Learn IC Decapping

Decapsulating ICs used to be an exotic technique. (I should know, I did that professionally for one of the big IC vendors back in the 1980s.) These days, more and more people are learning to take apart ICs for a variety of reasons. If you are interested in doing it yourself, [Juan Carlos Jimenez] has a post you should read about using acid to remove epoxy from ICs.

[Juan Carlos] used several different techniques with varying degrees of success. Keep in mind, that using nitric acid is generally pretty nasty. You need safety equipment and be sure to plan for bad things to happen. Have eyewash ready because once you splash acid in your eye, it is too late to get that together.

We never took apart MEMS devices, so that was especially interesting. A little metal lid keeps the epoxy out of the moving parts and that turns out to be hard to remove.

We were surprised he used 69% nitric. We used to use white fuming nitric, but that may be a bit harder to get now since it is associated with explosives. Then again, you can make it yourself. We used to also use copper bars to sit between the hot plate and the device to get better heat transfer. The acid reaction is much stronger when you heat it.

By the way, if you ever wanted to actually probe the device, you’d have to remove the passivation layer which is SiO2 glass. That takes hydrofluoric acid which is very nasty stuff, indeed. You can see the edge of the passivation around the bond pads under a microscope. You’ll notice the surface of the pad focuses differently than the edge where the thin passivation layer starts. When you can’t see that edge anymore, you’ve removed the glass.

Of course, you might not need to probe the actual circuit to get results. If you don’t care about the circuit’s survival, you can try more mechanical means.

15 thoughts on “Learn IC Decapping

    1. Yeah, at work we called it bic decap.
      I had a lot of luck with putting the chip on a milling machine and removing most of the encapsulation and then cooking off the bit remaining with a nearby flame. If I had multiple chips to practice on, I could usually mill down to 0.05mm above the die surface on the second try.
      I think if I had a bond diagram I could probably manage to mostly avoid the bond wires, too. We didn’t try that.
      There are papers on using ethylene glycol to do low temperature (200C) degradation of epoxy, that combined with careful mechanical removal, might end up with a fairly low toxicity, albeit slow, decap, that results in a functional chip.
      (For our chips, nitric acid isn’t really a great choice anymore if we need a functional decap since we’re not using gold bondwires anymore.)

    1. This looks pretty spectacular if you’ve already decapped the die and are looking at it through a thermal camera when you short the power. An unintentional accident while looking for something else

  1. Neither sulfuric nor hydrofluoric acid is stuff i want to mess with. But somewhere on the internet somebody experimented with “cooking” the IC in (melted) colophonium. Sure, you should do this outside and take some precautions, but it seems a lot less dangerous than acid. Did anybody try this?

    1. There is no huge scale decapping. Even FA labs only work on a few chips per project as this is very manual process. They will usually have a test case used to identify the worst performing chips to make whatever they’re looking for easiest to find and only need a few units.

      Beyond failure analysis, extreme investigation of counterfeits and idle interest in seeing the die I’ve never seen another market for decapping IC’s.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.