How do manufacturers remove oxygen from metal during melting? I melt aluminium with nichrome wires and during melting it, it forms lots of slag (aluminium oxide) and I have to remove it manually.

  • 1
    $\begingroup$ If the metal reacts with oxygen at high temperatures, then the processing needs to be done under an inert atmosphere (i.e. nitrogen, argon, etc.). $\endgroup$
    – airhuff
    Jun 28 '17 at 3:47
  • 3
    $\begingroup$ Putting a lid on the crucible might already help a lot. Also: Thicker wires, larger batch size. ;-) And yes, the oxide swims on top, and you remove it mechanically, if necessary. $\endgroup$
    – Karl
    Jun 28 '17 at 4:29

They use flux, that's how. That is, an inert substance which melts easier than the metal and floats on top, preventing it from oxidation in the air. Common fluxes for aluminum seem to be based on eutectic mixtures of various chlorides (say, $\ce{NaCl + KCl}$).

Covering your crucible with a lid does not offer complete protection, and working in inert atmosphere is much more costly (though sometimes they do that, too).

  • $\begingroup$ what does "lid does not offer complete protection" mean? It implies a lid offers partial protection. $\endgroup$
    – johny why
    Dec 28 '18 at 7:43
  • 1
    $\begingroup$ Well, yes, that's right. $\endgroup$ Dec 28 '18 at 7:53
  • $\begingroup$ Ok, cool. I assume you've noticed that a lid helps some. Can say a bit about how much it helps? Thanks! $\endgroup$
    – johny why
    Dec 28 '18 at 8:12
  • 2
    $\begingroup$ I don't think we can say anything more specific in the general case. There are different molten metals and different lids. $\endgroup$ Dec 28 '18 at 8:16
  • $\begingroup$ Can we say a tight lid will seal out more oxygen than a leaky one? $\endgroup$
    – johny why
    Dec 28 '18 at 9:10

There are two aspects to this, firstly is actually reducing the pure metal from its oxides, usually called smelting, most metal ores are metal oxides so this is clearly an important part of primary metal production For example in iron production iron oxides are reduced by carbon, usually from coke which is also the fuel which provides the energy for the reaction.

In the case of remelting metals for casting it is usually a case of limiting oxide formation rather than reversing it. For decent quantities it is usually self limiting to and extent as the oxides tend to float to the surface of the crucible, limiting and further oxidation. Cover fluxes can also be used which sit on the surface and provide a soft seal to the crucible.

This is aided by use of fluxes which are able to trap the oxides which do form and trapping them in a semi-fluid glass mass which is easier to remove. In iron casting limestone is used, other additives (such as sawdust) can be used as slag coagulants to make the resulting slag less sticky and easier to handle.

In the case of aluminium especially thin foil and wires surface oxide layers can make of a significant volume of the material before you even start melting it and there is not a lot you can do about that. Aluminium has a particularly high affinity for oxygen and the only practical way to reduce it is by electrolysis.

If you are melting a range of sizes of material (eg a mixture of scrap and ingots) it may help to melt the larger pieces first as they have a lower surface area to volume ratio and then adding the smaller bits so they melt while submerged in an established pool of molten metal.

Strictly speaking the oxides you get on the surface of molten aluminium are called dross, this term applies to metal oxides produced during melting while slag is a glassy material which is either a mixture of metal oxides and flux or silicates from the ore.

In industrial production there are various approaches to slag removal. It can either be skimmed off the surface of the ladle/crucible or you can tap from the bottom of the bottom (as in a cupola furnace). Some casting applications also use ceramic slag filters to prevent any stray bits from getting into the mould.

  • $\begingroup$ awesome. You listed about 6 different methods, where other answers only offer 1. $\endgroup$
    – johny why
    Dec 28 '18 at 7:41
  • 1
    $\begingroup$ Re the self-limiting layer: isn't this a good reason NOT to STIR the pot? I've noticed some youtubers continually stir the pot, as if it's soup :D $\endgroup$
    – johny why
    Dec 28 '18 at 9:34
  • $\begingroup$ @johnywhy maybe they stir the pot to ensure no bubbles are trapped or something. $\endgroup$
    – gfdsal
    May 25 '20 at 20:31

Generally speaking, manufacturers remove oxygen from metal by using deoxidizers. For instance, a manufacturer will melt steel, and pour liquid aluminum (or various other viable substances that act as reducing agents) into the liquid steel. This aluminum will react with the oxygen in the steel to form aluminum oxide (Al2O3). Because the oxide is lighter than the steel, it flows to the top, and is collected from there. There is more satisfying detail here.

There are more ways one would go about the deoxidation of the metal, like blowing the oxygen out with noble gases like Argon (since having a nonreactive gas in the metal won't be a problem). There are more methods like the vacuum method which is described extensively online. Regarding the deoxidation of aluminum as it pertains to your specific question, there is great stuff to read about on this site. Some of the information on that website is about copper, but the concepts are the same.

I just realized that I misunderstood your question, and that Ivan's answer is the correct one. However, I will still leave this explanation here. Hopefully it serves as good context for the process of metallurgy.

  • 4
    $\begingroup$ Surely the oxygen is the oxidiser. Wouldn't the alumunium would be a reducer in this senario? $\endgroup$
    – Richard
    Jun 28 '17 at 9:17
  • $\begingroup$ Yes, my mistake. When Aluminum reacts with oxygen, oxygen gets reduced and aluminum gets oxidized. Therefore, aluminum is the reducing agent and oxygen is the oxidizing agent. $\endgroup$ Jun 28 '17 at 20:03

Although the original question is How to deoxidize oxidized melt?, in practice, deoxidizing may be more difficult than prevention or removal.

This answers the separate questions:

  • How to prevent oxidation in melt?
  • How to deoxidize oxidized melt?
  • How to remove oxidation slag (dross) from the melt?

This is a collection of options, from numerous sources and commenters. Although it overlaps some answers already given, hopefully it's a comprehensive answer. And hopefully, more clearly organizes these three separate questions. (In some cases, it's unclear whether the method will prevent, deoxidize, or remove oxidized metal.)

Be aware that some methods, such as fluxes, may vary depending on which metals you're working with.

How to prevent/reduce oxidation?

  • Not Oxidation: Your slag might not be oxidized metal at all-- it might be impurities, or contamination by other metals.

  • Self-Limiting: As mentioned in another answer, it is self-limiting to an extent, as the oxides float to the surface, limiting further oxidation.

  • Electrolysis: Aluminium has a particularly high affinity for oxygen and the only practical way to reduce it is by electrolysis.

  • Boric Acid: materials like borax fluorite, silica, sodium carbonate are used. As mentioned in another answer, these sometimes form undesirable "Spinels" in the liquid metal. Boric acid should melt and then (at least partially) decompose into boric oxide and water. Water isn't as dangerous as you'd might think as long as it's at the surface, so as long as you give it ample time to decompose before thrusting anything through the layer it shouldn't be a problem.

  • Flux: When the metal fully melts into a liquid, carefully sprinkle a small amount of flux onto the molten metal. 'Marvelux flux' is an example (not an endorsement). It will sit on the surface and provide a soft seal to the crucible. After pouring, allow the piece and the mold to cool completely. Then remove the piece, and pickle it to clean off any leftover flux. Then thoroughly rinse it. Alcohol and evaporation seems a better choice than vinegar and water, as vinegar can cause corrosion.

  • Glass: On brass, you'd use broken glass bottles.

  • Charcoal: With low temps, crunched charcoal is the usual remedy.

  • Kitty Litter: some bullet casters use it.

  • Remove the Oxygen: possibly replacing it with an inert gas. Nitrogen is commonly used, and it's non-toxic (it's already in the air we breathe). Watch that you don't use silicone tubing to connect your N2 cylinder - it is quite O2 permeable. Ditto silicone gloves. There are several ways to do this-- all require an airtight environment. Oxygen must be removed before the melting metals are exposed to the atmosphere. Some methods listed are used in manufacture, some might work for small-scale production or DIY, and some are speculative:

    • Burn off the oxygen with a flame. Depending on volume, this may cause a pressure difference between the volume and the outside environment, which in turn may require replacing the oxygen with an inert gas.

    • Pump out all the air, and operate in a vacuum.

    • Adsorb it with an oxygen "getter". A getter is a deposit of reactive material that is placed inside a vacuum system. When oxygen molecules strike the getter material, they combine with it chemically or by absorption, thus reducing the amount of oxygen in the space.

      • Commercial gloveboxes use 10% H2 in N2 (formier gas) over heated Pd catalyst to convert O2 to water, which is then removed by drying train (molecular sieve). This may produce CO2, which may need to be removed.
      • Pass the air through a heated column of copper powder. Ideally, 63,55 g copper powder should remove 24,5 L of oxygen from the air - which leaves us with >80 L of nitrogen left. Now that's not a whole lot of copper for a 30-40 L glovebox! However, depends on column temperature, flow rate, mesh size etc. The atmosphere in the glovebox circulates all the time, and the catalyst can go from 20% O2 to 0ppm in like 15 minutes.
      • Food-safe deoxidizers: dry iron powder, sodium and sometimes activated carbon. Two 500cc of such oxygen absorbing agents can absorb up to 500cc of oxygen. Normal dry air contains 20.95% oxygen, so you'd need a volume of oxygen absorbing agent equal to 21% of your insulated workspace. More numbers here.
      • Pass it through a heated tube (circa 500 degreesC) stuffed with steel wool. Unsure if this can work with air, or just water.
      • Other recipes comprise a porous reducing support such as activated carbon, carbon black, coal, petroleum coke and titania which is impregnated with about 10% to about 90% by weight of an alkali metal oxide or an alkaline earth metal oxide.
      • A simple tray of vinegar and steel wool. The vinegar dissolves the protective layer on the steel wool, accelerating the rusting process, which sequesters oxygen from the local atmosphere.
      • Get some of those chemical hand warmers and cut them open and pour the powder in a dish in your glove box. That stuff is reduced iron combined with catalysts that make it more prone to oxidation.
      • A Zirconium-based getter.
      • An open petri dish of sodium/potassium eutectic, maybe with mag stirrer.
      • Get a hold of this stuff. Researchers from the University of Southern Denmark have synthesized crystalline materials that can bind and store oxygen in high concentrations. Just one spoon of the substance is enough to absorb all the oxygen in a room. Wow!
  • Lid: as mentioned in another answer, a crucible lid may reduce oxidation.

  • Pipes: Connect your crucible directly to your mold via a pipe, so the molten metal is transferred without exposure to the air.

  • Reduce area: A skinny crucible, or one with a narrowed upper section, will reduce the surface area exposed to the air. This will be impossible if your solids are thicker than the neck of your crucible. Consider starting with metal powder, instead of ingots.

  • Eliminate the crucible: Put dry metal powder directly into the mold, pack it down, seal the mold, and put it in an oven. Known as "Compacting and sintering".

  • Eliminate the crucible, 2: Put a thin layer of metal powder on a flat work surface. With a CNC laser, fuse the powder in the shape of the bottom of your pattern. Add another layer of metal powder, and fuse it. Repeat until the whole part is built up in layers. Known as Direct Metal Laser Sintering.

How to deoxidize oxidized metal?

  • Heat: turn the pot up really hot and it dissolves back in for the most part.

  • Charcoal: To deoxidize the layer of dross/oxide sprinkle with powdered charcoal.

  • Deoxidizers. As mentioned in another answer, a manufacturer may pour liquid aluminum (or other viable reducing agents) into liquid steel. The aluminum will react with the oxygen in the steel to form aluminum oxide (Al2O3). Because aluminum is lighter than the steel, it flows to the top, and is collected from there.

  • Apply electric current. In 1906, English engineer A. G. Bloxam used direct current in a vacuum, to produce filaments for incandescent lamps. The applied current was particularly effective in reducing surface oxides. (they might mean "prevent", not "reduce"). This has applications in powder-metal sintering.

How to remove oxidized metal?

  • Sawdust: Stir it and scrape the sides of the pot with a wooden stick. Pick out the trash and leave a layer of ash on top. As mentioned in another answer, sawdust may add hydrogen to the liquid Al causing a new set of problems.

  • Limestone: Used in iron casting, as mentioned in another answer.

  • Skim: It can be skimmed off the surface of the crucible.

  • Filters: use ceramic slag filters to prevent any stray bits from getting into the mould.

  • Bottom Tap: Extract from bottom of pot, instead of top.



Charcoal is added to the flux. Things like sawdust can add hydrogen to the liquid Al causing a new set of problems. Better check references for Al flux ; materials like borax fluorite, silica, sodium carbonate are used but sometimes form "Spinels" in the liquid metal = more problems.

  • $\begingroup$ what's the role of the charcoal? thx $\endgroup$
    – johny why
    Dec 30 '18 at 6:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.