IISc Researchers Use Machine Learning to Boost Magnesium Battery Performance

Researchers at the Indian Institute of Science (IISc), led by Assistant Professor Sai Gautam Gopalakrishnan, have developed a computational approach to enhance ion movement in magnesium batteries, potentially increasing their energy density. Unlike lithium ions, magnesium ions can transfer two electrons per atom, offering nearly double the energy storage per ion.

The team focused on cathodes, the battery’s positive electrodes, which need to absorb and release magnesium ions efficiently. Traditional crystalline materials slow magnesium movement, limiting performance. By modeling amorphous (disordered) vanadium pentoxide cathodes, the researchers found that breaking the crystal structure allows magnesium ions to move up to 100,000 times faster than in crystalline forms.

To achieve this, the team combined density functional theory (DFT) for small-scale accuracy with machine learning-driven molecular dynamics simulations for larger-scale modeling. This hybrid approach allowed them to predict magnesium mobility more efficiently and at greater scale than traditional methods.

The findings suggest amorphous materials could be key to developing faster, higher-capacity magnesium batteries. The next step is experimental validation to test stability and performance in real-world devices.

“Our work offers a completely different pathway to identify electrode materials for batteries and takes us a step closer to commercialisation of magnesium batteries,” said Gopalakrishnan.