Single-Pot Extraction-Analysis of Dyed Wool Fibers with Ionic LiquidsKatherine S. Lovejoy,† Alexander J. Lou,† Lauren E. Davis,† Timothy C. Sanchez,‡ Srinivas Iyer,‡

Cynthia A. Corley,§ John S. Wilkes,§ Russell K. Feller,† David T. Fox,‡ Andrew T. Koppisch,‡,⊥

and Rico E. Del Sesto*,†

†Materials Chemistry, Los Alamos National Laboratory, Los Alamos, NM, USA‡Biosciences Division, Los Alamos National Laboratory, Los Alamos, NM, USA§Department of Chemistry, United States Air Force Academy, USAFA, CO, USA

*S Supporting Information

ABSTRACT: Analytical capabilities to identify dyes associated with structurallyrobust wool fibers would critically assist crime-scene and explosion-sceneforensics. Nondestructive separation of dyes from wool, removal ofcontaminants, and dye analysis by MALDI- or ESI-MS, were achieved in asingle-pot, ionic liquid-based method. Ionic liquids (ILs) that readily denaturethe wool α-keratin structure have been identified and are conducive to smallvolume, high-throughput analysis for accelerated threat-response times. Wooldyed with commercial or natural, plant-based dyes have unique signatures thatallow classification and matching of samples and identification of dyestuffs. Woolreleased 0.005 mg of dye per mg of dyed wool into the IL, allowing for analysis of single-thread sample sizes. The IL + dyemixture promotes sufficient ionization in MALDI-MS: addition of common MALDI matrices does not improve analysis ofanionic wool dyes. An inexpensive, commercially available tetrabutylphosponium chloride IL was discovered to be capable ofdenaturing wool and was determined to be the most effective for this readily fieldable method.

New analytical methods for rapid identification of organiccompounds found on surfaces, in solids, or in aqueous

systems, are critical for environmental and forensic sampling.Ideally, analytical procedures should be designed to reduce thenumber of analytical steps, eliminate extensive samplepreparation, provide for analysis of diverse samples, and reducesample size requirements, thereby improving threat-responsetimes. Forensic analysis of dyes from textile fibers can provideinformation on dye batch and origin of the fabric or identifymatching fibers and dyes from different sources. Our researchwas driven by the need to identify textiles and dyes found atcrime scenes and forensic sites. The required analyticalprocedures need to extract dyes from fibers without alteringthe chemical nature of the dyes. Wool-derived threads or fibersamples pose a particular challenge, as they require a solventharsh enough to denature the robust and complex wool keratinstructure and disrupt the dye−keratin interactions. Thedenaturing agent must also be sufficiently nondestructive todyes and allow their release into the solvent without alterationor degradation. The dye must have significant solubility in thisdenaturing system to allow their separation from degradedwool keratin. Finally, the extraction solution needs to be readyfor analysis of the dissolved species with minimal samplepreparation.Wool is a protein-based fiber comprised of coiled coils of α

helix pairs, with strong hydrogen bonding interactionsenforcing the helical structure, while hydrophobic interactionsgenerate a strong core between each pair of helices. Dyes canpenetrate deep into this structure, thus making them difficult to

remove once incorporated into the wool fibers. Ionic liquids(ILs) are solvents that can inherently achieve the multiplefunctionalities necessary to disrupt the wool structure andextract the organic dyes. The nonvolatile, organic salts can bedesigned to introduce multiple types of intramolecularinteractions, including hydrogen bonds and hydrophobiceffects, while their charged nature presents nonsolvated ionicand polar components to disrupt intermolecular interactionsand extract or solvate dyes. Methods of wool dissolution usuallyinvolve ionic agents, such as the highly effective, yet extremelydestructive alkaline hydrolysis protocol.1 ILs also have a highionic strength but are less destructive to dyes because they lackthe chemical reactivity of OH−. The solvating power of ILs hasbeen utilized for dissolution of wood pulp,2 cellulose,3,4 lignin,5

and silk,6 in addition to standard organic and inorganiccompounds. Herein we present the extraction of typical aciddyes, used in wool fiber dyeing processes, from aqueoussolutions as well as from dyed wool fibers. The extractionsolutions can then be analyzed directly via MALDI-TOF orelectrospray mass spectrometry to identify components ofcommercial dyes, as well as more challenging natural dyes, fromwool fibers.Using chloride-based ILs and capitalizing on their hydrogen

bond-disrupting properties, we have succeeded in breakingdown the alpha and beta structures of keratin in wool fibers and

Received: July 5, 2012Accepted: September 27, 2012

Article

pubs.acs.org/ac

© XXXX American Chemical Society A dx.doi.org/10.1021/ac301873s | Anal. Chem. XXXX, XXX, XXX−XXX

in animal hoof material. We coupled this disruption capability,also known with imidazolium chloride-based ionic liquids, suchas 1-butyl-3-methylimidazolium chloride ([BMIM][Cl]),7 inseveral new ILs with direct analysis by MALDI-TOF, and aninvestigation of dye degradation and dye limit-of-detection toyield a single-pot forensic method. Various ionic liquids are ableto perform this task, though we found the tetrabutylphospho-nium chloride [Bu4P][Cl] ionic liquid to be most effective.With dyed wool fibers (both commercial and those prepared in-house), the dyes and associated process chemicals arenondestructively extracted into the IL solvent upon disruptionof the keratin structure. Additionally, we have determined thatthis method of extraction and analysis of anionic dyes bymatrix-assisted laser desorption ionization mass spectrometry(MALDI-MS) in the negative ionization mode, can reliablydetect as little as 25 pmol of dye molecule across a wide varietyof samples.

■ EXPERIMENTAL SECTION

Materials. Sheep wool was obtained as predyed yarn(Brown Sheep Inc., Lincoln, NE) or undyed from local sources.Keratin was obtained as pig hoof powder (USB Corp.,Cleveland, OH). Trihexyltetradecylphosphonium chloride, orCYPHOS 101 (CY101) was a gift from Cytec Industries, Inc.(Woodland Park, NJ) and was purified prior to use by washingwith water and extracting with hexanes until the UV−visabsorption beyond 300 nm disappeared, followed by dryingunder vacuum at 90 °C. MALDI matrices were obtained fromAldrich Chemical Co. (2,5-dihydroxybenzoic acid, DHB) andAcros Organics (9-aminoacridine, AA).MALDI-MS. Matrix assisted desorption laser ionization mass

spectrometry (MALDI-MS) was performed on a 4800 PlusMALDI-TOF/TOF system (Applied Biosystems, Framingham,MA), equipped with a 200 Hz ND:YAG laser operating at 355nm, in both positive (fixed laser intensity of 3700 arbitraryunits) and negative ion (fixed laser intensity of 4300) modesusing 400 shots per spectrum and with the final detectorvoltage set at 1.818 kV. Samples for determining limit ofdetection in the negative mode were prepared using kiton reddissolved in [R3R′P][Cl] (trihexyltetradecylphosphoniumchloride) at 0.24 M, and deposited in 1 μL spots containing0.4, 4, 25, 40, 400, 1200, 2400, or 4800 pmole, For determininglimit of detection in the positive and negative modes, coumarin540 was used in the same manner as for kiton red.Sodium Dodecylsulfate-Polyacrylamide Gel Electro-

phoresis (SDS-PAGE). The molecular weights of IL-dissolvedand NaOH-degraded keratin samples were determined by SDS-PAGE on Bis−Tris gradient gels (4−12%, NuPAGE). Briefly,samples were filtered through a 0.45 μm syringe filter, heated to90 °C in a loading buffer (26 mM Tris-HCl, pH 8.5, 35 mMTris base, 0.5% SDS, 2.5% glycerol, 0.13 mM EDTA, 0.05 mMbromophenol blue, plus 1 μL of 0.5 M DTT), eluted for 45 minat 200 V (running buffer: 2.5 mM MOPS, pH 7.7, 2.5 mM Trisbase, 0.005% SDS, and 0.05 mM EDTA), and stained withcoomassie blue.ESI-MS. Samples of dye extracted from wool were analyzed

in the negative mode on a time-of-flight mass spectrometer(Agilent 6210 series) with electrospray ionization. Sampleswere introduced by HPLC (Agilent 1100 series) with a shortreverse-phase column and eluted in 5 mM ammonium formatebuffer using a gradient of 30% to 90% methanol.

Protein Concentration. The concentration of solubilizedprotein was determined using modification of the Bradfordassay8 with bovine serum albumin as a standard.

Quantification of Dissolved Protein. Wool piecesweighing 2 mg and measuring ∼1 mm in length were incubatedin 1 dram glass vials containing 200 μL of MeOH (48 h, RT),200 μL of 0.75 M NaOH (1 h, 50 °C), or 8 mg of[(C4H9)4P][Cl] (tetrabutylphosphonium chloride), (6 h, 120°C).

Powder X-ray Diffraction. X-ray diffraction patterns werecollected at 25 °C using a Bruker D8 Advance diffractometerequipped with Cu Kα radiation (λ = 0.15418 nm) and an areadetector. Data were collected from 2θ = 4 to 39°.

FT-IR and UV−vis Spectroscopy. Infrared spectra weremeasured using an Avatar 360 FT-IR ESP (Thermo Electron-Nicolet, Waltham, MA) equipped with a Smart Dura SamplIRhorizontal attenuated total reflectance (HATR) accessory.Infrared spectra were fit with Gaussian line shapes to yieldpeak shape information. Absorption spectra were collected on aHewlett-Packard 8453 diode array spectrophotometer (AgilentTechnologies, Inc., Santa Clara, CA) in a 1 cm path lengthquartz cuvette. NMR data sets were collected on a 300 MHzBruker instrument using sample concentrations of 50 mM inCDCl3.

Wool Dyeing. Wool was obtained from local sources anddyed using an acid process. The preparation of dyed wool usingcommercial dyes was performed by adding 0.5 g of undyedmerino wool to a solution of 0.01 g acid dye, 0.13 mL of 4%acetic acid, and 5.6 mL water at 80 °C for 2 h. Wool was alsodyed with cota (often used locally as a dye from the plantThelesperma megapotamicum) or sandalwood (Pterocarpussantalinus) by placing 0.5 g of merino wool in a cota orsandalwood solution at 80 °C for 1 h. The dyed wool wasremoved and washed with hot water until the water ran clearfor both natural and commercial dyes, typically about threewashes of 50 mL water. The wool was then washed once withmethanol.

Extraction of Dye from Dyed Wool. A solutioncontaining 5 wt % wool dyed in our laboratory or commercialwool yarn (Brown Sheep Blood Red 80 Dyed Wool, Lincoln,Nebraska) and the [Bu4P][Cl] ionic liquid (CYPHOS 164) washeated to 130 °C. A bright red solution with visible yarn in itwas produced after 2 h and a cloudy red solution with nostructurally intact yarn was produced after 24 h. The solutionwas filtered through a 0.45 μM syringe filter and spotted on theMALDI-MS plate in 1 μL aliquots either neat or diluted 10000-fold in methanol. The diluted samples showed improvedionization and produced better mass spectrometry data.Extraction of dyes from brown, yellow, and orange-dyed woolwas performed in an identical manner.

Liquid−Liquid Extraction of Dyes from AqueousSolution. Orange, brown, yellow, or natural cota-derivedwool dyes were prepared as 5 mM solutions in 1 mL water. Onthe basis of the extinction coefficient, dyes were diluted to 250μM (orange, brown, and yellow) or 2.5 mM (cota) in 2 mL ofwater at room temperature. A 2 mL portion of neat IL(butylmethylpyrrolidinium bis(trifluoromethylsulfonyl)imide([C4MPy][Tf2N]), trihexyltetradecylphosphonium chloride([R3R′P][Cl]), trihexyltetradecylphosphonium bromide([R3R′P][Br]), or trihexyltetradecylphosphonium bis-(trifluoromethylsulfonyl) imide ([R3R′P][Tf2N])) was addedto the aqueous dye solution and the 2-layered sample wasshaken vigorously by hand for two minutes. The commercially

Analytical Chemistry Article

dx.doi.org/10.1021/ac301873s | Anal. Chem. XXXX, XXX, XXX−XXXB

available CY101 was purified prior to use as described in thegeneral procedures section. The pH of water after coming incontact with the trihexyltetradecylphosphonium-derived ILswas invariably between 1 and 2 under these conditions.Samples were centrifuged at 2000 rpm (515g) for 30 min in a

Thermo IEC Centra CL2 centrifuge (Thermo Fisher Scientific,Waltham, MA) equipped with a 4-hole fixed angle rotor (partnumber 804SF) to obtain clean separations of the ionic liquidand aqueous layers. The layers were separated and separatelycentrifuged for 1 min at 1000 rpm (130g) to improve clarity.Absorbance over a range from 200 to 900 nm was quantifiedusing an HP 8453 UV−visible Spectrophotometer. The pH ofthe aqueous layer before and after extraction was determinedfor all samples.

■ RESULTS AND DISCUSSIONProtocols for dissolution of wool using imidazolium-based ionicliquids7 were followed with the ionic liquids presented herein.We then screened a series of other ionic liquids in search ofwool-solubilizing ILs that were liquids at room temperature andinexpensive. Other desirable IL characteristics were immisci-bility with water, to allow later separation of wool keratin fromdyes, and promotion of satisfactory analyte signals by MALDI-MS. Ionic liquids containing halide ions seem to be especiallyeffective at disrupting hydrogen bonding in strongly networkedstructures.2,7 This is likely due to the highly exposed negativecharge of chloride, which cannot interact strongly with theircounterions due to the typical steric protection of phospho-nium and ammonium cations. The resulting strong interactionsof this exposed halide with solutes that are hydrogen bonddonors can completely disrupt even the most rigid of networks.Using chloride-based ionic liquids, we have partially denaturedthe alpha and beta structures of keratin to solubilize wool fibersand animal hoof material.Two types of keratinaceous materials were studied. Pig hoof

keratin, which contains more β-keratin content than wool, aswell as dyed and undyed sheep wool were analyzed. We initiallyevaluated a panel of ionic liquids for possible hoof keratinsolubilization (Figure 1). ILs that dissolved at least 1 wt %keratin included those with chloride anions, such as the[BMIM][Cl], [(C4H9)4P][Cl], and [C4MPy][Cl] ILs. Thesesame three ILs also dissolved predyed wool fibers, and thechloride-based disruption of the fiber structure was sub-stantiated by UV−vis absorption, powder X-ray diffraction(XRD), and infrared (FT-IR) characterization of textiles beforeand after addition of ILs. Figure 2 shows the results of theBradford protein assay for solutions containing proteinsolubilized in ILs, filtered through a 0.45 μm syringe filter,and diluted in buffer. Care was taken to use exactly 100 mM ILin each calibration and test sample because of a slight deviationin the assay response in the presence of IL. Results from theBradford assay, indicating that 12 mg/mL protein could befound in 0.75 M NaOH, 6 mg/mL in [(C4H9)4P][Cl] and 0.9and 0.8 mg/mL in [C4MPy][Cl] and [BMIM][Cl] parallel theraw absorption at 280 nm (Supporting Information Figure S-1).SDS-PAGE on protein precipitated from filtered IL samplesusing acetone and redissolved in loading buffer showed a widedistribution of protein molecular weights for NaOH-degradedprotein from 220 to 10 kDa. Protein precipitated from ILsolution remained in the wells and did not migrate into the gel(Supporting Information Figure S-2), indicating that theprotein content was sufficiently denatured to pass through a0.45 μm filter, but was larger than ∼220 kDa.

The X-ray diffraction pattern (Figure 3) of wool keratin showtwo large peaks centered at 2θ = 9.8°, d = 9.03 Å and 2θ =20.2°, d = 4.40 Å. The wool solubilized in [(C4H9)4P][Cl]shows two large peaks centered at 2θ = 9.02, d = 9.80 Å and 2θ= 22.08, d = 4.03 Å. Classical diffraction patterns for woolkeratin, which led to elucidation of the β-sheet and α-helixstructures, stipulate d values of 5.1 on the meridian and 9.8 Åon the equator for the α-helix and 4.65 and 9.8 Å, both on theequator, for the β-keratin form of stretched wool.9,10 All ofthese correspond to the center of broad, diffuse peaks,complicating definite attribution of peak shape changes. Weobserve a decrease in the intensity of the d = 9.80 peak relativeto the d = 4.03 peak after dissolution of the wool in ionic liquid,which we interpret as a decrease in structural integrity, in whichnearest-neighbor interactions between residues are mostlyundisturbed but longer-range order between residues is

Figure 1. Table (top) indicates the solubility of 1 wt % keratin in ILswith cations that include [(C6H13)3(C14H29)P] ([R3R′P]), and 1-butyl-, 1-octyl-, and 1-hexyl-3-methylimidazolium (BMIM, OMIM,and HMIM). Photo shows commercially dyed wool yarn dissolved in[R3R′P][Cl] (left), [(C4H9)4P][Cl] (center), and [BMIM][Cl](right). The degree of solubilization shown on the left is alsoobserved with organic solvents including chloroform, isopropanol, andethyl acetate. The CYPHOS names for ILs are [P(C4H9)4][Cl],CY164; [P(C14H29)(C6H13)3][N(CN)2], CY105; [P((CH3)2CH−CH2)3CH3][OTs], CY106; [P(C14H29)(C6H13)3][Cl], CY101; [P-(C14H29)(C6H13)3][Tf2N], CY109; and [P(C14H29)(C6H13)3][BF4],CY111.

Figure 2. Solubilization of hoof keratin and adjustment of theBradford assay for use with [(C4H9)4P][Cl].

Analytical Chemistry Article

dx.doi.org/10.1021/ac301873s | Anal. Chem. XXXX, XXX, XXX−XXXC

disrupted. The 9.8 Å reflection is described as “a fine oscillationin the transform due to external interference between themicrofibrils superimposed on a coarser oscillation attributableto the geometrical arrangement of the protofibrils”.11 Upondissolution in IL, we also observe a shift from d = 4.40 Å(ordered wool) to d = 4.03 (IL-degraded wool) and a slightnarrowing of this peak. We recognize in this shift a change inthe wool structure and speculate that some part of the peaknarrowing arises from loss of the shoulder of α-helical reflectionat 5.1 Å.The FT-IR spectra also demonstrate the disruption of

protein interactions following hoof keratin dissolution in[(C4H9)4P][Cl], as evidenced by changes in the absorptionsof the CO stretch near 1650 cm−1 (amide region I) and theN−H bend near 1530 cm−1 (amide region II) (Figure 4).Absorption for the CO stretch increased relative to the N−Habsorption and the peak shifted 15 cm−1 higher, indicating analtered environment for the carbonyl upon dissolution in ionicliquid. Shift of the carbonyl absorption to higher wavenumbersis associated with disordered keratin conformations followingby treatment with m-bisulfite and subsequent precipitation.12

When the amide I and II regions were fit with Gaussian lineshapes, relative contributions of the α-helix- and β-sheet-based

carbonyl absorptions before and after IL dissolution could bedetermined. The α-helix component decreased from 34 to 11%after treatment with IL and the β-sheet component decreasedfrom 25 to 11%, resulting in the peak shifting to higherwavenumbers The amide II N−H bend region did not shiftsignificantly and did not reflect the dissolution in IL as stronglyas the CO stretch region. No peaks in the sulfoxide regionbetween 1150 and 1000 or in the thiol region between 2600and 2500 appeared after IL treatment, confirming that cleavageof the disulfide bonds in keratin is not a significant contributorto dissolution of keratin in IL.When this treatment is performed on dyed wool fibers, in

addition to observing changes in the XRD and IR spectra andeither partial of complete dissolution of the fiber, the dyes arealso readily extracted into the IL solution. ILs used forextraction included [(C4H9)4P][Cl], and butylmethylpyrrolidi-nium bis(trifluoromethylsulfonyl)imide ([C4MPy][Tf2N]), aswell as trihexyltetradecylphosphonium chloride ([R3R′P][Cl]),bromide ([R3R′P][Br]) and bis(trifluoromethylsulfonyl)imide([R3R′P][Tf2N]). Extraction efficiencies were determinedaccording to the formula E% = Ci − Cf/Ci, where Ci and Cfare the initial and final concentrations of the dye in the aqueouslayer, and reported in Table 1. Extraction efficiencies were high

for all combinations of dye and IL, with the exception of cota in[C4MPy][Tf2N]. These high extraction efficiencies comple-ment previous results showing that anionic, neutral, andcationic dyes are all extracted with high efficiency into[R3R′P][Cl].13 Taken together, the data demonstrate very

Figure 3. XRD patterns for untreated wool and wool dissolved in ionicliquid. Arrows indicate peak centers of canonical α-keratin (d = 5.15and 9.80 Å) and β-keratin (d = 4.65 and 9.80 Å).

Figure 4. Disruption of wool keratin structure as determined by FT-IR.

Table 1. Extraction Efficiencies (%) Found for Commercialand Natural Dyes Dissolved in Water and Extracted intoOne of Five Ionic Liquids

dye[(C4H9)4P]

Cl [C4MPy]·[Tf2N][R3R′P]

Cl[R3R′P]

Br[R3R′P][Tf2N]

yellow 95.2 97.5 98.9 92.5 98.7brown 92.3 93.0 97.1 98.5 98.5orange 95.5 86.7 100 95.3 99.1cota 100 0.00 93.9 96.5 n.d.

Analytical Chemistry Article

dx.doi.org/10.1021/ac301873s | Anal. Chem. XXXX, XXX, XXX−XXXD

efficient partitioning of highly conjugated dye molecules intothe ILs chosen for our analytical method.Ultimately, this extraction technique can be used for the

sampling of wool fibers followed by direct evaluation of theextraction solution by MALDI-MS mass spectrometry. MALDI-MS instruments are in use in forensic laboratories and havebeen applied to dye analysis for pen ink.14,15 We havepreviously reported that ionic liquids are useful for MALDI-MS analysis of organic dyes.13 For analysis of dyes extractedfrom wool, samples are analyzed directly from an ionic liquidsolution without requiring additional matrix elements to assistin the ionization mechanism (Figure 5, and SupportingInformation Figures S-3−S-6). Most wool dyes are alreadycharged and are almost always anions. The negative chargegenerally arises from sulfonate functional groups, which interactstrongly with acid-pretreated wool. Because of this inherentcharge, we find that wool dye samples do not require theaddition of a standard matrix compound, such as 2,5-dihydroxybenzoic acid (DHB), which forms cations byprotonation of neutral or anionic analytes, or 9-aminoacridine(9AA), which forms anions by deprotonation of acidic analytes.We have found that analysis of the anionic wool dyes, whichabsorb to some degree near the wavelength of the MALDI-MSlaser, is hindered by the addition of 9-aminoacridine or 2,5-dihydroxybenzoic acid to the sample. A more robust analysiscan be achieved by spotting samples of the neat ionic liquidextraction solution without the additional matrix components.Another advantage of ionic liquids for this application is that

they are nonvolatile and do not evaporate to inhomogeneouscrystalline domains on the analysis plate during measure-ments.16,17 The commonly used solid matrices are generallydissolved in a volatile solvent such as methanol or water tofacilitate sample plate spotting, resulting in heterogeneouscrystalline spots when the solvent dries. Use of nonvolatile ILsassures spot-to-spot reproducibility from the homogeneousdistribution of the analyte and reduces the need for many lasershots at a single sample.16,17 The ionic liquids are notfragmented in the soft MALDI-MS technique, as shown bythe negative ion mode, IL-only, spectrum in SupportingInformation Figure S-7, minimizing interferences from the ILmatrix in the spectrum.To examine the potential of direct MALDI-MS analysis of

IL-dye solutions extracted from wool, we tested solutions ofknown dyes in ionic liquid, and compared those to solutions ofdyes extracted from wool fibers. We used commerciallyavailable kits of yellow, brown, and orange dyes, as well asthe natural dyes from cota and sandalwood. These dyes aremixtures of several dye molecules, and we used obtainedMALDI-MS and high pressure liquid chromatography-electro-spray ionization time-of-flight mass spectrometry (LC/MS-TOF) “fingerprint” of these dye mixtures and to identifycomponents from a list of 11 possible ingredients. Under highMALDI-MS laser power, dye peaks were observed at massesequal to [M]−, as well as occasional peaks for [M − SO3]

−. Theloss of sulfate upon ionization by the MALDI laser is known forseveral classes of dyes and can facilitate dye identification.14

The yellow dye contains several acid dyes of yellow andorange colors (Table 2). Analysis of the extraction solution byMALDI exhibited mass peaks associated with acid yellow 49and acid orange 7. ESI analysis of the extraction solutions alsoshowed the acid yellow 49 component but not the acid orange7. For both techniques, the extraction solutions were used

directly, thereby minimizing sample workup and processing orfurther techniques to isolate the analytes prior to analysis.The orange dye mixture was found to contain acid yellow 49

and acid orange 7, as well as acid red 151 and acid red 4 (Table3). The brown dye mixture contained acid orange 7, acid blue62, and acid red 151 (Table 4). Only the acid orange 7 signalappeared in the ESI.Cota, also known as Navajo Tea, is a traditional dye extracted

from a plant (Thelesperma megapotamicum) that grows in aridclimates. Several flavonoid molecules have been previouslyidentified in the extracted dyes, including luteolin, querciting,marein, and 2′,3,3′,4,4′-pentahydroxychalcone.18 These con-stituents, identified as likely dye components, are all neutraland, although they weakly absorb light at the wavelength of theMALDI laser, do not fly well. We have tentatively assignedsome peaks as belonging to quercitin (MW = 302 g/mol),which is a flavonoid that exists in many plant species. There arealso higher molecular weight species that are yet to beidentified, and require possibly MS-MS to fully characterize.Our method relies on the fact that most wool dyes are anions,and we recognize the limitation that neutral dyes, which

Figure 5. Negative mode ESI-TOF (top, A) and MALDI-MS (bottom,B) spectra of orange dye extracted from wool. Peaks in (A) areidentified (m/z) as: 327.0500, Acid Orange 7 [M]−; 424.0098, AcidYellow 49 [M]−. Peaks in (B) are identified (m/z) as acid orange 7 [M− SO3]

−, 248.18; acid red 4 [M − SO3]−, 278.21; acid orange 7 [M]−,

327.00; acid red 151 [M − SO3]−, 352.16; and acid yellow 49 [M]−,

423.99. Calculated m/z values are given in Table 3.

Table 2. Analysis of Yellow Commercial Dye by MALDI andESI-MS

dye componentobserved MALDI ILextraction solution

observed ESIextraction solution

acid yellow 49 [M]−

calculated = 424.00379[M]− = 423.99 [M]− = 424.0128

acid orange 7 [M]−

calculated = 327.04395[M]− = 327.00

Analytical Chemistry Article

dx.doi.org/10.1021/ac301873s | Anal. Chem. XXXX, XXX, XXX−XXXE