No abstract

Efficient wide-bandgap perovskite solar cells (PSCs) enable high-efficiency tandem photovoltaics when combined with crystalline silicon and other low-bandgap absorbers. However, wide-bandgap PSCs today exhibit performance far inferior to that of sub-1.6-eV bandgap PSCs due to their tendency to form a high density of deep traps. Here, we show that healing the deep traps in wide-bandgap perovskites—in effect, increasing the defect tolerance via cation engineering—enables further performance improvements in PSCs. We achieve a stabilized power conversion efficiency of 20.7% for 1.65-eV bandgap PSCs by incorporating dipolar cations, with a high open-circuit voltage of 1.22 V and a fill factor exceeding 80%. We also obtain a stabilized efficiency of 19.1% for 1.74-eV bandgap PSCs with a high open-circuit voltage of 1.25 V. From density functional theory calculations, we find that the presence and reorientation of the dipolar cation in mixed cation–halide perovskites heals the defects that introduce deep trap states.

Impressive progress in halide perovskite solar cells motivates further work to improve operating stability. It is known that ion-migration-driven decomposition represents a degradation pathway in perovskite solar cells and that it can occur within the perovskite material even in well-encapsulated devices. Here we find that quasi-two-dimensional (2.5D) perovskites suppress this ionmigration-induced degradation. Using TOF-SIMS, we confirm that iodide migration occurs in bulk perovskite photovoltaic devices operating at their maximum power point (MPP). We observe that iodine ions migrate across the spiro-OMeTAD layer to the spiro/ gold contact interface, oxidizing and deteriorating the gold at the interface. In contrast, we find that large ⟨n⟩ 2.5D perovskites exhibit a significantly reduced rate of ion migration compared to 3D devices and exhibit less than 1% relative PCE loss in over 80 h of continuous operation at MPP, whereas the PCE of 3D devices diminishes by more than 50% within the first 24 h.