Processing gold plated transistors in AR

[orvi]

Maybe this would be helpful for someone in some way
Material: 240 g. old transistors and logic IC, casing is usually made from ferrous alloy (magnetic) which didnt rust, so probably some FeNi alloy (i´ll get them XRF analyzed when it will be possible). bottoms are heavily gold plated, wires also. bottom parts seem to be made from non ferrous material. gold bonding wires.
yield vary widely from piece to piece, logic ICs with more than 3 legs use to have some little more goodie, also the ones without resin-cast bottoms (solid metal ones). usually between 8-13g/kg.
Main issue: lots of waste metal - as usually seen in videos on YT, guys cut the top of the transistor to reduce the ammount of iron. this is seriously tedious job, this particular type is very difficult to decap... so i wanted to do the test batch with whole untouched pieces, to see if the consumption of acid is too high.

Testing few pieces with only HCl shown very slow dissolution, so with this quantity, i opted for test run in AR to see how deep would be reduction of HNO3 in this enviroment, thus directly - how much nitric would it consume to dissolve it. Doing some math (very rough), if nitric is solely present as an one electron oxidant (giving off NO2), about 750 ml (50 % tech grade) is needed to fully oxidize iron to +3, alongside with copper legs, nickel and maybe other metals involved in casing + tin/lead to +2. For elements to stay in solution, 1250 ml of hydrochloric acid (30 %) was used.
Plan was to add HCl at the start, then onset the rxn with small addition of HNO3 and slight heat, than maintain the rolling dissolution with increments of nitric that way the temperature dont exceed 70°C (for deeper reduction of nitric and reduce HCl evaporation losses).

So, to get things started, 170 ml of nitric was added and beaker was heated on a plate to temperature of 45°, when clearly the rxn started to go its way. Beaker was placed to catching "pan" and covered. Temperature steadily gone up to 85 °C (next time less acid to start obviously), and fumes had more NOCl colour than NO2 colour. When reaction started to die down, i added 40-50 mls portions of nitric acid to maintain the rolling dissolution. At the end (375 ml total added, required about 6 hours to see temperature not rising after addition), temp was 60 °. I decanted the deeply brown-yellow solution from solids. Some small leftover material (mainly epoxy/resin bottoms with "trapped" pieces of legs inside) was boiled in small beaker with few ml of nitric/HCl for hour or so, and combined with main liquid (when everything dissolved).
Cooling down to room temp, adding sulfamic acid (only 2,5-3 teaspoons required, so not very much nitric/NOCl was unreacted), adding like 3 ml of sulfuric acid just to be sure with lead... Filtering on the Buchner, then dropping gold with sodium sulfite solution.
Gold drop was quite unusual. After adding the sulfite, solution was clear with no obvious ppt (checking with strong LED light) for like 10 minutes, but scratches from stirring rod were golden after like 20 seconds... Everything very slowly started to settling. Very dense ppt. Maybe only the cold solution was causing this, or the composition of the mixture overloaded with iron... Im not sure. Maybe its common, but for me - first time

Tomorrow, i will finish it and get some yield of this, but after all, the process for 240g needed "only" 375 ml of nitric (50%). This was very satisfying observation for me, because my access to nitric is very limited. If the metal content was about 220g (original weight minus plastic bottoms+some silicon), rough molar ratio of metal to nitric was 1:1,06. So clearly, deeper than one electron reduction overall occured. HCl must have also helped a lot with oxidation of iron portion.

 

[orvi]

transistors and ICs, various manufacturers, sizes and gold content

onset of the reaction after first nitric dose (170ml)

clearly golden fumes before adding second portion of nitric

first dose of nitric "decapped" most of the smaller transistors

solid remains of epoxy/resin cast bottoms and silicon chips

resulting solution before filtration and gold drop - deeply yellowish brown colour

 

[vgecas]

I would suggest using sulfuric and nitric acid mix to remove Fe an Ni. Cover transistors with sufuric and add nitric in small quantities, heating is good at first, later - not required.

 

[orvi]

yes, definitely this could be a better solution to remove ferrous metals, but unfortunately my access to sulfuric acid is also limited. so i opted for HCl as its cheap and accessible here.

 

[orvi]

so after drop, nice and heavy precipitate of gold formed. liquid was decanted very easily, precipitate moved to the smaller beaker and firstly washed with water, then boiled in HCl two times, washed with water and dryied on a hotplate.
result is 2.19 grams. yield near 9g/kg, as expected.
i dont know exactly why, but the purity was lowest that i ever achieved, 99.75%. so for the future, one more boil with HCl will be added (if it is AgCl what is the source of silver as inpurity).
for the material of the casing, test was unconclusive because top parts of the transistors (5 that i randomly picked up) are plated/coated with tin layer. unfortunately, i did not have time to sand it off in my few min access to the machine.
so thats it

first wash with HCl


XRF of the gold powder

 

[vgecas]

one can use car battery acid, which essentially is diluted sulfuric. it works fine.
to lower AgCl, AR solution should be cooled and diluted, that way AR is no more capable of dissolving AgCl and it would precipitate.

 

[orvi]

because of the EU regulations, one cannot purchase the battery acid... strange days. they say it is precursor for making explosives and i dont have much batteries left to open and drain. maybe trip to the junkyard will help with this, but spent acid is usually very diluted.
i usually add few ml of H2SO4 at the end to assure Pb precipitation, then cool the solution to room temperature and add bit of ice to it, so it goes down to like 5°.
then I filter it, and the whole point is to minimaize HCl evaporation from liquid when Buchner filtering - nothing precipitates after filtration.
but as i think about it, when dropping gold with sulfite, HCl must be used up to liberate SO2, so maybe this is messing things up...
also when drop yields more "powdery" precipitate, it is usually much more clean than these nice clumps. next time, i will be boiling in HCl for longer and one time extra

 

[orvi]

Mystery solved: after the ability to zapp the sanded casings, these are made of mostly nickel. some Fe in the alloy. But whole casing after pressing the cap in place is dipped in tin. That is why XRF read tin and nothing much else
And also, as the lid is soldered to the "base" of the transistor, decaping these is that painful... Good to know for future

 

[Junior2021]

View attachment 47305
transistores e CIs, vários fabricantes, tamanhos e teor de ouro
View attachment 47306
início da reação após a primeira dose nítrica (170ml)
View attachment 47307
fumos claramente dourados antes de adicionar a segunda porção de nítrico
View attachment 47308
primeira dose de nítrico "descapeado" a maioria dos transistores menores
View attachment 47309
restos sólidos de fundos fundidos epóxi/resina e chips de silício
View attachment 47310
solução resultante antes da filtração e gota de ouro - cor marrom profundamente amarelada

Muito bom essa ideia de digerir o metal de base no nitrico e separar do material nobre. Isso pode ser aplicado em outros materiais do mesmo gênero.