Effects of bond pad probing for Cu wire bond packages

While copper or palladium coated copper wire bonding in IC packaging have shown tremendous surge recently driven by huge cost saving and competitiveness across OSAT and IDM, there are new challenges when trying to qualify copper wire bond packages especially for fine pitch devices. This paper will study on the probe mark and its effects for copper wire bond integrity and stress test reliability performances. The increased probe touchdowns on bond pad exhibited greater risk resulting to deeper probe depth. Recent studies using gold wire bonding have reported that the limit for probe mark area damage should be limited to maximum of 30% [1] for 45μm ball size whilst other paper suggested maximum of 20% [2] of the probe mark area in order to achieve reliable gold ball bonding. While this paper focused on palladium coated copper (PdCu) wire bonding, the 10% ~ 50% range of probe mark area studied did not exhibit any related wire bonding failures across all legs during assembly. Failure analysis data results i.e intermetallic (IMC), pad crater test and ball bond cross-section for aluminum remains did not show much significant difference. Remarkably stress reliability test followed by destructive test did show some degradation results from unbiased HAST while HTS and temp cycle test did not reveal any failure during ball pull test. Reliability test data analysis concluded that the probe mark area should be limited to max of 30% for 35±2μm ball bond size while probe depth maximum at 80% from the original Al pad thickness or 20% of the Al thickness remains after probe as shown in figure 5 in order to achieve a reliable copper wire bonding process. This paper will share those probe marks criteria for considerations, challenges and recommendations when qualifying new packages for copper wire bonding process.