BRITE Pivot: Accelerating Manufacturing and Realization of Perovskite Micro-Light Emitting Device (Micro-LED) Displays through Data-driven Learning

University Of Washington, Seattle WA

Investigators

Abstract

This Boosting Research Ideas for Transformative and Equitable Advances in Engineering (BRITE) Pivot grant supports research aiming to advance augmented- and virtual-reality (AR/VR) display technologies through exploring new materials and manufacturing driven by machine learning approaches. AR/VR technologies connect people with the digital world through immersive experiences, and present unbounded applications in many industries. However, these technologies face many obstacles which hinder their prevalence, with one notable impediment being optical clarity and resolution. To this end, the industry is actively investing in micro-light emitting device (micro-LED) technology, which preserves the high optical quality of emissive LEDs while reducing the pixel size by nearly an order of magnitude over the current state-of-the-art. Nevertheless, manufacturing of micro-LED displays through conventional semiconductor fabrication processes is extremely time consuming and labor intensive, which leads to prohibitive costs, yields, and production times. This award supports fundamental research on new micro-LED technologies utilizing solution-processed perovskite materials to overcome these challenges and optimizing the device performance through machine learning. The project facilitates the acquisition of knowledge in a new field, machine learning, to supplement established expertise in perovskite optoelectronics. The vast scope of applications of AR/VR ranges from education and healthcare to manufacturing and defense. Therefore, results from this research benefit several U.S. industrial sectors, thus enhancing national security and prosperity. The multi-disciplinary research entails hands-on education and training experiences closely related to industrial applications, while promoting diversity, equity, and inclusion in project participation. While perovskite devices have quickly made significant impacts in photovoltaics, their progress for light-emitting applications has been lagging due to degradation of the materials and devices under high electric fields. Enhancing perovskite LED performance requires simultaneous optimization of material synthesis, device structures and operating conditions. Machine learning (ML) offers the possibility of quickly converging to a globally optimized result through a multi-variate data-driven learning process. This project aims to develop a neural network-based ML approach to facilitate manufacturing of perovskite micro-LED displays with high optical quality and stability. Furthermore, the ML process is designed and executed with the aim of optimizing perovskite LED output luminance and stability, with proper input features. Through the researched project, features pertaining to perovskite LEDs are identified and collected from various databases, and, along with additional experimental data, are fed into the ML models, which drive the solution processing of the perovskites. Additionally, a micro-patterning technology is developed to address the challenge of patterning perovskites with high resolution through photolithography. 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.