Experimental Study of Chalconatronite: From Its Identification to the Treatment of Copper Alloy Objects

Abstract

On the occasion of the reopening of the Dobrée Museum (Nantes, France), two statuettes of Egyptian origin, representing Harpocrate and Isis, were studied to shed light on the presence of the blue-green efflorescence on their surface. The efflorescence on the Harpocrate statuette was identified as being chalconatronite, while that which was present on the Isis statuette corresponded to sodium copper formate/acetate, probably due to the evolution of chalconatronite in an environment containing VOCs. The efflorescence appeared to be sensitive to the cyclic variation in relative humidity whereas it seemed stable. An experimental curative treatment to halt the reappearance was carried out. A series of pure water baths extracted a significant quantity of sodium. The treatment appeared effective and reduced the risk of a recrudescence of the efflorescence for both statuettes. However, when the efflorescence was dissolved on the Isis statuette, other compounds appeared to react with water, leading to acidification and a potential reaction with the lead in the alloy. A layer of lead carbonate/acetate on the surface appeared. The objects were then dried and protected with a highly concentrated acrylic varnish. They are currently being monitored to identify any new efflorescence that may appear during display.

1. Introduction

Active corrosion of copper alloys, commonly known as “bronze disease”, is frequently encountered by heritage conservators and restorers [1,2,3]. Formed by the presence of chlorides, this type of corrosion has a light green powdery appearance and a waxy consistency. Although problematic, its reactivation and treatment are increasingly well understood. However, another type of powdery corrosion appears cyclically and is often confused with bronze disease. It is described as blue-green or pale blue to greenish blue [4]. This type of efflorescence was first observed on Egyptian antiquities and bronze heritage artifacts as a result of the corrosion of copper alloys [5]. However, their presence extends to a wider range of objects [6,7,8]. This corrosion product is related to a hydrated double carbonate of copper and sodium, with the semi-developed formula Na₂Cu(CO₃)₂, 3H₂O, known as chalconatronite (or trihydrated sodium dicarbonatocuprate) [5,9,10,11]. First synthesized in 1852, chalconatronite was identified in 1955 as a natural corrosion product affecting certain buried Egyptian bronze antiquities [5]. The existing literature deals with the identification of chalconatronite, but few studies have examined its physicochemical and crystallographic properties. In 1978, A. Mosset et al. described its crystal lattice as a stack of sheets—formed by NaO₆ octahedra—parallel to the (010) plane; the connection between the sheets is ensured by Cu-CO₃-Cu chains. Each metal atom is linked to three carbonate groups (Figure 1) [12]. This compound grows preferentially in environments where specific conditions are met: very arid soil and in the presence of alkali carbonates like natron (Na₂CO₃, 10H₂O) or trona (Na₂CO₃, H₂O) [4,5,13]. In 1981, C.V. Horie and J.A. Vint observed this product on a collection of objects from the Roman site of the city of Chester (Northwest England) [14]. Following X-ray diffraction (XRD) analysis, sodium dicarbonatocuprate tri-hydrate was identified with a crystal structure identical to that of chalconatronite [15]. As these objects have no direct connection with Egypt or an environment similar to that of its “natural” formation, the authors assume that its formation has been caused by sodium sesquicarbonate (Na₂CO₃/NaHCO₃), traditionally used to treat copper alloy objects [1,16,17,18,19]. This corroborates the information given by R.J. Gettens and C. Frondel [5]. It would seem that the formation of chalconatronite originates in a medium rich in sodium hydrogen carbonate. The same authors put forward the hypothesis that a sesquicarbonate treatment may partially dissolve the copper corrosion products, leading to the re-deposition in cracks or on the surface of copper salts in the form of small crystals of hydrated sodium cupro-carbonate, identified as chalconatronite. Thus, the presence of chalconatronite does not necessarily depend on exposure to a desert environment but can provide valuable information on the nature of the environment and the position of the object during the corrosion phase [20,21]. It has also been specified that the formation of its filiform structure is only possible in air or water with a sufficient supply of sodium carbonate. In addition, chalconatronite has also been identified on composite objects combining copper alloys and glass [22,23]. Due to the high concentration of alkaline salts such as sodium, the development of a blue-green efflorescence on glass-associated objects does not appear to be uncommon. Indeed, during the hydrolysis of alkaline glass surfaces, carbon dioxide is absorbed from the air, forming lead and copper carbonate compounds. In the literature, it is reported that chalconatronite originates from a reaction between water and alkaline carbonates with copper alloys or in the presence of intermediate copper corrosion products such as malachite or atacamite. However, R.J. Gettens and C. Frondel refer to a direct reaction with a copper alloy [5,24]. Recently A. Fischer et al. have specified that this compound is the only stable copper phase after the drying of sodium sesquicarbonate solutions in the presence of Cu²⁺ [25,26,27]. The evolution of the carbonate concentration thus favors the formation of the bis(carbonate)cuprate(II) ion, Cu(CO₃)²⁻, resulting in the formation of chalconatronite when there is excess sodium. The work of K. Leyssens et al. has shown that a 5–10% w/v concentration of sesquicarbonate leads to the formation of chalconatronite [28].

In certain conditions, chalconatronite may evolve into other compounds. K. Trentelmann has identified a new pale blue corrosion product containing sodium, copper and formate [29]. Analyses of several hundred objects led A. Fischer et al. to determine the above compound as copper formate hydroxide oxide hydrate, formerly well known as sodium copper formate acetate or socoformacite [22]. The compound was characterized by a crystal structure determination; there is no acetate present in the crystal lattice, but it might occur as contamination. Other studies have focused on this type of efflorescence, in addition to that composed of chalconatronite [30,31]. In the case of composite objects (glass–bronze), the degradation of the glass component results in a reactive medium for the formation of formates from formaldehyde (H₂CO) due to the Cannizzaro reaction [22]. Although this reaction is considered very slow, half of the 250 objects analyzed from the same collection showed corrosion products based on sodium copper formate (Cu₂(OH)₃HCOO), in addition to chalconatronite, sodium copper acetate carbonate (NaCu(CO₃)(CHCO₂)) and others. The formation of formates and acetates on copper alloy objects can be attributed to their being stored or displayed in wood or materials containing wood products [32,33,34]. However, these compounds appear to be strongly linked to the initial presence of chalconatronite. Tests carried out on synthetic chalconatronite led to the formation of sodium copper formate when exposed to formic acid and a formaldehyde environment. Copper coupons moistened with a 1 mol/L sodium hydroxide solution showed that chalconatronite formation was followed by formate-based compounds after 6 months of exposure (in 4 ppm of formaldehyde at 75% RH and 3 L of air per minute) [25]. In the case of glass–bronze composites, the consequences of this corrosion may be irreversible since it is necessary to consume sodium, which is the constituent element of glass. On the other hand, the question remains relevant for objects where sodium comes from an exogenous environment (a burial medium or conservation treatment). Copper corrosion products evolve into chalconatronite, but has the metallic core been consumed? Only F. El Baz, R.J. Gettens and C. Frondel have introduced this question [5,24].

In 2022, the Arc’Antique laboratory was sent two bronze archaeological Egyptian statuettes (the Harpocrate statuette and the Isis statuette) from the Dobrée Museum (Nantes, France) displaying a pale blue-green powdery efflorescence. The Harpocrate statuette is a regular visitor to the laboratory, where it has been treated cyclically since 1996 according to the following protocol: removal of corrosion using a scalpel or micro-sandblasting, followed by protection with an acrylic varnish (5% w/w Paraloid B72^® in acetone). The decision to display them in the museum’s future permanent exhibition alongside artifacts made of organic materials caused us to seek to stabilize these powdery remnants in a more durable way. This choice of exhibition environment entails a number of constraints since the relative humidity is due to be calibrated at 55% to ensure the preservation of organic materials, and these two artifacts are required to have a similar appearance to the other bronzes in the display case (the color of patina and brilliance). Once again, recurring corrosion prompted us to find an ad hoc treatment solution. In this study, we will investigate the reactivation of the corrosion, particularly with regard to future display conditions, and we will seek to propose a treatment protocol based on the extraction of salts. The limitations of this treatment protocol, particularly in terms of monitoring, will also be discussed.

2. Materials and Methods

2.1. Corpus

2.1.1. Egyptian Objects

Two Egyptian copper alloy objects from the Dobrée Museum collection (Mediterranean archaeology) were studied: statuettes of Harpocrate and Isis. The history of these objects remains relatively obscure (Figure 2). Little information is available on the Musée Dobrée inventory and the Micromusée database. Owned by French explorer Frédéric Caillaud (1787–1869) between 1814 and 1822, these artifacts were bequeathed to the museum in 1869. The first traces of their ‘museum life’ date back to 2008, when they were loaned out for exhibitions, one of which seems to have caused a major setback in terms of corrosion, although the Harpocrate statuette had already been treated by our services back in 1996. Apart from that, we have no information on either exhibition or storage conditions, except for the relative humidity, which does not seem ever to have been controlled.

2.1.2. Synthesized Chalconatronite

To investigate the dissolution of chalconatronite, and in view of the small quantities available on the objects, we tried to purchase synthetic chalconatronite or obtain it from various institutions. Unfortunately, all requests to mineralogical, geological and gemological institutions, as well as to museums, proved unsuccessful. The only solution was, therefore, to synthesize chalconatronite. A.M. Pollard et al. studied the synthesis of georgeite, a copper carbonate [26,27]. In this synthesis process, the formation of malachite and chalconatronite is observed when certain parameters are modified. A.K. Sengupta and A.K. Nandi described a synthesis of chalconatronite with copper (II) acetate and a solution of hydrogenate carbonate and sodium carbonate [35]. The synthesized chalconatronite was produced according to the protocol defined by G. Eggert et al. [9]: 20 g of anhydrous sodium carbonate is mixed with 50 g of sodium hydrogen carbonate in 700 g of cold water. A total of 8 g of copper sulfate, dissolved in 20 g of water, is added. The mixture was stirred vigorously and filtered overnight through a Büchner flask.

2.2. Analytical Techniques

2.2.1. Scanning Electron Microscopy Coupled to Energy Dispersive X-ray Spectroscopy (SEM-EDS)

Energy dispersive spectrometry (EDS) coupled to a scanning electron microscope (SEM) was used for the elemental identification of the micro-samples (of the order of a few micrograms). The SEM-EDS is a JEOL^® model 5800 LV. The current was 300 nA, and the voltage was 20 kV. The samples were taken with a scalpel by a conservator under a binocular magnifier and then deposited on a conductive double-sided carbon tape on an aluminum support. A carbon coater was then used to ensure electron conduction, enabling an analysis of the sample. Spectra were acquired and processed using IDFix software. The resulting spectra displayed peaks that are characteristic of the elements detected.

2.2.2. X-ray Fluorescence

X-ray fluorescence (XRF) spectra were recorded with a Tracer III-SD pXRF spectrometer (Bruker^®, Am Studio 2d, 12489 Berlin, Germany), with the following parameters: 40 kV, 13 µA, no filter, Spectral Mode, 60 sec and a window aperture of 3 × 4 mm. The spectra were processed with Bruker 7.4 software. Each peak was characteristic of the relevant elements. The detectable elements range from aluminum (Al) to bismuth (Bi), and the analysis is purely qualitative. Artifacts are produced for rhodium (Rh) and palladium (Pd) from the X-ray tube, nickel (Ni) from the window used for analysis, titanium (Ti) from the window of the tube emitting the primary X-rays and argon (Ar) from the ambient environment of the analysis.

2.2.3. Raman Spectroscopy

Raman spectra were acquired with a Renishaw InVia^® (15 rue Albert Einstein Champs sur Marne 77447 Marne la Vallée Cedex 2), with the following parameters: 514 nm laser, 100–1400 cm⁻¹ and 2900–3800 cm⁻¹ ranges, 100x objective and from 1 to 10% of laser power (0.15 mW to 1.5 mW at the surface of the sample). The time acquisitions were 6 accumulations of 10 s for each spectrum.

2.2.4. X-ray Diffraction

Structural analysis by X-ray diffraction (XRD) was conducted to characterize the synthetized chalconatronite. The measurements were carried out on a Bruker^® D8 ADVANCE using the 1.5406 Å wavelength of the Kα line of Cu. Measurements lasted for 2 h, between 5 and 60°, with a resolution of 0.016°. XRD powder samples are prepared by manual grinding using a mortar and pestle. A few milligrams of powder are placed in a sample holder and then gently pressed against the sample holder with a glass slide. The surface of the powder should be smooth. Excess powder is removed from the edges of the sample holder and then carefully placed in the XRD. Data analysis was carried out using EVA 9.0 software with the PDF-2004 database.

2.2.5. Infrared Spectroscopy

Infrared analytical analyses were carried out using attenuated total reflectance (ATR) equipped with a diamond crystal and mounted on a Bruker^® Alpha spectrometer (Am Studio 2d, 12489 Berlin, Germany). Spectra were recorded with the following settings: 100 scans per measurement with a spectral resolution of 4 cm⁻¹ and a wavenumber range of between 400 and 4000 cm⁻¹. The samples were pressed against the crystal with a built-in pressure applicator. Spectra were managed using OPUS 7.5 software (Bruker Optics).

2.2.6. Conductivity, pH and Temperature Measurements

The pH was measured using a pH-education PHM210 pH meter (Radiometer Analytical^®, 72, rue d’Alsace 69627 Villeurbanne France) and a pH electrode (Electrod-Swiss^®, Grosseron, 4 Rue des Entrepreneurs, 44220 Couëron France) with a relative margin of error of 5%. Conductivity and temperature were measured using a VWR WTW 330i conductivity meter and a TetraCon^® 325–325/C combination cell (VWR, Europarc, 26 avenue Leonard de Vinci, 33608 Pessac cedex). Conductivity values were corrected according to the temperature, with a relative error of 2.5%. Three conductivity values were compared: the first was measured by the device; then, the theoretical values were calculated from the dosed sodium concentration, taking only the sodium in the solution into consideration; and the final value, integrating the presence of sodium (Na) and hydrogen carbonate ion (HCO₃⁻). The calculations are based on this equation σ = [X1]xλ1 + [X2]xλ2 + [X3]xλ3 + … + [**]xλi; with σ being the conductivity of the solution (µS.cm⁻¹); [**] being the molar concentration of the ions (mol.cm⁻³) determined by the sodium titration and, λi being the ionic molar conductivity of the ions ** (µS.cm².mol⁻¹), i.e., for sodium (5.01 µS.cm².mol⁻¹) and hydrogen carbonate ion (4.45 µS.cm².mol⁻¹).

2.2.7. Sodium Titration by Atomic Absorption Spectrometry

Sodium quantification was carried out on an atomic absorption spectrometer (AAS) equipped with a conventional slit burner head for an air–acetylene flame (flow: 10 L/min for air and 1.1 L/min for acetylene) and a hollow cathode lamp for sodium. The absorbance values were taken following an average of a few seconds of investigation time. These values were reported on a calibration curve obtained using a commercial standard solution of 1000 ppm of sodium (Plasmanorm, Prolabo^®, VWR International, France, Europarc, 26 avenue Leonard de Vinci, 33608 Pessac cedex). The standard range of 0–1 mg/L results in a linear function. The relative error for this analysis, which was estimated at 5%, was based on the disparity between the sample preparations and the standard results.

3. Results

3.1. Analysis of Corrosion Products

The XRF analyses of the Egyptian copper alloy objects determined their basic composition to be lead bronze. Observations of the objects revealed the presence of blue-green corrosion products in the form of efflorescence. The SEM-EDS analyses indicated the presence of copper (Cu) and sodium (Na). The Raman spectroscopic results of the Harpocrate and Isis statuettes are presented in Figure 3 and Figure 4. The bands at 327 cm⁻¹ (CuO stretch) and those at 1053 and 1072 cm⁻¹ (symmetric carbonate stretch) on the spectrum of the Harpocrate statuette correspond to those of chalconatronite synthesized according to the instructions published by M. Gröger [25,26,36,37]. The corrosion products on the Harpocrate statuette were therefore identified as chalconatronite.

However, the Raman results for the Isis statuette were different (Figure 4). The spectrum shows bands at 789 cm⁻¹ (formate OCO in-plane bend), 1066 cm⁻¹ (formate CH out-of-plane bend), 1354 cm⁻¹ sh (formate OCO sym. Stretch) and 1370 cm⁻¹ (formate CH in-plane bend), which are characteristic for formates. A weak band at 940 cm⁻¹ (acetate C-C sym. Stretch) indicates the presence of acetate [29]. The acetate symmetric CH3 deformation mode appears near 1350 cm⁻¹, but it can be obscured by the formate OCO symmetric stretch. Other bands, observed at 444 and 308 cm⁻¹, were similarly assigned to the copper–oxygen vibrational modes associated with a carboxylate. The bands at 148, 174 and 256 cm⁻¹ were assigned to vibrations of the carboxylate lattice. Moreover, copper–oxygen vibrations appeared in the range of 300–430 cm⁻¹ for copper formate and 320–380 cm⁻¹ for copper acetate [29,38]. All of this indicates that the blue corrosion product observed on the Isis statuette corresponds to sodium copper formate or acetate rather than chalconatronite.

The same corrosion products have been identified in several collections and are not always of Egyptian origin but may be Greek, Chinese, etc. [10,22,29,32,39]. K. Trentelman et al. specify that two parameters are necessary for the formation of these compounds: a source of sodium and a source of organic acids [29]. They also point out that the ratio of sodium to copper (1:1) is a key factor in the formation of these compounds. A. Fischer et al. experimented with the formation of sodium copper acetates/formates from chalconatronite [22,25]. The latter is transformed into sodium copper formate when exposed to formic acid or formaldehyde in a humid environment. Tests carried out on the copper coupons, moistened with 1 mol/l of soda and exposed to 4% formaldehyde in 75% and 85% RH and 3L air/min for 6 months, detected the presence of chalconatronite rather than formate. The most likely process would, therefore, be the formation of chalconatronite from malachite, which, in the presence of organic vapors and with sufficient sodium, evolves towards sodium copper carboxylate compounds. The environment in which objects are stored and exhibited is a common issue. Indeed, all the authors agree that the appearance of these carboxylates is due to exposure to organic acid vapors, particularly acetic, formic and formaldehyde acids [10,31,38]. The origin of acetic acid is often associated with the presence of wood, such as the wood in storage boxes. The environment in which objects are stored and exhibited becomes a common issue. Copper alloys interact with these vapors to transform and ultimately form acetate- and/or formate-based compounds. Unfortunately, our lack of information on the storage conditions of the Harpocrate and Isis statuettes means that we are unable to make any kind of correlation.

3.2. Stability of the Objects

The two statuettes in the collection are currently stored in the Dobrée Museum’s storerooms and are due to be displayed in 2024 in a display case that will also include wooden objects. This raises issues of both environmental recommendations and the reactivation of efflorescence. To observe this, tests were carried out to see what relative humidity conditions the efflorescence reappears under. To assess the stability of the objects, the surface was first brushed to remove any efflorescence, then placed in a hermetically sealed enclosure at high relative humidity (around 90%) to observe the oxygen consumption. If oxygen is consumed, we can deduce that there is a real risk of active corrosion. If not, these salts can be assimilated to efflorescence, whose crystallization depends on hygrometric variations. In one month, only 0.55% per day of oxygen was consumed, which was not considered significant, and no efflorescence appeared. These data can be compared with those for copper alloy objects suffering from bronze disease (with active chloride corrosion) [40]: the oxygen consumption of active objects was 15% per day, while a less reactive object consumed around 4% per day, and an object protected with a varnish layer consumed 1% per day. In view of the non-evolution of the salts, a second test was carried out by applying a cyclic relative humidity variation to these objects. This experiment was carried out simply by placing the artifacts alternatively in one of two storerooms for a three-day period, one storeroom with 20–25% relative humidity and the second with 53–58% (Figure 5). After one month, the macroscopic observations (DMS1000, LEICA microsystems^®, (Société française de Microscopie, CS 9605–28130 Pierres, France) revealed a recrudescence of blue-green efflorescence, leading to the migration of salts to the surface in previously observed areas. These tests confirmed that the salts are confined to the interior of the object and that a variation in relative humidity in the air brings them to the surface. Because no conservation treatment by sesquicarbonate was carried out on these objects, it would also seem to confirm that these salts may have originated in the burial soil. The reactivation of the efflorescence took place in relative humidity conditions similar to the future display case settings, hence the need to define a curative treatment.

3.3. Synthesis of the Chalconatronite

The synthesized powder was analyzed by Raman and X-ray diffraction (XRD) and compared with the literature reference spectra (Appendix A, Figure A1 and Figure A2). These were characteristic of chalconatronite (PDF-832253 of PDF-2004 database) [5,25,32,36]. A second synthesis was carried out in 2023 with a lower yield. Raman spectra were compared. Although the same bands were identified, confirming the formation of chalconatronite, the spectra differ according to the orientation of the electric field applied during analysis. This can be seen in the spectra by the height changes of the characteristic bands. In addition, the low-frequency bands have shifted. This may indicate that the product may have different geometric structures. A. Mosset et al. studied the crystal structure of synthetic chalconatronite, revealing different geometries depending on the carbonate groups involved [12].

3.4. Dissolution of the Chalconatronite

Previous publications indicate that chalconatronite is soluble and/or partially soluble in water and completely soluble in cold dilute acids [5,25]. Some authors even indicate that for objects with this type of efflorescence, rinsing with water is sufficient [26,32]. Because no physicochemical data were available on the solubility constant of this compound, we decided to carry out dissolution tests to verify the solubility of chalconatronite. A test was carried out on the chalconatronite samples taken from the synthesis. The synthetic powder was dissolved in pure water. The solution was stirred overnight and then filtered by a Büchner flask. The filtered solution was then prepared and analyzed for sodium ions by atomic absorption spectroscopy (AAS). The aim was to quantify the sodium dissolved in the water and compare it with the sodium theoretically present in the chalconatronite as a function of its chemical formulation. In principle, sodium represents ~16% of the total mass of the chalconatronite molecule. A total of 0.10 g of synthetic powder was dissolved in 0.05 L of pure water, i.e., 2.00 g/L. The sodium concentration should, therefore, correspond to 0.33 g/L. The AAS assay detected 0.36 g/L. This experiment was repeated once, yielding the same values. This seems to indicate that the sodium has dissolved in the solution and that the water is indeed dissolving a part of the product (copper carbonate being insoluble in water).

3.5. Curative Treatment

In practice, and notably for the Harpocrate statuette, conservation treatment by mechanical action and varnish-based protection were not sufficient to limit new efflorescence. While preventive action may not be effective, conservation treatment could put a stop to efflorescence by eliminating sodium—the primary cause of the formation of the corrosion product. Given the solubility of chalconatronite in water, it was decided to immerse the artifacts in pure water [5,25,32]. The purpose of the pure water (and not deionized water) was to limit the external contribution of sodium and thus optimize the measurement and evaluation of the concentration of sodium coming from the object (and not coming from the solution). The objects were placed in a hermetically sealed container at room temperature in a volume of water, which was kept to a minimum to fully immerse the objects. A 10 mL sample of the solution was taken after stirring every day for the duration of the treatments (except weekends). A 2 min nitrogen flush was then applied to reduce the presence of oxygen in the bath, thus limiting the risk of alloy corrosion. The baths were renewed when a plateau was observed, i.e., until the analytical data reached stable values or the object showed significant changes. Monitoring was carried out by measuring the conductivity, pH, temperature and sodium ion concentrations in the solution. The conductivity and the quantity of sodium extracted were the two reference measurements on which the monitoring and observation of treatments were based.

3.5.1. Harpocrate Statuette

The Harpocrate statuette was immersed for 35 days in three successive baths of 1 L of pure water. No visual changes were observed apart from the attenuation of efflorescence. The extraction of sodium ions in pure water was progressive until a plateau was reached over three successive baths. The solution conductivity measurements followed the same pattern. The second bath was coupled with a brushing of the statuette’s surface, which explains the slight increase in sodium in the solution. The third bath did not extract a significant amount of sodium, given the stability of the sodium concentration in the solution right up to the end. Over the entire immersion period, 60 mg of sodium was extracted, half of which was extracted during the first bath (Figure 6). The pH of the water remained stable throughout the treatment period (pH~6.5). At the end of the extraction process, the statuette was dried slowly by progressively opening its container to reach ambient relative humidity.

Some observations can be made following this treatment. White efflorescence appeared. EDS and Raman analyses did not detect the characteristic elements of chalconatronite. The constitutive elements of the efflorescence were lead (Pb), chlorine (Cl), calcium (Ca) and copper (Cu). The absence of sodium (Na) in these analyses is of particular interest. With regard to bath monitoring, it should be noted that the conductivity measured can be defined as a representative means of monitoring the sodium ion extraction. Indeed, even if the measured conductivity is not only due to the dissolution of sodium ions, the evolution of the values also remains similar to that of the sodium extraction, mainly for the first baths. The conductivity values that were calculated from the sodium concentration were lower than those measured. On the other hand, the opposite was true for the conductivity values calculated from the sodium concentration and the hydrogen carbonate ion. An equilibrium is certainly created in the solution with Na⁺/CO₃²⁻/HCO₃⁻/Cu(CO₃)₂²⁻ ions, as copper (Cu²⁺) can associate with a carbonate but is known to be insoluble. This may suggest that the sodium is “totally” dissolved and that there is a subsequent equilibrium between the hydrogen carbonate ions, not to mention the fact that there may be other elements in the solution originating from the object’s archaeological context (sediment, for example).

Indeed, from the third bath onwards, when the amount of sodium stagnates in the solution, conductivity continues to rise, indicating the passage of other ions into the solution. The water solution was evaporated, and the residue was analyzed by SEM-EDS. This revealed the presence of chlorine (Cl), potassium (K), calcium (Ca), magnesium (Mg), lead (Pb) and copper (Cu). Sodium (Na) was absent. The last two elements may have come from the object’s base metal or from corrosion products in pure water due to mechanical action. The dissolution equilibrium of chalconatronite remains to be studied in greater detail, and in the case of archaeological objects, other salts or elements may interact with the solution.

3.5.2. Isis Statuette

For the treatment of the Isis statuette, the quantification of extracted sodium increases slightly from the first days of the bath (only 7.0 mg total extracted sodium). The pH of the solution was acidified very quickly (from 6.5 to 4.5 in three days). In addition, a “film” appeared to form on the surface of the statuette. Immersion was halted to understand the reactions involved. The conductivity measured did not correspond to that calculated in relation to the amount of sodium in the solution. The calculated/measured conductivity ratios were unreliable (Figure 7). It should be noted that the analyses of the efflorescence did not show the presence of chalconatronite but rather a sodium copper formate/acetate compound. Because these compounds are soluble in water, it is, therefore, likely that this product decomposes into formic/acetic acid, thus acidifying the treatment water [32]. As the statuette is made of lead bronze, this acid-sensitive element may have reacted to form lead acetate on the surface. The solution was evaporated to determine the extracted salts. The SEM-EDS results indicated the presence of lead (Pb), chlorine (Cl) and other recurring elements in smaller proportions, such as sodium (Na), potassium (K), calcium (Ca) and copper (Cu) (Figure 8a). Analyses of the “film” predominantly detected lead, along with a few traces of copper (Cu). Infrared bands at 1480–1300 cm⁻¹, 1045 and 680 cm⁻¹ (vibration of the CO₃²⁻ group) are assigned to carbonate (Figure 8b). In the case of the Isis statuette, the transformation of chalconatronite into acetate and/or formate compounds did not prevent sodium from being extracted, considering these compounds to be water-soluble, nor has there been a recurrence of this efflorescence to date. On the other hand, the acidification of the water during treatment is a constraint that must not be overlooked. New solutions need to be developed, such as localized treatments to limit the risk of altering the object’s patina.

3.5.3. Preventive Conservation

The Harpocrate statuette had previously undergone mechanical treatment to remove efflorescence. The protection applied to it was based on 5% w/w Paraloid B72^®. Given that the planned display and environmental conditions for the statuettes imposed the presence of organic objects, it was decided to apply a highly concentrated acrylic varnish (10% w/w Paraloid B72^® in acetone) to create an effective barrier. To maintain a uniform appearance with the other bronzes displayed in the same case, we also applied a coat of diluted wax (Renaissance microcrystalline wax) to dull the shine of the varnish. The effectiveness of the protocol will be assessed by regular observation of the objects in their display case over the coming years.

4. Conclusions

Chalconatronite, a sodium copper carbonate, can be identified not only on bronze Egyptian antiquities but also on copper alloy objects treated in sesquicarbonate. These compounds take the form of green-blue efflorescence and are sometimes confused with bronze disease. It is, therefore, important to identify the type of efflorescence, as we did for the two Egyptian objects studied in this article. Although chalconatronite is relatively stable, it can evolve under very specific conditions to form acetates and formates. Our study showed that variations in relative humidity caused the efflorescence to develop. While a stable, high relative humidity (90% RH) does not seem to create new efflorescence, a variation in humidity (dry atmosphere (˂30% RH) to humid atmosphere (≈55% RH)) causes salts to migrate upwards and create new efflorescence. Partial dissolution of water-sensitive chalconatronite is already a well-known method, but our tests show that complete dissolution is, in fact, feasible. We can, therefore, propose an initial protocol for stabilizing chalconatronite by extracting the sodium in water by means of a dissolution effect. However, several limitations remain. It is difficult to monitor extraction since we do not know the quantity of salts remaining in the object or the minimum level of salts that must be reached in the bath. Furthermore, while stabilizing chalconatronite by immersion in water is a promising method, stabilizing its evolved form when acetates and formates are present is more challenging, particularly in the case of bronzes that are strongly bound to lead. However, stabilizing these objects without undertaking preventive conservation would be insufficient. We have seen that fluctuations in relative humidity are the cause of these surges in corrosion, in line with cycles of salt solubilization and surface recrystallisation. Consequently, exposure to a high relative humidity is not harmful as long as the values remain stable. The storage conditions of objects affected by chalconatronite should also be studied in more detail to better control the influence of organic vapor sources. Maintaining a stable atmosphere, even in a display case, remains difficult, and given our uncertainties about whether the sodium had been completely extracted, it was also necessary to provide surface protection for these objects. Finally, the question remains as to the efflorescence formation cycle and whether, under conditions of fluctuating humidity, sodium reacts with copper corrosion products (malachite) and/or with the copper alloy. And if the metal is involved, then objects are affected by active corrosion, which will be important to stabilize.