A ceramic vacuum chuck is used to hold large detector arrays flat while being attached parallel to a rigid substrate. Once held in the vacuum chuck, the component is typically seized by epoxy against a rigid substrate. The problem that interferes with this operation happens when the epoxy spreads to places where it is not wanted, even into the gap between the component and its vacuum chuck, and over electrical contacts that are intended for wire bonding.
In this innovation, the spreading epoxy problem has been solved by a combination of three essential countermeasures. First, a capture and spread (CAS) moat is cut into the rigid substrate, which defines the perimeter of the contact region. Second, the precise amount of epoxy needed to fill the space between the rigid substrate and the component is carefully metered out. In this step, the volume of the space is calculated down to the precision of 1 mm3. In the case of a large imaging array, each 1 mm3 will lift the surface by an additional half micron. Simply estimating the volume by a rectangular box is not adequate when the component contains unexpected undulations. Surface metrology can reveal the size of the undulations. The mass and density of the component give its volume, and that may provide information about the possible volume of any undulations.
The third countermeasure is the spatial placement of the epoxy so that two conditions are met: one is that the arrangement of epoxy flows without making voids. The other condition is that the expanding front of flowing epoxy reaches the CAS moat evenly. The CAS moat provides a huge safety factor by distributing any invading epoxy laterally through powerful capillary action. A great deal of excess epoxy can be captured in the CAS moat near the point of incursion.
The CAS moat is created in the substrate by carving out a trench roughly 400 microns wide by 400 microns deep. The substrate material in this case is a 1.5-mm thick aluminum nitride plate, which can be conveniently removed using a wafer dicing saw. In this work, four dicing lines were made in the substrate to form a rectangle. The ends of the dicing lines were sealed with tiny amounts of a ceramic epoxy so that the moat would not leak out of the edges of the substrate. The ceramic epoxy was cured and carved by a razor blade to make corners for the moat. Making the moat with an end mill would get around having to do this. The width and depth of the moat could be reduced if circumstances dictate.
This work was done by Todd J. Jones of Caltech for NASA’s Jet Propulsion Laboratory. NPO-48725