Pressure Builds on Gate First High-k

[Martimus]

Pretty cool, the hybrid approach, removing just the PMOS gates as a gate-last effort.

I never did understand why the hybrid method wasn't as effective as the gate last only method, since you could have specific metal types for each the NMOS and PNOS gates. There must be another advantage to gate last than just picking the metal type that I am not aware of.

 

[Idontcare]

I never did understand why the hybrid method wasn't as effective as the gate last only method, since you could have specific metal types for each the NMOS and PNOS gates. There must be another advantage to gate last than just picking the metal type that I am not aware of.

One of the MAJOR benefits of gate-last is that you can select and implement HK and/or MG material compositions that would be unusable because of the heat cycles (referred to as the "thermal budget") when doing dopant activation/anneals in the front-end.

I'm trying to make this sound as uncomplicated as possible and am probably failing with aplomb.

There are HK dielectrics as well as metal-gate materials that cannot withstand the temperatures used/needed to form the S/D and channel regions in the xtor. Having an integration scheme that lets you build the HK and MG after the fact is beneficial in that it removes this restriction.

Am I making any sense?

 

[Martimus]

One of the MAJOR benefits of gate-last is that you can select and implement HK and/or MG material compositions that would be unusable because of the heat cycles (referred to as the "thermal budget") when doing dopant activation/anneals in the front-end.

I'm trying to make this sound as uncomplicated as possible and am probably failing with aplomb.

There are HK dielectrics as well as metal-gate materials that cannot withstand the temperatures used/needed to form the S/D and channel regions in the xtor. Having an integration scheme that lets you build the HK and MG after the fact is beneficial in that it removes this restriction.

Am I making any sense?

That does make sense. Thank you. You are saying that there are materials that could be used if inserted after the doping process that could not survive the doping process. And those materials may be better for a particular gate than those that would survive the process.