We present a new wet-cleaning method to remove strongly adhered 90-nm, Si3N4 contaminant test particles from a smooth silica surface. The proposed cleaning solution consists of a foamed aqueous dispersion of solids. Contaminant removal increases with solids concentration, foam quality, process time, and shear rate and can approach 100 %.

Particle-removal experiments in a simple dip cell indicate that dispersed solids are necessary for contaminant removal. Exponential decline of surface concentration of adhered particles with number of immersion/withdrawal events and a linear increase in the rate of particle removal with solids concentration indicate that the removal mechanism is binary collision between Si3N4 contaminant particles and the solids. Foam or surfactant solution or aqueous solvent alone are not effective in particle removal.

A binary-collision rate model is derived to predict particle detachment. Comparison of the rate model with experimental removal data is good. Three adjustable parameters arise when no foam is employed: two first-order rate constants and the fraction of particles with each rate constant. Prediction of particle removal with a foamed-solids dispersion requires an additional influence parameter. The new binary-collision rate law provides a powerful design tool to assess the role of cleaning parameters including process time, shear rate, dispensing flow rate, system geometry, solids concentration, and foam quality.