Sciencemadness Discussion Board - SiO2 + Na thermite reaction?

SiO2 + Na thermite reaction?

NickBlackDIN - 11-7-2023 at 05:40

I was curious, the typical amateur synthesis of making elemental silicon is via the thermite process using aluminum. But then you have to remove the aluminum products, so typically people add sulfur to make that easier, but end up making some nasty fumes.

I was thinking, could I just melt some sodium metal, mix in the SiO2 to make the thermite mix (that way I don't have to deal with milling sodium and have to fuss around with argon atmospheres and such)

That way all the products I don't want should be water soluble.

How many issues do you see with this before I start investing any time on it?

j_sum1 - 11-7-2023 at 14:41

You will probably have a problem with the Na boiling away.
I have not checked the thermodynamics either. But I think it will be energetically favourable.
Nothing wrong with trying though.

I have used Mg powder to make Si. It gets around the problem of boiling. The products are soluble in water. And use of Mg has a better delta H than Na. Not to mention that it is cheaper and more accessible. You might want to consider that option.

Rainwater - 11-7-2023 at 14:54

Not sure sodium metal will work for silicon. Its the primary component of glass and does not get attacked it by sodium in anhydrous conditions.

Code:

SiO2 + 4 Na = 2 Na2O + Si
T = 0C
ΔH  69.5381  kJ/mol 
ΔS  -108.8677  J mol/K 
ΔG  99.2753  kJ/mol  

T=1000c
ΔG  208.1430  kJ/mol

Admagistr - 11-7-2023 at 15:13

I would try it,but I see a problem in that the sodium would ignite in the air before attacking the SiO2.Therefore I would cover the reaction mixture with a layer of NaCl.

NickBlackDIN - 11-7-2023 at 17:24

J_sum1:
I did think about Mg, but I didn't think MgO was water soluble.

the Na boiling away is a good point, I personally thought the bigger problem would be it just not getting hot enough for it to start taking the Oxygen from the silicon. (Which I guess is the same problem)

Rainwater;
Ohh, good point, I didn't think about it that way. Though in glassmaking, isn't it usually sodium ions that are added? Not the metal?

Admagistr:
Ohh, that's not a bad idea to try and increase yield (assuming it even works in the first place)

We'll see how busy I am next long weekend whether or not I have the time to try it

-----

Thank you folks for all the replies!

BromicAcid - 11-7-2023 at 18:57

Lithium will undergo a sustained reaction with glass.

B(a)P - 11-7-2023 at 20:52

Quote: Originally posted by NickBlackDIN

J_sum1:
I did think about Mg, but I didn't think MgO was water soluble.

It converts to hydroxide with the addition of water. You could wash your silicon with an appropriate acid to get rid of the magnesium hydroxide/oxide.

Edit
Typo

[Edited on 12-7-2023 by B(a)P]

Admagistr - 11-7-2023 at 21:09

Quote: Originally posted by B(a)P

Quote: Originally posted by NickBlackDIN

J_sum1:
I did think about Mg, but I didn't think MgO was water soluble.

It converts to hydroxide with the addition of water. You could wash your silicon with an appropriate acid to get ride of it.

Magnesium works very well, but the purity of the silicon obtained this way will not be high, some of the Si will certainly form some silicide with Mg, which will be very difficult to clean from silicon with acid.Sodium would probably provide a purer product.

[Edited on 12-7-2023 by Admagistr]

[Edited on 12-7-2023 by Admagistr]

B(a)P - 11-7-2023 at 21:24

Quote: Originally posted by Admagistr

Quote: Originally posted by B(a)P

Quote: Originally posted by NickBlackDIN

J_sum1:
I did think about Mg, but I didn't think MgO was water soluble.

It converts to hydroxide with the addition of water. You could wash your silicon with an appropriate acid to get ride of it.

Won't magnesium silicide react rapidly with a mineral acid to form silane? Making sure you do not have an excess of magnesium should reduce the magnesium silicide.

Admagistr - 11-7-2023 at 21:49

Won't magnesium silicide react rapidly with a mineral acid to form silane? Making sure you do not have an excess of magnesium should reduce the magnesium silicide.
[/rquote]

If the Mg silicide is dispersed in the silicon, the reaction will probably not be very fast...But the silicon could be finely pulverized...Another impurity in Si will be unreacted SiO2.

averageaussie - 15-8-2023 at 17:54

[/rquote]
Another impurity in Si will be unreacted SiO2.[/rquote]

Could melting the bulk silicon in an inert atmosphere make the silica sink in the liquid silicon? silica has a density of 2.65 g/cm3, and liquid silicon has a density of 2.57 g/cm3.
I know that melting the silicon might be a pain, but I see no other way to remove SiO2 impurities without turning your silicon to dust and picking it out.

even then, SiO2 impurities aren't even that visible I think, nor do they interfere with a reaction.

metalresearcher - 16-8-2023 at 07:10

I tried reacting SiO2 with Mg powder, it is a (rather slow) sustainable reaction, it gets very hot and then I dump it into HCl solution to dissolve MgO.
But indeed, melting the reaction result is a pain, not particularly because of the required 1410 C temperature (can be reached in a propane / air furnace or with a carbon arc of a welder), but the air. It oxidizes very quickly, hence the chip manufacturers use vacuum or inert gas (Ar) furnaces to handle liquid Si in the Czochralski process to purify it.

averageaussie - 16-8-2023 at 15:24

Quote: Originally posted by metalresearcher

I tried reacting SiO2 with Mg powder, it is a (rather slow) sustainable reaction, it gets very hot and then I dump it into HCl solution to dissolve MgO.
But indeed, melting the reaction result is a pain, not particularly because of the required 1410 C temperature (can be reached in a propane / air furnace or with a carbon arc of a welder), but the air. It oxidizes very quickly, hence the chip manufacturers use vacuum or inert gas (Ar) furnaces to handle liquid Si in the Czochralski process to purify it.

what setup do you have to melt the silicon? do you have access to an air free method of melting it?

also, can you post an image of the silicon obtained? I am quite curious about what it looks like.

Admagistr - 16-8-2023 at 16:15

Quote: Originally posted by averageaussie

[/rquote]
Another impurity in Si will be unreacted SiO2.[/rquote]

Could melting the bulk silicon in an inert atmosphere make the silica sink in the liquid silicon? silica has a density of 2.65 g/cm3, and liquid silicon has a density of 2.57 g/cm3.
I know that melting the silicon might be a pain, but I see no other way to remove SiO2 impurities without turning your silicon to dust and picking it out.

even then, SiO2 impurities aren't even that visible I think, nor do they interfere with a reaction.

An interesting aspect of fusing Si with SiO2 is the formation of SiO,which is volatile,hence one could probably remove SiO2 from silicon in this way!

metalresearcher - 17-8-2023 at 12:15

Quote: Originally posted by averageaussie

what setup do you have to melt the silicon? do you have access to an air free method of melting it?

also, can you post an image of the silicon obtained? I am quite curious about what it looks like.

No, I don't have access to it, but here a video of the reaction of Mg powder with silica sand. Afterwards I dumped it into HCl solution to dissolve the MgO.

https://www.metallab.net/jwplayer/video.php?f=/forums/MgO-Si...

averageaussie - 17-8-2023 at 18:43

[/rquote]

No, I don't have access to it, but here a video of the reaction of Mg powder with silica sand. Afterwards I dumped it into HCl solution to dissolve the MgO.

https://www.metallab.net/jwplayer/video.php?f=/forums/MgO-Silicon-20230817.mp4[/rquote]
with all those sparks, that would make quite the sparkler, no?

Rainwater - 9-2-2025 at 16:12

Quote:

could I just melt some sodium metal, mix in the SiO2 to make the thermite

I want to say I have confirmed that this reaction will not take place. But dont trust me, im an idiot

So today i tried sodium metal and finely ground SiO2 that i made from my stock of waterglass.
At about 875ish c under argon the sodium began to boil and made one hell of a mess.
But after boiling to dryness, i had no noticeable change in the weight of my crusable + sand.
Just the 5.02g loss which matches the weight of sodium added.

One observation was the mass of sand had fused together inside the crusable.
Which tends to happen with finely powered materials.
this is my first time heating amorphous silicate, i am not familiar with its properties
or behaviors. But with a melting point almost twice the reaction temperature i
wonder what flux like effect the molten sodium had. As some carbon from the crusable
did get mixed into the sand, it was easily removed by heating under oxidizing flame and
produced no noticeable change in mass after being removed with fire

Did i mention how much of a mess this made. Like how do i without a water hose even attempt... fKit i like fireworks

BromicAcid - 9-2-2025 at 16:36

Pretty sure magnesium silicides are made easily using Mg as a reluctant for SiO2 or perhaps that was how they were originally made. Might want to look into it.

teodor - 10-2-2025 at 06:12

Hm, is it exactly how Humphry Davy first get it? By heating SiO2 whith Na or K if I am not mistaken. I have his book in a different house, I will update here with the detais.

Update:
Yes, I was right, it was Davy who first get the Si element by adding K to white-heated amorphouse modification of silica, detailed method of preparation of which from a quarz sand takes a page in a book. But "compound is formed consisting of silica and potassa, and black particles not unlike plumbago are found diffused through the compound". So, the yield of this method is not good and purity is not the best.
The first practical process (end of 19 century) used K2SiF6 instead of SiO2 to fuse it sith K or Na and leaching fluorides with water leaving Si as a powder.
So, the answer, you cannot get it from SiO2 except in very small quantity.

[Edited on 10-2-2025 by teodor]

metalresearcher - 10-2-2025 at 09:42

Magnesium is easier. Recently I did it again, a kind of thermite with Mg powder and silica sand.
The output can just be dissolved in an HCl solution.
https://www.metallab.net/jwplayer/video.php?f=/forums/Silico...

teodor - 10-2-2025 at 10:19

J.W.Mellor, vol. VI. You can find it in the sciencemadness library. Contains description of many methods including Mg. But Mg route is not as simple if you need 90%+ Si and not Si+MgSi. Also sometimes you can get an explosion with Mg. Read in Mellor.

Rainwater - 10-2-2025 at 10:29

First hit on google

Quote:

French chemists Joseph Louis Gay-Lussac and Louis Jacques Thénard probably prepared impure amorphous Silicon by heating potassium with silicon tetrafluoride.

I was using sodium metal and silicon dioxide

teodor - 10-2-2025 at 11:42

Quote: Originally posted by Rainwater

First hit on google

Quote:

French chemists Joseph Louis Gay-Lussac and Louis Jacques Thénard probably prepared impure amorphous Silicon by heating potassium with silicon tetrafluoride.

I was using sodium metal and silicon dioxide

Do you know what is amorphous silicon?

The first crystalline silicon was prepared and analysed by the same man who first prepared potassium and sodium metals and exactly from silicon dioxide, by the same principle, but with more success his experiment was with isolating boron which he get in a good yield from boron oxide. He tried to use the same principle, also his goal was to show that silica is not an element but a complex compound with oxigen. Don't credit Lussac with all that.
And I cited his own explanation why the result with isolating Si from SiO2 was not satisfactory by his own opinion as well as that he succeeded with amorphous SiO2, not sand.

Well, of course you think that Google knows more.

Rainwater - 10-2-2025 at 14:32

Sounds rude.

Quote: Originally posted by teodor

Well, of course you think that Google knows more.

Google does know more than me.
No arguments to be made there.
As i said, im an idiot

So for a little more detail.
I slowly bubbled garage grade CO₂ gas into a diluted solution of sodium rich sodium silicate containing the equilivent of 3 grams of SiO₂.
I misplaced a decimal point and added 10x times the amount i wanted to.
I was attempting to lower the free NaOH content.
this produced amorphous silicon dioxide as a very fine almost colloidal suspension.
Dangerous stuff, like the kind that causes silicosis.
I had nothing better to do, and needed to test out my new Frankensteined smelter, now with controled atmosphere. Trying to breach 2000c.
And remembered this discussion, so test I did, results i shared

teodor - 11-2-2025 at 00:34

I think you can get some reduction by dropping of small portions of 2000C+ amorphous silica into molten alkali or NaK. The mass of alkali will dessipate the heat. Otherwise you should react SiO2 with gaseous metal, which is intetesting, of course, but requires a lot of work of appropriate apparatus design. At this temperature usage of H2 or other gazeous staff could be as well effective and dangerous.
But taking into account the danger of explosion as well as consideration that amorphous calcined SiO2 is a very fine dust, K2SiF6 + Na route looks more friendly. If your most interest is a thermite reaction, I think F2 is not less active as an oxidiser for this.

[Edited on 11-2-2025 by teodor]

Rainwater - 11-2-2025 at 12:52

Not really intrested in silicon or thermite, just curious to if the OP reaction would take place.
When thermodynamic calculations are so close to the 45 unit margin, I have noticed the reaction can go ether way.
This is due to errors/non-linearity in the listed values or reaction conditions.
Being a value of 100, and something i was easily able to do at the time.
I have another datapoint for my margin of error dataset

teodor - 11-2-2025 at 13:24

It's definitely will take place but the math is wrong because it can take place only on excess of cold alkali metal reacting with a small piece of 2200C amorphous SiO2. There is no experimental evidence that the reaction can start on a lower temperature and I think it is the main reason why the practical way was to react fluoride with Na/K.

[Edited on 11-2-2025 by teodor]

Acetyl - 11-2-2025 at 14:34

Recently I carried out a reduction of SiO2 using Mg.
The amount of Mg used for the reaction was a few percent higher than it would result from stoichiometry. Among the products you can see inclusions of a white substance, most likely unreacted SiO2. The SiO2 came from fuses.

I confirm that we will not get pure Si using this method. The thermite reaction is pretty for show: it proceeds calmly, shining with a strong white light. A lot smoke is generated.

I purified the post-reaction mixture in diluted HCl and washed it several times with water.

I show what I got in the photo. Among the reaction products was magnesium silicide: when pouring Si into the acid, you could hear a lot of shots and sparks above the solution.

The reaction of magnesium silicide with HCl was identical to that in the article:
https://en.wikipedia.org/wiki/Magnesium_silicide