Copper ball bond recipe cliff test on two pad aluminum thicknesses

Abstract

“Cliff test” studies with thermosonic wirebond of 25μm copper (Cu) wire were done on integrated circuit pads having either 0.8μm or 3.0μm aluminum (Al) thickness to investigate the effect on intermetallic (IMC) formation and shear test results. One half of the pads had been probed, leaving relatively large and invasive probe marks, while the other pads were unprobed prior to wirebond. Manufacturing recipes already in use for Cu bonding on each pad thickness were designated as the “nominal” conditions. The so-called “USG cliff test” investigates wirebond results when changing the ultrasonic generator (USG) current away from the nominal setting, in increments of 10% difference, until a non-stick-on-pad (NSOP) condition occurs or until the ball has become so flat that a shear test cannot be performed. Experimental USG settings ranged from −40% up to +120% from nominal. NSOP occurred at −40% USG current. In separate tests, a similar “Force cliff test” changed the bond force away from nominal, in increments of 10% difference. NSOP did not occur when varying the force between −80% and +150% from nominal. Probe mark effects on Cu wirebond shear test results are evaluated for nominal Cu wirebonding conditions on both Al thicknesses.