A Coffee Can Foundry... (making bronze, part 1)

As a chemist I've always wondered how early humans, without any knowledge of modern Chemistry, managed to make metals as versatile as bronze and iron. Although these discoveries have been understood for quite a while now, it remains to me quite amazing that the earliest tin bronze has been dated back as far as the late 4th millennium BC!

And so I decided to go on a bit of an "amateur chemico-archaeological dig", to find out how hard or easy it is to make bronze, armed with sound knowledge of the chemistry involved but only equipped with a kitchen laboratory (much to my wife's perennial chagrin). This is the first instalment of this little (?) journey...

I will not bore my readers with much chemical jargon or chemical formulas. The relevant details can very easily be found on t'Interwebs anyway. It suffices to know and understand three rather simple notions:

1. Traditional bronze is an alloy of copper and tin, in varying ratios, depending on desired properties of the alloy.


2. Copper can be made by chemical reduction of its oxide (copper oxide) to the elemental copper metal, using carbon (charcoal or cokes) as a reducing agent. The carbon "burns" off as carbon dioxide, leaving the metal behind. This reaction requires considerable heat to get started.


3. Tin can be made by chemical reduction of its oxide (tin dioxide) to the elemental tin metal, using carbon (charcoal or cokes) as a reducing agent. The carbon "burns" off as carbon dioxide, leaving the metal behind. This reaction requires considerable heat to get started.

The operative word here is definitely heat, read furnace. As the early bronze makers didn't have natural gas, propane or butane at their disposal, their furnaces would have been charcoal or wood-fired. And so the starting point of my journey became the design and construction of a small, simple and inexpensive charcoal-fired furnace.

After a bit of surfing I came across a very simple but rather perishable design for a so-called Coffee Can Foundry, here and its more robust follow-up, CCF Mark II, here. I decided to go for design II, using a 2.5 l empty paint can.

Simply put, the construction involves casting a cheap refractory material which is essentially a composite of Perlite (volcanic ash and also a common gardening material) and common fire cement, into a home-made mould. I used my paint can as the outside and an empty squash bottle as the inside of the mould. The rest involved some drying and baking (while the wife wasn't in sight) in our kitchen oven. Here are some pictures of my creation, shortly after the first tests today:







Before attempting to reduce the metal oxides, I needed to run some tests and get a bit of experience with this primitive piece of kit. One good way of getting to know the Paint Can Foundry, is to attempt to melt aluminium. Aluminium is readily available in the form of empty beer or pop cans and has a melting point of 660 degrees centigrade (DC). So, basically, if your foundry can melt aluminium, then at it has least reached that temperature.

My first attempt at getting a fire going involved a hairdryer and ended up with... no fire and a hairdryer, battered, broken and in the bin. Needless to say, the ancient furnaces would probably have been fanned by means of bellows (possibly powered by water, wind, animals or slaves).

So instead I got myself a second hand Hoover, more or less dismantled it to get access to the airflow (previously to the now defunct bag) and started the first serious trials.

The vac now blows into a cardboard funnel at the end of the tuyère (air inlet, in foundry jargon) and I can regulate the flow by increasing/decreasing the distance between the funnel and the (very strong) airflow from the vac.

It didn't seem to bode very well at first with having difficulty getting the charcoal to light, using methanol. But I kind of got the hang of it after a bit.

Eventually, after some mucking about, I got the fire to burn a bright amber-red (about 800 DC, at a guess). The first attempt at melting some aluminium can material failed because it turned out to be a steel can!

The second attempt also failed, using some aluminium from a coke can. Even the third attempt, using household aluminium foil, failed. Then I decided to try and put the lid of the paint can on the furnace and bingo, the foil melted alright! I added some coke can aluminium and that started to melt too but by then the charcoal had burnt out. So I definitely got past the 660 DC mark... Some more tests with smaller, cut up pieces of can material, rather than the quite large chunks I'd been using so far, were also successful but excessive dross from the paint and various can coatings didn't really yield anything castable, so no aluminium ingots today.

The main problem remains getting the airflow right. The other problem is that the charcoal really burns fast and the charge (about a good handful) burns out in about 5 minutes. Recharging is possible but doesn't really solve much because by the time I've got a good glow going again, the crucible (well, a fashioned steel can, actually...) has then cooled down almost completely. Propane, here I come, I guess. The foundry was actually designed with a propane burner (to be inserted through the tuyère) in mind, should later I want to experiment further.

I'll probably have to fool around a bit with the size of the charcoal chunks too...

Nonetheless, with some practice I feel it should be possible to run the copper oxide and tin dioxide reductions with charcoal without too many problems (famous last words!) The furnace might not reach the 1084 DC required to melt copper (or keep it molten) but the reaction enthalpy (heat released during the reaction) should help. Tin melts at 232 DC only, so no problem there.

The furnace itself held up very well though: no discernable cracks, just a little shrinkage, probably due to driving off some moisture.

Next time I'll try and cast some aluminium and melt some copper coins. Questions in the comment section, please...

---

Update:

Pictures of the operating furnace and the results of the first melts can be found in this post.