Statistical parametric study of non-parallel inductive reactors
- Univ. of Illinois, Urbana, IL (United States). Dept. of Electrical and Computer Engineering
Parameterization of new plasma processing is a very time consuming process. By combining statistical techniques with computer simulation, one can construct a numerical design of experiment (NDOE), which minimizes the time required for the investigation. Here the authors present the results of such an NDOE, applied to a generic non-parallel inductively coupled reactor, as there has recently been interest in inductive reactors with a dome shaped quartz top. They use a statistical design of experiment software too, Echip{copyright}, to construct the experiment and then perform the experiment with the Hybrid Plasma Equipment Model (HPEM). Echip{copyright} is also used in the post-experiment analysis. The goal is to determine functional relationships between ion and neutral flux uniformity to the wafer and geometrical parameters such as focus ring height, reactor height, and reactor radius. Additionally, they consider several non-geometrical variables such as substrate bias and pressure. By combining numerical modeling with statistics software they have been able to determine optimal parameters sets for different gas systems. Of the geometric parameters, focus ring height appears to have the strongest correlation with the ion and neutral flux uniformity. Correlation becomes worse at higher pressures for most parameters.
- Sponsoring Organization:
- National Science Foundation, Washington, DC (United States)
- OSTI ID:
- 323618
- Report Number(s):
- CONF-970559--
- Country of Publication:
- United States
- Language:
- English
Similar Records
Optimization of an ionized metal physical vapor deposition reactor
Modeling of etch profile evolution including wafer charging effects using self consistent ion fluxes