Scaling up IC chip processing

[Liquidau]

After a couple of years of study (Hoke and others linked here) and practice at a lab scale, in an equipped industrial lab with hood within an eWaste facility, I’m ramping up to pilot plant processing of around 100 lbs/day of boards. This is yielding about 5 lbs of chips, which are then pyrolyzed at 800C. The residual material is then placed into a cement mixer with 20 steel balls of 1” and ground up to perhaps 30 mesh. I can go finer or coarser.

This is where I’d appreciate specific advice.
Should this material be heated again?
Is it ready for gold cube separation?
Should it all be treated with HCl, washed, Nitric, washed, then AR?
Is the gold found in the connecting wires as well as on the Si dies? Does the gold “stick” to any magnetic or non-ferrous material, or does it mostly get separated out in the cube or blue bowl?

Any tips or suggestions are hugely welcome and appreciated.

Many thanks,
Len

 

[Martijn]

At 30 mesh and black carbon still present, there will be some bonding wires still caught in carbon. Incineration of the carbon to white ash will ensure liberation of all bonding wires.
What exactly did you study and practice? Sounds you still have too much uncertainties to ramp up anything.
Do you have a microscope in the lab? Did you operate a blue bowl before? Have you ever separated bonding wires in water using gravity separation?
Make sure surface tension is broken. Gold can float! Yes, it sure can!
Test and investigate after each step. Change one variable and test for a difference in results.
The ball mill loosens/breaks the gold wires from the pins and dies, but can leave traces behind. A microscope can help a lot.
Bonding wires can clump together in with the pins so check after magnetic separation. Some treat the pins in AP to get the last bit of gold from them.

Martijn.

 

[Liquidau]

At 30 mesh and black carbon still present, there will be some bonding wires still caught in carbon. Incineration of the carbon to white ash will ensure liberation of all bonding wires.
What exactly did you study and practice? Sounds you still have too much uncertainties to ramp up anything.
Do you have a microscope in the lab? Did you operate a blue bowl before? Have you ever separated bonding wires in water using gravity separation?
Make sure surface tension is broken. Gold can float! Yes, it sure can!
Test and investigate after each step. Change one variable and test for a difference in results.
The ball mill loosens/breaks the gold wires from the pins and dies, but can leave traces behind. A microscope can help a lot.
Bonding wires can clump together in with the pins so check after magnetic separation. Some treat the pins in AP to get the last bit of gold from them.

Martijn.

Thanks Martijn, I’ve been reading your replies for a while and appreciate your experience. I’ve studied Hoke and this Forum a lot but know there’s always so much more to learn.
I pyrolysed at 800C and can go higher. I understand that any black residue must be carbon which traps bonding wires. I have a 30X loupe but no microscope. I will buy one.
I have operated the blue bowl and now a gold cube but it sounds like these are more useful in mining operations for gold particles/dust and not bonding wires, especially with all of the ferrous wires also present. I use JetDri to break surface tension.
After crushing and grinding and sieving, is it not just easier to acid treat, wash, and then AR in order to chemically capture all gold?

 

[FrugalRefiner]

Remember that carbon will adsorb gold in solution. It is used in some cases to remove gold from solution. So if you have carbon present when you dissolve your gold in AR, the carbon will steal some of your gold. If you still see carbon after pyrolization, follow with incineration with plenty of oxygen present to fully convert the carbon to ash.

Dave

 

[Martijn]

Thanks Martijn, I’ve been reading your replies for a while and appreciate your experience. I’ve studied Hoke and this Forum a lot but know there’s always so much more to learn.

Thanks. I have a bit of experience with some processes. Compared to the pro's here i think i'm at level 3 on a scale of 10. Maybe.
The most I've learned is from theory here on the forum and searching answers to questions new members ask.
I've built a funace and pyrolysing tank myself but the afterburner still needs work.
The first test with 50 gram of chips still gave off some smoke.
For me it's a great hobby and keeps my mind busy.

Back on topic, you can get the most gold out by carefull incineration, milling, sieving, magnetic separation and panning. The amount of gold still left on the pins wil be very small.
If you have large quantities, you can test a kilo of 'cleaned' pins in acid, i would go for AP, and see whats left.

Then decide if it's worth chasing, considering the amount of toxic waste produced, acids and resources used, risk exposure and the neccessary control measures if you have employees and hours put in, including treating the waste.
The base metal scrap value might be a lot more than the profit to be made on that tiny bit of gold.

By the way, I bought a usb microscope on internet for less than 20 euro and it works great.

 

[samuel-a]

Hi,

Maybe it's just a figure of speech, but there's a difference.
Scaling up usually mean you have the pilot process/plant up running flawlessly with very little unknowns.

From your post, it seems that you are in a point where you are done with your laboratory scale processes. However, your questions would suggest otherwise.

Is this is intended to be a money making process?
Regardless of the answer above, scaling up to a physical separation and hydrometallurgical process is counterintuitive (for me at least).

 

[vgecas]

i would suggest to do couple test runs and calculate, with bigger amount of stuff maybe it is worth to go after silver also?

 

[nickvc]

Len my advice would be as others have said which is to remove all the carbon before you do anything else , the next step should be removing the ash which can be done using water and dish soap leaving all the heavier material .
Separation of the magnetIcs could be problematic as using the steel balls may well crush some of your values onto and in between the other metals , here your new microscope should give you an answer as to whether the amount lost is worth further effort on your part , it may pay to save it and ship once you have decent volumes for others to process, your next step is to run your material on your cube checking the tailings to ensure no values are escaping, once you then have clean material the choice is then whether to refine or simply melt your material.

 

[Geo]

After incineration and milling, gravity and magnetic separation will remove nearly all of the iron and carbon and most of the silica. At this point, I leach copper, silver and any small amount of iron out that the magnet didn't pull. I have tried on a small scale to concentrate the gold even further by adding lead and melting the lead to a liquid state. The lead acts as a collector and will absorb any gold and left over copper, if there is any left. The lead can be dissolved in HCl and recovered by adding a small amount of sulfuric acid and collected as lead sulfate. This can be smelted at relatively low temps with charcoal and sodium carbonate to recover lead metal to re-use in the process. The gold will be unaffected by the dissolution of lead and will remain behind free of carbon and silica.

 

[speed]

Personally i wonder if pyrolise-wet crush-gravity wash-dry-magnetic separation-smelt-copper cell is the way to go rather than messing with shaker tables or whatever? Usually the non magnetics will end up melted for copper anyway so is there any point in separating out bond wires to process separately? also if there is carbon left in the mix will it not be oxidized out during the smelt?

Im not particularly knowledgeable with smelting processes but from an engineering standpoint it seems that a rather large and expensive part of the process can be excluded.

 

[Geo]

Personally i wonder if pyrolise-wet crush-gravity wash-dry-magnetic separation-smelt-copper cell is the way to go rather than messing with shaker tables or whatever? Usually the non magnetics will end up melted for copper anyway so is there any point in separating out bond wires to process separately? also if there is carbon left in the mix will it not be oxidized out during the smelt?

Im not particularly knowledgeable with smelting processes but from an engineering standpoint it seems that a rather large and expensive part of the process can be excluded.

Ideally, you would want something a little faster than that. A copper parting cell is the preferred method of refining copper. In large operations, copper is the target metal and gold (and other precious metals) is considered a byproduct of copper refining. The trade off for using lead is speed. Also, lead melts at a much lower temp. Instead of melting copper and gold together, Melting lead can be accomplished with a burner from a gas grill or camp stove. When the lead melts, everything besides the gold will float. The lead picks up any gold and absorbs it. Turn off the heat and cast into the smallest area possible. The lead will harden and the ash and silica can be removed. You can use the same lead to absorb as much gold as possible before processing the lead to recover the gold.

As an aside, the lead can be cupelled if there is no interest on recovering the lead. It can still be smelted out but the CaO makes it more difficult.

 

[kurtak]

The residual material is then placed into a cement mixer with 20 steel balls of 1” and ground up to perhaps 30 mesh. I can go finer or coarser.

First of all you NEED to mill it MUCH finer then 30 mesh --- it NEEDS to be at least 80 mesh &/or finer

Also - you need to run more then just one size balls in your "ball charge" --- 1 inch balls alone are not big enough &/or heavy enough to do "effective" crushing - they are good for grinding

you want a mixed ball charge of 2 inch balls 1 & 1/4 balls & 1 inch balls

Also - when running a cement mixer for a ball mill - you need to run the angle of the mixer at different angles at different stages of the milling depending on what you want the mill to do - crushing (to start with) or grinding (at the end)

Here is one thread that I posted some info about processing chips in --- go through the thread & look for things I posted

https://goldrefiningforum.com/threads/pyrolysis-reactor.22581/#post-237121
I will "try" (no promise) to find some more things (links) about processing chips this weekend

Kurt

 

[kurtak]

I posted in this thread how to use a cement mixer for milling IC chips

https://goldrefiningforum.com/threads/new-process-for-separating-ash-from-bond-wires.25489/#
Kurt

 

[Liquidau]

Thank you all, guys, for your input!! Kurt, Géo, Speed, Nick, Samuel, Martijn—-I have gleaned your wisdom and experience. Permit me to expand on my situation.
After perfecting the refining process (on fingers and some ceramic CPUs—-yes Kurt, the east low hanging fruit!) I’m now recovering gold, silver and palladium from pcbs and RAM card ICs. My lab is well equipped and within an eWaste recycling centre, so I have access to thousands of pounds of boards at cost. The plan is to get a pilot plant operation running smoothly (100-200 lbs/week of boards generating an ounce or two). Once up and running, I have the space and resources to double and then hopefully double again. So not really industrial (certainly not refinery level) but definitely next level beyond hobbyist.
Thank you for the link to the (37)! page thread detailing the incineration/grinding/sieving/flotation, magnetics, and hydrometallurgical treatment. Combined with all of the recent messages, I’m more confident in now being able to efficiently do IC chip extraction. Key takeaways that others can learn:
Eliminate all carbon! Pyrolyse and incinerate to burn it all away because carbon steals gold!
Pulvérise to 80 mesh or smaller.
Get a microscope to see and study your material, and to verify and measure the amount of bonding wires.
Consider lead leaching in a larger operation.
And read, reread, and print out and read again the wisdom in this Forum!

 

[Martijn]

Geo, very interesting idea to collect with lead. I have some questions. Do you know at what temerature the gold forms an alloy with the lead, or is that the idea? Like inquarting?
I've read the thread
Thread 'Molten Lead Panning??' Molten Lead Panning??
Given the fact of cupelling i think gold is completely soluble in lead so you could collect quite a bit, but the HCL may not reach all the lead at higher purity, like silver in gold needs 6K to get it all out.

 

[Geo]

Geo, very interesting idea to collect with lead. I have some questions. Do you know at what temerature the gold forms an alloy with the lead, or is that the idea? Like inquarting?
I've read the thread
Thread 'Molten Lead Panning??' Molten Lead Panning??
Given the fact of cupelling i think gold is completely soluble in lead so you could collect quite a bit, but the HCL may not reach all the lead at higher purity, like silver in gold needs 6K to get it all out.

As soon as lead becomes liquid. Most liquid metals acts as a "solvent" for other metals and will dissolve metal much like acid does. It forms a molecular bond just by the atoms being in contact with each other. It also does the same in solid state but at a very, very slow rate. An experiment, that I can't remember the details to, was conducted hundreds of years ago to see what would happen if pure gold and pure silver were placed in close proximity in a vacuum. The two samples were certified and placed in the chamber in close proximity, but not touching. The experiment lasted years. When the two samples were removed and assayed, each was contaminated by the other. That means that each metal had migrated to the other without the benefit of a physical bridge. It was a vacuum, so no atmosphere. The metal had migrated atom by atom over a span without benefit of a bridge. This is an example of how metals are attracted to one another.

 

[orvi]

certainly agree with Geo. on few hundred lb/kg scale, hydrometallurgy processing is completely impractical. mostly, the properly incinerated chips are by some means pulverized, screened to preferred mesh to liberate most of the bonding wire (which is then collected by molten lead or directly smelted), and then metal parts could be easily recovered by gravitational separation (shaker table most easily). Magnetic and non-magnetic metal parts are separated, mostly copper non magnetics are melted and refined electrochemically to squeeze last bit of precious metals and get mostly pure copper. with magnetic stuff, pathways vary

for big operation lead is practical, and if proper ventilation/PPE/scrubbers are in place, it is relatively safe to cupell the lead out despite fumes of PbO. on the other hand, there is possibility of doing exactly the same thing with bismuth.
bismuth isn´t nearly as toxic as lead, melt in the similar temperature range and dissolve silver/gold very well. but is far more expensive recycling of used cupels is the option, Bi2O3 is very easily reduced when smelted, but this add one extra thing to already relatively time consuming routine.

 

[Geo]

I've considered using tin metal as a collector and then electro-refine the tin in an electrolytic cell. HCl would be the electrolyte.

 

[kurtak]

I've considered using tin metal as a collector and then electro-refine the tin in an electrolytic cell. HCl would be the electrolyte.

The problems (there are two) with running tin in a parting cell (like a silver or copper cell) is tin whiskers

the reason they are called whiskers is that they grow VERY long & VERY fast like whiskers on a cat

because of the short distance between anode & cathode (about 4 inches) the whiskers will grow very quickly between anode & cathode & short the cell out --- so you have to knock the whiskers down VERY OFTEN

I watched a video (like 6 or 7 years ago) where some students played with the idea (tin cell) --- in order to overcome the FAST whisker growth problem they set up a roller that would automatically roll over the cathode to knock down the whiskers - the roller would have to knock the whiskers down about every 10 minutes

In other words - unless you have something set up to knock the whiskers down automatically & often - unlike a silver or copper cell - you can't just turn on the power supply & walk away - you have to stay on top of the cell to knock the whiskers down OFTEN

The other problem is the anode slimes --- as you know - with a copper or silver cell the "other" PMs (gold &/or PGMs) end up in the slimes which then need to be treated to recover those "other" PMs

in a copper or silver cell those slimes will be MOSTLY copper or silver (plus/minus 80%) so you still need to part the copper/silver from the "other" PMs --- that in it's self is not an easy task --- there is a reason Nick "often" refers to them as swine's (instead of slims) they are a pig to deal with

You would/will have the same problem with a tin cell - tin slimes to part your gold/PGMs from

Slime can be a real pain in the butt especially if other metals are involved (like lead, iron etc. etc. etc.)

If/when "other" metals are involved it can/will take "multiple" treatments to part the PMs from the slimes

Generally speaking - anodes from a smelt have other metals involved - other then the "target" metals (silver/gold - copper/gold or in this case tin/gold) which is why processing slimes are OFTEN a pain in the butt (or swine's)

There is the old saying --- tin is a refiners nightmare - IMO - cleaning up the slimes from a tin cell would be nothing short of a nightmare

If your dore metal (from the smelt) was ONLY tin & gold - then why not just dissolve the tin away with HCl & let the gold settle --- if other metals are involved the tin is going to be a problem whether you run it in a cell or run it through otherwise normal processes

Not saying a tin cell wouldn't work --- just that you will have the same problem with the slimes as you would have processing dore metal (with tin from a smelt) by other normal processes --- parting PMs from other metals including tin

It might be a fun experiment but a don't think there would be any advantage over already proven methods - more likely additional problems

IMO

Kurt

 

[kurtak]

The plan is to get a pilot plant operation running smoothly (100-200 lbs/week of boards generating an ounce or two).

First of all - unless you are processing nothing but some of the very highest grade of boards (like tele comm - military - aerospace or medical boards) you are NOT going to see an ounce - much let alone two ounces from 100 - 200 pounds of run of the mill average everyday "consumer product" boards

Example; - it takes "about" one ton (2000 pounds) of "modern" day mother boards to get "about" 3.5 to 4 OZT gold

We are talking about whole boards here - so 100 - 200 pounds of boards per week (unless they are VERY high grade boards) is NOT going to pay the bills let alone make a profit

IC chips - there are two "basic" types - BGAs & then all others (excluding ceramics which are there very own type chip)

BGAs are the highest grade of IC chips which can run from 2 - 4 grams per pound --- all the other chip are going to run less then a gram per pound - some of them MUCH less then a gram per pound

So you should process BGAs separate from & on there own from all the other chips --- all the other chip can & should be processed together - call it a "mixed batch" of IC chips - other then BGAs

It will take about 50 pounds of "mixed" IC chips (other then BGAs) to get "about" (plus/minus) 1 OZT gold

Older boards will have FAR more chips on them then newer boards

That is because one chip on a modern board can process MANY times more information then one chip from an older board --- so an older board had to have MORE chips on it to process the same amount of information than just a few chips on a modern board

100 - 200 pound of OLD boards will give you (on average - about) 5 - 10 ("maybe" 12 - 13) pounds of mixed IC chips) & likely no BGAs

100 -200 pounds of MODERN boards is only going to give you (on average - about) 3 - "maybe" 5 pounds of IC chips - but - that will include BGAs as well as "other" mixed chips

Do the math - how long is it actually going to take you & how many pounds of boards do you actually need to get the 50 pounds of chips you need to get that 1 OZT of gold

That is - when talking about average run of the mill everyday consumer product boards - there are of course higher grade boards (as mentioned above)

So - from there - you can (somewhat) figure how many pounds of boards you "may" actually need to get an ounce of gold from "boards" let alone from IC Chips

If you are talking about 100 - 200 pounds of NOTHING but HIGH grade boards - that is a different story

Also - concerning IC chips - you are NOT going to find PGMs &/or silver in them --- unless they are the ceramic type chips - in which case the PGMS &/or silver is found in the braze for brazing the pins & Kovar metal lids to the ceramic IC chips

Kurt