Durability of Aluminum-Copper Laser Welds for EV Battery Applications

Abstract

Battery packs in electric vehicles (EVs) are typically composed of lithium-ion cells with aluminum and copper as the positive and negative terminals, respectively. These terminals are interconnected in series through conductive tabs to deliver sufficient power output for EV operation. However, joining thin, dissimilar materials like aluminum and copper, each with unique thermo-mechanical and reflectivity characteristics, presents considerable challenges. Among the available joining techniques, laser welding stands out as as a promising method due to its high precision, minimal heat input, and capacity to achieve low electrical contact resistance and high mechanical strength. Additionally, laser welding can effectively limit the formation of brittle intermetallic compounds, which are common in dissimilar metal joints. This thesis investigates the feasibility of laser welding thin aluminum and copper sheets for use in electric vehicle (EV) battery packs, with a focus on understanding both the quasi-static and fatigue behavior of Al-Cu laser weld joints. To address the challenges of dissimilar metal welding, a process window was developed to produce consistent, robust joints using laser spot sizes of 0.6 mm and 0.3 mm. Through parametric optimization, an optimal combination of laser welding parameters was identified to maximize tensile shear strength and joint reliability. Microstructural analysis revealed the presence of two primary brittle intermetallic phases, Al + θ-Al2Cu and θ-Al2Cu, at the weld interface. Mechanical testing of cross-tension specimens under quasi-static and fatigue loading conditions was conducted, particularly in shear-dominant orientations of 90° and 67.5°. As anticipated, specimens tested at the 90o loading orientation generally exhibited a longer fatigue life under the same or similar load levels. Fractographic analysis showed that fatigue cracks typically initiated on the copper side of the weld interface, often at sites of weld porosity or at sharp notches where intermetallic compounds were prevalent. These regions acted as stress concentrators, influencing the crack initiation and propagation mechanisms. To assess the fatigue behavior, this study employed nCode Rupp’s model to transform load-life data into stress-life plots, effectively consolidating fatigue data and reducing scatter in life predictions. Statistical optimization using the generalized reduced gradient (GRG) method further refined the model, enhancing its precision by minimizing residual error and improving the coefficient of determination. With this approach, the study achieved a robust master curve that accurately predicted fatigue life and a R90C90 design curve for Al-Cu laser welds under shear-dominant loading conditions. Subsequent fatigue tests verified the reliability of this predictive model, supporting its applicability to real-world EV battery pack environments.