Publication date: 4th October 2024
Before any perovskite solar cells (PSCs) are released for commercial use, their outdoor behaviour must be measured. This is an incredibly time consuming process that for relevant and slow degrading devices can last several months. In this work we demonstrate a flexible machine learning driven pipeline that can accurately determine the lifetime and degradation behaviour of any device, given short accelerated indoor tests. Additionally, by training our algorithm on different indoor tests, we are able to automatically determine the most relevant aging protocols, a process which so far relied heavily on intuition. Lastly, by identifying the most relevant stress factors, we can shed light on the outdoor degradation pathways. Our model has been tested on sollar cells fabricated in nip. Configuration of FTO/c-TiO2/m-TiO2/CsMAFAPb(IBr)3/Spiro-OmeTAD/Au in 6 different annealing temperatures. We verified the results by incorporating two a-priori known facts, namely that indoor aging under air will bear little relevance to outdoor degradation of encapsulated panels and aging under 1.4 sun and nitrogen is highly correlated to aging under 1.0 sun and nitrogen. Both effects were succesfully reconstructed by the model, therefore increasing the certainty that it can be applied at a larger scale. Finally, it must be noted that our methodology is not specific to PSCs and can be extended to other PV technologies where degradation and its mechanisms are a crucial element of their widespread adoption.
we acknowledge funding from ProperPhotoMile, under SOLAR-ERA.NET Cofund 2 by The Spanish Ministry of Science and Education
and the AEI under the project PCI2020-112185 and CDTI project number IDI-20210171;
the Federal Ministry for Economic Affairs and Energy on the decision by the
German Bundestag project number FKZ 03EE1070B and FKZ 03EE1070A; and the Israel
Ministry of Energy with project number 220-11-031. SOLAR-ERA.NET is supported by the
European Commission within the EU Framework Programme for Research and Innovation
15 HORIZON 2020 (Cofund ERA-NET Action, N° 786483). From TUM Innovation Network for Artifi-
cial Intelligence powered Multifunctional Material Design (ARTEMIS) framework of Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 2089/1 – 390776260 (e-conversion).The Blaustein postdoctoral fellowship at BGU. The Krietman Post- doctoral fellowship at BGU. The Swiss Inst. of dryland environmental
and energy research postdoctoral fellowship at BGU. Part of this work is under Materials
Science Ph.D. Degree for K.T. of the Universitat Autonoma de Barcelona (UAB, Spain).
ICN2 is supported by the Severo Ochoa Centres of Excellence programme (grant no. SEV-
2017-0706) and was funded by the CERCA Programme/Generalitat de Catalunya, Grant
CEX2021-001214-S, funded by MCIN/AEI/10.13039.501100011033.