Comparison of deal grove model growth rate with dry thermal oxidation process for ultra-thin silicon dioxide film

Abstract

The capability to grow the manufacturing grade ultra-thin dielectrics hold the key to the continuous miniaturization of semiconductor devices. While the thermally grown SiO 2 has been used as a gate dielectric ever since the decades of silicon device began, it appears that the electrical and physical properties of pure SiO 2 are not good enough to provide acceptable performance for ultra-thin gate dielectric film. There are many available methods to control the growth of the ultra-thin film. In this paper, we control the growth rate of dry thermal oxidation by incorporating nitrogen gas during the process. The Deal Grove model is universally accepted as the physical model to describe the thermal oxidation process. However, at the initial oxidation stage, this model cannot predict the process satisfactorily. The incorporated nitrogen will neutralize the growth sites at the oxide-silicon interface, which significantly slows down the oxidation process when N 2 gas is used as an oxidizing ambient. It would also affect the linear rate constant for the Deal Grove equation where, linear rate (B/A) is the reaction at the Si/SiO 2 regent that depend on the oxygen and nitrogen gases. The result was compared with the calculated growth rate, which based on the Deal-Grove model to investigate their correlation. Where for the result, it shows that the linear rate constant (B/A) of deal grove model for the dry oxidation of thickness <25nm, the deal grove model are not accurate, linear rate (B/A) response we're becoming less this is because the deal grove model can't predict the initial stage of the dry oxidation growth.