What Can Text Mining Tell Us About Lithium‐Ion Battery Researchers’ Habits?

Abstract

Artificial Intelligence (AI) has the promise of providing a paradigm shift in battery R&D by significantly accelerating the discovery and optimization of materials, interfaces, phenomena, and processes. However, the efficiency of any AI approach ultimately relies on rapid access to high‐quality and interpretable large datasets. Scientific publications contain a tremendous wealth of relevant data and these can possibly, but not certainly, be used to develop reliable AI algorithms useful for battery R&D. To address this, we present here a text mining study wherein we unravel lithium‐ion battery researchers’ habits when reporting results, reason on how these habits link to issues of lacking reproducibility and discuss the remaining challenges to be tackled in order to develop a more credible and impactful AI for battery R&D.

Understanding Battery Interfaces by Combined Characterization and Simulation Approaches: Challenges and Perspectives

“…In that respect, the battery field, when compared to other technologies such as photovoltaics or organic chemistry applied to pharmaceutics, suffers from a major disadvantage: the lack of standardization for cell assembly, testing protocols, data acquisition, and analysis. Hence, despite the enormous volume of scientific publications (more than 30000 currently [ 42 ] ) and internal reports on battery technologies, a scattering of results, when indeed sufficient information is provided for comparison, is widely observed for given chemistries. [ 42 ] This difficulty is not new, but can hardly be solved when cell format, cycling protocols, and consequent performance metrics differ for each battery chemistry and application.…”

“…Hence, despite the enormous volume of scientific publications (more than 30000 currently [ 42 ] ) and internal reports on battery technologies, a scattering of results, when indeed sufficient information is provided for comparison, is widely observed for given chemistries. [ 42 ] This difficulty is not new, but can hardly be solved when cell format, cycling protocols, and consequent performance metrics differ for each battery chemistry and application. Mature technologies such as alkaline primary batteries and MnO 2 electrode materials were developed via the standardization of cell formats, but more recent technologies, notably Li‐ion batteries, still do not benefit from the implementation of such similar standards.…”

Perspectives on manufacturing simulations of Li-S battery cathodes

“…Indeed, such approaches have been successfully applied to the field of materials science [133,134]. However, El-Bousidy et al from our group [135] showed, through similar text mining techniques, that LIB research articles show a general lack of consistent reporting of electrode properties, such as thicknesses, porosities, loadings, electrolyte volumes or surface areas. This poses serious problems for the completeness of the reported data, which hampers the correct prediction of battery properties through ML models, and brings out yet another example of the reproducibility crisis common to various scientific disciplines [136].…”

Data Reporting Checklists to Maximize your Impact

“…There are many papers published on the subject; while they offer sage advice from recognized authorities, they aren't necessarily themselves solutions. [1][2][3][4] As a service to our authors, reviewers and readers, we introduced the Emerging PV Reports (Figure 1), a collaboration with the Helmholtz Institute Erlangen-Nuremberg for Renewable Energy, Germany. This initiative aims to accelerate the development of emerging PV materials towards market readiness and to improve the reproducibility and quality of the associated data.…”