NIST researcher's findings influence industry acceptance of physical model for breakdown. (News Briefs).

A NIST researcher has played an important role in influencing the semiconductor industry to refute the previously accepted thermochemical model (or "E-model") of dielectric breakdown of reliability of ultrathin silicon dioxide.

The correct physical mechanism for dielectric breakdown has been controversial for over three decades. Until recently, there were three models that give vastly different projections for oxide life: the thermochemical model, the anode hole injection model, and the hydrogen release model. It is crucial that the correct physical model be used when making reliability projections; otherwise large errors in lifetime extrapolation will result.

The NIST researcher's work on substrate-hot electron injection showed that tunneling electrons were necessary for thin oxides to break down. This is significant because the thermochemical model does not require tunneling electrons to damage the oxide (only the applied gate field is required). Therefore, the thermochemical model can not be correct. This research is recognized as one of the most convincing demonstrations of the involvement of tunneling electrons in causing breakdown.

The NIST researcher recently created holes in thin oxides to show that holes do not efficiently cause defects that would eventually lead to dielectric breakdown. These results are significant in that they strongly suggest that the anode hole injection model either is incorrect or must be significantly modified as a physical model responsible for dielectric breakdown in ultrathin gate dielectrics.

CONTACT: David Blackburn, [phone]; david. [e-mail].