With the efforts of scientists around the world, the power conversion efficiency (PCE) of perovskite solar cells (PSCs) has reached 25.7%. To further improve the efficiency and break through the Shockley-Queisser (S-Q) limit, it is promising to construct all-perovskite tandem solar cellsvia combining wide-bandgap and narrow-bandgap perovskites^[1-5]. As the key light-harvesting material for the bottom cell in all-perovskite tandem devices, the narrow-bandgap Pb–Sn mixed perovskites have attracted increasing interest in recent years^[6-8]. However, the Pb–Sn mixed perovskites suffer from uncontrollable crystallization, easy oxidation of Sn²⁺ and high defect density, which significantly limit PCE improvement^[9,10]. Organic ammonium halides can improve the efficiency and stability of Pb–Sn mixed PSCs.

Besides forming 2D structures, organic ammonium halides are also effective passivators. Leeet al. introduced butylammonium (BA) halide (2%) into FA_0.83Cs_0.17Pb_0.5Sn_0.5I₃ precursor, and they found that BA cations allow self-aggregation on the surface and at the bottom of perovskite films^[18]. The BA cations on the film surface can suppress the diffusion of Ag into perovskite film, while BA cations at the buried interface can suppress the perovskite degradation induced by sulfonic acid groups in PEDOT:PSS. Most importantly, BA cations can induce a more phase-pure perovskite film with (100)-preferred orientation, reducing the trap states. With these merits, BA-based FA_0.83Cs_0.17Pb_0.5Sn_0.5I₃ solar cells gave a PCE of 18.66%. Lianget al. proposed a selective targeting anchor (STA) strategyvia jointly employing PEAI and ethylenediamine diiodide (EDAI₂) to passivate surface defects of Pb–Sn perovskite films^[19]. Combining DFT calculations and optoelectronic techniques, they demonstrated that PEA⁺ and EDA²⁺ cations can selectively anchor [PbI₆]⁴⁻ and [SnI₆]⁴⁻ octahedron, respectively, through filling A-site vacancies (Fig. 2(a)). The STA strategy yielded a champion PCE of 22.51%, and the cells can retain 80% of the initial PCE after being stored in N₂ glovebox for 2700 h. Huet al. modified the bottom surface of Cs_0.1FA_0.6MA_0.3- Pb_0.5Sn_0.5I₃ filmvia adding glycine hydrochloride (GlyHCl) into the perovskite precursor, and GlyH⁺ cations can self-assemble at the bottom surface to passivate trap states^[20]. They further passivated the surface defects by using EDAI₂. GlyH⁺ and EDA²⁺ cations can form interface dipoles to facilitate charge extraction. A PCE of 23.6% was obtained, which is the record for Pb–Sn PSCs. In our previous work, we introduced a trace amount of propanediamine diiodide (PDAI₂) into FA_0.7MA_0.3Pb_0.5Sn_0.5I₃ precursor to induce oriented crystal growth^[21]. PDA cations can anchor onto the nuclei to induce (100)-preferred orientation through strong interactions between PDA cations and [PbI₆]⁴⁻ and [SnI₆]⁴⁻ octahedron. Then, the nuclei would act as a surface template to modulate the crystal growth along (100) orientation, and PDA cations kept on the crystal surface (Fig. 2(b)). The DFT calculations and experimental characterizations confirmed that the (100)-oriented perovskite crystals have reduced trap states and higher carrier mobilities. Consequently, the PCE for PDA-based solar cells was improved from 16.62% to 20.03%. Owing to the relatively low light absorption coefficient, highly efficient Pb–Sn PSCs require thick (~1μm) perovskite films. However, the short carrier diffusion lengths caused by severe trap states limit the carrier transport in thick Pb–Sn perovskite films. Tan et al. developed a ~1.2μm thick Pb–Sn perovskite film with long diffusion length exceeding 5μm^[22]. PEA, phenylammonium (PA) and 4-trifluoromethyl-phenylammonium (CF3-PA) cations were employed as passivators in the precursor solution. DFT calculations revealed that electrostatic potentials at the -NH₃⁺ terminal of the three cations are different: PEA < PA < CF3-PA. High electrostatic potential is beneficial for molecules anchoring onto perovskite crystals. During the annealing process of perovskite films, CF3-PA cations showed the strongest tendency to anchor on the crystal surfacevia filling A-site vacancies (Fig. 2(c)). Meanwhile, the strong interaction between CF3-PA and perovskite can also suppress the formation of iodine vacancies, I_Sn and I_Pb antisite defects. The significantly reduced trap states contributed to enhanced carrier lifetimes and diffusion lengths, yielding a PCE of 22.2% for single-junction solar cells and a PCE of 26.4% for tandem cells.

In short, organic ammonium halides can modulate crystal growth, passivate trap sites and modify interfaces of Pb–Sn mixed perovskite films to enhance the device performance. Rational molecular design of organic ammonium halides can help to develop efficient Pb–Sn mixed PSCs, applicable in all-perovskite tandem solar cells.

Organic ammonium halides can be employed as additives to modulate the crystallization of Pb–Sn perovskite films, or passivate surface defectsvia post-treatment. Tonget al. found that the incorporation of guanidinium thiocyanate (GuaSCN) in (FASnI₃)_0.6(MAPbI₃)_0.4 perovskite film can significantly improve the optoelectronic properties^[11]. SCN^– can increase grain size and improve film morphology, while Gua⁺ can participate in constructing 2D phases at grain boundaries to prevent Sn vacancy diffusion and protect the film from oxygen erosion. The optimized film offered a prolonged carrier lifetime (>1μs) (Fig. 1(a)), yielding PCEs of 20.5%, 25% and 23.1% for single-junction solar cell, four-terminal (4-T) and two-terminal (2-T) tandem devices, respectively. Similarly, Zhouet al. added 12% guanidinium bromide (GABr) into FA_0.7MA_0.3Pb_0.7Sn_0.3I₃ perovskite film with a bandgap of 1.34 eV^[12]. They found that GABr can effectively reduce the defect density and facilitate charge transport. Consequently, GABr-modified Pb–Sn PSCs gave a PCE of 20.63% with enhanced environmental and thermal stability. To achieve vertically aligned crystals, Liet al. introduced 2-(4-fluorophenyl)ethylammonium iodide (FPEAI) into (MAPbI₃)_0.75(FASnI₃)_0.25 to form 2D/3D structure to induce oriented growth^[13]. FPEAI-based perovskite film presented (110)-preferred orientation (Fig. 1(b)), which is beneficial for effective charge transport and extraction. Meanwhile, the 2D/3D structure can suppress Pb–Sn phase segregation. As a result, the 2D/3D hybrid Pb–Sn PSCs delivered a PCE of 17.51% with superior stability. Compared with Ruddlesden-Popper (R-P) 2D phases formed by monovalent organic cations, the interlayer distance of Dion-Jacobson (D-J) 2D phases achieved by divalent organic cations is much shorter, and D-J 2D phase is less resistant to charge transfer^[14]. Keet al. developed a D-J 2D structure using a divalent 3-(aminomethyl)piperidinium (3AMP) spacer for MA_0.5FA_0.5Pb_0.5Sn_0.5I₃ perovskite films, affording a longer carrier lifetime of 657.7 ns and a PCE of 20.09% with an open-circuit voltage (V_oc) of 0.88 V^[15]. To avoid excess formation of 2D phases which could block charge transfer, Weiet al. dexterously designed an ultrathin 2D layer capping Pb–Sn perovskite film surface^[16]. Ethyl acetate (EA) with 0.5 mg/mL phenethylammonium iodide (PEAI) was applied as anti-solvent in film preparation. This approach not only passivated surface defects, but also avoided excess formation of 2D phases, yielding PCEs of 19.4% and 23.7% for single-junction and 2-T tandem solar cells, respectively. Low-dimensional structure treatment by using organic ammonium halides has been widely used to improve the performance of PSCs. However, most post-treatment results in the formation of one layer (1L) 2D structure on film surface, which could severely impede charge extraction. Ninget al. designed a new molecule 2-thiopheneethylamine thiocyanate (TEASCN) to construct bilayer (2L) quasi-2D structure that allowed effective charge transfer^[17]. For comparison, 2-thiopheneethylamine iodide (TEAI) was also used for surface treatment. Interestingly, though both TEAI and TEASCN treatment formed 1L structure during spin-coating process, the 1L structure of TEASCN film transformed to 2L structure after annealing process, while 1L structure of TEAI film remained on the surface (Figs. 1(c) and1(d)). The density functional theory (DFT) calculation revealed that the formation energy from TEA₂SnI₂SCN₂ to TEA₂FASn₂I₅SCN₂ is close to zero, suggesting the easy transformation from 1L to 2L, which explains the formation of the 2L structure. The 2L structure on the surface can not only ensure effective charge transfer, but also improve thermal stability of Pb–Sn perovskite films. TEASCN-treated Pb–Sn PSCs offered a PCE of 21.1%.