| The International Technology Roadmap for Semiconductors projects that for the 32nm technology node, the drain extension junction depth, sheet resistance, and lateral abruptness will be in the range 5–9nm, 940Δ/□, and 1.4nm/decade respectively. The current approach to scaling the source/drain extension profiles has been to reduce the ion implantation energy. However, as the implant energy is reduced and the implant dose is increased, the peak dopant and damage concentrations are increased, leading to dopant clustering or precipitation, which in turn leads to solubility levels that severely limit the electrical activation that can be achieved. This dopant clustering can also contribute to the anomalous diffusion of dopants.; Thus, in order to successfully continue the scaling trend, we need to gain a better understanding of dopant clustering dynamics, and in particular the activation mechanisms. Once the activation mechanisms are better understood, we can pursue techniques to reduce junction depth and at the same time obtain high levels of dopant activation.; In this work, we have extensively studied the electrical activation of boron, phosphorus, and arsenic implanted into silicon. In particular, we have experimentally investigated how the activation mechanism of boron is different than arsenic and phosphorus, and what implications this difference has from a process technology point of view. Furthermore, modeling needs for simulation of dopant activation and diffusion have been explored, and the dynamics of boron clustering have been investigated with help of Kinetic Monte Carlo based atomistic simulations.; Moreover, we have analyzed the electrical activation behavior of boron in pre-amorphized samples, the interaction between fluorine and boron, and the role of annealing ramp-rates on boron and arsenic electrical activation and diffusion. Finally, by utilizing flash-assisted RTA, we have demonstrated how the understanding of dopant activation can be utilized to form ultra-shallow, highly active junctions. |
