May_EDFA_Digital

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 22 NO. 2 10 side of the puck is due to the conductive filler. Config2 has several regions with clear transparent epoxy that allow the sample to be visible even though it is embedded in the epoxy puck. Figure 6b and c are low magnification SEM images of the cross-section at the left and right die corners, respectively. The die, C4 bumps, laminate, and BGA bumps are distinctly visible. No obvious charging or image distortion are observed. Figure 6d shows a higher magnification image of three C4 bumps closest to the die corner in Fig. 6b. The image is very crisp and all layers appear to be planar. Unlike the case of Fig. 4c and 5c, there is no additional outline or changes in contrast surrounding themetal features. In the highmagnification image of the die-C4 interface in Fig. 6e, the BEOL metal stack is clearly visible. No obvious charging, as seen in Fig. 4d and 5d, is induced in the underfill region. The considerable improve- ment in imagequality for Config2over Config0andConfig1 shows that the presence of the conductive epoxy sections provides adequate grounding and significantly reduces surface charging in theSEM. The same results have been observed in the right corner die in Fig. 6c, which is relatively further away from the conductive epoxy section compared to the left die corner in Fig. 6b. MODIFIED PUCK DESIGN ENHANCEMENT As previously mentioned, after the sample is impregnated in epoxy, it is difficult to perform subsequent analysis on it, as it is very challenging to remove the sample from the epoxy without damaging it. To resolve this, a further enhancement was proposed to the modified puck design. Rather than impregnating the sample directly into the epoxy, the epoxy puck was redesigned to have an empty slot in the central rib. The sample was inserted into the slot and then a lowmelting temperaturemountingwax was meltedand injected into the slot. Themount- ing wax used is very conformal, solidifies on cooling, and is visually transparent. Thewax material providesmechanical stability to the sample during the polishing process. After the sample is impregnated in the wax, it can bemechanicallypolishedusingconventional methods. After the cross-section polishing is completed and its characterization is com- pleted, the wax can be melted and removed by heating or by dissolving in acetone. This way the sample can very conveniently be removed with minimal to no damage from the epoxy puck. Figure 7 shows results of mechanically polishing a sample impregnated inmountingwax into a slottedmodi- fied epoxy puck incorporated with sections of conductive epoxy, referred to as Config3. The slotted modified puck with the sample is shown in Fig. 7a. The sample is housed in a slot in the central rib and the immediate vicinity of the sample is filled in with mounting wax. The darker contrast of the epoxy on the left side of the puck is due to the conductive filler, as in Fig. 6a. Figure 7b and c are low magnification SEM images of the cross-section at the left and right die corners, respectively. In addition to the die, C4bumps, laminate, andBGAbumps, twodistinct horizon- tal lines are observed that span the entire cross section: one above the die and the other below the BGA bumps. These lines are the top and bottom edges of the slot in the epoxy; the region within the two lines is filled in with mounting wax. Figures 7d and e are higher magnification Fig. 6 (a) Optical image of the sample in the modified puck with conductive epoxysections, Config2. LowmagnificationSEMimagesof the (b) leftand (c) right corners of the die in the cross-section. (d) Highermagnification SEM image of three C4 bumps closest to the die corner in (b). (e) High magnification image of the die-C4 interface. INNOVATIVE PUCK DESIGN FOR MECHANICAL CROSS-SECTIONING (continued from page 8)