A new study by researchers at the University of Toronto (USA) and Northwestern University has used machine learning to create better building blocks when assembling wireframe materials.
The new work says artificial intelligence (AI) can help develop new materials for a variety of applications. For example, when carbon dioxide is released in industrial combustion processes. AI can speed up material design cycles.
In order to improve the separation of chemicals in industrial processes, the research team has identified the best reticular scaffolds (eg, organometallic scaffolds, covalent organic scaffolds) to use.
Such scaffolds can be regarded as specially designed molecular “sponges”: they are formed on the basis of self-assembly of molecular building blocks in various configurations. This creates a new family of crystalline porous materials that can be used to solve many technological problems.
We have created an automated material discovery platform that generates designs for various molecular structures. This greatly reduces the time required to determine the optimum materials to use in that particular process. In our case, we used a scaffold detection platform that competes strongly with some of the most efficient materials used for CO2 separation.
Name Zhengpeng Yao, Research Fellow, Department of Chemistry and Computer Science, Department of Humanities and Science, University of California, and lead author of the study
The researchers say the model demonstrates excellent predictive and optimization capabilities when developing new reticular structures, especially when combined with those already known. Also, the platform is fully customizable in its application to solve many modern technological problems.