Analyzing the behavior and shear strength of common adhesives used in temporary wafer bonding

Creating 3D integrated circuits (3DICs) by stacking electronic devices optimized for their substrate has the potential to increase system functionality while decreasing the overall footprint. Wafer thinning is a critical element in 3D integration and complicated by the fragile nature of a thin wafer. Thinning and post-process handling can be potentially damaging to the wafer. For this reason, the device wafer is often temporarily bonded to a support, or carrier, wafer to prevent fracturing or warping of the wafer. The adhesive used to create this bond has to demonstrate certain qualities to be practical and efficient. There are tradeoffs in material qualities and bonding/debonding processes. A lack of quantitative data published in the literature makes comparison of adhesives currently on the market difficult. Once process constraints are determined, an adhesive can be chosen and depending on the constraints, few options may be available. Due to the fact the wafer thinning process involves rotation of a vacuum test jig containing the wafer stack above another rotating platen, the shear strength associated with this wafer stack is of interest. The primary objective of this project is to provide quantitative data on the strength of various adhesives that may be used in the temporary bonding process to determine the best adhesive for our particular application. To meet this objective, a test procedure was designed to acquire data related to adhesive shear strength. In our test procedure, two wafers are bonded together with the adhesive and mechanically pulled on each end with a tensile tester that records the load versus extension. The load value drops sharply once the adhesive fails. The peak load is collected and represents the maximum amount of shear force the adhesive can handle before failure. With the data acquired from this test, one can compare adhesives to determine which material may be best suited for a particular application. Data was collected to determin- the range of load values at which certain adhesives fail. For each adhesive, care was taken to apply the adhesive systematically and to observe the characteristics of the adhesive during and after testing in order to determine the practicality of the adhesive for the application of interest in temporary wafer bonding. The shear strength of the adhesive needs to be large enough to withstand the thinning process, however one must debond the adhesive easily as well. Adhesives that exhibited similar behavior were grouped into different classes (e.g. waxes, tapes, films). In addition, a computer based model, through a simulation of the wafer thinning process, attempts to verify the experimentally determined characteristics of adhesives that appear to be best suited for temporary wafer bonding.