DIY refractory from aluminium foil

A simple method of making DIY refractory from aluminium foil and sodium silicate.

I chanced upon this refractory while attempting to use DIY sodium silicate as a refractory glue to stick onto aluminium cooking foil. Well, it certainly became very good refractory glue, but to my surprise, it fizzed, bubbled, got very hot, blew off hydrogen gas and tuned into new refractory material.

I reported on this refractory and many others in a rather large DIY refractory post where it may not be found. So I am revisiting it here with a few more ideas as I think it deserves its own post. I hope that this simple and versatile refractory get the attention that it deserves, through the SEO process, so that it can be added to the refractory toolkits of those who tinker with fire.

This DIY aluminium refractory is most suited to situations where it can be put or pressed into a cavity between two surfaces in order to expand to fill or seal a void and form a strongly bonded refractory fill that is very lightweight.

The DIY refractory aluminium and sodium silicate

The silicate render. The DIY sodium silicate used in this refractory is quite alkaline and its preparation is described in the original DIY refractory post. I make a sodium silicate render with this Recipe; 100g DIY sodium silicate concentrate, 10g baby talcum powder (2 tsp.), 9g iron oxide ( Concrete coulouring pigment 2 tsp.) and 40g water. The quantities of each are not critical and the sodium silicate/water ratio will vary according to the strength of your particular concoction of silicate.

The aluminium foil. There are many ways to shape and prepare the foil to make it into the refractory depending upon the task. My most popular method is to coat the foil with a brushed layer of the above render (An old tooth brush works well). Then I fold the foil sheet over on itself and apply more render and then fold again. When the desired shape is achieved, I render the two remaining surfaces with render before the foil sandwich is rolled up around a fitting or simply stuffed into a void between two fittings (or a mixture of both).

Aluminium refractory reaction time. The initial reaction between the silicate and the aluminium is slow at first and gets quicker as the exothermic reaction gathers pace with the generated heat. Consequently, only a small amount of aluminium foil should be prepared for placement before the reaction progresses too far. “When left alone too long the prepared aluminium laminates will blow up like puff pastry pillows with hydrogen gas and be difficult to manage. So no naked flames unless you want some extra heat!”

Slowing the aluminium reaction speed. The reaction can be slowed somewhat by less vigorous brushing of the render and more particularly by not including the aluminium oxide grits in the render. “I think these actions slow the eventual destruction of the protective oxide film on the aluminium foil that then causes the reaction rate to rapidly build up. Fewer layers of aluminium in the refractory laminate also slows the reaction, presumably because there is less build-up of heat.

Speeding the aluminium reaction time. Adding some aluminium oxide (fine automotive ‘sandblasting’ grits) to the render and more vigorous brushing can greatly speed up the reaction. Also, the crinkling action of stuffing the aluminium/silicate composite into its final resting place or void also speeds up the reaction*.

“Bringing a butane torch to play on the new refractory makes the ultimate accelerant. As a bonus, there will be a display of hydrogen combustion and the sweet rose perfume of from the baby powder. That’s if it is not overpowered by the smell of burning hair! I think this is another wonderful example of ‘being born in fire to survive in fire.”

*Note: As with some of my other refractory materials, I am learning that the more obvious reaction of a refractory mix into a solid item is just the start of the journey for most refractories. Most undergo substantial chemical, physical or phase changes during deliberate firing to high temperatures or exposure to those temperatures during use. Such changes are often invisible to us but are beneficial. A virtue of this refractory is that it instantly becomes stable and fire-ready once heated to a moderate temperature of about 200C. Then it can withstand the heat of forced air charcoal combustion.

Layering up the aluminium refractory composite. As with all good DIY tinkering methods, this refractory is very tolerant to being layered up with successive batches of the refractory. The fresh material bonds well to the older material that has just ‘set up’ or has been cured at high temperatures.

This means that if there is insufficient refractory composite to fill a void, or the composite does not expand as much as expected then an extra batch can be made and added sooner or later. It also means that a small amount of the composite can be used to pack around a component that is held in a precise location with a jig. After a quick flaming and a fast set-up of the composite, the jig can be removed and the rest of the void can be filled with more composite.

Lastly, where the light weight of a refractory component is critical, it means that very light and foamy refractory can be formed in the middle of an item and the exposed faces can be made much stronger with the denser refractory to protect the inner core.

Aluminium refractory curing. The first stage of the refractory composite curing process happens by itself. and the liquid sodium silicate first goes to a toffee texture and then eventually to a solid.

During the liquid phase, hydrogen is produced and some of this forms bubbles in the liquid phase and will cause the black liquid to spew out excess foam. Eventually, the liquid will harden with many bubbles trapped in the matrix, leaving an ultralight refractory.

The next phase of curing requires an input of external heat. If heated hot enough the silicate seems to turn into a glassy state.

The density of packing of the aluminium refractory. If the coated aluminium composite is packed in lightly it will expand considerably and become a very lightweight, delicate and thermally insulating refractory with less strength. It can be packed in more tightly to give a denser and stronger refractory. In either case, the void will be filled and probably have an overflow of bubbly silicate refractory foam.

Aluminium refractory finishing. The foamy cellular nature of the refractory means that excess foam that leaks during the process can easily be removed with sandpaper. Sometimes a stronger finish may be required, so I find that the same render, with a little extra aluminium oxide added, makes a compatible and mechanically strong hot face that is easy to apply.

Compressive strength. The light and foamy nature of the aluminium refractory means that it can be crushed or frittered away when it is not contained within a protective shell. However, when a flimsy shell is filled with the refractory, together they become immensely strong and rigid.

Why is iron oxide in the mix? Although refractory, iron oxide has no special refractory properties in the mix (as far as I know). It has two purposes for my applications. Firstly, it pigments the sodium silicate liquid refractory mix so that it is easy to judge the depth of cover when applying the render.

The second reason is that the resulting dark refractory colour is highly emissive and is great for maximising the radiant heat that is emitted by tiny winter cooking stoves. “As a bonus extra, its colour transformation from dark grey to reddish colours indicates where a hot device is getting the hottest. It leaves a heat map over time.”

Why is baby powder in the mix? The talcum powder is brilliantly fine, is safe and pleasant to use and is about as refractory as asbestos without the risks. Consequently, I use it as a refractory filler to increase the viscosity of the render so that it can increase the application depth during brush rendering while minimizing ‘running’.

The refractory reactions

When I first posted on the aluminium refractory, I thought that the reaction was simply aluminium reacting with sodium hydroxide from the alkaline sodium silicate to produce aluminium oxide* with a melting point of 2,072C. The equations are scratched onto the first photo above from some time ago.

Then I thought that the aluminium oxide simply became cemented with some silica gel crystals (MP 1,710C) that came from the degradation of the silicate. “In other words, I thought that this refractory was simply aluminium oxide glued together in a matrix of refractory silica gel.” However, this is probably not the true story as the reactions involved are more numerous and complex.

*Note: My description of the reaction is probably incorrect as indicated in this reference. They indicate that in alkaline conditions, the reaction progresses to sodium aluminate (anhydrous NaAlO2 or hydrated NaAl(OH)4 or polymerised forms Na2O. Al2O3 or Na2Al2O4). However, in our aluminium refractory reaction, the neutralising effect, on the hydroxyl ions, of the added aluminium and later absorbed atmospheric carbon dioxide probably will drive the reactions toward oxide formation. Hydrogen gas liberation supports this later conjecture.

Aluminium sodium hydroxide reaction. This is the first and most obvious reaction is the formation and precipitation of aluminium hydroxide (or sodium aluminate or whatever) as the aluminium foil reacts with excess sodium hydroxide from the alkaline sodium silicate solution. The early release heat from the exothermic reaction to aluminium hydroxide and the release of hydrogen is evidence of this early reaction. This reaction, by consuming hydroxyl ions, makes the solution less basic and I think this causes the next reactions.

Precipitation of silicates/silica gel. Having made the solution less basic (lowering the pH), the silicate ions become less soluble and start to precipitate as condensed sodium silicate polymers and or as silica gel. This is evidenced by the rapid and total hardening of the liquid refractory mix.

Silicate reaction with carbon dioxide. There also is another very common reaction between sodium silicate and carbon dioxide. “Without quite knowing quite why or how I started to make this reaction happen quickly by putting my beloved bespoke refractory items in the carbon dioxide-rich gases in my wood heater. I had the mistaken belief that the process was just one of drying.” The method dries very well and also greatly accelerates this inevitable reaction with carbon dioxide. I have not found a direct reference to this reaction, but I know the reaction is commonly used in making refractory metal casting molds from sand for casting metals such as iron at very high temperatures.

This article titled; Sodium Silicate Binders for Green Sand Metalcasting Foundry has a nice description of the reaction:

“The basic hydroxyl component of strongly alkaline sodium silicate reacts with atmospheric carbon dioxide which acts as When a sodium silicate solution is mixed on sand and then cured with CO2, complex chemical and physical changes occur that contribute to the hardening of the core or mold. The bonding phase that holds the sand together consists of (1) precipitated silica gel formed as a result of the reaction between CO2 and sodium hydroxide and the accompanying pH change and (2) dehydrated sodium silicate formed by removal of water as air passes through the sand.”

Hill and Griffith

With this in mind, I think that in the aluminium refractory situation, the aluminium rapidly uses up the most of hydroxide from the alkaline render. If there is any leftover hydroxide it will slowly be neutralized by atmospheric carbon dioxide. Consequently, the pH will rise sharply to a point where the silicate will become insoluble and form crystals of various silicates and some water will be held in the crystal structure.

Further reactions and phase changes of silicates with high heat treatment. During the strong heating of the already set aluminium refractory matrix, there will be dehydration and more changes to the crystalline silicates, the formation of amorphous silica gel and at the highest temperatures the formation of silica gel glass. References to this process are difficult to find.

However, I found a paper titled; ‘Phase evolution on heat treatment of sodium silicate water glass‘. It gives what I think is a good description of the changes, that I observe while tinkering with refractories made with sodium silicate as they are fired (or used) at progressively higher temperatures. Here is the abstract from the paper:

“Heat treatment of sodium silicate water glass of the nominal composition Na2O/SiO2=1:3 was carried out from 100°C up to 800°C and the advancement of the resulting phases was followed up by powder X-ray diffraction, scanning electron microscopy and thermogravimetry along with differential thermal analysis. The water glass, initially being an amorphous solid, starts to form crystals of β-Na2Si2O5 at about 400°C and crystallizes the SiO2 modification cristobalite at about 600°C that coexists along with β-Na2Si2O5 up to 700°C. At 750°C Na6Si8O19 appears as a separate phase and beyond 800°C, the system turns into a liquid.”

R. Subasri & Helfried Näfe

Well, I have never seen my refractory turn to a liquid at a temperature of 1,000C or more. So I guess that the incorporation of aluminium in the mix may be forming some sort of sodium alumino silicate that avoids this melting.

Possible aluminosilicate formation? In this aluminium-based render, we will have abundant aluminium ions available for reaction with the various forms of sodium silicates, so sodium aluminosilicate (that is common in natural mineral formation) may also be formed during high-temperature firing.

My earlier description of this complex series of reactions was incomplete and probably incorrect, so I thought it would be a good time for a reflective ode about being wrong:

Do we all have the right to be wrong?
The same book be the source of our song?
Real truth is a progressive and evolving thing,
Not dangling on one influential persons string,
Adjust, we must, when the old truth just don’t belong.

Conclusion

Maybe the chemistry theory about the curing of this aluminium refractory is a bit sketchy, but I am sure that I have never managed to destroy it or detrimentally melt it with a butane torch with a temperature of 1,430C or 1970C.” More understanding of the changes that happen in the formation of this aluminium refractory should make its application better, even if that knowledge is imperfect.

I hope this makes a small improvement to our understanding of this handy DIY aluminium refractory and gives the method/s a little more chance to be found by others to add it to their fire tinkering tool kits.

Tim