From Lab to Lamp: Building a Phosphor Discovery Pipeline

NON-TECHNICAL SUMMARY:

The pursuit of energy-efficient lighting has led to the emergence of phosphor-converted light-emitting diodes (pc-LEDs), offering the promise of reducing global energy consumption and carbon dioxide emissions while revolutionizing lighting technologies. However, a significant challenge remains in finding new luminescent phosphors that efficiently convert blue LED light into high-quality white light, while also meeting the necessary chemical and thermal stability requirements for long-term consumer and commercial use.

Traditional methods have struggled to uncover suitable phosphors due to the difficulty in predicting crucial optical and chemical properties a priori. The proposed research project addresses this challenge by integrating machine learning with materials synthesis to expedite the discovery of efficient and stable phosphor powders compatible with blue LEDs.

By developing machine learning models, the goal is to predict the optical response and chemical stability of bulk ceramic phosphors, thereby identifying promising candidates for further synthesis and testing. Additionally, the academic team will collaborate with industry leaders to ensure that our phosphors meet the rigorous requirements for commercial applications.

This interdisciplinary effort not only advances lighting technology but also has broader impacts across various research fields, such as solar cells, by establishing a framework for predicting material optical properties. Furthermore, this program delivers hands-on research experiences for high school and undergraduate students, facilitated through initiatives like the ACS Project SEED and the Welch Summer Scholars Program, as well as an upper-level undergraduate course that integrates coding and statistics skills with materials science education.

By uniting materials chemistry, ceramic science, and data science, while enhancing the connections between academia and industry, this project will be pivotal in equipping the next generation of scientists to confront the challenges of future industries. TECHNICAL SUMMARY:

Replacing traditional fluorescent light bulbs with energy-efficient phosphor-converted light-emitting diodes (pc-LEDs) represents a direct approach to reducing electricity consumption. This technology relies on bulk ceramic phosphor powders, which, when applied to blue-emitting InGaN LED chips, absorb and partially convert the LED emission to produce a broad-spectrum white light.

The central hypothesis of the project is that machine learning can guide the synthesis of novel phosphors predicted to possess an optimal set of optical properties, thus enhancing LED adoption rates. Objectives include the development of machine learning models to reconstruct a phosphor's excitation spectra and multi-objective regression to predict emission color.

Validation will involve synthesizing rare-earth-substituted phosphors using high-temperature ceramic or solution-based "soft-chemical" methods, followed by characterization using X-ray and neutron scattering and photoluminescence spectroscopy. Moreover, the project addresses the critical need for improved chemical and thermal stability of phosphors in pc-LED-based lighting, crucial for extended operational lifetimes in harsh environments.

Traditional stability assessment methods are time-consuming, often taking years. Therefore, the project proposes support vector machine and time-series machine learning modeling methods to analyze and extrapolate chemical stability information, expediting testing and materials improvements. Ultimately, the project aims to produce phosphors with enhanced optical properties, including strong absorption of blue LED radiation, desirable emission colors, and stability in challenging conditions, facilitating advanced high-quality LED lighting technologies.

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.