What is N-diffusion in NMOS?

Question

A NMOS contains source, drain, and gate. Under the source and drain contains N+ implantations and under the gate is a polysilicon layer with the P substrate beneath it:

But I heard from a VLSI workshop that: "... When the Polysilicon crosses the N-diffusion, then an NMOS is created...".

First is what is "N-diffusion"? And second, AFAIK underneath the polysilicon is the P substrate but from what I understood from the statement above it seems that under the polysilicon is not the P substrate but an "N-diffusion"?

Answer 1

In older manufacturing processes the source and drain were diffused into the silicon rather than implanted. Wherever you read "diffusion" in older literature, think "implant".

When a transistor is drawn the top view looks like there is a single large N implant region with a stripe of polysilicon running down the middle. The polysilicon (just called "poly" sometimes) is created before the implant is done so it acts as a mask for the implant and effectively creates two separate regions: the source and the drain. So, looking at the drawing of an NMOS, where the poly crosses the N implant you will have the P body remaining underneath, thereby creating the channel region.

EDIT: See the images in Multiple fingers vs single finger layout (MOSFET transistor) for examples.