ET06.01.06 : Long-Term Calendar Degradation in Li-Ion Batteries

11:00 AM–11:15 AM Nov 26, 2018 (US - Eastern)

Hynes, Level 3, Room Ballroom A

Aziz Abdellahi1 Berislav Blizanac1 Brian Sisk1

1, A123 Systems LLC, Waltham, Massachusetts, United States

With the increased penetration of electrified vehicles in the automotive market, requirements pertaining to battery durability are becoming increasingly stringent. To meet the requirements of the automotive industry, lithium-ion batteries must exhibit extensive life before reaching a terminal state of capacity loss and impedance growth. For battery designers and manufacturers, it is therefore of paramount importance to understand and predict long-term battery cell degradation based on a necessarily limited set of accelerated degradation tests.

Long-term calendar aging, defined as the temperature-induced cell degradation in the absence of current, is especially difficult to predict at relevant battery operating temperatures (25oC – 45oC). Unlike cycling tests, which can be rapidly conducted to the end-of-life by removing rest periods between cycles, calendar tests cannot be directly accelerated. To this end, a variety of empirical and physics-based models have been developed to predict the long-term storage behavior of battery packs based on a set of accelerated storage tests conducted at high temperatures. However, the validity of these calendar predictions has not, to the best of our knowledge, been extensively studied against actual long-term storage data surpassing the 4 year mark.

In this presentation, we present a set of long-term storage experiments performed over the course of 4-to-6 years on LiFePO4/graphite cells, at various states of charge and temperatures. Analysis of the storage data sheds light on the long-term degradation mechanism in the cell, and demonstrates a transition between a reaction-controlled to a diffusion-controlled growth of the anodic solid electrolyte interphase (SEI). The dependence of state of charge and temperature on the degradation rate is clarified, and the predictive performance of empirical calendar life models is assessed. This work provides a mechanistic analysis of the nature of long-term degradation mechanisms in Li-ion batteries and paves the way towards an improvement of the predictive ability of empirical calendar life models. The conclusions of this study can also serve to understand long-term calendar degradation in higher-voltage NMC/graphite batteries, in which both the anode and the cathode may experience calendar degradation at high states-of-charge.