Tackling Instability in Perovskite Solar Cells through Machine Learning

非技术性：基于金属卤化物钙钛矿的太阳能电池由于其高效率和低制造成本，是一种很有前途的技术，可以提高可再生能源的利用率。然而，它们的性能在光照下会下降，这给这项技术的广泛采用带来了障碍。因此，迫切需要确定哪些因素会影响设备的性能和稳定性，但这个问题很复杂。影响设备性能的因素有很多，包括湿度、温度、暴露在氧气中以及运行中的电气条件。像这样的复杂问题是一个挑战，需要创新的解决方案。机器学习是一种人工智能，其中系统在没有明确编程的情况下从经验中学习和改进。机器学习(ML)非常适合解决钙钛矿型太阳能电池的稳定性等复杂问题。将机器学习技术应用于钙钛矿型太阳能电池的稳定性研究是本课题的主要研究内容。这项研究将通过实施ML例程来确定在钙钛矿太阳能电池中实现长期功率转换效率的理想条件，从而促进对钙钛矿太阳能电池的了解。ML将用于通过整体方法关闭改变游戏规则的稳定光伏设备的开发循环。这项研究建立在研究团队在太阳能电池制造和表征以及机器学习方面的经验基础上。该项目的外展/教育影响将为女学生在STEM领域的突出职位做好准备，为她们提供密集的指导和发展材料科学与工程(MSE)尖端研究的机会。PI将与数学工程科学成就(MESA)大学预科项目合作，吸引女性学生参加STEM。该项目的研究成果将丰富加州大学戴维斯分校的课程和课程，特别是在可再生能源领域。技术：本研究旨在确定卤化物钙钛矿太阳能电池设备休息和回收所需的条件。其目标是实现机器学习(ML)例程，以加速稳定设备的开发。首先，系统的光致发光(PL)测量和时间分辨微波电导率(TRMC)将在一组明确定义的环境条件下进行，以阐明水、氧、偏压、温度和光对器件性能的单独和联合影响。其次，数据将被用作ML的输入，以解决最大功率转换效率和增强寿命所需的条件。第三，对器件性能的最佳条件进行了实验论证。最后，ML所获得的知识将被用于开发稳定的钙钛矿设备的整体方法。这项研究将对光伏产生革命性的影响，导致一个快速诊断设备稳定性的新框架，并导致未来太阳能电池的设计具有全面增强的性能和寿命。除了光伏，这些方法还可以促进钙钛矿用于其他光电设备的科学知识，包括发光二极管(LED)和光电探测器。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Nontechnical:Solar cells based on metal halide perovskites are a promising technology to increase the adoption of renewable energy due to their high efficiency and low fabrication cost. However, their performance degrades under illumination, presenting a roadblock to widespread adoption of this technology. There is accordingly an urgent need to identify what factors impact device performance and stability, but this problem is complex. Many factors can impact device performance, including humidity, temperature, exposure to oxygen, and electrical conditions under operation. Complex problems such as this present a challenge that requires innovative solutions. Machine learning is a type of artificial intelligence in which systems learn and improve from experience without being explicitly programmed. Machine learning (ML) is well suited to complex problems such as the stability of perovskite solar cells. The application of machine learning to the stability of perovskite solar cells is the primary focus of this research project. This research will advance the state-of-knowledge of perovskite solar cells by implementing ML routines to identify the ideal conditions to achieve long-term power conversion efficiency in perovskite solar cells. ML will be used to close the loop in the development of game-changing, stable photovoltaic devices through a holistic approach. This research builds upon the research team’s experience in fabrication and characterization of solar cells and machine learning. The outreach/education impacts of this project will be preparing female students for prominent positions in STEM fields by providing them with intensive mentoring and the opportunity to develop cutting-edge research in materials science and engineering (MSE). The PI will work with the Mathematics Engineering Science Achievement (MESA) Pre-College program to engage female students in STEM. The research results of this project will enrich the curricula and programs at UC Davis, especially in the field of Renewable Energy.Technical:This research aims at determining the conditions required for device rest and recovery in halide perovskite solar cells. The goal is to implement machine learning (ML) routines to accelerate the development of stable devices. First, systematic photoluminescence (PL) measurements and time-resolved microwave conductivity (TRMC) will be performed under a set of well-defined environmental conditions to elucidate the individual and combined effects of water, oxygen, bias, temperature, and light on device performance. Second, the data will be used as an input for ML to resolve the conditions required for maximum power conversion efficiency with enhanced lifetime. Third, the optimal conditions for device performance will be experimentally demonstrated. Finally, the knowledge gained by ML will be used in a holistic approach to develop stable perovskite devices. This research will have a transformative impact on photovoltaics, leading to a new framework to quickly diagnose device stability and to the future design of solar cells with overall enhanced performance and lifetime. Beyond photovoltaics, these methods could advance the scientific knowledge of perovskites for other optoelectronic devices, including light-emitting diodes (LEDs) and photodetectors.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.