Issue link: https://iconnect007.uberflip.com/i/1288481
98 PCB007 MAGAZINE I SEPTEMBER 2020 convection reflow assembly simulation (IPC- TM-650 2.6.27B) and current-induced thermal cycling (CITC, IPC-TM-650 2.6.26A) are ex- plored in this study. Dr. Cauwe began by updating the results of Phase I testing in the ESA program. The re- sults indicated that for the three build-up lay- ers, the semi-stacked inside (L1-L2/l3 stacked) was superior (no failures) compared to the semi-stacked outside (L1/L2-L3) construction for both the 0.8-mm pitch D-coupons and for the 0.5-mm D-coupons using polyimide mate- rials. The modified high-Td epoxy SI materi- als passed at both pitches, but he noted that the single-prepreg used was 25 microns thin- ner. This was verified in failure analysis that the high Td SI materials strain was 9960 ppm at 190°C compared to the polyimide's 11340 ppm at 210°C. Theoretically, it is 22% lower than the polyimide due to CTE-Z and thick- ness, verified by modeling. Also, IMEC used fi- nite element analysis to help determine the ef- fects of microvia interfacial stress. Jerry Magera What was the source of the weak microvia inter- face? According to Jerry Magera, senior staff prin- cipal engineer at Motoro- la Solutions, it all began with the microvia target pad in his presentation on "The Complete Path to Least Resistance." He focused on the often maligned electroless copper process and proclaims that the IPC- 6012E performance specifications for metalli- zation for PCBs of "sufficient for subsequent plating" is too vague, sets low expectations for deposit quality, and is the reason optical mi- croscope views of well-formed electrolytically copper filled microvias fracture during solder reflow thermal excursions. Continuous resistance measurements dur- ing component reflow assembly revealed ther- mally induced microvia failures that were sub- sequently located by cross-section analysis at and in the vicinity of that electroless cop- per deposit. The fascinating elegance of elec- troless copper is not appreciated, relegated to- day as a transient step in the copper metalliza- tion process, with function reduced to a con- ductive liner within the laser-ablated microvia cavity that bridges the target pad to the adja- cent copper layer to support electrolytic copper fill plating. However, it must form a metallurgi- cal bond between the target pad and electrolyt- ic copper plating to survive reflow assembly. Four-wire resistance measurements of micro- via daisy-chains confirmed substantial chain to chain variation attributed to the electroless copper deposit in the microvia. The results of four-wire resistance measure- ments completed on simple L1L2 and L3L4 mi- crovia daisy-chains at ambient temperature are presented for samples prepared with produc- tion-ready processes by PWB manufacturers. The measurements objectively revealed the ac- tual variation in the quality of the electroless copper deposit lining the microvias that was missed by weight-gain and backlight assess- ments. Published electroless copper deposit thickness ranged from 0.3–3.0 µm for immer- sion times of 4–30 minutes, depending on the process implemented. Structural variations exist in microvia field- failure interfaces on products that had passed existing IPC and OEM validation tests and mi- cro-section inspections but had retroactive- ly failed IPC 2.6.27A testing and been inves- tigated using micro-sections produced by fo- cused-ion-beam (FIB) trench-machining tech- niques. These revealed defects that had been missed by conventional optical microscopy. Three modes of interface failure had been ob- served: between electroless copper and target pad, between electrolytic copper fill and elec- troless copper, and within the electroless cop- per. There was often a mix of all three failure modes. Scanning electron microscope (SEM) exami- nation of the surfaces of target pads after laser drilling and electroless copper showed some interesting variations in grain structure, which Magera explained in terms of the different rates of growth of (100) and (111) planes in the face- centered cubic crystal structure of copper. The evidence suggested that substrate morphology,