INTRODUCTION
Lead carbonates—cerussite (PbCO3) and hydrocerussite (2PbCO3Pb (OH)2)—were employed as cosmetics, eye remedies and white pigment (known as lead white) from the 4th century BC to the beginning of the 20th century (Welcomme et al. Reference Welcomme, Walter, Van Elslande and Tsoucaris2006; Stols-Witlox Reference Spring2011; Pérez-Arantegui Reference Pérez-Arantegui2021; Messager et al. Reference Messager, Beck, Blamart, Richard, Germain, Batur, Gonzalez and Foy2022). Craddock (Reference Craddock2009), was the first to suggest the use of the 14C content in lead white to distinguish past and modern productions, based on the difference between the carbon sources, derived from plants or fossil fuels. The first 14C dates of manufactured lead carbonates were obtained on cosmetics and paintings by two different groups in France and in Switzerland (Beck et al. Reference Beck, Caffy, Delqué-Količ, Moreau, Dumoulin, Perron, Guichard and Jeammet2018, Reference Beck, Messager, Coelho, Caffy, Delqué-Količ, Perron, Mussard, Dumoulin, Moreau and Gonzalez2019; Hendriks et al. Reference Hendriks, Hajdas, Ferreira, Scherrer, Zumbuhl and Kuffner2019; Beck et al. Reference Beck, Messager, Caffy, Delqué-Količ, Perron, Dumoulin, Moreau, Degrigny and Serneels2020; Messager et al. Reference Messager, Beck, de Viguerie and Jaber2020, Reference Messager, Beck, Germain, Degrigny, Serneels and Cano2021; Sà et al. Reference Sá, Hendriks, Pombo Cardoso and Hajdas2021) after a first application on a corroded lead coffin by Van Strydonck et al. (Reference Van Strydonck, Boudin, Van den Brande, Saverwyns, Van Acker, Lehouch and Vanclooster2016). These successful results clearly indicate that organic carbon in one form or another—not fully identified yet—is incorporated during the corrosion process when organic matter is present or during the lead white manufacturing process described in the ancient recipes (Welcomme et al. Reference Welcomme, Walter, Van Elslande and Tsoucaris2006; Stols-Witlox 2014; Gonzalez et al. Reference Gonzalez, Wallez, Calligaro, Gourier and Menu2019; Photos Jones et al. Reference Photos-Jones, Bots, Oikonomou and Hamilton2020).
The production of lead white was based on a corrosion process, combining the actions of a corrosive liquid and carbon dioxide on metallic lead. This production process was first mentioned by Theophrastus in the 4th–3rd centuries BC (in Caley and Richards Reference Caley and Richards1956) and was used up to the 19th century with minor modifications (Stols-Witlox 2014). From the 19th century on, new industrial processes were developed for a larger scale production using CO2 derived from fossil materials (Villon and Guichard Reference Villon and Guichard1898–1902).
In order to document the incorporation and pathway of carbon in the corrosion mechanism, lead carbonates were produced in the laboratory under monitored experimental conditions. Various materials, with different carbon isotope signatures in 14C and 13C were selected as corrosive reagents and CO2 sources. Six experimental setups were defined combining vinegar, acetic acid, horse manure and fossil CO2 gas.
The corrosion products were characterized by X-ray diffraction. 14C content and δ13C values of the initial reactants and the final products were measured by accelerator mass spectrometry (AMS) and isotopic ratio mass spectrometry (IRSM).
MATERIAL AND METHODS
Samples: Lead White Produced at LMC14
The corrosion process is based on three basic compounds: metallic lead, corrosive liquid, and CO2 source. To monitor the carbon pathway in the reaction, reactants with various carbon isotope signatures were selected (Table 1). The corrosive reagents were the natural vinegar (containing 14C) used in the traditional manufacture and its counterpart depleted in 14C, acetic acid. The two CO2 sources selected were horse manure (containing 14C), one of the ingredients described in the ancient literature (Stols-Witlox 2014), and its counterpart depleted in 14C, CO2 gas.
Table 1 Materials used for lead white produced by corrosion at LMC14.
Six experimental conditions were defined (Figure 1), using lead (Pb) and:
-
1. acetic acid (Aa)
-
2. vinegar (V)
-
3. acetic acid and horse manure (Aa+Hm)
-
4. vinegar and horse manure (V+Hm) [ = historical process]
-
5. acetic acid and fossil CO2 (Aa+ CO2B)
-
6. vinegar and fossil CO2 (V+ CO2B)
Figure 1 Experimental setups for the production of lead white by corrosion. 1: lead and acetic acid (Aa) (and photo 1b); 2: lead and vinegar (V); 3: lead, acetic acid and horse manure (Aa+Hm); 4: lead, vinegar and horse manure (V+Hm) (and photo 4b); 5: lead, acetic acid and fossil CO2 (Aa+ CO2B); 6: lead, vinegar and fossil CO2 (V+ CO2B) (and photo 6b).
To expose the lead to the vapors of the corrosive liquid, lead foils were hung over a 500 mL beaker containing 100 mL of acetic acid (10 vol%) or of vinegar. The whole was placed in a not tightly closed 20 L bucket, either empty (setups 1 and 2) or containing horse manure (setups 3 and 4). The temperature was between 20°C and 30°C throughout the process.
For setups 5 and 6, the beakers containing acetic acid (10 vol%) or vinegar were placed in an airtight Plexiglas box in which fossil CO2 gas was injected under a slight overpressure. The box was regularly refilled to maintain a pressure of ∼2 atm.
After a few days, a white film formed at the surface of the lead. After several weeks of exposure, complete consumption of the metal foil was observed, and the lead turned into thick, porous and brittle white flakes that no longer adhered to the metal (Figure 2). Samples were taken at different times of exposure and the final products were selected for X-ray diffraction (XRD) and AMS 14C dating. For some samples, IRMS δ 13C measurements were also carried out.
Figure 2 Lead white produced after one month of exposure of lead to vinegar and horse manure (setup 4). The corresponding diffractogram is presented in Figure 3b.
METHODS
X-Ray Diffraction (XRD)
Corrosion products were analyzed by XRD using a RU-200B (Rigaku) rotating anode X-Ray generator equipped with a Mo anode (λ = 0.7093 Å). The beam was focused on the sample with a beam size of 100 µm and a photon flux of about 2.107 ph/s. XRD data were collected on a Pilatus 300K (Dectris) hybrid pixel detector in transmission mode with a 2Ө range from 2 to 35°. 2D images were circularly integrated using the FIT2D software and phase identification was carried out with Diffrac-EVA software (Bruker) integrating a reference database from ICDD.
AMS 14C
Between 25 to 35 mg of white powder were thermally decomposed at 400°C on a dedicated CO2 collection line (Beck et al. Reference Beck, Messager, Coelho, Caffy, Delqué-Količ, Perron, Mussard, Dumoulin, Moreau and Gonzalez2019; Messager et al. Reference Messager, Beck, de Viguerie and Jaber2020; Dumoulin and Moreau, Reference Dumoulin and Moreausubmitted). CO2 was graphitized at 600°C using H2 and iron catalyst (Vogel et al. Reference Vogel, Southon, Nelson and Brown1984). The 14C/12C ratio was measured by AMS using the LMC14/ARTEMIS facility, a 3MV NEC Pelletron (Moreau et al. Reference Moreau, Messager, Berthier, Hain, Thellier, Dumoulin, Caffy, Sieudat, Delqué-Količ and Mussard2020; Beck et al. Reference Beck, Caffy, Delqué-Količ, Dumoulin, Goulas, Hain, Moreau, Perron, Setti, Sieudat and Thellier2023). Oxalic acid II was used for normalization and the international intercomparison samples FIRI H and FIRI I for validation. 14C ages were calculated using the Mook and van der Plicht (Reference Mook and van der Plicht1999) recommendations and expressed in pMC (percent of Modern Carbon). Radiocarbon results were calibrated using Oxcal 4.4.4 with the Bomb NH1 21 curve (Hua et al. Reference Hua, Turnbull, Santos, Rakowski, Ancapichún and De Pol-Holz2022).
IRMS δ13C
δ 13C values were measured offline with an IsoPrime 100 mass spectrometer. Between 200 and 350 µg of sample were loaded in homemade vials. Before use, they were pumped under vacuum to reach a pressure greater than 5.10–12 mbar. Samples were then treated with 0.2 mL of anhydride phosphoric acid solution at 90°C under vacuum. After normalization to international standards, the results were expressed in ‰, with respect to the Vienna Pee Dee Belemnite (V-PDB) reference.
For some samples, it was not possible to obtain IRMS δ13C due to limited access to the mass spectrometer. For these cases, AMS δ13C values were reported after checking that the results obtained by the two methods on some samples were reasonably comparable (see Table 3 in the results section). However, under no circumstances should these AMS δ13C values be used as absolute values.
RESULTS
14C Content and δ13C Value of the Initial Reactants
The results of the initial reactants are given in Table 2. The carbon contents of horse manure (100.4 ± 0.3 pMC) and natural vinegar (100.8 ± 0.3 pMC) show compatible results within 2σ, indicating that these two elements are contemporary. These values are also consistent with the contemporaneous atmosphere recorded in 2020 and measured on nuts collected in France (100.8 ± 0.4 pMC). CO2 gas, usually used as a blank, gives a mean 14C content of 0.206 ± 0.041 pMC. The industrial acetic acid 14C content is very low, close to the 14C detection limit.
Table 2 14C content in pMC and AMS δ13C of the reactants. For comparison, the contemporaneous 14C content of the atmospheric CO2 is indicated. It was estimated by measuring the 14C content of nuts collected in France in 2020.
The δ13C values of horse manure and vinegar, –29‰ and –18‰, are in the range of organic matter, whereas the δ13C values of CO2 gas and acetic acid, –35‰ and –43.6‰ respectively, are characteristic of materials derived from fossil fuels.
Chemical Structure of the Final Products
A white powder was obtained for all the setups due to the corrosion of the metallic lead foil (Figure 2). Two main products were determined by XRD after 2 months of exposure: lead acetate trihydrate, Pb(CH3COO)2.3(H2O) (Figure 3a) for setups 1, 2, and 5 and cerussite, PbCO3 (Figure 3b) for setups 3, 4, and 6 (Table 3).
Table 3 Composition, 14C content in pMC, IRMS and AMS δ13C of the reaction products after 53 to 120 days of exposure (120 days for setups 1 and 2; 73 days for setups 3 and 4; 53 for setups 5 and 6). Composition was determined by XRD after 64 days for setup 1; 53 days for setups 2, 5, 6; 10 and 73 days for setups 3 and 4.
Figure 3 X-ray diffractograms of the white powder collected from: (a) setup 2 showing the production of lead acetate trihydrate (setups 1 and 5 show the same pattern), (b) setup 4 showing the production of cerussite (setups 3 and 6 show the same pattern).
Lead acetates were detected for setups 1 and 2 where lead was submitted to the vapors of vinegar or acetic acid (CH3COOH), as also observed by Gonçalves et al. (Reference Gonçalves, Pires, Carvalho, Mendonca, Cruz and Urbano Alfonso2010). The formation of the trihydrate form was promoted when water vapor condensed (Gonzalez et al. Reference Gonzalez, Wallez, Calligaro, Gourier and Menu2019) following the reaction:
$${\rm{Pb + 2 C}}{{\rm{H}}_{\rm{3}}}{\rm{COOH + 1/2 }}{{\rm{O}}_{\rm{2}}}{\rm{\ +\ 2}}{{\rm{H}}_{\rm{2}}}{\rm{O}} \to {\rm{Pb}}{\left( {{\rm{C}}{{\rm{H}}_{\rm{3}}}{\rm{COO}}} \right)_{\rm{2}}}{\rm{.3}}\left( {{{\rm{H}}_{\rm{2}}}{\rm{O}}} \right)$$
In presence of horse manure and vinegar or acetic acid (setups 3, 4), pure cerussite (PbCO3) was formed from the second day of exposure. The formation of lead carbonate indicates that the supply in CO2, produced here by fermentation of the manure, was enough for the reaction
$${\rm{Pb + C}}{{\rm{O}}_{\rm{2}}}{\rm{\ +\ 1/2 }}{{\rm{O}}_{\rm{2}}} \to {\rm{PbC}}{{\rm{O}}_{\rm{3}}}$$
to be complete. No other intermediate compounds described by Gonzalez et al. (Reference Gonzalez, Wallez, Calligaro, Gourier and Menu2019), such as plumbonacrite (Pb5(CO3)3O(OH)2) and hydrocerussite (2PbCO3.Pb(OH)2) were observed, even in the early stages of the reaction.
When CO2 gas was injected, two different results were obtained. For setup 5 with acetic acid, corrosion took place very slowly and only a thin layer of lead acetate was detected. In contrast, in presence of vinegar (setup 6), lead carbonate (cerussite) was obtained. Several hypotheses can be put forward to explain this difference: experimental failure (lack of CO2 in setup 5) and/or a possible role played by the microorganisms contained in vinegar (setup 6) as an additional source of CO2 as suggested by Sanchez-Navas et al. (Reference Sanchez-Navas, Lopez-Cruz, Velilla and Vidal2013) and Photos-Jones et al. (Reference Photos-Jones, Bots, Oikonomou and Hamilton2020).
14C Content and δ13C Value of the Final Products
The results obtained for the final products are presented in Table 3. For setups 1 and 2, the samples were taken after 4 months of exposure. Lead acetate formed from lead and vinegar has a 14C content of 103.394 ± 0.205 pMC and a δ13C value of –17.4‰. Lead acetate formed from lead and acetic acid has a 14C content of 10.42 ± 0.05 pMC and a δ13C value of –39.6‰. The δ13C values are of the same order of magnitude as those of the initial ingredients (vinegar δ13C value = –18.8‰; acetic acid δ13C value = –35±3‰), but the 14C contents are slightly higher (vinegar 14C content = 100.8 ± 0.3 pMC; acetic acid 14C content = 0.206 ± 0.041 pMC). These results are probably due to external contamination by environmental CO2 or by the other setups nearby, occurring during the long time of exposure. For setup 1, measurement after one week of exposure only seems to confirm this interpretation with a 14C content of 1.558 ± 0.125 pMC and a δ13C value of –36 ± 4‰ (SacA61010, not reported in Table 3) very close to those of acetic acid. For setup 2, the result is above the current atmospheric level, but no other source of 14C has been identified so far. For setups 1 and 2, it can be inferred that Pb(CH3COO)2.3(H2O) has a similar carbon isotope signature to that of the corrosive liquid. No significant δ13C fractionation was observed. These results confirm that lead acetate was produced by the reaction of the acid vapors on lead according to reaction (1).
Lead actetate was also obtained in the case of setup 5 when acetic acid and fossil CO2 gas were used. The 14C content is 0.359 ± 0.033 pMC and the δ13C value –36.43‰. These values are similar to those of the initial ingredients, but as both acetic acid and fossil CO2 gas have a very similar carbon isotope signature, it is not possible to estimate the contribution of each component to the formation of lead acetate in that case.
For setups 3, 4, and 6, the samples were taken after ∼2 months of exposure. The cerussite produced in setup 3 from lead, acetic acid and horse manure has a 14C content of 94.4 ± 0.2 pMC and a δ13C value of –50.3‰. In setup 4 where vinegar is present, the 14C content is 101.1 ± 0.3 pMC and the δ13C value is –49.9‰. In setup 6, the cerussite produced from lead, vinegar and fossil CO2 gas has a 14C content of 4.59 ± 0.06 pMC and a δ13C value of –46.8‰. The 14C content of the cerussite produced shows values around 100 pMC when horse manure was involved whatever the corrosive reagent used (vinegar or acetic acid), and a low value when fossil CO2 gas was used with vinegar (Table 3). These results clearly indicate the predominant influence of CO2 when lead carbonate is produced according to reaction (2).
In these three setups, carbonatation was achieved and pure cerussite was obtained, mostly carriyng the 14C signature of the initial CO2. However, the influence of the corrosive reagent can be seen when looking in detail at the pMC values for setups 3 and 4. The 14C content of cerussite from setup 4 was 101.1 ± 0.3 when horse manure was used with natural vinegar, whereas the 14C content of cerussite from setup 3 was 94.4 ± 0.2 when horse manure was used with acetic acid (<0.25 pMC). The latter value reflects the presence of traces of acetic acid between the flakes of lead white or minor phases (<5%) not seen by XRD, carrying the 14C signature of acetic acid. The same observation can be made in the reverse situation (setup 6) where the cerussite produced by fossil CO2 gas (0.21 ± 0.04 pMC) with natural vinegar (100.8 ± 0.3) has a 14C content of 4.6 ± 0.1 pMC.
About the δ13C results, contrary to lead acetate, cerussite produced using horse manure shows a strong fractionation: very depleted δ13C values of –50.3‰ and –49.9‰ were obtained although the initial ingredients had δ13C values of –18.8‰ (vinegar) or –43.6‰ (acetic acid) and –28.9‰ (horse manure). The δ13C signature is lower than expected for the organic matter from an animal’s diet, which is generally between –32‰ and –10‰ (Amelung et al. Reference Amelung, Bol and Friedrich1999; Sponheimer et al. Reference Sponheimer, Robinson, Ayliffe, Passey, Roeder, Shipley, Lopez, Cerling, Dearing and Ehleringer2003; Inacio et al. Reference Inácio2013). Hendriks et al. (Reference Hendriks, Caseri, Ferreira, Scherrer, Zumbuhl, Kuffner, Hajdas, Wacker, Synal and Guntet2020) and Messager et al. (Reference Messager, Beck, Blamart, Richard, Germain, Batur, Gonzalez and Foy2022) obtained similar results, with δ13C values between –45 and –40‰ for lead whites prepared with the same process but provided by other suppliers.
Low δ13C values are not well documented for carbon dioxide. Botz et al. (Reference Botz, Pokojski, Schmitt and Thomm1996) reported δ13C values up to –49.7‰ for CO2 emitted during biologic methane formation and Whiticar (1999) indicated values up to –46‰ for CO2 produced by methane oxidation. As gas generated by the degradation of organic matter in manure contains both methane and carbon dioxide (Wartell et al. Reference Wartell, Krumins, Alt, Kang, Schwab and Fennell2012), the low δ13C values obtained for cerussite probably reflect the fermentation process. A direct analysis of the δ13C of gas would be necessary to confirm this hypothesis.
14C Dates for Lead White (Cerussite)
When cerussites were produced from acetic acid and/or fossil CO2 (setups 3, 5, 6), apparent old ages were found due to the presence of dead carbon in the initial ingredients. The calculation of absolute dates is therefore not relevant in these cases.
For cerussite produced according to the process described in the ancient recipes using natural vinegar and horse manure (setup 4), the 14C content was converted into calendar dates using the post-bomb atmospheric NH1 21 calibration curve (recorded up to 2019 only). Two dates were obtained due to the shape of the bomb peak: 1955 and 2017-… Despite the limited expansion of the calibration curve, the second interval can be considered to be in agreement with the manufacturing date in 2020. This result is consistent with the level of 14C recorded in the atmosphere for this given year.
Together with the pioneering works of Hendriks et al. (Reference Hendriks, Hajdas, Ferreira, Scherrer, Zumbuhl and Kuffner2019) and Beck et al. (Reference Beck, Messager, Coelho, Caffy, Delqué-Količ, Perron, Mussard, Dumoulin, Moreau and Gonzalez2019) on other lead white reproductions, this result demonstrates that the radiocarbon dates for these pigments produced following the historical corrosion process are accurate and meaningful.
CONCLUSION
Lead white was produced in the laboratory according to ancient recipes involving metallic lead, vinegar and horse manure. For comparison, alternative materials depleted in 14C—acetic acid and fossil CO2 gas—were also selected to discriminate the function of each ingredient. The reaction between lead and vinegar or acetic acid resulted in lead acetates with a carbon isotopic signature close to that of the corrosive reagent. In the presence of CO2, the carbonatation reaction occurred and the lead carbonates produced (cerussite here) had a predominant 14C signature of the carbon dioxide source.
As a result, lead white formed by the corrosion process involving vinegar and horse manure has the 14C content of its initial ingredients. This experimental result confirms the possibility of dating lead white by the 14C method and definitely supports the studies conducted on cosmetics or paintings (see introduction). However, a significant fractionation in 13C was observed, which is not fully explained yet.
In conclusion, when all the ingredients are biogenic, as was the case before the 19th century, the date of manufacture of lead white can be reliably determined.
ACKNOWLEDGMENTS
The authors thank the LMC14 colleagues for their help during sample preparation, graphitization and AMS 14C measurements, LSCE colleagues for IRMS δ13C measurements obtained on the Panoply analytical platform Dual Inlet mass spectrometer and Eddy Foy from LAPA (Saclay) for XRD experiment. We are also grateful to C. Oberlin from the CDRC (Lyon) for providing dated vinegar.
This work was partially supported by the Paris Seine Graduate School Humanities, Creation, Heritage, Investissement d’Avenir ANR-17-EURE-0021 – Foundation for Cultural Heritage Science.
Open access
