Travelling through textiles: The exploration of fragile textiles through 2D and 3D imaging.

(1)

Travelling through textiles

The exploration of fragile textiles through 2D and 3D imaging

Paulien Coopmans 10321411

Master’s thesis July 2017

Supervisor: Emmy de Groot Second assessor: drs. Femke Prinsen

(2)

Photo cover: Snapshot of a 3D reconstruction of sample uva-0213. For this reconstruction, X-ray computed tomography was used.

(3)

Abstract

The 2014 discovery of seventeenth century textile objects in a shipwreck in the Wadden Sea, near the Island of Texel (The Netherlands) is a challenge to researchers and textile conservators. Many of the very fragile artefacts do not allow handling. This makes examination of the shape and construction of these objects difficult. In this thesis, penetrative imaging techniques of which some were so far hardly used in textile conservation are studied and tested, as to assess their added value to the research on textile objects. The evaluation of the tests is based on a methodological approach that takes into account the problems that arise when scientists in non-related fields have to communicate. The work process and the outcomes were evaluated on the aspects effectiveness, user-friendliness and efficiency. The selected techniques, 2D X-radiography, X-ray computed tomography (CT) and neutron tomography, prove to be useful tools. The output of the test with neutron tomography is promising. There are however many practical limitations to its applicability in the field of textile conservation. 2D X-radiography is an accessible technique that is fit to orientate on the features of an object. The most powerful of the selected techniques is X-ray computed tomography. For the tests with X-ray computed tomography, a unique set-up at CWI in Amsterdam was used. The so-called FleX-ray scanner was used to scan several textile test samples and eventually one of the fragments of a garment from the Texel collection. The real-time scanning procedure with the FleX-ray resulted in datasets consisting of about 1500 X-radiographs per sample that together were combined to make digital 3D reconstructions of the samples. These reconstructions enable one to obtain information about layers of the sample that cannot be seen in visible light. However, this technique too has some serious limitations. It was concluded that the combination of textile and metal in one sample poses a challenge for imaging with X-rays. This limited the usefulness of the output data of the Texel fragment. In addition, the tests showed that mounting the samples is problematic in this set-up. As was experienced during the tests, the communication process between scientists of non-related fields is extremely time-consuming especially on the side of the textile conservator. Nevertheless, in this thesis the argument is made that textile conservators should increase the effort to explore and apply a range of surface and penetrative imaging techniques as these techniques can contribute to the preservation and accessibility of fragile textile artefacts for scientific research and the public.

(4)

Samenvatting

De vondst van zeventiende eeuwse textiele voorwerpen in een scheepswrak in de Waddenzee bij het eiland Texel die in 2014 werd gedaan, is een uitdaging voor onderzoekers en textielrestauratoren. Veel van de zeer kwetsbare voorwerpen kunnen niet worden gehanteerd. Dat maakt onderzoek naar de vorm en constructie van deze objecten moeilijk. In deze scriptie worden penetrative imaging technieken onderzocht en getest op hun toegevoegde waarde voor het onderzoek van objecten van textiel. Deze technieken worden op dit moment nauwelijks toegepast binnen textielconservering. De evaluatie van de tests is gebaseerd op een onderzoeksaanpak die rekening houdt met de problemen die zich voordoen bij de communicatie tussen wetenschappers van niet-verwante onderzoeksterreinen. De werkprocessen en de uitkomsten werden beoordeeld op de aspecten effectiviteit, gebruikersvriendelijkheid en efficiëntie. De geselecteerde technieken - tweedimensionale röntgenfotografie, röntgentomografie (CT) en neutronentomografie - blijken bruikbare instrumenten. De resultaten van de tests met neutronentomografie zijn veelbelovend. Er zijn echter veel praktische beperkingen in de toepasbaarheid van deze techniek binnen het terrein van textielconservering. Tweedimensionale röntgenfotografie is een toegankelijke techniek die bruikbaar is ter oriëntatie op de kenmerken van een object. De meest krachtige van de geselecteerde technieken is röntgentomografie. Voor de tests met röntgentomografie werd gebruik gemaakt van een unieke opstelling bij het CWI in Amsterdam. Deze zogenaamde FleX-ray scanner werd gebruikt voor het scannen van verschillende textiele proefstukken en uiteindelijk ook voor een fragment van de Texel collectie. De real-time scanningprocedure met de

FleX-ray leverde datasets op van circa 1500 röntgenopnamen per proefstuk die gecombineerd werden om er

vervolgens 3D reconstructies van te maken. Deze reconstructies maken het mogelijk om informatie over de lagen in de proefstukken te verkrijgen die niet door middel van zichtbaar licht zijn waar te nemen. De techniek heeft echter enkele aanzienlijke beperkingen. Er werd geconcludeerd dat de combinatie van textiel en metaal in een proefstuk een uitdaging vormt voor het in beeld brengen van het object met behulp van röntgenstralen. Dit beperkt de bruikbaarheid van de output data van het fragment van de Texel collectie. Voorts toonden de tests dat het installeren van het object in deze testopstelling problematisch is. Zoals bleek tijdens de tests, is het communicatieproces tussen wetenschappers van niet verwante terreinen buitengewoon tijdrovend, in het bijzonder aan de kant van de textielrestaurator. Desalniettemin wordt in deze scriptie betoogd, dat het zinvol zou zijn voor textielrestauratoren om het terrein van penetrative en surface imaging technieken verder te verkennen en te onderzoeken. Toepassing van deze technieken kan bijdragen aan het behoud en de toegankelijkheid van kwetsbare textiele voorwerpen voor wetenschappelijk onderzoek en voor het algemene publiek.

(5)

Acknowledgements

I would like to thank all the researchers and experts of the various institutions for thinking along and sharing their knowledge with me: Jenny Tiramani (School of Historical Dress, London), dr. Annelies van Hoesel (Rijksmuseum, Amsterdam), dr. Lambert van Eijck (TU Delft), the researchers of the computational imaging group of Centrum Wiskunde & Informatica (CWI, Amsterdam), prof.dr. Joost Batenburg, prof.dr.ir Robert van Liere and Alexander Kostenko. I would also like to thank Ana Serrano (RCE) for the SEM and HPLC analyses.

I am especially indebted to dr. Sophia Coban (CWI) for spending days in the scanner lab with me, trying to obtain the best possible output, even when both hardware and software conspired against us.

I would also like to thank my supervisor, Emmy de Groot, and Prof.dr. Maarten van Bommel (UvA).

Finally, I am grateful to Lambert Teerling for his patience and help and for staying optimistic during my struggle to finish this thesis.

Paulien Coopmans Amsterdam, July 2017

(6)

1. Introduction

In the summer of 2014, members of the amateur diving club Duikclub Texel found a wide variety of objects in a shipwreck in the Waddenzee, near the island of Texel. This seventeenth century shipwreck, named the Palmhoutwrak or BZN17, was discovered in 2009 and is located in the western part of the Waddenzee (Burgzand Noord) in a part of the area called De Rede van Texel. 1

Amongst the objects that were brought to the surface by the divers, were over a hundred textile fragments. The discovery of these maritime archaeological textiles is extraordinary, considering the age of the objects in combination with the conditions in which they have resided over more than 350 years and the remarkably good state some of the objects are in. However, on the other hand, the condition of some of the objects is very poor. In general, the objects are fragile and need to be handled with extreme care.

Part of this textile collection was handed over to the University of Amsterdam for further research into, among others, their origins, condition and possible conservation treatments. This will be done in close cooperation with the Rijksdienst voor het Cultureel Erfgoed (RCE), the Rijksmuseum, Museum Kaap Skil, Provincie Noord-Holland (owner), Duikclub Texel and Gemeente Texel. The textile collection has already been inventoried in January 2016 by two conservators-restorers in training at the University of Amsterdam, Sjoukje Telleman and Marijke de Bruijne. Now, further research into the specific textile objects is requested. Part of this research will be conducted by the master students in Conservation and Restoration of Cultural Heritage (specialisation textiles) of the UvA. This thesis must be viewed in this context.

1.1. Observation and problem

Some of the textile objects are in a relatively good state and have a recognisable shape and function (e.g. dress or stockings), other objects are/have not and can merely be called textile fragments. An example of this is a group of six silk fragments: BZN17-6263-019, BZN17-6263-060, BZN17-6263-082, BZN17-6263-101, BZN17-6263-123 and BZN17-6263-136 (Fig. 1). This group of fragments together appears to be the remnants of a garment. Preliminary research, conducted by costume historians, suggests that it is a kaftan; a long, male costume originating from the Middle East. In this thesis, the term kaftan will therefore be used to refer to the group of fragments as a 2

whole.

Translated: The Texel Roads. Bruijne and Telleman 2016b: 2, 3.

1

Stillman n.d: n. pag.; See section 4.1.4. for an elaboration on this subject.

(10)

To get a better understanding of what this must have looked like, research is needed. However, the state they are in (distorted, crumpled, creased, folded, incomplete, torn apart and degraded) makes research questions about the original shape nearly impossible to answer. Moreover, the condition of these fragments does not allow any handling (or restricts this to an absolute minimum). Turning or unfolding the fragments to see how they are constructed, what the back or the inside looks like or finding out how many layers of fabric are present, is practically impossible.

Fig. 1. Collage of photos of the six fragments: BZN17-6263-019, -060, -082, -101, -123 and -136. Photos:

Margareta Svensson, 2017. Collage: Paulien Coopmans. 060 019 082 101 136 123

(11)

For this thesis, two considerations are of importance. First of all, it has become clear that this shipwreck found is exceptional and provides a unique opportunity to document and examine the textile objects that were discovered. However, it has been established that many interesting 3

objects are so fragile, that the handling, needed to document and examine these objects, is difficult and unwanted. A detailed description of the objects and research into, among others, the original shape, dimensions and connection between separate fragments can therefore only partly be executed. Methods that normally are used in textile conservation (visual inspection, making measurements, etc.) are in this case insufficient as to get a complete image of the objects and to be able to make objective statements about the original appearance of the dress.

Secondly, the experience that is developed in other fields of science related to conservation and restoration, such as archaeology and other conservation specialisations, can be used. In these fields of science it is common practice to apply imaging techniques to get maximum information on objects in a poor condition. These techniques were developed for application in the field of 4

medicine, design and engineering. These techniques have scarcely been applied in textile conservation.

1.2. Relevance and challenge

The use of imaging techniques might help overcome the problems that arise when using the traditional techniques of analysing fragile objects like the kaftan. Imaging techniques could possibly provide crucial information, as to better understand the shape and construction of the object. It might provide answers to questions like: How many layers of fabric are there? What shape will appear if we could detect all the individual parts and virtually ‘remove’ creases and folds? What connecting techniques have been used to assemble the kaftan: seams, hems, darts, inserts, etc.? Can any fastening systems be detected?

Under the most favourable circumstances, the imaging technique could eventually be used to digitally reconstruct the complete artefact and tell us more about its overall original ‘uncreased’ shape and appearance. If successful, the images that are produced can be used for documentation and presentation of the objects. The images can also play a role in further art historical research.

Imaging techniques are a field of technology that is widely used especially in engineering, architecture and medical science.All of these techniques are developing rapidly, finding a growing range of applications. At the basis of this rapid development is specialised research in the fields of

Bruijne and Telleman 2016a: 1.

3

See section 2.2.: The use of imaging techniques in cultural heritage.

(12)

physics, maths and IT. One of the major challenges in this thesis is to find a mutual basis of communication with specialists in relevant fields of imaging techniques.

1.3. Research question

The central interest of this thesis is to identify imaging techniques that are suitable for the collection of data of fragile textile objects in the field of textile conservation and restoration. This means that a number of imaging techniques that look promising for this purpose must be studied and described focussing on the deliverables of these techniques. The description must furthermore give information about the applicability of the imaging technique in the field of textile conservation and restoration. Therefore three criteria will be regarded: effectiveness, user-friendliness and efficiency.

This thesis is a first reconnaissance in a field of science — imaging techniques — that is not familiar to most professionals in textile conservation and restoration. The results of this thesis may deliver a modest contribution to a new field of interdisciplinary research in conservation & restoration. The experience in related fields of conservation have shown that successful interdisciplinary research demands intense and long-lasting interaction between various fields of expertise. Therefore an interest of this thesis regards the organisation of the interaction between the textile conservator-restorer and the scientists that are competent in the various fields of imaging techniques. It is essential to assess how much knowledge about the non-related scientific field of imaging is required of the conservator-restorer in order to act as a professional user.

(13)

2. Current scientific knowledge

This chapter aims to give an overview of what has been written so far on the use of imaging techniques in textile conservation and the subjects related to this topic. It also discusses how this thesis can draw from this already existing information and where it can add knowledge to the research into BZN17 and into imaging techniques in textile conservation.

The first paragraph will discuss what research already has been done on the kaftan and how this can be useful for this thesis. In the second paragraph, this is followed by an inventory of what has been written about the use of imaging techniques in the field of cultural heritage. After this, the use of imaging in textile conservation will be touched upon. Finally, the position of this thesis within the current scientific knowledge will be discussed.

2.1. Previous research on the kaftan

In two ways, research has been done into the kaftan to get a better understanding of the object. However, so far, the information available on the object is still far from complete.

Firstly, a first attempt has been made to interpret the object from an art historical point of view. A quick identification and description of the object and its condition were made by two conservators in training, S. Telleman and M. De Bruijne, during the first inventory of the shipwreck found. Later, unpublished research has been done into the art historical aspects of the object by 5

researchers from the School of Historical Dress in London. Visual examination of the fragments 6

was done, as well as taking measurements and determining the interrelationship between the fragments, to establish a possible provenance of the object. This is still ongoing research. Although not fully developed yet, the hypotheses made during this examination are valuable for this thesis because they can add to the object description and they can be of help when it comes to the question how this group of fragments should be appreciated. Besides this, the art historical information also may be relevant for future research and for interpretation of the output data. Art historical background information could simplify this task and could help to make a final 3D reconstruction of the object.

Bruijne and Telleman 2016b.

5

A presentation on their latest research findings was given on 9 February 2016 by Jenny Tiramani, Melanie Braun and

6

Claire Thornton, in Amsterdam. On 9 May 2017 a conversation between the author and Jenny Tiramani took place, in which more information on this research was shared.

(14)

Secondly, research into the material side of the object was conducted. This technical 7

analysis was done by Ineke Joosten and later by Ana Serrano, both researchers from the RCE (Rijksdienst voor het Cultureel Erfgoed). SEM-EDX and HPLC analysis add to the current knowledge by providing information about the used dyestuffs and mordants, the identification of the metals present and the condition of the fibres. The results from this analysis will be discussed more elaborately in chapter 4. The knowledge about the elements present (detected with SEM) is valuable for this thesis in the sense that for some techniques it is important to know what materials are present, since this can be a determining factor in the choice of imaging techniques. In addition, knowledge about the condition of the fibres and the metal components present are crucial in determining the overall condition of the object and therefore the extent to which it is safe to handle this object. HPLC analysis may become useful in a later phase of the research, when an actual 3D reconstruction of the object is made and one wants to know what it looked like (in terms of colour) before it disappeared into the sea.

2.2. The use of imaging techniques in cultural heritage

In recent years, the topic of imaging techniques is much-discussed within the field of cultural heritage. More and more research is conducted into the application of such techniques to cultural materials as the development of these techniques in other scientific fields, such as engineering and the medical world, goes very fast. A lot of research has been done into the various techniques and many case-study based articles have been published. Because of the multitude of publications, it is impossible to discuss them all here. Moreover, the speed with which these techniques progress makes many techniques outdated within a few years.

Next to case studies, a couple of books have been published that attempt to give a broader overview of the most widely used techniques in cultural heritage imaging. The book 3D Recording, Documentation and Management of Cultural Heritage is the most recent publication on this subject. It discusses in great detail different ways of documenting and recording cultural heritage. 8

By listing the latest innovations and giving an extensive treatment of the basics of photography and modelling they provide a helpful handbook to professionals in the field of cultural heritage. Although at first sight this publication might seem useful, on closer inspection it becomes clear that

This analysis was not conducted specifically for the research described in this thesis. Serrano 2016abcd; Bruijne and

7

Telleman 2016b.

Stylianidis and Remondino 2016.

(15)

it mainly focusses on immobile cultural heritage (buildings, landscapes, historic sites, etc.) and that only techniques that register the surface of objects are included.

A publication that partly fills this gap is the 2005 publication Radiography of Cultural Material. This book discusses the radiography of cultural artefacts, with chapters dedicated to 9 metals, ceramics, paper, paintings and human/animal remains. Unfortunately, textiles are not included in this overview. Apart from a useful introductory chapter on the theory of X-rays and its practical application on cultural artefacts in general, this book does not offer much to those interested in textiles. This publication makes clear that research into imaging techniques applied to textile conservation and reservation is lagging behind the use of these techniques in other fields of cultural heritage. 10

As will be substantiated in chapter 5, penetrative imaging techniques seem most suitable for the objectives of this thesis. Therefore, scientific publications, dedicated to these types of imaging techniques, although not directly related to textiles, may be of interest. Especially research that deals with artefacts that are made of materials similar to textile (e.g. in composition and density) or that include textiles in another way, may be of interest.

One of the case-study based articles that focuses on one technique, that is useful for this thesis, is the article written by Seales et al. about the virtual unwrapping of a parchment scroll. 11

This research builds upon a lot of research that has been done into scanning other scrolls. The 12

article describes how a multidisciplinary team of researchers has managed to read a text written on a fragile scroll without physically unrolling it, using a micro CT scanner and a sequence of computational transformations. Their problem looks similar to the challenge that this thesis is confronted with regarding the kaftan. The kaftan, on one side, seems to pose more difficulties since the shape is (partly) unknown and it consists of different materials, whereas the paper scroll has a predictable shape and is made of one material, that ‘only’ needs to be unrolled. However, they have the additional problem of reading the text that is written on the scroll.

Another field that provides a lot of experience with imaging techniques is Egyptology. There, mainly X-ray computed tomography is used to detect the internal anatomic features of

Lang and Middleton 2005.

9

A box text in this book written by O’Connor says: ‘(…) the textiles incorporated in them [PC: = cultural artefacts] are

10

often considered as an incidental component or even as a hindrance to gaining a clear image of the intended subject’ and ‘(…) textiles have generally not been the primary subjects of radiographic studies.’ At the time of writing, research into the radiography of textiles had just started. In: Lang, Middleton 2005: 10, 11.

Seales, Parker, Segal, Tov, Shor and Porath 2016: 1-9.

11

Among others: Seales and Lin: 2004; Lin and Seales 2005; Baumann, Porter and Seales 2008; Seales, Griffioen and

12

Jacobs 2010; Seales, Griffioen and Baumann 2011; Allegra, Ciliberto, Milotta, Petrillo, Stanco and Trombatore 2015; Mocella, Brun, Ferrero and Delattre 2015.

(16)

mummies. Most of these publications mainly focus on the human/animal remains inside and not the fabric that is wrapped around it. In an article by Hoffman also the linen around the body is reconstructed. 13

Apart from X-ray computed tomography, there is one other penetrative technique used in archaeology/Egyptology that may be interesting, called Terahertz (THz) imaging. Jackson et al. and Cosentino describe the application of THz imaging in cultural heritage science. Both of them 14

shortly discuss its use on objects made of textile, but come to the conclusion that research into this technique is still in its infancy. An article by Fukanaga et al. discusses the use of THz imaging on the linen wrappings around a mummy. In this article they seem to be able to make the distinction 15

between the several layers of fabric.

A last penetrative technique is neutron imaging. Among others, Mongy, Dinca and Mandescu, and Mannes et al. describe the use of this technique on cultural artefacts. Neutron 16

tomography proves to be a very effective technique for those objects. Research into the neutron imaging of textile objects has not been carried out yet.

2.3. The use of imaging techniques in the domain of textiles

So far, little has been written about the application of imaging techniques on cultural materials made of textile, and therefore their use in textile conservation. An overview is not available, only a few case study-based articles discussing one technique have been published. The one technique that has a considerable amount of literature dedicated to it in the field of textile conservation is X-radiography. Most of the publications on this topic are written by Sonia O’Connor and Mary M. Brooks. The most complete overview of the application of X-rays on textile objects can be found in the book X-Radiography of textiles, dress and related objects. In this book, Brooks and O’Connor 17 discuss what the technique encompasses and in which cases it can be proven useful. Other articles written by them elaborate on this research. This offers a lot of general knowledge on the working 18

of X-rays as well as information on the practical application in the form of methods, X-radiography set-ups, exposure settings and data interpretation. This publication is a must-read for those interested in gaining information from two-dimensional images of textile objects. O’Connor and

Hoffman, Torres and Ernst 2002.

13

Cosentino 2016; Jackson, Bowen, Walker, Labaune, Mourou, Menu and Fukunaga 2011.

14

Fukanaga, Cortes, Cosentino, Stünkel, Leona, Duling III and Mininberg 2011.

15

Mongy 2014; Dinca and Mandescu 2015; Mannes, Schmid, Frey, Schmidt-Ott and Lehmann 2015.

16

O’Connor and Brooks 2007.

17

For example: O’Connor and Brooks 2008; Brooks and O’Connor 2012.

(17)

Brooks barely touch upon the subject of computed X-radiography that provides three-dimensional images. This is unfortunate since this is the focus of this thesis, but since the underlying principle, i.e. x-rays, is the same for both techniques, this publication is still considered valuable in terms of the interaction of the technique with textile.

An article that also deals with techniques used for the imaging of textiles is State of the art of 3D scanning systems and inspection of textile surfaces. It gives an overview of a range of 19

imaging techniques that are widely used in the textile industry. These techniques are mostly applicable to weaving, printing and quality control in industrial textile production. They give, among others, vital information about the surface of the material. Since all the techniques are not penetrative this publication is of little use for this thesis. However, many of these techniques can prove to be very useful in the field of textile conservation and restoration (see chapter 5).

2.4. The position of this thesis

In conclusion, much has been written about the use of imaging techniques in cultural heritage, but most of this literature does not focus on objects made of textile. This thesis may therefore be considered a first reconnaissance into the application of imaging techniques in the field of textile conservation, mainly focussing on 3D imaging techniques. The major aim of this thesis is to identify imaging techniques that have the potential to give textile conservators more information about layers and constructions in fragile textile samples that do not allow any handling. If this effort is successful, these techniques could in the near future be of use for research into the Texel collection, and more specifically for the kaftan that is part of this collection.

Although one of the initial ambitions was to contribute to the research into the Texel collection by making a 3D reconstruction of the original shape of the kaftan, it is not possible to do this within the time given to write this thesis.

Montilla, Orjuela-Vargas and Philips 2014.

(18)

3. Methodology

3.1. General methodology

In this thesis, a method is needed to be able to cooperate systematically with scientists of other fields of science (e.g. of mathematics, computer science and physics). At the beginning of such a cooperation between two non-related fields of science and technology there is always the problem of the lack of understanding of the other scientific field. In such a case an interface (a link between the two systems) will be needed. This interface will help to understand and share methods, vocabulary and practical problems. Therefore, in this thesis the focus is on:

- The input side: the articulation of the conservation problem and the description of the object, which the professional of the imaging technique needs to understand.

- The output side: the interpretation of the data that is generated in the imaging process. To assess the potential of the imaging techniques three criteria are used:

1) Effectiveness. How large is the contribution of the imaging technique to the understanding of the object? Is the object (and its condition) suitable for analysis with the imaging technique?

2) Efficiency. How time-consuming and complex are preparation, handling and imaging of the object and interpretation of the data?

3) User-friendliness. How much understanding and insight in the procedures and working of the imaging technique is needed for the conservator/restorer to fulfill its role at the input and output side? How accessible is this technique for the conservator-restorer?

A topic list, based on these three criteria, was developed for the dialogues with the experts on imaging techniques, see the topic list in Appendix I.

3.2. Selection of the research object: the kaftan

It was decided to take the kaftan as a test case in this thesis, because it is one of the objects of BZN17 that probably holds a lot of information that can only be extracted with the use of advanced imaging techniques. The complexity of the fragments (multiple layers and materials) and their condition make it possible to evaluate the techniques not only on their deliverables, but also on the practical problems that arise when using these techniques. Fragment 060 was chosen

(19)

for testing, because this is the smallest and therefore most manageable fragment. Nevertheless, it still can be considered representative for the whole group in terms of complexity (multiple layers and combination of materials) and condition (fragile and presence of creases, folds and holes). The research object will be analysed and described in order to be able to draw conclusions about handling and about the need for new techniques to detect missing information.

3.3. Phases of the research

1. The first step in this research was to select, analyse and describe the research object. See chapter 4: Object description.

2. The second step was to establish a brief overview of relevant techniques. Chapter 5: 20

Selecting imaging techniques for the kaftan gives an account of the overview that was made. Important is the distinction between techniques that scan the surface of an object and penetrative techniques that can reveal information about deeper layers.

3. A limited number of techniques was selected for further investigation. The selection was based on the three criteria mentioned in 3.1. The selected techniques belong to two groups of techniques: Neutron imaging and X-radiography. In chapter 6 and 7 the results of the research of these techniques are presented.

3.4. Investigation of the selected imaging techniques

During the preparation and execution of the actual experiments the following questions were answered:

- What are the basic technical/scientific principles of the selected techniques? - Which deliverables may be expected from the selected techniques?

o Is it possible to apply software tools on this data in order to produce a 2D or 3D reconstruction that can be manipulated (unwrapping, unfolding, stretching, etc.)? - What are the standard operating procedures of the selected techniques? What should be

the role of the conservator-restorer in this procedure?

o Interface on the input side: What preparations must the conservator-restorer make to be able to deliver the input data which the colleague in the non-related

scientific field needs to perform the selected techniques?

In this phase of the research an outline of the research or flyer was developed by the author to send to the experts of

20

(20)

o Interface on the output side: What knowledge does the conservator-restorer need in order to evaluate the acquired data in a dialogue with the colleague in the non-related scientific field?

- Which contribution to the field of conservation and restoration can be expected from the application of the selected techniques?

(21)

4. Object description

In this chapter the group of fragments that was used as a point of departure for the research question formulated in this thesis is described. First, the kaftan is discussed as a whole to be able to place the test fragment in context and to point out the problems that arise when one wants to image larger and more complex objects. After this, fragment BZN17-6263-060, that will be used as a test item, is described in more detail, in order to be able to judge what the imaging techniques can add to the understanding of the fragment. Also the condition of the fragment is discussed, in order to make clear what the implications of the condition for the handling and mounting of the object are.

4.1. Description of the kaftan

The object that will be used as a test case in this thesis is one of a group of textile fragments excavated from the shipwreck BZN17. This kaftan consists of six fragments, each varying in size and shape (see Table 1-6). In March 2016, during the first inventory of all the textile fragments found in BZN17, conservators in training, Sjoukje Telleman and Marijke de Bruijne established that the largest fragment and three smaller ones belong to each other. The two other fragments were 21

later added to the collection and were then also identified as being part of the same group of textiles. The fragments were identified as belonging to each other on the basis of the similarities of the characteristics of the fabrics. These characteristics will be discussed later in this chapter.

Together, the fragments are likely to be part of one garment. However, the condition of the fragments makes it hard to tell whether a complete garment can be composed of these fragments or if there are still pieces of a substantial size missing, that were never excavated. In addition, it is very likely that little snippets of a few centimetres of the same fabric are still somewhere amongst the other textile fragments excavated from BZN17. At this moment at least one snippet of a couple of centimetres large is kept together with the fragments of this group. It also still needs to be considered a possibility that the fragments of this group — perhaps complemented with non-excavated fragments — together make up more than one garment.

Bruijne and Telleman, Conditiebeschrijving 2016.

(22)

Table 1 BZN17-6263-101

Short

characterisation Largest fragment of the six (Fig. 2 and 11). Consists of multiple pieces of fabric attached to each other. Consists of two layers of base fabric: back without seams with

folded edges of 0.5 cm along the upper, lower and side edges. Along the lower edge, a silk facing of 4.5 cm is present. Attached are two front pieces that are both made up of two parts. An insert is attached to the front pieces along the neckline. Attached to the main front pieces are two triangular parts. Along the front centre opening there are facings present as well. Loops are present on the left and right front part along the front centre opening. On the right front part remnants of (possibly) metal buttons can be seen. The loops and buttons are most likely part of a fastening system.

Dimensions Height from shoulder to hem: 160 cm. Shoulder width: 46 cm. Insert at neckline: 27 x

10.5 cm. Image

(23)

Table 2 BZN17-6263-019

Short

charachterisation Ragged, rectangular piece with hems of 0.5 cm along the edges (Fig. 3). Remnants of silk facing are present. There is a sharp fold along the length across the middle of the

fragment. Has a lot of creases. A black imprint, probably caused by another object, is present on this fragment. Probably a loose sleeve, with a so-called dog ear shape (Fig. 11).

Dimensions Approximately 51 x 23 cm.

Image

!

Fig. 3. Fragment BZN17-6263-019. Photo: Margareta Svensson, 2017.

Table 3 BZN17-6263-060, test object

Short

characterisation Square piece that is folded double (Fig. 4). With hems of about 0.5 cm wide along at least two edges. There are remnants of silk facing present on the inside. Probably the

right sleeve (Fig. 11).

Image

(24)

Table 4 BZN17-6263-082

Short

characterisation Triangular piece consisting of two layers of the base fabric attached to each other along the longest side (Fig. 5). 0.5 m wide hems are present along three sides and a

silk facing of 4.5 cm wide can be found on the inside. Probably a fragment of the right triangular side part (Fig. 11).

Image

!

Table 5 BZN17-6263-123

Short

characterisation Single-layered, rectangular fragment that consists of two separate pieces (Fig. 6). Has two folded edges of 0.5 cm wide. Probably a fragment of the left triangular side

part (Fig. 11).

Dimensions Approximately 65 x 23 cm ( 14 x 23 and 51 x 23 cm).

Image

!

(25)

4.1.1. Materials

• The pieces of textile from the group discussed here, are all made of a silk base fabric. The fibres have been identified by SEM-EDX analysis that was carried out by Ineke Joosten from the RCE on 4 March 2016. For this analysis a sample was taken from BZN17-6263-019. SEM-EDX 22

analysis was carried out a second time, by Ana Serrano from the RCE in September 2016. 23

Samples were taken from the four fragments present at the Ateliergebouw in Amsterdam at that time (BZN17-6263–019, -082, -060, -101). The results from this analysis confirm that the textile fibres are silk (the main constituents of silk were detected: sulphur, carbon, nitrogen and oxygen). • In the silk base fabric also metal strips are present. These strips are probably made of silver.

However, silver was not the only element detected with SEM-analysis. Both silver and sulphur, 24

suggesting silver sulphide (the corrosion product of silver), were found in almost all samples. Also small peaks of gold and occasionally copper and iron were detected in certain areas. This suggests that the metal strips were gilded silver.

Table 6 BZN17-6263-136

Short

characterisation Single-layered triangular fragment with folded edges of 0.5 cm wide (Fig. 7). Probably a fragment of the left triangular side part (Fig. 11).

Dimensions Approximately 40 x 37 x 18 cm.

Image

!

Ibidem.

22

Serrano 2016abcd: n. pag.

23

Ibidem.

(26)

• Attached to this base fabric are remnants of facings present in all objects. These facings are also made of silk. The sewing thread that was analysed using SEM was also identified as being silk. 25

4.1.2. Technique

• All the fragments are made of the same base fabric. This fabric has a complex weave. Metal strips were added in the weft direction (Fig. 8 and 9). It can therefore be considered a silver (or gold, see 4.1.1. Materials) brocade. 26

• Some of the fragments have remnants of a facing fabric. This fabric is a very fine, plain weave 27

silk that is thinner than the base fabric. The facings are bias cut.

• The fabric features a floral pattern. This is created by extra weft yarns that float over one or multiple warp yarns, thus giving a flushing effect. 28

• Construction: The fabric along the edges of the fragments that once were connected to each other with seams, is folded (seam allowance). The two sides of the seam are folded into opposite direction (so, the edge of one side of the seam is folded to the back, the other edge is folded to the front). This implies that these pattern pieces were connected using a type of flat-felled seams. With this type of seam, the edges of the fabric are encapsulated and cannot unravel.

Ibidem.

25

Diehl, De Graaf and De Jonghe 1991: 120.

26

Upon first inspection, it was thought that this second type of fabric was a lining fabric. However, later it was

27

established that this fabric was only present along the edges. It therefore can be considered a facing fabric (fabric on the inside of a garment that is used to finish the fabric edges).

Diehl, De Graaf and De Jonghe 1991: 22.

28

Fig. 8. Detail of fragment 101. Clearly visible are the metal

strips (silver) that run in vertical direction, that are still very intact in this area.

Fig. 9. Detail of fragment 060. The

metal strips run in vertical direction. They have turned black in this area.

(27)

4.1.3. Colour

• At the moment, the fragments have a brown/salmon colour. This colour is not homogeneous throughout the fragments; some areas are lighter or darker. Some areas appear to have a darker coloured fabric, but this is caused by the metal strips that are severely corroded and have blackened, and the presence of dirt.

• HPLC analysis was done by Ana Serrano in September 2016. Samples were taken from 29

BZN17-6263-101, -019, -060 and -082. Primarily, small amounts of cochineal were detected. In the sample taken from 082 also madder was detected. It is possible that a combination of these dyes was used for this kaftan. Both dyes have a red colour, so it is likely that the fragments were redder originally and have discoloured to a more salmon colour over time. However, to be sure of this, more research is required.

• During SEM-EDX analysis that was done by Ana Serrano also aluminium was detected in all samples. The explanation for this is that it comes from an alum mordant. This mordant probably 30

also has had influence on the colour.

4.1.4. Context and relation between fragments The combination of fragments initially was thought to be a dress (worn by a woman) during the first inventory. A second examination done by costume historians of The School of Historical Dress in London, has contradicted this. Based on examination 31

of the shape of the different pieces and the remnants of a fastening system that are present on fragment 101, they have come to the conclusion that those fragments together are more likely to be a kaftan, worn by a man. In Fig. 11 it can be seen how the pieces can be positioned relative to each other to approach the typical shape of a kaftan (see Fig. 10 for an example). This research is still ongoing and therefore the position of the fragments is merely a provisional conclusion.

Serrano 2016abcd.

29

Ibidem.

30

Examination was done by Jenny Tiramani, Melanie Braun and Claire Thornton. Their findings are not (yet)

31

published.

Fig. 10. Example of a kaftan with a similar shape.

This is a nineteenth century silk kaftan from the collection of the Topkapi Saray Museum, Istanbul. Photo: Rogers 1986: 78.

(28)

Fig. 12. Top (a): Front side of fragment 060. D: Tear/hole in the fabric. Bottom (b): Back side

of fragment 060. A, B, C: Areas of loss.

A B D C Fold Fold

(29)

From this figure it becomes clear that still a lot of fabric is missing. At least the left sleeve, the second detachable sleeve and multiple pieces constituting the triangular side parts (some fragment are only one layer, where there should be two) are not present. The figure also shows that, according to their examination, the smallest fragment 060 (the test object) is the right sleeve of the kaftan. This is based on the position of the seams, the size and the shape of the fragment.

4.2. Description of test fragment BZN17-6263-060

4.2.1. Description

The fragment that is used as a test object (BZN17-6263-060) is the smallest fragment of the six and is thought to be the right sleeve of the garment (Fig. 11). It has a rectangular shape and is folded double along one of the short edges (Fig. 12a, b). In this state it is approximately 17 x 15 cm. The outer edges are folded (0.5 cm). Some are folded inwards, the others are folded outwards to work as a flat-felled seam. There are remnants of bias cut silk facing present on the inside. At some places they stick out between the two layers of base fabric.

060 019 136 123 082 101

Fig. 11. Left (a): Collage of the photos of the six fragments in their presumed position. Photos: Margareta Svensson.

Collage: author. Right (b): Schematic drawing of the front of a kaftan. This is possibly what the pattern pieces of the BZN17 kaftan also look like. Drawing: author.

(30)

4.2.2. Condition

This description of the condition will focus only on that part that has consequences for the way the fragment is treated during testing. The condition of this fragment is exemplary for the other fragments belonging to this group. Therefore, the condition of the other fragments is not discussed in detail. In general, the fragment is in a poor condition. There are a lot of areas of loss and the metal is severely corroded.

a. Structure/loss

-

Throughout the entire fragment, there are losses in the form of holes, tears, raveling edges, and whole areas of silk that are missing. These losses are primarily situated along the edges. In Fig. 12a,b it can be seen that the most prominent areas of loss are situated along one of the long edges of the back (A) and along the fold (B) and the surrounding area in both the front and the back of the fragment. The back is more severely damaged there; an area of approximately 2 x 5 cm has disappeared (C). The other edges are almost intact. There is a hole/tear in both the back and the front layer of fabric (D).

-

Although the condition of the object itself does not look very good at first sight and the many losses imply that the fibres have disintegrated, SEM-EDX shows that the condition of the fibres is good (Fig. 13). They are still intact, show flexibility and only occasionally show some dents. 32

These dents might have been caused by microorganisms. The fabric also still feels flexible 33

when touching it.

-

In between the front and back layer of the base fabric, there are remnants of thin silk facing present. This fabric is severely damaged and is in a disintegrated state. No sample was taken from the facing present in 060, but a sample was taken from the weft of the facing of fragment 019. This shows very damaged fibres that are brittle and have a lot of dents (Fig. 14). 34

Regarding the similarity of the condition of the fragments, it is likely that this is the same for the facing silk of 060.

-

The metal strips that are woven through the base fabric are in a very poor condition. The majority of the strips has corroded, which caused the metal to go completely black and break in very small pieces (Fig. 9). Breakage of the strips mainly occurs where the strip goes under one of the silk

Serrano 2016b: n. pag.

32

The SEM images made from the samples taken from the other fragments show a similar condition, however the

33

sample taken from the sewing thread in fragment 101 shows fibres in a poor state with a lot of dents. It is possible that this is also the case in the sewing thread in fragment 060, but no sample was taken from this.

Serrano 2016a: n. pag.

(31)

yarns. Therefore, the little pieces of metal are not properly woven into the fabric anymore and lay loose on top of the fabric. When touching the fragment, some of these pieces fall off. In some parts of the fabric it looks like there are no metal strips left at all. However, there are also some small areas where the metal is in a better condition (i.e. more intact and less blackened) (Fig. 8).

b. Creasing/distortion

-

Throughout the entire fragment, there is a lot of distortion visible. The shape of the entire fragment has changed overtime. Especially around the large area of loss and along the edges, the fabric is crumpled.

-

There are a lot of creases and sharp folds in the fabric.

c. Soiling/staining

-

The fabric has several areas that are darker. This is probably caused by a combination of black corrosion products that lay on top of the fabric and soiling coming from external sources. SEM-EDX analysis shows that besides small contaminations of silver sulphide, also small amounts of sodium, chlorine, silicon, magnesium, phosphorus and calcium are present on the fibres’ surface. These are elements of compounds that can be connected to the sandy and salty 35

conditions in which the fragments were found.

Serrano 2016b: n. pag.

35

Fig. 13. SEM-EDX image of sample taken from

BZN17-6263-060. The majority of the fibres is in a good condition (although there is a lot of

contamination present). Photo: Ana Serrano 2016b: n. pag.

Fig. 14. SEM-EDX image of sample taken from the silk

facing fabric present in BZN17-6263-019. The fibres are in a poor condition; brittle and with a lot of dents. Photo: Ana Serrano 2016a: n. pag.

(32)

4.2.3. Consequences of condition on handling

The before discussed condition has complications for the way this fragment can be treated in terms of handling and mounting. The silk fabric is still quite flexible and does not fall apart immediately when upon touching. Turning it over is possible without damaging the silk too much. Unfolding it and stretching creases however is not recommended, because it could cause breakage of the fibres. The main problem is the metal. When turning, moving or otherwise touching the fragment, some pieces of metal fall off. Unfolding or moving the fragment too much is therefore not possible. Preferably the fragment has to stay flat on a rigid support.

4.2.4. The need for new techniques to obtain missing information

The information discussed above has been obtained by employing simple observation techniques that are used in the everyday practice of a textile conservator. This includes non-invasive methods such as using the senses; look (with visible light), feel, smell and taking measurements. Normal light microscopy is also a method that is employed to identify materials. More advanced technical analysis, such as SEM-EDX and HPLC analysis is often also available for conservators in a research environment. Although for these last techniques a tiny sample is required, the object does not need to be handled a lot for these techniques. However, when one wants to get more detailed information about the construction of an object, the exact shape (both in crumpled and in stretched/ original state), the location of the tears and holes and position of seams and sewing thread that are not visible on the surface, these techniques fall short. This is why different techniques are needed to makes this visible.

(33)

5. Selecting imaging techniques for the kaftan

In the search for techniques that have been proven successful in the imaging of cultural heritage objects, it shows that a wide variety of techniques exists. In general there are two ways to make a division into groups of imaging techniques:

- A division between techniques for two-dimensional and for three-dimensional imaging (or even four-dimensional, i.e. 3D + time)

- A division into techniques that scan the surface of an object — surface techniques — and those that can detect deeper layers beneath the surface: penetrative techniques.

It will be clear that for the analysis of the research object of this thesis — the kaftan — penetrative techniques are most promising since all noninvasive methods that are limited to the scanning of the surface fall short in disclosing essential characteristics of the artefact. This includes even the most sophisticated surface imaging techniques. Nevertheless, surface imaging techniques may contribute considerably to the field of textile conservation and restoration.

5.1. The importance of surface imaging techniques for the field of textile conservation

A lot of research and development — driven by the soft- and hardware industry — has been done into this group of techniques. This development primarily lies in the improvement of the quality of the data in terms of resolution. In general, these techniques are used extensively in the imaging of cultural heritage, since they are widely available and rely on relatively inexpensive and user-friendly equipment. This group of techniques is often further divided into two groups.

• The techniques that make use of a passive sensor. The passive sensor does not emit radiation, but only detects the existing light that is being reflected by the object of examination.

• The group of techniques that make use of an active sensor. The active sensors emit a form of radiation (e.g. light or laser) and a detector detects its interaction with the object of examination.

Of both groups techniques will be highlighted here, illustrated with examples from its application in cultural heritage.

(34)

5.1.1. Passive sensor surface imaging techniques

Photogrammetry is an image-based modelling (IBM) technique. This technique gathers 36

measurements from objects by analysis and processing photographs. The quality of the resulting 3D-model depends on the quality of the images (in terms of

resolution, contrast, sharpness and lighting conditions), the amount of photos taken and the percentage of overlap between them. Due to the rise of consumer friendly apps and software this technique is increasingly used in museums. Photogrammetry offers the possibility to show digital 3D reconstructions of textile objects to the public in a permanent digital exhibition. It therefore has a lot of potential for use in textile conservation since most textile artefacts are very vulnerable, and this way, the objects do not have to go on display. An example of its use on textile objects is the digitisation of a part of the fashion collection of the Santa Cruz Museum of Art and History, California, by Abby Crawford. With the idea to show more of the collection than would be possible with a physical exhibition in the museum, Crawford made 3D models of costumes that were in storage, to make a digital exhibition on Sketchfab.com. The garments were 37

mounted onto a mannequin and photographed from all sides. In Fig. 15 one of the 3D models by Crawford can be seen.

Another example of the use of photogrammetry in textile conservation is an experiment to visualise the strain within

tapestry, using optical markers (optical fibre sensors) on the tapestry in combination with an advanced method of photogrammetry (digital image correlation). 38

5.1.2. Active sensor surface imaging techniques

A large range of techniques exists, that work with an active sensor. Best known is laser scanning. This technique is based on the principle of range-based modelling (RBM). It uses laser light to

Remondino and El-Hakim 2006: 270.

36

‘Santa Cruz Museum of Art and History.’ <https://sketchfab.com/santacruzmah>.

37

Khennouf, Dulieu-Barton, Chambers, Lennard, Eastop 2010.

38

Fig. 15. Screenshot of one of the

3D models made by Crawford for the Santa Cruz Museum of Art and History, California, USA. Photo: ‘Blue and white floral dress.’ 2015. <https://sketchfab.com/models/ 1d7d0681a2f4462788a8a343f2667 1d9>.

(35)

measure 3D coordinates of points on surfaces in a regular pattern. The time it takes for the laser 39

to travel and the orientation of the laser beam is measured. The result of this is a so-called point cloud. There are several laser imaging techniques that produce 2D and 3D images. 3D scanners 40

are widely used in cultural heritage to capture the features of artefacts. Another type of scanning is laser triangulation. This technique enables one to measure the irregularities of the surface. It 41

is therefore a widely applied in the textile industry. This could also be interesting for measuring the surface of textile artefacts.

One of the techniques that works with this triangulation principle is structured light scanning. A pattern of light is projected on an object and the deformation of the lines is then measured. This technique is used in cultural heritage for conservation and restoration purposes. 42

It was for instance used to make a three dimensional model of the damaged statue of The Thinker by Rodin, to compare the damaged cast with other casts of this statue, thus giving valuable clues for the restoration. Structured light could also be of interest for textile conservation and 43

restoration since this technique is especially appropriate to scan flexible or delicate objects.

5.1.3. Relevance of surface imaging techniques

Surface imaging techniques have the obvious disadvantage, that they cannot give information about layers beyond the surface. Since these techniques are widely used in the entertainment industry and in visual art, the quality of the images is improving fast. This makes surface imaging techniques interesting in a later phase of the research on the kaftan, since they might provide additional information on the surface of the test object (e.g. colour) that can be used in combination with the penetrative techniques to make an even more complete 3D reconstruction.

5.2. Choice for penetrative techniques

As is substantiated in section 5.1, the surface techniques do not have the properties that are required from an imaging technique for the goal described in this thesis. In areas where the fabric is distorted and parts of the fabric are not visible because of severe creases or sharp folds, a technique is needed that can look through the top layer(s) of fabric to see how the fabric that is obscured by this top

Stylianidis and Remondino 2016: 305.

39

Stylianidis and Remondino 2016: 306.

40

Montilla, Orjuela-Vargas and Philips 2014: 2, 3.

41

Beentjes and Van der Molen 2013: 149, 150.

42

Beentjes and Van der Molen 2013.

(36)

layer is arranged in the fold. Only this way it is possible to detect all the material that is present and a reconstruction of all the fabric that is present can be made.

Preliminary literature research (see chapter 2) has identified three techniques that have proved themselves and are used in the field of cultural heritage and conservation for similar purposes to the one in this thesis. These techniques are X-radiography, neutron tomography and Terahertz imaging. These techniques are all able to look through multiple layers of material and are all non-invasive and non-destructive.

X-radiography is the most well-known technique amongst conservators of all specialisations and therefore also the most widely used. The technique is divided in ‘traditional’ 2D X-radiography and X-ray computed tomography. As described in chapter 2, already quite some successful research has been done into 2D X-radiography on textiles by Brooks and O’Connor, and it has proved to be a useful technique in imaging a parchment scroll that was later virtually unrolled. 44

Neutron tomography is a technique similar to X-ray computed tomography in the sense that it produces output data similar to that of X-ray computed tomography. The underlying physical principle, however, is different and more complex. This also makes the technique a lot less easily accessible to conservators and therefore is also less well known in the field of conservation of cultural heritage. Nevertheless, neutron tomography has been used successfully on cultural artefacts and is often used complementary to X-ray computed tomography. 45

Terahertz imaging is also a technique that most conservators are not familiar with. This technique uses Terahertz pulses (electromagnetic radiation that is lower in energy than for example X-rays) to look through layers of material. Although Terahertz has been used to distinguish layers of textile and seems promising in the field of conservation of textiles, research into Terahertz is relatively new and is still in its infancy. It was therefore decided not to further examine this 46

technique in this thesis.

5.3. Selection of two techniques for further research

Based on these findings, it was decided to select two techniques, X-radiography and neutron tomography, for a more in-depth research. The next challenge was to find institutions in the Netherlands that are renowned for expertise and equipment in these fields of technology. For neutron tomography, the Reactor Institute (RID) of TU Delft (Delft University of Technology) was

O’Connor and Brooks 2007; Seales, Parker, Segal, Tov, Shor and Porath 2016.

44

Mongy 2014; Dinca and Mandescu 2015; Mannes, Schmid, Frey, Schmidt-Ott and Lehmann 2015.

45

Fukanaga, Cortes, Cosentino, Stünkel, Leona, Duling III and Mininberg 2011.

(37)

approached, since this is the only neutron imaging facility that exists in The Netherlands. Tests with 2D X-radiography were executed at the Rijksmuseum Amsterdam (RMA). For X-ray computed tomography the Centrum voor Wiskunde & Informatica (CWI, National Research Institute for Mathematics and Computer Science) was approached, because they installed a new state-of-the-art CT scanner at the time of doing preliminary investigation for this thesis. The scientists of these 47

three institutions — RID, RMA and CWI — showed interest to contribute to this thesis and thereby enabled the experiments that are described in the next two chapters.

The Academisch Medisch Centrum (AMC, Academic Medical Centre, UvA) was also approached to do tests with a

47

medical CT scanner. Although they were interested in collaborating, in practice it turned out that this facility was hardly accessible for the experiments, since there always is a priority for the patients.

Travelling through textiles: The exploration of fragile textiles through 2D and 3D imaging.