Optimization of the thermal reliability of a four-tier die-stacked SiP structure using finite element analysis and the Taguchi method. (March 2018)

• Record Type: Journal Article
• Authors: Tang, Y.; Luo, S.M.; Li, G.Y.; Yang, Z.; Chen, R.; Han, Y.; Hou, C.J.
• Is Part Of: Microelectronics journal. Volume 73(2018)
• Page Start: 18
• Page End: 23
• Publication Date: 2018-03
• Subjects: System in package -- Taguchi method -- Finite element analysis -- Optimization -- Reliability; Microelectronics -- Periodicals; Microélectronique -- Périodiques; Microelectronics; Electronic journals; Journals - contents and abstracts; Periodicals; 621.3805
• DOI: 10.1016/j.mejo.2018.01.011
• Languages: English
• ISSNs: 0959-8324

Abstract

Abstract: We used finite element analysis (FEA) and the Taguchi method to investigate the reliability of a four-tier die-stacked system-in-package (SiP) structure and to achieve the optimal design of the SiP structure under thermal cycling. A finite element model of a four-tier die-stacked SiP structure was developed to simulate the thermal stress distribution in the stacked die during the thermal cycling test. Scanning electron microscopy (SEM) was used to observe the failure sites of the SiP structure. The results show that the thermal stress is concentrated in the top die and that the maximum thermal stress occurs at the outmost four corners of the bottom of the top die. A micro-crack appears in the top die, and the crack starts from the bottom of the die close to the corner of die 4. Nine geometric parameters of the SiP structure were chosen as the variable factors, and the volume-averaged maximum thermal stress Δ σ max was chosen as the quality factor. An L 12 (2 9 ) orthogonal array was applied in a Taguchi experiment to estimate the effects of the nine factors and to reveal the optimal design of a four-tier die-stacked SiP structure based on the results of the thermal cycling test. The thickness of the top die has an important effect on the reliability; more than 21.2% of the volume-averaged maximum thermal stress is reduced using the optimal design.