Popping the Top of A Ceramic IC

If you’ve ever wanted to open up an IC to see what’s inside it, you have a few options. The ceramic packages with a metal lid will succumb to a hobby knife. That’s easy. The common epoxy packages are harder, and usually require a mix of mechanical milling and the use of an acid (like fuming nitric, for example). [Robert Baruch] wanted to open a fully ceramic package so he used the “cooler” part of a MAP gas torch. If you like seeing things get hot in an open flame, you might enjoy the video below.

Spoiler alert: [Robert] found out the hard way that dropping the hot part isn’t a great idea. Also, we are not sure what the heat does if you want to do more than just inspect the die. It would be interesting to measure a junction on the die during the process to see how much heat actually goes to the device.

The process is really fast: only about 20 seconds. We wondered if a larger part might take a little longer. However, compared to chemical methods, this looked very fast and easy, as long as you don’t mind the heat.

If you get the urge to start opening parts and want to actually probe the surface of the die, don’t forget there is a thin layer of glass over almost the entire chip. This layer–the passivation–is relatively thick and usually only has cutaways around the bonding pads. Getting rid of that layer requires hydrofluoric acid (nasty stuff). You can tell when you got it all by focusing a microscope up and down the edge of bond pad. When you can’t find the edge of the passivation, you are done.

Some people expose ICs dies to study, and some are looking for fake chips. Other times, it is electronic archeology. The last time we saw [Robert] he was building a CPU on an FPGA, so he’s clearly a hacker of wide-ranging interests.

29 thoughts on “Popping the Top of A Ceramic IC

    1. The rest of the Hackaday community is not as nasty as you man. Take it easy on the pills.

      Nah man, I saw the same thing, just didn’t want anyone else to think I did.

  1. Seems like a lot of work. Last time I did this I just laid the package on its side on a piece of scrap wood and put a wood chisel (that I didn’t care much about) on the glue line and one nice rap with a hammer and it cleaved right open. Like the hollywood image of the guy splitting a million dollar diamond. Ceramic packages are easy.

  2. I should really just bite the bullet and try the chisel. I just use the flame approach because I don’t care about a live die, and I have a failure rate (where I ruin the chip) of zero.

  3. Rather than a chisel, I used a small drill bit and a vise. There’s less chance of damaging the silicon or the bond wires that way.
    Take a drill bit that’s just slightly thicker than the width of the glue line.Place it along the short edge of the chip on the glue line, and clamp them both together in the vise. Tighten the vise to force the drill bit into the glue line and the top will pop right off cleanly

  4. My final year project was an interface bus board for a minicomputer in the Electronics department; it would protect the computer from any student projects, by giving them an isolated bus to plug in to. To demonstrate it worked, my professor got me to hook up a board for programming EPROMs (2708s) and to write some programming software for students to use. If you forget the 50ms delay in the main loop, EPROMs light up like a candle.

    But the only time I have ever automatically de-lidded something was a power transistor in a TO3 case. One went *foom* and the lid – this tiny flying saucer – went past my ear, leaving two posts and a small charred square.

  5. The way I did it was to put the chip lengthwise between the jaws of a bench vise at a slight angle so that one jaw is pushing the top one way, and other is pushing the bottom the other way in shear, and then tighten the jaws. The halves separate cleanly.

  6. Obligatory:

    That’s not “MAPP” (methylacetylene-propadiene propane) gas. That’s “MAP/Pro” which is actually propylene gas. MAPP gas has not been sold in the US since 2008.

  7. Be careful about this folks. I once wanted to see inside of an MRF151 RF transistor to see if I could determine exactly what had failed in the device. At the time they were running about $150 each. The transmitters I worked on contained about 240 of them, so the exact cause of failure was reasonably important. I though I’ll just grind the top off…. White ceramic… Suddenly realized that I was grinding on beryllium! No dust mask! Needless to say I immediately quit what I was doing. Just luck I didn’t get a lung full of beryllium dust! NEVER GRIND ON WHITE CERAMIC!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s