Fast and Flexible 3D Molecular Design Framework for Novel Organic Optoelectronic Materials

Area of Science:

• Materials Science
• Computational Chemistry
• Organic Electronics

Background:

• Organic optoelectronic materials (OOMs) are crucial for technologies like organic photovoltaics and LEDs.
• Current discovery methods are slow and limited in exploring the vast chemical space.
• There is a need for efficient computational tools to accelerate OOM discovery.

Purpose of the Study:

• To introduce O2-GEN, a novel AI framework for generating and screening organic optoelectronic materials.
• To enable comprehensive exploration of chemical space for novel OOMs.
• To facilitate the design of OOMs with tailored multiproperty profiles.

Main Methods:

• Utilized a 3D pretraining backbone trained on over 10 million molecules.
• Implemented global and local generation modes for molecular design.
• Integrated a property selector fine-tuned with density functional theory (DFT) data for multiproperty screening.

Main Results:

• O2-GEN effectively generates novel fused-ring systems and conjugated fragment assemblies.
• Demonstrated superior speed and chemical structural validity compared to existing models, especially for larger molecules.
• Enabled the construction of tailored datasets for specific optoelectronic applications.

Conclusions:

• O2-GEN significantly accelerates the discovery of novel organic optoelectronic materials.
• The framework offers precise control over molecular generation and property selection.
• O2-GEN facilitates the creation of application-specific material datasets, advancing organic electronics research.