce trap-states in the perovskite films.   When aged at room temperature (25 ºC), the efficiency decreased to ~75% after 1056 h, which was state-of-art. When aged at 85 ºC, the efficiency decreased to ~75% after 800 h, while, it reduced to ~50% when aged at 140 ᵒC. However, it decayed quickly to 0 when aged at 195 ºC. Besides, it reduced to ~70% after 500 h under the light soaking. The C-FAPbI 3 showed a poor thermal stability than that of FAI-20/FAI-30 devices because the defects and δ-phase in the films.

Device fabrication
Devices with theFTO\TiO 2 \FAPbI 3 \Spiro-OMeTAD\Au structure is fabricated. FTO glasses were firstly etched with zinc powder and hydrochloric acid (HCl) and then washed by detergent, deionized water, ethanol and acetone The cleaned FTO glasses were placed on a hotplate at the temperature of 500 °C. To deposit TiO 2 layer, 10 mL of an acetonitrile/ethanol (with 95:5 volume ratio) solution, nickel acetylacetonate (with 15 mol% magnesium acetatetetrahydrate and 5 mol% lithium acetate, totally 0.02 mol L-1) was sprayed, and TiO 2 layer was obtained as a compact layer. After annealing at temperature of 460 °C for 20 min, TiO 2 paste (the weight ratio of TiO 2 and ethanol is 1:6) was spin-coated on the compact layer at the spin-speed of 5000 rpm as to deposit mesoporous scaffold layer. Later, it was annealed at the temperature of 510°C for 30 min. The perovskite layer is fabricated as mostly reported. Firstly, PbI 2 /MAI (1:1) was dissolved in 1.2 mol DMSO/DMF (4:1) solution and the mixed solution was dropped on the mesoporous layer and then spun at a speed of 2100 rpm for 10 seconds and then 30 seconds. During the 2 nd spinning, chlorobenzene (CB) was dropped. Then, 40 µL solution of FAI dissolved in the IPA solution was dropped on the as prepared film. Subsequently, the as-prepared films were heated at 100 °C for 15 min and 140 °C for 60 min. Finally, the Spiro-OMeTAD was spin-coated on the perovskite layer, after which the gold electrode was thermally evaporated. The conventional method was the two-step sequential deposition process. A thin layer of PbI 2 was first deposited on the substrate and formamidinium iodide (FAI) was then applied to the predeposited PbI 2 to enable conversion to the perovskite phase, which was then annealled at 150 ᵒC.

Device characterization
The PCE and J-V curves were measured with a Keithley 2400 source-meter with a sunlight simulator (XES-300T1, SAN-EI Electric, AM 1.5), which is calibrated using a standard silicon reference cell. The solar cells were masked with a black aperture cover to a define active area of 0.09 cm 2 . Incident photon to current efficiency (IPCE) was tested as a function of wavelength from 300 nm to 900 nm (Enli Technology), with dual Xenon/quartz halogen light source, measured in DC mode with no bias light.
The setup was calibrated with a certified silicon solar cell prior to measurement.
Atomic force microscopy (AFM) images were acquired in tapping mode with a 5500 Co., LTD), which equipped with a Class AAA solar simulator; <420 nm UV light was cut off with an optical filter. The thermal stability was tested with controlled temperature at 85 °C in an electronic constant temperature/humidity chamber (THR030FA, Advantec Co., Ltd.). The absorption spectra were recorded using UV/Vis spectrometer (Shimadzu, UV-3600) in the 300 nm ~ 900 nm range. Confocal PL mapping was carried out with a laser confocal Raman spectrometer (Princeton Instruments, Acton Standard Series SP-2558) and a 485 nm laser (PicoQuant LDH-P-C-485, 0.4 mW with a 1% optical density filter), using a home-built confocal microscope on a 10 × 10 μm 2 sample area. The light intensity dependence measurements and transient photo-voltage decay measurements were performed on an electrochemical workstation (Zahner). A white LED with an intensity of 1000W m -2 was used as the light source in the experiments, devices were soaked in light for 2s before the LED light was turned off.