Non-Destructive Top Layer Bond Pad Cross Section

Abstract

Contaminated and corroded aluminum bond pads can lead to a number of reliability failures. Poor wirebondability, false failures at electrical probe due to high contact resistance, opens on RDL layers due to incomplete sputter, and various early life failures during JEDEC testing can all be attributed to excess oxidation and corrosion on aluminum bond pads. Understanding what layers exist on the bond pad surface is critical. Oxides, hydrates, xyfluorides and various soups of materials can be confirmed from combinations of Auger, SIMMS and FTIR but there is a chance the die can be damaged and the spot sizes can be problematic picking up information outside the bond pad target. Layer analysis of the pad can also be difficult since sputtering rates vary based on the contaminants on the pad and any thermal processing. How does one evaluate bond pads without a destructive test? Is there a process for accurate bond pad evaluation while minimizing die damage? Sure. The cycle time and costs of these tests can also be a concern particularly for a customer wanting a quick response from failure analysis. A better method is to bond the die with a gold stud bump followed by selectively etching away the gold. Etching the gold highlights the intermetallics and shows how much of the gold alloyed with aluminum versus blocked oxides. Voiding can be demonstrated, that might be misconstrued as Kirkendall voiding, by cross sectional analysis.