Shedding light on diafanoramas. A research into the causes of paint delamination on a set of three diafanorama plates in the collection of the Rijksmuseum

(1)

Shedding light on diafanoramas

A research into the causes

of paint delamination on a

set of three diafanorama

plates in the collection of

the Rijksmuseum

Master thesis

Conservation and Restoration of Cultural Heritage, Paintings University of Amsterdam

20-06-2017

Supervisor: dr. Maartje Stols-Witlox External advisor: Margot van Schinkel Student: Noa Kollaard

(2)

Noa Kollaard –UvA – 2017

2

Shedding light on diafanoramas

A research into the causes of paint delamination on a set of three diafanorama

plates in the collection of the Rijksmuseum

Master thesis

Conservation and Restoration programme, Paintings Specialisation June 2017

Word count: 17991

Noa Kollaard

Student number: 10351388

Supervisor: dr. Maartje Stols-Witlox

External advisor: Margot van Schinkel

Photos front page:

Anonymous, Set of three diafanorama plates of a battle on a village square, wood, glass and paint, 36 cm x 43 cmx 0.8 cm. Amsterdam: Rijksmuseum, c. 1750-1830. Photos: Rijksmuseum.

(3)

3

Table of contents

Summary 5

Samenvatting 6

Preface and Acknowledgements 7

Chapter 1 Introduction 8

1.1 Research question and scope of research 9

1.2 Current scientific knowledge 10

1.3 Subquestions, structure of thesis and methodology 12

1.4 Relevance of research 14

Chapter 2 Diafanoramas: history, use and manufacture 15

2.1 Historical context 15

2.2 Use of diafanoramas 15

2.3 Manufacture of diafanorama plates 18

2.3.1 Glass and preparation of the glass before painting 18

2.3.2 Painting technique and materials 20

2.4 Conclusion 22

Chapter 3 Three diafanorama plates of the Rijksmuseum: technique 24 and materials

3.1 Object description and history 24

3.2 Examination of materials and techniques 26

3.2.1 Methods 26

3.2.1.1 XRF 26

3.2.1.2 Analysis of paint samples 26

3.2.1.3 Dendrochronology 27

3.2.2 Results 27

3.2.2.1 Glass and preparation of the glass 27

3.2.2.2 Painting technique, paint layer build-up and painting materials 35

3.2.2.3 Frames 43

3.3 Conclusion 44

Chapter 4 Condition of the three diafanorama plates of the

Rijksmuseum and condition of diafanoramas in other collections 45

4.1 Front and middle plate 45

4.2 Back plate 47

4.3 Other diafanorama plates 50

4.4 Conclusion 53

Chapter 5 Discussion 55

5.1 Delaminating trees on the front and middle plate 55

(4)

4

5.3 Network of cracks on the back plate 58

Chapter 6 Conclusion 60

6.1 General conclusions 60

6.2 Further research 61

6.3 Advice for restoration and preventive conservation 61

Bibliography 63

Appendix

I Overview photos 69

II Additional information on artists 78

III Sample forms 80

IV SEM-EDX spectra and BSE images 87

V Reports transmission FTIR 97

VI Report THM-Py-GC/MS 100

VII XRF results 103

VIII Report dendrochronological analysis 109

IX Condition mappings 110

X Overview of condition of diafanoramas in and outside the

(5)

5

Summary

Shedding light on diafanoramas

A research into the causes of paint delamination on a set of three diafanorama plates in the collection of the Rijksmuseum

Noa Kollaard, University of Amsterdam, June 2017

This master thesis focuses on a set of three diafanorama plates (1750-1830) in the collection of the Rijksmuseum. Diafanorama plates are sets of cold-painted glass sheets, each of which depicts a part of the whole image. The three plates suffer from paint delamination. The aim of the research is to gain more insight into the possible causes of the delamination, which is needed for designing suitable conservation treatment and measures. Literature research was carried out to the original use and manufacture of diafanoramas and to the history of the three plates in particular. This provided insight into the (environmental) conditions the plates have been exposed to. Important findings are: heat-producing candles or oil lamps were placed behind the plates in order to view them and the plates have been subjected to uncontrolled climatic conditions in the past. Supplementary, technical research (visual examination, XRF and analysis of paint samples by SEM-EDX, transmission FTIR and THM-Py-GC/MS) was carried out to determine how and with which materials the plates have been painted and to establish where and how the delamination has manifested itself.

From the research it could be concluded that the delamination is manifested in primarily three ways, which most likely have different causes. Although it was not possible to designate one (or more) cause with certainty, hypotheses of possible causes could be formulated based on the research results. Firstly, the paint of the trees on the front and the middle plate has cracked into small, delaminating islands. This condition may be the result of the fact that the trees have been painted directly on the glass and not on top of a preparatory layer that functioned as a ‘bridging’ layer between the glass and the paint. Secondly, the blue paint of the sky on the back plate is brittle and crumbles from the support. The discovery of an irregular, whitish layer underneath the blue paint and the indicated presence of lead soaps in the paint led to the hypothesis that the whitish layer may be an efflorescence of crystalline lead soaps. This may have weakened the adhesion of the paint to the glass and made the paint more brittle. Thirdly, a large part of the painting on back plate shows networks of cracked wrinkles and cracks with delaminating edges along which the image does not align anymore. It is suggested that possible elevated temperatures caused by the candles or oil lamps may have resulted in the Tg of the paint having been reached. This may have allowed for movement and the formation of wrinkles. Finally, fluctuating environmental conditions might also have contributed to the cracked condition of the paint.

(6)

6

Samenvatting

De scriptie richt zich op een drietal diafanorama platen (1750-1830) in de collectie van het Rijksmuseum. Diafanorama platen zijn sets van koud-beschilderde glasplaten, waarvan elke plaat een gedeelte van een afbeelding weergeeft. De drie platen vertonen verf delaminatie. Het doel van het onderzoek is het verkrijgen van meer inzicht in de mogelijke oorzaken van het delamineren van de verf, wat nodig is voor het conserveren en restaureren van de objecten.

Onderzocht is hoe diafanorama platen werden gemaakt en gebruikt. Daarnaast is onderzoek gedaan naar de geschiedenis van de drie platen in kwestie, om inzicht te krijgen in de omgevingsfactoren waaraan de platen zijn blootgesteld. Belangrijke bevindingen zijn: warmte-producerende kaarsen of olielampen werden geplaatst achter de platen om deze te bekijken en de drie platen in kwestie zijn in het verleden blootgesteld aan ongereguleerde klimaatomstandigheden. Verder is technisch onderzoek uitgevoerd (visueel onderzoek, XRF en analyses van verfmonsters met SEM-EDX, transmissie FTIR en THM-Py-GC/MS) om vast te stellen hoe en met welke materialen de platen zijn geschilderd en om vast te stellen hoe en op welke manier de delaminatie zich heeft gemanifesteerd.

Uit het onderzoek kan geconcludeerd worden dat de delaminatie zich heeft gemanifesteerd op hoofdzakelijk drie verschillende manieren, die waarschijnlijk verschillende oorzaken hebben. Op basis van de onderzoeksresultaten was het niet mogelijk om één of meerdere oorzaken met zekerheid vast te stellen. Echter, hypotheses over mogelijke oorzaken konden wel worden gevormd. Ten eerste, de verf van de bomen op de voorste en middelste plaat is gebroken in kleine, delaminerende verfeilandjes. Deze conditie zou het resultaat kunnen zijn van het feit dat de bomen direct op het glas zijn geschilderd en niet op een voorbereidingslaag die fungeerde als een overbruggende laag tussen het glas en de verf. Ten tweede, de blauwe verf van de lucht op de achterste plaat is bros en brokkelt van de glasplaat. De vondst van een onregelmatige, witte laag onder de blauwe verf en een aangetoonde aanwezigheid van loodzepen in de verf, heeft geleid tot de hypothese dat de witte laag een kristallisatie van loodzepen zou kunnen zijn. Dit zou de adhesie van de verf op het glas kunnen hebben verzwakt en de verf bros hebben gemaakt. Ten derde, een groot gedeelte van de verflagen op de achterste plaat vertoont netwerken van gescheurde rimpels en scheuren met delaminerende randen waarlangs de afbeelding niet goed meer aansluit. Het is mogelijk dat eventueel verhoogde temperaturen, veroorzaakt door de kaarsen of olielampen, hebben geresulteerd in het bereiken van de Tg van de verf, waardoor de verf misschien heeft kunnen bewegen en rimpelen. Ten slotte, fluctuaties in klimaatomstandigheden kunnen hebben bijgedragen aan de gescheurde conditie van de verf.

(7)

7

Preface and Acknowledgements

This thesis is written in the context of the master Conservation and Restoration of Cultural Heritage, specialisation Paintings, at the University of Amsterdam. The research is carried out in collaboration with the Rijksmuseum Amsterdam (RMA) and the Cultural Heritage Agency of the Netherlands (RCE).

Many people have helped me throughout this project and I would like to thank all of them. Many thanks to Margot van Schinkel (RMA), Eveline Sint Nicolaas (RMA) and Lisette Vos (RMA), for making it possible to investigate the diafanorama plates, for providing me with documentation of the objects and for helping with the examination of the objects in the museum depot.

I want to thank my thesis supervisor Maartje Stols-Witlox (UvA) for her wonderful guidance and for carrying out SEM-EDX analysis. Furthermore, I want to thank other UvA teachers Rene Peschar, Mandy Slager, Nienke Besijn, Maarten van Bommel and Ella Hendriks for their useful advise.

This research could not have been done without the help of RCE researchers Suzan Catucci - de Groot and Henk van Keulen. I want to thank them both for carrying out the necessary technical analysis and trying different techniques to investigate the incredibly small samples! Many thanks as well to Guus Verhaar (UvA) for his explanation of glass degradation and for carrying out XRF analysis.

I also want to express my gratitude to Margreet Wagenaar-Fischer, who was kind enough to show me her own diafanorama collection and demonstrate how they were used. Furthermore, I thank the staff of the Museum Rotterdam and the Theatre collection of the University of Amsterdam for letting me examine the diafanoramas in their collection.

Last but not least, many thanks to my amazing fellow students Alma Jongstra, Madeleine Vaudremer, Marjolein Hupkes, Paulien Coopmans, Rosa Boute and Sepha Wouda for their great support!

(8)

8

Chapter 1 Introduction

This thesis describes a research carried out on a set of three diafanorama plates in the collection of the Rijksmuseum (fig. 1.1). Diafanorama plates are objects that were used for visual entertainment in the eighteenth and nineteenth century in Holland. The plates are cold painted glass sheets that usually come in sets of three. Each plate depicts the back-, the middle-, or the foreground of an image. Placed behind each other, the plates form the total image with a remarkable illusion of depth. The image was supposed to be viewed in a burning mirror (a concave mirror), which stood in front of the plates, while a light source was placed behind the plates (fig. 1.2). The mirror enlarged the image and enhanced the spatial effect of the three plates (Kuijper 1990, 39-44). Not surprisingly, the subjects that are most frequently depicted are subjects in which perspective and depth play an important role, such as landscapes, town views and stage plays (Kuijper 1990, 67).

The Rijksmuseum has in total twenty-one diafanorama plates, which were all acquired in 1887. A number of these plates is on permanent display in the museum. Some of the remaining plates, including the three plates on which this research focuses, show severe paint delamination. In multiple areas the adhesion between the glass support and the paint is poor, which in some cases has led to paint loss. Paint delamination is not an uncommon problem for diafanorama plates, and of cold painted glass in general. Glass is not an ideal support for painting as it is smooth and non-porous. Consequently, the binding medium of the paint is not absorbed by the glass and this limits the mechanical adhesion between the paint and the support (Baumer et al. 2012, 9; Bretz 1999, 183). Yet, quite a number of the diafanorama plates in the collection of the Rijksmuseum is still in a good condition, showing that paint delamination does not necessarily occur. The subject of this thesis concerns a diagnostic research into the problem of paint delamination in diafanorama plates, with the aim to gain more insight into how and why the delamination did occur at the three plates in question. This insight is important for the restoration of the plates and for preserving them in order that future generations can admire them as well.

Fig. 1.1 Anonymous, Set of three diafanorama plates of a battle on a village square, wood, glass and paint, 36 cm x 43 cmx 0.8 cm, c. 1750-1830. Amsterdam: Rijksmuseum. Inventory numbers from left to right: BK-NM-8462-J, BK-NM-8462-K and BK-NM-8462-L. Photos: Rijksmuseum.

(9)

9

1.1 Research question and scope of research

The Rijksmuseum would like to include the three plates in the permanent exhibition because their depiction, a seventeenth-century battle on a village square, is quite unique (Kuijper 1990, 73). However, the plates are too fragile to be exhibited and the museum faces the question how to stabilize the objects. Treating delaminated paint on glass can be challenging, as it falls between the field of glass conservation and paintings conservation. A better understanding of the problem of delaminating paint on the glass plates and the possible cause(s) will help design a suitable conservation treatment and measures.1_{For this}

reason, the following research question is formulated:

What could be the cause(s) for the paint delamination on the three diafanorama plates in the collection of the Rijksmuseum?

In literature on the conservation of cold painted glass various factors are mentioned that can promote the delamination of paint layers (see paragraph 1.2). Some of the factors concern the glass support in particular. Degradation of the (unstable) glass is regularly mentioned as a possible cause for paint delamination (Blewett, 2005, 5; Blewett 2012a, 112; Bretz 1999, 183, 211; Caldarado 1997, 157; Jablonski 2000, 6; Neelmeijer and Mäder 2004, 131). It has been established by Margot van Schinkel, glass and ceramic conservator of the Rijksmuseum, that the glass of the three plates in question shows no visual signs of degradation. 2 Therefore, it is not expected that a degradation of the glass has played a major role in the paint delamination. For this reason, and in order to limit the scope of the research, the primary focus of this thesis will lie on the investigation of causes related to the paint materials, rather than on the glass support.

1_{Schinkel, Margot van. Oral communication, 22 November 2016.}

2_{Schinkel, Margot van. Oral communication, 26 January 2017.}

3_{The magic lantern is composed of a light source, a translucent (painted glass) slide and a positive lens in a}

wooden or metal case (Wagenaar, Duller and Wagenaar-Fischer 2014, 7). Magic lantern slides are

2_{Schinkel, Margot van. Oral communication, 26 January 2017.}

Fig. 1.2 Original set-up of diafanorama plates with a burning mirror and a light source (a pair of oil lamps) behind. Zeist: Christiaan Huygens Theater. Photo: Wagenaar, Duller and Wagenaar-Fischer 2014, 94.

(10)

10

1.2 Current scientific knowledge

Not many publications can be found in which the original function and construction of diafanoramas is discussed in depth. The most detailed study on diafanoramas is a PhD research by H. Ch. Kuijper, titled Het diafanorama. Ofte fraaije stellen van fyne geschilderde glazen voor de brandspiegel. Kuijper discusses the historical social context of diafanoramas, explains how they were used and describes the iconography of diafanoramas from an art historical perspective (Kuijper 1990). Her research is mainly based on visual examination of diafanoramas in collections of different museums, on archival research into auction catalogues in which diafanoramas are offered for sale and on contemporary handbooks on optical instruments. To a lesser extent, Kuijper discusses the painting techniques of diafanoramas, based on visual examination. Only one (unspecified) technical analysis of the paint binding medium of a diafanorama is mentioned by the author. The analysis was performed in 1988 at the Centraal Laboratorium voor Onderzoek van Kunst en Wetenschap and showed that both egg and oil were used as binding medium (Kuijper 1990, 54).

Academic, technical research on how diafanorama plates were made, what materials were used, and on their conservation issues is scarce. Therefore, literature with information on technical and conservational aspects of other types of cold painted glass objects, such as magic lantern slides and reverse-glass paintings, has to be consulted.3

A relevant source on magic lantern slides is Dutch Perspectives, written by Willem Albert Wagenaar, Annet Duller and Margreet Wagenaar-Fischer (all collectors and experts in the field of historical optical instruments). The book describes (cold) painting techniques for glass lantern slides. However, the authors do not specify their source(s) of information and how they investigated the slides in order to determine the painting technique and materials (Wagenaar, Duller and Wagenaar-Fischer 2014, 38-40). One article was found that describes a research, carried out in 2008 at the University of Torino, into the binding media of two nineteenth-century magic lantern slides Different materials were identified by means of mid-infrared fibre-optic spectroscopy. Mainly linseed oil was found, sometimes in combination with natural resins. A natural gum was also found (possibly Arabic gum), which was either used as a binder or as a transparent preparatory layer (Ploeger, Scalarone and Chiantore 2010, 41).

Literature dealing with reverse-glass painting is more extensive in comparison. Paint delamination is a common conservation issue for this type of objects and various articles can be found that focus on conservation treatment cases. In these texts reverse-glass painting techniques are often briefly discussed and various causes for paint delamination are mentioned. These causes can be divided into two categories. First of all, there are causes related to the (painting) technique and materials. A change in the surface

3_{The magic lantern is composed of a light source, a translucent (painted glass) slide and a positive lens in a}

wooden or metal case (Wagenaar, Duller and Wagenaar-Fischer 2014, 7). Magic lantern slides are comparable to diafanorama in terms of materials and techniques and in terms of use: they had to be lit from behind and with the aid of optical instruments the painted image on the glass was projected. It should be kept in mind that this is different for reverse-glass paintings; these were used differently and viewed differently and consequently the painting techniques may be different.

(11)

11

topography of the glass as a result of cold flow, degradation of unstable glass (whereby alkali ions leach out from the glass structure), insufficient cleaning of the glass before painting, the absence of a preparatory layer, a pigment-binder ratio that is too high and the application of thick paint layers are all listed as aspects that can contribute to paint delamination (Agnini 1999, 259-260; Blewett 2005, 5; Blewett 2012a, 112; Bretz 1999, 183-211; Caldarado 1997, 157; Coppieters-Mols 1996, 34; Jablonski 2000, 6; Neelmeijer and Mäder 2004, 131; Schott 1999, 247). Secondly, there are factors that are related to external conditions. Photochemical oxidation of the binding medium and also high temperatures can lead to powdering and delamination of the paint. Fluctuations in temperature and relative humidity can cause contraction and expansion of paint layers. Upon a rigid support like glass this can induce mechanical stress within the paint layers with delamination as a result. Finally, moisture can decrease the adhesion between the preparatory layer and glass or between the preparatory layer and paint layer if one of these layers is hydrophobic and the other is hydrophilic. (Agnini 1999, 259-260; Blewett 2005, 5; Blewett 2012a, 112; Bretz 1999, 183, 211; Caldarado 157; Coppieters-Mols 1996, 34; David 2009, 223; Davison 2003, 339-340; Jablonski 2000, 6; Schott 1999, 247; Rydlová et al. 2015, 186). In most of these articles the exact (chemical and physical) mechanisms behind paint delamination are not much discussed. In the publications of conservators Morwenna Blewett and Simone Bretz, however, degradation phenomena of reverse-glass painting are more elaborately addressed. Many of the above mentioned causes for paint delamination can be found in their publications. However, they relate more directly the findings of their research to the techniques and the materials used to observed paint delamination. In her research into reverse-glass paintings by William Nicholson (1872-1949), Blewett could relate observed paint delamination to the degradation of a translucent preparatory layer. Furthermore, by means of gas chromatography-mass spectrometry, she found that different colours may contain a different binding medium. Brown and yellowish colours contained oils, while blue areas were painted with egg yolk or casein. The blue paint had problems with mould, had become powdery and showed the most delamination (Blewett 2004, 17). Bretz presents in a collaborative article an in depth analysis of a sixteenth-century reverse-glass painting with severe paint delamination. Gas chromatography-mass spectrometry of the paint showed a high amount of oxidation products of the resinous binding medium, which could be related to the brittle and delaminated state of the paint. From the type of oxidation products found it could be determined that the resin had been preheated before use and was severely oxidised in time (Bretz et al. 2008, 219-220).

Relevant research in the same field has also been carried out during a recent collaborative project of the Doerner Institut, the Museum Schnütgen and the Bundesanstalt für Materialforschung und –prüfung, in which Simone Bretz has participated as well. In this project, thorough studies were carried out on reverse-glass painting techniques and conservation issues. Technical examination of reverse-glass paintings, dating from the Middle Ages to the nineteenth-century, was performed in combination with investigation of contemporary art technological sources on painting on glass (Bretz 2003; Bretz et al., 2016; Oltrogge 2016). Analyses of the glass supports and the composition of the paint have

(12)

12

been carried out (Baumer et al. 2012; Baumer and Dietemann 2016; Baumer, Dietemann and Koller 2009; Hahn 2003; Koller and Baumer 2003; Müller 2003). An important part of the research was the identification of the binding media used. According to the researchers not all binders were equally suitable for painting on glass and choosing a less suitable binder may have consequences for potential degradation processes. The binder needs to have good wetting properties and have similar polar characteristics as the glass (Baumer et al. 2012, 10, Baumer and Dietemann 2016, 74-76, Koller and Baumer 2003, 114) Baumer and Dietemann explain that the polarity of binders increases with the amount of amount of oxygen-containing functional groups in their molecular structure. Their theory is that commonly used binders such as boiled oil or oil/resin mixtures are therefore more polar (in comparison to for example unboiled oil). Consequently, these binders adhere better to the glass as the functional groups can have strong interactions with the polar silica groups of the glass (Baumer and Dietemann 2016, 76). The researchers do not discuss possible influences of pigments or other additives on the polarity of paint. Koller and Baumer mention another chemical characteristic of binders (in this case oil) that can have consequences for potential degradation processes. Oil binders can interact with certain pigments to form metal soaps. This can eventually lead to a brittle and delaminating paint layer (Koller and Baumer 2003, 114). The researchers do not explain this further and do not provide examples of cases where this problem has occurred. In the consulted literature, no other remarks about metal soap formation as a cause for delaminating paint on glass was found.4

Finally, a research into the adhesion properties of oil paint on glass, carried at the University of Lincoln and the University of Chester, should be mentioned (Bayle et al. 2015). According to the researchers, the wetting of the paint – indicated by the contact angle between the paint and glass – is an important parameter for the adhesion of the paint to the glass. The study showed that the wetting of the paint depends on the type of pigments. This was shown by means of experiments in which the contact angles between a glass substrate and drops of nine different linseed oil paints were measured (Bayle et al. 2015, 297). Furthermore, the same test was performed with glass surfaces that were prepared with two different transparent preparatory layers (a mixture of Arabic gum and water and a mixture of ox-gall and water) and this showed that the layers had an enhanced wetting effect on the paint (Bayle et a. 2015, 297).

1.3 Subquestions, structure of thesis and methodology

From the current scientific research it can be deduced that in order to answer the research question it is necessary to determine how and with what materials the plates have been painted and what the condition is of these materials, both in delaminating areas and in non-delaminating areas. This information gives insight into the chemical and physical

4_{However, in the context of easel paintings multiple studies have been carried out, and are still ongoing, to}

metal soaps in oil paint. Since roughly the late 1990’s the damaging effect of metal soap formation, including brittleness and delamination, has increasingly been recognized and studied (Hermans 2017, 8). In the past years, progress is booked in the examination and characterisation of metal soap issues and in the understanding of the physical and chemical processes involved, of which a PhD research of J.J. Hermans is an important and very recent example (promotor: prof. dr. P.D. Iedema, copromotor: dr. K. Keune) (Hermans 2017).

(13)

13

characteristics of the materials present (and possible differences between delaminating and non-delaminating areas), which in turn should be related to the original use of the objects and the (environmental) conditions the plates have been exposed to. It is expected that this allows for an in-depth discussion about the mechanisms behind the observed paint delamination, which has been done scarcely in earlier research into cold painting on glass. This has led to a number of subquestions, which will be answered in individual chapters.

Chapter 2 will address the questions: what are diafanoramas, how were they used and how were they made? The answers to these questions give insight into what type of conditions and mechanical actions diafanorama plates were exposed to and it provides a context for the findings of the research into three plates of the Rijksmuseum (chapter 3). The questions will be answered by means of literature research. In order to gain more insight into how diafanorama plates were viewed exactly, the original set-up of diafanoramas was reconstructed together with Margreet Wagenaar-Fischer (manager and owner of the Christiaan Huygens Theatre in Zeist and collector of diafanoramas).

In chapter 3 the focus is narrowed down to the three plates of the Rijksmuseum. The history of the plates will be discussed first in order to gain insight into the environmental conditions the plates have been subjected. Furthermore, the following questions will be addressed: what are the painting techniques and materials of the three plates and are there differences between delaminating and non-delaminating areas in terms of materials used and (paint) layer build-up?

Chapter 4 will focus on the questions: what is the condition of the plates, how is the paint delamination manifested and do similar delamination manifestations occur at other diafanoramas as well (in and outside the collection of the Rijksmuseum)? This provides a detailed image of the problems of paint delamination on the three plates, where it occurs and in what form, and whether these problems are unique to the three plates. The glass of the three plates will receive a condition check as well, to establish whether signs of degradations are still not present.

The questions addressed in chapter 3 and chapter 4 will be answered by means of visual examination in normal light, raking light and UV, complemented by different technical analyses of paint samples that were taken from the plates (table 1.1).

Chapter 5 contains a final discussion of the possible causes for the paint delamination. The discussion is based on the results of the technical analysis of the three plates (as described in chapter 3 and 4) and literature research directed by these results. The underlying question is: what can be said about the potential relation between the observed delamination phenomena, the painting technique and materials of the plates and the environmental conditions the plates have been exposed to? With the answers to these questions a conclusion can be formulated whether certain causes for the paint delamination can be designated or excluded and where further research is needed.

(14)

14

1.4 Relevance of research

To date, no technical research as outlined above has been carried out on the three plates.5 As the material composition and degradation phenomena influence conservation decisions, both active and passive, this thesis project has a direct relevance for their conservation. The technical analyses are expected to give more insight into the painting materials. Knowing for example the binding medium/media is important for choosing suitable materials to consolidate the delaminating paint (Baumer et al. 2012, S9). Furthermore, a better understanding of what could have caused or promoted the paint delamination will help establish future preventive conservation measures.

As has become clear from the discussion of the current level of scientific knowledge, technical research into diafanorama plates has barely been performed up to now. This research will therefore contribute to the knowledge of the method of manufacture of these objects. In turn, this insight in the material properties of diafanoramas could aid in their conservation. As part of the research, other diafanorama plates in and outside the collection of the Rijksmuseum have been visually examined and similar delamination phenomena were observed in some occasions (see section 4.3). It is hoped that the research not only aids in the conservation of the three diafanorama plates in question but also of other diafanorama plates showing similar problems of paint delamination.

5_{Schinkel, Margot van. Oral communication. 22 November 2016.}

Technique Purpose

X-ray Fluorescence spectrometry (XRF) Analysis of the plates to study the composition of the glass and the pigments used in multiple areas.

Zeiss AXIO Imager. A2m Microscope (software: ZEN 2)

Microscopic analysis of paint cross-sections in order to study the layer build-up and the pigments present in bright field, dark field and UV.

Scanning Electron Microscopy - Energy Dispersive X-ray Spectroscopy (SEM-EDX)

Analysis of paint cross-sections in order to gain more insight in the layer build-up and the inorganic components present.

Transmission Fourier Transform Infrared spectrometry (transmission FTIR)

Analysis of paint samples to determine the type of binding medium of the paint.

Thermally assisted Hydrolysis and Methylation Pyrolysis Gas Chromatography – Mass Spectrometry (THM-Py-GC/MS)

Analysis of paint samples to identify the binding medium of the paint (in case of inconclusive results from transmission FTIR).

(15)

15

Chapter 2 Diafanoramas: history, use and manufacture

The chapter will focus on the historical context of diafanoramas, on how these objects were used and on how diafanorama plates (could) have been made. It will become clear below that sometimes literature regarding reverse-glass painting and magic lantern slides was consulted for this, in cases when literature regarding diafanoramas was not sufficient. 2.1 Historical context

Diafanoramas were popular in the Netherlands from roughly the second half of the eighteenth century to the end of the first half of the nineteenth century (Kuijper 1990, 45). Their popularity can be placed in the context of the Enlightenment (characterised by a raised interest in optics and other natural sciences within a wider audience) and a corresponding demand for optical games and indoor light shows (Kuijper 1990, 19-20). Kuijper discovered Dutch newspaper advertisements and auction catalogues, of which the earliest originate from 1770, in which sets of diafanorama plates were offered, often in combination with the necessary equipment (a separate frame to hold the plates, a burning mirror and a light source) (Kuijper 1990, 44-45). The sale of diafanoramas increased from this time on until around 1830. In this time sometimes the name of the artist who painted the plates is given in the catalogues. Based on these names, Kuijper suggests that painting diafanorama plates was commonly done by painters who were specialised in producing theatre sets (Kuijper 1990, 83, 90). After 1830, diafanoramas became less popular, their artistic value decreased and their production and sale eventually came to an end (Kuijper 1990, 78, 81-82).

On the basis on sales data in auction catalogues, Kuijper shows in her thesis that diafanoramas were expensive products and were only bought by wealthy citizens (Kuijper 1990, 87-89). People who could not afford diafanoramas could see them at funfairs or at traveling performers who gave shows with optical instruments (Kuijper 1990, 92-95). 2.2 Use of diafanoramas

Probably the most important feature of the diafanorama was the creation of an image with a remarkable illusion of depth. This was first of all achieved by painting the fore-, middle- and background of the image on three different plates. The plates were placed behind each other in a wooden outer frame, at a distance of two or three centimetres. The depictions on the plates fit together and form an image with an enhanced perspective. In addition, the burning mirror enhanced the illusion of depth even more. The operation of a burning mirror is explained in a Dutch translation of the book Nouvelles récréations physiques et mathématiques, written by M. Guyot in 1768:

(16)

16

6

The second effect mentioned was important for diafanoramas. The enlargement of the image and the illusion of the image floating in front of the mirror created a spatial effect. Furthermore, due to the curvature of the mirror and the stage like arrangement of different parts of the picture behind each other, the viewers would observe a changing image when they changed their viewing angle (Wagenaar, Duller and Wagenaar-Fischer 2014, 93). The diameter of the mirrors ranged from 25 centimetres to no less than 50 centimetres (Wagenaar, Duller and Wagenaar-Fischer 2014, 93) (fig. 2.1).

In order to see the image in the mirror, the plates were illuminated from behind by a light source. As discussed by Kuijper, from auction catalogues it becomes clear that the light source consisted of a number of candles placed in a wooden cabinet (Kuijper

1990, 43). In figure 2.2 such a ‘light box’ is shown. The inside of the box is covered with tin foil. The reason for this is unclear but is seems likely that the tin functioned as protection for the wood and its reflective surface might have enhanced the light intensity from the candles. Beside candles, oil lamps have been used as well. The Christiaan Huygens Theater has a series of 10 diafanorama plates in its collection and these plates are accompanied by a wooden cabinet with two oil lamps inside.7

6_{Translated from Dutch by Noa Kollaard.}

7_{Wagenaar-Fischer, Margreet. Oral communication. 6 March 2017.}

Fig. 2.1 Burning mirror, glass, wood and brass, h. 60 cm x w. 58 cm x d. 24.5 cm, ca. 1880-1910. Rotterdam: Museum Rotterdam. The mirror is part of a collection of diafanorama plates of the museum. Photo: Museum Rotterdam.

The different phenomena that this kind of mirrors create, are

First of all, they bring together all the rays of fires or lights that fall onto them to one focal point, in such a way that they heat up or kindle flammable substances, and melt and burn metal and the hardest stones.

Furthermore, these mirrors show objects, depending on the distance, either remarkably enlarged or reversed, or even entirely outside and in front of the mirror.

Thirdly, when one places a light source in front and close to the surface of the mirror, the rays that fall onto the mirror will be united and radiate continuously when they are reflected back, and together with the rays that fall directly on the objects, they bring a lot of light (Arrenberg 1774, 223-224). 6

(17)

17

Figure 2.3 show the manner in which the mirror, the frame holding the plates and the light source were placed in order to view the image.8 As the afore given citation from M. Guyot’s nieuwe natuur- en wiskonstige vermaaklykheden makes clear, the distance at which the mirror was placed was important for the desired effect. If the mirror is placed too far from the plates, the image is turned upside down and when the mirror is placed too close it becomes more difficult to look into the mirror and see the total image. The burning mirror and the light source in figure 2.3 are not original but it gives an impression of the distances between the different components in order to create a satisfying image (shown in figure 2.4).

8_{As was demonstrated by Margreet Wagenaar-Fischer during a visit to the Christiaan Huygens Theater in}

Zeist at 6 March 2017.

Fig. 2.2 Light box with candles, wood and metal, h. 32 cm x w. 44 cm x d. 28 cm, ca. 1750-1800. Rotterdam: Museum Rotterdam. The light box is part of a collection of diafanorama plates of the museum. The frontside could be closed (left) and opened (right). The front side was most likely opened in order to illuminate the diafanorama plates. Photo left: Museum Rotterdam. Photo right: Noa Kollaard.

Fig. 2.3 The plates are pushed from above in the wooden frame, in which they are held at a distance of 2 or 3 centimetres from each other (right). The light source was placed behind the frame holding the plates and the burning mirror was placed in front (left). The viewer stood behind the light and looked over the light and the plates into the burning mirror. In this set-up, the distance between the mirror (25 centimetres in diameter) and the plates is approximately 30 centimetres. The distance between the plates and the light source is approximately 20 centimetres. Photos: Noa Kollaard.

(18)

18

According to Kuijper, privately owned diafanoramas were mainly shown at special occasions such as birthday parties or when influential people from upper classes came to visit (Kuijper 1990, 89-90). She gives a description of how such a show might have taken place, based on a conversation with a man whose grandfather used to own diafanoramas. The diafanorama plates, the mirror and the light source were placed on a table at the right distance from each other. Then the candles were lit and the room was darkened. Kuijper describes that the viewers often took ample time to walk around the table to view the image from different angles (she does not specify the exact duration) (Kuijper 1990, 90-91).

2.3 Manufacture of diafanorama plates

2.3.1 Glass and preparation of the glass before painting

The supports of diafanorama plates are flat sheets of colourless glass with measurements ranging from around 25 by 30 centimetres up to 60 by 80 centimetres. Glass is an inorganic material that is composed of silica (SiO2), which forms the main network of the

glass. In this network positively charged cations are present that are bound to the negatively charged oxygens of silica. The most common cations are the alkali metal ions potassium (K+) and sodium (Na+) and alkaline earth metal ions like calcium (Ca2+) and magnesium (Mg2+) (added to the glass as potash, soda, lime and magnesia respectively). In addition, elements such as aluminium, phosphorous and iron, originating from the raw materials or from the environment, are present in minor quantities (Davison 2003, 178-177). According to Kuijper, for the manufacture of diafanorama plates flat glass sheets were imported from France (Kuijper 1990, 53). These sheets were produced by the cylinder method. This is a glass blowing technique whereby the glass blower makes an elongated glass balloon from around 1.5 metres long and with a diameter of around 45

Fig. 2.4 Image of the diafanorama plates as seen in the burning mirror. Photo: Noa Kollaard.

(19)

19

centimetres. The cylinder is cut open lengthwise and the glass is flattened. In this way, relatively cheap glass sheets of 2 to 3 mm in thickness were made (Coppieters-Mols 1996, 15; Kuijper 1990, 52).

Kuijper does not write about if and how the glass sheets were prepared before painting. However, it seems likely that in order to improve the adhesion of an organic material (the paint binder) on an inorganic material (the glass support) – which are inherently incompatible – the glass did needed to be prepared. In literature on painting on glass (both primary and secondary sources) different proceedings are mentioned for this. First of all, it is regularly stated that the glass should be thoroughly cleaned or degreased (Baumer and Dietemann 2016, 76; Blewett 2012a, 112; Bretz 2013, 24). A method to degrease the glass is found in Ars vitraria exerimentalis oder Vollkommene Glasmacher-Kunst, written by Johannes Kunckel in 1689. The author recommends rubbing the glass with grated lime or lead white. Secondly, the application of a transparent preparatory layer is advised in several art technological sources on reverse-glass painting dating from the fifteenth to the eighteenth century (Davison 2003, 56).9 The materials most often mentioned for this are drying oils such as linseed oil and nut oil, natural glues, glair, gelatine and a range of natural resins (Blewett 2012a, 112). Also in nineteenth-century handbooks on painting glass lanternslides the application of a transparent ground layer can be found. Two interesting sources are How to colour photographs and lanternslides by aniline dyes, water and oil colours, crystoleum and other processes, published by Richard Penlake in 1895 and The Book of the Lantern. Practical Guide to the Working of the Optical (or Magic) Lantern. With Full and Precise Directions for Making and Colouring Lantern Pictures, by H.C. Hepwarth in 1888. According to Penlake, when the lanternslides are painted with watercolours it is better to apply a varnish layer before painting (Penlake 1895, 29). When painting with oil paint the slides only need to be bone dry. A varnish is not necessary, as Penlake says, although some painters do apply it (Penlake 1895, 34). H. C. Hepwarth describes the application of a transparent ground layer, and also the application of the outlines of the image preparatory to painting:

Another method described by Hepwarth entails roughening the glass surface by rubbing a mixture of flour emery10_{with water in between the slide and another piece of}

glass (Hepwarth 1888, 101). After this, the preparatory drawing can be applied:

9_{For example, Kunckel recommends a linseed oil layer (Bretz 1999, 198). In the Marciana Manuscript it is}

recommended to first apply a coat of nut or linseed oil on the places that you want to paint. After this the pigments, ground in the same oil, can be applied. Another option is to coat the glass with warm glue (Merrifield 1849, 616-618).

10_{Emery is a powdered form of corundum (a mineral containing aluminium oxide) that was used as an}

abrasive (Oxford Dictionaries 2017, n. pag).

It will be found that such an outline is easier to produce if the glass be first covered with a layer of varnish. Some use a weak solution of gelatine in water, to give the glass the necessary surface for taking the pigment. A solution of sugar has also been recommended for the same purpose (Hepwarth 1888, 98).

He works with a hard pencil on a piece very finely-ground glass ; afterwards filling the outlines thus made with water colours, and applying a coat of varnish so as to give the necessary transparency to the picture (Hepwarth 1888, 99).

(20)

20

The varnish used to make the glass transparent after roughening is pale Canada balsam in benzene (Hepwarth 1888, 100). As an alternative to roughening the glass, Hepwarth advises to apply a matt varnish, made of sandarac, mastic, ether and benzole. When this varnish is dry, the preparatory sketch can be drawn with a pencil (Hepwarth 1888, 102-103).

2.3.2. Painting techniques and materials

The image to be painted had to be divided into different perspective planes, which were to be applied on separate glass plates. In accordance with atmospheric perspective, the foreground was painted in the brightest colours and the middle- and background in decreasingly saturated colours. Considering the fact that light needed to shine through the plates, the image was never painted completely opaque. The areas that had to be transparent in order to let the image of the plate behind it show through received no paint (Kuijper 1990, 55-56).

The transparency of the glass was exploited to create various visual effects. The spatial illusion of the image was enhanced by painting elements of the depiction on the reverse side of the glass. Such an element could be painted in a reverse-glass technique (by which the paint layers are applied in reverse order, so starting with the uppermost layer) or in the ‘normal’ non-reversed manner. Sometimes an element was painted partially on the front side while certain parts were left open and filled by a coloured layer on the reverse side (fig. 2.5). Furthermore, the suggestion of light-emitting parts of the image was achieved by applying a light translucent paint layer on the reverse side and scratching away the paint on the frontside so that more light could shine through (fig. 2.6-2.7). In general, both sides of the back plate were often entirely covered with paint. A neutral layer was usually applied on the reverse side of the plate. This layer would scatter the light coming from behind, which created a more diffuse lightening (Wälde 2014, 9).

Fig. 2.5 Anonymous, Diafanorama of a scene from Gijsbrecht van Aemstel, fifth part, fourth scene, paint on glass, 35 cm x 42 cm, 1750-1800. Amsterdam: Theatre collection of the University of Amsterdam. On the left a detail of the reverse side of the glass, showing a translucent yellow paint layer (indicated by red arrows). The yellow layer remains visible at the frontside (right photo) were parts of the image have been left open. Photos: Noa Kollaard.

(21)

21

Fig. 2.6 Anonymous, Diafanorama of fountains on the Hofvijver, paint on glass, 30 cm x 28 cm, 1750-1800. Rotterdam: Museum Rotterdam. On the left a detail of the reverse side of the glass, showing translucent yellow and blue paint layers. On the front side the sparkling water of the fountains is suggested by scratching away the black paint. Photos: Noa Kollaard.

Fig. 2.7 Anonymous, Diafanorama of river landscape in moonlight, paint on glass, 30 cm x 39 cm, 1750-1800.

Rotterdam: Museum Rotterdam.

Detail of subtle reflections of the moonlight in the water, created by making wavy scratches in the paint (indicated by red arrow). Photo: Noa Kollaard.

(22)

22

Not much information is available about the types of paint used for diafanoramas. As mentioned in the current scientific knowledge (section 1.2), a single analysis of the paint of a diafanorama plate has been performed by the Centraal Laboratorium voor Onderzoek van Kunst en Wetenschap in 1988. The paint on the reverse side was identified as oil paint, while on the front side both oil and egg were identified as binding media (Kuijper 1990, 60). 11 Such a combination of oil and egg might not have been uncommon, as it also occurs in other types of paintings on glass. Technical analyses of twenty-five reverse-glass paintings dating from the fourteenth to the twentieth century carried out by the Doerner Institut showed that both oil and egg can be present in one painting (sometimes found in different colours, sometimes found together in one colour). In addition, different types of natural resins (mainly larch turpentine and pine resin) were also identified, often mixed with (boiled) linseed oil (Baumer and Dietemann 2016, 71).

For lanternslides, both Hepwarth and Penlake mention the use of oil paint and watercolours. According to Hepwarth, sometimes both types of paint were used for one slide (Hepwarth 1888, 98). They do not mention the exact nature of the oil paint and watercolours. In this context it is interesting to refer to a technical examination of two nineteenth-century lanternslides carried out at the University of Torino. This showed that the binding medium of the paint was linseed oil, in some areas in combination with a natural resin. Natural gums were also found but it was not clear whether these originated from the paint or from a transparent ground layer that was found underneath the paint (Ploeger, Scalarone and Chiantore 2010, 41). For painting with oil paint, Penlake advises to use a suitable medium to dilute the paint. He recommends using megilp – which he believes is a mixture of turpentine and drying oil – for Prussian blue paint that is used for painting skies. For the painting of details he recommends using a mixture of copal or mastic and turpentine (Penlake 1895, 34).12

2.4 Conclusion

In this chapter the questions what diafanoramas are, how they were used and how they were made have been addressed. In the context of the research to the causes of paint delamination an important point is the use of candles or oil lamps to view the diafanorama plates. As these light sources produce warmth, the plates may have been exposed to elevated temperatures, which could have had an effect on the chemical and physical behaviour of the paint layers on the glass. According to Kuijper, diafanoramas may have only been viewed at special occasions. However, this is based on the experience of one person, meaning that no general conclusions should be drawn from this.

11_{More specific information about the results could not be obtained. The RCE library has been contacted for}

a report of the analysis but no documentation on the analysis has been archived.

12_{The materials mentioned by Penlake were common painting materials. Leslie Carlyle (Associate Professor}

of Paintings Conservation in the Conservation and Restoration Department at the Faculty of Sciences and Technology of the Universidade Nova de Lisboa) has carried out research into (British) oil painting instruction manuals and handbooks from the eighteenth and nineteenth century. In her book The Artists’ Assistant, nineteenth-century recipes for megilp and for paint media with oil, copal and/or mastic and (spirit of) turpentine are presented and discussed (Carlyle 2001, 103-105,125-128, 392-394, 411-416). These recipes show that megilp is not a mixture of turpentine and drying oil but a mixture of mastic varnish (or sometimes copal varnish) and turpentine. These media could be bought ready-made (Carlyle 2001, 105, 128-129).

(23)

23

An important point is the fact that in manuals on painting on glass it was often recommended to first clean the glass thoroughly and apply a preparatory layer before applying the paint. The exact function of the preparatory layer is less clearly stated in the sources, but according to H.C. Hepwarth it gives the glass a surface that can better take the preparatory painting or drawing. In any case, the result is the presence of an additional layer between the paint and the glass, possibly influencing the adhesion between the two. Many different materials have been advised for the preparatory layer (drying oils, natural glues, glair, gelatine and natural resins) and the same goes for the binding medium of the paint layers that were applied on top (drying oils, egg, natural resins and vegetable gums). This information provides an idea of what materials could be present on the three Rijksmuseum plates, which will be the focus of the following chapter.

(24)

24

Chapter 3 Three diafanorama plates of the Rijksmuseum: technique and materials This chapter will focus on three diafanorama plates of the Rijksmuseum. First a description of the objects and their history will be given. After this follows a discussion of how and with what materials the three plates have been made.

3.1 Object description and history

The three diafanorama plates are each composed of a glass sheet of approximately 29 centimetres wide, 24.5 centimetres high and 0.3 centimetres thick. The sheets are set in an oak wooden frame of 43 by 36 by 0.8 centimetres. The plates are not signed.13 They are dated by the Rijksmuseum between 1750 and 1830 (the period in which diafanoramas were most popular).

All three plates were cold-painted on both sides. Together, the plates depict a battle on a village square (fig. 3.1).14 On the plate depicting the foreground15, three out of ten figures are painted on the reverse side of the glass. At the middle plate, eight out of sixteen figures and two out of three trees are painted on the reverse side. The front side of the back plate is covered entirely with paint, as well as the reverse side (which contains an even, pale yellow paint layer).

13_{See appendix II for additional information on artists of diafanoramas, with a focus on the three}

Rijksmuseum plates.

14_{See appendix I.a for large-size photos of each individual plate.}

15_{From here onward, this plate will be referred to as the front plate. The plate depicting the middle ground}

and the plate depicting the background will be referred to as the middle plate and the back plate respectively. Fig. 3.1 Digital reconstruction of the three plates placed behind

each other. On the front plate, ten fighting and fleeing figures and one tree are depicted. On the middle plate, sixteen figures and three trees are depicted. The back plate depicts a large battle in between two small houses and some trees at the left and a church at the right. Photos: Rijksmuseum. Digital reconstruction: Noa Kollaard.

(25)

25

The plates were acquired by the Rijksmuseum in 1887 along with eighteen other diafanorama plates. At the time when they were acquired, the plates were not accompanied by an outer frame, nor by a light source and a burning mirror. The museum has no documentation on the former owner(s) or other information that could give insight into the original set-up of the plates and what type of light source and mirror have been used for them.16

The three plates have most likely never been exhibited in the museum. They have been stored in the attic of the Rijksmuseum along with the other plates for a long period of time. In this storage, all plates were stacked on top of each other. The climate of the attic was not controlled. The exact duration of this period cannot be determined with certainty but it could have started already in 1888, when the plates are listed in the museum’s inventory for the first time. In 2002, the plates were moved to an external, climate-controlled depot.17_{Special boxes were made to transport the plates. The plates that were}

stable, including the front plate of the set in question, were placed vertically in wooden chests. In the depot the chests are left open for ventilation.18 The plates that were not stable enough, including the middle and back plate, were placed horizontally in acid-free boxes (fig. 3.2). In 2012, the plates received a cleaning treatment, which entailed dry and wet cleaning of the glass (not of the paint). After this treatment the plates were returned to the depot.19

16_{The diafanorama plates are listed in the inventory of the Rijksmuseum of 1888 but no mention is made of a}

light source or a burning mirror.

17_{Schinkel, Margot van. Oral communication. 22 November 2016.}

18_{It is planned to store these plates in an acid-free environment in the future.}

19_{Besijn, Nienke. Oral communication. 19 January 2017. For wet cleaning of the glass a mixture of}

deionised water and ethanol was used. A report of the treatment is present in the archive of the Rijksmuseum. Fig. 3.2 Wooden chest in which the most stable

plates are stored (upper left) and individual boxes for the plates that are less stable (upper right and lower left). Photos: Noa Kollaard.

(26)

26

3.2 Examination of materials and techniques 3.2.1 Methods

In order to gain insight into how and with which materials the plates have been made, they were investigated by means of visual examination in normal light, raking light, UV and with the aid of a stereomicroscope. This was complemented with different technical analyses. In addition, the wood of the frames was investigated by means of dendrochronology. The applied analyses will be justified first, followed by the results of the research.

3.2.1.1 XRF

XRF was carried out on to the glass, in order to assess whether the composition shows any abnormalities that might have influenced the current condition of the plates.20_Spot

analyses of the painted surface were made to study the elemental composition of the pigments.

3.2.1.2 Analysis of paint samples

In order to study the paint layer build-up and to determine the composition of the paint it was necessary to take paint samples. Samples were taken from both delaminating and non-delaminating areas, in order to compare the paint stratigraphy and material composition in these locations. Table 3.1 gives a concise overview of the sample locations.21 Some tiny, loose paint flakes were collected from the back plate with a small brush.22 From both the front and middle plate a sample was taken with the aid of a scalpel.

Of all three samples a part was embedded as paint cross-section. The cross-sections were analysed using light microscopy and by SEM-EDX.23 The remaining sample material

20_{The analysis was carried out with a DELTA Professional Handheld XRF Analyser of Olympus by Guus}

Verhaar (UvA). Two spot sizes of 0.8 cm2 _{and of 0.07 cm}2_{were possible.}

21_{See appendix III for the complete sample forms.}

22_{This was done in two times because the material taken at the first time was not enough for}

THM-Py-GC/MS.

23_{SEM-EDX was carried out by Maartje Stols-Witlox (UvA). It was performed with a JSM-5910 lv scanning}

electron microscope of JEOL with a Thermo Fisher Scientific SDD Ultradry detector and Noran System Seven software. BSE images were taken at 20 kV acceleration voltage.

Plate Sample name Material Delaminated area

Front plate: BK-NM-8462-J

J-01 Brown paint of the

ground

No Middle plate:

BK-NM-8462-K

K-01 Dark brown paint of a tree trunk

No acute

delamination but a lot of paint loss Back plate:

BK-NM-8462-L

L-01 & L-02 Light blue paint of the sky

Acute delamination and a lot of paint loss

(27)

27

was analysed by transmission FTIR, with which both organic and inorganic functional groups in paint samples can be identified (Stuart 2007, 126-127).24 The main aim of FTIR was to characterise the type of paint binding media used. The technique is non-destructive and it requires only tiny amounts of sample material. However, FTIR cannot always give conclusive answers, as some materials have overlapping IR-absorption bands. This was the case with the samples from the back plate. Therefore, the samples were analysed with THM-Py-GC/MS, with which different organic components present can be separated and identified individually (Stuart 2007, 302-303).25

3.2.1.3 Dendrochronology

Because there were doubts about the originality of the wooden frames (see paragraph 3.2.2.3), dendrochronological analysis was carried out on the frame of the front plate, which is in the most stable condition of all three plates.26_{The growth rings of the frame}

were visible at the left, right and lower sides and were measured with a magnifying glass with an integrated scale.

3.2.2 Results

3.2.2.1 Glass and preparation of the glass

In total five different XRF spot analyses were carried out on the plates (three on the front plate, two on the middle plate and one on the back plate).27 No unusual elements were detected and the measurements of all three plates gave similar spectra, showing peaks for silicon, potassium, iron and calcium in corresponding ratios.28

The glass is not perfectly flat but shows slight undulations. Multiple tiny air bubbles trapped inside the glass indicate that the sheets have been created by means of a glass blowing technique.29 The surface of the glass is rather smooth. There are no signs that the glass has been roughened in preparation for painting.

By means of visual examination of the front and middle plates a translucent white layer with a streaky appearance underneath the paint was detected (fig. 3.3-3.4). The layer extends below the painted parts of the brown ground and the figures. However, a similar layer was not seen underneath the trees or underneath the brown and yellow paint of the two leftmost figures on the front plate. At the same locations where the translucent white layer is visible in normal light, a yellow fluorescence with a streaky appearance is apparent in UV (fig. 3.5-3.6).30_{Furthermore, the cross-section of the brown ground on the front}

plate shows a very thin layer underneath the paint layer – with a distinct yellow

24_{The analysis was carried out with a PERKIN Elmer 100 FTIR spectrometer by Suzan Catucci-de Groot}

(RCE).

25_{THM-Py-GC/MS was carried out by Henk van Keulen (RCE). See appendix VI.a for technical}

specifications.

26_{The analysis was carried out by Peter Klein (wood biologist at the University of Hamburg).}

27_{See appendix VII.a for an overview of the analysed spots and all the spectra.}

28_{Appendix VII. See also tables 3.2-3.4 later on in this chapter.}

29_{Van Schinkel, Margot. Oral communication. 20 December 2016. This could be the described cylinder}

method, however this cannot be determined with certainty.

(28)

28

fluorescence in UV – while no such layer is visible in the cross-section of the tree trunk of the middle plate (fig. 3.7-3.8).

Fig. 3.3 Detail of the front plate seen through the glass. The image on the right is a magnified detail of the image on the left. A thin white layer underneath the paint is visible. Photos: Noa Kollaard.

Fig. 3.4 Detail of the middle plate seen through the glass. The image on the right is a magnified detail of the image on the left. A white layer

underneath the paint is visible. The border of the layer is indicated with a red arrow. Photos: Noa Kollaard.

(29)

29

Fig. 3.5 Front plate. The image on the right shows a mapping that indicates (in blue) which figures are painted on the reverse side.

The image above is a photo of the front plate in UV light. A yellow fluorescence is visible at the figures and the ground that have been painted on the reverse side. The two leftmost figures show only partially fluorescence (only the heads and the cape of the figure on the right).

NB: Lighter coloured areas in the depiction painted on the front side (such as the smoke coming from the gun) show a weak fluorescence as well. This might be due to the presence of the pigment lead white (De la Rie 1982, 1). Photos: Rijksmuseum.

(30)

30

Fig. 3.6 Middle plate. The image on the right shows a mapping that indicates (in blue) which figures are painted on the reverse side.

The image above is a UV light photo of the middle plate. A yellow fluorescence is visible at the figures and the ground that have been painted on the reverse side. Below the single tree painted on the reverse side no fluorescence is visible.

NB: Lighter coloured areas in the depiction on the front side show a weak fluorescence as well. This might be due to the presence of the pigment lead white (De la Rie 1982, 1).

Photos: Rijksmuseum.

(31)

31

Because the whitish layer has no aesthetic function – it is entirely covered with paint and does not play a role in the colour scheme of the image – it seems likely that it is a preparatory layer that has been applied to the glass in order to receive the paint. As has become clear in chapter 2, this step was sometimes recommended in historical manuals for painting on glass. The streaky appearance suggests that the layer has been applied by brush. Interestingly, the brushstroke pattern sometimes continues throughout two adjacent but separately painted areas (fig. 3.9). Thus, it seems as if first a larger area of the glass was covered with the preparatory layer and that in a later stage the excess of material was removed.31

31_{This could provide an explanation for the absence of a preparatory layer underneath the trees. Because the}

trees have an intricate shape, removing the excess material around them may have been a difficult task. The painter might have omitted to apply a preparatory layer underneath the trees for this reason.

Fig. 3.7 Cross-section of the brown ground on the front plate (sample J-1), UV radiation (365 nm)

(top) and dark field

(bottom). In UV radiation a small layer with bright

yellow fluorescence is

visible underneath the paint layer.

Photos: Noa Kollaard.

Fig. 3.8 Cross-section of the tree trunk on the middle plate (sample K-1), UV radiation (365 nm) (top) and dark field (bottom). Only one paint layer is present.