Bayesian molecular optimization for accelerating reverse intersystem crossing

Area of Science:

• Materials Science
• Organic Electronics
• Quantum Chemistry

Background:

• Spin conversion in molecular excited states is vital for advanced optoelectronic devices.
• Current methods for optimizing molecular structures rely on slow, iterative experimental processes.
• Understanding structure-property relationships for spin conversion is challenging.

Purpose of the Study:

• To develop a Bayesian molecular optimization approach for accelerating virtual screening.
• To identify molecules with rapid triplet-to-singlet reverse intersystem crossing (RISC).
• To establish a data-driven method for designing optoelectronic materials.

Main Methods:

• Bayesian molecular optimization framework.
• Accelerated virtual screening for RISC.
• Machine learning model for analyzing structure-property relationships.

Main Results:

• Identified a molecule with a RISC rate constant of 1.3 × 10^8 s^-1.
• Achieved a high external electroluminescence quantum efficiency of 25.7%.
• Maintained high efficiency (22.8%) at a luminance of 5000 cd m^-2 in organic light-emitting diodes.

Conclusions:

• Bayesian optimization significantly accelerates the discovery of molecules for efficient spin conversion.
• The developed method facilitates informed materials design for optoelectronics.
• Machine learning analysis provides insights into structural determinants of spin conversion.