Crucial to the development and characterization of thermal interface materials (TIMs) is an understanding of the squeeze flow process that is commonly used to form thin bond layers in micro-electronic assemblies. A single model TIM, a dense, fairly monomodal suspension of submicron alumina particles suspended in a silicone-based resin, is first characterized as a Bingham fluid using a parallel disk rotational viscometer. Next, the model TIM is squeezed from ∼1 mm initial thickness to ∼.01 mm limiting thickness under nominally constant applied load (68 to 345 kPa) between 20 mm diameter aluminum plates in an axial compression test apparatus (the type commonly used for materials testing). The test plates are flat (∼10 μm flatness deviation over the plate) and smooth (Ra ∼ 20 nm), and are fixed in the test column with epoxy for optimum parallelism. Bond layer thickness is estimated using the LVDT built into the compression tester. The thickness measurement resolution is limited by LVDT noise of 10–20 microns. Squeezing forces are well above the ∼.02 N noise level of the 100 N load cell. Of the test system compliance, inertia, and friction, only the compliance is significant to our testing, and is corrected for. Squeeze flow tests of Newtonian standards are used to qualify the test process. In the case of the model paste, Bingham fluid model parameters from rotational viscometry are used in a lubrication model of squeeze flow that shows good agreement with the measured gap vs. time behavior during squeezing. Improved agreement is obtained by including plate flatness deviation and time-dependent force in the lubrication model. Parallel disk viscometry and squeeze flow testing of the base resin of the model TIM shows Newtonian behavior.
Reply to Discussion on “Analysis of Bingham fluid radial flow in smooth fractures”
