Artificial Ageing Study and Evaluation of Methods for Oil Removal on Decorative Plaster in Artistic Hispano-Muslim Artworks

Abstract

This study investigates Hispano-Muslim plasterworks, exemplified by the Cuarto Real de Santo Domingo, the Madraza, and the Alhambra in Granada, focusing on cleaning methods to remove oil-based repaintings without damaging the original polychromies. To this end, samples replicating traditional materials (gypsum coating, pigments, and binders) and techniques (tempera painting) were prepared and subjected to an artificial ageing protocol (AAP). Subsequently, cleaning tests aimed at removing the oil repaintings were performed to recover the original polychromies. Analytical techniques, including Fourier Transform Infrared Spectroscopy (FTIR), Gas Chromatography–Mass Spectrometry (GC–MS), and colorimetry, were employed to evaluate ageing effects and cleaning efficacy. Results revealed significant chromatic alterations in vermilion and azurite bound with animal glue, while ochre remained comparatively stable. Chemical analyses indicated marked binder deterioration, including protein denaturation in animal glue and oxidation/polymerization in linseed oil. Cleaning tests demonstrated that both a heptane–acetone gel and a novel polyamidoamine–glycine (PAAGLY) treatment effectively removed oil-based repaintings while preserving the original layers. These findings highlight the critical role of pigment–binder interactions in conservation strategies and advocate for selective, minimally invasive restoration methods.

1. Introduction

Over the centuries, plaster has been used for construction, as it is an easy and inexpensive material to work with, but it has also been widely used for decoration. In this last aspect, the Islamic tradition stands out in the elaboration of polychrome stucco coverings, made of gypsum, giving rise to the so-called Hispano-Muslim plasters, developed between the 12th and 15th centuries. This type of polychrome relief decoration, known as “yesería”, began with the arrival of the Muslims in the Iberian Peninsula and spread with Mudejar art into the Christian kingdoms. Among the earliest surviving examples are the “yeserías” found in the palatine city of Madinat al-Zahra (Córdoba, Spain) [1]. However, other very representative examples also stand out, such as those found in Granada: the Dar al Manjara l-kubra (Royal Room of Santo Domingo, Granada) Qubba from the last quarter of the 13th century [2], the Oratory of the Madrasa, or the palaces of the Alhambra, an example of Nasrid architecture (XIV–XV century) [3]. In Seville, it is possible to mention, among others, the plaster decorations of the Real Alcázar, including those of the Plaster Courtyard, the façade of Pedro I, or those of the Maidens’ Courtyard [4], or even in the Balearic Islands, where this type of tempera polychromy on a wood panel can also be found [5]. The traditional Islamic coverings were made with a characteristic technique, which implies the use of different materials in different layers: a first layer of plaster; a second layer with a preparation based on lime, gypsum, and the addition of an organic binder; and finally, a third layer made up of polychromy [6,7]. This third layer is applied with the animal glue tempera technique, used in the stuccos of the Alcazar of Seville [8] and in the Madraza of Granada [9], with Arabic gum identified in the Cuarto Real de Santo Domingo or in the Madraza of Granada [7,9], or with the egg tempera used in the Alhambra [6]. These binders were then mixed with a wide variety of pigments, such as cinnabar and vermilion as red, azurite as blue, malachite and verdigris as green, or carbon black as black, among others [10]. However, gypsum is a very porous and permeable material, a feature that does not favor its long-term conservation [11]. Added to this is the combination of this material with the polychromatic top layers, which contain organic binders and therefore can show poor stability over time. This means that the plaster can undergo alterations due to various factors such as biodeterioration, salts, humidity, and repolychromies, among others. These alterations can lead to material loss [6] (Figure 1).

One of the most common alterations is the repolychromies carried out in the 18th–19th centuries. In conservation and restoration, the application of these secondary layers involves a significant ethical responsibility, considering them from a dual perspective: on the one hand, their preservation as a historical part of the object, providing a faithful record of the transformations the work has undergone over the centuries, and on the other hand, their removal through cleaning treatments, due to the damage these layers may cause to the preservation of the work, in terms of both its material and technical integrity, as well as its interpretation and overall meaning [12]. In the case study presented in this research, we are faced with the second situation, the need to remove the applied layers. This decision is supported by two main reasons: on the one hand, these repolychromies were applied with different techniques and materials to the originals: linseed oil was used as a binder, with typical pigments of the 18th–19th centuries, such as synthetic ultramarine blue, synthetic azurite, ochre, or white lead [8] (Figure 2). The application of these overpaints is an important aspect in the long-term preservation of the plasters due to the different nature of the materials used in the original layers (tempera technique) and in the overpaintings (oil technique). This reduces the compatibility of both the polychrome techniques and the materials used, resulting in greater alteration and deterioration of the original layers [8]. On the other hand, the interventions were carried out according to the aesthetic preferences of the 18th and 19th centuries, and as such, the colors used in these layers often do not correspond to the original ones. As a result, the current appearance of the artwork differs significantly from its original state, leading to errors in dating and interpretation of this type of heritage [10]. This is the reason why the research presented in this study has been undertaken: to investigate cleaning treatments that allow for the removal of these layers in order to recover the original appearance of these artworks, and thereby reveal their authentic materials and techniques. However, there is a scarcity of studies on this specific issue. Therefore, the focus of this research is based on the necessity to study the cleaning treatments aimed at eliminating the superimposed layers on the original polychromies and preserve the original layer of the Hispano-Muslim plaster and to suggest a new protocol for cleaning.

Generally, the cleaning of gypsum-based polychrome artworks is a delicate process that requires thorough knowledge of materials and restoration techniques. Commonly used cleaning methods include solvents such as water or non-ionic surfactants (e.g., Triton X-100, which contains hydrophilic polyethylene oxide and hydrophobic groups) [13,14], as well as organic solvents such as alcohols (ethanol or isopropanol for removing natural resin-based varnishes), ketones (acetone, effective for synthetic resins but highly aggressive), and aliphatic hydrocarbons (heptane, white spirit, useful for removing dirt and some varnishes) [15,16]. The use of alkaline solutions like sodium hydroxide (NaOH) is generally discouraged due to their high reactivity; however, highly diluted solutions may be cautiously applied to remove specific types of dirt or patinas. Caustic soda can react with many pigments, altering their color and structure, while gypsum, commonly used as a base for polychromy, is soluble in alkaline media. Moreover, NaOH can damage organic binders, such as animal glues or oils, leading to paint layer detachment.

Therefore, less aggressive solvents such as aliphatic hydrocarbons (heptane, white spirit) are often preferred for dissolving non-polar materials, with the added benefit of high volatility, which limits solvent contact time with the artwork’s surface. These solvents are frequently used to remove altered varnishes, overpainting, or dirt [17]. The selection of solvents depends primarily on the solubility of the material to be removed and the sensitivity of the original substrates. This consideration is particularly important for gypsum, as its porous nature makes it prone to solvent penetration and potential damage. Traditionally, these treatments are applied using swabs, brushes, or poultices [13,14], which provide limited control over solvent penetration and action. Consequently, alternative application vehicles such as gels have emerged, allowing for more controlled and localized solvent delivery, thereby reducing the risk of damage to sensitive materials. Gels retain solvents and enable slower, more gradual removal of unwanted substances. Various gel formulations exist, based on cellulose derivatives, synthetic polymers, or other materials, which can be engineered to release solvents in a controlled manner [18,19]. Some of these materials have been studied for their application in cleaning treatments on gypsum works, such as cellulose ethers like methylcellulose and hydroxypropyl cellulose [20]. PVA–borax gels have also been shown to produce good cleaning results on surfaces as porous and hygroscopic as gypsum, such as in mural paintings on plaster mortar and polychrome tempera, which therefore have characteristics similar to those of Hispano-Muslim plaster [21]. Complex polysaccharide gels, especially agar-based gels, have also been widely studied as cleaning agents for removing overpainting and dirt from both polychromed [19] and unpolychromed plaster works [11].

Regarding all different types of gels, agar-based gels are increasingly used for the removal of surface patinas such as dust and soluble salts, with a gentle action that makes them particularly suitable for cleaning sensitive surfaces such as paintings or frescoes. Furthermore, it can be easily removed from the artwork’s surface after cleaning, without leaving undesired residues. In this research, many cleaning methods known in the literature, some of which have just been described, were studied and compared. In particular, the synergistic use of very promising materials for their respective intrinsic characteristics is proposed: polyamidoamine and agar-based gels. The use of glycine-functionalized polyamidoamines (PAA) for the capture of pigments from paint layers to be removed has been studied. Being an evolving field of research, the applications of this material in the restoration of artworks and in the removal of contaminants from delicate surfaces are being explored, although the chelating abilities of PAA have already been the subject of study [22,23]. The advantage of using PAAs lies in the possibility of creating gels with controllable properties, such as viscosity and swelling capacity. Furthermore, their chemical structure can be modified to optimize the interaction with specific materials. PAA gels act as matrices that trap pigments through chemical and physical interactions. The removal of pigments is achieved by applying the gel to the surface to be cleaned, followed by the removal of the gel that has incorporated the pigments.

Finally, it can be considered that this type of treatment can be used for the selective removal of undesired pigments from paintings, frescoes, and other polychrome artifacts, carefully selecting certain areas and cleaning delicate surfaces without damaging the original paint layers. To develop appropriate cleaning treatments, it is necessary to simulate the actual ageing of this type of plasterwork. For this reason, an innovative ageing protocol was proposed to subject the samples to climatic stress conditions that can occur in the Iberian Peninsula. Possible alterations undergone by materials during ageing were evaluated, especially for organic materials that tend to denature and lose physical properties. The samples were characterized before and after artificial ageing protocol (AAP), focusing on the organic components (animal glue, Arabic gum, and linseed oil). In this way, it is possible to establish a specific protocol that simulates the evolution and deterioration that occurs in the original materials. These characterizations were carried out by colorimetric evaluation, Fourier Transform Infrared Spectroscopy (FTIR), and Gas Chromatography coupled with Mass Spectrometry (GC–MS), both original polychromy and repaintings. Furthermore, using the same techniques, it is possible to evaluate the effectiveness of different methodologies for the removal of oil repainting from decorative plasters, minimizing damage to the original polychromes.

2. Materials and Methods

2.1. Sample Preparation

In this study, two samples were made to reproduce plasterwork of the Hispano-Muslim tradition of the 12th–15th centuries, called “P1” (sample 1) and “P2” (sample 2). The three layers that make up this type of plaster have been reproduced: the first layer corresponds to 80–90% Vigor^® (© VIGLIETTA Matteo SpA, Parma, Italy) plaster coating on a ceramic support of dimensions 30 × 15 × 1.5 cm, followed by the second layer, preparation layer, made with Vigor^® (© VIGLIETTA Matteo SpA, Parma, Italy)plaster mixed with lime and CTS^® (© Copyright C.T.S. España S.L., Madrid, Spain) animal glue as organic binder (v:v- 1:1:1/4) [7]. Finally, there is the third layer of polychromy, for which two representative pigments were selected: vermilion (42000 Kremer^®, © Kremer Pigmente, Aichstetten, Germany) and azurite (10200 Kremer^®,© Kremer Pigmente, Aichstetten, Germany) [10]. Both pigments were mixed with animal glue and Arabic gum as a binder, giving rise to a total of 4 evaluation samples (“Original layers” in

Table 1. Description of simulating Hispano-Muslim plaster samples.

Sample	Original Layer				Repainting
	Pigment	Binder	Band	Name	Pigment	Binder	Band	Name
P1	Azurite (a)	Animal Glue (Ca)	A	P1aCa(A)	Ultramarine Blue + Titanium White	Linseed Oil	B	P1aCa(B)
					Ochre	Linseed Oil	C	P1aCa(C)
		Arabic Gum (Ga)	A	P1aGa(A)	Ultramarine Blue + Titanium White	Linseed Oil	B	P1aGa(B)
					Ochre	Linseed Oil	C	P1aGa(C)
P2	Vermilion (v)	Animal Glue (Ca)	A	P2vCa(A)	Ultramarine Blue + Titanium White	Linseed Oil	B	P2vCa(B)
					Ochre	Linseed Oil	C	P2vCa(C)
		Arabic Gum (Ga)	A	P2vGa(A)	Ultramarine Blue + Titanium White	Linseed Oil	B	P2vGa(B)
					Ochre	Linseed Oil	C	P2vGa(C)

) (Figure 1): azurite with animal glue (v/v 1:1), azurite with Arabic gum (v:v −1:1), vermilion with animal glue (v:v −1:1/2), and vermilion with Arabic gum (v:v −1:1/2)

Subsequently, two oil colors were applied, simulating the types of alteration studied in this research: linseed oil binder (Ferrario^®, ©2025 Art & Colour, Kavala, Greece) mixed with ultramarine blue (45000 Kremer^®,© Kremer Pigmente, Aichstetten, Germany) and titanium white (46200 Kremer^®,© Kremer Pigmente, Aichstetten, Germany) (v:v- 1:1/3:1/4) and linseed oil mixed with yellow ochre pigment (40301 Kremer^®,© Kremer Pigmente, Aichstetten, Germany) (v:v −1:1/2). Each support was divided into three bands: band B, which corresponds to ultramarine blue and titanium white with oil binder; band A, which maintains a check of the original polychrome layer; in band yellow, ochre was applied with oil binder on top of the original polychrome layer (“Repainting” in Table 1) (Figure 3). The samples are stored under controlled laboratory conservation conditions, with 45–50% relative humidity, at a temperature of 20–21 °C, without natural or artificial exposure.

2.2. Artificial Ageing Protocol (AAP)

Each material responds differently to any type of environmental stress, such as climatic variations, extreme temperature exposures, and crystallization of salts dissolved in water, producing characteristic effects of a degradation process [24]. Nowadays, through experimental laboratory treatments and an artificial ageing test on samples, it is possible to observe and study the response of this material to the processes of degradation in progress. The different artificial ageing tests, coupled with physico-mechanical tests, could be a valuable tool for the study of the causes, processes, and effects of degradation on ornamental materials, like plasterwork, used in historic and modern buildings. Artificial ageing also allows the evaluation of the effectiveness of conservative treatments by reducing the risks of unsuccessful choices of products that are not appropriate to the real needs of historical works. Therefore, an appropriate ageing diagnostic program is a valuable tool for the quality and sustainability of conservative intervention.

At the present time, however, bibliographic research has revealed the fragmentary nature of such ageing tests and the lack of a unified protocol aimed at identifying and characterizing the degradation of ornamental stones. Moreover, although there are a limited number of norms and protocols, the literature offers some interesting tips and suggestions about the test models to be adopted and invites continuous testing to develop unified procedures. Considering these previous studies [25,26,27], a specific ageing protocol is proposed. Furthermore, this protocol is based on the study of the climatic conditions of temperature and relative humidity of those regions where the Hispano-Muslim plaster is found, in southern Spain. The ageing protocol is carried out after the application of the altered oil layers on the polychromed samples, which have previously undergone a 1-month natural curing process. The reason for performing the artificial ageing at this final stage of sample preparation, rather than before and after the application of this layer, lies in the nature of the materials used to reproduce the original stratigraphy. Both the gypsum and the binders used are highly sensitive to humidity above all, but also to temperature, which directly affects the preservation of this specific type of works [28]. Previous studies have also been taken into consideration, as they highlight the aggressive nature of applying an ageing protocol to these materials, which in this case must be artificial and accelerated, potentially causing destructive effects on the samples [25,26,27]. Furthermore, there is the added challenge of fully reproducing, both materially and technically, the typology of Hispano-Muslim gypsum artworks and subjecting them to ageing without any existing precedent. For this reason, the artificial ageing process is applied only after completing all the layers of the samples, including the alteration layer. In order to observe how each layer responds, the samples are divided into three study bands: original polychromy (B) and original polychromy with overpainting (A and C), as shown in Figure 3. Consequently, the developed protocol is divided into three phases to be applied daily for 1 week:

• Step 1: Freezer. The first phase consists of exposing the samples to sub-zero temperatures to simulate the temperature fluctuations typical of southern regions of Spain. To achieve this, the samples were placed in a freezing chamber for 3 h per day at an average temperature of −18 °C, with a temperature fluctuation range of ±2 °C.
• Step 2: Humidification chamber. The second phase provides for the introduction of the samples into a humidification chamber, for a total of 2 h per day, subjecting them to an environment with a saline humidity of 35 g/L of NaCl. This salinity concentration per liter is equivalent to that present in the Mediterranean Sea, which is found in the areas historically associated with the locations of the Hispano-Muslim plasterworks. The distribution of the saline mist in the humidity chamber is carried out in the form of water vapor, through two distribution systems located on the sides of the chamber and positioned above the samples. An amount of 1 L of saline solution is used over the course of 2 h, which corresponds to an administration rate of 0.39 g/L of saline solution per minute.
• Step 3: UV light lamp. The last stage reproduces the effect of UV light, using an OSRAM^® Ultra Vitalux 300 W 230 V E27 lamp. This lamp simulates sunlight and specializes in material ageing, with a UVA radiated power of 13.6W and a UVB of 3.0W. The samples were subjected to UV light for 5 h a day. The OSRAM Ultra Vitalux 300 W 230 V E27 UV lamp (© 2025, OSRAM GmbH, Getafe, Spain) emits ultraviolet radiation in two specific spectral ranges: UVB (280–315 nm): 3.0 W; UVA (315–400 nm): 13.6 W.

The ageing was stopped after completing 60 h in the first phase of the freezer, 40 h in the humidification chamber, and 100 h in the UV light lamp, for a total of 20 ageing cycles.

2.3. Test for Oil Repainting Removal

The choice of treatments focuses on the porous and elastic nature of the plaster and on the irregular nature of the materials used in the layers of original polychromy (tempera technique) and repainting (oily technique). Plaster cleaning studies are taken into consideration, and the use of 5% Fluka^® (© 2025 Thermo Fisher Scientific Inc, Alcobendas, Spain) hard agar gel is selected as a vehicle for the controlled application of solvents [11,29]. The application of solvents that allow the elimination of oily overpaintings, such as hydrocarbons, surfactants, bases, and the incorporation of the agar gel prepared with a 0.125 M solution of PAAGLY (polyamidoamine (© 2025 Merck KGaA, Darmstadt, Germany) and glycine (© 2025 Merck KGaA, Darmstadt, Germany) 50:50) is studied (

Table 2. Selected treatments applied to repainted plaster.

Medium	Agar 5% (w/v)								Test 5
Treatment	Test 1		Test 2		Test 3		Test 4
Cleaning agent	Heptane + Acetone (50:50 v/v)		Triton- X 100®		Heptane + Triton X 100® (50:50 v/v)		NaOH		PAAGLY Agar gel 5% (w/v)
Method	a	b	a	b	a	b	a	b	a	b
Quantity	150 µL	100 µL	150 µL	100 µL	150 µL	100 µL	150 µL	100 µL	-	-
Time	30’	20’	30’	20’	30’	20’	30’	20’	30’	20’
Wash solution	Ammonium hydroxide buffered with carbonic acid wash solution—medium swab

). PAAGLY solution was synthesized by a 1,4-conjugate addition reaction of glycine (GLY) to N,N′-methylenebisacrylamide (MBA), in water, in a 1:1 molar ratio. Briefly, the synthesis was performed in a flask, with a stirrer, thermometer, and dropper funnel. An appropriate amount of MBA powder was dissolved in double-distilled water and kept at 40 ± 1 °C for 15 min. GLY was diluted in a small amount of solvent with NaHCO₃ in equimolar ratio and slowly dropped into the MBA solution. The reaction was left under vigorous stirring at 55 °C for 4 h, until the solution became clear. The pH of the solution was 9.2.

For the treatments, two methods are evaluated: In method (a), all tests (1–5) were applied once for 30 min; specifically, tests 1, 2, 3, and 4 consist of applying 150 µL of cleaning agent using agar gel prepared at 5% in distilled water as a medium, while test 5 consists of preparing an agar gel in PAAGLY at 5% (Figure 4). Method (b) consists of reducing the action time to 20 min only for all tests (1–5), and specifically for tests 1–4, the amount of cleaning agent was reduced to 100 µL with the same medium (5% agar in distilled water), while test 5 was carried out using the same procedure, agar gel in PAAGLY at 5% (Figure 4). After the application of both methods, the gel residues were removed mechanically, and an ammonium hydroxide buffered with carbonic acid wash solution was applied with a swab (Table 2).

2.4. Methodology for Assessing the Effectiveness of Artificial Ageing Protocol (AAP) and Treatments for Oil Repainting Removal

2.4.1. Colorimetric Characterization

The changes in the surface appearance of the samples were determined by colorimetric measurements made with a TECHKON SpectroDens spectrodensitometer (TECHKON GmbH 2022 ©, Königstein, Germany). To determine the total color difference ΔE*_Lab according to UNE-EN 15886:2011 [30], 12 points of a few mm² of area were examined, and the average was calculated on each area, before and after the artificial ageing. The CIELab reference system was considered, using D65 (6500 K) as the standard illuminant and CIE 1964 (10°) as the standard observer. The obtained measurements were carried out excluding the specular component (SCE). The color study has been applied as an evaluation methodology to characterize the polychrome layers and repainting layers before and after the application of the artificial ageing protocol (AAP).

2.4.2. Fourier Transform Infrared Spectroscopy (FTIR)

The characteristic functional groups of the samples were studied with infrared spectroscopy (FTIR) in attenuated total reflectance (ATR) mode. Each spectrum was collected in the range 4000–400 cm⁻¹ using the PerkinElmer Spectrum Two™ spectrometer (Waltham, MA, USA). An average of 16 scans and a resolution of 2 cm⁻¹ were also used to obtain high-definition spectra. The samples of Band B (linseed oil with ultramarine blue pigment and titanium white pigment), control Band A of the original polychrome layer (vermilion or azurite with animal glue or Arabic gum as organic binder), and Band C (linseed oil with ochre) (Figure 1 and Table 1) are studied, before and after the ageing process. Data analysis was carried out with OMNIC 7.1 software. This analytical technique has been used for the characterization of the original polychrome layers and repainting layers before and after the application of the artificial ageing protocol (PAA), and to determine the effectiveness of this applied ageing methodology.

2.4.3. Gas Chromatography Coupled with Mass Spectrometry (GC–MS)

Separation and identification of organic compounds was performed using a gas chromatograph (7820A GC–MS, Agilent Technologies, Palo Alto, CA, USA) equipped with a Split/Splitless injection port and a mass spectrometer (GC/MSD 5977B, Agilent Technologies). An SLB^®-5MS capillary GC column (L × I.D. 30 m × 0.25 mm, df 0.25 μm; Sigma-Aldrich, Supelco, Darmstadt, Germany) was used for chromatographic separation. The carrier gas (He, purity 99.995%) was used in constant flow mode at 20 mL/min. The mass spectrometer was operated in the EI-positive mode (70 eV). The analytical procedure based on the GC–MS used for the analysis of lipids and proteinaceous fractions has already been described in the literature and is a widely regarded technique nowadays across the heritage science community for the detection and identification of organic components of paintings [31,32]. For the analysis of monosaccharides, the procedure used is described in the work of Bonaduce et al. [33]. The study for the determination of the linseed oil present in all repainted samples was possible thanks to the identification of the main components of polyunsaturated fatty acids, Linolenic C₁₈H₃₀O₂ (16–20%) and Oleic C₁₈H₃₄O₂ (22–42%), and saturated fatty acids, Palmitic C₁₆H₃₂O₂ (19–25%) and Stearic C₁₆H₃₆O₂. The characterization of the animal glue binders by GC–MS was performed, identifying the main amino acids present in the samples P1aCa(A) and P2vCa(A), such as glycine (32–35%), proline (15–21%), hydroxyproline (19–24%), glutamine (5–7%), and alanine (10–13%). For the Arabic gum study, the study was aimed at characterizing the monosaccharides that make up this type of gum in the analyzed samples P1aGa(A) and P2vGa(A). This analytical technique has been used for the characterization of the original polychrome layers and repainting layers before and after the application of the artificial ageing protocol (AAP) and to determine the effectiveness of this applied ageing methodology. This analytical technique has been used for the effectiveness of the application of cleaning treatments for the removal of oil repainting as well.

3. Results

3.1. Artificial Ageing Protocol (AAP)

3.1.1. Colorimetric Analysis

The results obtained from colorimetry after ageing are shown in

Table 3. Colorimetric measurements.

Sample	Before AAP			After AAP			Comparative
	L*(D65)	a*(D65)	b*(D65)	L*(D65)	a*(D65)	b*(D65)	ΔE*Lab	%ΔL*	%ΔC*	%ΔH*	total
P1aCa(A)	41.06	−10.32	−16.16	32.21	−8.18	−14.39	9.27	91.07	7.96	0.98	100
σ	0.83	0.60	0.38	2.28	0.71	0.91	2.11	5.79	6.64	1.25	-
P1aCa(B)	67.61	−4.37	−20.78	64.78	−7.31	−14.59	7.42	14.63	43.96	41.41	100
σ	3.06	0.54	4.05	3.40	0.80	4.31	0.86	17.36	17.02	5.32	-
P1aCa(C)	37.88	12.00	24.49	37.87	11.24	22.42	2.21	0.01	98.83	1.16	100
σ	3.71	1.19	4.95	3.70	0.93	3.81	3.31	26.71	32.46	25.97	-
P1aGa(A)	49.78	−8.24	−19.92	49.90	−8.63	−21.62	1.75	0.48	96.93	2.59	100
σ	3.04	0.63	1.17	1.75	0.95	1.23	1.80	36.14	35.02	4.10	-
P1aGa(B)	53.88	4.25	−44.60	52.66	3.35	−43.67	1.77	46.90	32.18	20.92	100
σ	0.85	0.64	0.74	0.82	0.67	0.71	0.46	29.81	22.03	11.53	-
P1aGa(C)	46.19	15.43	30.94	44.16	14.73	29.22	2.75	54.66	45.07	0.27	100
σ	0.79	0.49	1.13	1.77	0.92	1.63	1.48	25.10	23.76	2.02	-
P2vCa(A)	43.54	23.75	5.36	43.96	19.83	3.61	4.32	0.93	94.40	4.67	100
σ	6.15	12.34	13.11	6.51	12.79	13.20	3.06	25.33	36.52	31.80	-
P2vCa(B)	67.03	−1.99	−17.49	64.65	−4.10	−13.36	5.21	20.85	48.41	30.74	100
σ	3.07	0.53	3.79	3.10	0.91	4.54	1.42	27.82	28.39	9.33	-
P2vCa(C)	39.02	13.79	24.11	38.83	12.88	23.25	1.27	2.28	89.29	8.43	100
σ	3.20	1.54	4.67	4.21	1.50	3.93	1.89	33.60	34.75	5.85	-
P2vGa(A)	43.18	36.00	17.18	47.08	33.44	16.19	4.77	66.82	32.97	0.20	100
σ	2.08	2.27	1.94	4.82	3.72	2.45	5.14	24.63	20.97	8.91	-
P2vGa(B)	53.78	5.35	−45.50	51.55	3.32	−41.46	5.04	19.65	70.06	10.29	100
σ	1.51	0.99	2.44	1.47	1.30	2.93	2.73	21.62	19.77	10.06	-
P2vGa(C)	46.23	16.00	32.17	43.97	15.26	28.98	3.98	32.11	63.85	4.05	100
σ	0.73	0.35	0.71	1.45	1.04	2.07	2.47	13.30	15.54	11.67	-

. The data in the table show the mean values obtained for each band (A, B, and C) and their respective standard deviation values for each specimen before and after applying the ageing process.

These results show the values of ΔE*_Lab, referring to the chromatic variation experienced by the samples studied before and after applying the ageing process. Of the twelve samples analyzed, nine of them show ΔE*_Lab values higher than 2, which indicates that the color variation produced is perceptible and easily appreciable [30,34,35]. The results obtained from colorimetry after ageing (Table 3 and Figure 5) show that, in the case of the original pigments, the vermilion pigment shows a noticeable chromatic alteration, with both Arabic gum binder and animal glue, as the ΔE*_Lab values are higher than 2 in both cases: P2vCa(A) with animal glue binder presents ΔE*_Lab = 4.32, while with Arabic gum binder (P2vGa(A)), ΔE*_Lab = 4.77. However, the chromatic variations have occurred for different reasons: in the case of the sample with animal glue, the chromatic change is mainly due to variations in the chroma coordinate C* (%Δ C*); in this case, it indicates a variation in the intensity of the color, specifically a decrease in the intensity of the red component of the positive a-coordinate (a*(D65)). In the case of Arabic gum-bonded vermilion, the changes are mainly due to changes in color clarity (%ΔL*), as there is an increase in the L-coordinate, which shows that there is an increase in color clarity. Additionally, to a lesser extent, there is also a change in chroma (%ΔC*) due to a decrease in the intensity of the red color, as was also the case with Arabic gum (Table 3).

In the case of the azurite pigment mixed with Arabic gum, it has a stable behavior to ageing (ΔE*_Lab = 1.75), while mixed with animal glue, the polychromy is chromatically altered (ΔE*_Lab = 9.27). In the case of animal glue mixed with azurite (P1aCa(A)), the changes are mainly due to changes in color clarity (%ΔL*), as there is a decrease in the L-coordinate, which shows that there is a decrease in color clarity. In the case of the alterations studied, the repaint layer applied with linseed oil and ultramarine blue mixed with titanium white, except for being applied on azurite agglutinated with Arabic gum (ΔE*_Lab = 1.77), the rest of the samples show values of ΔE*_Lab higher than 2: P1aCa(B) ΔE*_Lab = 7.42; P2vCa(B) ΔE*_Lab = 5.21; and P2vGa(B) ΔE*_Lab = 5.04. In all three cases, the chromatic changes in this overpainting are due to variations in chroma (*%ΔC*), specifically a decrease in color saturation or intensity, mainly caused by a reduction in the negative b parameter (b*(D65)), which indicates a decrease in the intensity of the blue component. Finally, in the case of the ochre repainting with linseed oil, except for the sample P2vCa(C) with the value ΔE*_Lab = 1.27, the rest of the samples show values greater than 2: P1aCa(C) ΔE*_Lab = 2.21; P1aGa(C) ΔE*_Lab = 2.75; and P2vGa(C) ΔE*_Lab = 3.98. In all these three cases, the chromatic changes in this overpainting are also due to variations in chroma (%DC*), specifically a decrease in color saturation or intensity, mainly caused by a reduction in the positive b parameter (b*(D65)), which indicates a decrease in the intensity of the yellow component (Table 3 and Figure 5).

3.1.2. Fourier Transform Infrared Spectroscopy (FTIR)

The study of the samples by FTIR spectroscopy allowed for identifying the main absorption bands and peaks of the binders before applying the artificial ageing protocol. In the samples made on animal glue (samples P1aCa(A) and P2vCa(A)), the results show the characteristic groups of amide A at 3642 cm⁻¹ (P2vCa(A)) and 3418 cm⁻¹ (P1aCa(A)), amide I at 1623 cm⁻¹ (P1aCa(A)) and 1618 cm⁻¹ (P2vCa(A)), and amide II at 1565 cm⁻¹ (P1aCa(A)), which are signed with arrows in Figure 6 [36,37]. For Arabic gum samples studied before ageing (P2vGa(A) and P1aGa(A)), the main functional groups were identified, such as the hydroxyl group at 3425 cm⁻¹ (P1aGa), the bands 1457 cm⁻¹ (P2vGa(A)) and 1463 cm⁻¹ (P1aGa(A)) attributed to C=C conjugated double bond stretching, and glycosidic linkage indicated by the band at 1015 cm⁻¹ (P1bGa) and at 1009 cm⁻¹ (P1aGa(A)) (pink shaded area in Figure 7a,c) [38,39]. For linseed oil repainting samples before ageing (P1aGa(B), P1aCa(B), P1aGa(C), P1aCa(C), P2vGa(B), P2vCa(B), P2vGa(C), and P2vCa(C)), the signals assigned to the lipid fraction are evident in the doublet bands at 2920 cm⁻¹–2857 cm⁻¹ due to the long-chain aliphatic group bands, referring to long hydrocarbon chains, such as those found in triglycerides of unsaturated fatty acids (like linolenic acid), which make up linseed oil. Additionally, the characteristic band at 1742 cm⁻¹ of the carbonyl group has been also identified (green rectangles in Figure 6a–d and Figure 7a,c) [40].

Following the artificial ageing protocol, FTIR analysis revealed significant chemical changes in the binders used in the simulated polychrome samples. For the animal glue binder, particularly in the aged vermilion sample (P2vCa(A) aged), a marked reduction in the absorbance of the amide II band was observed, indicating partial denaturation or degradation of proteinaceous components. In the case of the aged azurite sample bound with animal glue (P1aCa(A) aged), the response was comparatively more stable, although a reduction in both amide I and II bands was still evident, suggesting structural alterations of the protein matrix, albeit to a lesser extent [36,37] (Figure 6).

In the case of Arabic gum, the P1aGa(A) aged sample with azurite, a decrease in the hydroxyl group of 3425 cm⁻¹ can be observed, which may be attributed to dehydration or partial loss of a hydrogen-bonded structure. A reduction in the stretching vibrations of the methyl and ethyl groups ν(CH3) and ν(CH2) at 2988 cm⁻¹ (P1aGa(A)) is also observed. The bands 1457 cm⁻¹ (P2vGa(A)) and 1463 cm⁻¹ (P1aGa(A)) attributed to C=C conjugated double bond stretching and glycosidic linkage at 1015 cm⁻¹ (P1bGa), at 1009 cm⁻¹ (P1aGa(A)), remain stable [39,41]. However, in the aged vermilion sample with Arabic gum (P2vGa(A) aged), the results were inconclusive due to interference from the pigment, which hindered the detection of gum-related spectral variations (Figure 7).

Regarding linseed oil, aged samples based on Arabic gum (P1aGa(B) aged and P1aGa(C) aged) showed a notable reduction in the carbonyl band at 1742 cm⁻¹. Furthermore, a decline in the alkyl group bands (2930–2855 cm⁻¹) was recorded in aged samples with animal glue as the underlying binder (P2vCa(B) aged and P2vCa(C) aged). A reduction in the aforementioned functional groups can be attributed to the partial evaporation of the organic medium, resulting from the decomposition of triglycerides of unsaturated fatty acids and the formation of volatile compounds induced by UV irradiation [36]. In other samples, an increase is observed in the 1750–1700 cm⁻¹ region, as seen in P1aCa(C), P2vCa(B), P2vCa(C), and P1Ga(C). This may be attributed to the formation of carbonyl groups, specifically ketones, esters, and carboxylic acids, during oxidative polymerization. The formation of ester bonds, in particular, plays a crucial role in the oxidative polymerization processes of lipid-based materials such as linseed oil [42]. In addition, new functional groups related to the hydroxyl group have been identified, 3550 cm⁻¹ and 3600 cm⁻¹ (P1aCa(C) aged, 3600 cm⁻¹ (P1aGa(C), 3620 cm⁻¹ (P2v Ga(C), 3490.1 cm⁻¹ (P1aGa(C), and 3423 cm⁻¹ (P1aGa(B), arising as a result of the oxidation and polymerization processes characteristic of linseed oil ageing (Figure 6 and Figure 7) [36,43]. These chemical transformations were particularly pronounced in samples containing ultramarine blue and titanium white, whereas ochre-based samples demonstrated greater stability under the same ageing conditions.

These groups of the Arabic gum binder and animal glue binder have been identified in a real polychrome of plasterworks, with examples such as those shown in Figure 8 from the Oratory of the Madraza. In the case of the Arabic gum binder, the components identified are the hydroxyl group at 3432 cm⁻¹, attributed to C=C conjugated double bond stretching at 1419.76 cm⁻¹, and the carbonyl group at 1031 cm⁻¹ at 988 cm⁻¹ (Figure 8a). An animal glue binder is possible to identify the characteristic groups of amide I at 1618.54 cm⁻¹ and amide II at 1543.90 cm⁻¹ (Figure 8b). The last binder analyzed, linseed oil, is also possible to identify in a real repolychrome of plasterworks from the Oratory of the Madraza and the Courtyard of the Maidens at the Royal Alcázar of Seville, mixed with artificial ultramarine blue pigment. As seen in the studied cases, the hydroxyl group has been identified due to oxidation and polymerization at 3522.84 cm⁻¹ and 3400.35 cm⁻¹, the characteristic band of oil referent to the carbonyl band at 1683 cm⁻¹ (Figure 8c).

3.1.3. Gas Chromatography Coupled with Mass Spectrometry (GC–MS)

The chemical stability of the animal glue binder was further evaluated by Gas Chromatography–Mass Spectrometry (GC–MS), focusing on its amino acid composition. The samples P1aCa(A) and P2vCa(A), analyzed before and after artificial ageing, consistently revealed glycine (32–35%), proline (15–21%), hydroxyproline (19–24%), glutamine (5–7%), and alanine (10–13%) as the dominant amino acids (

Table 4. Comparison of GC–MS results of animal glue amino acids before and after ageing.

Amino Acids	P1aCa(A)	P2vCa(A)	P1aCa(A) Aged	P2vCa(A) Aged
	µg/100 µg
ALA	3.6	1.6	2.6	2.7
GLY	10.8	4.1	4.1	5.8
LEU	0.5	0.1	0.8	0.6
PRO	6.7	2.0	3.0	3.8
HPRO	6.4	2.9	5.3	3.9
ASP	0.5	0.3	0.2	0.9
GLU	2.4	0.8	0.9	1.3
PHE	0.0	0.0	0.0	0.0
TOT	30.9	11.7	16.8	19.0
MED	22.0		22.7

). These amino acids are characteristic of collagen-derived materials and are indicative of the binder’s proteinaceous origin [44]. Following 1 week of artificial ageing, the overall proportion of amino acids remained relatively stable, averaging 22% across all samples. This suggests that, despite the FTIR evidence of partial protein denaturation (i.e., reduction in amide I and II bands) (Figure 6) [36,37], the primary amino acid profile of the binder was largely preserved. This apparent contradiction can be interpreted as surface-level degradation or conformational changes in the protein structure, rather than a complete breakdown of amino acid chains. It is because it has been demonstrated that the main amino acids composing animal glues, such as alanine (Ala), glycine (Gly), leucine (Leu), and proline (Pro), remain stable against long-term ageing [45].

In the case of Arabic gum binder, the components detected before and after ageing revealed the following: the P2bGa(A) sample was galactose (75%), arabinose (20%), rhamnose (2%), and glucose (3%), while in the P1aGa(A) sample, galactose accounted for 40%; arabinose, 33%; rhamnose, 23%; and glucose, 3% (

Table 5. Comparison of GC–MS results of Arabic gum monosaccharides before and after ageing.

Monosaccharides	P1aGa(A)	P2vGa(A)	P1aGa(A) Aged	P2vGa(A) Aged
	µg/100 µg
Arabinose	0.1	0.4	0.1	0.2
Rhamnose	0.1	0.0	0.1	0.1
Galactose	1.2	1.6	1.2	1.6
Glucose	0.0	0.1	0.0	0.0
TOT	1.4	2.1	1.4	1.9
MED	1.6		1.7

). These monosaccharides are typical constituents of arabinogalactan polysaccharides, which form the structural backbone of Arabic gum. Regarding ageing, a general trend was observed in which galactose remained the predominant monosaccharide, maintaining levels between 40% and 50% [46]. However, other sugar components, particularly arabinose and rhamnose, showed a reduction, suggesting selective degradation or loss of specific sugar units within the polysaccharide matrix. This pattern indicates that, while the overall polysaccharide framework remains relatively intact, certain labile components are susceptible to degradation under ageing conditions [47]. These findings align with FTIR results showing decreased absorbance in hydroxyl group regions (around 3425 cm⁻¹), consistent with the dehydration or breakdown of hydroxyl-rich polysaccharide chains [41].

Finally, for linseed oil, there is a chemical change in its composition, with the reduction in monounsaturated fatty acid components, especially in the case of oleic acid. This reduction reflects the susceptibility of these unsaturated components to oxidative degradation processes. Concurrently, there was an increase in saturated fatty acids such as palmitic acid, as well as in dicarboxylic acids, including azelaic, suberic, and sebacic acids. The accumulation of these dicarboxylic acids is a well-established marker of oxidative cleavage and polymerization within drying oils, indicating progressive ageing and chemical transformation. This chemical evolution influences not only the mechanical properties but also the optical characteristics of the paint layer, potentially contributing to embrittlement and color changes observed in aged samples [43] (

Table 6. Comparison of GC–MS results of linseed oil before and after ageing.

Fatty Acids	P2aGa(B)	P2aGa(C)	P2aGa(B) Aged	P2aGa(C) Aged
	% Rel Lip Tot	% Rel Lip Tot	% Rel Lip Tot	% Rel Lip Tot
C8D (suberic acid)	2.81	0.68	-	2.59
C9D (azelaic acid)	20.23	9.10	8.53	12.28
C10D (sebacic acid)	0.43	0.15	-	-
C16 (palmitic acid)	25.79	19.43	32.60	20.88
C18:1 (oleic acid)	22.30	42.26	18.98	40.50
C18 (stearic acid)	20.90	16.52	28.65	17.52
C18:2 (linolenic acid)	20.90	16.52	28.65	17.52

). These results complement FTIR findings that show new hydroxyl group bands and reductions in carbonyl and alkane groups, collectively confirming oxidative polymerization [36,43]. The integrated data thus provide a comprehensive picture of the degradation pathways affecting linseed oil binders under ageing conditions.

3.2. Test of Repainting Remotion

The results obtained from the application of the selected treatments (Figure 9) are shown in

Table 7. Results and comparison of the best cleanings to remove the repainting.

Test		Medium Cleaning—Agar 5% DI Water								Test 5 PAAGLY Agar 5%
		Test 1 Heptane + Acetone (50:50)		Test 2 Triton- X 100®		Test 3 Heptane + Triton X 100® (50:50)		Test 4 NaOH 0.1 M
Method		b	a	b	a	b	a	b	a	b	a
		100 µL	150 µL	100 µL	150 µL	100 µL	150 µL	100 µL	150 µL	-	-
		20’	30’	20’	30’	20’	30’	20’	30’	20’	30’
Samples	P1aCa(B)	R	✓✓	✓	R	✓	✓	✓✓	✓	✓	✓
	P2vCa(B)	R	✓✓	✓	X	✓	✓	R	✓	R	✓✓
	P1aGa(B)	✓	✓✓	✓	X	R	✓	X	R	R	✓
	P2vGa(B)	X	✓✓	R	R	R	X	X	R	X	X
	P1aCa(C)	✓	✓✓	X	X	✓	X	X	✓	✓✓	R
	P2vCa(C)	✓	✓✓	R	R	X	R	✓	R	✓	✓✓
	P1aGa(C)	✓✓	X	R	X	X	X	X	X	R	X
	P2vGa(C)	✓	✓✓	R	R	X	R	✓	R	✓	✓✓
	General	✓	✓✓	R	X	X	X	X	R	✓	✓✓

✓✓: effective cleaning removes repainting without damaging the original polychromy. ✓: effective cleaning removes repainting but damages the original polychromy. R: very aggressive cleaning, removes original layer. X: ineffective cleaning.

. These results make it possible to establish the most effective tests to remove both overpaintings (Figure 10):

• Test 1: agar gel 5% DI water medium with heptane and acetone (50:50) applied at 20 min and 30 min.
• Test 5%: agar gels in PAAGLY, 20 and 30 min of application.

For the specific case of removing oil paint with ultramarine blue and titanium white, in addition to the results outlined above, test 4 can also be highlighted: agar gel 5% + 150 µL of NaOH (0.1 M) applied at 30 min and agar gel 5% + 100 µL of NaOH (0.1 M) applied for 20 min (Figure 11).

Gas Chromatography Coupled with Mass Spectrometry (GC–MS)

Once the cleaning was completed, the mechanically removed remains of the repaint, collected by swabbing, were studied by Gas Chromatography coupled with Mass Spectrometry (GC–MS) to quantitatively assess the efficacy and selectivity of the treatments. This analysis focused on determining the relative amounts of fatty acids, constituents of the linseed oil-based repaint, and compounds originating from the original polychromy, specifically amino acids characteristic of animal glue and polysaccharides indicative of Arabic gum. By quantifying these molecular markers, this study was able to differentiate between the removal of the overpaint and any unintended loss of the original binder materials. The data presented in

Table 8. Results of the cleanings in removing the repaintings.

Treatments	Fatty Acid Removed (%)	Amino Acid Removed (%)	Polysaccharides Removed (%)
Test 1: Heptane/Acetone	98	6	13
Test 2: Triton-X100	99	56	60
Test 3: Heptane/Triton-X100	95	52	45
Test 4: NaOH	86	21	30
Test 5: PAAGLY	88	6	15

report the percentage of removal for both repaint and original components, providing critical insight into the cleaning performance. A high percentage of fatty acid removal indicates effective elimination of the oil-based repaint layers, whereas minimal removal of amino acids or polysaccharides reflects a conservative action that preserves the integrity of the underlying original polychromy. Table 8 shows the results expressed in percentage of removal.

This approach allows for an objective evaluation of cleaning protocols in terms of selectivity and preservation, which is fundamental in conservation treatments where the balance between effective repaint removal and safeguarding original material is essential. The results guide the optimization of cleaning methods to maximize paint layer recovery while minimizing damage to historic substrates. Accordingly, these results show that the treatments that allow the removal of the greatest amount of repaint (fatty acids), and without altering the original layer corresponding to animal glue (amino acids) or Arabic gum (polysaccharides), are test 1 using a hydrocarbon mixture of heptane and acetone at 50% and test 5 using polymer PAA (0.125 M) and the amino acid glycine (0.125 M) at 50%.

4. Discussion

The authors should discuss the results and how they can be interpreted from the perspective of previous studies and of the working hypotheses. The findings and their implications should be discussed in the broadest context possible. Future research directions may also be highlighted.

4.1. Artificial Ageing Protocol (AAP)

4.1.1. Colorimetric Analysis

With regard to the color, after ageing, in the case of the original polychrome bands (Band A), it is possible to establish with respect to the pigments used, azurite and vermilion, that both are altered by the ageing process, especially when they are bound with animal glue [36]. In the case of azurite, the chromatic alteration occurs when agglutinated with animal glue and is due to a loss of clarity (L*(D65)), indicative of a discoloration of the pigment, due especially to the relative humidity applied during the ageing process. In the case of the vermilion pigment, there is a loss of clarity, and especially chroma (loss of red), with both binders; this is due to the combination of UV light and relative humidity applied during the ageing process, as both parameters generate a process of photodegradation and oxidation in this pigment [48,49]. In the case of the band referring to the linseed oil repainting, it is possible to appreciate that the ochre pigment, due to its composition FeO(OH) + Al₂O₃ + SiO₄ + CaCO₃), is more stable to the ageing process than the mixture of artificial ultramarine blue pigment (Na_8-10Al₆Si₆O₂₉S_2-4) with titanium white. This is because the ultramarine pigment can be affected by the acidity of the binder used [49].

4.1.2. Fourier Transform Infrared Spectroscopy (FTIR) and Gas Chromatography Coupled with Mass Spectrometry (GC–MS)

The study of chemical alterations through Fourier Transform Infrared Spectroscopy (FTIR) and Gas Chromatography coupled with Mass Spectrometry (GC–MS) shows that, in the case of the pigments used (azurite, vermilion, ochre, and the mixture of ultramarine blue and titanium white), they have not been chemically affected by the ageing process. As previously stated, the change experienced is chromatic and is directly related to the binders used. Therefore, it is necessary to analyze what potential chemical changes these organic components may have undergone.

In the case of the animal glue binder, the results obtained through Fourier Transform Infrared Spectroscopy (FTIR) revealed a marked decrease in the absorbance of the amide II band in the aged sample containing vermilion and animal glue (P2vCa(A) aged). This suggests partial denaturation or degradation of the proteinaceous components and, therefore, a chemical alteration. In comparison, the aged sample with azurite and animal glue (P1aCa(A) aged) exhibited greater chemical stability, although a reduction in both amide I and II bands was still observed. This indicates that, although the structural damage to the protein matrix was less severe, there was still an alteration in the binder’s conformation. Such chemical structural changes are consistent with the expected effects of accelerated ageing, particularly under conditions of UV radiation and elevated temperatures, as commonly observed in animal glue binders [36,37]. On the other hand, amino acid composition analyses using Gas Chromatography Mass Spectrometry (GC–MS) after artificial ageing revealed notable stability in the proportion of the main amino acids, glycine, proline, hydroxyproline, glutamine, and alanine, with an average preservation rate of 22% across all samples. This stability suggests that, despite the chemical changes indicated by Fourier Transform Infrared Spectroscopy (FTIR), no significant degradation of the amino acid chains comprising the animal glue occurred. The apparent contradiction between Fourier Transform Infrared Spectroscopy (FTIR) and Gas Chromatography coupled with Mass Spectrometry (GC–MS) data may be interpreted as surface-level degradation or conformational changes in the protein structure, rather than complete breakdown of the peptide chains. This hypothesis is supported by previous studies that demonstrated the relative stability of certain amino acids, such as alanine, glycine, leucine, and proline under ageing conditions [43,50]. Therefore, animal glue shows initial signs of chemical deterioration, evidenced by the loss of components such as amide B and amine III, along with the reduction in the amide A, I, and II bands. This is mainly attributed to exposure to heat and UV radiation during phase 3 of the artificial ageing process [51]. A physical alteration is also evident, as shown in Figure 10, where cracking and fissures caused by thermal and UV exposure can be observed.

In the case of Arabic gum, Fourier Transform Infrared Spectroscopy (FTIR) analyses show a decrease in the hydroxyl group band at 3425 cm⁻¹, which was observed in the aged sample P1aGa(A) with azurite. This reduction may be attributed to dehydration or to the partial loss of a structure stabilized by hydrogen bonding. Meanwhile, the bands at 1457 cm⁻¹ (P2vGa(A)) and 1463 cm⁻¹ (P1aGa(A)), associated with the stretching of conjugated C=C double bonds, as well as the glycosidic linkage bands at 1015 cm⁻¹ (P1bGa) and 1009 cm⁻¹ (P1aGa(A)), remained stable [39,41]. The decrease in hydroxyl signal is mainly related to UV exposure during artificial ageing, since Arabic gum has demonstrated chemical stability under thermal ageing conditions [39,41,52]. However, in the aged vermilion sample with Arabic gum (P2vGa(A)), the results were inconclusive due to pigment interference, which hindered the detection of spectral changes attributable to the binder (Figure 7). Gas Chromatography coupled with Mass Spectrometry (GC–MS) results showed that the main components identified before and after ageing in the sample P2bGa(A) were galactose (75%), arabinose (20%), rhamnose (2%), and glucose (3%), while in the sample P1aGa(A), the composition was galactose (40%), arabinose (33%), rhamnose (23%), and glucose (3%) (Table 5). These monosaccharides are typical of arabinogalactan polysaccharides, the principal structural components of Arabic gum. Regarding ageing effects, galactose remained the predominant monosaccharide, with consistent levels between 40% and 50% [46], while other sugars, particularly arabinose and rhamnose, showed a decrease, suggesting selective degradation or loss of specific sugar units within the polysaccharide matrix. This pattern indicates that, while the overall polysaccharide structure remains relatively intact, certain components may be more susceptible to degradation, especially under UV exposure [46]. Taken together, the spectroscopic and chromatographic data suggest that Arabic gum exhibits chemical stability under accelerated ageing conditions [46], undergoing primarily physical changes, specifically, the loss of hydroxyl groups associated with hydrogen bonding due to the removal of hydration water used during its preparation. Thus, the observed alteration in Arabic gum is best interpreted as a drying process following step 3 of the ageing protocol, corresponding to the UV exposure phase [39,41,46,52].

The results obtained from Fourier Transform Infrared Spectroscopy (FTIR) analyses indicate that linseed oil undergoes both chemical and physical changes following artificial ageing. Specifically, the samples aged with Arabic gum as the binder (P1aGa(B) aged and P1aGa(C) aged) showed a notable reduction in the carbonyl band (~1742 cm⁻¹), suggesting a loss of functional groups characteristic of the triglycerides present in the oil. Additionally, a decrease in alkyl group bands (2930–2855 cm⁻¹) was observed in samples with animal glue as the underlying binder (P2vCa(B) aged and P2vCa(C) aged). This reduction can be attributed to partial evaporation of the organic medium, resulting from the decomposition of triacylglycerols of unsaturated fatty acids (linolenic, linoleic) and the formation of volatile compounds under UV radiation exposure [36]. In contrast, other samples showed an increase in the 1750–1700 cm⁻¹ region, associated with carbonyl groups such as ketones, esters, and carboxylic acids. This increase, identified in P1aCa(C), P2vCa(B), P2vCa(C), and P1Ga(C), is indicative of oxidative polymerization processes, in which ester bond formation plays a fundamental role in the chemical transformation of linseed oil [42]. The appearance of these groups can also be attributed to the formation of carboxylic and dicarboxylic acids (e.g., azelaic acid), which are typically generated during linseed oil oxidation [36]. This chemical evolution was further confirmed by the presence of new hydroxyl-related bands in the 3550–3423 cm⁻¹ region, linked to advanced oxidation products. These bands were observed in multiple aged samples, including P1aCa(C), P1aGa(C), P2vGa(C), and P1aGa(B), demonstrating the progressive oxidation of the oil (Figure 6 and Figure 7) [36,42,43]. From a compositional standpoint, Gas Chromatography coupled with Mass Spectrometry (GC–MS) analysis revealed a decrease in monounsaturated fatty acids, particularly oleic acid, along with an increase in saturated fatty acids, such as palmitic acid, and in dicarboxylic acids, including azelaic, suberic, and sebacic acids. These dicarboxylic acids are well-established markers of oxidative degradation in drying oils, confirming a substantial chemical transformation of the binder during the ageing process [53,54]. Therefore, the combined findings from Fourier Transform Infrared Spectroscopy (FTIR) and Gas Chromatography coupled with Mass Spectrometry (GC–MS) analyses confirm that linseed oil undergoes a progressive chemical deterioration. These chemical processes have direct implications for the physical behavior of the paint material. The advancement of oxidative polymerization and the accumulation of degradation products ultimately lead to physical alteration, manifested as cracking and craquelure on the paint surface.

These findings emphasize the complexity of binder ageing, where molecular rearrangements and loss of structural integrity can occur without significant compositional loss, and underline the importance of employing both spectroscopic and chromatographic techniques for a comprehensive evaluation. In addition, the results of the ageing process have a direct impact on the selection and application of cleaning treatment. For all of the three binders analyzed, the changes they have undergone, both chemically and physically, as can be seen in Figure 5, where cracking, fissures, and even losses of the polychrome layer have occurred have a direct impact on the proposed cleaning treatments. For linseed oil, the chemical changes that occur during ageing result in modifications to its solubility and polarity properties, which in turn affect the range of compatible solvents that can aid in its removal [17]. In the case of animal glue and Arabic gum binders, denaturation and drastic drying make them more sensitive to cleaning treatments, requiring careful control over the application time and the vehicle used to apply these treatments. Hence, it is importance to understand these changes in order to select appropriate cleaning treatments—in this case, solvents compatible with the removal of aged linseed oil and vehicles such as gels that allow control over the action of these solvents to ensure the preservation of the original polychromy.

4.2. Test for Oil Repainting Removal

The elimination of the two types of repainting were test 1 and test 5. In the case of the use of organic solvents such as acetone (low polarity solvent) and the hydrocarbon heptane (non-polar) of the alkane family (test 1), they are more effective for the dissolution of the oil repaint, as the mixture of both is within the solubility range of the aged and unaged oil repaints. In addition, both are organic solvents of neutral character, which allows the interaction with the overpaint to be due only to these solubility parameters, and thus, the removal of the overpaint is due to the swelling and solubility of the overpaint by physical means, without having a chemical interaction with the overpaint [17]. In the case of the other treatments (test 2 and test 3), the use of the surfactant Triton-X100^®, even though it is non-ionic and allows the removal of fatty elements in an aqueous environment, due to its chemical composition, it is very toxic. Moreover, in the case of oil removal, it can be very aggressive, and residues may even remain, even if a subsequent washing is carried out (as was performed in this study), which may alter the appearance of both layers to be removed and the original surface [55,56].

For the treatment using sodium hydroxide (Test 4), the impact was very aggressive due to the nature of the selected treatment, a strong base. Although the alkaline aqueous environment can interact with the acid molecules that make up the linseed oil used for oil repainting, working in an environment with a pH > 9 causes ionization on the oil to be removed, penetrating and acting on a chemical level also on the original layer (polychromies with Arabic gum and animal glue) [17]. Finally, the proposal to apply a mixture of the polymer polyaminoamide (0.125 M) and the amino acid glycine (0.125 M) at 50% has proven to be effective for the removal of the oil repaint as it can interact with the fatty acids that make up the linseed oil.

5. Conclusions

The findings of this study demonstrate that the artificial ageing protocol (AAP) exerts a significant impact on the organic binders used to reconstruct the original polychromy layers (Band A), namely, Arabic gum and animal glue. In contrast, the inorganic pigments employed in the samples, azurite, vermilion, ochre, and the mixture of artificial ultramarine blue with titanium white, did not undergo detectable chemical alterations. The observed changes in these pigments were exclusively chromatic in nature, indicating that the degradation processes primarily affect the optical properties of the materials, rather than their molecular structure.

• Chromatic Level: In the case of the original polychromy (Band A), azurite bound with animal glue exhibited a marked loss of lightness (L* parameter), which can be attributed to the effects of relative humidity during the ageing process. Vermilion, on the other hand, showed a reduction in both lightness and, more notably, chroma (saturation of the red component) when combined with either Arabic gum or animal glue. This behavior is associated with combined photodegradation and oxidation processes induced by exposure to ultraviolet light and high humidity conditions. In the oil-based overpaint layers (Band B), the ochre pigment demonstrated greater stability against artificial ageing, a result likely due to its mineral composition. In contrast, the mixture of artificial ultramarine blue with titanium white proved more susceptible to degradation, showing sensitivity to the acidity of the binder, which highlights its lower resistance under simulated deterioration conditions.
• Chemical Level: Fourier Transform Infrared Spectroscopy (FTIR) and Gas Chromatography coupled with Mass Spectrometry (GC–MS) analyses allowed for the characterization of chemical changes in the binders. In the case of Arabic gum, a reduction in reflectance peaks corresponding to hydroxyl and stretching of conjugated C=C double bonds were observed; however, no modifications were detected in its monosaccharide composition, suggesting that the changes are primarily due to a drying process rather than structural chemical alteration. Therefore, this is a physical change due to the loss of hydration water, rather than a chemical one. For animal glue, the analysis revealed both chemical and physical alterations. Fourier Transform Infrared Spectroscopy (FTIR) showed a decrease in the amide I and II bands, indicating partial denaturation of the protein matrix. However, Chromatography coupled with Mass Spectrometry (GC–MS) analysis revealed a 22% preservation of the main amino acids, suggesting no significant degradation of the peptide chains. The discrepancy between both methods points to surface-level degradation or conformational changes. Additionally, cracks and fissures were observed, likely caused by heat and UV radiation, confirming physical deterioration of the binder as well. In linseed oil, the analyses identified advanced oxidation and polymerization processes, evidenced by the appearance of new hydroxyl groups and a decrease in carbonyl and alkane groups. Additionally, the formation of dicarboxylic fatty acids such as sebacic, azelaic, and suberic acids, typical markers of aged oil, was confirmed by Fourier Transform Infrared Spectroscopy (FTIR) and Chromatography coupled with Mass Spectrometry (GC–MS). These chemical changes resulted in visible physical alterations, including craquelure. Therefore, linseed oil exhibited significant chemical and physical degradation following artificial ageing.

Regarding the removal of overpaint layers, the conclusions drawn from this study indicate that Treatment 1, based on a mixture of heptane and acetone, was highly effective in dissolving the oil-based repaint, including its aged form. This effectiveness is attributed to the neutral solvents acting solely through solubility parameters without affecting the original layer. Similarly, Treatment 5, which utilizes a combination of polyaminoamide (PAA) polymer and glycine, also demonstrated effective removal with minimal impact on the underlying original polychromy layers. In contrast, other treatments such as the use of the surfactant Triton-X100^® (Tests 2 and 3) and sodium hydroxide (Test 4) showed adverse effects: the former due to its toxicity and potential residue interference on the treated surface, and the latter due to its chemical aggressiveness as a strong base capable of affecting both the repaint and the original polychromy layer.

According to what has been stated, the importance of selecting cleaning treatments that are not only effective in removing overpaint layers but also respectful of the original materials’ integrity is evident. In this regard, according to the obtained results and conclusions, further exploration of polymer- and amino acid-based solutions, such as PAA-Gly, is recommended due to their potential selectivity and low aggressiveness. Furthermore, future research could focus on optimizing artificial ageing protocols to more accurately simulate the real conditions faced by historical artworks, as well as on developing conservation methodologies that balance effectiveness with minimal intervention. To this end, a new line of investigation emerging in this field involves the study of actual artworks in order to incorporate the real limitations these works present, which have been attempted to be reproduced in this study, to determine the true effectiveness of these treatments on this type of artwork.