Development of Whole-Cell Diagnostic Techniques and Tools for Lithium-ion Batteries

Abstract

Whole-cell diagnostic methods and analysis tools are critical for characterizing lithium-ion batteries as we aim to increase the performance and lifetime of these devices while also minimizing safety concerns and cost. Diagnostics of whole-cells can be significantly more complicated than their half-cell counterparts because of the lack of a reference electrode, and complex way two active electrodes interact with each other to yield a whole-cell response. The complexity of whole-cell electrochemical methods adds a further burden to the quality and reproducibility of the experimental data used to validate the performance of whole-cell analysis tools. We create a dataset used in all subsequent analysis that is well replicated and is used to showcase the statistical attributes of a testing regime carried out using Samsung INR 18650-15M cells with NMC | Graphite chemistry aged to different states-of-health (SoH) at different charging rates and temperatures. The dataset includes measurements of open-circuit voltage (OCV) from low C-rate scanning along with differential analysis of OCV and capacity, electrochemical impedance (EIS) and nonlinear electrochemical impedance (NLEIS) measurements. Quadruplicate measurements were taken for nearly all conditions. Using data from our well-characterized cells, we adapt the half-cell Multi-Species, Multi-Reaction (MSMR) model into a whole-cell diagnostic tool via inclusion of whole-cell design parameters and cell charge balance constraints. The whole-cell model is first compared to experiments using literature reference values for the MSMR thermodynamic parameters. To improve fit quality, the MSMR thermodynamic parameters and electrode capacities are simultaneously fit to the OCV and differential voltage data, producing low error, high quality fits to experiments. Bootstrap analysis is performed to show the robustness of the fitting software to experimental noise and data sampling. The MSMR results quantify which insertion reactions are most responsible for capacity loss in each electrode, while also showing how slippage in the lithiation window, changes in useable capacity, and other properties evolve as the cell ages. Finally, in this work, we provided an experimental framework for nonlinear electrochemical impedance spectroscopy (NLEIS). Increasing the input AC signal from the classic small-amplitude linear limit to a moderate amplitude that produces a second harmonic in the output signal (but no other harmonics), then the first-harmonic signal remains a valid representation of the linear response, while the second harmonic signal introduces new physics to the analysis. We show how the second harmonic NLEIS spectra build from, but complements, the Warburg and interfacial charge transfer response of the cell, providing unique insights into the evolution of charge transfer symmetry at low SOC as the cathode ages during cycling. These results launched two additional studies, where we collected the linear and nonlinear impedance response over much tighter SoC ranges to try and explore the emergence of these second harmonic charge-transfer kinetics and higher-order thermodynamic properties. We use traditional equivalent circuit elements to analyze the linear EIS, and then derive nonlinear equivalent circuit elements to model the NLEIS. Here, we also show that with inclusion of thermodynamic information achieved through the MSMR model, these new nonlinear circuit elements can capture the behavior we see in the charge-transfer asymmetry as well as the direction and quadrant that these nonlinear low-frequency may extend into. Finally, we also employ a full-physics pseudo-2-dimensional model, to show the general validity of the results we see from using the simpler, empirical equivalent circuit models.