A Coffee Can Foundry... (making bronze, part 5)

This is Part 5 of my mini-saga on attempting to make home-made bronze, using a method similar to Bronze Age technology. Parts 1, 2, 3 and 4 can be found here:

Part 1: Goal, building and initial tests
Part 2: First attempts at metal melting
Part 3: Improving the foundry
Part 4: Lump charcoal is as lump charcoal does...

Summarising the goal of the project...

I'm now close to achieving the stated goal of the project and it's worth summarising what the object of this little adventure is. Basically I'm trying to produce small amounts of castable bell bronze (20 - 25 w% tin) using a method close to what our Bronze Age ancestors would have used in all likelihood, although they would have done this on a large scale. Bronze is an alloy of copper and tin and both metals can be obtained by reducing their oxides by means of a chemical reduction with carbon. The oxides occur readily in nature as ores, in the case of tin a mineral called Cassiterite and in the case of copper, Cuprite (for both metals also other ores are also available). Cassiterite is mainly tin oxide, with small amounts of iron oxide as a contamination, while Cuprite is mainly copper (I) oxide. Carbon is of course most commonly used in the form of coal or charcoal.

By glowing the oxides with the right amount of carbon, the carbon recombines chemically with the oxygen present in the oxide, thereby being carried off as carbon dioxide and leaving the "pure" metals behind. I say "pure" because this is a technical process and the metals are therefore always slightly contaminated.

The process of extracting metals from their oxides by means of heat and a reducing agent (beside carbon there are many others such as aluminium, magnesium, calcium, lithium and others) is known as pyrometallurgy and our Bronze Age (and Iron Age) ancestors were master pyrometallurgists. We still are today, even more so...

My objective is to carry out both reductions at the same time on an appropriate mix of Cassiterite and Cuprite. Both constituent components of bronze would then form at the same time, fuse together and bronze metal would be the result.

The Cassiterite reduction

Cassiterite, a mineral once found and mined in abundance in Cornwall, South England, is a little hard to come by today, especially in small amounts. Instead I'm using the purer yet cheaper form which is essentially technically pure tin oxide (SnO₂). This is an off-white powder with a slightly greenish hue. It contains about 79 w% of tin metal.

Mixed with the theoretical, chemically required amount of carbon (I used the same charcoal that I use as fuel for the charcoal-fired paint can foundry but added some 20% extra to the mix to account for some that would simply burn off at these temperatures) and glowed to about 900 C (1650 F) or higher, the reduction reaction proceeded fast and easily.

It's interesting to watch, as the liberated by-product (carbon dioxide) puffs up the hot, powdered mix a little, then proceeds actually to bubble out of the mix. The powder is behaving therefore much like a boiling liquid that reduces steadily in volume, until only a small puddle of molten tin metal remains (MP 232 C/450 F). It's perfectly possible to keep topping the crucible with fresh "Cassiterite"/charcoal mixture, which then continuously reduces to pure tin metal, thereby filling up the crucible with the desired metal. The crucible in question here was a stainless steel former sugar bowl...

During the early stages of the reaction, the mixture darkens somewhat, probably due to the formation of finely divided tin metal, bubbles of carbon dioxide start to form and then small pearls of tin metal start appearing, which then later settle at the bottom of the crucible as a puddle of molten metal.

The "pearling effect" is apparently something that can be witnessed by putting some Cassiterite (you can find it in small quantities in mineral shops or around the now mostly abandoned Cassiterite mines in Cornwall) on the fire of a hot charcoal fired barbecue: the combination of heat and carbon causes some of the Cassiterite to reduce to tin metal. This is one of the most likely routes of discovery of this useful metal by our ancestors...











Right, two tin buttons (29 g) made by reducing tin oxide, left a copper blob (13 g) made by reducing copper (I) oxide. The copper is porous and "bubbly", probably because carbon dioxide got trapped during solidification. The white spot on the copper is a shard of porcelain from the broken crucible.


The Cuprite reduction

For basically the same reasons as above, I substituted the Cuprite ore with technically pure copper (I) oxide (Cu₂O), which is a heavy, wine-red powder. It contains about 89 w% of copper metal.

Other copper ores contain significant quantities of iron and sulphur and it's much harder to pyrometalurgically extract the metal from these. It is reasonable to assume that our ancestors did eventually master the art of extracting copper from the more unyielding ores but that they preferred (and started out with) the copper (I) oxide based ores, like Cuprite. Even today, I bet Cuprite ore still fetches a better market price than the other ores, because of its ease of conversion to the desired metal...

Mixing the "Cuprite" with charcoal and glowed along the same lines as above, this reaction proceeds also steadily although it all looks quite differently. The mixture starts darkening and glowing very quickly with the occasional green flame (typical of copper) erupting. At first I thought that the copper (I) oxide (Cu₂O) was merely oxidising further to copper (II) oxide (CuO is black) but at the end of the test I prised open the contents of the cooled crucible (breaking it in the process!) pure, slightly spongy and porous copper metal was clearly there. A quick estimate based on the weight of the crucible and its contents before and after firing seemed to indicate that the reaction copper (I) oxide + carbon --> copper metal + carbon dioxide had proceeded stoichiometrically. That requires some corroboration though: it's possible that some oxidation of the copper (I) oxide to copper (II) oxide did indeed take place and that not quite enough carbon was present to reduce the copper (II) oxide to copper metal...

Although the metal almost certainly formed in the liquid state, I didn't manage to keep it molten: copper's melting point is 1084 C (1984 F) and although I've melted both brass and copper in my dinky toy furnace before, it requires maximum running temperatures to achieve that. I may have been slightly short of that and it was only the very first test, cut slightly short because it started raining...















The paint can furnace during firing...

So, on the Cuprite side of things some problems remain to be resolved. The main problem is to achieve the high temperature needed to keep the copper metal molten and that entails a highly temperature resistant crucible. On this first test I used a 60ml chemlab ceramic crucible. Whilst this definitely resists that kind of temperatures (and higher), the material is also very fragile (hence the breakage when prizing the metal out) and not really suitable for this kind of thing. Some reading and some advice by the foundry enthusiasts at the ABYMC forum would indicate that in all likelihood a clay-graphite crucible is in order for more experimentation. This is definitely the material of choice for bronze temperatures. Steel can survive these temps as well, but copper is said to dissolve some chemical components of steel, making the material vulnerable in this usage.

Next stop: actually making bronze metal or so I hope...

Anticipation: three roads to bronze...

It's possible to envisage four possible routes to bronze, during the Bronze Age:

1. Two reductions and one alloying smelt:

This route would involve running the Cassiterite and Cuprite reductions separately, resulting in ingots of both tin and copper metal. Melting these together would yield the desired bronze alloy. This would require three furnace firings.

2. Two reductions, one of them being also an alloying smelt:

This would involve making either copper or tin, then loading some of either metal into a reducing crucible and adding the reducing mix for the other metal. The pre-loaded metal and the newly formed metal would then result in bronze. This would require two furnace firings.

The reduction of copper oxide in the presence of the right amount of tin metal would appear to be the most attractive here because of the lower melting point of tin.

3. Two reductions carried out simultaneously:

By reducing a mix of Cassiterite and Cuprite (mixed in the correct ratio) a bronze alloy could be obtained using one firing only. It's my contention that that is possible and that our ancestors would have cottoned onto that too. Wait and see...