One-year outdoor testing reveals degradation paths in perovskite tandem solar cells

Researchers from the Netherlands Organization for Applied Scientific Research (TNO) and Germany’s Fraunhofer Institute for Solar Energy Systems (Fraunhofer ISE) have conducted a 1-year test to measure the outdoor performance of peroskvsite solar technologies and have found that several factors contribute to significant perovskite degradation.

For the testing, the team used triple junction perovskite/perovskite/silicon solar cells with an active area of 1 cm × 1 cm. “We chose triple junction devices as they have a theoretical higher performance limit,” corresponding author Petra Manshanden told pv magazine. “These devices are fairly uncommon and long time outdoor exposure data has not yet been shown.”

The device tested is a monolithic triple-junction tandem combining a p-type heterojunction silicon bottom cell with two stacked perovskite subcells. The silicon base is rear-textured with a closed metal rear contact, acting as the near-infrared absorber. On top, a 1.56 electronvolt perovskite middle cell is deposited on indium tin oxide (ITO), with polybis(4-phenyl)(2,4,6-trimethylphenyl)amine and poly(9,9-bis(3′-(N,N-dimethylamino)propyl)-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene) as hole transport layers, and fullerene (C60) with tin oxide (SnOx) as the electron transport stack.

A second indium tin oxide (ITO) interconnect separates the middle cell from the top cell, which uses a self-assembled monolayer known as 2PACz and a wide-bandgap perovskite absorber. The top cell again employs C60 and SnOx for electron extraction and is completed with an ITO layer and evaporated silver (Ag) contacts. A magnesium fluoride (MgF₂) anti-reflection coating finalizes the stack.

The outdoor test station was located in Petten, the Netherlands. It consists of rooftop-installed with south-facing modules fixed at a 30° tilt under a local albedo of 10% for optimized annual yield.

Outdoor measurements at the station showed that, over the first month, morning and afternoon performance was largely comparable, with only early transient differences disappearing after the initial stabilization phase. Long-term outdoor monitoring, meanwhile, revealed a two-step degradation behavior, with efficiency dropping from around 17–18% initially to about 15% in March and approximately 13–14% in April, followed by continued decline. The first degradation phase was mainly attributed to voltage loss, while the second coincided with encapsulation delamination that reduced current collection and light in-coupling.

Microscopy confirmed delamination occurring within the encapsulation stack rather than at the active junctions, indicating mechanical or interlayer adhesion failure rather than moisture ingress. EQE and J–V analysis further showed that the loss in performance could not be explained by bandgap changes or intrinsic absorber degradation, but rather by interface-related losses and shunt pathways.

Photoluminescence and electroluminescence imaging after extended exposure showed strong spatial inhomogeneity, with the middle perovskite layer dominating current flow while the top junction was significantly weakened. This supported the conclusion that partial shunting and non-uniform degradation were key failure modes in the stack.

Indoor reliability tests confirmed good damp-heat stability with edge sealing but significant losses under thermal cycling and UV exposure, the latter causing up to around 65% degradation. Overall, despite degradation and hysteresis effects, the device achieved an estimated average annual efficiency of around 10%, with performance strongly dependent on irradiance conditions and spectral variability.

“The samples achieved 80% of the initial power conversion efficiency, after five months of outdoor operation, and 50% of after seven months,” Manshanden explained. “Early-stage degradation analysis indicates that the top junction is the least stable component of the device stack. In parallel, the middle junction, which initially limited current generation, developed edge-localized shunting during operation. Additional losses were attributed to degradation of the charge transport layers, most likely driven by sustained exposure to elevated operating temperatures.

The results of the testing were presented in “One Year of Outdoor Performance of Perovskite/Perovskite/Silicon Triple-Junction Solar Cell,” published in RRL Solar. “These findings helped us to understand degradation mechanisms and develop more stable next generation devices which are currently under testing,” concluded Manshanden.