Formation of lead dioxide and the development of a dark red ‘patina’
on outdoor lead sculpture
Master thesis, metals specialization
MA Conservation and Restoration of Cultural Heritage
Author: Manuela Toro (UvA) Supervisor: Tonny Beentjes (UvA) Second Reader: Bas van Velzen (UvA)
Advisors: Susanne Kensche (KMM), Luc Megens (RCE), Ineke Joosten (RCE), Arie Pappot (RMA), Maarten van Bommel (UvA), Ella Hendriks (UvA)
Date: 19/06/2018
Toro, UvA 2018 2
Table of contents
Abstract 4
English 4
Nederlands 5
1. Object metadata 7
2. Introduction 8
2.1. Relevance to the field of conservation 8
3. The problem 9
3.1. Treatment history 10
3.2. Current scientific knowledge 13
3.3. Hypotheses 15 3.3.1. Research questions 16 4. Historical context 16 4.1. The object 16 4.2. The artist 19 4.3. The foundry 20 5. Literature review 20
5.1. The Queluz case 20
5.2. Formation of lead dioxide 24
5.3. Is it a patina? 25
6. Methodology 28
6.1. Optical analysis 28
6.1.1. Visual analysis 28
6.1.2. Dino-‐Lite optical microscopy 29
6.2. Sampling and embedding 29
6.3. Microscopic analysis of cross-‐sections 31
6.4. SEM-‐EDX 31
6.5. pH measurements 32
7. Results 33
7.1. Optical analysis 33
7.1.1. Visual analysis 33
7.1.2. Dino-‐Lite optical microscopy 33
7.2. Sampling and embedding 35
7.3. Microscopic analysis of cross-‐sections 36
7.4. SEM-‐EDX 38
7.5. pH measurements 46
7.6. Discussion 47
7.7. Health and safety considerations 51
7.8. Answers to hypotheses and research questions 52
7.8.1. Stratigraphy 54
7.8.2. Further research 54
Toro, UvA 2018 3
9. Acknowledgements 56
10. Bibliography 57
11. Appendix I: SEM-‐EDX results of antimony-‐rich areas 61
12. Appendix II: additional SEM-‐EDX results 63
All figures and photos were created by Manuela Toro unless stated otherwise.
Toro, UvA 2018 4 Abstract
English
Formation of lead dioxide and the development of a dark red ‘patina’ on outdoor lead sculpture.
Keywords: Aristide Maillol, lead, outdoor sculpture, water, patina, lead dioxide, lead carbonate, lead whiskers.
This thesis for the Master of Arts in Conservation and Restoration of Cultural Heritage at the University of Amsterdam for the spring of 2018 researches the formation of a dark red layer at the surface of Aristide Maillol’s lead sculpture L’Air, exhibited outdoors at the Kröller Müller Museum in Otterlo (The Netherlands).
Dark red stains started to appear at the surface of L’Air in the mid-‐1990’s, indicating that a chemical change is occurring on the surface of the object apart from the expected behavior of lead oxidation in an outdoor environment. X-‐Ray diffraction (XRD) and scanning electron microscopy with energy dispersive X-‐Ray spectroscopy (SEM-‐EDX) analyses of the dark red corrosion layer were carried out in 2016 by the Cultural Heritage Agency of The Netherlands (RCE). Lead dioxides plattnerite (β-‐PbO2) and scrutinyite (α-‐PbO2) were identified as the red compounds. However, further research was needed in order to better understand the corrosion mechanism taking place. The process by which lead dioxide forms at the surface of the sculpture is still under discussion. This is the main focus of the thesis.
Three hypotheses were created pondering possible ways in which plattnerite and scrutinyite form at the surface of the object. Visual analysis and microscopy were used to better understand the patterns in which the stains formed and also to observe the morphology of the corrosion layer. Samples were taken and analyzed under SEM-‐EDX in order to carry out elemental analysis and to be able to investigate the stratigraphy of the cross sections. Last, pH measurements were also carried out in order to evaluate possible pH differences between the dark red and gray areas of the sculpture, and also to compare with available literature.
Visual analysis helped to confirm how the dark red stains follow the rain patterns. Microscopic analysis with a handheld Dino-‐Lite helped to observe the uneven, scale-‐like development of the corrosion layer. It also allowed to better distinguish how white spots of lead carbonate are mixed within the dioxide layer. Microscopic analysis of the cross sections with an optical microscope helped to see how the compounds are mixed at the surface. SEM-‐EDX analysis confirmed previous elemental composition results, and provided images of lead whiskers forming at the interface between the bare metal and the corrosion layer. The pH measurements indicated a slightly acidic environment at the surface of L’Air.
Toro, UvA 2018 5
The results showed that the surface of the sculpture is reacting in various ways to oxygen and water in its environment. The corrosion layer forms by at least two possible chemical mechanisms, which may in turn be continuously reversed by further reactions with water and fluctuations in pH. Also, changes in morphology such as surface cracking and whisker growth allow for these mechanisms to continue in a cyclic manner. Additional pH measurements, coupon tests, metallographic analysis, and electrode potential measurements are recommended for further research.
Nederlands
Vorming van looddioxide en de ontwikkeling van een donkerrode 'patina' op een sculptuur in de buitenlucht.
Sleutelwoorden: Aristide Maillol, lood, beeldhouwwerk buitenshuis, water, patina, looddioxide,
loodcarbonaat, loodhaartjes.
Deze scriptie voor de Master of Arts in Conservation and Restoration of Cultural Heritage aan de Universiteit van Amsterdam voor het voorjaar van 2018 onderzoekt de vorming van een donkerrode laag aan het oppervlak van Aristide Maillols sculptuur L'Air dat buiten tentoongesteld is in het Kröller Müller Museum in Otterlo (Nederland).
Aan het oppervlak van L'Air verschenen halverwege de jaren negentig donkerrode vlekken, wat aangeeft dat er een chemische verandering optreedt op het oppervlak van het object naast het verwachte gedrag van lood oxidatie in een buitenomgeving. X-‐Ray Diffraction (XRD) en scanning elektronenmicroscopie met energie-‐dispersieve röntgenspectroscopie (SEM-‐ EDX) analyse van de donkerrode corrosielaag werd in 2016 uitgevoerd door de Rijksdienst voor het Cultureel Erfgoed (RCE). Lead dioxides platnerite (β-‐PbO2) en scrutinyite (α-‐PbO2) werden geïdentificeerd als de rode verbindingen. Er was echter verder onderzoek nodig om het corrosie mechanisme dat plaatsvindt beter te begrijpen. Het proces waarbij looddioxide aan het oppervlak van het beeldhouwwerk wordt gevormd, is nog steeds onderwerp van discussie. Dit is de belangrijkste focus van deze scriptie.
Er werden drie hypotheses opgesteld over mogelijke manieren waarop plattnerite en scrutinyite zich aan het oppervlak van het object vormen. Visuele analyse en microscopie werden gebruikt om de patronen waarin de vlekken vormen en de morfologie van de corrosielaag beter te begrijpen. Monsters werden genomen en geanalyseerd onder SEM-‐EDX om elementen analyse uit te voeren en om de stratigrafie van de dwarsdoorsneden te kunnen onderzoeken. Als laatste werden ook pH-‐metingen uitgevoerd om mogelijke pH-‐verschillen tussen de donkerrode en grijze gebieden van het beeldhouwwerk te evalueren, en ook om deze te vergelijken met beschikbare literatuur.
Toro, UvA 2018 6
Visuele analyse bevestigde dat de donkerrode vlekken de regenstroom patronen volgen. Microscopische analyse met een handheld Dino-‐Lite hielp bij het waarnemen van de oneffenheden, schubachtige ontwikkeling van de corrosielaag. Het liet ook toe om beter waar te nemen hoe witte vlekken van loodcarbonaat zich mengen in de kooldioxidelaag. Microscopische analyse van de dwarsdoorsneden met een optische microscoop hielp om te zien hoe de verbindingen aan de oppervlakte gemengd zijn. SEM-‐EDX-‐analyse bevestigde eerdere resultaten van de elementaire samenstelling en verschafte beelden van loodhaartjes die zich vormen op het grensvlak tussen het kale metaal en de corrosielaag. De pH-‐metingen duiden op een enigszins zure omgeving aan de oppervlakte van L'Air.
De resultaten toonden aan dat het oppervlak van het beeldhouwwerk op verschillende manieren reageert op zuurstof en water in zijn omgeving. De corrosielaag vormt zich door ten minste twee mogelijke chemische mechanismen, die op hun beurt continu kunnen worden omgekeerd door verdere reacties met water en fluctuaties in de pH. Ook zorgen veranderingen in de morfologie, zoals oppervlakte-‐scheuren en de groei van de loodhaartjes ervoor dat deze mechanismen op een cyclische manier kunnen doorgaan. Aanvullende aanvullende pH-‐ metingen, coupontests, metallografische analyse en elektrodepotentiaalmetingen worden aanbevolen voor verder onderzoek.
Toro, UvA 2018 7 1. Object metadata Title: L’Air
Author: Aristide Maillol
Date: Conceived 1939, this cast 1962 Material(s): Lead, iron armature Technique: Casting
Foundry: Georges Rudier Foundry (Paris)
Signature: A. Maillol (lower left), 3/6 (lower left), Georges Rudier Fondeur Paris (side left) Dimensions: 177 x 239 x 99,5 cm
Owner/collection: Kröller Müller Museum (Otterlo) Inventory number: KM 127.576
Acquisition date: 1962
Date of examination: November 2017 through May 2018
Toro, UvA 2018 8
2. Introduction
Lead has been used in outdoor sculptures in the past because of its economical and material qualities, including resistance to corrosion in different environments. However, the dark red stains that have appeared on Maillol’s L’Air at the Kröller Müller Museum (Otterlo) indicate that there has been a change of chemical nature at the surface of the sculpture (Fig. 1.1). The topic was put forward by thesis supervisor Tonny Beentjes after previous attempts by the museum to undertake research about this object.
The goal of this research is to increase our understanding about how the already identified corrosion products (lead dioxides plattnerite and scrutinyite) form at the surface of the sculpture. This was done by taking samples of the object and analyzing them via SEM-‐EDX and microscopy. An experiment to measure the pH of the surface of the sculpture has also been carried out. Principally, the aim is to provide a stratigraphy of the corrosion layer once the research is concluded. By further understanding the degradation mechanisms of lead ‘patinas’, this research could help to better diagnose the condition of the artwork and serve as a basis for future observations, treatments, and scientific research.
The thesis will be divided in five main sections. First, a description of the problem, where current scientific knowledge is discussed and hypotheses are brought forward. Second, a description of the historical context around the object, the artist and the foundry. Third, a literature review about the formation of lead oxides, past conservation literature and whether the dark corrosion present in L’Air can be classified as a patina or not. Fourth, methodology, describing the process by which samples of the object were taken and analyzed via SEM-‐EDX, microscopy, and also the pH measurement experiment. And fifth and last, results, discussing the outcome of the research and leading towards a conclusion.
2.1. Relevance to the field of conservation
While a lot has been researched about the formation of lead carbonates, the formation of lead corrosion layers of a dark red color involving plattnerite and scrutinyite have been widely understudied in relation to cultural heritage objects. In that sense, this research will help improve the state of the art of said condition and also raise more questions for further research. This is especially relevant today given the recent interest in the degradation processes of lead brought up by similar observations in objects from various parts of the world.
On the other hand, to understand the corrosion mechanism by which the dark red layer is produced would signify locating the source of the problem, and in turn help to make more informed decisions if the object undergoes a conservation treatment in the future.
Toro, UvA 2018 9
3. The problem
A dark red color can be seen on most of the surface of L’Air in the form of stains and drips (Fig. 3.1). Its thin composition has a patina-‐like effect on several areas of the sculpture. Upon close examination, white powdery spots can be seen in the darkened areas.
The corrosion products in Maillol’s L’Air are not only distracting to the eye, but also pose risks to the sculpture’s structure and integrity if the corrosion is active1. Additionally, they become a risk for visitors since lead and its products are known to be toxic to humans. For these reasons, there is an impending interest in learning more about the formation of the dark red corrosion layer and its cause, thus helping conservators to potentially propose an appropriate treatment for the artwork in the near future.
Even though the dark red layer is not the most commonly known, the corrosion products that form it were identified in 2016 as plattnerite (β-‐PbO2) and scrutinyite (α-‐PbO2) by SEM-‐EDX and XRD analysis performed by the Rijksdienst voor het Cultureel Erfgoed2 (Cultural Heritage Agency of the Netherlands). These are polymorphs of lead (IV) oxide, (also known as lead dioxide, PbO2), which have also been found on lead water-‐supply pipes. The corrosion mechanisms by which plattnerite and scrutinyite form at the surface of the sculpture are yet to be explained. This is the main focus of the research.
Other cases of outdoor lead sculptures showing dark red stains have appeared in recent years in other parts of the world. The most relevant example for this research is that of the sculptures at the Palace of Queluz in Portugal, where plattnerite was also found3. The sculptures
1 Ongoing, new corrosion.
2 Joosten, I.; van Hoesel, A. L’Air, Aristide Maillol, 1930, Kröller Müller Museum, Onderzoek naar corrosieproducten. Rijksdienst voor het Cultureel Erfgoed, 2016. pg. 6.
3 Bernard, M.-‐C. ; Costa, V. ; Joiret, S. On unexpected colour of lead sculptures in Queluz: degradation of lead white. Corrosion Engineering, Science and Technology, 01 October 2010, Vol.45(5), pg. 341-‐344.
Toro, UvA 2018 10
at Queluz were treated in 2012, and have since developed the dark red corrosion layer again4, indicating that the process is still active and possibly as a cyclic mechanism.
3.1. Treatment history
The treatment history of the object has also yielded some clues. The sculpture was cared for in the same way as outdoor bronzes until the early-‐1990’s by undergoing light cleaning with soapy water and receiving applications of a beeswax coating once a year. It has not been described what kind of soap, beeswax, or tools were exactly used.
The period in which the treatment stopped seems to correspond with the approximate date in which the dark red stains were first spotted. This could partially confirm that the process involves water, since the wax coating would have isolated the sculpture from water during most of its life. The drip-‐like shape of the stains also strongly suggests that water is a key factor in the formation of the corrosion products. At this point, there seem to be no traces of beeswax left on the sculpture.
Condition reports begin to mention the dark corrosion stains around 19935, but no further descriptions are provided until 2001. From that year on, the sculpture was surveyed almost yearly. Photographic documentation has provided some indication of how and when the red stains started to appear and how quickly they developed (Table 3.1). A major increase in the amount of red corrosion can be observed between the years 2001 and 2005.
Date Image Description
1988 Only gray oxidation layer
present.
4 The author visited the Palace of Queluz in December 2017 and observed red stains with white spots on several of the sculptures.
Toro, UvA 2018 11 Unknown
Apparent dark red drip lines, perhaps from 1993 condition report(?).
1995
Difficult to tell because of poor image quality.
2001
Very thin layer of red corrosion on the head and chest, other areas not visible or referenced.
2003
Visible red corrosion on the top arm and leg, also on the right side of the base. Possiby increase of dark red stain on the chest.
Toro, UvA 2018 12
2005 Significant increase in the
amount of dark red corrosion.
2006 Clear presence of large, dark
red stains.
2012 Dark red stains continue to
develop in more areas of the sculpture.
2014 Apparent darkening of red
Toro, UvA 2018 13
2018 Dark red stains continue to
develop.
3.2. Current scientific knowledge
While lead is often thought to be resistant to corrosion, it is also known that it does corrode under particular environments and circumstances. In her book Metals and Corrosion: a
handbook for the conservation professional (2004), Lyndsie Selwyn describes outdoor lead
corrosion in the following way:
When exposed outdoors, lead usually corrodes until it develops an adherent film of insoluble lead compounds that protect the surface from further attack.
“When lead is first placed outdoors, it forms a thin layer of lead oxides (predominantly the yellow massicot but also some litharge). Over time, carbon dioxide that dissolves in any surface water on the lead reacts with the lead oxides to form cabinet compounds (mainly hydrocerussite but sometimes also cerussite). (…) If sulphur dioxide dissolves in the surface water, it reacts with hydrocerussite to form lead sulphite and lead sulphate. (…) If outdoor lead undergoes creep, its corrosion rate will increase because of the continued exposure of a fresh lead surface. (Selwyn, pg. 121-‐22).”
This behavior is not at all similar to what has been observed on L’Air since the mid-‐1990’s, when the sculpture started to show brown/red stains on the surface that developed and darkened in the relatively short time span of approximately 20 years.
The alloy composition doesn’t seem to be the cause of the problem. XRF analysis performed in 2016 by Arie Pappot (Rijksmuseum Amsterdam) showed that the alloy consists of lead with 4-‐6% of antimony. This type of alloy shows the same corrosion rates as pure lead, with the only benefit being the increase in hardness and resistance to creep.6 In Pappot’s analysis phosphorus and chlorine showed as trace elements in the dark gray areas (Fig. 3.2). He suggests
6 Corrosion of lead and lead alloys. ASM Handbook, Volume 13B: Corrosion: Materials. S.D. Cramer, B.S. Covino, Jr., editors, pg. 197.
Toro, UvA 2018 14
to study their presence more closely in order to determine whether they take part in the formation of the red corrosion products or not.7
Pappot also affirms that the dark spots do not correspond to the listed lighter elements but rather show a higher intensity of lead (Table 3.2).8
SEM-‐EDX and XRD analysis (Fig. 3.3) also performed in 2016 by the Rijksdienst voor het Cultureel Erfgoed (RCE) determined that the powdery white spots of corrosion were composed of lead carbonates cerussite (PbCO3) (predominant), and hydrocerussite (Pb3(CO3)2(OH)2) and
that the dark red products were lead dioxides plattnerite (β-‐PbO2) (predominant) and scrutinyite (α-‐PbO2).9 .This is the most important piece of information so far, since it will allow for a more detailed investigation of the compounds involved and also enable to compare with other case studies. It also confirms Pappot’s analysis that the dark corrosion products are not as closely linked to trace elements, but rather to polymorphs of lead (IV) oxide (PbO2). Given this identification, the red and white corrosion products will be referred to in this thesis according to these results.
7 Pappot, A. Corrosion phenomenon on L’Air by Aristide Maillol (KM 127.576). Rijksmuseum report, 2016. pg. 2. 8 Ibid, pg. 1.
9 Joosten, I. et al, 2016. pg. 6.
Fig. 3.2: “Graph SEQ Graph \* ARABIC 1: Deconvoluted spectra of test #31 showing reliable attribution of P and Cl peaks. S remains problematic.” (Pappot, 2016).
Table 3.2: “Table SEQ Table \* ARABIC 1: semi quantitative results of 10kV XRF spectra of eight different spots of corrosion according to their visual appearance: green is a relatively high value, red is low. Colors are comparable within one element only.” (Pappot, 2016).
Toro, UvA 2018 15 3.3. Hypotheses
Since it has been observed that the the dark red stains formed in the shape of drips after the waxing treatment was stopped, the first hypothesis would be that:
a) Lead dioxides plattnerite and scrutinyite are formed in a corrosion mechanism involving water at the surface of the sculpture.
Based on the literature and observations made by the author in December 2017 at the Palace of Queluz, where the problem recurred after treatment, a second hypothesis would be that:
b) Lead dioxides plattnerite and scrutinyite form via a cyclic corrosion mechanism at the surface of the sculpture.
The presence of white, powdery spots only in the areas where dark red staining is observed yields a third hypothesis:
Fig. 3.3: XRD results (Joosten et al, 2016, revised 2018).
Toro, UvA 2018 16 c) Lead dioxides plattnerite and scrutinyite’s formation processes are related to the presence of the lead carbonates cerussite and hydrocerussite.
3.3.1. Research questions
● How are lead dioxides plattnerite and scrutinyite formed at L’Air’s surface? ○ What is the role of water in the process?
○ What is the role of pH in the process? ○ What is the role of trace elements?
○ Is there a relationship with lead carbonates? if so, how?
○ Does creep play a role by continuously exposing a fresh lead surface? ○ Is the dark red layer a patina?
○ Is it a protective layer or does it harm the metal underneath? ○ Can the process be stopped?
○ Can the process be reversed?
4. Historical context
4.1. The object
Aristide Maillol’s L’Air is a larger-‐than-‐life size sculpture of a woman in the nude, the figure is leaning back and resting on her right hip on top of a rectangular base. The sculpture is located outdoors in the Museum’s gardens, and it is often touched by visitors. Both the figure and the base are made of lead. The design for this sculpture dates from 1939, but the sculpture was cast in 1962 after the death of the artist. It exists in an edition of six copies, the sculpture at the Kröller Müller Museum is numbered at the lower left of the base: 3/6. The artwork is signed “A. Maillol”, also at the base in the lower left. On the side, there is an inscription from the foundry were it was cast: “Georges Rudier Fondeur Paris”.
Today, the sculpture is of a dull gray-‐blue color overall, which is the expected appearance of a natural oxidation layer in lead, but it exhibits other less expected corrosion products as well that will be discussed in extensive detail throughout this thesis.
The first version of L’Air was realized in stone in 1938 as a commission by the city of Toulouse to make a memorial for French pilots (Fig. 4.1). Marlborough Gallery’s catalogue for an extensive Maillol exhibition in 2003 describes the project as “… a memorial to pilots in the Aéropostale killed in the line of duty.”10, however, in John Rewald’s account ‘Maillol
Remembered’ for the 1975 retrospective at the Guggenheim Museum (New York), he mentions
Toro, UvA 2018 17
that the commission honored French aviation pioneers, particularly Jean Mermoz.11 Rewald also writes that L’Air was previously conceived as a study dating around 1900, sketched and executed as a small terracotta figure.12 These first two versions include drapery unfolding around the figure as if being left behind, and carried away by the wind.
Maillol’s designs stand on the verge between abstraction and realism. Drawing both from nature and classical archetypes, his sculptures seem at once animated, tactile, modeled after reality, yet the softness of the lines, the gestures and inert facial expressions make them appear more classical, ageless and withdrawn. According to Rewald (1975), Maillol explained:
“What I am after […] is architecture and volume. Sculpture is architecture, is equilibrium of masses. This architectural aspect is hard to achieve. I endeavor to obtain it as did Polycletus. My point of departure always is a geometric figure – square, lozenge, triangle—because those the shapes which stand up best in space (pg.16).”
While drawing his contours almost exclusively from models, he combined those studies from nature with preconceived ideas. Also in Maillol’s words according to Rewald: “What I want […] is that the young girl depicted in a statue should represent all young girls, that the woman and her promise of maternity should represent all mothers.”13
11 Rewald, J. Maillol Remembered, in Artistide Maillol: 1861-‐1944. The Solomon R. Guggenheim Foundation, New York, 1975. pg.11.
12 Ibid. 13 Ibid. pg. 22.
Fig 4.1: L’Air, 1938. Stone. Ville de Toulouse. Source: retrieved from Wikimedia commons (2015).
Toro, UvA 2018 18
How exactly L’Air was transformed from the stone version into the equally well known bronze and lead counterparts is unknown (Fig.4.2). It is likely that Maillol’s patrons requested casts. Rewald (1975) writes:
“He told me that Vollard, in the early days, would buy his small terra cottas to have them cast in bronze. The artist would specify that the editions should be limited to ten casts; he added with resignation: ‘Well, he made ten casts, all right, except they turned to be ten thousand!’ (pg.11).”
It is also likely that material properties played a role in the design. The brittleness of stone may have required to include the drapery in order to help distribute the weight of sculpture at various points, while the hardness and malleability of metal allows the figure to balance on only one point. However, there is no literary evidence that supports this claim. Another outstanding remark is that the lead cast at the Jardin des Tuileries in Paris dates to 1932 (Fig. 4.3). A date six years earlier than the monument in Toulouse indicates that the drape-‐less version existed first, and was probably already being cast in lead -‐-‐perhaps even as a preliminary example-‐-‐ and bronze.
On the other hand, this date does not correspond with most written sources. Rewald’s comment about L’Air during his 1938 visit to the artist suggests that a large version was not even finished, pointing out to a possible discrepancy in the dating of L’Air: “His project having been approved, he was now faced with the problem of deriving from this piece of only a few inches an over life-‐size sculpture.”14
14 Ibid, pg. 11.
Fig 4.2: L’Air, design 1938, cast 1962. Lead. The J. Paul Getty Museum. Source: retrieved from The J. Paul Getty Museum.
Toro, UvA 2018 19
With all this shortage of information, it is also challenging to determine what the original surface of the lead casts must have been like. However, given photographic documentation of L’Air and other examples of lead sculptures, it is most likely that the dark gray surface was chafed and polished to an even sheen.
4.2. The artist
Aristide Maillol was born in France in 1861. His hometown, Banyuls-‐sur-‐Mer is part of the coast, but it is also close to the Albères Mountains and to the Spanish border. Its Mediterranean environment, port and vineyards were a source of inspiration for the artist throughout his entire life. He spent his early years between Perpignan, where he attended school and took drawing lessons and Banyuls, until he moved to Paris at age 20 in order to continue his studies as an artist.
In Paris, he studied both in the École des Arts Decoratifs and the École des Beaux-‐Arts. His circle of friends included Auguste Rodin, Édouard Vuillard, Maurice Denis and Paul Gauguin. During this time he worked with a variety of media, including drawing, painting, tapestry making, woodcarving, and sculpture. Even after moving to the outskirts to Paris with his wife Clotilde, he maintained a house and tapestry workshop in Banyuls where the family spent long seasons.
Ambroise Vollard, a Paris gallerist, gave Maillol his first solo show in 1902. They worked together until Vollard’s death in 1939. By the 1930’s, Maillol was well known and had exhibited his work in Germany, France, the Netherlands and the United States. In 1943, Maillol was injured in a car accident, and died in his house in Banyuls a few days later. Since then, major retrospectives have taken place in museums around the world, including the Basle Kunsthalle,
Fig 4.3: L’Air, 1932, Jardin des Tuileries. Source: retrieved from Wikimedia commons (2010).
Toro, UvA 2018 20
the Musée National d’Art Moderne (Paris), and the Guggenheim Museum (New York). In 1983 the Foundation Dina Vierny – Musée Maillol was created in Paris, and in 1994 a Maillol Museum opened in Banyuls.
4.3. The foundry
The history of the Rudier Foundries is notoriously complex and full of uncertainty. Alexis Rudier ran a well-‐known foundry in Paris during the late 1800’s. He worked very closely with artist Auguste Rodin, many of his casts are signed ‘Alexis Rudier Fondeur Paris’. Alexis had two brothers, François and Victor Rudier. Upon his death in 1897, François is left in charge of the foundry. In 1902, Alexis’ son, Eugène Rudier takes over from his uncle and the Rodin production. Eugène also casts many of Maillol’s sculptures.
When Eugène died in 1952, Georges Rudier (son of Victor Rudier) was encouraged to open a foundry, using equipment from his cousin’s business and thus preserving the family’s tradition and knowledge, and so he did. Around 1976 the Maillol Museum stopped working with George Rudier because of rumors about suspicious activities in the Foundry, the Rodin Museum soon followed suit. The Georges Rudier Foundry was connected to the the Hain scandal during the late 1980’s-‐early 1990’s, when it was discovered that collector Guy Hain, one of the foundry’s clients, had sold many unauthorized Rodin casts signed Alexis Rudier. The rest is still unknown.
5. Literature review
5.1. The Queluz case
As previously mentioned, the most similar (and therefore comparable) example of the
formation of a dark red layer on outdoor lead sculpture is that of the Palace of Queluz in Portugal. There, a group of mid-‐eighteenth century sculptures made by John Cheere developed a dark red layer mixed with a white ‘patina’ made of lead oxide and lead hydroxycarbonate, probably shannonite (PbCO3.PbO)15.
In Bernard et al’s article On unexpected colour of lead sculptures in Queluz: degradation
of lead white (2010), lead IV oxides were identified but no specific crystalline forms (i.e.
plattnerite or scrutinyite) due to the poor resolution of the Raman spectrum16. They were able to identify plattnerite in the dark red layer of the sculptures in Queluz for the final report of an
15 Bernard, M.-‐C. et al , 2010, pg. 341. 16 Ibid, pg. 343.
Toro, UvA 2018 21
extensive conservation project of the Palace’s gardens by the World Monuments Fund in 2012.17 They also tested the formation of the lead dioxide layer on coupons pretreated with acetic and sulphuric acid, and concluded that lead carbonates were necessary for the process to occur. According to their observations “It can then be proposed that the red colour of the white patina is due to oxidation of hydrocerussite by a strong oxidant [...].”18
While visually the dark red layers in both L’Air and the Cheere sculptures are strikingly similar, the analyses also show several differences between the two cases. First, there was no shannonite detected in the XRD results from L’Air acquired by the RCE in 201619. Second, according to the same results, the lead carbonate cerussite was more predominant in L’Air than hydrocerussite. Even though this does not rule out Bernard et al’s theory about the oxidation of hydrocerussite to form lead dioxide, the relationship between both lead carbonates and lead dioxide should be further investigated. For example, this difference could be related to the pretreatment of coupons in Bernard et al’s experiment, which based on current knowledge do not correspond with the treatment history or environmental conditions of L’Air.
Also, Bernard et al’s description of a ‘white patina’ does not correspond with past descriptions or the photographic history of L’Air (Table 3.1). Instead of a ‘white patina’, the red stains seem to have appeared amongst a dark gray oxide layer in the case of Maillol’s sculpture, questioning the carbonate oxidation theory.
Last but not least, although an oxidative environment must be present in both cases, the Palace of Queluz and the Kröller Müller Museum have very different environmental conditions. The Palace, once a summer retreat, is now within the metropolitan area of Lisbon and its gardens are surrounded by a major highway20. The Kröller Müller Museum is located within the Hoge Veluwe National Park, its closest town being Otterlo, a small village21. Only by these observations it can be assumed that the atmospheric conditions vary between both places. Therefore, it can also be said that while environmental conditions are an important factor in lead dioxide formation, the dark red stains are not directly related to a single environmental scenario. In the end, the presence of plattnerite in Cheere’s sculpture group and Maillol’s L’Air continues to draw a strong link between the two cases.
17 Os Jardins do Palácio Nacional de Queluz: intervenção de conservação, ed. A. Elena Charola and José Delgado Rodrigues. Associação World Monuments Fund Portugal e World Monuments Fund, 2012. pg. 97.
18 Bernard, M.-‐C. et al, 2010, pg. 344. 19 Joosten, I. et al, 2016.
20 Google Maps: https://www.google.nl/maps/place/Queluz+National+Palace/@38.7531326,-‐
9.2669359,15z/data=!4m5!3m4!1s0x0:0x38b1d71128e15bb!8m2!3d38.7507229!4d-‐9.2590824 visited 10-‐06-‐
2018.
21 Google Maps:
https://www.google.nl/maps/place/Otterlo/@52.0736776,5.7148315,12z/data=!3m1!4b1!4m5!3m4!1s0x47c7b18
Toro, UvA 2018 22
Conservator Rupert Harris, who was also involved in the Queluz project in 2012, has written about his observations in Portugal and regarding other sculptures and historic buildings in the United Kingdom. He is of the opinion that the phenomenon of the red staining comes from the disruption of the natural chemical process of lead patina formation by “...some form of abnormal chemical reaction”. 22 Harris points out that:
As the areas of staining do not exhibit any of the characteristics of active corrosion such as pitting, powdering of the surface or loss of metal, it is assumed that the chemical process being witnessed is one of the conversion of existing white to grey coloured lead patina compounds to red-‐brown lead dioxide.23
This partially agrees with Bernard et al’s theory that the red staining develops from the oxidation of hydrocerussite. However, the presence of white spots within the red stains both in the Queluz case and L’Air (Figs. 5.1 & 5.2), contrasts with Harris’s observations since they are powdery and may possibly be signs of active corrosion24.
He also mentions that:
“Cleaning tests using both dilute acetic acid and hydrochloric acid have been shown to remove the staining and reveal that no discernable damage to the underlying metal surface has occurred,
22 Harris, R. Summary Report on the Incidence and Possible Causes of the Occurrence of Red-‐Brown Staining on
Outdoor Lead Sculpture and Historic Buildings, 2017. pg. 1-‐9. Unpublished manuscript.
23 Rupert Harris: https://rupertharris.com/pages/additional-‐information-‐lead-‐dioxide-‐formation-‐on-‐lead visited 10-‐06-‐2018.
24 Active corrosion: fresh or new corrosion; often seen as spalling, cracking or flaking. (Selwyn, L., 2004, pg.195). White spots White spots
Fig 5.1: Detail of The Spring, Palace of
Toro, UvA 2018 23
further supporting the theory that it is a conversion process that is producing the surface colour change and not corrosion in the true sense25. “26
However, the sculptures at Queluz received a passivation treatment with oxalic acid (C2H2O4) in
201227, and the red stains fully reappeared by 2017 (Fig. 5.3). This not only highlights a downside of the treatment, but also the possibility that the formation of the dark red stains is the product of a cyclic corrosion process instead.
Finally, another point in common between the sculptures at Queluz and the Maillol at the Kröller Müller Museum is the role of light. Conservation scientist Virginia Costa relates the formation of the dark red layer not only to humidity levels, rain or air pollution, but also to sunlight exposure.28 While the role of light as a possible catalyzer of the reactions will not be investigated in this research, it deserves further attention in the future.29
25 It can be inferred from this comment that Harris alludes to corrosion ‘in the true sense’ as a mechanism by which metal is degraded into a different chemical form (in itself a conversion). In this case, he refers to the color change being the product of the conversion of an already existing compound/corrosion product into lead dioxide, instead of from metal into lead dioxide.
26 Rupert Harris: https://rupertharris.com/pages/additional-‐information-‐lead-‐dioxide-‐formation-‐on-‐lead visited 10-‐06-‐2018.
27 Charola, E.; Delgado, J-‐D. (Eds) 2012. pg. 107-‐8. 28 Ibid, pg. 97.
29 According to the electrochemical series in the CRC Handbook of Chemistry and Physics (2000), the reaction from cerussite to lead dioxide is non-‐spontaneous.
Toro, UvA 2018 24
5.2. Formation of lead dioxide
Extensive research has been done regarding the behaviour of lead in water supply systems. Some of this research includes the formation of lead dioxides. Even though in most experiments a strong oxidant was introduced in order to produce lead dioxide (usually chlorine, to this date not significantly present in L’Air’s environment), the results provide good insights into the possible corrosion processes.
In the article Formation of Lead(IV) Oxides from Lead(II) Compounds, (2010) Wang et al consider how the formation of PbO2 is affected by water chemistry. Starting from the notion that “Previous research showed that lead(II) carbonates formed as intermediates in the formation of PbO2. However, it is not known whether carbonates are required for PbO2 formation”30, they investigated possible corrosion mechanisms. In order to do so, they evaluated different pathways by looking at four different factors: the effect of dissolved inorganic carbon (DIC)31, the effect of pH on PbO2 formation, the effect of free chlorine concentration and, the effect of precursor identity on PbO2 formation (referring to whether lead carbonates are intermediate solids in the process or not)32.
They concluded that lead dioxides plattnerite and scrutinyite form from massicot (β-‐PbO) and dissolved lead(II) chloride (PbCl2) in the presence of free chlorine, but no DIC, at pH = 7.5 and pH = 10. “These results demonstrate that PbO2 formation does not require lead(II) carbonates as precursors or intermediate phases”.33 According to this, Bernard et al’s theory that lead dioxide comes from the oxidation of hydrocerussite is challenged, and it becomes important to consider whether lead dioxide could also form at the surface of lead sculptures directly from lead oxide.
If the presence of carbonates is not necessarily the main precursor of the reaction, then, what could be causing the right conditions for lead dioxide to form? Going back to Wang et al’s influential factors, both the concentration of dissolved inorganic carbon (DIC) and the concentration of free chlorine are related to the pH of a solution (to be recreated around L’Air as environment + water), thus altering the effect of pH on PbO2 formation. In Formation of Pb(IV)
oxides in chlorinated water (2005), Lytle and Schock observed that “The aging of lead to PbO2 was indicated by a change in solid color (white to dark red or brownish red, depending on pH) and lead solubility[...]”.34
30 Wang, Y.; Xie, Y. et al. Formation of Lead(IV) Oxides from Lead(II) Compounds. Environmental Science & Technology 2010 44 (23), pg. 8950.
31 Dissolved Inorganic Carbon (DIC): sum of inorganic carbon species in a solution, these include carbon dioxide, carbonic acid, and carbonates. (https://en.wikipedia.org/wiki/Total_inorganic_carbon visited 10-‐06-‐2018). 32 Wang, Y. et al, 2010.
33 Ibid, pg. 8951.
34 Lytle, D.; Schock, M. Formation of Pb(IV) oxides in chlorinated water. Journal American Water Works Association, Vol. 97, No. 11, 2005. pg. 112.
Toro, UvA 2018 25
In their experiments, they tested for the formation of lead dioxide as a precipitate in chlorinated solutions, including DIC. They found that at pH = 6.5-‐8 the sequence of the corrosion products that formed was: from hydrocerussite to cerussite, from cerussite to plattnerite, and from plattnerite to scrutinyite.35 They also noted that “Water pH dictated mineralogical presence. High pH favored hydrocerussite and scrutinyite; low pH favored cerussite and plattnerite ”.36 This suggests that even the type of lead dioxide polymorph being formed is influenced by changes in pH.
According to previous XRD results, the corrosion layer of L’Air has a higher concentration of plattnerite and cerussite, pointing to a lower pH when compared with Lytle and Schock’s conclusions. However, scrutinyite and hydrocerrusite are still present.37 It is possible that changes in the environment, such as the varying composition of the rainwater or the varying concentrations of pollutants in the air, are responsible for changes in pH and therefore for the formation of all four corrosion products. It is also possible that changes in the water chemistry of the rainwater from being in contact with an already corroded surface is creating the fluctuations in pH, further supporting the theory that a cyclic corrosion mechanism is present.
While the concentration of dissolved inorganic carbon (DIC) will not be measured at the surface of the sculpture, its presence can be accounted for in the lead carbonates present and the simple and ongoing reaction of carbon dioxide and water resulting in carbonic acid (likely to be involved in the oxidation process). And, while the nature of the corrosive environment is still under discussion since chloride was not significantly present in previous analyses, it brings up the question of whether oxygen, sunlight and water are oxidative enough for lead dioxide to form. Furthermore, optical analysis, pH measurements, and SEM-‐EDX analysis of the surface of L’Air will provide more information about the relationship between the lead dioxides and the lead carbonates, as well as the material’s relationship to its surroundings, in a way that is comparable to previous results.
5.3. Is it a patina?
The debate regarding the definition of a patina versus a corrosion layer has occupied conservators for several decades. In the case of the dark red stains appearing on lead sculpture, they have often been described as a patina. However, the use of the word ‘patina’ will be revised in this thesis because of its potentially controversial meaning.
The Merriam-‐Webster dictionary defines the word patina as:
35 Ibid.
36 Ibid, pg. 102.
Toro, UvA 2018 26
a. A usually green film formed naturally on copper and bronze by long exposure or artificially (as by acids) and often valued aesthetically for its color.
b. A surface appearance of something grown beautiful especially with age or use.
c. An appearance or aura that is derived from association, habit, or established character. d. A superficial covering or exterior.38
This extensive interpretation includes all the aspects that make the notion of patina confusing to begin with. First, it mentions that it can be developed naturally or artificially. Second, it attributes an aesthetic value to an object grown old or used by employing the word ‘beautiful’. Third, it introduces the notions of time and perception in a kind of ‘aura’. And at last, it simplifies the previous meanings into a ‘covering or exterior’.
On the other hand, the Merriam-‐Webster dictionary defines the word corrosion as ‘the action of corroding’39, which in turn is:
a. to eat away by degrees as if by gnawing; especially : to wear away gradually usually by chemical action.
b. to weaken or destroy gradually.40
The relationship of corrosion with deterioration has often been mixed up with the electrochemical notions of active41 and passive42 state of a metal. Further adding to the confusion by implying that a patina provides a protective layer to a metal surface while a corrosion layer does not, which is not necessarily true.
In A Review of the History and Practice of Patination, Phoebe Dent Weil explains the historical evolution of the concept of patina. She observes that the first printed use of the word was in 1681 in relation to paintings, as a dark ‘tone’ developed by the passing of time which could often be flattering to the objects.43 This first mention already combines a physical change (oxidation, interaction with the environment), with an additional aesthetic value.
The word patina in order to describe green corrosion products in bronze objects seems to develop from an increasing interest in archaeology during the eighteenth century44, and it is linked to the question of discovering what the original finish of an object was. Later on, references to the word patina and the finishing of bronzes appear during the nineteenth century, when “...artificial patination of bronzes by chemical means, with or without heat, was generally and
38 Merriam-‐Webster: https://www.merriam-‐webster.com/dictionary/patina visied 15-‐06-‐2018. 39 Merriam-‐Webster: https://www.merriam-‐webster.com/dictionary/corrosion visited 15-‐06-‐2018. 40 Merriam-‐Webster: https://www.merriam-‐webster.com/dictionary/corroding visited 15-‐06-‐2018.
41 “Condition in which a metal reacts with its environment (i.e. freely corrodes) because it is thermodynamically unstable and the corrosion products are soluble.” (Selwyn, 2004, pg.195).
42 “Condition in which a thermodynamically unstable metal has a low corrosion rate because the metal surface has reacted with the environment to form a protective film of corrosion products.” (Selwyn, 2004, pg. 202).
43 Weil Dent, P. “A Review of the History and Practice of Patination.” In Historical and Philosophical Issues in the
Conservation of Cultural Heritage, edited by Stanley Price, N. et al. The J. Paul Getty Trust, 1996. pg.398-‐99.
Toro, UvA 2018 27
widely practiced.”45 On the other hand, the additional value of ‘antiqueness’ a bronze object can acquire due to its color as a testament of age, is a concept that evolved during the Renaissance period. 46
Weil summarizes the topic by explaining that:
The problem of treating metallic objects is complicated by the fact that patina, i.e., corrosion is formed at the expense of the substance of the object itself, even occasionally to the point of complete mineralization, and once severe alteration or corrosion has occurred it is impossible to determine original coloration of finish from physical evidence. (Weil, 1996, pg.395).
“Once again, several conflicting and confusing aspects are implied in the definitions. First, Weil pairs together the words ‘patina’ and ‘corrosion’ almost as synonyms. Second, she brings forth the idea that the patina (or corrosion) may completely penetrate/alter the object. And third, Weil relates the concept back to the historical concern over what the original surface characteristics of an object were.”
Today, thanks in part to developments in conservation science and the conservation profession as a whole, there is a better understanding of the complexity of visual, chemical and morphological changes that can happen at the surface of a metal object. In Metals and corrosion
[...] (2004), Lyndsie Selwyn defines corrosion as “Electrochemical reaction between a metal and
its environment that causes the metal to deteriorate; the electrochemical process involves the transfer of electrons from the metal to another species during simultaneous oxidation and reduction reactions.47 Selwyn then defines patina as “Corrosion products on the surface of a metal; patina may occur naturally as a result of long-‐term exposure to weather, pollution, etc., or may be artificially induced through the application of various chemicals.”48 According to this, it seems that the words ‘corrosion layer’ and ‘patina’ can be used almost interchangeably. Nevertheless, it is important to note that even though Selwyn defines corrosion as a deterioration process, she does not specify whether the corrosion products that form a patina can be damaging to the rest of the material or not. This possibly separates the idea of patina from the discussion about the ‘stability’ of the corrosion layer (whether it is active or passive).
However, the concept of patina still carries with it historical ideas about physical properties and aesthetic value, and even experts in the field struggle with the use of the word. As written by David A. Scott:
“Strictly speaking, ‘patina’ and ‘corrosion’ are different words for the same surface alteration. [...] Corrosion may be termed the process of a chemical attack of an environment on a material, while patina could be defined as the accumulation of corrosion products and other materials
45 Ibid.
46 Ibid, pg. 403.
47 Selwyn, L. Metals and corrosion: a handbook for the conservation professional. Canadian Conservation Institute, 2004. pg. 197.
Toro, UvA 2018 28
from the environment. One person’s patina, however, may be another person’s corrosion, so a certain ambiguity is inherent in both words. (Scott, 2002, pg.10).”
Given this overview, it can be said that the dark red stains in L’Air are technically a part of its patina. However, in order to avoid attaching additional aesthetic value/character or assumptions of chemical ‘stability’ to the formation of lead dioxide on L’Air, the dark red stains will not be referred to in this thesis as a patina, but as a corrosion layer instead.
6. Methodology
6.1. Optical analysis
6.1.1. Visual analysis
Visual analysis was conducted first by the naked eye. It served to understand the patterns in which the stains are present, the occurrence of red areas (lead dioxides) with white spots (lead carbonates) and red areas without, the areas where rainwater runs and the areas in which it sits, and the places that receive the most sunlight.
In order to observe and confirm the water dripping patterns throughout the sculpture, drops of water were released and recorded as they oozed, dripped, and evaporated from the object (Fig. 6.1). This was done by using a plastic pipette, deionized water, and a camera.
With the aim of obtaining a more detailed, magnified image of the corrosion products, a Dino-‐Lite digital microscope was used on site.
Water drop and matching stain pattern