TodayToday's edition
University of Surrey researchers tested 28 iron, titanium and vanadium variants of a sodium-ion cathode; some titanium compositions kept over 90% capacity, but the dataset is too small for AI predictions.
2 min readAI-assisted
Beakers with different colour liquids for testing, as shown in the University of Surrey release. The image is not captioned beyond that and may not show the exact cathode samples.
Image: University of Surrey
Source · Press image from the University of Surrey newsroom, used for news reporting.
The University of Surrey has built a library of 28 sodium manganese oxide cathode materials for sodium-ion batteries and found that iron, titanium and vanadium each shift battery performance in a different direction, the university said on Oct. 21.
The team swapped part of the manganese in the base material for varying amounts of iron, titanium and vanadium, then tested each version in sodium-ion cells. According to the Surrey release, iron was linked to higher energy-storage capacity, vanadium improved stability over repeated charging and discharging, and small amounts of titanium improved performance at higher charge and discharge rates. Some titanium-containing compositions kept more than 90% of their capacity after 100 cycles.
The library of 28 compositions showed different effects per metal, but the dataset is too small for reliable machine-learning predictions, the University of Surrey said.
Illustration (AI-generated)
The study appears in the Journal of Materials Chemistry A, and the full paper is at DOI 10.1039/d6ta04588j. Dr Daniel Commandeur is the corresponding author. Eden Marks and Lawrence McCaul co-led the work.
The cathode is the part of a cell that largely sets how much energy it stores and how well it survives repeated cycling. Sodium is far more abundant than lithium, which the researchers describe as relatively scarce and unevenly distributed. The barrier to wider use of sodium-ion cells, they say, is limited energy capacity and performance over repeated charging.
Because each added metal pulls the material in a different direction, the useful question is how to balance them. Marks said small changes in composition could shift the balance between stored energy, rate performance and lifetime. Mapping those relationships, Marks said, gives a much clearer picture of where to look next.
The group built an automated method to analyse the large amount of data its experiments produced, and used the resulting dataset to train a machine-learning model. The library is still too small for reliable predictions, the university said.
The aim is to grow the dataset. Commandeur called the work a first step towards the larger datasets machine learning needs to guide the search. Expanding the approach to a much wider range of cathode materials could eventually let AI suggest promising compositions before anyone makes and tests them in the lab.
For grid and data-center operators weighing sodium-ion storage, this is materials groundwork, not a product. Sodium-ion cells are already moving toward the market, as Voltgrid reported when EVE Energy rolled out a sodium-ion pack and when Peak Energy raised $108M for sodium-ion storage. A systematic map of which metal does what could shorten the search for better cathodes, but the researchers give no timeline.
The release gives the 90% retention figure only for some titanium-containing compositions and does not give absolute capacity values for any of the 28 materials. It does not say how the results compare with commercial sodium-ion cathodes. The machine-learning model is not yet predictive, and the researchers have not announced a next-stage dataset size or date.
