Sciencemadness Discussion Board

Reduction of aluminium via Deville process

Titi - 15-6-2023 at 08:17

Good morning,

I read a lot about the first tentatives of reduction of alumina to aluminium. Before the Hall Heroult process, it was done in a purely chemical way. Although it is difficult to find a precise explanation of the process, what I understood is that it was done in two steps.
First, the aluminium oxide has to be transformed to (anhydrous) aluminum chloride, via the reaction
2 Al2 O3 + 3C + 6Cl2 -> 4Al Cl3+ 3CO2
Looking at enthalpy for this reaction gives that it should happen spontanously, however the text suggested heating was necessary. Maybe because it is a reaction between 3 reagents (and 2 of them solids). It seems that the important point is that carbon loves oxygen but not so much chlorine. Would subtitute carbon by anything else would make the reaction easier? I do not know what temperature is necessary to achieve this reaction, and in the forum anhydrous aluminum chloride is said difficult to produce.
Also, it seems that hydrated aluminium chloride is completly different, and is much more some kind of acidified alumina (covalent bounding occurs between Al and O, and not with Cl), which makes impossible to dehydrate it.

The other phase of the process is the reduction of sodium by the reaction
2NaOH + C -> 2Na + CO2 + H2
This reaction does not proceed naturally, and only works when Na becomes gaseous. Here also I wonder what kind of temperature is necessary. Partial pressure of Na becomes non negligeable at reasonnable temperature, so maybe with a vacuum pump this reaction could be facilitated. Also would not be this reaction easier with potassium? Deville worked with tens of kgs of sodium metal at once, so security of the reaction was important, but initial preparations of aluminum used potassium.
I also wonder if iron powder instead of carbon would be better, or any other metal (not aluminium obviously).

Then mixture of sodium and aluminium chloride reacts vigorously to make aluminium metal, so this should not be difficult.

Were these reactions already attempted by amateur chemists? Another possible route would be using magnesium, which can reduce alumina directly. Reduction of magnesium similar to sodium should be possible, but would require even higher temperatures. I was considering trying this with some copper tubing, so I have concerns about required temperatures.

Texium - 15-6-2023 at 09:46

First off, welcome to the forum.
Quote: Originally posted by Titi  
Looking at enthalpy for this reaction gives that it should happen spontanously, however the text suggested heating was necessary. Maybe because it is a reaction between 3 reagents (and 2 of them solids). It seems that the important point is that carbon loves oxygen but not so much chlorine. Would subtitute carbon by anything else would make the reaction easier? I do not know what temperature is necessary to achieve this reaction, and in the forum anhydrous aluminum chloride is said difficult to produce.
As has been stated many times on this forum: thermodynamics ≠ kinetics. A reaction can be spontaneous but not proceed without significant input energy simply because it has a very high activation energy. Diamond should spontaneously transform into graphite, but it doesn't, because breaking up the C-C covalent bonds is a very energy intensive process, even if graphite is a lower-energy structure than diamond. There always needs to be a reasonable pathway for the reaction to occur. In the case that you mention, I suspect the temperatures required would be extremely high, and it would likely need to be carried out in a sealed bomb reactor under a high pressure of Cl2. Very difficult and dangerous. If it wasn't, we'd already be very invested in that process as a method to make AlCl3, which is one of the most useful and interesting compounds of aluminum, and not too easy to come by.
Quote: Originally posted by Titi  

The other phase of the process is the reduction of sodium by the reaction
2NaOH + C -> 2Na + CO2 + H2
This reaction does not proceed naturally, and only works when Na becomes gaseous. Here also I wonder what kind of temperature is necessary. Partial pressure of Na becomes non negligeable at reasonnable temperature, so maybe with a vacuum pump this reaction could be facilitated. Also would not be this reaction easier with potassium? Deville worked with tens of kgs of sodium metal at once, so security of the reaction was important, but initial preparations of aluminum used potassium.
I also wonder if iron powder instead of carbon would be better, or any other metal (not aluminium obviously).
Once again, extreme temperatures will inevitably be required for carbothermal reduction. Vacuum is unlikely to help, as it is the temperature of the reaction melt, not the state of matter of the sodium produced, that is important to it proceeding. If you are interested in doing this as a recreation of a historic method (which is what I suspect since why else would you be going to all this trouble to make aluminum by an extremely inefficient and archaic method) it's going to be extremely difficult in a home lab. If you just want to get sodium metal, there are much better ways. Look up NurdRage's video on reducing sodium hydroxide with magnesium using menthol as a catalyst. Alternatively, even electrolysis of molten sodium salt mixtures is considerably more practical than carbothermal reduction, and there are multiple discussions of that here too.

Same applies if you want to use potassium. Amateur chemists have been accessing potassium metal with ease for a long time now. It's likely that they switched from potassium to sodium for this aluminum process because sodium is cheaper and easier to produce on an industrial scale.
Quote: Originally posted by Titi  

Then mixture of sodium and aluminium chloride reacts vigorously to make aluminium metal, so this should not be difficult.

Were these reactions already attempted by amateur chemists? Another possible route would be using magnesium, which can reduce alumina directly. Reduction of magnesium similar to sodium should be possible, but would require even higher temperatures. I was considering trying this with some copper tubing, so I have concerns about required temperatures.
A reaction being vigorous doesn't mean it isn't difficult. In this case, the reaction may be so vigorous that it is difficult to do in a controlled manner. You'd be lucky if the aluminum produced doesn't go up in smoke.

And overall, no, I don't think that this sequence has been done by amateur chemists before, because there is no reason to go through all that effort to make aluminum, and there are far better ways to access sodium and potassium.

Titi - 15-6-2023 at 11:45

Yes the idea was trying to recreate this old method for making aluminium. Yes I saw NurdRage video about making sodium from magnesium. This is a very interesting way, however, for aluminium reduction this is completly pointless as magnesium itself would be better for reducing aluminium.
About making sodium/potassium by electrolysis, I saw several videos doing it, it seems doable. By the way, would it be preferable to electrolyse an eutectic mixture of NaOH and KOH, because temperature would be lower?
About AlCl3, is there any way to make it reasonnably without coming from Al itself?

Admagistr - 15-6-2023 at 12:53

Quote: Originally posted by Titi  
Yes the idea was trying to recreate this old method for making aluminium. Yes I saw NurdRage video about making sodium from magnesium. This is a very interesting way, however, for aluminium reduction this is completly pointless as magnesium itself would be better for reducing aluminium.

About AlCl3, is there any way to make it reasonnably without coming from Al itself?


AlCl3 can be made from Al2O3 using carbon and chlorine.When you reduce aluminium compounds with magnesium,you get aluminium alloy with magnesium and magnesium is not cheap,as it is very energy intensive to produce.But the supply of Mg in nature is almost inexhaustible.

Titi - 16-6-2023 at 11:45

But how easy is this way of making AlCl3? Is it really necessary that chlorine be under pressure?
For magnesium reduction, I would guess that if magnesium is the limitant reageant, this would not be a problem. Just put an excess of Al2O3, and then melt the aluminium to separate it.
Remark that a constant supply of magnesium is not necessary, because MgO is produced when Al2O3 is reduced.

For a reversed question, is it possible to make sodium (or potassium) via aluminium, like
6NaOH + 4Al -> 6Na + 2Al2O3 + 3H2
Heating would be required, as Al is not as reducing as magnesium, but how much?

Admagistr - 16-6-2023 at 13:11

Quote: Originally posted by Titi  
But how easy is this way of making AlCl3? Is it really necessary that chlorine be under pressure?
For magnesium reduction, I would guess that if magnesium is the limitant reageant, this would not be a problem. Just put an excess of Al2O3, and then melt the aluminium to separate it.
Remark that a constant supply of magnesium is not necessary, because MgO is produced when Al2O3 is reduced.

For a reversed question, is it possible to make sodium (or potassium) via aluminium, like
6NaOH + 4Al -> 6Na + 2Al2O3 + 3H2
Heating would be required, as Al is not as reducing as magnesium, but how much?


Converting Al2O3 to AlCl3 will require very high temperature and chlorine under pressure.Aluminium has a really strong bond to oxygen which is difficult to break.I don't know if just adding Al2O3 is enough to prevent the formation of an Al-Mg alloy.Reducing NaOH with aluminium will also require very high temperature and an argon environment or high vacuum. In addition,the reaction is not so simple,but NaAlO2 is also formed and this complicated the reaction.It would probably be better to use anhydrous Na2CO3,because of the H2O content of NaOH.Even trace amounts of H2O can be a problem and make this reaction difficult.



[Edited on 16-6-2023 by Admagistr]

Rainwater - 16-6-2023 at 16:46

Quote: Originally posted by Titi  

By the way, would it be preferable to electrolyse an eutectic mixture of NaOH and KOH, because temperature would be lower?

Sodium metal is soluble in molton NaOH above a very narrow temperature window. This makes the process ineffective and difficult to reliably reproduce. At least for me, tho i was able to get a few grams

I've found molten NaCl is much easier work with but requires a lot of heat and special materials. A small 900c kiln can be made for about 150 bucks from common items from the hardware store. Kilewool wrapped around a stovetop heating element with a 1000w light dimmer can maintain around 950-1000c for over 100 hours before the element fails. The trick is low power and lots of insulation. Then a split cell connected at the bottom, formed from clay and fired in the above mentioned kiln. Heat, apply power, cool, collect. Can usually get 25-30g from 2amps over 24 hours.
Quote:

About AlCl3, is there any way to make it reasonnably without coming from Al itself?

Search the forms by clicking here
Not that I seen, I did a study on AlCl3 a while back. Needed a pure sample to compare to my experiments, Cl2 and aluminum foil was by far much easier to perform and less damaging to glassware than the other stuff I tried. But i was successful in making a crude AlCl3 with small amounts NH4Cl contamination. Everything I tried started with Al metal.

Titi - 17-6-2023 at 09:39

So I am beginning to think that using AlCl3 should be avoided. I also was thinking about another route using magnesium as intermediary
Al2O3 + 3MgCl2 + 6Na -> 2Al + 3MgO + 6NaCl
The idea is that MgCl2 can be reduced by sodium, and then Mg reduces alumina. This avoids the difficult to obtain AlCl3. By the way, making magnesium from sodium could be interesting
MgCl2+2Na -> Mg +2NaCl
As I understood, magnesium chloride can be dehydrated by heating contrary to aluminium chloride. So making it (or regenerating it from MgO) should be easier.

About electrolysis of NaCl, I have hard time to imagine this feasible. Isn't sodium a gas at these temperatures? Isn't the reaction with atmosphere immediate. I am not sure how the cell is build. Industrially they use an eutectic mixture with calcium chloride, but already this seems difficult.

Titi - 17-6-2023 at 09:58

After some additional readings, it seems that also MgCl2 cannot be dehydrated easily... What are the chlorides which can be dehydrated by simply heating? Maybe calcium chloride could work? But maybe calcium is not enough to reduce alumina.
It seems in general all these oxygen loving metals are very hard to reduce, harder than alcaline metals...

Rainwater - 17-6-2023 at 10:16

This post may be helpful
https://www.sciencemadness.org/whisper/viewthread.php?tid=15...

Admagistr - 17-6-2023 at 11:26

@TITI:
Sometimes it's good not to deal with thermodynamics in the first place and do the experiment.And deal with thermodynamics afterwards,just don't have preconceived ideas and don't give up beforehand.Maybe you will be the first one to do it successfully here on the Forum and you will reap great success.If you have an electric arc,you can get a bottle of argon or a good vacuum pump,go for it! I noticed that you have the right chemical thinking,I like some of your equations,they show complex thinking and talent for chemistry.Even a failed experiment shows that this is not the way,so it has value and makes you find another solution,I see it as a challenge;)