Smart Battery Perceiving Future Lifetime at Ultra-Early Life Stage by Decoding SEI Evolution

Accurate lifetime prediction is critical for the reliable use of lithium-ion batteries. However, batteries at early stage commonly show consistent electrical behaviors impeding distinguishing their future life trends. In this work, the internal strain is monitored in-situ by implanting fiber Bragg grating (FBG) sensors into 18 cells belonging to 3 different electrochemical systems. We find the internal strain at the formation stage exhibits a biphasic feature attributed to the dynamic competition between solid-phase growth and gas-phase accumulation. The SEI formation strain (SFS) is further found to be linked closely to the thickness and compactness of SEI which influence the degradation path of battery. The high correlation between SFS and capacity loss is further explored to build an elastic net model for battery lifetime prediction, achieving a mean absolute percentage error (MAPE) of 5.40%. This work presents a new avenue for understanding the mechanism of SEI formation and predicting the battery lifetime at the ultra-early formation stage.