LETTER pubs.acs.org/ac
Direct Identification of Dyes in Textiles by Direct Analysis in Real Time-Time of Flight Mass Spectrometry Cathy Selvius DeRoo† and Ruth Ann Armitage*,‡ † ‡
Detroit Institute of Arts, Conservation Department, 5200 Woodward Avenue, Detroit, Michigan 48202, United States Department of Chemistry, Eastern Michigan University, Ypsilanti, Michigan 48197, United States ABSTRACT: We present here a method requiring no sample preparation for direct identification of the organic dye compounds quercetin, indigotin, and alizarin in reference materials, in solution, and also in situ in dyed fibers by use of direct analysis in real time (DART) ionization and highresolution time-of-flight mass spectrometry. Exact mass determinations on small samples of dyed textiles were completed in less than 1 min. With the ability to identify flavonoid, indigoid, and anthraquinone classes of dyes, this technique shows early promise as an additional analytical tool in the challenging analysis of organic dyes in rare cultural heritage materials and possesses the unique advantages of sensitivity and simplicity without the preparatory procedures required by other methods.
O
rganic dyes derived from plant and animal sources have been used by diverse cultures since antiquity in the dying of fibers and in painting as lake pigments in conjunction with various binding media carriers. Red anthraquinone-based dyes such as alizarin and purpurin are derived primarily from globally distributed species of the Rubiaceae. Indigo-yielding plants are also widely distributed and include primarily species of the genera Indigofera, Isatis (woads), and knotweed from the Polygonaceae family. Numerous flavonoid-based dyes originate from many plant genera and are diverse in chemical structure and in color, ranging from yellow to red to blue. Additional yellow dye sources include the carotenoid dyes of which saffron (Crocus sativus), annatto (Bixa orellana), and turmeric (Curcuma longa) are among the most common. Dyes derived from animals include the purple shades of indigotin derived from mollusks and the red anthraquinones from red scale insects, most notably kermes and cochineal.1 Because these organic dyes have very powerful tinting strength at nanogram to microgram levels, they continue to present an analytical challenge, requiring techniques of sufficient sensitivity. Compounding the analytical challenges are the restrictions posed by the inherent limitations on the size of samples removed from rare cultural heritage objects. Thus, the analysis of dyes in cultural heritage materials is confronted by the dual challenge of detecting an ultralow chromophore concentration in a very small sample. A highly sensitive technique for the identification of organic dyes is high-performance liquid chromatography (HPLC),2 which is typically coupled to additional analytical tools such as diode array detection (DAD),3 fluorescence detection (FD),4 and/or mass spectroscopic detection (MS).5 7 Identification of organic dyes by Raman spectroscopy is stymied by fluorescence which obscures the Raman spectrum with the exception of indigo. In an effort to surmount the fluorescence effect, surface-enhanced Raman spectroscopic (SERS) methods have been extensively employed in the analysis of organic dyes in fibers, paints, and glazes.8 11 r 2011 American Chemical Society
Both of these analytical methods, HPLC and SERS, while successful in certain applications, have limitations. SERS requires preparation of various colloids, frequently a hydrolysis pretreatment of the dye metal complex, and plasmon conditions that must be carefully determined in order for the Raman scattering enhancement effect to occur; however, SERS does require comparatively smaller samples than does HPLC. HPLC presents its own challenges in the relatively greater amount of sample material required and in the extraction procedures required which have the potential to alter the chemical structure of the dyes. In short, analytical techniques of the requisite sensitivity for dye analysis remain limited, and those that do exist are labor intensive, the optimization of analysis conditions can be challenging, and the very limited sample material must be painstakingly marshaled. An extensive review of analytical methods for the characterization of dyes in cultural heritage materials provides a broad overview of the existing literature; of particular note is the section on direct mass spectrometry (including direct exposure-MS, matrix assisted laser desorption ionization (MALDI)- and laser desorption ionization (LDI)-MS, and time-of-flight-secondary ion mass spectrometry (TOF-SIMS)).12 In most cases, the direct mass spectrometric methods utilized yielded only unit mass resolution. High-resolution MS has only rarely been applied to such studies to confirm the nature of the dye chromophores.13,14 The research presented here demonstrates success in preliminary studies in which direct analysis in real time-time-of-flight mass spectrometry (DART-TOF-MS) is used to identify quercetin, indigo, and alizarin in reference materials, in solution, and also in situ in dyed fibers. With the ability to identify flavonoid, indigotin, and anthraquinone classes of dyes, this technique shows early promise as an additional analytical tool in the Received: July 6, 2011 Accepted: August 16, 2011 Published: August 16, 2011 6924
dx.doi.org/10.1021/ac201747s | Anal. Chem. 2011, 83, 6924–6928
Analytical Chemistry
LETTER
Table 1. Observed Exact Masses for Pure Dye Compounds compound
exact mass
observed
difference,
formula
(M + H)
mass
mmu
quercetin
C15H10O7
303.0503
303.0507
0.26
alizarin
C14H8O4
241.0501
241.0496
0.49
purpurin
C14H8O5
257.0450
257.0457
0.70
nitroalizarin
C14H7NO6
286.0352
286.0348
0.40
indigotin
C16H10N2O2
263.0821
263.0823
0.20
challenging analysis of organic dyes and possesses the unique advantages of sensitivity and simplicity without the preparatory procedures required by other methods. The rapid analysis made possible by ambient ionization methods like DART has led to forensic science applications for identifying colorants in other materials. Jones et al.15 differentiated 43 different blue and black inks using DART-MS, creating a highly specific searchable database. A DART-MS method for identifying 1-methylaminoanthraquinone in bank security dyes was reported recently by Pfaff and Steiner,16 with a 5 ppm lower limit of detection for the method when using a simple solvent extraction. Colorants were among the many components in food packaging that were identified by DART methods.17 To date, little has been published on applications of DART in conservation science. The Library of Congress has made significant use of AccuTOF-DART for characterizing the myriad cultural heritage materials in those collections. Of particular note is Adams’ paper on differentiating papers using DART.18 We find no previous studies that have been published on natural dyes identified in textiles or cultural heritage materials by DART-MS.
’ MATERIALS AND METHODS Quercetin (as quercetin dihydrate), indigo, alizarin, and purpurin reference materials (all 99+%) were purchased from VWR. Dry yellow onion skins were obtained from a local grocery store. Ground madder root was obtained from Kremer Pigments (New York). Food grade alum and tartaric acid (as cream of tartar) were used for the mordant solution. Additional reagents for the indigo vat, including sodium dithionite and sodium hydroxide, were of reagent grade and obtained from existing stocks. Dyebaths19 and aluminum mordant solution20 were prepared according to standard recipes, reduced in volume to accommodate small samples. Pure quercetin and alizarin (in water) were also used directly as a dye on both mordanted and unmordanted textile samples, following the procedure of Ferreira et al.21 Four natural fiber types, cotton, linen, wool, and silk, were employed in this study. Each was used either as an undyed, unsized natural fiber textile from TestFabrics (provided by K. Jakes) or as small, dyed, single fiber samples. Single fibers of a known dye process were obtained from Traite des Matieres Colorantes du Blanchiment et de la Teinture du Coton, a late 19th-century treatise on fiber, dye chemistry, mordants, and dying techniques which conveniently included an appendix of alizarin-dyed cotton fiber skein examples of several dying techniques from which small fiber samples, approximately 0.5 cm in length and weighing less than 1 mg, were taken in order to assess the in situ analytical capability of DART-TOF. For consistency, we focus primarily on the results for cotton fibers. DART-TOF Analysis. Analysis was conducted with a JEOL AccuTOF mass spectrometer (JEOL USA, Peabody, MA)
Figure 1. DART mass spectra for quercetin and dye material and textiles containing quercetin.
equipped with a DART ion source (Ionsense, Saugus, MA) in positive ion mode, with helium as the ionization gas under the following conditions: flow rate of 2.5 L/min, gas temperature of 300 °C, and grid voltage of +350 V. Orifice 1 was held at 120 °C and 30 V, orifice 2 was 5 V, and the ring lens voltage was set to 5 V. The peaks voltage was held at 1500 V. The instrument was calibrated daily with a solution of PEG-600 in methanol. Whole textile or fiber samples were introduced directly with forceps into the gap between the DART source and the mass spectrometer orifice. No sample preparation was necessary. The dyes were introduced as powders or solutions on the closed end of a capillary melting point tube. Spectra were collected over 6925
dx.doi.org/10.1021/ac201747s |Anal. Chem. 2011, 83, 6924–6928
Analytical Chemistry
LETTER
Figure 2. DART mass spectra for indigo and indigo-dyed cotton samples.
0.5 2 min. The mass resolution of the AccuTOF was approximately 6000 throughout the analyses. The DART ionization process has been described elsewhere in detail;22,23 ions were observed at M + 1+, due to the addition of a proton.
’ RESULTS AND DISCUSSION Quercetin, alizarin, and indigo reference materials produced DART spectra at M + H+, as indicated in Table 1. The pure dye compounds were identified with mass accuracy of less than 1 mDa. Dyes on textiles showed somewhat more mass variation, with mass accuracy in the range of 3 5 mDa on average. This is likely the result of lower signal intensity for the dyed textiles due to the textile matrix. Quercetin was identified at m/z 303.049 Da in all of the textile samples dyed with the pure quercetin solution; signal intensity was higher for all of the mordanted textiles (both proteinaceous and cellulosic) than for unmordanted ones, most likely because the mordant facilitates fixing of the colorants into the textiles. The exact process by which mordants aid in color fixation remains unclear.24 Quercetin was also readily observed in dry onion skins, in the dye solution made from the onion skins, and in all of the textiles dyed with onion skin. Figure 1 shows DART mass spectra for a selection of the quercetin-containing samples.
Figure 3. DART mass spectra for alizarin, madder root, and dyed cotton samples indicating dyes containing alizarin.
Indigotin at m/z 263.083 Da was the predominant peak in all of the textiles dyed with the vat process. A sample of a 128 year old indigo-dyed cotton skein from the 19th-century dyeing treatise was also confirmed to contain indigo. Comparisons of the indigo spectra are shown in Figure 2. The 19th-century cotton sample showed a large 262.079 Da peak compared to the M + H+; this may be related to decomposition of the indigotin, 6926
dx.doi.org/10.1021/ac201747s |Anal. Chem. 2011, 83, 6924–6928
Analytical Chemistry
LETTER
Table 2. Summary of Alizarin-Dyed Textiles from 1883 French Book Traite des Matieres Colorantes du Blanchiment et de la Teinture du Coton French descriptor
English translation
color
alizarin confirmation at m/z 241.050
Lilas d’alizarine a l’huile tournante
alizarin lilac with rancid oil
bluish-purple
yes, primary peak
Mordore d’alizarine (grenat) Rose d’alizarine a l’huile tournante
alizarin pomegranate brown alizarin rose with rancid oil
maroon red pink
yes, primary peak yes
Orange d’alizarine (nitroalizarine)
alizarin orange (nitroalizarine)
orange
yes (primary peak is for nitroalizarin)
Rouge d’alizarine a l’acide sulforicinique
alizarin red with sulfuric acid on castor oil
crimson
yes
Palliacat d’alizarine
alizarin in the style of Pulicat
plum
yes, primary peak
Fleur de p^echer a l’alizarine
peach blossom
mauve
yes, primary peak
Rouge d’alizarine a huile tournante
alizarin red
red
yes, primary peak
Violet d’alizarine a huile tournante
alizarin violet with rancid oil
blue-violet
yes, primary peak
Violet d’alizarine sans huile
alizarin violet without oil
nearly black
yes, primary peak
differences in the indigo source material, or some combination of analysis conditions and the nature of the material. Alizarin (as alizarine) was indicated as the primary colorant in a selection of cotton skein fibers obtained from the 19th century dyeing treatise. Alizarin was identified at m/z 241.050 Da in all of the tested samples. The 19th-century samples ranged in color from mauve (fleur de p^echer) to nearly black (violet d’alizarine sans huile), yet all yielded strong alizarin signals within a few seconds of exposure to the DART gas stream. Representative DART spectra are shown in Figure 3. The complex dyeing techniques used to achieve the colors described in Table 2 included specific mordants, synthetic alizarin dyes, and modifications in the laborious teinture en rouge d’Andrinople (also known as rouge de Turc or Turkey red) dye process. In these, too, the alizarin peak was clearly observed. The orange colored sample, orange d’alizarine (nitroalizarine), was confirmed to contain primarily nitroalizarin at m/z 286.035 (Table 1). The results for the 128 year-old French cotton samples are summarized in Table 2. For comparison purposes, a small (0.3 mg) piece of madder root, the predominant natural source of alizarin, was also analyzed. As expected, the primary compounds observed were alizarin and purpurin, which was confirmed through comparison to the pure reference material. In many of the dyed textile samples, both old and new, and regardless of the alizarin-based dye process employed, a peak was observed at 257.24 Da. With unit mass resolution, this compound might be identified as purpurin. Because of the high mass resolution of the AccuTOF-MS, we could clearly differentiate this unknown compound from purpurin (257.045 Da, Table 1). Exact mass was also used to identify two of the minor anthroquinone components of madder: pseudopurpurin at m/z 301.0353 Da and rubiadin at m/z 255.0667 Da, with mass differences of 0.36 and 0.86 mDa, respectively, for the M + H+. This preliminary work demonstrates that DART-MS is sufficiently sensitive for the qualitative identification of selected anthraquinone, flavonoid, and indigoid classes of dyes in biological source materials (madder root and onion skin), in aqueous dye solutions, and in situ in dyed natural fibers at concentrations relevant to the tinting strength of the chromophores. Analysis of historical fiber samples dyed with alizarin and its derivatives and with indigo showed that DART-MS confirms the presence of the relevant compounds rapidly, accurately, and without need for extraction, hydrolysis, or derivatization. This method shows excellent potential for further applications in characterizing organic colorants in cultural heritage materials.
’ CONCLUSIONS DART-MS is a rapid, sensitive, preparation-free method for identifying the primary organic dye chromophores in natural fiber textiles, both proteinaceous (wool and silk) and cellulosic (cotton and linen) fibers. Dyes were readily identified in freshly dyed textiles and in cotton skeins that were prepared more than a century ago. This method may not identify the specific source of the dye or be able to distinguish different chromophores possessing identical masses, but DART-MS can provide sufficient information from a small investment of time and material to determine what additional analytical approaches will be needed, if any. Further studies of historic textiles are currently underway. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Phone: (734) 487-0290. Fax: (734) 487-1496.
’ ACKNOWLEDGMENT The authors gratefully acknowledge financial support from the National Science Foundation (Award MRI-R2 No. 0959621) as well as the Eastern Michigan University (EMU) Chemistry Department and Provost’s Office. We thank EMU students John Hopkins, Payge Shelton, and Christina Varney and colleague Dr. Deborah Thompson for their help in dyeing and solution preparation and Dr. Kathryn Jakes for providing raw textile samples. ’ REFERENCES (1) deGraaf, J. H. H. The Colourful Past; Abegg-Stiftung: Riggisberg and Archetype: London, 2004. (2) Wouters, J. Stud. Conserv. 1985, 30, 119–128. (3) Zhang, X.; Corrigan, K.; MacLaren, B.; Leveque, M.; Laursen, R. Stud. Conserv. 2007, 52, 211–220. (4) Surowiec, I.; Nowik, W.; Trojanowicz, M. Microchim. Acta 2008, 162, 393–404. (5) Zhang, X.; Laursen, R. Int. J. Mass Spectrom. 2009, 284, 108–114. (6) Manhita, A.; Ferreira, V.; Vargas, H.; Ribeiro, I.; Candeias, A.; Teixeira, D.; Ferreira, T.; Dias, C. B. Microchem. J. 2011, 98, 82–90. (7) Mantzouris, D.; Karapanagiotis, I.; Valianou, L.; Panayiotou, C. Anal. Bioanal. Chem. 2011, 399, 3065–3079. (8) Leona, M. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14757–14762. (9) Leona, M.; Stenger, J.; Ferloni, E. J. Raman Spectrosc. 2006, 37, 981–992. (10) Jurasekova, Z.; Domingo, C.; Garcia-Ramos, J. V.; SanchezCortes, S. J. Raman Spectrosc. 2008, 39, 1309–1312. 6927
dx.doi.org/10.1021/ac201747s |Anal. Chem. 2011, 83, 6924–6928
Analytical Chemistry
LETTER
(11) Casadio, F.; Leona, M.; Lombardi, J. R.; Van Duyne, R. Acc. Chem. Res. 2010, 43, 782–791. (12) Degano, I.; Ribechini, E.; Modugno, F.; Colombini, M. P. Appl. Spectrosc. Rev. 2009, 44, 363–410. (13) Van Elslande, E.; Guerineau, V.; Thirioux, V.; Richard, G.; Richardin, P.; Laprevote, O.; Hussler, G.; Walter, P. Anal. Bioanal. Chem. 2008, 390, 1873–1879. (14) Papageorgiou, V. P.; Mellidis, A. S.; Assimopoulou, A. N.; Tsarbopoulos, A. J. Mass Spectrom. 1998, 33, 89–91. (15) Jones, R. W.; Cody, R. B.; McClelland, J. F. J. Forensic Sci. 2006, 51, 915–918. (16) Pfaff, A. M.; Steiner, R. R. Forensic Sci. Int. 2011, 206, 62–70. (17) Ackerman, L. K.; Noonan, G. O.; Begley, T. H. Food Addit. Contam., Part A 2009, 26, 1611–1618. (18) Adams, J. Int. J. Mass Spectrom. 2011, 301, 109–126. (19) Schweppe, H. Practical Hints on Dyeing with Natural Dyes: Production of Comparative Dyeings for the Identification of Dyes on Historic Textiles; Smithsonian Institution: Washington, DC, 1986; p 38. (20) Cannon, J.; Cannon, M. Dye Plants and Dyeing; Timber: Portland, OR, 1994; pp 13 15. (21) Ferreira, E. S. B.; Quye, A.; McNab, H.; Hulme, A. N. Dyes History Archaeol. 2002, 18, 63–70. (22) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297–2302. (23) McEwen, C. N.; Larsen, B. S. J. Am. Soc. Mass Spectrom. 2009, 20, 1518–1521. (24) Cardon, D. Natural Dyes: Sources, Tradition, Technology and Science; Archetype: London, 2007; pp 5 6.
6928
dx.doi.org/10.1021/ac201747s |Anal. Chem. 2011, 83, 6924–6928
