I want to extract elemental silicon from pure silicon dioxide powder.

From what I've read, one way to do it is by adding carbon from graphite at very high temperatures, which is difficult.

A better way to do it is to mix with magnesium powder and heat it to about $800 ^\circ \rm C$.

Some places mention the use of a hydrogen atmosphere, is that necessary?

What is the best way to do it, the best temperature? If I use a torch, wouldn't the exhaust gases purge the oxygen and serve as an inert atmosphere?

How do I know the reaction is over?

  • 3
    $\begingroup$ The exhaust gases, mainly CO2 will react with Mg at 800C. Nitrogen reacts with Mg at 800C to form Magnesium Nitride en.wikipedia.org/wiki/Magnesium_nitride. This is why an unreactive and reducing atmosphere of hydrogen is used. Argon might would probably be OK. This is not something to be attempted without the proper equipment $\endgroup$
    – Waylander
    Dec 19, 2019 at 13:40

3 Answers 3


Making elemental silicon is easy, but not for amateurs. Since it is cheap (~$2/kg) there is little reason to do this yourself.

Industrial processes do indeed mix silica with a source of carbon like coal. These react when very hot to give silicon. But the process usually uses an electric arc furnace and controlled conditions with carefully controlled use of gasses (including extra oxygen late in the process to enhance the amount of impurities that gather in the surface slag). It is unlikely that an amateur can easily reproduce this control.

You might get a result at lower temperatures using magnesium. But you will struggle to do this safely without some control of the atmosphere above the mix as magnesium will burn violently in air (or even in carbon dioxide). You might even need to use argon to blanket the mix as even nitrogen will react with magnesium.

So, basically, this is not something to contemplate as an amateur outside a well-equipped lab.

  • $\begingroup$ Assuming you do not burn the magnesium, it can still react with the silicon forming $\ce{Mg2Si}$. Carbon is less prone to a similar side reaction. $\endgroup$ Dec 19, 2019 at 17:38

Silicon dioxide is a very thermodynamically stable compound. In most labs I have been in, $\ce{SiO2}$ is equated to sand whenever we aren’t explicitly talking about silica gel (or sometimes celite, but that tends to get referred to as celite). Considering how hard it is to make anything out of sand that isn’t essentially still a $\ce{SiO2}$-derived compound should point to how difficult it is to break the $\ce{Si-O}$ bonds.

To make elemental silicon from the dioxide, you need to reduce it, so in theory all your standard reduction agents should work. Carbon is considered to be one of the cheapest, easiest to use and most readily available reducing agents (except maybe the cation which requires liquid phases). That, the equally high stability of $\ce{CO2}$ and the fact that the transformation is entropically favourable as a gas is generated from two solids make carbon an obvious choice in industry. (Note: I did not go through the trouble of looking at the actual processes, I am just deriving the obvious constraints from my chemical background knowledge.)

However you are going to synthesise elemental silicon, you need to ensure that it doesn’t get reoxidised rapidly. The higher the temperature, the more activation energy is supplied and the more likely it is for the back reaction to occur. A hydrogen atmosphere serves a double purpose: not only does it displace oxygen preventing reoxidation but it is also a reductant in itself and may aid (given appropriate conditions) the reduction of $\ce{SiO2}$. In case of a highly reactive reductant such as magnesium, it also prevents oxidative side reactions of magnesium to a certain extent.

All that said, industry (especially semiconductor industry) has a huge demand for silicon, meaning a lot is synthesised industrially; the raw materials (sand and coal) are almost literally dirt cheap and thus so is the price of industrial silicon. As very high purity silicon is needed in certain applications, you can likely acquire it from crude unpurified to extremely high purity, depending on what your situation requires. There is really little point in trying to make it oneself.


Another way of getting Silicon, is first to prepare SiCl4 by having a current of Chlorine gas crossing a horizontal tube filled with a mixture of silica SiO2 + charcoal C in a flame, but not at a as high temperature as the electric furnace mentioned by Matt black. The following reactions happens : $$\ce{SiO2 + 2 C + 2 Cl2 ->SiCl4 + 2 CO}$$ SiCl4 is a gas above 57°C. It is condensed by cooling the gas mixture getting out of the reaction tube, than purified from impurities like AlCl3.

Later on, SiCl4 reacts with magnesium at high temperature, according to the reaction : $$\ce{SiCl4 + 2 Mg -> Si + 2MgCl2}$$ This produce pure Silicon, and MgCl2 which liquid à 112°C.

  • 4
    $\begingroup$ Magnesium chloride is not liquid at 112°C. Wikipedia gives a melting point of 117°C for the hexahydrate if heated rapidly, but this system is clearly anhydrous. The relevant melting point, again from WP, is then 714°C. $\endgroup$ Dec 19, 2019 at 16:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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