Mehergarh cire perdue spoked-wheel of copper alloy is a eraka arā hypertext, signifies moltencast alloy work

Mirror: Mehergarh cire perdue spoked-wheel of copper alloy is a eraka arā hypertext, signifies moltencast alloy work

This note details the archaeo-metallurgical and underlying language (Meluhha) framework related to a remarkable artifact which is 5 mm. dia. It is a spoked-wheel made of copper alloy. How did the artisans among Bharatam Janam of Sarasvati-Sindhu (Harappa) civilization accomplish this technological innovation? I suggest that this ‘amulet’ is a metalwork catalogue in Harappa (Indus) Script, comparable to a compartmental seal of BMAC or Gilund. The spoked-wheel is a recurrent hieroglyph on Harappa (Indus) Script corpora. It occurs four times within a 10-hieroglyph hypertext message on Dholavira advertisement board. The spoked-wheel hieroglyph is a proclamation: eraka ‘knave of wheel’ rebus: erako ‘moltencast copper’ PLUS arā ‘spokes’ rebus: āra ‘brass’. Read on…

This 6,000-year-old amulet is the oldest example of a technology still used by NASA

By Sarah Kaplan November 15
The amulet from Mehrgarh. (D. Bagault/C2RMF)

The amulet doesn’t look like much: A lopsided, six-spoke wheel barely an inch across, swollen and green from corrosion.

But the 6,000-year-old object, uncovered from the ruins of a Neolithic farming village in Pakistan, holds clues about the ancient world it came from. And the effort to decipher those clues required some of the most sophisticated technology of today.

In the journal Nature Communications on Tuesday, scientists describe how they used a powerful synchrotron beam to analyze the tiny amulet on a microscopic level, revealing secrets about its origins that were once thought lost.

The mystery of the amulet

Play Video6:57

Scientist carried out a detailed study to find out how amulets were made 6,000 years ago. (NPG Press)

// Scientists Unravel The Mysteries Behind 6,000-Year-Old Amulet

Published on Nov 16, 2016

Technology has enabled scientists to figure out how a 6,000-year-old amulet resembling a wheel was created.

Technology has enabled scientists to figure out how a 6,000-year-old amulet resembling a wheel was created. The copper object was found decades ago in present-day Pakistan, and it has been traced back to an innovative Neolithic settlement called the Mehrgarh. As the Washington Post reports, researchers gained a unique view into the artifact with a method known as full-field photoluminescence which causes electrons to activate and emit light. Scientists used this technique to measure different factors such as the type of metal used, the levels of oxygen that seeped in, and the temperature at which the substance melted and set. Based on these results, they ultimately determined that the artifact was likely made through lost-wax casting where wax and clay are used to make a mold for metal objects. As such, the amulet has become the earliest known piece made from this method. Researchers think the small piece may have had some significance at the time but have not been able to confirm its true purpose.

Peering through the corrosion, “we discovered a hidden structure that is a signature of the original object, how it was made,” said lead author Mathieu Thoury, a physicist at Ipanema, the European center for the study of ancient materials. “You have a signature of what was happening 6,000 years ago.”

The study relied on an imaging technique called full-field photoluminescence. The researchers shined a powerful light at the amulet, exciting electrons in the atoms that compose it so that they emitted their own light in response. By analyzing the spectrum of this emission, the researchers could figure out the shape and composition of parts of the amulet they couldn’t see.

The technique revealed something surprising: countless tiny, bristle-like rods of copper oxide scattered throughout the interior of the amulet. Their structure was very different from the copper-oxygen compounds that pervade the rest of the object as a result of heavy corrosion over the course of thousands of years.

Thoury believes that ancient metallurgists were trying to craft the amulet out of pure copper, but inadvertently allowed some oxygen in during the production process. Those early copper oxides hardened into the microscopic bristles in the amulet’s interior.
Photoluminescence revealed tiny, bristle-like rods of copper oxide (top right) in the amulet’s interior. (T. Séverin-Fabiani, M. Thoury, L. Bertrand, B. Mille/Ipanema CNRS MCC UVSQ/Synchrotron Soleil/C2RMF)

Their existence, paired with the fact that the amulet is not symmetrical, also suggests that the amulet was made via a process called lost-wax casting — one of the most important innovations in the history of metallurgy. The age-old process, which is still used to make delicate metal instruments today, involves crafting a model out of wax, covering it in clay, and baking the whole thing until the wax melts out and the clay forms a hard mold. Then molten metal is then poured into this cavity and cooled until it hardens. When the mold is broken open, a perfect metal model of the original wax structure remains.

At 6,000 years, the amulet is the oldest known example of this technique. Eventually, lost-wax casting would be used to produce countless functional objects — knives, water vessels, utensils, tools — as well as jewelry, religious figurines, impressive metal statues of gods, kings and heroes. The technique helped societies transition from the Stone Age to the ages of copper and bronze and gave rise to new and powerful types of culture. We have it to thank for the incredible bronze Buddha at Tōdai-ji temple in Japan and Faberge eggs. Investment casting, which is based on the process, is now used to produce equipment for NASA that has flown to the International Space Station and Mars.

In terms of beauty or sophistication, the amulet cannot rival its more famous successors. But Thoury finds it impressive in other ways. Not only did the amulet’s creators use a new casting technique, they also opted to craft the amulet entirely from copper — a rare and unusual choice, since pure copper is hard to acquire and corrodes more easily than an alloy.

“It is not the most beautiful object, but still it holds so much history,” he said. “It shows how the metalworkers at the time were so innovative and wanted to optimize and improve the technique.”
The archaeological site MR2 at Mehrgarh, where the amulet was found. (C. Jarrige/Mission Archéologique de l’Indus)

Mehrgarh, the ancient settlement where the amulet was uncovered 35 years ago, is already known as a “crucible” of innovation, Thoury added. The first evidence of proto-dentistry was uncovered at the site, which is more than 600 miles southwest of Islamabad. It also contains some of the most ancient evidence of agriculture and the oldest ceramic figurines in South Asia. It’s thought that this small farming community was a precursor to the entire Indus Valley civilization, one of the most important cultures in the ancient world.

“I’m really impressed that these people at the time were so keen on experimenting,” Thoury said. As a scientist, that’s an impulse he knows well.

Sarah Kaplan is a reporter for Speaking of Science.

 Follow @sarahkaplan48

The earliest lost-wax cast object is by Bharatam Janam. Metallurgy explained — M. Thoury et al (March 2016). Harappa Script & Language explained.

Harappa (Indus) script hieroglyph: eraka ‘knave of wheel’ rebus: eraka ‘moltencast, metal infusion’; era ‘copper’. āra ‘spokes’ arā ‘brass’ erako molten cast (Tulu) Ka. eṟe to pour any liquids, cast (as metal); n. pouring; eṟacu, ercu to scoop, sprinkle, scatter, strew, sow; eṟaka, eraka any metal infusion; molten state, fusion.Tu. eraka molten, cast (as metal); eraguni to melt (DEDR 866)  agasāle, agasāli, agasālevāḍu 89,1]m. ( √ अर्च्) , Ved. a ray , flash of lightning RV. &cthe sun RV. &c Rebus: copper L.அருக்கம்¹ arukkam, n. < arka. (நாநார்த்த.) 1. Copper; செம்பு. 2. Crystal; பளிங்கு. அக்கம்&sup4; akkam

, n. < arka. An ancient coin = 1/12 காசு; ஒரு பழைய நாணயம். (S. I. I. ii. 123.)

అగసాలి (p. 23agasāli or అగసాలెవాడు agasāli. [Tel.] n. A goldsmith. కంసాలివాడు.

Kannada Glosses

erka = ekke (Tbh. of arka) aka (Tbh. of arka) copper (metal); crystal (Ka.lex.) cf. eruvai = copper (Tamil)


See: The cire perdue spoked wheel of copper+lead alloy was NOT an amulet, it was a metal artifact, a metal coin, akkam; it was a compartmental Harappa seal with Harappa (Indus) Script hieroglyph. May or may not have been used as a coin to value and exchange goods but a proclamation of the metallurgical excellence achieved by Bharatam Janam of 4th millennium BCE.

Artisans at work in Burma making Karen drum


Sun motif in the centre of the tympanum, Karen drum.

arká1 m. ʻ flash, ray, sun ʼ RV. [√arc] Pa. Pk. akka — m. ʻ sun ʼ, Mth. āk; Si. aka ʻ lightning ʼ, inscr. vid — äki ʻ lightning flash ʼ.(CDIAL 624) rebus: erako ‘moltencast’ arka, eraka ‘gold, copper’.

Detail of the tympanum of Karen drum.
ayo ‘fish’ rebus; aya ‘iron’ ayas ‘metal alloy’

Frog on the Karen bronze pancaloha ‘five metal alloys’ drum.

Kur. mūxā frog. Malt. múqe id. / Cf. Skt. mūkaka– id. (DEDR 5023) Rebus: mū̃h ‘ingot’ mũhe ‘ingot’ mũhã̄ = the quantity of iron produced at one time in a native furnace.

Elephant motif.  karba, ibha ‘elephant’ rebus: karba, ib ‘iron’.Ta. ayil iron. Ma. ayir, ayiram any ore. Ka. aduru native metal. Tu. ajirda karba very hard iron. (DEDR 192)                                                                                      “The town of Nwe Daung, 15 km south of Loikaw, capital of Kayah (formerly Karenni) State, is the only recorded casting site in Burma. Shan craftsmen made drums there for the Karens from approximately 1820 until the town burned in 1889.  Karen drums were cast by the lost wax technique; a characteristic that sets them apart from the other bronze drum types that were made with moulds. A five metal formula was used to create the alloy consisting of copper, tin, zinc, silver and gold. Most of the material in the drums is tin and copper with only traces of silver and gold. The Karen made several attempts in the first quarter of the twentieth century to revive the casting of drums but none were successful.”


In ancient Indian texts, such as Manasollasa, Silparatna, Manasara,the cire perdue technique is referred to as madhucchiṣṭa vidhānam.  मधु madhu -उच्छिष्टम्,-उत्थभ्,-उत्थितभ् 1 bees’-wax; शस्त्रासवमधूच्छिष्टं मधु लाक्षा च बर्हिषः Y.3.37; मधूच्छिष्टेन केचिच्च जध्नुरन्योन्यमुत्कटाः Rām.5.62.11.-2 the casting of an image in wax; Mānasāra; the name of 68th chapter. This technique was clearly attested in the Epic Rāmāyaa. मधुशिष्ट madhuśiṣṭa ‘wax’ (Monier-Williams, p. 780).

karaṇḍa ‘duck’ (Sanskrit) karaṛa ‘a very large aquatic bird’ (Sindhi) karaDa ‘safflower’ rebus:karaḍa ‘double-drum’ Rebus: करडा [ karaḍā ] Hard from alloy–iron, silver &c kharādī = turner (Gujarati)

कारण्डवः, पुं, स्त्री, (ञमन्ताडड इति रमेर्ड । रण्डः । ईषत् रण्डः । “ईषदर्थे” ६ । ७ । १०५ । इति कोः कादेशः । कारण्डं वाति । वा गतौ + “आतोनुपेति” । ३ । २ । ३ । कः । करण्डस्येदं कारण्डं तदाकारं वाति वा ।) हंसविशेषः इत्यमरः । २ । ५ । ३४ ॥ खडहाँस इति भाषा (यथा ऋतुसंहारे । शरद्वर्णणे ८ । “कारण्डवाननविघट्टितवीचिमालाः कादम्बसारसकुलाकुलतीरदेशाः” ॥)शब्दकल्पद्रुमः

कारण्डव पुंस्त्री रम–ड तस्य नेत्त्वम् रण्डः ईषत् रण्डः कारण्डः तं वाति वा–क करण्डस्येदं कारण्डं तदाकारं वाति वा–क वा । हंसभेदे “हंसकाण्डवोद्गीताः सारसाभिरुतास्तथा” भा० व० ३८ अ० । स्त्रियां जाति- त्वात् ङीष् । अस्य अजिरादि० पाष्ठात् मतौ संज्ञायामपि न दीर्घः कारण्डववती नदीविशेषः । “हससारसक्रौञ्च- चक्रवाककुररकादम्बकारण्डवेत्युपक्रमे” प्लवाः सधचारिणश्च” इति सुश्रुते तस्य प्लवत्वं स घचारित्वञ्चोक्तम् ।वाचस्पत्यम्
kuṭhi ‘tree’ rebus kuṭhi ‘a furnace for smelting iron ore, to smelt iron’) tALa ‘palm trees’ rebus: DhALa ‘large ingot (oxhide)’

Hieroglyphs of Indus Script Cipher are sitnified on the Shahi Tump leopard weight which has been produced using the lost-wax casting method. The hieroglyphs are: 1. leopard; 2. ibex or antelope; 3. bees (flies). The rebus-metonymy readings in Meluhha are:

karaḍa  ‘panther’; karaḍa tiger (Pkt); खरडा [ kharaḍā ]  A leopard. खरड्या [ kharaḍyā ] m or खरड्यावाघ m A leopard (Marathi). Kol. keḍiak  tiger. Nk.  khaṛeyak  panther.  Go. (A.) khaṛyal tiger; (Haig) kariyāl panther Kui kṛāḍi, krānḍi tiger, leopard, hyena.  Kuwi (F.) kṛani tiger; (S.) klā’ni tiger, leopard; (Su. P. Isr.) kṛaˀni (pl. -ŋa) tiger. / Cf. Pkt. (DNM) karaḍa- id. (DEDR 1132).Rebus: करडा [karaḍā] Hard from alloy–iron, silver &c. (Marathi)  kharādī ‘ turner, a person who fashions or shapes objects on a lathe’ (Gujarati)

Hieroglyph: miṇḍāl ‘markhor’ (Tōrwālī) meḍho a ram, a sheep (Gujarati)(CDIAL 10120) Rebus: mẽṛhẽt, meḍ ‘iron’ (Munda.Ho.) mr̤eka, melh ‘goat’ (Telugu. Brahui) Rebus: melukkha ‘milakkha, copper’. If the animal carried on the right hand of the Gudimallam hunter is an antelope, the possible readings are: ranku ‘antelope’ Rebus: ranku ‘tin’.

Ka. mēke she-goat; mē the bleating of sheep or goats.  Te. mē̃ka,  mēka goat.

Kol. me·ke id. Nk. mēke id. Pa. mēva, (S.) mēya she-goat. Ga. (Oll.)mēge, (S.) mēge goat. Go. (M) mekā, (Ko.) mēka id. ? Kur. mēxnā (mīxyas) to call, call after loudly, hail. Malt. méqe to bleat. [Te. mr̤ēka (so correct) is of unknown meaning. Br. mēḻẖ is without etymology; see MBE 1980a.] / Cf. Skt. (lex.) meka- goat. (DEDR 5087). Meluhha, mleccha (Akkadian. Sanskrit). Milakkha, Milāca ‘hillman’ (Pali) milakkhu ‘dialect’ (Pali) mleccha ‘copper’ (Prakritam).

The bees are metaphors for wax used in the lost-wax casting method.

Hieroglyph: माक्षिक [p= 805,2] mfn. (fr. मक्षिका) coming from or belonging to a bee Rebus: ‘pyrites’: माक्षिक [p= 805,2] n. a kind of honey-like mineral substance or pyrites MBh. उपधातुः An inferior metal, semi-metal. They are seven; सप्तोपधातवःस्वर्णं माक्षिकं तारमाक्षिकम् । तुत्थं कांस्यं च रातिश्च सुन्दूरं च शिलाजतु ॥ उपरसः uparasḥउपरसः 1 A secondary mineral, (red chalk, bitumen, माक्षिक, शिलाजित &c).(Samskritam)

mákṣā f., mákṣ — m. f. ʻ fly ʼ RV., mákṣikā — f. ʻ fly, bee ʼ RV., makṣika — m. Mn.Pa. makkhikā — f. ʻ fly ʼ, Pk. makkhiā — f., macchī — , °chiā — f.; Gy. hung. makh ʻ fly ʼ, wel. makhī f., gr. makí f., pol. mačin, germ. mačlin, pal. mắki ʻ mosquito ʼ,măkīˊla ʻ sandfly ʼ, măkīˊli ʻ house — fly ʼ; Ash. mačī˜ˊ ʻ bee ʼ; Paš.dar. mēček ʻ bee ʼ, weg. mečīˊk ʻ mosquito ʼ, ar. mučəkmučag ʻ fly ʼ; Mai. māc̣hī ʻ fly ʼ; Sh.gil.măṣīˊ f., (Lor.) m*lc̣ī ʻ fly ʼ (→ Ḍ. m*lc̣hi f.), gur. măc̣hīˊ ʻ fly ʼ (ʻ bee ʼ in gur. măc̣hi̯kraṇ, koh. măc̣hi — gŭn ʻ beehive ʼ); K. mȧchi f. ʻ fly, bee, dark spot ʼ; S. makha,makhi f. ʻ fly, bee, swarm of bees, sight of gun ʼ, makho m. ʻ a kind of large fly ʼ; L. (Ju.) makhī f. ʻ fly ʼ, khet. makkīˊ; P. makkh f. ʻ horsefly, gnat, any stinging fly ʼ, m. ʻ flies ʼ, makkhī f. ʻ fly ʼ; WPah.rudh. makkhī ʻ bee ʼ, jaun. mākwā ʻ fly ʼ; Ku. mākho ʻ fly ʼ, gng. mã̄kh, N. mākho, A. mākhi, B. Or. māchi, Bi. māchī, Mth. māchī,mã̄chīmakhī (← H.?), Bhoj. māchī; OAw. mākhī, lakh. māchī ʻ fly ʼ, ma — mākhī ʻ bee ʼ (mádhu — ); H. māchīmākhīmakkhī f. ʻ fly ʼ, makkhā m. ʻ large fly, gadfly ʼ; G. mākhmākhī f. ʻ fly ʼ, mākhɔ m. ʻ large fly ʼ; M. mās f. ʻ swarm of flies ʼ, n. ʻ flies in general ʼ, māśī f. ʻ fly ʼ, Ko. māsumāśi; Si. balu — mäkka, st. — mäki — ʻ flea ʼ, mässa, st. mäsi — ʻ fly ʼ; Md. mehi ʻ fly ʼ.
*makṣātara — , *mākṣa — , mākṣiká — ; *makṣākiraṇa — , *makṣācamara — , *makṣācālana — , *makṣikākula — ; *madhumakṣikā — .
Addenda: mákṣā — : S.kcch. makh f. ʻ fly ʼ; WPah.kṭg. mákkhɔmáṅkhɔ m. ʻ fly, large fly ʼ, mákkhi (kc. makhe) f. ʻ fly, bee ʼ, máṅkhi f., J. mākhī, Garh. mākhi. (CDIAL 9696) mākṣiká ʻ pertaining to a bee ʼ MārkP., n. ʻ honey ʼ Suśr. 2. *mākṣa — . [mákṣā — ]
1. WPah.bhad. māċhī ʻ bee ʼ, khaś. mākhī; — Pk. makkhia — , macchia — n. ʻ honey ʼ; Ash. mačimačík ʻ sweet, good ʼ, mačianá ʻ honey ʼ; Wg. mác̣imäc̣ ʻ honey ʼ, Kt. mac̣ī˜, Pr. maṭék, Shum. mac̣hī, Gaw. māc̣hī, Kal.rumb. Kho. mac̣hí, Bshk. mē̃c̣h, Phal. mn/ac̣hīmḗc̣hī, Sh. măc̣hīˊ f., S. L. mākhī f., WPah.bhiḍ. māċhī n., H.mākhī f.
2. K. mã̄ch, dat. °chas m. ʻ honey ʼ, WPah.bhal. māch n. — For form and meaning of Paš. māšmōṣ ʻ honey ʼ see NTS ii 265, IIFL iii 3, 126.
*mākṣakulika — , *mākṣikakara — , *mākṣikamadhu — .Addenda: mākṣika — : Kho. mac̣hi ʻ honey ʼ BKhoT 70.(CDIAL 9989)*mākṣikakara or *mākṣakara — ʻ bee ʼ. [Cf. madhu- kara — m. ŚārṅgP., °kāra — m. BhP., °kārī — f. R.  mākṣiká — , kará — 1]
Ash. mačarīk°čerīˊk ʻ bee ʼ, Wg. mac̣arīˊk, Kt. mačerík NTS ii 265, mac̣e° Rep1 59, Pr. mučeríkməṣkeríkmuṭkurīˊk, Shum. mã̄c̣hāˊrik, Kal.rumb. mac̣hḗrik, Bshk.māˊc̣ēr, Phal. māc̣hurīˊ f.; Sh.koh. măc̣hāri f. ʻ bee ʼ, gil. (Lor.) m*lc̣hari ʻ bee, wasp, hornet ʼ (in latter meaning poss. < *makṣātara — ); P. makhīr m. ʻ bee ʼ, kgr. ʻ honey ʼ; — Gaw. mã̄c̣(h)oṛík with unexpl. —  — . (CDIAL 9990)  *mākṣikamadhu ʻ honey ʼ. [mākṣiká — , mádhu — ]
P. mākhyō̃ f., mākho m. ʻ honey, honeycomb ʼ.(CDIAL 9991) مچئِي mac̱ẖaʿī, s.f. (6th) A bee in general. Sing. and Pl. سره مچئِي saraʿh-mac̱ẖaʿī, s.f. (6th). Sing. and Pl.; or دنډاره ḏḏanḏḏāraʿh, s.f. (3rd) A hornet, a wasp. Pl. يْ ey. See ډنبره (Pashto) माक्षिक [p= 805,2] mfn. (fr. मक्षिका) coming from or belonging to a bee Ma1rkP. मक्षिकः makṣikḥ मक्षि makṣi (क्षी kṣī) का kāमक्षिकः मक्षि (क्षी) का A fly, bee; भो उपस्थितं नयनमधु संनिहिता मक्षिका च M.2.-Comp.-मलम् wax.  madhu

मधु a. -मक्षः, -क्षा, -मक्षिका a bee. (Samskritam) माक्षिक [p= 805,2] n. a kind of honey-like mineral substance or pyrites MBh. उपधातुः An inferior metal, semi-metal. They are seven; सप्तोपधातवः स्वर्णं माक्षिकं तारमाक्षिकम् । तुत्थं कांस्यं च रातिश्च सुन्दूरं च शिलाजतु ॥ उपरसः uparasḥउपरसः 1 A secondary mineral, (red chalk, bitumen, माक्षिक, शिलाजित &c).(Samskritam) மாக்கிகம் mākkikam, n. < mākṣika. 1. Bismuth pyrites; நிமிளை. (நாமதீப. 382.) 2. Honey; தேன். (நாமதீப. 410.) செம்புத்தீக்கல் cempu-t-tīkkal

, n. < செம்பு +. Copper pyrites, sulphide of copper and iron; இரும்புஞ்செம்புங்கலந்த உலோகக்கட்டி. Loc.
Leopard weight. Shahi Tump. H.16.7cm; dia.13.5cm; base dia 6cm; handle on top.  Seashells inlays on frieze. The pair of leopard and ibex is shown twice, separated by stylized flies.

“The artefact was discovered in a grave, in the Kech valley, in eastern Balochistan. It belongs to the Shahi Tump – Makran civilisation (end of 4th millennium — beginning of 3rd millennium BCe). Ht. 200 mm. weight: 13.5 kg. The shell has been manufactured by lost-wax foundry of a copper alloy (12.6%b, 2.6%As), then it has been filled up through lead (99.5%) foundry. The shell is engraved with figures of leopards hunting wild goats, made of polished fragments of shellfishes. No identification of the artefact’s use has been given. (Scientific team: B. Mille, D. Bourgarit, R. Besenval, Musee Guimet, Paris).”
Source: Tump Leopard weight of Shahi Tump (Balochistan), National Museum, Karachi. The artefact was discovered in a grave, in the Kech valley, in Balochistan. ca. 4th millennium BCE. 200 mm. h. 13.5kg wt. The shell has been manufactured by lost-wax foundry of a copper alloy (12.6% Pb, 2.6% As), then it has been filled up through lead (99.5%) foundry. The shell is engraved with figures of leopards hunting wild goats, made of polished fragments of shellfishes. No identification of the artefact’s use has been given. (Scientific team: B. Mille, D. Bourgarit, R. Besenval, Musee Guimet, Paris.

Meluhha hieroglyphs:

karaḍa  ‘panther’ Rebus: karaḍa ‘hard alloy’. mlekh ‘goat’ Rebus: milakkhu ‘copper’ (Pali)


The pinnacle of achievement in Bronze Age Revolution relates to the invention of cire perdue technique of metal castings to produce metal alloy sculptures of breath-taking beauty. This achievement is exemplified by Nihal Mishmar artifacts dated to ca. 5th millennium BCE.

Mehergarh. 2.2 cm dia. 5 mm reference scale. Perhaps coppper alloyed with lead. [quote]Bourgarit and Mille (Bourgarit D., Mille B. 2007. Les premiers objets métalliques ont-ils été fabriqués par des métallurgistes ? L’actualité Chimique . Octobre-Novembre 2007 – n° 312-313:54-60) have  reported the finding (probably in the later still unreported excavation period) of small Chalcolithic “amulets” which they claim to have been produced by the process of Lost Wax. According to them, “The levels of the fifth millennium Chalcolithic at Mehrgarh have delivered a few amulets in shape of a minute wheel, while the technological study showed that they were made by a process of lost wax casting. The ring and the spokes were modelled in wax which was then coated by a refractory mould that was heated to remove the wax. Finally, the molten metal was cast in place of the wax. Metallographic examination confirmed that it was indeed an object obtained by casting (dendrite microstructure). This discovery is quite unique because it is the earliest attestation of this technique in the world.” They then, further on, state that “The development of this new technique of lost wax led to another invention, the development of alloys…Davey (Davey C. 2009.The Early History of Lost-Wax Casting, in J. Mei and Th. Rehren (eds), Metallurgy and Civilisation: Eurasia and Beyond Archetype, pp. 147-154. London: Archetype Publications Ltd.) relies only upon these Mehrgarh findings , as well as on the Nahal Mishmar hoard, to claim that Lost Wax casting began in the Chalcolithic period before 4000 BCE.” [unquote]  (Shlomo Guil)

Shahi Tump. Kech valley, Makran division, Baluchistan, Pakistan (After Fig. 1 in Thomas et al)Benoit Mille calls the bronze stamps of Shahi-Tump ‘amulets’ (made from copper alloyed with lead). Mehrgarh is well recognised as a centre for early pyrotechnologies.The wax models of the stamps would have   been solid and     may have had a simple core inserted.This is perhaps the first stage in the technology:”Small copper-base wheel-shaped “amulets” have been unearthed from the Early Chalcolithic levels at Mehrgarh in Balochistan (Pakistan), dating from the late fifth millennium B.C. Visual and metallographic examinations prove their production by a lost-wax process—the earliest evidence so far for this metalworking technique. Although a gap of more than 500 years exists between these ornaments from Mehrgarh and the later lost-wax casts known in the Indo-Iranian world, the technological and compositional links between these artefacts indicate a similar tradition. We already know that the lost-wax process was commonly used during the second half of the fourth millenium B.C, as exemplified by figurative pinheads and compartmented seals, the latter of which were produced and distributed across the region until the early second millennium B.C. Most, if not all, of these artefacts were made using the lost-wax technique. This intensive practice of lost-wax  lasting certainly stimulated the technical development of the process, allowing the elaboration of more complex and heavier objects. The “Leopards Weight” (Balochistan, late fourth or early third millennium B.C.) is one of the best examples of these developments: the lost-wax copper jacket, with its opened hollow shape, constitutes an extraordinary technical achievement.(Mille, B., Bourgarit, D., and Besenval, R. 2005. ‘Metallurgical study of the ‘Leopards weight’ from Shahi-Tump (Pakistan)’, in C. Jarrige and V. Lefevre, eds., South Asian Archaeology 2001, Editions Recherches sur les Civilisations, Paris: 237-44) True hollow casting does not appear until the third millennium B.C., as illustrated by the manufacture of statuettes, including the Nausharo bull figurine (Balochistan, 2300–2100 B.C.), or those from BMAC sites in Central Asia (based upon analyses of items in the Louvre collections). The birth of the lost-wax casting process can also be paralleled with the first emergence of alloying in South Asia, as many of these early lost-wax cast artefacts were made of a copper-lead alloy (c. 10–40 wt% Pb and up to 4 wt% As). Significantly, it seems that the copper-lead alloy was solely dedicated to artefacts made using the lost-wax technique, a choice no doubt driven by the advantageous casting properties of such an alloy.” (Mille, Benoit, On the origin of lost-wax casting and alloying in the Indo-Iranian world, in: Lloyd Weeks, 2007, The 2007 Early Iranian metallurgy workshop at the University of Nottingham)
(Source: B. Mille, R. Besenval, D. Bourgarit, 2004, Early lost-wax casting in Balochistan (Pakistan); the ‘Leopards weight’ from Shahi-Tump. in: Persiens antike Pracht, Bergbau-Handwerk-Archaologie, T. Stollner, R Slotta, A Vatandoust, A. eds., pp. 274-280. Bochum: Deutsches Bergbau Museum, 2004.


Mille, B., D. Bourgarit, JF Haquet, R. Besenval, From the 7th to the 2nd millennium BCE in Balochistan (Pakistan): the development of copper metallurgy before and during the Indus Civilisation, South Asian Archaeology, 2001, C. Jarrige & V. Lefevre, eds., Editions Recherches sur les Civilisations, Paris, 2005.)


“Benoit Mille has drawn attention to copper alloy ‘amulets’ discovered in the early Chalcolithic (late 5th millennium) levels of Mehrgarh in Baluchistan, Pakistan. He reported that metallographic examination established that the ornaments were cast by the lost-wax method (Mille, B., 2006, ‘On the origin of lost-wax casting and alloying in the Indo-Iranian world’, in Metallurgy and Civilisation: 6th international conference on the beginnings of the use of metals and alloys, University of Science and Technology, Beijing, BUMA VI). The amulets were made from copper alloyed with lead. Mehrgarh is well recognised as a centre for early pyrotechnologies. The wax models of the amulets would have been solid and may have had a simple core inserted. This is understandably the first stage in the technology. Mille also draws attention to the ‘Leopards weights’ from Baluchistan, dating to about 3000 BCE which were made using a complex core keyed into the investment mould.”(Davey, Christopher J., The early history of lost-wax casting, in: J. Mei and Th. Rehren, eds., Metallurgy and Civilisation: Eurasia and Beyond Archetype, London, 2009, ISBN 1234 5678 9 1011, pp. 147-154; p. 151).

Remarkable evidences of the excellence achived in cire perdue metal catings are provided by bronze or copper alloy artifacts kept in the British Museum, said to have been acquired from Begram, and dated to ca. 2000 to 1500 BCE.




Six bronze stamps (a-b) circular with pin-wheel design recalling a svastika (c) square with heart-shaped pattern; broken lug on the back (d-f) broken with radiating spokes; one with broken lug.
Cast, copper alloy, circular, openwork seal or stamp, comprising five wide spokes with projecting rims, radiating from a circular hub also encircled by a flange. The outer rim is mostly missing and two spokes are broken. The back is flat, with the remains of a broken attachment loop in the centre.
2000BC-1500BC (circa) Copper alloy. Pierced. cast.
Made in: Afghanistan(Asia,Afghanistan)

Found/Acquired: Begram (Asia,Afghanistan,Kabul (province),Begram)


Curator’s comments
IM.Metal.154: ‘Six bronze stamps for impressing designs’.

  1. Fabrègues: Together with 1880.3710.b-c, the object belongs to the large class of compartmented seals. Such partitioned seals are characteristic of the Bactria-Margiana Archaeological Complex (BMAC, also known as the Oxus Civilization), the modern archaeological designation for a Bronze Age culture located along the upper Amu Darya (Oxus River) in present-day Turkmenistan, Afghanistan, southern Uzbekistan and western Tajikistan. The BMAC may have extended as far as southern Afghanistan and Baluchistan, which have also yielded artefacts typical of the culture.




Copper alloy.


Cast, copper alloy, circular, openwork seal or stamp, comprising five wide spokes with projecting rims, radiating from a circular hub also encircled by a flange. The outer rim is mostly missing and two spokes are broken. The back is flat, with the remains of a broken attachment loop in the centre.


1880.3710.a IM.Metal.154: ‘6 bronze stamps for impressing designs’.

  1. Fabrègues: Together with 1880.3710.b-c, the object belongs to the large class of compartmented seals. Such partitioned seals are characteristic of the Bactria-Margiana Archaeological Complex (BMAC, also known as the Oxus Civilization), the modern archaeological designation for a Bronze Age culture located along the upper Amu Darya (Oxus River) in present-day Turkmenistan, Afghanistan, southern Uzbekistan and western Tajikistan. The BMAC may have extended as far as southern Afghanistan and Baluchistan, which have also yielded artefacts typical of the culture.

Compartmented seals have been found in large numbers in these areas, both from clandestine diggings in the 1970s (Pottier 1984, Tosi 1988, fig.11, Salvatori 1988) and from scientific excavations. Known sites where examples have been excavated are: Namazga on the banks of the Murghab river (Masson and Sarianidi 1972) Togolok (Sarianidi 1990) and Gonur Tepe in Margiana (Sarianidi 1993, 2002), Dashly Tepe (Masson and Sarianidi 1972) and Mundigak (Casal 1961) in Afghanistan, Dabar Kot, Rana Gundai and Shahi Tump (Amiet 1977, p.117), and the Mehrgarh-Sibri complex (Sarianidi 1993, p.37) in Baluchistan.
These seals depict geometrical motifs, like 1880.3710.a–c, and also floral motifs, crosses, animals such as goats, snakes and scorpions, birds (primarily eagles with spread wings), human figures and fantastic dragons. 1880.3710.a, c closely resemble some examples from plundered tombs in Bactria, now in the Louvre Museum (Amiet 2002, p.168, fig.13.h, l) and 1880.3710.c an example said to come from southern Bactria, now in a private collection (Salvatori 1988, p.183, fig.49, bottom right).
Impressions of such seals have been found on pottery. Scholars disagree about their use. It has been suggested that they were used for administrative control of trade and production (Hiebert 1994, p. 380); were related to a well organised trade system which involved transporting and transacting goods over long distances (Salvatori 1988, p.163); were symbols of power and property, or, since a large number have similar images, they may have served as amulets protecting their owners from evil rather than as symbols of ownership (Sarianidi 2002, p.41).
Compartmented seals have been variously dated to the end of the 3rd/beginning of the 2nd millennium (Amiet 1977, p.119, Salvatori 1988), or to the first half of the 2nd millennium BC (Tosi 1988, p.123, Sarianidi 1993, p.36). According to Amiet (1977, p.117, 1988, pp.166, 169), they originated in Iranian Sistan: at Shar-i-Sokhta their development can be charted throughout the 3rd millennium BC from steatite prototypes and it is only here and at Shahdad, on the other side of the Lut desert in the Kerman region, that they are known to have been used as marks on pottery (Hakemi and Sajjadi 1988, pp.145, 150). Sarianidi considers this a purely local invention (2002, p.41).
The Begram seals add to the number of examples already available, provide an exact provenance for some varieties and evidence that the Begram plain had interaction with the BMAC.
Amiet, P. (1977) ‘Bactriane proto-historique’, Syria LIV, pp.89–121.
Amiet, P. (1988) ‘Antiquities of Bactria and outer Iran in the Louvre collection’, in Ligabue G. and Salvatori, S. eds. Bactria. An Ancient Oasis from the Sands of Afghanistan, Venice, pp.159–80.
Casal, J.M. (1961) Fouilles de Mundigak, Mémoires de la Délégation archéologique française en Afghanistan XVII, Paris.
Hakemi, A. and Sajjadi, S.M.S. (1988) ‘Shahdad excavations in the context of the Oasis civilization’, in Ligabue G. and Salvatori, S. eds. Bactria. An Ancient Oasis from the Sands of Afghanistan, Venice, pp.143–53.
Hiebert F. (1994) ‘Production evidence for the origin of the Oxus Civilization’, Antiquity 68, pp. 372-87.
Masson, V.M. and Sarianidi V.I. (1972) Central Asia. Turkmenia before the Achaemenids, New York– Washington.
Parpola, A. (1997) ‘Seals of the greater Indus Valley’, in Collon, D. ed. 7000 Years of Seals, London, pp.51, 53, nos.3/16, 3/17.
Salvatori, S. ‘Early Bactrian objects in private collections’, in Ligabue G. and Salvatori, S. eds. Bactria. An Ancient Oasis from the Sands of Afghanistan, Venice, pp.181–7.
Sarianidi, V. (1993) ‘Excavations at Southern Gonur’, Iran XXXI, pp.25–39.
Sarianidi, V. (2002) ‘The palace and necropolis of Gonur’, in Rossi-Osmida, G. (ed.) Margiana. Gonur Depe Necropolis. 10 Years of Excavations by Ligabue Study and Research Centre, Florence, pp.17–49.
Tosi, M. (1988) ‘The origin of early Bactrian civilization’, in Ligabue G. and Salvatori, S. eds. Bactria. An Ancient Oasis from the Sands of Afghanistan, Venice, pp. 109–23.

High spatial dynamics-photoluminescence imaging reveals the metallurgy of the earliest lost-wax cast object

  • Nature Communications7, Article number: 13356 (2016)
  • doi:10.1038/ncomms13356
  • Download Citation


01 March 2016


26 September 2016

Published online:

15 November 2016


Photoluminescence spectroscopy is a key method to monitor defects in semiconductors from nanophotonics to solar cell systems. Paradoxically, its great sensitivity to small variations of local environment becomes a handicap for heterogeneous systems, such as are encountered in environmental, medical, ancient materials sciences and engineering. Here we demonstrate that a novel full-field photoluminescence imaging approach allows accessing the spatial distribution of crystal defect fluctuations at the crystallite level across centimetre-wide fields of view. This capacity is illustrated in archaeology and material sciences. The coexistence of two hitherto indistinguishable non-stoichiometric cuprous oxide phases is revealed in a 6,000-year-old amulet from Mehrgarh (Baluchistan, Pakistan), identified as the oldest known artefact made by lost-wax casting and providing a better understanding of this fundamental invention. Low-concentration crystal defect fluctuations are readily mapped within ZnO nanowires. High spatial dynamics-photoluminescence imaging holds great promise for the characterization of bulk heterogeneous systems across multiple disciplines.


For the last 15 years, specific cutting-edge developments have led to considerable improvements in photoluminescence-based analysis. Life sciences and semiconductor physics have been the main drivers strongly influencing instrumental choices1,2. In particular, monitoring target biomolecules with fluorescence imaging has led to major breakthrough in biomedical research3. A critical development has been specific antibody tagging, which provides the specificity and high quantum yield required to map and dynamically follow proteins within tissues at cellular level4. In solid-state physics, high-resolution low-temperature (helium) photoluminescence micro-spectroscopy has become the preferred technique to assess intrinsic electronic properties from individual nanostructures, such as the early state of chemical doping in single-walled carbon nanotubes5. Interpretation of spectral signatures collected at room temperature is challenging as emission bands are thermally broadened, particularly owing to the temperature-dependent phonon-coupling factors. Ultra high analytical sensitivity, great ease of use and emergence of super-resolved imaging have been instrumental to further establish photoluminescence as an essential tool in these fields. These optimizations have been driven by specific constraints; for instance, attaining nanoscale spatial resolutions has triggered near-field scanning at the expense of narrow fields of view and stringent requirements in sample surface roughness and slope. However, if major developments including near-field configuration, specific labelling and cryogenic environment have strongly enhanced the capability of characterizing specific biomolecules and semiconductor nanostructures, they are not directly applicable to imaging much of the very large range of mixed-compositional materials that are heterogeneous at bulk, such as those encountered in environmental, material, earth or planetary sciences, engineering and so on. In these samples, significant areas need to be studied at high spatial resolution to attain a statistically significant representation of materials’ heterogeneity. Even for materials where specific staining would be applicable, it is often not an option owing to the alteration induced on the analyte. Characterization therefore needs to resort to autoluminescence. However, the high contrast in luminescence yields between intrinsic luminophores becomes a limiting factor. In addition, many samples cannot tolerate mechanical stress or chemical transformation induced by large temperature changes when placed in a cryogenic environment6. To tackle the characterization of such materials, the ideal system would allow covering all length scales from micrometric resolution to centimetres, providing wide tunability in excitation energy and detection from the deep ultraviolet to the near infrared to collect autoluminescent signatures, while being efficient at room temperature. Here we demonstrate the great benefit of gigapixel luminescence images obtained from coupling full-field imaging and optimized raster scanning. Versatile characterization of complex low-intensity photoluminescence signatures from crystallite sizes to whole macroscopic objects opens a new possibility for the study of polycrystalline semiconductors and other heterogeneous materials. For these materials, ensuring the best compromise between full tunability in excitation and emission, high spatial dynamics, that is, a high ratio between field of view and lateral resolution, and convenient room-temperature operation, is often more critical than reaching nanometric resolution. This means, for example, that we were able to study fluctuations in crystal defect density at the submicrometric scale while imaging this behaviour over centimetres. The wide tunability of the excitation, owing to the ability to switch between conventional and synchrotron sources, allows selecting an optimized excitation of luminophores above 200 nm.

We demonstrate this improved capability on two applications. Although use of advanced photoluminescence imaging has never been reported in archaeology, imaging reveals a hidden microstructure across a particularly challenging archaeological artefact. In a fully corroded 6,000-year-old small amulet identified as the earliest lost-wax cast and discovered in Mehrgarh (Baluchistan, Pakistan), one of the most important archaeological sites from the early Neolithic period, the clue to the entire metallurgical process of the earliest lost-wax cast amulet is provided by multiscale photoluminescence imaging. The methodology identifies the coexistence of two hitherto indistinguishable non-stoichiometric cuprous oxide phases and allows visualization of the spatial distribution of a ghost fossilized eutectic system, which reveals the innovative process they developed. All the images were collected on a fully customized synchrotron full-field microscope equipped with multispectral detection. The overall data cube results from the mosaicking of 414 tiles collected in three emission bands at three excitation energies, totalling 1.5 gigapixels. Using the same strategy, we could image structured crystal defects fluctuation within individual ZnO nanowires across populations of hundreds, from their low-yield photoluminescence. The continuous tunability of the synchrotron beam allows excitation down to the shortwave ultraviolet (UVC). We therefore demonstrate the exceptional potential of high spatial dynamics-photoluminescence imaging to study nano- and polycrystalline materials for applications within a variety of fields, ranging from quality control in semiconductor solid-state physics to geophysics, archaeology and environmental sciences.


The Mehrgarh amulet is the earliest known lost-wax cast object

To highlight the novelty of our approach, we report the information revealed by high spatial dynamics-photoluminescence imaging on a six-millennia old amulet discovered at Mehrgarh (Baluchistan, Pakistan), one of the most important archaeological sites from the early Neolithic period in the Ancient Near East (Fig. 1 and Supplementary Fig. 1).

Figure 1: The amulet MR. from Mehrgarh.

(a) Map indicating the major Indo-Iranian archaeological sites dated from the seventh to the second millennia BC. Scale bar, 200 km. (b) View of the MR2 archaeological site at Mehrgarh (sector X, Early Chalcolithic, end of period III, 4,500–3,600 BC). (c) View of the front side of the wheel-shaped amulet. Scale bar, 5 mm. (d) Dark-field image of the equatorial section of the amulet.

Full size image

The ornament with inventory number MR. was studied in detail (Fig. 1c,d). A visual inspection indicates that its ‘spoked wheel’ shape consists of six small rods lying on a ring of 20 mm diameter. At the centre of the wheel, the spokes were clearly pressed on each other until a junction was obtained by superposition; the base of each spoke was attached to the support ring using the same technique. Both the spokes and the support ring are circular in section. Only a wax-type material, that is, easily malleable and fusible, could have been used to build the corresponding models. This wheel-shaped amulet cannot result from casting in a permanent mould: this shape could not have been withdrawn without breaking the mould, as no plane intercepts jointly the equatorial symmetry planes of the support ring and of the spokes without inducing an undercut. The artefact was therefore cast using a lost-wax process (Supplementary Fig. 2).

A first campaign of measurements was performed 10 years ago but the wheel-shaped amulet could only be exhaustively described through novel advanced imaging. X-ray radiographs showed that it is corroded from its surface to its core. SEM examination of the equatorial section of the amulet corroborated the complete corrosion of the artefact, yet showed locally a fossilized dendritic structure, confirming a casting process. X-ray microanalyses on small areas highlighted Cu, O and Cl in the dendrites and Cu and O in the interdendritic space. Raman spectra allowed identifying the corrosion compounds: clinoatacamite Cu2(OH)3Cl in the dendrite and cuprous oxide Cu2O in the interdendritic space. However, full corrosion of the metal to cuprous oxide Cu2O precluded any further understanding of the manufacturing and metallurgical processes.

Macroscale imaging confirms casting in a single piece

Photoluminescence imaging shows the continuity of the spatial distribution and orientation of the remnant dendritic structure all across the equatorial section (Figs 1d and 2a,bSupplementary Fig. 3). This demonstrates that the artefact was cast in a single piece and does not consist of soldered parts (Supplementary Fig. 4). The lack of any crystal deformation shows that the object was made with very little, if any, subsequent work on the object, such as hammering. In addition, in the amulet three-dimensional morphology, no plane intercepts jointly the equatorial symmetry planes of the support ring and of the spokes without inducing an undercut. These observations therefore designate lost-wax casting as the procedure used for its fabrication. This is in agreement with the history of metallurgy in Baluchistan that shows evidence of an important development of lost-wax casting as demonstrated by finds such as the ‘Leopards Weight’, an extraordinary decorated ovoid ball of copper and lead weighing more than 15 kg dated end of the fourth millennium BC (ref. 7), and by the absence of any tradition of casting intricate shapes using piece-moulds as for instance reported in China8.

Figure 2: Fossil microstructure of the eutectic revealed in the 6,000-year-old Mehrgarh amulet.

Images reveal a typical eutectic morphology. The regular rod-like pattern is observed over millimetres in the interdendritic spaces. (a) Low magnification photoluminescence (PL) image of the wheel under 420–480 nm excitation and 850–1,020 nm bandpass emission (× 40 objective, NA=0.6). Scale bar, 500 μm. (b) Close-up view of the wheel (× 100 objective, projected pixel size: 155 nm, NA=1.25). den, dendrite; eu, rod-like eutectic in the interdendritic space. Scale bar, 100 μm. (c) Dark-field microscopy image of the same area of a. (d) Dark-field microscopy image of the same area of b. Note that the dendritic microstructure is more clearly evidenced in a than in c, and that the eutectic microstructure in b is not visible in d.

Full size image

Mesoscale imaging reveals atypical metallographic structure

Between corroded dendrites, hundreds of micrometres wide interdendritic spaces are observed in photoluminescence imaging. So-called ‘ghost’ dendritic structures are frequently observed in highly corroded ancient copper alloys9. On alloys, an interdendritic structure only occurs in the solidification of a two-phase system with alloying element such as Pb, As or Sn in ancient copper alloys. Extensive investigation by optical microscopy, scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS) and Raman spectroscopy reveals no alloying element at the 100 μm length scale: red cuprous oxide Cu2O is ubiquitous in the extended interdendritic spaces, while green clinoatacamite Cu2(OH)3Cl has formed in the corroded dendrites (Figs 2c,d and 3). The chemical composition of the interdendritic spaces is extremely homogeneous throughout the entire artefact (Fig. 3b–d, and Supplementary Fig. 4). Apart from copper and oxygen, only Ag and Fe are identified as traces with SEM-EDS (Supplementary Fig. 5).

Figure 3: Mapping of Cu2O species in interdendritic spaces.

(a) Image of dendrites and homogeneous interdendritic spaces (SEM-BEI, 10 kV). Scale bar, 300 μm. (b) RGB false colour image (SEM-EDS) of Cu (red), Cl (green) and O (blue) from the area denoted by a rectangle in a. Interdendritic spaces contain only Cu and O as major elements, while Cl is found in the corroded dendrites. Scale bar: 30 μm. (c,d) Identification of Cu2O in interdendritic spaces in the area denoted by a rectangle in b(c) Typical Raman spectrum from a Cu2O region. The spectrum was obtained by averaging 12 scans within the zone imaged in d (using four pixels in three separate areas). (d) RGB false-colour image of Raman vibrational bands characteristic of Cu2O: 632 (red), 416 (green) and 218 cm−1 (blue). Raman spectroscopy mapping does not show any variation in the characteristic vibrational features of Cu2O that would allow evidencing the rod-like eutectic structure. Scale bar, 4 μm.

Full size image

Microscale imaging reveals an invisible eutectic microstructure

The intense photoluminescence signal within the interdendritic spaces appears to result from the presence of an exceptionally well-fossilized microscopic pattern, invisible with the other methods used (SEM, EBSD, white light OM, Raman spectroscopy). The ∼1 μm lateral resolution allows the clear observation of a rod-like structure of high-yield luminescent Cu2O in the near infrared within a distinctly emitting Cu2O matrix (Fig. 2a,b). Such rod-like pattern, which has been preserved through corrosion, is a direct signature of a eutectic growth. The interdendritic spaces therefore correspond to eutectic areas that were initially composed of Cu0 with rod-like Cu2O, and result from the hypoeutectic solidification of the binary system Cu0–Cu2O in which initial Cu0 dendrites were formed. During long-term corrosion at ambient temperature, the original Cu0 has been oxidized to Cu2O, while the rod-like eutectic Cu2O phase has been preserved. These two distinct cuprous oxides Cu2O observed today are hereafter designated as co-Cu2O (corrosion) and eu-Cu2O (eutectic), respectively. Strikingly, this micrometric structure was completely preserved over centimetres during six millennia (Supplementary Fig. 3). Due to the aggressive role of chlorides in the archaeological soil, dendritic Cu0 was more affected by corrosion than eutectic Cu0 in contact with eu-Cu2O, inducing the progressive formation of Cu2(OH)3Cl in the dendrites11–13.

Pure Cu2O is a semiconductor whose spectroscopic properties are highly sensitive to intrinsic or extrinsic crystal defects14,15. Although uniquely consisting today of Cu2O (Fig. 3b–d), the different nature of atomic-scale crystal defects within eu-Cu2O and co-Cu2O of the interdendritic spaces allows visualization of the 6,000-year-old metallographic structure. The associated photoluminescence signal of the eu-Cu2O is dominated by emission in the near infrared from copper vacancies (VCu), while the excitonic emission near the band-edge transition at 2.1 eV is quenched16,17. The formation of eu-Cu2O at high temperature (the eutectic reaction occurs at 1,066 °C, Supplementary Note 1), must have led to the creation of a high density of stable VCu.

The oldest lost-wax cast

The ability to cover all length scales continuously from crystallite sizes to macroscopic sample dimensions allows deciphering invisible patterns that provided the key for a complete understanding of the manufacturing of the Mehrgarh artefact. From the visual inspection of the artefact, we show that the 20 mm wheel-shape model was prepared in a waxy material: the spokes were brought together by pressing each other at the wheel centre, and the base of each spoke was pressed on the support-ring (Fig. 4aSupplementary Fig. 2). Once made, the wax model was invested into a clay mould. The clay mould was heated upside down to run out the wax; baking was extended at higher temperature to harden the mould and drive out any moisture. Copper was poured in the mould, taking the place of the wax to cast the artefact in a single piece (Fig. 4b). The absence of any alloying element or significant impurity except low traces of Au, Hg and Ag in the amulet points to the use of a very pure copper, possibly native copper, that was melted in air above 1,085 °C. Had arsenic been present, as in most coeval cast alloys known so far18, the eutectic could not have formed, as oxidation of liquid copper is mitigated by the greater affinity of arsenic for oxygen19. The Cu0–Cu2O phase diagram can be exploited to trace the metallurgical sequence. During casting, the furnace atmosphere was inevitably oxidizing, and the copper melt absorbed ∼0.3 wt% of oxygen (∼1.1 at.%, Supplementary Fig. 6 and Supplementary Note 1), leading to the observed hypo-eutectic structure. The solidification of the dendrites started at about 1,070–1,074 °C (Fig. 4eSupplementary Fig. 6) while the eutectic formed at 1,066 °C (Fig. 4f). After cooling, the mould was broken and the casting was finished by cold working such as cutting the sprue and polishing (Fig. 4c,g). After burial, slow alteration took place in a sandy clayey soil and in a relatively dry environment (Fig. 4d,hSupplementary Fig. 7). The ghost fossilization of the metallographic structure took several centuries to complete in a comparatively dry environment—at typically about one micrometre per year20,21—leading to a final uniform presence of Cu2O within the eutectic.

Figure 4: Manufacturing of the amulet from Mehrgarh by the lost-wax casting process.

(a) The model was shaped by manufacturing small rods circular in section in a very ductile material that melts at low temperature, such as beeswax. Each wax piece was welded to the other by a slight heating of their extremities. (b) The wax model was invested by a clay mixture to form a mould. The mould was heated to run out the wax, and copper was poured in the mould, taking place of the wax. (c) The final copper artefact was extracted by breaking the mould after cooling. (d) Totally corroded artefact after its 6,000-year burial. (eh) Schematic representation of the solidification process and its evolution at a microscale: (e) 1,085 °C>T>1,066 °C. Dendritic growth of metallic copper (oxygen content in dendritic Cu<0.03% at). (f) Formation of the Cu-Cu2O eutectic at 1,066 °C. The liquid phase solidifies into Cu0 (0.03%at O) and a eu-Cu2O rod-like structure. (g) Final metallurgical structure of crystals of dendritic copper (low in oxygen) surrounded by oxygen saturated Cu0(Cu0.97O0.03) and rod-like Cu2O. (h) Current state of the artefact with the formation of the Cu2Cl(OH)3 phase within dendrites, while Cu0 fully oxidizes to co-Cu2O within the eutectic. eu-Cu2O is fully preserved.

Full size image


The discovery of the wheel-shaped amulets from Mehrgarh is an extraordinary evidence of the first attempts to manufacture precision casts by a lost-wax process. This innovation did not replace casting in permanent moulds but engendered a novel lineage of objects, whose complex shapes can only be obtained by this method. We can now state not only that metallurgists invented a totally new technique for casting, but also that control of the metal composition was part of their innovative research. By choosing a very pure copper rather than the usual arsenical copper22, they used a metal whose origin was probably considered to be of higher value and quality. The traces of mercury, silver and gold identified in the corroded amulet form a typical pattern for native copper23. The use of high-purity copper turned out to be a dead end: this did not improve the casting properties of the melt but caused unfamiliar problems to the founder: the melting point is not decreased, whereas the metal castability is severely reduced24. Although the lost-wax process proved to be an irrefutable and permanent success, selecting very pure copper for casting has not been retained as a valid innovation. Looking for improvements, Baluchistan founders soon discovered that the addition of a large proportion of lead to copper (Pb: 10–30 wt%) vastly increased the metal fluidity. During the fourth millennium BC and up to the end of the third millennium BC, this new Cu–Pb alloy was extensively used, and solely dedicated for lost-wax casting7,25. Lost-wax casting and Cu–Pb alloy were therefore widely adopted in the Ancient Near East, and used to manufacture artefacts of the highest symbolic and ceremonial significance. The use of Cu–Pb alloy was only challenged at the beginning of the second millennium BC, when Cu–Sn bronze became widely used within this geographic area owing to its improved metallurgical properties.

Mehrgarh is a crucible for technological innovation during Neolithic and Chalcolithic times in the ancient South Asia from lithics, pottery, ornaments, clay figurines, glazed materials as well as textiles and early practice of dentistry25,26,27. The emergence of the lost-wax technique at Mehrgarh could have been triggered by several factors. The availability of beeswax is attested in the Near East at this period28. Second, recent works have proposed that lost-wax casting has been adopted more for the central role of beeswax as a ritually important material than for a technical need29. It is also significant that the very first objects made by lost-wax casting did not fully exploit the potential of lost-wax casting. The amulet here in question is practically flat, and arguably a rather similar one could have been cast more easily using an open mould. The wax rods used to shape the metal amulet closely resemble the small clay coils used to model hundreds of clay figurines and amulets discovered in the Neolithic and Chalcolithic levels of Mehrgarh, and possibly associated with a magical and/or religious function. With lost-wax casting, it was now possible to produce these traditional adornment artefacts in metal, by simply working wax in place of clay, maintaining the long-established way in which they were modelled. The specific context at the site (resources, ritual, know-how) nurtured metallurgical invention, while other sites, possibly contemporaneous, such as Nahal Mishmar in the Levant that may have led to independent invention of lost-wax casting30 did not provide the incubating context allowing dissemination to the entire ancient Near East. Lost-wax casting tested for the first time with the Mehrgarh artefact is still the premier technique for art foundry. It is also today the highest precision metal forming technique—under the name ‘investment casting’—in aerospace, aeronautics and biomedicine, for high-performance alloys from steel to titanium31. Today, rapid prototyping technique such as three-dimensional printing offers revolutionary capabilities to design plastic, polymer or wax models used in investment casting32,33. New templating approaches for nanocasting semiconductor structures are among the latest evidence of the fundamental character of the lost-wax concept34,35.

We demonstrate the potential of gigapixel photoluminescence imaging to study the response of materials at micrometric resolution over centimetre-size fields within desired spectral bands. The exploration of the spatial distribution of the electronic density of state within polycrystalline semiconductor materials is then possible. The proposed approach goes far beyond collection of point or average luminescence signal of great complexity, towards determination of the representative elementary areas in which the measured photoluminescence response in a heterogeneous matrix becomes continuous quantities. Here, high-definition images of crystal defect contrasts provide a direct probe of stoichiometry fluctuations, which in turn record information on the materials’ manufacturing process. This approach can conversely prove to be extremely effective in optimizing the synthesis route of systems that are far less expected to be heterogeneous, such as batches of semiconductor nano-structures. We have therefore extended our proof of concept to a modern synthetic material by mapping and characterization of crystal defects density within a batch of nanowires. High signal-to-noise ratio images of zinc oxide nanowires of 0.5–1 μm in diameter and 14 μm in length deposited on a substrate were collected in nine spectral bands ranging from the deep ultraviolet to the near infrared using an excitation wavelength of 275 nm. The images reveal both unexpected spectral-dependent spatially variable emission from crystal defects along the length of individual nanowires and the statistical variability of the distribution of those defects within the entire population where a limited number of typical nanowire behaviours is observed (Fig. 5). Deep ultraviolet-optimized multispectral collection strategy allows ‘à la carte’ adaptation of integration times to each spectral emission range, to collect extremely low-yield responses that would otherwise go undetected through hyperspectral data collection. The ability to collect emission from single grains or crystallites to centimetres of samples at room temperature with tuneable source over the whole deep ultraviolet to near infrared range therefore provides unprecedented capability to image the intrinsic complexity of heterogeneous materials from nanosciences, engineering, geophysics, archaeology and environmental sciences.

Figure 5: Spatial distribution of crystal defect and band edge emission of ZnO nanowires.

Full-field photoluminescence image of a batch of ZnO nanowires (ultraviolet excitation: 275 nm, 4.50 eV). False colour overlays of signal in the 850–1,020 nm (red), 499–529 nm (green) and 370–410 nm (blue) bands. The image is corrected in each channel from collection time, quantum efficiency of the CCD camera, transmission of emission filters and theoretical point spread function of the objective. Scale bar, 10 μm.

Full size image


Photoluminescence imaging

Photoluminescence micro-imaging was performed on a full-field inverted microscope (Axio Observer Z1 microscope, Zeiss) at the DISCO beamline (SOLEIL synchrotron)36. The microscope is equipped with custom quartz lenses instead of the original glass ones, to ensure transmission of excitation and emission above 80% and allow collecting luminescence images down to 200 nm. The beamline exploits the tunability of the bending magnet source, with an energy bandwidth ΔE/E of 2 × 10−2 at 275 nm (100 grooves per millimetre grating, iHR320 monochromator, Jobin-Yvon, Longjumeau, France).

In the frame of this work, specific developments were implemented to optimize excitation tunability, high-throughput detection and spatial dynamics required to detect and spatially resolve the multi-scale luminescence pattern in the amulet (Supplementary Fig. 8a,b in comparison with Supplementary Fig. 8c–h). Two sources were coupled to attain the respective excitation ranges 220–400 nm (synchrotron radiation source)36,37 and above 400 nm (halogen lamp coupled to an interference bandpass filter). In the deep ultraviolet (synchrotron) range, energies greater than 1.2 eV are blocked using a cold finger of thickness 7.5 mm that intercepts a vertical angle of 1.5 mrad in the middle of the beam. As a result, the spatial distribution of the beam at the exit of the monochromator is composed of two longitudinal sheets. To obtain a homogeneous field of illumination down to the deep ultraviolet, an optical set-up using micro-array lenses and a rotating diffuser was developed and positioned ahead of the microscope.

High-grade optical elements were used all along the optical path to minimize all optical distortions, particularly field and chromatic aberrations, and allow image stitching. A × 40 / NA 0.6 and × 100 / NA 1.25 glycerine Zeiss ultrafluar apochromatic immersion objectives were used to excite and collect images from ultraviolet-C to near infrared ranges. High spatial dynamic images were gathered by collecting mosaics of tiles with an XY motorized stage (PI) allowing to image areas of hundreds of micron side. For instance, Fig. 2a is made of overlapping tiles, each of 1.4 × 104 μm2 (774 × 759 pixels), in a 414 images matrix that creates a final 4.0 mm2 image (14,888 × 11,415 pixels). The projected pixel size of 155 nm is 2.4 times smaller than the theoretical diffraction limit of 374 nm (=935 nm/2/1.25) at 935 nm. Measurement of the optical point spread function across an ∼400 nm CdS particle shows that spatial resolution is ∼1 μm (Supplementary Fig. 9). During the optimization procedure of our set-up, the experiment was replicated four times on the amulet. For each measurement, the eutectic pattern could clearly be visualized in the images collected in the near infrared (Supplementary Fig. 10). In addition, all the tiles collected showed a similar reproducible pattern.

High-throughput spectral detection from UVC up to near infrared was achieved by using a multi-spectral detection using high-transmission interferential bandpass filters positioned in front of a back-illuminated 1,024 × 1,024 pixels CCD (PIXIS:1024BUV, Princeton Instrument 13 × 13 μm2 pixel size)38. The images shown in this work were collected using 370–410, 499–529 and 850–1,020 nm interference bandpass filters (transmission >90%). The collection time is adjusted for each set of excitation/emission conditions to optimize the signal-to-noise ratio (up to a few minutes per tile).

Optical microscopy

Dark-field microscopy was performed using a Zeiss Axio Imager M2m microscope coupled to an AxioCam ICc5 camera, with × 5 and × 20 objectives (C2RMF). The images collected on an XY motorized stage were mosaicked to cover a large field of view.

Raman spectroscopy

Raman spectroscopy was performed at an excitation of 532 nm and on-sample power of 2 mW with a × 100 objective (SOLEIL, SMIS). The spectra were collected using an integration time of 2 s, accumulation of two spectra per point and a 25 μm spectrograph aperture slit.

Scanning electron microscopy

SEM and EDS were performed on a Zeiss Supra 55 VP coupled to a Bruker EDS system (Quantax 800, 30 mm2 silicon drift detector (SDD); IPANEMA).

Electron backscatter diffraction

EBSD was conducted on a JSM 7100F apparatus equipped with an Oxford AztecHKL and NordlysNano with 4 FSD detector (Centre de Microcaractérisation Raimond Castaing, Toulouse, France). For this analysis, the surface was prepared using vibratory polishing (Buehler VibroMet 2, ChemoMET polishing cloth) with 50 nm colloid alumina suspension. A carbon coating a few nanometres was applied (Leica EMACE600). The experiments were performed at 20 kV (70° tilt) and data were processed using the Channel 5 Tango software.

Sample preparation

The wheel-shaped amulet inventory number MR. was collected in 1985 at the MR2 site of Mehrgarh during the excavations of the ‘Mission Archéologique de l’Indus’ (dir. Jean-François Jarrige) in collaboration with the Department of Archaeology and Museums of Pakistan. A section was prepared in the equatorial plane, embedded in epoxy resin (Epofix, Struers) and polished with diamond pastes up to 0.25 μm grain size (C2RMF).

Preparation of the ZnO nanowires

ZnO nanowires were grown at 850 °C by metal–organic chemical vapour deposition (MOCVD) on a (0001) sapphire substrate using diethylzinc and nitrous oxide as zinc and oxygen precursors (GEMaC, Versailles, France).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

How to cite this article: Thoury, M. et al. High spatial dynamics-photoluminescence imaging reveals the metallurgy of the earliest lost-wax cast object. Nat. Commun. 7, 13356 doi: 10.1038/ncomms13356 (2016).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


Gilmore, A. M. Luminescence: The Instrumental Key to the Future of Nanotechnology Pan Stanford Publishing (2013).

  • + Show context

Negrin, R. S. & Contag, C. H. In vivo imaging using bioluminescence: a tool for probing graft-versus-host disease. Nat. Rev. Immunol 6, 484–490 (2006).

Fernández-Suárez, M. & Ting, A. Y. Fluorescent probes for super-resolution imaging in living cells. Nat. Rev. Mol. Cell Biol. 9, 929–943 (2008).

Marks, K. M. & Nolan, G. P. Chemical labeling strategies for cell biology. Nat. Methods 3, 591–596 (2006).

Crochet, J. J., Duque, J. G., Werner, J. H. & Doorn, S. K.Photoluminescence imaging of electronic-impurity-induced exciton quenching in single-walled carbon nanotubes. Nat. Nanotechnol 7, 126–132 (2012).

Blasse, G. in Solid State Luminescence (ed. Kitai A. H. 21–53Springer (2012).

  • + Show context

Mille, B., Bourgarit, D. & Besenval, R. in South Asian Archaeology 2001 (eds Jarrige C., Lefèvre V. 237–244Éditions Recherches sur les Civilisations (2005).

  • + Show context

Zhou, W. & Huang, W. Lost-wax casting in ancient China: new discussion on old debates. J. Min. Met. Mat. Soc 67, 1629–1636 (2015).

  • + Show context

Chase, W. T. in Ancient and Historic Metals: Conservation and Scientific Research (eds Scott D. A., Podany J., Considine B. B. 85–118Getty Publications (1994).

  • + Show context

Stefanescu, D. M. & Ruxanda, R. in Metallography and Microstructures, ASM Handbook Vol. 9 (ed. Vander Voort G. F. ASM International (2004).

  • + Show context

Pourbaix, M. Lectures on Electrochemical Corrosion Plenum (1973).

  • + Show context

Lucey, V. Developments leading to the present understanding of the mechanism of pitting corrosion of copper. Br. Corros. J. 7, 36–41 (1972).

  • + Show context

Scott, D. A. Copper and Bronze in Art: Corrosion, Colorants, Conservation Getty Publications (2002).

  • + Show context

McShane, C. M. & Choi, K.-S. Photocurrent enhancement of n-type Cu2O electrodes achieved by controlling dendritic branching growth. J. Am. Chem. Soc. 131, 2561–2569 (2009).

Pelant, I. & Valenta, J. Luminescence Spectroscopy of Semiconductors Oxford Univ. Press (2012).

  • + Show context

Li, J. et al. Probing defects in nitrogen-doped Cu2O. Sci. Rep 4, 7240 (2014).

Solache-Carranco, H. et al. Photoluminescence and X-ray diffraction studies on Cu2O. J. Lumin. 129, 1483–1487 (2009).

  • + Show context

Thornton, C. P., Lamberg-Karlovsky, C., Liezers, M. & Young, S. M.On pins and needles: tracing the evolution of copper-base alloying at Tepe Yahya, Iran, via ICP-MS analysis of common-place items. J. Archaeol. Sci. 29, 1451–1460 (2002).

  • + Show context

Azakami, T. & Yazawa, A. Activity measurements of liquid copper binary alloys. Can. Metall. Quart. 15, 111–122 (1976).

  • + Show context

Shreir, L. L., Jarman, R. A. & Burstein, G. T. Corrosion. Vol. 1, Metal / Environment reactions Newnes-Butterworths (1976).

  • + Show context

Robbiola, L. & Portier, R. A global approach to the authentication of ancient bronzes based on the characterization of the alloy–patina–environment system. J. Cult. Herit. 7, 1–12 (2006).

  • + Show context

Weeks, L. in A Companion to the Archaeology of the Ancient Near East (ed. Potts D. T. 295–316John Wiley and Sons (2012).

  • + Show context

Pernicka, E. et al. Prehistoric copper in Bulgaria: its composition and provenance. Eurasia Antiqua 3, 41–180 (1997).

  • + Show context

Neff, D.-V. in Casting, ASM Handbook Vol. 15 ed. Stefanescu, D. M.964–1084ASM International (1988).

  • + Show context

Jarrige, J.-F. in Villes, villages, campagnes de l’Âge du Bronze (ed. Guilaine, J.) 9–22 (Errance, 2008).

  • + Show context

Coppa, A. et al. Palaeontology: early neolithic tradition of dentistry. Nature 440, 755–756 (2006).

Moulherat, C., Tengberg, M., Haquet, J.-F. & Mille, B. First evidence of cotton at Neolithic Mehrgarh, Pakistan: analysis of mineralized fibres from a copper bead. J. Archaeol. Sci. 29, 1393–1401 (2002).

  • + Show context

Roffet-Salque, M. et al. Widespread exploitation of the honeybee by early Neolithic farmers. Nature 527, 226–231 (2015).

  • + Show context

Martinón-Torres, M. & Uribe-Villegas, M. A. Technology and culture in the invention of lost-wax casting in South America: an archaeometric and ethnoarchaeological perspective. Camb. Archaeol. J 25, 377–390 (2015).

  • + Show context

Gilead, I. & Gošić, M. Fifty years later: a critical review of the stratigraphy, chronology and context of the Nahal Mishmar Hoard. J. Israel Prehist. Soc 44, 226–239 (2014).

  • + Show context

Backman, D. G. & Williams, J. C. Advanced materials for aircraft engine applications. Science 255, 1082–1087 (1992).

  • + Show context

Cheah, C. M., Chua, C. K., Lee, C. W., Feng, C. & Totong, K. Rapid prototyping and tooling techniques: a review of applications for rapid investment casting. Int. J. Adv. Manuf. Technol 25, 308–320 (2005).

  • + Show context

Ryan, G. E., Pandit, A. S. & Apatsidis, D. P. Porous titanium scaffolds fabricated using a rapid prototyping and powder metallurgy technique. Biomaterials 29, 3625–3635 (2008).

Lou, X. W., Archer, L. A. & Yang, Z. Hollow micro-/nanostructures: synthesis and applications. Adv. Mater 20, 3987–4019 (2008).

Goldberger, J., Fan, R. & Yang, P. Inorganic nanotubes: a novel platform for nanofluidics. Acc. Chem. Res. 39, 239–248 (2006).

Giuliani, A. et al. DISCO: a low-energy multipurpose beamline at Synchrotron SOLEIL. J. Synchrotron Radiat 16, 835–841 (2009).

  • + Show context

Jamme, F. et al. Synchrotron UV fluorescence microscopy uncovers new probes in cells and tissues. Microsc. Microanal. 16, 507–514 (2010).

Bertrand, L., Réfrégiers, M., Berrie, B., Echard, J.-P. & Thoury, M. A multiscalar photoluminescence approach to discriminate among semiconducting historical zinc white pigments. Analyst 138, 4463–4469 (2013).

  • + Show context

Download references


This article is dedicated to the memory of Jean-François Jarrige (1940–2014), former director of the musée Guimet in Paris, who discovered Mehrgarh in 1974 and directed the ‘Mission Archéologique de l’Indus’ from 1975 to 2014. Claudie Josse is warmly acknowledged for providing EBSD results (Centre de microcaractérisation Raimond Castaing, CNRS UMS 3623, Toulouse, France). We acknowledge SOLEIL for provision of synchrotron radiation under projects no 20120848 and 20130920. We thank Christophe Sandt at the SMIS beamline for access to Raman microscopy (SOLEIL synchrotron), Pierre Gueriau for complementary synchrotron XRF mapping (IPANEMA) and Frédéric Jamme (SOLEIL synchrotron) for providing support to generate the point spread function (PSF). We thank Pierre Galtier, Alain Lusson and Vincent Sallet (GEMaC UMR8635) for preparing and providing the ZnO nanowires. We thank Sebastian Schoeder (synchrotron SOLEIL) for the representation of the amulet in three dimensions. We especially thank Catherine Jarrige, Gonzague Quivron, Aurore Didier and Jérôme Haquet who provided complementary information about the metal artefacts from Mehrgarh. We thank Barbara Berrie, Catherine Perlès, Denis Gratias and Uwe Bergmann for critical re-reading of the manuscript.

Author information

Author notes

    • J-F Jarrige



  1. IPANEMA, CNRS, ministère de la Culture et de la Communication, Université de Versailles Saint-Quentin-en-Yvelines, USR 3461, Université Paris-Saclay, 91128 Gif-sur-Yvette, France
    • Thoury
    • , T. Séverin-Fabiani
    • & L. Bertrand
  2. Synchrotron SOLEIL, 91128 Gif-sur-Yvette, France
    • Thoury
    • , T. Séverin-Fabiani
    • , M. Réfrégiers
    • & L. Bertrand
  3. C2RMF, Palais du Louvre, 75001 Paris, France
    • Mille
  4. PréTech, CNRS, Université Paris Nanterre, UMR 7055, 92023 Nanterre, France
    • Mille
  5. TRACES, CNRS, ministère de la Culture et de la Communication, Université Toulouse—Jean Jaurès, UMR 5608, 31100 Toulouse, France
    • Robbiola
  6. ArScAn, CNRS, Université Paris Nanterre, Université Paris 1, ministère de la Culture et de la Communication, UMR 7041, 92023 Nanterre, France
    • J-F Jarrige
  7. Institut de France, 23 quai de Conti, 75006 Paris, France
    • J-F Jarrige


M.T. and L.B. designed the experiments. L.B., M.T. and T.S.-F. coordinated and drafted the manuscript. T.S.-F., M.T., L.B., B.M. and L.R. wrote the manuscript and prepared the figures. B.M. selected the artefact and provided the archaeometallurgical interpretation. L.R. provided the corrosion interpretation. The experiments and data analysis were performed at the DISCO beamline at synchrotron SOLEIL (M.T., T.S.-F., L.B., M.R.), SEM-EDS (L.R., B.M.), Raman (M.T., L.R.), EBSD (L.R.) and OM (B.M., T.S.-F.). J.-F.J. provided the archaeological information.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to M. Thoury. Supplementary information

PDF files

  1. Supplementary Information

Supplementary Figures 1-10, Supplementary Note 1, Supplementary Methods and Supplementary References2.Peer Reew File

  1. Kalyanaraman

Sarasvati Research Center

November 16, 2016


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s