AVS 50th International Symposium
    Plasma Science and Technology Monday Sessions
       Session PS-MoM

Paper PS-MoM7
On the Roughness of Etched Silicon

Monday, November 3, 2003, 10:20 am, Room 315

Session: Critical Dimension Etching
Presenter: A.A.E. Stevens, Eindhoven University of Technology, The Netherlands
Authors: A.A.E. Stevens, Eindhoven University of Technology, The Netherlands
H.C.W. Beijerinck, Eindhoven University of Technology, The Netherlands
Correspondent: Click to Email
The smaller the etched features, the more important the roughness of the etched feature surface becomes. Not only for integrated circuits, but also for, e.g., photonic crystals, the surface roughness might limit the device quality. By using in situ (spectroscopic) ellipsometry and mass spectrometry, the effect of Ar@super +@ ions and XeF@sub 2@ etch precursor on the etch mechanism has been studied, simultaneously collecting information regarding the roughness evolution. Analysis of the XeF@sub 2@ reactivity and SiF@sub x@ products in contrast to the reaction layer composition shows that the reaction layer thickness, i.e. the surface fluorine content, scales with the roughness. This implies, that the SiF@sub x@ layer build up suggested in literature can be explained by the increasing roughness of the surface, thus the effective surface area increase of the etched Si samples. Furthermore, the etched Si samples have been analyzed with an AFM for comparison with the in situ ellipsometry results. The same trends in roughness evolution have been observed for the ellipsometry measurements and the AFM measurements, corroborating the used ellipsometry models. To learn about the role of the etch precursor and ions in the roughening of the surface during etching, the Family-Vicsek scaling theory of the surface roughness has been applied. The scaling coëfficients @alpha@ and @beta@, representing lateral correlations on the surface and time dependent roughness evolution, respectively, have been derived from the AFM data analysis. Since specific @alpha@ and @beta@ values are related to the surface processes occurring during the etching, the application of the scaling theory aids in understanding the role of the ions and etch precursor in the roughening process.