Archaeologia Islandica - 01.01.2007, Page 58

Arne Espelund
MnO FeO Fe203 Si02 CaO Na20 K20 P205 Sum R
%wt 2.1 68.9 (75.4) 18.1 2.0 0.5 0.7 0.52 92.82 3.3
Table 4. Analysis of slagfrom L’Anse aux Meadows by Rosenqvist (1977/85:425).
were roasted at 600 °C in the laboratory
prior to analysis. Due to a high content
of Si02, sample No. 1 would not give any
iron upon reduction. The content of MnO
is very high, reminiscent of the content in
slag from Rauset in Hallingdal. Sample
No. 2 is comparable to the ore presented
above. It is low in P205- The three sets
show a considerable variation of analyses,
especially for the important oxide Si02. A
good ore should contain next to 80-90%
Fe203 some 5-10% Si02. The analyses,
falling into two groups, are not too differ-
ent from the values given for 10 samples
of ore by Unglik & al (1999:220). The
spreading may have contributed to the
problems of amateur smelters. Both ores
presented above, one useless, the other
very rich, will look alike.
The total amount of slag found
at L’Anse aux Meadows is 15 kg. The
pictures labelled as figures 46 a and b in
Rosenqvist indicate slag that had solidi-
fied from a homogeneous melt and there-
fore in principle could be the waste mate-
rial after proper smelting. The analysis
runs as shown in Table 4 (sample I 21
-LaM 283 (Rosenqvist 1977/85:425). It
is unlikely that iron is present as Fe,03.
In the table therefore the amount 75.4
for Fe203 has been replaced by the cor-
responding amount of FeO. The sum - a
bit low - is also based upon FeO as the
constituent. Again A1203 is missing. This
slag is shown as a polished section. It
is claimed that the slag contains wustite
FeO and iron in a glassfayalite matrix. The
picture was obtained by optical micros-
copy (Rosenqvist 1977/85).
An important find of iron was a
rivet LaM 60, presented as a micrograph
in fig. 21. It shows ferrite with a small
amount of pearlite, which may represent
a total of less than 0.1% carbon. Fig. 20
shows the same rivet with pure ferrite and
slag stringers, claimedto be single-phased.
In a table the same rivet contains about
10% A1 and 20% Si, obtained by optical
spectrography, telling that the arc of the
instrument hit oxides only, present as slag
inclusions (Rosenqvist 1977/85:408). The
method now is obsolete, replaced by elec-
tron optical methods.
The books by Unglik and Stew-
art (1999) represent a very comprehensive
approach to the problems of bloomery
ironmaking, related to the finds at L’Anse
aux Meadows. The contribution covers
361 pages (2 volumes) and more than
100 pages with illustrations (1 volume),
in addition some 100 pages about copper
alloys, conclusions and a list of cited lit-
erature with more than 700 entries. The
approach can be classified by archae-
ometry, rooted in physical metallurgy.
The investigations included microscopy,
chemical analysis, scanning electron
microscopy SEM, X-ray diffraction, meas-
urement of density, magnetism, hardness
etc. Analysis by means of SEM repre-
sents a major advancement for analysis of
solid samples. Unfortunately the authors
do not present the dual problem related
to carbon control of the resulting metal
in the presence of incandescent charcoal
and the risk of FeO-depletion from the
slag, rendering a solid Si02-rich slag. By
leaving out a thermodynamic reasoning,
56