Skip to Main Content
This paper presents the thermal cycling test results for edge and corner bonded 0.5 mm pitch lead-free package stackable very thin fine pitch ball grid arrays (PSvfBGAs) as package-on-package (PoP) bottom packages on a quasi-standard JEDEC thermal cycling test board. Thermal cycling tests were carried out based on the JEDEC JESD22-A104C standard, -40-125°C temperature range with 32 minutes/cycle test condition was applied, and hold times in each extreme temperature were 10 minutes. The daisy chain resistance of each PSvfBGA was measured by an event detector and a 1000 Ω, rise (or more) from initial resistance was considered as the failure criterion. Three materials used in this study were a UV-cured acrylic edge bond adhesive, a thermal-cured epoxy edge bond adhesive and a thermal-cured epoxy corner bond adhesive, and PSvfBGAs without bonding were also tested for comparison. Statistics of the number of cycles-to-failure for the 45 components of each test group are reported. The test results show that the characteristic lives of PSvfBGAs with edge bond acrylic, edge bond epoxy and corner bond epoxy are respectively about 0.6, 0.9 and 1.3 times of the PSvfBGAs without bonding. The corner bond epoxy with lowest CTE, highest storage modulus and smallest adhesive volume yielded best performance. Because it can slow down the CTE mismatch and reduce effective action area of interactive forces between the adhesive, BT substrate and PCB during thermal cycling test. So if we want to improve the thermal cycling reliability of PSvfBGA with adhesives, the thermal-mechanical material properties and dispending patterns of adhesives should be thoroughly considered. Failure analysis was performed using dye-and-pry, cross section scanning electron microscope (SEM) and energy dispersive X-ray (EDX) methods. The common failure mode observed is solder cracking near package and/or PCB pads.
Date of Conference: 8-11 Aug. 2011