MERV, an oasis in the Kara Kum desert in present day Turkmenistan, was a
major trading post on the Silk Road and has been settled for more than 2500
years. Destroyed and abandoned several times, each new city was built alongside
its predecessors, and archaeologists had little doubt that excavations would
provide brilliant insights into the ancient urban lifestyles of Western
Asia.
But when Georgina Herrmann of the Institute of Archaeology at University
College London started excavations in 1992, the last thing she expected to find
was early evidence for steel production鈥攎uch less the first hard evidence
of a mysterious technique known only from early Islamic manuscripts.
Fragments of furnace walls and clay crucibles were uncovered at the site by
Herrmann鈥檚 International Merv Project (IMP), and identified as relics of the
earliest known Islamic metalworking foundry by John Merkel, an
archaeometallurgist also from UCL. With other researchers he discovered tiny
droplets of steel, frozen in a glassy slag in the crucible fragments. This
indicated that the 9th-century inhabitants of Merv were manufacturing steel
several hundred years before it became widespread among the metalworkers of
India and Syria in the Middle Ages.
Advertisement
Steel is made of iron and up to about 1.5 per cent carbon. This level of
carbon requires skill to produce, since it falls between the levels of carbon
found in the two natural products of iron ore smelting鈥攃ast iron and
wrought iron. Cast iron is produced during smelting, when iron exposed to
carbon-rich charcoal for long periods runs off as molten liquid (at a
temperature of 1400 掳C, falling to 1200 掳C as carbon levels rise). It
contains between 2 per cent and 4.5 per cent carbon. Wrought iron, which
contains virtually no carbon, requires temperatures above 1500 掳C to melt
completely, so was traditionally extracted from smelting forges as a semimolten
mixture of metal globules and slag, and worked (wrought) when white hot to expel
the slag. In Europe, cast iron production was not well understood until AD 1500,
making it secondary in importance to wrought iron.
The high carbon content of cast iron makes it hard and brittle鈥攁nd
virtually impossible to work. Wrought iron is softer, but this limits its use
for weapons and tools. Steel鈥檚 combination of strength and workability made it a
prized material for armourers and tool-makers across the centuries.
The antiquity of the steel production at Merv was not the only remarkable
aspect of the find, because, says Herrmann, the city 鈥渨as not a natural place to
make steel鈥�. There are no nearby iron ore deposits and no suitable kaolin clay
for making the production crucibles. Equally intriguing was the discovery that
the broken crucibles contained the first tangible evidence for a unique method
of steel production previously known only from 11th-century Islamic and
6th-century Chinese manuscripts.
Two into one
Known as 鈥渃o-fusion鈥�, the technique involved heating small pieces of solid
wrought iron with cast iron, until the latter melts. Some of the cast iron鈥檚
carbon diffuses from the melt into the wrought iron (which only melts
partially), producing a workable, durable steel. The final carbon content and
quality of the steel depends on the extent to which the metals melt, or fuse
together. Co-fusion produces steel at temperatures several hundred degrees lower
than the 1500 掳C melting point of true cast steel, because you only need
initially to liquefy the cast iron in the crucible.
The discovery that co-fusion was used in Merv during the 9th century confirms
that Asians were aware of the melting and hardening properties of cast iron in
steel production several centuries before Europeans. 鈥淢erv had a long tradition
of steel production, but nothing like this has ever been found before,鈥� says
Herrmann.
Other early metallurgists of Asia employed various methods to create steel.
One of the best known was the Indian technique used to make 鈥渨ootz鈥�, a
high-carbon steel, suitable for forging thick, hard swords. India was exporting
wootz in large quantities to Europe by AD 1000. It was made by packing pieces of
wrought iron and other carbon-containing compounds, such as charcoal, wood and
leaves, into small, conical crucibles. These were stacked, sometimes 50 or more
at a time, into furnaces and heated to around 1500 掳C for about nine hours,
transforming the contents into a uniform liquid steel.
Another method, from medieval China and Japan, involved removing some of the
carbon from cast iron鈥攄ecarburization鈥攖o reduce its brittleness and
hardness. This was achieved by blowing air over the molten metal, but was
notoriously difficult to control. A more popular technique, known as case
hardening or cementation, involved heating solid wrought iron with material
containing carbon鈥攕uch as powdered charcoal, leather or shavings of
animals鈥� hooves. During heating, the carbon diffused into the surface of the
wrought iron. This method was widely used to make files and swords. Yet the
products of case hardening were often contaminated with impurities from the
wrought iron, and diffusion of carbon into the surfaces produced steel of uneven
quality.
The co-fusion process is more efficient, in terms of the size of furnace and
speed of production, than any of these techniques鈥攁nd can produce a more
uniform product. The wrought iron takes up carbon much more readily because its
entire surface is in contact with carbon-containing molten cast iron. Yet
Herrmann鈥檚 discovery at Merv is the only evidence for it, and confirms for the
first time the accuracy of the references in various texts.
Before the recent discoveries at Merv, virtually everything that
archaeologists knew about co-fusion was gleaned from the writings of al-Biruni,
an Islamic scholar who lived between AD 973 and AD 1050. His first
reference to co-fusion reads: 鈥淭he mixture of narm-ahan and its water, which is
the substance which flows when the narm-ahan is purified, is fuladh. The area of
Herat [in former Persia, some 300 kilometres south of Merv] is especially noted
for it and it is called baidat (eggs) on account of its shape. The eggs are long
and round-bottomed, following the shape of the crucibles, and from them Indian
swords and others are fashioned.鈥�
Liquid iron
It is known from other texts that narm-ahan refers to wrought iron, and
fuladh to steel, but experts were confused over what al-Biruni meant by the
鈥渨ater of narm-ahan鈥�, a substance he also referred to in his book as 鈥渄us鈥�.
Most scholars are now convinced that 鈥渄us鈥� is cast iron, and that 鈥渢he water
of narm-ahan鈥� refers to it in the molten state. Metallurgists suspect that
al-Biruni is referring to one of the two products of iron ore smelting: iron
kept in close contact with the fuel for a long period will increase its carbon
content, lower its melting point and run off as liquid cast iron鈥攚ater of
narm-ahan.
Al-Biruni鈥檚 texts describe the manufacture of two types of co-fusion steel.
鈥淓ither the narm-ahan and its water melt equally in the crucible and unite
so that one cannot distinguish the one from the other, in which case it is good
for files and the like . . . or alternatively, the melting qualities of what is
in the crucible vary, so that the two do not mix completely but on the contrary
are separate in their parts from one another . . . This is called damask.鈥�
As its name implies, the product of this second process was the source
material for the famous Damascus blades, which carried patterns of bright and
dark bands that reflect the variable carbon content achieved by stopping the
fusion before the reactants mixed fully.
Thanks to the Merv finds, metallurgists have been able to piece together more
information about the co-fusion process. Merkel and his colleague Ann Feuerbach
at the Institute of Archaeology found abundant steel droplets up to 0.3
millimetres in diameter and of varying composition caught up in the glassy slag
on the inner surface of the crucible fragments. They also found fragments of
cast iron trapped within the slag. Merkel says these 鈥渄on鈥檛 prove that the
process used was co-fusion, but along with other observations they are highly
suggestive that it was鈥�.
The other evidence that co-fusion was used rather than another crucible steel
technique is the total absence in the slag of any hint of organic material. In
cementation 鈥渢here are always, without fail, at least impressions of organic
material鈥�, says Feuerbach, who presented a paper on the findings at a meeting of
the US Materials Research Society in Boston in December.
Merkel thinks that the wrought iron needed for steel production must have
been taken from scrap brought in along the Silk Road, given the lack of iron ore
deposits near Merv. But Herrmann believes the steel makers of Merv imported the
iron as blooms (or ingots) rather than rely on scrap. 鈥淭hey were particular
about the type of woods and clays they imported for the furnaces, and I don鈥檛
see why they would have relied on scrap to provide their iron,鈥� she says. 鈥淭hey
went to tremendous lengths to obtain all the other correct raw materials.鈥�
More evidence is emerging from the site about these unusual steel makers. The
researchers have collected enough fragments to reconstruct several crucibles.
They were made of fine clay, about 6 centimetres in diameter with a single hole
in the centre of the lid, used either for periodically checking the contents of
the crucible or for letting out excess carbon dioxide. The team has also
discovered the remains of several circular furnaces, around 70 centimetres in
diameter and capable of holding some 40 crucibles.
Apart from the evidence for early steel production, the IMP鈥攎ade up of
experts from Britain, Russia and the Academy of Sciences of
Turkmenistan鈥攈as found that cotton was cultivated at Merv as long ago as
the 5th century AD, many centuries earlier than thought from earlier excavations
in the region. The team is using satellite imagery, aerial photography and
ground mapping techniques to map 1000 hectares of the ancient site and is
helping the Turkmen government to secure Merv as a UNESCO World Heritage
site.
Last year Herrmann won a Rolex Award for Enterprise for her work at Merv. The
award, designed to encourage 鈥渢he spirit of enterprise鈥� in science and
exploration, acknowledges the innovative nature of the project and the potential
for future discoveries at Merv. 鈥淚t is an outstanding site, and very relevant to
archaeologists,鈥� says Herrmann. 鈥淚t links the archaeology of Iran, which we
already know about, to that of central Asia, about which we know relatively
little.鈥� And she suspects that Merv is hiding plenty more archaeological
surprises.
