so we encountered a really strange thing in the lab yesterday
we've brought up a few times to our coworkers that the lab's fluorescent lighting is going to be a problem when we get to doing tests on bare dice
well yesterday we were only working with packaged parts, and it was a problem anyway!
to explain why, let me go into a little detail on electronics:
-F
on semiconductors and light, long
All semiconductors are photosensitive, and any light of energy equal or greater than their bandgap will generate free charge carriers--valence-band holes and conduction-band electrons. This is how photodiodes, solar cells, and even LDRs/photocells work (and how LEDs work in reverse).
But when this is undesired, it can cause problems. This is one of several reasons most electronics are packaged in opaque plastic housings and only the metal leads stick out, no exposed semiconductor. But plastic doesn't work for everything: metal is used for things that need to conduct heat out of them (and old parts that predate the development of plastics that are suitable for this purpose, and a few specialty parts that need to be metal for other reasons), and ceramics are used for things that need to work at high temperatures or withstand higher voltages or just have very repeatable electrical characteristics (too much manufacturing variance in plastic)
Metal and plastic cases, unless damaged or intentionally designed that way, are completely 100% opaque, at least as far as can be measured
most ceramic cases also are, but *this specific device* that we're dealing with is in a hybrid metal-ceramic case, and apparently the alumina ceramic used for it (chosen for its electrical characteristics for reasons we don't fully understand--that's not our area) is at least a little bit transparent to some frequencies of light--probably not visible light because you can't see through it, but possibly IR or UV
because we were measuring off-state leakage current in a FET, and noticed that when our hand was on the controls, putting our shadow on the device, the leakage dropped from about 300 nA to about 100 nA
then we turned off the lights, and it dropped all the way to 7 nA.
This is going to make measuring leakage currents Difficult.
until we get some better solution we'll probably have to cover parts in electrical tape for measurement, just to block the light
-F&A
on semiconductors and light, long
Well that last bit is a pain in the rear. XD (I knew the first bit already. :P)
@zetasyanthis we did tell our coworkers that we would probably want to do something about the light... and they said it wouldn't be a problem, but i feel like we were just proven right
-F
I feel like you were too. :P
@zetasyanthis the problem is, figuring out how the hell we'll deal with this
maaaaybe put a little enclosure around the probe station with blackout curtains or something? do they make blackout curtains that do a good job blocking UV and IR too?
-F
I don't suppose you could pot the devices in resin yourself for testing? (Granted, you'd have to check the resin types yourself.)
@zetasyanthis no, we're talking about doing tests on a wafer level
-F
@zetasyanthis can't really pot an entire wafer, and even if it was just individual dice we'd still need to put in a leadframe or at least some sorta wire bonds
-F
And I'd go with Aluminum foil if you want to reflect all that.
on semiconductors and light, long
@Felthry you get to put a fun note in the datasheet maybe?
@Felthry Wow, that's a hell of a physics glitch y'all found.
on semiconductors and light, long
@Felthry That's wonderfully strange and I'm eager to hear how this story turns out.
@dodec i mean the transparency of alumina is strongly dependent on sintering conditions, so it probably varies between devices
it could also be that this is perfectly opaque and just has a crack in it that's too small to see, too!
we're probably going to end up with a brute force solution of just, cover the thing with something opaque
-F
also ping @starkatt and @Kyresti because i think you might find this interesting
-F