- Email: [email protected]
Identification of inorganic dyeing mordant in textiles by surface-enhanced laser-induced breakdown spectroscopy
Microchemical Journal 139 (2018) 230–235
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Identification of inorganic dyeing mordant in textiles by surfaceenhanced laser-induced breakdown spectroscopy Beatrice Campanella a, Ilaria Degano b, Emanuela Grifoni a,c, Stefano Legnaioli a,c,⁎, Giulia Lorenzetti a, Stefano Pagnotta a, Francesco Poggialini a, Vincenzo Palleschi a,c a b c
Applied and Laser Spectroscopy Laboratory, Institute of Chemistry of Organometallic Compounds, Research Area of CNR, Via Giuseppe Moruzzi, 1, 56124 Pisa, Italy Department of Chemistry and Industrial Chemistry, University of Pisa, Via Giuseppe Moruzzi, 13, 56124 Pisa, Italy National Interuniversity Consortium of Materials Science and Technology (INSTM), Italy
a r t i c l e
i n f o
Article history: Received 28 August 2017 Received in revised form 31 January 2018 Accepted 27 February 2018 Available online 2 March 2018 Keywords: Textiles Mordant Surface-enhanced LIBS HPLC-DAD
a b s t r a c t The identification of both the organic and inorganic fraction of dyes and pigments is fundamental for their complete characterization and to assess the technologies used in their production. In this work, the feasibility of determining metallic elements used as mordant for dyed textiles was tested using Laser-Induced Breakdown Spectrometry (LIBS) in combination with liquid micro-extraction. Both reference laboratory-dyed and historic textiles were analysed in this study. Samples were first analysed without any preparation. Then, the chromophores-containing molecules were separated using a sample preparation procedure based on aqueous hydrolysis, and analysed by high-pressure liquid chromatography coupled with a diode array detector. The same extracts, containing also the inorganic fraction, were analysed by Surface-Enhanced Laser-Induced Breakdown Spectroscopy (SENLIBS) after drying on a solid substrate. Compared to the direct analysis, the SENLIBS method improved the sensitivity of the measurements. The procedure presented here allowed for the characterization of both organic and inorganic fraction of a single textile micro sample, thus avoiding further sampling. © 2018 Elsevier B.V. All rights reserved.
1. Introduction Until 150 years ago, the dyes for the textiles industry were obtained from natural sources. The dyeing process requires that the chromophores, obtained from the natural sources, are bonded as strongly as possible to the fibers. For the binding process, dyeing with the vast majority of natural dyes entails the treatment of the textile fibers with a solution of inorganic-based constituents, called mordants. The mordant is complexed by polar groups on the surface of protein fibers, and during the dyeing process the dye interacts with the mordant–fiber complex to form an insoluble brightly coloured species [1]. Alum (K2O4Al2(SO4) 3), iron (FeSO4·7H2O), and copper (CuSO4·5H2O) were the most commonly used mordants in antiquity, while tin (SnCl2·2H2O) and chrome (K2Cr2O7 or Na2Cr2O7) were introduced in the 18th century. The mordant ensures the brightness and wash-fastness of the dye, and it has great influence on the final colour obtained. For different cultures and periods, different mordanting systems and procedures were employed to produce a wide variety of hues. Determining which ⁎ Corresponding author at: Applied and Laser Spectroscopy Laboratory, Institute of Chemistry of Organometallic Compounds, Research Area of CNR, Via Giuseppe Moruzzi, 1, 56124 Pisa, Italy. E-mail address: [email protected] (S. Legnaioli).
https://doi.org/10.1016/j.microc.2018.02.034 0026-265X/© 2018 Elsevier B.V. All rights reserved.
substance was used to mordant or dye archeological or historical textiles provides a better understanding of the ancient mordanting or dying technology and assists in the dating and authentication of a specimen. Furthermore, the source of metal cations and dye/cation ratio, together with the dye and the substrate, are crucial parameters affecting the light-fastness and chemical features of the textiles [2,3]. Several methods have been developed to analyse dyes in textiles matrices in the last twenty years, exploiting different analytical techniques and testing different sample pre-treatment procedures. Most of them are based on High Pressure Liquid Chromatography coupled with Diode-Array, Fluorescence or Mass Spectrometric Detectors (HPLC-DAD-FD/MS) [4–12]. 3D Fluorescence [13,14], Surface Enhanced Resonance Raman Scattering (SERS) [15–17] and Matrix Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS and LDI-MS) are also employed [18,19]. Together with the identification of the organic dye, the analysis of the elemental composition of the inorganic fraction of dyed yarns can provide useful information, such as a relative dating. For example, developments in the use of different mordants for textile dyeing, notably the use of tin salts with cochineal dyestuff to give scarlet, were not taken up by the dyers until the very end of the 18th century, or beginning of the 19th century [20]. The first known commercial use for chromium began in the late 1700s as a paint pigment, and as early as 1820
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
the textile industry was using large amounts of chromium compounds as mordants [21]. However, literature on the characterization of inorganic mordants is still relatively poor [22–26]. The use of scanning electron microscopy (SEM) along with energy dispersive X-ray spectrometry (EDS) appears to be suitable for the identification of metallic mordants on fibers, whenever such instrumentation is available; however, most SEM-EDS applications (except Variable Pressure SEM) require the samples to be either conductive or coated in gold, graphite or another conductive material. Together with SEM-EDS, also energy-dispersive X-ray fluorescence (ED-XRF) is a common elemental technique for the mapping of mordant metals. Both techniques, however, are limited by low sensitivity (XRF is unsuitable for the analysis of light elements such as aluminum). The basic idea of this work is to demonstrate the feasibility of analysing by Laser-Induced Breakdown Spectroscopy (LIBS) a small part of the aqueous fraction from the micro-extraction aimed at dye characterization, for trace elements identification [27,28]. LIBS is a multi-elemental analytical technique that does not rely on sample treatment, as inductively-coupled plasma–optical emission spectroscopy, it does not require the samples to be conducting, as spark-OES does, and is not limited to the analysis of heavy elements, as X-ray fluorescence is. The required ablated mass is about few ng. Laser induced plasma formation on or inside liquids is characterized by large energy losses due to liquid evaporation [29]. This effect dramatically reduces the limits of detection of the technique. One of the best strategy in this cases is to deposit the liquid sample on a proper substrate and wait for a completely evaporation (SENSLIBS, Surface-Enhanced LIBS) [30]. In this work, a double approach has been developed to characterize reference dyed woolen and historic textiles, containing different metal cations. A previously developed extraction procedure based on EDTA/ DMF was applied to separate the mordant from the fiber, thus releasing the dyestuff in the extract. The extract was analysed both by HPLC-DAD, to identify the chromophore-containing molecule, and by SENLIBS to determine the mordant. 2. Materials and methods 2.1. Reagents and samples Disodium EDTA (Sigma-Aldrich, USA), acetonitrile (HPLC grade, Sigma-Aldrich, USA), ethyl acetate (AcOEt; Carlo Erba, Italy) and dimethylformamide (DMF, 99.8%, J.T. Baker, Holland) were used for sample pre-treatment and HPLC-DAD analyses. PTFE filters (4 mm thickness and 0.45 μm pore diameter, Sigma Aldrich) were chosen for the filtration. Ultrapure water prepared with an Elga Purelab-UV system (Veolia Environment, France) was used throughout. Glass slides (76 mm × 26 mm, 1.1 mm thickness microscope slides, from Sigma-Aldrich), Teflon sheets (fiberglass coated with PTFE), and
Table 1 Description of the references and the historical samples collected from Giuseppe fugge dalla moglie di Putifarre. Sample
Mordant Dye
References Wool Cu Wool Silk Historical samples I Ro1⁎ I G1⁎ I N1⁎ III Be2⁎
Fe Al
European buckthorn (“giallo di Spincervino”, i.e. Rhamnus cathartica) Gall nut Cochineal
⁎ Unknown dye and mordant.
Colour
Sample weight (mg)
silicon bulk wafers (single side polished, without dopant, 2 in. diameter and 0.5 mm thickness, from Sigma Aldrich) were tested as substrate materials for SENLIBS analysis. Two references raw wool (Italian commercial raw wool, 4/10 “pura lana”, cream colour from Campolmi, Italy) and a reference raw silk yarn (“seta schappe”, provided by Opificio delle Pietre Dure) were mordanted and dyed in laboratory. The detailed description of the raw materials and the treatment process is reported in Degano et al. [31]. Four historical samples from the tapestry “Giuseppe fugge dalla moglie di Putifarre” (Florence, Salone dei 200, Palazzo Vecchio), kindly provided by the Opificio delle Pietre Dure of Florence, were studied. The Italian artist Angnolo Bronzino designed this tapestry's cartoon around the half of XVI century, and the weaving is due to Karcher's atelier [32]. References and samples are listed in Table 1. 2.2. Sample preparation for HPLC and SENLIBS analysis The samples collected from reference specimens and the historical tapestry (see Table 1) were suspended in 400 μL and 200 μL, respectively, of 0.1% EDTA/DMF solution (1:1), and sonicated for 1 h at 60 °C. The extract in EDTA/DMF was filtrated through PTFE filters and then injected in the HPLC-DAD system. For the liquid-liquid extraction experiment, the extract in EDTA/DMF was diluted with 200 μL of water and extracted three times with 200 μL of ethylacetate. For SENLIBS analysis, the extract in EDTA/DMF or the aqueous residue from liquid/liquid extraction was deposited on the substrate and evaporated to dryness. Mild warming with a hot plate was used to facilitate the process. Upon dry out, the substrate was cooled to room temperature and the surface analysed by LIBS. 2.3. Instrumentation 2.3.1. HPLC-DAD Chromatographic separation was achieved using an analytical reverse-phase TC-C18 (2) column (4.6 × 150 mm, particle size 5 μm, Agilent) with a TC-C18 (2) precolumn (4.6 × 12.5 mm, particle size 5 μm, Agilent). The HPLC system consists of a PU-2089 quaternary pump (Jasco International Co., Japan) equipped with a degasser, an AS-950
Table 2 Results of the HPLC-DAD analysis (complete chromatograms are reported in the Supporting Information). Sample
Retention time (min)
Compound
Oak gall dyed wool, FeSO4 mordant
11.5 3.8 13.2
American cochineal dyed wool, alum mordant
9.2 10.2 13.5 15.3 20.6 21.2 18.9 22.1 25.2 29.2 32.2 11.6 4.2 11.5 18.9 21.6 22.2 10.1 9.1
Ellagic acid Gallic acid Gallic acid-like compound dcII Carminc acid dcIV dcVII Kermesic acid Flavokermesic acid Quercetin Kaempferol Isorhamnetin Rhamnetin Emodin Ellagic acid Gallic acid Ellagic acid Luteolin Apigenin Chrysoeriol Carminic acid dcII
Rhamnus, copper mordant
Yellow 6.0 Brown Red
5.4 5.0
Red Yellow Black Beige
1.4 1.6 0.7 0.4
231
I N1 III Be2 I G1
I Ro1
232
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
Fig. 1. Comparison of LIBS signal from a solution containing 250 mg/L of Cu, Al, Sn and Fe, deposited on Teflon® (black line), glass (red line) and silicon wafer (blue line).
autosampler (Jasco International Co., Japan) and an MD-2010 spectrophotometric diode array detector (DAD) (Jasco International Co., Japan). Data were processed with ChromNav software. The working conditions were: room temperature, injection volume 10 μL, flow rate 1 mL/min. The two eluents were water (A) and acetonitrile, ACN (B), both with 0.1% (v/v) trifluoroacetic acid (TFA). The elution program was 85% A for 5 min, then to 50% A in 25 min, then 30% A in 10 min, then to 0% A in 1 min and hold for 4 min, and re-equilibration took 7 min. DAD spectra acquisition in the range of 200–650 nm every 0.8 s with a resolution of 4 nm.
2.3.2. SENLIBS All experiments were performed with the Modì double pulse LIBS instrument (Marwan Technology, Italy) [33]. LIBS spectra were acquired in collinear double-pulse mode (70 mJ per pulse at 1064 nm in 15 ns FWHM); the focusing lens had a focal length of 10 cm; the signal was collected with an optical fiber, placed at an angle of 45° with respect to the beam direction. The tip of the optical fiber was placed about 2 cm from the surface to collect the signal from the whole laser-induced plasma. The signal was acquired with a twochannels spectrometer (AvaSpec-2048-2, Avantes, The Netherlands) covering the spectral range from 180 to 900 nm; the spectral resolution of the spectrometer is 0.1 nm between 180 and 450 nm and 0.3 nm from 450 to 900 nm. LIBS spectra were acquired 300 ns after the second laser pulse, with an interpulse delay of 1 μs. The integration time was 2.5 ms.
3. Results and discussion 3.1. HPLC analysis Both references and historical samples were first characterized by HPLC-DAD, a classic technique for the identification of chromophorescontaining molecules. In this work, the HPLC analyses were carried out on textiles extracts obtained following a mild extraction procedure [34]. The adopted extraction treatment (DMF–Na2EDTA method) is able to preserve the glycoside compounds fixed on the fiber, along with several labile components of the dyestuffs. The identification of these glycosides and minor components is important in order to unequivocally characterize the dye source and to obtain information on the dyeing recipe applied. Table 2 shows the results of the HPLC-DAD analysis, reporting the sample list and molecular markers identified. As expected, the three references all show the signals relative to the main molecular markers of the dyes used in their production. In sample I Ro1 carminic acid, the main component, and dcII, a minor constituent of cochineal dye, were detected. HPLC chromatograms permit to identify luteolin, apigenin and traces of chrysoeriol in sample I G1, suggesting the use of weld. In samples I N1 and III Be2 only gallic and ellagic acids were detected, suggesting the use of gall nuts or alder bark to obtain a black shade.
3.2. SENLIBS analysis A small part (5 μL) of the EDTA/DMF extracts of the textiles was analysed by LIBS for trace elemental identification after drying on a
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
233
Fig. 2. Insets of LIBS spectra relative to each inorganic mordant of cochineal dyed silk (A, tin), European buckthorn dyed wool (B, copper), and oak gall dyed wool (C, iron) (black line: sample analysis as-is; red line: aqueous extract dried on silicon wafer).
solid support. In their work Hidalgo et al. [27] demonstrated that the analysis by direct laser irradiation of micro-droplets imply low precision and sensitivity when compared to the laser irradiation of micro-droplets dried on solid substrates (SENLIBS). The choice of the substrate material is critical to the success of the analysis. An ideal substrate should be inert towards both the solvent and the solute, should possess a simple electronic configuration to minimize spectral interferences, and should be inexpensive, replaceable, and easy to handle. Metallic supports (e.g. copper or aluminum) are generally inexpensive, readily available and they have excellent conductive properties to guarantee significant signal enhancement, but they can interfere with the identification of metal-based pigments. Preliminary SENLIBS experiments were conducted with materials that fully or partially satisfied the above mentioned criteria, i.e. glass, silicon wafer and Teflon®. Fig. 1 shows the results from the analysis of a standard solution of 250 mg/L Cu, Al, Sn, and Fe in 2% HNO3 deposited on the three substrates. Teflon® was tested because it is relatively free from transition metals, and for its hydrophobic nature, which allows the deposited drop to retain a small impact surface for the laser. Sample dispersion on the substrate for the analysis reduces indeed drastically the intensity of the LIBS signal, as it is a surface technique. However, Teflon® turned out to be unsuitable for LIBS analysis, because the absence of conduction electrons negatively affected the plasma formation and the intensity of the LIBS signal. On the other hand, we observed how silicon wafers for their semi-conductive nature provided the best results in terms of signal enhancement. Moreover, in the region of interest for the transitions of the elements in the analysed samples (250–350 nm) this substrate has a low number of emission lines, corresponding to silicon and to minor contaminants.
Different amounts of samples (5, 10 and 15 μL) were evaluated for SENSLIBS analysis, and 5 μL proved sufficient for sample characterization. This is a remarkable result, since the amount of extract could vary from tens to hundreds of microliters, depending on the initial amount of textile. In the case of historical tapestries, usually the amount of sample is relatively abundant, while it reduces drastically for historical embroideries. SENLIBS analyses were then applied on the same extracts used for HPLC-DAD characterization. Fig. 2 shows the insets of the LIBS spectra corresponding to the peaks of the selected mordants (Fe, Al and Cu), obtained by depositing 5 μL of the residues from reference laboratory-dyed samples on the silicon plate. Comparison with the spectrum from the clean substrate allowed straightforward identification of elemental emission lines arising from the analytes. For comparison, LIBS analyses were also carried out on each sample as-is (Fig. 2, black line). For Al and Cu mordanted textiles, a clear enhancement was obtained from SENLIBS technique with respect to the application of LIBS analysis directly on untreated samples, while in the case of Fe no appreciable increase in the signal intensity was observed. SENLIBS was able to determine the elements involved in the mordanting process, as the results agree with the recipes used for the preparation of the reference materials (mordanted with Al, Cu and Fe respectively). The same procedure of analysis was finally applied to the historical samples. Fig. 3 shows the comparison among the samples for the selected spectral range. To verify the presence of contaminants introduced during the preparation step, an undyed and not mordanted wool sample was subjected to the same pre-treatment and analysed as blank. The analysis of undyed wool revealed that neither Al or Fe are present in the blank, while low levels of Cu can be detected. Copper could be
234
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
Fig. 4. LIBS spectra for the aqueous extract of European buckthorn Cu-mordanted.
solvent to purify and pre-concentrate the dye components. The objective is to eliminate as much contaminants and solid particulate as possible, as they could enhance the background noise and/or damage the HPLC system. The feasibility of our approach was demonstrated also in this case, by the analysis of a reference textiles (European buckthorn Cu-mordanted) subjected to liquid-liquid microextraction with ethylacetate. The sample in DMF:EDTA was further diluted with water and, after being sonicated, was extracted three times with 200 μL of ethylacetate. The aqueous residue, which is usually discarded, was here reduced to 50 μL under nitrogen flow, dried on the solid support and analysed by LIBS. Even in this case the exact identification of the inorganic mordant was attained (Fig. 4). 4. Conclusions
Fig. 3. Comparisons of LIBS spectra from the dried extracts of blank wool (black), I Ro1 (red), I G1 (blue), I N1 (pink) and III Be2 (green). The inset relative to the spectral zone for each inorganic mordant (Sn, Cu or Fe) is shown.
present in the sample itself or could have been introduced in the sample preparation step. The sample I G1 results clearly Al-mordanted, while the signal related to Cu is comparable to the blank. In sample I R1, the predominant mordant is copper, even if small quantities of Al were detected. Iron is present only in samples I N1 and III Be2. This is consistent with the HPLC data that identified gallic and ellegic acid as chromophores. Sample III Be2 contains less Fe respect to I N1, as different shades of black were obtained in the past by modulating the Fe content. In sample III Be2, however, also a high content of Cu was detected. In our procedure, the same extract was used for both organic and inorganic analysis. In the case of archeological samples, the aqueous extracts usually undergo a liquid-liquid microextraction with an organic
Analysis of natural dyes is still an active area of research in cultural heritage studies. The major drawback associated with the identification of the original source of the historical colours arises from the fact that, due to their poor lightfastness, some chromophores are highly degraded and only minimal amounts can be extracted from the samples. Thus, considering the complexity of the matrix, a multi-analytical approach able to provide the highest number of information from a single sample is desirable. The multi-analytical approach developed in this work allowed us to highlight the potentialities and the limits of the tested techniques in the investigation of dyes and inorganic mordants. Concerning the determination of the organic dyes, the HPLC-DAD analyses detected the molecular markers of dyeing materials, confirming the identity of the studied samples. SENLIBS technique performed on the same samples used for HPLC analysis, dried on a silicon substrate, provided better results in terms of sensitivity respect to the analysis on textile samples as-is. The results of this work prove that the analysis of the inorganic fraction from the aqueous residue is a promising technique, which could be used in place of the direct analysis, to minimize sample destruction. The determination of metallic mordants was successfully achieved using SENLIBS on textiles extracts. Nevertheless, a careful interpretation of the results is necessary since the presence of some elements can be due to environmental contamination rather than mordanting. Further studies will extend the applicability of this micro-destructive approach to other kind of matrices, such as lakes in paintings. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.microc.2018.02.034.
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
References [1] S. Grierson, The Colour Cauldron: The History and Use of Natural Dyes in Scotland, Grierson, Perth, 1986. [2] I. Degano, et al., Analytical methods for the characterization of organic dyes in artworks and in historical textiles, Appl. Spectrosc. Rev. 44 (5) (2009) 363–410. [3] R.L. Feller, Artist's Pigments: a Handbook of Their History and Characteristics, Vol. 1, National Gallery of Art, Washington, Arctype Publications, London, 1986. [4] G.C.H. Derksen, H.A.G. Niederländer, T.A. van Beek, Analysis of anthraquinones in Rubia tinctorum L. by liquid chromatography coupled with diode-array UV and mass spectrometric detection, J. Chromatogr. A 978 (1) (2002) 119–127. [5] S.M. Halpine, An improved dye and lake pigment analysis method for high-performance liquid chromatography and diode-array detector, Stud. Conserv. 41 (2) (1996) 76–94. [6] J. Kirby, R. White, The identification of red lake pigment dyestuffs and a discussion of their use, National Gallery Technical Bulletin, 17, 1996, pp. 56–80. [7] L. Rafaëlly, et al., Optimisation of ESI-MS detection for the HPLC of anthraquinone dyes, Dyes Pigments 77 (1) (2008) 191–203. [8] J. Sanyova, Mild extraction of dyes by hydrofluoric acid in routine analysis of historical paint micro-samples, Microchim. Acta 162 (3–4) (2008) 361–370. [9] I. Surowiec, B. Szostek, M. Trojanowicz, HPLC-MS of anthraquinoids, flavonoids, and their degradation products in analysis of natural dyes in archeological objects, J. Sep. Sci. 30 (13) (2007) 2070–2079. [10] M.R. van Bommel, The analysis of dyes with HPLC coupled to photodiode array and fluorescence detection, Dyes in History and Archaeology, 20, 2005, pp. 30–38. [11] X. ZhanG, R.A. Laursen, Development of mild extraction methods for the analysis of natural dyes in textiles of historical interest using LC-diode array detector-MS, Anal. Chem. 77 (7) (2005) 2022–2025. [12] O. Deveoglu, et al., The characterisation by liquid chromatography of lake pigments prepared from European buckthorn (Rhamnus cathartica L.), Pigm. Resin Technol. 41 (6) (2010) 331–338. [13] A. Claro, et al., Identification of red colorants in van Gogh paintings and ancient Andean textiles by microspectrofluorimetry, J. Cult. Herit. 11 (1) (2010) 27–34. [14] A. Claro, et al., The use of microspectrofluorimetry for the characterization of lake pigments, Talanta 74 (4) (2008) 922–929. [15] B. Doherty, et al., A detachable SERS active cellulose film: a minimally invasive approach to the study of painting lakes, J. Raman Spectrosc. 42 (11) (2011) 1932–1938. [16] M. Leona, J. Stenger, E. Ferloni, Application of surface-enhanced Raman scattering techniques to the ultrasensitive identification of natural dyes in works of art, J. Raman Spectrosc. 37 (10) (2006) 981–992. [17] C. Lofrumento, et al., SERS detection of red organic dyes in Ag-agar gel, J. Raman Spectrosc. 44 (1) (2013) 47–54. [18] F. Sabatini, et al., A matrix-assisted laser desorption/ionization time-of-flight mass spectrometry method for the identification of anthraquinones: the case of historical lakes, J. Am. Soc. Mass Spectrom. 27 (11) (2016) 1824–1834. [19] N. Wyplosz, Laser Desorption Mass Spectrometric Studies of Artists' Organic Pigments, 2003.
235
[20] J. Kirby, M. Spring, C. Higgitt, The technology of red lake pigment manufacture: study of the dyestuff substrate, National Gallery Technical Bulletin, 26, 2005, pp. 71–87. [21] R.D. Morrison, B.L. Murphy, Environmental Forensics: Contaminant Specific Guide, Academic Press, 2010. [22] D. Montegut, et al., Examination of metal threads from some XV/XVI century Italian textiles by scanning electron microscopy-energy dispersive X-ray spectrometry, MRS Online Proc. Libr. Arch. 267 (1992). [23] L. Dussubieux, M.W. Ballard, Using ICP-MS to detect inorganic elements in organic materials: a new tool to identify mordants or dyes on ancient textiles, MRS Online Proc. Libr. Arch. 852 (2004). [24] L.G. Troalen, et al., A multi-analytical approach towards the investigation of subarctic Athapaskan colouring of quillwork and its sensitivity to photo-degradation, Microchem. J. 126 (2016) 83–91. [25] A. Manhita, et al., Enlightening the influence of mordant, dyeing technique and photodegradation on the colour hue of textiles dyed with madder–a chromatographic and spectrometric approach, Microchem. J. 98 (1) (2011) 82–90. [26] R.J. Koestler, R. Sheryll, N. Indictor, Identification of dyeing mordants and related substances on textile fibers: a preliminary study using energy dispersive X-ray spectrometry, Stud. Conserv. 30 (2) (1985) 58–62. [27] A.W. Miziolek, V. Palleschi, I. Schechter, Laser Induced Breakdown Spectroscopy (LİBS) – Fundamentals and Applications, Cambridge University Press, 2006. [28] D. Anglos, V. Detalle, Cultural heritage applications of LIBS, Laser-Induced Breakdown Spectroscopy, Springer Series in Optical Sciences, Volume 182, SpringerVerlag, Berlin Heidelberg 2014, p. 531 ISBN 978-3-642-45084-6. [29] V. Lazic, LIBS analysis of liquids and of materials inside liquids, Laser-Induced Breakdown Spectroscopy, Springer Series in Optical Sciences, Volume 182, SpringerVerlag, Berlin Heidelberg 2014, p. 195 ISBN 978-3-642-45084-6. [30] M.A. Aguirre, et al., Elemental analysis by surface-enhanced laser-induced breakdown spectroscopy combined with liquid–liquid microextraction, Spectrochim. Acta B At. Spectrosc. 79 (2013) 88–93. [31] I. Degano, et al., Historical and archaeological textiles: an insight on degradation products of wool and silk yarns, J. Chromatogr. A 1218 (34) (2011) 5837–5847. [32] A. Andreotti, et al., A Multi-Analytical Approach to Determine Organic Dyes in Tapestries, Vol. 1, 2007 21 Papers. [33] A. Bertolini, et al., Modì: a new mobile instrument for in situ double-pulse LIBS analysis, Anal. Bioanal. Chem. 385 (2) (2006) 240–247. [34] A. Manhita, et al., Extracting natural dyes from wool—an evaluation of extraction methods, Anal. Bioanal. Chem. 400 (5) (2011) 1501–1514.
Contents lists available at ScienceDirect
Microchemical Journal journal homepage: www.elsevier.com/locate/microc
Identification of inorganic dyeing mordant in textiles by surfaceenhanced laser-induced breakdown spectroscopy Beatrice Campanella a, Ilaria Degano b, Emanuela Grifoni a,c, Stefano Legnaioli a,c,⁎, Giulia Lorenzetti a, Stefano Pagnotta a, Francesco Poggialini a, Vincenzo Palleschi a,c a b c
Applied and Laser Spectroscopy Laboratory, Institute of Chemistry of Organometallic Compounds, Research Area of CNR, Via Giuseppe Moruzzi, 1, 56124 Pisa, Italy Department of Chemistry and Industrial Chemistry, University of Pisa, Via Giuseppe Moruzzi, 13, 56124 Pisa, Italy National Interuniversity Consortium of Materials Science and Technology (INSTM), Italy
a r t i c l e
i n f o
Article history: Received 28 August 2017 Received in revised form 31 January 2018 Accepted 27 February 2018 Available online 2 March 2018 Keywords: Textiles Mordant Surface-enhanced LIBS HPLC-DAD
a b s t r a c t The identification of both the organic and inorganic fraction of dyes and pigments is fundamental for their complete characterization and to assess the technologies used in their production. In this work, the feasibility of determining metallic elements used as mordant for dyed textiles was tested using Laser-Induced Breakdown Spectrometry (LIBS) in combination with liquid micro-extraction. Both reference laboratory-dyed and historic textiles were analysed in this study. Samples were first analysed without any preparation. Then, the chromophores-containing molecules were separated using a sample preparation procedure based on aqueous hydrolysis, and analysed by high-pressure liquid chromatography coupled with a diode array detector. The same extracts, containing also the inorganic fraction, were analysed by Surface-Enhanced Laser-Induced Breakdown Spectroscopy (SENLIBS) after drying on a solid substrate. Compared to the direct analysis, the SENLIBS method improved the sensitivity of the measurements. The procedure presented here allowed for the characterization of both organic and inorganic fraction of a single textile micro sample, thus avoiding further sampling. © 2018 Elsevier B.V. All rights reserved.
1. Introduction Until 150 years ago, the dyes for the textiles industry were obtained from natural sources. The dyeing process requires that the chromophores, obtained from the natural sources, are bonded as strongly as possible to the fibers. For the binding process, dyeing with the vast majority of natural dyes entails the treatment of the textile fibers with a solution of inorganic-based constituents, called mordants. The mordant is complexed by polar groups on the surface of protein fibers, and during the dyeing process the dye interacts with the mordant–fiber complex to form an insoluble brightly coloured species [1]. Alum (K2O4Al2(SO4) 3), iron (FeSO4·7H2O), and copper (CuSO4·5H2O) were the most commonly used mordants in antiquity, while tin (SnCl2·2H2O) and chrome (K2Cr2O7 or Na2Cr2O7) were introduced in the 18th century. The mordant ensures the brightness and wash-fastness of the dye, and it has great influence on the final colour obtained. For different cultures and periods, different mordanting systems and procedures were employed to produce a wide variety of hues. Determining which ⁎ Corresponding author at: Applied and Laser Spectroscopy Laboratory, Institute of Chemistry of Organometallic Compounds, Research Area of CNR, Via Giuseppe Moruzzi, 1, 56124 Pisa, Italy. E-mail address: [email protected] (S. Legnaioli).
https://doi.org/10.1016/j.microc.2018.02.034 0026-265X/© 2018 Elsevier B.V. All rights reserved.
substance was used to mordant or dye archeological or historical textiles provides a better understanding of the ancient mordanting or dying technology and assists in the dating and authentication of a specimen. Furthermore, the source of metal cations and dye/cation ratio, together with the dye and the substrate, are crucial parameters affecting the light-fastness and chemical features of the textiles [2,3]. Several methods have been developed to analyse dyes in textiles matrices in the last twenty years, exploiting different analytical techniques and testing different sample pre-treatment procedures. Most of them are based on High Pressure Liquid Chromatography coupled with Diode-Array, Fluorescence or Mass Spectrometric Detectors (HPLC-DAD-FD/MS) [4–12]. 3D Fluorescence [13,14], Surface Enhanced Resonance Raman Scattering (SERS) [15–17] and Matrix Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS and LDI-MS) are also employed [18,19]. Together with the identification of the organic dye, the analysis of the elemental composition of the inorganic fraction of dyed yarns can provide useful information, such as a relative dating. For example, developments in the use of different mordants for textile dyeing, notably the use of tin salts with cochineal dyestuff to give scarlet, were not taken up by the dyers until the very end of the 18th century, or beginning of the 19th century [20]. The first known commercial use for chromium began in the late 1700s as a paint pigment, and as early as 1820
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
the textile industry was using large amounts of chromium compounds as mordants [21]. However, literature on the characterization of inorganic mordants is still relatively poor [22–26]. The use of scanning electron microscopy (SEM) along with energy dispersive X-ray spectrometry (EDS) appears to be suitable for the identification of metallic mordants on fibers, whenever such instrumentation is available; however, most SEM-EDS applications (except Variable Pressure SEM) require the samples to be either conductive or coated in gold, graphite or another conductive material. Together with SEM-EDS, also energy-dispersive X-ray fluorescence (ED-XRF) is a common elemental technique for the mapping of mordant metals. Both techniques, however, are limited by low sensitivity (XRF is unsuitable for the analysis of light elements such as aluminum). The basic idea of this work is to demonstrate the feasibility of analysing by Laser-Induced Breakdown Spectroscopy (LIBS) a small part of the aqueous fraction from the micro-extraction aimed at dye characterization, for trace elements identification [27,28]. LIBS is a multi-elemental analytical technique that does not rely on sample treatment, as inductively-coupled plasma–optical emission spectroscopy, it does not require the samples to be conducting, as spark-OES does, and is not limited to the analysis of heavy elements, as X-ray fluorescence is. The required ablated mass is about few ng. Laser induced plasma formation on or inside liquids is characterized by large energy losses due to liquid evaporation [29]. This effect dramatically reduces the limits of detection of the technique. One of the best strategy in this cases is to deposit the liquid sample on a proper substrate and wait for a completely evaporation (SENSLIBS, Surface-Enhanced LIBS) [30]. In this work, a double approach has been developed to characterize reference dyed woolen and historic textiles, containing different metal cations. A previously developed extraction procedure based on EDTA/ DMF was applied to separate the mordant from the fiber, thus releasing the dyestuff in the extract. The extract was analysed both by HPLC-DAD, to identify the chromophore-containing molecule, and by SENLIBS to determine the mordant. 2. Materials and methods 2.1. Reagents and samples Disodium EDTA (Sigma-Aldrich, USA), acetonitrile (HPLC grade, Sigma-Aldrich, USA), ethyl acetate (AcOEt; Carlo Erba, Italy) and dimethylformamide (DMF, 99.8%, J.T. Baker, Holland) were used for sample pre-treatment and HPLC-DAD analyses. PTFE filters (4 mm thickness and 0.45 μm pore diameter, Sigma Aldrich) were chosen for the filtration. Ultrapure water prepared with an Elga Purelab-UV system (Veolia Environment, France) was used throughout. Glass slides (76 mm × 26 mm, 1.1 mm thickness microscope slides, from Sigma-Aldrich), Teflon sheets (fiberglass coated with PTFE), and
Table 1 Description of the references and the historical samples collected from Giuseppe fugge dalla moglie di Putifarre. Sample
Mordant Dye
References Wool Cu Wool Silk Historical samples I Ro1⁎ I G1⁎ I N1⁎ III Be2⁎
Fe Al
European buckthorn (“giallo di Spincervino”, i.e. Rhamnus cathartica) Gall nut Cochineal
⁎ Unknown dye and mordant.
Colour
Sample weight (mg)
silicon bulk wafers (single side polished, without dopant, 2 in. diameter and 0.5 mm thickness, from Sigma Aldrich) were tested as substrate materials for SENLIBS analysis. Two references raw wool (Italian commercial raw wool, 4/10 “pura lana”, cream colour from Campolmi, Italy) and a reference raw silk yarn (“seta schappe”, provided by Opificio delle Pietre Dure) were mordanted and dyed in laboratory. The detailed description of the raw materials and the treatment process is reported in Degano et al. [31]. Four historical samples from the tapestry “Giuseppe fugge dalla moglie di Putifarre” (Florence, Salone dei 200, Palazzo Vecchio), kindly provided by the Opificio delle Pietre Dure of Florence, were studied. The Italian artist Angnolo Bronzino designed this tapestry's cartoon around the half of XVI century, and the weaving is due to Karcher's atelier [32]. References and samples are listed in Table 1. 2.2. Sample preparation for HPLC and SENLIBS analysis The samples collected from reference specimens and the historical tapestry (see Table 1) were suspended in 400 μL and 200 μL, respectively, of 0.1% EDTA/DMF solution (1:1), and sonicated for 1 h at 60 °C. The extract in EDTA/DMF was filtrated through PTFE filters and then injected in the HPLC-DAD system. For the liquid-liquid extraction experiment, the extract in EDTA/DMF was diluted with 200 μL of water and extracted three times with 200 μL of ethylacetate. For SENLIBS analysis, the extract in EDTA/DMF or the aqueous residue from liquid/liquid extraction was deposited on the substrate and evaporated to dryness. Mild warming with a hot plate was used to facilitate the process. Upon dry out, the substrate was cooled to room temperature and the surface analysed by LIBS. 2.3. Instrumentation 2.3.1. HPLC-DAD Chromatographic separation was achieved using an analytical reverse-phase TC-C18 (2) column (4.6 × 150 mm, particle size 5 μm, Agilent) with a TC-C18 (2) precolumn (4.6 × 12.5 mm, particle size 5 μm, Agilent). The HPLC system consists of a PU-2089 quaternary pump (Jasco International Co., Japan) equipped with a degasser, an AS-950
Table 2 Results of the HPLC-DAD analysis (complete chromatograms are reported in the Supporting Information). Sample
Retention time (min)
Compound
Oak gall dyed wool, FeSO4 mordant
11.5 3.8 13.2
American cochineal dyed wool, alum mordant
9.2 10.2 13.5 15.3 20.6 21.2 18.9 22.1 25.2 29.2 32.2 11.6 4.2 11.5 18.9 21.6 22.2 10.1 9.1
Ellagic acid Gallic acid Gallic acid-like compound dcII Carminc acid dcIV dcVII Kermesic acid Flavokermesic acid Quercetin Kaempferol Isorhamnetin Rhamnetin Emodin Ellagic acid Gallic acid Ellagic acid Luteolin Apigenin Chrysoeriol Carminic acid dcII
Rhamnus, copper mordant
Yellow 6.0 Brown Red
5.4 5.0
Red Yellow Black Beige
1.4 1.6 0.7 0.4
231
I N1 III Be2 I G1
I Ro1
232
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
Fig. 1. Comparison of LIBS signal from a solution containing 250 mg/L of Cu, Al, Sn and Fe, deposited on Teflon® (black line), glass (red line) and silicon wafer (blue line).
autosampler (Jasco International Co., Japan) and an MD-2010 spectrophotometric diode array detector (DAD) (Jasco International Co., Japan). Data were processed with ChromNav software. The working conditions were: room temperature, injection volume 10 μL, flow rate 1 mL/min. The two eluents were water (A) and acetonitrile, ACN (B), both with 0.1% (v/v) trifluoroacetic acid (TFA). The elution program was 85% A for 5 min, then to 50% A in 25 min, then 30% A in 10 min, then to 0% A in 1 min and hold for 4 min, and re-equilibration took 7 min. DAD spectra acquisition in the range of 200–650 nm every 0.8 s with a resolution of 4 nm.
2.3.2. SENLIBS All experiments were performed with the Modì double pulse LIBS instrument (Marwan Technology, Italy) [33]. LIBS spectra were acquired in collinear double-pulse mode (70 mJ per pulse at 1064 nm in 15 ns FWHM); the focusing lens had a focal length of 10 cm; the signal was collected with an optical fiber, placed at an angle of 45° with respect to the beam direction. The tip of the optical fiber was placed about 2 cm from the surface to collect the signal from the whole laser-induced plasma. The signal was acquired with a twochannels spectrometer (AvaSpec-2048-2, Avantes, The Netherlands) covering the spectral range from 180 to 900 nm; the spectral resolution of the spectrometer is 0.1 nm between 180 and 450 nm and 0.3 nm from 450 to 900 nm. LIBS spectra were acquired 300 ns after the second laser pulse, with an interpulse delay of 1 μs. The integration time was 2.5 ms.
3. Results and discussion 3.1. HPLC analysis Both references and historical samples were first characterized by HPLC-DAD, a classic technique for the identification of chromophorescontaining molecules. In this work, the HPLC analyses were carried out on textiles extracts obtained following a mild extraction procedure [34]. The adopted extraction treatment (DMF–Na2EDTA method) is able to preserve the glycoside compounds fixed on the fiber, along with several labile components of the dyestuffs. The identification of these glycosides and minor components is important in order to unequivocally characterize the dye source and to obtain information on the dyeing recipe applied. Table 2 shows the results of the HPLC-DAD analysis, reporting the sample list and molecular markers identified. As expected, the three references all show the signals relative to the main molecular markers of the dyes used in their production. In sample I Ro1 carminic acid, the main component, and dcII, a minor constituent of cochineal dye, were detected. HPLC chromatograms permit to identify luteolin, apigenin and traces of chrysoeriol in sample I G1, suggesting the use of weld. In samples I N1 and III Be2 only gallic and ellagic acids were detected, suggesting the use of gall nuts or alder bark to obtain a black shade.
3.2. SENLIBS analysis A small part (5 μL) of the EDTA/DMF extracts of the textiles was analysed by LIBS for trace elemental identification after drying on a
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
233
Fig. 2. Insets of LIBS spectra relative to each inorganic mordant of cochineal dyed silk (A, tin), European buckthorn dyed wool (B, copper), and oak gall dyed wool (C, iron) (black line: sample analysis as-is; red line: aqueous extract dried on silicon wafer).
solid support. In their work Hidalgo et al. [27] demonstrated that the analysis by direct laser irradiation of micro-droplets imply low precision and sensitivity when compared to the laser irradiation of micro-droplets dried on solid substrates (SENLIBS). The choice of the substrate material is critical to the success of the analysis. An ideal substrate should be inert towards both the solvent and the solute, should possess a simple electronic configuration to minimize spectral interferences, and should be inexpensive, replaceable, and easy to handle. Metallic supports (e.g. copper or aluminum) are generally inexpensive, readily available and they have excellent conductive properties to guarantee significant signal enhancement, but they can interfere with the identification of metal-based pigments. Preliminary SENLIBS experiments were conducted with materials that fully or partially satisfied the above mentioned criteria, i.e. glass, silicon wafer and Teflon®. Fig. 1 shows the results from the analysis of a standard solution of 250 mg/L Cu, Al, Sn, and Fe in 2% HNO3 deposited on the three substrates. Teflon® was tested because it is relatively free from transition metals, and for its hydrophobic nature, which allows the deposited drop to retain a small impact surface for the laser. Sample dispersion on the substrate for the analysis reduces indeed drastically the intensity of the LIBS signal, as it is a surface technique. However, Teflon® turned out to be unsuitable for LIBS analysis, because the absence of conduction electrons negatively affected the plasma formation and the intensity of the LIBS signal. On the other hand, we observed how silicon wafers for their semi-conductive nature provided the best results in terms of signal enhancement. Moreover, in the region of interest for the transitions of the elements in the analysed samples (250–350 nm) this substrate has a low number of emission lines, corresponding to silicon and to minor contaminants.
Different amounts of samples (5, 10 and 15 μL) were evaluated for SENSLIBS analysis, and 5 μL proved sufficient for sample characterization. This is a remarkable result, since the amount of extract could vary from tens to hundreds of microliters, depending on the initial amount of textile. In the case of historical tapestries, usually the amount of sample is relatively abundant, while it reduces drastically for historical embroideries. SENLIBS analyses were then applied on the same extracts used for HPLC-DAD characterization. Fig. 2 shows the insets of the LIBS spectra corresponding to the peaks of the selected mordants (Fe, Al and Cu), obtained by depositing 5 μL of the residues from reference laboratory-dyed samples on the silicon plate. Comparison with the spectrum from the clean substrate allowed straightforward identification of elemental emission lines arising from the analytes. For comparison, LIBS analyses were also carried out on each sample as-is (Fig. 2, black line). For Al and Cu mordanted textiles, a clear enhancement was obtained from SENLIBS technique with respect to the application of LIBS analysis directly on untreated samples, while in the case of Fe no appreciable increase in the signal intensity was observed. SENLIBS was able to determine the elements involved in the mordanting process, as the results agree with the recipes used for the preparation of the reference materials (mordanted with Al, Cu and Fe respectively). The same procedure of analysis was finally applied to the historical samples. Fig. 3 shows the comparison among the samples for the selected spectral range. To verify the presence of contaminants introduced during the preparation step, an undyed and not mordanted wool sample was subjected to the same pre-treatment and analysed as blank. The analysis of undyed wool revealed that neither Al or Fe are present in the blank, while low levels of Cu can be detected. Copper could be
234
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
Fig. 4. LIBS spectra for the aqueous extract of European buckthorn Cu-mordanted.
solvent to purify and pre-concentrate the dye components. The objective is to eliminate as much contaminants and solid particulate as possible, as they could enhance the background noise and/or damage the HPLC system. The feasibility of our approach was demonstrated also in this case, by the analysis of a reference textiles (European buckthorn Cu-mordanted) subjected to liquid-liquid microextraction with ethylacetate. The sample in DMF:EDTA was further diluted with water and, after being sonicated, was extracted three times with 200 μL of ethylacetate. The aqueous residue, which is usually discarded, was here reduced to 50 μL under nitrogen flow, dried on the solid support and analysed by LIBS. Even in this case the exact identification of the inorganic mordant was attained (Fig. 4). 4. Conclusions
Fig. 3. Comparisons of LIBS spectra from the dried extracts of blank wool (black), I Ro1 (red), I G1 (blue), I N1 (pink) and III Be2 (green). The inset relative to the spectral zone for each inorganic mordant (Sn, Cu or Fe) is shown.
present in the sample itself or could have been introduced in the sample preparation step. The sample I G1 results clearly Al-mordanted, while the signal related to Cu is comparable to the blank. In sample I R1, the predominant mordant is copper, even if small quantities of Al were detected. Iron is present only in samples I N1 and III Be2. This is consistent with the HPLC data that identified gallic and ellegic acid as chromophores. Sample III Be2 contains less Fe respect to I N1, as different shades of black were obtained in the past by modulating the Fe content. In sample III Be2, however, also a high content of Cu was detected. In our procedure, the same extract was used for both organic and inorganic analysis. In the case of archeological samples, the aqueous extracts usually undergo a liquid-liquid microextraction with an organic
Analysis of natural dyes is still an active area of research in cultural heritage studies. The major drawback associated with the identification of the original source of the historical colours arises from the fact that, due to their poor lightfastness, some chromophores are highly degraded and only minimal amounts can be extracted from the samples. Thus, considering the complexity of the matrix, a multi-analytical approach able to provide the highest number of information from a single sample is desirable. The multi-analytical approach developed in this work allowed us to highlight the potentialities and the limits of the tested techniques in the investigation of dyes and inorganic mordants. Concerning the determination of the organic dyes, the HPLC-DAD analyses detected the molecular markers of dyeing materials, confirming the identity of the studied samples. SENLIBS technique performed on the same samples used for HPLC analysis, dried on a silicon substrate, provided better results in terms of sensitivity respect to the analysis on textile samples as-is. The results of this work prove that the analysis of the inorganic fraction from the aqueous residue is a promising technique, which could be used in place of the direct analysis, to minimize sample destruction. The determination of metallic mordants was successfully achieved using SENLIBS on textiles extracts. Nevertheless, a careful interpretation of the results is necessary since the presence of some elements can be due to environmental contamination rather than mordanting. Further studies will extend the applicability of this micro-destructive approach to other kind of matrices, such as lakes in paintings. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.microc.2018.02.034.
B. Campanella et al. / Microchemical Journal 139 (2018) 230–235
References [1] S. Grierson, The Colour Cauldron: The History and Use of Natural Dyes in Scotland, Grierson, Perth, 1986. [2] I. Degano, et al., Analytical methods for the characterization of organic dyes in artworks and in historical textiles, Appl. Spectrosc. Rev. 44 (5) (2009) 363–410. [3] R.L. Feller, Artist's Pigments: a Handbook of Their History and Characteristics, Vol. 1, National Gallery of Art, Washington, Arctype Publications, London, 1986. [4] G.C.H. Derksen, H.A.G. Niederländer, T.A. van Beek, Analysis of anthraquinones in Rubia tinctorum L. by liquid chromatography coupled with diode-array UV and mass spectrometric detection, J. Chromatogr. A 978 (1) (2002) 119–127. [5] S.M. Halpine, An improved dye and lake pigment analysis method for high-performance liquid chromatography and diode-array detector, Stud. Conserv. 41 (2) (1996) 76–94. [6] J. Kirby, R. White, The identification of red lake pigment dyestuffs and a discussion of their use, National Gallery Technical Bulletin, 17, 1996, pp. 56–80. [7] L. Rafaëlly, et al., Optimisation of ESI-MS detection for the HPLC of anthraquinone dyes, Dyes Pigments 77 (1) (2008) 191–203. [8] J. Sanyova, Mild extraction of dyes by hydrofluoric acid in routine analysis of historical paint micro-samples, Microchim. Acta 162 (3–4) (2008) 361–370. [9] I. Surowiec, B. Szostek, M. Trojanowicz, HPLC-MS of anthraquinoids, flavonoids, and their degradation products in analysis of natural dyes in archeological objects, J. Sep. Sci. 30 (13) (2007) 2070–2079. [10] M.R. van Bommel, The analysis of dyes with HPLC coupled to photodiode array and fluorescence detection, Dyes in History and Archaeology, 20, 2005, pp. 30–38. [11] X. ZhanG, R.A. Laursen, Development of mild extraction methods for the analysis of natural dyes in textiles of historical interest using LC-diode array detector-MS, Anal. Chem. 77 (7) (2005) 2022–2025. [12] O. Deveoglu, et al., The characterisation by liquid chromatography of lake pigments prepared from European buckthorn (Rhamnus cathartica L.), Pigm. Resin Technol. 41 (6) (2010) 331–338. [13] A. Claro, et al., Identification of red colorants in van Gogh paintings and ancient Andean textiles by microspectrofluorimetry, J. Cult. Herit. 11 (1) (2010) 27–34. [14] A. Claro, et al., The use of microspectrofluorimetry for the characterization of lake pigments, Talanta 74 (4) (2008) 922–929. [15] B. Doherty, et al., A detachable SERS active cellulose film: a minimally invasive approach to the study of painting lakes, J. Raman Spectrosc. 42 (11) (2011) 1932–1938. [16] M. Leona, J. Stenger, E. Ferloni, Application of surface-enhanced Raman scattering techniques to the ultrasensitive identification of natural dyes in works of art, J. Raman Spectrosc. 37 (10) (2006) 981–992. [17] C. Lofrumento, et al., SERS detection of red organic dyes in Ag-agar gel, J. Raman Spectrosc. 44 (1) (2013) 47–54. [18] F. Sabatini, et al., A matrix-assisted laser desorption/ionization time-of-flight mass spectrometry method for the identification of anthraquinones: the case of historical lakes, J. Am. Soc. Mass Spectrom. 27 (11) (2016) 1824–1834. [19] N. Wyplosz, Laser Desorption Mass Spectrometric Studies of Artists' Organic Pigments, 2003.
235
[20] J. Kirby, M. Spring, C. Higgitt, The technology of red lake pigment manufacture: study of the dyestuff substrate, National Gallery Technical Bulletin, 26, 2005, pp. 71–87. [21] R.D. Morrison, B.L. Murphy, Environmental Forensics: Contaminant Specific Guide, Academic Press, 2010. [22] D. Montegut, et al., Examination of metal threads from some XV/XVI century Italian textiles by scanning electron microscopy-energy dispersive X-ray spectrometry, MRS Online Proc. Libr. Arch. 267 (1992). [23] L. Dussubieux, M.W. Ballard, Using ICP-MS to detect inorganic elements in organic materials: a new tool to identify mordants or dyes on ancient textiles, MRS Online Proc. Libr. Arch. 852 (2004). [24] L.G. Troalen, et al., A multi-analytical approach towards the investigation of subarctic Athapaskan colouring of quillwork and its sensitivity to photo-degradation, Microchem. J. 126 (2016) 83–91. [25] A. Manhita, et al., Enlightening the influence of mordant, dyeing technique and photodegradation on the colour hue of textiles dyed with madder–a chromatographic and spectrometric approach, Microchem. J. 98 (1) (2011) 82–90. [26] R.J. Koestler, R. Sheryll, N. Indictor, Identification of dyeing mordants and related substances on textile fibers: a preliminary study using energy dispersive X-ray spectrometry, Stud. Conserv. 30 (2) (1985) 58–62. [27] A.W. Miziolek, V. Palleschi, I. Schechter, Laser Induced Breakdown Spectroscopy (LİBS) – Fundamentals and Applications, Cambridge University Press, 2006. [28] D. Anglos, V. Detalle, Cultural heritage applications of LIBS, Laser-Induced Breakdown Spectroscopy, Springer Series in Optical Sciences, Volume 182, SpringerVerlag, Berlin Heidelberg 2014, p. 531 ISBN 978-3-642-45084-6. [29] V. Lazic, LIBS analysis of liquids and of materials inside liquids, Laser-Induced Breakdown Spectroscopy, Springer Series in Optical Sciences, Volume 182, SpringerVerlag, Berlin Heidelberg 2014, p. 195 ISBN 978-3-642-45084-6. [30] M.A. Aguirre, et al., Elemental analysis by surface-enhanced laser-induced breakdown spectroscopy combined with liquid–liquid microextraction, Spectrochim. Acta B At. Spectrosc. 79 (2013) 88–93. [31] I. Degano, et al., Historical and archaeological textiles: an insight on degradation products of wool and silk yarns, J. Chromatogr. A 1218 (34) (2011) 5837–5847. [32] A. Andreotti, et al., A Multi-Analytical Approach to Determine Organic Dyes in Tapestries, Vol. 1, 2007 21 Papers. [33] A. Bertolini, et al., Modì: a new mobile instrument for in situ double-pulse LIBS analysis, Anal. Bioanal. Chem. 385 (2) (2006) 240–247. [34] A. Manhita, et al., Extracting natural dyes from wool—an evaluation of extraction methods, Anal. Bioanal. Chem. 400 (5) (2011) 1501–1514.