Abstract
A phenomenological model has been developed to simulate the feature profile evolution for nanometer-scale control of the profile and critical dimension during plasma etching. Attention was focused on the feature profile evolution of infinitely long trenches etched in Si with chlorine chemistries. The model takes into account the transport of ions and neutrals in microstructures, multilayer surface reactions through ion-enhanced etching, and the resulting feature profile evolution, where the transport is analyzed by a two-dimensional particle simulation based on successively injected single-particle trajectories with three velocity components. To incorporate an atomistic picture into the model, the substrates are taken to consist of a large number of small cells or lattices in the entire computational domain of interest, and the evolving interfaces are modeled by using the cell removal method; the Si atoms are allocated in the respective two-dimensional square lattices of atomic scale. Moreover, the Monte Carlo calculation is employed for the trajectory of incident Cl+ ions that penetrate into substrates. The present model has a prominent feature to phenomenologically simulate the multilayer surface reaction, the surface roughness, and also the feature profile evolution during etching. The etching of planar Si substrates was simulated for a test of validity of the present model, showing the structure of surface reaction layers, the distribution of Cl atoms therein, and the surface roughness that depend on incident neutral-to-ion flux ratio and ion energy. The etch yield as a function of neutral-to-ion flux ratio for different ion energies gave a similar tendency to the known experimental data, indicating that the present model properly reflects synergistic effects between neutral reactants and energetic ions in the ion-enhanced etching. The feature profile evolution during etching was then simulated for sub-100 nm line-and-space patterns of Si, exhibiting the reactive ion etching (RIE) lag that occurs depending on neutral-to-ion flux ratio and ion energy. The degree of RIE lag was found to be more significant at higher flux ratios and higher energies, being associated with the difference in surface chlorination at the feature bottom; in effect, for narrow pattern features of the order of sub-100 nm, the bottom surfaces tend to starve for neutral reactants owing to severe effects of the geometrical shadowing.