Investigation of Cu-Al Bonding Interface: Eliminating Bimetallic Corrosion Failures, and Enabling Next-Gen Cu-Cu Wire-Bonding by Nanometer Interfacial Chemistry Control

Abstract

The first part of this dissertation explores the chemistry of an inhibitor complexation with Cu. First, the Cu oxidation state of the complex was +1. Second, identified by differential RAIRS, one source of Cu(I) for the Cu(I)-inhibitor complex could be Cu(I) oxide. The characteristic Cu(I) oxide peak at 650 cm⁻¹ was observed to decrease after CVD coating process was applied. This led to a major hypothesis that in order for the reaction between Cu(I) oxide and the inhibitor to proceed, protons from the inhibitor and oxygen from Cu₂O are stabilized by reacting to form water. The applicability of the passivation nature of Cu(I)-inhibitor films was explored for Cu-Al wire-bonded devices in its ability to protect from Cu-Al peripheral galvanic corrosion and the galvanic corrosion of the Cu-Al intermetallic compounds in their roles for corrosion-induced liftoff. The second part of this work studied the effect of replacing Al bond pad with Cu on the corrosion induced liftoff of wire-bonds when exposed to low ppm levels of chloride contamination. Applying protective coating to the Cu pad surface before wire-bonding was found to suppress the thermally induced oxidation of Cu in air, helping to enable successful Cu-Cu direct wire-bonding. Compared to Cu-Al devices with passivation coating, which has a few wires liftoff with 6 hours, the Cu-Cu bonded devices survived much longer, over 40 days, with almost no liftoff observed. This demonstrates that removing the galvanic contact, the root cause of the corrosion induced failure, is a more robust and permanent solution to the corrosion experienced by these devices.