My results of specific types of IC chips, flatpacks and BGA

[arisas]

that what i call a POST about gold recovery bravo to you sir you give to the user here what they came to to ask for bravo again...

 

[Tzoax]

5. Repeating steps 3-5 until all of the white/grey powder goes through the sieve, and the resin consists of wires, silicon dies and heat spreaders

Do you happen to know the mesh size of the sieve you were using? I've been using a flour sifter, approximately 25-30 mesh? I believe I counted 27/inch. (The square holes are a bit smaller than 1mm by 1mm) I was wondering if I should keep going until everything is around -100 or smaller, either with my pipe crusher, mortar/pestle, or just re-burn it. I just sifted out 397g of black/grey powder from a batch I did yesterday.

I ran a small batch on an old scrap BBQ before, panned it a bit, then ground it up small enough to have the powder go through a coffee filter. I ended up with some copper and bonding wires, maybe 15 of each. Took a good while to get down to just the wires though.

I was wondering if it's possible to crush the pieces up enough, so that you wouldn't need to incinerate it at all? I have to buy coal/charcoal, but I can use my pipe crusher or mortar/pestle for free. I'm just seeing if I can save a couple dollars.


Edit - Nov 10th (Answered my own questions)

Over the last 2 days, I very slowly went through approximately 200 grams of the powder I had. The bonding wires must be around -150 or -200 mesh. I knew they were really small, but I didn't think they were quite this small. Just messing around with a makeshift goldpan, I spent atleast 10+ hours going through it. I used a hard drive magnet, and even went as far as using tweezers to pick out the copper wires. It might end up weighing 0.1g? (it's still wet and is probably around 80% gold, the rest is black). It was all mixed chips "even the 3 legged ones with the copper tops that have a hole in them", just the lowest recovery, no Ram, CPU, N/S bridge etc.

Yes, it is possible to crush the chips enough, that you wouldn't need to incinerate them. (Just incase you live in an area that doesn't allow fires) It'll take you a while if you use a pipe crusher and mortar/pestle though.

On a side note, panning bonding wires is probably about the same as panning "gold flour", except you might want to use an eye dropper instead of a snuffer bottle. I may just melt this down into a tiny picker, then dissolve and clean it after saving up more.

Grelko, Sorry for the late answer, I was busy these days... I use simple plastic sieve like in the picture. I've done 30+ kg of chips with just this one, this sieve is very strong and it perfectly fits to jars where I shake and sieve the chips. I don't know the mesh size but it does job for me. I am considering to buy a bigger one with the same mesh size to save some time.

I don't use a pipe crusher and mortar/pestle. I shake the burnt chips inside of jar, sieve them, and the rest of the chips or parts of the chips that wasn't burned enough/didn't go through the sieve - I burn them once again. That way the silicon dies stays in one piece and at final step i remove the wires, plates and silicon dies by hand and with magnet. There will still be some little pieces of chips there and I burn them once again or if they are soft i crush them with spoon and then sieve that again. At the end I sieve the whole powder again and remove the much of little copper/kovar wires i can. Then i rinse the powder with water and remove magnetic parts with magnet.
I haven't try to grind the chips, but researching the forum I found out it is a not good way because of many reasons.

 

[mls26cwru]

for what its worth, I pass my material through 40, 100, & 400 mesh sieves...The bonding wires will pass through even the 400 mesh... They do have a tendency to clump together with other copper wires, so I pull the clumps out during the screening and process them later.

 

[Grelko]

Grelko, Sorry for the late answer, I was busy these days... I use simple plastic sieve like in the picture. I've done 30+ kg of chips with just this one, this sieve is very strong and it perfectly fits to jars where I shake and sieve the chips. I don't know the mesh size but it does job for me. I am considering to buy a bigger one with the same mesh size to save some time.

I don't use a pipe crusher and mortar/pestle. I shake the burnt chips inside of jar, sieve them, and the rest of the chips or parts of the chips that wasn't burned enough/didn't go through the sieve - I burn them once again. That way the silicon dies stays in one piece and at final step i remove the wires, plates and silicon dies by hand and with magnet. There will still be some little pieces of chips there and I burn them once again or if they are soft i crush them with spoon and then sieve that again. At the end I sieve the whole powder again and remove the much of little copper/kovar wires i can. Then i rinse the powder with water and remove magnetic parts with magnet.
I haven't try to grind the chips, but researching the forum I found out it is a not good way because of many reasons.

Mine was just a stainless steel flour sifter. Yesterday, I ordered a couple more sieves, a gold panning kit, and a few other items. The sieves are atleast 100-200 mesh.

I started trying to incinerate my chips, but couldn't get rid of the plastic smell "because my neighbors live about 15 feet away", so I was going to crush all of them into powder "I know about the silicon dust etc" and I'll be wearing some type of filter mask. I had all the powder in a jar to shake/break the powder up, then got everything wet and took out the magnetic pieces. Picked out the copper wire with tweezers, then broke it up more with a mortar/pestle until everything looked to be the consistancy of flour or finer. (I basically used most of the same steps as Patnor's 1st E-book)

for what its worth, I pass my material through 40, 100, & 400 mesh sieves...The bonding wires will pass through even the 400 mesh... They do have a tendency to clump together with other copper wires, so I pull the clumps out during the screening and process them later.

I noticed that the bonding wires were EXTREMELY small. :lol: 200 mesh seems like it's good enough to get them out of the powder. If I see any clumping together, I put that part back into the mortar/pestle and crush it again. I'll be processing the copper wires later to recover any gold that's still mixed in.

Even with a 10x jeweler's loupe, some of the bonding wires are hard to see.

I'm about to crush up close to a pound of TO-92's? the small round, but flat on 1 side, with 3 legs "not gold plated legs" to see if any of those have gold wires in them.

 

[mls26cwru]

remember to save the silicon dies and process them for PMs as well... there is a bit of brazing that gets left behind where the wire brakes off from the die during grinding. It is a substantial amount and worth going after if you have enough... its usually a silver/gold alloy (for the chips i have processed) so you will have to take that into account when you process them.

 

[Nickperl]

Very helpful. I have work to do. My result from just the ceramics was .77g of 139.5g chips (again, not the fiber bottoms). My process was as follows:
1. 139.5 grams of ceramic tops of Northbridge IC chips.
2. Pyrolized in SS pan over burner with torch over (covered)
3. Mortar and pestal
4. Meshed through grease screen used in cooking (sorry, no pic - but can get one).
5. Pan for gold. (all other material separated and processed via nitric including copper, heatsinks, even the wafers - then merged with all nitric results before AR)
6. Nitric, then rinse, then dry , then AR, filter and put in separatory funnel
7. Add Butyl Diglyme and shake for five minutes.
8. Rinse 2-3 times with five percent HCL
9. Drop with dissolved Oxylic Acid.
10. Put on heat, nearing boiling, stirring often, takes 2-3 hours.
11. Filter (coffee) and rinse. Then put back in beaker and boil with water, again with HCL, again with water.
12. Filter and then let dry in filter.
13. Move to crucible and melt.
RESULT. This is just the ceramics. The fiber still needs to be done. .77 gram gold from the 139.5 grams. With all my gold, I eventually re-refine to gain purity.
PROBLEM. Any time I use the AR the solution always has GREEN in it. The only yellow AR I get is when I'm processing gold soaked from china and dishware using Acid-peroxide process).
PROBLEM, My yields are not as good as yours, although I did not do the fiber halves. I understood there wasn't much in those anyway but I will eventually process those.

 

[g_axelsson]

There is plastic tops on the BGA chips. Not ceramics.

Göran

 

[Nickperl]

Plastic versus ceramic I guess I can't be sure. They seemed to crush all the same. Thanks.

 

[Tzoax]

Nickperl, thank you for sharing your results. There could be many reasons why the results are varying, some of them are:
1. If you have a pile of all kind of BGA chips (only PLASTIC tops, there are no ceramic BGA chips as i know) there is a big difference between the ones with and without internal heatsink, the more of chips with internal heatsinks you have - the less gold per weight there will be. I am intending to test them separately one day.
2. When you sieve the chips, the longer gold bonding wires will get tangled together and will not go through the sieve, on the first picture there is that residue that didn't gone through sieve, the tangled gold bonding wires are visible with naked eyes, on the second picture are the same gold bonding wires - magnified. Make sure to when processing BGA chips, all of that residue mix together with your concentrate (previously washed with water) and then continue with nitric acid or whatever process you are using, Before everything, when chips are burned i am only remove the internal heatsinks, everything else goes to process, and this steps only relates to BGA chips since they do not have any magnetic metals, they also have very little basic metals.


3. The results may varying because the chips are not burned to ash as well, i do not use mortar and pestle, i just shake them in jar and the parts that stays hard, i burn them again until they all becomes white and soft, that way there is less risk that gold bonding wires stays trapped inside of unburned clumps of plastic chips.
4. maybe you processed everything well and that is a true yield for that chips, there are many types of BGA chips and the yield is depending of type, where are they built, when are they built, etc.

I don't know the exact yield of green fiber bases, i tried once to process them, and obtained a very little of gold, about 0.1g for a bunch of them.

"PROBLEM. Any time I use the AR the solution always has GREEN in it. The only yellow AR I get is when I'm processing gold soaked from china and dishware using Acid-peroxide process)."
That means you still have the basic metals inside of solution, when you (before AR) properly process the material with nitric acid (and washed with dist. water) there should be no green/blue color or very little in your AR solution.

 

[sokon]

I don't know the exact yield of green fiber bases, i tried once to process them, and obtained a very little of gold, about 0.1g for a bunch of them.

In green bottom bases is 0.7g / kg. Analysis made in Umicore.

 

[Tzoax]

I don't know the exact yield of green fiber bases, i tried once to process them, and obtained a very little of gold, about 0.1g for a bunch of them.

In green bottom bases is 0.7g / kg. Analysis made in Umicore.

Thank you very much for that information. I am collecting the green bases and one day I will process them. And there is a gold inside the BGA solder balls too, i read somewhere there is 2g of gold per 1kg of solder balls, i never tried to test the yield, but I've just done a gold presence test and i am sure there is gold. I have read about it in a Patnor's posts and i think that some people were saying it is not worth processing because of the spending too much acids. If someone tested the BGA solder balls it would be nice to share the results. Thank You.

 

[g_axelsson]

Most BGA have flash gold on the surface of the copper. (For example the small gold corner on the base.) When the solder balls are soldered to the fiber base the gold is dissolved in the solder so there is no surprise that there is gold in the solder balls.
The same is true for solder from all gold plated boards that are soldered.

Göran

 

[Tzoax]

Thank you very much for explanation Göran. I will make a test for gold content inside of BGA solder balls as soon i collect several hundred grams of them.

 

[vgecas]

hello,

maybe someone have tried to strip gold plating by putting plated items in molten solder bath and later dissolving solder in HCl ?

 

[mls26cwru]

That is a very good result! I was afraid that the short small gold wires in the BGA chips would mean less gold. It's quite a lot actually, compensating for the smaller component size.

I've just started going through boxes of memory sticks and was going to do a test as yours, but you beat me with a few weeks.
I will add my results when I have them.

Thanks!

Göran

I did a test too, but my numbers are a decent amount lower... I ended up with 1.1g Au./lb of chips (1.9g of Au from 770g of chips). I have not checked the residue from dissolving the solder balls for gold, and I am going to rerun my ash through the sluice to see if I may have missed some stuff in there. I will update if something significant changes :/

 

[Tzoax]

This will be the next type of IC chips for gold content testing - small and middle CERAMIC IC chips like shown on the picture:



I scrapped this chips from telecom boards from middle 90's.
They DO NOT HAVE gold plated wires, gold plated cap, and golden bonding wires. They ONLY HAVE gold based alloy used for BRAZE. So, this test is basically for gold content inside of the braze used to solder chip to the ceramic.



According to this link http://goldrefiningforum.com/phpBB3/viewtopic.php?f=49&t=21617&p=223791&hilit=braze+pins+times#p223791 and informations that archerytech1 posted, I am expecting relatively good yield. This is the text from the archerytech1 post:

(1) Plane areas. 4-8 micro". An example of "Plane" areas are large gold plated areas covering a circuit card. The gold's main purpose is to provide some corrosion protection and that only takes about 5 micro". There may be some shielding going on also.
(2) Wear areas. 25-35 micro" (usually closer to 30). These are areas that might be connected and disconnected 100's of times during the life of the instrument. The best examples are connector pins and circuit board fingers.
(3) Heated areas. 45-55 micro". On some all gold plated parts, such as old side-braze IC packages, the chip is brazed to the base at about 700F and the lid is brazed to the seal ring at about 560F. The entire package is heated. If the gold were much thinner, there would be a tendency for the gold to discolor during the heating. Also, there could be brazing problems if it were thinner.

I've separated the IC chips with knife and hammer and this is the look of the chips now, they are ready for AR treatment.



According to the above data the thickness of the braze is 45-55 micro inches
That is approximately 50 micro inches = 0.000127 cm

The volume of gold that fits on 1 square cm of the braze field is: 0.000127 * 1 * 1 = 0.000127cm3

The density of Au is 19.3 g / cm3

So, the theoretical weight of PURE gold per 1 square centimeter of braze field is:
0.000127 * 19.3 = 0.0024511 g

The weight of test IC chips is 153.9g, there is 38 pieces, average weight per one chip is 4.05g.

The approximate surface area of 1 chip braze is 4mm x 4 mm = 16mm2 = 0.16cm2

And finally, total estimated gold content for 153.9g of small and middle IC chips (with gold braze) would be 38 * 0.0024511 = 0.0931418
Minus 20% = 0.07451344 g of gold

The 20 percents was subtracted because braze alloy is made with 20% of tin and 80% of gold.

That would be 0.4841 grams of gold per 1 kg of those chips.

But, it is just my guessing, we will see when i bring this to the end, i will post my results when i finish the test. I hope that i will get some detectable gold weight with my scale with only one decimal.

 

[Anonymous]

If, as you are suggesting, there is 80/20 gold braze on these then you'll be extremely pleasantly surprised when you process them.

 

[Tzoax]

If, as you are suggesting, there is 80/20 gold braze on these then you'll be extremely pleasantly surprised when you process them.

That is the data that i found on several sources, i hope that it is a truthful. Yes, and considering that the braze "maybe" have some tin and the volume of chips is large comparing to the gold ratio, i have a bad feeling that there is also a possibility to be extremely pleasantly surprised when i face the problem with monoatomic gold. :mrgreen: It happened to me couple of times but with some other materials, and i haven't mastered the solving the problem of monoatomic gold. This is a first time i process this kind of chips. And i've read it have some nasty glue used to glue the two halves of chip. I hope that it won't cause me much trouble in reaction.

 

[Anonymous]

Colloidal gold? Never had a problem with it mate. Make sure you're not playing with tiny amounts in the first place.

 

[Tzoax]

Thank you spaceships, next time when i have very small amount of gold in very diluted solution i will use evaporation to prevent creating colloidal gold (not monoatomic, thank you for correction).