Control of Phosphorus Transient Enhanced Diffusion using Co-implantation

Abstract

The production of Ultra Shallow Junctions (USJ) in silicon devices requires controlling the Transient Enhanced Diffusion (TED) of electrical dopants. USJ development has focused on boron because hole mobility is lower than electron mobility in silicon and because arsenic has such excellent diffusion properties. However, the advent of strain enhanced mobility in P‐type silicon has created the need to study higher solubility N‐type dopants like phosphorus and find methods to control their diffusion. Co‐implants have proven effective in controlling the interstitial diffusion mechanisms of boron TED. In this work the effectiveness of some co‐implants on phosphorus to form high performance USJ is reported. It has been found that carbon and fluorine co‐implants reduce phosphorus diffusion. As work with boron has shown, this is due to the carbon Kick‐out mechanism and Fluorine‐Vacancy clusters, both of which consume the interstitials driving TED. It has also been found that record levels of phosphorus diffusion control can be obtained if boron and carbon are co‐implanted. In this junction diffusion control increases as the boron implant energy decreases; even as low as 0.5 KeV. However, this may be activating Uphill diffusion. The data also shows that the carbon implant energy has very little effect on phosphorus diffusion. The boron and carbon co‐implants also produce the steepest phosphorus USJ yet reported at 2.5nm/decade with a solubility >1.0E21 atoms/cm3. Counter intuitively it has been found that the boron and carbon USJ is shallower with a higher solubility if the phosphorus implant energy is increased from 2 to 3 KeV. These boron and carbon co‐implant findings are quite novel even if they are not technologically useful. They strongly support the widely held model that phosphorus TED occurs via an interstitial diffusion mechanism and that techniques to block this mechanism can control it. The boron implanted below the phosphorus is probably consuming interstitials very efficiently in Boron Interstitialcy clusters (BI+) causing boron TED rather than phosphorus TED. The electrical characteristics of a phosphorus USJ with high doses of boron below it may be undesirable; but, it does demonstrate that phosphorus diffusion can be well controlled if the right co‐implants are found.