Self-Diffusion of Electrolyte Species in Composite Battery Electrodes Using PFG-SE MAS NMR for Better Understanding of Their Electrochemical Performance

Abstract

Long charging times and low power to mass output of current battery systems limit their widespread use for automotive applications. The maximum charging rate is believed to be ultimately limited by diffusion of the lithium ion in the pores of the electrode. The interplay between electrode microstructure, electrochemical performance and lithium ion diffusion must be further studied to improve battery engineering. Electrochemical modelling combined with complementary experimental measurements will enable the development of a tool that will allow battery engineers to tailor battery architectures and in turn increase their performance. Herein, the electrochemical performance of NMC532-based positive electrodes designed for EV application were analyzed with the penetration depth model. A very simple relation was postulated by Gallagher et.al. [1] relating the depth of penetration of the lithium salt in the electrode pore matrix with the ratio of the capacity of the battery at high C-rates (typically C or higher) against its capacity at low rates (typically C/100 or lower). DoD = (Qhigh C-rate/Qlow C-rate) = Ld/L = [(ε⋅D0⋅C0⋅F)/(TL(1-t+)I)] Where Ld is the salt penetration distance, L is the thickness of the electrode, ϵ is the porosity, T is the tortuosity, D0 is the salt self-diffusion coefficient in the electrolyte, C0 is the electrolyte concentration, F is the Faraday’s constant, I is the applied current density and t+ is the cation transference number. The penetration depth model was then used to reproduce experimental electrochemical data of the NMC532-based electrodes. Half-cells made with NMC532 positive electrodes and Li metal negative electrodes were assembled. The NMC532 electrodes were designed with varying degrees of loading, mass percentages of electrode components and calendaring to evaluate electrode microstructure effects against electrode performance. LP30 was chosen as electrolyte and the cells were cycled thru a power test. Three formation cycles of C/10 followed by a power series that includes charge and discharge steps ranging between C/10 and 10C. The temperature was also varied from 0°C to 40°C to evaluate its influence on the electrochemical performance. Electrochemical data shows that denser (less porous) electrodes exhibited less gravimetric capacity than more porous electrodes at higher C-rates whereas the opposite is observed at low C-rates. It was also observed that electrodes having increased active material content and decreased carbon additive content have increased gravimetric capacity than their counterparts. Results from these electrochemical tests were then compared to the fits obtained from the penetration depth model and the resulting fits were in good accordance with each other. The fitted self-diffusion coefficient D0 (which probes the ability of the lithium salt to diffuse into the electrode pores) was retrieved at various temperatures, displaying a trend parallel to the one measured for D(Li+) obtained from PFG-SE NMR for comparison, indicating similar activation energies. However, the effective D0 is smaller compared to the bulk electrolyte, even after accounting for tortuosity and porosity, highlighting other additional effects. Pulsed Field Gradient Spin - Echo NMR (PFG-SE NMR) probes the self-diffusion of spin-bearing species in liquids. Although the technique has been applied to measure self-diffusion coefficients of the electrolyte species when confined into separators and/or membranes, it has never been applied to battery electrodes.[2,3] Paramagnetic electrode materials prevent the measurement of PFG-SE signals, however, composite model electrodes with alumina can be studied by PFG-SE NMR if Magic Angle Spinning is applied to increase resolution. Therefore, the role of binder materials on lithium diffusion was probed by PFG-SE NMR with binder films wetted with LP30 electrolyte, or with alumina/binder/carbon black composites, shedding a new light on these complex systems. Acknowledgments Financial funding from the ANR program no. ANR-15-CE05-0001-01 is acknowledged.