BUTSER ANCIENT FARM ARCHIVE 1973-2007 Archivist Christine Shaw
Home History of the Project Research Publications Image Archive
Back to top

Extract from Butser Monograph :

The Discovery and Utilisation of Iron

The late A G Hamlin B.Sc.


Research into the discovery and utilisation of iron has to answer two important questions:

1) Why did the discovery and utilisation of iron lag a thousand years or so behind the discovery and utilisation of the other metals of antiquity, lead, copper, tin, and the bronzes?

2) How did the early experimenters discover that iron, a potentially abundant but soft metal inferior in many ways to the bronzes, could be transformed into a range of vastly superior products by the controlled introduction of small amounts of carbon into the metal?

This Report deals with the first of these questions.

Experimental work at Butser Ancient Farm has, for some time, been researching possible routes for the discovery and utilisation of metals. Since native metals are rare, and have no obvious connection with their ores, man, at the start of the metal ages, would have had no general knowledge of metals and their sources, and it would seem to be unlikely that he would have attempted directly to convert ore to metal. The approach at Butser has therefore been to consider how man might have produced metals while processing their ores for some other purpose. The ores of the metals of antiquity are mostly highly coloured and are potentially useful pigments, but they need to be converted to a fine powder in order to have effective colouring and,covering properties. Heat can be a convenient means of doing this, and it seems probable that heat treatment of metal ores to produce pigments, using the pottery technology available at the start of the metal ages was the "other purpose".

Research at Butser on this type of operation has proved experimentally that it could have led to the discovery and application of lead, silver, copper, and lead bronze, and this work has been described previously. It is not possible to prove from such chemical experimental archaeology that this was the route that actually led to the discovery of these metals, but its experimental feasibility shows that it is not necessary to assume that metallurgical technology was discovered at, and spread from, a single source. The discovery could have been made wherever the production of pigments from coloured ores using pottery kilns was undertaken. Some producers would have made better metals than others, not only through better technological competence, but also because their ores would have contained, without their knowing it,desirable alloying elements such as arsenic in copper, but there would seem to be no doubt that metallurgical competence would have been widespread before the start of the Iron Age.

Iron ore is one of the most abundant and highly coloured ores, and there is ample evidence of the use of iron oxide pigments, ranging from dark brown to bright red, throughout antiquity. Chemically, it is not difficult to convert iron ore to iron. Therefore, if the theory that metals were discovered as a development of the heat treatment of ores to give pigments is correct, iron ores might have been expected to yield their metal at around the same time as a competence in producing the non- ferrous metals was achieved. Although a few examples of smelted iron have been dated to the Bronze Age, there was clearly some fundamental difference between iron and the other metals of antiquity that delayed the utilisation of iron by some thousand years or so.

Butser Ancient Farm relates to the Iron Age, and therefore it is important to understand how iron came to be the metal of the age, and to draw conclusions as to whether its use spread from a point of invention, or whether, ultimately, its use could have developed anywhere and everywhere that ores were processed for metal or pigments.

In order to achieve this understanding, it is necessary to draw upon modern chemical knowledge of the chemistry of iron, and this has been done in the following Section. The discussion has been kept as simple as possible, and any technical terms used are defined in the Glossary, Appendix 1.


The smelting process is essentially the removal of the oxygen that is combined with a metal in its ore, to leave the metal on its own. The means of doing this is to expose the ore to a reducing gas at high temperature.

The metals of antiquity can each combine with different proportions of oxygen to form more than one specific compound. For each metal it is possible that one or more of these compounds (oxides) will not react with the reducing gas under certain conditions in the smelting process. Moreover, if conditions in the smelting process have been favourable for the conversion of ore to metal, but then change to conditions under which one or more of the oxides concerned cannot be converted to metal, the metal already formed may revert to these oxides. Such changes are known as reversible reactions, and whether they go from ore to metal or from metal back to ore depends entirely upon the temperature and composition of the reducing gas in the smelting process.

These concepts may seem complicated, but consideration of them allows the metals of antiquity to be divided into three distinct groups.

Group 1 Lead, copper, silver.

Group 2 Tin

Group 3 Iron

The metals of Group 1 have relatively unstable oxides, which are converted to metal by very low concentrations of reducing gas - carbon monoxide or hydrogen - at fairly low temperatures. They are easily smelted under pottery kiln conditions, and because their oxides are easily converted by minimal concentrations of reducing gas, they are unlikely to be reconverted to ore as the kiln or furnace cools. If they have been melted after smelting, they are even more stable. Melting greatly reduces the surface area of metal exposed to the gases in the kiln, and so minimises the reconversion of metal to ore.

Experimental work at Butser Ancient Farm previously reported (1) has confirmed that these metals are the most easily produced of the metals of antiquity, and they form the basis of the Copper and Bronze Ages.

Tin, in Group 2, is intermediate between Groups 1 and 3. It has two oxides, which are reduced straightforwardly by either carbon monoxide or hydrogen, but, unlike the oxides of the metals of Group 1, tin oxides are converted only under rather specific conditions. These conditions are most easily shown in the form of equilibrium diagrams (Fig. 1 a, b) relating the temperature and composition of the reducing gas to the conversion of the metal oxide to the metal. If the reducing gas has a temperature and composition lying in the shaded area of the diagram, the oxide will not be converted. The oxide will be converted to metal only if the temperature and composition of the reducing gas lies in the clear area.

If the reducing gas contains both carbon monoxide and hydrogen, combination of the diagrams (Fig. 1 c) shows that the clear area is slightly reduced, but still leaves a range of temperatures and compositions of the reducing gas that should be easily obtainable by kiln or shaft furnace technology.

Equilibrium diagrams are obtained by theoretical calculations, Although they define clearly the conditions under which metal can be produced (the clear area), in practice, the useful part of the clear area may be reduced by considerations of gas composition and reaction rates. In the case of tin, it is known that for practical rates of conversion, gas compositions and temperatures lying towards the top right hand corners of the diagrams of Fig. 1 are required.

The position of tin in the prehistory of metals has not yet been fully worked out at Butser. Its ore (cassiterite, stannic oxide) is dirty brown in colour, although it does give a white streak if rubbed on an abrasive surface, and therefore it would have been of little pigmental attraction. The high temperature and high concentration of reducing gas required for its conversion are both at or beyond the limits of kiln technology. It has not been possible to make any measurement of the ratio R% (Fig. 1) in kilns at Butser, but experience has indicated that the maximum temperature that can be expected with any appreciable reducing condition is about 1000 deg C, corresponding to the peak shown in Fig. 4 when practical heat losses are involved. The value of R% at this point would be very small and to increase it would involve moving along the downward slope to the left of the peak in Fig. 4 towards a temperature of perhaps about 800 deg C when practical heat losses are taken into account. Kiln firing cannot be controlled well enough to give consistently high values of R%, and therefore kiln smelts are likely to give conditions lying towards the lower left part of the clear areas of Fig. 1 rather than the upper right hand area known to be required for success in modern technology. If tin were to be removed from the smelting reaction as soon as it is formed, then better results might be obtained under the conditions available from kiln technology as this would reduce the rate of any reconversion of tin to oxide. Such an effect could have been achieved by co-smelting tin with copper, the tin being converted to bronze on formation. However, co- smelting of tin and copper ores attempted at Butser has been notably unsuccessful, the tin ore remaining unchanged while the copper ore was converted to metal. Further practical difficulty is presented by the fact that tin oxide combines so readily with ceramic crucible materials well below 1000 deg C, that it is difficult to find a practical container.