Measurement of Polishing Rate as a Function of Pad Independent Abrasive Friction for Chemical Mechanical Polishing

Optimization of Chemical Mechanical Planarization (CMP) on integrated circuits requires an accurate model of material removal. The current model of polishing is two-body abrasion of particles trapped at the pad asperity-substrate junction, however, this is not supported by friction measurements. A modified friction model is applied to data collected from a pin-on-disk experiment preformed on a fused silica wafer with various polymer polishing pads, flooded with a silica nanoparticle slurry. This model indicates that the friction arises predominantly in the swept region preceding the asperity contact. We will add to this model the measurements of the volume of the wear-track, to identify the rate of material removal. Our aim is to develop a Preston-style relationship between friction and the material removal rate that is independent of the pad material. Preliminary data suggests that this pad independent wear coefficient is two orders of magnitude greater than that from a pad dependent example. The long-term goal of this investigation is to aid in the optimization of CMP for structures of decreasing size.