Strategies for Consistent Alignment of Next-Generation Battery Electrochemical Data with Physics-Based Models

Abstract

Lithium-ion batteries (LIBs) have played a critical role in modern society and the clean energy transition by enabling electrified transportation, improving grid resilience, and powering consumer devices. However, not all demands for energy storage are likely to be met by current LIBs, motivating the swift development of next-generation batteries that offer enhanced performance metrics such as higher energy densities, wider operating temperature windows, improved safety, and lower costs than LIBs. To accelerate the pathway to commercialization for these emerging technologies, rational design strategies and data-driven methods can be utilized, where innovation is guided by the predictive relationship between a strong set of quantitative experimental descriptors and the enhanced performance metrics being optimized. As the only directly measurable variables in a sealed battery are voltage, current, and temperature, we often turn to physics-based models that can further describe these measured electrochemical signals with quantitative physicochemical parameters which can then serve as predictive features. Unfortunately, there is a lack of consensus on how electrochemical data for next-generation chemistries are generated and analyzed, leading to a lack of consistency on how physics-based models are applied. This lack of strategy for model-data alignment, combined with a dearth of high-quality electrochemical datasets, hinders the accelerated development of next-generation batteries. In this work, we explore experimental and modeling strategies to improve the consistency of how physics-based models are aligned with diverse electrochemical data for three promising next-generation battery chemistries. First, we demonstrate how C/20 galvanostatic cycling data from a homologous series of differentially-processed hard carbon anodes can be consistently parametrized using the Multi-Species, Multi-Reaction (MSMR) thermodynamic model, where the optimization landscape is filled with several local minima. Insights from this study were used to quantify reversible sodiation and clarify mechanistic considerations in the development of hard carbon anodes for sodium-ion batteries, a low-cost emerging technology. Next, we introduce second-harmonic nonlinear electrochemical impedance spectroscopy (2nd-NLEIS) as a characterization technique that is highly sensitive to subtle differences in the dynamic electrochemical response of the nearly symmetric anode-free electrochemical cell Li(bulk)||Li(plate)|Cu, enhancing consistent mechanistic interpretation and parametrization of EIS data generated from the same experiment. Here, we show the first-ever 2nd harmonic NLEIS measurement from bulk and plated lithium metal interfaces and compare features from a low-performing carbonate and high-performing localized high-concentration ether electrolyte to explore the applications of 2nd-NLEIS in high energy density lithium metal battery development. Finally, we establish a standardized EIS testing protocol and software data analysis pipeline for performing high-volume screening of solid-electrolyte materials for possible use in aerospace missions considered by NASA Langley Research Center, where all-solid-state batteries might meet stringent safety and thermal performance metrics. The resulting testing and software tool was utilized to down-select among 13+ candidate electrolytes from 4 crystal families of sulfide/halide materials for their performance over a temperature range of –20 ˚C to 150 C and cell stack pressure from 2-200 MPa. Cumulatively, the efforts to consistently align electrochemical data and physics-based models of next-generation batteries work towards generating robust quantitative physicochemical descriptors for data-driven methods, paving the way for accelerated development and eventual widespread electrification.