Electronics Weekly
Sodium-ion batteries are seen as a lower cost alternative to lithium-ion batteries, for grid storage, for example, but the large sodium ions are difficult to store.
The university is looking into a class of compounds called transition-metal layered oxides of the form NaMeO2, where ‘Me’ represents a mixture of metals such as manganese, titanium, zinc, nickel, iron or tin. The crystal structure of interest is called ’03’.
These “are powerful materials for the positive electrode of sodium-ion batteries, offering exceptional energy density and capacity,” according to the university. “However, for multi-element layered oxides composed of several transition metals, the sheer number of possible combinations makes finding the optimal composition both complex and time-consuming. Even minor changes in the selection and proportion of transition metals can bring about marked changes in crystal morphology and affect battery performance.”
Raw data was provided by 11 years of research over which 68 compositions have been tested in 100 experiments with different charge and discharge voltage limits. Results included initial discharge capacity, average discharge voltage and capacity retention after 20 cycles.
Several machine learning algorithms and Bayesian optimisation were used to learn how operating voltage, life-time and energy density related to the detailed composition of the different layered oxides, and then to predict metal ratios that would produce a good balance of voltage, life and density.
The model predicted Na(Mn0.36Ni 0.44Ti 0.15Fe0.05)O2 would have a high energy density, which turned out to be true after the researchers made samples of the material and tested coin cell batteries made from them.
“The measured values were, for the most part, consistent with the predicted ones, highlighting the accuracy of the model and its potential for exploring new battery materials,” said the university.
Tokyo University of Science worked with Chalmers University of Technology and the Nagoya Institute of Technology.
Findings are published as ‘Na(Mn0.36Ni 0.44Ti 0.15Fe0.05)O2‘ predicted via machine learning for high energy Na-ion batteries‘ in the Journal of Materials Chemistry A. This can be read in full without payment, and includes descriptions of the algorithms as well as electrochemical results.
Images: Tokyo University of Science
See also: Scalable, fully-coupled annealing processor with 4096 spins accelerates problem-solving
