Probing particle removal in brush scrubber cleaning with fluorescence technique

Brush scrubber cleaning is widely used for post chemical mechanical polishing (CMP) cleaning in semiconductor manufacturing. In this study, an experimental system based on fluorescence technique and particle-tracking velocimetry (PTV) technique was employed to characterize the particle removal displacement and velocity in the interface between a transparent copper film and a porous polyvinyl alcohol (PVA) brush during the cleaning process. Several different cleaning conditions including rotation speeds, loading pressure and cleaning agent were examined and the particle removal rate was compared. Elastic and friction removal, hydrodynamic removal and mixed-type removal are the three types of particle removal. Particles with an arc trace and uniform velocity curves were removed by friction and elastic force which were related to the brush load. Particles with a random trace and fluctuant velocity curves were removed by hydrodynamic force which was determined by the brush rotation speed. The increase of particle removal rate (PRR) with brush rotation speed is a logistic function. It is easier to improve PRR by increasing the brush load or by adding surfactant than by increasing the brush rotation speed.