The Reliability of Soldered or Epoxy Bonded Chip Capacitor Interconnections on Hybrids

Abstract

The influence of capacitor chip-to-substrate assembly variables on interconnect reliability was evaluated using several conventional environmental tests. The variables investigated included: capacitor chip size, capacitor end metallization, substrate pad size, substrate metallization, different solders, and a nonconduetive and a conductive epoxy. Environmental tests consisted of mechanical shock, temperature cycling, and high temperature storage. The test specimens included three capacitor sizes: 0.050 X 0.080 in, 0.080 X 0.180 in, and 0.080 X 0.270 in, all 0.050 in high; and three end terminations: silver, palladium-silver, and an Sn60-Pb36-Ag4 solder coating. Substrate pad sizes were varied to represent large, medium, and small bonding areas. Capacitors having all three end terminations were reflow soldered on Sn10-Pb88-Ag2 or Sn96-Ag4 tinned pads, by reflowing Sn62-Pb-36-2Ag paste, or by Silver conductive epoxy bonding with Ablefilm 606-2 (commonly used for bonding active devices). Also, capacitors with palladium-silver end terminations were bonded with nonconductive epoxy, while the electrical interc0nnection from substrate to chip was made by thermocompression bonded gold wire. Two types of substrate metallization were used: a) thin film nichrome, nickel, gold;b) thick film platinum-gold. Testing consisted of the sequential exposure:· i) to 3000-G mechanical shock (5 pulses), ii) temperature cycling from -65 to 150°C (100 cycles), iii) 3000-G mechanical shock (5 pulses), and iv) high temperature exposure at 150°°C for one week (168 h). All chips were visually and electrically checked after each test. Capacitor attachment was more reliable when small chips were bonded to thick f'dm metallization and when the bonding pads were large. Snl0 solder or nonconductive epoxy were the most reliable inter, connect materials. When soldering, the best electrical and mechanical end terminations were palladium-silver; the appearance of silver end terminations changed during environmental testing, but solder coated or palladium-silver terminations were visually unaffected. When conductive epoxy bonding, no differences were noted between silver or palladium-silver terminations. Visual examination was of minimal value in detecting marginal chip attachments. The most effective environmental test combination was thermal cycling followed by mechanical shock.