Reactive Passivation of Wide-Bandgap Organic-Inorganic Perovskites with Benzylamine

Abstract

While amines are widely used as additives in metal-halide perovskites, our understanding of the way amines in perovskite precursor solutions impact the resultant perovskite film is still limited. In this paper, we explore the multiple effects of benzylamine (BnAm), also referred to as phenylmethylamine, used to passivate both FA_0.75Cs_0.25Pb(I_0.8Br_0.2)₃ and FA_0.8Cs_0.2PbI₃ perovskite compositions. We show that, unlike benzylammonium (BnA⁺) halide salts, BnAm reacts rapidly with the formamidinium (FA⁺) cation, forming new chemical products in solution and these products passivate the perovskite crystal domains when processed into a thin film. In addition, when BnAm is used as a bulk additive, the average perovskite solar cell maximum power point tracked efficiency (for 30 s) increased to 19.3% compared to the control devices 16.8% for a 1.68 eV perovskite. Under combined full spectrum simulated sunlight and 65 °C temperature, the devices maintained a better T₈₀ stability of close to 2500 h while the control devices have T₈₀ stabilities of <100 h. We obtained similar results when presynthesizing the product BnFAI and adding it directly into the perovskite precursor solution. These findings highlight the mechanistic differences between amine and ammonium salt passivation, enabling the rational design of molecular strategies to improve the material quality and device performance of metal-halide perovskites.

Figure 1

(a) UV–vis absorption spectra (left) and PL emission spectra (right) of FA_0.75Cs_0.25Pb(I_0.8Br_0.2)₃ with 0.1–5.0 mol % BnAm additive. (b) Statistical box plot of photoluminescence quantum yield (PLQY) of FA_0.75Cs_0.25Pb(I_0.8Br_0.2)₃ with 0.1–5.0 mol % BnAm additive measured with a 532 nm laser at 100 mW/cm² (AM 1.5) intensity. Four points were measured on each film. (c) The PL spectra of the reference sample over time. (d) The PL spectra of perovskite with 0.3 mol % BnAm additive over time. Samples were measured with a 532 nm laser under 1300 mW/cm² intensity.

Figure 2

(a) ¹H NMR spectrum of the reaction product of a BnAm and FAI mixture in DMSO-d₆. (b) ¹³C{¹H} NMR spectrum of the BnAm and FAI reaction mixture in DMSO-d₆. (c) HSQC spectrum of the solution. (d) Schematic of reaction between FAI and BnAm, yielding the Z-BnFA⁺ and E-BnFA⁺ isomers. (e) ToF-SIMS molecular distribution of perovskite film with 0.6 mol % of BnAm additive. See the molecular distribution of a control perovskite film in Figure S28. Note that the Pb–NH₃ fragment is absent in the control film.

Figure 3

¹H magic-angle spinning (MAS) NMR spectra of FAPbI₃ + 5 mol % BnAm fabricated mechanosynthetically by liquid-assisted grinding (a), FA_0.75Cs_0.25Pb(I_0.8Br_0.2)₃ + 0.3 mol % BnAm thin films (b), and FAPbI₃ thin films (c) (23.5 T, 55 kHz). Insets highlight peaks corresponding to the BnFA⁺ additive. Spectra shown in full are quantitative. Insets are acquired using a short recycle delay to highlight the more rapidly relaxing additive nuclei. ¹H-decoupled ¹³C MAS spectra of FAPbI₃ + 5 mol % BnAm fabricated by liquid-assisted grinding (d) and BnFAI (e) (11.7 T, 15 kHz). Spectra were acquired either by ¹H–¹³C cross-polarization (CP) or using a Hahn echo pulse sequence, to highlight the rigid and mobile organic species, respectively. (f) ¹H–¹H spin diffusion (SD) spectrum of FAPbI₃ + 5 mol % BnAm fabricated by liquid-assisted grinding. Sections of cumulative 1D projections shown by dotted lines are enlargements of the corresponding region. (g) ¹²⁷I nuclear quadrupole resonance (NQR) spectra of the radiofrequency region containing NQR transitions of ¹²⁷I nuclei within α-FAPbI₃ phase. All thin film materials are mechanically exfoliated to give a powder prior to measurement. (h) Schematics highlighting the modes of BnFA⁺ incorporation into perovskite materials (left) and the additive concentration-dependence of these incorporation modes. Molecular structures are adapted from VESTA.3 Software.^,

Figure 4

(a) Schematic of the p–i–n device structure used. (b) Statistical histogram of 1.68 eV perovskite p–i–n devices parameters without additive, perovskite with 0.3 mol % BnAm additive, and perovskite with 0.3 mol % BnFAI additive (15 devices each with Me-4PACz as hole-transporting layer (HTL) and antireflective coating). (c) 30 s of maximum power point tracking efficiencies (η_MPP) of FA_0.75Cs_0.25Pb(I_0.8Br_0.2)₃ champion devices for the three conditions: control, perovskite with 0.3 mol % BnAm, and perovskite with 0.3 mol % BnFAI replacing FA⁺. (d) 65 °C AM 1.5G aging of p–i–n devices with the 1.68 eV FA_0.75Cs_0.25PbI_0.8Br_0.2 (with Poly-TPD as HTL and encapsulated). The colored area denotes the standard deviation. The BnAm bulk condition is perovskite with 0.3 mol % additive. Devices were aged under open-circuit voltage conditions. Each data point is the average η_MPP of 6 devices for each condition. The data is normalized to show T₈₀.