Issue link: https://iconnect007.uberflip.com/i/1520213
82 SMT007 MAGAZINE I MAY 2024 is graph is difficult to interpret due to the high sample-to-sample shape variation at room temperature. However, if we subtract the matrix of the original room temperature shape of each sample, the warpage becomes clearer. e "PT0005" comes out as a clear outlier in this data set. e relative warpage data is plotted against the 3S warpage gauge in Figure 18. Figure 18 shows that analysis of the data via relative change can be helpful to interpret the data. Analysis and Discussion Further analysis requires a closer look sam- ple by sample, considering the construction of the part, die-to-mold ratios, substrate thick- ness, and relative shape change of the samples. 20 x 15 mm sample Starting with the simplest case, the 20 x 15 mm sample showed no discernable change from the different moisture exposure levels tested in this study (Figure 13). is is an unsurprising outcome, given that it is also the only bare die sample and contains no mold compound. e die is also quite large relative to the package size. While the package does show shape change as the substrate expands at a faster rate than the die, the shape change is perceivably unaffected by moisture, with no mold to soak up moisture and the relatively thin substrate not holding enough water to affect the warpage. 27 x 27 mm and 40 x 40 mm samples e two larger samples in the study exhibit the most obvious cases of moisture affect- ing warpage. Both samples have prominently thicker molded areas and small die-to-mold ratios. e higher warpage is clear for MSL3 and MSL4 for the 27 mm package and clear for MSL3, the worst-case moisture exposure, on the 40 mm package. It is unclear specifically why the MSL4 exposure matched better with the control and MSL3 + bake warpage than the MSL3. Figure 18: Relative 3S warpage, 14 x 14 mm, MSL3.