Here you can find one of our last publications, from Aalto University and the University of Turku teams, entitled: “Density Functional Theory and Machine Learning for Electrochemical Square-Scheme Prediction: An Application to Quinone-type Molecules Relevant to Redox Flow Batteries”
In CompBat we are interested of molecules redox potentials (Ered) but depending the pH or the charge state of the molecule the molecules protonation state can change.
We have developed a machine learning (ML) model to predict Ered and acidity constants (pKa) of a rather large set of molecules. Quantum chemical DFT methods has been used to compute Ered and pKa values of more than 8200 quinone-type organic molecules.
Supervised om Forest Regression model is used to teach the ML method. All studied molecules underwent two-proton and two-electron transfer process. Both structural and chemical descriptors are used.
The HOMO of the reactant and LUMO of the product participating in the oxidation reaction appeared to be inversely associated with Ered. Trained models using a SMILES-based descriptor can efficiently predict the pKa and Ered with a mean absolute error of less than 1 and 66 mV, respectively.
High prediction accuracy of R2 > 0.76 and > 0.90 was also obtained on the external test set for Ered and pKa, respectively. This hybrid DFT-ML study can be applied to speed up the screening of quinone-type molecules for energy storage and other applications.
Here you can find the full text, available in ChemRxiv (loaded 22.5.2023):
https://doi.org/10.26434/chemrxiv-2023-wfv75