退火(玻璃)
材料科学
热冲击
热的
触电
计算机科学
工程物理
复合材料
机械工程
热力学
工程类
物理
作者
Haozhong Wang,Chunchao Lai,Bingxu Ma,Xiaodong Jian,Si Chen,Hongtao Chen,Laiwang Fang,Xiaofeng Yang
标识
DOI:10.1109/icept63120.2024.10668558
摘要
2.5D advanced packaging technology is a promising solution for “more than Moore”. TGV interposer has attracted growing attention for the high-frequency applications in 2.5D advanced packaging due to its low dielectric constant and loss tangent, etc. Some typical failure modes of TGV structure were discussed in the previous literatures. However, there is a lack of systematic research on the failure mechanisms of TGV under annealing and thermal shock condition. This paper focused on the failure mechanisms of TGV under different thermal conditions using both experimental and simulation methods. Annealing experiment was carried on the TGV sample at 400 °C for different periods. Crack appeared in the glass at the top corner of TGV after annealed at 400 °C for 5 min. With the increasing annealing period from 5 min to 30 min at 400 °C, the protrusion height of TGV increased from $0.61\ \mu\mathrm{m}$ to $1.21\ \mu \mathrm{m}$, and the simulated protrusion height increased from $0.50\ \mu \mathrm{m}$ to $0.89\ \mu \mathrm{m}$. The FEA simulation result showed that stress concentration was located in the top corner of TGV, the stress was 1456.63 MPa at 400 °C for 5 min. TGV samples were subjected to 0, 300, 500 and 1000 cycles under the condition of $-55\ ^{\circ}\mathrm{C}$ to 150 °C. Glass will crack around TGV-RDL after 300 thermal shock cycles. The stress was found concentrated around TGV-RDL according to FEA, which was 370.34 MPa. Crack appeared at the bottom corner of the TGV after 1000 cycles. The equivalent plastic strain at the bottom corner of TGV increased from 1.09 to 3.64 with the cycles increased from 300 to 1000. This article analyzes failure mechanisms of TGV and provides design guidelines for the pretreatment process of TGV. The simulation result offers information for the reliability evaluation of TGV.
科研通智能强力驱动
Strongly Powered by AbleSci AI