Can lasers modify the threshold voltage of power-off transistors?

[Aghiles]

Hi,
I am a phd student in hardware security,
I want to know what is the effect of the laser on power-off transistors and can it modify the threshold voltage of these transistors? And is there an equation that links the laser and the threshold voltage degradation?

Sincerely

 

[f95toli]

That question is way to broad to answer. What type of laser (wavelength, power) and what type of transistor? Also, are you referring to individual, discrete. transistors? Or transistors i CMOS circuit? And in what kind of package?

That said, the answer is of course in principle yes, if you shine a high powered laser on a transistor it will heat up and the threshold voltage will decrease. That said, you can achieve the same thing using a hot-air gun or a flamethrower so I suspect this is not what you have in mind...

 

[Baluncore]

Welcome to PF.

It will depend on how the transistor is packaged, and the wavelength and power of the laser. The effect will probably be due to thermal migration of the chemistry, with the transistor being progressively aged or degraded at a higher rate than normal due to higher temperature.
https://en.wikipedia.org/wiki/Arrhenius_equation

 

[Delta Prime]

What are these transistors you are speaking of, more information is needed please.

 

[Aghiles]

sorry for the inaccuracies, basically I want to know the effect of the laser (whatever its parameters) on MOSFETs (65 nm or 55 nm), the important thing for me is to know if for example the heat generated by the laser can change the characteristics of the transistor even if it is turned off. That's why I'm looking for an equation that makes the link between the laser parameters and the threshold voltage deviation for unpowered transistors. To know if we attack a transistor with a laser when it is not powered and that we turn it on again afterwards, if its characteristics have changed.
for attacks on the back side, i.e. directly on the silicon
thanks

 

[f95toli]

Just to be clear, you do know that you can't realistically target an individual transistor in a 55nm CMOS circuit, right? The minimum "spot-size" of even UV laser will be ~um in size, and that is in a highly controlled setting.
Essentially, the best you could hope for would be to locally heat a part of the circuit.
Could you apply enough power to permanently change the parameters? Probably (with enough heat you would essentially locally anneal the Si), but the most likely outcome would be that you would simply destroy the circuit.
And again, this assumes you have direct access to the unpackaged circuit which realistically only happens during its manufacture,.

So, if you are thinking about this in terms of it being a cybersecurity threat the answer is no.

 

[Aghiles]

yes it's clear that we can't target a single transistor, but the goal for me is to understand the effect on a single transistor to start with, then I can use the conclusions on circuits like PUF.

Thanks