Old CPU caps

[MiltonFu]

I recently had some old CPU caps whicj I soaked in 35% nitric and heated. I recovered the gold flakes and washed several times with hot water.
These were then dissolved in AR and the gold was recovered. The solution was yellow when dissolution took place and remained yellow after the gold was recovered. In the past the solution would be clear and colorless after the gold was recovered.
I tested the solution with stannous chloride and the result was red.
Any ideas?.

 

[goldsilverpro]

Ah! It's so good to see a pile of those old (actually, they still use them) all-gold 40 lead (and, other sizes) side-braze packages. They add up so fast and they're so easy to run. I've literally run many 100s of 1000s of them - as many as 50,000/week for 3 or 4 years (they were my life at that time). I selectively removed all the gold and braze, replated them, and shipped them back to Intel or AMD, for re-use (new chip, wires, lid, and brazes). Besides the unit price we charged, the gold braze(s) we removed was all gravy and basically paid for the process.

I have no idea what the red is.

The nickel/gold plated Kovar (54Fe/29Ni/17Co) leads are brazed to the Ni/Au plated Mo/Mn side pads with Ag/Cu braze. The Ni/Au plated Kovar lid is attached to a Ni/Au plated Kovar sealing ring (attached to a Mo/Mn pad) with 80/20, Au/Sn braze. The gold backed chip is attached to a Ni/Au plated Mo/Mn chip pad with a Au/Si eutectic (about 97% Au). The chip is wire bonded to Ni/Au plated Mo/Mn inside fingers with either 999.9 gold wires or pure Al wires. So, you have Au, Ni, Cu, Ag, Sn, Si, Mo, Mn, and, maybe, Al wires and/or W traces (instead of Mo/Mn). That's about the size of it, except for the 3 layer Al2O3 ceramic and the 5% glass and the internal Mo/Mn traces it contains.

The reason Kovar is used is because it has a similar coefficient of expansion as the glass that holds the ceramic and the fired Mo/Mn thick film traces together. That way, no leaks.

When everything you see on the outside is gold plated, the gold value should run about $200 - $240/pound for the 40 lead packages, at a $1200 market. The smaller packages (28, 24, 18, 16, 14 lead) will run higher. The small 14 and 16 leads can run double that of the 40 leads. Good stuff!

As you did, I would cook them in 50/50 hot nitric until all the Kovar is dissolved. Then, dissolve the residue in hot AR until the chip will slide around freely (test them with a probe) on the pad (this assures all the Au/Si braze under the chip is dissolved). At this point, 100% of the Au should be in AR. Some people eliminate the nitric and just dissolve everything directly in AR. That works fine, also, but you have to deal with some AgCl.

 

[teabone]

Very good scrap indeed , and easy to process. I always used AR with good results without the extra step of nitric.
I too sometimes ended up with a yellow or orange tinged solution after precipitating the gold out. I always assumed it was either iron or kovar.

 

[goldsilverpro]

I've done them both ways and both are about equally simple. I prefer using nitric to eliminate what I can, before using the AR. That way, the gold drop ends up purer and there's a lot less gold solution to deal with. You don't have to filter and rinse after using the nitric. Just carefully pour off all the solution you can, add some water and pour off again, and then use the AR.

 

[goldsilverpro]

MiltonFu,

I would be interested to know the weight of packages you started with and the exact weight of the gold. Just to see if the numbers I think I remember are right. I'm guessing it took about 35 pounds to get the 10 oz.

 

[teabone]

Goldsilverpro,

Your numbers are about spot on. You had a older post stating those numbers and I used them as reference when bidding on a lot on ebay.
Needless to say , I won the lot with money to spare (about a year ago).
After processing (28 lead packages ) the numbers were very close , if not better then what you stated in your post.

 

[MiltonFu]

GSP,
The nugget (10.65 ozs ) was acheived with close to 42 lbs of CPUs. I have the residual AR and I believe there is still some gold in it. I will have it tested by atomic absorption just to make sure. Not sure about the silver content since it was only the caps I ran in nitric.
It is wonderful to share and confirm information with the experienced folks.

Teabone, I believe you are right about the iron causing the yellowish tinge and the redness on the stannous. I will dissolve some iron in AR and try it on stannous to make sure.
This is great material for recovery.

Thanks for the feedbacks.

 

[goldsilverpro]

I have seen some stannous coloration from cobalt, but I am thinking it was light blue. I'm having problems seeing the iron producing the red color from stannous chloride. Maybe, the color comes from the Mo, Mn, or W.

To me, the secret of getting all of the gold with the AR is to try to push the chip around with a probe and make sure it will slide around freely. When the gold is gone, the surface tension holds the chip (or, chip pieces) onto the pad but it will still slide around when you push it. The gold/silicon braze is the last thing to dissolve because the AR has to penetrate from the edges of the chip. This penetration is slow, even with a lot of heat - it can take all day, or more, to completely get the gold from under a whole chip. Although I usually left the package whole, it can help to break up the package a bit so that the chip is broken into a few pieces. That way, there are more chip edges for the AR to penetrate. However, leaving the package whole doesn't take that long to process and I felt that the labor of breaking up a zillion parts wasn't worth the effort. I definitely don't like to pulverize the package because I like to be able to move the chip around to tell me when all the gold is dissolved. Also, it's a pain to efficiently filter and rinse all of that ceramic powder.

 

[MiltonFu]

GSP,

Perhaps you are familiar with the old Northern Telecom interrupters that were made by johnson matthey?. Could you comment on the content of the contacts?.

 

[goldsilverpro]

GSP,

Perhaps you are familiar with the old Northern Telecom interrupters that were made by johnson matthey?. Could you comment on the content of the contacts?.

I would need a photo or a link with a photo and a description.

 

[Harold_V]

I tested the solution with stannous chloride and the result was red.
Any ideas?.

The color is reminiscent of platinum, but unless the test has had ample time to settle, I would have expected it to be uniform throughout, not concentrated at the bottom. Assuming it is platinum, with a greater concentration, it would be even darker, leaning towards brown.

What I am wondering is if you have iron in solution. Please apply a drop of HCl to the test, to see if the precipitate dissolves.

It has been a long time, so memory fails me, but I believe that in Hoke you can find a test for iron. It's very handy when working with residues, and you're not quite sure of what you're seeing in other tests.

Harold

 

[Higashi]

Dude that is some pretty antique and vintage CPUs you got there... not old. 8)

I'd keep one in a glass box/frame/display.

 

[MiltonFu]

Thanks harold. Will try the HCl.