Optimizing Perovskite Thin‐Film Parameter Spaces with Machine Learning‐Guided Robotic Platform for High‐Performance Perovskite Solar Cells

Abstract

Simultaneously optimizing the processing parameters of functional thin films remains a challenge. The design and utilization of a fully automated platform called SPINBOT is presented for the engineering of solution‐processed functional thin films. The SPINBOT is capable of performing experiments with high sampling variability through the unsupervised processing of hundreds of substrates with exceptional experimental control. Through the iterative optimization process enabled by the Bayesian optimization (BO) algorithm, the SPINBOT explores an intricate parameter space, continuously improving the quality and reproducibility of the produced thin films. This machine learning (ML)‐guided reliable SPINBOT platform enables the acceleration of the optimization process of perovskite solar cells via a simple photoluminescence characterization of films. As a result, this study arrives at an optimal film that, when processed into a solar cell in an ambient atmosphere, immediately yields a champion power conversion efficiency (PCE) of 21.6% with satisfactory performance reproducibility. The unsealed devices retain 90% of their initial efficiency after 1100 h of continuous operation at 60–65 °C under metal‐halide lamps. It is anticipated that the integration of robotic platforms with the intelligent algorithm will facilitate the widespread adoption of effective autonomous experimentation to address the evolving needs and constraints within the materials science research community.

Toward Self-Driven Autonomous Material and Device Acceleration Platforms (AMADAP) for Emerging Photovoltaics Technologies

“…To demonstrate the exceptional experimental capabilities, this DAP executed 5 optimization steps including a total of 61 variables for perovskite thin film fabrication. The step-by-step optimization process identified an optimal condition, which resulted in a champion perovskite device with a PCE of 21% . The exceptional manufacturing process control and data reproducibility offered by this automated DAP were exemplified through repeating device fabrication under optimal conditions.…”

“… 2023 13 2302594 . 3 An ML-driven automated platform, SPINBOT, efficiently explored a complex parameter space of perovskite thin films. The solar cells based on the optimized film achieved an impressive 21.6% efficiency under ambient conditions, along with excellent long-term stability.…”

“…Autonomous acceleration research platforms have been introduced to explore the realms of materials science research and development (R&D). , While the MGI expressed an acceleration of a factor of 2 and a reduction of costs by a factor of 2, some of the modern MAPs are proposing larger acceleration factors, ideally reducing the entire materials R&D cycle from typically 10 to 20 years to only 1 or 2 years. , Such acceleration factors become possible when automated laboratory setups are equipped with artificial intelligence (AI). , These setups build on recent scientific breakthroughs and the ability to program machines so that they can make independent and autonomous decisions to design and optimize materials and processes, moving away from the traditional Edisonian method of discovery. , Such breakthroughs hold the promise of a true digital twin of a technology, allowing science to predict optimized materials and processes according to the requirements of a target application. , The methodology of integrating automation and autonomous operation into emerging photovoltaics research, such as perovskite solar cell (PSC) and organic photovoltaic (OPV) materials, is drawing broad attention from the community. ,,− MAPs and DAPs offer tremendous opportunities for accelerating functional materials discovery and device optimization with respect to performance, environmental stability, and cost.…”

Accelerating Photostability Evaluation of Perovskite Films through Intelligent Spectral Learning-Based Early Diagnosis

“…To establish and train the machine learning toolkit for early diagnosis of film stability, we first carried out Experiment I. First, 160 perovskite films with different compositions were fabricated through an automatic platform . Subsequently, these films are exposed under light illumination over 1700 h to evaluate their lifetime, as shown in Figure a.…”

“…First, 160 perovskite films with different compositions were fabricated through an automatic platform. 38 Subsequently, these films are exposed under light illumination over 1700 h to evaluate their lifetime, as shown in Figure 1a. T 80 lifetime is employed to evaluate the stability that represents the time required to decay to 80% of the initial absorbance value, which considers the decline in absorbance exhibited by perovskite films during the decomposition process.…”