open-source-silicon.dev

Has anyone attempted and had success with extracting the die from QFN packaged part? We are exploring some marketing campaign to show off the die, but did not receive bare dies.

<@U02TGPV63FA> It has definitely been done: <https://www.youtube.com/watch?v=zUv6sdxOaFE> -- just be aware that working with these chemicals is extremely dangerous and definitely not recommended. Vaporised H2SO4 is extremely corrosive to the lungs. HF acid is painlessly absorbed by the skin with zero damage, but dissolves nerves and bone. Scary! More info here: <https://www.youtube.com/watch?v=UTXApNlG4Yw>

Anecdotally, I once managed to decap a chip (it was a DIP40) by heating it with a torch and then dropping it in ice-cold water. It cracked open in such a way that most of the epoxy simply popped off the die. Don't know if that was a fluke though, or if it would not work with a more-modern QFN.

This is a tinytapeout from CI2311 that was in a QFN64 and was decapped and imaged :
<https://siliconpr0n.org/map/efabless/sky130-ci2311-tt05/infosecdj_mz_nikpa40x/>

Oh cool! I didn't know that was possible, or that this image had been released. Thanks <@U016HU5HK8V>!

<@U06H50VMDPU> This is a chip that ended up damaged, shorts between power rails.  TBH I'm not entirely sure _how_ I damaged it, it was during the testing of the experimental analog designs on TT05. I was curious if there would be any obvious damage so I asked someone to decap and image it :sweat_smile:

Oh right, interesting. Nothing obvious showed up in the decap then?

Was there power gating on the experimental analog designs in TT05?

( <https://dev-zzo.github.io/blarg/2023/02/25/decapping-with-rosin.html> )

<@U06H50VMDPU> No, nothing I could really see in the decap.

Yes, there was power gating for that design.

Fascinating, thanks Sylvain :slightly_smiling_face:

Mmm, I had never scrolled there, but this doesn't look good :