Pigments and Pictorial Style Used in the Artworks of the Romanian Painter Theodor Aman

Abstract

This work presents the first in-depth investigation of Theodor Aman’s paintings that focuses on three of his heritage artworks: “Hora de peste Olt” (1866), “Teleleice în Harem” (1879), and “Regimul vechi” (1881), and that relies on both elemental and spectroscopic analytical techniques. Non-destructive Raman spectroscopy was employed on all three works of art to identify the pigments used by the Romanian master. In addition, micro-samples were available from “Hora de peste Olt” and “Teleleice în Harem”, which were further analyzed using XRD, micro-Raman, ATR-FTIR, and SEM-EDS techniques to provide complementary information on the pigments. SEM-EDS was also applied to study the structure of the preparation layers. The analyses revealed significant differences between the artworks in terms of both the pigments employed and the preparation of the canvas, suggesting that the earlier artwork belongs to one creative phase, while the newer pieces can be attributed to a later phase in the artist’s career.

1. Introduction

Since the first radiographs of paintings were recorded by Walter König, the study of cultural heritage objects has always benefited from the advancement of new spectroscopic techniques applicable in materials science. The primary distinction between the use of these methods in other domains is that the items that are studied have special cultural value and need careful preservation and handling. As such, researchers generally rely on in situ, non-destructive or micro-destructive techniques in the analysis of cultural heritage artifacts [1]. These techniques include X-ray-based methods such as XRD, XRF, XPS, PIXE, etc., UV-VIS and fiber-optic reflectance spectroscopy, vibrational spectroscopy such as Raman and IR, to complex hyphenated chromatographic and mass spectrometry techniques, etc. The landscape of analytical techniques is always advancing, allowing for greater understanding at the macro-, micro-, and nano-levels of the various components of the works of art. Current techniques have allowed researchers to study works of art in depth, permitting them to investigate their composition, provenance, the interaction between their composing materials, and the degradation processes that affect them. However, the intrinsic complexity of most artifacts forces researchers to rely on a multitude of investigative techniques that complement each other in order to fully characterize them.

Theodor Aman (1831–1891) can be considered the architect of modernity in Romanian painting, as he was the one who ensured the synchronization of national visual arts with the European value system through his dual mission as a visionary artist and an institutional founder. His work constitutes an effervescent junction between the rigors of Parisian academic canons and the first pulsations of modernity, foreshadowing plastic solutions that would become benchmarks for subsequent generations [2]. Through his masterful command of traditional genres—ranging from historical composition to genre painting and engraving—Aman succeeded in crystallizing a national visual identity under the auspices of exceptional technical refinement. His originality lies in his capacity to infuse descriptive realism with a unique luminous sensibility, transforming historical or social documents into artistic facts of high esthetic standing. Despite his significant impact on modern Romanian painting, the choices of pigments and materials used by Aman have largely gone unexplored. Existing studies have primarily focused on stylistic analysis, often overlooking the material aspects of his work. This investigation aims to fill that gap by identifying the specific pigments he used across various genres and periods. Such an analysis can provide concrete evidence of his technical development, influences, and potential sources of his materials. Understanding Aman’s palette involves not just chemical analysis but also an interpretation of his artistic influences, ranging from his academic background in Paris to the emerging modernist trends, and helps clarify his role in shaping a national visual identity.

This study examines Aman’s work through an analysis of three key paintings, i.e., “Hora de peste Olt” (1866), “Teleleice în Harem” (1879), and “Regimul Vechi” (1881), highlighting his thematic and chromatic versatility (see Figure 1, Section 2.1. Art Historical Context and Selection of Heritage Paintings). The goal was to explore the variety of materials he used and identify differences in technique related to subject matter and period. A comprehensive suite of analytical methods was employed, including portable Raman spectroscopy (p-Raman), micro-Raman spectroscopy (μ-Raman), Fourier transform infrared spectroscopy (FTIR-ATR), X-ray diffraction (XRD), scanning electron microscopy with energy dispersive X-ray spectrometry (SEM-EDS), and gas chromatography with mass spectrometry (GC-MS). This selection aimed to fully characterize the pictorial materials by identifying pigments, determining crystalline phases, analyzing elemental composition, and characterizing organic compounds. In line with the principle of minimal intervention in conservation, the investigation started with non-destructive methods and then progressed to micro-invasive analyses of specific areas of the paintings.

2. Materials and Methods

2.1. Art Historical Context and Selection of Heritage Paintings

“Hora de peste Olt” (1866) (Figure 1a) is part of the famous “Hore” cycle and, in this work, Aman idealizes rural life as a pillar of national identity. Unlike his salon scenes, the master adopts a solar and dynamic perspective here, marking his opening toward plein-air painting [3]. From a technical standpoint, one observes a more fragmented freedom of brushwork, intended to capture the circular movement of the dance. The artist partially abandons the rigor of the line in favor of patches of color that vibrate under natural light. The chromatic palette is dominated by warm, earthy tones, punctuated by the brilliant white of the traditional folk costume (i.e., catrință), rendered with a precision that betrays the artist’s love for authentic picturesque detail. Through this luminous effervescence, Aman foreshadows the interest in light that would later be perfected by Nicolae Grigorescu [4].

“Teleleice în Harem” (1879) (Figure 1b) is an interior scene that reflects Aman’s fascination with the “Near East,” transposed into an atmosphere of siesta and controlled sensuality. Unlike his historical paintings, the composition here is relaxed, often unfolding horizontally, suggesting the characters’ surrender to the enclosed space of the harem [5]. Aman does not seek documentary truth, but rather creates an atmosphere of oneiric reverie, emphasizing tactile contrasts. His refined technique succeeds in differentiating the density of Oriental carpets from the fineness of transparent veils and the metallic reflections of copper vessels. Color thus becomes the primary instrument for suggesting a warm and mysterious ambiance, transforming the “exotic” into a demonstration of technical modernity and chromatic virtuosity [6].

“Regimul Vechi” (1881) (Figure 1c) is a piece of strong narrative and critical character, reflecting the structural transformations of 19th-century Romanian society through the contrast between the old feudal-Phanariote order and the new modernizing currents [7]. Aman utilizes a technique based on chiaroscuro and a rigorously structured composition to emphasize the ideological message. The artist employs successive glazes (thin layers of color) to achieve depth and a faithful rendering of materiality, from the heavy texture of brocade to the sheen of silk. Stylistically, the canvas aligns with Academic Realism, where drawing precedes and disciplines color [8]. The characters are strategically placed in space, with light being used as a psychological element to isolate or highlight the anachronism of figures who refuse to adapt to the country’s new European rhythm.

2.2. Materials and Methods Used for the Investigation of Heritage Paintings

This study relied on on-site analyses with portable equipment performed at the “Moldova National Muzeum Complex”, and on investigations of collected microscopic fragments using tabletop instruments. Portable Raman measurements were performed using R-3000CN from Raman Systems, Woburn, MA, USA equipped with a 785 nm diode laser. Spectral resolution was set to 2 cm⁻¹, while integration time was 10 s. Laser power was in the 10–50 mW range, and the laser spot was about 0.1 mm in diameter. The spectrometer was calibrated before each experimental session using the Raman peak of a silicon crystal at 520.5 cm⁻¹. The portable Raman analysis points, as highlighted in Figure 1, are referenced further on with the initials of the work of art, H for “Hora de peste Olt”, T for “Teleleice în Harem”, and R for “Regimul Vechi”, followed by the number of each analysis point.

Classical tabletop analytical techniques were employed for microscopic samples collected from areas under the picture frames, as highlighted in Figure 1. Grains, generally less than 1 × 1 mm in size, were collected from blue and green colored areas for “Hora de peste Olt”, referenced from now on as H_b and H_g, and from red and green areas from “Teleleice în Harem”, referenced from now on as T_r and T_g. ATR-FTIR spectra were produced with a Bruker Vertex 70 instrument (Billerica, MA, USA), in the 4000–600 cm⁻¹ region, with a resolution of 2 cm⁻¹, by performing 64 scans at room temperature, using the ATR technique and the Opus 5 FTIR Software. Due to the brittleness of the samples, grains selected for ATR-FTIR cracked into 2 or more pieces when removing them from the sample bags with tweezers. When possible, the individual pieces were collected, and a spectrum was recorded for each individual one. For this reason, there are one recorded spectra for H_b and T_r, three for T_g (from now on called T_g_1, T_g_2, T_g_3) and three for H_g (from now on called H_g_1, H_g_2, H_g_3). For the microscopic grains collected from the rims of the paintings, Raman spectra were recorded using a Renishaw InVia Reflex spectrometer with a 785 nm diode laser through a 50× objective. The laser power density on the samples was minimized to 5 mW to avoid any thermal damage. Each spectrum was recorded using a 1s accumulation time for 15 spectral accumulations to achieve an acceptable signal-to-noise ratio. The resolution of the recorded spectra was 2 cm⁻¹. The spectrometer was calibrated before each experimental session using the Raman peak of a silicon crystal at 520.5 cm⁻¹. The SEM-EDS analyses were performed with a Verios G4 UC Scanning electron microscope (Thermo Scientific, Brno, Czech Republic) coupled with an energy dispersive X-ray spectroscopy analyzer (EDAX Octane Elite, Mahwah, NJ, USA) with the following settings: 4 mm working distance, and an accelerating current of 20 KeV and high vacuum. Grains from all 4 samples were mounted with conductive double adhesive tape on Al stubs. One H_g grain was encased in Struers EpoFix Resin and sectioned with a Microtome CUT4050 (microTec Laborgeräte GmbH, Walldorf, Germany) prior to SEM-EDS investigations. XRD diffractograms were recorded on a Rigaku SmartLab X-ray diffractometer, Cu anode, 2θ 2° to 90°, scan step 0.02°, 3°/min acquisition time, and phase identification was performed with Match! 4 software version 4.1 build 311 (Crystal Impact, Bonn, Germany). OM images were collected with a Leica DM2500 M microscope (Leica Microsystems, Wetzlar, Germany) and LAS Interactive Measurement software (https://www.leica-microsystems.com/products/microscope-software/p/leica-las-x-ls/ accessed on 12 April 2026) for image acquisition. Graph creation for the FTIR, Raman, and µ-Raman spectra, as well as background subtraction, were performed using Origin software (https://www.originlab.com/ accessed on 12 April 2026).

The samples, approximately 1 mg each, underwent organic phase extraction using 1 mL of n-hexane (99.1%, Merck, GC ECD and FID grade). This was followed by ultrasonication at 25 kHz with continuous excitation for 5 min. After centrifugation at 1000 rpm, the solids settled at the bottom, and the organic supernatant was collected. This extraction procedure was repeated three times. In a 20 mL vial with a tight cap, the samples were derivatized by adding 1 mL of freshly prepared 2% sodium methoxide solution in methanol. The reaction occurred for 1 h at 55 °C in a thermal orbital shaker at 250 rpm, with cycles of 5 s on and 1 s off, rotating in both directions. After the reaction, the liquid phase was washed with 2 mL of 20% NaCl solution. The mixture was centrifuged at 1000 rpm, and the upper organic phase was separated and collected. The aqueous saline solution was then re-extracted with 1 mL of hexane by vortexing at 1000 rpm, and this organic fraction was combined with the first fraction. The combined fractions underwent a second derivatization step, which involved adding 1 mL of 10% BF₃ solution in methanol. This reaction was carried out for 30 min at 55 °C, under the same stirring conditions. After cooling, the reaction was stopped by adding water and stirring for 10 min. The mixture was treated with 2 mL of 20% NaCl solution, and the organic phase was extracted with n-hexane through vortexing and centrifugation. All organic fractions were combined, and any traces of water were removed by adding 0.5 g of anhydrous magnesium sulfate (MgSO₄). The dried organic phase was then transferred to a vial insert and brought to a final volume of 100 µL with GC-grade n-hexane (99.1%).

A volume of 1 µL of the final solution was injected into a GC-MS gas chromatography system (Shimadzu GC-2010 coupled to GCMS QP2010 plus, equipped with a Shimadzu AOC5000 automatic injector). The injection was performed in split mode, with a split ratio of 1:10, and the injector temperature was set to 260 °C. Separation was achieved using a Thermo Scientific TRACE TR-Wax capillary column (60 m × 0.25 mm × 0.25 µm). Helium (6.0 grade) served as the carrier gas, with a constant column flow rate of 2.43 mL/min and a linear velocity of 40 cm/s. The oven temperature profile was as follows: 50 °C held for 0.25 min, followed by a ramp to 250 °C at a rate of 3 °C/min, and held for 8.08 min. The transfer line temperature to the mass spectrometer was 250 °C, and the ion source temperature was set to 200 °C. Ionization was accomplished by electron impact at 70 eV, and data acquisition was performed in scan mode over the range of 50–350 Da. Compound identification was achieved by comparing mass spectra with the NIST 14, Wiley 229, and SZTERP libraries, utilizing the similarity index and corroborating with retention times in relation to a standard Supelco 37 Component FAME mixture (certified reference material, TraceCERT).

3. Results

The complementary use of portable techniques, p-Raman, and micro-invasive analyses, μ-Raman, ATR-FTIR, XRD, and SEM-EDS, has revealed the presence of multiple pigments that compose the color palette of the artist.

Table 1. Pigments identified in the works of Theodor Aman based on the employed spectroscopic techniques.

Pigment	Artwork
	“Hora de Peste Olt”			“Teleleice în Harem”			“Regimul Vechi”
	Painting	H_b	H_g	Painting	T_r	T_g	Painting
Lead white	a (H_2)	c, e	c, e	a (T_11, T_18, T_28)	c, e	c, e	a (R_19, R_20)
Barium white		c	c
Chalk (calcite + dolomite)		c, e	c		e	e
Vermillion	a (H_14, H_16)	b		a (T_6, T_7, T_17, T_18, T_28)	b, c	b, c
Red ochre		b				d
Realgar		b *
Chrome yellow		b, e	e	a (T_13, T_14)		b, e
Zinc yellow		e			d	d, e
Yellow ochre		b
Indian yellow		b, e	e		e	e
Ultramarine		b	b		b	b	a (R_9)
Prussian blue		b	b
Carbon black		b	b		b	b

(a) portable Raman measurements collected on the paintings with the specified analysis point where the pigments were observed, specified in parentheses, (b) μ-Raman, numbers indicate the analysis point, (c) XRD, (d) EDS, (e) FTIR, *—identified from degradation products.

summarizes the list of identified pigments and the corresponding analytical techniques that allowed for their identification.

3.1. Portable Raman Spectroscopy (p-Raman)

p-Raman spectroscopy was utilized for the in situ identification of the main pigments in Aman’s palette, effectively complementing and validating the findings obtained through laboratory analysis. Vermilion (HgS) was detected in both “Hora de peste Olt” (points H_14, H_15) and “Teleleice în Harem” (points T_6, T_7, T_17, T_18, T_28), thanks to the high quality of the recorded spectra (Figure 2a,b). Its characteristic bands were observed in the ranges of 253–255 cm⁻¹, 278–286 cm⁻¹, and 343 cm⁻¹, aligning with the existing literature [9].

Lead white (2PbCO₃·Pb(OH)₂) was detected in all three paintings at points H_2, T_11, T_18, T_28, R_19, and R_20, characterized by a strong absorption band at 1047–1049 cm⁻¹ (Figure 2a,c) [10]. Chrome yellow (PbCrO₄) was found in “Teleleice în Harem” at points T_13 and T14, identifiable by its distinctive band at 839-840 cm⁻¹ (Figure 2b) [11]. Ultramarine was identified in “Regimul Vechi” at point R_9, based on its primary band at 545 cm⁻¹ (Figure 2c) [12].

3.2. μ-Raman Spectroscopy

µ-Raman analysis revealed a remarkable diversity of pigments used by Aman, demonstrating both a commitment to academic tradition and an openness to contemporary materials of the era. In T_g, H_b, and H_g samples, ultramarine was identified by the main spectral band at approximately 546 cm⁻¹, along with additional ones around 1095 cm⁻¹, 258 cm⁻¹, 585 cm⁻¹, and 805 cm⁻¹ [12] (Figure 3a and Figure S1). In the H_b and T_g samples, the µ-Raman spectra indicated the presence of chrome yellow (PbCrO₄), which can be recognized by its distinct bands near 835, 358, and 839 cm⁻¹ [11] (Figure 3a). Although the weak signal of chrome yellow indicates a low concentration, its presence is significant for understanding Aman’s color palette during this period.

Vermilion (HgS) was identified in T_r, and T_g samples based on its characteristic spectral bands, confirming the artist’s frequent use of this pigment [9] (Figure 3a and Figure S4). Goethite (α-FeOOH), which gives yellow ochre its color, was detected at a single analysis point in the H_b sample (Figure 3b). The strong fluorescence of the spectrum limited the identification to the main band at 387 cm⁻¹ [13]. Hematite (α-Fe₂O₃), the active component of red ochre, was also identified in the T_r sample, based on its multiple characteristic vibrational modes (at 228, 248 cm⁻¹, 294 cm⁻¹, 298 cm⁻¹, 412 cm⁻¹, 500 cm⁻¹, and 610 cm⁻¹) [14]. Raman spectroscopy revealed that pararealgar was formed in sample H_b as a result of the photochemical transformation of realgar (As₄S₄) under light exposure. Although fluorescence affected the spectrum, the band at 234 cm⁻¹ (likely results from overlapping the two characteristic bands at 230 and 236 cm⁻¹) allowed for its identification [15].

Carbon black, which is a form of amorphous carbon, was detected in all samples due to its distinctive D band (approximately 1355–1365 cm⁻¹) and G band (approximately 1582–1587 cm⁻¹), as shown in Figure 3c and Figure S3 [16]. Prussian blue (Fe₄[Fe(CN)₆]₃·×H₂O) was found only in the samples from “Hora de peste Olt,” characterized by a specific doublet in the 2090–2150 cm⁻¹ range, which is attributed to the C≡N vibration [10] (Figure 3c and Figure S2).

A high-fluorescence spectrum obtained from the H_b sample displayed bands at 1385 cm⁻¹, 1430 cm⁻¹, and 1624 cm⁻¹ (Figure 2c). This spectrum suggests a mixture of Indian yellow [17], an organic pigment of animal origin made from natural magnesium and calcium salts of euxenic acid derived from cow urine fed on mango leaves, which was widely used in painting until the early 20th century, and disordered graphite [18]. The literature indicates that the Raman spectrum of Indian yellow typically shows five main bands at 1346, 1368, 1425, 1441, and 1626 cm⁻¹ [17]. In our analysis, the sharp band at 1624 cm⁻¹ corresponds to the most intense band of this pigment, while the broader band at 1430 cm⁻¹ likely results from the overlapping of the characteristic bands at 1425 and 1441 cm⁻¹, which could not be distinctly separated due to high fluorescence and a poor signal-to-noise ratio. This phenomenon is well documented in Raman analyses of Indian yellow using 785 nm excitation, where strong fluorescence interferes with spectral quality [17]. While other materials, such as certain resins, may also show absorption bands near 1624 cm⁻¹, the obtained spectrum is notable for lacking additional strong bands in the 1500–1700 cm⁻¹ range that are typical of these materials. The presence of metal carboxylates at approximately 1535 cm⁻¹, along with oil oxidation products in the 1700–1750 cm⁻¹ range, is expected in the context of an aged sample. The lack of secondary bands, combined with the broad band at 1430 cm⁻¹, further supports the identification of Indian yellow. Additionally, the band at 1385 cm⁻¹ can be attributed to highly disordered graphite (amorphous carbon), likely associated with carbon black present in or near the analyzed area. Disordered graphite usually displays an intense D band in this spectral region [18].

3.3. XRD

The minerals present in the four samples were identified based on their XRD patterns with their most important peaks highlighted (Figure 4) and the complete list of entries for each sample being presented in Supporting Information (Figures S5–S8). In Figure 4 the peaks of hydrocerussite, and only the clear distinctive peak of cerussite at 25.47 are highlighted as most peaks of these two compounds overlap.

For quartz, whose main peak at 2θ = 26.64 overlaps with the second most intense peak of cinnabar, it was chosen to indicate the position of this peak.

The analysis of the diffraction pattern shows that all samples contain both hydrocerussite and cerussite, the two main components of the lead white pigment. Their presence can be attributed to both the use of lead white as a pigment in itself and as a component in the preparation layer of the canvas.

For “Teleleice în Harem”, XRD analysis showed the presence of kaolinite in both samples. Impurities of kaolin, such as muscovite, were found T_g, while quartz was identified in T_r. The presence of quartz cannot be solely attributed to the use of kaolin, as cinnabar and red ochre, other minerals identified by XRD and Raman spectroscopy, are also known to contain quartz impurities.

The identification of cinnabar, the mineral from which the red pigment vermilion is produced, in both “Teleleice în Harem” samples, while other pigments identified via Raman or EDS spectroscopy were not found, is likely due to its higher crystallinity rather than the amount used in the artwork.

For “Hora de peste Olt”, the XRD investigations were able to identify only the materials used in the preparation layer of the painting. As such, alongside the hydrocerussite and cerussite, previously mentioned barite, calcite and dolomite were also encountered. The presence of barite can be attributed to the use of barium white pigment, while calcite and dolomite are components of chalk.

3.4. OM and SEM-EDS

SEM-EDS measurements of the samples proved difficult as they were very brittle, and simple handling in order to mount them with conductive double adhesive tape on Al stubs resulted in them breaking apart. As such, fractured grains were investigated for the T_r, H_b and H_g samples. Section investigations were possible for a T_g, as handling resulted in a clean break, and for H_g, a grain was encased in resin and sectioned. Elemental compositions of the investigated areas are presented in Tables S1–S4.

Prior to SEM-EDS investigations, OM was performed on the samples chosen for analysis and areas with different color grains were recorded in order to determine the materials used by the painter by correlating feature observed in OM with ones notices in SEM.

Red grains, designated as 1, 2, 3 and 4 in the OM investigations (Figure 5a–c) of T_r, were analyzed by EDS (Figure 5d). The measurements showed that significant amounts of Fe and Hg were present (Table S1, lines 1–4). The presence of Hg can be straightforwardly attributed to the use of vermillion, HgS, which has a bright intense red color. The rather low amounts of Hg compared to Fe are due to the fact that vermillion has high tinting power, and generally, even a small amount allows for developing the required shade.

The presence of Fe in these red grains can be correlated to the use of red ochres, which was observed in the μ-Raman investigations. Ni is found in most cases in reduced amounts in the area of the red grains, and its presence can be correlated to Ni being an impurity in iron-bearing ores that contain hematite [19]. The investigations also showed the presence of As (Table S1, line 5); however, its identification in only one analysis point indicates that the painter made little use of realgar, As₄S₄.

The investigation of several grains that were dislodged from the original sample upon mounting on Al stubs (Table S1 lines 6–8, Figure S9) further confirms the use of red ochre and vermillion, with atomic percentages reaching over 26% for Fe and 13% for Hg. In most analysis points, Pb is found in significant amounts, indicating that Aman made extensive use of the lead white pigment. Its presence also greatly influenced the quantification of S due to the overlap of the Mα₁ at 2.3455 of Pb, which is intense due to the high amount of Pb, and the Kα₁ at 2.30784, Kα₂ at 2.30664 and Kβ₁ at 2.46404 of S. Due to this, the stoichiometry for both realgar and vermillion is not respected in the atomic percentages. The analysis of these grains also showed important amounts of Zn and Cr, which can be correlated to the use of zinc yellow (4ZnCrO₄·K₂O·3H₂O). Other elements that were encountered include Al, Si, Ca, K, Mg, which can be correlated to the presence of alkali feldspars and kaolin.

OM investigations of the section of the T_g grain showed that the stratigraphy of the painting consists of a white preparation layer, followed by a red underlayer on top of which the final pigment and oil layer is situated (Figure 6). In the EDS mapping of the section for Pb, Al, Si and Fe, we have highlighted two distinct areas A and B that correspond to the white preparation layer and the red underlayer. In area A, there are higher amounts of Pb as opposed to B, while Fe is completely missing, and Al and Si are found in only a few random grains. This indicates that the preparation layer was made exclusively with lead white. In area B, we find that Pb, Al, Si, and Fe are present, indicating that the red underlayer was composed of lead white, with the addition of red ochre. The kaolin identified in the XRD measurements likely comes from this layer as red ochre is based on siliceous clays, such as kaolin, rich in hematite (Fe₂O₃). The difference in the composition of the layer is obvious when looking at individual grains. As such, the most obvious is grain 3 highlighted in Figure 6e,h, where Pb is completely absent, and the EDS mapping of Fe shows a large, well-defined grain. For the grain indicated with 1 in Figure 6d–g,i, again we see a depletion of Pb well-defined edges appear for O, Al, Si and K, indicating the presence of a K feldspar, which is a common impurity in kaolin. Another impurity of kaolin, quartz was observed due to the presence of Si (Figure 6g) and absence of other elements in grains numbered 6 and 7. The grain highlighted with 2 in Figure 6d,e,i–k provided conclusive evidence of the use of zinc yellow, in use since the beginning of the 19th century, as in that area we observe a depletion of Pb and instead O, K, Zn and Cr are high. In the grains highlighted with 5 and 4 in Figure 6e,l,m, we observe low Pb while Hg and S are high, indicating the presence of vermillion on the very surface within the oil layer of the painting which is high in C and O. The EDS mapping also showed the presence of several Cu-rich grains, highlighted with 8, 9 and 10, which we could not correlate with any particular pigment, as they were not noticed during OM, and the presence of C and O in the binder and in the oil makes it impossible to attribute them to blue or green or other colored pigments, such as azurite, malachite, copper resinate, etc. Only a few random individual analysis points were also recorded in the section, and only the most important ones are presented (Table S2). Their composition mirrors the one in the EDS mapping, with high amounts of Fe from red ochre and significant amounts of Zn, Cr, and K from zinc yellow (4ZnCrO₄·K₂O·3H₂O). Both in the measurements for the red grains and the green grains, the stoichiometry of the zinc yellow is not respected, as Zn is in higher amount, indicating that an excess of ZnO was used in the manufacture of the pigment. Since no blue grains were clearly observed by OM, and ultramarine is a silicate mineral, its presence could not be clearly confirmed.

In the case of the H_b grain chosen for OM and EDS analysis, the sample cracked when handling, and the two largest pieces were then collected in order to be analyzed. Unfortunately, again, during mounting on Al stubs for EDS measurements, the samples cracked, and as such, no correlation with OM images could be performed. The investigations (Table S3) showed that the most abundant metals are Ca, Ba, Pb and Mg. The presence of Ca can be correlated with calcite, and that of Mg in dolomite, both found in chalk. Distinct analysis points where Ca is in very high amounts, over 14% and 9%, contain only much smaller amounts of S, indicating that CaCO₃ is present and not gypsum, CaSO₄. Ba and Pb, on the other hand, can be attributed to barium white and lead white, respectively. While Ba is also found in lithopone (a mixture of BaSO₄ and ZnS), the small amounts of Zn in the analysis point indicate that this pigment is not present. Al and Si were found in several areas, indicating the presence of aluminosilicate impurities in the white pigments. In an area (Table S3, line 5), Fe is encountered in large quantities (over 5.%), but its presence cannot be clearly attributed to any specific pigment as OM analysis provides no support for the color of the individual analysis point, and Raman analysis proved the presence of Prussian blue, red ochre, and yellow ochre. Other elements are generally in small abundance, and no clear correlations can be made to their presence.

The analysis of H_g grains, mounted on Al stubs, showed a similar composition to the blue ones (Table S4). Again, the most abundant metals are Pb, Ca, Ba, and Mg that can be correlated to the use of lead white, barium white and chalk. The analysis of the green sample encased in resin and sectioned (Figure 7) showed that the stratigraphy of the painting consists of only the preparation layer (highlighted area in Figure 7e–g,j,k and the paint layer. The paint layer could not be highlighted as the encasing resin has a similar composition to the oil rich in C and O. The EDS mapping of the sectioned grain clearly confirms previous results, indicating that in the preparation layer Aman used a mixture of white lead, barium white and chalk (calcite with dolomite impurities).

Unfortunately, the paint layer was very thin and no pigments could be observed. Several distinct grains associated with impurities are observed, and the one highlighted as 1 (Figure 7h) contains high amounts of Si. This is obviously a grain of quartz, even though in some areas the EDS mapping shows Ba and Ca. This indicates that the grain is not on the surface, but just below, and Si was detected, as EDS is a volume technique that can reach up to several μm in depth. The same can be said for the grains highlighted as 2 in the Al EDS mapping (Figure 7j), which are likely feldspars. Distinct grains containing Fe were noticed (Figure 7l), and their presence can be simply attributed to ochre residues on the paint brush when the artist created the preparation layer. Noticeably, no distinct grains of Hg (Figure 7m) and Cr (Figure 7n) are absent, indicating that in the section, no pigments associated with red or yellow are present in the analyzed area.

3.5. ATR-FTIR Spectroscopy

FTIR-ATR analysis provided additional molecular insights that complemented data from other analytical techniques, enhancing our understanding of the materials’ composition and Theodor Aman’s execution technique. The infrared spectra (Figure 8) illustrate the complexity of heritage materials, arising from aging and interactions among the original components. A complete assignment of the absorption bands for each sample is provided in Table S5 and the ESI. Identification of absorption bands in the FTIR spectra was based on correlations with pigments previously identified using complementary techniques, such as Raman, XRD, and EDS.

Lead white was identified in all samples by the characteristic bands of cerussite (PbCO₃) and hydrocerussite (2PbCO₃·Pb(OH)₂) [20,21], reflecting the mixed nature of the pigment and its use in both primers and color mixtures. Quartz (SiO₂) was identified in the T_g_1, T_g_2, and H_b samples [22,23] and is commonly associated with red ochre or with texturing agents of various colors, i.e., black (black chalk/graphite), white (calcium carbonate, lead white, kaolin), or colored (cinnabar, green earth) [24]. Kaolinite, identified in the T_g_2 sample, was confirmed by the characteristic hydroxyl and Si–O bands, consistent with the literature [25]. Chrome yellow (PbCrO₄) was confirmed in the T_g_1, T_g_3, H_b, H_g_1, and H_g_3 samples [26], and zinc yellow (ZnCrO₄) in the T_g_2 and H_b samples [27,28], based on the FTIR absorption bands characteristic of each pigment. Calcium and magnesium carbonates, such as calcite (CaCO₃) and dolomite (CaMg(CO₃)₂), were identified in the T_r, T_g_1, and H_b samples [29]. Differentiation of calcite from dolomite was based on the shift in the ν₄ band (711 cm⁻¹ for calcite vs. 721–729 cm⁻¹ for dolomite) and the splitting of the ν₃ band (a component at around 1445 cm⁻¹ attributed to calcite) [29]. Although XRD did not confirm their presence in all samples, the detection of these carbonates suggests that mineral impurities may be associated with kaolinite or other earth pigments.

The FTIR spectra of samples T_r, T_g_3, H_b, and H_g_2 show bands in the region of the characteristic spectral signatures of Indian Yellow. These include 1622–1627 cm⁻¹ (euxanthic ring), 1580–1584 cm⁻¹ (carboxyl groups), and 1445–1459 cm⁻¹ (combination of vibrational modes) [30]. The bands in the range 981–1550 cm⁻¹ were assigned to specific vibrations of the euxanthic ring, carboxyl groups, and C–O–C bonds. Other bands reported in the literature for Indian Yellow include 1487, 1423, 1345, 1258, and 1016 cm⁻¹ [30]. Unfortunately, these absorption bands overlap with the broad absorption bands of aged linseed oil, and the presence of pigments and degradation products prevents their clear assignment to this pigment. However, given that the pigment was identified by Raman spectroscopy, the assignment of Indian Yellow vibrational modes is presented in Table S5 (ESI).

Vermilion (HgS) and realgar (As₄S₄) show no absorption bands within the analyzed FTIR range (4000–600 cm⁻¹), so they were not detected. Barium white (BaSO₄) [31] and ultramarine (Na₈–₁₀Al₆Si₆O₂₄S₂–4) [32], although confirmed by other techniques, could not be definitively identified by FTIR because of spectral overlaps. FTIR also did not confirm the presence of Prussian blue, which may indicate either its absence at the sampling points or a concentration below the method’s detection limit.

FTIR analysis confirms that all samples are complex mixtures of aged oil, mineral pigments, and metallic soaps, reflecting both the original materials and post-execution processes. The characteristic spectral signatures (3000–2800 cm⁻¹, 1700–1750 cm⁻¹, 1200–1000 cm⁻¹) are typical of an aged, oily drying binder, most likely linseed oil, given the historical context [33]. Although hydrolytic degradation may be more chemically precise, we use the traditional term “aged linseed oil” in this paper to encompass both oxidative and hydrolytic processes characteristic of heritage materials. Metallic soaps, namely zinc azelates (T_r, H_b, H_g_1–H_g_3 samples) [34], calcium stearates (T_r, T_g_3, H_g_2 and H_g_3 samples) [35] and lead stearates (T_g_1, T_g_2, and H_g_1 samples) [36], were identified in the works “Hora de peste Olt” and “Teleleice în Harem” (see Table S5, ESI), indicating active reactions between oxidized fatty acids and metal ions released from the pigments. In samples taken from under the frame (T_r, H_b, H_g_2, and H_g_3), calcium oxalate (whewellite) was identified [37], suggesting a complex degradation involving both biological factors (such as microbial colonization) and oxidative processes.

Identifying natural resins in FTIR spectra is challenging because bands from aged oil, pigments, and metallic soaps overlap. However, a comparative analysis with reference spectra suggests that sandarac resin is present, especially in the H_g_3 sample [38,39], with 16 of 19 characteristic bands coinciding, albeit with minor shifts and partial masking of some bands. Clear signatures of a protein binder (animal glue) with an unmodified triple helix structure were identified in the H_b and H_g_3 samples, indicated by amide bands I, II, and III [40,41]. The presence of β-sheets (associated with the aging process) was observed in all samples except the T_r one. Nevertheless, FTIR cannot differentiate between specific animal sources (such as rabbit skin, bone, or hide glues).

The chromatographic profiles of the examined samples include unsaturated fatty acids such as oleic acid (C18:1) and linoleic acid (C18:2), saturated fatty acids such as palmitic acid (C16:0) and stearic acid (C18:0), and specific degradation products (C13:0, C15:0) (see Figure S10, Table S6, ESI) [42,43]. The calculation of the palmitic/stearic (P/S) ratio provided valuable insights only for sample T_r, which had a P/S ratio of 0.83. For samples T_g and H_g, the absence of stearic acid rendered this ratio meaningless. In contrast, for sample H_b, the calculated P/S ratio of 34.4 is significantly higher than the range reported in the literature.

4. Discussion

Multi-analytical investigations of three representative works by Theodor Aman, namely “Hora de peste Olt” (1866), “Teleleice în Harem” (1879), and “Regimul Vechi” (1881), have enabled a detailed reconstruction of the artist’s palette and working techniques. This analysis highlights the specific materials used in each artwork and the evolution of his practice over time. These three works illustrate distinct moments in Aman’s career, and a comparative analysis of the materials reveals a clear trajectory, i.e., from the experimental diversity of his early period to a refined palette and a more complex stratigraphic technique in his mature work.

In “Hora de peste Olt” (1866), Aman’s palette showcases a remarkable diversity, featuring pigments such as lead white (used both as a pigment and in the primer), barium white, chalk (a mixture of calcite and dolomite), vermilion, red ochre, realgar, chrome yellow, zinc yellow, yellow ochre, Indian yellow, ultramarine, Prussian blue, and carbon black. This extensive range reflects a period of exploration and experimentation, during which the artist tested various materials, from traditional pigments (like ochres, realgar, and ultramarine) to modern ones (such as Prussian blue and chrome yellow). Notably, the use of Indian yellow signifies Aman’s access to a diverse European market for artistic materials and his interest in achieving special chromatic effects, especially in warm and bright areas. The primer for this early work consists of lead white, barium white, and chalk, with the paint layer applied directly over it, without visible intermediate layers. This suggests a direct technique characteristic of Aman’s early period. In contrast, “Teleleice în Harem” (1879) features a considerably refined palette, retaining only lead white, chalk, vermilion, chrome yellow, zinc yellow, Indian yellow, ultramarine, and carbon black. This simplification indicates that Aman had developed a set of preferred materials, allowing him to achieve desired visual effects in a more controlled manner, while abandoning some experimental or less crucial pigments. From a stratigraphic perspective, “Teleleice în Harem” reveals a more sophisticated technique compared to the earlier work. Over the primer (composed of lead white and chalk), Aman applies a red base layer (made with either vermilion or red ochre) before building the final pictorial layer on top. This approach, typical of European painting, enhances depth and chromatic vibrancy, effectively rendering complex textures, from the richness of oriental carpets to the delicacy of fabrics and the warmth of skin tones, thus significantly contributing to the warm and mysterious atmosphere of the piece. For “Regimul Vechi” (1881), the analysis was limited to in situ portable Raman measurements, which confirmed the presence of lead white and ultramarine. The absence of other pigments in the data does not necessarily indicate they are missing from the painting; rather, it reflects the limitations of surface analysis and the limited number of accessible measurement points.

Beyond the inorganic palette, FTIR analysis provided crucial insights into the organic materials employed by Aman, leading to a more nuanced understanding of his technique. Animal glue, a protein binder known for its distinctive amide bands, was identified in H_b and H_g_3 samples. Its presence suggests a mixed technique, likely involving protein materials either in the primer layer (to isolate and stiffen the canvas) or as a binder in certain areas of the painting, a practice commonly observed in 19th-century European artwork. Additionally, in the H_g_3 sample, sandarac resin was tentatively identified based on the overlap of sixteen out of nineteen characteristic bands. This finding raises the possibility that a protective varnish was applied over the pictorial layer, a practice documented during that period.

Regarding the conservation status, analyses have revealed various post-execution degradation phenomena affecting both pigments and binders. Raman spectroscopy identified two significant degradation phenomena in the H_b sample. First, the formation of pararealgar was detected as a result of the photochemical transformation of realgar (As₄S₄) when exposed to light, which is a rapid and irreversible process. Although the chromatic changes (from orange-reddish to yellow) are not visible macroscopically in the analyzed areas (beneath the frame), such transformations likely occurred in the exposed sections of the painting [44]. Second, early transformations of lead white were observed, even in protected areas (under the frames). The absence of ν₁ band splitting in the Raman spectrum of hydrocerussite indicates the formation of sulfur-substituted lead carbonates (xPbCO₃·yPbS), which coexist with a fraction of unreacted pigment [45]. This indicates the pigment’s sensitivity to atmospheric pollutants, particularly sulfur compounds. The initial transformations of lead white, even in seemingly protected regions, suggest a reactive environment in which atmospheric pollutants have initiated chemical reactions on a micro-level, implying that degradation could be further advanced in the exposed areas. The presence of metallic soaps, such as zinc azelates, and calcium and lead stearates indicates an active, acidic chemical environment within the paint layer. Zinc azelates form when Zn²⁺ ions (derived from the degradation of zinc-based pigments) react with azelaic acid, which is a byproduct of oil oxidation. This reaction serves as a marker for the degradation of both the pigment and the binder. Calcium stearate results from the reaction of fatty acids in degraded oil with the calcium ions from calcite. Similarly, lead stearates result from the reaction of Pb²⁺ ions, sourced from lead white or chrome yellow, with saturated fatty acids that are typically found in aged oil. The formation of these compounds has direct implications for the mechanical integrity of the painting, leading to increased fragility, a higher risk of localized detachment of the painting layer, modifications in surface texture (including protrusions and crystallizations), and alterations in optical properties and color perception. The identification of calcium oxalate (whewellite) in the samples taken from under the frame indicates a complex degradation history involving both biological factors (such as microbial colonization in the humid microclimate between the canvas and the frame) and advanced oxidative processes [37,46]. Its presence confirms that degradation processes were also active in the seemingly protected areas of the painting, highlighting the vulnerability of the materials to environmental and biological influences. Overall, the presence of these degradation products points to an active chemical environment in the painting layer, where interactions between the original components and environmental factors have resulted in significant structural and chromatic changes. These observations are crucial for assessing the conservation status and for informing future restoration and protection strategies.

In the case of Indian yellow, the difficulty in identifying specific markers (euxanthone [47,48]) via FTIR highlights the general challenge of analyzing organic materials in complex, aged matrices. The degradation of this pigment likely manifests as a gradual fading and loss of chromatic intensity, rather than through dramatic chemical transformations, a phenomenon that can only be inferred indirectly from the layer’s overall composition.

The data from chromatographic measurements, including the profiles of unsaturated fatty acids and degradation products, along with the historical context, suggest that a drying oil-type binder, most likely linseed oil, is involved. This type of binder is commonly found in 19th-century paintings. According to the literature, the palmitic-to-stearic acid (P/S) ratio in linseed oil can widely vary, ranging from 0.6 to 1.6, depending on the pigment type present in the mixture [42]. The low P/S ratio value of 0.83 in T_r sample can be linked to the presence of hematite in red ochre, which has been shown to decrease the P/S ratio in aged linseed oil [42]. The absence of stearic acid in T_g and H_g samples, coupled with the high P/S ratio in the H_b sample, indicates a depletion of stearic acid. This depletion may be related to the compositional variability of the pictorial layer, as interactions between the binder and pigments can significantly alter the detectable fatty acid profile. According to the literature, this behavior is not an anomaly, but rather an expected outcome of complex aging processes. The absence of stearic acid could also be attributed to the formation of metal soaps, identified by ATR-FTIR investigations that do not completely undergo esterification.

5. Conclusions

The present study conducts the first in-depth material investigation of the works of Theodor Aman (1831–1891), a pioneer of modernity in Romanian painting. It analyzes three representative works from different stages of his career: “Hora de peste Olt” (1866), “Teleleice în Harem” (1879), and “Regimul Vechi” (1881). Through a multi-analytical approach that includes complementary Raman spectroscopy (both portable and micro-Raman), FTIR-ATR, XRD, SEM-EDS, and GC-MS, the study aims to reconstruct the artist’s palette and execution techniques in detail.

The results reveal a clear evolutionary trajectory in Aman’s artistic practice. The early period, exemplified by “Hora de peste Olt,” features a diverse palette where traditional pigments coexist with modern materials such as Prussian blue and chrome and zinc yellows, along with exotic pigments like Indian yellow. This period is marked by a direct application technique. In contrast, the mature period, particularly represented by “Teleleice în Harem,” reflects a conscious refinement of the palette, simplifying it to its essentials, alongside a more sophisticated stratigraphic technique. This is highlighted by the introduction of an intermediate red base layer, indicating increased artistic control and the maturation of Aman’s plastic language. GC-MS analysis confirmed the presence of a drying oil-type binder in the samples, with a fatty acid profile compatible with linseed oil. Additionally, the identification of animal glue in some samples suggests the use of a mixed technique involving a primer or protein binder. The provisional presence of sandarac resin indicates the possible use of a protective varnish.

The study also highlights several post-execution degradation phenomena, including the formation of pararealgar from realgar, early transformations of lead white, and the appearance of metallic soaps (zinc azelates, and calcium and lead stearates) due to the interaction between the oxidized oily binder and the pigments. Furthermore, the presence of calcium oxalate (whewellite) serves as a marker of biological and oxidative degradation. These observations provide a detailed understanding of the material history of the works and the environmental factors that have affected them over time.

Although the study has some inherent limitations, such as microsampling being restricted to marginal areas (under the frame) and less detailed characterization of “Regimul Vechi”, the results obtained serve as an essential reference for art historians interested in Aman’s creative process and its integration into the European context, as well as for specialists in conservation and restoration. This research paves the way for more extensive investigations into the artist’s entire oeuvre, which could confirm and refine the trends identified in this study.