A group of scientists at the Dutch Institute for Fundamental Energy Research ( DIFFER ) in Eindhoven, the Netherlands, has created a database of 31,618 molecules that could be used in future redox flow batteries. To do this, the researchers used artificial intelligence and supercomputers to identify the properties of the Molecule Database.
With the increase in potential molecules that have been designed for use in flow batteries for energy storage, DIFFER researchers thought about creating a database that would be quick and easy to access to know the properties of these molecules.
The problem was knowing the properties of many of the molecules such as redox potential and solubility in water, relevant information since they are related to the energy generation capacity and energy density of redox flow batteries.
AI and supercomputers for the creation of the database
To solve this problem, the researchers carried out four steps to develop the database . The first involved using a desktop computer and intelligent algorithms to create thousands of virtual variants of two types of molecules, quinones and aza-aromatics. The researchers incorporated backbone structures of 24 quinones and 28 aza-aromatics plus five different chemically relevant side groups into the computer. From that, the computer created 31,618 different molecules.
In the second step, the researchers used supercomputers, which apply quantum chemical equations, to calculate almost 300 different properties of each molecule. The next step was to incorporate machine learning to predict whether the molecules would dissolve in water.
Finally, the human- and machine-readable RedDB (from Redox DataBase) database was created, containing the molecules and their properties with convenient names and descriptions.
This database is public and researchers around the world have the ability to easily search for potentially interesting molecules for redox flow batteries. Additionally, researchers can use the database to improve their machine learning models to accelerate the design of high-quality molecules for energy storage.