Reverse Vbe exceeded, long term gegradation

Hot-carrier light emission is super fast, and is used for diagnostics--search for PICA. It was invented by a couple of my old colleagues.

Cheers

Phil Hobbs

Ultra violet light is emitted by ordinary MOSFET transistors when in saturation. Impact ionization makes the UV emission from silicon. Those are hot carriers from accelerated electrons hitting solid silicon. Surface areas can accumulate damage that makes Vt changes in pmos.

Bulk pnp and npn silicon bipolar devices have no vulnerable surface areas for damage to affect anything. Bulk silicon has advantages over poly base, over Ge-Si heterojunction bipolar transistors. They are more reliable in this regard.

when in saturation. Impact ionization makes the UV emission from silicon"

Hold on here, does that mean that they actually emit UV, or is it reflected back to the die by the case ? Most transistor cases look UV proof but I do not make assumptions. Notnymore. Of course the metal ones are, but when wa s the last time you used one of them ?

ed back to the die by the case ? Most transistor cases look UV proof but I do not make assumptions. Notnymore. Of course the metal ones are, but when was the last time you used one of them ?

Yes, UV is emitted by silicon. I was also surprised when I worked at AMD, Advanced Micro Devices Inc. in Sunnyvale, California, USA. An internal conference at AMD in 1993 has papers written by AMD employees for AMD employees to hear the words and see the pictures on a rectangular screen. The audience was treated to a surprising fact: MOSFET silicon transistors emit UV light !

I also was surprised. Do you want a link to that amazing fact?

ed back to the die by the case ? Most transistor cases look UV proof but I do not make assumptions. Notnymore. Of course the metal ones are, but when was the last time you used one of them ?

Here is the documentation of UV from MOSFET :

Quote, "Photon emission spectra of the grounded-gate nMOSFET biased into th e snapback region showed that the spectra are a combination of the spectra for a forward-biased and reverse-biased p-n junction. A very weak emission at a wavelength of 375 nm (3.3 eV) was also detected, and this was suspecte d to be due to the injection of hot electrons into the gate oxide. "