Reconstructing a medieval recipe for red lead (Pb₃O₄): an archaeometry approach

1 Reconstructing a medieval recipe for red lead (Pb₃O₄): an archaeometry approach Amélie Demoulin1, Carine Perrin2, Odile Guillon3, Valérie Gontero-Lauze4, Jean Rouquerol1 and Florence Boulc’h1* 1 Aix-Marseille Université, CNRS UMR 7246 Matériaux Divisés, Interfaces, Réactivité, Électrochimie (MADIREL) 2 Aix-Marseille Université, CNRS UMR 7334 Institut Matériaux Microélectronique Nanosciences de Provence (IM2NP) 3 Centre Interdisciplinaire de Conservation et de Restauration du Patrimoine (CICRP) 4Aix-Marseille Université, EA 4235 Centre Interdisciplinaire d'Étude des littératures d’Aix Marseille (CIELAM) * Corresponding author [email protected] ABSTRACT Red lead (Pb₃O₄), or minium, was one of the most common pigments in medieval manuscript illumination. However, the exact procedures used for its preparation and the thermal pathways underlying its formation remain insufficiently documented. In this study, we experimentally reconstructed a medieval synthesis of red lead based on recipes from Theophilus’s 12th-century treatise. Controlled transformation rate thermal analysis (CRTA) combined with in situ X-ray diffraction provided insights into the phase transitions occurring during heating. Our results indicate that the pigment referred to as “minium” in historical sources does not always correspond to pure Pb₃O₄, but may involve mixtures of PbCO₃, PbCO₃·2PbO, α-PbO, and β-PbO, depending on the thermal and atmospheric conditions. This variability reflects both the metastability of Pb₃O₄ and the technological diversity of historical lead-based pigments. By bridging experimental archaeology, historical reconstruction, and advanced materials characterization, this work contributes to a better understanding of medieval pigment production and its underlying chemical mechanisms. It highlights how archaeometry approaches can illuminate the interplay between craft knowledge and material science in the Middle Ages. KEYWORDS Red lead (minium); Medieval pigments; Historical recipes; Experimental archaeology; Archaeometry; Controlled Transformation Rate Thermal Analysis (CRTA); In situ X-ray diffraction; 2 INTRODUCTION Our research project focuses on the second half of the 15th century and, more specifically, on the manuscripts produced during this period. Its aim is to better understand how seigneurial libraries were built at the end of the Middle Ages. In this context, we seek to address several key questions: Which authors and texts were read during this period? What iconographic programs accompanied them? How were images and texts related? Which materials were used for illumination, and did these choices depend on the iconographic or textual content of the manuscript? In other words, we are trying to understand (i) how the image fits into the text, (ii) how the text and the image fit into this manuscript, and (iii) how this manuscript fits into the library. To understand this, we chose to focus all these questions on the collection of manuscripts assembled by Tanguy IV du Chastel and Jeanne Raguenel de Malestroit, a couple of noble bibliophiles who belonged to the royal court of Charles VII and later of Louis XI. Our working methodology, combining medievalists and chemists, brings together two approaches: one consists in analysing the manuscript in situ within its place of conservation, the other in reconstructing ex situ, in the chemistry laboratory, the gesture of the illuminator when preparing the colours on his palette. The reproduction or reworking of historical processes has only recently gained recognition in the history of chemistry 1 , 2 . Today, it is increasingly used and acknowledged as a valuable complement to historical inquiry. In the field of heritage science, one notable precedent is the work of Gonzalez et al. (2019), who proposed the first reasoned, experiment-based reconstruction of a corrosion process described in historical sources — the 17th-century Dutch process 3 . To our knowledge, the present study is the first to apply an experimental approach specifically to the practices of medieval illuminators. The synthesis and use of red lead (Pb₃O₄) has a long history 4 , dating back to medieval manuscript Despite more than sixty years of studies on the thermal decomposition of lead carbonate, the precise conditions required for Pb₃O₄ formation remain poorly understood. Most previous investigations have focused on the pathways leading to PbO, leaving open questions regarding the completeness of the reaction, the temperature range of Pb₃O₄ stability, and the reproducibility of its 5 . Addressing these questions is essential not only for solid-state chemistry, but also for cultural heritage studies, where red lead played a central role as a pigment. In this work, we adopt an experimental approach inspired by medieval pigment, notably those compiled in the Schedula diversarum artium by Theophile 6 . By revisiting these historical sources through modern characterization techniques, our aim is twofold: to reconstruct the processes that illuminators may have employed in the preparation of red lead, and to advance the understanding of the fundamental chemical mechanisms involved in lead carbonate decomposition. 3 To achieve this, we employed Controlled Transformation Rate Thermal Analysis (CRTA) combined with in situ high-temperature X-ray diffraction (XRD). CRTA enables quasi-equilibrium conditions, allowing us to disentangle overlapping reactions and propose a general mechanism for Pb₃O₄ formation 7 , 8 . In situ XRD further confirmed the sequence of structural transformations and provided direct evidence for the phases involved. To our knowledge, this combination of advanced thermal and structural methods is applied here for the first time in the field of heritage science. Our results offer new insights into both the chemistry of Pb₃O₄ synthesis and the material practices of medieval colour preparation. EXPERIMENTAL Materials. The preparation of red lead is described in the section devoted to ceruse as follows6: “Grind ceruse […] place it into two or three new vessels, set them on hot coals; take a thin curved iron […] with which you can stir the ceruse from time to time. Continue until the minium becomes completely red.” Ceruse (lead white) designates a pigment based on lead carbonate, present either as cerussite (PbCO₃), hydrocerussite (2PbCO₃·Pb(OH)₂), or, more frequently, as a combination of both phases. As a reference system for our study, we selected the simpler phase, PbCO₃, obtained from Kremer Pigments (ref. 46000). Thermal analysis. Thermogravimetric analyses (TGA) were performed on ~72mg samples using a Q500 analyzer (TA Instruments). Conventional TGA experiments were conducted under flowing air (40mL·min⁻¹), heating to 650°C at 4°C·min⁻¹. Controlled Transformation Rate Thermal Analysis (CRTA) measurements were performed on the same instrument in constant reaction rate mode, under flowing air or argon (40mL·min⁻¹). X-ray diffraction. Ex situ: Powder XRD patterns were collected on a Siemens D5005 diffractometer with Cu-Kα radiation (20–40° 2θrange, step size 0.04°, step time 2s). In situ: High-temperature XRD was carried out in Bragg–Brentano geometry with a PANalytical X’Pert MPD diffractometer equipped with an X’Celerator detector. Measurements (20–40° 2θ range, step size 0.04°, step time 50s) were performed in air using an Anton Paar HTK 1200 chamber, heating from room temperature to 650°C at ~1.5°C·min⁻¹. Photography. Samples were documented with Nikon D800E (36MP) and Nikon D850E (45MP) cameras equipped with 60mm objectives. Images were acquired in the visible range with a Schott 4 BG39 filter, illuminated by two Elinchrome Ranger lamps (1100 W). An X-Rite ColorChecker was used for calibration. RESULTS AND DISCUSSION 1. Thermal analysis of the decomposition of lead carbonate Figure 1 presents CRTA curves recorded for neutral lead carbonate (PbCO₃) under flowing air and argon. The experiments were conducted at a controlled reaction rate of 18.0±0.1mg·h⁻¹·g⁻¹, with a total acquisition time of ~10h. The advantage of CRTA over conventional thermogravimetry is evident when comparing the thermal analysis curves obtained in air (Fig.2): the dashed line corresponds to conventional TGA, whereas the solid red line represents CRTA. In the latter, four distinct decomposition steps are clearly resolved, enabling mechanistic interpretation. According to the CRTA data, the decomposition of PbCO₃ occurs between 20°C and 650°C, with a mass loss of 16.5±0.1% under both air and argon. This value matches the theoretical loss expected from the following reaction, confirmed by the XRD phase analysis (Fig. 3): PbCO₃(s) → β-PbO(s) + CO₂(g) (1) Figure 1 (a) Thermal decomposition of lead carbonate monitored by Controlled Transformation Rate Thermal Analysis (CRTA) in flowing air (solid curve) and argon (dashed curve). (b) Experimental conditions correspond to a reaction rate of 18.0±0.1mg·h⁻¹·g⁻¹. (c) Temperature profile recorded under constant decomposition rate conditions; analyses required approximately 10 hours. 5 Figure 2 Thermal decomposition of lead carbonate monitored by conventional thermogravimetry (dashed curve, heating rate = 4°C·min⁻¹) and by CRTA (solid curve). Four well-resolved steps are identified (1–4). Figure 3 X-ray diffraction patterns of the Kremer sample recorded at room temperature: (a) before CRTA and (b) after CRTA. The initial phase corresponds to pure cerussite, PbCO₃ (PDF 01-76-2056), while the final phase is pure massicot, β-PbO (PDF 00-038-1477). Identical results were obtained under flowing argon. All diffraction patterns recorded up to 240±10°C display only the PbCO₃ phase, characterized by intense reflections between 24° and 26° (Fig. 4). Notably, from 180±10°C onwards, these reflections gradually shift to lower angles, most likely due to lattice expansion upon heating. Above 240±10°C, an additional reflection emerges at ~29°, indicative of the formation of the PbCO₃·2PbO phase, in agreement with the following reaction: 3PbCO₃(s) → PbCO₃·2PbO(s) + 2CO₂(g) (2) 20 25 30 35 40 Intensity (a.u.) 2 theta (°) (a) (b) 6 Figure 4 In situ X-ray diffraction of lead carbonate during heating (1.5°C·min⁻¹) from room temperature to 650°C, illustrating the four steps observed in the CRTA curve. The temperature of 240±10°C identified for this reaction in the X-ray diffraction chamber is consistent with the features of the thermal analysis curve. On the CRTA trace, an inflection point is observed at 250±2°C. This profile, commonly described as saddle-shaped—characterized by a transient decrease in temperature following the onset of decomposition—is indicative of a nucleationand-growth mechanism 9 . As the reaction interface propagates, a lower temperature is sufficient to maintain the reaction rate at the imposed level. The reaction is also found to be zero-order, as the decomposition temperature is independent of the extent of conversion. A similar phenomenon has been reported for the thermal decomposition of gibbsite, where the isothermal stage arises from the advance of a reactive interface parallel to the largest crystal face. By analogy, the zero-order behaviour observed here can be assigned a clear physical significance. Importantly, diffraction patterns reveal that the reaction does not proceed to completion (Fig. 4). Even at 260°C, a substantial fraction of lead carbonate remains in the sample. This thermodynamic limitation is corroborated by CRTA data: while the reaction equation predicts a theoretical mass loss of 11%, the experimental value is only 9.0±0.1% under both air and argon atmospheres. 7 In contrast to step one, which is characterized by a constant decomposition temperature of 220°C, step two exhibits a decomposition temperature that gradually increases from 220°C to 330°C in air, as shown by the CRTA curve. This behaviour suggests that the underlying mechanism is more complex than that proposed for step one. The X-ray diffraction pattern recorded at 290±10°C reveals residual lead carbonate (PbCO₃) together with the PbCO₃·2PbO phase, whereas the pattern collected at 300 ± 10°C shows only PbCO₃·2PbO and α-PbO. The slope change observed in the CRTA curve around 220°C can therefore be attributed to the two simultaneous reactions described below. 3PbCO₃(s) → PbCO₃·2PbO(s) + 2CO₂(g) PbCO₃·2PbO(s) → 3α-PbO(s) + CO₂(g) (2) (3) As shown by the diffraction pattern recorded at 330±10°C in the X-ray diffraction chamber, step two concludes with the single transformation of PbCO₃·2PbO into α-PbO. Notably, this step is complete in the thermal analysis performed under argon, emphasizing the influence of the surrounding atmosphere on the thermodynamic parameters—a key factor in understanding and reproducing the medieval preparation of red lead. Moreover, the reaction order under argon remains zero, as previously demonstrated for step one. We therefore conclude that the second stage of decomposition is governed by two parallel reactions (reactions (1) and (2)), with reaction (2) remaining incomplete. At 360±10°C, XRD patterns show PbCO₃·2PbO, α-PbO, and Pb₃O₄, confirming that two reactions take place concurrently. PbCO₃·2PbO(s) → 3α-PbO(s) + CO₂(g) 3α-PbO(s) + ½O₂(g) → Pb₃O₄(s) (3) (4) Subsequently, the intensities of the characteristic peaks of PbCO₃·2PbO and α-PbO decrease rapidly, and the Pb₃O₄ phase becomes dominant after one hour at 380°C±10°C. These observations are corroborated by the CRTA curve: reaction (3) is associated with a theoretical mass loss, whereas reaction (4) corresponds to a theoretical mass gain. The plateau observed on the CRTA curve therefore indicates that these two reactions occur simultaneously during the third stage. Furthermore, the diffraction patterns demonstrate that Pb₃O₄ remains stable over a wide temperature range, suggesting that the plateau may also reflect the intrinsic thermal stability of Pb₃O₄ within this interval. 8 A slope break is observed on the CRTA curve at 410°C±2°C, indicating that Pb₃O₄ is stable over a narrow range of only 30°C±2°C under controlled transformation rate thermal analysis conditions. To elucidate this fourth stage, we turn to the diffraction patterns recorded in the X-ray chamber, where the sample was heated at ~1.5°C/min to reduce the experimental duration. As noted earlier, CRTA requires more than 10 hours—for instance, reaching 300°C takes about 500 minutes, whereas the same temperature is reached in 220 minutes in the X-ray chamber. This compromise inevitably alters the thermodynamic parameters, leading to a shift in the temperature range of the thermal decomposition of Pb₃O₄ into α-PbO. Specifically, the Pb₃O₄ transition temperature is observed at 535°C±10 °C in the X-ray chamber, compared to 410°C±2°C on the CRTA curve. Nevertheless, the in situ X-ray diffraction study demonstrates that red lead (Pb₃O₄), initially formed from α-PbO, reverts to this phase above 535°C±10°C. We therefore conclude that Pb₃O₄ is metastable in air. Otto noted that the Pb₃O₄–PbO–O₂ system is readily reversible under controlled atmospheres, though he provided no experimental evidence 10 . In our case, this reversibility is a key factor for both the preparation of red lead and its long-term stability as a pigment. Accordingly, the fourth stage observed on the CRTA curve can be ascribed to the transformation of Pb₃O₄ into α-PbO. In addition, at 560°C±10°C, the X-ray diffraction patterns reveal the transformation of α-PbO into β-PbO. Taken together, these results indicate that the fourth decomposition stage corresponds to the following two transformations. Pb₃O₄(s) → 3α-PbO(s) + ½O₂(g) α-PbO(s) → β-PbO(s) (5) (6) The first reaction pathway is consistent with the thermal analysis, which shows a clear mass loss between 410°C±2°C and 460°C±2°C. The second pathway most likely accounts for the plateau observed on the CRTA curve, extending from 460°C±2°C to 650°C±2°C. Cultural interpretation Theophile’s 12th-century recipe describes preparing Pb₃O₄ by heating ceruse on hot coals until the powder turns completely red. But what does “red” correspond to in terms of crystalline phases? To address this question, we heated PbCO₃ samples in a furnace at temperatures ranging from 200 to 650°C at a high rate of 10°C·min⁻¹. Each sample was subsequently analysed by photography and ex situ X-ray diffraction to characterize its colour and to identify the crystalline phases formed as a function of temperature (Table 1). 9 Table 1 Colours and crystalline phases of cerussite (PbCO₃) samples as a function of annealing temperature. 25°C-260°C 260°C-300°C 300°C-400°C 450°C-500°C 600°C-650°C PbCO3 PbCO3.2PbO PbCO3 PbO PbO (maj.) Pb3O4 Pb3O4 (maj.) PbO βPbO βPbO (maj.) PbO The sequence of phases observed with increasing temperature is fully consistent with those identified in the in situ X-ray diffraction study. Specifically, we observe the successive formation of PbCO₃·2PbO, litharge (α-PbO), minium (Pb₃O₄), its reversion to litharge (α-PbO), and finally massicot (β-PbO). The stability ranges of these phases closely match those revealed in situ. Table 1 further highlights that the samples display various red to orange hues between 300 and 600°C. These colors may arise not only from Pb₃O₄, which appears orange, but also from mixtures of phases such as PbCO₃·2PbO and α-PbO. This complexity is critical for interpreting the red tones found in medieval pictorial works. Indeed, our results show that the “red” coloration of lead compounds is not unique: in situ analyses of lead-based red pigments must account for both the presence and the absence of Pb₃O₄. It should also be recalled that the present study focused on cerussite (PbCO₃). However, the historical term “ceruse” also designates hydrocerussite (2PbCO₃·Pb(OH)₂), or more commonly, mixtures of cerussite and hydrocerussite. We therefore plan to extend this investigation to the thermal decomposition of 2PbCO₃·Pb(OH)₂ and of PbCO₃/2PbCO₃·Pb(OH)₂ mixtures, in order to better understand the phases actually present in medieval lead-based red pigments. CONCLUSION We have investigated the chemical mechanisms underlying pigment preparation using advanced thermal and structural characterization techniques. Following a 12th-century recipe, we examined the synthesis of minium by combining Controlled Transformation Rate Thermal Analysis (CRTA) with in situ X-ray diffraction. Our results demonstrate that “minium,” commonly identified with red lead (Pb₃O₄), can in fact correspond either to pure Pb₃O₄ or to mixtures of PbCO₃, PbCO₃·2PbO, α-PbO, and β-PbO. In some cases, lead-based red pigments may even be entirely devoid of Pb₃O₄. This finding is crucial for interpreting in situ analyses of medieval red painting materials.