Smelting

Gold Refining Forum

Help Support Gold Refining Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

4metals

Moderator
Staff member
Supporting Member
Moderator
Joined
Apr 5, 2009
Messages
6,070
Location
northeast USA
OK smelting it is.

First let me say that the majority of medium to the small side of large refiners today are running circuit boards today by smelting and refining the copper. The more time you can put into segregating and separating the material the better the payoff. That is why this is done in countries where labor is cheap. Having 15 or 20 sorters picking through the boards and cutting off components is effective at producing a feedstock which lends itself to electrolytic copper refining to concentrate the PM's in slimes. Poor sorting results in smelting lower grades of copper with metals like lead which are a nuisance metal in the copper refining process and they are a reason copper smelters will charge you extra when these metals are in the mix.

So assuming we all have the sorting process down pat, we end up with boards and components which we need to burn. The ideal scenario is controlled pyrolysis, followed by incineration, and then either sifting or smelting. If we go the smelting route generally circuit boards need copper added to get up to about 15% metal weight for the best collection.

When smelting the ideal system involves melting in a rotary kiln. These are pricey and not very versatile.

So let's start out with me showing you a photo of a small tilting pour gas furnace, this is a center pivot furnace and this is about as small as they get.

T160 gas furnace.jpg

The center pivot refers to the pivot points on the tripod stand which means the unit pivots from the center when pouring. These are the least expensive units to produce and they are challenging to pour bars with. As you tip the furnace more and more when pouring, the spout where the metal comes out moves closer to the furnace center line. At full tilt, you are under the furnace. So you have to move your mold to catch all of the pour. To define the challenge a little better, if you have a furnace 24" in diameter, the pour spout moves in an arc which will drop 12" down and 12" towards the centerline of the furnace. A hydraulic melt table will make this do-able but it has to be moved in and down as the pour proceeds. It's a skill you acquire with practice, or maybe spills and practice.

The alternative is a nose pour furnace which has the pivot point from the stand on the outside edge in a line tangent to the outside diameter of the furnace body. This results in less travel of the pour point when pouring bars.

this is a catalog cut from a Baker gas nose tilt furnace.

crucible_tilt_150.jpg

So you might ask, why not a lift and pour crucible? My answer is capacity. I tend to work on bigger stuff so i think along those lines. Plus for a lot of members here on GRF who work alone, lifting and pouring a large crucible is a 2 man proposition. A tilting pour is a one man show.

You could build a furnace into the ground. Just like the refiner in Thailand from this thread posted by Gaurav.http://goldrefiningforum.com/phpBB3/viewtopic.php?f=38&t=23609

To pour from this furnace they need to lift the crucible out and pour it. Generally a crucible is lifted with a pair of these tongs.
crucible tongs.jpg

These can lift 2 ways, one by having one man on each end, and the other by using a hoist on the ring in the center, the hoist pulls up and the tongs cradle the crucible from the sides tapering to the bottom so it doesn't slip. Then to pour it has to be transferred to a pouring cradle.

That is why I like a tilt pour setup.

The unit I am working on for my client in Mexico is a hybrid of 3 different pieces of equipment. It has a nose tilt crucible furnace which is poured by tilting the furnace with a hoist pulling on a cable attached to the lower half of the furnace 180 degrees opposite the pour spout. Pretty basic.

Then there are different lids. The regular melting furnace lid, the limit the oxygen and deflect the smoke into the afterburner lid, and the actual afterburner.

Now here is a catch. These lids are refractory and quite heavy so you will need an overhead lift. I prefer this type of crane but an overhead I beam would do as well. In a smelting shop an overhead hoist is a blessing. When you shot metal into water, you need a good quantity of water and you don't want to change it often. (climate change, drought, California is a desert, you know the drill) So the shot tank has a rigid metal liner on chains that fits the bottom of the tank and after you pour, you hoist out the liner and all of your nice shot comes with it.

This is the style of hoist crane I prefer;
hoist rane.jpg

I made this furnace representation in Paint on my laptop so some of this makes sense. The red is obviously the flame. The bottom flame heats the crucible and its contents. There is no crucible shown in the sketch. I prefer nose pour crucibles but it will also take a standard crucible and you can add a launder to direct the molten metal. The top section is an afterburner. When using the unit for pyrolysis, a lid is added over the charge so all of the smoke comes out the center. The afterburner section is just an extended burn chamber with its own flame which I like to run with excess air as the combustion smoke is air starved at this point. This is pretty effective at burning the smoke off from the pyrolysis providing the hole in the top between the top and bottom (also not shown) limits the smoke and uncombusted fume discharge rate.

2 section gs tilt furnace.jpg

The rings on top of the top unit have a chain which is picked up by a hoist to remove the afterburner. A gas quick disconnect is helpful here.

So to operate the unit with a batch of unburnt circuitry first the crucible is filled. Then the air restricting lid is placed on top with the hoist, then the afterburner is put in place. The afterburner fires first and when it is hot (glowing is nice) the lower unit is lit. For pyrolysis this is a low heat not full blast but warm enough to get the pyrolysis going. I toyed with the idea of putting a bottom on the afterburner unit to eliminate the need for the lid to control the smoke entering the afterburner. The bottom would obviously have a small hole to emit the smoke. I chose to not do that because I may need the afterburner for mild smoke conditions at some point and no lid would be needed, so for versatility I chose to keep it separate.

When the pyrolysis is done, a # 150 crucible should take about 1 hour, the burners are shut off and the afterburner is removed, the inner lid is removed, and the furnace (bottom) is re-lit. Then the carbon is burned off to complete the pyrolysis/incineration process. I would be putting the afterburner back on for this process until you learn how long to burn to complete pyrolysis, but that's me, I've inhaled enough crap in my life already.

The furnace can now be either emptied out to sift and flux the powders or flux can be added to the charge in place. A second lid for melting is placed on top of the furnace, it too is placed with the hoist. I added a pin at the rear opposite the pour spout so the lid doesn't slip off while pouring. The pin slides into an oversized sleeve on the furnace so it is easy to place the lid on. For the smelting part of the cycle, copper is added to get the metallic fraction over 15% and I like to rock the furnace. Remember it pivots, so getting the molten pool swirling by slow rocking using the tilt hoist will aid in collection. The molten charge is periodically slagged off by pouring into a slag mold. Large slag molds are sold by Legend and they attach to a buggy so you can roll them around.

slag mold buggy.jpg

After the furnace has melted all of the flux/sweeps charge, the majority of the slag is poured off until the metal starts to flow into a sample ladle then the charge is shot into water to make grain.

The copper based grain is cast into anodes. The furnace is too large to make the high surface area thin anodes needed for the copper cells so that is why grain is cast. It is easily weighed out and cast into individual anodes.

So that is smelting, we do need to cover fluxes and a few details but this is the hardware, the rest is chemistry.

The actual furnace I just described is currently being cast, the 1/4 scale model worked very well, It takes a # 40 crucible. They plan on using the little guy as well as the one being built. The first unit was built by the maintenance crew at the shop from some rather crude drawings I had made, they completely understood the concept (not from my Spanish, from the boss's translation!) and went with it. The bigger unit is more of the same. So for a moderately handy person, this is definitely within reach.
 
One can smelt many types of material.
As I understand it, the success and profitability of the smelting depends a lot on material preparation.

Is this aspect of material preparation will be covered depending on the feedstock ?

A few examples of various types of scrap one can get :

-Circuit boards
-Sulfide ores
-low grade PGM bearing material (car spark plugs mentioned earlier)
 
I know I've seen lead mentioned several times as a problem metal for the copper refinery, and spoke of again here by 4metals. What are the other problem metals?

That might help me wrap my head around a little thought bubble that's trying to percolate, but right now my lips keep covering my eye teeth and i can't quite see what I'm trying to say.
 
FrugalRefiner said:
Splitting the thread is not a problem. I've already PMd 4metals about it and I'll split the smelting portion off after I get his response.

Dave

Thanks Dave --- I will be adding to the smelting thread after you have split it off to its own discussion so as not to clutter this thread

Kurt
 
I always prefer to remove the incinerated material from the furnace to crush and screen it before fluxing and smelting. This is for a number of reasons, first, there is a fraction that is metallic and if it can be separated by crushing and sifting, it can be melted with copper directly to make anodes. Often the highest percentage of values is in this "metallic" fraction. For years I used to incinerate boards and crush and sift them to make copper based bullion (which I assayed and shipped to a copper refinery) and shipped the powder to low grade refiners as sweeps. For the purpose of this thread, we are going to use the approach that we will process this material through the copper electrolytic process to concentrate and recover all of the values and pure electrolytically refined copper.

Some guys just flux the load in place and melt. I do not like to do this because for one, we do not know an after burn weight so we do not know how much flux to add. Second, without mixing the flux and ash well, a good amount of the benefit of fluxing is never realized. I have seen refiners pour melted sweeps only to see a nice smooth flux pour off first followed by a lump that never melted because it wasn't fluxed properly.

Another benefit of crushing the lot is you can do test fusions in an assay crucible in a kiln to determine the best flux mixture. Then you can flux the entire lot and smelt it.

In the prototype furnace I described earlier it was easy to onload a lot after incineration by lifting off the lid and tilting the furnace 90 degrees and raking the ash into a tray positioned in front of and below the crucible. A hoe fashioned out of steel and having a radius to match the crucible does a nice job. The larger furnace won't be as easy to lift off the lid for additions and emptying because it is much heavier and larger so a stand is being fashioned to allow the lid to swing open and rest on while charging the furnace while it is hot. This top down view drawing made in Paint may make this easier to understand. A removable steel rod will slide into a sleeve on the lid to make a handle to lift and swing the furnace open. The green circle represents the cover opening allowing material to be fed into the furnace.

2 section gas furnace lid.jpg

Originally I was planning on making a lift mechanism similar to what Mifco uses but considering this lid needs to come off entirely when using the afterburner, I opted for simple. Remember heavily used equipment in refineries break, simple is better. The mechanism for a Mifco lifting lid can be seen in this drawing;

Mifco lid lift configuration.jpg

Once you have your ash out of the furnace it is a good idea to do a magnetic separation, a hand magnet like this can de-mag a lot of ash quickly and drop it into your magnetics pile. This is a nice hand magnet in action;

hand-magnet.jpg

Now you can ball mill the ash without any magnetics in it, this is the best option but most small refiners do not have a ball mill. Some use cement mixers with a few balls to break up the ash. It works, kind of, remember it isn't really a ball mill.

Another option for nice crumbly ash is to sift it on a mechanical sifter and a hand full of 1" steel balls to crush and help the material go through the screen. A -20 mesh screen would be the largest I would use, a -40 mesh is better. A nice sifter for a small refiner is a drum top sifter. This little beauty sits on a 30 gallon drum and sifts right into the drum.

kason drum top sifter.jpg

Another sifter for a small operation is a Ro Tap sieve shaker. These come in a size to shake 8" screens or 12" screens. This actually does not need balls as the Tap portion of the name implies, there is a weighted arm that pounds on the top of the sieve frame every few seconds to help break things up.

Ro-tap sifter.jpg

It is always a good idea to keep an eye out for these at auctions, there are lots of industries that use them and you never know when the people bidding have no idea what it is used for.

This pretty much describes what you will need to smelt. Next we should discuss fluxes. The right flux will make the melt more fluid and allow the metallic fraction to collect in a pool of molten metal.
 
What would be the targeted temperature for pyrolysis ? If the charge is heated too high, some metals are going to melt obviously.
Aluminum, Tin, Lead... Any problems resulting from melting these metals in the pyrolysis stage or does the right fluxing will get rid of them ?
 
alexxx said:
What would be the targeted temperature for pyrolysis ?

between 600 & 750 F

If the charge is heated too high, some metals are going to melt obviously.
Aluminum, Tin, Lead... Any problems resulting from melting these metals in the pyrolysis stage or does the right fluxing will get rid of them

Per the under lined above - the answer is both yes & no --- in order to have metals go off in the slag (flux) they need to be in an oxidized state --- this can be "somewhat" effected/influenced with fluxing - furnace flame environment also plays "somewhat" of a factor here (whether the flame is an oxidizing flame or reducing flame)

in other words though some metal may be oxidized & carried off in the slag it wont be complete

Kurt
 
I am understanding the furnace design correct in that is acts as both the furnace and crucible? And the pour is directly from the furnace as a crucible?
 
That makes more sense, and now that you explain it, I can see it in the Mifco design. This is looking to be a great post, Thanks for all the input.
 
If I remember correctly Harold made a furnace that was a direct melt furnace. I seem to remember him posting pictures of it on the forum. He is quite the craftsman!

Maybe he will chime in here with some pictures and his experiences with his furnace.
 
As Alexx mentioned, there are many materials where smelting might be the best process. I'll add to his list by suggesting posts on MLCCs and ICs.

I don't deal with the quantities of e-waste that many of you do, but I do accumulate a bit of it. I'd be interested to know what those with experience would consider minimum batch sizes. Everyone collects MLCCs, and as I understand it, smelting is the best way to process them. If someone has an appropriate sized furnace and crucible, could they smelt a few ounces, a pound, a few pounds?

I'm sure some of our prospectors will be looking forward to coverage of ores. Keep it coming guys. I see a great Library thread in the making here.

Dave
 
I posted this exact post on a thread about melting gold but I also realize it is pertinent here as well, so at the risk of being banned for double posting I have copied it here as well.

Types of flames for melting

When melting the 2 fuel sources mix together to form a specific type of flame depending on their ratio. Those 3 types are the Oxidizing flame, the Neutral flame, and the Reducing flame (also called the carburizing flame)

flame types.jpeg


So by controlling the airflow to the fuel and the quantity of fuel you control the flame. Some torches burn oxygen and a pressurized gas but blower type furnaces use air but the thing in the air that is doing the work is the oxygen.

A reducing flame is a flame that is starved of oxygen, when a flame is starved of oxygen it tries to make it up wherever it can. A potential source is metals that are oxidized in the melt. The flame wants that oxygen bad and it grabs it from the oxidized metals in the melt and in the process reduces those metals. A reducing flame is made by cutting back the oxygen supply by throttling down the air supply. This makes the flame start to roar. A reducing flame is the loudest flame in a melt furnace.

A neutral flame is a flame burning with a balanced quantity of oxygen and fuel. For metals melting it will get your charge hot enough to melt but will neither oxidize or reduce any of the charge.

An oxidizing flame is one that has an excess of oxygen and it gives up oxygen to any metals in the melt that can be oxidized. Fortunately for precious metal refiners the Noble Metals are noble because they resist corrosion and oxidation so they will not come out of the charge as an oxide.

In practice you never have the gas or the air wide open, usually you set both to light the furnace and get it pre heated and then start lowering each independently a little at a time to get the flame you desire.

The catch in all of this is that in a crucible melt, it really has little effect because the flame does not contact the charge other than passing over the surface of the melt which usually has some sort of flux covering it.

When incinerating and having an open flame on the charge it does make a difference as the flame (and the excess or lack of oxygen) reacts directly with the charge. I have also done quite a bit of sweeps melting using a rotary kiln which actually uses the kiln as the crucible and the flame heats the charge directly. For this type of melting the ability to control the reducing power or oxidizing power of the flame matters.

For crucible melting if you desire to remove metals from a melt and encourage them to collect in the flux, an oxygen wand immersed in the molten pool with a slow steady flow of oxygen rising through the melt is very effective at removing a lot of metals. The exception being obviously the noble metals and copper. Copper doesn't like to gas off into the flux with oxygen. While in theory it should, in practice it just doesn't. The trick here is to have an open center of molten metal in the crucible and a donut shaped pool of flux to collect the oxides when gassing oxygen with a wand.

I have found most melt and incineration setups lack the sophisticated gauges and valves to precisely control either the gas or the oxidized. So you learn to work the valves and listen to the sound of the burner. Remember, louder is reducing, a more quiet roar is oxidizing. You can also judge from the flame color coming out of the top of the furnace when it is running.

This furnace is running an oxidizing flame
oxidizing flame indication.jpeg


And this furnace is running a lean or a reducing flame, and quite likely more noisy than the previous furnace.
reducing flame indication.jpg



I have seen large gas melt furnaces that have the mass to absorb the sound (my guess) combined with a large top hole that only produce a gentle roar regardless of the type of flame. But the flame color coming out the top is still an indication.
 
Great post again, wow, I love smelting already.

I can see the uses around an oxidizing flame.
Again, with a prepared feed stock of boards like you mentioned earlier, if you want to remove as much base metals as possible from the charge, injecting oxygen into the melt and with the right flux recipe, one could achieve a good copper anode quality ? For that specific scenario, it's a great thing that copper doesn't react like the other base metals.

Do you have examples of when a neutral or reducing flame would be used when dealing with precious metals bearing materials ?
 
Choosing a reducing or oxidizing preference is almost like asking a kid if he wants vanilla or chocolate today. It all depends.

I believe in the initial separation of the values and metals from the ash in circuit boards you want to assure you don't lose any metals to volitalization, you want to catch them all in the metal pool so you don't lose any values. I do this using a reducing flame and a reducing flux mixture. Notice I said I believe, I am sure there are other opinions and good discussion is healthy!

In all of this type of material I have processed in my day, the metals I've had to contend with are;

Zinc
Lead
Iron
Tin
Antimony
Nickel
Aluminum
Copper

And of course silver, gold, and the PGM's but they stay with the copper.

I went back to the data I personally collected from 3 setups I have done for this and came up with a range of each of the above metals. Since this is scrap after all, we can never control the source and the manufacturer or the age of the circuitry so these numbers vary but in the analyses I've done on the collected metals after smelt these are the percentages determined by atomic absorption on the metal poured out and sampled. I've tabulated this on Excel so it is quasi understandable. They all represent a % of the base metals in the final pour less the weight of copper added and, of course based on the pre burn weights.

Screen Shot 2016-03-26 at 9.41.36 PM.png

So from my experience the metals I want to collect in the metal pool are listed above. Less the precious stuff obviously.

So my flux needs to have the power to reduce the metals. Now most metals that are metallic to start only need to be melted but often from things like resins that have been roasted you want to reduce the metal which may have been captivated as a salt. One example of this are refining residues with chlorides or burnt cloths from spill clean ups of aqua regia or test papers. By melting them in a reducing environment you are assured better collection.

I also tabulated some information about different fluxes for our discussion.

Screen Shot 2016-03-26 at 10.01.06 PM.png

What this all tells us is how different ingredients in our basic flux will react to our charge. It has been my experience that PCB ash is pretty much a pH neutral charge and its components are really pretty easily reduced into the pool. For some ores and different materials some of the reducing agents will need to be added or possibly some desulfurization aids. But for PCB's I always just used a flux made up of Borax glass, Soda ash, and Silica. Together they will give you a fluid melt. To start off I would add a mixture with 10 pounds of Soda Ash, 5 pounds of Borax and 2 pounds of silica. Mix that up well and add it to the ash in the ratio of 1 pound mixed flux to 1 pound ash. Experience will teach you to tweak the formulations or you can simply add a half pound of flux and a half pound of ash to a melt crucible and fuse it in a gas melter to assure you get the fluidity in the melt and that everything is fluid so it stirs easily and pours smooth. If it is lumpy that says more flux to ash in the mix. This flux will be effective at collecting all of the metals in the molten pool. Generally printed circuit boards have enough metal in the overall percentage that they do not need additional copper.

Once molten, the melt is allowed to rest at full heat so the metal in the pool collects in the bottom of the crucible. The slag is poured off and when you get down to the metal pool it is either cast into bars or shot. This metal has mostly copper, all of the PM's and the other metals listed above if they are there. Remember, the better you clean up your boards before smelting, the less you have to deal with in the alloy you just poured. It's just like a computer, garbage in garbage out.
 
Before we can recover our precious metals from the slimes in the anode bags of our copper cells, we need to make sure we can get the feed copper to be as pure as we can up front. If you did your work up front and removed a lot of the iron and aluminum this part is easier.

I have cleaned my copper up while it is molten in a crucible by slowly bubbling oxygen into the melt to form oxides of the metals we are looking to remove. I have tabulated the metals we are looking to remove in this table. Copper is on the list but it comes off last so all of the other contaminants will react first. This is also done with compressed air as the source of oxygen. When refining copper electrolytically, Tin, Bismuth, Antimony and Lead all dissolve and end up in the slimes. Keeping this in mind it may be more desirable to skip the oxygen sparging (which is an extreme step) and allow the undesirable metals either end up in the slimes or dilute them with clean copper so you can go directly to the anode phase here.

Screen Shot 2016-03-26 at 9.57.21 PM.png

This shows us that the metals will form oxides which we can collect in a flux ring which forms when there is not enough flux to cover the entire melt surface. This thin donut of flux is made up of borax, silica, and soda ash 1:1:1 to gather any oxides that form and capture them in the flux. Sometimes the flux gets so thick with the metals you need to skim it off and add new. This process is not fast and usually it goes 6 to 8 hours to make the copper 99% copper. (actually 99% copper plus precious metals content)

So this is the point where all of the others who have done this can comment and make this more interesting. For now we should limit the discussion to printed circuit scrap, maybe we can get to ore concentrates later if Deano gives his opinions.

Have a Happy Easter!
 
A few questions regarding the table that was posted;

- The last 2 columns "oxide melting points in C / F"
Are those temperatures the temperatures needed to form an oxide or the temperature needed to melt the oxided elements in order for them to report into the slag? In other words, do you need to reach the oxide melting point in order to drive the oxide into the slag?

- since fluxes are also playing an important role in lowering the melting points of various elements, are they also lowering the melting points of their respective oxides?
 
Back
Top