A Novel Approach for Classifying Battery and Pseudocapacitor Materials Using Capacitive Tendency and Supervised Machine Learning

Abstract In recent decades, there have been more than 100,000 scientific articles dedicated to developing electrode materials for supercapacitors and batteries. A heated debate nonetheless persists surrounding the standards for determining electrochemical behavior involving faradaic reactions, since the electrochemical signals produced by the various electrode materials and their different physicochemical properties often complicate matters. The difficulty lies in determining which group these materials fall into through simple binary classification as there can be an overlap between battery and pseudocapacitor signals and because both materials are faradaic in origin. To solve this conundrum, we applied supervised machine-learning toward a statistical analysis of electrochemical signals, and consequently developed a new standard which we called capacitive tendency. This predictor not only surpasses the limitations of human-based classification but also provides statistical tendencies regarding electrochemical behavior. Notably, and of particular importance to the electrochemical energy storage community publishing over a hundred articles weekly, we have created an online tool for easy classification of their data.