Gold recovery from smelter slags

[Deano]

The gold in the slag is there because the metalloid coating prevents the beads from coalescing into the smelt pool.
If you separate the beads from the slag and smelt them you will have exactly the same conditions which got those beads into the slag in the first place.
Best of luck in trying to assay the slag by conventional methods.
The coatings prevent the beads entering the lead pool in a fire assay and they stop the beads dissolving in any form of wet chemistry type assay.
I did a major project on gold assaying of beads in slag for a company which was interested in such.
The bulk of the work was carried out on slag generated from the smelting of gold resulting from stripping of carbon in a commercial CIP stripping plant.
The slag was jaw crushed and passed over a 2mm screen, gold pieces larger than 2mm were picked from the screen and all the rest of the slag was re-milled to less than 2mm.
Both fire assay and wet chemistry showed values of around 130ppm gold in the minus 2mm slag on multiple subsamples of the slag.
Subsamples of the minus 2mm slag were then tabled with the average gold recovered from the gold line on the table weighing in at around 1500ppm.
Subsamples of the table mids and tails were then run through 24 hour bottle rolls in dilute leaches with high cyanide levels in the bottles.
The average gold recovered from these bottle rolls was around 1100ppm.
This meant that the gold recovered by gravity and leaching was 1500 + 1100 = 2600ppm, this compares with the 130ppm assay values.
Thus 95% of the gold did not report to standard assay methods but was recovered by gravity and modified bottle roll.
Similar results were gained for slags from e-waste smelting and other sources when run through the same processes.
If the gravity gold was smelted as a separate operation the bulk of the gold prills did not report to the gold pool, they had distributed themselves through the slag.
If the prills were given the HCl treatment then virtually all of the prills reported to the gold pool on smelting.
Deano

 

[4metals]

The slag was jaw crushed and passed over a 2mm screen, gold pieces larger than 2mm were picked from the screen and all the rest of the slag was re-milled to less than 2mm.

I think the effects of the metalloids we are discussing here has different effects based on a lot of possible factors but most prominent is size. Deano is sizing his fractions down to minus 2 mm. All of my experience with crushed fractions happened with much smaller fractions. In the US, sampling of the prepared powders requires the incoming material be at least minus 40 mesh. (Here in the US, we go by a US Standard sieve series based on that archaic standard of inches.) Be that as it may, a particle passing through a 2 mm mesh screen is 4.7 times larger than a sample passing through a 2 mm sieve. I always crushed to pass through a 60-mesh screen (8 times smaller than a 2 mm sieve) and when samples of higher values were shipped for sampling, they were often sieved for 100-mesh. (16 times smaller than a 2 mm sieve). Often the samples needed additional processing in a shatter mill to get them to pass the 100-mesh screen. The reason for smaller particles is reflected in the assay. A sample made up of smaller particles more accurately reflects the actual composition of the materials being assayed.

Fire assays for powdered materials are processed in a 2-step process, first a fusion in a litharge-based flux, followed by a cupellation of the resulting lead cone. The great difference between simply re-melting slags and fire assaying slags (or anything for that matter) is the flux used in a fire assay has lead in it, which provides metal to serve as a collector. Maybe that is why this speak of metalloid coatings comes off as strange to me because the collector metal in the flux was able to break up this metalloid coating and allow the values to report to the molten pool. The thought that a metalloid coating could prevent a proper fusion and allow values to escape detection quite honestly can throw the entire refining industry into disarray.

While I am not disputing the entire metalloid claim, it is my conviction that the effect of said metalloid coating is directly proportional to particle size. The smaller the particle size, the greater the effect. Possibly this metalloid effect is par with quantum mechanics, which really goes un-noticed in the everyday world but lurks in the sub-atomic world!

I’ve put all of this down without the benefit of my pile of books to fact check any of the things that just flowed out of my aging brain. But to go further and delve into the metalloid question I’d like to review some of the concepts floating about my brain now, so I don’t put my foot into my mouth, sort of fact checking before the fact. (Quite unusual in today’s world but it works for me.)

 

[Deano]

I was writing the assay preparation method on the basis of assumed knowledge, so let me expand on how sample preparation was conducted.
Before any fire or wet chemical assay is conducted the ore sample is milled to, generally, 100% passing through a 100 micron or 150 micron screen.
Most milling is done in a ring mill but other milling types can be employed.
If the sample has coarse gold particles present then these will become smeared through the sample and also form a contaminating surface in the mill itself.
This requires passing cleaning samples of barren ore through the mill between each sample, this is a major pain.
By removing the coarser gold particles before milling occurs the degree of smearing is minimised.
Most of the coarse particles are liberated by jaw crushing to 2mm and can be separated from the slag by manual picking and tabling.
The minus 2mm table tails are then ring milled after drying.
That gold is present in slag is attested to by firms who buy and retreat slag of all types, the retreatment procedures are confidential to these firms.
Deano

 

[4metals]

Let me start for thanking you for the detail of your posts. Some threads can suffice with a simple explanation, and some are better explained with details. Your posts typically come with details. This particular topic, which ends up here discussing slags and analytics, is one where members may not have experienced the methods we are discussing, so details are good because members can learn something they may not know. We never stop learning and I, for one, believe you can teach an old horse new tricks! Again, thank you.

I believe the ring mill you mention is just another name for what I referred to as a shatter mill, it is also referred to as a puck mill. A series of heavy cylinders surrounding a heavy puck with spaces between the cylinders for sample which is subjected to a base that causes the container the cylinders and puck sit in to slide back and forth causing collisions in the spaces and effectively crushing the samples. Cleaning is very important as you mention but the samples I processed were usually minus 60 to start so I just made it a habit to process a cup full of assay sand to clean it every 5th run.

It is interesting to note that, as you mentioned, there are companies that make a living buying slags to recover what was left behind. There is always something left behind, we never get it all, but we strive to get as much as is economical to recover.

When I smelted jewelers sweeps in quantity the slags were remelted once and then they were shipped to a copper smelter who took them off of our hands for free. We usually shipped 2 tons at a clip and our assay indicated we were always under 0.25 OPT. Since the copper smelter electrolytically refined his copper, he got the gold from his standard process and benefited from the “free flux” we gave him.

This begs the question “how anyone can process 2 tons of anything, including shipping, for 1 ounce of gold. And that was some time ago so $300 gold! And now, years later, you mention what I’ll call the metalloid factor and it leaves me scratching the part of my head that once had hair saying, “what did we miss?”

I would like to think we missed very little, as our in-house assays always jived with our recovery, but it makes me want to discuss and possibly figure out just what the metalloid factor is and when it is an issue.

Classically metalloids are a group of elements (typically thought to be 7 of them) that exhibit properties somewhere between metals and non-metals. The catch is when they exhibit these properties and when they don’t. The most abundant of these troublemakers on the planet is silicone and boron is also on the list. Hmmmmm two things I intentionally add to my flux recipe for fusion assays.

More head scratching while Deano sits whispering in my ear.
“Gold which is present in slag is there because it has a metalloid coating on the gold particles.”

Maybe we can get to some understanding of the effect if I describe some of the materials I have smelted and assayed which gave me problems until I could work out the best flux formulation.

Hands down the hardest material to produce a good fusion with were prepared green rouge sweeps. These were made up from jewelers polishing sweeps which were polished with green rouge. The green rouge contained chromium and even after the material was incinerated crushed and sieved to minus 100 mesh it was a dull green in color and, if fused with a typical flux from a red rouge sweep the fusion was never complete so a special flux was prepared just for green rouge sweeps. When a fusion is poured an assayer can get a good idea about how well the fusion went because any beads remaining in the crucible are very visible as bright glowing points against the orange glow of the crucible. I never thought of the possibility of the actual gold beads having a metalloid coating preventing, or hindering coalescing into the pool.

Have you ever had similar situations where the metalloid coating was worse than others? Is that attributable to other metals in the mix or the concentration of the boron or the silicone?

Whenever I had a pool of metal, be it gold alloy from karat refining, or copper-based bullion from e-scrap smelting, I never had losses due to excessive bead formation. Do you think this is from the effect of the size of the metal in the pool? Even in fire assays, the size of the lead pool vs the sample size is considerable. Where I’m going with this is figuring out if there are situations, be it within the feedstock or in the flux additions, that can make this condition worse?

Is it possible your work with CIP carbons produces this effect more than other processes? My experience with CIP is limited but yours is not. I wish a few more people with experience would chime in here. More eyes on the same question always produces a better discussion.

 

[Shark]

I wish a few more people with experience would chime in here. More eyes on the same question always produces a better discussion.

I for one am enjoying following along. Seeing more people involved would be even better as this is an area I know practically nothing about.

 

[goldshark]

Just curious as to the average particle size of the uncoalesced beads are. It sounds to me like possibly finishing the crush/impact particle reduction process with possibly a very close tolerance rolling mill. Final recovery would be via a screen to sift out oversize, flattened particles. Just a thought though. Never seen such a set of rolls for such fine crushing, but doesn't mean it couldn't be built.

 

[Deano]

The worst coating offender was lanarkite, it showed up on the XRF as sulfur deficient lead sulfate. There are no known published quantitative solvents for this material despite claims to the contrary.
I did find that a long term HCl leach would slowly remove this coating to the degree that normal gold solvents would dissolve the gold but I suspect that the coating was not totally removed, just enough coating was removed to allow leach penetration through weakened areas.
The most problematic coatings which produced the largest prills were from ores treated as CIP, many of these prills weighed in excess of 5 grams.
It is no secret in the gold industry that these prills exist, the usual treatment method is to throw the slag into the milling circuit as a quick and easy remedy.
My test work demonstrated that remilling these prills had the effect of turning them into much smaller prills with the same coatings but much less visible as being a problem.
The cleaner the source material the smaller the prills, I put this down to the lower levels of base metals available to form coatings as all the coating materials had to be sourced from the flux components without any input from the gold source.
Many of the prills were at or below detection limit by optical means, the only simple method of demonstrating their presence was by the bottle roll method.
The best recoveries from bottle rolling were from the minus 2mm slag, despite adding ball made from steel, ceramics or glass I could not get good recoveries if I milled the ore finer.
It appeared that the weight of the particles gave an abrasive effect which would momentarily puncture the coatings and allow leaching of the gold to occur.
The coatings appeared to be self healing because if the cyanide level was low even extended leach times would not give good recoveries, you needed to have a lot of solubilising agent available to take advantage of any short term coating weakness.
Deano

 

[4metals]

I think we are getting somewhere. First let me state a few assumptions I’ve made from the thread. The gold from ore is processed, milled, and leached in cyanide. The gold cyanide is passed through a column of activated carbon which traps the gold cyanide in any of its many crevices until the carbon is loaded with the gold.

Next, the gold is eluted from the activated charcoal using hot caustic cyanide which re-dissolves the gold and is recovered from the concentrate. The carbon, because some of its deeper crevices, holds on to some of the gold but it is still useful to absorb more gold cyanide from the leach field. The carbon is returned to the owner and the process repeats.

I do not know what percentage of the gold the carbon traps, but I assume after multiple uses (whatever that number is I do not know) it is holding more gold than the owner cares to return to the process and a new carbon is used. It can be an entire thread as to how the activated carbon works and the logic to determine when to elute and return vs elute and recover the remaining gold.

The gold eluted from the carbon is recovered by electrowinning or some other process and now the issue is the gold trapped in the little crevices of the carbon.

Since we are discussing values as beads being trapped in the slags from processing the exhausted CIP, it is here that the problem is born.

And just when we thought there was progress, Deano reveals that it not only is a metalloid, but it is also an angstrom thick self-healing metalloid coating.

Somehow, from a feedstock of gold (possibly silver and copper too) dissolved in cyanide and trapped in the tiny pores of activated carbon, a fluxed melt generates this metalloid coating which envelopes the gold particles. I’m sure it’s not all of them but enough for it to be noticed.

Fortunately, Deano has determined that at least one culprit is Lanarkite. Probably one of many causing this issue but, starting with this one, Lanarkite, we can discuss its effect and maybe it will lead to other culprits in other feed streams. If nothing else, we are learning something new here. Maybe not new to gold refiners down under, but their input is also welcomed here.

Lanarkite, PbSO4.PbO, Lead sulfate PbSO4 is called Anglesite. Both of these compounds melt below 1200ºC. What this says to me is even if this ominous self-healing metalloid exists and is coating all of the gold particles in a fire assay, the lead pool and fire assay temperature will most likely dissolve the Sulfate fraction and render the lead to the pool and allow the gold to be detected in a fire assay. Even the smaller particles Deano crushed beyond detection before assay will surrender themselves to a fire assay. That is a sigh of relief for me!

Still, if these smaller coated particles resist typical recovery methods, we still, from a refining standpoint, have an issue.

Something in the flux, combined with lead will form this crystal (Larkanite is a crystal) which make sense because in the right conditions it will grow. And growing, in this case, is what a self healing coating is doing.

Well, my day is complete knowing that at least a fire assay can tame this metalloid. More thinking and details about processing will get us closer to the answer. I think I'll go down to the pond and catch a catfish for dinner, great way to celebrate the end of summer.

 

[Deano]

I have obviously not explained the origin of the gold well enough.
The problem is with gold retained in carbon which is ashed for recovery as well as gold recovered by electrowining as part of the standard carbon stripping process.
When the slag from smelting the electrowon gold is treated by the process I developed you will recover gold beads which do not report into either fire or wet chemical assays.
Many of these beads report to a jaw crushed plus 2mm fraction but most report to the minus 2mm fraction.
The beads can, dependent on size, be separated from the slag by manual picking off a 2mm screen or by tabling the minus 2mm screen undersize.
These separated beads are, in the main, resistant to leaching by cyanide or aqua regia solutions.
They are also resistant to being fire assayed and report to the fire assay slag.
The table tails also contain non-assayable gold particles which are retained in the slag particles, most of these particles are not locatable by optical methods. The gold in these particles is only registered by using the bottle roll technique, they do not report to fire or wet chemistry methods.
Deano

 

[fiquett11]

I think the effects of the metalloids we are discussing here has different effects based on a lot of possible factors but most prominent is size. Deano is sizing his fractions down to minus 2 mm. All of my experience with crushed fractions happened with much smaller fractions. In the US, sampling of the prepared powders requires the incoming material be at least minus 40 mesh. (Here in the US, we go by a US Standard sieve series based on that archaic standard of inches.) Be that as it may, a particle passing through a 2 mm mesh screen is 4.7 times larger than a sample passing through a 2 mm sieve. I always crushed to pass through a 60-mesh screen (8 times smaller than a 2 mm sieve) and when samples of higher values were shipped for sampling, they were often sieved for 100-mesh. (16 times smaller than a 2 mm sieve). Often the samples needed additional processing in a shatter mill to get them to pass the 100-mesh screen. The reason for smaller particles is reflected in the assay. A sample made up of smaller particles more accurately reflects the actual composition of the materials being assayed.

Fire assays for powdered materials are processed in a 2-step process, first a fusion in a litharge-based flux, followed by a cupellation of the resulting lead cone. The great difference between simply re-melting slags and fire assaying slags (or anything for that matter) is the flux used in a fire assay has lead in it, which provides metal to serve as a collector. Maybe that is why this speak of metalloid coatings comes off as strange to me because the collector metal in the flux was able to break up this metalloid coating and allow the values to report to the molten pool. The thought that a metalloid coating could prevent a proper fusion and allow values to escape detection quite honestly can throw the entire refining industry into disarray.

While I am not disputing the entire metalloid claim, it is my conviction that the effect of said metalloid coating is directly proportional to particle size. The smaller the particle size, the greater the effect. Possibly this metalloid effect is par with quantum mechanics, which really goes un-noticed in the everyday world but lurks in the sub-atomic world!

I’ve put all of this down without the benefit of my pile of books to fact check any of the things that just flowed out of my aging brain. But to go further and delve into the metalloid question I’d like to review some of the concepts floating about my brain now, so I don’t put my foot into my mouth, sort of fact checking before the fact. (Quite unusual in today’s world but it works for me.)

That's exactly what I'm having to learn because I can't get the small bb's of gold to meet up together but seems like no matter what proportion use, I get to the same point in melting with all kinds of flux : flurospar, soda ash, borax, anthracite, wood ash and petroleum, even. Lead of various sizes (currand the ratio I've experimented with is still the same result. Not any of the Gold in my experience came from old tech parts or jewelry i have been working with natural gold and excavated areas with heavy oxides and either way the same thing happens with gold needing further melting and flux pretty much not really helpful any experience moving forward?

 

[Deano]

Your gold has coatings on it and will not coalesce until it has been cleaned with a HCl leach.
Separate the gold pills from the slag by crushing and panning and then give them a leach in 20% HCl.
It can take up to a month or so to remove heavy coatings, heating the leach speeds up the process.
Deano

 

[4metals]

While I have never experienced this phenomenon, I have always worked with e-scrap or karat gold as both sweeps and bullion. But I have no doubt it is an issue.

The OP was dealing with beads in karat gold bars and I do not think (or I have never seen) beads that refused to coalesce. Beads that hung up yes, but all were coaxable by either fluxing or collector metals.

Fiquette above is having the issue with ore and it seems to be the same as the situation Deano describes. Was the gold mixed with heavy oxides leached to concentrate on activated carbon?
If the beads can be separated I have no doubt Deano's suggestion is valid and can solve your problem. I am trying to get to the point where we can determine where this metalloid coating came into the equation.

The problem is with gold retained in carbon which is ashed for recovery as well as gold recovered by electrowining as part of the standard carbon stripping process.

The beads recovered from ashing the end of life activated carbon can cause this issue but now gold electrowinned from the leach from activated carbon is also mentioned. Are these concentrates electrowinned on to a copper cathode for recovery or are they electrowinned on to a carbon cathode?

I have processed a lot of cyanide stripped e-scrap and plated it out on to a copper cathode, this is refined in aqua regia or in a copper cell depending on which equipment is available. But I have never employed a carbon cathode to concentrate the values.

I remember a while back Jon (Anachronism) built a device to electrowin gold from cyanide which I believe Deano designed and posted on the forum. He was using a carbon felt cathode to collect his values. Is this the type of cathode used to recover the gold that we are discussing here which has a tendency not to coalesce?

 

[Deano]

The standard method of electrowinning in the gold mining industry is to use steel wool as a cathode, it allows you a relatively large surface area but with the ability to load high metal levels without blocking the strip[ liquor flow.
Deano

 

[4metals]

The standard method of electrowinning in the gold mining industry is to use steel wool as a cathode

The solution used to elute the gold from the activated carbon is passed through a steel wool cathode (certainly economical) then, once the gold has all been electrowinned to the steel wool, the well rinsed cathode can be either (in an exhaust hood) treated with Hydrochloric Acid to leach away a good amount of the iron leaving just electrowinned gold and possibly silver as well. Then the filtered insolubles are digested in aqua regia normally. One could forego the Hydrochloric Acid iron leach as well, not sure what would be more effective.

After the gold has been removed from the aqua regia by Sulfur Dioxide or Metabisulfite, the rinsed gold is melted.

as well as gold recovered by electrowining as part of the standard carbon stripping process.

Is this gold also subject to the hung up beads? Have I listed the process accurately? At this point the gold dropped from aqua regia is not unlike refined karat and, in my experience, does not suffer from hung up beads.

I get it that gold from the CIP process recovered by ashing causes these beads of gold to remain in slags, I'm just not sure why. But the gold electrowinned on to steel wool and refined behaves the same way? Just trying to understand what causes the problem and what does not.

 

[Deano]

The cathode steel wool and attached gold, silver and copper as well as any other base metals reduced in the electrowin step are digested in HCl.
After the digestion step the remnants are dried and smelted with various fluxes, each mine has its own flux formulation and the same problem with bead formation.
The beads report to the smelting slag.
No aqua regia is used to digest the HCl residue, the whole approach is to get the gold in a bar by smelting as quickly as possible to minimise theft.
It is an article of faith that these beads are recovered by putting the slag in the mill and reprocessing it, my test work showed that any recovery from this approach is minimal.
Deano

 

[4metals]

No aqua regia is used to digest the HCl residue, the whole approach is to get the gold in a bar by smelting as quickly as possible to minimise theft.

Ahhh the human factor. And the extreme effect it has on recovery!

It would be interesting to take all of the insolubles from a lot after the Hydrochloric Acid and split it in 2 (evenly to make this as analytical as possible). Then smelt one half and digest the other half in aqua regia. Then compare the yields.

 

[HarryLee1506]

I am having a problem with my slags. Fire assay is 25g per ton, I only have hammer mill with 1-2mm material output, I use Jinchan as leach solution (main ingredient is Thiourea as declared by manufacturer), I recover 15% gold compared to fire assay (it seems faster and stronger than cyanide). The problem started here, I used activated carbon to precipitate the gold but it only helped me recover 30% of the gold in the pregnant solution, the solution assay still had a significant amount of gold (15000 liters of solution circulating 100 kg of activated carbon in 2 weeks). Can anyone help me recover all the gold in this solution? Thank you very much...

 

[Yggdrasil]

I am having a problem with my slags. Fire assay is 25g per ton, I only have hammer mill with 1-2mm material output, I use Jinchan as leach solution (main ingredient is Thiourea as declared by manufacturer), I recover 15% gold compared to fire assay (it seems faster and stronger than cyanide). The problem started here, I used activated carbon to precipitate the gold but it only helped me recover 30% of the gold in the pregnant solution, the solution assay still had a significant amount of gold (15000 liters of solution circulating 100 kg of activated carbon in 2 weeks). Can anyone help me recover all the gold in this solution? Thank you very much...

Usually the pure chemical is better than these "green" concoctions,
the issue is that that it is the "main" ingredient, what else is in there?
They strip well, but can be a nightmare to get the Gold out of.
There often are a lot of extra chemicals in these, blocking access to the Gold.
I have never used ThioUrea but I believe I have seen that one of the drawbacks of ThioUrea is exactly what you describe, it is hard to get the Gold out.

Deano will know more on this topic.

 

[4metals]

I am having a problem with my slags.

I agree that Deano will know more about this topic for sure. From reading your post it seems you are having trouble capturing all of the gold in the carbon rather than recovering all of the gold from the carbon.

This thread is specifically about recovering beads from slag and your question is really not about gold in slags. I suggest you start a new thread and name it appropriately so it can be referenced in the future based on the thread title. I have no doubt you will receive valuable advice here and it would be best if future members can find it based on the title.

 

[HarryLee1506]

Usually the pure chemical is better than these "green" concoctions,
the issue is that that it is the "main" ingredient, what else is in there?
They strip well, but can be a nightmare to get the Gold out of.
There often are a lot of extra chemicals in these, blocking access to the Gold.
I have never used ThioUrea but I believe I have seen that one of the drawbacks of ThioUrea is exactly what you describe, it is hard to get the Gold out.

Deano will know more on this topic.

Ingredients of Jinchan as announced by the manufacturer: SC(NH2)2 (Thiourea), NaOH (oxidant), Na2SiO3 and (NaPO3)6 (stabilizer). Besides being environmentally friendly, I feel it is faster and stronger than the previous cyanide I used.
According to some documents I have read, ion exchange resin can help me solve this problem, I will start with it. There are four types of ion exchange resins: strong and weak cations, strong and weak anions. Do you have any advice for me?
Thank you very much...