Monday, May 04, 2009

L'Anse aux Meadows Smelt - Ore Analog

A proposed mix for the DARC Dirt 2 - LAM analog:

A current project now in its initial stages is DARC working towards a possible full scale interpretive presentation at L'Anse aux Meadows NHSC in summer of 2010. July of that year marks the 50th anniversary of the archaeological site by Ingstad and Stine. Part of the presentation by DARC will include an iron smelt, replicating the one undertaken by Leif Eirikson's crew some time about 1000 AD (the first iron production in North America).

Readers can expect to see a number of commentaries and reports on this specific experimental series as the work progresses from the theoretical to the practical.

There are a number of elements that will frame the reconstruction smelt:

- Work inside the confines of the 'Furnace Hut' structure found at LAM. (a roughly 3 x 3 metre space, roofed and open one side )
- Furnace construction of free standing stone slabs.
- Air provided by human powered double bag bellows type.
- Ore used is local primary bog iron ore
- Determination of tuyere type and layout (unknown)
- (on site production of charcoal as related demonstration / project)
- (on site use of stone surface for consolidation process)

The core of the smelt process revolves around the ore. If you have been following the many notes related to experimental iron smelting, you have seen how historically it was primarily the ore that determined the location of the iron smelting. Ore also shapes the dynamics of individual furnaces.

The largest element of the Norse decision to attempt to smelt iron at Vinland around 1000 AD is the ore itself. When the boat crews cut and pulled up the peat blocks to build their over wintering houses, they would have exposed a large quantity of primary bog ore. Now, there are some differences of opinion of just why Leif's crew undertook an iron smelt, but the important fact is that they in fact did.

I had been provided with a small amount of the ore off the archaeological layer, as uncovered by Dr. Birgitta Wallace during excavations in the early 1970's. Of course there are a number of potential problems attempting to match this sample to what might have actually been available (even at exactly the same spot) 1000 years ago. As the formation of primary bog ore is a chemical process with a very large organic component, a change in physical environment could also change the material. More important for this specific series, we certainly will not use actual bog ore for the many tests. (There is a very good chance that even for the on site demonstration this may not be possible. Parks Canada does maintain a very strict 'no environmental impact' policy on all of its sites.)

The original LAM ore sample provided by Birgitta (tested by R. Hansen for Arne Espelund) shows:

Fe2O3 - 89.5
( Fe - 62 )
SiO2 - 1.24
Al2O3 - 2.45
MnO - 5.33

One of my largest concerns right now is about silica content - the components that make the slag bath. Arne had commented back then (and I totally agree) that the archaeological sample was lacking in silica - and would make for a 'dry' smelt. The silica / slag is important for controlling the final carbon content of the metal. Too little slag, and there is a good chance the metal will absorb way too much carbon, resulting in a cast iron material (which can not be forged). Our past experience is with clay structures, which at smelting temperatures melt to create a surplus of slag (if anything). The rock available in Ontario is not a perfect match for the basalt type stone in north Newfoundland. (Even at that, that source stone has to be gathered up off the Canadian Shield, considerably North from Wareham!) The temperatures inside a working smelter are high enough to melt even basalt, but the rate of melting on Ontario types, thus slag creation from the stone, may both vary and be important.

The obvious solution to the problem of available ore materials is to turn to an analog. This is in fact something we have already worked considerably on, building on Robert Gissing's early work on DARC Dirt. I have also been trying to pull together reports and data detailing the chemical components of the various ore types that have been used in past smelts. Comparing it with some of our other successful ore / smelt combinations, it looks like we have got the best results with a silica contribution from the ore of something in the range of 5 plus %, with the combined SiO2 and Al2O3 at roughly 8 %.

Utilizing a 'Black Iron Oxide' from the pottery supply:

Fe3O4 - 93
( Fe - 67.3 )
SiO2 - 3.5
Al2O3 - 3.5
MnO - 0.0

Using that as a base, and if we disregard the 10 % flour mixed in as a binder (as it will cook off anyway and does not effect the iron chemistry), I'm suggesting we add 5 % silica to make for a bit more 'juice. That would leave us with something :

Fe3O4 - 84.4
( Fe - 64.1 )
SiO2 - 8.8
Al2O3 - 3.3
MnO - 0.0

When we mix, the ratio would be 10 % flour, 85 % black oxide, 5 % silica - not counting the water weight. Past experience has shown that a good 8 kg or so of ore is required to establish a working slag bowl inside the type of iron smelting furnace that will be used here. The target yield will be roughly 3kg (plus) as that is the estimate of what was produced originally. With a fairly normal yield expected at approximately 25 - 30%, a minimum of 15 plus kg of ore analog will be required for each full test smelt.

Take a look at a table showing the major components of past ore types

No comments:


February 15 - May 15, 2012 : Supported by a Crafts Projects - Creation and Development Grant

COPYRIGHT NOTICE - All posted text and images @ Darrell Markewitz.
No duplication, in whole or in part, is permitted without the author's expressed written permission.
For a detailed copyright statement : go HERE