The Identification of Slags from Archaeological Sites

Hans-Gert Bachmann

• Editore: Routledge
• Anno: 2016
• ISBN:
  9781315418193
• Edizione: 1
• Argomento: Sciences sociales
• Categoria: Archéologie

This is a guide to the field identification and laboratory analysis of metallic slags found in archaeological sites.

Part I: Slags in General

INTRODUCING SLAGS

In almost every survey or excavation, the field archaeologist will encounter finds which he may tentatively label ‘slags’ or ‘vitrified material’. In very general terms, slags are waste, ie. discarded or left behind evidences of human activity, difficult to classify and to date, but remarkably resistant to weathering. Do slags reveal anything but the fact that fire and heat were somehow involved in their making? Is there a difference at all between ‘man-made’ slags and the molten material (lava, obsidian, etc.) from volcanism? Are the answers to these questions relevant to archaeological research at large or is the identification of slags just another topic to which undue importance is paid by a trespassing natural scientist?

The effect of heat on inorganic material, whether accidental or intentional, will initiate reactions: the firing of clay, the burning of top soil in a domestic hearth, or the smelting of ores in a furnace. These reactions need not necessarily result in slag formation. The degree of heating plus refractory properties of the material exposed to higher temperatures very often result in products, which – though they may have suffered transformations – have not reached the molten state; they do not belong to the type of material classified as slags. A slag in its true sense is a once molten silicate or silicate mixture, sometimes including oxides, phosphates, borates, sulphides, carbides, pure metals, etc. Some of these admixtures may not have reached their melting points during the formation of the bulk slag material. They are embedded in a slag matrix like raisins in a cake. A slag can represent a simple compound consisting of just one single slag mineral or phase after solidification, but more often slags are complex mixtures of many constituents with a wide range of composition.

In the course of archaeological field work, one has to decide whether slags found at a site justify further investigation or whether they are just ‘finds’ hardly worth recording. For instance, a mud brick building reinforced with timber, burnt down by invaders or by accident may turn into a mass of slag, sometimes preserving the original wall structure of the building to a considerable height. Two other groups of non-metallurgical slag-like materials come to mind. There are the vitrified forts of the Celtic Iron Age, whose glass compositions are reported (with a good bibliography) by Youngblood et al. (1978), and the cinder mounds of the South Indian Neolithic, believed to be burned cow dung, cf. Zeuner (1959) and Allchin (1963). What information can be deduced from the analyses of such slags? Probably very little, except perhaps an indication of the temperature of the fire that caused the destruction, as derived from the formation temperatures of the slag minerals. On the other hand, small fragments of inconspicuous slag adhering to a crucible sherd discovered in a workshop area are likely to tell us something about the metals or alloys used at the site in question and may therefore be of paramount significance. How can one distinguish between significant and insignificant slag material? What specimens should be selected for further investigation by a chemist, and what – after all – is the archaeological implication of the laboratory results? These are only a few of the questions underlying the co-operation between archaeology and archaeometry. In order to avoid the expansion of excavation reports with endless and not always very informative tables of analytical data, some clarification and classification is necessary. In doing so, the chemist is fully aware that ancient slags are not always well-defined, homogeneous materials, easily placed in particular categories, nor can he do without the terminology and the reference to physico-chemical principles when explaining his results to the archaeologist.

However, archaeologists are becoming more and more conscious of the facilities modern analytical methods can offer. If slags are submitted for identification and analysis, the archaeologist should supply the necessary information about the site, any accompanying material, stratification, etc., and – very important – the amount of slag found. The joint venture between archaeologist and chemist requires some understanding of the possibilities and limitations offered by physico-chemical methods. None of these is a panacea to all questions. Neither does the combined application of numerous methods always give the answers the archaeologist is expecting. The following chapters are intended as a brief outline of presently available methods and their significance to archaeological research. As the archaeologist normally is not so much concerned with the practical side of analytical work, but rather interested in the results and their interpretation, importance is given more to the latter than the former.

FIELD IDENTIFICATION OF SLAGS

In an attempt to classify the various possible origins of ancient slags, the following sources have to be considered:

Geological material (‘natural’ slags):

– lava, pumice, obsidian (frequent in volcanic areas)

– tektites (generally rare, but quite abundant in certain regions; resembling obsidian)

Domestic and destruction slags:

– destruction of buildings by fire eg. molten and vitrified mud brick walls), vitrified forts

– slagged structures and remains of household ovens and heating installations (eg. slagged fuel ashes, bricks, linings, etc.)

Technological, non-metallurgical slags:

– brick and tile manufacturing

– lime kilns

– pottery kilns

– glass-making and glazed stone production (eg. in early dynastic Egypt)

Metallurgical slags:

– non-ferrous metal smelting (eg. copper, lead, tin)

– non-ferrous metal melting (eg. purification of raw metals, alloying)

– bloomery hearths for wrought iron production

– iron smithing

This brief and incomplete summary leaves out slags from modern processes, like blast furnace slags. A full treatment of this is given in Slag Atlas, published recently by the Verein Deutscher Eisenhüttenleute (1981). According to Needham (1958) cast iron was made in China in blast furnaces at least as early as the Warring States period in the 4th century BC. There are also a number of probably accidental examples of cast iron from the Roman period, as reported by Coghlan (1956). This means that blast furnace slag phases might occasionally be found on Roman bloomery sites. Though not mentioned specifically, the summary above includes all sorts of material somehow affected by fire and heat, such as partly molten or vitrified material (cinder, dross, etc.), as well as glassy and crystalline slags proper. Perhaps one day all these relics of ‘pyrotechnology’ will be equally significant to archaeological research. If, for instance, fission track dating is applicable to all glasses, whether they are of geological or human origin, glassy slags may become important objects for dating. Rather high uranium concentrations (perhaps 1000 ppm) would be required for the fission track dating of materials as young as slags, although the presence of unabsorbed zircon inclusions might make the process feasible. The same is true of other dating techniques, like thermoluminescent (TL) and archaeomagnetic methods. TL dating of slags has not been a conspicuous success so far, probably because of the high concentration of transition metal ions which they usually contain, but attempts to use the most common slag mineral, fayalite, for TL dating are promising (Wagner, 1981). For the time being, slag identification has preference over dating. Learning more about slags may initiate future developments in this field. In limiting ourselves to methods which are at present available, slag identification will essentially be treated as a criterion for distinguishing between slags and ‘non-slags’, and further between metallurgical and non-metallurgical slags.

As will be shown in this chapter, identification should start in the field. The ‘macroscopic’ description of a sample as found is not only a prelude to further ‘microscopic’ treatment; very often, a thorough field identification makes the more elaborate laboratory work superfluous.

Contrary to the case of glass and pottery making, where broken remains and fragments of the finished product have nearly always been left behind at the production site, slags are very often the only witnesses of archaeometallurgical processes that once took place. In antiquity, non-ferrous metals were precious. Great care was taken to recover and salvage even the smallest bits and pieces from the byproducts of smelting and melting. Further, in metal production, slagging was an essential part of the pyrometallurgical process. This is the reason why metal slags serve as process indicators. Metallurgical slags are not just unwanted waste; they were made intentionally and their identification is more important than that of slags from other sources.

Slags vary in colour, appearance, texture, specific gravity, etc. Fortunately, nearly all ancient metallurgical slags are ferrous silicates, no matter whether they are the byproducts of iron or non-ferrous metal production. This facilitates field identification and distinction from non-metallurgical slags. In order to identify a find, tentatively classified as slag or slagged material, the following external properties and circumstances of occurrence should be checked:

Colour:

Due to their high content of iron silicates, slags from metal smelting or melting tend to be black or dark grey on freshly exposed surfaces. An iron tool, preferably a geologists’ hammer or the like, should be used to chip off a fragment from the slag piece to be examined.

Streak:

Streak is the colour of a powdered specimen. This identification method, well known in mineralogy, is equally applicable to artificial minerals, ie. slags. The most common method of determining streak is to draw the fresh surface of the specimen across a plate of white, unglazed porcelain, called a streak plate. As a substitute, the protruding, unglazed rim of the reverse side of a saucer, cup or plate can be used. Ferrous slags give a dark brown to black streak, while, for instance, slagged mud bricks – even if they are dark grey externally and may therefore be mistaken for metallurgical slags – only give a very faintly brown or uncoloured streak. However, some metallurgical slags are extremely hard and do not powder readily.

Texture:

Many slags show flow textures, often resembling volcanic lava. Other samples are striated or layered. Glassy slags will reveal conchoidal fractures, while crystalline slags exhibit reflections or lustre from crystals (needles or laths) of slag minerals.

Porosity:

Though a great number of naturally occurring minerals resemble slags, a look at the fresh surface, preferably with a magnifying lens (magnification of 5 to 8 times is sufficient) will nearly always answer the question: mineral or slag? Unlike most minerals, slags show gas holes (‘gassiness’). They vary in size from minute cavities to large circular or elliptical vesicles. Because some dark volcanic rocks, like basalts, are also occasionally porous, the geology of the site where the slags were found should be taken into consideration. The Roman preference for millstones hewn out of black or dark grey basalt lava, often transported over long distances, may sometimes be puzzling. Experience will help to overcome such occasional obstacles. A word of warning: slag identif...