• Journal of the Microelectronics and Packaging Society
• /
• v.15 no.2
• /
• pp.37-45
• /
• 2008

In this study, reliabilities of Cu (60 um)/SnAg (20 um) double-bump flip chip assemblies were investigated for the flip chip interconnections on organic substrates with 100 um pitch. After multiple reflows at $250^{\circ}C\;and\;280^{\circ}C$, bump contact resistances were almost same regardless of number of reflows and reflow temperature. In the high temperature storage test, there was no bump contact resistance change at $125^{\circ}C$ up to 2000 hours. However, bump contact resistances slightly increased at $150^{\circ}C$ due to Kirkendall voids formation. In the electromigration test, Cu/SnAg double-bump flip chip assemblies showed no electromigration until about 600 hours due to reduced local current density. Finally, in the thermal cycling test, thermal cycling failure mainly occurred at Si chip/Cu column interface which was found out the highest stress concentration site in the finite element analysis. As a result, Al pad was displaced out under thermal cycling. This failure mode was caused by normal compressive strain acting Cu column bumps along perpendicular direction of a Si chip.

• Journal of the Korean GEO-environmental Society
• /
• v.19 no.4
• /
• pp.5-16
• /
• 2018

In the current study, the engineering behaviour of prebored and precast steel pipe piles was examined from a series of full-scale field measurements by conducting static pile load tests, dynamic pile load tests (EOID and restrike tests) and Class-A and Class-C1 type numerical analysis. The study includes the pile load - settlement relations, allowable pile capacity and shear stress transfer mechanism. Compared to the allowable pile capacity obtained from the static pile load tests, the dynamic pile load tests and the numerical simulation showed surprisingly large variations. Overall among these the restrike tests displayed the best results, however the reliability of the predictions from the numerical analysis was lower than those estimated from the dynamic pile load tests. The allowable pile capacity obtained from the EOID tests and the restrike tests indicated 20.0%-181.0% (avg: 69.3%) and 48.2%-181.1% (avg: 92.1%) of the corresponding measured values from the static pile loading tests, respectively. Furthermore, the computed results from the Class-A type analysis showed the largest scatters (37.1%-210.5%, avg: 121.2%). In the EOID tests, a majority of the external load were carried by the end bearing pile capacity, however, similar skin friction and end bearing capacity in magnitude were mobilised in the restrike tests. The measured end bearing pile capacity from the restrike tests were smaller than was measured from the EOID tests. The present study has revealed that if the impact energy is not sufficient in a restrike test, the end bearing pile capacity most likely will be underestimated. The shear stresses computed from the numerical analysis deviated substantially from the measured pile force distributions. It can be concluded that the engineering behaviour of the pile is heavily affected if a slime layer exists near the pile tip, and that the smaller the stiffness of the slime and the thicker the slime, the greater the settlement of the pile.