A study conducted at the CiTOS Lab potentially reshapes the preparation of high-value compounds in organic chemistry. This innovative approach combines Quantum Chemistry and Artificial Intelligence, enabling the generation of optimized process conditions within minutes, completely waste-free. In silico conditions are then instructed to an automated lab-scale flow synthesis system, which executes them within record time. The project establishes a robust foundation for a radical shift in the approach to preparing high-value organic compounds. The results of this study are now published in Angewandte Chemie International Edition.
S
et up just over ten years ago by Jean-Christophe Monbaliu, the Center for Integrated Technology and Organic Synthesis (CiTOS) is based on an interdisciplinary strategy combining synthetic organic chemistry, physical and computational organic chemistry and micro/mesofluidic process technology. "Over the past decade, we have gained a renowned expertise in developing innovative, versatile, and efficient fluidic processes for the production of high-value compounds. Our work aims to provide practical solutions to redefine chemical production in the current environmental and safety context," continues the Director of CiTOS. While the adoption of new micro/mesofluidic reactor technologies has often been presented as a de facto solution for agile and safe processes, their implementation remains limited, in part due to the difficulty of anticipating the feasibility of a reaction in these reactors. "It is not easy to change decades worth of expertise with current production tools, and a too radical paradigm shift often faces a form of conservatism despite technological promises," adds the CiTOS director and investigator at the WEL Research Institute.
Fundamental research in organic chemistry primarily relies on the development of new methods and reagents that exploit specific reactivities, allowing to access to molecular diversity and the properties of a wide range of high-value products. This research involves iterative processes of optimization and selection to maximize yields. These iterative processes are both time-consuming, resource-intensive, and waste-generating, even with new micro/mesofluidic technologies. This is particularly concerning in cases where resources are limited or when toxic, highly active, or unstable compounds are involved.
To further accelerate the adoption of micro/mesofluidic technologies, a predictive model combining quantum chemistry and artificial intelligence (ML-DFT) has been developed at CiTOS. "This model predicts experimental data with high precision, such as optimal conditions for temperature, pressure, and reaction time to achieve a desired yield," explains Pauline Bianchi, first author of the study and F.R.S.-FNRS doctoral candidate at CiTOS. The optimized conditions are obtained in silico, solely through computational methods, and do not generate any chemical waste. "These calculated optimal conditions are then executed by an automated flow system. Artificial Intelligence-reinforced Quantum Chemistry protocols swiftly provide non-generic conditions for the synthesis of a wide range of organic derivatives. This is a distinctive advantage for creating molecular diversity when no pre-existing experimental data exist" Pauline Bianchi further elaborates. This ML-DFT requires computational power that is readily available in many chemistry laboratories, facilitating its widespread adoption.
This tool has been successfully applied to electrophilic aminations, a class of reactions particularly relevant to pharmaceutical compound manufacturing. This research holds both fundamental and applied significance, as evidenced by its use in physical organic chemistry to understand reactivity and its application in an industrial collaboration with Mithra Pharmaceuticals.
Bianchi, J.-C. M. Monbaliu, Revisiting the Paradigm of Reaction Optimization in Flow with a Priori Computational Reaction Intelligence, Angew. Chem. Int. Ed. 2023, e202311526. https://doi.org/10.1002/anie.202311526
Full article is available in Open Access
This project represents an innovative approach to urban planning and management. For example, it will enable the City of Liège to monitor tree planting as part of its Canopée plan.
A study that used the MAR climate model developed by ULiège researchers to predict the risk of meteorite loss due to the melting of the ice sheet.
Xavier Raick a brillamment réussi son doctorat et a reçu la pleine reconnaissance de sa qualité au travers du Prix “Oceanography and Oceans Protection”.
