February_EDFA_Digital

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 20 NO. 1 24 (Fig. 5). Such fine spacings between Ag 3 Sn precipitates were observed at essentially all points in the tin matrix of the copper pillar solder joints. In contrast to larger lead-free solder joints (see, for example, Fig. 1), where recalescence causes much of the joint to solidify at or near the melting temperature of the alloy, these much smaller copper pillar solder joints would not increase in temperature as much after initial solidification. Thus, a morehomogeneous distributionof fineprecipitateswould be expected to be observed (Fig. 5, 6). Aging these copper pillar solder joints resulted in a dramatic decrease in the number of Ag 3 Sn precipitates in the solder joint. Aging was conducted at a temperature of 125 °C for a period of 1000 h, resulting in significant coarsening of the precipitates (Fig. 6). The number of precipitates in a cross section of a typical solder joint decreased from several hundred to less than ten for all silver concentrations (Fig. 7). Measurements of shear strength were conducted on copper pillar/SnAg solder joints to examine separately Fig. 5 SEM micrographs of copper pillar pad diameters of 100 µm with different solder compositions, as noted in the figure on the silicon chip. The Ag 3 Sn precipitate morphology was observed to change with silver composition in the solder. Fig. 6 Backscattered SEM images of two category copper pillars on the left after reflow. The right is after 1000 h (6 weeks) of aging at 125 °C. The number of Ag 3 Sn precipitates has decreased, while their size is increased. Pad diameter of pillars is 30 µm on the silicon chip. Fig. 7 Quantititave analysis for the number of Ag 3 Sn preciptitates as a function of aging time for pad diameters of (a) 100 µm and (b) 30 µm (b) (a)

RkJQdWJsaXNoZXIy MjA4MTAy