May_EDFA_Digital

edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 22 NO. 2 6 The same polishing conditions such as rotational veloc- ity, vertical down force, and consumables were applied Fig. 1 (a) Top-view schematic of the flip-chip packaged sample, the dicing locations, and the polishing direction. (b) Cross-section viewof the sample. (c) Magnified view of the region highlighted by the red box in (b), showing the C4 bumps connecting the die to the laminate. (d) The diced sample impregnated in an epoxy puck. Fig. 2 Optical images of the (a) conventional and (b) modified epoxy pucks, respectively, after polishing. (c) Thickness of the material removed for the conventional pucks and modified pucks as a function of polishing time. (a) (b) (c) (d) (a) (b) (c) consistently across all samples. Polishing abrasives and consumables used were 600 grit SiC sandpaper, polish- ing cloths, 1 μm, and 0.3 μm alumina suspensions, with water as the polish- ing lubricant. Prior to SEM imaging, all cross-sections were sputter-coatedwith an identical thin layer of Cr to reduce surface charging. RESULTS AND DISCUSSION The design of the epoxy puck was modified to remove sections of theepoxy that were not in the immediate vicinity of the sample. The circular outer shape with a low aspect ratio was retained for stability andpreservationof a flat polish- ing profile. The modified design was a continuous epoxy puck that consists of an outer ring for stability, a central epoxy rib to house that sample, and two semi- circular cut out sections in the epoxy on either side of the rib. POLISHING RATE COMPARISON FOR CONVENTIONAL VS. MODIFIED BLANK EPOXY PUCKS To validate the hypothesis that reducing the cross-sectional area and introducing outer ring edge regions would indeed improve the MRR and reduce the polishing time, polishing rate experiments were set up using epoxy-only pucks, without impreg- nated samples. Six conventional blank pucks were polished, three pucks at a time, using 600 grit sandpaper. The pol- ishing times were carefully monitored and the puck thicknesses removedwere recorded at multiple points on each puck at time intervals of 60 min. The same procedure was repeated using the modified blank pucks. Figures 2a and b are optical images of the conventional andmodified epoxy pucks, respectively, after polishing. The graph in Fig. 2c compares the thickness of the material removed for the conventional pucks and modified pucks as a function of time.