Plasma etch is currently used for virtually all critical pattern transfer steps in semiconductor device manufacturing. The advantages of plasma-etch subtractive processing to form patterns are considerable, not the least of which is the existence of decades of expertise and knowledge of plasma etch tool design and control. However, the relentless shrinking of device critical dimensions (CD) and the associated stringent demands on reducing CD variability prompts the question of ultimate etch limits: how small a feature can be created by plasma etch? Not only is etch itself becoming more challenging, but the etch step is now used to minimize lithographic limitations. Further, novel etch masks at nanometer length scales are emerging, including di-block and tri-block co-polymer self assembled structures and various nano-imprint methods. The limitations and characteristics of plasma etch with these masks has hardly begun to be explored and much remains to be learned about controlling nano-structural features during etching with these novel masks.

Our strategy for studying the science underlying nano-scale plasma etch is to develop and apply atomistic simulation methods. In this talk, I will present recent results using molecular dynamics simulations of etching features (holes and trenches) on the order of several nanometers, both with and without explicit masks. One aspect that is obvious even from the simplest simulations is the role of ion bombardment in creating sidewall 'passivation' layers, even without depositing neutral species. Other important questions that challenge current MD simulations of feature shape evolution include etch mask treatment, slow neutral transport and reaction dynamics within the feature, and feature charging and ion deflection.