Summary

Systems

FNS

EPFL, iCoSys, iSIS, iPRINT

Laurent Donato
Skills directory

January 2022 - December 2025

Automated Photonic-pulses Processing for Thin Solar Energy Devices

Current situation

According to the Shell Sky 2050 scenario, photovoltaic energy will need to reach a production capacity of 1,000 GW per year from 2035 onwards to comply with the Paris Agreement and stay below a 2-degree rise in global temperatures. This represents a phenomenal growth rate compared with today's total installed capacity of around 600 GW (for reference, the capacity added in 2020 was around 140 GW).

Project objective

The aim of this project is to take the Flash Infrared Annealing (FIRA, a technology invented by EPFL Professor Christophe Ballif) technique for curing perovskite thin-film materials to the next level by demonstrating a fast, continuous process with low environmental impact for the production of high-efficiency photovoltaic devices. The FIRA method is being applied to improve and accelerate the processing time of perovskite-silicon tandem solar cells, a solar cell technology that has the potential to accelerate the deployment of photovoltaic energy thanks to its high-performance potential and low manufacturing costs.

Role of the research institutes

As part of this project, HEIA-FR researchers are responsible for developing the automated pulsed photonic annealing system, which will be synchronized with a solution deposition system. To this end, SeSi researchers have developed and built a crystallization chamber capable of accommodating large-scale perovskite samples (over 400 cm2) for future commercial production. As the technology for manufacturing perovskite films is not yet mature, numerous tests are being carried out to find the most efficient production method by measuring the intensity, duration and frequency of photon flashes. To operate this crystallization chamber, iSIS researchers have developed an electronic control system from scratch, including an electronic board and on-board software to manage the machine's operation. By applying the MBSE model (model-based systems engineering approach), the researchers were able to simulate the operation of the electronic system and its interactions with the upstream crystallization chamber virtually, before doing so in the real world.

A large number of images are taken in the crystallisation chamber. The iCoSyS institute has taken on the task of developing the system for analysing these images using machine learning (ML) technologies. ML-based image segmentation is trained to automatically detect the edges of perovskite crystal grains, from which the distribution of crystal grain areas, vacancies and secondary phases are calculated. These calculations are important for characterizing film quality and conformal deposition of perovskite on textured silicon substrates, key factors in the subsequent manufacture and scale-up of high-efficiency tandem solar cells. The availability of these algorithms significantly accelerates research into perovskite solar cells by automating a process that scientists previously carried out by hand.

Advantages and impact of this technology

The FIRA production method enables perovskite films to be cured in less than 2 seconds, resulting in single-junction solar cell efficiencies of up to 20%. It's also important to note that this approach has just 10% less environmental impact than the more widely-used antisolvent (AS) method, based on life-cycle assessment. The FIRA method therefore offers unique potential for industrial manufacturing, since it can be adapted to plate-to-plate and roll-to-roll (R2R) processing, and replaces hour-long annealing steps.

The project will have an impact beyond the field of photovoltaics, as its findings on optimal layer deposition parameters will find application in research fields such as sensors, light-emitting devices, detectors and flexible electronics, etc.