Adhesion of Alumina Particles to Thin Films

Abstract

An atomic force microscope was used to study the adhesion of alumina particles to copper, SiO2, and tungsten films in aqueous solutions used during integrated circuit manufacture. The solutions included deionized (DI) H2O and aqueous solutions containing dilute H2O2, or NH4OH. When different particles with the same nominal diameter and roughness contacted the same surface, variations in the removal force were observed. Although the particles were similar, the precise contact area and roughness distribution at the point of contact varied from particle to particle, and from location to location on the same substrate, and these effects altered the measured force. When the same particle was brought into contact with different substrates in the same aqueous solution, variations in the removal force were seen. This is due to differences in the roughness, surface chemical composition, and mechanical properties of the different substrates. Alumina particles adhere roughly five times more strongly to copper substrates in DI water than in 0.2 wt % H2O2 in DI water. Alumina particles adhere roughly three times more strongly to SiO2 substrates in 0.2 wt % H2O2 in DI water or 0.2 wt % NH4OH in DI water than in DI water alone. Alumina particles adhere roughly two times as strongly to tungsten surfaces in 0.2 wt % H2O2 in DI water than in DI water alone, and they adhere roughly two times as strongly in 0.2 wt % NH4OH in DI water than in DI water alone. © 2002 The Electrochemical Society. All rights reserved.

Force measurements with the atomic force microscope: Technique, interpretation and applications

“…Lowest adhesion was found in a solution with citric acid, benzotriazole and NH 4 OH which also exhibited the best cleaning performance. Cooper et al [605] measured the adhesion of alumina particles to copper, SiO 2 , and tungsten films in solutions common in semiconductor processing. Adhesion of polystyrene and glass particles on a copper electrode during the electrodeposition of copper was studied by Dedeloudis and Fransaer [606].…”

Hamaker Constants in Integrated Circuit Metalization

“…VDW forces are important to quantify in adhesion studies because they are always present and are always attractive. VDW forces dominate the interaction of particles in contact with thin films (6)(7)(8)(9) and can also be the controlling forces in film adhesion, contributing to adhesive or cohesive film failure (9). Within the equation that describes VDW forces, the Hamaker constant is a key parameter that describes how strongly two materials interact.…”

“…The new method involved bringing the spheres into contact with the surface and measuring the interaction force as they are pulled out of contact. The Hamaker constant is calculated from the force measured as the cantilever is removed from the surface using an adhesion model developed previously (6)(7)(8)(9) which accounts for the effects of surface roughness, particle geometry, and elastic properties of the interacting materials in the measured adhesion force. In this work Hamaker constants were determined for interactions involving copper, silver, titanium nitride, silicon dioxide, parylene-N, and PTFE.…”

“…This model has been validated through measurements of the interaction forces between polystyrene latex (PSL) spheres and silicon dioxide substrates in aqueous solution (6-9). The model has also been validated for asymmetrical alumina particles interacting with substrates of silicon dioxide and copper in both water and air (7)(8)(9). Previously, this model has been used with Hamaker constants from the literature to predict adhesion forces in various systems (9).…”

“…The model has also been validated for asymmetrical alumina particles interacting with substrates of silicon dioxide and copper in both water and air (7)(8)(9). Previously, this model has been used with Hamaker constants from the literature to predict adhesion forces in various systems (9). If the removal force and model parameters are measured, then this model can be used to predict the system Hamaker constant.…”