Investigating Failure of Li-Ion Batteries Using in Operando Electrochemical-Acoustic Time-of-Flight Analysis

Wednesday, October 14, 2015: 09:00
106-B (Phoenix Convention Center)
A. Hsieh, S. Bhadra, P. J. Gjeltema (Princeton University), and D. A. Steingart (Princeton University)
Electrochemical-Acoustic Time-of-Flight (EAToF) analysis takes advantage of the fact that in all closed electrochemical energy storage systems, regardless of the reaction mechanism (intercalation, dissolution/reprecipitation, phase change, etc.) the density and elastic modulus of each electrode changes as a function of its state of charge (SOC). This, in turn, affects how sound waves behave as they travel through the cell. Additionally, as a battery is cycled, progressive formation/degradation of critical surface layers, mechanical degradation of electrodes, consumption of electrolyte, etc. also affect the behavior of acoustic waves passing through the cell. Thus, the distributions of density and modulus within the cell as well as the rates of change of these distributions (and the resulting effect on the acoustic echoing behavior) can act as a fingerprint for the state of health (SOH) of the cell. In a previous study, we have demonstrated this in operando EAToF analysis experimentally using ultrasonic pulses during normal cycling of a wide variety of battery chemistries and form factors.

Here, we focus on the effect of long-term cycling on the electrochemical-acoustic fingerprints, particularly for commercially-available Li-ion 18650 type cells. We examine how the EAToF patterns evolve as a function of depth of discharge and cycling rate, and also investigate the effect of “abusive” cycling conditions (e.g., elevated temperatures, over-charge/discharge, etc. as prescribed by UN 38.3 and IEC 62133 standards) on acoustic behavior. Two failure modes that we study in detail are electrolyte consumption and lithium plating in Li-ion batteries, and we verify failure modes with post-mortem analysis. Our results show that the fingerprints from EAToF analysis, perhaps in combination with various numerical analysis methods, can be correlated not only to SOC during normal cycling, but also to SOH during extended cycling and catastrophic failure events.