Use of Wire Bonding to Study Bond Pad Damage from Wafer Probe

Abstract

Harsh wafer probing and harsh wire bonding have been used in previous work by our group to study cracking and ripple effect on various bond pad designs in technologies having aluminum (Al) alloy based metallization in SiO2 dielectric. One outcome of the previous experimentation is the confirmation that cracks due to wafer probe may not be visible even in a careful cratering test. Probing stress can bend the top SiO2 of the pad downwards, causing high tensile stress to the underside of the film. Cracks initiating in this region easily go undetected in routine monitor procedures, unless they propagate to the top surface and break the barrier film above. Latent probing cracks can result in reliability issues for circuit under pad (CUP) bond pad designs. A typical probe crack in a CUP pad's top SiO2 layer, detected in a cratering test, may actually be accompanied by one or more hidden cracks at the bottom of the film, which may only be detected in a focused ion beam (FIB) cross section. This study uses harsh probing to create cracks in various bond pad test structures and then follows them through the wirebond process, including harsh bonding experiments. Previously undetected probe cracks may be enhanced, lengthened, and made easier to detect in wirebond monitoring and in a cratering test after wirebond. Cracks due to wirebond have different characteristics than probe cracks. Distinguishing features of both crack types are compared after bonding. Monitoring of the ripple effect in both probe and bond helps to predict and track bond pad cracking tendencies. Methods to reduce cracking from both probe and bond are reviewed.