Using AI, China Develops Lithium Injection to Treat Batteries
Lithium-ion batteries, commonly found in laptops, cars, and phones, gradually lose their ability to store energy due to the inactivity of lithium ions, rendering them incapable of charging and discharging.
Instead of replacing and recycling the battery, a team of researchers at Fudan University in China approached the problem as a medical condition and sought to find a cure. They called it “precision therapy for electric battery cells,” drawing inspiration from how the human body is treated with injections to restore vitality.
To achieve this, the researchers needed to discover a suitable molecule that could carry lithium ions back into the battery and reactivate its performance without causing issues due to interactions with the materials. Additionally, the solution needed to be cost-effective for practical application. Given the complexity of this research, they used artificial intelligence to suggest compounds that might meet these criteria. Eventually, they settled on a new molecule, lithium tri-fluoro-methane-sulfonate (LiSO2CF3), which they synthesized in the lab.
After proving successful in lab battery models, they tested the molecule on widely used lithium iron phosphate (LFP) batteries. They found that injecting these molecules increased the battery’s lifespan to 12,000 cycles, more than double the typical lifespan of 4,000 to 6,000 cycles.
Notably, high-quality LFP batteries exist that can achieve a lifespan of around 12,000 cycles without any injections. Therefore, the question of cost arises: is it cheaper to inject the battery with these molecules or purchase a high-quality battery? The team stated that the injection costs about 10% of the battery’s price.
The team is now working on improving the production process of the lithium-carrying molecule and collaborating with battery manufacturers to commercialize and implement the technology. They aim to extend battery life, reduce resource waste, and minimize pollution.
The research was published in the journal Nature in February 2025.
