In recent years, perovskite solar cells (pero-SCs) have attracted significant attention for their potential application in photovoltaics, with single-junction pero-SCs achieving power conversion efficiencies (PCEs) over 26%. To further enhance PCE and overcome the limits of single-junction devices, researchers have exploited the tunable bandgap perovskites to develop tandem solar cells (TSCs), such as perovskite/silicon, perovskite/perovskite and perovskite/organic TSCs. Among these, perovskite/organic TSCs have emerged as a promising alternative and offered advantages in fabrication and stability, which benefit from fully solution-processable active layers in comparison with the perovskite/silicon TSCs, and avoid oxidation issues typically associated with narrow-bandgap perovskites in the perovskite/perovskite TSCs. Additionally, short-wavelength UV light is known to accelerate the degradation of organic solar cells, while perovskites serve as effective UV filters protecting the organic layer in the perovskite/organic TSCs. Furthermore, the organic subcell functions as an encapsulation layer, shielding the perovskite from moisture and oxygen, which further enhances stability of the devices. Consequently, the perovskite/organic TSCs exhibit superior stability compared to their individual single-junction counterparts, significantly broadening their potential for practical applications. However, the wide-bandgap (WBG) perovskite usually suffers higher energy loss, which hinders the efficiency increase of TSCs and needs to be tackled with.

In perovskites, crystal imperfections and defects usually act as charge recombination centers and cause severe nonradiative energy loss. Therefore, defect passivation is an effective approach to eliminate these charge recombination centers and enhance the voltage. Similarly, special passivation strategies are required to reduce the voltage loss of WBG pero-SCs. This study draws inspiration from previous passivation agents for WBG perovskites, which typically feature diammonium groups, such as 1,2-ethylenediammonium iodide and 1,3-propylenediammonium iodide. However, the ammonium groups in the reported molecules are not spatially fixed due to the rotational flexibility of the alkyl chains. To overcome this, we explored the molecular structure of cyclohexane, which naturally has rigid cis and trans isomers. Herein, we introduced cyclohexane-1,4-diammonium iodide (CyDAI₂) as a passivating agent for the WBG perovskite layers, with ammonium groups positioned either on the same side (cis-CyDAI₂) or opposite sides (trans-CyDAI₂) of the cyclohexane ring.

We thought that the cis-CyDAI₂ with ammonium groups on the same side, would result in higher passivation density, leading to improved PCE. As expected, after fabricating different perovskite devices, we observed significant performance differences. Following this discovery, we performed a series of characterization techniques to analyze the behavior differences between the two isomers.

Fig. 1. Structures of CyDAI₂ isomers and their interaction with the perovskite surface. 

First, we conducted theoretical calculations to analyze the passivation molecules (Fig. 1a). As predicted, the rigidity of the molecular conformation in the cis- and trans-isomers allowed the passivating agent to maintain a large dipole moment at the molecular level. We further performed first-principles calculations to evaluate the interaction between the different passivating agents and the perovskite surface, revealing that the larger dipole moment of cis-CyDAI₂ favored more stable contact with the perovskite (Fig. 1b, c). Notably, there were slight differences in the preferred adsorption sites of the two isomers on the perovskite surface, which caught our attention.

To bridge the theoretical calculations with experimental data, a lot of techniques are utilized to study the different behaviors of the passivation molecules with cis-trans isomerism on the perovskite surface. Interestingly, we observed the formation of a horizontally oriented 2D perovskite and a 3D/2D perovskite heterostructure on the surface for the trans-CyDAI₂ treated perovskite films (Fig. 1d, e). While, no obvious 2D perovskite was observed in the cis-CyDAI₂ treated perovskite films.

This phenomenon reveals different behaviors of the two passivation agents on the perovskite surface. Since 2D perovskite formation typically destroys the surface, we performed stability energy tests for the passivation molecules in the bulk phase. The results showed that the trans-CyDAI₂ with ammonium groups on opposite sides, encounters less steric hindrance, making it more likely to penetrate the bulk perovskite.

The reduced steric hindrance of trans-CyDAI₂ lead to surface reconstruction, hindering efficient charge contact with the electron transport layer. In contrast, the cis-CyDAI₂ provides higher surface passivation density and a more favorable energy alignment for electron transfer to the electron transport layer. 

Fig. 2. Photovoltaic performance of the perovskite/organic TSCs based on cis-CyDAI₂ treated WBG perovskite. 

Due to the different passivation mechanisms, the WBG pero-SCs treated with cis-CyDAI₂ exhibited reduced interfacial energy loss. In a 1.88 eV wide-bandgap perovskite single-junction cell, we achieved an open-circuit voltage of 1.36 V. Eventually, we constructed a perovskite/organic TSC (Fig. 2a) using the low-energy-loss WBG perovskite as the front cell and a narrow-bandgap organic solar cell as the rear cell, achieving a PCE of 26.4% with a certified value of 25.7% (Fig. 2b), which is the highest reported PCE for the perovskite/organic TSCs.

Overall, our work not only offers new insights into the design of passivation agents but also provides a valuable reference for future researchers working on direct contact between perovskites and electron transport layers.