Sequential closed-loop Bayesian optimization as a guide for organic molecular metallophotocatalyst formulation...

Area of Science:

• Organic chemistry
• Catalysis
• Materials science

Background:

• Organic photoredox catalysts (OPCs) are versatile tools in chemical synthesis.
• Predicting OPC activity is difficult due to complex property interdependencies.
• Catalyst discovery often relies on trial-and-error methods.

Purpose of the Study:

• To develop a data-driven strategy for targeted synthesis and optimization of OPCs.
• To apply this strategy to metallophotocatalysis, specifically for decarboxylative cross-coupling reactions.
• To accelerate the discovery of efficient and selective organic photocatalysts.

Main Methods:

• Employed a two-step approach combining Bayesian optimization with molecular descriptors.
• Screened a virtual library of 560 candidate OPC molecules.
• Optimized reaction conditions for metallophotocatalysis.

Main Results:

• Identified highly effective OPC formulations through efficient exploration of the catalyst space.
• Achieved catalytic performance competitive with iridium-based catalysts.
• Explored only 2.4% of the total reaction condition space to find optimal formulations.

Conclusions:

• The data-driven approach significantly accelerates the discovery and optimization of OPCs.
• This method offers a powerful alternative to traditional trial-and-error catalyst development.
• Highlights the potential of computational strategies in advancing photoredox catalysis.