Evaluation of raw substrate variation from different suppliers and processes and their impact on package warpage

Thin substrates have been used in more and more package-on-package (PoP) designs to meet the overall package thickness requirement. Low CTE cores are becoming more popular to reduce thin package warpage. On the other hand, substrates used in the same product are often sourced from multiple suppliers. Packages built with thin substrates sourced from different suppliers were found to have different end-of-line (EOL) package warpage. In this paper, 5 legs of substrates from 3 different suppliers were studied and compared with regard to raw substrate warpage, raw substrate modulus and CTE properties, and their reactions to 1× reflow thermal conditioning in order to understand any correlation to end-of-line package warpage. It was found that raw substrates sourced from different suppliers, or different processes in the same supplier, could have different levels of initial bare substrate warpage due to residual stress. Simulation results showed clear correlation between bare substrate warpage and EOL package warpage. However, such correlation was not observed with the limited measurement data collected. It was also found that properties (CTE and modulus) of finished composite substrates from different suppliers and processes could vary significantly, especially in the high temperature range. The difference in properties could be correlated to the difference at end-of-line package warpage in some cases. Further more, the substrates from different suppliers or processes could change their warpage, modulus and CTE properties in different ways after 1× reflow temperature conditioning. The study shows that it becomes more and more important to have better quality control of substrates sourced from different suppliers as substrate becomes thin and low CTE core is used.