Ultrasonic Characterization of Ethylene Vinyl Acetate (EVA) Crosslinking for Quality Assurance and Lamination Process Control

To increase throughput and reduce cost, module manufacturers seek to minimize lamination processing times. However, too short lamination result in poor Ethylene-Vinyl Acetate (EVA) properties due to unfinished crosslinking and a high content of residual aggressive reaction starters in the completed module. These modules are prone to aging, delamination, mechanical fatigue, corrosion, and yellowing, severely reducing long-term reliability. Identification of unfinished crosslinking is crucial for lamination process surveillance, but since material and process conditions vary locally within the module (and over time), destructive characterization techniques (e.g. differential scanning calorimetry (DSC), Soxhlet extraction or peel testing) on some selected samples only provide limited insight. For that reason, we propose a non-destructive method that uses ultrasound to characterize crosslinking within the solar module. High-frequency longitudinal ultrasound is transmitted through the thickness of the solar module and analyzed. By knowing the thickness of the polymer, the speed of sound and the frequency-dependent ultrasonic attenuation, the degree of crosslinking can be determined. In this paper, we demonstrate an experimental correlation between the DSC degree of crosslinking and the speed of sound, as well as the frequency-dependent attenuation coefficient. After an initial drop in velocity during initial melting (or humidity evaporation), sound velocity and attenuation increased with curing duration. This correlation can be utilized to determine the degree of crosslinking. However, a calibration of the material-specific model with conventional methods is required (e.g. via DSC analysis). The proposed ultrasonic method then allows a non-destructive characterization of the module lamination process, capable of detecting inhomogeneities and deficiencies in crosslinking, to ultimately drive process optimizations, increase production yield and reduce the LCOE.