An Early‐Stage Bond Wire Lift‐Off Defect Identification Method Based on Surface Temperature Rise Characteristics

Bond wire lift‐off is one of the primary failure modes in power modules and early identification of its degradation state and lift‐off ratio is crucial for ensuring reliable operation. Conventional methods have limitations in quantifying early‐stage lift‐off defects. To address this issue, this study begins with the equivalent resistance variation caused by lift‐off, establishes a coupling model between resistance and temperature characteristics based on the equivalent circuit of insulated gate bipolar transistor (IGBT) modules and validates the feasibility of using the temperature field to characterize bond wire states. Furthermore, a refined finite element model is developed to reveal the influence of lift‐off defects on surface temperature distribution and a correlation model between lift‐off ratio and temperature rise is established. On this basis, a defect identification method based on rapid reconstruction of surface temperature using discrete monitoring point data is proposed. Experimental results show that the proposed method achieves an average relative error below 0.11% with a root mean square error (RMSE) within 0.5 K and improves computational efficiency by approximately 114 times compared with the full‐order model. Moreover, the predicted lift‐off ratios are consistent with experimental measurements. The proposed method provides an effective approach for bond wire lift‐off defect identification.