Impact of Process Improvements on Reliability of RCP Technology During Scale-up

Abstract

The start-up of the 300mm RCP prototype production line using automated tools was challenged by significant differences in tool sets when compared to the manual and semi-automatic lab based equipment used to develop the baseline RCP fabrication process. Thus, process transfer to automated tools required key parameter adjustments in an attempt to match the process output and reliability performance previously achieved with the 200mm lab based flow. During reliability testing of early parts from the 300mm prototype line, failures in the build-up layers were observed after thermal cycling test. The failure mechanisms were previously observed and resolved during the technology development phase using the 200mm lab based tools. An experiment was developed to identify the root cause of the build-up failures. A single die, 9x9mm, 258 I/O, 0.5mm pitch package with 2 metal routing layers was chosen as a test vehicle for the experiment. Moisture preconditioning and air to air thermal cycling at 2 different conditions followed by electrical test and visual inspection for dielectric cracks were used to gage reliability performance. Several factors were evaluated related to the dielectric and metallization processes as well as the backend interconnect process. The results from the experiment indicated the appropriate tool parameters that allowed the 300mm prototype line process output to closely match that of the 200mm lab based tools. The optimized process combined with the use of an alternate solder material allowed the 300mm RCP prototype line to successfully achieve line certification status and demonstrated RCP reliability performance beyond standards.