Determination of Lithium Diffusion Coefficient in FeS2 through Improved Galvanostatic Intermittent Titration Technique (GITT) Modeling

Abstract

Significant, recent efforts have been directed towards modeling lithium-ion batteries to optimize their performance and design. These models depend highly on physical parameters that must be evaluated experimentally for the system of interest. One of the most critical of these parameters is the chemical diffusion coefficient for lithium inside the active cathode material. Galvanostatic Intermittent Titration Technique (GITT) is a commonly used technique that can evaluate the diffusion coefficient, as well as other important information about the cell such as the reaction rate constant and open circuit potential. Despite its popularity, there remain several issues with the existing method of analyzing GITT data. Notably, the conventional method for extracting diffusion coefficients from GITT data relies on numerous assumptions, applies only at short discharge times, and lacks evidence validating the obtained coefficients.In this work, we describe a new method of extracting diffusion coefficients from GITT data. This method involves directly fitting a GITT pulse to a one-dimensional, single particle electrochemical model and can be applied for both ideal and non-ideal solution theory. We apply this approach to experimental measurements at two temperatures for the intercalation regime of FeS2, a conversion cathode material. Results are compared to the conventional method of evaluating diffusion coefficients from GITT data as well as diffusion coefficients obtained from electrochemical impedance spectroscopy. Diffusion coefficients are found to be higher for direct fitting, and more than an order of magnitude difference is observed between ideal and non-ideal diffusion coefficients. The modeling is also extended to a distribution of particles, and we show that the distribution behaves as a single particle with an effective particle radius proportional to the total volume to surface area ratio. Finally, the new diffusion coefficients are validated against discharge data for the cell, and it is shown that a reasonable prediction to the experimental data can be achieved using a single particle model. However, this is only possible with a non-ideal diffusion model, and high variation is observed across predictions obtained using other diffusion coefficient evaluation methods. This work demonstrates the susceptibility of electrochemical modeling to transport parameters and offers a simple, yet improved, method for evaluating diffusion coefficients from experimental GITT measurements.Supported by the Laboratory Directed Research and Development program at Sandia National Laboratories, a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-NA-0003525.