Ways to minimize scattering as T increases

[ZeroFunGame]

TL;DR

as temperature rises in a MOSFET, are there tricks to prevent the Coulomb/phonon scattering? Or at least minimize it?

as temperature rises in a MOSFET, are there tricks to prevent the Coulomb/phonon scattering? Or at least minimize it?

This is in reference to Fig 4 here:
https://pdfs.semanticscholar.org/f8a0/b9b7030f7201ef17c6ff66ec660fd75c7aae.pdf

The more overarching question is related to ring oscillators and circuits in general, and whether there is a way to minimize frequency degradation as temperature is increased.

 

[Fred Wright]

Dip the circuit in liquid nitrogen.

 

[ZeroFunGame]

Added the caveat that you cannot cool the device :)

I was thinking semiconductor material engineering, or circuit compensation, etc. Not sure if these actually exist as solutions

 

[ZeroFunGame]

Added the caveat that you cannot cool the device :)

I was thinking semiconductor material engineering, or circuit compensation, etc. Not sure if these actually exist as solutions

Improve surface state density via improved dielectrics, reduce phonon scattering via strain and isotropic doping, reduce surface scattering by reducing surface roughness. There's a few more, but I guess standard channel engineering tricks to boost mobility is the way to go

 

[Dr_Nate]

Improve surface state density via improved dielectrics

Could you explain what you mean by this quote above?

What is isotropic doping?

Could you also talk about your limitations? For example, must you use silicon?

 

[ZeroFunGame]

Could you explain what you mean by this quote above?

What is isotropic doping?

Could you also talk about your limitations? For example, must you use silicon?

My understanding is that you want to reduce interface states and charge traps, as this could lead to greater scattering at the dielectric semiconductor interface, thus reducing mobility (please correct me if I am mistaken)

Isotropic doping, I assume means uniform doping as a function of depth, but my understanding here could be mistaken.

Limitations? No, you don't have you use silicon. For example, why not use GaN since they have a wide bandgap. But then the limitations wouldn't be interface states or doping since it's a HEMT structure. Diamond or GaO or SiC could be another option. I'm not sure what you mean by limitations -- for Si, it's the dominance of intrinsic carriers as T increases