Artificial Intelligence deciphers the oxidation states within crystal compositions
In a groundbreaking development, a team of researchers led by Kevin Jablonka, a PhD student in Smit's group at EPFL, have published a paper in Nature Chemistry that introduces a machine-learning model to classify the oxidation states of metal-organic frameworks (MOFs). The paper, titled "Using collective knowledge to assign oxidation states of metal-cations in metal-organic frameworks," was published on July 5, 2021.
Oxidation states, also known as oxidation numbers, play a crucial role in the fundamentals of chemistry. They represent the number of electrons an atom must gain or lose to form a chemical bond with another atom. However, determining the oxidation state in complex materials can be a complicated task.
The current state-of-the-art in predicting oxidation states is based on the bond valence theory, developed in the early 20th century. But this method has its limitations, particularly when dealing with large databases of materials. To address this challenge, the research team, which includes Daniele Ongari, Seyed Mohamad Moosavi, and Berend Smit, created a machine-learning model that functions similarly to the TV game "Who Wants to Be a Millionaire?".
The model asks a chemist for the oxidation state if they do not know, and provides the most likely oxidation state based on the collective knowledge of the chemistry community. This collaborative approach ensures a more accurate and consistent assignment of oxidation states, especially for MOFs.
The Cambridge structural database, a repository of crystal structures, was used as the primary source of data for training the machine-learning model. However, it's important to note that the database contains many errors and inconsistencies due to a mixture of experiments, expert guesses, and different variations of the bond valence theory used to assign oxidation states.
The machine-learning model developed by the team captures the collective knowledge of the chemistry community, providing a more reliable and consistent method for assigning oxidation states of metal-cations in MOFs. The paper was published by Nature Chemistry, and the DOI for the paper is 10.1038/s41557-021-00717-y.
The atomic number of an element, which shows how many protons there are in the element's nucleus, serves as the element's ID card. It determines how many electrons orbit the nucleus, which essentially makes the element what it is and gives it its chemical properties. The periodic table, as of 2016, includes 118 elements, each represented by a one- or two-letter abbreviation and an atomic number. However, it's worth mentioning that the periodic table does not currently include the element's oxidation state.
This innovative research not only addresses a long-standing challenge in the field of chemistry but also paves the way for future advancements in materials science and the development of new technologies.
