AVS 47th International Symposium
    Dielectrics Thursday Sessions
       Session DI+EL+MS-ThM

Paper DI+EL+MS-ThM9
Investigation of Fluorine in Dry Ultrathin Silicon Oxides

Thursday, October 5, 2000, 11:00 am, Room 312

Session: Ultrathin Dielectrics and Interfaces
Presenter: G. Vereecke, IMEC, Belgium
Authors: G. Vereecke, IMEC, Belgium
E. Röhr, IMEC, Belgium
R.J. Carter, IMEC, Belgium
T. Conard, IMEC, Belgium
H. Dewitte, IMEC, Belgium
M.M. Heyns, IMEC, Belgium
Correspondent: Click to Email
As critical dimensions of integrated circuits continue to decrease, insulators with dielectric constants higher than silicon dioxide will be introduced in capacitors and transistors. However an ultrathin (< 1 nm) silicon oxide layer will generally be needed at the interface between Si and high-k layers. In cluster tools integrating surface preparation and dielectric deposition, vapor HF chemistries are envisaged to etch the native oxide prior to oxide growth. This HF etch leaves F on the surface of the Si wafer, which gets incorporated into the growing oxide. In addition, the silicon subsurface has also been proposed as a source of the F found in these oxides. The presence of F may be beneficial or detrimental for the properties of these layers depending on application, F location, and layer thickness. We have evaluated the sources of F in ultrathin oxides grown by UV/O2 at room temperature in an experimental vapor phase cleaning tool. Surface pretreatment was either by in-situ HF/methanol vapor process or by ex-situ wet HF dip followed by a DIW rinse. Evidence was found for F cross-contamination from the tool gaspanel when the HF etch step was performed in-situ. After correcting for this, F atomic concentrations in oxides grown on vapour HF and wet HF treated surfaces were of about 5 % and 2 %, respectively. The former would lower the dielectric constant of the layer if homogeneously distributed. The level of F contributed by the subsurface was estimated with oxides grown on wet HF treated surfaces in a specially built quartz chamber with no F contamination. No F was detected in these oxides, which indicates that the level of subsurface F is lower than previously reported. Ultrathin oxides continued to grow when exposed to air. This raises concern about their stability during the deposition and annealing of high-k layers. XPS results suggest that this is related to the exchange of labile F groups in the films.