Experimental investigation of tensile and adhesive properties in gesso and egg tempera systems

Abstract

This study presents a systematic mechanical characterization of traditional materials employed in egg tempera painting techniques, focusing on gesso grounds and pigmented egg tempera layers. Gesso samples were prepared using rabbit skin glue, calcium carbonate, and titanium white pigment, while tempera paints were formulated with egg yolk binder and four representative mineral pigments: titanium white (TW), zinc white (ZW), yellow ochre (YO), and ultramarine blue (UB). Mechanical performance was assessed through micro-tensile and peeling tests. For gesso, the results demonstrate a strong dependence on Pigment Volume Concentration (PVC), with higher PVC values leading to reductions in both elastic modulus and ultimate tensile strength. For tempera paints, marked variability in elastic modulus was observed, with yellow ochre exhibiting the highest stiffness, whereas Titanium-rich formulations showed the greatest strain capacity at failure. In addition, adhesive fracture energy was quantified, providing the first dataset on the peeling resistance of these traditional materials. The results offer new insights into the fracture and adhesion behavior of tempera systems and establish a robust experimental framework that can support the development of conservation strategies for historical paintings.

Introduction

The preservation of painted artworks represents a central priority in art conservation. Egg tempera, a technique practiced since antiquity, is among the most historically significant painting media. Formed by dispersing pigments in egg yolk and applied over gesso-prepared substrates, it is characterized by luminous visual qualities and remarkable longevity. Yet, tempera also displays pronounced mechanical vulnerability, being especially sensitive to environmental fluctuations. Over time, degradation arises from both physicochemical instability and biological activity, threatening the cohesion and integrity of the paint layers^1,2,3,4. One of the most distinctive manifestations of these processes is craquelure, the network of fine cracks that alters the artwork’s appearance and reveals the cumulative impact of mechanical stress and material aging^5,6. Understanding the origins and evolution of such features requires closer examination of the multilayered nature of paintings and the interactions among their constituent materials⁷.

Paintings are stratified systems composed of heterogeneous materials, each with distinct physicochemical and mechanical properties^8,9. Typically, these include a structural support (wooden panels or canvas), animal glue sizing, preparatory grounds (chalk or gypsum bound in proteinaceous or oil-based media), pigmented paint layers, and, in many cases, protective surface varnishes^10,11. The mechanical stability of the whole system emerges from the interplay among these layers, whose mismatched responses to stress, moisture, or temperature variations can generate significant internal tensions. Consequently, the preservation of such works requires not only knowledge of individual material properties but also an understanding of how external agents trigger and exacerbate these internal mismatches over time.

In this context, heritage science has increasingly emphasized the influential role of environmental factors–particularly fluctuations in temperature and relative humidity–in driving deterioration of these layered structures^{12,13,14,15,16,17,18}. Among the environmental effects, the hygroscopic response of materials such as wood, canvas, and glue is particularly critical: as they absorb and release moisture in response to humidity changes, they undergo dimensional cycling—expansion and contraction—that alters their intrinsic material properties and induces strain within the layered system^{1,9,19,20,21,22,23,24}. Because different layers respond unevenly to these environmental stimuli, large internal stresses accumulate, often leading to cracking and delamination^16,25,26. Damage in these systems typically arises not from isolated extreme events but from cumulative fatigue under repeated hygric and mechanical cycling, to which gesso grounds are particularly susceptible. The resulting progressive loss of mechanical integrity diminishes their capacity to accommodate stress, thereby increasing the likelihood of cracking and interfacial adhesive failure^27,28,29,30. This progressive weakening underscores why long-term fluctuations, even if moderate, can be as damaging as acute environmental shocks, as accumulated fatigue generates localized stress concentrations that initiate or extend cracks, exacerbating the deterioration process^{10,31,32,33,34}. Interestingly, however, a fully developed craquelure network can also reduce vulnerability: by redistributing internal stresses more evenly across the surface, it may help stabilize the system against further environmentally induced cracking³⁴. Thus, the presence of craquelure represents both a symptom of past deterioration and a potential moderator of future stress.

Beyond structural and environmental factors, the intrinsic properties of artistic materials themselves exert a decisive influence on mechanical behavior. In the case of egg tempera paints, mechanical response is strongly dependent on variables such as pigment composition, binder-to-pigment ratio, and curing conditions, which govern key properties including elasticity, tensile strength, and strain-to-failure^7,26,35. Among the compositional parameters that influence mechanical behavior, piment volume concentration (PVC) together with the mineralogical and morphological characteristics of pigment particles are particularly critical¹¹. Within a paint film, pigments function as rigid inclusions embedded in a comparatively flexible binder matrix, and an increase in PVC generally leads to higher stiffness but reduced ductility^6,36. The geometry of pigment particles further modulates this response: plate-like or acicular pigments promote mechanical interlocking and raise elastic moduli, whereas equiaxed or aggregated particles tend to lower stiffness^35,37,38. As a result, even subtle variations in pigment type, concentration, or distribution within a single medium such as egg tempera can produce pronounced differences in fracture resistance and overall mechanical performance. Experimental studies provide direct support for these observations. Uniaxial tensile tests on model paint films have shown that increasing pigment-to-binder ratios enhances stiffness while reducing flexibility^11,33. Coatings with different mineral additives likewise display distinct elastic moduli and fracture thresholds, even when prepared under otherwise comparable conditions^39,40.

Complementing bulk testing, localized techniques such as nanoindentation have revealed significant heterogeneity within paint layers, identifying stiffness gradients that correlate with pigment distribution and binder–matrix interactions^24,41. Complementary approaches, including uniaxial and biaxial tensile testing^12,42,43 and dynamic mechanical analysis (DMA), have further enabled systematic characterization of elastic and viscoelastic behavior under controlled humidity and temperature^24,44. Notably, humidity-regulated DMA (DMA-RH) has revealed moisture-induced softening and identified critical RH thresholds that trigger plasticization in both proteinaceous and oil-based coatings⁴⁵. Using time–temperature superposition and master-curve techniques, DMA studies also provide access to long-term creep, stress relaxation, and fatigue responses, extending insights beyond the directly measurable experimental timescales⁴⁶.

Despite these advances, significant knowledge gaps remain, as most research has focused on oil- and acrylic-based systems, while studies on egg tempera paints are scarce and investigations of gesso grounds, largely confined to the 1980s and 1990s, have not been substantially updated with modern analytical approaches. Pioneering work from that period reported parameters such as Young’s modulus and ultimate tensile strength, particularly in relation to PVC and humidity effects^47,48,49,50. While these studies provided valuable baseline data, they were restricted to specific preparation methods and testing protocols that limit their general applicability. Revisiting these properties with contemporary methodologies is therefore essential, both to validate adequate sample preparation and measurement techniques as well as to establish a more robust mechanical reference. In contrast, adhesive and fracture-related properties remain virtually unexplored: peeling resistance of gesso ground and tempera layers has, to the best of our knowledge, never been quantitatively assessed. The knowledge gap is even more pronounced for egg tempera paints. To date, only a single systematic study has investigated their uniaxial tensile behavior and moisture-related properties, demonstrating how pigment type influences elastic modulus, strain at break, and hygroscopic response¹. However, this work did not consider the effects of PVC or binder composition, nor did it address fracture-related properties. In particular, peeling resistance and substrate detachment behavior—critical for understanding failure mechanisms in historical tempera paintings—remain entirely unexplored.

To address these gaps, the present study undertakes a systematic mechanical characterization of both gesso grounds and tempera paints prepared following traditional recipes. For gesso, we revisit bulk mechanical properties (elastic modulus and ultimate tensile strength) and extend the scope to include peeling resistance. Beyond their intrinsic interest, these measurements also serve to validate the experimental setup and sample preparation procedures, providing a solid baseline for the subsequent characterization of tempera paints. For tempera, we examine the influence of pigment type and PVC not only on uniaxial tensile properties but also on peeling and fracture resistance. In order to isolate material-specific behavior, peeling tests were performed directly on wooden substrates, with each layer–ground or paint–tested independently. This approach avoids the additional complexities introduced by full stratigraphic assemblies (sizing–ground–paint) and allows clearer attribution of mechanical behavior to material composition. At this initial stage, neither artificially aged nor historical samples were employed, as the study focuses on methodology, aiming to establish robust and reproducible tensile and peeling protocols using standardized mechanical testing on traditional painting materials, while also providing the first quantitative dataset of their mechanical and adhesive properties. By providing the first systematic measurements of peeling resistance in both gesso and tempera, by expanding the tensile characterization of tempera beyond the very limited existing literature, and by establishing validated datasets for gesso as a reference, this study contributes a new and more complete basis for understanding the mechanical performance of traditional painting materials. These results provide a foundation for future investigations involving aged samples, complete stratigraphies, and additional complementary experimental techniques that link microstructural analyses to bulk material properties, ultimately supporting the development of more effective conservation strategies for tempera-based artworks.

Materials and methods

Materials

Two primary material systems were investigated in this study: gesso ground layers and pigmented egg tempera paints. In the context of the Italian tempera tradition, the term gesso refers specifically to a calcium-carbonate preparation bound with animal glue (gesso sottile), as distinct from gypsum-based layers (gesso grosso) typically employed for preliminary leveling. While the present study adopts this calcium-carbonate formulation in accordance with Cennini’s description⁵¹ and standard Renaissance practice, it is important to acknowledge that alternative preparation layers, particularly those incorporating silicate-rich materials such as kaolin or gypsum, are documented in specific historical contexts⁵². The choice of a calcium-carbonate gesso therefore follows the established Italian tradition and provides a coherent historical framework for the materials investigated in this work. The preparation of gesso ground and egg tempera paints followed historical methods derived from recipes described by Cennini^51,53, with systematic variations introduced in the PVC which quantifies the volumetric ratio of pigment relative to the total solid content of pigment and binder, is defined as:

$$\begin{aligned} PVC=\frac{P}{P+B}\times 100 \end{aligned}$$

(1)

where P and B represent the volumes of pigment and dried binder, respectively. The quantities P and B were determined from mass measurements and converted to volumes using the known densities of the pigments and the dried binder (rabbit-skin glue for gesso and egg-yolk solids for tempera), ensuring consistent and comparable PVC values across all formulations. In this calculation, it was assumed that all the water present in the egg yolk and added during sample preparation evaporated completely during drying, providing a consistent basis for comparing specimens with different pigment concentrations.

Gesso preparation

Gesso ground samples were prepared using rabbit skin glue, calcium carbonate, titanium white pigment (all sourced from Antichità Belsito, Italy), and deionized water. The preparation protocol followed the procedure described by Bratasz et al.³⁴. A 6.7 wt% glue solution was obtained by hydrating 6.7 g of rabbit skin glue in 93.3 g of water overnight (Fig. 1a), allowing complete swelling of the collagen-based binder (Fig. 1b ). The hydrated mixture was subsequently heated in a double boiler at 60 $^{\circ }$C for 60 minutes (Fig. 1c) until a fully dissolved, homogeneous solution was obtained (Fig. 1d). Calcium carbonate was then gradually incorporated into the warm solution until a cohesive, paste-like consistency was achieved, with no visible free liquid. To enhance optical brightness, a small amount of titanium white pigment was added during the incorporation of calcium carbonate (Fig. 1e). The study employed two PVC values for gesso (92% and 94%), selected within the range documented for historical preparations. These two formulations were chosen as representative cases to establish a robust experimental workflow and to assess the responsiveness of the methodology to compositional differences.

Fig. 1

(a) Glue–water mixture immediately after preparation; (b) fully hydrated glue following 24 h water immersion; (c) double-boiler setup used for controlled heating; (d) homogeneous glue solution after complete dissolution; (e) final gesso paste incorporating calcium carbonate and titanium white to enhance brightness and optical uniformity.

Egg tempera preparation

Four pigments were selected for this study: Titanium White (TW), Zinc White (ZW), Yellow Ochre (YO), and Ultramarine Blue (UB). These pigments were selected to encompass a wide range of mineralogical classes and expected mechanical behaviors while maintaining historical relevance. This set was sufficient to validate the mechanical testing framework, with future work planned to expand the pigment range to include other historically significant materials, such as azurite, malachite, and cinnabar. Fresh hen egg yolk was used as the binder for all tempera formulations. The yolk was separated from the egg white and carefully cleaned to remove chalazae and fibrous membranes, yielding a homogeneous binder phase. The preparation procedure followed the historical recipe adapted by Poznańska et al.¹, in which the relative proportions of egg yolk, water, and pigment were adjusted to produce mixtures with workable consistency that could be cast into molds and retain their geometry after drying (Fig. 2a). For each pigment formulation, the PVC was calculated, as reported in Table 1. The PVC values were selected to obtain mixtures that were sufficiently workable during application and that, after drying, resulted in stable, homogeneous specimens suitable for mechanical testing. Pigments were gradually incorporated into the binder using a porcelain mortar and pestle to ensure uniform dispersion and to minimize agglomeration. This procedure provided consistent and reproducible mixtures, enabling reliable preparation of specimens for mechanical characterization.

Table 1 PVC of the egg tempera formulations.

Samples preparation

Although the materials formulation is well defined, obtaining mechanically testable specimens presented significant difficulties, particularly in maintaining consistent geometry and reproducibility during drying. Consistent geometry is essential in mechanical testing, as deviations can lead to stress concentrations and unreliable measurements. To avoid geometric irregularities such as pronounced shrinkage and warping, customized casting molds were designed and fabricated from polylactic acid (PLA) using 3D printing technology (Fig. 2a,b). The mold geometry was specifically optimized to accommodate the expected volumetric shrinkage of gesso and tempera layers during drying, while enabling the reproducible production of dog-bone-shaped specimens with accurately defined cross-sectional dimensions (Fig. 2c). Prior to casting, the inner surfaces of the molds were treated with a thin, uniform layer of neutral, plant-based oil to act as a chemically inert release agent, preventing adhesion to the mold walls and allowing clean demolding without surface contamination or alteration of intrinsic mechanical properties. After filling, the molds were stored under controlled ambient conditions (21 $^{\circ }$C, 50% RH) to dry for a minimum of one week, ensuring equilibrium moisture content. Once drying was complete, the specimens were demolded and subsequently polished using a fine-grit abrasive sponge to remove surface irregularities and to ensure consistent edge quality (Fig. 4a). Particular consideration was given to cavity dimensions and edge profiles: by reducing lateral constraint during drying, the molds permitted uniform shrinkage and effectively prevented warping, curling, and thickness irregularities, thereby facilitating the preparation of mechanically robust samples that could be consistently aligned and mounted for both tensile and peeling tests. Final specimen thicknesses were measured at 1.4 mm for samples prepared with 92% PVC and 2.0 mm for those prepared with 94% PVC.

Fig. 2

(a) Customized 3D-printed mold filled with freshly prepared gesso ground paste prior to drying; (b) Customized 3D-printed mold filled with freshly prepared yellow ochre (YO) egg tempera mixture prior to drying; (c) Schematic illustration of the dog-bone sample dimensions.

The evaluation of fracture behavior required methodological refinements in substrate selection. Preliminary trials employing conventional polymeric supports, such as polybutylene terephthalate (PBT) substrates with polypropylene (PP) strips, proved unsuitable for peeling experiments due to premature detachment during drying, failure during mounting, and irreproducible force–displacement curves. To address these limitations, beech wooden substrates (150 mm $\times$ 14 mm $\times$ 1.8 mm) and wood strips (7.0 mm $\times$ 0.5 mm), supplied as thin beech sheets intended for smoking applications, were adopted. Beech was selected for its homogeneity, stability, and its close mechanical relevance to commonly used historical supports, providing adequate adhesion during preparation and consistent mechanical behavior during peeling tests. In addition to these practical advantages, the choice of beech wood was also motivated by experimental design considerations, as it offered a configuration more consistent with historical supports. For each pigment and for the gesso ground formulation, specimens were prepared by applying the corresponding mixture as an adhesive layer between the wooden substrate and veneer strip. At this stage, peeling tests were not conducted on multilayer stratigraphies (sizing–ground–paint), which would more closely reproduce real paintings; instead, each layer–gesso or tempera–was applied and tested independently on wood to isolate its intrinsic fracture and adhesion properties, without the effects of interfacial interactions among multiple layers. Representative prepared specimens, including those for Ultramarine Blue (UB) and Yellow Ochre (YO), are shown in Fig. 3.

Fig. 3

Representative peeling test specimens prepared with Ultramarine Blue (UB) and Yellow Ochre (YO) tempera formulations.

Methods

The experimental program was designed to provide a systematic characterization of both gesso grounds and egg tempera paints prepared following traditional recipes. Two complementary testing approaches were adopted: micro-tensile testing to probe bulk elastic and fracture behavior, and peeling tests to quantify adhesive fracture energy.

Micro-tensile testing

Tensile testing was performed on the polished dog-bone-shaped samples (Fig. 4a) according to principles adapted from ASTM D1708, employing a DEBEN micromechanical testing stage housed at the MUSAM-Lab, IMT School for Advanced Studies Lucca (Fig. 4b). Tests were conducted at a constant displacement rate of 0.1 mm/min. A minimum of three dog-bone-shaped specimens were evaluated for each formulation to account for sample variability and to ensure the statistical reliability of the results.

Fig. 4

(a) Polished gesso samples after drying; (b) DEBEN micromechanical tensile stage.

Peeling test

Peeling tests were performed in accordance with ASTM D1876, using a Zwick/Roell universal testing machine available at the MUSAM Lab, IMT School for Advanced Studies Lucca (Fig. 5a). Each specimen was mounted between the clamp and base body of the peel kit, with the free strip secured in the grip attached to the moving crosshead. Tests were carried out at a constant crosshead displacement rate of 50 mm/min. The peeling force was continuously recorded by a load cell integrated into the crosshead, while peel extension was determined from the absolute crosshead displacement. For each condition, a minimum of two specimens were tested to verify repeatability and ensure reliability of the results. For a $90^{\circ }$ peeling configuration, assuming a linearly elastic and inextensible strip, the adhesive fracture energy G can be calculated directly from the steady-state peeling force F normalized by the specimen width w, according to the Rivlin relation⁵⁴:

$$\begin{aligned} \frac{F}{w}=\frac{G}{1-cos(\theta )} \end{aligned}$$

(2)

where ($\mathcal {\theta }$) is the peeling angle as illustrated in Figure 5b.

Fig. 5

Experimental setup for peeling tests: (a) Peeling test setup: laboratory apparatus; (b) schematic configuration of the $90^{\circ }$ peeling test.

Results

This section presents and analyzes the results of the experimental investigations. The discussion focuses on the tensile properties, fracture mechanisms, and interfacial adhesion behavior of gesso and egg tempera formulations, with particular attention to the influence of pigment composition and PVC, as well as their interaction with the wooden substrate. The data are discussed in relation to the intrinsic mechanical performance of traditional painting materials, establishing a foundation for future investigations on aged or naturally degraded samples, as well as on materials intended for restoration and conservation applications.

Mechanical properties of gesso grounds

Tensile behavior

The tensile response of gesso specimens with PVC of 92% and 94% was investigated using micro-tensile testing. A total of three specimens were tested for each PVC concentration to ensure reproducibility of the mechanical measurements. Representative stress–strain curves for each formulation are shown in Fig. 6b, illustrating the characteristic mechanical behavior of gesso as a ground material. In these plots, the nominal stress ($\sigma$) was calculated as the applied load divided by the initial cross-sectional area of the specimen, while strain ($\varepsilon$) was defined as the ratio of elongation to the initial gauge length. Both formulations exhibited a quasi-brittle response, characterized by an initial linear elastic regime, the progressive accumulation of microdamage, and abrupt brittle fracture. This response is consistent with fracture characteristics previously reported for collagen-based gesso grounds²¹. The quasi-brittle nature of gesso was also reflected in the rupture morphologies observed during tensile testing. As shown in Fig. 6a, the fracture surfaces reveal sudden, catastrophic failure with negligible evidence of plastic deformation. Furthermore, the stress–strain curves demonstrated a pronounced dependence of mechanical behavior on PVC: variations in pigment concentration significantly influenced stiffness, strength, and the overall failure process, underscoring the sensitivity of gesso performance to compositional parameters.

Fig. 6

(a) Micro-tensile test results of gesso specimens; (b) ruptured gesso samples (PVC = 92%) after micro-tensile testing.

The tensile test results for gesso ground specimens are reported in Table 2. The elastic modulus was determined from the slope of the initial linear region of the stress–strain curve. The 92% PVC specimens displayed an average modulus of 3919 ± 134 MPa, approximately 7% higher than the 94% PVC specimens (3660 ± 362 MPa). This inverse relationship between PVC and stiffness reflects the compositional balance between pigment and binder: at lower PVC values, the greater proportion of rabbit skin glue enhances cohesive interactions between calcium carbonate particles, thereby improving stress transfer across the microstructure. A similar trend was observed in ultimate tensile strength ($\sigma _u$). The 92% PVC gesso exhibited 9.84 ± 0.57 MPa, compared with 7.25 ± 0.63 MPa for the 94% PVC formulation, corresponding to an approximate 36% reduction in strength with higher pigment content. The rupture strain followed the same pattern, with values of $5.3 \times 10^{-3}$ for 92% PVC and $3.8 \times 10^{-3}$ for 94% PVC. These results demonstrate that even modest changes in pigment-to-binder ratio can substantially alter the fracture resistance and overall mechanical integrity of gesso grounds. The values obtained for both elastic modulus and tensile strength align closely with those reported in the literature for gesso samples tested under comparable relative humidity ranges (30–65%)^16,21,27, thereby confirming the reliability of the present preparation and testing methodology.

Table 2 Mechanical properties of gesso specimens with different PVC. Reported values represent mean ± standard deviation.

Fracture behavior

The peeling behavior of traditional painting materials has not previously been characterized through dedicated mechanical testing, as earlier studies have focused primarily on the bulk mechanical response of preparatory layers under tensile or bending loads. In this work, a systematic $90^{\circ }$ peeling test was developed to quantify the adhesive fracture energy of gesso and pigmented egg tempera, thereby providing direct insight into their interfacial failure mechanisms. For the gesso ground, only the 92% PVC formulation was tested, as this composition demonstrated superior tensile performance (higher elastic modulus, ultimate tensile strength, and rupture strain) compared to the 94% PVC formulation. Focusing on the mechanically more robust formulation enabled a clearer assessment of fracture behavior with direct relevance to conservation applications.

As shown in Fig. 7, the peeling force–displacement curve exhibited a characteristic profile consisting of an initial rising region associated with crack initiation, followed by a relatively stable plateau corresponding to controlled crack propagation along the interface. From these data, the average adhesive fracture energy (G) was determined as 0.235 ± 0.020 N/mm, representing the energy required to propagate a delamination front between the gesso and the wooden substrate. This value reflects an intermediate adhesive performance: strong enough to maintain stable attachment to wooden substrates under moderate stresses, but limited by the brittle nature of the gesso matrix, which restricts its cohesive strength. Furthermore, the relatively small standard deviation (± 0.020 N/mm) demonstrates that both preparation and testing procedures were highly reproducible, ensuring that the reported values constitute a robust reference for the reconstruction of historical recipes and the development of conservation strategies.

Fig. 7

Peeling force results of gesso ground specimen (PVC = 92%). Labels $\#1$ and $\#2$ refer to individual replicate specimens tested under identical conditions.

Mechanical characterization of egg tempera pigments

Tensile behavior

The micro-tensile test results for pigmented egg tempera specimens are presented in Figure 8, with the corresponding mechanical properties summarized in Table 3. The data reveal pronounced differences in tensile response depending on the pigment employed:

• Yellow Ochre (YO) exhibited the highest stiffness of all formulations, with an elastic modulus of 970.7 ± 89.6 MPa and an ultimate tensile strength of 1.27 ± 0.07 MPa. This combination of exceptionally high stiffness with only moderate strength indicates a highly rigid yet brittle behavior, consistent with previous findings for iron oxide–based pigments¹.

• Ultramarine Blue (UB) showed an elastic modulus of 400.1 ± 35.5 MPa and tensile strength of 0.72 ± 0.02 MPa. The relatively high stiffness suggests a rigid structure, likely associated with the fine particle size and strong interfacial adhesion between pigment and egg yolk binder.

• Zinc White (ZW) displayed intermediate properties, with an elastic modulus of 176.4 ± 29.0 MPa but the lowest tensile strength of all pigments (0.39 ± 0.05 MPa). This comparatively weak performance may be related to large pigment particle size and reduced binder interaction, resulting in poor internal cohesion.

• Titanium White (TW) demonstrated the lowest stiffness (29.6 ± 4.1 MPa), a moderate ultimate strength of 0.55 ± 0.04 MPa, and the highest strain at failure (0.04 ± 0.002). This unusually ductile response suggests that titanium dioxide particles promote greater deformability of the paint matrix, possibly due to their fine dispersion and relatively low density compared to other mineral pigments.

Overall, the measured mechanical properties are in strong agreement with those reported by Poznańska et al.¹, who examined tempera paints under varying relative humidity conditions. In particular, the elastic modulus, tensile strength, and failure strain for Yellow Ochre, Ultramarine Blue, and Titanium White closely match the values reported under high-humidity conditions. For Zinc White, however, direct comparative data on egg tempera are lacking. Available literature instead concerns aged oil paints prepared with cold-pressed linseed oil^1,33, which show markedly different mechanical responses. These discrepancies underscore the significant influence of binder type on the long-term mechanical performance of historical paint formulations and point to the need for further targeted investigations.

These findings confirm that pigment composition plays a decisive role in determining the stiffness, strength, and deformability of egg tempera films. Formulations with high stiffness and limited tensile strength, such as Yellow Ochre, are likely to be more prone to cracking and delamination when subjected to fluctuations in temperature and relative humidity that drive dimensional changes in the substrate. Conversely, more deformable formulations, such as Titanium White, may accommodate strain more effectively but at the expense of reduced structural rigidity. Recognizing these pigment-specific behaviors is directly relevant to conservation practice: it informs risk assessments for environmental control (e.g., establishing safe RH fluctuation ranges), supports the design of preventive conservation strategies aimed at stabilizing fragile paint passages, and guides the selection of consolidation treatments by identifying where adhesion reinforcement may be most necessary. Furthermore, understanding the mechanical incompatibilities between brittle gesso grounds and variably compliant tempera paints provides a scientific basis for predicting delamination risks and tailoring interventions to the specific mechanical profiles of different layers.

Fig. 8

Representative stress–strain curves obtained from micro-tensile tests of egg tempera specimens prepared with different pigments: (a) UB; (b) YO; (c) ZW; (d) TW. Labels $\#1$, $\#2$ and $\#3$ refer to individual replicate specimens tested under identical conditions.

Table 3 Mechanical properties of egg tempera specimens prepared with different pigments. Reported values represent mean ± standard deviation calculated from the results presented in Fig. 8.

Fracture behavior

The peeling tests on egg tempera formulations yielded adhesive fracture energy (G) values in the range of 0.032 ± 0.006 N/mm (Ultramarine Blue) to 0.048 ± 0.001 N/mm (Zinc White), as summarized in Table 4 and illustrated in Fig. 9. These values are approximately one order of magnitude lower than those obtained for gesso, reflecting the fundamentally different adhesion mechanisms between binder–substrate interfaces in egg tempera and in gesso.

Fig. 9

Peeling force results of pigmented egg tempera specimens: (a) UB; (b) YO; (c) ZW; (d) TW. Labels $\#1$ and $\#2$ refer to individual replicate specimens tested under identical conditions.

Examination of the peeling force–displacement curves highlights pigment-specific differences in interfacial behavior. Ultramarine Blue exhibited the lowest adhesive fracture energy, which may be linked to its high PVC (78%) and the limited binder–pigment interaction, leading to a relatively brittle interface with reduced resistance to crack initiation. Zinc White demonstrated the highest fracture energy among all tested formulations. Its peeling curves displayed a remarkably stable plateau, indicative of uniform crack propagation. The very low standard deviation further supports the interpretation of a homogeneous pigment distribution within the egg yolk matrix and consistent interfacial bonding to the wooden substrate. Yellow Ochre and Titanium White, despite exhibiting comparable adhesive fracture energies (0.046 ± 0.006 N/mm and 0.042 ± 0.004 N/mm, respectively), showed contrasting failure mechanisms. The Yellow Ochre specimens presented pronounced fluctuations in the plateau region, suggesting microstructural heterogeneities or localized bonding inconsistencies, likely related to the intrinsic variability of ochre pigment particles. By contrast, Titanium White displayed a moderately stable plateau with a progressive increase in peeling force, pointing to a more ductile failure process consistent with the enhanced deformability observed in its tensile behavior.

Notably, while the elastic moduli and tensile strengths of egg tempera paints varied widely across pigment types, the adhesive fracture energies remained within a narrow interval (0.032–0.048 N/mm). This observation indicates that interfacial adhesion in tempera films is largely governed by the egg yolk binder, with pigment properties exerting a secondary influence that shapes the detailed fracture mechanisms rather than the overall adhesion strength.

Table 4 Adhesive fracture energy (G) values for egg tempera formulations (mean ± standard deviation calculated from the results presented in Figure 9).

To complement the peeling force–displacement results, Fig. 10a presents representative peeled specimens for each pigment formulation. These macroscopic observations reveal the characteristic modes of adhesion loss between the egg tempera layers and the wooden substrate, providing qualitative confirmation of the distinct pigment–binder–substrate interactions discussed above. Furthermore, a microscopic image of the Ultramarine Blue specimen from test #2 (Fig. 10b) provides a detailed view of the peeled interface, where alternating adhesive and cohesive failure regions are evident. These zones of exposed substrate and residual pigment indicate heterogeneous interfacial bonding and correspond closely to the fluctuations observed in the peeling force curve, thereby reinforcing the interpretation of pigment-dependent interfacial mechanisms.

Fig. 10

(a) Peeled samples of egg tempera formulations (ZW, YO, UB, TW); (b) Microscopic detail (10$\times$ magnification) of the Ultramarine Blue peeled interface.

Discussion

The experimental results presented in this study provide a comprehensive mechanical picture of traditional egg tempera and gesso systems, revealing a complex interplay between material composition, microstructural organization, and interfacial behavior. The findings confirm that both bulk and adhesive mechanical responses are governed by the pigment-to-binder ratio and by the mineralogical characteristics of the pigments, which jointly determine stiffness, strength, and fracture resistance. Beyond their intrinsic material significance, these results have broader implications for understanding failure mechanisms in historical paintings and for guiding preventive and interventive conservation strategies.

The results demonstrate a clear dependence of both stiffness and tensile strength on PVC in gesso and on pigment type in egg tempera. This relationship can be explained within the micromechanical framework of particle-reinforced composites, where rigid inclusions (pigment or filler) are dispersed in a comparatively compliant matrix (glue or egg binder). In gesso, a moderate increase in PVC (from 92% to 94%) resulted in a 7% reduction in stiffness and a 36% decrease in ultimate strength, indicating that the reduction of continuous glue bridges weakens the load transfer between calcium carbonate grains and promotes microcrack initiation. The quasi-brittle stress–strain response observed is consistent with earlier findings for collagen-based preparatory grounds subjected to low humidity^21,34. In tempera paints, variations in particle type and morphology are reflected at the macroscopic level by a pronounced pigment-dependent mechanical response, with the elastic modulus spanning over an order of magnitude. Iron-oxide-based Yellow Ochre produced rigid yet fragile coatings, whereas titanium dioxide formed more ductile films capable of sustaining strains up to 4%, consistent with enhanced stress redistribution and crack-bridging effects within the binder matrix. The thicknesses of the specimens tested (1.4–2.0 mm) exceed those of historical paint layers. However, this choice was necessary to ensure stable gripping, precise alignment, and reproducible measurements in both tensile and peeling tests. While thicker than typical historical coatings, these specimens provide reliable insight into the intrinsic macroscopic mechanical response of the gesso and tempera formulations. A key aspect emerging from the data is the pronounced stiffness mismatch between gesso (E $\approx$ 3.6–3.9 GPa) and egg tempera films (E $\approx$ 30–970 MPa), corresponding to elastic modulus ratios up to 100:1 (Fig. 11). Such disparity is sufficient to generate stress concentrations at the ground–paint interface under hygrothermal fluctuations. The gesso, being both stiff and brittle, acts as a stress concentrator when the more compliant paint layer expands or contracts. Even modest humidity variations can therefore induce tensile stresses approaching the measured fracture limits ($\sigma _u \approx$ 8–10 MPa) in the gesso, explaining the frequent occurrence of crack initiation and delamination in historical tempera artworks. Once initiated, fracture propagation follows the weakest path, either through rupture of the paint layer or adhesive failure at the paint–ground interface, depending on local material composition and degree of degradation. Future work employing microscale tensile devices, nano-indentation, or micro-peel geometries will allow extension of the methodology to layer thicknesses closer to those found in historical artworks, enabling direct comparison with in situ conditions.

Fig. 11

Normalized results from tensile tests: (a) Elastic modulus; (b) Ultimate stress.

The peeling experiments provided the first quantitative dataset of adhesive fracture energy for both gesso and tempera systems (Fig. 12). The gesso exhibited a mean fracture energy of 0.235 ± 0.020 N/mm, an order of magnitude higher than that of the tempera films (0.032-0.048 N/mm). This hierarchical adhesion strength is highly significant: it implies that delamination is energetically more favorable within the paint layer than at the gesso–wood interface, thereby confining delamination to the uppermost strata and reducing the risk of catastrophic substrate detachment. Among tempera formulations, the narrow range of G values despite large differences in stiffness and strength indicates that interfacial adhesion is largely governed by the egg yolk binder, while pigment characteristics primarily modulate fracture mechanisms. The stable plateau observed for zinc white, for instance, points to homogeneous interfacial bonding, whereas yellow ochre exhibited force fluctuations consistent with microstructural heterogeneity. These subtle differences underline the complex role of pigment morphology in determining not only mechanical stiffness but also local fracture pathways.

Fig. 12

Normalized results from peeling tests.

From a conservation perspective, the measured mechanical properties and fracture energies offer valuable quantitative benchmarks. The low adhesive strength of tempera layers (G < 0.05 N/mm) suggests that even minimal mechanical or hygroscopic stresses may initiate paint lifting or flaking, explaining the high sensitivity of tempera paintings to environmental instability. Conversely, the relatively high cohesion of gesso grounds (G $\approx$ 0.23 N/mm) supports the use of consolidants or adhesives of comparable fracture energy to achieve mechanical compatibility without overstrengthening the interface. Reinforcement treatments that produce adhesion values significantly exceeding those of the original system risk generating new stress concentrations and secondary cracking upon humidity cycling. Moreover, understanding pigment-specific stiffness variations enables more targeted interventions such as reinforcing ochre-rich brittle passages while minimizing intervention on titanium white areas that exhibit natural compliance.

Although the present study focused on static testing, the observed mechanical profiles provide insights into the expected fatigue behavior under environmental cycling. Given its high stiffness and limited ductility, gesso is expected to accumulate damage progressively under repeated humidity-induced expansion–contraction cycles, leading to microcrack coalescence and eventual delamination. In contrast, more ductile tempera formulations, such as titanium white, may buffer substrate movements more effectively but could experience long-term viscoelastic creep and plasticization under sustained moisture exposure. Beyond their immediate relevance to conservation science, the present results provide essential input parameters for numerical models describing crack initiation and interfacial delamination in stratified cultural heritage materials. Integrating experimentally derived mechanical properties into numerical analyses will enable predictive simulations of craquelure development and restoration outcomes. Furthermore, the methodological framework established here, combining traditional recipes, controlled preparation, and micromechanical testing, can be extended to aged samples and alternative binders (e.g., casein, linseed oil, protein-oil emulsions). Eventually, the integration of complementary characterization techniques, such as chromatic measurements and optical or SEM imaging, will provide a more comprehensive understanding of pigment morphology, color properties, and microstructural features, enabling subsequent investigations to link these microstructural characteristics to the macroscopic mechanical and adhesive properties presented in this study, thereby enriching the established mechanical framework. Such studies will deepen our understanding of the mechanical evolution of artistic materials and support the design of conservation treatments that respect the original structural balance of historical paintings.

Conclusion

This work provides a comprehensive experimental framework for understanding the mechanical behavior of traditional painting materials used in egg tempera systems. By combining micro-tensile and peeling tests on gesso grounds and egg tempera paints prepared according to historical recipes, the study establishes quantitative relationships between composition, microstructure, and fracture performance.

Results demonstrate that the mechanical integrity of gesso is highly sensitive to PVC. In egg tempera paints, mechanical properties vary by more than an order of magnitude depending on pigment type, with iron-oxide-based ochres producing the stiffest and most brittle films, and titanium-white formulations showing exceptional ductility. Adhesive fracture energy measurements reveal a clear hierarchy between layers–gesso grounds exhibit far greater adhesion than tempera layers- indicating that delamination preferentially occurs within or above the paint layer rather than at the ground–support interface. Together, these findings clarify the mechanical incompatibilities inherent in traditional multilayer tempera systems and explain their characteristic susceptibility to cracking and flaking under environmental fluctuations. The dataset also provides quantitative benchmarks for assessing condition, diagnosing mechanical risks, and selecting consolidants or adhesives whose stiffness and fracture energies are compatible with original materials, thereby minimizing stress concentrations introduced during restoration. Such parameters offer a scientific basis for designing preventive conservation strategies–particularly in the optimization of climate control–and for guiding restoration treatments that preserve the structural compatibility of historical media.

In conclusion, this work establishes quantitative links between composition and mechanical performance in traditional painting materials, highlighting how pigment type, binder content, and ground composition govern stiffness, strength, and adhesion. Beyond immediate conservation applications, the results create a foundation for predictive modeling of stress evolution, crack formation, and interfacial failure in stratified artworks. Future research should expand this methodology to a broader range of historical pigments and binders, incorporate both naturally aged and artificially aged samples, and integrate experimental data into computational models capable of simulating long-term degradation. Together, these efforts will advance the integration of materials science into cultural heritage conservation, enabling data-driven approaches to preservation and restoration.