SPECTROSCOPIC STUDIES of COPPER-BASED PIGMENTS By

SPECTROSCOPIC STUDIES

COPPER-BASED PIGMENTS

Marcie B. Wiggins

A dissertation submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Chemistry & Biochemistry

Spring 2019

SPECTROSCOPIC STUDIES

COPPER-BASED PIGMENTS

Marcie B. Wiggins

Approved: ______Brian Bahnson, Ph.D. Chair of the Department of Chemistry and Biochemisty

Approved: ______John Pelesko, Ph.D. Interim Dean of the College of Arts and Sciences

Approved: ______Douglas J. Doren, Ph.D. Interim Vice Provost for Graduate & Professional Education

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______Karl S. Booksh, Ph.D. Professor in charge of dissertation

Signed: ______Jocelyn Alcántara-García, Ph.D. Professor in charge of dissertation

Signed: ______Steven D. Brown, Ph.D. Member of dissertation committee

Signed: ______Cecil Dybowski, Ph.D. Member of dissertation committee

Signed: ______Robert Opila, Ph.D. Member of dissertation committee ACKNOWLEDGMENTS

I have had many great teachers and advisors who have guided me and encouraged me on the way. Mr. Cowder, my high school chemistry teacher, made the material in the classroom and in the laboratory finally connect. Dr. Eric Breitung, my undergraduate research advisor, gave me my first true research experience and taught how to take ownership of my projects. Dr. Meredith Gill, my undergraduate art history advisor, supported my dual interest in art history and chemistry and helped me to combine them. I would like to thank my graduate co-advisors, Dr. Jocelyn Alcántara- García and Dr. Karl Booksh. Dr. Alcántara-García continues to push me so I can improve in every phase of research as an excellent advisor and friend. Dr. Booksh has allowed me to pursue and advance my interests, and he has never shied away from a new project or challenge. I would also like to thank my committee members for their guidance and feedback during my studies. I want to thank the UD’s Office for Graduate and Professional Education and the UD’s Department of Chemistry and Biochemistry for financial support of my research. Thanks needs to be given wholeheartedly to all of the Booksh research group

(past and present), including its honorary members Zachary Voras and Chris Goodwin. Dr. Josh Ottaway, Dr. Zachary Voras, Dr. Anna Murphy, and Chris Goodwin have been terrific mentors and friends during the process, even if most conversations stray to cover any topic under the sun (and beyond!). I want to thank the faculty and staff in the UD’s Department of Art Conservation and Winterthur Museum, Garden and Library for their support. I especially thank Winterthur’s

iv Scientific Research and Analysis Laboratory staff for the opportunities they gave me and the knowledge they shared. Catherine Matsen, Dr. Rosie Grayburn, and Dr. Judy Rudolph gave endless help and were great role models. Portions of this work would not be possible without the contributions of my undergraduate students, Lizzy van Winkle and Emma Heath, who have also made me a better teacher. My friends and family have provided never-ending reassurance, comfort, and motivation during this period. I give special thanks to my parents, Mike and Margie Wiggins, for continuing to support me despite the fact I am “always going to be in school,” and I hope to continue to make them proud. There are no words to thank Stephen Christian enough for always supporting and encouraging me during every high and low, whether it is in the form of keeping me on schedule or taking a holiday. However, his most important task, by far, was distracting the cat, Boe, long enough for me to write this dissertation. Boe has provided an endless supply of smiles (and some tears…) for entertainment and distraction alike, but, honestly, he was quite jealous and angry over the amount of attention I paid this thesis instead of him.

v TABLE OF CONTENTS

LIST OF TABLES ...... x LIST OF FIGURES ...... xi ABSTRACT ...... xv

Chapter

1 INTRODUCTION ...... 1

1.1 Motivation ...... 1 1.2 Overview of Dissertation ...... 3 1.3 Importance ...... 6

REFERENCES ...... 8

2 REVIEW OF COPPER SALTS IN CULTURAL HERITAGE ...... 12

2.1 Introduction ...... 12 2.2 Use in Cultural Heritage ...... 12

2.2.1 Paintings ...... 13 2.2.2 Paper ...... 14

2.3 Preparation and Characterization ...... 14

2.3.1 Verdigris (Copper Acetates) ...... 14 2.3.2 Basic Copper Chlorides ...... 16 2.3.3 Characterization Methods ...... 17

2.4 Alteration and Treatment Studies ...... 18

2.4.1 In Paintings ...... 19 2.4.2 On Paper ...... 20

2.5 Conclusion ...... 22

REFERENCES ...... 24

vi 3 EXPERIMENTAL OVERVIEW ...... 35

3.1 Introduction ...... 35 3.2 Instrumentation ...... 35

3.2.1 X-ray Fluorescence ...... 35 3.2.2 Scanning Electron Microscopy-Energy Dispersive X-ray Spectroscopy ...... 36 3.2.3 X-ray Diffraction ...... 37 3.2.4 X-ray Photoelectron Spectroscopy ...... 40 3.2.5 Time of Flight-Secondary Ion Mass Spectrometry ...... 42 3.2.6 Fourier-Transform Infrared Spectroscopy ...... 45

3.2.6.1 Transmittance FTIR ...... 45 3.2.6.2 Attenuated Total Reflectance FTIR ...... 46

3.2.7 Raman Spectroscopy ...... 46

3.2.7.1 Multivariate Curve Resolution-Alternating Least Squares ...... 49

REFERENCES ...... 51

4 CHARACTERIZATION OF GREEN PAINTS IN MING & QIAN DYNASTIES LIN’XI PAVILION ...... 53

4.1 Introductory Remarks ...... 53 4.2 Introduction ...... 53 4.3 Materials ...... 56 4.4 Methodology ...... 59

4.4.1 Visible and Ultraviolet Microscopy ...... 59 4.4.2 Polarized Light Microscopy (PLM) ...... 59 4.4.3 Scanning Electron Microscopy-Energy Dispersive X-ray Spectroscopy (SEM-EDX) ...... 60 4.4.4 Time of Flight-Secondary Ion Mass Spectrometry ...... 60 4.4.5 X-ray Diffraction (XRD) ...... 61 4.4.6 Raman Spectroscopy ...... 62 4.4.7 Fourier-Transform Infrared Spectroscopy ...... 62

4.5 Results ...... 62 4.6 Discussion ...... 68

4.6.1 Organics Materials ...... 68

vii 4.6.2 Inorganic Materials ...... 69

4.7 Conclusions ...... 71 4.8 Acknowledgements ...... 72 4.9 Concluding Remarks ...... 72

REFERENCES ...... 74

5 POLYMORPH IDENTIFICATION IN GREEN CHINESE ARCHITECTURAL PAINTS USING RAMAN IMAGING AND MCR- ALS ...... 77

5.1 Introductory Remarks ...... 77 5.2 Introduction ...... 77 5.3 Materials ...... 79 5.4 Methodology ...... 82 5.5 Results ...... 86 5.6 Discussion ...... 91 5.7 Conclusions ...... 94 5.8 Acknowledgements ...... 95 5.9 Concluding Remarks ...... 95

REFERENCES ...... 96

6 MULTI-ANALYTICAL CHARACTERIZATION OF COPPER-BASED PIGMENTS’ PREPARATION AND ALTERATION PRODUCTS ...... 100

6.1 Introductory Remarks ...... 100 6.2 Introduction ...... 100 6.3 Experimental ...... 103

6.3.1 Preparation of Pigments ...... 103 6.3.2 Preparation of Samples ...... 104 6.3.3 Aging Studies ...... 106 6.3.4 Instrumentation ...... 107

6.4 Results & Discussion ...... 108

6.4.1 Characterization of Prepared Pigments ...... 108 6.4.2 Degradation of Pure Powder Pigments and Pigment-Cellulose Mixtures ...... 111 6.4.3 Degradation on Paper ...... 114 6.4.4 Degradation on Paper with Gum Arabic ...... 116

viii 6.5 Conclusion ...... 122 6.6 Acknowledgements ...... 123 6.7 Concluding Remarks ...... 123

REFERENCES ...... 125

7 MULTIDISCIPLINARY LEARNING: REDOX CHEMISTRY AND PIGMENT HISTORY ...... 133

7.1 Introductory Remarks ...... 133 7.2 Introduction ...... 133

7.2.1 Copper-Containing Artists’ Materials ...... 135 7.2.2 Learning Objectives ...... 138

7.3 Experimental Summary ...... 139

7.3.1 Synthesis of Verdigris ...... 139 7.3.2 Application ...... 141 7.3.3 Aging Study ...... 143

7.4 Hazards and Safety ...... 145 7.5 Outcomes and Critical Assessments ...... 145 7.6 Assessment ...... 147 7.7 Limitations ...... 148 7.8 Acknowledgements ...... 149 7.9 Concluding Remarks ...... 149

REFERENCES ...... 151

8 CONCLUSIONS ...... 156

8.1 Summary ...... 156 8.2 Future Directions ...... 158

REFERENCES ...... 160

Appendix

PERMISSION FOR MATERIAL REPRINT ...... 162

ix LIST OF TABLES

Table 5-1 List of paint cross-section samples analyzed in this study. This includes with their locations and time periods, as well as images of the cross- sections...... 81

Table 5-2 The relative ratios of polymorphs in the cross-sections analyzed by comparing the concentrations of components are reported, as well as the number (N) of contributing spectra and the number of factors used to build the MCR model. All of the compound identities were verified with XRD...... 91

Table 6-1 Summary of FTIR, XRD, and XPS key features of pigments prepared according to historical recipes and their corresponding overall characterization...... 109

x LIST OF FIGURES

Figure 3-1 Schematic of x-ray diffraction, in which incident x-rays on a rotating sample result in a diffraction pattern. The plot shows a sum of the intensities of scattered x-rays versus the angle of scattering...... 39

Figure 3-2 a) Diagram of incident x-rays ejecting a core electron and b) schematic of XPS with incident x-rays hitting the sample and emitted electrons being separated in a hemispherical analyzer...... 41

Figure 3-3 a) Schematic of ToF-SIMS with a primary ion gun targeting the sample and secondary ions traveling to a time of flight mass analyzer to be recorded as a spectrum. b) Illustration of primary ion hitting the surface to eject secondary ions and radiating damage into the material. 43

Figure 3-4 Jablonski diagram of the energy transitions for Rayleigh and Raman scattering...... 48

Figure 3-5 Schematic representation of multivariate curve resolution-alternating least square algorithm used for Raman spectroscopic post-processing. . 50

Figure 4-1 Exterior of Lin’xi Pavilion, which is located in the Forbidden City...... 54

Figure 4-2 (a) The Lin’xi Pavilion ceiling painting (297 x 297 cm) prior to treatment. The yellow box indicates the region (b), where samples were taken...... 55

Figure 4-3 Cross-section images of LXT-04 (i) and LXT-05 (iii) in normal light and LXT-04 (ii) and LXT-05 (iv) in ultraviolet light. The layered structure of both the first and second generation is: a) paper support, b) silk substrate, c) green paint, d) gold leaf, and e) white paint...... 58

Figure 4-4 Dispersed sample of the pigments from LXT-04 at 500x magnification in a) plane polarized transmitted light and b) crossed-polarized light. ... 64

Figure 4-5 SEM-EDX SEM-EDX false color elemental maps of elements for LXT- 04 (i and ii) and LXT-05 (iii and iv) with layers for silk (b), green paint (c), and white paint (e)...... 65

Figure 4-6 ToF-SIMS maps of mass fragments for LXT-04 (i) and LXT-05 (ii)...... 66

xi Figure 4-7 XRD patterns of LXT-04 and references botallackite and atacamite...... 67

Figure 4-8 Raman spectra of LXT-04 and LXT-05 compared to reference spectra of atacamite and botallackite...... 68

Figure 5-1 A selection of raw spectra from YHD13 maps are seen in (a) and those same spectra are rubberband baseline corrected in (b). These are compared to the calculated spectra for the components of interest (atacamite and botallackite) generated by MCR-ALS in (c)...... 85

Figure 5-2 MCR-ALS generated atacamite (a) and botallackite (b) component spectra from each data set compared to RRUFF database reference spectra (R050098 for atacamite and R070066 for botallackite)...... 87

Figure 5-3 Example of cross-section, YHD13, with the yellow rectangle indicating mapped regions (a). The pure component spectra and its reconstructed contour heat maps of its concentrations for epoxy (b), atacamite (c), and botallackite (d). A normalized overlay of atacamite and botallackite concentrations, which is not contoured (e)...... 89

Figure 6-1 Progress of verdigris synthesis using an historic method: One of the copper plates suspended over vinegar (a) and documented after 6 months (b & c). The neutral verdigris crystals are present in (b) denoted as A1, and basic blue verdigris crystals are present in addition to the green neutral verdigris crystals in (c) denoted as A2...... 103

Figure 6-2 Paper samples treated with commercial (C, top), neutral (N, middle), basic (B, bottom) verdigris applied with 10, 5, 1, and 0% gum Arabic (left to right), which has been aged for 0 (a), 3 (b), and 7 (c) days...... 106

Figure 6-3 XRD patterns of basic, Cu(CH3COO)2Cu(OH)25H2O (left), and neutral, Cu(CH3COO)2H2O (right), verdigris pigments at 0 (black) and 7 (blue and green) days of aging (50ºC, 60% RH). Conversion from both starting forms to Cu(CH3COO)23Cu(OH)22H2O is evidenced in both the absence (blue) and presence (green) of cellulose: new peaks at 9.41° and 18.73° (red lines)...... 112

Figure 6-4 FTIR spectra of basic, Cu(CH3COO)2Cu(OH)25H2O (left), and neutral, Cu(CH3COO)2H2O (right) verdigris pigments. Red lines indicate a shift of the carbonyl bands, associated with aging...... 113

xii Figure 6-5 XRD patterns of basic (Cu(CH3COO)2Cu(OH)25H2O) verdigris pigment applied to Whatman filter paper No. 1 as dye (aqueous solution). Measurements were taken after 0 (black), 3 (blue), and 7 (green) days. Peaks at 35.5° and 38.6° indicate formation of copper (II) oxide (red lines)...... 115

Figure 6-6 XPS spectra of the copper 2p region for basic verdigris (Cu(CH3COO)2Cu(OH)25H2O) mixed with different amounts of gum Arabic prior to degradation. Red lines highlight Cu (I) increasing as gum Arabic concentration increases...... 117

Figure 6-7 XPS spectra of the copper 2p region for neutral verdigris with of 10% gum Arabic during accelerated degradation. The red lines indicate the increase of copper (II) species as the copper is oxidized...... 119

Figure 6-8 XRD patterns of paper with basic (Cu(CH3COO)2Cu(OH)25H2O) verdigris after 7 days of aging. Note that verdigris (peaks at 9.45 and 18.96°) are still present in the sample with 10% gum Arabic but not in the control...... 120

Figure 6-9 A flowchart schematic of the proposed degradation pathway for verdigris pigments with regards to cellulose and gum Arabic. Blue indicated the organics present during that stage of the reaction...... 121

Figure 7-1 Objects showing signs of decay and discoloration due to verdigris’ degradation. Leaves show some green intact verdigris as well as brown discolored verdigris. (Left) Courtesy, Winterthur Museum, Fraktur: Birth and baptismal certificate of Julyanna Biehl by George Peter Deisert, 1792, Adams County, PA, Watercolor, Ink, Laid paper, 2013.31.2.1. (Right) Courtesy, Winterthur Museum, Fraktur: Birth and baptismal certificate of Anna Maria Biehl by George Peter Deisert, 1792, Adams County, PA, Watercolor, Ink, Laid paper, 2013.31.2.2...... 137

Figure 7-2 A copper sheet suspended on plastic cups over 50% (v/v) acetic acid. It was sealed in a glass container for 3 weeks...... 139

Figure 7-3 The copper solution after the addition of ammonia, which prompted the formation of CuSO4•3Cu(OH)2...... 141

Figure 7-4 Copper (I) oxide forms for Fehling’s reaction upon addition of excess heat to the verdigris and gum Arabic mixtures...... 142

xiii Figure 7-5 Whatman filter papers No. 1 aged over 7 days (60°C, 85%RH) matted on white board to observe various verdigris pigments applied as a dye or watercolor: A) no pigment, B) multi-step synthesized verdigris in water (dyed by students), C) commercially available verdigris in water, D) simple synthesized verdigris, E) multi-step synthesized verdigris in water, F) gum Arabic, G) commercially available verdigris in gum Arabic, H) simple synthesized verdigris in gum Arabic, and I) multi-step synthesized verdigris in gum Arabic...... 144

Figure 7-6 FTIR spectra of aged Whatman paper dyed with simple synthesized verdigris, with noted bands at 1316 cm-1 (solid arrow), 1425 cm-1 (small dashed arrow), and 1560 cm-1 (big dashed arrow)...... 145

Figure 7-7 The appearance of the dissolved verdigris changed upon the addition of gum Arabic from a light, semi-insoluble blue crystal (left) to a deep- green solution (right)...... 147

xiv ABSTRACT

Copper-based pigments are common in both Western and Eastern artwork as green and blue colorants, even though they are commonly associated with degraded organic materials as well. Spectroscopic analyses of copper-based compounds provide information on both the origin and manufacturing of these blue and green pigments. Furthermore, they provide insight into the preservation of cultural heritage objects as these pigments interact and degrade with the surrounding organic media. For example, Chinese architectural paints are composed of several different “copper green” pigments. Understanding their application and usage prompted investigation into green pigments across China using samples from the 12th to 19th centuries. Hyperspectral Raman images of the paint layers of several Chinese architectural paint cross-sections, including several from the Forbidden City, Beijing, China, were analyzed with multivariate curve resolution-alternative least squares (MCR-ALS) to distinguish component spectra of copper-based species of interest and their relative ratios. The differences in the relative ratios of copper chloride trihydroxide polymorphs indicate the historical synthesis of these pigments. Through this study, the transition to synthesized pigments for Chinese architectural materials is beginning to be understood, as well as the pigments’ role in the stability of the paint. Furthermore, to complement this investigation into copper-based pigments, accelerated degradation studies were carried out to study the interaction of copper in many historical objects. Common copper-based pigments were prepared following adapted historical recipes and then extensively characterized to identify the starting materials of various

xv artworks. Also, these methodologies were successfully adapted for educational purposes to connect the art and sciences in the classroom as well as the research laboratory. The accelerated degradation of the most common of these pigments, both neutral and basic verdigris (copper acetates), was studied in relation to organic substrates. Through X-ray diffraction, infrared, Raman, and X-ray photoelectron spectroscopies, the alteration products of verdigris on cellulose with and without gum Arabic as a binder, show an intermediate species in the degradation process prior to the final product of copper oxide. Additionally, the changes of copper oxidation state were observed to relate to the amount of gum Arabic added to the samples. With these complementary studies, we better understand the usage and effects of copper-based pigments in both Western and Eastern cultural heritage, and this information can be utilized by conservators preserving the objects in question.

xvi Chapter 1

INTRODUCTION

This thesis consists of spectroscopic studies on copper salts that degrade in artworks. It provides chemical information for understanding the formation, preservation, and conservation of copper-based pigments. Through spectroscopic studies of these copper-based pigments, it is possible to identify them, distinguish their origin, and study their degradation pathway. Several copper-based pigments are common in both Western and Eastern artwork as green and blue colorants; however, they are also associated with degraded organic materials.1-5 Their presence indicates the need for preservation of cultural heritage objects as these pigments interact and degrade with the surrounding organic media.

1.1 Motivation Cultural heritage objects generally are complex materials, containing both inorganic and organic compounds.5-8 These heterogeneous mixtures are subject to many inorganic and organic interactions over their long lifetimes. Conservators, curators, and caretakers of these objects prefer to minimize these reactions to preserve the objects for future generations. Therefore, the inorganic and organic materials cannot be studied independently; the entire environment must be considered when studying cultural heritage objects. The copper salt studies presented in this thesis can be extrapolated to similar materials and other cases, to provide information on several cultural heritage challenges. Specifically, the focus of the following work is copper

1 acetates and basic copper chlorides. Understanding the fundamental processes and reactions of these materials is essential for being able to treat, conserve, and ultimately preserve these materials. Due to the complex materials found in cultural heritage, a range of methodologies and techniques are necessary to address questions, such as those related to the history and origin of the object or to the condition and possible alteration of the object. Spectroscopic techniques are ideal for these studies, because they can be used non-invasively in situ or on carefully preserved samples from the artifacts. Analysis that requires samples, but does not change the sample, can be considered minimally- invasive. For this thesis, primary analysis methods were Raman spectroscopy, Fourier- transform infrared spectroscopy (FTIR), x-ray photoelectron spectroscopy (XPS), and x-ray diffraction (XRD).9-13 Additional techniques were occasionally utilized to complement these results. In this dissertation, I develop methodologies and use spectroscopic techniques to address questions related to copper salts used both in painting and on paper media. Answering these questions can provide insight into the history and origin of the pigments, and these procedures allow predicting and monitoring of their condition and alteration over time. Ultimately, method development and chemical studies are invaluable for the field of cultural heritage research. Because copper-based pigments are very common, they are essential to analyze.4, 14-15 Previously, studies have highlighted how they have been used in cultural heritage, and some explore their possible degradation pathways.5, 7, 16-17 Basic copper chlorides, especially their origins, are not as well-studied, and studies into verdigris’s degradation with organics, specifically paper-related media, are less fundamental and primarily treatment-

2 motivated.18-20 Hopefully, the following studies will contribute scientific knowledge to both these subjects.

1.2 Overview of Dissertation In Chapter 2, I discuss the literature on copper salt pigments in relation to cultural heritage, specifically their usage, identification, and alteration. Copper acetates (verdigris) and basic copper chlorides as the most relevant materials to this thesis. Both pigments are synthesized from copper corrosion products, and they have been used for centuries in painting.4, 15 Verdigris has also been extensively used on paper-based objects as well,2, 12 although basic copper chlorides have not be found on cellulose-based media. Studies into verdigris alteration, degradation, and treatment methods have been summarized in this section to understand their interaction with the organic surroundings better;3, 7, 21 however, there is very little published on time- dependent alteration of basic copper chlorides. While x-ray fluorescence (XRF) is initially helpful in finding these copper salts on objects, normally XRD, FTIR, and/or Raman spectroscopies are necessary to characterize the copper salts in the artifacts.9-13 Chapter 3 explains the experimental procedures of this body of work and details sample preparations and instrumentation for this research. It includes preparation of paint cross-sections from several Chinese paintings, and the additional treatments for use with various instrumental techniques. Furthermore, the preparation of verdigris pigments based on historical recipes is detailed.1, 22-23 Those pigments were then applied to cellulose substrates as suspensions in water and gum Arabic. These samples, as well as individual powdered samples, were subject to artificial aging, or accelerated degradation, to study the degradation mechanisms with regard to the effects of the organic surroundings. The instrumentation used during the course of

3 this work includes XRF, scanning electron microscopy-energy dispersive x-ray spectroscopy (SED-EDX), XRD, XPS, time of flight-secondary ion mass spectrometry (ToF-SIMS), FTIR spectroscopy, and Raman spectroscopy. Subsequent data processing for Raman imaging analysis included multivariate curve resolution- alternate least square (MCR-ALS). The fundamentals and parameters for these methods are briefly presented in Chapter 3. Chapter 4 highlights several of these complementary techniques being utilized to identify the materials used in a Chinese architectural painting. The identification of the materials used in the two paint generations of the ceiling of the Lin’xi Pavilion in the Forbidden City relied on ToF-SIMS, SEM-EDX, XRD and Raman spectroscopy. These showed similar components used in both paint generations, such as a calcium sulfate ground on silk and copper salt pigments, and additional decorations of lead white and gold leaf in the second generation. The most interesting finding from this case study was the use of basic copper chloride polymorphs, atacamite and botallackite, in the different paint generations from the Ming and Qing dynasties.24 Copper carbonate pigments would have been expected for this massive ceiling painting, but these results suggest synthetic basic copper chlorides of the 15-18th centuries might have been more popular than previously believed.

Chapter 5 expands on the questions raised in Chapter 4 regarding the use and manufacturing of basic copper chlorides in Chinese architectural painting. Because botallackite is naturally less stable and less common as an artists’ material, it was interesting to find a predominantly botallackite mixture in the Lin’xi Pavilion.25 This prompted additional investigation into green-blue paints in Chinese architecture to understand the ratios of copper chloride trihydroxide polymorphs better. Raman

4 mapping and MCR-ALS analysis were used to determine relative ratios of the polymorphs non-invasively, while maintaining spatial resolution for the paint cross- sections.26-27 These findings were subsequently related to the synthesis and possible preparations of these pigments for use in Chinese art to the benefit of historians, curators, and conservators.28-29 Chapter 6 focuses on verdigris, copper acetate pigments, and its alteration related to paper-based media. While verdigris is studied more than basic copper chlorides, there are still gaps in the studies of the alteration of the pigment and how the inorganic-organic interactions could affect the objects or guide treatment. This study started by reproducing verdigris and related pigments by historical recipes and then characterizing them to determine common starting compounds of cultural heritage materials.30 Next, to determine the effects of cellulose and gum Arabic (a common binding media), the neutral and basic verdigris were artificially aged with and without organics. Alterations were evaluated using XRD, FTIR, and XPS. These findings revealed that both forms of verdigris convert to an intermediate species with or without organics, and ultimately form the degradation product of copper oxides.9, 31-32 Monosaccharides of gum Arabic reduced the copper ions prior to accelerated degradation, and the slower formation of copper oxide in those samples suggest reduction stabilizes the verdigris species on paper. This evidence may guide future scientists and conservators on treatments for objects suffering from copper-induced degradation. Chapter 7 is a description of how the work in Chapter 6 provides an opportunity to teach students about chemistry and art. The laboratory activity provides a novel introduction to the preparation of verdigris by two methods (simple and multi-

5 step).33-34 After that synthesis, students create samples to analyze on paper. Subsequently, they age the materials and monitor them by instrumental techniques. This lesson teaches such subjects as redox chemistry, stoichiometry, crystal structures, and several instrumental methods. Its versatility allows instructors to adapt the activity for the audience. This real-world and hands-on activity has been shown to be successful at teaching complicated scientific concepts to non-science majors through relevant cultural heritage material.

1.3 Importance Collectively, this thesis reports how the instrumentation and analysis techniques that can be used to understand the material nature of cultural heritage objects. Thus, it provides valuable scientific information to humanities disciplines and those caring for these objects. Specifically, copper salts were addressed due to their extensive use in Western and Eastern artworks and their reactivity in organic media.1-2, 4-5, 15 By utilizing complementary analyses, the characterization of copper chloride trihydroxides and other materials puts them into the larger context of history, specifically in Chinese architectural painting.29, 35-38 Furthermore, additional studies with Raman spectroscopy and chemometrics provide information on the manufacturing and origins of some copper-based pigments. The reactive nature of copper salts as pigments has previously been demonstrated and studied. This dissertation builds on these findings by documenting the effects of organic surroundings, cellulose and gum Arabic, on copper acetates. The studies indicated a degradation pathway consisting of an intermediate state and stabilizing in the copper (I) state. These studies were adapted to educating students, both science and non- science majors, about challenging chemistry topics by connecting them to real-world

6 applications. Thus, this work provides answers using spectroscopic studies to chemical problems on copper-based pigments manufacturing, preservation, and conservation of cultural heritage objects.

7 REFERENCES

1. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

2. Naumova, M. M.; Pisareva, S. A., A Note on the Use of Blue and Green Copper Compounds in Paintings. Studies in Conservation 1994, 39 (4), 277- 283.

3. Zoleo, A.; Nodari, L.; Rampazzo, M.; Piccinelli, F.; Russo, U.; Federici, C.; Brustolon, M., Characterization of Pigment and Binder in Badly Conserved Illuminations of a 15th‐Century Manuscript. Archaeometry 2014, 56 (3), 496- 512.

4. Scott, D. A.; Taniguchi, Y.; Koseto, E., The verisimilitude of verdigris: a review of the copper carboxylates. Studies in Conservation 2013, 46 (2), 73- 91.

5. Santoro, C.; Zarkout, K.; Le Ho, A.-S.; Mirambet, F.; Gourier, D.; Binet, L.; Pagès-Camagna, S.; Reguer, S.; Mirabaud, S.; Le Du, Y.; Griesmar, P.; Lubin- Germain, N.; Menu, M., New Highlights on degradation process of verdigris from easel paintings. Applied Physics A 2014, 114, 637-645.

6. Prati, S.; Bonacini, I.; Sciutto, G.; Genty-Vincent, A.; Cotte, M.; Eveno, M.; Menu, M.; Mazzeo, R., ATR-FTIR microscopy in mapping mode for the study of verdigris and its secondary products. Applied Physics A 2015, 122 (1), 10.

7. Gunn, M.; Chottard, G.; Rivière, E.; Girerd, J.-J.; Chottard, J.-C., Chemical Reactions Between Copper Pigments and Oleoresinous Media. Studies in Conservation 2002, 47 (1), 12-23.

8. Dedecker, K.; Pillai, R. S.; Nouar, F.; Pires, J.; Steunou, N.; Dumas, E.; Maurin, G.; Serre, C.; Pinto, M. L., Metal-Organic Frameworks for Cultural Heritage Preservation: The Case of Acetic Acid Removal. ACS Applied Materials & Interfaces 2018, 10 (16), 13886-13894.

8 9. Salvadó, N.; Butí, S.; Cotte, M.; Cinque, G.; Pradell, T., Shades of green in 15th century paintings: combined microanalysis of the materials using synchrotron radiation XRD, FTIR and XRF. Applied Physics A 2013, 111 (1), 47-57.

10. Švarcová, S.; Čermáková, Z.; Hradilová, J.; Bezdička, P.; Hradil, D., Non- destructive micro-analytical differentiation of copper pigments in paint layers of works of art using laboratory-based techniques. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2014, 132, 514-525.

11. Schmidt, B. A.; Ziemann, M. A.; Pentzien, S.; Gabsch, T.; Koch, W.; Krüger, J., Technical analysis of a Central Asian wall painting detached from a Buddhist cave temple on the northern Silk Road. Studies in Conservation 2016, 61 (2), 113-122.

12. Chaplin, T. D.; Clark, R. J. H.; Jacobs, D.; Jensen, K.; Smith, G. D., The Gutenberg Bibles: Analysis of the Illuminations and Inks Using Raman Spectroscopy. Analytical Chemistry 2005, 77 (11), 3611-3622.

13. Lee, L. R.; Thompson, A.; Daniels, V. D., Princes of the House of Timur: conservation and examination of an early Mughal painting. Studies in Conservation 1997, 42 (4), 231-240.

14. Kanngießer, B.; Hahn, O.; Wilke, M.; Nekat, B.; Malzer, W.; Erko, A., Investigation of oxidation and migration processes of inorganic compounds in ink-corroded manuscripts. Spectrochimica Acta Part B: Atomic Spectroscopy 2004, 59 (10), 1511-1516.

15. Scott, D. A., A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation 2000, 45 (1), 39- 53.

16. Calvini, P.; Gorassini, A., The Degrading Action of Iron and Copper on Paper A FTIR-Deconvolution Analysis. In Restaurator, 2002; Vol. 23, p 205.

17. Šelih, V. S.; Strlič, M.; Kolar, J.; Pihlar, B., The role of transition metals in oxidative degradation of cellulose. Polymer Degradation and Stability 2007, 92 (8), 1476-1481.

18. Kolar, J.; Možir, A.; Balažic, A.; Strlič, M.; Ceres, G.; Conte, V.; Mirruzzo, V.; Steemers, T.; de Bruin, G., New Antioxidants for Treatment of Transition Metal Containing Inks and Pigments. In Restaurator, 2008; Vol. 29, p 184.

9 19. Malešič, J.; Kolar, J.; Anders, M., Evaluation of Treatments for Stabilization of Verdigris and Malachite Containing Paper Documents. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 283.

20. Ahn, K.; Hartl, A.; Hofmann, C.; Henniges, U.; Potthast, A., Investigation of the stabilization of verdigris-containing rag paper by wet chemical treatments. Heritage Science 2014, 2 (1), 12.

21. Cartechini, L.; Miliani, C.; Brunetti, B. G.; Sgamellotti, A.; Altavilla, C.; Ciliberto, E.; D’Acapito, F., X-ray absorption investigations of copper resinate blackening in a XV century Italian painting. Applied Physics A 2008, 92, 243- 250.

22. Pliny; Rackham, H.; Jones, W. H. S.; Eicholz, D. E., Natural history, Volume IX: Book 34.26. Harvard University Press ; W. Heinemann: Cambridge; London, 1979.

23. Thompson, D. V., The Materials and Techniques of Medieval Painting. Dover Publications: New York, 1956.

24. Martens, W.; Frost, R. L.; Williams, P. A., Raman and infrared spectroscopic study of the basic copper chloride minerals -implications for the study of the copper and brass corrosion and "bronze disease". Neues Jahrbuch für Mineralogie - Abhandlungen: Journal of Mineralogy and Geoche 2003, 178 (2), 197-215.

25. Pollard, A. M.; Thomas, R. G.; Williams, P. A., Synthesis and stabilities of the basic copper(II) chlorides atacamite, paratacamite and botallackite. Mineralogical Magazine 1989, 53 (373), 557-563.

26. Smith, J. P.; Smith, F. C.; Ottaway, J.; Krull-Davatzes, A. E.; Simonson, B. M.; Glass, B. P.; Booksh, K. S., Raman Microspectroscopic Mapping with Multivariate Curve Resolution–Alternating Least Squares (MCR-ALS) Applied to the High-Pressure Polymorph of Titanium Dioxide, TiO2-II. Applied Spectroscopy 2017, 71 (8), 1816-1833.

27. Offroy, M.; Moreau, M.; Sobanska, S.; Milanfar, P.; Duponchel, L., Pushing back the limits of Raman imaging by coupling super-resolution and chemometrics for aerosols characterization. Scientific Reports 2015, 5, 12303.

28. Li, M. Analysis and Research on Copper Green Pigment. Northwest University, Xi’an, Shaanxi,China, 2013.

10 29. Hu, K.; Bai, C.; Ma, L.; Bai, K.; Liu, D.; Fan, B., A study on the painting techniques and materials of the murals in the Five Northern Provinces’ Assembly Hall, Ziyang, China. Heritage Science 2013, 1 (1), 18.

30. Wiggins, M. B.; Alcántara-García, J.; Booksh, K. S., Characterization of copper-based pigment preparation and alteration products. MRS Advances 2017, 2 (63), 3973-3981.

31. San Andrés, M.; de la Roja, J. M.; Baonza, V. G.; Sancho, N., Verdigris pigment: a mixture of compounds. Input from Raman spectroscopy. Journal of Raman Spectroscopy 2010, 41 (11), 1468-1476.

32. Masciocchi, N.; Corradi, E.; Sironi, A.; Moretti, G.; Minelli, G.; Porta, P., Preparation, Characterization, and ab initio X-Ray Powder Diffraction Study of Cu2(OH)3(CH3COO)·H2O. Journal of Solid State Chemistry 1997, 131 (2), 252-262.

33. Solomon, S. D.; Rutkowsky, S. A.; Mahon, M. L.; Halpern, E. M., Synthesis of Copper Pigments, Malachite and Verdigris: Making Tempera Paint. Journal of Chemical Education 2011, 88 (12), 1694-1697.

34. Wiggins, M. B.; Heath, E.; Alcántara-García, J., Multidisciplinary Learning: Redox Chemistry and Pigment History. Journal of Chemical Education 2019, 96 (2), 317-322.

35. Fan, Y.; Chen, X.; Li, Z.; Hu, Z., Micro diffraction analysis of the rare green pigment botallackite in ancient wall paintings. Journal of Lanzhou University 2004, 40 (5), 52-55.

36. Xia, Y.; Wang, W. F.; Liu, L. X., Study on mural painting pigments of Fuxi Temple, Tianshui, Gansu Province. Science Conservation Archaeology 2011, 23, 18-24.

37. Yong, L., Copper trihydroxychlorides as pigments in China. Studies in Conservation 2012, 57 (2), 106-111.

38. Egel, E.; Simon, S., Investigation of the painting materials in Zhongshan Grottoes(Shaanxi, China). Heritage Science 2013, 1 (1), 29. https://doi.org/10.1186/2050-7445-1-29

11 Chapter 2

REVIEW OF COPPER SALTS IN CULTURAL HERITAGE

2.1 Introduction Copper salts are of great interest to the field of cultural heritage, either as degradation products, or as constituent materials. For millennia, they were common pigments used by artists for their blue or green color and were applied to various subject surfaces via organic binders. Common copper salts found in cultural heritage are copper carbonates (azurite, malachite), copper acetates (verdigris), basic copper chlorides (atacamite, botallackite) and copper sulfates (posnjakite, langite).1,2 In all cases, each group of pigments includes the various crystal structures of these salts. The term “verdigris” is nowadays applied to pigments composed of (a mixture of) copper (II) acetates. However, throughout history, the term could also refer to any of the previously mentioned copper salts. Historians, conservators and scientists are interested in how they were used, how to identify them, and how they degrade. This chapter only focuses on the studies of verdigris and basic copper chlorides.

2.2 Use in Cultural Heritage

Verdigris is an umbrella term for copper acetates and it has been a common pigment used since antiquity. Its preparation was reported by Pliny the Elder of the early Roman Empire, and it has been found on several ancient Egyptian artifacts.2,3 Likewise, basic copper chlorides have been found on Egyptian objects, and a recipe for the pigment was recorded by Theophilus in the 12th century.4 Daniels found copper

12 salts, possibly verdigris, in waxy paints on four different Egyptian antiques dating from 1186 BCE to after 664 BCE.5 Verdigris was also found, alongside other colorants, in drawing illustrations on 19th dynasty Egyptian papyri of the Book of the Dead.6 Copper chloride trihyrdoxides were found on Egyptian painted objects, such as coffins7,8 and papyri,9 as well. These examples highlight the use of these copper salts as pigments for millennia.

2.2.1 Paintings Verdigris is used predominantly as green shades in both decorative easel painting and architectural painting. Cennino Cennini discussed its use in painting in his artists’ treatise, thus highlighting its popularity amongst artists, specifically in the medieval and Renaissance periods.10 Shades of green in paintings during a wide range of periods were commonly copper-based, both as verdigris or atacamite, as seen in a selection of survey works analyzed by Salvadó et al.11 and Švarcová et al.12 Additionally, architectural and wall paintings containing verdigris includes those dating from both the Roman Empire13 and 16-18th centuries14,15 by different application techniques. Various isomers of copper chloride trihydroxides have also been identified in paintings. Scholars have referred them to as “salt green”, a pigment related to verdigris. They are most often found used in Eastern artwork, such as Chinese architectural and mural paintings.16-24 Additional examples include 16th century Russian frescoes,25 medieval Swedish murals,26 Gothic Catalan altarpieces,27 etc.12,28- 33

13 2.2.2 Paper Copper-based pigments and compounds have been popular colorants for paper- based artwork and documentation. While painting analysis is a major part of technical art history, identification of copper materials on documentation and manuscripts also has large ramifications for the stability of the objects. Although basic copper chlorides are not widely found on these types of objects, verdigris and copper inks are often found. Atacamite, a copper chloride trihydroxide, has been found on maps, its role and origin being uncertain.34,35 Verdigris has been found in illustrations of books, such as the Gutenberg Bibles,36 and in manuscripts, from Islamic to German manuscripts.37-48 Maps are another common document type where verdigris was used as a colorant.40,49- 51

2.3 Preparation and Characterization

Conservators and scientists have shown consistent interest in these copper salts and their use in artwork since 2010. A portion of these studies has often been focused on the preparation of these compounds using historical methods, and characterizing those products, which are comparable to the ones artists would have used on objects. This characterization provides crucial insight into the materiality of the artifacts, both to learn about the creation of an object and to assess its condition. Often by-products of synthesized pigments affect the degradation of the artifact; thus, it is important to understand the starting materials before degradation.

2.3.1 Verdigris (Copper Acetates) The first record of verdigris preparation methods was by Pliny the Elder. Pliny describes the simplest recipe as:2,3

14 [I]t is scraped off the stone from which copper is smelted or by drilling holes in white copper and hanging it up casks of strong vinegar which is stopped with a lid; the verdigris is of much better quality if the same process is performed with scales of copper.

The metal copper reacts with the acetate ions to form a blue-green crust of copper acetates. Variations on this recipe have been reported, and those recipes are well consolidated in David Scott’s Copper and Bronze in Art: Corrosion, Colorants, Conservation. These variations include mixing in salt, alum, soda, or using oak boxes.2 Many of the studies characterizing modern reproductions of this recipe show that the copper (II) acetate mixtures that compose “verdigris” differ in their hydration and basicity. The formulas for these components follow [Cu(CH3COO)2]x

[Cu(OH)2]y nH2O. For “neutral verdigris”, the structure is (x=1, y=0, n=1). For basic copper acetates, these are several structures: (x=2, y=1, n=5), (x=1, y=1, n=5), (x=1, y=2, n=0), (x=1, y=3, n=2).2,52 Neutral verdigris is the most common, as basic copper acetates have often been treated with vinegar before being used as a pigment.53 Reference spectra for these various crystal structures for Raman, FTIR, and XRD can be found in San Andrés et al. and Salvadó.11,54 De la Roja et al. have used other historical recipes from treatises to create verdigris, and characterized all four products as dominantly neutral verdigris with traces of compounds that could not be identified.55 Additional references pertaining to copper acetates, useful for conservation scientists identifying these materials, can be found.53,56-61 San Andreas et al. highlights the heterogeneity of the verdigris crystals that can found on a single sheet reinforcing the complex nature of the pigment.54

15 2.3.2 Basic Copper Chlorides In comparison to copper acetates, discussion of the use of copper chloride trihydroxides as artists’ materials is far more limited. Basic copper chlorides are found as natural minerals, as synthesized pigments, and as corrosion products. The polymorphs of basic copper chlorides are atacamite (orthorhombic), paratacamite (rhombohedral), clinoatacamite (monoclinic), and botallackite (monoclinic).62,63 Atacamite is the most abundant and naturally stable form, and is the most frequently discussed copper chloride trihydroxide. The limited information on these artists’ materials may be due to the assumption copper chloride trihydroxides were not used as original material but, instead, are an alternation product of other copper-based pigments. Copper chloride trihydroxides could have originated as natural or as synthetic pigments, while verdigris would only have been an actively synthesized pigment. In actuality, it does seem that these basic copper chloride pigments were originally used in some paintings, especially Chinese works, and there is a historical recipe available for the compounds. Theophilus describes taking copper sheets, smearing them with honey, and then rolling them in salt. Those sheets were then suspended over vinegar and sealed until a blue-green crust had formed. For this reason, this pigment is frequently referred to as “salt green”.64 This “salt green” recipe was repeated by several studies, although they are not well detailed. David Scott found the predominant product of his prepared samples to be atacamite.62 Kühn’s characterization was not extensive, and concluded the mixtures were either basic copper chlorides and copper carboxylates or basic copper chlorides and copper acetates.52 Naumova and Pisareva also found a mixture of atacamite and copper acetates in their samples.28 This suggests “salt green” pigments would generally have been the copper chloride trihydroxide, atacamite, and possibly varying

16 amounts of verdigris. Additional recipes appear to be available in Chinese primary sources that produce mixtures of atacamite and botallackite polymorphs, but these recipes require subsequent investigation and analysis.65

2.3.3 Characterization Methods The characterization of copper-based pigments in cultural heritage objects is dependent on a range of instrumentation. These techniques should preferably be non- or minimally-invasive in order to preserve the condition of priceless artifacts; however, destructive methods are used for small amounts of samples when necessary. X-ray fluorescence (XRF) is an extremely useful technique for identifying the presence of copper non-invasively; however, it can only identify and occasionally quantify the element, but not identify the form of copper.5,42,46,66,67 It is generally useful for initial investigations, but additional methods are necessary to determine the copper compound present. Primary laboratory-based methods to characterize the copper salt present on an object include Fourier-transform infrared spectroscopy (FTIR),5,11,12,14,22,23,27,42,66,67 Raman spectroscopy,13,18,20,22,32-34,36,37,39,44-46,49,68-71 and x- ray diffraction (XRD).11,12,19,21,22,24,27,41,66 Each techniques includes challenges for distinguishing and identifying these materials in complex mixtures with spectral interferences. While FTIR and Raman can be performed in situ on objects, the structure and conformation of the object may limit the information one can obtain. XRD requires sampling to characterize the materials. Less common techniques, such as x-ray absorption spectroscopy (XAS),72,73 micro-particle induced x-ray emission (PIXE),6,49 and differential pulse voltammetry,29 have been used to identify compounds in more challenging cases. Unfortunately, these techniques have problems regarding their limited availability

17 (some being available at synchrotron facilities) and interpretation (requiring specialists). For most conservation science laboratories and museums, there are limited resources to investigate these materials.

2.4 Alteration and Treatment Studies Copper salts, especially verdigris, are commonly associated with the alteration of their organic surroundings,41,74 and can induce severe degradation that is visually disruptive.38,40,42,50,51,75 Frequently, reacted copper products are found on cultural heritage objects as ultimate degradation products. The presence of copper oxides is associated with copper-based pigment starting materials, such as the reactive verdigris pigment. This has been the case in both paintings and painted objects, such as those reported by Cartechini et al.72 Additionally, copper carboxylates and copper resinates are found on objects;14,49,69,72,76 these are associated with verdigris which has reacted with the paint binders (intentionally or not) to produce these compounds. Other organic copper salts have been considered to be verdigris by several technical studies, such as in Chaplin et al.’s work on the Gutenberg Bible36 and Tanevska et al.’s work on Islamic manuscripts.37 Typically, verdigris studies have monitored mock-up samples that have undergone accelerated degradation to understand this process. Generally, such mock- up studies are carried out by conservators with some scientific assistance in order to evaluate and study possible treatment and conservation methods for objects with verdigris. Such similar studies for basic copper chlorides are not published, possibly because many researchers associate their presence with a degradation product and not original artists’ material.

18 2.4.1 In Paintings In painting, copper salts can be found mixed in oil, egg, or resin Subsequently, these copper salts react with the organics to form alteration products, which have been the focus of several scientific studies. It is worth noting that all of the following studies are on “mock-ups”, or surrogate samples. Early work by Rasti and Scott indicated verdigris might be a light stabilizer for oil paints because it is reduced and acts as an antioxidant.77 Work by Prati et al. used FTIR have been used to monitor the curing of verdigris and lead tin yellow (type I) in linseed oil to show that during the early stages of curing, copper formates form, with subsequent formation of copper carboxylates and oxalates. The copper oxalates, commonly found alteration products, might have been the result of reactions with the binder or be the result of biochemical activity.53 Zoppi et al.’s work shows that copper pigments, especially verdigris, have a high affinity and react readily with available oxalic acid.78 Ortiz-Miranda et al. studied the effects of fungi and bacteria on verdigris samples in egg and oil mixtures. They found the fungi and bacteria modified the protein and released Cu (II) ions, which then reacted with carboxylate groups. However, these effects were less effective in emulsions that contained more oil.79 Studies of protein-based paints, like egg tempera, show copper pigments, including verdigris, decrease the retrievable amount of intact proteins for analysis and promote degradation.80 Generally, only neutral verdigris is used to study oil paint aging, and those studies indicate that copper ligands form. Santoro et al. found the ligand-to- metal charge transfer is triggered by light absorption, which releases carboxylates that reduce Cu(II) to Cu(I). Even with a small formation of Cu(I), the optical properties and color are drastically changed. Following this reduction, dimers of Cu(I)-Cu(I) are formed. The authors hypothesized the copper is oxidized by peroxide to form a

19 74 stabilizing bridge, Cu(II)-O2-Cu(II). Ioakimoglou et al. had found that the formation of peroxides in linseed oil was accelerated by verdigris, compared to other copper- based pigments.81 Doménech-Carbó et al. showed that copper ions also react with terpenoid resins, which had already been established in the commonly found product, copper resinate.52,82 In the presence of copper, the original diterpenoids decrease while their oxidized products increase, until light exposure forms metal-resin complexes. Of the pigments incorporated in Doménech-Carbó et al.’s study, verdigris shows the strongest effects on the resin.82 Research by Gunn et al. established the ability of component acids of oil and resin binders to extract copper ions from verdigris. Linoleic and abietic acids (both in free acid and carboxylate forms) have been used to represent oil and resin binders. Both acids have an equal rate of extraction, which starts immediately, as the carboxylic groups replace the acetate ligands. Like Santoro et al., the authors associate this phenomena with the color, but elaborate that the extraction of copper ions from copper carbonates, like malachite and azurite, is not as easy and makes them more stable.76

2.4.2 On Paper In the case of paper-based objects, copper salts are found to react with the cellulose support. The copper ions appear to play the largest role in this process. However, copper ions on paper objects do not just originate from pigments, but also from copper-rich inks, which are what many studies focus on. The role of copper ions, either as copper-rich inks or pigments, in the degradation of cellulose has been extensively studied in relation to environmental effects and treatments. These studies

20 are less in-depth than the studies of painting because they focus on conservation treatments for cellulose being damaged by copper. Fundamental studies regarding copper focus on its hydrolysis depolymerization in an acidic or alkaline environment, or its oxidation. Depolymerization results in the breakdown of the cellulose substrate, and the cellulose’s oxidation is the source of discoloration of the paper.83,84 Frequently, copper is studied in relation to iron on cellulose due to the extensive use of iron gall inks, as copper was occasionally a component of the iron gall ink mixtures. Bicchieri and Pepa found copper ions catalyze the oxidation of the glucopyronase ring, while iron cleaves the 1-4-β-glucosidic bond.84 Calvini and Gorassini expand on these findings, stating that Cu(II) oxidizes the glucopyronase ring to aldehyde and carboxyl groups, which open the polymer.83 Most of the transition metals relevant to paper artifacts will catalyze the oxidation of the cellulose, but copper, as Cu(II), is found to be “far more catalytically active”.85-87 Copper produces the most hydroxyl radicals when artificially aged with cellulose through Fenton-like reactions, and copper’s activation energy for producing oxidizing species was calculated to be the highest at 108 ± 2 kJ mol K.85,86,88 Strlič et al. also states that copper and iron do not interact when mixed together and artificially aged on paper.85 However, Kanngießer et al. contradicts this by stating copper ions catalyze the photoreduction of iron in iron gall inks.87 For conservators and curators of the objects, the focus of the studies on copper degradation is on evaluating treatment methods. Kolar et al. have established that antioxidants are needed to treat paper degradation to stop the oxidation-induced decay.89 Over the past decade, many studies have evaluated antioxidants in order to inhibit the hydrolysis and oxidation of cellulose caused by copper ions.

21 Overwhelmingly, the preferred treatment when copper is present is tetrabutylammonium bromide; it is best at stabilizing the paper and minimizing color changes.40,51,90-93 Copper and zinc ions are the most mobile on paper, which facilitates damage.87 Interestingly, many artists were aware of the detrimental effect of verdigris, and some of them took precautions to limit the possible damage. Persian artists would add saffron, a natural dye, to verdigris, and objects using this mixture are in far better condition than many other similar objects. This noticeable trend is because saffron has an unsaturated dicarboxylic acid, ester, and nitrogen groups, which act as a pH buffer to stabilize the verdigris.94,95 An additional concern with either copper salt in this discussion is the anion present in the pigments, which has the potential to catalyze degradation. For verdigris, it is the acetate group, and basic copper chlorides contain halide ions. Vapors of acetic acid, a common volatile organic compound (VOC), are known to cause damage to inorganics and organics alike,96-100 so there is the potential for acetate groups within verdigris to contribute to the depolymerizaton of organics as well as the copper ions themselves.101 The behavior of chloride ions in cultural heritage related organics is significantly less studied. Overall, it is associated with degradation and the uptake of water in paint films. Some halides, including chlorides, have been shown to be successful in treating degrading cellulose.102

2.5 Conclusion This chapter summarizes the use of two common groups of copper salts as they are studied in cultural heritage: copper acetates and basic copper chlorides. These groups contain variations of crystal structures that change some properties of the

22 pigments. Identifying these has primarily been done using FTIR, Raman, and XRD. Verdigris has been found in a range of paint- and paper-based materials created throughout the centuries. In attempts to describe verdigris’s reactive nature, degradation studies have been carried out on both oil and cellulosic media to determine how to treat these materials. Basic copper chlorides are less common, and they have only been found in paintings. They are frequently considered corrosion products, but they can be present as original material. The fact likely explains the presence of fewer degradation studies on these materials. However, the studies on verdigris give some insight into the behavior of basic copper chlorides as well.

23 REFERENCES

1. Valadas, S.; Freire, R. V.; Cardoso, A.; Mirão, J.; Dias, C. B.; Vandenabeele, P.; Candeias, A., On the Use of the Unusual Green Pigment Brochantite (Cu4(SO4)(OH)6) in the 16th-Century Portuguese-Flemish Paintings Attributed to The Master Frei Carlos Workshop. Microscopy and Microanalysis 2015, 21 (2), 518-525.

2. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

3. Pliny; Rackham, H.; Jones, W. H. S.; Eicholz, D. E., Natural history, Volume IX: Book 34.26. Harvard University Press ; W. Heinemann: Cambridge; London, 1979.

4. Theophilus; Dodwell, C. R.; Theophilus; Theophilus; Lessing, J. R. R. C., The various arts. T. Nelson: London; New York, 1961.

5. Daniels, V., Analyses of Copper- and Beeswax-Containing Green Paint on Egyptian Antiquities. Studies in Conservation 2007, 52 (1), 13-18.

6. Olsson, A. M. B.; Calligaro, T.; Colinart, S.; Dran, J. C.; Lövestam, N. E. G.; Moignard, B.; Salomon, J., Micro-PIXE analysis of an ancient Egyptian papyrus: Identification of pigments used for the “Book of the Dead”. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 2001, 181 (1), 707-714.

7. Scott, D. A., Ancient Egyptian Pigments: The Examination of Some Coffins from the San Diego Museum of Man. MRS Bulletin 2010, 35 (5), 390-396.

8. Abdelaal, S.; Mahmoud, N.; Detalle, V., A technical examination and the identification of the wood, pigments, grounds and binder of an ancient egyptian sarcophagus. Int. J. Conserv. Sci. International Journal of Conservation Science 2014, 5 (2), 177-188.

9. Scott, D. A., A review of ancient Egyptian pigments and cosmetics Studies in Conservation 2014, 61 (4), 185-202.

24 10. Cennini, C.; Thompson, D. V., The craftsman's handbook. Dover Pubications: New York, 1954.

11. Salvadó, N.; Butí, S.; Cotte, M.; Cinque, G.; Pradell, T., Shades of green in 15th century paintings: combined microanalysis of the materials using synchrotron radiation XRD, FTIR and XRF. Applied Physics A 2013, 111 (1), 47-57.

12. Švarcová, S.; Čermáková, Z.; Hradilová, J.; Bezdička, P.; Hradil, D., Non- destructive micro-analytical differentiation of copper pigments in paint layers of works of art using laboratory-based techniques. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2014, 132, 514- 525.

13. Villar, S. E. J.; Edwards, H. G. M., An extensive colour palette in Roman villas in Burgos, Northern Spain: a Raman spectroscopic analysis. Analytical and Bioanalytical Chemistry 2005, 382 (2), 283-289.

14. Samanian, K., Identification of Green Pigment Used in Persian Wall Paintings (ad 1501–1736) Using PLM, FT–IR, SEM/EDX and GC–MS Techniques. Archaeometry 2015, 57 (4), 740-758.

15. Alderson, C., Re-Creating a 19th Century Paint Palette. Bulletin of the Association for Preservation Technology 1984, 16 (1), 47-56.

16. Yong, L., Copper trihydroxychlorides as pigments in China. Studies in Conservation 2012, 57 (2), 106-111.

17. Piquè, F.; Courtauld Institute of, A.; Conservation of Wall Painting, D.; Getty Conservation, I. Scientific examination of the sculptural polychromy of Cave 6, Yungang. 1992.

18. Xia, Y.; Wang, W. F.; Liu, L. X., Study on mural painting pigments of Fuxi Temple, Tianshui, Gansu Province. Science Conservation Archaeology 2011, 23, 18-24.

19. Fan, Y.; Chen, X.; Li, Z.; Hu, Z., Micro diffraction analysis of the rare green pigment botallackite in ancient wall paintings. Journal of Lanzhou University 2004, 40 (5), 52-55.

20. Hu, K.; Bai, C.; Ma, L.; Bai, K.; Liu, D.; Fan, B., A study on the painting techniques and materials of the murals in the Five Northern Provinces’ Assembly Hall, Ziyang, China. Heritage Science 2013, 1 (1), 18. https://doi.org/10.1186/2050-7445-1-18

25 21. Sultan, S.; Kareem, K.; He, L.; Simon, S., Identification of the authenticity of pigments in ancient polychromed artworks of China. Analytical Methods 2017, 9 (5), 814-825.

22. Schmidt, B. A.; Ziemann, M. A.; Pentzien, S.; Gabsch, T.; Koch, W.; Krüger, J., Technical analysis of a Central Asian wall painting detached from a Buddhist cave temple on the northern Silk Road. Studies in Conservation 2016, 61 (2), 113-122.

23. Egel, E.; Simon, S., Investigation of the painting materials in Zhongshan Grottoes(Shaanxi, China). Heritage Science 2013, 1 (1), 29. https://doi.org/10.1186/2050-7445-1-29

24. Wei, S.; Schreiner, M.; Guo, H.; Ma, Q., Scientific Investigation of the Materials in a Chinese Ming Dynasty Wall Painting. International Journal of Conservation Science 2010, 1 (2), 99-112.

25. Naumova, M. M.; Pisareva, S. A.; Nechiporenko, G. O., Green Copper Pigments of Old Russian Frescoes. Studies in Conservation 1990, 35 (2), 81-88.

26. Nord, A. G.; Tronner, K., The Frequent Occurrence of Atacamite in Medieval Swedish Murals. Studies in Conservation 2018, 63 (8), 477-481.

27. Salvadó, N.; Pradell, T.; Pantos, E.; Papiz, M. Z.; Molera, J.; Seco, M.; Vendrell-Saz, M., Identification of copper-based green pigments in Jaume Huguet's Gothic altarpieces by Fourier transform infrared microspectroscopy and synchrotron radiation X-ray diffraction. Journal of Synchrotron Radiation 2002, 9 (4), 215-222.

28. Naumova, M. M.; Pisareva, S. A., A note on the use of blue and green copper compounds in paintings. Studies in Conservation 1994, 39 (4), 277- 283.

29. Doménech-Carbó, M. T.; Casas-Catalán, M. J.; Doménech-Carbó, A.; Mateo-Castro, R.; Gimeno-Adelantado, J. V.; Bosch-Reig, F., Analytical study of canvas painting collection from the Basilica de la Virgen de los Desamparados using SEM/EDX, FT-IR, GC and electrochemical techniques. Fresenius' Journal of Analytical Chemistry 2001, 369 (7), 571- 575.

26 30. Richardin, P.; Mazel, V.; Walter, P.; Laprévote, O.; Brunelle, A., Identification of Different Copper Green Pigments in Renaissance Paintings by Cluster-TOF-SIMS Imaging Analysis. Journal of The American Society for Mass Spectrometry 2011, 22 (10), 1729-1736.

31. Kim, M.; Lee, J.; Doh, J.-M.; Ahn, H.; Kim, H. D.; Yang, Y.; Lee, Y., Characterization of ancient Korean pigments by surface analytical techniques. Surface and Interface Analysis 2016, 48 (7), 409-414.

32. Proietti, N.; Di Tullio, V.; Presciutti, F.; Gentile, G.; Brunetti, B. G.; Capitani, D., A multi-analytical study of ancient Nubian detached mural paintings. Microchemical Journal 2016, 124, 719-725.

33. Holakooei, P.; Karimy, A.-H., Micro-Raman spectroscopy and X-ray fluorescence spectrometry on the characterization of the Persian pigments used in the pre-seventeenth century wall paintings of Masjid-i Jāme of Abarqū, central Iran. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2015, 134, 419-427.

34. Sommer, D. V. P.; Mühlen Axelsson, K.; Collins, M. J.; Fiddyment, S.; Bredal-Jørgensen, J.; Simonsen, K. P.; Lauridsen, C. B.; Larsen, R., Multiple Microanalyses of a Sample from the Vinland Map. Archaeometry 2017, 59 (2), 287-301.

35. Kogou, S.; Neate, S.; Coveney, C.; Miles, A.; Boocock, D.; Burgio, L.; Cheung, C. S.; Liang, H., The origins of the Selden map of China: scientific analysis of the painting materials and techniques using a holistic approach. Heritage Science 2016, 4 (1), 28. https://doi.org/10.1186/s40494-016-0098-x

36. Chaplin, T. D.; Clark, R. J. H.; Jacobs, D.; Jensen, K.; Smith, G. D., The Gutenberg Bibles: Analysis of the Illuminations and Inks Using Raman Spectroscopy. Analytical Chemistry 2005, 77 (11), 3611-3622.

37. Tanevska, V.; Nastova, I.; Minčeva-Šukarova, B.; Grupče, O.; Ozcatal, M.; Kavčić, M.; Jakovlevska-Spirovska, Z., Spectroscopic analysis of pigments and inks in manuscripts: II. Islamic illuminated manuscripts (16th–18th century). Vibrational Spectroscopy 2014, 73, 127-137.

38. Zoleo, A.; Nodari, L.; Rampazzo, M.; Piccinelli, F.; Russo, U.; Federici, C.; Brustolon, M., Characterization of Pigment and Binder in Badly Conserved Illuminations of a 15th‐Century Manuscript. Archaeometry 2014, 56 (3), 496-512.

27 39. Clark, R. J. H.; Gibbs, P. J., Analysis of 16th Century Qazwīnī Manuscripts by Raman Microscopy and Remote Laser Raman Microscopy. Journal of Archaeological Science 1998, 25 (7), 621-629.

40. Hofmann, C.; Hartl, A.; Ahn, K.; Faerber, I.; Henniges, U.; Potthast, A., Studies on the Conservation of Verdigris on Paper. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 147.

41. Scott, D. A.; Khandekar, N.; Schilling, M. R.; Turner, N.; Taniguchi, Y.; Khanjian, H., Technical Examination of a Fifteenth-Century German Illuminated Manuscript on Paper: A Case Study in the Identification of Materials. Studies in Conservation 2001, 46 (2), 93-108.

42. Faubel, W.; Staub, S.; Simon, R.; Heissler, S.; Pataki, A.; Banik, G., Non- destructive analysis for the investigation of decomposition phenomena of historical manuscripts and prints. Spectrochimica Acta Part B: Atomic Spectroscopy 2007, 62 (6), 669-676.

43. Arias Teresa, E.; Montes Ana, L.; Bueno Ana, G.; Benito Adrián, D.; García Rosario, B., A Study about Colourants in the Arabic Manuscript Collection of the Sacromonte Abbey, Granada, Spain. A New Methodology for Chemical Analysis. In Restaurator, 2008; Vol. 29, p 76.

44. Chaplin, T. D.; Clark, R. J. H.; McKay, A.; Pugh, S., Raman spectroscopic analysis of selected astronomical and cartographic folios from the early 13th century Islamic ‘Book of Curiosities of the Sciences and Marvels for the Eyes’. Journal of Raman Spectroscopy 2006, 37 (8), 865-877.

45. Gilbert, B.; Denoël, S.; Weber, G.; Allart, D., Analysis of green copper pigments in illuminated manuscripts by micro-Raman spectroscopy. Analyst 2003, 128 (10), 1213-1217.

46. Wehling, B.; Vandenabeele, P.; Moens, L.; Klockenkämper, R.; von Bohlen, A.; Van Hooydonk, G.; de Reu, M., Investigation of pigments in medieval manuscripts by micro raman spectroscopy and total reflection X- ray fluorescence spectrometry. Microchimica Acta 1999, 130 (4), 253-260.

47. Brown, K. L.; Clark, R. J. H., Analysis of key Anglo-Saxon manuscripts (8–11th centuries) in the British Library: pigment identification by Raman microscopy. Journal of Raman Spectroscopy 2004, 35 (3), 181-189.

28 48. Brown, K. L.; Clark, R. J. H., Three English manuscripts post-1066 AD: pigment identification and palette comparisons by Raman microscopy. Journal of Raman Spectroscopy 2004, 35 (3), 217-223.

49. Mendes, N.; Lofrumento, C.; Migliori, A.; Castellucci, E. M., Micro‐ Raman and particle‐induced X‐ray emission spectroscopy for the study of pigments and degradation products present in 17th century coloured maps. Journal of Raman Spectroscopy 2008, 39 (2), 289-294.

50. Miller, A. M.; Hanson, L., The maker and the monk: conservation of the Mercator Atlas of Europe. Journal of the Institute of Conservation 2010, 33 (1), 29-39.

51. Hofmann, C.; Hartl, A.; Ahn, K.; Druceikaite, K.; Henniges, U.; Potthast, A., Stabilization of Verdigris. Journal of Paper Conservation 2016, 17 (3- 4), 88-99.

52. Kühn, H., Verdigris and Copper Resinate. Studies in Conservation 1970, 15 (1), 12-36.

53. Prati, S.; Bonacini, I.; Sciutto, G.; Genty-Vincent, A.; Cotte, M.; Eveno, M.; Menu, M.; Mazzeo, R., ATR-FTIR microscopy in mapping mode for the study of verdigris and its secondary products. Applied Physics A 2015, 122 (1), 10.

54. San Andrés, M.; de la Roja, J. M.; Baonza, V. G.; Sancho, N., Verdigris pigment: a mixture of compounds. Input from Raman spectroscopy. Journal of Raman Spectroscopy 2010, 41 (11), 1468-1476.

55. De la Roja, J. M.; San Andrés, M.; Cubino, N. S.; Santos-Gómez, S., Variations in the colorimetric characteristics of verdigris pictorial films depending on the process used to produce the pigment and the type of binding agent used in applying it. Color Research & Application 2007, 32 (5), 414-423.

56. Pereira, D. C.; de Faria, D. L. A.; Constantino, V. R. L., CuII hydroxy salts: characterization of layered compounds by vibrational spectroscopy. jbchs Journal of the Brazilian Chemical Society 2006, 17 (8), 1651-1657.

57. Masciocchi, N.; Corradi, E.; Sironi, A.; Moretti, G.; Minelli, G.; Porta, P., Preparation, Characterization, and ab initio X-Ray Powder Diffraction Study of Cu2(OH)3(CH3COO)·H2O. Journal of Solid State Chemistry 1997, 131 (2), 252-262.

29 58. Bette, S.; Kremer, R. K.; Eggert, G.; Tang, C. C.; Dinnebier, R. E., On verdigris, part I: synthesis, crystal structure solution and characterisation of the 1–2–0 phase (Cu3(CH3COO)2(OH)4). Dalton Transactions 2017, 46 (43), 14847-14858.

59. Bette, S.; Kremer, R. K.; Eggert, G.; Dinnebier, R. E., On verdigris, part II: synthesis of the 2-1-5 phase, Cu3(CH3COO)4(OH)2·5H2O, by long-term crystallisation from aqueous solution at room temperature. Dalton Transactions 2018, 47 (25), 8209-8220.

60. Chaplin, T. D.; Clark, R. J. H.; Scott, D. A., Study by Raman microscopy of nine variants of the green–blue pigment verdigris. Journal of Raman Spectroscopy 2006, 37 (1‐3), 223-229.

61. Scott, D. A.; Taniguchi, Y.; Koseto, E., The verisimilitude of verdigris: a review of the copper carboxylates. Studies in Conservation 2013, 46 (2), 73-91.

62. Scott, D. A., A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation 2000, 45 (1), 39-53.

63. Pollard, A. M.; Thomas, R. G.; Williams, P. A., Synthesis and Stabilities of the Basic Copper(II) Chlorides Atacamite, Paratacamite and Botallackite. Mineralogical magazine 1989, 53 (373), 557-563.

64. Thompson, D. V., The Materials and Techniques of Medieval Painting. Dover Publications: New York, 1956.

65. Li, M. Analysis and Research on Copper Green Pigment. Northwest University, Xi’an, Shaanxi,China, 2013.

66. Lee, L. R.; Thompson, A.; Daniels, V. D., Princes of the House of Timur: conservation and examination of an early Mughal painting. Studies in Conservation 1997, 42 (4), 231-240.

67. Buti, D.; Rosi, F.; Brunetti, B. G.; Miliani, C., In-situ identification of copper-based green pigments on paintings and manuscripts by reflection FTIR. Anal Bioanal Chem Analytical and Bioanalytical Chemistry 2013, 405 (8), 2699-2711.

30 68. Chaplin, T. D.; Clark, R. J. H.; Martinón-Torres, M., A combined Raman microscopy, XRF and SEM–EDX study of three valuable objects – A large painted leather screen and two illuminated title pages in 17th century books of ordinances of the Worshipful Company of Barbers, London. Journal of Molecular Structure 2010, 976 (1), 350-359.

69. Conti, C.; Striova, J.; Aliatis, I.; Possenti, E.; Massonnet, G.; Muehlethaler, C.; Poli, T.; Positano, M., The detection of copper resinate pigment in works of art: contribution from Raman spectroscopy. Journal of Raman Spectroscopy 2014, 45 (11-12), 1186-1196.

70. Hibberts, S.; Edwards, H. G. M.; Abdel-Ghani, M.; Vandenabeele, P., Raman spectroscopic analysis of a ‘noli me tangere’ painting. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 2016, 374 (2082), 20160044.

71. Wang, N.; He, L.; Egel, E.; Simon, S.; Rong, B., Complementary analytical methods in identifying gilding and painting techniques of ancient clay- based polychromic sculptures. Microchemical Journal 2014, 114, 125-140.

72. Cartechini, L.; Miliani, C.; Brunetti, B. G.; Sgamellotti, A.; Altavilla, C.; Ciliberto, E.; D’Acapito, F., X-ray absorption investigations of copper resinate blackening in a XV century Italian painting. Applied Physics A 2008, 92, 243-250.

73. Quartieri, S., Synchrotron Radiation in Art, Archaelogy and Cultural Heritage. In Synchrotron Radiation: Basics, Methods and Applications, Mobilio, S.; Boscherini, F.; Meneghini, C., Eds. Springer Berlin Heidelberg: Berlin, Heidelberg, 2015; pp 677-695.

74. Santoro, C.; Zarkout, K.; Le Ho, A.-S.; Mirambet, F.; Gourier, D.; Binet, L.; Pagès-Camagna, S.; Reguer, S.; Mirabaud, S.; Le Du, Y.; Griesmar, P.; Lubin-Germain, N.; Menu, M., New Highlights on degradation process of verdigris from easel paintings. Applied Physics A 2014, 114, 637-645.

75. Stanley, T., A Conservation Case Study of Persian Miniatures. In Restaurator, 2006; Vol. 27, p 162.

76. Gunn, M.; Chottard, G.; Rivière, E.; Girerd, J.-J.; Chottard, J.-C., Chemical Reactions Between Copper Pigments and Oleoresinous Media. Studies in Conservation 2002, 47 (1), 12-23.

31 77. Rasti, F.; Scott, G., Mechanisms of antioxidant action: The role of copper salts in the photostabilization of paint media. European Polymer Journal 1980, 16 (12), 1153-1158.

78. Zoppi, A.; Lofrumento, C.; Mendes, N. F. C.; Castellucci, E. M., Metal oxalates in paints: a Raman investigation on the relative reactivities of different pigments to oxalic acid solutions. Analytical and Bioanalytical Chemistry 2010, 397 (2), 841-849.

79. Ortiz-Miranda, A. S.; Doménech-Carbó, A.; Doménech-Carbó, M. T.; Osete-Cortina, L.; Bolívar-Galiano, F.; Martín-Sánchez, I., Analyzing chemical changes in verdigris pictorial specimens upon bacteria and fungi biodeterioration using voltammetry of microparticles. Heritage Science 2017, 5 (1), 8.

80. Ren, F.; Atlasevich, N.; Baade, B.; Loike, J.; Arslanoglu, J., Influence of pigments and protein aging on protein identification in historically representative casein-based paints using enzyme-linked immunosorbent assay. Analytical and Bioanalytical Chemistry 2016, 408 (1), 203-215.

81. Ioakimoglou, E.; Boyatzis, S.; Argitis, P.; Fostiridou, A.; Papapanagiotou, K.; Yannovits, N., Thin-Film Study on the Oxidation of Linseed Oil in the Presence of Selected Copper Pigments. Chemistry of Materials 1999, 11 (8), 2013-2022.

82. Doménech-Carbó, M. T.; Kuckova, S.; de la Cruz-Cañizares, J.; Osete- Cortina, L., Study of the influencing effect of pigments on the photoageing of terpenoid resins used as pictorial media. Journal of Chromatography A 2006, 1121 (2), 248-258.

83. Calvini, P.; Gorassini, A., The Degrading Action of Iron and Copper on Paper A FTIR-Deconvolution Analysis. In Restaurator, 2002; Vol. 23, p 205.

84. Bicchieri, M.; Pepa, S., The Degradation of Cellulose with Ferric and Cupric Ions in a Low-acid Medium. In Restaurator, 1996; Vol. 17, p 165.

85. Strlic, M.; Kolar, J.; Selih, V. S.; Kocar, D.; Pihlar, B., A Comparative Study of Several Transition Metals in Fenton-like Reaction Systems at Circum-Neutral pH. Acta Chimica Slovenica 2003, 50 (4), 619-632.

86. Šelih, V. S.; Strlič, M.; Kolar, J.; Pihlar, B., The role of transition metals in oxidative degradation of cellulose. Polymer Degradation and Stability 2007, 92 (8), 1476-1481.

32 87. Kanngießer, B.; Hahn, O.; Wilke, M.; Nekat, B.; Malzer, W.; Erko, A., Investigation of oxidation and migration processes of inorganic compounds in ink-corroded manuscripts. Spectrochimica Acta Part B: Atomic Spectroscopy 2004, 59 (10), 1511-1516.

88. Strlic, M.; Kolar, J.; Pihlar, B., The effect of metal ion, pH and temperature on the yield of oxidising species in a fenton-like system determined by aromatic hydroxylation. Acta Chimica Slovenica 1999, 46, 555-566.

89. Kolar, J., Mechanism of Autoxidative Degradation of Cellulose. In Restaurator, 1997; Vol. 18, p 163.

90. Ahn, K.; Hartl, A.; Hofmann, C.; Henniges, U.; Potthast, A., Investigation of the stabilization of verdigris-containing rag paper by wet chemical treatments. Heritage Science 2014, 2 (1), 12. https://doi.org/10.1186/2050- 7445-2-12

91. Malešič, J.; Kolar, J.; Anders, M., Evaluation of Treatments for Stabilization of Verdigris and Malachite Containing Paper Documents. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 283.

92. Malešič, J.; Kočar, D.; Balažic Fabjan, A., Stabilization of copper- and iron-containing papers in mildly alkaline environment. Polymer Degradation and Stability 2012, 97 (1), 118-123.

93. Kolar, J.; Možir, A.; Balažic, A.; Strlič, M.; Ceres, G.; Conte, V.; Mirruzzo, V.; Steemers, T.; de Bruin, G., New Antioxidants for Treatment of Transition Metal Containing Inks and Pigments. In Restaurator, 2008; Vol. 29, p 184.

94. Barkeshli, M.; Ataie, G. H., pH Stability of Saffron Used in Verdigris as an Inhibitor in Persian Miniature Paintings. In Restaurator, 2002; Vol. 23, p 154.

95. Davies, B. H., Did the Medieval Illuminator Know How to Prevent Oxidative Damage? Dyes in History and Archaeology 2003, 19, 24-31.

96. Dedecker, K.; Pillai, R. S.; Nouar, F.; Pires, J.; Steunou, N.; Dumas, E.; Maurin, G.; Serre, C.; Pinto, M. L., Metal-Organic Frameworks for Cultural Heritage Preservation: The Case of Acetic Acid Removal. ACS Applied Materials & Interfaces 2018, 10 (16), 13886-13894.

33 97. De Laet, N.; Lycke, S.; Pevenage, J. V.; Moens, L.; Vandenabeele, P., Investigation of pigment degradation due to acetic acid vapours: Raman spectroscopic analysis. European Journal of Mineralogy 2013, 25 (5), 855- 862.

98. Grzywacz, C. M.; Tennent, N. H., Pollution monitoring in storage and display cabinets: carbonyl pollutant levels in relation to artifact deterioration Studies in Conservation 1994, 39 (sup2), 164-170.

99. Prosek, T.; Taube, M.; Dubois, F.; Thierry, D., Application of automated electrical resistance sensors for measurement of corrosion rate of copper, bronze and iron in model indoor atmospheres containing short-chain volatile carboxylic acids. Corrosion Science 2014, 87, 376-382.

100. Tétreault, J.; Sirois, J.; Stamatopoulou, E., Studies of lead corrosion in acetic acid environments. Studies in Conservation 1998, 43 (1), 17-32.

101. Fenech, A.; Strlič, M.; Kralj Cigić, I.; Levart, A.; Gibson, L. T.; de Bruin, G.; Ntanos, K.; Kolar, J.; Cassar, M., Volatile aldehydes in libraries and archives. Atmospheric Environment 2010, 44 (17), 2067-2073.

102. Malesic, J.; Kolar, J.; Strlic, M.; Polanc, S., The Influence of Halide and Pseudo-Halide Antioxidants in Fenton-Like Reaction Systems. ACTA CHIMICA SLOVENICA 2006, 53, 450-456.

34 Chapter 3

EXPERIMENTAL OVERVIEW

3.1 Introduction In this chapter, the experimental methods used to obtain the results are described. The instrumental theory, the type of information collected, and the instrument parameters are found in Section 3.2. The instrumentation played an important role in identifying copper-based pigments and in studying degradation phenomena. Extensive characterizations of these materials, resulting in the identity of the copper salts, provided clues on the nature of the object and the condition of the object during the degradation. It was necessary to use several types of instrumentation to collect the variety of chemical data seen in this dissertation.

3.2 Instrumentation

3.2.1 X-ray Fluorescence X-ray fluorescence (XRF) is a non-destructive technique for identifying elemental compositions in solid and liquid samples. This quantitative technique measures the secondary, or fluorescent, x-rays an element emits, which are unique to each element.1 High-energy incident x-rays hit the sample, and if the energy is sufficient (greater than the atom’s K or L shell binding energy), an inner orbital electron is ejected. An electron from a higher energy orbital then fills the vacancy of the inner orbital. This process releases energy in the form of secondary, or fluorescent,

35 x-rays. The energy of these x-rays is unique to the element, because it is equal to the energy difference between the two quantum states of the electron. However, limitations of this technique mean, at best, only elements with Z greater than 11 can be detected. The amount of emitted radiation is also generally indicative of the element’s concentration in the sample. XRF analysis for Chapter 4 was performed with the ArtTAX µXRF spectrometer using a rhodium tube (600µA current, 50kV voltage, 100 seconds live time irradiation) with a spot size of approximately 70-100 micron. The element detection range is potassium (Z=19) to uranium (Z=92). The emitted x-rays are detected based on the silicon drift principle. Spectra were interpreted using the Intax version 4.5.18.12 software; an integrated CCD camera allowed a magnified image of the region of analysis to be acquired.

3.2.2 Scanning Electron Microscopy-Energy Dispersive X-ray Spectroscopy Scanning electron microscopy (SEM) uses an electron beam to generate high- magnification images of samples. A backscattered electron (BSE) image is generated in gray-scale at sub-micron resolution.2 The backscattered electrons from the sample provide contrast based on the value of Z for the atom. The heavier elements have more collisions and backscattering of electrons, thus appearing brighter. This analysis is generally done in a vacuum to improve the resolution of the image; however, partial vacuum can be used to reduce charging effects on organic materials. Secondary electrons from the sample indicate the topography, because they originate from nearer to the surface of the material. Additionally, the incident electrons interact with matter for energy dispersive x-ray spectroscopy (EDX), which is frequently coupled with SEM. When the incident

36 electrons hit a sample, there is a transfer of energy. That energy can cause an atomic electron to “jump” to a higher orbital or be ejected from the atom altogether. An electron from a higher orbital in the atom then “jumps” to fill the hole left. The difference in energy states releases characteristic x-rays for that electron’s transition for that element. This is a concept similar in theory to XRF. The electromagnetic radiation is indicative of the elements in the sample. The output for this analysis is a spectrum of the intensity with regard to the energy of the x-rays. This method can be used for both qualitative and quantitative analysis. These images show the spatial distribution of these elemental signals to relate to the BSE images. For the analysis in Chapter 4, the embedded cross-sections were first reduced in width for better imaging once mounted in the SEM chamber. The excess casting medium was cut off the back of the sample with a jeweler’s saw. The cross-sections were mounted to SPI Supplies Zeiss aluminum slot head stubs (12.7×3.1mm) with SPI Supplies double-sided carbon tabs (12mm diameter). SPI Supplies conductive carbon paint (20% colloidal graphite in isopropanol) was applied on the side and top surfaces of the casting epoxy, without covering the cross-section itself, to prevent charging. The sample was examined using a Zeiss EVO MA15 scanning electron microscope with LaB6 source at an accelerating voltage of 20kV for the electron beam, working distance of approximately 8.5 mm, and sample tilt of 0°. The EDS data was collected with the Bruker Nano X-flash® detector 6│30 and analyzed with Quantax 200/Esprit 1.9 software.

3.2.3 X-ray Diffraction X-ray diffraction (XRD) is used to identify and characterize crystalline compounds based on how their crystal structure diffracts x-rays.3 A collimated,

37 incident beam of monochromatic x-rays is focused on the sample, and the crystal structure and its atoms interact with and scatter the radiation. The regular structure of a crystal means Bragg’s Law describes the scattering:

� � = 2� sin (�) where n is an integer, λ is the wavelength of the incident x-rays, d is the interplanar spacing between rows of atoms, and θ is the angle of incident (or diffracted) light with respect to the crystal planes. The result of the constructive interference is the diffraction pattern that is dependent on the size and shape of the unit cell. From the distances between the angles obtained from the scattering pattern, the structural features of the relative positions of atomic centers can be determined, giving the so- called x-ray diffraction structure. µXRD usually requires the small samples to be rotated on an axis by a goniometer. This is because powder and similar types of samples contain crystals of all possible orientations. If the experimental angle is systematically changed, all possible diffraction angles from the powder are detected. The resulting 2D pattern can be summed into a line pattern plot of intensity with respect to angle. Figure 3-1 illustrates this process. These line patterns serve as a fingerprint of the crystal, because each structure produces a unique pattern. Mixtures of crystals produce a pattern that is a sum of all of the patterns that would be seen independently. These patterns were compared to a reference database of powder diffraction files from the International Center for Diffraction Data (ICDD).

All XRD analyses in this thesis were performed with a Rigaku D/max Rapid II diffractometer with a copper anode x-ray tube and 0.3mm collimator. The anode was held either at 40kV and 30mA or at 45kV and 40mA based on the materials being analyzed. Powdered material was adhered to a rotating goniometer either by static to the tip of a wood toothpick (~2cm length) or with Parabar 10312 (Hampton Research) to a glass loop. The sample was analyzed in spin mode (0-360° rotation) at a speed of 10°/sec. The total collection time was verified as reported in each chapter. Rigaku RAPID/XRD software (v.2.4.2) was used for instrument operation and data collection, and Rigaku 2DP software (v.2.0.1.1) was used to select the portion of diffraction rings

39 for interpretation. Rigaku PDXL 2 software (v.2.3.1.0) was used to interpret the diffraction pattern using the ICDD reference database.

3.2.4 X-ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy (XPS) is also referred to as electron spectroscopy for chemical analysis (ESCA) and is used for identifying elements in the surface layers of materials.4 The technique is based on the photoelectric effect, where a material emits electrons when it is irradiated. This surface-sensitive technique only probes the top few nanometers of a sample, depending on the energy of the incident x- rays and the escaping photoelectrons. XPS is used to identify the elements and the oxidative states of a sample. Monochromatic x-rays are generated from a source (such as aluminum), and then strike a sample and eject characteristic core electrons from the atoms. Only electrons at the surface are able to escape, while bulk electrons are attenuated, undergoing either inelastic scattering, excitation, or recombination. The emitted photoelectrons travel to a hemispherical analyzer, traditionally, in an ultra-high vacuum chamber. There have been recent advances in ambient XPS systems.5 By varying the voltage on the hemisphere, the photoelectrons are separated and isolated based on their kinetic energy (as seen in Figure 3-2). The kinetic energy of the electron is directly related to its binding energy in the original atom. The binding energy is calculated from the electron’s kinetic energy, the energy of incident x-rays, and the work function of the instrument.

Figure 3-2 a) Diagram of incident x-rays ejecting a core electron and b) schematic of XPS with incident x-rays hitting the sample and emitted electrons being separated in a hemispherical analyzer.

These analyses produce XPS spectra that plot the counts (per second) of electrons as a function of binding energy in electron volts (eV). Usually, a survey spectrum scanning a wide range (~1000 eV) is collected first to gather possible elements present. Next, high-resolution spectra of smaller ranges (~30 eV) are collected for each possible core electron of an element. The binding energy of the photoelectron indicates the element, the orbital it was ejected from, and the chemical environment. Shifts in the binding energy specify the oxidation state of the elements at the surface of the sample. XPS can be used as a quantitative tool to report the relative ratios of atoms on the surface. The intensities of their signals relate to the atomic sensitivities, which are determined by the instrument’s detection efficiency and the x-ray source. Additionally, XPS can be used in a mapping mode, where the stage is rastered to

41 collect spectra at several points along a sample, giving an overall understanding of the homogeneity of the sample. In Chapter 6, we report XPS analyses of powdered samples and cuttings (1x1 mm) from artificially aged filter papers, which were adhered to carbon tape (Nisshin EM Co., Ltd.) for analysis with a Thermo Scientific K-Alpha+ XPS instrument. This instrument was equipped with a monochromatic Al Kα source (hυ = 1486.6 eV), and it operated at a base pressure of 8x10-9 mbar. For each sample, measurements were taken in three different spots with a 400 µm X-ray spot size and the counts were summed.

An electron flood gun was used to reduce charging effects. Cu 2p high-resolution spectra were collected with 25 scans across the 965-925 eV range. The pass energy was 20 eV at 0.1 eV/step and a 50 second dwell time. Data were collected with Thermo Avantage (v5.962) and analyzed with CasaXPS (version 2.3.18PR1.01), where the spectra were calibrated using the C 1s peak at 284.6 eV.

3.2.5 Time of Flight-Secondary Ion Mass Spectrometry Time of flight-secondary ion mass spectrometry (ToF-SIMS) is a surface- sensitive technique for characterizing a material. It allows mapping of molecular mass fragments from the top layers of a sample. A focused primary ion beam bombards the sample surface. This absorption of energy causes secondary ions, both atomic and molecular fragments, to be ejected from the top nanometers of the sample. ToF-SIMS can be operated in either positive or negative mode, in which either only positively or negatively charged species, respectively, are accelerated into the analyzer. Using the time of flight mass analyzer, all of the incoming ions have the same kinetic energy, but depending on their masses, individual ions take different amounts of time to reach the detector. For example, smaller fragments reach the detector faster than larger

42 fragments traveling slower. This process is illustrated in Figure 3-3. This analysis must be done in ultra-high vacuum to ensure ions can reach the analyzer with minimal collisions and to reduce possible surface contaminations. The output of ToF-SIMS is a mass spectrum at each analysis point plotting the ion counts compared to the mass-to- charge ratio (m/z). By rastering the ion beam, a spectrum is collected at several points across the surface. From this dataset, maps can be generated for specific mass-to- charge ratios corresponding to molecular fragments of interest.

43 SIMS may be used in the static or dynamic mode. Static SIMS is used for surface analysis, because only the uppermost layers emit ions and the damage to the material is insignificant. Dynamic SIMS is used for bulk analysis, because it is coupled with a sputter beam that removes material, then analyzed, and repeated. This is done in order to get an understanding of the three-dimensional composition, or “depth profile”, of the sample. Generally, ToF-SIMS is considered to be a qualitative technique and is challenging to use quantitatively. This characteristic of ToF-SIMS is partially due to the matrix effect, where the matrices in the materials affect the ion yield creating large variances. For the studies reported in Chapter 4, a ToF-SIMS IV with the upgraded capabilities of a ToF-SIMS V (ION-TOF, GmbH) was used to image the paint cross- sections. A bismuth/manganese primary ion source was used to collect all images and spectra in the analysis chamber with a pressure of 5.0 × 10−8 mbar or less. The mode

+ consisting of 25 keV Bi3 clusters of a pre-bunched pulse width of 640 ps and target current of ∼0.18 pA was used. The image pixel density was 128 × 128 pixels with the ion dose density set to the static SIMS limit of 1 × 1012 ions/cm2.6 At the static SIMS limit, <0.1% of the sample surface was removed or damaged. To dissipate a charge build-up on the sample surface, a low-energy (75 eV) electron flood gun was used.

Each mass spectrum was calibrated using ubiquitous ions for both positive and

+ + + + + negative ion modes. The positive ion mass calibrations used: H , H2 , H3 , C , CH ,

+ + + + + + + + CH2 , CH3 , C2H3 , C3H5 , C4H7 , C5H5 , C6H5 , and C7H7 . The negative ion mass

− − − − − − − − − − − − calibrations used: H , H2 , C , CH , CH2 , CH3 , C2 , C2H , C3 , C4 , C5 , C6 , and

− C7 . ION-TOF Measurement Explorer (version 6.2) software was used for all collection and data analysis, and the ion images were normalized by total ion intensity.

44 3.2.6 Fourier-Transform Infrared Spectroscopy Infrared spectroscopy (IR) is a common vibrational spectroscopic technique used in all fields of chemistry. It is used to determine the types of bonds and functional groups of materials.7 Fundamentally, electromagnetic radiation in the infrared region interacts with the materials. Those materials can be solids, liquids, or gases based on the instrument setup of the spectrometer. Because molecular bonds with non-zero dipole moments have characteristic resonant frequencies, the IR radiation with the matching frequency is absorbed by the molecular bonds. The IR light continues on to a detector, which determines the loss of intensity due to absorption by the sample. Spectra are reports of intensity of absorption as a function of the wavelength (or frequency) of the radiation. Frequencies are usually reported as inverse wavelength of the radiation (wavenumber or cm-1). A major improvement to IR spectroscopy resulted from the incorporation of the Michelson interferometer. This results in a faster system that does not rely on scanning individual monochromatic beams of light. In the interferometric technique, the incident light is split so one half interacts with the sample and the other half does not. With a set of moving mirrors, polychromatic light is used but modulated, which results in an interferogram. Using a Fourier transformation, the interferogram is converted into an absorbance or transmittance plot. This adaption is called Fourier- transform infrared spectroscopy (FTIR).

3.2.6.1 Transmittance FTIR

The simplest form of FTIR is done in transmittance mode. In this mode, it measures the amount of incident light transmitted through a sample. For the studies in Chapter 4, the sample was acquired with a stainless steel scalpel with the aid of a

45 stereomicroscope and then placed directly on a diamond cell. The material was rolled flat on the cell with a steel micro-roller to decrease the thickness and increase transparency and signal. The samples are analyzed using a Thermo Scientific Nicolet 6700 FT-IR spectrometer with a Nicolet Continuµm FT-IR microscope in transmission mode. Data were acquired in 128 scans from 4000 to 650 cm-1 at a spectral resolution of 4 cm-1. Spectra were collected with Omnic 8.0 software and analyzed by comparison to various reference spectral libraries.

3.2.6.2 Attenuated Total Reflectance FTIR Attenuated total reflectance (ATR) FTIR is a variant of IR spectroscopy, which generally improves the sample signal. The sample is pressed into a crystal, where the IR light passes into the crystal, then into the sample, and then back out of the sample. At the interface of the crystal and sample, the IR radiation goes through multiple internal reflections based on the materials’ reflective indexes. This improves the signal-to-noise ratio, and it limits the penetration depth of the IR radiation, making it more surface sensitive. ATR-FTIR spectra for the studies in Chapters 6-8 were collected using a Bruker Optic Vertex 70 FTIR spectrometer and Hyperion 2000 Microscope with a single-point ATR crystal attachment made of germanium. Spectra were collected using OPUS 6.0 (v.6.0.72) software, averaging 128 scans in the 4000–

600 cm−1 region with a resolution of 4 cm-1.

3.2.7 Raman Spectroscopy

Raman spectroscopy is a vibrational spectroscopy that provides information on the molecular and crystal structure of materials. It is used for characterization, polymorph differentiation, tracking molecular and crystalline charges, and

46 determining orientation of molecules. This technique can be used for qualitative or quantitative analysis. Raman scattering occurs from the changes in the polarization, where symmetric molecular vibrations have large distortions of the electron cloud induced by laser light.8-9 For Raman spectroscopy, first a laser light is focused on the sample. This produces an amount of Raman scattered light. Generally, most light scattered from matter results from elastic (or Rayleigh) scattering, where there is no change in energy from the incident light. A small percentage (0.000001%) of the light is inelastically scattered and, therefore, has a different energy from the incident light.10 The incident light interacts with the molecules’ electrons and promotes them to a “virtual state”. However, this state is unstable, and the electrons relax, releasing light. Rayleigh scattering occurs when the electrons (promoted from the ground state) return directly to the ground state; thus, there is no difference in the absorbed and emitted photons’ energies. To prevent this form of scattering from saturating the detector when measuring for the inelastic scattering, a notch filter is typically used. Stokes scattering occurs when the electrons (promoted from the ground state) relax to a vibrational state; thus, the photons have less energy. Anti-Stokes scattering occurs when the electrons (promoted from a vibrational state) relax to the ground state, thus the photons have more energy. These different types of scattering that can occur are schematically represented in a Jablonski diagram in Figure 3-4.

Figure 3-4 Jablonski diagram of the energy transitions for Rayleigh and Raman scattering.

Raman spectroscopic measurements generate a spectrum of the intensity of light as a function of the frequency difference, or Raman shift. Peaks in these spectra correspond to specific molecular or lattice vibrations. These can be shifted based on the chemical environment, thus creating spectral “fingerprints” for the materials. The width of the peaks is indicative of the crystalline nature of the sample. For the investigations discussed in Chapter 4, Raman measurements were performed using a single grating RM2000 Renishaw spectrometer coupled with a diode laser source emitting light at 785 nm. The cross-section samples were analyzed with a 50x magnification objective, which provides a laser spot of about 2 µm diameter. A spectral range of 100-1400 cm-1 and a spectral resolution of about 3 cm-1 was used. A laser power of 1% of the total laser power (500 mW) and an acquisition time of 30 seconds with 3 accumulations was used. Spectra were collected using Renishaw WiRE 3.4 software. For the investigations reported in Chapter 5, Raman

48 measurements were performed using a Senterra Raman microscope (Bruker Optics, MA, USA) coupled with a frequency-doubled Nd:YAG laser at 785 nm. The samples were analyzed with a 50× objective lens (Olympus, New York, USA), which provided –1 ∼2 µm spot in diameter. A spectral range of 70–3700 cm and spectral resolution of 3–5 cm–1 was used. A laser power of 1 mW and acquisitions of 240 seconds with 8 co- averages were employed. Raman imaging was accomplished by generating a rectangular grid of Raman spectra on the surface of each sample. The grid dimensions and step size varied with each cross-section. The specific x and y locations of spectral acquisitions within this grid were achieved using the controllable stage in the Raman microscope and the z direction was held constant. OPUS 7.2 software was used for collecting Raman spectra.

3.2.7.1 Multivariate Curve Resolution-Alternating Least Squares Multivariate curve resolution (MCR) is a post-processing technique used in this thesis on Raman spectral mapping data sets.11-12 Using this procedure, one identifies pure signal profiles, such as spectra, of chemical components in an unresolved multicomponent mixture without any spectral input. This process is represented by the matrix equation:

� = � × �! + � where D is the raw dataset, ST are the calculated pure component spectra, C are the related concentration profiles, and E is the error. The method is iterated to minimize the error between the calculated model and the raw dataset. Principal component analysis is used to generate the initial profiles, which are optimized. Requirements for this processing technique are twofold. The data must be two-dimensional and can be reasonably explained using a bilinear model with a limited number of components.

49 Alternating least squares (ALS) is the constraining algorithm assuming non-negative spectra and a linear additive model, for the iterative MCR process. A visual representation of this process is found in Figure 3-5. For analysis done in Chapter 5, MATLAB R2017a 9.2.0.538062 (MathWorks) was used with the PLS_Toolbox 8.2.1 (Eigenvector Research).

Figure 3-5 Schematic representation of multivariate curve resolution-alternating least square algorithm used for Raman spectroscopic post-processing.

50 REFERENCES

1. Glocker, R.; Schreiber, H., Quantitative Röntgenspektralanalyse mit Kalterregung des Spektrums. Annalen der Physik 1928, 390 (8), 1089- 1102.

2. Goldstein, J.; Newbury, D. E.; Michael, J. R.; Ritchie, N. W. M.; Scott, J. H. J.; Joy, D. C., Scanning electron microscopy and x-ray microanalysis. Springer: New York, NY, 2018.

3. Harris, K. D. M.; Tremayne, M.; Kariuki, B. M., Contemporary Advances in the Use of Powder X-Ray Diffraction for Structure Determination. Angewandte Chemie International Edition 2001, 40 (9), 1626-1651.

4. Ratner, B. D.; Castner, D. G., Electron Spectroscopy for Chemical Analysis. In Surface Analysis – The Principal Techniques, Vickerman, J. C.; Gilmore, I. S., Eds. 2009; pp 47-112.

5. Salmeron, M.; Schlögl, R., Ambient pressure photoelectron spectroscopy: A new tool for surface science and nanotechnology. Surface Science Reports 2008, 63 (4), 169-199.

6. Vickerman, J. C., Molecular Surface Mass Spectrometry by SIMS. In Surface Analysis – The Principal Techniques, Vickerman, J. C.; Gilmore, I. S., Eds. 2009; pp 113-205.

7. Smith, B. C., Fundamentals of Fourier transform infrared spectroscopy. CRC Press: Boca Raton, FL, 2011.

8. Smekal, A., Zur Quantentheorie der Dispersion. Naturwissenschaften 1923, 11 (43), 873-875.

9. Gardiner, D. J.; Graves, P. R.; Bowley, H. J., Practical Raman spectroscopy. Springer-Verlag: Berlin; New York, 1989.

10. Long, D. A., The Raman effect : a unified treatment of the theory of Raman scattering by molecules. Wiley: Chichester; New York, 2002.

51 11. Offroy, M.; Moreau, M.; Sobanska, S.; Milanfar, P.; Duponchel, L., Pushing back the limits of Raman imaging by coupling super-resolution and chemometrics for aerosols characterization. Scientific Reports 2015, 5, 12303.

12. Smith, J. P.; Smith, F. C.; Krull-Davatzes, A. E.; Simonson, B. M.; Glass, B. P.; Booksh, K. S., Raman microspectroscopic mapping with multivariate curve resolution-alternating least squares (MCR-ALS) of the high-pressure, α-PbO2-structured polymorph of titanium dioxide, TiO2-II. Chemical Data Collections 2017, 9-10, 35-43.

52 Chapter 4

CHARACTERIZATION OF GREEN PAINTS IN MING & QIAN DYNASTIES LIN’XI PAVILION

4.1 Introductory Remarks

This chapter details the case study of the painted ceiling of Lin’xi Pavilion in the Forbidden City. The ceiling decoration consists of two distinct paint generations, which serves as a unique case study into architectural paint materials during both the Ming and Qing dynasties. Paint samples and cross-sections from both paint generations were analyzed with SEM-EDX, ToF-SIMS, XRD, FTIR, and Raman spectroscopies. In particular, the green pigments were identified to better understand the copper salts used in Chinese architecture at the time to inform historical records and conservation treatments. In both paint generations, botallackite and atacamite polymorphs were identified. This evidence puts the case study in a larger context, where a shift from natural mineral sources to synthetic copper-based pigments was taking place for these larger architectural projects. This work has been submitted to the journal Studies in Conservation in 2019. The authors of this work were Marcie B. Wiggins, Liu Mengyu, Catherine Matsen, Liu Chang, and Karl Booksh.

4.2 Introduction Lin’xi Pavilion (Lin Xi Ting) is a landscape building located in the southern part of the Cining Palace Garden, or the Garden of Compassion and Tranquility, found in the northwestern part of the Forbidden City, Beijing (Figure 4-1). It served as the

53 residence and leisure area for the dowager empress and concubines of the Ming and Qing dynasties. The garden was built in the mid-1500s (Jiajing Era, Ming Dynasty), but the Lin’xi Pavilion was constructed later. According to Sun Chengze's Chun Ming Meng Yu Lu (春明梦余录): “Lin Xi Hall was built in the 6th year of Wanli Era, and was renamed as Lin Xi Ting in the May of the 11th year of Wanli Era.”1 This corresponds to 1578 when the pavilion would have been completed. The History of the Palace of the State (Guo Chao Gong Shi, 国朝宫史), written in the 7th year of

Qianlong Era (1742) also referred to this building: “There is a pool in the garden and

Lin’xi Pavilion is in front of the pool.”2 The building also appeared in a map of the Forbidden City of Qianglong Era.3 Thus, it can be inferred that the pavilion survived in the original location through the Ming and Qing Dynasties.

Figure 4-1 Exterior of Lin’xi Pavilion, which is located in the Forbidden City.

54 According to the official historic archives of the Qing Dynasty, there were several restorations in the Cining Palace area during the Qing Dynasty. The largest renovation took place in the 30th year of Qianlong Era (1765). In this year, the Qianlong Emperor intended to restore Lin’xi Pavilion.4 Additionally, there were some smaller maintenance projects during the 17-19th centuries.3 During treatment in 2015, the polychrome ceiling decoration of Lin’xi Pavilion was preserved according to extensive historical research. It can be deduced from the pattern style that the ceiling decoration was originally painted in the Ming Dynasty.

However, the details of the paintings also have some distinctive features of the mid- Qing Dynasty architectural paintings. The ceiling decoration contains various motifs, including dragons, phoenixes, clouds, and flames (Figure 4-2a). Green paints were predominantly found in the background of the dragon area, as well as in the green clouds and flames. Green paints from different areas are similar in color; they are not bright but rather have a saturated, pronounced, and uniformed appearance.

Figure 4-2 (a) The Lin’xi Pavilion ceiling painting (297 x 297 cm) prior to treatment. The yellow box indicates the region (b), where samples were taken.

55 The identity of the green pigments used in this ceiling painting was of interest to the conservators and art historians. Literature on artists’ materials from the 16th-18th centuries suggest green pigments sources were either “mineral green” or “copper green”. The term “mineral green” was used for mined basic copper carbonates, malachite (Cu2CO3(OH)2) and azurite (Cu3(CO3)2(OH)2), while “copper green” was an umbrella term for other copper-based pigments, such as verdigris

5,6 (Cu(CH3COO)2H2O), atacamite (Cu2Cl(OH)3), etc. Determining which pigment was used in this ceiling painting would be informative of the artists’ materials available and used for this large-scale architectural project in both the Ming and Qing dynasties, thus informing the source and cost of the ceiling paint materials. To identify and study these materials, we first utilized scanning electron microscopy-energy dispersive x-ray spectroscopy (SEM-EDX) and time of flight- secondary ion mass spectrometry (ToF-SIMS) to map the elemental and molecular fragment distribution across the cross-sections. These techniques identify layered materials within paintings and other cultural heritage objects, first for inorganics and then for organics.7-9 Following these analyses, characterization of pigment particles and some organic material was carried out using X-ray diffraction (XRD) and infrared and Raman spectroscopies, with specific interest in the green pigments. Again, these methods are commonplace in studying cultural heritage objects for the identification of artists’ materials to guide further treatments and studies.

4.3 Materials As part of a recent conservation project, the decorated ceiling fabric of the Lin’xi Pavilion was removed from the structure and moved to the conservation lab of the Department of Architectural Heritage at the Palace Museum. The existing ceiling

56 decoration was separated into two paint layers suggesting it was comprised of two different paint campaigns (generations). Therefore, this object provided a good opportunity for historical study for comparing the two generations of Ming and Qing painting. The two generations have a similar structure: a supporting substrate layer followed by a paint layer, and finally a gilded layer in certain areas, as seen in Figure 4-3. The supporting substrate layer was made of many layers of paper and textile, pasted together. After painting and gilding, the entire sheet was pasted on a wooden panel, and then installed onto the timber structure of the building. This is a traditional technique to make ceiling decorations for Chinese architecture.10 In the Qing dynasty restoration campaign (c. 1765), the second generation’s supporting layer was made of silk and paper and then painted. The goal was not to remove old layers, but to paste the new supporting layer on to the old surface, thus creating the second-generation painting on top of the first generation, which is how it remained until the 21st century.11 As a part of that addition, the same patterns were painted on the new supporting layer with some slight variations. In some later, small-scale conservation efforts, patches were made on the second generation of painting to cover damage and deterioration.11

Two green paint samples were taken from adjacent areas of the east side of the ceiling but from different generations for comparison. The samples were cast separately in mini-cubes, of approximately half inch widths, with polyester resin (Extec polyester clear resin (methyl methacrylate monomer) with a methyl ethyl ketone peroxide catalyst (10mL : 8 drops), Extec Corporation®, Enfield, CT). The resin was allowed to cure for 24 hours at room temperature under ambient light. Excess casting medium was removed from the cube just up to the surface of the sample with a jeweler’s saw (Rio Grande saw blades, laser gold). The cubes were then dry, hand-polished successively with 400- and 600-grit Buehler Carbimet paper

58 (silicon carbide) and 1500- to 12,000-grit Micro-Mesh Inc. polishing cloths (silicon carbide or aluminum oxide) to expose the cross-section. Sample LXT-04 was taken from the first generation paint layer of the ceiling at an area of loss of the second- generation that exposed the underlying, original painting (Figure 2b). Sample LXT-05 is from an area of green and gilded clouds (Figure 2b), taken from the second- generation painting surface, corresponding to the later 18th c. restoration layer.

4.4 Methodology

4.4.1 Visible and Ultraviolet Microscopy

To establish finish stratigraphies of the paint cross-sections from the above- mentioned sample locations, the samples were examined and digitally photographed using a Zeiss Axio Imager M2m binocular microscope (20× objective) equipped with a Kübler Codix HXP 120C mercury lamp for reflected visible and ultraviolet light. The samples were viewed in dark-field reflected light using the Zeiss 02 cube (excitation 365nm, barrier 420nm, beam splitter 395nm). Images were taken with the Zeiss AxioCam HRc digital camera using Zeiss AxioVision software (v.4.9.1.0).

4.4.2 Polarized Light Microscopy (PLM) Polarized light microscopy (PLM) is a common method of pigment identification that is based on the optical properties of the crystals. For PLM, pigment particles were scraped from the bulk paint with a scalpel, transferred to a glass slide, and covered with a coverslip. The heat-reversing mounting medium, Meltmount™ (refractive index of 1.662 at 25°C), was melted at approximately 70°C and then injected between the cover slip and the slide. After cooling, the pigment particles were dispersed in the mounting media. A Nikon LV100ND microscope with a polarized

59 attachment and a DS-Ri2 microscope camera was used for sample observation and photomicrography, using NIS-Elements software.

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

For SEM-EDX analysis, the cross-sections were mounted to an SPI Supplies Zeiss aluminum slot head stub (12.7 × 3.1 mm) with SPI Supplies double-sided carbon tabs (12 mm diameter). SPI Supplies conductive carbon paint (20% w/w colloidal graphite in isopropanol) was applied on the side and top surfaces of casting media, without covering the cross-section surface itself, to prevent charging effects. The samples were examined using a Zeiss EVO MA15 scanning electron microscope with a LaB6 source at an accelerating voltage of 20kV for the electron beam, a working distance of approximately 8mm, and a sample tilt of 0°. The EDS data were collected with the Bruker Nano X-flash® detector 6│30 and analyzed with Quantax 200/Esprit 1.9 software.

4.4.4 Time of Flight-Secondary Ion Mass Spectrometry As previously established by Voras et al.8 and deGhetaldi et al.,7 a ToF-SIMS IV with the upgraded capabilities of a ToF-SIMS V (ION-TOF, GmbH) was used to image the paint cross-sections. A bismuth/manganese primary ion source was used to collect all images and spectra with a pressure of 5.0 × 10−8 mbar or less. The mode

+ consisting of 25 keV Bi3 clusters of a pre-bunched pulse width of 640 ps and target current of ∼0.18 pA was used. The image pixel density was 128 × 128 pixels, with the ion dose density set to the static SIMS limit of 1 × 1012 ions/cm2. At the static SIMS limit, < 0.1% of the sample surface was removed or damaged.12 To dissipate a charge build-up on the sample surface, a low-energy (75 eV) electron flood gun was used.

60 Each mass spectrum was calibrated using ubiquitous ions for both positive and

+ + + + + negative ion modes. The positive ion mass calibrations used: H , H2 , H3 , C , CH ,

+ + + + + + + + CH2 , CH3 , C2H3 , C3H5 , C4H7 , C5H5 , C6H5 , and C7H7 . The negative ion mass

− − − − − − − − − − − − calibrations used: H , H2 , C , CH , CH2 , CH3 , C2 , C2H , C3 , C4 , C5 , C6 , and

− C7 . ION-TOF Measurement Explorer (version 6.2) software was used for all collection and data analysis, and the ion images were normalized by total ion intensity.

4.4.5 X-ray Diffraction (XRD) XRD was carried out using a Rigaku D/max Rapid II diffractometer with a copper anode x-ray tube and 0.3 mm collimator. Isolated material from the original paint layer (LXT04) was adhered to a glass loop by Parabar 10312 liquid (Hampton Research) and secured to the sample stage. The sample was analyzed by rotating phi (0-360° rotation) at a speed of 10°/sec with omega held at 0°. Collection time was 3 hours with an x-ray tube held at 40kV, 30mA. The XRD pattern for the second- generation layer (LXT05) was collected from the cast cross-section. The x-ray beam (45kV, 40mA) was focused on the green paint layer of the cross-section and the sample was oscillated phi from 42 to 49° at a speed of 1°/sec with omega held at 45° for a collection time of 8 hours. Rigaku RAPID/XRD software (v.2.4.2) was used for instrument operation and data collection, and Rigaku 2DP software (v.2.0.1.1) was used to select the portion of diffraction rings for interpretation. Rigaku PDXL 2 software (v.2.3.1.0) was used to interpret the diffraction pattern, and the powder diffraction file from the International Center for Diffraction Data (ICDD) was used as a reference database.

61 4.4.6 Raman Spectroscopy Raman measurements were performed using a single grating Invia Renishaw spectrometer coupled with a diode laser source emitting light at 785 nm, determining a spectral resolution of about 3cm-1. The cross-section samples were analyzed with a 50x magnification objective, which provides a laser spot of about 2 µm diameter. The spectra were recorded in the range 100-1400 cm-1, with a laser power on the sample of 1% of the total laser power, 500 mW, and an acquisition of 30 sec. with 3 accumulations. Spectra were collected using Renishaw WiRE 3.4 software.

4.4.7 Fourier-Transform Infrared Spectroscopy

The Fourier-transform infrared microspectroscopy (FTIR) sample was acquired with a stainless steel scalpel from the backing silk of a paint scraping from LXT05 with the aid of a stereomicroscope and then placed directly on a diamond cell. The material was rolled flat on the cell with a steel micro-roller to decrease thickness and increase transparency. The sample was analyzed using the Thermo Scientific Nicolet 6700 FT-IR with Nicolet Continuµm FTIR microscope in transmission mode. Data was acquired for 128 scans from 4000 to 650 cm-1 at a spectral resolution of 4 cm-1. Multiple scrapings of the sample were taken from the sample and multiple spectra were taken from different areas within each scraping. Spectra were collected with Omnic 8.0 software and analyzed in this program with various IRUG and commercial reference spectral libraries.

4.5 Results To establish the layered structure of the paintings, the cross-sections were examined and documented (Figure 4-3). In both LXT-04 and LXT-05, textile substrate is clearly seen just below the paint layer, denoted in Figure 4-3 as (b). With LXT-05,

62 the cellulosic paper support, denoted by (a) in Figure 3, is seen below the textile substrate. This sequence corroborates the above-mentioned process of the second generation paint and textile being applied to a paper support prior to being adhered to the ceiling. The green paint layers in both generations contained spherical green particles, approximately 40 µm in diameter. Specific to the newer generation (LXT- 05), thin layers of gold leaf (d) exist both above and below the green paint layer. Also, the uppermost paint layer is thin (~20 µm) and white (e). In ultraviolet light illumination, an organic material, such as shellac or a natural resin, just below the green paint layer is evident by its orange fluorescence in some areas of the cross- section. It is more pronounced in regions with fewer green pigment particles above it (not shown). Scrapings from the pigmented layers were analyzed with polarized light. In plane-polarized light, the pigment particles appear as a fine mesh of translucent green, rounded crystals. They were in high relief with a lower reflective index (RI) than the medium (Figure 4-4a). Under crossed polars, the samples have moderate birefringence. Some particles appear to be colors produced by second-order interference (Figure 4-4b). This observation indicates the presence of atacamite.13,14

Figure 4-4 Dispersed sample of the pigments from LXT-04 at 500x magnification in a) plane polarized transmitted light and b) crossed-polarized light.

The cross-sections were studied with SEM-EDX to determine the elemental distribution across the layers (Figure 4-5). The data show the presence of copper and chlorine mapping together within the pigment particles from the two generations. In both samples, calcium and sulfur were detected in the textile fibers. In addition, for LXT-04, potassium and silicon were found in the areas of some slightly transparent particles in the paint layer. For LXT-05, the upper white paint layer is rich in lead.

Figure 4-5 SEM-EDX SEM-EDX false color elemental maps of elements for LXT-04 (i and ii) and LXT-05 (iii and iv) with layers for silk (b), green paint (c), and white paint (e).

To identify molecular species in the stratigraphies, ToF-SIMS was used to map

- mass fragments in the cross-sections. The mass fragment CaSO4 , found in the fiber substrates in negative mode, indicates a calcium sulfate ground and is consistent with the EDX elemental maps of both paint generations (Fig. 4-6). Copper chloride

- - trihydroxides were identified in the green paint layer by the presence of CuCl2 , CuCl ,

- + + Cl , Cun , and Cu2OH fragments. These fragments correspond to the pigment particles in both cross-sections, which were previously attributed to the colorant atacamite by Richardin et al.15 and Kim et al.16 Atacamite is a common copper

65 chloride trihydroxide pigment, but ToF-SIMS cannot determine the specific polymorphs alone.

Figure 4-6 ToF-SIMS maps of mass fragments for LXT-04 (i) and LXT-05 (ii).

Additionally, positively charged amino acid mass fragments of glycine

+ + + + + (CH4N ), alanine (C2H6N ), proline (C4H8N ), and valine (C4H10N and C5H7O ) were detected in the substrate region due to the silk textile. The silk fibers, composed of fibroin, contain glycine, alanine, and serine.17 The presence of the remaining amino

+ acids and the unique hydroxyproline mass fragment (C4H8NO ) indicates the source of amino acids in the low layers is from an additional protein, animal glue.8 The source of hydroxyproline would not be the silk substrate, but a fish or mammalian glue.18

66 This practice agrees with known artistic methods for ceiling fabric paintings.10 Trace

- amounts of fatty acid markers, such as for palmitic acid (C16H31O2 ), were found, but this is likely just from the glue. However, organic fragment markers for protein and oils were not detected in regions that corresponded to the green and white paint layers in the cross-sections.7,8 Therefore, the paint binder could not be directly identified. XRD and Raman spectroscopy were used to specify the copper chloride trihydroxide polymorph (Cu2(OH)3Cl). XRD patterns for both LXT-04 and LXT-05 show a mixture of botallackite and atacamite crystals.19 Both are polymorphs, although botallackite is less common in nature.6,20,21 Figure 4-7 shows the botallackite and atacamite comparison of XRD patterns for LXT-04. The presence of botallackite was verified as well with Raman spectroscopy.22 Compared to the RRUFF database standard for botallackite (R070066), characteristic bands at 400, 450, and 510 cm-1 were found for both samples. Additionally, weak characteristic bands at 820, 910, and 975 cm-1 for atacamite (R050098) were also found for LXT-04 (Figure 4-8).

Figure 4-7 XRD patterns of LXT-04 and references botallackite and atacamite.

Figure 4-8 Raman spectra of LXT-04 and LXT-05 compared to reference spectra of atacamite and botallackite.

4.6 Discussion

4.6.1 Organics Materials

The organic materials were identified spatially on the cross-sections using ToF-SIMS and with supplemental verification on isolated material by GC-MS and FTIR spectroscopy. Both generations of paint were applied to the ceiling on a textile substrate, which was identified as silk. These silks were both treated with a common ground material, calcium sulfate, and animal glue as a binder. The cross-section of the second generation, LXT-05, shows the paper substrate placed between the two generations of paint.10 However, neither paint layer shows any unique mass fragments for protein or oil. With GC-MS (not shown) and ToF-SIMS, fatty acids were identified but these are likely just trace amounts resulting from the animal glue.

68 The microscopic analysis using ultraviolet light showed one material in the second generation of paint had an orange fluorescence. This phenomenon suggested the presence of shellac just below the paint layer. FTIR spectroscopy verified a natural resin on the back of the fibers, based on a carbonyl band at 1708 cm-1 and C-H stretching at 2930 and 2857 cm-1.23 However, the presence of shellac was not verified with GC-MS (not shown). Shellac is surprising to find as a material used in the second-generation, 17th century paint.24,25 It is possible this natural resin found at the edge of the cross-section with little or no paint was a part of a later restoration.

4.6.2 Inorganic Materials Calcium sulfate was detected in the silk substrate and must have served as a ground material for paint layers. This practice is consistent with traditional painting on silk from the 15th to the 18th centuries.10 The ceiling was predominantly painted green in both the first and second generations. Based on the previously mentioned analyses with Raman spectroscopy and XRD, the Lin’xi Ting ceiling is painted with a mixture of botallackite and atacamite in both generations. To our best knowledge, botallackite has rarely been identified in Chinese architectural painting. Botallackite has been found in wall paintings in the Furi Temple in Gansu Province and the Five Northern Provinces’

Assembly Hall in Shaanxi Province.13,26 Atacamite and another polymorph, paraatacamite, have also been found in Chinese works.5,27 Previously, botallackite has been found along with atacamite in cave murals.28,29 The challenges in identifying botallackite have previously been: 1) the low sample population of Chinese architecture, 2) the difficulty to differentiate botallackite and atacamite visually or with PLM, 3) the lack of botallackite reference Raman spectra, and 4) the small

69 sample size for XRD.13 As a result, copper chloride crystals cannot always be correctly distinguished as atacamite or botallackite and, more commonly, the copper chloride trihydroxide crystals are categorized as only atacamite or “copper green”. Because of this lack of distinction in previous studies, there is likely more botallackite present in Chinese architectural paints than is currently documented. The identification of both botallackite and atacamite in the 16th c. (first generation) and 18th c. (second generation) samples for the Lin’xi Pavilion indicates the historical usage of “copper green” pigments. The term “copper green” refers to many types of copper-based pigments, such as copper chlorides and copper acetates (commonly referred to as verdigris).5 Alternatively, historic green and blue pigments are referred to as “mineral green” or “mineral blue” for basic copper carbonates, malachite (Cu2(CO3)2(OH)2) and azurite (Cu2CO3(OH)2), respectively. Mineral green was commonly used before the 18th century in Chinese architectural painting.5 “Copper green” pigments also appeared at that time, but after the 18th century they were more common than “mineral green”.14 In this case study, both generations from the 16th and 18th centuries lack any “mineral green” pigments, and there is only atacamite and botallackite used for the large ceiling paintings. The mineral reserves of malachite decreased with time, which caused the price to increase. Hence, craftsmen would have been more inclined to use copper green instead. Primary literature suggests the prices for “copper green” were cheaper than for “mineral green”,30 and since a large quantity of the pigment was needed in the restoration, the less expensive option was chosen. During the 15th century, recipes for synthesizing copper chloride trihydroxide pigments appear in the literature, which is the likely cause for decreased costs and increased usage by artisans of that period.5

70 The two generations of green paint are similar in composition, but the first generation of paint (LXT-04) contains silicon-containing particles equal in size to the copper-based particles. The SEM-EDX image also shows smaller potassium- containing particles above the green paint particles. This is possibly a result of dirt accumulated over the years and then trapped under the second generation of paper, silk, and paint. It is also possible the silicon and/or potassium were originally an additive to the green paint layer. The second generation of paint (LXT-05) involved more decoration, which can be seen in the layers of white paint and gold leaf. From the microscopic images, the thin layers of gold were identified both above and below the green pigment particles. These materials correspond to the golden cloud motif on the edges of the ceiling design. The thin lead white layer was mixed with some green pigment to lighten it and painted directly on top of the upper most gold. This lighter outline around the motif is referred to as Tuiyun. This resulted in a more decorative and elaborate second- generation ceiling to Lin’xi Pavilion.

4.7 Conclusions Using SEM-EDX, ToF-SIMS, FTIR, Raman, and XRD, an extensive characterization of the Lin’xi Ting ceiling painting has documented the materials used in both the Ming and Qing dynasty campaigns. The materials were very similar across the two paint generations. Both contained a silk and paper substrate, which had been treated with calcium sulfate and animal glue. The green paint pigments were identified as a mixture of atacamite and botallackite polymorphs. The second paint generation includes gold leaf and lead white paint, as well.

71 The presence of atacamite and, predominantly, botallackite opens a discussion regarding the copper-based green pigments used in China at this time. As atacamite and botallackite would be considered “copper green”, this suggests this was a more common material for architectural painting during the Ming and Qing dynasties than was previously thought. Malachite would have decreased in usage during this time period, and new syntheses of atacamite and botallackite suggest they would have been a cheaper, preferred option for craftsmen.

4.8 Acknowledgements I would like to thank the staff in Winterthur Museum, Garden & Library’s Scientific Research & Analysis Laboratory for their instrumentation and knowledge and also the Research Fellowship Program. I would also like to thank Dr. Zachary Voras and Chris Goodwin for their assistance with ToF-SIMS analysis and the National Science Foundation (NSF) for support of the ToF-SIMS instrument (DMR- 9724307) and the UD Surface Analysis Facility (NIH NIGMS COBRE (P30- GM110758)).

4.9 Concluding Remarks

The characterization of botallackite and atacamite, as well as other compounds, in this chapter shed important light on the material history and conservation needs of this historical landmark. Being able to distinguish these basic copper chlorides in artworks is of great interest in the field of conservation science due to their limited study. In addition to basic copper chlorides affecting conservation, these identifications bring to light the art historical shift from natural to synthetic copper

72 salts for painting. Additional analysis into atacamite and botallackite polymorphs in Chinese architectural paints can be found in the next chapter.

73 REFERENCES

1. Sun, C. 孙., Chun Ming Meng Yu Lu（春明梦余录). Beijing Guji Press: Beijing, China, 1992.

2. Ertai, History of the Palace of the State (Guo Chao Gong Shi, 国朝宫史). Beijing Guji Press: Beijing, China, 1992; Vol. 13.

3. Chang, X., Architecture in the Area of Cining Palace. In Proceedings of the Chinese Society of the Forbidden City, Book 8, 2012; pp 302-317.

4. Wen, M., Sightseeing in Cining Palace Garden. Forbidden City 2015, 07, 70-85.

5. Li, M. Analysis and Research on Copper Green Pigment. Northwest University, Xi’an, Shaanxi,China, 2013.

6. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

7. de Ghetaldi, K.; Wiggins, M. B.; Bertorello, C.; Voras, Z.; Norbutus, A.; Beebe Jr, T. P.; Baade, B., In-depth examination and analysis of Domenico Cresti's oil on wall paintings in Santa Maria della pace in Rome. Journal of Cultural Heritage 2017, 28, 48-55.

8. Voras, Z. E.; deGhetaldi, K.; Baade, B.; Gordon, E.; Gates, G.; Beebe, T. P., Comparison of oil and egg tempera paint systems using time-of-flight secondary ion mass spectrometry. Studies in Conservation 2016, 61 (4), 222-235.

9. Keune, K.; Boon, J. J., Imaging Secondary Ion Mass Spectrometry of a Paint Cross Section Taken from an Early Netherlandish Painting by Rogier van der Weyden. Analytical Chemistry 2004, 76 (5), 1374-1385.

10. Jiang, B., Materials and Techniques of Paste and Mounting in Ming and Qing Dynasties. Traditional Chinese Architecture and Garden 1992, 03, 12-16.

11. Personal communication with Y. Li. 2015.

74 12. Vickerman, J. C., Molecular Surface Mass Spectrometry by SIMS. In Surface Analysis – The Principal Techniques, Vickerman, J. C.; Gilmore, I. S., Eds. 2009; pp 113-205.

13. Hu, K.; Bai, C.; Ma, L.; Bai, K.; Liu, D.; Fan, B., A study on the painting techniques and materials of the murals in the Five Northern Provinces’ Assembly Hall, Ziyang, China. Heritage Science 2013, 1 (1), 18. https://doi.org/10.1186/2050-7445-1-18

14. Yong, L., Copper trihydroxychlorides as pigments in China. Studies in Conservation 2012, 57 (2), 106-111.

15. Richardin, P.; Mazel, V.; Walter, P.; Laprévote, O.; Brunelle, A., Identification of Different Copper Green Pigments in Renaissance Paintings by Cluster-TOF-SIMS Imaging Analysis. Journal of The American Society for Mass Spectrometry 2011, 22 (10), 1729-1736.

16. Kim, M.; Lee, J.; Doh, J.-M.; Ahn, H.; Kim, H. D.; Yang, Y.; Lee, Y., Characterization of ancient Korean pigments by surface analytical techniques. Surface and Interface Analysis 2016, 48 (7), 409-414.

17. Altman, G. H.; Diaz, F.; Jakuba, C.; Calabro, T.; Horan, R. L.; Chen, J.; Lu, H.; Richmond, J.; Kaplan, D. L., Silk-based biomaterials. Biomaterials 2003, 24 (3), 401-416.

18. Gustavson, K. H., The Function of Hydroxyproline in Collagens. Nature 1955, 175 (4445), 70-74.

19. Zheng, X. G.; Mori, T.; Nishiyama, K.; Higemoto, W.; Yamada, H.; Nishikubo, K.; Xu, C. N., Antiferromagnetic transitions in polymorphous minerals of the natural cuprates atacamite and botallackite Cu2Cl(OH)3. Physical Review B 2005, 71 (17), 174404.

20. Krivovichev, S. V.; Hawthorne, F. C.; Williams, P. A., Structural complexity and crystallization: the Ostwald sequence of phases in the Cu2(OH)3Cl system (botallackite-atacamite-clinoatacamite). Struct Chem Structural Chemistry : Computational and Experimental Studies of Chemical and Biological Systems 2017, 28 (1), 153-159.

21. Pollard, A. M.; Thomas, R. G.; Williams, P. A., Synthesis and stabilities of the basic copper(II) chlorides atacamite, paratacamite and botallackite. Mineralogical Magazine 1989, 53 (373), 557-563.

75 22. Martens, W.; Frost, R. L.; Williams, P. A., Raman and infrared spectroscopic study of the basic copper chloride minerals -implications for the study of the copper and brass corrosion and "bronze disease". Neues Jahrbuch für Mineralogie - Abhandlungen: Journal of Mineralogy and Geoche 2003, 178 (2), 197-215.

23. Derrick, M. R.; Stulik, D.; Landry, J. M., Infrared spectroscopy in conservation science. Getty Conservation Institute: Los Angeles, 1999.

24. McGowan-Jackson, H., Shellac in Conservation. AICCM Bulletin 1992, 18 (1-2), 29-39.

25. Returns of Trade at the Port of Canton, For the Year 1866. Custom, C. M., Ed. 1866; Vol. 2, p 847.

26. Xia, Y.; Wang, W. F.; Liu, L. X., Study on mural painting pigments of Fuxi Temple, Tianshui, Gansu Province. Science Conservation Archaeology 2011, 23, 18-24.

27. Cheng, X.; Yang, Q., Micro-Raman spectroscopy study of three green pigments containing Copper and Arsenic. Science of Conservation and Archaeology 2015, 27 (3), 84-89.

28. Fan, Y.; Chen, X.; Li, Z.; Hu, Z., Micro diffraction analysis of the rare green pigment botallackite in ancient wall paintings. Journal of Lanzhou University 2004, 40 (5), 52-55.

29. Egel, E.; Simon, S., Investigation of the painting materials in Zhongshan Grottoes (Shaanxi, China). Heritage Science 2013, 1 (1), 29.

30. Regulations and precedents on prices of pigments, issued by the Ministry of Revenue（户部颜料价值则例). 1741.

76 Chapter 5

POLYMORPH IDENTIFICATION IN GREEN CHINESE ARCHITECTURAL PAINTS USING RAMAN IMAGING AND MCR-ALS

5.1 Introductory Remarks

In this chapter, several Chinese architectural paint samples are analyzed from different locations and time periods to determine any trends indicative of the synthesis or origin of these materials. Due to their previous success, Raman imaging and MCR- ALS are used to distinguish the polymorphs present and the relative ratios of those polymorphs. The differing ratios of atacamite and botallackite indicate variations in their preparation methods related to their time periods. This work is being prepared for submission to a peer-reviewed journal, such as Applied Spectroscopy. The author list will consist of Marcie B. Wiggins, Liu Mengyu, Liu Chang, and Karl S. Booksh.

5.2 Introduction Differentiation of polymorphic forms of pigments is informative for conservation professionals about the origins of materials. Polymorphic forms have also been suggested to play a role in the stability of the paint.1 It is also beneficial to be able to distinguish materials in trace amounts or with weak signals within the mixture.2 Therefore, the methodology to distinguish component spectra of polymorphs is essential for defining the composition of paint mixtures.

Copper chloride trihydroxide (Cu2(OH)3Cl) is used as green colorants in artwork.3-6 This compound is known to have different crystal states, for example

77 orthorhombic atacamite and monoclinic botallackite. Atacamite is the naturally stable common polymorph of this material.7 Basic copper chlorides are frequently found in cultural heritage objects, but they are normally associated with corrosion of copper materials.4, 8-11 Artists’ treatises frequently refer to it as “salt green,”3, 12 and it has been found sporadically as a pigment in Western art.13-15 However, it is more widely used in Eastern art, specifically in China. Basic copper chlorides have been found in mural paintings and on painted objects throughout China.4-5, 16-17 Atacamite and, rarely, botallackite have also been found in Chinese architectural painting, which is the focus of this study.6, 18-19 Copper chloride trihydroxides fall under the umbrella term of “copper green” pigments used in Chinese cultural heritage. Additionally, there are “mineral green” or “mineral blue” pigments, which are the natural copper basic carbonates, malachite

20 (Cu2CO3(OH)2) and azurite (Cu3(CO3)2(OH)2), respectively. Both mineral pigments were in common use until around the 15-16th centuries. Artwork originating from later periods has been found to contain more copper chloride trihyroxides pigments.5 This suggests a shift to incorporation of synthetic pigments, like “copper green”, for Chinese architectural paints. Historians and conservators are interested in the objects’ exact materials in order to learn more about their origin and condition. Thus, there is a need to distinguish these polymorphs or other trace materials present in the paint to learn about the use of basic copper chlorides during this transition period. Additionally, techniques that can identify these polymorph mixtures with minimal or no damage to these priceless samples are preferred. Initial investigations of these pigments often only suggest the presence of the atacamite polymorph; however, subsequent investigations have shown the presence of

78 botallackite as well. These polymorphs can be distinguished using Raman spectroscopy and x-ray diffraction (XRD).6 Therefore, the presence of botallackite and the implications regarding the historical production of pigments in cultural artifacts warrant further investigation. Based on previous work on the green pigments at Lin’xi Pavilion, we seek in this study seeks to understand the polymorph mixtures in “copper green” pigments being used in this transition period.5, 20 To do so, Raman spectroscopic mapping was coupled with mixture resolution using multivariate curve resolution-alternating least squares (MCR-ALS) to distinguish the components. The method has previously been effective in distinguishing spectra of trace polymorphs, which suggests that it would be useful for identifying the polymorphs present and their locations within the architectural paint samples. 21-22 Furthermore, this methodology enables estimation of relative ratios of the polymorphs within the sample, which information would suggest possible synthesis methods for these pigments in China.

5.3 Materials Six green and blue architectural pigment samples were analyzed. The samples covered a range of locations and time periods. This information is summarized in Table 5-1, along with images of the paint cross-sections. Raman maps were first attempted on paint scrapings from several of the architectural features. However, due to instrumental limitations (principally the lack of autofocus capability of instrument used), the rough topography of the samples prevented automated collection of Raman spectral maps. Instead, the samples were analyzed as embedded, polished cross-sections. These leveled surfaces improved map collection.

79 Different laboratories prepared the paint samples, and therefore, their embedding practices varied slightly. YHD-13, LXT-04, and LXT-05 were cast separately in mini-cubes of approximately half-inch widths. Extec polyester resin (methyl methacrylate monomer with methyl ethyl ketone peroxide catalyst, at a 10 mL:8 drops ratio) (Extec Corporation®, Enfield, CT) was used for these samples. The resin was allowed to cure for 24 hours at room temperature and under ambient light. NHS02, YXD02, and W01 were cast separately into blocks, approximately half-inch by quarter-inch by quarter-inch. Epo-Fix polyester resin (bisphenol-a-diglycidylether with triethylenetetramine catalyst, at a 4:1 mL ratio) (Electron Microscopy Sciences, Hatfield, PA) was used for these samples. The resin was allowed to cure for 24 hours at room temperature and under ambient light. For YXD02, there was enough material to cast two separate cross-sections from that area, referred to as S1 and S2. For all the embedded samples, excess casting medium was removed from the cube just up to the surface of the sample with a jeweler’s saw (Rio Grande saw blades, laser gold). All resin blocks were then hand-polished successively with 400- and 600-grit Buehler Carbimet paper (silicon carbide) and 1500- to 12,000-grit Micro-Mesh Inc. polishing cloths (silicon carbide or aluminum oxide) to expose the cross-section. The cross-section samples were examined and digitally photographed using a

Zeiss Axio Imager M2m binocular microscope (5, 10×, 20×, and 50× objectives) equipped with a Kübler Codix HXP 120C mercury lamp for reflected visible and ultraviolet light. The samples were viewed in dark field reflected light and using the Zeiss 02 cube (excitiation 365nm, barrier 420nm, beam splitter 395nm). Images were taken with the Zeiss AxioCam HRc digital camera in conjunction with Zeiss AxioVision software version 4.9.1.0.

80 Table 5-1 List of paint cross-section samples analyzed in this study. This includes with their locations and time periods, as well as images of the cross-sections.

Sample ID Building (Location) Time Period Cross-Section Image LXT04 Lin’Xi Ting 16th century (Forbidden City)

LXT05 17th-19th century

NHS02 Nianhua Temple 18th-19th century (Beijing)

YHD13 Ying Hua Dian 16th-17th century (Forbidden City)

YXD02 Annex of Yangxin 15th-16th century Dian (Forbidden City)

S1:

S2: W01 Yanshan Temple 12th century (Shanxi Province)

81 5.4 Methodology A Senterra Raman spectrometer (Bruker Optics) coupled to a BX-51 microscope (Olympus) was used for the Raman spectroscopic mapping. An excitation source of 785 nm was used for all the samples. The laser was focused onto the sample using a 50× close-working-distance objective Olympus lens with numerical apertures of 0.75. This set of parameters resulted in an analysis spot size having a diameter of 2 µm. The Raman scattered light was collected through the objective lens, dispersed by a 1200 grooves/mm grating, and detected on a charge-coupled device (CCD), which was thermoelectrically cooled to -65° C. The laser power was 0.1 mW, and the spectra were collected for 240 s with eight co-averages. The spectra were collected over a range from 80 cm-1 to 1525 cm-1 with a spectral resolution between 3 cm-1 to 5 cm-1. Background measurements were collected after every 1000 s of acquisition. The Raman spectra were collected using the OPUS 7.2 program. Raman maps were collected by assigning a grid of analysis points across the sample surface. For all Raman maps, the objective height above the sample was held constant, while the x and y positions were rastered using a movable stage on the Raman microscope. The spectral grids were rectangular and determined by the dimensions of the particular paint layer. The step size between Raman spectral acquisition points ranged from 3 mm to 8 mm, which allowed a better overall visualization of the paint composition.

The data processing of these Raman spectral sets is critical to sample interpretation. Traditionally, analytes are mapped using spectroscopic techniques by univariate methods, where the scattering intensity of a single diagnostic band indicates its presence. However, this procedure leads to challenges when a single unambiguous diagnostic band does not exist, or the analytes have weak signals. Multivariate analysis techniques provide advantages over univariate techniques, such as greater

82 sensitivity (signal-to-noise ratio) and selectivity, and can reduce the effective interference of fluorescence.21-23 MCR-ALS was applied as the primary data-processing method for the Raman spectral data sets. MCR-ALS is a chemometrics technique for extracting pure component spectra from spectra taken of mixtures. The spectra collected from the Raman maps comprised the data cube to which MCR-ALS was applied. Based on a suggested number of pure components, MCR-ALS builds a model of two matrices. The first matrix contains the calculated pure component spectra contributing to the original data cube. The second contains the related concentrations of those pure spectra in each of the original spectra. Previous research has shown the success of coupling this processing tool to Raman mapping to rebuild the concentration matrix into maps of the pure components. MCR-ALS was performed using Matlab R2017a (The Mathworks, Inc.) The PLS Toolbox 8.2.1 (Eigenvector Research Inc.) was used with Matlab to perform MCR-ALS. To determine ratios of polymorphs in each sample, several Raman maps from throughout a paint layer were combined into a single spectral dataset. This action provided additional data for signal averaging and gave a better understanding of the entire paint layer. The spectra were preprocessed to correct baselines, before using

MCR-ALS to resolve component spectra. Figure 5-1 shows a selection of raw data; due to the fluorescence and non-linear background features of the spectra, these needed to be corrected subtracting a background baseline (seen in Figure 5-1b). This figure demonstrates how challenging it still is to differentiate bands, thus why MCR- ALS was needed to distinguish components (Figure 5-1c) and why a large number of factors were necessary to build the model for the datasets and resolve components of

83 interest. The relative concentration matrix for the appropriate MCR-ALS-generated component spectra was used to determine the ratio of atacamite spectral intensity to botallackite spectral intensity. First, the contributing spectra from epoxy, atacamite, and botallackite components were extracted and normalized to unit net spectral intensity, as in all cases these were the components of interest that could be identified. Then, the points with more than 50% epoxy were removed from the calculations to minimize possible distortions. The atacamite and botallackite concentrations were re- normalized to unit net spectral intensity and averaged across all pixels that were predominantly basic copper chlorides to indicate the relative ratio of atacamite to botallackite in the samples.

All the mineral compounds identified by Raman spectroscopy were verified by x-ray diffraction (XRD) performed on the loose scrapings of the samples. A Rigaku

85 D/max Rapid II diffractometer with a copper anode x-ray tube and 0.3mm collimator was used for XRD analysis. Scraped samples were adhered to a glass loop with Parabar 10312 liquid (Hampton Research) and then secured to the sample stage. The samples were analyzed by rotating phi (0-360° rotation) at a speed of 10°/sec with omega held contant at 0°. Collection time ranged from 30 to 90 minutes with an x-ray tube held at 40kV, 30mA. Rigaku RAPID/XRD software (v.2.4.2) was used for instrument operation and data collection, and Rigaku 2DP software (v.2.0.1.1) was used to select the portion of diffraction rings for interpretation. Rigaku PDXL 2 software (v.2.3.1.0) was used to interpret the diffraction pattern, and the Powder Diffraction File from the International Center for Diffraction Data (ICDD) was used as a reference database.

5.5 Results In this study six green paint samples from different locations and time periods of Chinese architecture were analyzed. Multiple Raman spectroscopy maps of varying sizes were collected across each paint layer (in the cross-sections). Using MCR-ALS, models were constructed with pure component spectra, and those components were visually identified as the compounds of interest, atacamite and botallackite. In all but one sample, atacamite and botallackite were identified as distinctive component spectra using MCR-ALS. Atacamite spectra were distinguished in all samples, except W01. Botallackite spectra were distinguished in all samples, except NHS02 and W01. The calculated component spectra by MCR-ALS for the sample sets can be seen in Figure 5-2 for both atacamite and botallackite, as compared to reference spectra from the RRUFF database (R050098 and R070066, respectively).

Figure 5-2 MCR-ALS generated atacamite (a) and botallackite (b) component spectra from each data set compared to RRUFF database reference spectra (R050098 for atacamite and R070066 for botallackite).

As Figure 5-2 shows, the similarities of each component spectrum to the reference spectrum of each polymorph identify the species within each sample. MCR- ALS does not require a reference spectrum as a guide to generate the calculated component spectra. However, more than three factors were required for these models to distinguish the components of interest, because of the large residual baselines and weak overall signals (as seen in Figure 1). The pure component spectra identified as atacamite had characteristic bands at 513, 818, 907, and 974 cm-1. The pure component spectra identified as botallackite had characteristic bands at 400, 448, 498, and 892 cm-1.8-11, 24-25 An additional spectral feature can been seen in both atacamite and botallackite component spectra for samples LXT04, LXT05, YXD02, and NHS02 at about 735 cm-1. This feature appears to be a part of the background matrix, which is correlated with the atacamite and botallackite components in these four samples. The pure component spectrum for W01 identified as azurite had characteristic bands at 250, 400, and 1095 cm-1 (not shown).24-25

87 Upon distinguishing the component spectra of interest, the images could be displayed as false-color maps for atacamite, botallackite, and other components. This method was previously demonstrated by Offroy et al.23 and Smith et al.21-22 to visualize and characterize samples. An example of this application to the paint cross- sections is seen in Figure 5-3 in the analysis of a region of YHD13’s paint layer. The three component spectra of interest were the embedding resin epoxy, atacmite, and botallackite as verified by comparison to reference spectra.8-11, 24-25 Maps were reconstructed for each of these components in the 8x10 grid. A small portion of epoxy can be seen in the upper left of the reconstructed epoxy image as well as in the analysis region in the visible-light photograph. Atacamite appears to be present throughout the region that was imaged, whereas botallackite was concentrated in the lower left of the region. In the overlay of atacamite and botallackite, it can be seen that, in this paint layer, botallackite corresponds to the larger dark green particles and atacamite corresponds to the smaller lighter green particles. The presence of basic copper chlorides in the paint layer were verified with scanning electron microscopy- energy dispersive x-ray spectroscopy, and the presence of these polymorphs were verified in loose scrapings from this painting by x-ray diffraction. The exact desperation of the polymorphs in the paint layers cannot be verified beyond this analysis. These findings demonstrate the use of this technique in identifying and locating different species, like polymorphs, in paint cross-sections.

89 For each sample, the relative ratios of atacamite-to-botallackite have been calculated from the combined dataset of all Raman images for that sample (Table 5-2). In the case of YXD02, there are ratios reported for the combined data of the two samples (S1 & S2), and for the combined data of S3 and S4, to determine the possible variance of different samples from the same source. All of these samples, along with both LXT04 and LXT05 from Lin’xi Ting, reported approximately a 50:50 ratio of atacamite to botallackite. It is interesting that Lin’xi pigments were similar, despite being painted hundreds of years apart. Therefore, this method is capable of identifying differences in the atacamite-to-botallackite ratios across multiple samples. For YHD13, the sample indicates predominantly atacamite is present with some botallackite. In the case of NHS01, only atacamite could be distinguished in the MCR- ALS model, and no botallackite signal could be identified. Finally in W01, neither atacamite nor botallackite could be identified. Instead, the only pigment that could be spectroscopically identified was azurite (Cu3(CO3)2(OH)2), another common copper- based pigment.

90 Table 5-2 The relative ratios of polymorphs in the cross-sections analyzed by comparing the concentrations of components are reported, as well as the number (N) of contributing spectra and the number of factors used to build the MCR model. All of the compound identities were verified with XRD.

Sample ID Location Time Ratio (Atacamite: N= # of Period Botallackite) Factors LXT04 Lin’xi Ting 16th century 47:53 119 13 LXT05 (Forbidden City) 17th-19th 57:43 411 12 century NHS02 Nianhua Temple 18th-19th 100:0 237 14 (Beijing) century YHD13 Ying Hua Dian 16th-17th 70:30 244 15 (Forbidden City) century YXD02 Annex of 15th-16th S3: 47:53 119 15 Yangxin Dian century S4: 50:50 104 14 (Forbidden City) Total: 49:51 216 12 W01 Yanshan 12th century Azurite found 79 10 Temple (Shanxi (only) Province)

Alongside the relative ratios reported in Table 5-2, the number (N) of image locations (pixels) contributing to that ratio calculation and the number of total component, or factor, spectra in the MCR-ALS model are reported. The factors incorporate component spectra of species not of interest, such as the epoxy, interferences, such as fluorescence and noise, and spectra that could not be attributed to specific materials. Some of these component spectra were representative of fluorescence, noise, and other background features. The number of factors was determined once component spectra matching both atacamite and botallackite could be distinguished (if they were present, as verified with XRD).26-27

5.6 Discussion The relative amounts of basic copper chlorides polymorphs highlight the different recipes that could have been used in the preparation of artists’ materials.20 In

91 Table 5-2, the six samples are reported with their relative ratios of atacamite to botallackite. Two samples (LXT04 & LXT05) from the ceiling painting in Lin’xi Ting have very similar compositions of atacamite and botallackite, despite being from two different generations of paint application. It is possible these materials were used intentionally; the pigments could have been purposely prepared to match that of the first generation. The almost 50:50 mixture of polymorphs was also found in the painting of the annex of Yangxin Dian, also located in the Forbidden City from about the sample time period. The analysis of both YXD02 samples (S1 & S2) shows similar results between Raman mapping and MCR-ALS analysis for two cross-sections of the same material. These ratios are close to an equal amount of atacamite and botallackite. On the other hand, compared to the others, YHD13 and NHS02 show more relative atacamite, which is geologically the more stable of the two polymorphs.7 The Nianhua Temple (NHS02) sample only has atacamite present in the pigment. Because this paint sample is from the latest period (18th-19th centuries) of those studied, it is possible that only finding atacamite shows the increase in effective synthesis process of atacamite by this point.3 In contrast, MCR-ALS could not distinguish either atacamite or botallackite in the Yanshan Temple (W01), the oldest paint sample

th studied (12 century). Instead, azurite (Cu2CO3(OH)2) was found both spectroscopically and with XRD. This agrees with previous studies indicating that in the 12th century, green or blue pigments were more commonly “mineral greens” instead of the “copper greens”.5, 20 Thus for a sample from this time period natural azurite or malachite are more likely than the synthesized copper chloride trihydroxides.

92 This study highlights the variations in composition of “copper green” pigments, specifically copper chloride trihydroxides. Historical recipes and written records of how artists prepared their pigments are still being investigated with regard to Chinese architectural painting.20 These variations in synthesis methods are likely the source of variation in relative polymorph ratios in the paints. The preparation method reported in Europe involves covering a copper plate in honey and salt (NaCl) and then suspending it over vinegar to produce the pigment. Studies that have replicated this process have found only atacamite present.3-4, 28 These varying pigment products warrants additional studies into “copper green”, as it is not clearly understood and it possibly plays a larger role in Chinese architecture than “mineral green”. Evidence, while limited, both from this study and others,4-6, 16-17 show a transition from the mined “mineral green” to synthesized “copper green” pigments around the 16th century. Coupling Raman spectroscopic imaging with MCR-ALS was successful in differentiating copper chloride trihydroxides in several historic paint samples. This method was used to find relative ratios that relate to synthesis processes and spatial resolution regarding the polymorphs location within paint layers. XRD is the only other common analysis technique for distinguishing these polymorphs, but with

Raman microscopy, the spot size was significantly smaller. Raman spectroscopy makes it easier to analyze samples within cross-sections and to minimize the amount of sample ultimately taken or damaged from the object. However, Raman spectroscopy does risk damaging the samples as well, with high laser powers focused on small spot sizes. Additionally, long acquisition times are needed to collect spectra with some characteristic bands and to overcome inferences, such as fluorescence. With

93 improvements to Raman spectroscopy, or using alternative spectroscopy techniques, MCR-ALS should be useful for distinguishing paint components within historical samples.

5.7 Conclusions This study finds Raman imaging and MCR-ALS are successful in differentiating and mapping pigment polymorphs in Chinese architectural paint cross- sections. The presence of the two copper chloride trihydroxide polymorphs, atacamite and botallackite, is indicative of a shift from natural “mineral green” pigments to synthesized “copper green” pigments in China at some time during the period from the 12th to 19th centuries. The relative ratios of the polymorphs show this trend in six samples from various locations and time periods. In addition to the relative ratios, the locations of the polymorphs within the paint layers could be visualized by reconstructing the MCR-ALS models. Coupling MCR-ALS with spectroscopic imaging has been very successful for a range of fields, but this study shows its implementation in cultural heritage. By distinguishing materials, new information can be gained on the material history of copper chloride trihydroxides as “copper green” pigments. This conclusion suggests additional studies are needed into their preparation and subsequent usages. However,

Raman imaging is challenging to use on soft materials such as paint. This methodology shows promise for future studies, especially for alternative spectroscopic mapping techniques.

94 5.8 Acknowledgements I would like to thank the staff in Winterthur Museum, Garden & Library’s Scientific Research & Analysis Laboratory for their instrumentation and knowledge and also the Research Fellowship Program.

5.9 Concluding Remarks

Raman spectroscopic imaging and chemometrics analysis were successful in distinguishing basic copper chloride polymorphs within a paint cross-section. This method was also used to report the relative ratios of atacamite and botallackite. Differences between the time periods can be seen, likely related to their preparation methods. Therefore, this study contributes chemical information for understanding the origin and different manufacturing practices of basic copper chloride pigments in Chinese architectural paints.

95 REFERENCES

If you want to number your bibliographic entries, change the style of the items to Bib Entry - numbered.

1. Allen, N. S.; Edge, M.; Sandoval, G.; Ortega, A.; Liauw, C. M.; Stratton, J.; McIntyre, R. B., Interrelationship of spectroscopic properties with the thermal and photochemical behaviour of titanium dioxide pigments in metallocene polyethylene and alkyd based paint films: micron versus nanoparticles. Polymer Degradation and Stability 2002, 76 (2), 305-319.

2. Prati, S.; Bonacini, I.; Sciutto, G.; Genty-Vincent, A.; Cotte, M.; Eveno, M.; Menu, M.; Mazzeo, R., ATR-FTIR microscopy in mapping mode for the study of verdigris and its secondary products. Applied Physics A 2015, 122 (1), 10.

3. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

4. Scott, D. A., A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation 2000, 45 (1), 39-53.

5. Yong, L., Copper trihydroxychlorides as pigments in China. Studies in Conservation 2012, 57 (2), 106-111.

6. Hu, K.; Bai, C.; Ma, L.; Bai, K.; Liu, D.; Fan, B., A study on the painting techniques and materials of the murals in the Five Northern Provinces’ Assembly Hall, Ziyang, China. Heritage Science 2013, 1 (1), 18. https://doi.org/10.1186/2050-7445-1-18

7. Pollard, A. M.; Thomas, R. G.; Williams, P. A., Synthesis and Stabilities of the Basic Copper(II) Chlorides Atacamite, Paratacamite and Botallackite. Mineralogical magazine 1989, 53 (373), 557-563.

96 8. Martens, W.; Frost, R. L.; Williams, P. A., Raman and infrared spectroscopic study of the basic copper chloride minerals -implications for the study of the copper and brass corrosion and "bronze disease". Neues Jahrbuch für Mineralogie - Abhandlungen: Journal of Mineralogy and Geochemistry 2003, 178 (2), 197-215.

9. Frost, R. L., Raman spectroscopy of selected copper minerals of significance in corrosion. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2003, 59 (6), 1195-1204.

10. Frost, R. L.; Martens, W.; Kloprogge, J. T.; Williams, P. A., Raman spectroscopy of the basic copper chloride minerals atacamite and paratacamite: implications for the study of copper, brass and bronze objects of archaeological significance. Journal of Raman Spectroscopy 2002, 33 (10), 801-806.

11. Bongiorno, V.; Campodonico, S.; Caffara, R.; Piccardo, P.; Carnasciali, M. M., Micro-Raman spectroscopy for the characterization of artistic patinas produced on copper-based alloys. Journal of Raman Spectroscopy 2012, 43 (11), 1617-1622.

12. Thompson, D. V., The Materials and Techniques of Medieval Painting. Dover Publications: New York, 1956.

13. Naumova, M. M.; Pisareva, S. A.; Nechiporenko, G. O., Green Copper Pigments of Old Russian Frescoes. Studies in Conservation 1990, 35 (2), 81-88.

14. Nord, A. G.; Tronner, K., The Frequent Occurrence of Atacamite in Medieval Swedish Murals. Studies in Conservation 2018, 63 (8), 477-481.

15. Salvadó, N.; Pradell, T.; Pantos, E.; Papiz, M. Z.; Molera, J.; Seco, M.; Vendrell-Saz, M., Identification of copper-based green pigments in Jaume Huguet's Gothic altarpieces by Fourier transform infrared microspectroscopy and synchrotron radiation X-ray diffraction. Journal of Synchrotron Radiation 2002, 9 (4), 215-222.

16. Xia, Y.; Wang, W. F.; Liu, L. X., Study on mural painting pigments of Fuxi Temple, Tianshui, Gansu Province. Science Conservation Archaeology 2011, 23, 18-24.

17. Fan, Y.; Chen, X.; Li, Z.; Hu, Z., Micro diffraction analysis of the rare green pigment botallackite in ancient wall paintings. Journal of Lanzhou University 2004, 40 (5), 52-55.

97 18. Egel, E.; Simon, S., Investigation of the painting materials in Zhongshan Grottoes(Shaanxi, China). Heritage Science 2013, 1 (1), 29.

19. Schmidt, B. A.; Ziemann, M. A.; Pentzien, S.; Gabsch, T.; Koch, W.; Krüger, J., Technical analysis of a Central Asian wall painting detached from a Buddhist cave temple on the northern Silk Road. Studies in Conservation 2016, 61 (2), 113-122.

20. Li, M. Analysis and Research on Copper Green Pigment. Northwest University, Xi’an, Shaanxi,China, 2013.

21. Smith, J. P.; Smith, F. C.; Krull-Davatzes, A. E.; Simonson, B. M.; Glass, B. P.; Booksh, K. S., Raman microspectroscopic mapping with multivariate curve resolution-alternating least squares (MCR-ALS) of the high-pressure, α-PbO2-structured polymorph of titanium dioxide, TiO2-II. Chemical Data Collections 2017, 9-10, 35-43.

22. Smith, J. P.; Smith, F. C.; Ottaway, J.; Krull-Davatzes, A. E.; Simonson, B. M.; Glass, B. P.; Booksh, K. S., Raman Microspectroscopic Mapping with Multivariate Curve Resolution–Alternating Least Squares (MCR-ALS) Applied to the High-Pressure Polymorph of Titanium Dioxide, TiO2-II. Applied Spectroscopy 2017, 71 (8), 1816-1833.

23. Offroy, M.; Moreau, M.; Sobanska, S.; Milanfar, P.; Duponchel, L., Pushing back the limits of Raman imaging by coupling super-resolution and chemometrics for aerosols characterization. Scientific Reports 2015, 5, 12303.

24. Eremin, K.; Stenger, J.; Li Green, M., Raman spectroscopy of Japanese artists' materials: The Tale of Genji by Tosa Mitsunobu. Journal of Raman Spectroscopy 2006, 37 (10), 1119-1124.

25. Bouchard, M.; Smith, D. C., Catalogue of 45 reference Raman spectra of minerals concerning research in art history or archaeology, especially on corroded metals and coloured glass. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2003, 59 (10), 2247-2266.

26. Krivovichev, S. V.; Hawthorne, F. C.; Williams, P. A., Structural complexity and crystallization: the Ostwald sequence of phases in the Cu2(OH)3Cl system (botallackite-atacamite-clinoatacamite). Struct Chem Structural Chemistry : Computational and Experimental Studies of Chemical and Biological Systems 2017, 28 (1), 153-159.

98 27. Zheng, X. G.; Mori, T.; Nishiyama, K.; Higemoto, W.; Yamada, H.; Nishikubo, K.; Xu, C. N., Antiferromagnetic transitions in polymorphous minerals of the natural cuprates atacamite and botallackite Cu2Cl(OH)3. Physical Review B 2005, 71 (17), 174404.

28. Wiggins, M. B.; Alcántara-García, J.; Booksh, K. S., Characterization of copper-based pigment preparation and alteration products. MRS Advances 2017, 2 (63), 3973-3981.

99 Chapter 6

MULTI-ANALYTICAL CHARACTERIZATION OF COPPER-BASED PIGMENTS’ PREPARATION AND ALTERATION PRODUCTS

6.1 Introductory Remarks

In this chapter, the focus is on the preparation of verdigris and verdigris-like pigments by reproducing historical recipes. Neutral and basic verdigris were further studied by artificially aging them with and without cellulose and gum Arabic. Analyzing these materials with FTIR spectroscopy, XPS, and XRD indicated a pathway of degradation through an intermediate state to copper oxides. Gum Arabic appears to stabilize the copper species and slow the overall degradation. These fundamental studies create ways to understand better the damaging effects of verdigris pigments on paper-based media and to guide future treatments of cultural heritage artifacts. The first part of this work regarding the pigment preparation and characterization was published in the journal MRS Advances in 2018. The second part focusing on the alteration and degradation studies is under review for the journal Applied Spectroscopy. Figures and text in this chapter were reprinted with permission. The authors were Marcie B. Wiggins, Emma Heath, Karl S. Booksh, and Jocelyn Alcántara-García.

6.2 Introduction There is a growing fascination with the research of historic and artistic materials, as well as their decay. Areas of interest include the artists’ processes and

100 techniques (technical art history),1-2 the development and evaluation of art conservation methodologies,3-5 and research into constituent materials’ decay.4, 6-9 The degradation of pigments is particularly interesting, given that new findings can change the course of action of conservation professionals.10 Verdigris is one of the pigments that falls into this category. Commonly present in works of art dating back as far as antiquity, this copper-based pigment has been identified in various states of degradation in leather, parchment,11-12 fabric,13 waxy paint,14 (oil and canvas) paintings,7, 9, 15-17 and paper.8, 18-21

The term “verdigris” is nowadays applied to a blue-green pigment composed of (a mixture of) copper (II) acetates. However, throughout history, the term could also refer to copper carbonates (malachite, azurite), chlorides (atacamite, brochantite),2 and sulfates (posnjakite, langite).22 Being a synthetic pigment, verdigris’ chemical composition is the direct result of its synthesis, and its various degradation mechanisms are strictly associated with the media in which it was applied. Its application can be as a pure pigment or as a mixture. To all these considerations, conservation professionals must add the conditions in which the object was, is, and will be treated, stored, and/or displayed to evaluate it. For the purpose of this chapter, we refer to verdigris as the blue-green pigment obtained by reacting copper and acetic acid. Studies have shown that obtaining a single phase of this pigment is challenging.23-24 Therefore, one can expect to find more than one form of verdigris in both works of art and in experimental preparations. Chemically, verdigris can predominantly be classified as neutral and basic forms, with the most common structures being (Cu(CH3COO)2•H2O) and

25 (Cu(CH3COO)2•(OH)2•5H2O), respectively. The pigment has been identified in

101 numerous oil paintings, and its degradation has been the subject of detailed studies. For instance, it has been reported that ligand-exchange reactions take place between fatty acids and verdigris.26 Degradation products include copper oxalates, carboxylates, and even formates, which were previously thought to be intermediate products from fatty acids carboxylates to oxalates.7 Similarly, research into the discoloration observed when verdigris is mixed with linseed oil suggested a light- induced mechanism that favors decarboxylation and reduction from Cu(I) to Cu(II).9 Every artifact containing verdigris can be affected by the pigment’s established reactivity. Those damaging effects are particularly concerning for more fragile materials, such as paper. The pigment has been extensively used on illuminated manuscripts and maps.11, 18, 20-22, 27-30 Although many seem in good condition, it has been demonstrated that soluble copper ions, present in inks, watercolors, and pigments like verdigris, are particularly active in the so-called “corrosion” of paper.31 That is, copper ions induce hydrolysis, oxidation, and radical driven reactions that contribute to the subsequent degradation of cellulose.5, 32-34 Much of the research on this pigment and the effects copper ions have on paper has focused on treatment, and research on copper ions has been commonly associated with iron gall ink research.33-36 Detailed studies on the development and use of antioxidants, complexating agents, and deacidification processes have shown various degrees of success in slowing paper decay induced by copper ions.3-5, 8 However, similar studies to those made on verdigris with oil media remain scant. Herein, we report on the formation of this complex family of pigments (including verdigris-like pigments, such as salt and soap greens). Historic recipes of verdigris and verdigris-like pigments were replicated and characterized, both as pure pigments and on an organic paper substrate. Expanding on

102 these findings, detailed accelerated degradation studies of verdigris using gum Arabic on cellulose, a common binding agent and medium in writing and illumination, are reported.

6.3 Experimental

6.3.1 Preparation of Pigments

Historical recipes for preparing verdigris and other verdigris variants were reported in Pliny, Scott, and Thompson 37-39. Figure 6-1 shows the progress of replicating the simple verdigris recipe. Copper plates (15×10×0.3 cm) were suspended on top of inverted plastic cups placed at the corners (Samples A), see Figure 6-1a. X- ray fluorescence (XRF) was used to verify the plates were pure copper at the start of the process. Distilled white vinegar (Heinz®, 5% acetic acid) was placed at the base of a (19×14×5 cm) Rubbermaid Tupperware® container beneath the copper plate, which was then sealed and stored for 6 months under ambient conditions. Samples were unsealed to photodocument crystal formation approximately once a month.

103 Two variants of verdigris pigments were also prepared based on these reports 38, 40. Soap green, also known as Rouen green, has not been completely characterized, but it is assumed to be predominantly copper carboxylates. This pigment was synthesized by applying lye-based soap in a coating across the copper plate’s surface (Sample B). The coating thickness was assumed to be homogenous and less than 1 mm. Lye-based soap was synthesized, following Phanstiel et al., in the laboratory to minimize any contamination from commercial additives.40 Salt green was prepared by coating (<1 mm) a copper plate’s surface with organic honey and sodium chloride

(Sample C). The crystals that formed on the surfaces of the copper plates were scraped away after 6 months and collected. Both pigments retained some amount of moisture compared to the simple verdigris.

6.3.2 Preparation of Samples We studied the degradation of verdigris pigments, both dependent and independent of an organic substrate. First, both the independent powder pigments and the mixtures of pigment powders with cellulose powder were aged. For these studies, only the laboratory-synthesized pigments, neutral (A1) and basic (A2) verdigris, were used. The degradation related to the presence of cellulose was studied by mixing each pigment and microcrystalline cellulose (ACROS Organics, particle size 50 mm) in a

1:1 (w/w) ratio. Mixtures of pigment and powdered cellulose complement our study of paper colored with the pigments, as an organic substrate (cellulose) with pigment could be evaluated without the influence of any binding medium. This sample set was prepared by placing approximately 0.3 g of each pigment or pigment-cellulose mixture in 2 mL gas chromatography vials with caps (Sun-sri™). Septa from these caps were replaced with porous, nonwoven polyester fabric (Reemay®) to allow equilibrium

104 with the environmental conditions, while minimizing mass loss due to air circulation within the chamber. To study the effects of an organic substrate on the pigments, verdigris was applied to cellulose paper (Whatman filter paper No. 1), both as a dye and as a watercolor. Commercially available pigments (Kremer) and the neutral (A1) and basic (A2) verdigris pigments as previously described were dissolved in water at 5% (w/w). Various concentrations of gum Arabic were added to each solution. XRD and FTIR spectroscopy showed the Kremer verdigris was composed of neutral verdigris crystals.

The basic verdigris pigment was not readily soluble in water, so a small aliquot of 0.75 mL of 5% acetic acid was added to 15 mL of solution to aid in dissolution by slightly neutralizing the pigment, while still maintaining its basic character. Gum Arabic mixtures of each pigment (commercially acquired, neutral, and basic) were prepared at 1%, 5%, and 10% (w/w) against the total weight of the verdigris solution. Solutions of the verdigris pigments with no gum Arabic were drop-cast onto five strips of Whatman filter paper No. 1 (13 cm x 3 cm) in 20 µL aliquots. The process was repeated five times and allowed to dry between applications to concentrate the pigment. This process was repeated with the 1%, 5%, and 10% (w/w) gum Arabic concentrations for a total of twenty paper strips (4 x 5 replicates). Within the environmental chamber, these strips were hung from the grates by commercially available, metal binder clips attached to the edges of the verdigris-colored papers. Mylar strips were used as support and buffers along the edges of the papers to avoid contamination from the clips. One sample strip of each type was withheld from aging to act as controls, and the subsequent replicates were removed after 1, 3, 5, and 7 days (see Figure 6-2).

105

6.3.3 Aging Studies

Pigments were studied both as pure powder pigments and as mixed with organic matrices (mixed in cellulose powder and applied to cellulose paper) using an ESPEC BTL 433 test chamber. The accelerated degradation cycle ramped the chamber parameters to 60±2° C and 85±2% RH (15 min), then maintained these conditions for up to 21 days. Samples of the paper were taken at 1, 3, 5, and 7 days (Figure 6-2). The

106 pigment-cellulose mixtures were sampled at 3 and 7 days. Independent pigment powder samples were sampled at 3, 7, and 21 days.

6.3.4 Instrumentation FTIR spectroscopy was used to monitor changes made by the aging process to the pigment and cellulose through functional groups. The spectra were collected using a Bruker Optic Vertex 70 FTIR spectrometer and Hyperion 2000 Microscope with a single-point attenuated total reflectance (ATR) attachment. Spectra were collected using OPUS 6.0 (v.6.0.72) software, averaging 128 scans in the 4000–600 cm−1 region with a resolution of 4cm-1. X-ray diffraction (XRD) was used to identify the crystalline materials and their changes during the aging process. Diffraction was performed using a Rigaku D/max Rapid II diffractometer with a copper anode X-ray tube and 0.3 mm collimator. Samples adhered to a glass loop by Parabar 10312 (Hampton Research) were secured to the sample stage in spin mode (0-360° rotation), 10°/sec, 40 kV, 30 mA, and a total collection time ranging from 15 to 30 min. Pigment paper fibers used 45 kV, 40mA, and a total collection time ranging from 60 to 90 minutes. Rigaku RAPID/XRD software (v.2.4.2) was used to operate the instrument and collect the data, while Rigaku 2DP software (v.2.0.1.1) was used to select and process the diffraction pattern.

Rigaku PDXL 2 software (v.2.3.1.0) was used to interpret and compare to a reference database, the powder diffraction file from the International Center for Diffraction Data (ICDD). X-ray photoelectron spectroscopy (XPS) was used to determine the composition and oxidation states of elements, specifically studying copper changes. The spectroscopy was performed on cuttings (1x1 mm) from the artificially aged filter

107 papers adhered to carbon tape (Nisshin EM Co., Ltd.) for analysis with a Thermo Scientific K-Alpha+ XPS instrument. This instrument was equipped with a monochromatic Al Kα source (hυ = 1486.6 eV), and it operated at a base pressure of 8x10-9 mbar. For each sample, measurements were taken in three different spots with a 400 µm X-ray spot size and the counts were summed. An electron flood gun was used to reduce charging effects. Cu 2p high-resolution spectra were collected with 25 scans across the 965-925 eV range. The pass energy was 20 eV at 0.1 eV/step and a 50 sec dwell time. Data was collected with Thermo Avantage (v5.962) and analyzed with

CasaXPS (version 2.3.18PR1.01), where the spectra were calibrated using C 1s peak at 284.6 eV.

6.4 Results & Discussion

6.4.1 Characterization of Prepared Pigments

A summary of spectroscopic characterization of historically prepared pigments is found in Table 6-1. The XPS results list copper states observed, but the predominant state(s) found in each pigment are noted by an asterisk (*).

108

Table 6-1 Summary of FTIR, XRD, and XPS key features of pigments prepared according to historical recipes and their corresponding overall characterization.

Pigments FTIR (cm-1) XRD (2Θ) XPS Characterization 1421 (1) 12.8 (1) Cu (I) & (1) Neutral Verdigris

1444 (1) 14.3 (1) Cu (II)* [Cu(CH3COO)2 H2O] 1595 (1) 15.3 (1) A1 3269 (1) 16.5 (1) 3367 (1) 25.3 (1) 3460 (1) 28.6 (1) 1431 (2) 10.8 (2) Cu (I) & (2) Basic Verdigris

1537 (2) 12.5 (2) Cu (II)* [Cu(CH3COO)2

A2 1568 (2) 18.1 (2) Cu(OH)25H2O] 3325 (2) 19.3 (2) 25.3 (2) 1421 (1) 12.8 (1) Cu (I)* (1) Neutral Verdigris

1442 (1) 14.3 (1) [Cu(CH3COO)2 H2O] 1597 (1) 15.1 (1) (3) Copper Carboxylates 1739 (3) 16.4 (1) B 2922 (3) 25.2 (1) 2852 (3) 3269 (1) 3363 (1) 3456 (1) 845 (4) 16.1 (4) Cu (I)* & (4) Atacamite

893 (4) 17.6 (4) Cu (II)* [CuCl(OH3)]

949 (4) 31.5 (4) (5) Cuprite [Cu2O] 987 (4) 32.2 (4) C 3330 (4) 36.4 (5) 3438 (4) 39.6 (4) 42.3 (5) 61.3 (5) 73.5 (5)

109 The historical recipe used to make the plain verdigris produced predominantly blue-green neutral verdigris crystals (Sample A1), see Figure 6-1b. FTIR spectra of this sample showed C-O bands and broad O-H bands characteristic of neutral verdigris

7, 41-42 (Cu(CH3COO)2 H2O). Of the three verdigris plates prepared, one also had light blue crystals start to form in the center of the plate after 2 months. These crystals (Sample A2) continued to grow over the four remaining months (see Figure 6-1c) and were later identified as a basic form of verdigris (Cu(CH3COO)2 Cu(OH)2 5H2O) using FTIR and XRD methods of identification.7, 37, 41-42

Soap green was a mixture of copper soaps (copper stearates, palmitates, etc.) as well as neutral verdigris. The copper carboxylates were identified with FTIR spectroscopy in addition to the characteristic spectra for neutral verdigris, but they could not be verified with XRD due to their amorphous nature.43 Based on the interpretation of the XPS spectra, this pigment was predominantly a copper (I) species, which is different from the other pigments prepared. Salt green pigments were identified as mixtures of two different compounds, which agrees with Scott,44 who states that salt green is a mixture of copper chlorides.

FTIR spectroscopy and XRD identified the presence of atacamite (CuCl(OH3)), indeed a chloride commonly found in Asian and Western paintings and frescos, where synthetic atacamite is associated with medieval and Renaissance painting 44-46. Atacamite is thought to be the stable form of copper chloride trihydroxide.47 In addition, XRD identified a red layer of cuprite (Cu2O) between the copper plate and the crust of atacamite (Sample C).48 It is likely that honey, a substance that naturally contains monosaccharides, promoted the oxidation of copper at this layer via a redox process 50.

110 Characterization by XPS of the verdigris pigments shows they are mixtures of copper (I) and, mainly, copper (II) species. Salt green shows an equal ratio of copper (I) and copper (II).50 The larger amount of copper (I) species in soap green suggests reaction with fatty acids from soap reduces the copper pigments, as compared to prepared verdigris pigments.

6.4.2 Degradation of Pure Powder Pigments and Pigment-Cellulose Mixtures XRD and FTIR analyses of these verdigris pigments showed two products: neutral (Cu(CH3COO)2H2O) and basic (Cu(CH3COO)2Cu(OH)25H2O) verdigris (as seen in Table 6-1).25 As expected, neutral verdigris is a green-turquoise product, while basic verdigris is light blue in color. These two verdigris species represent two possible starting materials used by artists for their different coloring. 12, 27, 51-53 Due to our interest in the changes of the pigment in cellulose, accelerated degradation tests were performed on powdered pigments with no organic substrate/media; subsequently, degradation tests were performed on a mixture of the pigment and powdered cellulose; and, third, on pigments on paper samples applied as watercolors (with gum Arabic) and as dyes (dissolved in water). Studying the pigments’ degradation on paper was critical to the understanding of a near real-case degradation. Studying mixtures of powdered verdigris and cellulose, on the other hand, provided experiments in the absence of any binder, with a comparatively higher ratio of pigment to organics and with comparatively greater surface areas than on a paper sheet. The pure powdered pigments and the mixture of powdered pigments with powdered cellulose were aged for 7 days at 50 ºC and 60 % RH in open vials (for more details see the Experimental Section). Figure 6-3 shows XRD patterns of

111 pigments’ aging: control (with no organics present) of basic (left) and neutral (right) verdigris. After 7 days of aging, new peaks at 9.41° and 18.73° suggested the formation of the less common basic verdigris species Cu(CH3COO)23Cu(OH)22H2O as an intermediate before showing any visual changes pointing to degradation. 37, 41-42

At 7 days of aging, the pure basic verdigris, without organics, is almost entirely converted to the above-mentioned intermediate structure as shown by XRD. The conversion reaction for neutral verdigris seems to be slower, as

Cu(CH3COO)2H2O is still present along with the intermediate after 7 days. The samples of pigments mixed with cellulose showed similar trends. XRD of both basic and neutral verdigris mixtures showed a near-complete conversion to the intermediate

112 species Cu(CH3COO)23Cu(OH)22H2O, although that was slightly more pronounced for basic verdigris. The FTIR spectra of both pigments, displayed in Figure 6-4, further supports the alteration suggested by XRD, which, to the naked eye, is associated with a shift to paler hues of both basic and neutral pigments. Unreacted basic verdigris showed carbonyl bands at 1564, 1537, and 1420 cm-1, which shifted to 1527 and 1415 cm-1 when aged with cellulose powder. Unreacted neutral verdigris, on the other hand, exhibited the carbonyl bands at 1595, 1442, and 1421 cm-1, and shifted to 1547 and

1412 cm-1 when aged with cellulose powder, although these spectra are harder to interpret due to the cellulose signals. 7, 37, 41-42

Figure 6-4 FTIR spectra of basic, Cu(CH3COO)2Cu(OH)25H2O (left), and neutral, Cu(CH3COO)2H2O (right) verdigris pigments. Red lines indicate a shift of the carbonyl bands, associated with aging.

Altogether, FTIR and XRD evidence suggests both basic,

Cu(CH3COO)2Cu(OH)25H2O, and neutral, Cu(CH3COO)2H2O, verdigris degrade

113 into the same intermediate form: Cu(CH3COO)23Cu(OH)22H2O. The intermediate was present in the samples with and without cellulose. However, cellulose’s presence seems to promote and increase the production of the alteration product. The deleterious effects of copper-containing species with paper have been the focus of attention for paper conservators and heritage conservation scientists alike; however, in contrast to our research, the research has been heavily focused on treatment development and evaluation.4, 8, 52, 54-55 Although there are several established chemical structures for copper (II) acetates or verdigris, the present study documented that two starting species (basic and neutral) will transform into a different verdigris species, the intermediate Cu(CH3COO)23Cu(OH)22H2O during their degradation, as pigment, and in the presence of cellulose.

6.4.3 Degradation on Paper The preliminary work on verdigris on cellulose suggests both basic and neutral verdigris pigments degrade into an intermediate state of verdigris,

(Cu(CH3COO)23Cu(OH)22H2O), that continues to degrade into copper oxide. A set of aging experiments (50ºC, 60%, 7 days) was performed on unisized paper (Whatman filter paper No. 1), where commercially available (Kremer) and lab-prepared verdigris (basic and neutral) was applied as dye (mixed in water). XRD and FTIR results of freshly applied pigments on paper showed the presence of neutral

(Cu(CH3COO)2H2O) and basic (Cu(CH3COO)2Cu(OH)25H2O) verdigris species, respectively.7, 37, 41 In addition, samples of basic verdigris already contained the intermediate species (Cu(CH3COO)23Cu(OH)22H2O) prior to aging (Figure 6- 5). Figure 6-5 shows the intermediate species (peaks at 9.41° and 18.73°), at 0 days of aging, which seems to be connected to the rapid transformation of the basic pigment

114 into copper (II) oxide (peaks at 35.5° and 38.6°).48, 56 Since copper oxide is a common degradation product in cultural heritage objects where copper salts have reacted with organic materials (e.g. parchment, binders, etc.), its presence was expected.35, 57-58

Figure 6-5 XRD patterns of basic (Cu(CH3COO)2Cu(OH)25H2O) verdigris pigment applied to Whatman filter paper No. 1 as dye (aqueous solution). Measurements were taken after 0 (black), 3 (blue), and 7 (green) days. Peaks at 35.5° and 38.6° indicate formation of copper (II) oxide (red lines).

This set of experiments puts the decay of verdigris on paper into context for the conservation audience. It has been well established that both basic and neutral verdigris pigments degrade into copper oxides. However, the evidence presented here allowed the identification of the intermediate (Cu(CH3COO)23Cu(OH)22H2O) species. Its presence, more importantly, seems to be the indicator of a subsequent rapid degradation process, as decay was more pronounced when applied to paper as a dye (aqueous solution) and not as the cellulose-pigment (described in the previous section).

115 6.4.4 Degradation on Paper with Gum Arabic Gum Arabic was a common binder in works of art on paper, including illuminated manuscripts, where verdigris was used.52, 59-61 To evaluate the role of this substance in the degradation pathway, aging experiments (50ºC, 60%, 7 days) were performed on unisized paper (Whatman filter paper No. 1). Samples of commercially available (Kremer, neutral) and laboratory-synthesized verdigris (basic and neutral) were applied both as dye (mixed in water) and as watercolor (mixed in an aqueous solution of gum Arabic) in a range of gum Arabic concentrations (0, 1, 5, and 10%). Surface analyses of the samples were performed using XPS, which tracked the changes in oxidation states of copper-containing species. Because this technique provides information about chemical environments, but not exact species, the XPS discussion is limited to oxidation states only. Given its high-spatial resolution, attempting to sample the same site to run complementary analyses like XRD was impossible. However, other areas of the same mock-up sample were analyzed using XRD and FTIR spectroscopy. Monitoring of the cellulosic changes with the FTIR technique did not show any conclusive trends. XPS analysis of commercially available (neutral), and laboratory-synthesized (basic and neutral), both as unaged pigments and applied to paper as aqueous solutions, showed mixtures of Cu (I) and Cu (II) states. The presence of gum Arabic seems to be associated with the reduction of copper to Cu (I). Even at zero days of aging, basic verdigris was predominantly in a Cu (II) state without gum Arabic. Neutral verdigris, on the other hand, was approximately an even mixture of Cu (I) and Cu (II). The XPS high-resolution scan of the copper 2p region (Figure 6-6) suggested a correlation between higher concentrations of gum Arabic and more Cu (I). Higher concentrations of gum Arabic

116 yielded an increase in peaks at 953.3 eV (2p1/2) and 933.1 eV (2p3/2), indicating a shift to Cu (II) oxidation state. Therefore, there were more Cu (II) species in the sample. 56

Higher concentrations of gum Arabic imply higher concentrations of polysaccharides and reducing sugars. The results showed gum Arabic, mixed with all forms of verdigris, reduced some verdigris species from Cu (II) to Cu (I), with higher concentrations of gum Arabic leading to more Cu (I). Cu (I) species produce more hydroxyl radicals in cellulose than Cu (II), leading to faster cellulose decay.33, 62 Reducing sugars with copper (II) acetate, also known as Barfoed’s reaction,63 suggests objects, like illuminated manuscripts, contain all the starting reagentsthat will lead to a Cu (I) species as byproducts. This observation is critical for conservation treatments. Given that many historic ink recipes contain Cu (II) ions as copper sulfates or copper

117 chlorides,62 the development of treatments for copper-rich writing/illuminating materials for paper has targeted copper ions in this form. As a result, verdigris- containing objects are frequently treated in a similar way.4 Interestingly, pigments seem to react with gum Arabic to oxidize copper during the accelerated degradation process, as Cu (I) oxidizes into Cu (II) species. Figure 6-7 shows laboratory-synthesized neutral verdigris in 10 % gum Arabic, which showed the highest Cu (I) species at the start. Over the course of aging (3, 7 days), the peaks at 955.1 eV (2p1/2), 944.7-941.1 eV (satellite), and 934.9 eV (2p3/2) slightly increased, possibly indicating some conversion to a Cu (II) compound, such as copper (II) oxide.56 Commercially available neutral verdigris showed an oxidative change as well. Samples that started with higher amounts of Cu (II) (lower amounts of gum Arabic) remained in that state throughout the experiment, and it is possible they reach stoichiometric equilibrium or perhaps undergoing a different pathway to copper oxide. The spectra show XPS evidence of the oxidation of the copper species during the artificial aging process.

118

The presence of Cu (I) species, significant when the pigment is mixed with gum Arabic, seems to be associated with a slower pigment degradation. XPS results show Cu (I) and Cu (II) are present during the entire aging time in the presence of gum. In contrast, Cu (II)-containing species are overwhelmingly present in its absence. In addition, XRD data on the same mock-ups suggest a slower reaction in the presence of gum Arabic, by showing the same starting material

(Cu(CH3COO)23Cu(OH)22H2O) during the 7 days of aging. Figure 6-7 shows the XRD patterns (after 7 days of aging) for basic verdigris with 10% gum Arabic and a control sample, with no gum Arabic. While the sample with no gum Arabic only shows peaks for CuO (peaks at 35.5° and 38.6°), the sample with 10% gum Arabic shows the presence of the initial intermediate verdigris (peaks at 9.45° and 18.96°) as well as the final product, CuO.37, 41-42 Furthermore, visual assessment of these samples show that those with more gum Arabic are less discolored and brown, compared to the

119 controls, with no gum Arabic (see Figure 6-2). This comparison suggests that without gum Arabic, under these conditions, the reaction goes to completion quickly, while in the presence of gum Arabic, the reaction progresses slowly. The results discussed here seem to suggest the degradation from the intermediate verdigris (Cu(CH3COO)23Cu(OH)22H2O) to copper (II) oxide (CuO) occurs by more than one mechanism, and that degradation is slower in the presence of gum Arabic during paper aging, possibly because the gum converts verdigris to a Cu (I) species first. Feasible reasons include: (1) a stable equilibrium between the various saccharides present in gum Arabic and verdigris variants; (2) pigment simultaneously reacting with cellulose and gum Arabic; or (3) a different, specific, mechanism from that occurring in the absence of gum, yet to be studied. Our findings have significant implications for conservation, as they could help explain dissimilar degradation in artifacts and could potentially lead to novel treatments.

120 As a summary of the findings, a schematic of the proposed pathway is shown in Figure 6-9. Both neutral and basic verdigris transform to an intermediate state of copper acetate, whether there is cellulose or not. That intermediate continues to form copper (II) oxide on cellulose. During this degradation process, the copper ions are predominantly in a Cu (II) state. When gum Arabic is added as a binder, the copper ions reduce to Cu (I) and then oxidize to Cu (II) during the degradation process to copper oxide.

The presence of numerous transition metal ions in historic inks is a well- documented cause of cellulose degradation.33-34, 62, 64-66 In particular, the presence of Fe (II) has been the subject of numerous studies of iron gall inks and their treatment.5, 31, 34, 36, 65, 67-71 Of all ions under study, copper has been deemed the most efficient to promote auto-oxidation and depolymerization of cellulose either as Cu (I) in the

121 pseudo-Fenton’s reaction to form hydroxyl radicals,34, 64, 72-73 or as Cu (II) to act as an oxidation agent with cellulose carboxylates.32, 59, 75-78 Our findings show once more that copper is indeed a clear cause of decay, but in the presence of gum Arabic, Cu (I)- promoted degradation is slower than that of Cu (II) without gum.

6.5 Conclusion The spectroscopic study of verdigris pigments prepared by historical methods provides a greater understanding of verdigris mixtures as the starting materials used in works of art and archival materials with visual signs of decay. Verdigris pigments predominantly fall within neutral, Cu(CH3COO)2H2O, and basic verdigris,

(xCu(CH3COO)2yCu(OH)2nH2O), which has many variations of basic verdigris crystal structures of the pigment. In this study, we only identified basic verdigris

(Cu(CH3COO)2Cu(OH)25H2O) initially formed alongside neutral verdigris. On the other hand, atacamite was identified in salt green, while neutral verdigris and copper carboxylates were identified in soap green. The study also allows one to determine the alteration of these verdigris species. Both laboratory-synthesized neutral and basic verdigris were found to transform into the same intermediate verdigris species, Cu(CH3COO)23Cu(OH)22H2O, with and without cellulose being present. This alteration was put into the context of the pathway of verdigris to copper oxide on paper, as identified by XRD. The role of gum Arabic, a common binder for paper documentation, was evaluated as well. Gum Arabic formed reduced copper species in the verdigris watercolors on paper. Those copper species appear to slow the ultimate reaction to the degradation product, copper (II) oxide. These findings were documented by XPS and XRD with supporting FTIR analysis.

122 This research documents the alteration in various forms of verdigris during the aging process on cellulose, and the oxidation states with gum Arabic binders on cellulose. This information benefits conservators and conservation scientists trying to preserve and treat historical documents, where verdigris plays a role in the degradation.

6.6 Acknowledgements I thank Winterthur Museum, Garden & Library’s Scientific Research & Analysis Laboratory (SRAL) for the use of the XRD and accelerated aging systems. The Andrew W. Mellon Foundation provided funding for the accelerated aging unit. I also would like to thank Brain Baade at the University of Delaware for materials and assistance in preparing many of the pigments. I thank the National Science Foundation (CHM 1506853) and the University of Delaware’s Office of Graduate and Professional Education for support. I also thank the University of Delaware’s Surface Analysis Facility for XPS support, as well as the NSF (CHE 1428149) and the NIH NIGMS COBRE program (P30-GM110758) for their support of the acquisition of instrumentation.

6.7 Concluding Remarks

This chapter contributes to the overall understanding of verdigris’s alteration and degradation pathways on paper-based media. The role of gum Arabic in these mechanisms had not been well-studied previously. This study utilized a range of spectroscopic techniques for characterized pigments prepared by historic recipes to identify the starting materials of these reactions. Additional evidence that was presented in this study indicated an intermediate verdigris species and suggested gum

123 Arabic could stabilize or slow degradation of verdigris. Therefore, this chapter focuses on a common copper salt pigment’s fundamental degradation to inform conservation professionals.

124 REFERENCES

1. De la Roja, J. M.; San Andrés, M.; Cubino, N. S.; Santos-Gómez, S., Variations in the colorimetric characteristics of verdigris pictorial films depending on the process used to produce the pigment and the type of binding agent used in applying it. Color Research & Application 2007, 32 (5), 414- 423.

2. Valadas, S.; Freire, R. V.; Cardoso, A.; Mirao, J.; Dias, C. B.; Vandenabeele, P.; Candeias, A., On the Use of the Unusual Green Pigment Brochantite (Cu(4)(SO(4))(OH)(6)) in the 16th-Century Portuguese-Flemish Paintings Attributed to The Master Frei Carlos Workshop. Microsc Microanal 2015, 21 (2), 518-25.

3. Malešič, J.; Kolar, J.; Anders, M., Evaluation of Treatments for Stabilization of Verdigris and Malachite Containing Paper Documents. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 283.

4. Ahn, K.; Hartl, A.; Hofmann, C.; Henniges, U.; Potthast, A., Investigation of the stabilization of verdigris-containing rag paper by wet chemical treatments. Heritage Science 2014, 2 (1), 12. https://doi.org/10.1186/2050-7445-2-12

5. Jana Kolar, A. M., Aneta Balažic, Matija Strlic, Gabriele Ceres, Valeria; Conte, V. M., Ted Steemers, Gerrit de Bruin, New Antioxidants for Treatment of Transition Metal Containing Inks and Pigments Restaurator. International Journal for the Preservation of Library and Archival Material 2008, 29, 184- 198.

6. De Laet, N.; Lycke, S.; Pevenage, J. V.; Moens, L.; Vandenabeele, P., Investigation of pigment degradation due to acetic acid vapours: Raman spectroscopic analysis. European Journal of Mineralogy 2013, 25 (5), 855- 862.

7. Prati, S.; Bonacini, I.; Sciutto, G.; Genty-Vincent, A.; Cotte, M.; Eveno, M.; Menu, M.; Mazzeo, R., ATR-FTIR microscopy in mapping mode for the study of verdigris and its secondary products. Applied Physics A 2015, 122 (1), 10.

125 8. Hofmann, C.; Hartl, A.; Ahn, K.; Faerber, I.; Henniges, U.; Potthast, A., Studies on the Conservation of Verdigris on Paper. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 147.

9. Santoro, C.; Zarkout, K.; Le Hô, A.-S.; Mirambet, F.; Gourier, D.; Binet, L.; Pagès-Camagna, S.; Reguer, S.; Mirabaud, S.; Du, Y.; Griesmar, P.; Lubin- Germain, N.; Menu, M., New highlights on degradation process of verdigris from easel paintings. Applied Physics A 2014, 114 (3), 637-645.

10. Vermeulen, M.; Sanyova, J.; Janssens, K.; Nuyts, G.; De Meyer, S.; De Wael, K., The darkening of copper- or lead-based pigments explained by a structural modification of natural orpiment: a spectroscopic and electrochemical study. Journal of Analytical Atomic Spectrometry 2017, 32 (7), 1331-1341.

11. Zoleo, A.; Nodari, L.; Rampazzo, M.; Piccinelli, F.; Russo, U.; Federici, C.; Brustolon, M., Characterization of Pigment and Binder in Badly Conserved Illuminations of a 15th‐Century Manuscript. Archaeometry 2014, 56 (3), 496- 512.

12. Chaplin, T. D.; Clark, R. J. H.; Martinón-Torres, M., A combined Raman microscopy, XRF and SEM–EDX study of three valuable objects – A large painted leather screen and two illuminated title pages in 17th century books of ordinances of the Worshipful Company of Barbers, London. Journal of Molecular Structure 2010, 976 (1), 350-359.

13. Leese, V. D. a. M., The Degradation of Silk by Verdigris. Restaurator. International Journal for the Preservation of Library and Archival Material 1995, 16 (45-63), 45.

14. Liang, J.; Scott, D. A., Green-copper containing waxy paint on two Egyptian polychrome artifacts: A technical study Studies in Conservation 2014, 59 (6), 391-403.

15. Ren, F.; Atlasevich, N.; Baade, B.; Loike, J.; Arslanoglu, J., Influence of pigments and protein aging on protein identification in historically representative casein-based paints using enzyme-linked immunosorbent assay. Anal Bioanal Chem 2016, 408 (1), 203-15.

16. Villar, S. E. J.; Edwards, H. G. M., An extensive colour palette in Roman villas in Burgos, Northern Spain: a Raman spectroscopic analysis. Analytical and Bioanalytical Chemistry 2005, 382 (2), 283-289.

126 17. Samanian, K., Identification of Green Pigment Used in Persian Wall Paintings (ad 1501–1736) Using PLM, FT–IR, SEM/EDX and GC–MS Techniques. Archaeometry 2015, 57 (4), 740-758.

18. Arias, T. E.; Montes, A. L.; Bueno, A. G.; Benito, A. D.; García, R. B., A Study about Colourants in the Arabic Manuscript Collection of the Sacromonte Abbey Granada Spain. A New Methodology for Chemical Analysis. Restaurator 2008, 28, 76-106.

19. Mendes, N.; Lofrumento, C.; Migliori, A.; Castellucci, E. M., Micro‐Raman and particle‐induced X‐ray emission spectroscopy for the study of pigments and degradation products present in 17th century coloured maps. Journal of Raman Spectroscopy 2008, 39 (2), 289-294.

20. Chaplin, T. D.; Clark, R. J. H.; Jacobs, D.; Jensen, K.; Smith, G. D., The Gutenberg Bibles: Analysis of the Illuminations and Inks Using Raman Spectroscopy. Analytical Chemistry 2005, 77 (11), 3611-3622.

21. Chaplin, T. D.; Clark, R. J. H.; McKay, A.; Pugh, S., Raman spectroscopic analysis of selected astronomical and cartographic folios from the early 13th century Islamic ‘Book of Curiosities of the Sciences and Marvels for the Eyes’. Journal of Raman Spectroscopy 2006, 37 (8), 865-877.

22. Gilbert, B.; Denoël, S.; Weber, G.; Allart, D., Analysis of green copper pigments in illuminated manuscripts by micro-Raman spectroscopy. Analyst 2003, 128 (10), 1213-1217.

23. Bette, S.; Kremer, R. K.; Eggert, G.; Tang, C. C.; Dinnebier, R. E., On verdigris, part I: synthesis, crystal structure solution and characterisation of the 1-2-0 phase (Cu3(CH3COO)2(OH)4). Dalton Trans 2017, 46 (43), 14847- 14858.

24. Bette, S.; Kremer, R. K.; Eggert, G.; Dinnebier, R. E., On verdigris, part II: synthesis of the 2-1-5 phase, Cu3(CH3COO)4(OH)25H2O, by long-term crystallisation from aqueous solution at room temperature. Dalton Trans 2018, 47 (25), 8209-8220.

25. Kühn, H., Verdigris and Copper Resinate. Studies in Conservation 1970, 15 (1), 12-36.

26. Gunn, M.; Chottard, G.; Rivière, E.; Girerd, J.-J.; Chottard, J.-C., Chemical Reactions Between Copper Pigments and Oleoresinous Media. Studies in Conservation 2002, 47 (1), 12-23.

127 27. Scott, D. A.; Khandekar, N.; Schilling, M. R.; Turner, N.; Taniguchi, Y.; Khanjian, H., Technical Examination of a Fifteenth-Century German Illuminated Manuscript on Paper: A Case Study in the Identification of Materials. Studies in Conservation 2001, 46 (2), 93-108.

28. Davies, B. H., Did the Medieval Illuminator Know How to Prevent Oxidative Damage? Dyes in History and Archaeology 2003, 19, 24-31.

29. Buti, D.; Rosi, F.; Brunetti, B. G.; Miliani, C., In-situ identification of copper- based green pigments on paintings and manuscripts by reflection FTIR. Anal Bioanal Chem 2013, 405 (8), 2699-711.

30. Muralha, V. S.; Burgio, L.; Clark, R. J., Raman spectroscopy analysis of pigments on 16-17th c. Persian manuscripts. Spectrochim Acta A Mol Biomol Spectrosc 2012, 92, 21-8.

31. Strlič, M.; Kolar, J., Ageing and stabilisation of paper. National and University Library, Turjaska: Ljubljana, Slovenia, 2005; p 211.

32. Strlic, M.; Kolar, J.; Pihlar, B., The effect of metal ion, pH and temperature on the yield of oxidising species in a fenton-like system determined by aromatic hydroxylation. Acta Chimica Slovenica 1999, 46, 555-566.

33. Strlič, M.; Kolar, J.; Šelih, V.-S.; Kočar, D.; Pihlar, B., A comparative study of several transition metals in Fenton-like reaction systems at circum-neutral pH. Acta Chim. Slov. 2003, 50, 619-632.

34. Šelih, V. S.; Strlič, M.; Kolar, J.; Pihlar, B., The role of transition metals in oxidative degradation of cellulose. Polymer Degradation and Stability 2007, 92 (8), 1476-1481.

35. Kanngießer, B.; Hahn, O.; Wilke, M.; Nekat, B.; Malzer, W.; Erko, A., Investigation of oxidation and migration processes of inorganic compounds in ink-corroded manuscripts. Spectrochimica Acta Part B: Atomic Spectroscopy 2004, 59 (10), 1511-1516.

36. Malešič, J.; Kolar, J.; Strlič, M.; Polanc, S., The influence of halide and pseudo-halide antioxidants in Fenton-like reaction systems. Acta Chim. Slov. 2006, 53, 450-456.

37. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

128 38. Thompson, D. V., The Materials and Techniques of Medieval Painting. Dover Publications: New York, 1956.

39. Pliny; Rackham, H.; Jones, W. H. S.; Eicholz, D. E., Natural history, Volume IX: Book 3-35. Harvard University Press ; W. Heinemann: Cambridge; London, 1979.

40. Phanstiel, O.; Dueno, E.; Wang, Q. X., Synthesis of Exotic Soaps in the Chemistry Laboratory. Journal of Chemical Education 1998, 75 (5), 612.

41. San Andrés, M.; de la Roja, J. M.; Baonza, V. G.; Sancho, N., Verdigris pigment: a mixture of compounds. Input from Raman spectroscopy. Journal of Raman Spectroscopy 2010, 41 (11), 1468-1476.

42. Salvadó, N.; Butí, S.; Cotte, M.; Cinque, G.; Pradell, T., Shades of green in 15th century paintings: combined microanalysis of the materials using synchrotron radiation XRD, FTIR and XRF. Applied Physics A 2013, 111 (1), 47-57.

43. Robinet, L.; Corbeil-a, M.-C., The Characterization of Metal Soaps. Studies in Conservation 2003, 48 (1), 23-40.

44. Scott, D. A., A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation 2000, 45 (1), 39- 53.

45. Naumova, M. M.; Pisareva, S. A., A note on the use of blue and green copper compounds in paintings. Studies in Conservation 1994, 39 (4), 277-283.

46. Richardin, P.; Mazel, V.; Walter, P.; Laprévote, O.; Brunelle, A., Identification of Different Copper Green Pigments in Renaissance Paintings by Cluster-TOF- SIMS Imaging Analysis. Journal of The American Society for Mass Spectrometry 2011, 22 (10), 1729-1736.

47. Pollard, A. M.; Thomas, R. G.; Williams, P. A., Synthesis and Stabilities of the Basic Copper(II) Chlorides Atacamite, Paratacamite and Botallackite. Mineralogical magazine 1989, 53 (373), 557-563.

48. Pierson, J. F.; Thobor-Keck, A.; Billard, A., Cuprite, paramelaconite and tenorite films deposited by reactive magnetron sputtering. Applied Surface Science 2003, 210 (3), 359-367.

129 49. Kolar, J.; Strlič, M. In Stabilisation of ink corrosion, The Iron Gall Ink Meeting, Newcastle upon Tyne, University of Northumbria: Newcastle upon Tyne, 2000; pp 135-140.

50. Moulder, J. F.; Chastain, J.; King, R. C., Handbook of X-ray Photoelectron Spectroscopy : a Reference Book of Standard Spectra for Identification and Interpretation of XPS Data. Physical Electronics: Eden Prairie, Minn., 1995.

51. Švarcová, S.; Čermáková, Z.; Hradilová, J.; Bezdička, P.; Hradil, D., Non- destructive micro-analytical differentiation of copper pigments in paint layers of works of art using laboratory-based techniques. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2014, 132, 514-525.

52. Hofmann, C.; Hartl, A.; Ahn, K.; Druceikaite, K.; Henniges, U.; Potthast, A., Stabilization of Verdigris. Journal of Paper Conservation 2016, 17 (3-4), 88- 99.

53. Miller, A. M.; Hanson, L., The maker and the monk: conservation of the Mercator Atlas of Europe. Journal of the Institute of Conservation 2010, 33 (1), 29-39.

54. Ahn, K.; Hofmann, C.; Horsky, M.; Potthast, A., How copper corrosion can be retarded--New ways investigating a chronic problem for cellulose in paper. Carbohydr Polym 2015, 134, 136-43.

55. Zaccaron, S.; Potthast, A.; Henniges, U.; Draxler, J.; Prohaska, T.; McGuiggan, P., The disastrous copper. Comparing extraction and chelation treatments to face the threat of copper-containing inks on cellulose. Carbohydr Polym 2019, 206, 198-209.

56. Masciocchi, N.; Corradi, E.; Sironi, A.; Moretti, G.; Minelli, G.; Porta, P., Preparation, Characterization, and ab initio X-Ray Powder Diffraction Study of Cu2(OH)3(CH3COO)·H2O. Journal of Solid State Chemistry 1997, 131 (2), 252-262.

57. Špec, T.; Retko, K.; Ropret, P.; Meden, A.; Bernard, J., The influence of UV– Vis radiation, and oscillations of temperature and relative humidity, on malachite alteration in the presence of different organic binders and varnishes. Journal of Raman Spectroscopy 2014, 45 (11-12), 1068-1075.

58. Quartieri, S., Synchrotron Radiation in Art, Archaelogy and Cultural Heritage. In Synchrotron Radiation: Basics, Methods and Applications, Mobilio, S.; Boscherini, F.; Meneghini, C., Eds. Springer Berlin Heidelberg: Berlin, Heidelberg, 2015; pp 677-695.

130 59. G. Banik, H. S., F. Mairinger, A. Vendl, J. Ponahlo, Analytical investigations of the problem of "Kupferfrass" in illuminated manuscripts. Mikrochim. Acta 1981, I, 49-55.

60. Ricciardi, P.; Pallipurath, A.; Rose, K., ‘It's not easy being green’: a spectroscopic study of green pigments used in illuminated manuscripts. Analytical Methods 2013, 5 (16), 3819.

61. Barkeshli, M.; Ataie, G. H., pH Stability of Saffron Used in Verdigris as an Inhibitor in Persian Miniature Paintings. In Restaurator, 2002; Vol. 23, p 154.

62. Kolar, J.; Strlič, M., Iron gall inks: on manufacture, characterization, degradation and stabilization. National and University Library Ljubljana, Slovenia, 2006; p 253.

63. Barfoed, C., Ueber die Nachweisung des Traubenzuckers neben Dextrin und verwandten Körpern. Zeitschrift für analytische Chemie 1873, 12 (1), 27-32.

64. Kolar, J., Mechanism of Autoxidative Degradation of Cellulose. In Restaurator, 1997; Vol. 18, p 163.

65. Strlič, M.; Kolar, J.; Pihlar, B., The effect of metal ion, pH and temperature on the yield of oxidising species in a Fenton-like system determined by aromatic hydroxylation. Acta Chim. Slov. 1999, 46 (4), 555-566.

66. Kolar, J.; Malešič, J.; Kočar, D.; Strlič, M.; De Bruin, G.; Koleša, D., Characterisation of paper containing iron gall ink using size exclusion chromatography. Polymer Degradation and Stability 2012, 97 (11), 2212- 2216.

67. Strlič, M.; Kolar, J.; Pihlar, B., Some preventive cellulose antioxidants studied by an aromatic hydroxylation assay. Polym. Deg. Stab. 2001, 73, 535-539.

68. Šala, M.; Kolar, J.; Strlič, M.; Kočevar, M., Synthesis of myo-inositol 1,2,3- tris- and 1,2,3,5-tetrakis(dihydrogen phosphate)s as a tool for the inhibition of iron-gall-ink corrosion. Carbohydr. Res. 2006, 341 (7), 897-902.

69. Ceres, G.; Conte, V.; Mirruzzo, V.; Kolar, J.; Strlič, M., Imidazolium-Based Ionic Liquids for the Efficient Treatment of Iron Gall Inked Papers. ChemSusChem 2008, 1 (11), 921-926.

70. Neevel, J. G., Phytate: a potential conservation agent for the treatment of ink corrosion caused by irongall inks. Restaurator 1995, 16, 143-160.

131 71. Neevel, J. G., (Im)possibilities of the phytate treatment of ink corrosion. In Contributions to conservation: Research in conservation at the Netherlands Institute for Cultural Heritage (ICN), J. Mosk, N. T., Ed. James & James Science Publishers: London, 2002; p 76.

72. Valenzuela, R.; Contreras, D.; Oviedo, C.; Freer, J.; Rodríguez, J., Copper catechol-driven Fenton reactions and their potential role in wood degradation. International Biodeterioration & Biodegradation 2008, 61 (4), 345-350.

73. Malešič, J.; Kočar, D.; Balažic Fabjan, A., Stabilization of copper- and iron- containing papers in mildly alkaline environment. Polymer Degradation and Stability 2012, 97 (1), 118-123.

74. Beeson, W. T.; Vu, V. V.; Span, E. A.; Phillips, C. M.; Marletta, M. A., Cellulose Degradation by Polysaccharide Monooxygenases. Annual Review of Biochemistry 2015, 84 (1), 923-946.

75. Hosny, W. M.; Hadi, A. K. A.; El‐Saied, H.; Basta, A. H., Metal chelates with some cellulose derivatives. Part III. Synthesis and structural chemistry of nickel (II) and copper (II) complexes with carboxymethyl cellulose. Polymer International 1995, 37 (2), 93-96.

76. Zhongyue, Q.; Yuyue, C.; Peng, Z.; Guangyu, Z.; Yan, L., Structure and properties of Cu(II) complex bamboo pulp fabrics. Journal of Applied Polymer Science 2010, 117 (3), 1843-1850.

77. Faubel, W.; Staub, S.; Simon, R.; Heissler, S.; Pataki, A.; Banik, G., Non- destructive analysis for the investigation of decomposition phenomena of historical manuscripts and prints. Spectrochimica Acta Part B: Atomic Spectroscopy 2007, 62 (6), 669-676.

132 Chapter 7

MULTIDISCIPLINARY LEARNING: REDOX CHEMISTRY AND PIGMENT HISTORY

7.1 Introductory Remarks

In this chapter, I demonstrate the use of copper-based pigments, verdigris, for teaching purposes through synthesis, degradation, and analysis. The synthesis processes is explained, followed by the application and degradation being introduced from an educational viewpoint. This chapter thus shows the benefit of this thesis’s work for education and outreach activities. This work was published in the journal Journal of Chemical Education in 2018. Figures and text in this chapter are reprinted with permission. The author list was Marcie B. Wiggins, Emma Heath, and Jocelyn Alcántara-García.

7.2 Introduction Interdisciplinary classes are becoming increasingly popular, as numerous workshops,1 projects,2-4 and courses5−9 illustrate. These efforts merge the arts and humanities with science, which has proven successful at illustrating challenging topics in science curricula.10−12 Emerging art conservators, as students, are preparing themselves for the challenge of keeping our material cultural heritage safe. This preparation requires incorporating more science into art conservation curricula. Art conservation students learn hands-on and preventive conservation techniques, connoisseurship, (art) history, documentation, and applied conservation science, to

133 mention just a few areas of the vast curriculum. Conservation professionals must learn fundamental principles of chemistry and physics, as well as some characterization techniques. Because of students’ interest in applied material science, in addition to time constraints, accelerated degradation, often referred to as “accelerated aging” or “artificial aging”, allows us to replicate chemical phenomena that students can easily relate to historic, “naturally degraded” objects. Although the interest of this audience is very specific, similar experiments can be successfully adapted for college and high school students, who are not necessarily interested in cultural heritage. The experiments detailed in this chapter show successful and collaborative strategies to teaching science to both nonscientific and scientific audiences. The set of activities can be easily adapted to audiences from college-level general chemistry to art conservation students. The latter may use similarly tailored projects to address unique conservation-related questions throughout their careers. As such questions are often related to the original state of the material, synthesis and evaluation of those materials can and will likely affect treatment, housing, and display decisions. Additionally, the role of these materials on “problem-substrates”, or substrates such as paper and paint that react and degrade with the material, contributes to these decisions. The materials in question can also be used to illustrate simpler subject matter to younger chemistry students. Integrating artists’ materials as a part of a general chemistry laboratory experiment connects lessons to real world situations. Therefore, this project presents a format that can grow and expand from fundamental high school chemistry, to undergraduate science courses, to conservation-related research projects.

134 7.2.1 Copper-Containing Artists’ Materials Copper reactivity offers seemingly endless opportunities to teach chemistry. Besides their high reactivity, copper compounds are ideal candidates for teaching several science topics due to both their relatively low cost and toxicity. Copper compounds provide a wide range of colors, from various shades of blue to red, depending on their chemical state (oxidation states, crystal structures, ligands, etc.).13−18 Numerous historic pigments are copper-based, and artists’ treatises contain guidelines for their preparation and use (supported by modern analytical investigation), e.g., blue azurite (Cu3(CO3)2(OH)2) for Mary in Christian depictions.19−22 Because of the varying interpretation of these historical synthesis descriptions, these pigments provide interesting and challenging real world, artistic, and historic examples for students. Verdigris (copper(II) acetate) is possibly one of the most widely used pigments in wart since antiquity. This material can range in color from pale blue to a dark emerald, based on its crystal structure, which is related to its synthesis. The most common form is neutral verdigris (Cu(CH3COO)2·H2O). Historic recipes for verdigris are as simple as described by Pliny:19,23 [I]t is scraped off the stone from which copper is smelted or by drilling holes in white copper and hanging it up in casks of strong vinegar

which is stopped with a lid... Metallic copper reacts with acetate ions to form the oxidized blue-green corrosion crust, which is harvested for painting and/or writing material. Although suspending copper materials over acetic acid is the simplest of methods, more complicated syntheses are also described in Pliny’s and other treatises:19,24

135 ...made by grinding up in a mortar of true Cyprian copper with a pestle of the same metal equal weights of alum and salt or soda with the very strongest white vinegar. This preparation is only made on the very hottest day of the year, about the rising of the Dogstar. The mixture is ground up until it becomes of a green color... To remedy any that is blemished, the urine of a young boy to twice the quantity of vinegar that was used is added to the mixture. Verdigris has been extensively used in many works of art, including, but not limited to, 17th century illuminated manuscripts, where the pigment was identified using Raman spectroscopy.25 More commonly, however, verdigris is not unequivocally identified, but discolored areas corresponding to otherwise blue or green (sea in a map, leaves of a plant) are assumed to contain it (Figure 7-1). On paper, these remarkably brittle areas represent oxidized cellulose.26−30 Verdigris’ low stability was well- documented even in the 15th century. In “Il libro dell’arte”, which is pivotal to understand historic artistic practices, Cennino Cennini states that verdigris “makes a green for grass most perfect and beautiful to the eye, but not durable”.31 Objects with verdigris are of interest in cultural heritage and serve as an engaging case study for both fundamental chemistry and art conservation alike.

136

This set of activities is adaptable for different students and subject matter. Herein, we describe how we utilized copper-based pigment synthesis to teach redox reactions in a 200-level undergraduate general chemistry course. Building on this concept, we highlighted its historical use and relevance in material cultural heritage, i.e., watercolors on paper. We kept students engaged in the activities through the use of their own pigments to create their own artwork. We used the pigments and artwork

137 they prepared to work with emerging art conservators. The secondary part of the laboratory activity used extreme conditions of relative humidity and temperature to induce the artworks’ decay (artificial/accelerated degradation). Degradation of paper, a combination of acid hydrolysis and oxidation, is a visual effect and can be monitored by instrumental methods, as we later explain.

7.2.2 Learning Objectives Two methods of synthesis are described below. Each can be applied to specific chemistry lessons. The simple method or direct synthesis19,23 of verdigris illustrates redox chemistry, limiting reagents, stoichiometry, and the role of d-electrons in color, to name a few concepts. The multistep method,8 on the other hand, can be used for explaining ligand field theory and coordination chemistry, acidity, and crystallization methods, in addition to the concepts mentioned before. The follow-up activity (degradation of cellulose) is a prime example of organic redox chemistry, which is seldom covered. Last, the role of these pigments in the degradation processes shows the applicability of chemistry in day-to-day life. For an upper-level undergraduate chemistry or a graduate art conservation course, characterization techniques can be easily added to the experimental design. Namely, X-ray diffraction (XRD), Raman and Fourier-transform infrared (FTIR) spectroscopies, and X-ray photoelectron spectroscopy (XPS) were used to to track chemical changes.25,29,32,33

138 7.3 Experimental Summary

7.3.1 Synthesis of Verdigris

The synthesis of verdigris can be performed in one of two ways: the simple or the multistep synthesis. For the simple synthesis of verdigris, a copper plate, tube or wire is suspended over acetic acid inside a sealed container (preferably glass). It is advisable to do a demo and/or set up this experiment many weeks before the session where it is to be used, as the synthesis is slow. Depending on the acid’s concentration, this synthesis can take anywhere between 2 days and 6 months. Figure 7-2 illustrates a green-blue “crust,” which formed after 3 weeks with 50% (v/v) acetic acid.

Figure 7-2 A copper sheet suspended on plastic cups over 50% (v/v) acetic acid. It was sealed in a glass container for 3 weeks.

An alternative synthesis is the multi-step procedure similar to that of Solomon et al.8 Students split into groups of three to four to prepare their verdigris pigments

139 from copper sulfate. The first step in the procedure (Reaction 1) is the precipitation of basic copper(II) sulfate by adding drop-wise ammonia (29% v/v, Consolidated Chemical & Solvents LLC) to a stirring solution of copper sulfate pentahydrate (>99%, Sigma-Aldrich):

4(CuSO4•5H2O) (aq) + 6NH3 (aq) → CuSO4•3Cu(OH)2 (s) + 3(NH4)3SO4 (aq)

+ 14H2O (l) (1) The reaction is complete when the solution turns a deep royal blue color

(Figure 7-3). The precipitated CuSO4•3Cu(OH)2 is collected via gravitational filtration and rinsed with distilled water. It is then re-dissolved in water and reacts with a 5% w/w sodium hydroxide (≥97%, Sigma-Aldrich) solution to precipitate copper(II) hydroxide (Cu(OH)2), as indicated in Reaction 2.

CuSO4•3Cu(OH)2 (s) + 2NaOH (aq) → 4Cu(OH)2 (s) + Na2SO4 (aq) (2)

Cu(OH)2 is collected via gravitational filtration and rinsed with distilled water. It is then reacted with in acetic acid (5% v/v, Sigma-Aldrich), and allowed to crystalize at room temperature for approximately a week. The pigment crystals, copper acetate monohydrate, are neutral verdigris (Cu(CH3COO)2•H2O). Students then collected the product and continued with the applications during the next laboratory period.

140

Figure 7-3 The copper solution after the addition of ammonia, which prompted the formation of CuSO4•3Cu(OH)2.

7.3.2 Application Students used their own synthesized pigments to create their artwork either as dyes (using concentrated aqueous solutions of the pigment) or watercolors (using gum Arabic as a binder). When applied as a dye, students dissolved approximately 0.5 g of pigments into approximately 10 mL of water to achieve approximately a 5% (w/w) verdigris solution. The students prepared their watercolors by carefully warming up the remaining verdigris solution (40-50°C) and adding approximately 1.3 g of gum Arabic to prepare a 10-15% (w/w) thick solution. This watercolor solution was described as “syrupy.”

Each student chose his or her “ideal working viscosity” by adding less or more gum. It is important that these watercolors never boil, as they would promote reduction of the pigment, as illustrated in Figure 7-4, or simply burn the gum. To avoid potential reduction reactions in students’ watercolors, the gum Arabic should be added slowly and on low-to-medium heat (never exceeding 60°C). Also, “swelling” of the gum Arabic in the solution at room temperature, where the gum is concentrated at the

141 bottom of the container for extended periods of time, may facilitate this reduction reaction, so care should be taken to avoid this.

Figure 7-4 Copper (I) oxide forms for Fehling’s reaction upon addition of excess heat to the verdigris and gum Arabic mixtures.

Students prepared duplicates of all artwork, created using paintbrushes, on Whatman filter paper No. 1. The use of a 100% cellulose paper with no additives was crucial, as any sizing agent or whitener may interfere with chemistry we wanted to observe. Most students took some pride in their artwork, thus becoming personally invested in the outcomes of the aging experiment, including preparation “accidents” that we will discuss in detail later (Figure 7-4). Alongside these samples, several standardized samples were prepared with the same materials for the aging study (detailed in next section).

142 7.3.3 Aging Study Whatman filter paper No. 1 (3.0 cm x 1.0 cm) was dipped in the solutions to apply the pigment to the paper, both as a dye and a watercolor. The pigments used for the study were commercially available verdigris (Kremer), verdigris prepared by the simple method, and verdigris prepared by the multi-step method. The paper was allowed to dry before aging. Students were encouraged to make duplicates of their design, so one set could be aged and one served as a control for comparison. Accelerated aging was carried out with an ESPEC BTL 433 test chamber located at the Winterthur Museum. Both the pigmented Whatman paper and control uncolored Whatman paper were aged under the same conditions (60ºC, 85% RH).35,36 Samples of the standardized paper were taken at 0, 1, 3, 5 and 7 days to allow students to observe the changes induced by the presence of copper-based pigment on the cellulose. Students’ artworks were removed after 7 days to see the most drastic changes. Visible changes were clearly observed as discoloration, or “browning”, of the pigmented Whatman filter paper compared to the aged controlled paper. Figure 7-5 compares the aging of the pigmented papers with controls. At the high school and undergraduate level, these visual changes exemplify the interactions taking place between the copper-based verdigris and the organic substrate, cellulose and in some cases gum Arabic, as a result of accelerated degradation. This is related to why many historical documents, maps, etc. appear brown and discolored today, and these students can see this effect in a series of mock-ups, as well as in their own artworks.

143

Figure 7-5 Whatman filter papers No. 1 aged over 7 days (60°C, 85%RH) matted on white board to observe various verdigris pigments applied as a dye or watercolor: A) no pigment, B) multi-step synthesized verdigris in water (dyed by students), C) commercially available verdigris in water, D) simple synthesized verdigris, E) multi-step synthesized verdigris in water, F) gum Arabic, G) commercially available verdigris in gum Arabic, H) simple synthesized verdigris in gum Arabic, and I) multi-step synthesized verdigris in gum Arabic.

For art conservation graduate students, this sample served as a case study to observe the chemical changes taking place in paper painted with copper pigments. This case study was used as a part of the introduction to FTIR spectroscopy. The behavior indicated in Figure 7-6 is a direct consequence of the degradation promoted by the pigments via mostly acid hydrolysis and cellulose’s oxidation.37

144

Figure 7-6 FTIR spectra of aged Whatman paper dyed with simple synthesized verdigris, with noted bands at 1316 cm-1 (solid arrow), 1425 cm-1 (small dashed arrow), and 1560 cm-1 (big dashed arrow).

7.4 Hazards and Safety Most starting compounds used are inexpensive and nonhazardous for students of all ages. The greatest concern arises from the dilution of concentrated acetic acid and sodium and ammonium hydroxides to the desired concentrations. Undergraduate students performed this without any problems, but these solutions can be diluted by the instructor in future experiments.

7.5 Outcomes and Critical Assessments

Although the synthesis, application, and aging were the main elements of this activity, there were numerous spontaneous conversations because of unplanned student observations.

145 1. Yields were lower than expected. Although students followed a given procedure based on the report of Solomon et al.,8 not all teams obtained the same quantity. This framed discussions on limiting reagent and molarity, although they had not been covered at that point. 2. Students crystallized ammonium copper complexes by accident likely due to excess ammonia in the first step of the multistep synthesis. Students reported it was hard to know when to stop, as the “deep blue” mentioned in ref [8] occurred for some students at the beginning of the

addition, whereas others added almost twice as much of the reactant. Excess ammonia in the first step resulted in ammonium copper crystals in the final product (appearing dark blue compared to verdigris crystals). This framed discussions on Le Chatelier’s principle, “complex formation, acid–base chemistry and solubility”.13 3. Addition of the gum Arabic initiated a color change. Gum Arabic consists of natural gums (monosaccharide sugars), and, historically, it was commonly used as a paint binder for watercolors. Depending on the purity of the students’ final product, the verdigris solution appeared either lighter or deeper green-blue. In all cases, the addition of gum

Arabic caused the solutions to become a uniform deep blue-green (Figure 7-7). Some groups used more of the gum Arabic resulting in thicker watercolors. Excess gum Arabic or heating the solutions for too long or at a higher temperature caused the mixtures to change colors and within a few minutes to form a red precipitate (Figure 7-4). These “accidents” framed discussions on the formation of copper(I) oxide, the

146 Fehling’s reaction, and redox reactions on organic compounds. The test of the same name identifies reducing sugars,38,39 which have unique roles in verdigris-pigmented artwork decay.40

Figure 7-7 The appearance of the dissolved verdigris changed upon the addition of gum Arabic from a light, semi-insoluble blue crystal (left) to a deep- green solution (right).

7.6 Assessment Altogether, synthesis of pigments, preparation of watercolors, applications, and aging studies made an engaging activity in which students learned about artists’ materials and how chemistry and art can be studied in tandem. These laboratory-based activities proved to be characteristically attractive to a broad audience, because they allowed for a visual assessment of a real-life situation. Additionally, these activities illustrated various aspects of the chemistry of historical objects and fundamental chemistry topics through synthesis and application. We encourage a visit to a museum or archive after finalizing these activities, so students can compare the rapid chemistry they did in a laboratory to naturally aged cultural heritage.

147 Students were personally invested in the activities, because they synthesized and prepared their own artist’s materials. This translated into first-hand interest in the degradation processes: from generic objects to “my” objects. Students already interested in degradation processes (art conservation and related) took special interest in aging, as historic materials rarely exhibit the colors they present when freshly applied. Upper-level students benefited from the collaborative activities that provided an interesting and multidisciplinary context for learning characterization techniques. The samples’ wide span of visual results sparked other more in-depth discussions, such as the role various chemical substrates (fibers, fillers, sizing agents) and media (oil, gum, water) play into either promoting or slowing down degradation. These activities are versatile both for audiences and materials. The laboratory experiments, as written, can be easily adapted to teach redox chemistry of inorganic and organic compounds, coordination chemistry, and/or degradation mechanisms. Different aspects of the activity can be highlighted on the basis of the age group being instructed, from high school students to master’s-level art conservation students. The chemistry of numerous other artists’ materials is well-known for their effects on material culture degradation. Hence, with the interchangeability of various artists’ materials like Prussian blue (Fe4(FeCN6)3) or historic inks (copper- or iron-based), instructors can adjust the subject matter to better suit their lessons and materials.

7.7 Limitations Drawbacks regarding this activity include its time-consuming nature: from the verdigris synthesis to the accelerated aging. However, this limitation can be easily mitigated by performing some of the activities beforehand. For instance, the simple synthesis can be set up one semester/year in advance by students from a semester/year

148 ahead. Students would then use the products slowly synthesized for over a semester/year, in addition to setting up the experiment for the next class. Courses that meet once every week over the semester fit this activity well if several laboratory periods are available for this activity. Instrumentation for this activity, such as a commercial accelerated aging unit and an FTIR spectrometer, was available. However, alternative instrumentation can be used for similar results, such as custom accelerated degradation systems and fiber-optic and UV–vis spectrometers.41−43 These pigments are so reactive that alternative and simplified degradation chambers are effective. For instance, any sealed container with saturated ionic salts solutions will hold said relative humidity.44 If these boxes are placed in a warm environment (like in direct sunlight), the temperature will rise, resulting in similar changes. Another alternative is placing samples inside a dark container with UV lamps set with timers. The research aim is to speed up degradation in a laboratory setting “in order to elucidate the chemical reactions involved and the physical consequences thereof”.45

7.8 Acknowledgements I would like to thank University of Delaware’s ARTC/CHEM 210 Science of Color Phenomena classes (Spring and Fall 2017). I would also like to thank the Andrew W. Mellon Foundation and Winterthur Museum, Garden, and Library for the use of the ESPEC BTL 433 test chamber, and I would like to thank Dr. Karl S. Booksh for use of the ATR-FTIR for measurements.

7.9 Concluding Remarks

This chapter exemplifies how the use of verdigris as a simple, yet reactive pigment can be used to better educate a wide audience. The pigment synthesis,

149 degradation, and analysis utilize a real-world example to engage students and communicate complex scientific concepts across disciplines. As a whole, this work contextualizes the studies from the earlier chapters, especially Chapter 6, in the broader world, and connects it with possible outreach activities.

150 REFERENCES

1. Halpine, S. M., Introducing Molecular Visualization to Primary Schools in California: The STArt! teaching Science Through Art Program. Journal of Chemical Education 2004, 81 (10), 1431-1436.

2. Bopegedera, A. M. R. P., The Art and Science of Light. An Interdisciplinary Teaching and Learning Experience. Journal of Chemical Education 2005, 82 (1), 55-59.

3. Buechley, L.; Eisenberg, M.; Catchen, J.; Crockett, A., The LilyPad Arduino: using computational textiles to investigate engagement, aesthetics, and diversity in computer science education. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, ACM: Florence, Italy, 2008; pp 423-432.

4. Kesner, L.; Eyring, E. M., Service-Learning General Chemistry: Lead Paint Analyses. Journal of Chemical Education 1999, 76 (7), 920-923.

5. Esson, J. M.; Scott, R.; Hayes, C. J., Chemistry and Art: Removal of Graffiti Ink from Paints Grounded in a Real-Life Scenario. Journal of Chemical Education 2018, 95 (3), 400-402.

6. Giménez, J., Finding Hidden Chemistry in Ancient Egyptian Artifacts: Pigment Degradation Taught in a Chemical Engineering Course. Journal of Chemical Education 2015, 92 (3), 456-462.

7. Harper, C. S.; Macdonald, F. V.; Braun, K. L., Lipid Residue Analysis of Archaeological Pottery: An Introductory Laboratory Experiment in Archaeological Chemistry. Journal of Chemical Education 2017, 94 (9), 1309-1313.

8. Solomon, S. D.; Rutkowsky, S. A.; Mahon, M. L.; Halpern, E. M., Synthesis of Copper Pigments, Malachite and Verdigris: Making Tempera Paint. Journal of Chemical Education 2011, 88 (12), 1694-1697.

9. Hoffman, E. M.; Beussman, D. J., Paint Analysis Using Visible Reflectance Spectroscopy: An Undergraduate Forensics Lab. Journal of Chemical Education 2007, 84 (11), 1806-1808.

151 10. Bequette, J. W.; Bequette, M. B., A Place for Art and Design Education in the STEM Conversation. Art Education 2012, 65 (2), 40-47.

11. Walling, D. R., Rethinking Visual Arts Education: A Convergence of Influences. Phi Delta Kappan 2001, 82 (8), 626-28,30-31.

12. Labianca, D. A., Science for the nonscience major through interdisciplinary study: The interrelation of science and art. Science Education 1975, 59 (2), 187-190.

13. Johnson, A. R.; McQueen, T. M.; Rodolfa, K. T., Species Distribution Diagrams in the Copper-Ammonia System: An Updated and Expanded Demonstration Illustrating Complex Equilibria. Journal of Chemical Education 2005, 82 (3), 408-414.

14. Rorabacher, D. B., Electron transfer by copper centers. Chemical Reviews 2004, 104 (2), 651-697.

15. Earl, B. L., The colorful complexes of copper(II). Journal of Chemical Education 1985, 62 (9), 798.

16. Baker, A. T., The Ligand Field Spectra of Copper(II) Complexes. Journal of Chemical Education 1998, 75 (1), 98-99.

17. Cui, A.-L.; Chen, X.; Sun, L.; Wei, J.-Z.; Yang, J.; Kou, H.-Z., Preparation and Thermochromic Properties of Copper(II)-N,N-Diethylethylenediamine Complexes. Journal of Chemical Education 2011, 88 (3), 311-312.

18. Parmeggiani, F.; Sacchetti, A., Preparation and Luminescence Thermochromism of Tetranuclear Copper(I)–Pyridine–Iodide Clusters. Journal of Chemical Education 2012, 89 (7), 946-949.

19. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

20. Feller, R. L., Artists' pigments : a handbook of their history and characteristics. National Gallery of Art: Washington, 1986; Vol. 1.

21. Roy, A., Artists' pigments : a handbook of their history and characteristics. National Gallery of Art: Washington, 1986; Vol. 2.

22. Muller, N. E., Three Methods of Modelling the Virgin's Mantle in Early Italian Painting. Journal of the American Institute for Conservation 1978, 17 (2), 10-18.

152 23. Pliny; Rackham, H.; Jones, W. H. S.; Eicholz, D. E., Natural history, Volume IX: Book 34.26. Harvard University Press ; W. Heinemann: Cambridge; London, 1979.

24. Pliny; Rackham, H.; Jones, W. H. S.; Eicholz, D. E., Natural history, Volume IX: Book 34.28. Harvard University Press ; W. Heinemann: Cambridge; London, 1979.

25. Chaplin, T. D.; Clark, R. J. H.; Martinón-Torres, M., A combined Raman microscopy, XRF and SEM–EDX study of three valuable objects – A large painted leather screen and two illuminated title pages in 17th century books of ordinances of the Worshipful Company of Barbers, London. Journal of Molecular Structure 2010, 976 (1), 350-359.

26. Hofmann, C.; Hartl, A.; Ahn, K.; Faerber, I.; Henniges, U.; Potthast, A., Studies on the Conservation of Verdigris on Paper. In Restaurator. International Journal for the Preservation of Library and Archival Material, 2015; Vol. 36, p 147.

27. Hofmann, C.; Hartl, A.; Ahn, K.; Druceikaite, K.; Henniges, U.; Potthast, A., Stabilization of Verdigris. Journal of Paper Conservation 2016, 17 (3- 4), 88-99.

28. Miller, A. M.; Hanson, L., The maker and the monk: conservation of the Mercator Atlas of Europe. Journal of the Institute of Conservation 2010, 33 (1), 29-39.

29. Scott, D. A.; Khandekar, N.; Schilling, M. R.; Turner, N.; Taniguchi, Y.; Khanjian, H., Technical Examination of a Fifteenth-Century German Illuminated Manuscript on Paper: A Case Study in the Identification of Materials. Studies in Conservation 2001, 46 (2), 93-108.

30. Lee, L. R.; Thompson, A.; Daniels, V. D., Princes of the House of Timur: conservation and examination of an early Mughal painting. Studies in Conservation 1997, 42 (4), 231-240.

31. Cennini, C.; Thompson, D. V., The craftsman's handbook. Dover Pubications: New York, 1954.

32. Wiggins, M. B.; Alcántara-García, J.; Booksh, K. S., Characterization of copper-based pigment preparation and alteration products. MRS Advances 2017, 2 (63), 3973-3981.

153 33. Švarcová, S.; Čermáková, Z.; Hradilová, J.; Bezdička, P.; Hradil, D., Non- destructive micro-analytical differentiation of copper pigments in paint layers of works of art using laboratory-based techniques. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2014, 132, 514- 525.

34. Salvadó, N.; Pradell, T.; Pantos, E.; Papiz, M. Z.; Molera, J.; Seco, M.; Vendrell-Saz, M., Identification of copper-based green pigments in Jaume Huguet's Gothic altarpieces by Fourier transform infrared microspectroscopy and synchrotron radiation X-ray diffraction. Journal of Synchrotron Radiation 2002, 9 (4), 215-222.

35. Kolar, J.; Malešič, J.; Kočar, D.; Strlič, M.; De Bruin, G.; Koleša, D., Characterisation of paper containing iron gall ink using size exclusion chromatography. Polymer Degradation and Stability 2012, 97 (11), 2212- 2216.

36. Strlič, M.; Kolar, J.; Pihlar, B., Some preventive cellulose antioxidants studied by an aromatic hydroxylation assay. Polymer Degradation and Stability 2001, 73 (3), 535-539.

37. Pereira, D. C.; de Faria, D. L. A.; Constantino, V. R. L., CuII hydroxy salts: characterization of layered compounds by vibrational spectroscopy. jbchs Journal of the Brazilian Chemical Society 2006, 17 (8), 1651-1657.

38. Robyt, J. F., Essentials of carbohydrate chemistry. Springer: New York, 1998.

39. Scherz, H., Ein Beitrag zur Kenntnis der Fehling'schen Reaktion. Zeitschrift für Lebensmittel-Untersuchung und Forschung 1979, 168 (2), 91-98.

40. Wiggins, M. B.; Heath, E.; Booksh, K. S.; Alcántara-García, J., Verdigris alteration species on organic substrates by spectroscopic methods (Unpublished work). 2018.

41. Safford, H. W.; Westneat, D. F., An inexpensive, easily constructed spectrophotometer. Journal of Chemical Education 1953, 30 (7), 343-347.

42. Grasse, E. K.; Torcasio, M. H.; Smith, A. W., Teaching UV–Vis Spectroscopy with a 3D-Printable Smartphone Spectrophotometer. Journal of Chemical Education 2016, 93 (1), 146-151.

154 43. Albert, D. R.; Todt, M. A.; Davis, H. F., A Low-Cost Quantitative Absorption Spectrophotometer. Journal of Chemical Education 2012, 89 (11), 1432-1435.

44. Greenspan, L., Humidity fixed points of binary saturated aqueous solutions. J. RES. NATL. BUR. STAN. SECT. A. Journal of Research of the National Bureau of Standards Section A: Physics and Chemistry 1977, 81A (1), 89-96.

45. Feller, R. L., Accelerated aging : photochemical and thermal aspects. Getty Conservation Institute: Marina del Rey, CA, 1994.

155 Chapter 8

CONCLUSIONS

8.1 Summary Copper salts are widely used as pigments in Western and Eastern artworks, but suffer from alteration and degradation through interaction with their organic surroundings over time.1-5 Therefore, chemical analysis of these copper salts, such as verdigris (copper acetates) and basic copper chlorides, are needed to learn more about the origin and condition of the artwork. Spectroscopic studies can be used to provide evidence of the manufacturing and usage of the original pigments, and as a means to propose pathways for its alteration and degradation in cultural heritage objects. The literature review of Chapter 2 discusses previous studies on copper acetates’ and basic copper chlorides’ usage, characterization, and degradation in cultural heritage.4-6 This overview reveals some gaps in the knowledge of how basic copper chlorides were used and manufactured and highlights the lack of fundamental studies of verdigris with cellulose and gum Arabic containing media. This dissertation focuses on filling these voids with spectroscopic analyses. These analyses rely heavily on XRD, XPS, FTIR, and Raman spectroscopy. Analysis of green Chinese architectural paints revealed that the use of atacamite and botallackite, both copper chloride trihydroxide polymorphs, increased over two different dynasties. This increased usage suggests they were a part of a transitional period in Chinese art material during which artists moved from naturally available materials to cheaper synthetic ones.7-8 With Raman spectroscopic imaging

156 and MCR-ALS, different relative ratios of these polymorphs could be distinguished in several paint samples, suggesting different syntheses were used on the objects. These studies collectively report on the origin, usage, and manufacturing of basic copper chloride pigments in Chinese architectural paints. Preparation and alteration studies of verdigris inform the possible starting materials and degradation products expected on cultural heritage objects where it might have been used. Reproducing historical recipes of verdigris yielded two different crystal types, neutral and basic verdigris.9-11 Inducing degradation of these materials with and without cellulose revealed their alteration to an intermediate species and then to copper (II) oxides. Reactions with gum Arabic showed the reduction of the copper species, which might serve to stabilize copper on paper. These materials also proved effective teaching tools for science and non-science majors in a general chemistry course. The proposed lesson plan can be adjusted for subject matter and audiences, while maintaining the interesting, real-world application of degrading copper-salts on paper.12 Therefore, these studies contribute new information to the degradation pathway of verdigris on paper-based media. In the field of heritage science, goals range from characterizing materials to evaluating treatment methods. As a whole, the research presented in this dissertation utilized a range of spectroscopic techniques to contribute to these goals regarding copper salts, specifically verdigris and basic copper chlorides. These fundamental results about compositions and possible pathways will directly guide future historical research and conservation treatments.

157 8.2 Future Directions Moving forward with these projects, more can be done to study the effects on the organic surroundings of the copper salts to better understand possible inorganic- organic interactions. While studying the degradation of the cellulose was not a primary focus of these studies, it is necessary to explore in the future. FTIR spectroscopy had proved limiting in this study, but preliminary work using XPS and ToF-SIMS was promising. These results showed cellulose alterations co-localizing with copper ions and an increase in carboxyl groups. Additional benefits can come from implementing mapping and chemometric analyses for distinguishing components in complex mixtures related to cultural heritage. Raman imaging and MCR-ALS is useful in extracting component spectra of small amounts of materials, which can be poor Raman scatters, by signal averaging over several spectra.13-14 Then, those components can be re-mapped to spatially distinguish them within the sample. There is ample opportunity for this method to be used to differentiate degradation and alteration products within samples to monitor their formation and continued interactions. However, this requires expanding the work to include additional imaging techniques, such as FTIR, XPS, and ToF-SIMS. These techniques would provide additional chemical information for organic or oxidation state changes, which is necessary to monitor interactions and degradation within the samples. Copper salts play a very important role in the field of cultural heritage. These objects frequently contain challenging mixtures undergoing several reactions over their lifetime. Therefore, spectroscopic studies, with minimal invasiveness, can and have provided useful information to historians, conservators, and scientists. Additional implications of the methodologies and fundamental studies presented in this

158 dissertation could continue to contribute to the field in order to better understand and preserve cultural heritage artifacts.

159 REFERENCES

1. Naumova, M. M.; Pisareva, S. A., A Note on the Use of Blue and Green Copper Compounds in Paintings. Studies in Conservation 1994, 39 (4), 277- 283.

2. Zoleo, A.; Nodari, L.; Rampazzo, M.; Piccinelli, F.; Russo, U.; Federici, C.; Brustolon, M., Characterization of Pigment and Binder in Badly Conserved Illuminations of a 15th‐Century Manuscript. Archaeometry 2014, 56 (3), 496- 512.

3. Santoro, C.; Zarkout, K.; Le Ho, A.-S.; Mirambet, F.; Gourier, D.; Binet, L.; Pagès-Camagna, S.; Reguer, S.; Mirabaud, S.; Le Du, Y.; Griesmar, P.; Lubin- Germain, N.; Menu, M., New Highlights on degradation process of verdigris from easel paintings. Applied Physics A 2014, 114, 637-645.

4. Scott, D. A., Copper and Bronze in Art: Corrosion, Colorants, Conservation. Getty Conservation Institute: Los Angeles, 2002.

5. Scott, D. A.; Taniguchi, Y.; Koseto, E., The verisimilitude of verdigris: a review of the copper carboxylates. Studies in Conservation 2013, 46 (2), 73- 91.

6. Scott, D. A., A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation 2000, 45 (1), 39- 53.

7. Li, M. Analysis and Research on Copper Green Pigment. Northwest University, Xi’an, Shaanxi,China, 2013.

8. Yong, L., Copper trihydroxychlorides as pigments in China. Studies in Conservation 2012, 57 (2), 106-111.

9. Wiggins, M. B.; Alcántara-García, J.; Booksh, K. S., Characterization of copper-based pigment preparation and alteration products. MRS Advances 2017, 2 (63), 3973-3981.

160 10. San Andrés, M.; de la Roja, J. M.; Baonza, V. G.; Sancho, N., Verdigris pigment: a mixture of compounds. Input from Raman spectroscopy. Journal of Raman Spectroscopy 2010, 41 (11), 1468-1476.

11. Salvadó, N.; Pradell, T.; Pantos, E.; Papiz, M. Z.; Molera, J.; Seco, M.; Vendrell-Saz, M., Identification of copper-based green pigments in Jaume Huguet's Gothic altarpieces by Fourier transform infrared microspectroscopy and synchrotron radiation X-ray diffraction. Journal of Synchrotron Radiation 2002, 9 (4), 215-222.

12. Wiggins, M. B.; Heath, E.; Alcántara-García, J., Multidisciplinary Learning: Redox Chemistry and Pigment History. Journal of Chemical Education 2019, 96 (2), 317-322.

13. Offroy, M.; Moreau, M.; Sobanska, S.; Milanfar, P.; Duponchel, L., Pushing back the limits of Raman imaging by coupling super-resolution and chemometrics for aerosols characterization. Scientific Reports 2015, 5, 12303.

14. Smith, J. P.; Smith, F. C.; Ottaway, J.; Krull-Davatzes, A. E.; Simonson, B. M.; Glass, B. P.; Booksh, K. S., Raman Microspectroscopic Mapping with Multivariate Curve Resolution–Alternating Least Squares (MCR-ALS) Applied to the High-Pressure Polymorph of Titanium Dioxide, TiO2-II. Applied Spectroscopy 2017, 71 (8), 1816-1833.

161 Appendix

PERMISSION FOR MATERIAL REPRINT

Text extracts in Chapter 6 were reprinted with permission from Characterization of copper-based pigment preparation and alteration products by Marcie B. Wiggins, Jocelyn Alcántara-García, and Karl S. Booksh in MRS Advances (Cambridge University Press) January 14, 2018. Copyright © Materials Research Society 2018. See below.

162 Chapter 7 was reprinted with permission from Multidisciplinary Learning: Redox Chemistry and Pigment History by Marcie B. Wiggins, Emma Heath, Jocelyn Alcántara-García in Journal of Chemical Education (American Chemical Society) November 1, 2018. Copyright © 2018 American Chemical Society. See below.

11/30/2018 Rightslink® by Copyright Clearance Center

Title: Multidisciplinary Learning: LOGIN Redox Chemistry and Pigment If you're a copyright.com History user, you can login to Author: Marcie B. Wiggins, Emma Heath, RightsLink using your Jocelyn AlcántaraGarcía copyright.com credentials. Already a RightsLink user or Publication: Journal of Chemical Education want to learn more? Publisher: American Chemical Society Date: Nov 1, 2018 Copyright © 2018, American Chemical Society

PERMISSION/LICENSE IS GRANTED FOR YOUR ORDER AT NO CHARGE

This type of permission/license, instead of the standard Terms & Conditions, is sent to you because no fee is being charged for your order. Please note the following:

Permission is granted for your request in both print and electronic formats, and translations. If ﬁgures and/or tables were requested, they may be adapted or used in part. Please print this page for your records and send a copy of it to your publisher/graduate school. Appropriate credit for the requested material should be given as follows: "Reprinted (adapted) with permission from (COMPLETE REFERENCE CITATION). Copyright (YEAR) American Chemical Society." Insert appropriate information in place of the capitalized words. One-time permission is granted only for the use speciﬁed in your request. No additional uses are granted (such as derivative works or other editions). For any other uses, please submit a new request.

https://s100.copyright.com/AppDispatchServlet 1/1

163