Introduction

The development of the radiocarbon method has had a profound impact on archeology as it enables the dating of the remains of ancient civilizations. The method is usually applied to organic materials - wood, charcoal, bone, etc. - which incorporate radioactive carbon through photosynthesis or ingestion. However, a few studies have shown that radiocarbon dating can also reveal the age of inorganic manufactured materials. As early as the 1960s, 14C dating was extended to building materials1. In particular, mortar which incorporates atmospheric carbon dioxide as it hardens, has been extensively studied and used to date churches and Roman buildings2, 3.

More recently, the direct dating of iron was successfully achieved by combining chemical analysis and radiocarbon measurement. During the manufacture of the iron alloy, carbon from charcoal combustion is incorporated by diffusion into the metal and forms cementite (Fe3C)4. This approach has shed new light on the technical skill and intentions of cathedral builders in Europe5 and on the chronology of the Angkor temples in Cambodia6.

Lead carbonates were used as ingredients in paint and cosmetics. Lead white, a mixture of cerussite (PbCO3) and hydrocerussite (Pb3(CO3)2(OH)2), was the most important of all white pigments until the nineteenth century7. Lead carbonates were also employed as cosmetics from Antiquity to the eighteenth century8, 9. They were used by ancient Egyptians, Greeks, and Romans10,11,12,13,14,15. Understanding the synthetic process of cosmetics production in Antiquity is of great interest for human history and for our knowledge of the development of chemical techniques.

Here we demonstrate the ability of radiocarbon to date archeological and historical lead carbonates and to discriminate natural and synthesized cosmetics. Based on a purposely designed protocol (Methods), radiocarbon measurements are performed on ancient Egyptian and Greek make-up held at the Louvre museum. As they are considered as the earliest synthesized cosmetics, the application of the radiocarbon method is highly meaningful not only to date the compound but also to determine its natural or artificial origin.

We measure the 14C/12C ratio by accelerator mass spectrometry (AMS) in each cosmetic powder prepared by thermal decomposition and graphitization16, 17 (Methods). We show that atmospheric carbon dioxide was incorporated in some lead carbonates during their synthesis. In this way, the radioactive isotope 14C was fixed in the mineral matrix through the carbonate ion and decayed in time after its incorporation, making radiocarbon dating possible. To our knowledge, this study presents the first absolute dating of lead carbonate. We provide a tool to directly date lead carbonate-based materials such as ancient cosmetics and paintings. Recrystallization of lead carbonate in cosmetics or paintings is unlikely to be problematic for radiocarbon dating as for other carbonates such as shells and some mortar due to lack of exposure to water.

We also show that the quantity of radiocarbon (14C) is the criterion of choice to discriminate natural and synthesized lead carbonates. The results give new insight into the production of white make-up powders in Antiquity and confirm the expertize of ancient Egyptians and Greeks in chemically synthesizing lead carbonate make-up.

Results

Ancient Egyptian and Greek white cosmetics from the Louvre museum

To highlight the novelty of our approach we report here results obtained on well-preserved Egyptian and Greek ancient cosmetics. They are held at the Louvre museum in their original containers in wood, reed or alabaster (Fig. 1). The objects were associated to identified Egyptian or Greek tombs and dated from the 3rd millennium to the 3rd century BC (for more information on the samples, see Methods and Supplementary Table 1). The cosmetics were selected on the basis of their known archeological context and chemical composition. Small samples of powders (10–50 mg) containing two forms of lead carbonate were collected: phosgenite (Pb2Cl2CO3) and cerussite. These two compounds were pure or mixed together or associated to other lead minerals: galena (PbS), laurionite (PbCl(OH)), and anglesite (PbSO4). Galena is an abundant natural mineral, widely distributed in various environments. On the contrary, laurionite and phosgenite are very rare in nature and are considered as the earliest synthesized cosmetics15. Cerussite exists as a natural mineral, but has also been manufactured. Natural cerussite is found in the oxidation zone of lead deposits. As an artificial compound, the production of “psimythium” (the Greek term for cerussite) or “cerussa” (the Latin term) is described in several written sources from Antiquity, from the 4th century BC to the 1st century AD18,19,20. Two modern samples of cerussite made according to the historical process were also measured to further validate our methodology (Supplementary Table 5).

Fig. 1
figure 1

Egyptian and Greek make-up from the Louvre museum collection. (a) in a wooden cylinder (E 22326, 10 cm high and 4.4 cm diameter), one of the four channels is visible, (b) in an alabaster vase (E 23092, 4.9 cm high, 5.1 cm diameter) and, (c) in a 2.5 diameter wooden box (AGER-CA 508) © Musée du Louvre, LMC14, C2RMF

Full size image

Greek Cerussite and Egyptian phosgenite were synthesized in Antiquity

The radiocarbon dating of five samples of powder provided dates coherent with the expected age of the cosmetics (Table 1 and Supplementary Table 2). One sample is a pale pink blush powder (AGER-CA508, Fig. 1c) composed of cerussite. It was preserved in a small box found in a tomb discovered in Eretria (Greece). This tomb was dated on a numismatic basis from ca 330 to ca 266 BC21. The other four powders are mainly composed of phosgenite and were placed in four distinct channels of a wooden cylinder (E 22326, Fig. 1a) engraved with the cartouches of Amenhotep III and his wife Queen Tiye. The dates of Amenhotep III’s accession and the end of his reign are estimated to be 1423–1386 calBC and 1385–1348 calBC, respectively22, 23.

Table 1 Radiocarbon dates of ancient cosmetics made with cerussite or phosgenite
Full size table

For the Greek cerussite sample (AGER-CA508) we obtained a date range of 353 to 57 BC (95.4%) (Fig. 2a). This date is consistent with the numismatic result and in agreement with two previous radiocarbon dates measured on an almond seed (AGER-CA518) from the same tomb21 (Fig. 2b).

Fig. 2
figure 2

Radiocarbon dating of a Greek cerussite make-up powder. (a) The calibrated radiocarbon date (353–57 calBC) of the Greek cerussite powder (AGER-CA508 - SacA48856) found in the tomb of a child in Eretria (Greece) is compared to two previous radiocarbon dates (359–112 calBC and 375–203 calBC) measured on an almond seed (AGER-CA 518 - 1A214/SacA22810 and 2A214/Sac22811)40 of the same tomb21. Calibrated dates, represented by age probability distributions in gray, are calculated using the Intcal13 atmospheric calibration indicated in blue (see Methods), (b) The statistical combination of the three dates is represented in red before calibration (2175 ± 18 BP) and in gray after calibration (356–171 calBC, see details in the key); A χ2 test value of T = 3.7( < 6) shows the consistency of the dates

Full size image

The radiocarbon dates of the four phosgenite Egyptian powders (E 22326) are between 1763 and 1216 calBC (Fig. 3). The powders are preserved in four distinct channels of the container and do not have exactly the same composition24. The relative scatter on the dating results may suggest different production periods. If somewhat large, the range nevertheless includes the radiocarbon date of the container (1514–1412 calBC, see Table 1) and the dates ascribed to Amenhotep III. However, we have to keep in mind that the date obtained on the wood might not accurately reflect the period of use due to the inbuilt effect as well as a possible material reuse due to the scarcity of wood in Ancient Egypt25. The container may have been carved from reclaimed wood that had previously been used for other purposes.

Fig. 3
figure 3

Calibrated radiocarbon dates of Egyptian make-up. Four phosgenite powders and their wooden container are dated from 1763 to 1216 calBC. This large range can be due to the fact that the four powder samples may not have been manufactured from the same batch since they are preserved in four distinct channels of the container. As the cartouches of Amenhotep III and his wife Queen Tiye are engraved in the wood, the dates ascribed to the Amenhotep III’s reign (accession dates from 1423 to 1386 and end of reign from 1385 to 1348 calBC22, 23) are indicated by shaded region

Full size image

The fact that the measured radiocarbon dates of these cosmetics match with the expected dates is strong evidence that atmospheric carbon dioxide was incorporated at the time of formation of the lead carbonates and that the carbonate formation occurred in a period close to that of the use of the cosmetics. Through the absorption of CO2, the radioactive isotope 14C was fixed in the carbonate ion, and decayed over time. The present radiocarbon content of the phosgenite samples ranges from 65.7 ± 0.3 to 68.5 ± 0.2 pMC (percentage of modern carbon). The mean value of the four samples (66.8 ± 1.2 pMC) is consistent with the radiocarbon content of the wooden container (67.21 ± 0.26 pMC).

The radiocarbon measurements thus provide two key results: the age and the signature of the artificial production of the cosmetics.

Ancient Egyptian cerussite is a natural mineral

The radiocarbon dates of seven Egyptian powders composed of mixtures of galena, laurionite, anglesite, cerussite, and phosgenite are much older than the period considered (Table 2 and Supplementary Table 3). As galena, laurionite, and anglesite do not contain carbon, the radiocarbon measurements concern phosgenite and cerussite exclusively. The two samples containing galena and cerussite give dates from 8793 to 7142 calBC. As they are much older than the expected age (3000–1000 BC), these results evidence a natural origin of the Egyptian cerussite. Cerussite is abundant in nature as a secondary mineral formed by the natural weathering of galena exposed to air or water26. The measured radiocarbon dates thus correspond to its in situ formation about 10 kyears ago and indicate that these cosmetics are composed of a natural association of galena and cerussite.

Table 2 Radiocarbon dates of Egyptian cosmetics composed of various lead compounds
Full size table

The radiocarbon dates of the five samples containing galena, cerussite and phosgenite are also older than expected, but to a significantly lesser extent (4700–3300 calBC). These intermediate ages are coherent with a combination of the old ages of the natural cerussite and the “pharaonic ages” of the synthesized phosgenite. We interpret this result to mean an intentional combination of natural cerussite associated to galena with synthetic phosgenite.

Discussion

Radiocarbon measurements on ancient cosmetics provide two key results: the production process of the lead carbonate powders and, when synthesized, the date when they were produced.

Our results show two distinctive cases: we find much older radiocarbon dates than the period of use for the Egyptian cosmetics composed of mixtures of lead compounds, and radiocarbon dates in agreement with the expected dates for Greek and Egyptian make-up in pure cerussite or pure phosgenite.

In the first case, for the cosmetics composed of a mixture of cerussite and galena, radiocarbon dates give ages that are too old. This result is the fingerprint of a natural cerussite since geological carbonates contain very low amounts of 14C. Cerussite is a common alteration product of galena, associated to secondary minerals and other carbonates in oxidized zones of deposits. The weathering of galena exposed to air and water involves the formation of anglesite. In presence of calcareous soil or carbonated water, cerussite is also found. Galena deposits in Egypt occur mainly along the Red Sea coast27. The site of Gebel el-Zeit was one of the main mines exploited in Antiquity28. The mines are located on the calcareous slope of the wadi and in the superficial zones, galena is partially transformed into cerussite29. The mine of Gebel el-Zeit was exploited in the second millennium BC, from the 12th dynasty up to the end of the New Empire (ca 2000–1200 BC)30. Minerals containing galena and secondary minerals were collected at this period and used as cosmetics without chemical transformation. The color, from black to light gray, was probably adjusted by selecting various mineral blends, depending on the proportion of black galena and grayish natural cerussite and anglesite. White powder was thus difficult to obtain and it became necessary to search for pure white compounds.

Phosgenite is a solid white powder. Its presence is observed in the cosmetics as pure or mixed with galena and cerussite. Very few occurrences of natural phosgenite are reported. Contrary to lead carbonates and lead sulfates, phosgenite is, like all minerals of lead and chlorine, soluble and does not accumulate in the oxidized zones of lead deposits31. To obtain large quantities of pale powder, the production of an artificial white was thus required. The first evidence that phosgenite was chemically synthesized by ancient Egyptians was reported twenty years ago15. This statement was based on the small abundance of these minerals in nature, on ancient recipes and on their replication in modern laboratories. Here, by using radiocarbon, we provide an absolute method to directly identify a synthetic process for phosgenite. Dioscorides in the second half of the first century AD reported a recipe involving lead oxide, salt, and white soda to produce a white compound19, 32 (the corresponding chemical compounds are indicated in brackets):

“[…], mix with 80 Attic drachmai [1 drachme = 3.4 g] of litharge (PbO), one litra of mined salt (NaCl) and pound. Letting some time pass, add 47 holcai [1 holce = 3.4 g] of very white soda (also translated as natron) (Na2CO3) dissolved in water and again pound it until the litharge becomes quite white […]. That which was washed seems to be good for eye medications, for unseemly scars and for faces full of wrinkles and blemishes”.

Successful replication of this recipe showed that phosgenite is obtained after many repetitive operations15, 33. The first step in the recipe is to form laurionite by stirring PbO (litharge) with NaCl (salt): PbO + NaCl + H2O → Pb(OH)Cl + NaOH. When carbon dioxide is added in the solution, the chemistry of lead oxychloride is altered and phosgenite forms34: 2Pb(OH)Cl(s) + CO2(g) → Pb2Cl2CO3(s) + H2O. The CO2(g) provides the radioactive isotope 14C used for dating the cosmetics. According to our results, the carbon dioxide used to manufacture phosgenite is of atmospheric origin. In the recipe, CO2(g) is produced by the dissolution of natron in water. The term of natron is vague and can refer to many phases of sodium carbonates (Na2CO3 and its hydrated forms or NaHCO3 (bicarbonate))35. Natural sodium carbonates from the Wadi El Natrum depression were known by the ancient Egyptians34. They are formed during the drying up of salt lakes in summer. Sodium carbonates can be formed at the surface by absorption of carbon dioxide from the atmosphere or in the brine by combination of HCO32− with Na35. Over the years the lake builds up a stratigraphy of deposits which are complicated by partial dissolution during the winter when the water rises again. The resulting radiocarbon content of such materials is therefore likely to be complex. However, due to the seasonality of the phenomenon, natron may have been collected annually, limiting in time the incorporation of radiocarbon. The other way of incorporating 14C in phosgenite is during the synthesis as NaOH produced by the chemical reaction is prone to absorb atmospheric CO2. The precise source of CO2 is still an open question, but we can assert that 14C was fixed in the carbonate of the phosgenite through the incorporation of atmospheric carbon dioxide. Furthermore as the quantity of 14C is very close to that of the contemporaneous atmosphere measured in the wood of the container, we demonstrate that this phenomenon took place at the same period. The absorption of CO2 in sufficient quantity to produce grams of powders reveals an intentional manufacture. We thus confirm that the Egyptian phosgenite was chemically synthesized and that this synthesis was achieved between 1760 and 1200 calBC. Through this innovative process, artificial phosgenite powder was “mass-produced” and used pure to obtain white and cream colors on the Egyptian make-up palette. Artificial phosgenite was also blended with natural galena and cerussite to lighten the color. In this way, Egyptians of the 18th dynasty not only took advantage of natural resources of lead ore for their cosmetics industry but were also capable of developing targeted chemistry.

More than one millennium later, in Greece, another situation emerged. To satisfy the demand for white cosmetics, the Ancient Greeks artificially produced cerussite, as we now demonstrate. A recipe for “psimythium” is reported by Theophrastus, born around 372 B.C. at Eresos on the island of Lesbos18:

“Lead (Pb) about the size of a brick is placed in jars over vinegar (CH3COOH), and when this acquires a thick mass, which it generally does in ten days, then the jars are opened and a kind of mold is scraped off the lead, and this is done again until it is all used up. The part that is scraped off is ground in a mortar and decanted frequently, and what is finally left at the bottom is psimythium (PbCO3 and Pb3(CO3)2(OH)2)”.

This preparation was also reported by Dioscorides19 and Pliny the Elder20 in the 1st century AD. In this process, lead is exposed to acetic acid vapors (“placed over vinegar”) and corrodes. Corrosion products consisting of white flakes (“a kind of mold”) form at the surface and are collected. Vinegar vapor converts the lead to white lead acetate36. CO2 is required at this stage to produce plumbonacrite (PbCO3·3Pb(OH)2·PbO), hydrocerussite and then cerussite. A supply of CO2 is not mentioned in the recipe, but our results indicate that carbon dioxide of atmospheric or organic origin was the main reactant to produce the carbonate. CO2 probably came from organic impurities in the vinegar or could be supplied by the fermentation of additional, but not reported, organic matter. Later recipes mention dung or horse manure to produce CO237, 38. As for the artificial phosgenite, 14C was fixed in the carbonate and decayed over time. Its measurement provides direct evidence that the Greeks synthesized artificial cerussite to produce white make-up in the 4th century BC. Despite its toxicity, psimythium was preferred to other white pigments such as chalk, kaolin, or plaster because of its excellent concealing effects39.

We have demonstrated, using radiocarbon measurement, that natural and artificial cerussite can be unequivocally differentiated. This procedure can be extended to all other carbonates, as we also show for phosgenite. We confirm that ancient Egyptians and Greeks were capable of chemically synthesizing lead carbonate make-up and we provide a time scale showing that phosgenite was artificially produced one millennium before cerussite. During the Egyptian Kingdom cerussite was a natural mineral extracted from the galena mines.

In conclusion, we provide an absolute method to directly date lead carbonates. We extend the range of inorganic materials datable by the radiocarbon method. We anticipate that our methodology will be applied to lead white, which is the most important pigment in art history, paving the way for a new tool for the authentication of ancient paintings.

Methods

Selection and collection of the samples

We selected samples of cosmetics according to their archeological context and chemical composition (Supplementary Table 1). Powder of 20–50 mg were taken directly from their wood, reed or alabaster containers held in the Louvre museum, Paris, France (Département des Antiquités égyptiennes and Département des Antiquités grecques, étrusques et romaines). In order to avoid any contamination, we used sterile tools and receptacles. The selected objects are associated to identified Egyptian or Greek tombs and some containers are decorated with clear inscriptions. The wood cylinder of the white phosgenite make-up (E 22326) is engraved with the cartouches of Amenhotep III and his wife Queen Tiye. Two reed cases (E 11048c, E 11048e) were found during the excavations of Lady Touti’s tomb at Medinet el-Gorab (Fayoum). Her death dates back to the 18th dynasty, just before or during Tutankhamun’s reign. The small pyxide containing a pale pink powder (AGER-CA508) comes from a tomb discovered in Eretria (Greece) dated by coin identification from ca 330 to ca 266 BC21. An almond seed from the same tomb was previously radiocarbon dated between 360 and 196 calBC40. We collected samples of powders which were chemically analyzed in previous studies by scanning electron microscopy and quantitative X-ray diffraction12,13,14,15, 24. Pure phosgenite (E 22326a/b/c/d), pure cerussite (AGER-CA508), and different mixtures of phosgenite, cerussite, laurionite, and galena (AF 6772, E 11048c/e, E 21562, E 23092, and E 25256/eb) were identified and selected.

Preparation of modern replications of cerussite according to the historical process

Modern samples of cerussite were reproduced by a pigment supplier according to the historical process18,19,20. Pieces of lead were placed over a vessel filled with vinegar. Lead was exposed to acetic acid vapors in the presence of horse manure (sample MH) or sugar and yeast dissolved in water (sample MM). White flakes of lead carbonate were produced, collected and milled to obtain a fine pigment. For both samples, the pMC values (Supplementary Table 5) are in accordance with modern radiocarbon contents.

Sample preparation

We prepared samples for radiocarbon analysis using a purposely-designed protocol. To avoid any contamination during the sample preparation process, glassware and tools were thoroughly cleaned. First, glassware was boiled at 100 °C with lab detergent, washed with water, then placed overnight in a 10% HCl bath and washed again with deionized water. Each piece of glassware was covered with aluminum foil and baked at 450 °C for 5 h to remove any remaining organic contaminants. Stainless steel tools were hand washed with TFD4 (detergent, foam for the decontamination of radioactivity), rinsed with deionized water, cleaned again in ethanol in an ultrasonic bath and finally dried in a 60 °C oven.

Samples of cosmetics were first observed under the microscope to remove possible macro-contaminants. However, due to the small size of the samples, we did not apply the standard acid-base-acid method which is usually used to remove micro-contaminants. To control this procedure, modern cerussite samples were treated in the same way. For all the samples, we succeeded in extracting carbon from the lead carbonates. We obtained carbon contents of 0.2 to 1 mg. Samples were directly thermally decomposed in a sealed tube at 400 °C for 2 h on a vacuum line (5 × 10−6 mbar). Lead carbonate decomposes, producing lead oxide, carbon dioxide and water. CO2 was separated from H2O using a dry ice/alcohol trap (−78 °C) and the pressure of each CO2 sample was measured to determine the carbon content. We produced graphite targets by reducing CO2 with hydrogen over iron catalyst41. Before the reaction, the iron powder (Merck, particle size 10 μm) was degassed for 1 h at 600 °C under vacuum. Hydrogen was introduced in the reduction line with an H2/CO2 ratio of approximately 2.5. Iron was heated at 600 °C while the water was trapped at –78 °C. The pressure in the reaction lines was monitored and the reduction ended when the slope of the pressure became below 0.3% in 3 to 4 h. For some samples, due to the presence of Cl (from phosgenite) and/or S (from galena), we had to apply a Sulfix treatment to achieve the carbon formation reaction16.

The wood and reed materials of the cosmetics containers were prepared using the standard acid-base-acid method (0.5 M HCl at 80 °C,1 h/ 0.1 M NaOH at 80 °C, 1 h/ 0.5 M HCl at 80 °C, 1 h)16 and then dried under vacuum overnight (60 °C-0.1 mbar). CO2 was obtained by combustion (5 h, 850 °C) in a sealed tube with an excess of CuO (400–500 mg) and a 1-cm Ag wire. CO2 was separated from H2O using a dry ice/alcohol trap (−78 °C) and the pressure of each CO2 sample was measured to determine the carbon content.

Radiocarbon measurements

Carbon isotopes were measured by AMS on the LMC14/ARTEMIS facility17. We used the NEC9SDH-2 Pelletron coupled to a 134-position MC-SNICS cesium sputter ion source. Normalization of the measurements was done with the oxalic acid II (1 Ox-II per 10 unknown samples). International intercomparison sample (FIRI H) and blanks (C1) were measured with the samples. 14C ages were calculated using the Mook and van der Plicht42 recommendations. Calibration of the radiocarbon dates was against the IntCal13 atmospheric calibration curve43 using OxCal v4.244 (Supplementary Tables 2-5).

Data availability

Data that support the findings of this study are available within the article and its Supplementary Information files and from the corresponding author upon request.