the 'Espelund Method' - Saturday, Nov. 3

Next Iron Smelt at Wareham

Saturday November 3

I have just come off running two smelts over a single week (!!)

A film company from the UK was here on Oct 10. This was a standard 'Norse short shaft' using our bog ore analog. All went well, even if the furnace was pretty much pulled apart on the extraction. The bloom was a bit spongy, but at 8.6 kg from 30 kg ore, the yield was very good (28%) *

Standard 'Norse Short Shaft' on a stone plinth.

Resulting Bloom : Not well compacted, more work put to slicing when hot.

The second (3 days following!) was using a 'Roman' style tall stack, with passive air. Neil Peterson and I worked from suggestions in Rehder's 'Mastery and Uses of Fire'. The furnace had originally been constructed for Neil's ongoing research into Viking Age glass bead production. The body of this furnace was a typical 30 cm ID, but with a shaft height of 150 cm, extended to a total of 185 cm by a metal collar on top. This furnace body was fitted with a total of four steel pipe tuyeres, each at 3 cm ID, set at roughly our normal height above base (20 cm) and entry angle (22 down).

'Mother of All Furnaces - Iron' - during smelt

Neil had earlier invested in a multiple input data recorder and high temperature probes, so we were able to get temperature readings over the whole course of the smelt. Generally there were interior temperatures in our desired working range (1100 - 1200 C), with a high temperature recorded at 1285 C.

Temperature probes at tuyere +10 / +20 / +50 cm

As a passive furnace cycles over a much longer period, we ended up just leaving the furnace to burn down and cool overnight. This means we will need to cut the walls open to extract the now solid slag bowl mass - which will be hammered apart before we can see the results. Our plan is to make this the first work on Nov. 3. *

Neil and I intend to carefully (!) slice apart the much larger MoAFI furnace. We hope to be able to lift off the top section (about 60 cm of this) and set it down on new base plinth. We will make a second slice at roughly tuyere level, so we can observe and extract that slag mass.
This should give us at least one, maybe two, clay cylinder 'uppers' to convert back to our standard short shaft layout.

the 'Espelund Method'
I owe a long debt to historic metallurgist / researcher Arne Espelund. He was part of the small group gathered by Parks Canada at L'Anse aux Meadows NHSC in 2001, along with archaeologist Dr Birgitta Wallace, staff interpreter Mark Pilgrim and myself. The week long workshop was intended to research and discuss how the original iron smelt by the Norse at that site could be illustrated to the visiting public. My very first smelt was undertaken at that workshop. (In retrospect, what we attempted there would never have resulted in any actual iron production!) This then started me down the path I have followed in the many years since then.

Arne has certainly contributed to the development of a working understanding of historic bloomery furnaces in Europe, through his participation in conferences and workshop events, contributions to journals and publications, and through documenting his own ongoing field research. (1)

At the Heltborg Iron Symposium in 2008, he presented a concept, which he and I (attempted to) demonstrate at the workshop.
In simple outline, Arne had regularly found segregated piles of both iron rich tap slag and lacy metallic / slaggy 'gromps' at historic iron smelting sites throughout Norway. He suggested a mechanism where a two step process might have been used, a first smelt sequence from ore to partially refined 'gromps', with a second smelt sequence then undertaken to concentrate this 'semi product' into a higher quality solid bloom. (2)
At the time (and since) many of us working in Early Iron have questioned the duplication of (massive) effort and materials such a 'two smelt' method would require. However, this does not negate the existence of collected semi-products, or explain why ancient iron masters would have undertaken this sorting and collecting.

So - from an earlier message I had sent to Arne, related to his two step process:

" ... A thought on your theory of 'double smelt':
Micheal Nissen of Denmark has illustrated a sequence of adding iron rich tap slag as the first charges into a hot furnace - before the actual ore additions. In practice, adding about 3- 5 kg of tap slag acts to quickly create an effective slag bowl system inside the working furnace. This means the reduced iron almost immediately starts to collect into a bloom once ore is added. When you undertake this method, there is a significant increase in both bloom density and overall yield. (At least a 5% increase, sometimes more.) The tap slag used is being broken up to the same sized pieces as the ore is before addition to the furnace.
So we now are making our own collected piles of tap slag, held aside to add to the next smelt operation. Seems to conform (at least roughly) to the kind of practice / observed remains that you described much earlier?

I have gone back to collecting up the 'gromps' - those lacy iron foil and slag bits that don't stick to the bloom - or that get knocked off during the initial compaction of the bloom after extraction. Again I break these down to smaller pieces before adding. Normally any given smelt does not allow for much collection of these pieces. Maybe on a 30 kg ore smelt = 5 kg bloom I might end up with about 1 kg of gromps. Not really enough to be able to distinguish an effect on addition.
When would be the best time to add those? Again not sure, not enough tests yet to determine this experimentally.
As these gromp pieces are easy enough to sort, just based on relative weight and visual appearance, collecting them up does make sense. (We can use a powerful magnet, but I never have trouble sorting even without this fast method)... "

Experiment Profile - November 3

I have checked, and to date have accumulated a volume of iron gromps from previous smelts. The total is roughly enough to fill 1 1/2 standard 5 gallon / 20 litre sized pails. I have not weighed this all yet, but expect there is certainly more than enough to equal one of our normal 30 kg ore test smelts. The larger pieces will need to be broken down before addition.
So what I want to do on Nov 3 is run a complete smelt, using these collected gromps. Right now the thought is to use much the same addition sequence as we use for the ore analog. (So start with 1 kg per 2 kg charcoal charge, increasing as the furnace allows.)
I also have about 2 - 3 kg of iron rich tap slag on hand. In keeping with the proven effectiveness of the 'Nissen Method' mentioned above, we will start the smelt additions with this slag. (Should quickly establish the working slag bowl inside the furnace).


NOTE TO INTERESTED OBSERVERS

The fixed iron smelt experiments at Wareham are 'semi open' events.
By this I mean those interested in observing, possibly taking part, are welcome to attend.
Do please contact me before planning to come!
This primarily to organize things like parking and physical support.
I will reply with directions and some notes on dress, etc.


* Reports on both these October smelts are in process.

1) Most significantly :
Arne Espelund
'The Evidence and the Secrets of Ancient Bloomery Ironmaking in Norway'
Arketype, Norway : 2013
978-82-996953-3-6

2) Part of the argument here seems to be the failure (at that time at least) of modern experimenters to produce blooms of the same size and dense quality found in the historic artifact blooms recovered.
My own work, based on the pioneering work of Lee Sauder & Skip Willliams, suggests that increased air volumes inside the smelting furnace will also create historic type blooms.