Evaluating Thermal Quenching Temperature in Eu³⁺-Substituted Oxide Phosphors via Machine Learning

Area of Science:

• Materials Science
• Solid-State Chemistry
• Computational Materials Science

Background:

• Energy-efficient solid-state lighting offers a sustainable solution to reduce global energy consumption.
• Europium (Eu³⁺)-substituted inorganic oxide phosphors are crucial for warm white light but often suffer from thermal quenching.
• Discovering new thermally robust phosphors is typically a slow and resource-intensive process.

Purpose of the Study:

• To develop a machine learning (ML) model for rapid prediction of thermal quenching temperature (T₅₀) in Eu³⁺-doped phosphors.
• To accelerate the identification and synthesis of novel, thermally stable phosphor materials for solid-state lighting applications.

Main Methods:

• A regression ML algorithm was trained using 134 experimental data points of temperature-dependent Eu³⁺ emission.
• The model estimated the T₅₀ for over 1000 potential oxide phosphor hosts.
• Five promising compounds were synthesized and experimentally validated for thermal stability.

Main Results:

• The ML model successfully predicted the T₅₀ for various Eu³⁺-substituted oxide phosphors.
• Five selected phosphors (Sr₂ScO₃F, Cs₂MgSi₅O₁₂, Ba₂P₂O₇, LiBaB₉O₁₅, Y₃Al₅O₁₂) demonstrated good thermal stability with T₅₀ > 423 K.
• Experimental validation confirmed the predicted thermal robustness of the synthesized materials.

Conclusions:

• The developed ML methodology significantly accelerates the discovery of thermally stable phosphors.
• This approach enables efficient identification of materials suitable for high-performance, energy-saving solid-state lighting.
• The validated phosphors represent promising candidates for next-generation lighting technologies.