A R C H A E O L O G Y I N T E R N A T I O N A L

37

N

1

3

4

5

60 km 2000

2

n the West the earliest production ofliquid steel is usually credited to theBritish industrial revolution, andparticularly to the work by BenjaminHuntsman, who pioneered the pro-

cess in England in the 1740s. In Europeprior to that time, soft iron and mild steelwere worked only in their solid state,because of the exceptionally high temper-ature of more than 1400 °C required to meltthem. However, in several places in Asia,liquid steel had already been produced foralmost a millennium by exploiting partic-ular clays that made possible the produc-tion of pottery capable of withstandingthe necessary high temperatures. Recentresearch by archaeometallurgists has nowshown that, in Central Asia and in the areathat today comprises southern India andSri Lanka, two visually very different(white and black) traditions of making thepottery vessels or crucibles employed inthis process developed (Figs 1, 2).

South AsiaThroughout the eighteenth and nineteenthcenturies, India attracted the attention ofEuropean naturalists, ethnographers andother travellers interested in the exoticplants, animals, materials and traditionalcrafts of the subcontinent. Among the morehighly praised (and priced) materials werecakes or ingots of high-quality steel, free ofsuch impurities as slag and sulphur thatused to contaminate European steel at thattime. The metal made from Indian ingots,which were known as wootz (the name isan anglicized corruption of an old Indianword), was superior for making fine bladeswith exceptionally hard cutting edges; andthe surfaces of many artefacts manufac-tured from these ingots also displayed thedecorative pattern of very fine lines knownas damask in steel, textiles and other mate-rials. The quality of this damask (or Damas-cus) steel triggered much research acrossEurope – not least under the auspices of theBritish Royal Society, including such wellknown scientists of the day as MichaelFaraday and Henry Wilkinson – in an effortto solve the mystery of its production andmanufacture. Modern metallurgical tech-nology was stimulated by, and developedin parallel with, the study of these Indiansteel ingots, driven by the almost mysticalreputation for beauty and performance ofthe products manufactured from them.

Despite the former fame of Indian steel,little attention has been paid to the extra-ordinary ceramic material that enabled itto be melted. The special quality of the

material shows in the ability of the cruci-bles made from it to survive temperaturesgreater than 1400°C. Most pottery melts attemperatures as low as 1200°C, making ituseless for melting steel. During the 1980sand 1990s, Thelma Lowe, who was then adoctoral student at the University of Cali-fornia, Berkeley, investigated this cruciblematerial from several sites in southernIndia, including Konasamudram (Fig. 1),and found that a particular mineralogicalstructure gave the material its extraordi-nary strength.1 During the past ten years,several research students at the Institute ofArchaeology have carried out similar stud-ies in both southern India and Sri Lanka,for example at the sites of Gattihosahalliand Mawalgaha (Fig. 1), where this uniquematerial was made.2

The material is in general a highlyporous and very black pottery (Fig. 2). Theclay used in its manufacture was mixedwith roughly equal amounts of rice chaff,which, upon firing, burned away, leavingan open network of pores and minutesilica-rich remains of ash. A series ofanalyses recently undertaken at the Insti-tute showed that the clay is chemicallydifferent from that used for other pottery,indicating that it was specifically selectedfor this process. Clearly, the craftspeoplewere well aware of what they were doing,and they did it on a massive scale. Eachwootz ingot was quite small and probablyweighed between 100g and 1kg, but thissmall size was counterbalanced by the vastnumber of ingots produced. To this day,massive heaps of broken crucibles coverthe outskirts of several Indian villages, andare testimony to a time in the eighteenthand early nineteenth centuries when Indianwootz ingots were sold by the shipload toEuropean and Arabian merchants, servinginternational markets. But the antiquityand origins of this steelmaking technologyremain a matter for speculation. Writtenaccounts and trade records stretch back afew hundred years only, although one siteof steel-ingot manufacture, in the Knuck-les range north of the Central Highlands inSri Lanka (Fig. 1), has been reliably datedby radiocarbon to the middle of the firstmillennium AD.3

As similar as black and white: steelmakingcrucibles from South and Central Asia

Thilo RehrenIn recent years, fieldwork by archaeometallurgists, and labora-tory analysis of the materials found at sites of early iron- andsteelmaking, have led to the discovery that liquid steel was beingmade in parts of South and Central Asia a thousand years ago,long before it was manufactured in Europe. Research studentsand members of staff of the Institute of Archaeology have beenin the forefront of these investigations, some of the results ofwhich are described here.

Figure 1 Central and South Asia show-ing the approximate extent of areas of ancient crucible steelmaking and the loca-tion of sites mentioned in the text:(1) Merv, (2) Akhsiket, (3) Konasamudram, (4) Gattihosahalli, (5) Knuckles Range,(6) Mawalgaha.

cm 50

Figure 2 Cross sections through two ceramic fragments of the bases of steelmaking crucibles; left: a white-fired one from Akhsiket in Uzbekistan, right: a thicker black-fired one from Konasamudram in southern India.

A R C H A E O L O G Y I N T E R N A T I O N A L

38

Central AsiaUntil recently few Westerners were awareof, and even fewer had direct experienceof, the archaeological riches of CentralAsia. But since the dissolution of theSoviet Union it has become possible forarchaeologists from the West to undertakeresearch in the region by participating ininternational projects in the newly inde-pendent Central Asian republics. Severalmembers of the Institute of Archaeologyhave seized these new opportunities byundertaking fieldwork, including archaeo-metallurgical research, in Turkmenistan atthe ancient urban centre of Merv, as well asmuch farther east at the ancient site ofAkhsiket in Uzbekistan (Fig. 1). From bothplaces we now have well dated and com-prehensively studied evidence for theproduction of liquid steel in crucibles asearly as the eighth to twelfth centuries AD.4

Again, the scale of production was mas-sive, with hundreds of thousands of indi-vidual crucibles being made at Akhsiketalone, each of which would have producedabout 4kg of high-quality steel.

Several historical sources indicate thatthe steel ingots were traded within theIslamic world south and west, to majormarkets in Persia, but also farther west to

places such as Damascus in Syria. There iseven some indication that steel ingots wereproduced in this way in Islamic Spain, forexample in Andalucia, before the end ofthe Christian reconquest in 1492.5

The steelmaking processA detailed metallurgical study at the Insti-tute of the remains from many of the siteshas now demonstrated that the same tech-nical process was used in both Central andSouth Asia to produce the steel. In bothregions, iron was first smelted in the tra-ditional way, using small furnaces. Then,this iron was broken into small pieces,mixed with an equal amount (by volume)of plant matter, and placed in the crucible.The crucible was then sealed to prevent airgetting in during the operation, with a fewsmall holes left to allow pressure to bereleased as the vessel was heated. At hightemperature, the iron begins to absorb car-bon from the plant matter, up to about 2 percent by weight. This transforms the metalinto steel, and simultaneously reducesthe melting temperature from more than1530°C for pure iron to around 1400°C forsteel. The liquid steel collects at the bottomof the vessel until all the iron is trans-formed. By then, the remaining carbon and

any slag floats on top of the metal, and thereaction automatically comes to an end.After cooling, the crucible is broken open,the ingot removed and the vessel discarded.

Despite the similarities in the ancientsteelmaking processes employed in Cen-tral and South Asia, there is one aston-ishing difference – that in the cruciblesthemselves (Fig. 3). In Central Asia, theyare much larger, holding up to about half alitre of liquid steel, as well as enoughexcess volume to contain the initial plantmatter and eventually the slag. Also, theyare made from a very different type of clay,which, when fired, turns into pale creamto white pottery of almost porcelain-likequality (Fig. 2). This pottery is very dense,has very few pores, and despite the muchlarger vessel size it is only about half asthick as its black South Asian counter-parts. Chemical and mineralogical analy-ses of this pottery have shown that it wasmade from a very special type of clay, com-pletely different from the one used to makethe domestic pottery of the time. To mod-ern materials scientists, such clay (whichclosely resembles fire clay and china clay)is known to be even more heat resistantthan the type used in South Asia.

Figure 3 Reconstructions of (left) two large crucibles typical of Central Asia and (right) three smaller ones typical of South Asia.

cm 100

A R C H A E O L O G Y I N T E R N A T I O N A L

39

ConclusionFor more than a millennium, until thenineteenth century AD, steel produced inCentral and South Asia was of a qualitymuch higher than its European equivalent.At present, we cannot identify where thetechnology for making liquid steel origi-nated, but it is apparent that in most partsof Asia where steel was made the samemetallurgical principle (the method of car-burization) was used in the process. It wasonly when the European technology of theblast furnace introduced cast iron to Indiathat a new way of making steel, whichinvolved melting together the hard andbrittle cast iron with traditional soft iron,was developed there. The possibilityexists that the latter process was usedmuch earlier in China, where cast iron wasproduced more than two millennia ago,but at present we have no archaeologicalevidence pointing in that direction.

The metallurgy involved is the same forSouth and Central Asia, but the ceramictraditions are very different. In South Asia,a black-firing, porous and relatively thick-walled ceramic material was used, proba-bly at the limit of its stability when in use,which allowed only small vessels to bebuilt. In Central Asia a white-firing, denserand much more refractory (heat-resistant)ceramic material was employed, whichenabled more voluminous vessels withslender walls to be built, and hence muchlarger steel ingots to be produced. Thesimilarity, in composition, appearanceand date, of these Central Asian steelmak-ing crucibles to early Chinese porcelainmay well indicate that an as yet unknowntransfer of ceramic (and perhaps metallur-gical) technology took place between Cen-tral and East Asia. As archaeometallurgicalresearch continues, we can expect morediscoveries to throw new light on theancient history of iron- and steelmaking inAsia and the associated developments inceramic technology.

Notes1. See T. Lowe, N. Merk, G. Thomas, “An his-

torical mullite-fibre-reinforced ceramic composite: characterization of the ‘wootz’ crucible refractory”, in Materials Issues in Art and Archaeology II, P. B. Vandiver, J. Druzik, G. Wheeler (eds), 627–32 (Pitts-burgh: Materials Research Society, 1991).

2. See, for example, G. Juleff, “Early iron and steel in Sri Lanka: a study of the Samana-walewa area”, (Mainz: Philipp von Zab-ern, AVA-Materialien 54, 1998) and S. Srinivasan & D. Griffiths, “Crucible steel in south India – preliminary investiga-tions on crucibles from newly identified sites”, in Materials issues in art and archaeology V, P. B. Vandiver, J. Druzik, J. Merkel, J. Stewart (eds), 111–25 (Pitts-burgh: Materials Research Society, 1997).

3. M. Wayman & G. Juleff, “Crucible steel-making in Sri Lanka”, Historical Metal-lurgy 33, 26–42, 1999.

4. D. Griffiths & A. Feuerbach, “Early Islamic manufacture of crucible steel at Merv,

Turkmenistan”, Archaeology Interna-tional 1999/2000, 36–8; A. Feuerbach, Crucible steel in Central Asia: production, use and origins, PhD thesis, UCL Institute of Archaeology, University of London, 2002; and Th. Rehren & O. Papakhristu, “The Ferghana process of crucible steel smelting”, Metalla (Bochum) 7, 55–69, 2000.

5. M. Karlsson Dinnetz, “Literary evidence for crucible steel in medieval Spain”,Historical Metallurgy 35, 74–80, 2001.