PTAA (average M_n 7×10³−1×10⁴), BCP, lead iodide (PbI₂, 99.999% trace metals), dimethyl sulfoxide (DMSO), L-α-phosphatidylcholine (LP), benzylhydrazine hydrochloride, 2-methoxyethanol (2-ME), and toluene were purchased from Sigma-Aldrich and used without further purification. C₆₀ and PCBM were purchased from Nanjing Zhiyan Inc. Methylammonium iodide (MAI), formamidinium iodide (FAI), 4-fluoro-phenyethylammonium iodide (p-F-PEAI) and n-dodecylammonium iodide were purchased from GreatCell Solar. All chemicals without any notes were commercially available and used without further purification.

Patterned ITO glass substrates (1.5 cm × 1.5 cm) were first cleaned by ultrasonication with soap, deionized water and isopropyl alcohol, and then UV-ozone treated for 15 min before use. All perovskite solar devices were prepared by blade-coating at room temperature inside a fume hood with a relative humidity of 40%±10%. The PTAA solution at a concentration of 3.3 mg·mL⁻¹ dissolved in toluene was blade-coated onto ITO glass substrates at a speed of 20 mm·s⁻¹. The gap between blade-coater and ITO substrates was 150 μm. 1.35 mol/L MA_0.7FA_0.3PbI₃ precursor solution was prepared by dissolving corresponding organic halide salts and lead iodide into 2-ME in a N₂-filled glovebox with an O₂ level <10 ppm. 0.83 mg·mL⁻¹ n-dodecylammonium iodide, 0.27 mg·mL⁻¹ LP, 0.14 vol% MAH₂PO₂, 1.40 mg ml⁻¹ p-F-PEAI, 0.15 mg·mL⁻¹ benzylhydrazine hydrochloride, 2.8 vol% DMSO were added as additives before blade-coating. Subsequently, the precursor solution was blade-coated onto the PTAA-covered ITO glass substrates with a gap of 250 μm at a movement speed of 20 mm·s⁻¹. The air knife worked at 20 psi during blade-coating. After that, the perovskite films were annealed at 120 °C for 10 min in air. The filtered PCBM or PFBCP ETL solutions (20 mg·mL⁻¹ in 1,2-dichlorobenzene) were coated onto perovskite layers at a speed of 20 mm·s⁻¹. The optimal thickness of ETLs is around 30 nm. For PCBM-based PSCs, a thin layer of BCP (5 nm) was evaporated on the top of PCBM (0.1 Å·s⁻¹). The solar cells were completed by thermally evaporating 100 nm copper (1 Å·s⁻¹).

¹H-NMR and ¹³C-NMR spectra were obtained on a 400 MHz Bruker AVANCE III-400 spectrometer. High-resolution mass spectrometry (HRMS) was done on a Bruker Daltonics MALDI TOF UltrafleXtreme. Thermogravimetric analysis (TGA) was conducted using a Netzsch STA449F3 thermogravimetric-differential scanning calorimetric thermal synchronous analyzer. The sample was heated from 30 °C to 600 °C at a heating rate of 10 °C·min⁻¹ under a nitrogen atmosphere. UV-Vis absorption spectra were obtained with a Shanghai Metash UV-8000 spectrometer. Cyclic voltammetry was carried out on a ZAHNER IM6EX electrochemical workstation with a three-electrode configuration, using Ag/AgCl as the reference electrode, a Pt plate as the counter electrode, and a glassy carbon as the working electrode. The ETL solutions were drop-cast onto the electrodes to form thin films. 0.1 mol·L⁻¹ tetrabutylammonium hexafluorophosphate in anhydrous acetonitrile was used as the supporting electrolyte. Potentials were referenced to the ferrocenium/ferrocene couple by using ferrocene as external standards in acetonitrile solutions. The scan rate is 0.1 V·s⁻¹. Both steady-state PL and TRPL spectra were acquired on a HORIBA FL-3 fluorescence spectrophotometer at room temperature. The excitation wavelengths were 480 nm and 405 nm for PL and TRPL measurements, respectively. The TRPL curves were fitted with a bi-exponential formula:

$$y=y_0+A_1\exp (-(x-x_0)/t_1)+A_2\exp (-(x-x_0)/t_2)(t_2>t_1)$$

1

where the t₁ values of the two systems were found to be comparable (Table S1 in the electronic supplementary information, ESI), and here we used t₂ to represent the lifetime of the films.

Atomic force microscopy (AFM) images were scanned from a Bruker Icon Atomic Force Microscope. Scanning electron microscopy (SEM) images were taken on TESCAN MIRA3 LMH operating at 5.0 kV. The J-V characteristics of solar cells were performed using an LED solar simulator (BG-LED3A-100S, Class AAA Solar Simulator) and the power of the simulated light was calibrated to 100 mW·cm⁻² by a silicon reference cell (Newport 91150V). A metal mask with an aperture (6.8 mm²) aligned with the device area was used for measurements. All devices were measured using a Keithley 2400 source meter with a backward scan rate of 0.1 V·s⁻¹ in air at room temperature, and the delay time was 10 ms. There was no preconditioning before measurement. The PSCs were encapsulated with cover glass sealed by an epoxy encapsulant on the back. After curing, the stability of encapsulated cells was monitored with a 91 PVKSOLAR MSCLT-1 automatic maximum power point (MPP) tracker. After connecting the PSCs to the MPP tracker, the cells kept working at MPP conditions under simulated AM 1.5G one sun illumination (100 mW·cm⁻²) in air (relative humidity: 40%±10%). The stabilized PCEs were recorded every 5 s. No fan or cooler was used to control the cell temperature, and the temperature of the cells was measured to be about 40 °C.

The electron mobilities were measured using the space-charge-limited-current (SCLC) method, employing a device architecture of glass/ITO/ZnO/ETL/PDINN/Ag. The mobilities were obtained by taking current-voltage curves and fitting the results to a space charge limited form, where the SCLC is described by:

$$J=\frac{9{ϵ}_0{ϵ}_{\mathrm{r}}\mu (V_{\mathrm{a}\mathrm{p}\mathrm{p}\mathrm{l}}-V_{\mathrm{b}\mathrm{i}}-V_{\mathrm{s}})}{8L^3}$$

2

where ε₀ is the permittivity of free space, ε_r is the relative permittivity of the material (assumed to be 3), μ is the electron mobility, V_appl is the applied voltage, V_bi is the built-in voltage (0.7 V), V_s is the voltage drop from the substrate’s series resistance (V_s = IR, R is measured to be 10.8 Ω) and L is the thickness of the film. By linearly fitting J^1/2 with V_appl-V_bi − V_s, the mobilities were extracted from the slope and L:

$$\mu =\frac{{\mathrm{S}\mathrm{l}\mathrm{o}\mathrm{p}\mathrm{e}}^{2}8L^3}{9{ϵ}_0{ϵ}_{\mathrm{r}}}$$

3

tDOS spectra were obtained with an Agilent E4980A precision LC meter. The energetic profile of tDOS was derived from the angular frequency dependent capacitance using the equation:

$$N_{\mathrm{T}}\left(E_{\omega }\right)=-\frac{1}{q{k}_{\mathrm{B}}T}\frac{\omega \mathrm{d}C}{\mathrm{d}\omega }\frac{V_{\mathrm{b}\mathrm{i}}}{W}$$

4

where C is capacitance, ω is the angular frequency, q is the elementary charge, k_B is the Boltzmann’s constant and T is the temperature. V_bi and W are the built-in potential and depletion width, respectively, which were extracted from the Mott-Schottky analysis. The applied angular frequency ω defines an energetic demarcation,

$$E_{\omega }=kT\ln (\frac{{\omega }_0}{\omega })$$

5

where ω₀ is the attempt-to-escape angular frequency (ω₀ = 2πν₀, ν₀=1×10⁹ Hz). The trap states below the energy demarcation can capture or emit charges with the give ω and contribute to the capacitance.