Characterisation and performance evaluation of early-generation commercial sodium-ion batteries

Sodium-ion (Na-ion) batteries are a promising low-cost option for large-scale energy storage, yet practical deployment requires control-relevant characterisation of the commercial cells and their inherent variability. This work evaluates a population of early-generation 18650 layered-oxide Na(Ni,Fe,Mn)O2 (NFM) cells from a monitoring, control, and model parameterisation perspective, benchmarking against Na(Cu,Fe,Mn)O2 (CFM), polyanion Na4Fe3(PO4)2(P2O7) (NFPP), and Li-ion reference cells. Similar to Li-ion, Na-ion performance degrades as temperature decreases; however, all tested Na-ion chemistries retain measurable discharge capacity at −40∘C (51%–62% of the 25∘C value), whereas the Li-ion reference cells did not sustain discharge under the same protocol. Polarisation resistance from electrochemical impedance spectroscopy (EIS) testing across a range of operating temperatures exhibits the Arrhenius-type temperature dependence, with layered oxides showing higher thermal sensitivity (Ea≈70–75kJmol−1) than NFPP. Furthermore, temperature-indexed open-circuit voltage (OCV) measurements show that OCV–State of Charge (SOC) relations are both chemistry- and temperature-dependent. Layered oxides exhibit more pronounced low-temperature hysteresis and curve-shape changes, while NFPP remains more consistent within a mid-SOC window. Additionally, voltage-synchronous casing strain is robustly observable for layered oxides (NFM/CFM) under the present mounting configuration; by contrast, no resolved casing-level signature is observed for NFPP. Within NFM, charging at 0∘C exhibits a heavy-tailed constant-voltage duration distribution (outliers >1000min), indicating that fixed-voltage termination can induce pack-level SOC imbalance when the cell kinetics are becoming the limiting factor. In strain-enabled ageing measurements, an accumulating residual casing strain follows ∼t kinetics and correlates with capacity loss in NFM. Long-term cycling shows accelerated capacity fade relative to the other Na-ion chemistries. These results motivate chemistry-specific derating maps, adaptive end-of-charge supervision, pulse-horizon power calibration, and uncertainty-aware, multi-modal SOC and State of Health (SOH) estimation that are required in developing an adequate battery management system for Na-ion batteries.