Some oxides, like sodium and potassium are easily boiled off in furnaces. Unless the stack is very hot, they cool down and get deposited again. In blast furnaces this is a serious problem to refractories. Alkalies attack refractories. Most of the alkalies are deposited on the cool incoming charge, but as soon as they reach a hot zone they boil off again. Effectively they become trapped in the blast furnace, recirculating in the top and dissolving the refractories.

Silica, SiO₂ is first reduced to SiO, which is a gas. It does not condense at low temperature but continues its journey until it meets oxygen. Here is burns; combines exothermically with oxygen to form silica again. This white “smoke” is composed of very fine round particles about 0,2 micron in diameter. This “silica fume” has an enormous surface area per gram, and is a useful by product. It is added to refractories mostly at around 3%by weight. It behaves almost like water it is so fine and the water demand of castable allows less calcium aluminate to be used, and these castables are called “low cement castables”.

Zircon ZrO₂.SiO₂ can be converted to zircon by fusion in an arc furnace in a reducing atmosphere. The best silica fumes are from silicon smelting furnaces and zirconia production. Carbon can be purified by raising it to enormous temperatures, where virtually everything else is volatilised.

Refractories containing carbon are sometimes reduced. Carbon magnesia refractories in basic oxygen furnaces are prone to this reaction. The MgO is reduced. The MgO is reduced to magnesium gas, which travels until it reaches oxygen. It often builds up a solid wall where it is deposited. The solid wall of magnesia can be beneficial in stopping slag penetration, but it leaves behind a “decarburized zone” which is porous from loss of MgO and carbon. I first encountered this reaction about 30 years ago at Cullinan Refractories. We fired some pitch impregnated refractories at 1700°C. when they cam out of the kiln they were an amazing sight. The bricks had bloated, cracked and had hollow pale tubes growing out of the cracks. They resembled potatoes left in a bag too long which had sprouted. What made them look even more weird, was the fact that the tubes grew around the corners of the neighbouring bricks. The growth of the tubes was following air flow. While they were growing, they were filled with magnesium gas, which deposited MgO on the tube end where it encountered oxygen. Initially, the carbon burned out of the surface of the bricks. Once the temperature reached high enough to start decarburization, the magnesium deposited a solid layer on the edge of the unburned brick, sealing off the decarburized zone. The magnesium gas which formed the tubes.

I have seen this reaction in aluminosilicate bricks with organic matter added for insulation porosity. There was no bloating or tube formation, but the dense layer and decarburized zone are visible. More than one dense layer can be seen as the process continued.

Silica can also be volatilized by steam at high temperatures.