NSF Award
9960600
This Small Business Innovation Research Phase I project will develop a statistical (Monte Carlo) software model and software simulation tool for the pre-sheath and sheath regions of low pressure plasmas used in IC fabrication. The objective is to bridge the time/length scales between reactor scale phenomena and feature scale phenomena. The sheath models will be integrated with an existing reactor scale software model (CFD-ACE+) and feature scale software simulators (SPEEDIE and CATS). The work will fill a void currently faced by designers of plasma equipment and processes. It will evaluate the influence of macroscopic reactor conditions on feature scale profile evolution. Stanford University Center for Integrated Circuits will be a sub-contractor on this project. The Phase I effort will focus on an intermediate-scale model based on kinetic treatment of charged particle transport near the wafer surface. This model will be an interface between a hydrodynamic model in CFD-ACE+ and collisionless gas phase models in SPEEDIE and CATS for interstructure particle transport. The plasma-presheath model will provide spatially resolved distributions of ion flux, energy and angular distributions to SPEEDIE and CATS. Both SPEEDIE and CATS require die level models to resolve the impact of circuit layout/topography on species generation/loss, charged particle collection and currents paths to the substrate.<br/><br/>In Phase II, the models will be refined and validated (against experiments conducted at Stanford) for silicon etch processes in Cl2 and SF6 systems. CATS will be expanded to include the wafer charging circuit and will be merged with SPEEDIE. The commercial availability of the capability will allow process engineers to design better processes and identify equipment/process deficiencies before physical prototyping. The use of the model will enable the reduction of potential yield losses due to unsatisfactory gap/step coverage, film noncomformality and undesirable etch profiles. According to industry observers, even a 2% improvement in the fabrication yield will provide significant savings to the industry.
