Best practices for incremental capacity analysis

Abstract

This publication will present best practices for incremental capacity analysis, a technique whose popularity is growing year by year because of its ability to identify battery degradation modes for diagnosis and prognosis. While not complicated in principles, the analysis can often feel overwhelming for newcomers because of contradictory information introduced by ill-analyzed datasets. This work aims to summarize and centralize good practices to provide a strong baseline to start a proper analysis. We will provide general comments on the technique and how to avoid the main pitfalls. We will also discuss the best starting points for the most common battery chemistries such as layered oxides, iron phosphate, spinel or blends for positive electrodes and graphite, silicon oxide, or lithium titanate for negative electrodes. Finally, a set of complete synthetic degradation maps for the most common commercially available chemistries will be provided and discussed to serve as guide for future studies.

Differential Analysis of Galvanostatic Cycle Data from Li-Ion Batteries: Interpretative Insights and Graphical Heuristics

“…While the combination of processes in both electrodes within each full cell dQ/dV peak has been clearly explained through the "clepsydra analogy" and has been reiterated in best-practice literature, 24,25,34 we find this concept to still be routinely misunderstood or overlooked in applications reported by other practitioners. For this reason, we include here some supplementary demonstrations.…”

“…The above concepts are explored by simulated differential analysis for the LCO|graphite system (Figure 18). To explore a range of candidate degradation processes, 25 we consider two exemplar cases: LAM PE up to 10% (zero LAM NE and LLI) and LLI up to 10% (zero LAM PE and LAM NE ).…”

“…There exists a substantial amount of literature on the relationship between experimental protocol selection (e.g., charge/discharge rate, data collection frequency) and postprocessing factors (i.e., downsampling rate, smoothing etc.) that may affect the output peak shapes, and by extension any qualitative or quantitative understanding of performance losses ,,, or models derived therefrom. To generate robust, artifact-free differential analysis plots, carefully considered smoothing and/or downsampling of raw data is required. ,,,− Feng et al addressed practical issues specific to differential analysis methods .…”

“…A recent best-practice paper expressed the concerns that "[differential capacity] analysis can often feel overwhelming for newcomers" and "the [differential analysis] literature is now a minefield with frequent inaccuracies and fundamental errors". 25 One challenge arises because, despite the close relation of differential voltage and differential capacity analysis, many of the early works introducing these analytical techniques focused exclusively on one technique or the other (i.e., exclusively differential voltage 2,19 or differential capacity 6,23 ). Although it has become recognized that full cell analysis benefits from a combined analysis (IC-DV), not all necessary aspects of a unified theory have been effectively recapitulated.…”

“…6,20−24 Synthetic cycling data derived from the 'alawa toolbox 23 has been used to propose state-of-the-art best practice for incremental capacity (differential capacity) analysis for several cell chemistries. 25 Especially since 2010, differential analysis has extended to commercialscale full cell configurations and pack-level analyses, demonstrating its wide relevance to battery analysis. 17,21,23,26−29 Works since 2017 have emphasized the concurrent application of both differential capacity and differential voltage analysis (IC-DV technique).…”