| dc.description.abstract | Numerous experimental methods are available to predict the hazards associated with thermal runaway (TR) and subsequent catastrophic failure of lithium-ion batteries (LIB), but these methods are time-intensive and costly. The current study provides a thorough review of these experimental methods which include closed-vessel gas sampling, accelerating rate calorimetry, cone calorimetry, and Tewarson calorimeters. The strengths and weaknesses of each experimental method as applied by various researchers are critically analyzed, and several shortcomings of current approaches are identified. Key deficiencies in current approaches include lack of control of reactant gases (i.e., ambient air or similar), inadequate heating rates that are not comparable to realistic conditions, and failure to measure condensable reaction products (e.g., water or liquid electrolyte). In lure of experimental approaches, an a priori modeling approach based on chemical equilibrium analyses (CEA) is proposed herein. Standard CEA software is limited in applicability, so that several improvements are required for accurate modeling. These improvements include prediction of electrolyte solution densities; inclusion of key reactant and/or product species and their respective thermodynamic properties; and accurate representation of high-temperature oxygen release from metal oxide cathodes. The current study focuses on addressing the first of these two improvements, but additional work is still required to fully address them. Future work will encompass resolving the third improvement (i.e., metal oxide oxygen release), model validation against available experimental data, and modeling of LIB failure scenarios to inform future designs. | en |
