'so damn WET'... Ore Analog consideration

A potential correction on reported yields?

For a good long while, the iron smelt experiments have used an artificial preparation we call 'DARC Dirt'.
The origins of this preparation go back to 2008, when we started on the 'Vinland' series. Here at Wareham, there is no available natural iron ore of any kind. (As I have mentioned in previous posts, Wareham sits at one end of the massive block of limestone that forms the Niagara Escarpment - a block roughly 400 + km long and about 100 m thick!). Although 'primary bog iron ore' can be found 100 km north of here (up on to the Canadian Shield), I have never found any more than the faintest traces of iron locally.

Primary Bog Iron Ore - from L'Anse aux Meadows (early 1970's excavations)

For the Vinland series, we wanted to have a dependable quantity of an iron 'ore', which would also at least approximate the chemistry of that known to have been used at L'Anse aux Meadows by the Norse. Our solution was to mix up an analog, using commercially available 'Spanish Red' Fe2O3 oxide powder.
One huge advantage is that use of the analog has given us a very consistent baseline within our many variations on furnace design and air systems testing. There have been slight variations of this standard mix, usually by adding various amounts of forge scale (so Fe3O4) to 'enrich' the basic mixture.

DD-1 analog as added in a working smelt (Vinland 1)

Details of this mix  *

Chemical analysis
Fe2O3    81  %
SiO2      5  %
CaO      2.3 %
MgO2     2  %
Al2O3    2.6 %
Loss on ignition 7 %

As DD#1
Red Oxide          82 %
Wheat Flour          8 %
Water (after drying) ≠ 10%

Adjusted Fe2O3 content = 66 %

It is the WATER here that may be throwing off our reported yield numbers.
The iron oxide powders used as the basis for the analog mixtures are extremely fine. Originally, in an attempt to match the reported chemistry of the L'Anse aux meadows samples, whole wheat flour was added to create the 'organic' component. The physical process was to combine the dry materials, then add water to make a thick paste, roughly the consistency of butterscotch pudding. This was then ladled out on to trays to dry in the sun. It was found that the flour acted like a binder, producing loose lumps of finished 'ore', which also had about the same consistency (porosity) as the natural ore.

How much of the added weight was water?
Although sometimes a sample of the ore used in a specific experiment is pulled off to be completely dried (Usually baked in the gas forge) this 'loss on ignition' (LoI) is not usually recorded. Typical is about 10 - 13 % of the 'as used' weight.

This all occurred to me after the last smelt experiment ('Icelandic' stone block).
Both being somewhat rushed for time, and with an extremely rainy early Summer at Wareham, I had a lot of trouble getting that analog mix dried from the 'pudding' phase. On dry baking a measured sample, it was found that the LoI was actually almost double - at 18%.

I'm not sure exactly how best to apply is considerable difference (10 to 18 % LoI) to the recorded results. One thing would be to consider 'apples with apples', as most commonly the water weight is * not * considered in past recorded results.
- The difference in water weight on this last smelt is then about + 8%
- Reduce the ore weight to 'standard' (29.3 kg x 92% = 27 kg 'adjusted')
- Effects (increases) reported yield (7 kg from adjusted 27 kg = 26 %)


( * )  If you are interested in the long development sequence of our bog ore analog, the easiest way to read the many earlier commentaries is a simple search :
http://warehamforgeblog.blogspot.ca/search?q='bog+ore+analog'