Here is the abstract you requested from the DPC_2007_3D technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.
|Assembly and Cooling Technology for 3-D VLSI Chip Stacks|
|Keywords: 3-D, Chip stacking, 3-D cooling|
|Though a great deal of work is going into wafer stacking for 3-D VLSI ICs using through silicon via (TSV) processing, wafer assembly may not be the best way to build 3-D systems from VLSI elements. Systems need more than silicon. In the case of our 3-D RF program, we need to assemble both GaAs and silicon layers with TSVs, remove significant amounts of heat, provide integrated decoupling and power distribution, and avoid cumulative yield issues by pretesting individual die layers before assembly. As each layer is a miniature subsystem, Z-axis interconnect demand can be satisfied by connections on 100 - 200 micron pitch. This paper describes our progress in developing an integrated assembly and cooling method for 3-D VLSI systems. To deal with the CTE mismatch of different materials, plated compliant copper posts are used to make connections between layers. The copper is also used to form a fluid dam around the resulting channel between each pair of die. Circulating fluid is pumped through these channels for heat removal. Sequential attachment of pretested layers is accomplished by forming copper/tin intermetallics. We have developed a system for plating ~ 100 micron high, 30 micron diameter copper posts using KMPR photoresist. At the same time, we plate the fluid confinement dam that will be joined to the successive chip. These structures are topped with plated tin, and can be mated to a corresponding set of copper pads on another silicon layer using a flip chip bonder. Cooling fluid is circulated through the resulting gap. The results of test vehicles intended to demonstrate the effectiveness of these methods will be presented. This novel concept for packaging miniature 3-D VLSI systems appears well-suited to high power RF applications.|
|Leonard W. Schaper, Professor
University of Arkansas