MATCH-UP: MATerials disCovery for next generation pHotovoltaics via high-throUghPut synthesis, characterization and ai-based analysis
Coordinator: Philip SCHULZ
IPVF – Institut Photovoltaïque d’Ile de France
(UMR 9006 CNRS/Ecole Polytechnique/ENS Chimie Paris/IPVF/IP Paris/Univ. Paris Sciences et Lettres)
Keywords: Energy transition, Green technologies; Solar cells, Halide perovskite semiconductors, Automated thin-film deposition, High throughput, Multimodal in-situ characterization, Accelerated aging, Datamining, Physical modeling, AI-based methods
In the context of the energy transition, digital transformation, climate crisis and energy sovereignty, the accelerated development of reliable, low-energy, non-toxic and high-performance photovoltaic materials and devices is a key factor. This calls for greater collaboration between the various areas of expertise in the laboratories and the industrial R&D pilot lines (feeding into the planned French gigafactories), and improved circulation of experimental data made possible by process automation and robotization, high-throughput characterization and analysis by physical modeling and associated artificial intelligence.
To meet this need, the present project brings together a consortium with balanced skills in materials, synthesis processes and thin-film deposition, analysis by in situ characterization and advanced degradation, analysis of PV devices, automation and analysis by physical modeling and AI. This complementarity is a key element for this type of project, which not only requires strong collaboration but also enables collaborations to be accelerated thanks to the new platform developed locally, as well as the data generated and valorized.
Automating the various stages from synthesis to high-throughput analysis of solar devices speeds up the design of thin films, multilayers and solar cells. This enables accelerated discovery of new materials, prediction of material properties and figures of merit of solar devices, while drastically reducing the time
needed for exploration, analysis and knowledge creation. To solve the problems inherent in the search for new materials and solar devices, artificial intelligence methods combined with the automation of deposition, characterization and degradation analyses can increase the quantity/quality, reproducibility of samples and reliability of the data created. To date, only a tiny fraction of the data generated in laboratories is used, shared and sufficiently reliable to enable more global use and exploit the full potential of data science methods.
To this end, we propose to bring together in a single automated platform: high-throughput thin-film deposition processes, characterization and analysis of degradation, with coupled experiments at ex-situ sites such as large user facilities, e.g. synchrotron light sources. The use of infrastructures for automation, robotization and high-throughput analysis are very promising ways of digitizing experimental tests and taking advantage of AI methods. Pooling different deposition and analysis methods on the same automated platform has the advantage/constraint of focusing on unsophisticated characterization methods, while at the same time enabling us to take advantage of the sophisticated methods available on common measurement platforms. Another aspect of these methods is to better control the experimental process, improve reproducibility and quality, and gain precise knowledge of the processes to limit the environmental impact of the tests carried out.
Thanks to the diversity of possible compositions and the great prospects offered by the family of metal halide perovskites, this is a prime area for the development of these methods. A number of highly promising compositions have recently set records for performance and stability in the photovoltaic sector, which is undergoing exponential development.