No abstract

No abstract

No abstract

We have examined three types of die-attach adhesives by measuring curvatures in assemblies that are formed by adhesively joining silicon chips to alumina substrates. The adhesives differ mainly in their post-cure elastic moduli, which are 24 GPa for a solder-like adhesive, 2.7 GPa for an epoxy-like adhesive, and 0.003 GPa for a gel-lie adhesive. All of the as-prepared assemblies have residual bending resulted from the adhesive curing process. The assembly bonded by the gel-like adhesive has substantially lower, about a factor of ten less, bending than the assemblies attached by the other two types of adhesives. Thermal responses of assembly curvatures are measured by subjecting the assemblies to a temperature excursion from -15°C to 160°C. The measured temperature dependences of curvature in all three types of assemblies are found to be in good agreements with the theoretical derivations by Suhir [l] for thermally induced bending in finite m-material structures.

An experimental study was performed to investigate various heat sinks and attach methods commercially available for the high-performance 1 19-lead plastic ball grid array (PBGA) package. Attach methods investigated in this study include a pressure-sensitive adhesive, mechanical clips, and thermally conductive adhesives. This work will show the effects of heat sink attach methods on the package structural integrity and solder ball interconnect reliability. This work is necessary to better provide total solutions for customers who will need to dissipate high levels of power from this package.The heat sinks were attached to the 119-lead PBGA packages which were mounted on multi-layer printed-circuit boards. These assemblies were then subjected to accelerated-life testing (ALT) using air-to-air thermal cycling between 0 "C and 100 "C to determine the effect of attaching heat sinks to the packages on the electrical reliability of the solder joint connection. Besides interconnect reliability, ALT was used to determine the effect of thermal cycling on the reliability of the heat sink attach mechanism. The heat sink attach mechanisms were tested by three different methods. The first was to simply observe the heat sinks to see if they fall off during thermal cycling. Mechanical shock tests then were performed using a package drop tester. Samples that underwent 1 OOO, 2 OOO, and 11 400 thermal cycles, as well as no thermal cycling, were subjected to this test. Finally, shear strength tests using a constant strain-rate mechanical testing machine were performed on the packages with heat sinks attached using adhesives.Several modes of failure appeared in the experiments. These failures included extremely low interconnect life in thermal cycling, heat sink detachment during drop tests and during thermal cycling, and low adhesion strength in shear tests. While some attachment methods were obviously not suitable, others provided robust solutions to the heat sink attachment problem.