Navigation – Plan du site

AccueilHors séries EGGEGG 3DossierZinc oxide grounds in 19th and 20...

Dossier

Zinc oxide grounds in 19th and 20th century oil paintings and their role in picture degradation processes

Literary review, paint failure mechanisms and conditions of potential risk
Elena Pratali

Résumés

Lorsque l’oxyde de zinc est le composant principal d’une préparation à l’huile, sa rigidité et son comportement fragile pourrait affecter le système de peinture tout en produisant des dommages aux couches supérieures pigmentées. Des échantillons de blanc de zinc et blanc de plomb ont été préparés et vieillis artificiellement. Des données visuelles ont été recueillies et la spectroscopie IRTF s’est révélée efficace pour confirmer la formation de carboxylates de zinc. Des lignes directrices pour les traitements de conservation ont été établies en ce qui concerne les procédures techniques et le choix des matériaux.

Haut de page

Texte intégral

The authorwishes tothank the scientific staff of the CCR La Venaria Reale laboratories for technical support and FTIR measurements, the departments of Chemistry and of Earth Sciences of Turin University for Raman analysis and use of climatic chamber. Another thank goes also to Francesca Ghilardi for availability in sharing spectroscopic results and opinions.

Introduction

1Zinc oxide as a white pigment has been widely diffused across Europe for about a century, from the second half of  the 19th to the late sixties of the 20th century, when the introduction of titanium white progressively replaced it. Compared to lead white, its practical benefits are the great fineness, the less tendency to yellowing and the resistance to darkening when mixed with hydrogen sulphide pigments (Bersch 1901; Kühn 1986; Carlyle 2001).

2A wide literature survey conducted over both oil painting instruction manuals and texts addressed to industrial paintmakers (Pratali 2012) shows how the constant evolution and the variety of production procedures have determined the diffusion on the market of several zinc oxide brands very different both in physical features (shape, size, purity), esthetical qualities (whiteness, hiding power) and durability. Besides, a large amounts of additives were employed inside the ready-to-spread colour pastes which were sold by colourmen to artists, and the final properties of the paint film seem to have been affected by a great number of factors, such as the type of drying oil, the bleaching or refining treatments by which the oil medium has been processed, the inert pigment and liquid medium ratio, the presence of fillers and driers of different nature.

  • 1  Stress-strain tests made on zinc white in linseed oil paint samples naturally aged in standard con (...)

3In recent years Marion Mecklenburg and other researchers have examined art materials of canvas paintings from the mechanical point of view, and through a series of stress/strain tests performed on naturally aged samples the resistance and flexibility of different paint films has quantitatively been determined (Mecklenburg 2008; Rogala et al. 2010). Results show how zinc oxide paint films become extremely stiff and brittle - anyhow developing a certain strength - in a matter of few years after paint application. Compared to other white pigments commonly used as underpaints, zinc oxide films seem unable of suffering the least strain before failure (Figure 1)1.

Fig. 1  Stress-strain test performed over naturally aged white paint samples

Fig. 1  Stress-strain test performed over naturally aged white paint samples

Zinc oxide paint film proves particularly stiff and brittle compared to other white pigments commonly encountered as ground layers. The maximum strain suffered before paint failure appears in fact to be only a fifth of that imposed over paint by a stretching operation on a medium size canvas painting (Mecklenburg 2008).

Credit: Mecklenburg 2008

4When zinc white is found as main pigment inside a ground layer for canvas painting inner fractures may occur as a result of dimensional changes of underlying materials or of mechanical stresses due to picture transport and handling, producing evidences of delamination and chalking of layers painted over zinc white grounds (Figure 2).

Fig. 2 Example of selective cracking of carbon black painted over a zinc oxide ground of a painting by Franz Kline, (“Palladio”, 1961)

Fig. 2 Example of selective cracking of carbon black painted over a zinc oxide ground of a painting by Franz Kline, (“Palladio”, 1961)

A black paint failure was noticed over the zinc white underpaint only, while upper layers are intact where painted over lead white ground (Rogala et al. 2010).

Source: Rogala et al. 2010

5The unexpected recognition of significant amounts of oleic acid - that is of unsaturated chains which normally are converted to azelaic acid through oxidation - inside aged paint films has brought to assume an unique drying process of zinc oxide oil paint (Rogala et al. 2010; Maines et al. 2011). In an oil medium zinc oxide would form a tightly packed crystalline structure with a highly ordered lamellar distribution. Because of the layered structure the paint film would appear very stiff in a short time, and the tight packing of hydrocarbon chains would make double bonds more difficult to oxidize. As a result, unsaturated chains could remain trapped within the aged paint layer for years, this way leading to a lower polimerization degree. Cross-linking is therefore hindered and overall structural stability results minor than that of a normally oxidized oil paint film. This condition produces outwardly the effect of a film that prematurely appears hard and becomes brittle in a short time. The absence of an adequate number of cross-links together with the stiff lamellar structure would be responsible of episodes of planar fracture within zinc white ground layers.

  • 2  A protrusion is a globular-shaped agglomerate of lead soaps that can reach macroscopic dimensions. (...)

6In ancient literature the peculiar stiffening and embrittlement of zinc white paint layers can be found occasionally mentioned, mostly referring to conditions in which zinc white is the only mineral pigment and  it is applied  over a more resilient and flexible support material (Müller 1962). A relationship between this behavior and the pigment reactivity has been advanced inside scientific studies since the first half of the 20th century. In oil zinc oxide tends to react with fatty acids of siccative oils, forming metal carboxylates. The presence of zinc soaps has been observed inside zinc white aged paint and somehow linked to the film mechanical properties and its tendency to crack, though a precise assignment has not been made yet (Rhodes and Mathes 1926; Jacobsen and Gardener 1941).In recent years many scientific studies about painting materials have focused their attention on deterioration phenomena due to metal soaps formation and migration through the paint film, with a special reference to the case of lead pigments and the phenomenon of ‘protrusions’2. Lead soaps features and their role in picture degradation processes have been particularly investigated and behavioral patterns proposed to explain their causes and ways of development might be effective also in the case of zinc soaps.

7In the perspective of preservation of cultural heritage a better comprehension of which environmental parameters prove to be determinant in triggering the deterioration processes seems to be fundamental for the conservator. Since the evidence of decay specifically due to the presence and mechanical behavior of zinc white grounds has been so far only occasionally reported, such sporadicity could be attributed not only to the lower diffusion of zinc white as an underpaint, but also to the intervention of other flaming factors, both internal (type of zinc oxide, composition of mixtures, interactions between other materials) and external (climatic parameters of conservation, mechanical stresses). Conservative treatments could also potentially determine some risk conditions, as for instance the simultaneous use of heat, humidity and solvents during consolidation, lining, cleaning or varnishing operations, or the mechanical stresses applied in canvas stretching. The experimentation carried out within the present study is aimed to closely examine the film deterioration processes and to enlighten about which of these causes may represent a major critical factor in the conservation of canvas painting with zinc oxide oil grounds. Experimental data and literary records are thus compared as to outline general features of decay processes valuable for further investigations and for a major awareness in the conservative approach.

Experimental procedures

  • 3  Distributed as “Cremnitz White” (Kremer Pigmente 2012c)
  • 4  Maimeri has not released any information about treatments cast on poppy oil, but a certain reducti (...)

8Sample series and experimental conditions are summarized in Figure 4. Paint mixtures have been realized with 2 mineral pigments - zinc oxide (Zecchi) and lead white3 (Kremer) - and 3 siccative oils, different from each other both in physical and chemical features, and corresponding to a plausible variety of media employed by colourmen and artists: cold-pressed linseed oil (Kremer), refined poppy oil (Maimeri)4 and linseed standoil (Zecchi). Though linseed oil is far the most largely diffused medium in oil painting, for preparation of white paints poppy oil is often referred to thanks to its lower yellowing degree. On the other hand, use of standoil is recorded by Kühn (1986) precisely to counteract zinc oxide tendency to form hard and brittle films.

Fig. 3 SEM analysis of raw pigments: zinc white (a) and lead white (b)

Fig. 3 SEM analysis of raw pigments: zinc white (a) and lead white (b)

EDX analysis showed the presence of calcium carbonate (darker square particles) inside lead white composition. Size ratio of lead white particles is approximately 10 times higher than zinc oxide, this fact determining a higher oil-taking capacity of the latter.

Credits: Pratali 2012.

Fig. 4 Sample series, aging phases and analysis summary

Fig. 4 Sample series, aging phases and analysis summary

Legend: ZW=zinc white; LW=lead white; LIN=cold-pressed linseed oil; POP=refined poppy oil; STO=linseed standoil; PIG=Burgundy yellow ochre upper layer; ½PIG= Burgundy yellow ochre layer applied on half of the white ground surface; N=non-aged sample (natural aging). Number of “x” indicates how many time an analysis has been repeated on the same sample.

Credits: Pratali 2012.

9All mixtures have been applied on both microscope slides and on sized canvas stretched over small wooden frames, in order to find a relationship between paint film behavior and the presence of supports with a different reactivity (Figure 6).

Fig.5 Summary of oil medium compositions

Fig.5 Summary of oil medium compositions

Information have been collected from technical data sheets available online or directly provided by manufacturers via private communication (Kremer 2012a and 2012b).

Credits: Pratali 2012.

Fig. 6 Realization of paint samples

Fig. 6 Realization of paint samples

Paint films have been layered both on microscope slides (a) and on sized canvas (b).

Credits: Pratali 2012.

10A final layer of Burgundy yellow ochre (Kremer) in linseed standoil has then been applied on top of samples on canvas 07-18, to verify the possible repercussion of ground deterioration phenomena on upper paint layers.

11Pure pigments have been examined on a SEM microscope (Zeiss EVO60) equipped with EDX detector (Oxford Penta FET). Results of chemical analysis confirm a chemical composition of pure zinc oxide for the zinc white, with negligible traces of impurities, whereas lead white appears to be a mixture of lead carbonate and a significant amount of calcium carbonate, probably added as a filler. Size scale of lead white particles is several times higher than that of zinc oxide (Figure 3), thus confirming information about granulometry provided by Kühn (1986).

12A ratio of 75% of pigment and 25% of medium (parts by weight) has been kept for all mixtures, therefore providing an easily repeatable parameter in view of further experimentations. This ratio approximately reflects the indications about preparation of industrial ‘stiff paints’ contained inside paintmaker production manuals (Cruickshanck 1909; Uebele 1913; Petit 1907). It allows to obtain a dough texture, similar to that of oil grounds applied on canvas with the “priming knife” (Witlox and Carlyle 2005).

  • 5  For the preparation of canvas samples the CTS 1111 linen “boiled long linden tree” canvas has been (...)

13Pigments have been ground in a ceramic mortar both individually and together with the required amount of oil; mixtures have then been applied with a spatula, taking care to maintain the same thickness. Paint layers have been applied on 24 microscope slides, sequenced from A to L and realized in two twin series (series 1 and series 2), and on 18 sized canvas5, as shown in Figure 4. Sample series labeled with an “N” were intended for natural aging.

Accelerated aging

  • 6  According to Boon soap formation has been observed by Carlyle after about a month of exposure at 5 (...)

14Accelerated aging of samples was performed in three successive steps, using different parameters of climatic stress. Paint films have been allowed to completely dry in an oven, then subjected to cyclical variations in temperature and relative humidity, and finally to UV irradiation. Values ​​of temperature and RH were set according to literature data, in order to achieve an acceleration of soap formation processes6.

  • 7  Paint samples have entered the oven after a few days after white paint application. After 500h an (...)
  • 8  Each cycle lenght ranged from 6 to 12h.
  • 9  This last step has involved all samples on canvas previously aged in oven and climatic chamber (04 (...)

15In the first phase samples intended for artificial aging have been left in a Binder FD115 oven at 50 °C for 2200 h7. Second phase involved the use of an Heraeus HC 4020 climatic chamber, where paint samples were subjected to 80 cycles of thermo-hygrometric severe fluctuations, with the transition from a condition “A” (T = 50 °C and RH = 80%) to a condition “B” (T = 15 °C and RH = 45%)8. Last stage of aging has been carried out inside an Heraeus UV Suntest-CPS irradiation chamber, equipped with a Xenon 1500 W95 ultraviolet lamp, for a period of 500 h9. Ultraviolet lamp also produces heat, so that the inside temperature was approximately of 50-55 °C.

Analysis

16Visual data on surface physical changes were collected after each step of aging, while chemical processes were observed by ATR-FTIR and FT-Raman spectroscopy.

17FT-Raman spectra were acquired inside Chemistry Department of Turin University with a FTIR Bruker Vertex 70 spectrophotometer, equipped with a Raman Ramll module; exciter laser diode operated in the NIR at 1064 nm. Approximately 200 scans were collected for each spectrum. FTIR analysis was completed inside laboratories of the Centro di Conservazione e Restauro La Venaria Reale by a FTIR Bruker Vertex 70 spectrophotometer in ATR mode. Spectral analysis was recorded both after second and after third step of artificial aging, and compared with spectra obtained from fresh paint films measured 2 days after application. Standard spectra of raw materials (individual pigments and drying oils) were also collected.

18On most significant samples RTI technique (Reflectance Transformation Imaging) was performed in order to highlight physical changes of surface appearance, obtained in the case of zinc white and poppy oil paint films applied on canvas. The high resolution power and the possibility of image manipulation allowed in fact to clearly outline some physical degradation features and to collect detailed information on its evolution course.

Results

Visual analysis

19Examination of naturally aged samples showed a significant difference in drying rates of mixtures: paint samples made with lead white in standoil appeared dry to the touch a few days after application, while zinc white in poppy oil samples required over three weeks. Besides, all lead white samples showed a considerably higher drying rate than zinc white.

20After first stage of aging all white samples displayed signs of yellowing compared to naturally aged samples. This tendency was markedly evident in the case of linseed oil mixtures, and especially in its combination with lead white, which proved as the most yellowed. Standoil paint films were also slightly yellowed but mostly characterized by a gloss appearance and a soft and elastic consistency. Mixtures realized with poppy oil kept the whiter appearance and resulted as the most similar to corresponding naturally aged samples. A slight hint of cracking was observed on canvas sample 05 only (zinc oxide + poppy oil), in the form of thin sub-millimeter parallel cracks diffused over the surface. On the contrary, no visible sign of physical degradation has occurred on the same paint film applied on slide (samples E1 and E2).

21Same trends observed after first phase were confirmed also after second step of aging: linseed oil paint samples showed a marked yellowing, whereas poppy oil were fairly the whitest and most similar to naturally aged samples. Fine cracks diffused on sample 05 appeared more pronounced, and other thinner extra cracks occurred at right angles to the first ones. Their distribution seemed to reflect the underneath yarn pattern, and in fact major cracks were opened along warp direction. In this phase it was also observed how a crack network similar to that of sample 05 was beginning to show also on sample 17 (zinc oxide + poppy oil + yellow ochre), beneath the ochre layer, which nevertheless did not display any evidence of failure.

22After third step of aging the photochemical action of UV radiation determined a neat bleaching and matting of all paint surfaces and of sized canvas itself, such that all white grounds applied both on canvas and on slides resulted whiter than corresponding naturally aged samples. Since only series 2 of slide samples has undergone this stage of aging, a direct comparison between slide samples subjected to different aging phases can be proposed in Figure 10.

23At this stage cracks occurred on sample 05 became more severe and outlined, and clearly spread over the whole surface. On sample 17 no sign of upper paint failure was recorded yet, but the presence of several cracks within zinc white underlying layer is clearly detectable from the visual examination of surface. It is then remarkable to observe on N sample 02, after approximately six months of natural aging, some early traces of a physical decay similar to that observed on the corresponding artificially aged sample. Besides, also on sample 04 (zinc white + linseed oil) a very slight outset of the same cracking phenomenon was observed at this point, whereas it was not perceivable after second phase.

RTI analysis

24RTI analysis was performed on sample 05 and 17 after second and after third phase of aging, and permitted to characterize their physical decay. Sample 05: after second step of aging a thin cracking was already diffused over the whole paint surface, parallel to weft yarn direction. Areas in which cracks have occurred correspond to sites of film lower thickness, where fabric yarns overlap each other. Each crack length never exceeds 1 mm, and their distribution appears very regular (Figure 7 and 8).

Fig. 7 RTI images of paint sample 05 after second (a) and after third (b) phase of aging

Fig. 7 RTI images of paint sample 05 after second (a) and after third (b) phase of aging

RTI permits to enhance morphological features by manipulating light emission sources. Cracks appear deepened and more outlined after third stage of aging.

Credits: Pratali 2012.

Fig. 8 RTI detail of sample 05

Fig. 8 RTI detail of sample 05

Surface after third step of aging (left) and scheme of cracking network (right)

Credits: Pratali 2012.

25After third phase crack have deepened and paint flaps seem to have lifted and opened, implying the possibility of a forthcoming formation of a full cracking network, with effective risks of paint losses.  Crack position with respect to yarns structure is shown by scheme on the right: major cracks have opened at right angles to warp direction, and correspond to areas of paint film lower thickness.

26Compared to those at right angles, cracks parallel to warp direction are thinner, less diffused and not so deep. After UV irradiation all cracks became markedly more pronounced, and depth seemed to have increased. Furthermore flaps of paint film beside each crack seemed to have lifted and opened, assuming a lips-like appearance (Figure 8). Sample 17: by modulation of chromatic values it was possible to highlight morphologic features of paint surface - otherwise flattened and concealed by yellow ochre layer - thus permitting to recognize how cracking processes similar to those on sample 05 have appeared also below the upper paint layer, but without affecting its integrity. Their distribution and appearance essentially resembles what observed on sample 05 (Figure 9).

Fig. 9 RTI detail of sample 17 after third step of aging

Fig. 9 RTI detail of sample 17 after third step of aging

By reduction of chromatic values and light manipulation (b) it is possible to highlight the appearance of cracking underneath the yellow ochre layer.

Credits: Pratali 2012.

Spectroscopic analysis

27FT-Raman analysis performed on fresh and artificially aged paint samples indicated a substantial analogy with data collected within other studies (Oyman, Ming and Van der Linde 2005; Mallégol, Gardette and Lemaire 1999; Manzano et al. 2011). A comparison between spectra acquired from fresh samples of all three oil media showed a great similarity of peaks; cold-pressed linseed oil and poppy oil appeared almost identical, while some small differences were found in the case of standoil, but they disappeared as soon as the medium was mixed with pigment. FT-Raman spectroscopy unfortunately revealed itself not effective in determining the presence of zinc carboxylates; none of the spectra has permitted in fact to distinguish the characteristic peaks of zinc or lead soaps as they were reported in literature (Robinet e Corbeil 2003).

28Also FTIR spectroscopy showed small differences between standard spectra of oil media. The peak at 867 cm-1, which Mallégol attributes to linolenic acid skeletal vibrations, was well-outlined in the case of linseed oil but almost not resolvable in poppy oil (Mallégol, Gardette and Lemaire 1999). On the other hand, standoil spectrum displayed a significant lower intensity at 3013 cm-1 and a moderate signal at 1657, 1266 and 972 cm-1. These features, however, even if still recognizable within spectra of fresh mixtures, seemed to vanish along with the curing of oil paint films.

29A most remarkable result of FTIR-ATR spectroscopy was the evidence of a broad band (1530- 1620 cm-1), assigned to zinc carboxylates, displayed with no reasonable difference of intensity in all spectra of zinc white aged samples (Figure 11). Since standard spectra of zinc soaps expected inside an aged oil paint film are reported in literature (Robinet e Corbeil 2003; Mazzeo et al. 2008), bandwidth can be related to the presence of carboxylates with different molecular weight. On the other hand, within lead white aged samples the presence of carboxylates should be detected around 1541 cm-1 (lead stearate and palmitate), where in fact two small absorptions appear at 1548 and 1536 cm-1; the signal intensity, however, is extremely weak and peaks are hardly resolvable from background noise (Figure 12).

Fig.10 Comparison between samples on slides which have undergone different aging treatments

Fig.10 Comparison between samples on slides which have undergone different aging treatments

The linseed oil mixtures fairly appear as the most yellowed ones, especially in combination with lead white, while those prepared with poppy oil present the slightest changes in color. However, after UV irradiation these differences are cleared and all paint samples result neatly bleached and whiter than corresponding naturally aged samples.

Credits: Pratali 2012.

Fig.11 FTIR spectroscopy: comparison between zinc white in cold-pressed linseed oil aged spectra

Fig.11 FTIR spectroscopy: comparison between zinc white in cold-pressed linseed oil aged spectra

Appearance of a broad band between 1520 and 1630 cm-1 can be noticed in aged samples only.

Credits: Pratali 2012.

Fig. 12 FTIR spectroscopy: comparison between lead white in cold-pressed linseed oil aged spectra

Fig. 12 FTIR spectroscopy: comparison between lead white in cold-pressed linseed oil aged spectra

Carboxylate diagnostic absorptions, expected around 1541 cm-1, are hardly discernible from background noise.

Credits: Pratali 2012.

Discussion

  • 10  Linseed oil,  since it contains less than 0.4% of oleic acid, is almost completely devoid of this (...)

30Visual examination of paint sample drying process has confirmed poppy oil as the slowest in drying, especially when mixed with zinc white, which - unlike lead pigments - does not accelerate in any way the film oxidation (Carlyle 2010). This slowness, in comparison with linseed oil, is due to the lack of linolenic acid, and to the presence of a significant percentage of oleic acid (not siccative). The latter parameter is particularly interesting because it is a poppy oil typical feature10, and can also be put in relation with data reported within Mecklenburg studies, who had unexpectedly found significant amounts of free oleic acid in zinc aged paint samples (Rogala et al. 2010; Maines et al. 2011). A direct relationship between presence of oleic acid and film stiffening, however, has not been assigned yet.

31After first phase of aging the yellowing undergone by linseed oil paint samples is a typical feature of this medium, related to its high content of linolenic acid (C 18:3) - more than any other drying oil - that therefore implies a greater number of unsaturated bonds. The presence of double bonds increases reactivity and drying rates but also produces absorption centers that over time accentuate the yellowing of paints, as it is particularly evident in the case of white pigments. Heat action during thermal aging accelerates cross-linking and film oxidation, making it sooner appear dry to the touch but considerably yellowed as well. The great difference in yellowing rates observed between linseed oil and poppy oil paint samples is also explained by the fact that while the former has not undergone any chemical or physical treatment (it is in fact referred to as “cold-pressed” and contains trace amounts of mucilage, q. v. Kremer Pigmente 2012a), the latter has been somehow refined in a way that unfortunately was not released by manufacturers.

32After UV irradiation all white paints showed a substantial whitening, especially in case of very yellowed samples, such as those of white lead in linseed oil. It is actually a well known phenomenon, for which paint films turn yellow if stored in the dark, while fade if placed to direct sunlight, as it is particularly evident in the case of white pigments (Levison 1985). Photolytic action of ultraviolet radiation break the double bonds present in the exposed area and counteracts the yellowing and reinstates the former appearance. This result, however, is still a forced action, equal to a surface abrasion, involving phenomena of bond splitting and weakening of the film: a certain stiffening has been in fact noticed on all paint samples.

33As for what concerns degradation phenomena, after three stages of artificial aging only zinc white in poppy oil mixtures applied on canvas - and to a far lesser extent also those in linseed oil - have produced appreciable effects of paint failure. After first thermal aging several thin parallel cracks appeared along weft direction on sample 05 and increased after UV irradiation, when other extra fractures were produced at right angles. RTI technique has made possible to identify how crack positions correspond to sites of yarn overlapping, where a lower thickness of film is found. Since cracking pattern mostly reflects the underlying fabric structure and its appearance was found on paint samples on canvas only, a relationship between crack formation and reactivity of support is presumed. Canvas dimensional changes are usually more severe along warp direction and in facts major cracks were found at right angles to warp yarns, in areas where film thickness is lower and cohesion forces are weaker. Since RH values have never been a constant parameter during all aging phases, it might be reasonable to assume that dimensional changes undergone by canvas to accommodate RH fluctuations may have contributed to enhance the film mechanical degradation. Cracking phenomena observed in this study cannot however be attributed to RH changes alone, otherwise, if it did, it would not be possible to detect such a marked difference between sample 02 (naturally aged) and sample 05 (artificially aged) in terms of decay intensity. In fact RH variations suffered by aged samples were indeed intense and sudden, but not so different from those naturally occurring in standard conditions. It has therefore to be - as it always happens when studying art materials - the contribution of several factors.

34Crack formation was also detected below the yellow ocher layer on paint sample 17. This fact can be probably attributed to a restraint action caused by the ochre upper paint layer that, being mixed with standoil, was more flexible and durable and resulted unaffected by failure of zinc white underpaint, merely having reproduced the cracking texture on its surface (Figure 9). These observations thereby confirm a note recorded by Kühn (1986) on the practice of mixing zinc white with standoils in order to prevent chalking and delamination.

35After last stage of aging some very thin cracks were displayed also on sample 04, although initially it seemed to have an excellent resistance. This fact confirms the responsibility of zinc white in developing of failure processes, but also suggests the importance of oil medium in their progression rate.

36Soap formation processes were clearly recognized by FTIR-ATR spectroscopy, with appearance of a broad band between 1520 and 1630 cm-1 within aged samples of zinc white paint films alone. Bandwidth is attributed to the simultaneous presence of various types of carboxylates, including also stearate, palmitate and oleate of zinc, whose individual signals have however not been discerned. A quantitative evaluation has however not been possible and differences in signal intensity between various oil media were not recognized. On the contrary identification of lead soap on lead white paint samples, although far more documented inside scientific literature, was definitely less clear. Zinc soaps formation rate seems thus considerably higher than lead soaps, thus suggesting a greater reactivity of the pigment.

Indications for a conservative approach

37On the basis of both literature records and data collected within the present study some guidelines for the conservation of easel paintings can be defined. Since the main physical decay is related to underpainted ground layers, a full evaluation of its extension and intensity might not be easily given. A preventive facing of the whole paint surface is thus advised, in order to contain delamination and possible paint losses. The lack of cohesion detected within zinc white ground layers usually suggests an intervention with consolidating treatments, which requires adhesives with good penetration properties. In the case of synthetic polymers this means a material with low molecular weight and dimensions, which can penetrate a thin porosity. Besides, since zinc oxide powders are extremely fine and chalking processes can determine the formation of sub-micron vacuums, the penetration matter is of the utmost importance. A proper choice of solvents and of applicative methods is also important in order to reach a better penetration, and the use of the low pressure suction table - if compatible with size and features of the painting - could be an effective procedure. The lack of flexibility observed in zinc oxide paint films, which inhibits the possibility of following dimensional changes of canvas support, should dissuade from treatments involving significant amounts of moisture or water-based cleaning operations, which could promote dimensional changes in canvas support. For what concerns mechanical properties, the choice of adhesives with elastic properties and resistant to aging should be advised; a stiff and strong material insertion could in fact determine on a long term the building up of inner stresses, eventually leading to more severe failure; use of animal glues should thus be avoided.

38At the conclusion of any conservative operation it is then of the utmost importance to set a programmed monitoring plan, so as to recognize in advance a possible new evidence of decay within ground layers.

Conclusion

39Although many question are yet to be solved, some noticeable observations have been made about zinc white paint films and variables affecting their behavior:

  • even if spectral analysis did not evidenced a significant difference, paint film reactions showed a great difference both on the physical and macroscopic level according to the type of oil medium employed in the mixture;

  • climatic parameters chosen for accelerated aging were not so far from real conditions, and the ultimate effect of all three steps can be approximately compared to a few years of natural aging in standard conditions: the appearance of cracking on sample 05 and its enhancing in such a short time - whereas no physical damage has shown on lead white paint samples - confirms the pigment tendency to promptly become stiff and brittle;

  • the critical role played by the support and its reactivity to dimensional changes connected with climatic fluctuations is suggested by the appearance of cracking phenomena on canvas samples alone; thus stiffening and embrittlement do not represent dangerous conditions if not combined with the presence of a flexible and unstable support;

  • since major cracks have compared aligned to yarns directions, and especially at right angles to warp direction - along which major dimensional changes of canvas support are found - a critical role of RH fluctuations in enhancing the paint failure can be presumed;

  • although in literature the presence of lead soaps is far more recorded, the major intensity of carboxylate band displayed in FTIR spectra of zinc white aged samples confirms the thesis of a higher rate in soap formation processes: zinc white reactivity in oil seems higher than lead white, thus emphasizing its peculiar condition, while degradation phenomena related to the formation and migration of lead soaps - well accounted inside recent scientific literature - must develop in a greater lapse of time.

40Many new questions aroused at the end of this study, such as those concerning the causes of poppy oil peculiar behavior, the effective role played within aging processes by various climatic conditions, the unique drying and soap formation processes of zinc white paint films, the possible existence of a link between relative percentage of fatty acid within an oil medium and a greater tendency to embrittlement (i. e. a greater oleic acid content, as in the case of poppy oil). Variables observed within this experimentation certainly do not extinguish the possible variety of conditions that can be encountered in a real case. The presence of other materials, a different stratigraphy or a more complex one, an alternative composition of mixtures, could be all critical factors, leading to completely different results of decay. Execution of further experimentations and chemical analysis allowing a higher characterization of product carboxylates (i. e. via extraction processes) and of physical changes (i. e. a wider RTI campaign, along with microscopy observations on cross sections) could also be recommended.

Haut de page

Bibliographie

BOON, J. J., HOOGLAND, F., KEUNE, K., “Chemical processes in aged oil paints affecting metal soaps migration and aggregation, dans AICPaintings Specialty Group, Postprints of Annual Meeting in Providence. Rhode Island, 2006, V. 19, p. 18-25.

BOON, J. J., VAN DER WEERD, J., KEUNE, K., “Mechanical and chemical changes in Old Master paintings: dissolution, metal soaps formation and remineralization process in lead pigmented ground/intermediate paint layers of 17th century paintings”, dans Thirteenth Triennal Meeting, Rio de Janeiro, 22-27 September 2002: ICOM Committee for Conservation Preprints, V. 1, p. 401-406.

CARLYLE, L., The Artist’s Assistant: Oil painting Instruction Manuals and Handbooks in Britain 1800-1900 with references to Selected Eighteenth Century Sources, London, Archetype Books, 2001.

CRUICKSHANCK, S. J., Oxide of zinc: its nature, properties and uses, with special reference to the making and application of paint, London, The Trade Papers Publishing Co., 1909.

HIGGITT, C., SPRING, M., SAUNDERS, D., “Pigment-Medium Interactions in Oil Paint Films Containing Lead-based Pigments”, dans WAAC Newsletter, V. 27, n. 2, p. 12-16.

KREMER PIGMENTE (a), Linseed Oil, cold pressed, fiches tecniques consultable en ligne sur <http://www.kremer-pigmente.com/en/mediums--binders-und-glues/oils/natural-oils/linseed-oil--cold-pressed-73054.html>, [20-03-2012].

KREMER PIGMENTE (b), Leinöl-Standöl dick, fiches tecniques consultable en ligne sur <http://www.kremer-pigmente.com/media/files_public/73201.pdf>, [20-03-2012].

KREMER PIGMENTE (c), Bleiweiss, fiches tecniques consultable en ligne sur <http://www.kremer-pigmente.com/de/product/cremnitz-white-46000.html?info=4661&sorting=model>, [20-03-2012].

KÜHN, H., “Zinc White”, dans R. L. Feller (ed.), Artist’s pigments: a handbook of their history and characteristics, V. I, Cambridge, Cambridge University Press, 1986, p. 169-186.

LEVISON, H. W., “Yellowing and Bleaching of Paint Films”, dans Journal of the American Institute for Conservation, V. 24 (2), p. 69-76.

MAINES, C. A., et al., “Deterioration in Abstract Expressionist Paintings: Analysis of Zinc Oxide Paint Layers in Works from the Collection of the Hirshhom Museum and Sculpture Garden, Smithsonian Institution”, dans MRS Proceedings, vol. 1319, p. 275-284.

MALLÉGOL, J., GARDETTE, J., LEMAIRE, L., “Long-Term Behavior of Oil-Based Varnishes and Paints. I. Spectroscopic Analysis of Curing Drying Oils”, dans JAICS, V. 76, n. 8, p. 967-976.

MANZANO, E., et al., “Discrimination of aged mixtures of lipidic paint binders by Raman spectroscopy and chemometrics”, dans Journal of Raman Spectroscopy, consultable en ligne sur <http://onlinelibrary.wiley.com/doi/10.1002/jrs.3082/abstract>, [15-3-2012].

MAZZEO, R., et al., “Attenuated total reflection micro FTIR characterization of pigment-binder interaction in reconstructed paint films, dans Anal. Bioanal. Chem, V. 392, p. 65-76.

MECKLENBURG, M. F., Meccanismi di cedimento nei dipinti su tela: approcci per lo sviluppo di protocolli di consolidamento, Padova, Il Prato, 2008.

MÜLLER, H. G., Schoenfeld’s Malerfibel: Pigmente und Bindemittel, Dusseldorf, Dr. Fr. Schoenfeld & Co., 1962.

OYMAN, Z. O., MING, W., Van der Linde, R., “Oxidation of drying oils containing non-conjugated and conjugated double bonds catalyzed by a cobalt catalyst”, dans Progress in organic coatings, V. 54, p. 198-204.

PETIT, G., The manufacture and comparative merits of white lead and zinc white paints, traduit par D. Grant, London Scott, Grenwood & Son., 1907.

PRATALI, E., Le preparazioni pittoriche a base di bianco di zinco in olio: meccanismi di degrado e problemi di conservazione - Analisi conoscitiva e intervento di restauro su un’opera di Ugo Malvano del 1929, thèse, superviseur: P. Buscaglia, E. Diana, F. Rovati, Università degli Studi di Torino, année académique 2011-2012.

ROBINET, L., CORBEIL, M., “The Characterization of Metal Soaps, dans Studies in Conservation, V. 48, n.1, p. 23-40.

ROGALA, D., et al., “Condition problems related to zinc oxide underlayers: examination of selected Abstract Expressionist paintings from the collection of the Hirshhorn Museum and Sculpture Garden, Smithsonian Institution”, dans JAIC, n. 49, p. 96-113.

UEBELE, C. L., Paint making and color grinding. A practical treatise for paint manufacturer and factory managers, New York, The Painters Magazine, 1913.

VAN LOON, A., Color changes and chemical reactivity in seventeenth-century oil paintings, PhD Thèse, University of Amsterdam, p. 120-204.

WITLOX, M., CARLYLE, L., “‘A perfect ground is the very soul of the art’ (Kingston 1835): ground recipes for oil painting, 1600-1900”, dans ICOM-CC 14th Triennial Meeting Preprints, V. 1, London, James&James, 2005, p. 519-528.

Haut de page

Notes

1  Stress-strain tests made on zinc white in linseed oil paint samples naturally aged in standard conditions show a strain rate of only 0,3%. The strain caused by a simple stretching operation on a standard size painting was measured and compared to that endured by zinc white paint films before failure. Zinc white paint samples 14 years-aged are far from reaching even that minimum strain, proving themselves extremely stiff and less flexible than lead white paint films (Mecklenburg 2008; Rogala et al. 2010).

2  A protrusion is a globular-shaped agglomerate of lead soaps that can reach macroscopic dimensions. Clusters of lead soaps formed within lead rich paint layers can migrate toward surface where they accumulate and grow assuming a whitish and translucent appearance (q.v. Boon, Van der Weerd and Keune 2002; Higgitt, Spring and Saunders 2005; Boon, Hoogland and Keune 2006; Van Loon 2008).

3  Distributed as “Cremnitz White” (Kremer Pigmente 2012c)

4  Maimeri has not released any information about treatments cast on poppy oil, but a certain reduction in free acid content can be noticed from the composition data (see Figure 5).

5  For the preparation of canvas samples the CTS 1111 linen “boiled long linden tree” canvas has been chosen (structure 1:1 and count of 15x15 yarns/cm). The canvas has been stretched and sized with two layers of animal glue. The glue has been prepared the day before: dried flakes of rabbit skin glue and animal bone glue (CTS Europe) have been separately made swell in a proper amount of osmotic water (glue-water ratio of 1:10). Two coats of warm glue size have then been applied by brush on the stretched canvas. After complete drying the sized canvas has been cut in 18 squares and each one of them has been stretched on a smaller wooden frame.

6  According to Boon soap formation has been observed by Carlyle after about a month of exposure at 50 °C and 80% RH (Boon, Hoogland and Keune 2006).

7  Paint samples have entered the oven after a few days after white paint application. After 500h an upper yellow ochre layer has been layered on 07-18 paint samples on canvas, and thereafter they have been left again inside the oven for 1700h .

8  Each cycle lenght ranged from 6 to 12h.

9  This last step has involved all samples on canvas previously aged in oven and climatic chamber (04, 05, 06, 10, 11, 12), and series 2 of samples on slides (D2, E2, F2, J2, K2, L2), while series 1 has been excluded from this further stage. This has permitted to make a direct evaluation of effects caused by photochemical aging only, as shown in Figure 10.

10  Linseed oil,  since it contains less than 0.4% of oleic acid, is almost completely devoid of this non siccative unsaturated acid, while the poppy oil contains over 22% of the same. Triglycerides composition of standoil is not shown (see Figure 5), but since it is obtained from a linseed oil, similar rates of fatty acids could be presumed (q. v. Kremer Pigmente 2012b).

Haut de page

Table des illustrations

Titre Fig. 1  Stress-strain test performed over naturally aged white paint samples
Légende Zinc oxide paint film proves particularly stiff and brittle compared to other white pigments commonly encountered as ground layers. The maximum strain suffered before paint failure appears in fact to be only a fifth of that imposed over paint by a stretching operation on a medium size canvas painting (Mecklenburg 2008).
Crédits Credit: Mecklenburg 2008
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-1.jpg
Fichier image/jpeg, 68k
Titre Fig. 2 Example of selective cracking of carbon black painted over a zinc oxide ground of a painting by Franz Kline, (“Palladio”, 1961)
Légende A black paint failure was noticed over the zinc white underpaint only, while upper layers are intact where painted over lead white ground (Rogala et al. 2010).
Crédits Source: Rogala et al. 2010
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-2.jpg
Fichier image/jpeg, 28k
Titre Fig. 3 SEM analysis of raw pigments: zinc white (a) and lead white (b)
Légende EDX analysis showed the presence of calcium carbonate (darker square particles) inside lead white composition. Size ratio of lead white particles is approximately 10 times higher than zinc oxide, this fact determining a higher oil-taking capacity of the latter.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-3.jpg
Fichier image/jpeg, 112k
Titre Fig. 4 Sample series, aging phases and analysis summary
Légende Legend: ZW=zinc white; LW=lead white; LIN=cold-pressed linseed oil; POP=refined poppy oil; STO=linseed standoil; PIG=Burgundy yellow ochre upper layer; ½PIG= Burgundy yellow ochre layer applied on half of the white ground surface; N=non-aged sample (natural aging). Number of “x” indicates how many time an analysis has been repeated on the same sample.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-4.jpg
Fichier image/jpeg, 180k
Titre Fig.5 Summary of oil medium compositions
Légende Information have been collected from technical data sheets available online or directly provided by manufacturers via private communication (Kremer 2012a and 2012b).
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-5.jpg
Fichier image/jpeg, 68k
Titre Fig. 6 Realization of paint samples
Légende Paint films have been layered both on microscope slides (a) and on sized canvas (b).
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-6.jpg
Fichier image/jpeg, 420k
Titre Fig. 7 RTI images of paint sample 05 after second (a) and after third (b) phase of aging
Légende RTI permits to enhance morphological features by manipulating light emission sources. Cracks appear deepened and more outlined after third stage of aging.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-7.jpg
Fichier image/jpeg, 92k
Titre Fig. 8 RTI detail of sample 05
Légende Surface after third step of aging (left) and scheme of cracking network (right)
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-8.jpg
Fichier image/jpeg, 196k
Titre Fig. 9 RTI detail of sample 17 after third step of aging
Légende By reduction of chromatic values and light manipulation (b) it is possible to highlight the appearance of cracking underneath the yellow ochre layer.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-9.jpg
Fichier image/jpeg, 272k
Titre Fig.10 Comparison between samples on slides which have undergone different aging treatments
Légende The linseed oil mixtures fairly appear as the most yellowed ones, especially in combination with lead white, while those prepared with poppy oil present the slightest changes in color. However, after UV irradiation these differences are cleared and all paint samples result neatly bleached and whiter than corresponding naturally aged samples.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-10.jpg
Fichier image/jpeg, 272k
Titre Fig.11 FTIR spectroscopy: comparison between zinc white in cold-pressed linseed oil aged spectra
Légende Appearance of a broad band between 1520 and 1630 cm-1 can be noticed in aged samples only.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-11.jpg
Fichier image/jpeg, 100k
Titre Fig. 12 FTIR spectroscopy: comparison between lead white in cold-pressed linseed oil aged spectra
Légende Carboxylate diagnostic absorptions, expected around 1541 cm-1, are hardly discernible from background noise.
Crédits Credits: Pratali 2012.
URL http://journals.openedition.org/ceroart/docannexe/image/3207/img-12.jpg
Fichier image/jpeg, 100k
Haut de page

Pour citer cet article

Référence électronique

Elena Pratali, « Zinc oxide grounds in 19th and 20th century oil paintings and their role in picture degradation processes »CeROArt [En ligne], EGG 3 | 2013, mis en ligne le 10 mai 2013, consulté le 18 avril 2024. URL : http://journals.openedition.org/ceroart/3207 ; DOI : https://doi.org/10.4000/ceroart.3207

Haut de page

Auteur

Elena Pratali

The author has obtained the master degree in Conservation and Restoration of Cultural Heritage, University of Turin. Her area of expertice includes planning, intervention and preventive conservation of paintings on wooden or canvas support, and of objects of contemporary art. Contact: elena.pratali@libero.it

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search