Characterization techniques, when used on electrodes after cycling and/or post mortem, are crucial to the improvement of battery performance. They can indeed pinpoint which failure mechanism is involved and suggest ways to overcome it. Unfortunately, in order to be more precise or save time, objects of interest are rarely studied in large numbers. One might then wonder how representative of the “real” battery they are. In this presentation, examples based on electrochemistry, XPS and EELS spectroscopies will serve to resolve this problem.

In order to highlight of the importance of statistical treatment, we benefited from the analysis of variance (ANOVA) in order to identify pertinent parameters in the aging of Li-ion batteries. To this end, two battery systems: Li₄Ti₅O₁₂|LiFePO₄ and Li₄Ti₅O₁₂|LiNi_1/3Mn_1/3Co_1/3O₂ cells, were aged. The aging conditions, i.e. the rates of charge and discharge and the usage profile, were chosen within an experimental design. It is noteworthy that, contrary to expectations, the evolutions of capacity and resistance are not generally influenced by the aging conditions…

Additionally, how a rarely-used technique, such as XPS imaging, can improve our understanding of the true repartition of chemical species at the surface of electrodes will also be shown. Electrode mixes (binder, active material, electron conductor) are often so complex that a single spectrum, however precise it may be, can lead to misinterpretation. For example, it might imply an interaction between species even though some of them are too far apart to interact, or lead to a compound composition even though two compounds should be considered. This will be demonstrated by virtue of XPS imaging, and by using lithium metal electrodes [R. Grissa et al., ACS Appl. Energy Mater. 2018, 1, 5694−5702] and solid electrolyte polymers that allow one to discriminate between these possibilities. The importance of PCA treatment of the data will be also indicated.

Finally, since microscopy techniques are notorious for obscuring the big picture when it comes to the electrode as a whole, and all the more so when high-level spectroscopy (Electron Energy Loss spectroscopy) is used, silicon electrodes were analyzed after cycling thanks to a very rapid and optimized phase analysis technique developed by M. Boniface et al. (Nano Lett. 2016, 16, 7381−7388). Although a technique of this kind is able to identify the reaction mechanism and solid electrolyte interphase in an aggregate, the universality of this behaviour in the electrode still needed to be demonstrated. The purpose of this part of the presentation is to prove that such is indeed the case, a conclusion resulting from the analysis of hundreds of particles, as well as a careful and absolute quantification of the species in the sample. Notably, the macroscopic characteristics of the electrode can be obtained from the sum of the microscopic data.