Sunday, 13 October 2013

Q1: Polysilicon VS Metal contact.
Ans:There are a few reasons that poly was the preferred material for the past decade or so; metal gates were used when operating voltages were 3-5V. As operating voltages lowered, manufacturers transitioned to using polysilicon as the gate material. Recently, after 45nm (for Intel) and 28nm (for TSMC), gates are again made with metal in conjunction with high-k insulators.

One reason for the initial switch to polysilicon is that fabrication processes after the initial doping required very high temperature annealing. Metal gates would melt under such conditions whereas polysilicon would not. Using polysilicon allowed for a one-step process of etching the gates compared to elaborate multi-steps that we see today in metal-gate processes.

The other reasons is that the threshold voltage of the MOSFET inversion layer is correlated with the work-function difference between the gate and the channel. Using metal would result in a higher Vt compared to polysilicon since a polysilicon gate would be of the same or similar material composition as the bulk silicon channel.

As we reach smaller and smaller scales, the need for a higher Vt has become important again due to problems of leakage. Higher conductivity in the gate has also become important as the oxide dielectric layers cannot be shrunk any further to increase speed. Thus, metal gates with a high-k dielectric are used in modern CMOS transistors.