I have a question about the changing of Gate terminal material In the CMOS VLSI book of David Money Harris

Question about the changing of Gate terminal material

In the CMOS VLSI book of David Money Harris, there is information about the changing of Gate terminal material.
At the begin, the Gate is formed from Metal so the gate stack is called Metal-Oxide-Semiconductor (MOS).
After that, in the 1970s Gate material is changed to polycrystal silicon (polysilicon).
And nowaday, from 2007, it is switched back to metal "solve the meterials problems in advanced manufacturing process".

Could anyone give more specific about this changing? Thank you very much.

Re: Question about the changing of Gate terminal material

First MOS used the (single level of) metal interconnect.
N and/or P channel implants made up the work function
difference.

Poly has a zero work function difference (until doped).
Better when you want both N and P VTs to be low.

Poly is more resistive and prone to effects like poly
depletion (even doped poly has a nonuniform profile,
less dopant at the bottom where you need it, and
the thermal budget needed to fully activate (let
alone drive to uniform distribution) is higher than
CMOS gate ox really wants to see.

What other "advanced" issues it might solve (and
what that solution costs otherwise) I do not know;
I spend my time at the trailing edge doing stuff
that's "too hard" for other reasons than achievable
transistor count. But I have worked on some of the
oldest metal gate CMOS back when it was still in
production.