With demonstrated power conversion efficiencies close to 23%, perovskite-based photovoltaics is already able to compete with established technologies like silicon, CdTe and CIGS. However, next to high efficiencies, the potential low-cost fabrication of devices with sufficient stability under real-world conditions is of key importance for the future economic prospects of the perovskite technology.

In this contribution, we report on a novel inexpensive architecture for efficient and highly reproducible, all-evaporated perovskite solar cells. Our evaporated CH₃NH₃PbI₃ absorber is sandwiched in a p-i-n structure between inexpensive and highly stable nickel oxide as hole transport material and a combination of C₆₀ and bathocuproine (BCP) as electron hole transport material. In contrast to that, most of the approaches in the community employ highly expensive hole transport materials like Spiro-MeOTAD or PTAA with prices up to 1,000,000 $/kg, which would hamper the commercialization of the technology. Most importantly, the common organic hole transport material Spiro-MeOTAD shows low stability at elevated temperatures above 60 °C, making it an unsuitable choice for applications under typical outdoor conditions. By replacing the unfavorable Spiro-MeOTAD by electron-beam deposited nickel oxide and the gold back electrode by copper, we reduce the cost of materials on the lab-scale to one third of the price of common stacks (e.g., ITO/TiO₂/CH₃NH₃PbI₃/Spiro-MeOTAD/Au). At the same time, power conversion efficiencies of the devices reach stabilized values above 14% without hysteresis. Moreover, high thermal stability of the employed transport materials is demonstrated in extremely stable devices even at 80 °C, which is a typical operating temperature for solar modules under real-world situations as well as standard test condition in established performance tests. In contrast, our reference all-solution-based devices with Spiro-MeOTAD degrade fast to about 80% of their initial values under the same conditions.

Towards an industrialization of the perovskite technology, a highly controllable deposition and an easy upscaling is needed. Our all-evaporated approach is able to meet these criteria. Even on small lab-scale areas, 30% lower cell-to-cell variations in comparison to common spin-coating approaches are achieved. In terms of upscaling, homogenous and reproducible depositions up to areas of 8x8 cm² are demonstrated and investigated by light beam induced current mapping. Finally, as an inverted architecture with the anode deposited on top of the substrate the discussed layer stack is a promising candidate for two-terminal tandem cells and modules on top of CIGS or p-type silicon.