IC chip process without sulfuric acid

[orvi]

Orvi - per the bold print - the fillers in the black epoxy are for the most part FINE ceramics (like alumina powder) so not silica sand

Also the actual IC chip/die is "silicon" - silicon & silica are two different things - silicon will actually alloy with gold whereas silica does not & so is used as a (acidic) flux ingredient for slagging of metal oxides - silica will also dissolve silicon

Therefore - after milling IC chips & because you will end up small/fine particles of silicon in the concentrates you have to actually add some silica to your flux to dissolve & slag that silicon off or it can end up alloyed with your gold

IF (the BIG IF) you get a GOOD HIGH gold concentrate of bond wires you can direct smelt the gold without a collector metal - but - if you use a borax/soda ash flux (with a "bit" of fluorspar for thinning) you will likely end up with a gold/silicon alloy --- adding some actual silica to the flux/smelt will slag the silicon off

Concentrates from N/S bridge BGAs (the black epoxy top only) is a good example of where you can get a GOOD HIGH gold concentrate that can go to direct smelt without a collector

Kurt

I never came across any other filler than silica. In the manufacture of the classic IC chips as we process, silica is very helpful to cut the difference between thermal expansion coefficients of silicon and molded plastic. If there is too much of a difference, cooling and heating of the chips (eg during soldering, or simply dissipating heat when operating) could cause cracks in brittle silicon. Ceramics do not have that favourable properties and manufacture of ceramics in precise shape is also much more expensive.
It isn´t exactly sand, but small silica spheres.

 

[kurtak]

I never came across any other filler than silica. In the manufacture of the classic IC chips as we process, silica is very helpful to cut the difference between thermal expansion coefficients of silicon and molded plastic. If there is too much of a difference, cooling and heating of the chips (eg during soldering, or simply dissipating heat when operating) could cause cracks in brittle silicon. Ceramics do not have that favourable properties and manufacture of ceramics in precise shape is also much more expensive.
It isn´t exactly sand, but small silica spheres.

I don't mean to argue (as I certainly don't know everything) however quite a number of years ago - & because processing IC Chips on a regular basis in relatively large batches I wanted (needed) to know just what was/is in them for safety concerns --- so did my research

Thermal expansion is certainly a concern when it comes to IC packages for the prevention of failure

the fact is that they use a number of different fillers in IC epoxy to address the thermal expansion issue depending on thermal dynamics of the chip (Kovar lead frame, copper lead frame, amount of data the chip is processing etc.)

Here is a discussion you may find interesting about that very thing back in 2016

https://goldrefiningforum.com/threads/ic-powders.22638/page-3
according to my research (back then) alumina & aluminum nitride are two of the more common filers used - other fillers at least mentioned in my research are - silicon carbide, beryllium oxide, & yes silica

Note that my research came from both chip epoxy patents as well as chip manufacture web sites

Bottom line - there are "lots" of possible (if not likely) nasties to be encountered when processing IC chips

Not just the nasties in burning the epoxy - but nasties in the dust of the burned (carbon) epoxy

Kurt

 

[Alondro]

Yes, that is true. Cupelling lead is relatively hazardous procedure. On the other hand, in 180°C sulfuric acid, all the organic junk we want to avoid with proper incineration is evolved to the air. Unless you are wet ashing in the fume hood, this is nowhere safe. Quite a bit of volatile hazardous stuff can be evolved during the process, and I do not mean the stuff arising from the acid, but from the decomposing plastic.

On bigger scale, there are also methods how to scavenge the gold from the lead without direct need for cupelling (similar to Parkes process). And recycling the same lead (with leftover traces of gold) for the next batch. But this is more for the production, not for amateur with two or three few kilo batches.

I agree, that for amateurs, AR leaching of the incinerated remains is much easier than going with smelting. But all big players doing not only the IC chips, but whole trucks worth of PCBs are doing it by means of pyrometallurgy.

That's why I found the process from Researchgate that required heating only to 125C. And to only dissolve silver and palladium, it only requires heating the sulfuric to 80C for several hours to overnight, especially with just plating which will dissolve away quickly.

I burn the chips up first regardless, since it reduces the volume so much, meaning much less acid is needed to cover the material, AND it prevents the formation of the goo that results from sulfuric reacting with some plastics.

It's simple to burn the chips with a scrap steel mesh over a simple wood fire with plenty of air flow, and the sticks and scrap wood cost nothing! Doing the burn first saves on the use of acid and has no expense itself.

Big industrial places have scrubbers in the chimneys to grab the lead fumes. If you need to cupel for testing and assaying material, then it's not too bad. Just keep away from the kiln while the process is taking place and a respirator is needed when you need to approach and peek inside.
I watch how Mount Baker Mining does it quite a bit. I will need to do a cupel assay at some point on some ore samples, and I'll be making sure to do it right with the lowest amount of lead possible.

 

[orvi]

I don't mean to argue (as I certainly don't know everything) however quite a number of years ago - & because processing IC Chips on a regular basis in relatively large batches I wanted (needed) to know just what was/is in them for safety concerns --- so did my research

Thermal expansion is certainly a concern when it comes to IC packages for the prevention of failure

the fact is that they use a number of different fillers in IC epoxy to address the thermal expansion issue depending on thermal dynamics of the chip (Kovar lead frame, copper lead frame, amount of data the chip is processing etc.)

Here is a discussion you may find interesting about that very thing back in 2016

https://goldrefiningforum.com/threads/ic-powders.22638/page-3
according to my research (back then) alumina & aluminum nitride are two of the more common filers used - other fillers at least mentioned in my research are - silicon carbide, beryllium oxide, & yes silica

Note that my research came from both chip epoxy patents as well as chip manufacture web sites

Bottom line - there are "lots" of possible (if not likely) nasties to be encountered when processing IC chips

Not just the nasties in burning the epoxy - but nasties in the dust of the burned (carbon) epoxy

Kurt

I am also just discussing, because I don´t know it either. On the other side, I read a whole bunch of papers regarding mold manufacture of standard plastic ICs. And in all of them, only silica was stated as filler in the chips.

Alumina, aluminium nitride, beryllium oxide and silicon carbide are all used, but in special ceramics compositions for manufacture of quality high frequency chips. They all possess one feature (except plain alumina) - they are a lot more expensive than silica spheres. A lot, mainly size-defined AlN particles and BeO of course.
And they could be used for specialty applications, where heat transfer is needed even through the plastic package. This is not common nowdays, as we now have numerous other packages with heatsinks like SOT or TO packages.

And if it does not need to be special, it will be just silica. Natural pulverized silica, or even precipitated inorganic SiO2 for applications, where even traces of natural uranium and thorium are concerning because of "bit flip".

Below I attach a good small review about the manufacture process of our familiar black ICs, overall very interesting reading. So much work made, and now I just putting them into the furnance

 

[orvi]

I overall do not like sulfuric acid ashing, but it´s personal for me it is just wasting acid. But folks somwhere in India where it costs nothing compared to our prices may be happy with it

Classical litharge fire assay smelt should be OK and there is no need for big ammounts of lead. If lead is a problem, bismuth could be a solution. Similarly as litharge, you can use Bi2O3. But it is harder to find.

 

[Alondro]

I overall do not like sulfuric acid ashing, but it´s personal for me it is just wasting acid. But folks somwhere in India where it costs nothing compared to our prices may be happy with it

Classical litharge fire assay smelt should be OK and there is no need for big ammounts of lead. If lead is a problem, bismuth could be a solution. Similarly as litharge, you can use Bi2O3. But it is harder to find.

I managed to get 4 gallons of conc. sulfuric for $80. That'll last me for a LONG time, since I use it for very specific processes. That's MUCH lower than an equal amount of nitric would cost me, plus it will dissolve roughly double the amount of metal (bivalent vs monovalent anion).

Nitric will be used for silver (in a silver electrolytic cell) and palladium purification. I don't think I'll run into much platinum with the material I have, so that's not really a consideration.

Oh, speaking of chips, I don't know how many people know that the 90's SMD IC transistors (the ones with the solid copper backing) have gold bond wires through the plastic. I ripped a couple off to see what was in them, and one cracked just right to expose a fairly long gold wire! Naturally, I broke up the plastic of several more, and sure enough, they all had the wires.

It's a bit blurry, but you can see the reflection of the gold wire in the sunlight. Given how small these devices are, to see visible gold wires means they're really worth collecting! They have much less plastic material and base metal to remove versus large chips.

 

[Alondro]

Here's what those IC transistors look like on the boards.

 

[kurtak]

Here's what those IC transistors look like on the boards.

View attachment 49775

in as much as "some" of those have gold bond wires - some of them are made with aluminum bond wires

https://goldrefiningforum.com/threa...s-of-ic-chips-flatpacks-and-bga.22951/page-13
So you would need to know the part number of each & every type TO transistor package - test each type to determine which have gold bond wires & which have aluminum wires

Also - there are only two bond wires (which could be gold - or - aluminum) in each of these

to see visible gold wires means they're really worth collecting!

Therefore yield on these is likely to be less then .1 gram per kilo

The copper plate that the silico die is mounted on is likely worth more then then the bond wires

Because there is so little gold in them I never processed then "alone" for the bond wires

However - I did save them - but for the copper not the gold

When I had a couple 5 gallon (20 liter) buckets of them I incinerated them - milled them - then sifted them to separate the copper from the carbon - the carbon (which had very little gold in it) then went in with the carbon from higher grade incinerated IC chips - then run across my concentrator table

So yes - any (small) amount of gold in those was recovered - but in reality it was a "small" addition to processing "much" higher grade material --- in other words - the gold was a "small" by product in actually going for the copper

Personally - these are so low grade (only two bond wires) I would never go after these - in & of them selves - for the gold --- you will in fact need a couple 5 gallon (20 liter) buckets of them to get a "few" grams of gold

Kurt

 

[kurtak]

Alondro - to put it in perspective - read this -----------

https://goldrefiningforum.com/threa...ic-types-of-ic-chips-flatpacks-and-bga.22951/
Scrole down to post 8

Those are PROM type chips - they have many more bond wires in them (then the TO transistor chips/packages)

They (the PROMs) only pay (yield) .245 gram per kilo

So you can't expect much of a yield out of a kilo of chips that "may or not have gold bond wires (some having aluminum bond wires)

Kurt

 

[Geo]

Transistors can only have two bond wires. They are so low yield for gold, you would make a better profit to process them for their copper. I imagine sorting out the bond words couldn't hurt but I certainly would not process them for the gold value. In my opinion, they cost more in chemicals and time than you would make off of them. Mosfets like the ones in the picture have notorious low yields for gold.

 

[Alondro]

in as much as "some" of those have gold bond wires - some of them are made with aluminum bond wires

https://goldrefiningforum.com/threa...s-of-ic-chips-flatpacks-and-bga.22951/page-13
So you would need to know the part number of each & every type TO transistor package - test each type to determine which have gold bond wires & which have aluminum wires

Also - there are only two bond wires (which could be gold - or - aluminum) in each of these



Therefore yield on these is likely to be less then .1 gram per kilo

The copper plate that the silico die is mounted on is likely worth more then then the bond wires

Because there is so little gold in them I never processed then "alone" for the bond wires

However - I did save them - but for the copper not the gold

When I had a couple 5 gallon (20 liter) buckets of them I incinerated them - milled them - then sifted them to separate the copper from the carbon - the carbon (which had very little gold in it) then went in with the carbon from higher grade incinerated IC chips - then run across my concentrator table

So yes - any (small) amount of gold in those was recovered - but in reality it was a "small" addition to processing "much" higher grade material --- in other words - the gold was a "small" by product in actually going for the copper

Personally - these are so low grade (only two bond wires) I would never go after these - in & of them selves - for the gold --- you will in fact need a couple 5 gallon (20 liter) buckets of them to get a "few" grams of gold

Kurt

With these little IC transistors, the plastic is very brittle. It's easy enough for me to crush it off, which generally happens regardless when you wrench them off the boards. I can see the bond wires when I do that and tell at once which kind they are. The copper generally is still stuck the to the boards at that point, then I grab the edge of it and peel it off. So, it's just a matter of burning the plastic away to release the bond wires.

The BIG old IC transistors I'm just saving for copper.

 

[orvi]

Speaking of transistors, some old THT mount classical types in TO92 or larger casings, old manufacture, has gold plated pins under the plastic. Sometimes whole copper heatsink is plated from the inside of the transistor, plus gold bonding wires. More rated current for the transistor, thicker the bonding wires.
But yes... Low yield anyway. It all come to incineration, if you want to recover the copper. And it is fairly easy to separate copper heatsinks from anything else saving the ashes for sluicing or leaching. These would have bit thicker wires compared to most of the IC chips (I assume from larger rated current), so panning/sluicing will be a bit more effective

 

[patnor1011]

If the OP cant get nitric in Vienna all they need to do is to pop in north to slovakia and get few bottles in any hardware shop.

 

[kurtak]

Speaking of transistors, some old THT mount classical types in TO92 or larger casings, old manufacture, has gold plated pins under the plastic. Sometimes whole copper heatsink is plated from the inside of the transistor, plus gold bonding wires.

Per the bold print - true enough - as we know - generally speaking OLD stuff was often manufactured with more PMs then newer stuff

More rated current for the transistor, thicker the bonding wires

Also true enough the 2 bond wires in the MOSFETs are "a bit" longer & "may" be "a bit" thicker - BUT - they still don't add up to the yield of the PROMs I pointed out in the thread posted by Tzoax - "maybe" half (likely less) the yield of PROMs

So lets put that in perspective -----

PROMS run (about) .24 gram gold per kilo - lets call it .25 g (for easy math)

So (about) 4 kilos (or "about" 9 pounds) to recover 1 gram --- (and that is A LOT of chips - meaning you have to de-populate A LOT of boards to get that many chips)

A troy ounce gold = 31.1 grams (call it 31g)

So 31 X 4 = 124 kilos to recover 1ozt

A 5 gallon (20 liter) bucket will hold (about) 18 - 20 kilos of PROMs

That means you would need (about) 6 buckets of PROMs to recover a ozt gold

Now double that to recover a ozt gold from MOSFETs = 12 buckets to recover 1 ozt gold

Now think about how many boards you would need to de-populate just to get 1 bucket chips --- let alone 6 buckets (PROMs) or 12 buckets (MOSFETs)

My point --- when it comes to the PMs in e-waste - some things are worth chasing - other things (though they may well have PMs in them) simply are NOT worth chasing

At least (& for sure) not in & of them selves

Therefore -----------

Given how small these devices are, to see visible gold wires means they're really worth collecting!

Per the bold print - in the case of MOSFETs - the fact of the matter is that they are so LOW GRADE that they most certainly are not worth casing - at least not in & of them selves

Or as Geo said -------

Transistors can only have two bond wires. They are so low yield for gold, you would make a better profit to process them for their copper. I imagine sorting out the bond words couldn't hurt but I certainly would not process them for the gold value. In my opinion, they cost more in chemicals and time than you would make off of them. Mosfets like the ones in the picture have notorious low yields for gold.

To clarify - it is NEVER worth while - "chasing" - this kind of low grade stuff - in & of it's self - but -----

It does hurt if "some" of this LOW grade ends up in the process of casing the actual high grade

Example of how "some" of this "low" grade can end up "mixed in" with you chasing "high" grade ------

If you use a chisel to knock off a high grade chip & there is a MOSFET next to it that also comes off - toss them both in the bucket - BUT - what you DON'T do is go chasing after every MOSFET that is on the board --- they simply are NOT worth it !!!

Kurt

 

[kurtak]

then I grab the edge of it and peel it off. So, it's just a matter of burning the plastic away to release the bond wires.

Alondro --- you will need (about) 6 - 8 kilos of these just to get 1 gram of gold

That is (about) 1/4 - 1/3 of a 20 liter bucket of these

Do you have any Idea of how many circuit boards you would have to pull these off of - just to get that 1 gram of gold ------------

HUNDREDS of boards - & I do mean hundredS - as in likely approaching 1,000 (& maybe even more) boards

For what ? --- a gram of gold ?

For what it is worth Alondro - I used to process around 500 - 600 pounds (230 - 270 kilos) of chips per year

I would NEVER even chase the PROMs (they are so low grade) let alone chase MOSFETs --- they simply are NOT worth it

I chase the HIGH grade - & if in the process "some" low grade ends up in the mix - so be it

Kurt

 

[GoldUser]

I agree

The problem for some people (in some countries) is that can't get nitric acid for making AR & like in the case of the OP they can't even get nitrates (like sodium nitrate or potassium nitrate) for making poor man's AR

Which is why he asked -------



I have always had nitric so I can't really speak to using H2O2 or chlorine (bleach) for dissolving gold bonding wires --- I only know that they will both dissolve gold foils (which are VERY light/thin gold)

Gold bond wires are a much thicker/heavier gold - so I "assume" to dissolve gold bond wires with H2O2 you would need a HIGH concentrate H2O2 --- then (because I believe you would end up with LOTS of free/excess oxidizer) I am not sure how you would go about ridding that oxidizer in order to drop the gold

The problem with chlorine (bleach) is keeping the chlorine in solution long enough & in high enough concentration to dissolve the bond wire - it works fine on foils - but bond wires (being heaver) is a different story

Here is a part of what the OP said that I don't quite understand -----------



It sounds like he has looked into making his own nitric using sulfuric with nitrates (sodium or potassium nitrate) but he can't get some - or any of those things ether ?

GoldUser - is that right ? --- if so - is it only some of that stuff you can't get - or all of it you cant get

Kurt

Hello Kurt,
first thanks for all the work u put in here
I cant get ammonia,potassium,sodium nitrates or any products with could be used for explosives. Its really just HCl/H2O2/Bleach. There is even just one bleach brand that contains it. I looked into making it from air with a lightning generator to break uf the nitrogen & oxygen in the air to make NO2. And further reacts with water to HNO3. It runs on 40W i think. Its from Codyslab:



It seems like its strong enough to dissolve a penny? Could this be used for any good AquaRegia?

Im really disapointed at this point beacause I probably will have to use such methodes.

 

[dpgold]

Incinerate them and you will be fine. On my opinion the sulfuric acid ashing is unnecessarily dangerous and expensive. Burning the chips is much easier and you do not produce gallons of acidic waste.

Key is to incinerate the chips to just ashes - you need to have plenty of oxygen around to make them crispy grey-white, and not grey-black still inside. You must get rid of all carbon if you intend to leach the ashes.

Sifting and panning will recover most of the gold, but certainly not all. It is a nice thing for people who do not want to get involved in "chemicals", just to burn and sluice/pan the material.

is it better to grind after incineration?

 

[dpgold]

I agree

The problem for some people (in some countries) is that can't get nitric acid for making AR & like in the case of the OP they can't even get nitrates (like sodium nitrate or potassium nitrate) for making poor man's AR

Which is why he asked -------



I have always had nitric so I can't really speak to using H2O2 or chlorine (bleach) for dissolving gold bonding wires --- I only know that they will both dissolve gold foils (which are VERY light/thin gold)

Gold bond wires are a much thicker/heavier gold - so I "assume" to dissolve gold bond wires with H2O2 you would need a HIGH concentrate H2O2 --- then (because I believe you would end up with LOTS of free/excess oxidizer) I am not sure how you would go about ridding that oxidizer in order to drop the gold

The problem with chlorine (bleach) is keeping the chlorine in solution long enough & in high enough concentration to dissolve the bond wire - it works fine on foils - but bond wires (being heaver) is a different story

Here is a part of what the OP said that I don't quite understand -----------



It sounds like he has looked into making his own nitric using sulfuric with nitrates (sodium or potassium nitrate) but he can't get some - or any of those things ether ?

GoldUser - is that right ? --- if so - is it only some of that stuff you can't get - or all of it you cant get

Kurt

I used HCL and H2O2 of 3 percent, it is very slow, how do I get rid of oxidizer?

 

[Alondro]

Alondro --- you will need (about) 6 - 8 kilos of these just to get 1 gram of gold

That is (about) 1/4 - 1/3 of a 20 liter bucket of these

Do you have any Idea of how many circuit boards you would have to pull these off of - just to get that 1 gram of gold ------------

HUNDREDS of boards - & I do mean hundredS - as in likely approaching 1,000 (& maybe even more) boards

For what ? --- a gram of gold ?

For what it is worth Alondro - I used to process around 500 - 600 pounds (230 - 270 kilos) of chips per year

I would NEVER even chase the PROMs (they are so low grade) let alone chase MOSFETs --- they simply are NOT worth it

I chase the HIGH grade - & if in the process "some" low grade ends up in the mix - so be it

Kurt

I have about 15 pounds of chips. I'll be processing them all at once. I depopulate boards while watching TV, so it's not like I'm losing time! Rather that sitting there doing nothing, I'm getting materials ready.

I toss the low-grade ones in as well. It takes only a second or two to rip these off the boards, I depopulate everything except resistors and the small glass diodes.

I make it a point to waste as little as possible, especially since I'm dealing with much smaller amounts of material. Like I mentioned, I'm even processing the aluminum canister capacitors for as much aluminum as I can literally squeeze out of them with my cast iron pipe smelting tube and circular brick press units. The pipe cost nothing, just left behind in an empty lot when an old house was torn down. It's long enough for me to fashion 6 crucibles from. 1/4 inch thick cast iron. Excellent crucible material for lower-temp melts. I've melted Al in such tubes for years with just a wood fire in a kiln made from waste brick, which I have piles of recovered from old houses being torn down.

I use pretty much everything except the depopulated boards. They just have too much plastic to burn off.

If you're dealing with MASSIVE amounts of material, then it makes sense to only go for the highest grade or you'll never be able to get through it all. But a little guy like me can take his time and squeeze every drop of metal out. With the aluminum capacitors, LITERALLY squeeze every drop of metal out, lol.

 

[Alondro]

is it better to grind after incineration?

Yes, it's absolutely far easier to grind after incineration. The plastic will be all gone, leaving only ceramics and metal.