Precision interface chemistry pushes perovskite solar cells beyond 26% efficiency

An international research collaboration has developed a new molecular strategy for controlling one of the most critical interfaces in perovskite solar cells. The resulting solar cells reached a power conversion efficiency of 26.19% in the n i p architecture, together with strong operational stability under prolonged illumination and elevated temperature. The results have been published in the Journal of the American Chemical Society.

The work addresses a persistent challenge in perovskite photovoltaics: residual lead iodide, PbI₂, that remains at the surface of the perovskite after film formation. Although moderate amounts of PbI₂ can be beneficial during crystallization, an inhomogeneous distribution at the final interface can create local variations in surface potential, promote charge trapping, and increase nonradiative recombination. Now, an international collaboration, involving my own team, the Robotized Optoelectronic Material and Photovoltaic Engineering Group at Helmholtz Zentrum Berlin, and the research team of Professor Letian Dou at Purdue University and Emory University, has developed a new approach to tackle this problem.

We designed a new class of bidentate molecular ligands that interact selectively with residual PbI₂ through two anchoring sites. In contrast to conventional molecules that bind through a single interaction point, the new molecules reconstruct the residual PbI₂ into more stable and electronically favorable PbI₆ coordination structures while preserving the underlying three dimensional perovskite absorber. The most successful molecule, MeXT, produced a significantly more homogeneous electronic landscape across the perovskite surface. This reduced interfacial disorder and nonradiative voltage losses while improving the transport of photogenerated charge carriers toward the hole transport layer. The champion device reached 26.19% efficiency, with an open circuit voltage of 1.198 V, a fill factor of 83.2%, and a short circuit current density of 26.28 mA cm⁻². The device also delivered a stabilized efficiency of 25.65%. Under combined light and thermal stress at 75 °C, the treated devices retained more than 80% of their initial efficiency after 1000 hours.

Insights into charge transport

A central contribution from my team at HZB was the application of advanced transient and spatially resolved surface photovoltage measurements. These measurements provided direct insight into how the molecular treatment changes charge separation and extraction at the interface The optimised treatment did not simply passivate defects. It changed the interfacial charge selectivity itself. While insufficiently treated surfaces showed signatures of electron accumulation and trapping, the optimized bidentate treatment suppressed these electron trapping pathways and strongly promoted hole accumulation and extraction toward the hole transport layer. Measurements on complete perovskite, ligand, and hole transport layer stacks showed a faster and substantially stronger positive photovoltage response for the best treatment, consistent with enhanced hole extraction and reduced interfacial recombination.

Surface photovoltage allowed us to see what conventional efficiency measurements alone cannot reveal.We could directly distinguish how different molecular treatments change charge selectivity, defect activity, and extraction dynamics. This helped identify not only whether a treatment works, but why it works and where the optimum lies for the complete device.

The study demonstrates the strength of combining rational molecular design, advanced spectroscopy, spatial mapping, theoretical modelling, and complete device engineering. The chemical design and photovoltaic development were carried out in close collaboration with the group of Professor Letian Dou, with additional theoretical contributions from the team of Professor Brett M. Savoie. Together, we established a broader design principle for creating electronically homogeneous interfaces through selective chemical coordination rather than nonspecific surface treatment.

This work also points toward the next stage of photovoltaic research at HZB: autonomous materials and device optimisation. Over the coming three months, I will install a new fully robotized line for solar cell preparation, characterization, and optimisation will be installed at HySPRINT. The platform will combine automated device fabrication with rapid optoelectronic characterization and data driven optimisation. The goal is to accelerate experimental optimisation by approximately a factor of ten while generating deeper physical insight into the relationships between processing, interface properties, and final device performance.

The next step is to connect this type of fundamental interface understanding directly with autonomous experimentation. Instead of testing materials through long sequential optimisation campaigns, we want robotic systems to prepare devices, measure the relevant physical parameters, and use the results to decide which experiment should be performed next. With my team at HZB, we arepreparing to share the first photographs and videos from the new robotic laboratory in September and October 2026, marking the beginning of a new phase in automated discovery and optimisation of photovoltaic materials and interfaces.