Potential Loss With Ceramic Chips

[Purplesquish]

Let me start off by saying how much I appreciate the information contained in this forum and those who created and contribute to it. It has served as an invaluable source of information that has helped me successfully recover and refine 6.2 grams of gold to date with virtually no issues. I'm sure that 6.2 grams is small potatoes to some, but it makes me giddy beyond words, so Thank You!

Now my question...

I have accumulated ~500 grams of ceramic ICs, which I am beginning to prep for processing, and want to make sure that I am not throwing away potential gold. The lot is made up of Linear Amplifiers, SRAM and EEPROMs from the late 80's to mid 90's, all of which are either purple or black ceramic. None have gold plated legs, and none possess lids of any kind. All of the ICs are MIL spec, and visually appear similar in form factor to #1 in the image below.

With the top ceramic plane removed, it is obvious that the chips internal topographies differ between each type, but at the same time they are all very similar in that they all appear to have a solid backplane of gold with silicon wafers attached in varying configurations similar to #2 in the image below.

In preparation for processing, I have been removing all of the legs (see image #3), and cutting off the non-populated ceramic ends from the base of the chips (see image #4). I examined the removed edges using a jeweler's loupe, but failed to see anything that would suggest that there is any gold bonding wire, or anything for that matter, running through the base ceramic plane. I did the same with the legs, looking at the end closest to the heart of the chip, but did not see any visible gold bonding wire or gold of any sort.

I know that every IC is different, but generally speaking, do you believe that there is any gold hidden in the ceramic sections or legs that I am removing?

I intend to lift the gold backplane from the chip using the AP process, but I am concerned that if I process what's left of the ICs as is, as they are in image #4, that the acid will not be able to reach the base metals behind the gold plane, as the gold plating appears to have full coverage, thus not being able to free the gold foil, similar to issues some face when processing large MIL spec pins using the AP process. Is this something I should worry about dealing with, and if so, should I just roughly break them up?



Again, I really appreciate the free exchange of knowledge and experience that this forum provides. Thank You!

 

[the iron dwarf]

one thing to consider when breaking these up is that from each leg at the end nearest the gold is a tiny solid gold wire to connect the leg to the chip, this will be very hard to see and will be easy to lose.
it may stay on the leg or the chip or end up in the dust, it can be a tenth of the thickness of a human hair and only 2 or 3mm long

 

[macfixer01]

one thing to consider when breaking these up is that from each leg at the end nearest the gold is a tiny solid gold wire to connect the leg to the chip, this will be very hard to see and will be easy to lose.
it may stay on the leg or the chip or end up in the dust, it can be a tenth of the thickness of a human hair and only 2 or 3mm long

Purplesquish,
As Iron Dwarf said, they are very hard to see but there is a hair-thin wire connecting the inner end of each pin to the pads around the edges of the chip die. Those wires may be gold, or may only be aluminum? I don't believe I've ever heard of any white gold alloy being used for those, so their color alone should tell if the wires are worth saving. The lower ceramic layer may or may not have gold in the well underneath the chip die as yours do in the photos. So save the bottom ceramic layer and any loose bits of gold wire you can locate. Then once you get any remaining bits of gold wire off the ends of the pins the rest is pretty much useless. (although there may be some copper in the pins I suppose?)

The chips you have are a little unusual in that they seem to have 2 totally separate chip dies so they have 2 separate device functions (like a Dual Op Amp for example). Multiple dies in some chips such as hybrids aren't unusual, but there would usually be some wires internally interconnecting the two dies also. It's fuzzy but zooming in I think I can actually see wires dangling from the left-hand die in the chip that's placed to the left of the circled pins?

macfixer01

 

[goldsilverpro]

We called those parts Cerdips. The 2 ceramic halves (the ceramic contains about 5% glass, which bonds the ceramic together) are cemented together by a layer of molten glass during a firing step.

I developed a simple method to break them apart and expose the encapsulated gold. The parts were heated to about 800-900F and were then quenched in water. This fractured the glass and the 2 halves separated. Then, they were then tumbled alone in a cement mixer. This broke the Kovar legs loose, which were then separated with a magnet. This left the ceramics with the exposed gold, which was then leached in aqua regia.

In the mid-80's, when gold was around $400/oz, the Cerdips like yours were worth about $.05 each in gold. That is about $.15 each today.

 

[Purplesquish]

Following your comments I set out to dissect one of each type of these chips to get as close to an undisturbed view of the interiors as possible. After a few attempts I was able to crack open one of the linear amplifiers without damaging the interior, giving me a fairly good view of the wires connecting the legs to the chips (see image below). It appears that this particular chip variant does not contain gold bonding wires. I will post similar pictures for the other variants for documentation purposes.