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

[4metals]

Back in the 1970's I worked for a connector manufacturer and a common failure was due to brittle solder joints, this was due to gold contamination in the solder pots where the contacts were dipped to solder them. We had to routinely analyze the solder pots to keep the solder under the limit.

The pots could never exceed 0.2% gold or they had to be changed. We changed out the pots a few times a year by removing about half of the solder and putting in fresh solder.

Considering the solder pots were changed out at or before 0.2% (or 2 grams per kilogram) I wonder how likely it will be that the solder you have removed from the scrap will be at the 0.2% level? It is possible and will be interesting to see the result.

This is a website where they list the levels of contamination permitted to comply with the industry specification J-STD 001E.

http://solderlab.com/why-test-your-solder/

I remember it was never an easy task to find a refiner who wanted to process solder pot material, especially when we knew what was in there when we shipped it. But then again the gold price back then was in the $150 an ounce range.

 

[kernels]

I would be surprised and amazed if the solder balls give you 2g per kg. Surely it is only the ENIG plating that dissolves into the solder balls ? Can't wait to hear the results man, great test to run.

 

[Tzoax]

According to this document http://www.smta.org/chapters/files/SMTA_India_03_PCB_Surface_Finishes_-__Implication_on_the_SMT_Process_Yield_(Liya-Indium).pdf under "ENIG: Electroless Nickel/Immersion Gold" subject, i found that "- Au Thickness = 3-10 microinches" and about maximum gold content inside of solder by percent is "This threshold is ~ 3 wt% gold".

And here i found that gold content should not be more than 5% -
"The second coating is plated electroless nickel/immersion
gold over the copper pad. The thickness of the electroless
nickel layer is determined by the allowable internal material
stresses and the temperature excursions the board will be
subjected to throughout its lifetime. Even though the gold
metallization is typically a self−limiting process, the
thickness should be at least 0.05 microns thick, but not
consist of more than 5% of the overall solder volume.
Excessive gold in the solder joint can create gold
embrittlement. This may affect the reliability of the joint."
https://www.onsemi.com/pub/Collateral/SOLDERRM-D.PDF

This is interesting article about gold plating thickness: http://www.finishing.com/590/96.shtml

Anyway, the more i read, the more i am confused about this subject. Personally, i would be satisfied if there is even 1g of gold per kilogram. But, on the other hand it may happen that i dont see any gold from it. I guess there is only one way to find out.

I filtered the solution (while still very hot) and this is pictures of filtered and cooled solution. There should be the most of the tin (stannous chloride) and lead (lead chloride crystals).





And this is picture of remaining solids inside of filter paper.



Today i will incinerate this material with filter paper, i will measure the weight of it and most probably repeat HCl simmer boiling process one more time to be sure all of the tin has gone, because the last thing i want to deal with is colloidal gold.

Alex

 

[Tzoax]

I finally bought a scales with 2 decimals. Thanks to everyone pointing this out, now my results will be much more accurate.

Alex

 

[anachronism]

Well done Alex.

 

[Tzoax]

Thank you Jon.

I incinerated remaining material and after weighting it was lighter by 43.6%. There is still a 56.4% of total starting weight. So there is obviously more tin there. I will repeat the whole process once again (HCl simmer boiling/filtering/incinerating) and i will measure total weight again before AR process.

 

[kurtak]

Andrew

I could be wrong here - but from what I can see it looks like copper is causing a problem for you

In your OP showing your starting material I am seeing a fair amount of copper traces struck to the bottom of the solder balls (pic of solder balls still on green fiber part) & this copper is in with the solder balls after removing the solder from the green fiber

Then looking at the last set of pics after incineration (including magnified pics) it looks more like the material is copper then looking like tin (solder)

So here is what I "think" is happening --- because you are using a simmer boil - air - from the bubbles of the simmer boil is a factor in dissolving of the tin

As we know - air (oxygen) with HCl allows copper to dissolve as CuCl2

Therefore - you have some copper going into solution - but - because the tin is more reactive then the copper - the tin is cementing the copper back out - causing the copper to form a passivation layer on the tin

The reason the copper becomes a passivation layer is that as the HCl is loading with tin (& thereby cementing the copper) is the HCl looses it power (becomes weaker) to dissolve copper - the less reactive metal --- even though there may actually still be free HCl to dissolve more metal

I "think" (there I go - thinking again) :lol: what you are going to find is that you need to use an "excess" of HCl so that there is enough HCl to dissolve both the copper & the tin AND keep the copper in solution

Personally - if I were doing this I think I would set up an electrolytic cell for tin - so as not to detract from your project I won't go into detail about the tin cell - other then to say one of the problems with a tin cell is that the tin likes to grow "very" long "whisker" like tin crystals from the cathode & they grow "very" fast - so you have to really keep an eye on the cell to keep the whiskers knocked down to keep the cell from shorting between anode & cathode

Kurt

 

[Tzoax]

Great "thinking" Kurt! :shock: Thank you, now it makes sense. I knew that HCl (with a bit of air (oxygen)) will not dissolve (much of) copper either way, and i thought there was much less of copper in there so i was planning when finally getting rid of most of the tin and lead i could process that "little" copper (along with gold and other undissolved metals) in AR.

So, i am considering that i repeat this process once again (HCl simmer boiling/filtering/incineration) AND after that i could use old CuCl2 solution to dissolve most of the copper (and tin if left) from material. Then i would incinerate material and process with AR (if the most of Sn/Cu is dissolved - i will check by final weight of material).

Kurt, thank you once again, and i wish you a Happy Birthday!

Alex

 

[g_axelsson]

Incineration done correct will drive off water but add oxygen, so any exposed copper will be oxidized and is easily dissolved in straight HCl.

Since incineration adds oxygen the material actually gains weight. Just something to think about. Both copper(II) oxide and tin(IV) oxide is 25% heavier than pure copper or tin.

Göran

 

[patnor1011]

If you notice there was only 0.3% of copper in material I had scanned. It was because I melted small solder balls so any of those copper traces or lets say most of them were left out and not mixed much with tin. When I think about it more even harvesting method may be a factor. I usually do use heat and solder balls melt so there should be minimum copper present there. Scraping them off may introduce more it can be seen on your picture.

 

[kernels]

Hi Alex, any update yet ?

 

[Tzoax]

Thank you all for replies, i was busy last week so now i will continue to bring this test to the end.

I repeated HCl simmer boiling one more time.



This time the whole material becomes gray/white colored and soft. I filtered it while still very hot, but very soon filtering becomes very, very slow - one drop for about 10 seconds. After that i boiled material with water and filtered it.



This is a look of material through magnifying glass.



This gray/white colored sediment looks just like metastannic acid. Maybe there is some of the silver chloride too. But, after incinerating i thought that if metastannic was converted to metal form next time HCl simmer boiling will put tin in solution (stannous chloride)...but it looks like it produces even more of metastannic than first time (in case that this gray/white colored sediment it is metastannic).

The filtered solution (cooled) shows no crystals this time.

I am considering either to repeat incineration/HCl simmer boiling OR processing it with AR from this point

Alex.

 

[patnor1011]

One more incineration and wash does not hurt. I am wondering (and it is just an idea) that you may get better result if you try incinerate and leach it in hot sulfuric this time instead of HCl. If you are careful and brave enough to play with hot sulfuric it may speed up things for you.

 

[g_axelsson]

NaOH have also been used to fight the metastannic acid.
http://goldrefiningforum.com/phpBB3/viewtopic.php?f=52&t=20477

Göran

 

[geedigity]

I finally bought a scales with 2 decimals. Thanks to everyone pointing this out, now my results will be much more accurate.

Excellent decision.

 

[Tzoax]

Thanks Pat, i have no experience with hot sulfuric and i am not in hurry. I will repeat incineration/HCl simmer boiling couple more times if needed. I will incinerate this material and i will weight it. The filtered solution was yellow/green color and i think the most of the copper is separated from material.

And if somehow metastannic appears again and again after couple more of incineration/HCl simmer boiling, i am more interested about using NaOH like Goran said.

It looks like material weight is reduced, after i incinerate it i will post pictures and weight of it.

Alex

 

[Tzoax]

I finished second HCL simmer boiling/incineration treatment and this time overall weight was reduced by 14.6%.

This is a comparison between two treatments and overall weight:

41.8% -remaining material
43.6% -reduced weight after first treatment
14.6% -reduced weight after second treatment




It still looks like there is much of tin present, i guess i will have to repeat these steps couple more times.

 

[geedigity]

Is there any possibility that the silver that was present in the solder balls as been converted into silver chloride? Maybe you could take a small sample, mix it with an equal part of sodium carbonate and a little borax and melt to see if you get any metallic silver. Not sure if the tin would be reduced to metallic tin during the melting, but if not, then try to dissolve the reduced metal in nitric acid. Any residue would likely be your gold.

 

[Tzoax]

Is there any possibility that the silver that was present in the solder balls as been converted into silver chloride? Maybe you could take a small sample, mix it with an equal part of sodium carbonate and a little borax and melt to see if you get any metallic silver. Not sure if the tin would be reduced to metallic tin during the melting, but if not, then try to dissolve the reduced metal in nitric acid. Any residue would likely be your gold.

Thank you geedigity, i think there is some of silver chloride there, but the problem is that there is much of tin left in there. I think there is much of tin left in there based on total weight left, according to Patnor's test there should be about 87% of tin, 10% of lead, 2% of silver and 0.22% of gold. So i managed to get rid of 48.2% of base metals (tin and lead) but there is still 41.8% of material. Even if i count that tin(IV) oxide is 25% heavier than pure tin (like Goran said) it is about 30% of overall weight. So i think there must be much of tin there, little of silver and very little of gold. Adding nitric to any form of tin (with gold) will give much troubles like metastannic, gold loss and/or forming colloidal gold after, plus slow filtering (while colloidal gold is running through filter paper). It could be done that way, but it is much, much greater work with much of gold loss. So i am not processing tin with nitric or AR.

I will try to make few more times steps i already did with incinerating/HCl, and i will continue with it until there is about few percents left from overall weight.

Alex

 

[g_axelsson]

Tin can have several different oxidation states. The one that can create problem is Sn²⁺ that makes a gel that traps gold when mixed with nitric acid or colloidal gold when mixed with gold chloride. Sn⁴⁺ as an oxide might be a lot easier to filter and will not reduce gold chloride. You could always try on a small sample to see if you can put gold in solution without making an unfiltrable (is that a word?) gel.

Göran