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

[jason_recliner]

I would put my money on a HCl wash then a copper ii chloride leach, as the best process here. No nitric/nitrate means no risk of tinny slime. Incineration also would not not necessary, so you'd avoid a fine carbon mess to have to be entirely rid of.

Just as long as the leach is pre-used copper ii chloride and not a fresh batch containing peroxide. For often with AP [in the literal sense], if you dissolve a little gold it doesn't matter as it will cement out. But with these fibre boards, dissolved gold carries some risk of cementing on copper traces inside the fibre layers.

I am a little concerned that this cementing could also happen with your KNO3 PMAR. If that dog happens to bite, you still have the ability to incinerate and crush to reclaim it. Though then it starts becoming a lot more work for the small yield (as maybe 95 percent of the gold is in the black tops).

Best of luck

Edit to add:

But, on my surprise, the metastannic acid goo was starting to form, i thought that tin inside of AR solution should form soluble Stannic chloride, not metastannic (that is formed in a Nitric acid solution).

This may not be stoichiometrically correct, but something like:
HCl + KNO3 -> KCl + HNO3.
Your added nitrate gives you a little weak nitric acid in your hydrochloric, plus potassium chloride salt.

Tin in AP will form stannous chloride (and cements copper out first too). Tin in AR or PMAR will form goo.

 

[Tzoax]

Thank you Jason, once i tried to process several of fiber masks (with no solder balls) in a AP, and it did quite well - the foils separated and begin to flow, just like finger foils. And it was done in a 2-3 days with no air bubbler or heating.
Now i was thinking that because there was solder balls on fiber masks, it would be quicker to process it with AR after HCl boiling, but this way it is not only longer, than it is much complicated and it is definitely not a good way to process this kind of material.

 

[Tzoax]

After all, i made it. After boiling in HCl (in total 2.5 liters) i removed majority of tin. With PMAR (in total 1.5 liters HCl + 150g KNO3) added in portions (300-400ml) copper and gold finally started to color the solution with beautiful emerald green color.



This is a look of remained material.



I filtered the solution.
Since i had a lot of solution with little gold (1.5 liters) i started to evaporate the solution. But, after reducing volume by 1/3 (to 1 liter) the solution turned black.



There was some of the stannous chloride left in solution, so i was sure that it was colloidal gold. I stopped evaporating, neutralized the solution with urea until there was no foaming, tried to drop the gold with SMB but nothing happened, the gold was in colloidal state. I added about twice much hot water, heated and still nothing happened. After couple hours, when solution cooled the gold was forming at the bottom. *(I was surprised since i was expecting that gold will drop faster when warmed, but in this case it started to drop when it was cool. Maybe it just needed some time, or maybe there was much contaminants like AgCl that is settling only when cold)
Anyway, the gold looked dirty.



So i dissolved it again in PMAR and this is a look of AR now, after second dissolving. I pictured this picture with a flashlight so the contaminants like silver chloride would be more visible, but it looks like there is very little of contaminants.



So, it turned out pretty well, i did this with a wrong method, and after 5 days of work i managed to obtain some gold. I will wait 'till tomorrow that contaminants settle, i will drop the gold, wash, and post result.

 

[jason_recliner]

The final colour is looking good to me.

I advise you stock-pot all of your wastes from this exercise, with the exception of initial tin washes. I expect there is a probably a decent proportion of gold remaining in it. Note I said proportion, though maybe not quantity - it's all relative isn't it? But it will be good practise in your recovery techniques.

 

[Tzoax]

Thank you Jason, i will save the rest in stockpot. I dropped and washed the gold.




There is 0.1g of gold in 218.8g BGA (green) fiber masks with BGA solder balls still attached on them.
Gold content per 1kg of material - 0.45g.

*My scale have one decimal, so the result could be anywhere between 0.05-0.149g.

 

[Anonymous]

I think what you're doing here is great, however please excuse me when I politely suggest that using scales with this low level of reading makes the results posted pretty inaccurate all round especially when coupled with the small sample sizes.

It's almost at the point of not being useful because your errors in some cases could be over 50 percent yield. Might I suggest that it would be more pertinent to re-run all these with a valid set of scales so that your findings can be relied upon?

Again let me state that I think it's a great thing to do and report upon however the figures have to be more accurate to be of any use to members reading them. As they are, the margin for error is just too large.

Kind Regards

Jon

 

[Tzoax]

Jon, when i buy preciser scales (0.01) i will re-run all sort of material where the result was very low (0.1g, 0.2g...). I never processed materials like silicon dies or BGA fiber, solder balls, etc, so this was just a quick test to get familiar with process and see is it worth processing. In all cases i found it is worth processing even if the result was 0.05 instead of 0.1 for such a small sample sizes. When i was buying scales i thought that i wouldn't need preciser scales more than one decimal, but i also had no idea that one day i will test materials so precisely. Thank you for all of your help and advices, i will buy 0.01 scales as soon as possible. I am collecting silicon dies, bga fiber masks and bga solder balls to collect much larger sample sizes. Also, i already prepared a many types of IC chips (with minimum sample of 1kg) for re-running.

If you have read my posts about IC chips, you noticed that results was only from ashes, not including silicon dies, wires, plates, and possible bonding wires trapped in all that material.
So, even in that cases it is not so precise, because that chips contains even more gold.
Now i have idea of re-running many types of IC chips but with complete treatment, to get all the gold out of specific types of IC chips including silicon dies, wires, plates, and possible bonding wires trapped. All of that plus with my new 0.01 scales i think that results will become much, much valid and useful for GRF members.

 

[Anonymous]

Brilliant Tzoax- thanks for the answer- I'm looking forward to the results. 8)

 

[richard2013]

Tzoax thanks for the hardwork

Hope you can try cellphone small IC types, also please consider the recovery of gold powder dissolve in hot HCL or left for longer in cold HCL.

 

[Topher_osAUrus]

Tzoax,

On the green bottoms (see attached)

Is the spot that i have circled able to be scraped off with your fingernail?

It looks black, like cemented gold to me?
But, i honestly have *NO* idea, just a curiousity. As i would think the underlying material beneath the plating would be a different color.

Again, excellent work on this thread.

-me

 

[Tzoax]

Topher,

i thought too that it could be cemented gold on top of copper, so i boiled it again and again with fresh PMAR but nothing happened. When i looked it with microscope it looked like plastic, very smooth and shiny (light reflection) so i guess it is just a color of material beneath the copper. It looks exactly same like on fingers (when foils falls off), in the rest of processed residue fingers you will see a black line on many of them. I guess it is some kind of barrier between fiber and copper.

 

[Topher_osAUrus]

Could be a carbon barrier, but thats pure conjecture on my part.

I havent ever done these really..
Well, i have, but i stupidly threw them all in the fire pit then treated as ash.

Thats the last bucket of my mistakes to fix... Ill get to it... Eventually.

-topher

 

[Tzoax]

This is illustration of gold source of IC chip silicon dies, there are 4 possibilities for cracking, only when gold bonding wire/joint breaks, the gold stays on the silicon die.

 

[mls26cwru]

As promised, my numbers from my last run (finally ).

Wt. Si chips: 105g
Gold recovered: 0.40g*
Silver was recovered as well in the form of AgCl, but added to a larger batch to process later. Estimate between 0.5-1g of Ag.
total wt% gold: 0.38%

*my scale is accurate to 0.01g, but I rounded down from 0.44g since it was gold only refined one time.

I also have the luxury of a microscope to visually confirm when all the bonding wire stubs have been dissolved. While its not much, it adds up. Usually I wait until I have a pound or more of chips, so it is definitely worth saving them.

 

[Tzoax]

mls26cwru, it is great result! Thank you for sharing. I have many pounds of IC chips wires, internal heatsinks and silicon dies and it is all mixed up, all along with some gold bonding wires that escaped me there.

I made a test with mixed (barely magnetic) pins in AP and it worked quite well.
I was thinking to separate the iron/kovar based wires and plates from barely magnetic ones with magnet. And to process in AP copper/brass/bronze (barely magnetic) wires, plates and silicon dies with it. When the reaction completes in remaining material besides gold there would be Si dies, after filtering dies could be processed along with gold in AR. That would save me a time to manually separate Si dies.

I am very interested in the results because i am wondering how much gold bonding wires i missed, how mush gold is in brazed contacts and plates, possible gold plated wires and plates etc. all together.

 

[Tzoax]

I made a test with a different way of processing IC chips and i managed to get about 30% more gold than i usually do. I don't know exactly how (i have some thoughts how it is possible), but when i was processing ic chips this way (few times) i was always surprised with gold content. But i wasn't sure, because i wasn't testing, i expected approximate yields and it was always higher than expected. This time i decided to make some tests compared on my earlier test results (done in a regular/old way (diluted nitric/AR)) and on my surprise the more than 30% of gold showed up.

I tested multiple times specific types of IC chips and the results was always the same +- 5 or 10 percents maximum. But processing the new way, results was always better than processing the old way for about 30%.

Example:
Test:

Total: 931g

Small IC chips 313.7
Flatpacks 4-Sided Square + Rectangular THIN 350.7g
Flatpacks 4S Rectangular THICK 266.6g

From my earlier results i would calculate and expect something about 1.6g for this batch (look at my earlier tests in this thread). And i would get near expected, always. But processing a new way, the result was 2.5g, little more than 30% than expected.

Interesting thing is that anyone can do the same test with both methods and compare the results to be sure.

Some pictures with gold bonding wires doing this way (above batch):



Of course, I will share with you which process i used, but i wonder does anyone is willing to make a same tests that i did (the old way and the new way with a same type and weight of IC chips), compare the results and post it here? I see that people are very rarely share their results, it would be much more helpful if for example 10 experts make a test of a same type of IC chips, with the same process, and share the results here. That way the expected amount of gold for a certain type of IC chips would be much, much more reliable. After that, in my opinion, the next direction of discussion should be modifying used process for more efficiency in gold recovery for the certain type of element.

My conclusion is that based on my own previous tests of IC chips i always had plus/minus 5 to maximum 10 percents of exception of gold content processing the well known way of iC chips. When i process the same weight of same chips with different method i recover +30% more gold. Before talking what is really going on and chemistry behind it, first i would like to confirm that my statement is right. I need a couple of volunteers that are willing to make the same test i did with both methods (dilute nitric/AR and a new one) and share the results here. That way it would be reliable. Is anyone interested?

Alexander

 

[Tndavid]

Very,very nice post. Thank you!!

 

[Tzoax]

I made a detailed instructions of the process, if someone is interested to make a test and share results send me PM. I am preparing next batch of ic chips for testing, this time i will take pictures step by step.

 

[richard2013]

Are you smelting then inquarting the Concentrated gold bonding wires in the new method?

 

[Tzoax]

Several people are already testing and they will probably post results here, if anyone is interested to try and compare results, here is what i have done.



Also, i will make another version of my process since i wasn't mentioning copper I and copper II chlorides, most likely the solution became saturated with copper, not silver since copper is higher than silver in a reactivity series of metals.. But, either way, that is process i used, i just have to make a good chemical explanation, someone without experience could mix up white powders of copper I chloride with silver chloride, so i will have to make a remake of instruction.