1. Introduction
Red and yellow ochres constitute by far the most widespread and extensively used class of mineral pigments since the beginning of the artistic expression of humankind, being among the earliest pigments used from the Paleolithic era, and later in Egyptian, Persian, Greek, and Roman art, up to the Byzantine, Medieval, and modern period [
1]. They are either natural or partially treated pigments composed of a mixture of ferric oxo-hydroxide minerals as chromophore agents. Hematite and goethite constitute the most common chromophore phases in red and yellow ochre, respectively. Furthermore, they are generally associated with accessory silicate minerals like clays, quartz, and feldspars, whose amount and composition are mainly related to the source of extraction and processing of raw materials [
2]. They are usually found in lateritic soil horizons widely occurring in subtropical as well as temperate regions, making them the most readily available mineral pigments throughout the world. Ochres are characterized by exceptional aesthetic and technological properties, such as good pigmenting properties and relatively high chemical and physical stability, favoring their application on a wide range of organic and inorganic substrates and employing a wide variety of binding media [
2].
Ochre-based pigments were widely used during the Roman civilization [
3,
4,
5]. The almost ubiquitous presence of ochre in Roman wall painting practice is evident in archaeological remains, with countless examples in well-known sites like Pompeii, Herculaneum, and Oplontis, as well as sites across the Italian peninsula and the Mediterranean region. The widespread use of ochre pigments in evolved societies like ancient Rome deeply influenced the development of their technological application in wall paintings. Optimization of their aesthetic and functional properties led to the formulation of complex and/or physically modified pigment mixtures, as testified by the written sources of several Latin authors. Vitruvius and Pliny dedicated ample passages in their encyclopedic manuals
De Architectura and
Naturalis Historia to ochre-based pigments, called either
Sinopis,
Rubrica, or
Ochra, defining both sources of raw materials and technological practices to enhance their properties [
3] (pp. 283–291), [
4] (pp. 117–121).
The ubiquitous utilization of ochre-based pigments is related not only to the widespread occurrence of raw materials for their production but also to the existence of a wide range of simple operational processes to concentrate the active compounds and optimize their optical and aesthetic properties. Apart from direct extraction from naturally concentrated ores, an enrichment of chromophore ochre compounds from lateritic soils can be easily obtained through simple sieving and decantation procedures [
2].
Furthermore, due to the close chemical and crystallographic correlation between the chromophores in the main classes of ochre pigments, pyrotechnology has long been used to change the chromatic properties of the extracted raw materials, as in the case of producing red ochre from yellow ochre by heating [
6]. Despite being characterized by different unit cells (goethite α-FeOOH crystallizes in the orthorhombic space group
Pbnm, while hematite α-Fe
2O
3 crystallizes in the rhombohedral space group
R c), the phases are characterized by common anionic sub-lattices [
7]. This structural characteristic allows the transformation from goethite to hematite through a topotactic process after dehydroxylation at relatively low temperatures (below 300 °C), with ongoing debates in the scientific community related to the possible formation of metastable non-stoichiometric compounds such as hydrohematite [
8] and protohematite [
7] during the reaction process. Since prehistoric times, this technological property favored the treatment of yellow ochre to produce the red counterpart artificially through simple manufacturing processes to obtain the desired colors. Furthermore, apart from intentional thermal treatments of the pigments, the chromatic conversion process may also be induced by natural disasters such as fires or volcanic eruptions like the one of Mount Vesuvius in 79 AD, with associated alterations to original palettes, especially in wall paintings. During the catastrophic event, temperatures around 280–300 °C were reached in the city of Pompeii due to the action of the pyroclastic flow [
9], while the pyroclastic surge that struck the city of Herculaneum was characterized by even higher temperatures (around 400 °C), causing evident chromatic changes in the famous monochrome panels decorating the luxurious domus and other public and private buildings of the towns.
For these reasons, the determination of solid analytical protocols for the discrimination between natural and thermally treated red ochres constitutes an important archaeometric challenge to decipher both production processes and alteration phenomena. Several studies have been reported in the literature on monitoring mineral composition and crystallographic changes during the thermal transformation processes through X-ray powder diffraction (XRPD) [
6,
7], micro- and nano-structural changes in the pore networks of oxo-hydroxides through transmission electron microscopy (TEM) [
10,
11], or on employing other analytical techniques such as thermal and spectroscopic analyses [
12] and more articulated multi-analytical approaches [
13]. Furthermore, a recent study [
14] proposed an alternative approach based on portable X-ray fluorescence (p-XRF) that obtained promising results on Pompeian wall paintings using arsenic as a heavy metal tracer for the discrimination of original and converted red ochres. Nevertheless, the monitoring of heavy metals and other trace compounds for such purposes needs to be carefully evaluated considering the possible influence of adulterations and additions of other compounds to ochre pigments, widely used in the Roman world, as reported in the encyclopedic works of Vitruvius and Pliny. The addition of charcoal to ochres was quite common [
15], as was the addition of artificial lead-based pigments [
5]. Two of these mixed pigments,
Sandyx and
Syricum, are documented in the encyclopedic work of Pliny, and they were used for specific painting purposes, such as an under-coating for cinnabar and red lead [
3] (pp. 283–295).
The complex technological framework related to the utilization of ochre pigments in the Roman world requires rigorous scientific efforts to characterize their application in wall paintings. This study reports the results of a phased multi-analytical archaeometric study to characterize ochre-based roman monochrome frescoes in the tablinum of the House of the Bicentenary at the archaeological site of Herculaneum, starting with non-invasive investigations to provide preliminary information on the pigments used and targeting representative sites for limited sampling, followed by in-depth laboratory research to characterize the pigments and understand transformation processes. The study was aimed at understanding the original palette and color scheme of the monochrome portions of the tablinum, thus distinguishing the fields of original red monochrome background from the fields of red converted from yellow due to the heat from the 79 AD eruption of Mount Vesuvius.
The research was carried out from 2015 to 2017 in the context of a collaborative project of the Getty Conservation Institute, the Herculaneum Conservation Project, and the Archaeological Park of Herculaneum, to study and conserve the architectural surfaces in the tablinum of this house, with a central component focused on the in-depth scientific study of the wall paintings.
2. Herculaneum and the House of the Bicentenary
Herculaneum, an ancient Roman town located on the Bay of Naples at the base of Mount Vesuvius, was a thriving city containing a wealth of architectural features typical of the period, including wall paintings and mosaic pavements [
16].
The eruption of Mount Vesuvius in 79 AD had a devastating effect on the surrounding region. Entire towns and cities were destroyed and covered with hot ash and volcanic material. Herculaneum was buried under nearly 20 meters of pyroclastic flow that sealed it for centuries. Destroyed as a living city, it was preserved nearly intact for two millennia, until it was rediscovered in the 18th century, first excavated by tunneling, then as an open-air site from the late 19th century to the mid-20th century [
17], followed by limited excavations since then.
The House of the Bicentenary (
Figure 1), excavated in 1938, is one of the most sumptuous and richly decorated domus located along the decumanus maximus. The tablinum or formal reception room, which is the focus of this project, is located along the central axis of the house, leading from the atrium to the peristyle garden. It has three walls oriented approximately east, south, and west. The north end is open to the atrium, and the south end has a large opening leading to the peristyle garden [
18].
The tablinum exhibits a sophisticated decorative scheme dating to ca. 50 AD, with typical Fourth Style wall paintings and a mosaic pavement in red, yellow, and black. The painted scheme is divided into three horizontal sections (registers) that exhibit red and yellow backgrounds, separated by a frieze and vertical decorated bands with black backgrounds. A continuous upper register extends around the three walls, with a lower register further subdivided into three rectangular panels on the east and west walls and a base register along the three walls (
Figure 2).
The decoration is painted over a complex plaster stratigraphy typical of Roman wall painting techniques (
Figure 3), beginning with a series of progressively finer plaster layers composed of a lime-based binder with local silicate-based volcanic sand as aggregate [
19], covered in the lower and upper registers by a fine finishing plaster composed of lime and marble dust. The finishing plaster on the upper register is a coarser preparation with less comminuted marble particles with respect to the lower registers. Furthermore, this final plaster layer is tinted pink superficially in the lower registers and the upper register of the south wall for an approximate thickness of 1 mm through the addition of a diluted admixture of red ochre pigment. Finally, the finishing plaster on the lower south wall contains crushed brick as aggregate [
18]. The red and yellow backgrounds were painted in fresco technique, while mezzofresco or secco techniques were employed for the black backgrounds [
20]. Materials analyses identified yellow ochre, red ochre, and carbon black as main background pigments. The frieze along the three walls and delicate architectural, floral, and figurative elements are painted in mezzofresco or secco technique, including high-quality figurative scenes imitating portable paintings that are centrally located in the panels of the lower registers. They are rectangular in the central panels and round in the side panels. The palette is typical of Roman paintings of the period, including pigments such as green earth, Egyptian blue, cinnabar, and lime white [
18].
Due to their superb quality, the wall paintings are some of the most celebrated at the site. They have, however, suffered severe deterioration both during the eruption and since excavation. It appears that the pyroclastic surge led to at least the partial collapse and subsequent detachment of the wall paintings, especially on the east and south walls, while the west wall was less affected by this phenomenon. The detached and fallen fragments were collected during the mid-20th century excavations [
17] and recomposed on backing support panels. Then, the panels were remounted in their original position on the tablinum walls [
18]. The walls were subjected to significant deterioration processes even after the excavation and restoration phases, which related to severe water infiltration from the damaged roof of the tablinum and led to the formation of superficial deposits of secondary phases [
18]. This phenomenon particularly affected the east and south walls.
Furthermore, as a result of the eruption, the wall paintings show chromatic changes, notably in large swaths of yellow monochrome background converted to red when exposed to the heat generated by hot volcanic material deposited during the eruption, and likely by fire from burning architectural elements. This phenomenon is particularly evident in the reattached pieces that show heterogeneous chromatic variations related to a difference in exposure to high temperatures following the eruption (
Figure 4).
This color shift significantly changed the appearance of the monochrome decorative scheme, which is thought to have been an alternating pattern of red and yellow fields of color in the panels of the lower registers, a yellow background on the upper register of all three walls, and red, yellow, and black bands along the base of the walls, as described by Maiuri in his description of the wall paintings [
17]. Presently, the color of the monochrome backgrounds is predominantly red; however, the central panel of the east wall exhibits a yellow background around the figurative scene, and there are areas of yellow in the upper east wall as well.
A stylistic study conducted on several domus of the Vesuvian sites showed many examples of alternating yellow, red, and black schemes, with a good example in the Casa di Marco Lucrezio Frontone at Pompeii [
13]. Furthermore, a number of other wall paintings at the site of Herculaneum show the same yellow to red conversion phenomenon, with notable examples in the Villa dei Papiri, in a room of the Quartiere Abitativo, and in the Casa del Rilievo di Telefo, where large areas of yellow have converted to red.
5. Discussion
The initial onsite survey campaign by p-XRF on the Roman wall paintings in the tablinum of the House of the Bicentenary made it possible to perform a first chemical characterization of the decorated surfaces. The main information obtained indicated a significant chemical differentiation between central panels, side panels, and upper registers according to Pb:Fe/As:Fe ratios calculated from adjusted peak areas (central panels: low both Pb:Fe and As:Fe; side panels high Pb:Fe and medium As:Fe; upper registers: highest Pb:Fe and low As:Fe). Considering both hypothesized and observed original color patterns, the interpretation of such results indicated that As constitutes a useful chemical tracer for the discrimination between original red and transformed yellow backgrounds in the tablinum of the House of the Bicentenary in accordance with Marcaida et al. [
14]. Furthermore, the similar Pb:Fe/As:Fe ratios of TY sites with those of OY sites in side panels and upper registers constitutes further strong evidence that these areas were actually yellow transformed to red due to the heat of the volcanic eruption.
It is not immediately clear why the three areas of monochrome background should show distinctly and consistently different patterns of lead and arsenic levels. It is possible that lead and arsenic are naturally occurring impurities in the ochres, considering their widely known tendency to be adsorbed through surface interactions onto several classes of crystalline compounds, especially phyllosilicates [
38,
39] and transition metal oxides [
40,
41]. Furthermore, both elements have been shown to be useful indicators of provenance in the investigation of ochre pigments [
30,
42,
43]. As for this case study, the lack of proper comprehensive databases makes it difficult to formulate solid hypotheses on the provenance of the employed ochres, also considering the possibility that arsenic- or lead-based pigments were intentionally added to the ochre pigments to modify their color. If so, then arsenic would likely have been added as realgar (red; AsS) or orpiment (yellow; As
2S
3) and lead would likely have been added as minium (red lead, defined as
Usta by Pliny; Pb
2O
3), litharge/massicot (yellow lead, defined as
Sandaraca by Pliny; PbO), or hydrocerussite (white lead, defined as
Cerussa by Pliny; (PbCO
3)
2·Pb(OH)
2), although the latter was known to be unstable for lime-based wall paintings.
Published levels of naturally occurring Pb and As in ochre pigments are typically in the ppm range [
30,
42,
43,
44], suggesting that the detected levels in the tablinum samples are unlikely to be naturally occurring, at least for Pb, which is present in significantly higher amounts than As. To further investigate this hypothesis, a selection of 10 modern ochre and sienna pigments supplied by Kremer Pigmente (Kremer Pigmente, Aichstetten, DE), ranging from yellow to red, were also analyzed by p-XRF using the same methodology applied onsite, though with longer collection time to decrease noise in the spectra (
Table S4). While some level of As was detected in eight of the 10 pigment samples, none yielded an As:Fe ratio greater than 0.002, far below the levels obtained in the side panels of the tablinum. Traces of lead were detected in only three of the ochre samples, but they were at levels so low that the Pb:Fe ratios calculated were 0.0003 or less and orders of magnitude lower than in the tablinum spectra.
Mock-up samples were prepared to make a rough estimate of how much arsenic- and/or lead-based pigment would have been added to the ochre pigment in order to produce lead and arsenic results similar to those found in the tablinum. Various mixtures of red ochre and yellow ochre pigments (with little to no lead or arsenic content) were prepared by adding either one part to ten or one part to twenty (by weight) of minium, litharge, and/or realgar. Pigment mixtures in a dilute acrylic binder were applied in a wash to a substrate of thick lime plaster to imitate the conditions on the walls of the tablinum. In the mock-ups as a whole, the addition of the arsenic- and lead-based pigments at a proportion of 1:20 by weight (significantly less by volume) made very little difference in the observable color of the final paint wash. A similar p-XRF instrument was used to acquire spectra from the mock-up samples and the spectra were processed using PyMca, in the same manner as the tablinum spectra, to generate As:Fe and Pb:Fe ratios. The results of this crude experiment suggest that lead levels found in the originally red side panels would likely correspond to substantially less than one part in 20 of red lead in red ochre by weight. Four measurements of the 1:20 mock-up yielded an average Pb:Fe ratio of 0.29 compared to the tablinum group 1 average of 0.10. Similarly, the As:Fe ratios found in the same tablinum panels would appear to correspond to a very minor addition of an arsenic-based pigment such as realgar; the 1:20 mock-up panel with realgar yielded an average As:Fe ratio of 0.27 compared to the tablinum group 1 average of 0.05. Mock-ups of litharge and red lead in yellow ochre pigment at 1:20 by weight yielded similar Pb:Fe ratios as the spectra collected from the tablinum’s originally yellow upper register (group 3). With a Pb:Fe ratio averaging about 0.13, the tablinum spectra fell between the litharge and red lead mock-ups, whose Pb:Fe ratios averaged 0.10 and 0.15, respectively.
In conclusion, the amount of lead and arsenic found by p-XRF in the background monochrome panels of the tablinum seems higher than expected for natural occurrence in ochre pigments, but lower than would be expected if lead or arsenic pigments were added to adjust the color of the paint. Of course, modern, industrially refined pigment mixtures are not the optimal point of comparison and may be significantly purer than ancient Roman varieties. Direct comparisons to a range of surviving examples of ancient roman ochres would be valuable.
The mineralogical information obtained through XRPD analyses on micro-scrapings constituted a perfect analytical bridge to link broad onsite chemical data by p-XRF with the punctual outputs of microstructural and microchemical analyses by SEM-EDS, µ-Raman, and EMP-WDS on cross-section samples. As a first result, the analyses indicated significant perturbations of the original mineralogical assemblage related to alteration processes during eruptive, burial, and post-burial phases, in accordance with the complex conservation history of the building. This is shown both by the presence of phases of clear secondary origin, related either to the formation of superficial concretions (gypsum, weddellite), dust deposition (volcanic minerals, dolomite), or to the post-depositional transformation of chromophore compounds (bernalite), and by the mineralogical evidence of thermally-induced transformations of ferric oxo-hydroxides constituting the chromophore minerals of ochre-based pigments.
In this perspective, the study made it possible to determine an absolute mineralogical parameter for the discrimination between original and converted red monochromes, namely the level of crystallinity of chromophore hematite determined by the Lorentzian size of crystallites, higher than 60 nm in the original red ochres and lower than 30 nm in the original yellow ochres altered to red. This clearly indicates that the impulsive heat stimulation related to the ingress of the pyroclastic surge in the building and the fire of wooden architectural elements at temperatures lower than 500 °C, which did not allow a crystal growth process comparable with long-term lateritic alterations of soils or prolonged high-temperature thermal treatments like those applied for the manufacture of artificial red ochres. In terms of archaeometric significance, the data were consistent in confirming the chromatic pattern of the registers hypothesized by archaeologists and conservators.
Furthermore, the mineralogical profiling of the decorated surfaces clearly indicated the occurrence of significant fractions of accessory phases in the applied ochre pigments, sometimes in ratios comparable with the chromophore phases. This is coherent with the intrinsic limits of ancient technologies for the purification of lateritic sediments, constituting a further confirmation that direct comparisons with modern, industrially refined pigments should be carefully evaluated. As for the nature of the accessory phases, besides the occurrence of common minerals such as quartz and phyllosilicates, all the samples showed the presence of peculiar compounds such as talc and opal-CT. The presence of talc could be related also to the accessory mineralogical assemblage of the ochre pigments. Nevertheless, its voluntary addition as an additive for superficial treatment cannot be excluded, possibly as a luster enhancer, as already suggested by Angelini et al. [
22]. As for opal-CT, the occurrence of the phase could be either related to the accessory mineralogical assemblage of the ochres or to superficial restoration treatments with ethyl silicate-based consolidants. Nevertheless, the second hypothesis is in contrast with published experimental evidence [
45], indicating the absence of significant diffraction contributions from these classes of products after polymerization. Furthermore, opal-CT is present also in the materials sampled from the atrium of the domus, where superficial consolidation treatments were not reported. Assuming its correlation with the original mineralogical assemblage of the pigments, the presence of opal-CT would constitute a significant marker of common provenance for the ochres, likely ascribable to silcretized horizons of lateritic sediments, whose occurrence has been reported in France [
46]. Furthermore, this would imply an extraction of the ore materials from sites where both red and yellow ochres were available, such as in the famous Roussillon quarries [
47].
Also, this theory would strengthen the hypothesis that the heterogeneous elemental distributions in the tablinum walls determined by p-XRF are not actually related to trace contaminations in ochres coming from different extraction sources but rather are related to the addition of heavy-metal-based pigments to the pigmenting mixtures. This would be valid for Pb, also considering the significant detected amounts of the element, while it may be less likely for As, which is present in small amounts, compatible with trace contamination in the poorly purified ochre pigments applied in the tablinum wall paintings. As for the discrepancy of concentration of As between originally red and originally yellow ochres, this would be justified considering both the higher As sorption affinity of hematite with respect to goethite and its lower desorption tendency from hematite substrates at high pH levels, typical for lime-based binders [
48,
49]. Nevertheless, the hypothesis of the utilization of pigmenting mixtures containing both ochres and lead-based pigments contrasts with the fact that no traces of crystalline lead-based compounds were found in any of the samples analyzed, thus requiring a careful crosscheck with the results obtained from detailed microstructural and microchemical analyses on the cross-section samples analyzed.
As a first result, the multi-analytical microstructural characterization of the tablinum cross-section samples made it possible to confirm and further characterize the Pb chemical zoning detected by p-XRF. The element is absent in the central panels, while it is present in reduced amounts on the side panels and in higher concentrations in the upper registers, with the highest concentration of approximately 3 wt% in the upper register of the west wall. The element is homogeneously distributed in the paint layer matrices, regardless of iron concentrations, apart from slight increases in the chromophore-rich pigment lumps of one sample from the upper east register. Furthermore, it shows a certain degree of diffusion in the uppermost portions of the finishing plasters of the side panels and upper registers, and it is present in reduced amounts in the macroscopic pigment lump within the pink-tinted plaster of the east-central panel sample. Such ubiquitous Pb distribution is to be considered even more peculiar when compared with the varying compositional profiles of major elements in the analyzed paint layers, not correlated with any architectural register or the related ochre type. These are characterized by variable Si/Al ratios, with excess Al in some areas suggesting the occurrence of amorphous aluminum oxo-hydroxides as accessory phases, fairly common in lateritic sediments [
46], in association with phyllosilicates, quartz, and opal-CT, already determined by XRPD.
As a second significant result, a marked microstructural and qualitative difference of the painted plaster sequences applied in the three areas of the tablinum walls was observed, whose main characteristics are reported hereafter.
Base register. The ochre-based background pigment is incorporated in fresco technique within the uppermost volcanic sand-based preparatory plaster.
Lower register. The ochre-based background pigments are applied in fresco technique over a fine finishing plaster made with lime and marble dust aggregate. This finishing plaster is pink-tinted in the uppermost portions through the addition of a diluted red ochre fraction, to confer a homogeneous warm tone to the plaster. Over this pink-tinted layer, ochre particles are homogeneously dispersed in the supporting plaster forming a thin and compact paint layer, and they are characterized by a reduced particle size distribution. The black background on the lower register of the south wall is applied likely with a binder over a lime-based plaster containing crushed bricks as aggregate and with no pink-tinted plaster beneath.
Upper register. The ochre-based background pigment is applied in fresco technique over a coarser finishing plaster made with lime and marble dust aggregate with higher particle size distribution with respect to the lower registers. The pink-tinted plaster is absent on the east and west walls, while it is present on the south wall, and the paint layer is thicker, more irregular, and less homogeneous, with fractured and dislocated skins of carbonated lime forming pockets filled with poorly cemented pigment particles. These are characterized by significantly higher particle size distribution with respect to their counterparts in the lower registers.
The varying characteristics suggest greater care of the artisans in the preparation of the different areas. This is perfectly justified by the higher functional and artistic value of the lower registers of east and west walls, more visible and decorated with high-quality figurative scenes. This analytical evidence may imply either lower care of the painters in the preparation of the less important areas, as an efficiency measure, or the execution of such portions by less-skilled artisans. From this perspective, it is possible to hypothesize the availability in the painters’ palette of ochre batches with different grades of quality, related to differing post-extraction treatments such as decantation and comminution. Consequently, the ochres should have been characterized by different pigmenting properties that are obviously influenced by their fineness (the higher, the deeper, and homogeneous the color hue). To correct the hues during the application of the paint layer and guarantee an overall chromatic consistency, the artisans may have applied both simple proportioning procedures such as increasing the amount of applied pigment, as observed in the samples from the east upper register, and express variations of the pigment mixtures by the addition of specific compounds to the base ochre. Besides the hypothetical application of talc as a superficial additive, microanalyses highlighted the diffuse addition of carbon black pigment to the earth-based compounds, previously reported in scientific literature [
15] and possibly applied as an opacifier of the fairly transparent ochre pigments.
This widespread utilization of additives further suggests the hypothesis of the application of lead-based pigments in the wall paintings of the tablinum. Such compounds are characterized by bright and deep colors, close to those of ochres and lime white, but they are denser, more finely textured, and with better hiding power [
50]. Furthermore, they were artificially obtained since ancient times through simple chemical processing of scrap metals in acidic solutions, to obtain lead white, and subsequent calcination of lead white, to obtain red and yellow lead oxide pigments [
51], thus representing a readily available substitute and/or adulterant for earth-based pigments.
The first route of interpretation to justify the occurrence of lead-based compounds in the monochrome backgrounds of the tablinum could be that the red pigments used in the side panels were not pure ochres but artificial mixtures of ochre and lead white calcined together to obtain the
Sandyx (mixture of white lead and red ochre, then roasted) and
Syricum (mixture of
Sandyx and red ochre) pigments cited by Pliny [
3] (pp. 283–295). This would not, however, justify the presence of lead in the upper registers, which were originally yellow-tinted. A second hypothesis would be the utilization of either red (minium/red lead) or yellow lead oxides (litharge, massicot) as express color/hue correctors added to the earth-based pigments to confer chromatic homogeneity to the monochrome backgrounds when lower-quality ochre batches were used. Finally, a third hypothesis would be the utilization of lead white pigment as an express whitening agent, either before or in concomitance with the ochre application and in areas where the lower quality of finishing plasters impaired the desired aesthetic effect.
Nevertheless, these theories are not supported by the fact that no traces of crystalline lead-based compounds were found in any of the analyzed samples. Furthermore, lead is generally present in low amounts, so it raises the question if such reduced additions would be capable of conferring specific chromatic properties to the treated surfaces. Nonetheless, both issues may be justified considering the well-known tendency of lead-based pigments to degrade and dissolve under the influence of oxido-reductive and pH-dependent transformation processes [
51].
As for lead white, the strong instability of hydrocerussite in alkaline conditions typical for lime-based setting plasters is well-known [
52,
53,
54]. Awareness of this issue was clearly demonstrated by Pliny in his
Naturalis Historia, including
Cerussa in the list of pigments that are not suitable for wet-on-wet applications [
3] (pp. 297–299).
Other plausible degradation processes are the temperature-related ones [
55,
56]. Under standard conditions, thermal decomposition of lead oxides results in a progressive reduction of lead that is associated to a decrease in oxygen content. Such processes occur in the temperature range 330–570 °C, with the first conversion of plattnerite into minium through a series of stoichiometry intermediates, at temperatures around 375 °C, and the subsequent conversion of minium into the reduced forms, either into the low-temperature polymorph litharge stable at heating up to 490–540 °C or the high-temperature massicot. Furthermore, the transformation of lead monoxide into minium can be thermally induced both through oxidative heating of litharge at temperatures higher than 450 °C and through reverse oxidation of massicot at very slow cooling below 450 °C after the thermal decomposition of lead dioxide. Finally, the thermal decomposition of lead carbonates, both anhydrous (cerussite; PbCO
3) and hydrated (hydrocerussite), starts at temperatures below 300 °C with the formation of distorted forms of α-PbO and subsequent conversion into the most stable lead monoxide polymorphs. All these transformation processes are widely within the temperature ranges that would have been subjected to the painted plasters of the House of the Bicentenary when struck by the pyroclastic surge of Mount Vesuvius eruption. Therefore, they may have concurred, together with the goethite to hematite transformation, to the yellow to the red transition of several portions of the monochromes. Nevertheless, they would still not justify the total absence of crystalline lead compounds and the low aliquots of lead detected in several areas of the wall.
Another less known alteration phenomenon involving lead oxide-based pigments is their whitening/bleaching [
57,
58]. This essentially consists of solution-mediated carbonation and sulfatation processes of lead oxides, preceded for minium by a stage of photochemical reduction of Pb(IV) to Pb(II), causing a progressive discoloration of the chromophore phases after transformation into colorless compounds. As for lime-based mural paintings, carbonation occurs after progressive uptake of atmospheric CO
2 in pore solution, both in concomitance and after binder reaction, with the formation of hydrocerussite through the metastable carbonate-deficient compound plumbonacrite (3PbCO
3·Pb(OH)
2·PbO). In the same way, sulfatation could occur with the formation of anglesite (PbSO
4) in the presence of sulfate-rich pore solutions. It is plausible to hypothesize a strong influence of both processes in the alteration of the analyzed plasters, which were subjected before and after burial to abundant percolation of carbonate and sulfate-rich solutions related to coastal proximity, deposits from the eruption of Mount Vesuvius, and post-excavation pollution factors, as demonstrated by the abundant occurrence of secondary gypsum in the Herculaneum wall paintings. Furthermore, given the net solubility increase of lead-based compounds after transformation from oxides to carbonates [
53] and/or sulfates [
59], especially in the typical pH/Eh conditions and pore solutions chemical composition of lime-based plasters, it is plausible to hypothesize a progressive dissolution and leaching of the alteration products during burial conditions.
Such transformation processes would justify both the absence of crystalline lead-bearing phases and the lead depletion in the paint layer matrices of the tablinum, with partial percolation of the element in the uppermost portions of the finishing plasters and preferential re-adsorption of the free ions by the compounds with a higher chemical affinity toward the heavy metal, such as iron oxo-hydroxides. Furthermore, this hypothesis would be consistent with the higher Pb fractions found in the sample from the upper west register, less affected by collapses during eruption and alteration processes during burial and post-burial phases with respect to the analyzed counterpart in the east wall.
6. Conclusions
The analytical campaign on the painted surfaces in the tablinum of the House of Bicentenary made it possible to answer the main archaeometric questions formulated at the beginning of the study, namely, (a) to determine the original decorative scheme of the red and yellow-painted backgrounds of the tablinum, heated by the pyroclastic surge of the eruption, and (b) to characterize the compositional and textural characteristics of the painted plaster sequences, determining the nature of the ochre-based pigments used for the fresco-painted monochrome backgrounds.
As for the first question, the results obtained made it possible to reconstruct the original color scheme of the monochrome backgrounds, heavily perturbed by chromatic transformations during and following the eruption. This was achieved by applying a chemical discrimination methodology based on the determination of As and Pb levels through p-XRF. Furthermore, it has been proven that observations of hematite crystal size through Rietveld refinement on XRPD data constitutes a reliable mineralogical parameter to distinguish between natural hematite and hematite thermally transformed from goethite by the heat of the pyroclastic surge. These results show that the decorative scheme in this room, and potentially many other decorative schemes in the Vesuvian region, were markedly different than they appear today, with the popular “Pompeiian red,” in some cases, actually being yellow.
As for the characterization of the painted plaster sequences, the experimental results demonstrated marked differences in the quality of the preparations, showing greater care in the execution of the panels in the lower registers both in terms of applied materials and manufacturing techniques. This is coherent with the higher visibility and artistic value of these portions of the walls, confirming once more a structured organization of work by Roman artisans, respectful of both aesthetic and economic factors. As for the pigment mixtures applied, the analytical results indicated that the ochre-based pigments constituting the main raw materials for the execution of the famous Roman monochrome backgrounds were available in different grades of quality, which were mainly related to their fineness. Such differentiation must have been obtained through post-treatment processes such as decantation and comminution of raw materials possibly related to common extraction sources if considering opal-CT a valid mineralogical tracer not related to modern superficial restoration treatments. Furthermore, they may not have been applied pure, but rather mixed with a variety of additives to enhance specific properties of the painted surfaces, potentially varying from those reported in Roman technical treaties. If some of these additions have already been reported in the literature, such as the combined utilization of ochre and carbon black, other hypothesized treatments constitute a novelty in this field of archaeometric research, such as the superficial addition of talc.
The widespread use of additives and pigment mixtures in Roman painting techniques leads to alternative routes of interpretation of the observed heavy metal chemical zoning in the different registers of the tablinum. While the presence of As in the original red of the side panels could be related to the original chemical profile of the red ochre, considering both the low level of depuration of ancient pigments compared to modern, more highly refined pigment manufacturing processes, and the higher sorption affinity of hematite to the element, the Pb levels seen in the upper register appear to be higher than expected for trace contaminations in the lateritic sediments. One hypothesis is that lower-quality ochres were mixed with lead oxide-based pigments to compensate for the lower-quality and/or perhaps to correct or enhance the hue. The availability of hybrid mixes obtained by calcination of red ochre with lead white was reported in the Roman world, in the description of the Sandyx and Syricum pigments in Pliny’s Naturalis Historia, but there are no written sources reporting the practice of correcting the color hue of ochres during the execution of fresco decorations by the express addition of lead oxide pigments, either minium, litharge, or massicot, so this hypothesis would constitute a novel finding in the definition of Roman painting technology. Similarly, the addition of lead white during the execution of fresco decorations as a whitening agent to compensate for the poor chromatic aspect of lower-quality finishing plasters has not been recorded, and Pliny stated that it should not be used in fresco applications due to its instability with lime-based plasters, so this hypothesis would indicate the flexibility of Roman artisans in following the indications of the technical treatises in common decorative practice.
Given the low detected fractions of Pb in most of the analyzed samples and the total absence of crystalline lead oxides, it was not possible to confirm these hypotheses, particularly given the well-known instability of lead pigments in mural paintings, further exacerbated by the marked deterioration processes that affected the tablinum since the 79 AD eruption of Mount Vesuvius. Therefore, the presence of Pb detected in the wall paintings of the tablinum of the House of the Bicentenary raises more questions than answers at this point. Further research in the field, through comparisons with other case studies and analysis of Roman pigment samples is warranted to shed further light on this relevant archaeometric question.
As a concluding remark, the phased multi-scale and multi-analytical approach adopted, starting with non-destructive tests onsite and complemented by detailed laboratory investigations on a limited number of samples to preserve the high cultural value of the analyzed context, made it possible to overcome intrinsic difficulties of interpretation related to the varying information obtained through single analytical techniques, leading to a comprehensive and integrated interpretation of the archaeometric issues.