surface-enhanced raman spectroscopy: using nanoparticles ... · u ncorrect ed 1 proof...

UNCORREC

TEDPR

OOF

1 Surface-Enhanced Raman Spectroscopy:2 Using Nanoparticles to Detect Trace3 Amounts of Colorants in Works of Art

4 Federica Pozzi, Stephanie Zaleski, Francesca Casadio, Marco Leona,5 John R. Lombardi and Richard P. Van Duyne

6 Abstract In recent years, powerful physical processes occurring in the vicinity of7 nanoscale metal surfaces have been exploited in the art world for the detection of8 trace amounts of colorants with surface-enhanced Raman spectroscopy (SERS).9 With this technique, naturally occurring and man-made organic molecules used as

10 dyes and pigments in objects from antiquity to the present day are being detected11 with high molecular specificity and unprecedented sensitivity. This chapter reviews12 the broad spectrum of SERS analytical methodologies and instrumental improve-13 ments that have been developed over the years in the field of cultural heritage14 science, and discusses significant case studies within different types of works of art15 and archaeological artifacts.16

F. Pozzi (&)Department of Conservation, Solomon R. Guggenheim Museum, 1071 Fifth Ave,New York, NY 10128, USAe-mail: [email protected]

S. Zaleski � R.P. Van DuyneDepartment of Chemistry, Northwestern University, 2145 Sheridan Rd., Evanston,IL 60208, USAe-mail: [email protected]

R.P. Van Duynee-mail: [email protected]

F. CasadioDepartment of Conservation, Art Institute of Chicago, 111 South Michigan Ave,Chicago 60603, USAe-mail: [email protected]

M. LeonaDepartment of Scientific Research, Metropolitan Museum of Art, 1000 Fifth Avenue,New York, NY 10028, USAe-mail: [email protected]

J.R. LombardiDepartment of Chemistry, City College of New York, 160 Convent Ave,New York, NY 10031, USAe-mail: [email protected]

Layout: T1 Standard Unicode Book ID: 420585_1_En Book ISBN: 978-94-6239-197-0

Chapter No.: 6 Date: 19-4-2016 Time: 1:18 pm Page: 161/204

© Atlantis Press and the author(s) 2016P. Dillmann et al. (eds.), Nanoscience and Cultural Heritage,DOI 10.2991/978-94-6239-198-7_6

161

UNCORREC

TEDPR

OOF

17 1 Introduction

18 Since the dawn of time, color has been used by humanity to beautify the world and19 convey thoughts and feelings through various means of artistic expression.20 Among the colored materials employed since antiquity, pigments and colorants21 have played a key role in the dyeing of fabrics and creation of artworks, to22 embellish individuals, decorate objects of everyday use, and as a sign of hierar-23 chical status. The term “pigment” is used to describe insoluble materials, organic or24 inorganic, which are ground to fine powder and suspended in a binder to form paint.25 On the other hand, the terms “dye” and “colorant” generally refer to organic sub-26 stances that are soluble in the medium and show affinity for a substrate either27 inherently or through the action of additives. In textile dyeing, colorants may be28 applied to fiber substrates directly (direct dyes), after being solubilized with a29 reducing agent (vat dyes), or with the aid of bridging metal ions called mordants30 (mordant dyes). In painting, dyestuffs were typically precipitated onto an inert,31 insoluble, inorganic substrate to produce lake pigments with significantly improved32 lightfastness properties compared to the primary fugitive dyes.33 Prior to the introduction of synthetic colorants in the second half of the 19th34 century, dyes were derived from a variety of natural sources ranging from roots,35 berries, bark and leaves to different types of organisms such as lichens, insects and36 shellfish (Mills and White 1987; Hofenk de Graaff et al. 2004; Cardon 2007). Most37 natural red dyes owe their color to anthraquinone chromophores and can be38 extracted from insects, as in the case of cochineal, kermes and lac dye, or plants,39 among which madder and brazilwood are numbered. Several blue and purple40 colorants, including indigo and Tyrian purple, consist of a mixture of41 indigotin-related substances obtained from plants or sea snails, while other dyes of42 similar shades, like orchil, are derived from different sources, such as lichens. As far43 as yellow colorants are concerned, flavonoids—occurring as glycosides in a great44 variety of botanical species—are their main chromophores, although a number of45 other molecules such as carotenoids, curcuminoids, naphthoquinones and gal-46 lotannins may be found as well. Besides reds, blues and yellows, other hues could47 be produced by using two or more dyes in combination: green shades, for instance,48 were traditionally obtained as mixtures of a blue and a yellow colorant. For this49 reason, historic tapestries that have been extensively exposed to light often show50 blue foliage as a result of the fading of the more fugitive yellow component of the51 mixture.52 The fortuitous discovery of mauveine, the first synthetic dye, by William Perkin53 in 1856 prompted extensive chemical experimentation that led both to the synthesis54 of molecules previously derived from natural sources, such as alizarin and indig-55 otin, and to the introduction of several improved artificial colors that have no56 counterparts in nature, thus initiating a rapid decline in the dominance of natural57 dyes in world markets. In addition to their use in textiles and artworks, those58 synthetic colorants that could be tested as safe have found wide application as

162 F. Pozzi et al.



UNCORREC

TEDPR

OOF

59 printing and pen inks, paint components, dyes for food, plastic and rubber, and60 cosmetic constituents (Mills and White 1987; Lewis 1988).61 Scientific analysis of objects of artistic, historic and archaeological significance62 is key to reconstructing their story and elucidating the circumstances in which they63 have been created and, in some cases, have travelled through the centuries prior to64 ending up in the inquiring hands of conservators and scientists for in-depth tech-65 nical examination. In particular, the identification of natural and synthetic colorants66 may help to clarify attribution, provenance and dating, technology of fabrication,67 trade routes and commercial transactions that may have allowed the usage of certain68 dyes far away from their native locations. Moreover, investigating the origin, nature69 and chemical behavior of the colored materials employed in the production of70 artworks and historical artifacts may shed new light on their original color and71 appearance, thus offering insight into the artist’s actual intentions and choices, the72 techniques used, and the dates ante quem and post quem the object was created,73 possibly even leading to the uncovering of falsifications and forgeries. Most74 importantly, a clear understanding of the materials used is essential to enable the75 object’s long-term preservation and to inform conservation treatments, as well as76 museum display and lending policies.77 The identification of colorants from artworks and objects of archaeological,78 anthropological or historical value poses a whole set of analytical challenges.79 Recent advances in macro-X-ray fluorescence (macro-XRF) mapping (Dik et al.80 2008; Alfeld et al. 2013a, b) and infrared (IR), ultraviolet/visible (UV/vis), and81 hyperspectral imaging (Delaney et al. 2010, 2014; Ricciardi et al. 2012; Dooley82 et al. 2013; Rosi et al. 2013) have opened up exciting opportunities to track the83 distribution of pigments over the entire surface of a painted object. However, X-ray84 techniques are most successful at mapping inorganic pigments and are not able to85 map organic compounds, with the notable exception of eosin (tetrabromofluores-86 cein, also known as geranium lake, a pigment extensively used by Vincent van87 Gogh). In hyperspectral imaging, the signal from dyes is invariably compounded88 with the spectral contributions of the binding medium and other paint components,89 so the identification of the organic colorant is possible only in rare cases. At the90 micro-scale, although recent research has been aiming at detecting and identifying91 dyes non-invasively in situ, to date the greatest chances of success still rely on the92 removal of one or more samples from the object under study. To complicate the93 analysis even further, the extraordinary tinting strength of many organic colorants94 caused them to be used in minute amounts in both historical textiles and works of95 art, where they are typically found in complex chemical environments, bound to96 cloth fibers by means of metal ions or embedded in paint layers in the form of lake97 pigments. As a result, the application of intensive sample pretreatment procedures98 often becomes a necessary step to isolate the colorants from their matrixes and99 achieve selective molecular identification. An additional obstacle is posed by the

100 lack of permanence and susceptibility to deterioration of several colored organic101 materials that, upon exposure to air and light, undergo a number of chemical102 degradation processes leading to the formation of colorless species with a different103 molecular structure compared to the primary dye (Grosjean et al. 1988; Saunders

Surface-Enhanced Raman Spectroscopy: Using Nanoparticles … 163



UNCORREC

TEDPR

OOF

104 and Kirby 1994; Ferreira et al. 2001; Yoshizumi and Crews 2003; Ahn and105 Obendorf 2004, 2007; Cooksey and Sinclair 2005; Colombini et al. 2007; Clementi106 et al. 2007; Koperska et al. 2011; Manhita et al. 2011, 2013; Degano et al. 2011;107 Ramesova et al. 2012). Therefore, the amount of colorant available for analysis is108 often infinitesimal, challenging the limits of detection of many analytical109 techniques.110 The analysis of organic colorants, made possible by the introduction of chro-111 matographic and spectroscopic techniques, greatly benefited, in the past few dec-112 ades, from both the improvement of pre-existing analytical procedures and the113 development of new cutting-edge instrumental methods. Initial attempts at detecting114 and identifying dyes in historical artifacts involved the use of absorbance UV/vis115 spectrophotometry (Taylor 1983; Wouters 1985) and thin-layer chromatography116 (TLC) (Masschelein-Kleiner and Heylen 1968; Schweppe 1993). However, in both117 cases, a sizable amount of colorant needs to be removed from the sample and118 brought into solution for analysis, thus requiring that the sample itself be of sig-119 nificant dimensions or deeply colored. Alternative non-invasive approaches by120 means of reflectance UV/vis spectrophotometry (Bacci et al. 1991; Montagner et al.121 2011; Gulmini et al. 2013) and fluorimetry (Claro et al. 2008, 2010; Degano et al.122 2009; Melo and Claro 2010; Romani et al. 2010; Clementi et al. 2014; Amat et al.123 2015) have been also evaluated, occasionally leading to the successful characteri-124 zation of pigments and colorants from medieval illuminations, paint cross sections,125 ancient textiles and wall paintings. Yet, electronic methods typically provide broad,126 featureless spectra that often appear to be nearly identical for distinct compounds of127 similar molecular classes and, in some cases, even for totally unrelated chemical128 groups. This, along with the ubiquitous additional contribution of binding media,129 fillers, additives, and all other sample components to the spectra, makes the130 interpretation extremely challenging. High-performance liquid chromatography131 (HPLC) remains the technique of choice for the characterization of organic col-132 orants, allowing researchers to separate and identify dyes even in complex mixtures133 (Wouters 1985; Wouters and Verhecken 1989; Koren 1994; Halpine 1996; van134 Bommel et al. 2007). However, in this case, too, the amount of sample required for135 analysis may sometimes be prohibitive when dealing with museum objects for136 which permission to sample cannot be granted due to value, condition, or policy.137 While enabling the non-destructive characterization of colorants in minute samples138 or whole objects, vibrational methods such as Fourier-transform infrared (FTIR)139 (Gillard et al. 1994) and Raman/FT-Raman (Bell et al. 1997; de Oliveira et al. 2002;140 Edwards et al. 2003; Scherrer et al. 2009) spectroscopies were proven more suitable141 for the analysis of inorganic pigments and synthetic dyes. Raman spectroscopy, in142 particular, suffers from inherently weak signals and strong molecular fluorescence143 from natural dyestuffs that often obscures the Raman scattering signal even when144 using near-infrared (NIR) and IR excitation wavelengths to avoid fluorescence,145 such as 785 nm and above. Furthermore, because only sub-nanogram levels of dyes146 are needed to achieve intense coloration, Raman spectroscopy is often not sensitive147 enough to probe these materials, especially when they are embedded in complex148 matrixes such as textile fibers or artist’s paints.

164 F. Pozzi et al.



UNCORREC

TEDPR

OOF

149 Among the less sample-intensive techniques, surface-enhanced Raman spec-150 troscopy (SERS) has allowed researchers to overcome many of the drawbacks151 described above, enabling successful detection and identification of colorants from152 artworks and historical artifacts with high molecular specificity, unparalleled sen-153 sitivity, and the lowest known limits of detection (Chen et al. 2007; Wustholz et al.154 2009; Casadio et al. 2010). In SERS, organic molecules are adsorbed on noble155 metal nanostructures and can thus be probed due to the exceptional Raman scat-156 tering enhancement and fluorescence quenching provided by the substrate via157 chemical and electromagnetic effects (Fig. 1) (Lombardi and Birke 2009, 2012). It158 has been demonstrated that the SERS signal of ensemble-averaged molecules159 exhibits enhancements up to eight orders of magnitude over the normal Raman160 cross section. Since the initial discovery of SERS (Fleischmann et al. 1974) and the161 explanation of the observed phenomenon in the 1970s (Albrecht and Creighton162 1977; Jeanmaire and Vanduyne 1977), the field has grown enormously and has163 been enriched by a wealth of experimental studies. In particular, in the past ten164 years, the potential of SERS for dyestuff identification in various art and archae-165 ological applications has been increasingly harnessed, with a significant growth in166 the number of publications in this area.167 Initially, the greatest effort in the field of SERS for art and archaeology aimed at168 the characterization of reference materials, with special attention being paid to169 natural red colorants: alizarin, purpurin, carminic and laccaic acids, and related170 anthraquinones have been the subject of several studies in the scientific literature171 (Shadi et al. 2004; Cañamares et al. 2004, 2006a; Chen et al. 2006; Cañamares and172 Leona 2007; Baran et al. 2009; Rambaldi et al. 2015). Articles about the detection173 of flavonoids (Jurasekova et al. 2006, 2008, 2012; Wang et al. 2007; Teslova et al.174 2007; Cañamares et al. 2009; Corredor et al. 2009; Mayhew et al. 2013), indigoids175 (Bruni et al. 2010; Oakley et al. 2012), and dyes belonging to other molecular176 classes (Leona and Lombardi 2007; Cañamares et al. 2008a, b, 2010, 2014;

Fig. 1 Schematic of the surface-enhanced Raman scattering process. The molecule of interest,alizarin, is adsorbed onto an Ag nanoparticle substrate. The substrate is then irradiated with laserlight (green), and the enhanced Raman scattering (red) is detected




UNCORREC

TEDPR

OOF

177 Geiman et al. 2009; Chang et al. 2009; Xie et al. 2012; Bruni et al. 2011a; Mayhew178 et al. 2013; Benedetti et al. 2014; Greeneltch et al. 2012; Doherty et al. 2014;179 Cañamares and Lombardi 2015; Zaffino et al. 2015) can also be found in the180 literature. In addition to providing valuable reference spectra for identification181 purposes, these works have explored aspects such as the influence of pH on the182 resulting spectral patterns, and the binding geometry and relative orientation of the183 molecules examined with respect to the metal surface. Computational methods such184 as density functional theory (DFT) have also been used to calculate theoretical185 spectra and assign the normal modes of vibration to the Raman bands, thus sup-186 porting the interpretation of SERS data.187 At the same time, extensive research has focused on the improvement and188 optimization of the technique as an analytical tool to be used for the detection and189 identification of dyes in samples of artistic and historical value that may be very190 complex, altered or degraded. In this context, identifying faded colorants in art-191 works, and investigating the nature and causes of the corresponding degradation192 processes is essential to implement measures to slow down or altogether prevent193 future fading and to shed new light on how certain artifacts might have looked like194 right after completion. SERS has offered insight into important examples of such195 processes, such as the fading of madder- and cochineal-based purples and reds in196 watercolors by the American painter Winslow Homer (1836–1910) (Brosseau et al.197 2011) and in masterpieces by the French Impressionist Pierre-Auguste Renoir198 (1841–1919) (Collins et al. 2014; Pozzi et al. 2014a) and post-Impressionist199 Vincent van Gogh (1853–1890) (Vellekoop et al. 2013). In all these cases, the200 detection and identification of faded dyes with SERS allowed researchers to confirm201 that the current colorless or severely discolored appearance of the artworks is the202 result of a fading process, and to put forward hypotheses on how the masterpieces203 might have looked like right after the artists painted them.204 Recent advancements in the SERS technique that will be discussed in this205 chapter include the comparison of various metal substrates, such as silver colloids206 and silver films over nanospheres (AgFONs), and evaluation of their performances;207 the development of non-invasive approaches involving the use of SERS-active208 removable substrates; the adaptation of pre-existing methodologies of sample209 treatment and analysis, such as laser ablation and tip-enhanced Raman spectroscopy210 (TERS), to the specific needs of the cultural heritage field; and the combination of211 SERS with separation techniques, such as TLC and microfluidics, to resolve dye212 mixtures.213 Complemented by case studies drawn from important collections in the United214 States, this chapter will demonstrate the tremendous advances and high applica-215 bility of SERS, as well as outline areas of future development for a technique that is216 now solidly established as a powerful means of colorants’ investigation in the field217 of art and archaeology.

166 F. Pozzi et al.



UNCORREC

TEDPR

OOF

218 2 SERS Substrates and Analytical Methodologies

219 Since the discovery of the surface-enhanced Raman effect in the 1970s, metal220 substrates of different types and forms have been developed for various SERS221 applications, beginning with roughened silver electrodes, then moving onto col-222 loidal nanoparticles and solid surfaces with tailored thickness and nanostructured223 features. Accurate choice of the metal substrate is of utmost importance to ensure224 adequate measurement efficiency and high reproducibility of the spectra collected.225 While several metals including gold, copper, platinum, palladium and aluminum226 have been tested over the years, the best performances are expected from silver227 owing to its dielectric function and localized surface plasmon resonance (LSPR)228 spanning from the UV to the NIR/IR regions, i.e. the range of wavelengths typically229 used in SERS experiments for analyte excitation. Although incessant exploration in230 the field of high enhancing metal surfaces has provided practitioners with a con-231 tinuously growing number of choices, substrates used in cultural heritage applied232 research must meet additional requirements of quick preparation, relatively low cost233 and straightforward use, as few museum laboratories have the extensive sample234 preparation capabilities of university-based facilities. Additionally, researchers are235 also striving to develop SERS substrates that could be used to identify colorants236 in situ without the need of removing even the smallest of samples: this would be237 crucial, for instance, in the case of works on paper or modern and contemporary art,238 from which sampling is normally more problematic.239 The substrates and methodological approaches that have been used over the240 years for SERS analysis of art objects and archaeological artifacts are described in241 the following.

242 2.1 Colloidal Nanoparticles

243 Colloidal silver nanoparticles are by far the most widely used SERS substrate for art244 applications. The most popular method to synthesize silver colloids is by chemical245 reduction of a silver salt, typically silver nitrate, with trisodium citrate at boiling246 temperature (Lee and Meisel 1982). The Lee and Meisel method usually generates247 silver nanoparticles, mostly nanospheres and nanorods, in the 3–80 nm diameter248 range with a fairly broad visible absorption near 430 nm. Particle size, shape, and249 resulting plasmonic properties can be fine-tuned by suitable choice of the metal,250 reducing agent and stabilizer, chemicals concentration, addition rate, and temper-251 ature. In addition to testing alternative reducing agents, such as hydroxylamine252 (Leopold and Lendl 2003) and borohydride (Creighton et al. 1979), research groups253 all over the world have evaluated how different aggregants—including potassium254 nitrate (Cañamares et al. 2004), poly-L-lysine (Shadi et al. 2004), sodium chloride255 (Cañamares et al. 2008b) and perchlorate (Bruni et al. 2011a)—may promote the256 association of nanoparticles in clusters of various dimensions, where localized




UNCORREC

TEDPR

OOF

257 regions of intense electromagnetic fields, the so-called hot spots, are responsible for258 a giant enhancement of the Raman signal (Ringe et al. 2013). An alternative to259 Lee-Meisel colloids involves a synthetic protocol based on the glucose-assisted260 reduction of silver sulfate in the presence of sodium citrate as a capping agent261 (Leona 2009). In this case, the use of a microwave oven equipped with pressure and262 temperature control has afforded silver colloids with a significantly narrower par-263 ticle size range (3–10 nm) and absorption band, leading to increased stability of the264 substrate itself and more reproducible SERS performances.265 The main drawback of citrate-capped silver nanoparticles lies in the occasional266 occurrence of spurious bands in SERS spectra that are due to competitive267 adsorption of citrate ions and related oxidation products onto the metal surface268 (Leona et al. 2006; Brosseau et al. 2009a; Bruni et al. 2010). This issue may be269 circumvented by adopting alternative techniques for the production of colloids,270 such as laser ablation and photoreduction. More popular than the first, based on the271 extraction of metallic nanoparticles from a silver plate induced by high intensity272 laser pulses (Cañamares et al. 2008c), is the fabrication of Ag substrates by laser273 photoreduction of a silver nitrate solution (Cañamares et al. 2007; Jurasekova et al.274 2008, 2010; Retko et al. 2014). In both approaches, the absence of reducing or275 capping agents prevents disruptive interference (Cañamares et al. 2008c); however,276 the prolonged in situ irradiation required may raise concerns when these methods277 are applied to the analysis of samples from actual artworks due to the risk of278 thermal degradation and formation of carbon-rich phases.279 In order to increase the applicability of SERS-active colloids to the analysis of280 complex samples from works of art, a useful pretreatment step has also been281 introduced involving a gas-solid hydrolysis performed by exposing microscopic282 samples to hydrofluoric acid (HF) vapor for 5 min in a closed polyethylene283 microchamber (Fig. 2). Specifically developed for mordant dyes and lake pigments,284 such pretreatment aims at hydrolyzing the chemical bond that the organic colorant285 forms with the fabric or inorganic portion of the lake through bridging metal ions.286 Compared to previously common extraction approaches based on the use of heated287 acids or alkali (Wouters 1985; Tiedemann and Yang 1995), the HF hydrolysis288 exploits milder experimental conditions and may be considered extractionless in289 that it does not involve physical detachment of the target dye from the sample under290 study. Simply, a small amount of colorant is released, in situ, from the surface of the291 dyed object becoming available for coupling to the silver nanoparticles. This292 methodology has proven critical to attain highly reproducible, conclusive dye293 identification in many instances, and has delivered superior results for the analysis294 of paints and glaze layers (Pozzi et al. 2012a, 2014a; Zaffino et al. 2014).295 An alternative approach involves the use of colloidal pastes produced by cen-296 trifugation and concentration of nanoparticle suspensions. These pastes have been297 successfully used, in some cases, for the direct, extractionless, non-hydrolysis298 detection and characterization of organic colorants from textile fibers, pastels and299 watercolors (Brosseau et al. 2009a, b, 2011; Idone et al. 2013).

168 F. Pozzi et al.



UNCORREC

TEDPR

OOF

300 Although most of the studies mentioned so far have been carried out on loose301 samples removed from works of art, a limited number of proof-of-concept works302 recently published in the literature have also reported initial attempts at applying303 colloids directly on cross sections from paintings, with the goal of obtaining304 spatially-resolved data (Idone et al. 2014; Retko et al. 2014; Frano et al. 2014).

Fig. 2 SERS analysis of a red lake oil paint sample with Ag colloids upon HF hydrolysistreatment. Step (1) a microscopic paint specimen is removed from a painting by means of a scalpelblade and placed on a polyethylene holder; step (2) the sample is exposed to HF vapor at roomtemperature for 5 min in a closed polyethylene microchamber; step (3) 0.8 µL of Ag colloids and0.1 µL of aggregant, typically 0.5 M KNO3, are deposited onto the sample and SERS analysis isperformed before evaporation of the droplet




UNCORREC

TEDPR

OOF

305 2.2 Solid-State SERS Substrates: Silver Films Over306 Nanospheres (AgFONs)

307 The second most common SERS substrate aside from colloidal nanoparticles308 consists of lithographically fabricated metal surfaces. A variety of lithographic309 fabrication methods exist, such as electron beam lithography (EBL), focused ion310 beam (FIB) milling, and nanosphere lithography (NSL) (Fan et al. 2011). However,311 most of these techniques require costly equipment, which can be difficult to access312 in a museum setting, and have therefore seen limited use in the field. NSL is a313 cost-efficient and facile method of fabricating SERS substrates. NSL masks are314 obtained by dropcasting commercial polystyrene or silica microspheres onto a315 cleaned glass coverslip and allowed to dry in a hexagonal close packed array.316 A 200-nm layer of metal, typically Ag or Au, is then deposited through thermal317 vapor deposition on the microsphere mask to create the SERS substrate, commonly318 known as a film over nanospheres (FON) (Fig. 3). FONs have been shown to have319 high uniformity over a large area and their LSPR, readily tunable by changing the320 microsphere size, can be easily matched with the laser excitation wavelength321 (Sharma et al. 2013). When optimized, FONs have enhancement factors (EFs) on322 the order of 106–108.323 The ease of fabrication, tunability, and high EFs of FONs make them ideal324 substrates for the identification of artists’ materials with SERS. The first use of325 AgFONs in cultural heritage science was to identify and characterize the red326 dyestuffs alizarin, purpurin, lac, cochineal and their mixtures (Whitney et al. 2006).327 In this work, AgFONs were optimized in three excitation wavelength regimes328 relative to the electronic resonance of the dyes: pre-resonant (632.8 nm), resonant

Fig. 3 Representative scanning electron microscopy (SEM) image of a FON. A 200-nm Ag layeris deposited on the surface of hexagonal close packed silica microspheres. The high enhancementfrom the FON originates from the formation of nanopillar-like structures on the deposited metalsurface, which leads a high number of SERS hot spots on the Ag surface (SEM image courtesy ofDr. Anne-Isabelle Henry, Van Duyne group, Northwestern University)

170 F. Pozzi et al.



UNCORREC

TEDPR

OOF

329 (532 nm) and non-resonant (785 nm). It was found that the SERS signal was two330 orders of magnitude higher in the resonant regime compared to the non-resonant331 wavelength range, highlighting the importance of the resonance contribution to the332 SERS enhancement and the usefulness of a tunable SERS substrate. Additionally,333 the authors successfully resolved various binary dye mixtures in the pre-resonant334 wavelength regime. Whitney et al. (2007) demonstrated that exciting on electronic335 resonance with the dye in question can yield greater SERS signal, but it can also336 lead to photodegradation and possible fluorescence from binding media. Recent337 work examined the EFs of AgFONs optimized for NIR and IR wavelengths and Ag338 colloidal nanoparticles to identify eosin Y, an early modern synthetic dye, which339 was extensively used by Vincent van Gogh and sold under the name of geranium340 lake (Greeneltch et al. 2012). The use of NIR and IR laser excitation is beneficial341 for analyzing cultural heritage samples as compared to visible excitation because it342 minimizes photodegradation and interfering fluorescence. The authors observed343 that AgFONs EF using 1064 nm excitation was 20 times greater than that using344 785 nm excitation. In addition, when compared to Lee and Meisel Ag colloidal345 nanoparticles, FONs were found to provide EFs greater by two orders of magnitude346 and more uniform signal.347 Notwithstanding their ease of fabrication, high enhancement, uniformity, and348 tunability, AgFONs have been only occasionally used for practical applications of349 SERS to the identification of artists’ colorants. One of the main hurdles in their350 widespread use may be that the dyestuff needs to be brought into solution by351 extraction from the sample, while colloids and colloidal pastes can be applied352 directly onto a solid sample with or without HF pretreatment.

353 3 Innovative SERS Approaches

354 A significant portion of the recent research in the field of SERS for cultural heritage355 has focused on the development of new methods that can be deployed in situ or356 approaches that are endowed with enhanced spatial resolution. The most promising357 methodological developments to date are reported in the following.

358 3.1 Laser Ablation (LA)—SERS

359 A substantial drawback of most approaches to SERS for analyzing cultural heritage360 samples is their lack of spatial resolution. Samples smaller than 50–100 μm are361 difficult to extract in a controlled and precisely localized way, and, in the case in362 which drops of colloid are deposited directly on the sample by hand (Idone et al.363 2013) the spatial resolution is ultimately limited by the size of the drop, i.e. about




UNCORREC

TEDPR

OOF

364 0.5–1 mm in diameter with the most common pipettes. Even when using inkjet365 devices to deliver the colloidal nanoparticles (Benedetti et al. 2014) the spatial366 resolution does not go below 50–80 μm. Londero et al. improved the applicability367 of SERS analysis towards complex organic samples, while simultaneously reaching368 high spatial resolution, by combining laser ablation microsampling with SERS369 detection on a silver nanoisland film in a vacuum chamber (Londero et al. 2013).370 They successfully demonstrated spatial resolution of 5 µm and sensitivity down to371 120 attomol on a test sample created by vacuum depositing a film of copper372 phthalocyanine on a quartz disc. In their setup, a visible laser pulse from an optical373 parametric oscillator (OPO) is focused on a target inside a vacuum chamber with374 sufficient intensity to ionize the sample and produce a plasma. The plasma heats the375 ionized region via an inverse Bremsstrahlung process and causes the material to376 explode outward as a plume of vapor. The vapor deposits onto a SERS active377 substrate, the silver nanoisland-coated bottom surface of the chamber window,378 where it can be interrogated by a read laser (Fig. 4).379 Laser ablation has another advantage over the use of colloids besides spatial380 resolution: as it relies on vapor deposition rather than diffusion through water for381 the analyte to adsorb on the plasmonic substrate, it can be used on water insoluble382 pigments with excellent results. In their work, Londero et al. (2013) showed this on383 the pigments quinacridone and quinacridone quinone, and their mixture Pigment384 Orange 48.385 An improved version of the ablation setup (Cesaratto et al. 2014) features UV386 ablation. By ablating in the UV, specifically at 355 nm, i.e. at a wavelength where387 virtually all organic materials are highly absorbing, vaporization is confined to388 approximately 1-µm depth. Even with extremely simple UV optics, Cesaratto and389 coworkers were able to show significant advantages over visible ablation. A cross390 section taken from the red area of a 16th-century decorated dish shows that two391 separate red paint layers exist on the plate (Fig. 5). The top layer is composed of392 sparse carbon particle embedded in a red glaze, while the bottom one is a lighter red393 glaze. The cross section was ablated in three different areas, with a minimum spot394 size about 13-μm wide (as determined by atomic force microscopy after the anal-395 ysis). The two different layers could be characterized separately and without con-396 tamination one from the other, and the red colorant in the very top thin layer could397 be easily distinguished from the carbon particles. This is particularly critical given398 graphite’s ability to displace colorants on the SERS substrate, which can result in it399 being the only detectable component. The bottom layer was identified as a400 madder-based lake, while the upper one gave a SERS spectrum compatible with an401 unidentified synthetic organic colorant. The results reinforced suspicions of a402 modern addition raised from previous visual inspection of the object. This kind of403 work would not have been possible with the conventional approaches to SERS404 because of the complexity of the layering structure, or indeed, with any alternative405 technique, given that samples suitable for HPLC or liquid chromatography/mass406 spectrometry (LC/MS) could not have been obtained from such thin layers.

172 F. Pozzi et al.



UNCORREC

TEDPR

OOF

407 3.2 Inkjet Nanoparticle Delivery SERS

408 An interesting alternative to traditional approaches that involves removing a sample409 and treating it with a colloid in a separate stage was presented by Benedetti et al.410 (2014). Using inkjet technology to reproducibly deliver microdroplets of silver411 nanoparticles either on the specimen to be analyzed by SERS or on a work of art,

(c)

(a) (b)

Fig. 4 Experimental setup for LA-SERS. a The sample is placed on a holder within a vacuumchamber and is elevated to 200–300 µm from the SERS active quartz window. The ablation andread lasers are imaged onto the sample and SERS film, respectively, by the microscope objective.The signal generated from the read laser is collected by the same objective and filtered from anybackscattered laser light by the longpass filter, after which a spectrum is recorded. b Illustration ofthe two ablation geometries that can be utilized: geometry A optimizes spatial resolution, whilegeometry B increases the ablated surface area for situations such as representative sampling ofgranular mixtures. c Steps for a typical measurement: the sample is ablated in one of the twopossible ablation geometries and then collected on the SERS-active film, where it is excited by theread laser




UNCORREC

TEDPR

OOF

412 the authors of this work were able to demonstrate quasi non-destructive analytical413 capabilities. Droplets of approximately 50–80 μm were deposited directly onto a414 Japanese woodblock print with a piezoelectric inkjet head mounted on the Raman415 microscope and aimed at the focal point of the microscope objective. The suitability416 of either thermal or inkjet printing heads to deliver silver colloids for SERS analysis417 was demonstrated as well (Leona and Tague 2010). Although the droplets of silver418 nanoparticles cannot be removed from the substrate when dry, they are of such419 small diameter that they are nearly invisible to the naked eye.

Fig. 5 a A 16th-century decorated dish and microscope images of a cross section from thedecoration (sample DD), with the ablated craters A, B and C. b SERS spectrum of the UV-ablatedcrater A in the main layer. Signals of reference madder lake were detected at 1283 and 1321 cm−1,highlighted in the graph with an asterisk. c SERS spectra of the UV-ablated craters B and C in thetop layer

174 F. Pozzi et al.



UNCORREC

TEDPR

OOF

420 3.3 Detachable SERS Substrates

421 While SERS analysis, unlike ordinary Raman, cannot be carried out in an entirely422 non-invasive way, several researchers have investigated quasi non-invasive ana-423 lytical approaches based on the use of gels. Gels have been used for years in424 conservation to remove surface contamination from works of art in a localized and425 selective way. A gel confines the action of a solvent or cleaning agent to the area of426 the object with which it is in contact, preventing the spread of any material—either427 the cleaning agents or the materials targeted for removal—through and across the428 artwork (Carretti et al. 2010; Baglioni et al. 2015). For SERS applications, a gel429 could substitute the colloid drop traditionally used as the medium for the target dye430 and the plasmonic nanoparticles to interact; alternatively, a gel loaded with431 appropriate reagents could be used to perform a mild extraction of the target analyte432 from its substrate, for subsequent analysis with a silver colloid deposited on the gel.433 Gel-based colloidal substrates for SERS analysis were first explored by Bell and434 Spence (2001) and by Farquharson and Maksymiuk (2003) without, however, a435 specific focus on quasi non-invasive sampling of works of art. The first application436 of gels to SERS analysis of artworks is due to Leona, who used cross-linked437 hydroxyacrylate gels for non-destructive dye extraction from paper and textiles438 (Leona 2008; Leona et al. 2011). The cross-linked gels were effectively used as439 solid-phase microextraction substrates, and SERS analysis was performed by440 depositing silver colloids on the gels after removal from the work of art. The441 method was tested on the Metropolitan Museum’s Unicorn Tapestries (1495–442 1505), enabling the identification of madder, and on a 19th-century Japanese443 woodblock print, where methyl violet was detected (Fig. 6).444 A different approach was followed by Doherty et al. (2011) using methyl cel-445 lulose as the gel medium and Lofrumento et al. (2013), Platania et al. (2015) using

Fig. 6 a Use of hydroxygels to extract minute amounts of ink from a ball point pen tracing onpaper. The region of extraction is barely distinguishable from the area from which no extractiontakes place. b Sekigahara Homare no Gaika (A poem about the battle of Sekigahara) by ToyoharuKunichika, 1892. Woodblock print on paper, triptych, each sheet originally oban size(27 × 39 cm), slightly trimmed. Private collection




UNCORREC

TEDPR

OOF

446 agar. Both groups incorporated the silver colloid directly in the gel substrate to be447 put in contact with the work of art. The agar gel approach was further refined by448 Platania et al. (2014) by adding ethylenediaminetetraacetic acid (EDTA)—a449 chelating agent—to the gel, thus improving the extraction of a variety of dyes from450 substrates such as oil paints, printed and dyed textiles. The agar gel seems to be one451 of the most promising approaches because, due to its shrinking upon drying, it452 causes the nanoparticles to come into close contact, inducing the occurrence of453 SERS active hot spots. Additionally, agar gels are increasingly used for conser-454 vation treatments and are, therefore, widely accepted in the field.

455 3.4 Tip-Enhanced Raman Spectroscopy (TERS)

456 Tip-enhanced Raman spectroscopy (TERS) is a promising non-destructive analog457 to SERS, which combines the nanometer spatial resolution of scanning probe458 microscopy (SPM) and the chemical sensitivity of SERS. A TERS experiment can459 be performed using one of two variants of SPM: scanning tunneling microscopy460 (STM) and atomic force microscopy (AFM). In an STM experiment, a voltage is461 applied to the scanning probe tip that comes in tunneling contact with the sample.462 An STM image is produced by measuring the changes in the tunneling current as a463 function of tip position across the sample. Despite its sub-nanometer scale spatial464 resolution, STM is sample-limited due to the requirement that the substrate of465 interest must be conductive and therefore cannot be used to analyze most materials466 from works of art. Alternatively, in an AFM experiment, the scanning probe tip is467 on a cantilever which makes contact with the surface of interest; changes in tip468 height due to local surface topography are measured to produce an AFM image.469 While the AFM tip does make contact with the sample, the tip can be readily470 removed from the sample surface, which is advantageous when analyzing sensitive471 cultural heritage samples. The surface generality of AFM is an advantage over STM472 and has been used to probe a wide range of surfaces such as DNA strands, amyloid473 fibrils, cell membranes and polymer films, as highlighted in a recent review474 (Schmid et al. 2013). Moreover, the surface generality of AFM makes it amenable475 to analyze surfaces relevant to cultural heritage such as cross sections, paint films,476 works on paper including watercolors, and textile fragments.477 The scanning probe tip is at the heart of a TERS experiment, as it is both the478 source of the nanometer scale spatial resolution as well as the plasmonic enhancing479 surface. A TERS-active scanning probe is fabricated by electrochemical etching of480 Ag or Au wire for STM, or by thermal deposition of an Ag or Au metal film onto a481 commercial Si cantilever for AFM.482 The methodology of a TERS experiment is illustrated in Fig. 7. First, the483 plasmonic tip is brought in close contact with the sample surface of interest. Laser484 light is then directed towards and focused onto the tip-sample junction. Low laser485 powers on the order of microwatts are often used in TERS experiments because of486 the tight focusing of the laser at the probe tip apex, which can be a beneficial feature

176 F. Pozzi et al.



UNCORREC

TEDPR

OOF

487 when analyzing artworks prone to thermal and photodegradation. The illumination488 of the tip induces an enhanced electromagnetic field around the probe tip apex, and489 only molecules directly beneath the tip experience Raman scattering enhancement.490 Then, the probe can either remain stationary in a single location while TERS spectra491 are acquired, or it can scan across the surface of interest. The intensity of a specific492 Raman band as a function of tip position on the surface can be used to produce a493 TER spectral map. Lastly, in order to ensure the TERS signal acquired is not due to494 contamination of the scanning tip or normal Raman or SERS of the substrate, a495 ‘withdrawn’ or ‘retracted’ spectrum is typically acquired with the plasmonic tip far496 from the surface of interest.497 The first proof-of-principle AFM-TERS study for the non-invasive identification498 of dyestuffs of significance for the art field was recently reported (Kurouski et al.499 2014). In this work, the authors analyzed a reference sample consisting of Kinwashi500 paper dyed with indigo, and were able to successfully acquire TERS spectra of the

Fig. 7 Schematic of an experimental setup for AFM-TERS. The sample of interest is placedunderneath the AFM scan head (inset). After the sample is put in contact with the AFM tip, thelaser light is brought up through a microscope objective and focused at the apex of the Ag-coatedAFM cantilever tip, and a TERS approached signal is acquired. The sample is then withdrawnfrom the tip and spectra are collected to ensure that no normal Raman or SERS signal is present orany tip contamination occurred




UNCORREC

TEDPR

OOF

501 blue colorant on the paper support. When the dyed paper spectra were compared to502 a reference of indigo on an Au film, they observed missing vibrational bands and503 band shifts in the dyed paper spectral patterns, likely indicating the occurring of504 complex dye-substrate interactions. It was also found that signature peaks of cel-505 lulose and glucose, two major constituents of paper, were present in the TERS506 spectra. In order to demonstrate the feasibility of TERS for analysis of artworks,507 TERS spectra were obtained from a historic 19th-century manuscript fragment with508 iron gall ink and compared to that of Kinwashi paper freshly dyed with iron gall509 ink. The authors found that some vibrational modes of gallic acid, a primary510 component of iron gall ink, were only observed in either Kinwashi paper or the511 manuscript, which was attributed to the complex chemical composition of the ink512 and paper and to the interaction of gallic acid with the paper substrate. This work513 demonstrates the feasibility and strong future potential for TERS to non-invasively514 probe artworks with the high sensitivity of SERS. At present, the main limitations515 of the technique are that it is highly specialized and mostly available in university516 research labs (although commercial instruments have started to appear on the517 market in the past 5 years); in addition, when used in transmission mode, samples518 need to be semi-transparent to be able to be analyzed and they have to be limited in519 size to what the sample chamber can accommodate. These drawbacks can be520 overcome by working in reflectance mode, but to date no applications to cultural521 heritage materials have been described in the literature.

522 4 Resolving Dye Mixtures

523 Although SERS has successfully reduced the sample size requirements with respect524 to HPLC, it is still limited by its inability to physically separate dye mixtures. An525 immediate consequence of this is that, in the case of dye mixtures, only the signals526 of the main component are typically detected due to discrepancies in resonance,527 solubility, affinity for the metal surface, SERS cross section, or a combination of528 two or more of these factors (Shadi et al. 2004). This represents a real limitation, as529 it is well known from documentary sources that, throughout history, dyes were530 often used in combination to achieve particular color shades both in painting and531 textile dyeing (Cardon 2007). Several studies describing applications of the SERS532 technique to the analysis of artworks and historical textiles have described instances533 in which, while the presence of multiple colorants had been assessed by HPLC,534 only one could be detected by SERS (Leona and Lombardi 2007; Pozzi et al.535 2012b). On the other hand, recent contributions to the scientific literature have536 offered proof of the simultaneous SERS detection of binary mixtures of dyes in537 multiple-component commercial pigments (Londero et al. 2013), fibers from his-538 torical textiles (Idone et al. 2013), and oil paint samples (Pozzi et al. 2014a). This539 has prompted researchers to pursue an extensive investigation of the capabilities of

178 F. Pozzi et al.



UNCORREC

TEDPR

OOF

540 SERS to concurrently detect and identify several dyes in combination and, at the541 same time, has encouraged alternative, more efficient ways to accomplish analysis542 of mixtures of organic colorants by SERS upon physical separation of the indi-543 vidual constituents.

544 4.1 Binary Mixtures in Solutions

545 While a detailed evaluation of the SERS technique’s ability to detect single dyes546 has been the subject of comprehensive studies in the past decade, to date very little547 work has focused specifically on the identification of individual components of dye548 mixtures. Thus far, spectra of binary combinations of reference alizarin, purpurin549 and lac dye have been presented within very preliminary studies (Whitney et al.550 2007; Van Elslande et al. 2008; Jurasekova et al. 2010), some of which have offered551 an initial assessment of suitable metal substrates and ideal experimental conditions,552 along with summary considerations on marker bands, adsorption geometries, and553 relative detection limits of the colorants examined. These early studies have been554 significantly expanded by a more systematic work recently carried out at the Art555 Institute of Chicago. This work aims to determine the capabilities of SERS in terms556 of distinguishing a wider number of dyes, including alizarin, purpurin, carminic557 acid, laccaic acid and brazilein, in series of historically relevant mixtures, both for558 reference solutions and in actual samples such as mock-up red lake oil paints. First,559 SERS spectra of binary mixtures of the commercial colorants in various relative560 proportions and concentration comparable to that observed in museum objects were561 recorded on two different SERS substrates, i.e. Lee-Meisel citrate-reduced silver562 colloids and AgFONs. This aimed to qualitatively establish relative detection limits563 for each dye when present in solution alongside another colorant, and to gain a564 deeper understanding of how different metal substrates and analytical methodolo-565 gies may affect dye identification. Second, binary mixtures of madder, cochineal566 and brazilwood were analyzed with SERS in red lake oil paint reconstructions567 prepared according to 19th-century historical recipes. For this set of samples,568 analysis was performed after hydrolysis with HF and only employed Lee-Meisel569 silver colloids as SERS-active substrate. Preliminary results from experiments on570 reference solutions have shown that, in many cases, the spectral contribution of the571 second dye in the mixture remains undetected unless it is present in high relative572 proportions (usually 1:1). A typical example, in this regard, is offered by mixtures573 of alizarin and carminic acid: for both colloids and FONs, the latter component574 becomes detectable in the 1:1 mixture, while the contribution of alizarin is pre-575 dominant even in the 1:10 solution, i.e. in cases where the relative amounts of576 carminic acid are 10 times as high as those of alizarin. In addition, remarkable577 differences were observed between the two substrates both in terms of contribution578 of each dye in the mixture to the SERS spectrum and, in some cases, for what579 concerns the position and relative intensities of bands. This is likely due to the580 experiments’ different incubation times, enhancing in the case of colloidal




UNCORREC

TEDPR

OOF

581 nanoparticles the preferential adsorption of one of the two dyes possibly due to the582 factors listed above as well as to differences in the surface chemistry of the metal583 substrates used. Overall, these new results underscore the need for more extensive584 experimentation in the area of hyphenated techniques that may enable the combined585 separation of colorants (e.g. via TLC or microfluidics) and their SERS detection in586 artworks.

587 4.2 Thin-Layer Chromatography (TLC)—SERS

588 Combining efficient separation with the fingerprinting ability typical of vibrational589 spectroscopies, the coupling of SERS with chromatographic techniques such as590 TLC or HPLC provides an obvious means toward effective discrimination of591 multiple dye components in complex mixtures. In particular, TLC-SERS, whose592 first application was reported by Henzel in 1977 (Henzel 1977), has been used for a593 series of preliminary proof-of-concept studies recently undertaken in the cultural594 heritage field thanks to its ability to significantly reduce the amount of material,595 sophisticated equipment and time needed for analysis. The first example of use of596 TLC-SERS for the characterization of several organic colorants in combination was597 demonstrated by Brosseau et al. (2009b). In this work, the authors were able to598 discriminate single anthraquinoid dyes in a binary mixture in both reference solu-599 tions and extracts from a dyed wool fiber. In the same year, a second work pub-600 lished by a different research group described the separation and identification of601 dye components in ballpoint pen inks by TLC-SERS, highlighting the great602 potential of this technique in the field of forensics (Geiman et al. 2009). After these603 initial studies, applications of TLC-SERS were then extended to the analysis of604 mixtures of the main β-carboline alkaloids from the seeds of the Syrian rue plant605 (Peganum harmala), i.e. harmalol, harmaline, harmane and harmine, which are606 relevant as historical dyes and as drugs (Pozzi et al. 2013a), and to the separation607 and identification of the structurally related components of mauve, the first syn-608 thetic organic dyestuff (Cañamares et al. 2014). All these works share a common609 analytical procedure, in which solution aliquots of the dye extract or commercial610 material of interest were deposited onto a silica gel TLC plate using a glass capillary611 and eluted in a glass developing chamber by means of an appropriate solvent612 mixture. Then, the eluted spots were visualized and marked under ultraviolet light.613 SERS analysis was subsequently performed directly on the silica gel plate, upon614 deposition of a few microliters of Ag colloid with or without aggregant, after drying615 of the droplet or while the spots were still wet. In these approaches, the sometimes616 inefficient dye separation provided by the silica gel was compensated for by its617 chemical and vibrational inactivity compared to other substrates. Initial attempts to618 make the entire TLC plate SERS-active by both pre- and post-treatment with silver619 colloids resulted either in poor separation of the dyes and difficult spot identifica-620 tion, or in an uneven distribution of the nanoparticles on the plate surface. New621 avenues for methodological improvement include the application of AgI and other

180 F. Pozzi et al.



UNCORREC

TEDPR

OOF

622 silver halides onto gold-coated slides to form a new generation of SERS-active TLC623 substrates that have shown great promise thanks to their tunable structure, as well as624 superionic, optical, and complexing properties (Sciutto et al. 2015).

625 4.3 Microfluidics—SERS

626 An alternative method to simultaneously separate, detect and identify multiple627 analytes is offered by the development of a microfluidic-SERS platform. In this628 approach, the channels of the microfluidic-SERS device are functionalized with629 alkanethiol self-assembled monolayers (SAMs), all which have different function-630 alities. The polarity of the tail group can be adjusted to attract different classes of631 dyes, analogous to adjusting the mobile phase solvent polarity in chromatography.632 A microfluidic-SERS approach is not only tailorable, but also requires a small633 sample; the micrometer scale of the device channels typically involve sample634 volumes on the order of 100 µL to flow solution through the entire device. This is635 beneficial when working with samples that are limited in size or volume, which is636 often an issue in cultural heritage science. Previous work has shown that637 microfluidic-SERS devices can detect analytes with a concentration as low as 3 nM638 (Kim et al. 2014).639 A microfluidic-SERS device developed by the Van Duyne group at640 Northwestern University consists of a SERS-active substrate, i.e. metal FONs, and641 a polydimethylsiloxane (PDMS) patterned channel mold. Device fabrication is642 illustrated in Fig. 8. First, the microfluidic devices are produced by lithographically643 patterning the channel design onto the AuFON SERS substrate with UV light cured644 photoresist. Next, the photoresist is removed, leaving behind the channel pattern645 directly on the SERS substrate. The patterned SERS-active substrate and the cured646 PDMS mold are exposed to oxygen plasma and fused together. Holes are punched647 in the PDMS to create the inlets and outlets for the channels, where solutions can be648 injected in a controlled manner via an infusion pump. Each individual channel is649 then functionalized with the alkanethiol of interest, and the channels are rinsed to650 remove excess or unbound reagent. Lastly, the target analyte is flown through the651 entire device and SERS measurements are taken on an inverted microscope setup652 by focusing the laser light at the plasmonic substrate surface.653 In proof-of-principle experiments, the Van Duyne group has successfully fab-654 ricated microfluidic-SERS devices and explored the separation capability of alka-655 nethiol SAMs on a SERS substrate. First, SERS spectra of a 100 µM alizarin656 solution in both a non-functionalized device and a SAM-functionalized device657 channel were acquired, and the characteristic SERS spectrum of the dye was suc-658 cessfully detected in both devices. Additionally, it was found that the PDMS has a659 unique Raman spectrum, which can be subtracted out from the SERS spectra to660 avoid interference. Non-microfluidic device FONs were then prepared in order to661 explore the capability of different terminated alkanethiol SAMs to separate different662 classes of dyestuffs. Two FONs were functionalized with a –CH3 and –OH




UNCORREC

TEDPR

OOF

663 terminated alkanethiol SAMs, HS–(CH2)5–CH3 and HS–(CH2)6–OH. After thor-664 ough rinsing, the FONs were incubated in a 1:1 ratio mixture of dyes alizarin, an665 anthraquinone dye, and brazilein, a neoflavonoid dye. As shown in Fig. 8, the –CH3

666 terminated SAM only shows SERS of alizarin, but the –OH SAM displays SERS667 signal characteristic of the mixture of alizarin and brazilein. The two proof-of-668 principle experiments demonstrate that alkanethiol SAM-functionalized669 microfluidic-SERS devices can be designed to selectively separate and detect dif-670 ferent classes of dyes in mixtures.

671 5 Case Studies

672 Almost three decades after the first reported use of SERS for the analysis of art673 samples using a roughened silver electrode (Guineau and Guichard 1987), and674 particularly after the last decade of renewed and sustained interest in the technique,675 SERS has become an established and increasingly used method for the identifica-676 tion of organic colorants in art and archaeology. Following demonstration of the677 reproducibility of the results obtained on reference dyes, colored fibers, and lake678 pigments (Pozzi et al. 2013b), and upon publication of significant collections of679 reference spectra of anthraquinoids, flavonoids, carotenoids, and other classes of680 colorants widely encountered when analyzing art (Leona et al. 2006; Cañamares681 et al. 2006a, 2008b; Whitney et al. 2006; Cañamares and Leona 2007; Bruni et al.

Fig. 8 a Schematic of FON fabrication (left) and microfluidic-SERS device fabrication (right).b SERS on various alkanethiol SAM-functionalized FONs. Only alizarin was detected with HS–(CH2)5–CH3 SAM (red), while both brazilein and alizarin were identified with HS–(CH2)6–OHSAM (black), as compared to SERS of alizarin (purple) and brazilein (blue) on anon-functionalized FON. Experimental parameters for SERS measurements: excitation wave-length = 633 nm; acquisition time = 4 s; power at the sample ≈ 360 µW

182 F. Pozzi et al.



UNCORREC

TEDPR

OOF

682 2011a), SERS has become a powerful technique in the analyst’s toolbox to advance683 the knowledge and support the preservation of a wide range of art objects, as684 demonstrated in the following sections.

685 5.1 Textiles

686 Dyers and craftsmen have harnessed the coloring power of various insects and687 plants to dye textiles since the beginning of human history. Hence, it is natural that688 the first applications of SERS as a less sample-intensive alternative to HPLC for the689 analysis of colorants in art have been directed to the analysis of textiles. As early as690 1987, Guineau and Guichard identified madder on a woolen thread after extraction691 with hydrochloric acid (HCl) and acetone using a roughened silver electrode692 (Guineau and Guichard 1987). Borrowing from the sample preparation approaches693 developed for HPLC, the most commonly followed analytical procedures to analyze694 historical textiles initially involved pretreatment of the sampled threads with a hot695 solution of HCl and methanol (MeOH) in water. Such pretreatment aimed to release696 the dyestuff from the complex with the metal cation that acts as a linker between the697 colorant and the fiber. This type of hydrolysis, however, is rather aggressive and698 may affect the proteinaceous or glycosidic structure of the fiber itself, bringing in699 solution much more than just the free dye. While not much of a concern for700 separation techniques such as HPLC, this is an impediment to efficient SERS701 analysis of colorants due to interfering components being included in the analysis.702 Thus, very early on researchers started to experiment with milder extraction703 methods (Leona et al. 2006) for textile fibers, such as treatment with 1:1704 dimethylformamide and water with 1 % EDTA, as originally proposed by705 Tiedemann and Yang (1995). An extraction procedure based on immersion of the706 fiber in 4 M HF, coupled with SERS on NaClO4-aggregated silver colloids,707 allowed the identification of a yellow dye derived from the dried fruit rinds of708 pomegranate in archaeological woolen threads from the Libyan Sahara, a result709 confirmed by parallel analysis with gas chromatography/mass spectrometry710 (GC/MS) and HPLC (Bruni et al. 2011b). Moreover, Bruni and coworkers iden-711 tified madder on a red woolen thread from the same burial site and Tyrian purple on712 an ancient bone from a 4th-century tomb in the basilica of S. Ambrose in Milan,713 where the colorant had likely transferred from the textile used to wrap the body.714 While it was not possible, in the first case, to identify madder upon HCl:MeOH715 extraction due to interference of the peptide residues from the hydrolyzed wool716 protein, a milder method based on immersion in HF led to the successful identi-717 fication of the red anthraquinone colorant with SERS (Bruni et al. 2010).718 Researchers have also been able to circumvent the issue of extraction of the719 dyestuff from colored fibers with a variety of alternative methodological approa-720 ches. Jurasekova et al. (2010) used laser photoreduction of an aqueous solution of721 silver nitrate to perform in situ SERS identification of alizarin and carminic acid on722 reference wool and linen fibers dyed with madder and cochineal. In the same work,




UNCORREC

TEDPR

OOF

723 colorants belonging to the flavonoid family have also been identified in fibers dyed724 with dyer’s greenweed (Genista tinctoria L.), onionskin (Allium cepa), chilca725 (Baccharis sp.) and old fustic (Clorophora tinctoria L.) following traditional726 pre-Columbian recipes. This approach was also successfully applied to the identi-727 fication of alizarin in an archaeological Coptic textile from Egypt dated to the 6th–728 8th century. Although the authors were able to detect Al-complexes of the dye-729 stuffs, in many cases the spectra reported appear on top of the bands of carbona-730 ceous material between 1300 and 1600 cm−1, likely formed because of the laser’s731 thermal energy (dwelling for up to 20 min on the analyzed spot) and conductive732 silver nanoparticles.733 A milder, in situ, extractionless SERS method relies on the use of Ag734 nanoparticle pastes obtained by centrifugation of Lee-Meisel colloids that are735 deposited directly on textile fibers. These colloidal pastes were first described by736 Cañamares (Cañamares et al. 2006b). A few years later, Brosseau et al. applied737 them directly to reference wool fibers dyed with purpurin, carminic acid, madder,738 Cape Jasmine, as well as to samples from historical textiles (Brosseau et al. 2009a).739 Thus, they were able to identify purpurin on a silk thread from a 17th-century cover740 from Italy; cochineal in the red and pink silk fibers from a long shawl from France741 dated to the 19th century; and lac, native of India, on a wool thread from a Turkish742 carpet dated to the late 16th/early 17th century. Brosseau and coworkers also743 demonstrated the validity of this methodological approach by identifying curcumin744 in reference silk yarns dyed with turmeric, though a certain degree of pho-745 todegradation was observed in this case, as evidenced by the broad carbon bands on746 which diagnostic signals for the dyestuff were superimposed.747 Interesting studies of alizarin, purpurin and madder include the work of748 Wustholz et al. (2009), who used pastes of silver colloids applied directly on a749 minute fragment of a Peruvian textile (800–1350 A.D.) to identify alizarin, likely750 derived from the native plant Galium corymbosum L. Rambaldi et al. (2015) used751 silver colloids with and without HF pretreatment to record SERS spectra of wool752 threads tinted with various species of madder (Oldenlandia umbellata L. and Rubia753 tinctorum L.), which led to the preferential detection of alizarin or pseudopurpurin754 —a minor component of the dyestuff mixture present in madder—depending on the755 specific botanical species employed for dyeing. Additionally, the silver-doped agar756 gel micro-extraction procedure described above was successfully applied by757 Lofrumento and coworkers on a pre-Columbian textile to detect alizarin after758 training the method on laboratory-dyed textiles treated with alum and purpurin,759 alizarin and caminic acid (Lofrumento et al. 2013). In an effort to identify areas with760 the highest metal coverage by monitoring the morphology and distribution of the761 applied silver nanoparticles, Prikhodko et al. (2015) recently developed a762 hyphenated SEM-Raman system to conduct SERS analysis on reference samples763 dyed with cochineal and madder.764 A few works also concern themselves with lesser studied compounds. For765 example, in situ applications of silver colloids were exploited to characterize the766 colorants of wool threads dyed with materials traditionally used in Mexico, such as767 carminic acid from cochineal (Dactylopius coccus), achiote (Bixa orellana), muitle

184 F. Pozzi et al.



UNCORREC

TEDPR

OOF

768 (Justicia spicigera), zacatlaxcalli (Cuscuta sp.), brazilwood (Caesalpinia echinata),769 and cempazuchitl (Tagetes erecta) (Casanova-González et al. 2012). Doherty et al.770 successfully employed SERS to identify different orchil species, i.e. Roccella771 tinctoria and Lasalia pustulata, on wool fibers both on reference samples and on a772 16th-century purple dyed tapestry from Brussels. In this work, Lee-Meisel colloids773 aggregated with magnesium sulfate were directly deposited onto the threads,774 without any pretreatment (Doherty et al. 2014).775 Early approaches availing themselves of an in situ hydrolysis pretreatment step776 include Leona and Lombardi’s demonstration of the use of HCl vapor to enhance777 detection of berberine on a small thread from a faded and weathered 17th-century778 silk textile from China, which had been previously analyzed with HPLC to reveal a779 mixture of berberine and safflower (Leona and Lombardi 2007).780 The in situ, non-extractive hydrolysis with HF described above for the analysis781 of red lakes in paints (Fig. 1) has also been demonstrated to be of value for the782 analysis of textile samples, a first example of this being the identification of ber-783 berine on a Byzantine textile dated to the 11th century (Leona et al. 2006). The HF784 pretreatment has proven to be highly effective for the analysis of textile samples,785 with the notable exception of silk fibers (Pozzi et al. 2012a). These were found to786 undergo a significant degree of hydrolysis resulting in the release of proteinaceous787 by-products in the analyte solution, which interfere with the analysis by preferential788 adsorption on the silver nanoparticles. Despite the challenges encountered in the789 examination of silk fibers, the HF hydrolysis was still successfully applied to the790 analysis of both reference and historic silk samples. For example, a reference silk791 fiber dyed with weld, from the Reseda luteola dried plant, was treated with HF792 vapors and then analyzed with concentrated Ag colloids by Corredor et al. (2009),793 revealing the presence of luteolin. This method was also applied to the identification794 of reference silk dyed with extracts of Persian berries, demonstrating the identifi-795 cation of some diagnostic bands for quercetin, probably mixed with other flavo-796 noids such as rhamnetin and kaempferol (Teslova et al. 2007). Other examples797 include the examination of reference samples of unmordanted and mordanted silk,798 as well as mordanted wool fibers dyed with Cape Jasmine, leading to the SERS799 identification of the main chromophores of this dye, the carotenoids crocetin and800 crocin (Cañamares et al. 2010). The authors noted, though, that in this case SERS801 analysis allowed the detection of the diagnostic bands for the dyestuffs even without802 the HF hydrolysis pretreatment step, likely because the two colorants identified also803 display a strong normal Raman spectrum.804 Zaffino et al. (2014) extended the SERS analysis of colorants in textiles to NIR805 excitation, comparing HF pretreatment and non-hydrolysis procedures on a large set806 of reference dyes and dyed fibers, as well as historical samples from Chinese807 Ningxia (18th and 19th century) and Caucasian Kaitag (17th and 18th century)808 textiles. This work included the first instance of FT-SERS spectrum of an iron-gall809 dye on a sample from the latter (Zaffino et al. 2014). In one of the most extensive810 applications of the combined HF pretreatment and silver colloid SERS, Pozzi et al.811 examined an important corpus of Navajo blankets in the collection of the Art812 Institute of Chicago (Pozzi et al. 2014b), allowing the identification of the colorants




UNCORREC

TEDPR

OOF

813 in their various red woolen yarns in a selection of 8 19th-century blankets. While814 the presence of carminic acid was positively confirmed by SERS for fibers dyed815 with natural dyes, synthetic organic colorants were identified by normal Raman816 spectroscopy with 785 nm excitation, highlighting the use of Ponceau 4R and two817 beta-naphthol dyes. Results from this study had a major impact on the dating of818 some of the blankets examined, that had been given an earlier date based on stylistic819 grounds only (Fig. 9).820 In many cases, it is advantageous to combine SERS with complementary methods821 of colorant identification and materials characterization. This multi-technique822 approach was followed, for example, in the analysis of Kaitag textiles from Russia823 (Pozzi et al. 2012b), where SERS was used in association with SEM/energy dis-824 persive X-ray (EDX) analysis, visible reflectance spectroscopy, HPLC, and XRF825 spectrometry for a complete characterization of the textile fibers, with special focus826 on their colorants and deterioration products. The two dozen textiles analyzed, dating827 from the 17th to 18th century, were primarily made of cotton with silk embroideries828 and showed a variety of colorants. Of those, SERS identified madder for red; indigo829 for dark and light blue; weld (or another luteolin-based dye) for yellow; the same830 luteolin-based dye, or sometimes a different yellow one, in combination with indigo831 for green; tannins mordanted with iron for dark brown; again tannins, sometimes with832 the addition of indigo, for black. In this particular case study, the main limitation of833 SERS emerged during examination of those threads that were shown through the834 multi-technique analytical approach to be dyed with more than one dyestuffs, as only835 one component was detected by SERS.

Fig. 9 Chief Blanket (Third Phase), 1865/80. Navajo; northern New Mexico or Arizona, UnitedStates. Wool, single interlocking tapestry weave; twined edges; corner tassels; two selvagespresent; 149.6 × 173.9 cm (58 7/8 × 68 1/2 in.). Collins et al. (2001). The SERS spectra of twored fibers removed from the blanket (a, b) are in accordance with that obtained from a referencewool sample dyed with cochineal (c)

186 F. Pozzi et al.



UNCORREC

TEDPR

OOF

836 To date, only a handful of examples have been reported on the successful SERS837 detection of multiple dyestuffs on a single textile fiber. One of these accounts was838 provided by Idone and coworkers, who performed direct, extractionless identifi-839 cation of both cochineal carmine and brazilwood on animal and vegetable fibers840 using colloidal pastes (Idone et al. 2013). The study focused on the much admired841 textiles produced in Italy in the early 20th century by artist Mariano Fortuny y de842 Madrazo (Granada, Spain, 1871—Venice, Italy, 1949). The samples dyed with both843 carmine and brazilwood, as identified by SERS, included two red cotton panels844 produced by the Società Anonima Fortuny and a silk thread of a red velvet panel845 made at the Palazzo Pesara-Orfei in Venice, the latter also displaying a Mariano846 Fortuny Deposé stamp.

847 5.2 Paintings

848 Although many of the technical and methodological improvements in the field of849 SERS for cultural heritage have only been tested on mock-up paint layers (Doherty850 et al. 2011, 2014; Platania et al. 2014; Retko et al. 2014), SERS is becoming851 increasingly popular in real-world case studies to detect and identify red dyes and852 lake pigments in samples from actual paintings. The first application in this context,853 reported in 2009, concerns the analysis of a 50-µm red glaze sample from St. John854 the Baptist Bearing Witness (ca. 1506–07), a painting from the workshop of Italian855 artist Francesco Granacci (Leona 2009). In this case, SERS analysis was performed856 on glucose-reduced citrate-capped silver nanoparticles with HF hydrolysis pre-857 treatment of the sample, proving the effectiveness of this combined methodology858 for the ultra-sensitive analysis of medium-rich paint samples such as colorants859 dispersed in oil glazes. Remarkably, the identification of kermes in the specimen860 examined is consistent with this anthraquinoid dye being deemed as the main861 colorant for red glazes in Europe before the introduction of cochineal from the New862 World in the second decade of the 16th century. This analytical procedure, pro-863 gressively grown into a standard protocol at the Metropolitan Museum of Art and864 other institutions, has then proven successful in various routine applications,865 including the characterization of madder lake in oil paint samples from The Card866 Players (1623–24) by the Dutch painter Jan Lievens and from an homonymous867 painting (1890–92) by Paul Cézanne (Pozzi et al. 2013b).868 The first extractionless non-hydrolysis study of organic dyes in oil paintings was869 reported by Oakley et al. (2011), who employed colloidal pastes to examine red870 colorants in oil glazes from works belonging to the Colonial Williamsburg871 Foundation collection, i.e. Portrait of William Nelson (1748–50) by the earliest872 native-born American artist of European descent, Robert Feke, and Portrait of873 Isaac Barré (1766) by Sir Joshua Reynolds, a founder of the Royal Academy of874 Arts. In this case, in addition to providing a probe of local environment by means of875 correlated fluorescence measurements, the authors identified carmine lake in




UNCORREC

TEDPR

OOF

876 samples from the two paintings, albeit showing poor reproducibility due to877 unevenness of the silver nanoparticle coating.878 A systematic work comparing SERS spectra of anthraquinoid red dyes from a879 wide variety of substrates obtained with or without HF hydrolysis confirmed how,880 in many cases, highly reproducible, conclusive dye identification in lake-containing881 glazes can only be achieved upon HF hydrolysis (Pozzi et al. 2012a). The high882 effectiveness of the HF treatment for SERS analysis of oil paints has been initially883 demonstrated on the aforementioned samples from Cézanne’s The Card Players, as884 well as paints from Henri Matisse’s The Young Sailor (1906), and Rembrandt’s885 Aristotle With a Bust of Homer (1653). Further proof of the efficacy of the HF886 treatment is offered in another article by Pozzi et al. (2014a), which examines a887 wide selection of red lake oil paint reconstructions prepared according to888 19th-century historical recipes, proving that successful dye identification can be889 accomplished upon HF hydrolysis even when inorganic pigments, extenders,890 ground materials or binding media are associated with the red lake in the sample891 analyzed. This comprehensive work on reference materials was then extended, in892 the same paper, to the systematic analysis of samples from 19th-century French893 Impressionist and post-Impressionist paintings in the collection of the Art Institute894 of Chicago, including works by Manet, Pissarro, Renoir, Monet and Gauguin. This895 study demonstrated the pervasive use of mostly two types of red lakes by these896 artists: cochineal and madder. In the case of Renoir’s Chrysanthemums (1881–82),897 microscopic examination of the painting revealed the presence of a very thin,898 purple-colored overpaint on top of the original red lake. The two paint layers could899 be analyzed individually upon careful separation by means of a scalpel, revealing900 the use of cochineal as the original lake and Pigment Red 48:3 for the *20-μm901 thick retouching (Fig. 10).902 The first evidence of SERS identification of binary mixtures of red lakes in903 painting layers was offered in two instances, i.e. Renoir’s Woman at the Piano904 (1875–76) and Édouard Manet’sWoman Reading (1879–80). In the first case, a few905 particles with an orange fluorescence emission were spotted within a906 non-fluorescing red lake paint layer in a cross section sample removed from the907 painting. The identity of the two types of lake pigments could be determined by908 SERS: while the signals observed right after sample preparation were consistent909 with madder, after a few minutes the spectrum of carminic acid was recorded from910 the same sample. This observation most likely indicates that the alizarin and pur-911 purin molecules originating from the fluorescing madder particles, which have a912 higher affinity for the silver substrate (Cañamares et al. 2006a), were adsorbed first913 on the nanoparticle surface. As aggregation proceeded, carminic acid from cochi-914 neal—present in significantly higher amounts in the sample within the915 non-fluorescing layer—then occupied the remaining surface sites, resulting in the916 carminic acid signals becoming predominant in the SERS spectrum. As to Manet’s917 Woman Reading, while the SERS data acquired from the corresponding samples918 did not match the individual reference spectra of commercial madder or cochineal919 lakes, a perfect correspondence was found with the spectrum of a red lake paint

188 F. Pozzi et al.



UNCORREC

TEDPR

OOF

920 reconstruction prepared as a mixture of cochineal and Kopp’s purpurin (a921 purpurin-rich extract of the madder lake) on aluminum sulfate (Pozzi et al. 2014a).922 Beside successfully detecting red lakes in combination in a few cases, SERS was923 used by Oakley et al. (2012) upon treatment with sulfuric acid to probe mixtures of924 blue inorganic pigments and organic colorants, i.e. Prussian blue and indigo, in a925 single sample from the early 18th-century oil painting Portrait of Evelyn Byrd926 (1725–26).927 Most recently, the applicability of SERS to the analysis of oil paints and lake928 glazes has been further expanded to include the examination of paint layers in cross929 section samples. Idone et al. (2014) used an extractionless non-hydrolysis approach930 to characterize highly fluorescing red lake pigments in a cross section from a931 16th-century mural painting of Sant’Anna Metterza. Although the experiment was932 successful, the authors emphasized the need to repeat the experiment on a number933 of different areas in order to find suitable spots where clusters of aggregated934 nanoparticles give rise to a significantly improved signal enhancement. Frano et al.

Fig. 10 Pierre-Auguste Renoir’s Chrysanthemums, 1881/82 (Mr. and Mrs. Martin A. RyersonCollection, The Art Institute of Chicago 1933.1173), where sampling site is indicated by an arrowin the painting image and micrograph. The UV-induced fluorescence image (500×) of a samplemounted as a cross section shows the presence of a thin varnish application and a purplishretouching layer on top of the original paint. The SERS spectrum of the Renoir’s original red lake(a) is in accordance with those obtained from carmine oil paint historical reconstructions (b), whilethe spectral pattern of the retouching material (c) matches reference Pigment Red 48:3 (d)




UNCORREC

TEDPR

OOF

935 (2014) identified carmine and madder lakes in cross sections of 18th- and936 19th-century paintings by direct deposition of colloidal pastes. Although more937 complex and equipment-intensive, the coupling of UV laser ablation sampling with938 SERS detection on a vapor-deposited silver nanoisland film, described above, has939 recently led to the successful identification of kermes in samples from The940 Incredulity of Saint Thomas by Venetian master Luca Signorelli and Adoration of941 the Shepherds by Giorgione, both dated to the 15th century (Cesaratto et al. 2014).

942 5.3 Works of Art on Paper

943 Applications of SERS to the analysis of colorants used in works of art on paper are944 still rare. In graphic works, the pigment particles are typically bound with small945 quantities of aqueous media such as glue or gums, and mixtures of compounds tend946 to be simpler than for pigments used in paintings. Therefore, reported uses of SERS947 in this area have involved the direct deposition of colloidal pastes on small pigment948 particles removed from the artworks. To date, various authors have successfully949 performed SERS analysis on watercolors and pastels from the 19th century, illu-950 minated manuscripts, and Japanese screens and woodblock prints.951 Brosseau et al. (2009b) analyzed the colorants of late 19th-century pastel sticks952 that belonged to artist Mary Cassatt (1844–1926), ranging in color from pale pink to953 dark purple, and found dyestuffs that were also encountered on a pastel drawing by954 the artist in the collection of the Art Institute of Chicago. Analysis was performed955 with citrate-reduced Lee-Meisel colloids concentrated into a paste and deposited956 directly on individual pigment grains. Interestingly, the materials identified show957 the coexistence of traditional dyes with some newly introduced synthetic organic958 colorants in the palette of the artist, including carmine lake, madder, rhodamine B959 and 6G, and a beta-naphthol or monoazo dye.960 Regarding the watercolor medium, Pozzi et al. (2013c) used the HF pretreatment961 to analyze reference historical lake pigments in series of watercolor washes962 included in historical catalogs of the famed English colorman Winsor and Newton,963 and reported identification of madder on a sample removed from Silver Ball, Barge964 and Trees (1930), a watercolor by Arthur Dove made with gouache, ink and965 charcoal (Pozzi et al. 2013b). SERS analysis of a sample taken from an extremely966 faded area of the sky of Winslow Homer’s watercolor For to Be a Farmer’s Boy967 (1887) found cochineal, Indian purple (a precipitate of cochineal carmine on copper968 substrate) and purple madder by comparison with reference spectra taken on a969 Winsor and Newton antique book showing swatches of original, unfaded water-970 color samples (Brosseau et al. 2011). In Homer’s work, the red lakes were found to971 be mixed with vermilion and chrome yellow, which led to conclude that the sky972 must once have depicted a sunset or dawn scene. Based on this newly found

190 F. Pozzi et al.



UNCORREC

TEDPR

OOF

973 evidence, a decision was made to digitally recolorize an image of the work of art to974 offer scholars and the public the opportunity to view the work as the artist originally975 intended.976 Several researchers have recently added SERS to the toolbox of analytical977 methods deployed to investigate illuminated manuscripts, artifacts that are generally978 studied mostly non-invasively and for which miniaturization of the samples ana-979 lyzed is of paramount importance in light of the rarity of the illuminations and980 extremely thin layers of colored media. Castro et al. (2014) combined minimally981 invasive SERS with non-invasive microspectrofluorimetry to identify, upon HF982 hydrolysis, lac dye reds in 12th–13th-century medieval illuminations from Portugal.983 Aceto et al. (2015) successfully combined SERS and other micro-invasive tech-984 niques (such as matrix assisted laser desorption ionization-time of flight-mass985 spectrometry, and inductively coupled plasma-mass spectrometry) with FTIR,986 FT-Raman, fiber optic reflectance spectrophotometry, spectrofluorimetry, and XRF987 spectrometry to identify folium and orchil on various reference samples including988 dyed parchment—a mock up of the exceedingly rare purple codices. The authors989 claim that SERS was effective at identifying the colorants on the dyed parchment990 samples with and without extraction with formic acid, and report the first SERS991 spectrum of folium. El Bakkali et al. (2014) used UV/vis reflectance and fluores-992 cence spectroscopies alongside SERS to identify carminic acid in red and pink inks993 on 19th-century manuscripts from Morocco. In a work by Leona and coworkers994 (featured in Sgamellotti et al. 2014), SERS on silver colloids was used to identify995 lac dye in the center of the iris petals in Ogata Kōrin’s masterpiece Irises at996 Yatsuhashi (1709 or later), a pair of Japanese screens from the early 18th century in997 the collection of the Metropolitan Museum of Art. The identification of a triaryl-998 methane dye, either methyl violet or crystal violet (the two dyes cannot be differ-999 entiated by SERS) on a Japanese print from 1892, Sekigahara Homare no Gaika (A

1000 poem about the battle of Sekigahara) by Toyoharu Kunichika, was also reported by1001 Leona et al. as an example of both gel-sampling SERS and inkjet colloid delivery1002 SERS.1003 Additional work on Japanese woodblock prints of the Meiji period (1868–1912)1004 is being conducted at the Metropolitan Museum of Art. The two-step SERS pro-1005 tocol developed by Pozzi et al. (2012a) has been used to detect what are probably1006 some of the earliest uses of magenta, cochineal, and eosine in Japan (Fig. 11)1007 (Leona et al. 2015).1008 Of relevance to the analysis of graphic documents are the recent applications of1009 SERS to study historical inks. Roldán et al. (2014) combined SERS and other1010 analytical techniques to identify inks made with bistre. Among historic inks, sepia1011 can be challenging to identify in old and aged samples because of the similarity of1012 its spectral pattern with the profiles of carbon-based blacks. Centeno, Roldán, and1013 respective coworkers (Centeno and Shamir 2008; Roldán et al. 2014) have studied1014 this subject extensively and combined SERS with complementary analytical tech-1015 niques to successfully characterize the synthetic chromophore in both reference and1016 historical materials.




UNCORREC

TEDPR

OOF

1017 5.4 Other Applications

1018 The ability to detect minute amounts of colorants makes SERS an ideal technique1019 for the investigation of archaeological samples, which often contain only very small1020 remnants of the original material and are heavily affected by contamination and1021 aging. Hence, SERS may sometimes be the only analytical method that can be used1022 to obtain dye information from these very ancient and often fragmentary artifacts.1023 The first application of SERS to archaeological material was demonstrated by Van1024 Elslande et al. (2008), who detected purpurin in Roman cosmetics. Subsequently,1025 Leona (2009) described the earliest identification of madder dye in a 4000 years old1026 Egyptian leather fragment. The identification of madder in archaeological samples1027 including pink pigments from an excavation at Corinth (2nd century B.C.) and a1028 statue of Caligula (Virginia Museum of Fine Art, 1st century A.D.) were reported1029 by Pozzi et al. (2013b). Londero et al. (2013) also identified madder in a fragment1030 from the trappings of an ancient Egyptian chariot of Amenhotep III (New Kingdom,1031 1390–1352 B.C.) pretreated with HF and analyzed with LA-SERS.1032 Few studies exist in the literature in relation to lake pigments used to decorate1033 polychrome sculpture and furniture. Some notable examples include studies of red1034 glazes in medieval polychrome sculpture at the Metropolitan Museum of Art, i.e.1035 the detection of lac dye in a Spanish crucifix dating to 1150–1200, which is thought1036 to be the earliest example of the use of lac dye in European sculpture (Pozzi et al.1037 2013b). From a slightly later date, ca. 1175–1200, the French Romanesque1038 sculpture of a Virgin and Child in Majesty, made in Auvergne, France, is the1039 earliest documented example of lac dye use in France (Leona 2009). These findings1040 document trade from India, where the colorant originates from, in medieval times

Fig. 11 Hiroshige III, View of Benten on Nakanoshima in Shinobazu Pond, Ueno Park, 1881.Woodblock print triptych on paper. The violet is methyl violet, the lighter red on the robes iscochineal, and the pink of the cherry blossoms is eosine

192 F. Pozzi et al.



UNCORREC

TEDPR

OOF

1041 and are of great historical importance. Daher et al. (2014) also used non-hydrolysis1042 SERS on Ag colloids to identify carmine and madder lake in pink layers from1043 French decorative arts objects such as a weaving shuttle and a commode dating1044 from the 18th century.1045 Introduced in the mid-19th century as the first form of recording a photographic1046 image on a plate, daguerreotypes by their very nature owe to the image-forming1047 process the presence of silver/mercury nanoparticles on their surface. Hence,1048 Centeno et al. (2008) exploited this structural and compositional property of the1049 medium to diagnose the nature of a white degradation haze forming on1050 daguerreotypes originating from the famed Southworth and Hawes studio. Silver1051 chloride compounds and substituted aromatic compounds were identified by simply1052 focusing a 785 nm laser onto the naturally nanostructured silver surface of the1053 daguerreotypes, giving rise to the SERS effect. Because metallic silver can be1054 redeposited when silver chlorides are exposed to UV/vis illumination, ultimately1055 obfuscating the original image, the SERS data was instrumental to inform exhibi-1056 tion policies for the collection.

1057 5.5 The Interface with Forensics: Modern Inks

1058 The examination of questioned documents is an active area of interest for the1059 application of Raman spectroscopy and SERS, which are considered of value in1060 identifying and comparing the inks of questioned entries. Since it represents a1061 relatively non-destructive technique, Raman spectroscopy promotes the identifica-1062 tion of pigments and dyes, which often represent the most intense contributors to1063 the Raman signal. While several reviews on the analysis of questioned documents1064 discuss most of the recent literature concerning normal Raman spectroscopy (Braz1065 et al. 2013; Calcerrada and Garcia-Ruiz 2015), the special contributions of SERS1066 studies are presented here.1067 In recent years, the focus of forensic research on the characterization of pen inks1068 has turned to ball point and gel inks. These are especially amenable to SERS studies1069 since the dyes typically display a strong SERS enhancement, and the fluorescence1070 interference that usually hampers normal Raman spectroscopy is readily suppressed1071 by adsorption of the analyte on a metal substrate. A recent study on synthetic dyes1072 compared systematically the discrimination capabilities of SERS with various1073 excitation wavelengths (Geiman et al. 2009). Ten dyes representing classes com-1074 monly found in ink formulations were selected for the study: Acid Blue 1, Acid1075 Orange 10, Acid Red 52, Aniline Blue, Crystal Violet, Methyl Violet,1076 Pararosaniline, Rhodamine B, Sudan Black B, and Victoria Blue B. These dyes1077 were studied at several excitation wavelengths, i.e. 633, 785 nm (with and without1078 SERS), as well as FT-Raman 1064 nm. Among normal Raman techniques, only the1079 FT-Raman produced spectra with fairly good signal intensity and signal-to-noise1080 ratios, while 633 and 785 nm laser excitation caused high levels of fluorescence1081 overwhelming the dye signals. SERS spectra were obtained using dilute solutions




UNCORREC

TEDPR

OOF

1082 of the dye molecules with silver colloids, and although some differences were1083 encountered between SERS and the normal Raman spectra, presumably due to1084 differences in the selection rules, a considerable net increase in signal intensity was1085 observed. In a related study, Seifar et al. reported that methyl violet exhibited a1086 strong resonance effect (SERS) at 514.5 and 457.9 nm, while other dyes were most1087 easily observed with 785 nm excitation (Seifar et al. 2001). Bell and Spence (2001)1088 have compared the discriminant advantages of SERS using in situ deposition of1089 silver colloids and normal Raman spectroscopy combined with other analytical1090 techniques commonly employed in document examination. Compared with various1091 extraction methods such as TLC, they found that Raman techniques provided an1092 equivalent and sometimes better discrimination, while being both rapid and1093 non-destructive. SERS ensured an increased differentiation when compared with1094 normal Raman spectroscopic techniques, but problems of reproducibility were1095 obtained in several samples. Luo et al. (2013) employed gold nanoparticles as a1096 SERS active substrate in order to analyze pens and printers/copiers inks in situ at1097 633 nm. Raza and Saha employed silver-doped agarose gel disks as SERS substrate1098 for examining ballpoint inks, carrying out analysis at 785 and 514 nm excitation.1099 The gel disks have been shown to extract the ink on small sections of the pen1100 strokes. The dyes extracted into the gel medium were found to be stable for more1101 than thirty days and were furthermore examined by attenuated total reflection1102 (ATR)-FTIR in sequence (Raza and Saha 2013). White (2003) reported an inter-1103 esting SERS experiment on a three year old bank check that was coated with silver1104 colloid. Finally, Wagner and Clement (2001) described the use of SERS (silver1105 colloid, 633 nm) on ballpoint inks as well as fluid inks of varying colors.1106 A comparison of the discrimination power of normal Raman spectroscopy and1107 SERS showed significant enhancement of the Raman signal using SERS. However,1108 black ballpoint inks apparently lack characteristic features. They are most likely1109 composed entirely of graphite, which usually displays only several broad peaks in1110 normal Raman spectroscopy, and are difficult to characterize using SERS.1111 Only few reports have focused on ink jet/toner printers as analyzed by SERS.1112 Rodger et al. have obtained SERS on a set of four different ink jet dyes on five1113 different papers on silver colloid, at 514 and 633 nm, as well as FT-Raman at1114 1064 nm (Rodger et al. 2000). SERS was shown to be successful for identifying the1115 dyes, while FT-Raman produced interference from the paper and filler as well. This1116 was explained as due to the increased size of the area sampled, which most likely1117 included some of the paper. Additionally, SERS is relatively insensitive to the1118 ingredients in paper. Luo et al. (2013) have examined a gold colloid formulation on1119 ink from printers and copiers as well as pen inks, although no information on the1120 samples were provided.1121 Beside the popularity of pen and printers inks, security inks have a special1122 forensic interest. They are largely used on banknotes (either as the ink used in the1123 banknote itself or as an added stain during robberies), various official documents,1124 and checks, by doping the material of interest with pigments or nanoparticles that1125 exhibit unique optical properties such as fluorescence. Historically, security inks1126 were formulated using materials doped with rare earth lanthanides elements such as

194 F. Pozzi et al.



UNCORREC

TEDPR

OOF

1127 europium, terbium, ytterbium, thulium or erbium. These rare earths have unique1128 spectral features when added to a normal ink, which can be clearly discriminated1129 from any other attempted forgery or counterfeit. In recent years different patents1130 have been issued which utilize SERS taggants or SERS active sols in security inks.1131 SERS nanotags consist of a nanoparticle coated with a unique “reporter” molecule,1132 which is then encapsulated by an inert silica or polymer layer. The outermost layer1133 ensures nanoparticle stability over time and protection from environmental condi-1134 tions, so that the taggants can be analyzed even after many years.

1135 6 Conclusions and Future Outlook

1136 It has taken the science of SERS over thirty years since inception for it to become a1137 mature field of study. This is testament to the wide variety and applicability of its1138 scope, as well as the potential for numerous applications covering a wide variety of1139 disciplines. Furthermore, the theoretical underpinnings have engendered lively1140 controversy, and have also led to connections to numerous other areas of scientific1141 concern. Ultimately, the potent combination of high sensitivity and high resolution1142 afforded by SERS makes it unique as an analytical tool, and provides numerous1143 pathways for potential applications. In this chapter, we have reviewed the appli-1144 cation of SERS to cultural heritage objects and, due to the similarity of analytical1145 techniques and materials, to systems of interest to forensic science. It may be fairly1146 said that even the above compilation is likely incomplete, and subject to aug-1147 mentation in the near future.1148 Researchers have reached the hallmark of a robust field, which should be cel-1149 ebrated. However, this brings up the question as to where the field goes next.1150 Certainly, some of the techniques discussed above are still in their infancy, such as1151 TERS, laser ablation—SERS, and microfluidics—SERS. Therefore, considerable1152 intensification of effort and progress should be expected in these areas. More1153 common hybrid techniques such a TLC-SERS have the promise to allow SERS to1154 become more multidimensional in its approach. More well-developed areas, such as1155 substrate development or extraction methodology, have been of considerable1156 interest for some time, and may not see such rapid change in the future. In any case,1157 the exciting aspect of recent developments is that there is an increasing variety of1158 new locales in which SERS is applied to cultural heritage and forensic science. This1159 proliferation confirms that the applications described here are becoming more1160 widely accepted, and of interest to a wider scientific community for routine-based1161 use. We hope that this review will become a springboard for new researchers1162 interested in advancing the field through their own ideas and unique applications.

1163 Acknowledgments SZ, FC and RPVD would like to acknowledge the National Science1164 Foundation (MRSEC NSF DMR-1121262, NSF CHE-1152547, and NSF CHE-1041812) and the1165 Northwestern University/Art Institute of Chicago Center for Scientific Studies in the Arts1166 (NU-ACCESS) for their support. Additionally, this work made use of the EPIC facility (NUANCE




UNCORREC

TEDPR

OOF

1167 Center-Northwestern University), which has received support from the MRSEC program (NSF1168 DMR-1121262) at the Materials Research Center; the International Institute for Nanotechnology1169 (IIN); and the State of Illinois, through the IIN. JL and ML thank the National Science Foundation1170 (CHE-1402750) for funding. FC and FP are indebted to the Andrew W. Mellon Foundation for1171 funding a postdoctoral research position and providing support for Conservation Science at the Art1172 Institute of Chicago.

1173 References

1174 Aceto M, Arrais A, Marsano F, Agostino A, Fenoglio G, Idone A, Gulmini M (2015) A diagnostic1175 study on folium and orchil dyes with non-invasive and micro-destructive methods. Spectrochim1176 Acta Part A Mol Biomol Spectrosc 142:159–168. doi:10.1016/j.saa.2015.02.0011177 Ahn C, Obendorf SK (2004) Dyes on archaeological textiles: analyzing alizarin and its degradation1178 products. Text Res J 74:949–954. doi:10.1177/0040517504074011021179 Ahn C, Obendorf SK (2007) GGMS analysis of curcumin dye after selective degradation1180 treatment. Fibers Polym 8:278–283. doi:10.1007/BF028772701181 Albrecht M, Creighton J (1977) Anomalously Intense Raman-Spectra of Pyridine at a Silver1182 Electrode. J Am Chem Soc 99:5215–5217. doi:10.1021/ja00457a0711183 Alfeld M, De Nolf W, Cagno S, Appel K, Siddons DP, Kuczewski A, Janssens K, Dik J,1184 Trentelman K, Walton M, Sartorius A (2013a) Revealing hidden paint layers in oil paintings by1185 means of scanning macro-XRF: a mock-up study based on Rembrandt’s “An old man in1186 military costume”. J Anal At Spectrom 28:40–51. doi:10.1039/c2ja30119a1187 Alfeld M, Pedroso JV, van Eikema Hommes M, Van der Snickt G, Tauber G, Blaas J, Haschke M,1188 Erler K, Dik J, Janssens K (2013b) A mobile instrument for in situ scanning macro-XRF1189 investigation of historical paintings. J Anal At Spectrom 28:760–767. doi:10.1039/c3ja30341a1190 Amat A, Miliani C, Romani A, Fantacci S (2015) DFT/TDDFT investigation on the UV-vis1191 absorption and fluorescence properties of alizarin dye. Phys Chem Chem Phys 17:6374–6382.1192 doi:10.1039/c4cp04728a1193 Bacci M, Baldini F, Carla R, Linari R (1991) A color analysis of the Brancacci Chapel frescoes.1194 Appl Spectrosc 45:26–31. doi:10.1366/00037029143377131195 Baglioni P, Carretti E, Chelazzi D (2015) Nanomaterials in art conservation. Nat Nanotechnol1196 10:287–290. doi:10.1038/nnano.2015.381197 Baran A, Wrzosek B, Bukowska J, Proniewicz LM, Baranska M (2009) Analysis of alizarin by1198 surface-enhanced and FT-Raman spectroscopy. J Raman Spectrosc 40:436–441. doi:10.1002/1199 jrs.21471200 Bell SEJ, Spence SJ (2001) Disposable, stable media for reproducible surface-enhanced Raman1201 spectroscopy. Analyst 126:1–3. doi:10.1039/b009519m1202 Bell IM, Clark RJH, Gibbs PJ (1997) Raman spectroscopic library of natural and synthetic1203 pigments (pre- ≈ 1850 AD). Spectrochim Acta Part A Mol Biomol Spectrosc 53:2159–2179.1204 doi:10.1016/S1386-1425(97)00140-61205 Benedetti DP, Zhang J, Tague TJ, Lombardi JR, Leona M (2014) In situ microanalysis of organic1206 colorants by inkjet colloid deposition surface-enhanced Raman scattering. J Raman Spectrosc1207 45:123–127. doi:10.1002/jrs.44241208 Braz A, Lopez-Lopez M, Garcia-Ruiz C (2013) Raman spectroscopy for forensic analysis of inks1209 in questioned documents. Forensic Sci Int 232:206–212. doi:10.1016/j.forsciint.2013.07.0171210 Brosseau CL, Gambardella A, Casadio F, Grzywacz CM, Wouters J, Van Duyne RP (2009a)1211 Ad-hoc surface-enhanced Raman spectroscopy methodologies for the detection of artist1212 dyestuffs: thin layer chromatography-surface enhanced Raman spectroscopy and in situ on the1213 fiber analysis. Anal Chem 81:3056–3062. doi:10.1021/ac802761v

196 F. Pozzi et al.



http://dx.doi.org/10.1016/j.saa.2015.02.001

http://dx.doi.org/10.1177/004051750407401102

http://dx.doi.org/10.1007/BF02877270

http://dx.doi.org/10.1021/ja00457a071

http://dx.doi.org/10.1039/c2ja30119a

http://dx.doi.org/10.1039/c3ja30341a

http://dx.doi.org/10.1039/c4cp04728a

http://dx.doi.org/10.1366/0003702914337713

http://dx.doi.org/10.1038/nnano.2015.38

http://dx.doi.org/10.1002/jrs.2147


http://dx.doi.org/10.1039/b009519m

http://dx.doi.org/10.1016/S1386-1425(97)00140-6


http://dx.doi.org/10.1016/j.forsciint.2013.07.017

http://dx.doi.org/10.1021/ac802761v

UNCORREC

TEDPR

OOF

1214 Brosseau CL, Rayner KS, Casadio F, Grzywacz CM, van Duyne RP (2009b) Surface-enhanced1215 Raman spectroscopy: a direct method to identity colorants in various artist media. Anal Chem1216 81:7443–7447. doi:10.1021/ac901219m1217 Brosseau CL, Casadio F, Van Duyne RP (2011) Revealing the invisible: using surface-enhanced1218 Raman spectroscopy to identify minute remnants of color in Winslow Homer’s colorless skies.1219 J Raman Spectrosc 42:1305–1310. doi:10.1002/jrs.28771220 Bruni S, Guglielmi V, Pozzi F (2010) Surface-enhanced Raman spectroscopy (SERS) on silver1221 colloids for the identification of ancient textile dyes: tyrian purple and madder. J Raman1222 Spectrosc 41:175–180. doi:10.1002/jrs.24561223 Bruni S, Guglielmi V, Pozzi F (2011a) Historical organic dyes: a surface-enhanced Raman1224 scattering (SERS) spectral database on Ag Lee-Meisel colloids aggregated by NaClO4.1225 J Raman Spectrosc 42:1267–1281. doi:10.1002/jrs.28721226 Bruni S, Guglielmi V, Pozzi F, Mercuri AM (2011b) Surface-enhanced Raman spectroscopy1227 (SERS) on silver colloids for the identification of ancient textile dyes. Part II: pomegranate and1228 sumac. J Raman Spectrosc 42:465–473. doi:10.1002/jrs.27361229 Calcerrada M, Garcia-Ruiz C (2015) Analysis of questioned documents: a review. Anal Chim Acta1230 853:143–166. doi:10.1016/j.aca.2014.10.0571231 Cañamares MV, Leona M (2007) Surface-enhanced Raman scattering study of the red dye laccaic1232 acid. J Raman Spectrosc 38:1259–1266. doi:10.1002/jrs.17611233 Cañamares MV, Lombardi JR (2015) Raman, SERS, and DFT of Mauve Dye: adsorption on Ag1234 nanoparticles. J Phys Chem C 119:14297–14303. doi:10.1021/acs.jpcc.5b026191235 Cañamares MV, Garcia-Ramos JV, Domingo C, Sanchez-Cortes S (2004) Surface-enhanced1236 Raman scattering study of the adsorption of the anthraquinone pigment alizarin on Ag1237 nanoparticles. J Raman Spectrosc 35:921–927. doi:10.1002/jrs.12281238 Cañamares MV, Garcia-Ramos JV, Domingo C, Sanchez-Cortes S (2006a) Surface-enhanced1239 Raman scattering study of the anthraquinone red pigment carminic acid. Vib Spectrosc1240 40:161–167. doi:10.1016/j.vibspec.2005.08.0021241 Cañamares MV, Sanchez-Cortes S, Garcia-Ramos JV (2006b) In: IRUG7 proceedings, New York,1242 28–31 Mar 2006. Museum of Modern Art, p 191243 Cañamares MV, Garcia-Ramos JV, Gomez-Varga JD, Domingo C, Sanchez-Cortes S (2007) Ag1244 nanoparticles prepared by laser photoreduction as substrates for in situ surface-enhanced raman1245 scattering analysis of dyes. Langmuir 23:5210–5215. doi:10.1021/la063445v1246 Cañamares MV, Chenal C, Birke RL, Lombardi JR (2008a) DFT, SERS, and single-molecule1247 SERS of crystal violet. J Phys Chem C 112:20295–20300. doi:10.1021/jp807807j1248 Cañamares MV, Lombardi JR, Leona M (2008b) Surface-enhanced Raman scattering of1249 protoberberine alkaloids. J Raman Spectrosc 39:1907–1914. doi:10.1002/jrs.20571250 Cañamares MV, Garcia-Ramos JV, Sanchez-Cortes S, Castillejo M, Oujja M (2008c)1251 Comparative SERS effectiveness of silver nanoparticles prepared by different methods: a1252 study of the enhancement factor and the interfacial properties. J Colloid Interface Sci 326:103–1253 109. doi:10.1016/j.jcis.2008.06.0521254 Cañamares MV, Lombardi J, Leona M (2009) Raman and surface enhanced Raman spectra of1255 7-hydroxyflavone and 3′,4′-dihydroxyflavone. E-PS, pp 81–881256 Cañamares MV, Leona M, Bouchard M, Grzywacz CM, Wouters J, Trentelman K (2010)1257 Evaluation of Raman and SERS analytical protocols in the analysis of Cape Jasmine dye1258 (Gardenia augusta L.). J Raman Spectrosc 41:391–397. doi:10.1002/jrs.24621259 Cañamares MV, Reagan DA, Lombardi JR, Leona M (2014) TLC-SERS of mauve, the first1260 synthetic dye. J Raman Spectrosc 45:1147–1152. doi:10.1002/jrs.45081261 Cardon D (2007) Natural dyes: sources, tradition, technology and science. Archetype, London1262 Carretti E, Bonini M, Dei L, Berrie BH, Angelova LV, Baglioni P, Weiss RG (2010) New1263 Frontiers in materials science for art conservation: responsive gels and beyond Acc Chem Res1264 43:751–760. doi:10.1021/ar900282h1265 Casadio F, Leona M, Lombardi JR, Van Duyne R (2010) Identification of organic colorants in1266 fibers, paints, and glazes by surface enhanced Raman spectroscopy. Acc Chem Res 43:782–1267 791. doi:10.1021/ar100019q




http://dx.doi.org/10.1021/ac901219m





http://dx.doi.org/10.1016/j.aca.2014.10.057


http://dx.doi.org/10.1021/acs.jpcc.5b02619


http://dx.doi.org/10.1016/j.vibspec.2005.08.002

http://dx.doi.org/10.1021/la063445v

http://dx.doi.org/10.1021/jp807807j


http://dx.doi.org/10.1016/j.jcis.2008.06.052



http://dx.doi.org/10.1021/ar900282h

http://dx.doi.org/10.1021/ar100019q

UNCORREC

TEDPR

OOF

1268 Casanova-González E, García-Bucio A, Ruvalcaba-Sil JL, Santos-Vasquez V, Esquivel B, Falcón1269 T, Arroyo E, Zetina S, Roldán ML, Domingo C (2012) Surface-enhanced Raman spectroscopy1270 spectra of Mexican dyestuffs. J Raman Spectrosc 43:1551–1559. doi:10.1002/jrs.40861271 Castro R, Pozzi F, Leona M, Melo MJ (2014) Combining SERS and microspectrofluorimetry with1272 historically accurate reconstructions for the characterization of lac dye paints in medieval1273 manuscript illuminations. J Raman Spectrosc 45:1172–1179. doi:10.1002/jrs.46081274 Centeno SA, Shamir J (2008) Surface enhanced Raman scattering (SERS) and FTIR character-1275 ization of the sepia melanin pigment used in works of art. J Mol Struct 873:149–159. doi:10.1276 1016/j.molstruc.2007.03.0261277 Centeno SA, Meller T, Kennedy N, Wypyski M (2008) The daguerreotype surface as a SERS1278 substrate: characterization of image deterioration in plates from the 19th century studio of1279 Southworth & Hawes. J Raman Spectrosc 39:914–921. doi:10.1002/jrs.19341280 Cesaratto A, Leona M, Lombardi JR, Comelli D, Nevin A, Londero P (2014) Detection of organic1281 colorants in historical painting layers using UV laser ablation surface-enhanced Raman1282 microspectroscopy. Angew Chem Int Ed 53:14373–14377. doi:10.1002/anie.2014080161283 Chang J, Cañamares MV, Aydin M, Vetter W, Schreiner M, Xu W, Lombardi JR (2009)1284 Surface-enhanced Raman spectroscopy of indanthrone and flavanthrone. J Raman Spectrosc1285 40:1557–1563. doi:10.1002/jrs.22981286 Chen K, Leona M, Vo-Dinh KC, Yan F, Wabuyele MB, Vo-Dinh T (2006) Application of1287 surface-enhanced Raman scattering (SERS) for the identification of anthraquinone dyes used in1288 works of art. J Raman Spectrosc 37:520–527. doi:10.1002/jrs.14261289 Chen K, Leona M, Vo-Dinh T (2007) Surface-enhanced Raman scattering for identification of1290 organic pigments and dyes in works of art and cultural heritage material. Sens Rev 27:109–1291 120. doi:10.1108/026022807107316781292 Claro A, Melo MJ, Schafer S, de Melo JSS, Pina F, van den Berg KJ, Burnstock A (2008) The use1293 of microspectrofluorimetry for the characterization of lake pigments. Talanta 74:922–929.1294 doi:10.1016/j.talanta.2007.07.0361295 Claro A, Melo MJ, Seixas de Melo JS, van den Berg KJ, Burnstock A, Montague M, Newman R1296 (2010) Identification of red colorants in van Gogh paintings and ancient Andean textiles by1297 microspectrofluorimetry. J Cult Herit 11:27–34. doi:10.1016/j.culher.2009.03.0061298 Clementi C, Nowik W, Romani A, Cibin F, Favaro G (2007) A spectrometric and chromato-1299 graphic approach to the study of ageing of madder (Rubia tinctorum L.) dyestuff on wool. Anal1300 Chim Acta 596:46–54. doi:10.1016/j.aca.2007.05.0361301 Clementi C, Basconi G, Pellegrino R, Romani A (2014) Carthamus tinctorius L.: a photophysical1302 study of the main coloured species for artwork diagnostic purposes. Dyes Pigm 103:127–137.1303 doi:10.1016/j.dyepig.2013.12.0021304 Collins D, Groom G, Ireson N, Keegan K, Shaw J, Nichols K (2014) Renoir paintings and1305 drawings at the Art Institute of Chicago. Art Institute of Chicago, Chicago, USA1306 Colombini MP, Andreotti A, Baraldi C, Degano I, Lucejko JJ (2007) Colour fading in textiles: a1307 model study on the decomposition of natural dyes. Microchem J 85:174–182. doi:10.1016/j.1308 microc.2006.04.0021309 Cooksey C, Sinclair RS (2005) Colour variations in tyrian purple dyeing. Dyes Hist Archaeol1310 20:1271311 Corredor C, Teslova T, Cañamares MV, Chen Z, Zhang J, Lombardi JR, Leona M (2009) Raman1312 and surface-enhanced Raman spectra of chrysin, apigenin and luteolin. Vib Spectrosc 49:190–1313 195. doi:10.1016/j.vibspec.2008.07.0121314 Creighton J, Blatchford C, Albrecht M (1979) Plasma resonance enhancement of Raman-scattering1315 by pyridine adsorbed on silver or gold sol particles of size comparable to the excitation1316 wavelength. J Chem Soc, Faraday Trans 75:790–798. doi:10.1039/f297975007901317 Daher C, Drieu L, Bellot-Gurlet L, Percot A, Paris C, Le Hô A-S (2014) Combined approach of1318 FT-Raman, SERS and IR micro-ATR spectroscopies to enlighten ancient technologies of1319 painted and varnished works of art. J Raman Spectrosc 45:1207–1214. doi:10.1002/jrs.4565

198 F. Pozzi et al.





http://dx.doi.org/10.1016/j.molstruc.2007.03.026

http://dx.doi.org/10.1016/j.molstruc.2007.03.026


http://dx.doi.org/10.1002/anie.201408016



http://dx.doi.org/10.1108/02602280710731678

http://dx.doi.org/10.1016/j.talanta.2007.07.036

http://dx.doi.org/10.1016/j.culher.2009.03.006


http://dx.doi.org/10.1016/j.dyepig.2013.12.002

http://dx.doi.org/10.1016/j.microc.2006.04.002



http://dx.doi.org/10.1039/f29797500790


UNCORREC

TEDPR

OOF

1320 de Oliveira LFC, Edwards HGM, Velozo ES, Nesbitt M (2002) Vibrational spectroscopic study of1321 brazilin and brazilein, the main constituents of brazilwood from Brazil. Vib Spectrosc 28:243–1322 249. doi:10.1016/S0924-2031(01)00138-21323 Degano I, Ribechini E, Modugno F, Colombini MP (2009) Analytical methods for the1324 characterization of organic dyes in artworks and in historical textiles. Appl Spectrosc Rev1325 44:363–410. doi:10.1080/057049209029378761326 Degano I, Biesaga M, Colombini MP, Trojanowicz M (2011) Historical and archaeological1327 textiles: An insight on degradation products of wool and silk yarns. J Chromatogr A1328 1218:5837–5847. doi:10.1016/j.chroma.2011.06.0951329 Delaney JK, Zeibel JG, Thoury M, Littleton R, Palmer M, Morales KM, de la Rie ER,1330 Hoenigswald A (2010) Visible and infrared imaging spectroscopy of Picasso’s Harlequin1331 musician: mapping and identification of artist materials in situ. Appl Spectrosc 64:584–5941332 Delaney JK, Ricciardi P, Glinsman LD, Facini M, Thoury M, Palmer M, de la Rie ER (2014) Use1333 of imaging spectroscopy, fiber optic reflectance spectroscopy, and X-ray fluorescence to map1334 and identify pigments in illuminated manuscripts. Stud Conserv 59:91–101. doi:10.1179/1335 2047058412Y.00000000781336 Dik J, Janssens K, Van der Snickt G, van der Loeff L, Rickers K, Cotte M (2008) Visualization of1337 a lost painting by Vincent van Gogh using synchrotron radiation based X-ray fluorescence1338 elemental mapping. Anal Chem 80:6436–6442. doi:10.1021/ac800965g1339 Doherty B, Brunetti BG, Sgamellotti A, Miliani C (2011) A detachable SERS active cellulose film:1340 a minimally invasive approach to the study of painting lakes. J Raman Spectrosc 42:1932–1341 1938. doi:10.1002/jrs.29421342 Doherty B, Gabrieli F, Clementi C, Cardon D, Sgamellotti A, Brunetti B, Miliani C (2014) Surface1343 enhanced Raman spectroscopic investigation of orchil dyed wool from Roccella tinctoria and1344 Lasallia pustulata. J Raman Spectrosc 45:723–729. doi:10.1002/jrs.45431345 Dooley KA, Lomax S, Zeibel JG, Miliani C, Ricciardi P, Hoenigswald A, Loew M, Delaney JK1346 (2013) Mapping of egg yolk and animal skin glue paint binders in Early Renaissance paintings1347 using near infrared reflectance imaging spectroscopy. Analyst 138:4838–4848. doi:10.1039/1348 c3an00926b1349 Edwards HGM, de Oliveira LFC, Nesbitt M (2003) Fourier-transform Raman characterization of1350 brazilwood trees and substitutes. Analyst 128:82–87. doi:10.1039/b209052j1351 El Bakkali A, Lamhasni T, Lyazidi SA, Haddad M, Rosi F, Miliani C, Sanchez-Cortes S, El1352 Rhaiti M (2014) Assessment of a multi-technical non-invasive approach for the typology of1353 inks, dyes and pigments in two 19th century’s ancient manuscripts of Morocco. Vib Spectrosc1354 74:47–56. doi:10.1016/j.vibspec.2014.07.0081355 Fan M, Andrade GFS, Brolo AG (2011) A review on the fabrication of substrates for surface1356 enhanced Raman spectroscopy and their applications in analytical chemistry. Anal Chim Acta1357 693:7–25. doi:10.1016/j.aca.2011.03.0021358 Farquharson S, Maksymiuk P (2003) Simultaneous chemical separation and surface-enhanced1359 Raman spectral detection using silver-doped sol-gels. Appl Spectrosc 57:479–482. doi:10.1360 1366/000370203606260411361 Ferreira ESB, Quye A, McNab H, Hulme AN, Wouters J, Boon JJ (2001) Development of1362 analytical techniques for the study of natural yellow dyes in historical textiles. Dyes Hist1363 Archaeol 16(17):179–1861364 Fleischmann M, Hendra P, Mcquilla AJ (1974) Raman-spectra of pyridine adsorbed at a silver1365 electrode. Chem Phys Lett 26:163–166. doi:10.1016/0009-2614(74)85388-11366 Frano KA, Mayhew HE, Svoboda SA, Wustholz KL (2014) Combined SERS and Raman analysis1367 for the identification of red pigments in cross-sections from historic oil paintings. Analyst1368 139:6450–6455. doi:10.1039/c4an01581a1369 Geiman I, Leona M, Lombardi JR (2009) Application of Raman spectroscopy and1370 surface-enhanced Raman scattering to the analysis of synthetic dyes found in ballpoint pen1371 inks. J Forensic Sci 54:947–952. doi:10.1111/j.1556-4029.2009.01058.x1372 Gillard RD, Hardman SM, Thomas RG, Watkinson DE (1994) The detection of dyes by FTIR1373 microscopy. Stud Conserv 39:187–192. doi:10.2307/1506597




http://dx.doi.org/10.1016/S0924-2031(01)00138-2

http://dx.doi.org/10.1080/05704920902937876

http://dx.doi.org/10.1016/j.chroma.2011.06.095

http://dx.doi.org/10.1179/2047058412Y.0000000078

http://dx.doi.org/10.1179/2047058412Y.0000000078

http://dx.doi.org/10.1021/ac800965g



http://dx.doi.org/10.1039/c3an00926b

http://dx.doi.org/10.1039/c3an00926b

http://dx.doi.org/10.1039/b209052j



http://dx.doi.org/10.1366/00037020360626041

http://dx.doi.org/10.1366/00037020360626041

http://dx.doi.org/10.1016/0009-2614(74)85388-1

http://dx.doi.org/10.1039/c4an01581a

http://dx.doi.org/10.1111/j.1556-4029.2009.01058.x

http://dx.doi.org/10.2307/1506597

UNCORREC

TEDPR

OOF

1374 Greeneltch NG, Davis AS, Valley NA, Casadio F, Schatz GC, Van Duyne RP, Shah NC (2012)1375 Near-infrared surface-enhanced Raman spectroscopy (NIR-SERS) for the identification of1376 Eosin Y: theoretical calculations and evaluation of two different nanoplasmonic substrates.1377 J Phys Chem A 116:11863–11869. doi:10.1021/jp30810351378 Grosjean D, Whitmore P, Demoor C, Cass G, Druzik J (1988) Ozone fading of organic colorants1379 —products and mechanism of the reaction of ozone with curcumin. Environ Sci Technol1380 22:1357–1361. doi:10.1021/es00176a0171381 Guineau B, Guichard V (1987) Identification des colorants organiques naturels par microspec-1382 trometrie Raman de resonance et par effet Raman exalte de surface (SERS). The Getty1383 Conservation Institute, Sydney, pp 659–6661384 Gulmini M, Idone A, Diana E, Gastaldi D, Vaudan D, Aceto M (2013) Identification of dyestuffs1385 in historical textiles: Strong and weak points of a non-invasive approach. Dyes Pigm 98:136–1386 145. doi:10.1016/j.dyepig.2013.02.0101387 Halpine SM (1996) An improved dye and lake pigment analysis method for high-performance1388 liquid chromatography and diode-array detector. Stud Conserv 41:76–94. doi:10.2307/1389 15065191390 Henzel (1977) Journal of chromatography library. Elsevier, Amsterdam1391 Hofenk de Graaff JH, Roelofs WGT, Van Bommel MR (2004) The colourful past: origins,1392 chemistry and identification of natural dyestuffs. Archetype Publications, London ;1393 Abegg-Stiftung, Riggisberg, Switzerland1394 Idone A, Gulmini M, Henry A-I, Casadio F, Chang L, Appolonia L, Van Duyne RP, Shah NC1395 (2013) Silver colloidal pastes for dye analysis of reference and historical textile fibers using1396 direct, extractionless, non-hydrolysis surface-enhanced Raman spectroscopy. Analyst1397 138:5895–5903. doi:10.1039/c3an00788j1398 Idone A, Aceto M, Diana E, Appolonia L, Gulmini M (2014) Surface-enhanced Raman scattering1399 for the analysis of red lake pigments in painting layers mounted in cross sections. J Raman1400 Spectrosc 45:1127–1132. doi:10.1002/jrs.44911401 Jeanmaire D, Vanduyne R (1977) Surface Raman spectroelectrochemistry. 1. Heterocyclic,1402 aromatic, and aliphatic-amines adsorbed on anodized silver electrode. J Electroanal Chem1403 84:1–20. doi:10.1016/S0022-0728(77)80224-61404 Jurasekova Z, Garcia-Ramos JV, Domingo C, Sanchez-Cortes S (2006) Surface-enhanced Raman1405 scattering of flavonoids. J Raman Spectrosc 37:1239–1241. doi:10.1002/jrs.16341406 Jurasekova Z, Domingo C, Garcia-Ramos JV, Sanchez-Cortes S (2008) In situ detection of1407 flavonoids in weld-dyed wool and silk textiles by surface-enhanced Raman scattering. J Raman1408 Spectrosc 39:1309–1312. doi:10.1002/jrs.20531409 Jurasekova Z, del Puerto E, Bruno G, Garcia-Ramos JV, Sanchez-Cortes S, Domingo C (2010)1410 Extractionless non-hydrolysis surface-enhanced Raman spectroscopic detection of historical1411 mordant dyes on textile fibers. J Raman Spectrosc 41:1455–1461. doi:10.1002/jrs.26511412 Jurasekova Z, Domingo C, Garcia-Ramos JV, Sanchez-Cortes S (2012) Adsorption and catalysis1413 of flavonoid quercetin on different plasmonic metal nanoparticles monitored by SERS. J Raman1414 Spectrosc 43:1913–1919. doi:10.1002/jrs.41141415 Kim D, Campos AR, Datt A, Gao Z, Rycenga M, Burrows ND, Greeneltch NG, Mirkin CA,1416 Murphy CJ, Van Duyne RP, Haynes CL (2014) Microfluidic-SERS devices for one shot1417 limit-of-detection. Analyst 139:3227–3234. doi:10.1039/c4an00357h1418 Koperska M, Lojewski T, Lojewska J (2011) Vibrational spectroscopy to study degradation of1419 natural dyes. Assessment of oxygen-free cassette for safe exposition of artefacts. Anal Bioanal1420 Chem 399:3271–3283. doi:10.1007/s00216-010-4460-71421 Koren Z (1994) HPLC analysis of the natural scale insect, madder and indigoid dyes. J Soc Dye1422 Colour 110:273–277. doi: 10.1111/j.1478-4408.1994.tb01656.x1423 Kurouski D, Zaleski S, Casadio F, Van Duyne RP, Shah NC (2014) Tip-enhanced Raman1424 spectroscopy (TERS) for in situ identification of indigo and iron gall ink on paper. J Am Chem1425 Soc 136:8677–8684. doi:10.1021/ja50276121426 Lee PC, Meisel D (1982) Adsorption and surface-enhanced Raman of dyes on silver and gold sols.1427 J Phys Chem 86:3391–3395. doi:10.1021/j100214a025

200 F. Pozzi et al.



http://dx.doi.org/10.1021/jp3081035

http://dx.doi.org/10.1021/es00176a017

http://dx.doi.org/10.1016/j.dyepig.2013.02.010

http://dx.doi.org/10.2307/1506519

http://dx.doi.org/10.2307/1506519

http://dx.doi.org/10.1039/c3an00788j


http://dx.doi.org/10.1016/S0022-0728(77)80224-6





http://dx.doi.org/10.1039/c4an00357h

http://dx.doi.org/10.1007/s00216-010-4460-7

http://dx.doi.org/10.1111/j.1478-4408.1994.tb01656.x

http://dx.doi.org/10.1021/ja5027612

http://dx.doi.org/10.1021/j100214a025

UNCORREC

TEDPR

OOF

1428 Leona M (2008) Non-invasive identification of fluorescent dyes in historic textiles by matrix1429 transfer-surface enhanced Raman scattering. United States Patent and Trademark Office;1430 7,362,431 B2. Filed 9 Aug 2006; awarded 22 Apr 20081431 Leona M (2009) Microanalysis of organic pigments and glazes in polychrome works of art by1432 surface-enhanced resonance Raman scattering. Proc Natl Acad Sci USA 106:14757–14762.1433 doi:10.1073/pnas.09069951061434 Leona M, Lombardi JR (2007) Identification of berberine in ancient and historical textiles by1435 surface-enhanced Raman scattering. J Raman Spectrosc 38:853–858. doi:10.1002/jrs.17261436 Leona M, Tague T (2010) Method and apparatus for in situ measurement of material properties by1437 surface enhanced raman spectroscopy. United States Patent and Trademark Office; US1438 7,787,117 B1. Filed 24 June 2008; awarded 31 Aug 20101439 Leona M, Stenger J, Ferloni E (2006) Application of surface-enhanced Raman scattering1440 techniques to the ultrasensitive identification of natural dyes in works of art. J Raman Spectrosc1441 37:981–992. doi:10.1002/jrs.15821442 Leona M, Decuzzi P, Kubic TA, Gates G, Lombardi JR (2011) Nondestructive identification of1443 natural and synthetic organic colorants in works of art by surface enhanced Raman scattering.1444 Anal Chem 83:3990–3993. doi:10.1021/ac20070151445 Leona M, Smith HD II, Cesaratto A, Luo YB (2015) Synthetic dyes in the woodblock prints of1446 Meiji Japan. In: Technart 2015, non-destructive and microanalytical techniques in art and1447 cultural heritage, Catania, 27–30 Apr 20151448 Leopold N, Lendl B (2003) A new method for fast preparation of highly surface-enhanced Raman1449 scattering (SERS) active silver colloids at room temperature by reduction of silver nitrate with1450 hydroxylamine hydrochloride. J Phys Chem B 107:5723–5727. doi:10.1021/jp027460u1451 Lewis PA (1988) Organic pigments. Federation of Societies for Coatings Technology,1452 Philadelphia, PA, USA (1315 Walnut St., Philadelphia 19107)1453 Lofrumento C, Ricci M, Platania E, Becucci M, Castellucci E (2013) SERS detection of red1454 organic dyes in Ag-agar gel. J Raman Spectrosc 44:47–54. doi:10.1002/jrs.41621455 Lombardi JR, Birke RL (2009) A unified view of surface-enhanced Raman scattering. Acc Chem1456 Res 42:734–742. doi:10.1021/ar800249y1457 Lombardi JR, Birke RL (2012) The theory of surface-enhanced Raman scattering. J Chem Phys1458 136:144704. doi:10.1063/1.36982921459 Londero PS, Lombardi JR, Leona M (2013) Laser ablation surface-enhanced Raman microspec-1460 troscopy. Anal Chem 85:5463–5467. doi:10.1021/ac400440c1461 Luo Z, Smith JC, Goff TM, Adair JH, Castleman AW (2013) Gold cluster coatings enhancing1462 Raman scattering from surfaces: Ink analysis and document identification. Chem Phys 423:73–1463 78. doi:10.1016/j.chemphys.2013.06.0211464 Manhita A, Ferreira V, Vargas H, Ribeiro I, Candeias A, Teixeira D, Ferreira T, Dias CB (2011)1465 Enlightening the influence of mordant, dyeing technique and photodegradation on the colour1466 hue of textiles dyed with madder—a chromatographic and spectrometric approach.1467 Microchem J 98:82–90. doi:10.1016/j.microc.2010.12.0021468 Manhita A, Santos V, Vargas H, Candeias A, Ferreira T, Dias CB (2013) Ageing of brazilwood1469 dye in wool—a chromatographic and spectrometric study. J Cult Herit 14:471–479. doi:10.1470 1016/j.culher.2012.10.0161471 Masschelein-Kleiner L, Heylen JB (1968) Analyse des laques rouges anciennes. Stud Conserv1472 13:87–971473 Mayhew HE, Fabian DM, Svoboda SA, Wustholz KL (2013) Surface-enhanced Raman1474 spectroscopy studies of yellow organic dyestuffs and lake pigments in oil paint. Analyst1475 138:4493–4499. doi:10.1039/C3AN00611E1476 Melo MJ, Claro A (2010) Bright light: microspectrofluorimetry for the characterization of lake1477 pigments and dyes in works of art. Acc Chem Res 43:857–866. doi:10.1021/ar90018941478 Mills JS, White R (1987) The organic chemistry of museum objects. Butterworths, London1479 Montagner C, Bacci M, Bracci S, Freeman R, Picollo M (2011) Library of UV-Vis-NIR1480 reflectance spectra of modern organic dyes from historic pattern-card coloured papers.1481 Spectrochim Acta Part A Mol Biomol Spectrosc 79:1669–1680. doi:10.1016/j.saa.2011.05.033




http://dx.doi.org/10.1073/pnas.0906995106



http://dx.doi.org/10.1021/ac2007015

http://dx.doi.org/10.1021/jp027460u


http://dx.doi.org/10.1021/ar800249y

http://dx.doi.org/10.1063/1.3698292

http://dx.doi.org/10.1021/ac400440c

http://dx.doi.org/10.1016/j.chemphys.2013.06.021




http://dx.doi.org/10.1039/C3AN00611E

http://dx.doi.org/10.1021/ar9001894


UNCORREC

TEDPR

OOF

1482 Oakley LH, Dinehart SA, Svoboda SA, Wustholz KL (2011) Identification of organic materials in1483 historic oil paintings using correlated extractionless surface-enhanced Raman scattering and1484 fluorescence microscopy. Anal Chem 83:3986–3989. doi:10.1021/ac200698q1485 Oakley LH, Fabian DM, Mayhew HE, Svoboda SA, Wustholz KL (2012) Pretreatment strategies1486 for sers analysis of indigo and prussian blue in aged painted surfaces. Anal Chem 84:8006–1487 8012. doi:10.1021/ac301814e1488 Platania E, Lombardi JR, Leona M, Shibayama N, Lofrumento C, Ricci M, Becucci M,1489 Castellucci E (2014) Suitability of Ag-agar gel for the micro-extraction of organic dyes on1490 different substrates: the case study of wool, silk, printed cotton and a panel painting mock-up.1491 J Raman Spectrosc 45:1133–1139. doi:10.1002/jrs.45311492 Platania E, Lofrumento C, Lottini E, Azzaro E, Ricci M, Becucci M (2015) Tailored1493 micro-extraction method for Raman/SERS detection of indigoids in ancient textiles. Anal1494 Bioanal Chem 407:6505–6514. doi:10.1007/s00216-015-8816-x1495 Pozzi F, Lombardi JR, Bruni S, Leona M (2012a) Sample treatment considerations in the analysis1496 of organic colorants by surface-enhanced Raman scattering. Anal Chem 84:3751–3757. doi:10.1497 1021/ac300380c1498 Pozzi F, Poldi G, Bruni S, De Luca E, Guglielmi V (2012b) Multi-technique characterization of1499 dyes in ancient Kaitag textiles from Caucasus. Archaeol Anthropol Sci 4:185–197. doi:10.1500 1007/s12520-012-0092-51501 Pozzi F, Shibayama N, Leona M, Lombardi JR (2013a) TLC-SERS study of Syrian rue (Peganum1502 harmala) and its main alkaloid constituents. J Raman Spectrosc 44:102–107. doi:10.1002/jrs.1503 41401504 Pozzi F, Porcinai S, Lombardi JR, Leona M (2013b) Statistical methods and library search1505 approaches for fast and reliable identification of dyes using surface-enhanced Raman1506 spectroscopy (SERS). Anal Methods 5:4205–4212. doi:10.1039/C3AY40673C1507 Pozzi F, Lombardi JR, Leona M (2013c) Winsor & Newton original handbooks: a1508 surface-enhanced Raman scattering (SERS) and Raman spectral database of dyes from1509 modern watercolor pigments. Heritage Sci 1(23):1–8. doi:10.1186/2050-7445-1-231510 Pozzi F, van den Berg KJ, Fiedler I, Casadio F (2014a) A systematic analysis of red lake pigments1511 in French Impressionist and Post-Impressionist paintings by surface-enhanced Raman1512 spectroscopy (SERS). J Raman Spectrosc 45(11–12):1119–1126. doi:10.1002/jrs.44831513 Pozzi F, Chang LK, Casadio F (2014b) The Navajo blankets from the Art Institute of Chicago1514 collection: technical analysis of yarn and weavings coupled with dye identification by normal1515 Raman and surface-enhanced Raman spectroscopy (SERS). In: BridglandJ (ed) ICOM-CC1516 17th triennial conference preprints, Melbourne, 15–19 Sept 2014, art. 1808, 8pp. International1517 Council of Museums, Paris (ISBN 978-92-9012-410-8)1518 Prikhodko SV, Rambaldi DC, King A, Burr E, Muros V, Kakoulli I (2015) New advancements in1519 SERS dye detection using interfaced SEM and Raman spectromicroscopy (μRS). J Raman1520 Spectrosc 46:632–635. doi:10.1002/jrs.47101521 Rambaldi DC, Pozzi F, Shibayama N, Leona M, Preusser FD (2015) Surface-enhanced Raman1522 spectroscopy of various madder species on wool fibers: the role of pseudopurpurin in the1523 interpretation of the spectra. J Raman Spectrosc 46(11):1073–1081. doi:10.1002/jrs.47261524 Ramesova S, Sokolova R, Degano I, Bulickova J, Zabka J, Gal M (2012) On the stability of the1525 bioactive flavonoids quercetin and luteolin under oxygen-free conditions. Anal Bioanal Chem1526 402:975–982. doi:10.1007/s00216-011-5504-31527 Raza A, Saha B (2013) Silver nanoparticles doped agarose disk: highly sensitive surface-enhanced1528 Raman scattering substrate for in situ analysis of ink dyes. Forensic Sci Int 233:21–27. doi:10.1529 1016/j.forsciint.2013.08.0041530 Retko K, Ropret P, Korosec RC (2014) Surface-enhanced Raman spectroscopy (SERS) analysis of1531 organic colourants utilising a new UV-photoreduced substrate. J Raman Spectrosc 45:1140–1532 1146. doi:10.1002/jrs.45331533 Ricciardi P, Delaney JK, Facini M, Zeibel JG, Picollo M, Lomax S, Loew M (2012) Near infrared1534 reflectance imaging spectroscopy to map paint binders in situ on illuminated manuscripts.1535 Angew Chem Int Ed 51:5607–5610. doi:10.1002/anie.201200840

202 F. Pozzi et al.



http://dx.doi.org/10.1021/ac200698q

http://dx.doi.org/10.1021/ac301814e


http://dx.doi.org/10.1007/s00216-015-8816-x



http://dx.doi.org/10.1007/s12520-012-0092-5

http://dx.doi.org/10.1007/s12520-012-0092-5



http://dx.doi.org/10.1039/C3AY40673C

http://dx.doi.org/10.1186/2050-7445-1-23




http://dx.doi.org/10.1007/s00216-011-5504-3





UNCORREC

TEDPR

OOF

1536 Ringe E, Sharma B, Henry A-I, Marks LD, Van Duyne RP (2013) Single nanoparticle plasmonics.1537 Phys Chem Chem Phys 15:4110–4129. doi:10.1039/c3cp44574g1538 Rodger C, Dent G, Watkinson J, Smith WE (2000) Surface-enhanced resonance Raman scattering1539 and near-infrared Fourier transform Raman scattering as in situ probes of ink jet dyes printed1540 on paper. Appl Spectrosc 54:1567–1576. doi:10.1366/00037020019488171541 Roldán ML, Centeno SA, Rizzo A (2014) An improved methodology for the characterization and1542 identification of sepia in works of art by normal Raman and SERS, complemented by FTIR,1543 Py-GC/MS, and XRF. J Raman Spectrosc 45:1160–1171. doi:10.1002/jrs.46201544 Romani A, Clementi C, Miliani C, Favaro G (2010) Fluorescence spectroscopy: a powerful1545 technique for the noninvasive characterization of artwork. Acc Chem Res 43:837–846. doi:10.1546 1021/ar900291y1547 Rosi F, Miliani C, Braun R, Harig R, Sali D, Brunetti BG, Sgamellotti A (2013) Noninvasive1548 analysis of paintings by mid-infrared hyperspectral imaging. Angew Chem Int Ed 52:5258–1549 5261. doi:10.1002/anie.2012099291550 Saunders D, Kirby J (1994) Light-induced colour changes in red and yellow lake pigments. Natl1551 Gallery Tech Bull 15:79–971552 Scherrer NC, Stefan Z, Francoise D, Annette F, Renate K (2009) Synthetic organic pigments of the1553 20th and 21st century relevant to artist’s paints: Raman spectra reference collection.1554 Spectrochim Acta Part A Mol Biomol Spectrosc 73:505–524. doi:10.1016/j.saa.2008.11.0291555 Schmid T, Opilik L, Blum C, Zenobi R (2013) Nanoscale chemical imaging using tip-enhanced1556 Raman spectroscopy: a critical review. Angew Chem-Int Ed 52:5940–5954. doi:10.1002/anie.1557 2012038491558 Schweppe H (1993) Handbuch der Naturfarbstoffe: Vorkommen, Verwendung, Nachweis.1559 Nikol-Verl.-Ges, Hamburg1560 Sciutto G, Prati S, Bonacini I, Litti L, Meneghetti M, Mazzeo R (2015) In Technart 2015,1561 non-destructive and microanalytical techniques in art and cultural heritage, Catania, 27–30 Apr1562 20151563 Seifar RM, Verheul JM, Ariese F, Brinkman Ua T, Gooijer C (2001) Applicability of1564 surface-enhanced resonance Raman scattering for the direct discrimination of ballpoint pen1565 inks. Analyst 126:1418–1422. doi:10.1039/b103042f1566 Sgamellotti A, Brunetti B, Miliani C (2014) Science and art: the painted surface. The Royal1567 Society of Chemistry, Cambridge1568 Shadi QT, Chowdhry BZ, Snowden MJ, Withnall R (2004) Semi-quantitative analysis of alizarin1569 and purpurin by surface-enhanced resonance Raman spectroscopy (SERRS) using silver1570 colloids. J Raman Spectrosc 35:800–807. doi:10.1002/jrs.11991571 Sharma B, Cardinal MF, Kleinman SL, Greeneltch NG, Frontiera RR, Blaber MG, Schatz GC,1572 Van Duyne RP (2013) High-performance SERS substrates: advances and challenges. MRS1573 Bull 38:615–624. doi:10.1557/mrs.2013.1611574 Taylor GW (1983) Detection and identiflcation of dyes on Anglo-Scandinavian textiles. Stud1575 Conserv 28:153–1601576 Teslova T, Corredor C, Livingstone R, Spataru T, Birke RL, Lombardi JR, Cañamares MV,1577 Leona M (2007) Raman and surface-enhanced Raman spectra of flavone and several hydroxy1578 derivatives. J Raman Spectrosc 38:802–818. doi:10.1002/jrs.16951579 Tiedemann EJ, Yang Y (1995) Fiber-safe extraction of red mordant dyes from hair fibers. J Am1580 Inst Conserv 34:195–206. doi:10.1179/0197136958061246571581 van Bommel MR, Berghe IV, Wallert AM, Boitelle R, Wouters J (2007) High-performance liquid1582 chromatography and non-destructive three-dimensional fluorescence analysis of early synthetic1583 dyes. J Chromatogr A 1157:260–272. doi:10.1016/j.chroma.2007.05.0171584 Van Elslande E, Lecomte S, Le Ho A-S (2008) Micro-Raman spectroscopy (MRS) and1585 surface-enhanced Raman scattering (SERS) on organic colourants in archaeological pigments.1586 J Raman Spectrosc 39:1001–1006. doi:10.1002/jrs.19941587 Vellekoop M, Bakker N, van Dijk M, Geldof M, Hendriks E, Reissland B, Holberton P, Alkins T,1588 Hoyle M, Jackson B, van Gogh V, Van Gogh Museum A (2013) Van Gogh at work




http://dx.doi.org/10.1039/c3cp44574g

http://dx.doi.org/10.1366/0003702001948817








http://dx.doi.org/10.1039/b103042f


http://dx.doi.org/10.1557/mrs.2013.161


http://dx.doi.org/10.1179/019713695806124657

http://dx.doi.org/10.1016/j.chroma.2007.05.017


UNCORREC

TEDPR

OOF

1589 Wagner E, Clement S (2001) Surface enhanced resonance Raman scattering (Serrs) spectroscopy1590 —study on inks. Probl Forensic Sci 46:437–4411591 Wang M, Teslova T, Xu F, Spataru T, Lombardi JR, Birke RL, Leona M (2007) Raman and1592 surface enhanced Raman scattering of 3-hydroxyflavone. J Phys Chem C 111:3038–3043.1593 doi:10.1021/jp062100i1594 White PC (2003) In situ surface enhanced resonance Raman scattering (SERRS) spectroscopy of1595 biro inks—long term stability of colloid treated samples. Sci Justice 43:149–152. doi:10.1016/1596 S1355-0306(03)71762-61597 Whitney AV, Van Duyne RP, Casadio F (2006) An innovative surface-enhanced Raman1598 spectroscopy (SERS) method for the identification of six historical red lakes and dyestuffs.1599 J Raman Spectrosc 37:993–1002. doi:10.1002/jrs.15761600 Whitney AV, Casadio F, Van Duyne RP (2007) Identification and characterization of artists’ red1601 dyes and their mixtures by surface-enhanced Raman spectroscopy. Appl Spectrosc 61:994–1602 1000. doi:0003-7028/07/6109-0994$2.00/01603 Wouters J (1985) High performance liquid chromatography of anthraquinones: analysis of plant1604 and insect extracts and dyed textiles. Stud Conserv 30:119–128. doi:10.2307/15059271605 Wouters J, Verhecken A (1989) The coccid insect dyes: HPLC and computerized diode-array1606 analysis of dyed yarns. Stud Conserv 34:189–200. doi:10.2307/15062861607 Wustholz KL, Brosseau CL, Casadio F, Van Duyne RP (2009) Surface-enhanced Raman1608 spectroscopy of dyes: from single molecules to the artists’ canvas. Phys Chem Chem Phys1609 11:7350–7359. doi:10.1039/b904733f1610 Xie Y, Li Y, Sun Y, Wang H, Qian H, Yao W (2012) Theoretical calculation (DFT), Raman and1611 surface-enhanced Raman scattering (SERS) study of ponceau 4R. Spectrochim Acta Part A1612 Mol Biomol Spectrosc 96:600–604. doi:10.1016/j.saa.2012.06.0551613 Yoshizumi K, Crews PC (2003) Characteristics of fading of wool cloth dyed with selected natural1614 dyestuffs on the basis of solar radiant energy. Dyes Pigm 58:197–204. doi:10.1016/S0143-1615 7208(03)00065-21616 Zaffino C, Bruni S, Guglielmi V, De Luca E (2014) Fourier-transform surface-enhanced Raman1617 spectroscopy (FT-SERS) applied to the identification of natural dyes in textile fibers: an1618 extractionless approach to the analysis. J Raman Spectrosc 45:211–218. doi:10.1002/jrs.44431619 Zaffino C, Russo B, Bruni S (2015) Surface-enhanced Raman scattering (SERS) study of1620 anthocyanidins. Spectrochim Acta A Mol Biomol Spectrosc 149:41–47. doi:10.1016/j.saa.1621 2015.04.039

204 F. Pozzi et al.



http://dx.doi.org/10.1021/jp062100i

http://dx.doi.org/10.1016/S1355-0306(03)71762-6

http://dx.doi.org/10.1016/S1355-0306(03)71762-6


http://sites.northwestern.edu/vanduyne/files/2012/10/2007_Whitney.pdf

http://dx.doi.org/10.2307/1505927

http://dx.doi.org/10.2307/1506286

http://dx.doi.org/10.1039/b904733f


http://dx.doi.org/10.1016/S0143-7208(03)00065-2

http://dx.doi.org/10.1016/S0143-7208(03)00065-2




surface-enhanced raman spectroscopy: using nanoparticles ... · u ncorrect ed 1 proof...

Documents