Optimization of nanoscale liquid-assisted laser cleaning and elucidation of lower size limit of particle removal

Abstract

Since the development of the liquid-assisted laser cleaning method known as "steam laser cleaning (SLC)" several decades ago, its application in semiconductor manufacturing has been limited due to its wet process nature and potential damage caused by the electromagnetic field enhancement underneath targeted particles. However, as the primary target size has now been shifted towards the sub-10 nm range, the potential of SLC needs to be reassessed in the unexplored domain. Because previous studies have primarily been focused on particles over ~100 nm, the process has hardly been optimized for smaller particles. Moreover, the damage concern caused by relatively large dielectric particles has diminished, as the field amplification beneath transparent particles is weak in the case of nanoscale particles. Consequently, this study aims to evaluate the potential of the SLC technology in the size range of less than tens of nanometers. Specifically, the smallest particle size that can be removed from a silicon surface is identified with optimized SLC processes using a nanosecond pulsed laser. The investigation covers three distinct particle types: gold, polystyrene (PS), and alumina particles. After optimizing the process through numerical simulation and experiment, the damage threshold and the particle removal efficiency (PRE) are measured by varying the laser fluence and particle size. The results indicate that the lower size limit, as determined by a 90 % PRE criterion, is 3 nm for gold particles, 20 nm for PS particles, and 50 nm for alumina particles. Discussions are made on the mechanisms responsible for the relatively high cleaning performance observed for gold particles.