[Papers from the Institute of Archaeology 5(1994) 49-59]
Wootz crucible steel: a newly discovered production site in
South India
Sharada Srinivasan
Institute of Archaeology, VCL
Introduction
High-carbon iron alloys are known to have bee n produced in parts of Asia such
as India, where a traditional crucible steel has been produced, (Bronson 1986;
Smith 1960: 14-24) and in China, where cast iron was produced (Tylecote 1976:
85), long before they came into vogue in Europe (Lowe 1988). The repute of
Indian iron and steel can be traced to Classical Mediterranean accounts (Bronson
1986: 18). European travellers and geologists from the seventeenth century
onwards have described the production of steel ingots, in different parts of South
India (Fig. 1), by crucible processes. Such accounts were made in the former
province of Golconda, Andhra Pradesh (Voysey 1832), the former state of
Mysore (modem Kamataka) (Buchanan-Hamilton 1807) and Salem district in
Tamil Nadu (Buchanan-Hamilton 1807, Wood 1893); Coomaraswamy (1956:
192-3) has also described crucible steel processes at Alutnuvara in Sri Lanka.
This South Indian steel was named wootz, a European corruption of the Telegu
word for steel, ukku.
Indian wootz ingots are believed to have been used to forge the famed Oriental
Damascus swords which have been found to have a very high carbon content of 1.5-
2.0% (Smith 1960: 14-6). The Persian Damascus blades, made in Khorasan and
Isfahan from South Indian wootz from Golconda, were known to be the finest
weapons then made in Eurasia (Bronson 1986: 22-3, Smith 1960: 14-6) and were
reputed to cut even gauze kerchiefs (Bronson 1986: 1). The properties of South
Indian steel, which became synonymous with Damascus steel, were investigated
during the nineteenth and twentieth centuries by numerous European scientists,
chemists and metallurgists with the aim of reproducing it on an industrial scale
(Smith 1960: 25-9). A typical wootz ingot analysed in 1804 by Mushet (cited in
Smith 1960: 22) was found to contain about 1.3 % carbon and had a dendritic
structure (Smith 1960:22). European scientists who were successful in replicating
and forging wootz included Stodart who found that wootz steel had a superior
cutting edge that of any other steel (Bronson 1986: 30) while Zschokke in 1924
(cited in Smith 1960: 14) found that with heat treatment this steel had special
properties such as higher hardness, strength and ductility.
Recent investigations of the Indian wootz process have concentrated on
material from the known sites of Konasamudram, Nizamabad district or former
Golconda, Andhra Pradesh (Lowe 1990; Voysey 1832) and Gatihosahalli in the
Chitradurga district of Kamataka (Freestone and Tite 1986; Rao 1970). These
investigations have shown the existence of specialised, standardised and semi
industrial production techniques dating from at least the late medieval period.
During the course of field investigations of copper mining and smelting in South
India, the author of this paper came across a previously unrecorded
archaeometallurgical site in Mel-siruvalur, South Arcot district, Tamil Nadu,
50 S. Srill;vasall
which investigations have confinned was a production centre for wootz crucible
steel in the Deccan. The find of this production centre supports the idea that
wootz steel production was relatiYely widespread in SouthIndia, and extends the
known horizons of this technology further .
..
,.
11:11.01111111
Figure 1 Map of South India indicating sites mentioned in the text.
History of South Indian steel
Bronson (1986: 18) summarises eight mentions in Classical Mediterranean
literature of Indian iron or steel. The earliest of these is that of the Greek
physician Ctesia of the late fifth century BC; who mentions the wonderful
swords ofIndian steel presented to the King of Persia (Bronson 1986: 18; Schoff
1915). The import ofIndian iron and steel to the Roman world is suggested by
Pliny's Natural History which refers to iron from the Seres, identified with the
Southern kingdom of the Cheras, while the Periplus of the Erythraeoll sea
unequivocally mentions that iron and steel were imported from India (Bronson
1986: 18; Schoff 1915). Although the literary references have not yet been
corroborated archaeologically. excavation and investigations on the iron-rich
Wootz crucible steel 51
megalithic sites of Tamil Nadu and the Malabar (mid first millennium BC to
early centuries AD) could be revealing: these fall within the domain and period
of the Sangam Chera kingdom which may relate to Roman accounts of Seric iron
or Chera iron. Indeed recent excavations at an iron age megalithic site at
Kodumanal, Tamil Nadu (c. third century BC), close to Karur, the capital of the
Chera kingdom of the Sangam era (c. third century BC-third century AD) has
revealed furnaces stacked with vitrified crucibles which were found separated
from abundant iron slag (Rajan 1991: 98).
More concrete literary evidence of ancientIndian steel is found in later Arab
and Middle Eastern sources. Pre-Islamic Arabic literature of the sixth-seventh
centuries AD, such as Hamasa's collection of poems, refers to swords of AI-Hind
or Hinduwani fromIndia; whileIslamic writers such as JabirIbn Hayyar of the
8th century and AI-Biruni of the eleventh century AD make it clear that South
Asian steel fromIndia and Sri Lanka was used in many places for sword making
(Bronson 1986: 19). The Arab Edrisi (cited in Schoff 1915: 232) comments that
it was impossible to find anything to surpass the edge obtained fromIndian steel.
The first explicit documented evidence of the export of wootz steel from South
India to make Persian Damascus blades comes from Tavemier (cited in Bronson
1986: 23) who in 1679 mentions the trade in steel from former Golconda near
Hyderabad, Andhra Pradesh, which was the only sort which could be damascened
by Persian artists, by etching with vitriol.
The considerable European interest in the nineteenth century in wootz steel
and Damascus blades contributed greatly to the development of metallography
in Europe, as pointed out by Belaiew (1918) and Smith (1960). Such
metallographic interest was aimed at understanding the distinctive wavy duplex
pattern of the Damascus blades and their relation to the crystalline structure of
the wootz ingot from which they were produced. The mechanical properties of
wootz steel were also much speculated abaut, and indeed the steel was replicated
with success and used to make surgical and high-grade cutting tools by cutlers
like Stodart and Damernme (cited in Smith 1960: 25-6). Attempts to duplicate
wootz lead to important experiments, on wootz and alloyed steel, by Michael
Faraday in association with Stodart (Smith 1960: 25).
Past observers of the manufacture of wootz steel in India have commented
on the process of carburisation of iron to steel in crucibles where a batch of closed
crucibleswith the low carbon iron charge were stacked in a large furnace and fired
in a long 14-24 hour cycle at high temperatures up to 1200 °C in a strongly
reducing atmosphere (Percy 1860-1880: 773-6). Three different types of
crucible processes have been described by nineteenth century travellers varying
from region to region, i.e. the Deccani or Hyderabad process, the Mysore process
and the Tamil Nadu process. In the Tamil Nadu process and the Mysore process,
the charge consisted of wrought iron produced separately which was then
stacked in closed crucibles and carburised in a large furnace (Verhoeven 1987).
But while the Mysore process charged the wrought iron with carbonaceous
matter, Wood's (1893) observations on crucible processes in Salem and Arcot
districts in Tamil Nadu suggest that only iron was charged and the crucible
containing the ingot was not fast cooled in water as in the Mysore process
(Bronson 1986). The Deccani process was renowned for the best quality wootz
52 S. Srinivasan
(Bronson 1986) and the process followed here was not of carburisation of a
wrought iron bloom but of fusion of two separate pieces of cast iron (i.e., high
carbon iron) and an iron bloom (low-carbon iron) (Voysey 1837: 247) so
producing a homogenous alloy of intermediate composition (Bronson 1986: 43;
Rao 1970).
The known sites of crucible steel production in South India, i.e. at
Konasamudram and Gatihosahalli, date from at least the late medieval period,
16th century. But, although these may be earlier, systematic excavations have
not been carried out to determine their antiquity. The existing research on wootz
steel at these sites has been more concerned with metallurgical re-construction
of the wootz process based on surface finds. The investigation presented here
is also from surface finds at a mound in Mel-siruvalur village, South Arcot
district, Tamil Nadu.
Mel-siruvalur: location and history of the site
In November 1991 the author made field investigations of old workings at a
polymetallic copJlCr-lead-zinc deposit on Kanankadu hillock, 21-22km S.S.W.
o 0
of Mamandur (12 00" N; 79 OO"E) in Kallakurichi taluk, South Arcot district,
about 40 km south by road from the nearest town of Tiruvannamalai. This
polymetallic sulphidic mineralisation occurs in association with meta-anorthosites
in the granulitic terrain of the Archaean complex of South India. Iron ores are
also found in banded ferruginous quartzite formations in Kallakurichi taluk. The
area was visited in the hope of locating evidence for copper smelting based on
reports by geologists from the GSI in Madras �ho had noticed some unidentified
metallurgical debris near the village of Mel-siruvalur about 5km from the
Kanankadu hillock. However, investigations of some of the debris collected and
presented here, indicate that it is instead related to crucible steel processes; which
is nevertheless of significance to the history of metallurgy in the area.
The village of Mel-siruvalur comprises a cluster of two or three houses in this
very sparsely populated arid region. Evidence of metallurgical activity came
from a mound just behind the village of about 25m x 8-9m wide and up to Srn high
(Fig. 2) and from some trenches near the houses. However the. villagers had no
memory of recently undertaken metallurgical activity. Occupation of the area
in antiquity is indicated by pottery sherds collected adjacent to an old canal,
about 1I2km away from the mound. Roating slag debris and crucible fragments
were also found all around the canal site. Among the sherds were many large rim
fragments, about 3cm thick, belonging to huge storage jars about 60cm in
diameter. These had no slip, and were found to be tempered with rice hulls. C.
S. Patil (pers. comm.) of the Mysore Archaeological Survey has pointed out their
resemblance to megalithic storage jars of red ware without slip. The megalithic
occupation in Tamil Nadu starts around the fifth-fourth century BC and continues
to around the fifth century AD. Megalithic dolmens have been found in
Thiruvannamalaiand Tirukoilur taluks adjacent to Kallakurichi taluk, in South
Arcot district. Also among the finds were pottery sherds of painted ware with
a red slip, decorated with a chain or hatched design, which were identified by C.
Wootz crucible steel 53
S. PatiI as being of the late medieval period (c. sixteenth century AD). Several
hollow conical terracotta jars about 70cm long of indetenninate function were
also found stacked along the walls of the canal. Without more detailed survey
and investigations, the possibility of these pottery assemblages being related to
the metallurgical activity cannot be confirmed.
Figure 2 Mound near Mel-siruvalur village, South Arcot district, Tamil Nadu.
Figure 3 Lid fragment of broken wootz crucible.
54 S. Srinivasan
Description of archaeo-metallurgical debris at Mel-siruvalur
Numerous crucible fragments were found at the Mel-siruvalur mound together
with fragments of glassy slag, charge and debris. When re-constructed the
fragments of crucibles showed typical features of the aubergine-shaped closed
crucibles used for wootz steel production known from other sites in South India
such as Gatihosahalli and Konasamudram. Thick covering lids of a diameter of
about 7cm, which would have sealed the refractory vessel during firing with the
iron charge, were found. Pieces with interior glazed surfaces and distinctive
'fins' of glassy slag that would have formed the middle portion of the crucibles
were also located. Several curved bases of the crucibles, about O.8-1.5cm thick
were among the finds. The dimensions of the various fragments indicated that
the ingots were of a diameter of c. 2.5cm. Some of the crucible bases appear to
have remnants of the rusty charge attached to them. The exterior surface of the
crucibles was covered with thick black ash glaze.
About 70m away from the mound were a set of two trenches inter-connected
in a pinch and swell shape of about IOm long. One of these was clearly the
furnace area, as it contained several tapering tuyeres fragments (with an inner
diameterc. l.5cm, and varying from O.8-2cm thick), along with furnace remnants,
consolidated mud and slag. The other trench contained only blocks of slag 20cm
high and 20cm in diameter with a flow texture; indicating that it had been used
to tap out slag from the main furnace.
Analytical results and discussion
A mounted section of a lid of a crucible (Fig. 3) was examined microscopically
and tiny iron prills of a diameter less than lOOj4m were found along the glassy
edge of the lid. Analyses of the priUs using Electron Probe Microanalysis
(EPMA) confirmed that they are steel prills (Table 1). The prills were embedded
in the outer crucible lining of the lid probably due to splashing of molten liquid
due to overflow at high temperatures. Similar prills have previously been found
embedded in Deccan wootz crucibles (Scott 1991: 35).
The etched microstructure (Fig. 4) of the largest prill (Prill 1), of a diameter
of c. 80j4m, has a lameUar eutectoid structure of fine pearlite inside original
hexagonal grains of austenite. This suggests that it derives from a very good
quality hypereutectoid high-carbon (>0.8 per cent) steel. The prill had a hardness
of around 400 VPN which is within that for normalised steel of c. 1 per cent
carbon (Scott 1991: 82). The presence of much smaller amounts of a lightly
etched network of cementite (iron carbide) between grains near the boundaries,
and as occasional needles in the pearlite was also noted. Scanning Electron
Microscopy with Energy Dispersive Spectroscopy (SEM-EOS) suggested that
this lightly-etched cementite contained some phosphorus impurities, i.e. consisted
of cementite-phosphide. Iron phosphide tends to form a ternary eutectic of
steatite along with pearlite (Avner 1988: 439) and its presence may indicate a
slightly higher carbon content of about 1 per cent. Prill 2 and four tinier priUs
-
�
<.
�
�
s:
�
""
�
�
-
53 J4m
i
....J
Figure 4 Prill from lid of crucible showing pearlitic structure in prior
austenite grains interspersed with some cementite.
�
56 S. Srinivasan
inspected microscopically had a similar pearlitic structure, with the presence of
the interdendritic continuous lightly-etched network of cementite around the
pearlitic eutectoid in varying degrees.
It is interesting that the micro-structures of the priUs etched in nital, of darkly
etched pearlite surrounded by lightly-etched cementite, are somewhat reminiscent
of the macro-structures associated with the beautiful patterns formed on Damascus
swords. These patterns are thought to consist of well formed lamellar darkly
etched high carbon pearlitic steel interwoven with a network of lighty-etched
iron carbide or cementite; formed by the forging of a high carbon iron ingot
followed by etching (Smith 1960: 16; Tylecote 1962: 295).
I Sr.No. Fe (wt%) Cu (wt%) As (wt%) S (wt%l Total (wt%) I
: Prill1
I
97.615 0.038 0.023 0.565 98.241 I
I
i Prill2 102.385 0.075 0.251 102.385 I
Table 1 Analysis of priUs in a section of a crucible from Mel-siruvalur
Analysed by EPMA on a polished and carbon-coated cross-section using
JOEL Superprobe JXA-8600 at 20 KV with ZAF correction, within instrumental
accuracy of 1 % over 100%. Trace silicon and phosphorus were also noted
especially in Prill 2 but have not bee n analysed by EPMA; however these
elements were shown to be less than 1 % using SEM with EDAX analysis
(HITACHI S-570 with link AN-l000).
As has been reported in previous analyses of wootz-making crucibles in
South India (Lowe 1990: 237-50), the fabric of the Mel-siruvalur vessel consists
of a porous glassy matrix with distinctive cooked rice hull relics (Fig. 5)
dispersed in the matrix along with sand or quartz grains. The inclusion of rice
hulls in the refractory material is a distinctive feature of the manufacture of
Deccani wootz crucibles (Lowe 1990; Voysey 1832: 246). Lowe postulates that
these were added for their high silica and carbon content, making the crucible a
particularly effective re-inforced composite refrnctory material; both to withstand
very high temperatures over a very long firing cycle and to maintain a highly
reducing environment to enable carburisation of the iron charge.
Qualitative analysis of a few samples of the slag collected from the second
trench by SEM-EOS showed that the major constituents were iron and silicon,
suggesting that these may be fayalite (iron silicate) type iron slags. Hence it
appears that the iron charge was being smelted by the bloomer process in the
trenches. The iron bloom produced here may have formed part of the charge to
produce high carbon iron by the wootz crucible process in the area where the
mound with the crucibles was found. Further investigations are needed to verify
which of the crucible steel processes was followed here: the carburisation of a
bloom, i.e. the Mysore or Tamil Nadu processes, or the fusion of cast iron with
a bloom, i.e. the Deccani process. The crucible frngments found on the mound
appear to be from fired crucibles which had been broken to retrieve the finished
ingots.
�
n
�s:
�
�
l\
�
-
83 Jlm
Figure 5 Rice hull relic in glassy matrix of fired refractory.
�
58 S. Srinivasan
Conclusions
The preliminary investigations reported here indicate that crucible steel production
was carried out, in the pre-industrial era, at a hitherto unreported site at MeI
siruvalur, South Arcot district, Tamil Nadu. Analytical investigations indicate
that closed crucible fragments were fired to a high degree of vitrification with the
charge, to produce a high carbon steel. Use was made of refractory reinforced
with rice hulls in the manufacture of the crucible as observed in the Deccani
process of wootz steel production. Further archaeo-metallurgical investigations
and surveys are required to determine the extent of metallurgical activi ty and the
antiquity of the site.
Acknowledgements
I would like to thank Daffyd Griffiths, John Merkel, Thelma Lowe and R. A.
Charles for their comments pertaining to the investigations, K. Reeves for
technical assistance and R. Krishnamurthy for making the visit to the site
possible and Digvijay Mallah for his assistance.
References
Avner, S. H. 1988. Introduction to Physical Metallurgy. Singapore: Mc Graw
Hill.
Belaiew, N. T. 1918. Damascene steel. Journal of the Iron and Steel Institute,
97(1): 417-37.
Bronson, B. 1986. The making and selling of wootz - a crucible steel of India.
Archaeo-materia/s, 1(1): 13-51.
Buchanan-Hamilton, F. 1807. A Journeyfrom Madras through the Countries
of Mysore. Canara and Malabar. London.
Coomaraswamy, A. K. 1956. Mediaeval Sinhalese Art. New York: Pantheon
Books.
Free stone, I. C. and Tite, M. S. 1986. Refractories in the ancient and
preindustrial world. In Kingery, W. D. (ed.). Ceramics and Civilisation.
Pittsburgh: American Ceramic Society, 35-63.
Lowe, T. L. 1988. Solidification and the crucible processing of Deccani ancient
steel. In Proceedings, Indo-US Conference on Solidification and Materials
Processing at Defence Metallurgical Research Laboratory, Hyderabad, Andhra
Pradesh, India, 19-23 January 1988. Unpublished.
Wootz crucible steel
Lowe, T. L. 1990. Refractories in high-carbon iron processing: a preliminary
study of the Deccani wootz-making crucibles. In Kingery, W. D. (ed.).
Ceramics and Civilization. Cross-Craft and Cross-Cultural Interactions in
Ceramics. Pittsburgh: The American Ceramic Society. Vol. 4: 237-50.
Percy,J. 1860-1880. Metallurgy. Vol. 2, Part 3. Netherlands: De archaeologische
pers Nederland.
Rajan K. 1991. New light on the megalithic cultures of the Kongu region, Tamil
Nadu. Man and Environment. 16( 1): 93-102.
Schoff, W. H. 1915. The eastern iron trade of the Roman empire. Journal of
the American Oriental Society. 35: 224-39.
Scott, D. A. 1991. Metallography and Microstructure of Ancient and Historic
Metals. Singapore: The J. Paul Getty Trust.
Smith, C. S. 1960. A History of Metallography. The Development of Ideas on
the Structure of Metals Before 1890. Chicago: University of Chicago Press.
Tylecote, R. F. 1962. Metallurgy in Archaeology. London: Edward Arnold.
Tylecote, R. F. 1976. A History of Metallurgy. London: The Metals Society.
Verhoeven, J. D. 1987. Damascus steel. Part I: Indian steel. Metallography,
20: 145-5l.
Voysey, H. W. 1832. Description of the manufacture of steel in Southern India.
Journal of Asiatic Society of Bengal, 1: 245-7.
Wood, C. 1893. Discussion of Turner's paper on production of iron in small
furnaces in India. Journal of the Iron and Steel Institute. 44 (2): 177-80.
