Advancing the Forensic Analysis of Dyed Fibers by Time-of-Flight Mass Spectrometry

Materials

Fabrics used in this work were a scoured 100% polyester knit (Guilford Inc.) and a 100% spun nylon 6.6 (Testfabrics #361). Water for HPLC-TOF MS analysis was 18 MΩ from a Siemens Elga Purelab Ultra purification system. HPLC grade acetonitrile and methanol were from Burdick and Jackson. Formic acid and ammonium acetate (99.9%) were from Sigma Aldrich. Acetic acid was from Termo Fisher. Commercial dye samples were obtained from various manu-facturers and were used without purification.

Apparatus

Standard dyeing procedures ⁸ were used to dye fabrics in either a Datacolor Ahiba Nuance Top Speed II dyeing machine or a Roaches Pyrotec dyeing machine. A Pierce Reacti-Term Heating Module with an aluminum block was used for dye extraction in Termo Scientific (# 13220) Reacti-Vials. ⁸

Instruments

An Agilent 1200 LC SL with diode array detection (DAD) and 6520 Q-TOF MS with an electrospray interface were used for HPLC-TOF MS analysis. For disperse dyes, the mass spectrometer was operated in positive-ion mode, and in negative-ion mode for acid dyes with a capillary voltage of 4 k V, fragmentor voltage of 110 V, nebulizer pressure of 35 psig, and a drying gas flow rate of 12 L/min at 350 °C.

A TOF SIMS V (ION TOF Inc.) instrument equipped with a Bi _n ^m+ (n = 1 to 5, m = 1 or 2) liquid metal ion gun were used for TOF SIMS analyses. For the mass spectral images acquired in this study, a 256 × 256 pixel image of a 100 × 100 μm area was acquired using a Bi₃⁺ primary ion beam. An electron gun was used to prevent charge buildup on the insulating sample surfaces. The total accumulated primary ion dose for an image acquisition was less than 1 × 10¹³ ions/cm², an amount of ions that is within the static SIMS regime. C₆₀⁺ sputter cleaning and damage removal and Bi₃⁺ data acquisition were used to allow signal averaging to improve the signal-to-noise ratio. The analysis protocol consisted of sputtering the dyed nylon surface for 14.6 s with a 1 nA C₆₀⁺ ion beam, followed by spectrum acquisition with Bi₃⁺ at 1 shot/pixel for 20 frames. The C₆₀⁺ sputtered area and Bi₃⁺ analysis areas were 300 × 300 μm and 100 × 100 μm, respectively. This sputtering/acquisition sequence was then repeated until a useful signal-to noise ratio (S/N) was achieved or until the sample was consumed.

Methods

Extractions were performed on small (100–200 μg) samples of yarns pulled from the dyed fabrics. PET fibers were placed in a Reacti-Vial, 1,2-dichlorobenzene was added (typically 200 μL), followed by heating for 30 min at 120 °C. The solvent was then blown of from the vial with nitrogen, followed by dissolution in initial HPLC mobile phase (acetonitrile:water, 48:52), (200 μL). The solutions were filtered through Millex-GV 13 mm, 0.22 μm PVDF filters (Termo Fisher). Dry, disperse dye powders were dissolved in the initial mobile phase, mixed for 10 sec at 3000 rpm in a VWR vortex mixer, filtered, and diluted to give 5 ppm solutions for HPLC-MS analysis. Acid dye powders were dissolved in water, filtered, and diluted to 30 ppm prior to analysis. Each dye sample was spiked with 10 ppm uracil as an indicator of a completely unretained analyte.

For disperse dyes, a Poroshell 120 EC C18, (3.0 × 100 mm, 2.7 μm) column at 45 °C was used. For acid dyes, a Zor-bax Eclipse Plus C18 (2.1 × 50 mm, 3.5 μm) column with a Zorbax Eclipse Plus C18 narrow-bore guard column (2.1 × 12.5 μm, 5 μm) was used.

Mobile phases were water (A) and acetonitrile (B) (with 0.1% formic acid added to both for positive ion MS) for disperse dyes and 20 mM ammonium formate adjusted to pH 4 with formic acid (A) and 70:30 methanol:acetonitrile (B) for acid dyes. Gradients for the two methods are shown in Tables I and II.

OPEN IN VIEWER

Table I

Disperse Dye Gradient at 0.5 mL/min

Time (min)	%B
0	48
8	70
12	70
12.5	48
19.5	48

OPEN IN VIEWER

Table II

Acid Dye Gradient at 0.25 mL/min

Time (min)	%B
0	5
2	5
9	90
19	90
19.5	5
25.5	5

A hydraulic press was used to compress dye powder and swatches of dyed nylon fabrics to achieve relatively fat surfaces for efficient secondary ion extraction and improved mass resolution. Cross sections of dyed fibers, with a thickness of ∼500 nm to 700 nm, were prepared by cryo-microtomy. Fibers were embedded in Eponate 12-Aaraldite 502 (Ted Pella) in silicone molds. Embedded samples were then cured over 12 h at 45 °C, 24 h at 60 °C, and then for 8 h at 70 °C, followed by an oven cool down. Cross sections were then cut on a Leica UC7 (Leica Microsystems) with a cryo-attachment at a temperature of –40 °C using a 45° cryo-diamond knife. The cryo-temperatures were necessary to avoid the smearing of both the fiber surface and the embedding medium, preventing distortion or contamination of the fibers.

HPLC-TOF MS Analysis of Acid Dyes

Retention and separation of anionic acid dyes on a reverse phase HPLC column is more difficult than separation of neutral disperse dyes because the charged dye is not attracted to the non-polar octadecylsilyl stationary phase. To overcome this limitation, buffers and ion pairing agents can be used. In this work, adequate separation was provided by using a pH 4 formic acid/ammonium formate buffer. Fig. 1 shows mass spectra extracted from the total ion chromato-gram of three separate dyes, C.I. Acid Blue 113, C.I. Acid Blue 25, and C.I. Acid Green 16.

OPEN IN VIEWER

Fig. 1

Mass spectra of (a) Acid Blue 113, (b) Acid Blue 25, and (c) Acid Green 16.

In the dye database under development, dye identification was accomplished by using retention time (indexed to the unretained uracil), visible spectrum, and exact mass.

HPLC-TOF MS Analysis of Disperse Dyes

Fig. 2 shows an example chromatogram (a) showing three peaks absorbing in the blue region and a representative mass spectrum (b) from the peak at 7.53 min for a C.I. Disperse Blue 60 sample, included because it is the only test disperse dye comprised of three dyes.

OPEN IN VIEWER

Fig. 2

HPLC (DAD) chromatogram of (a) three components of Disperse Blue 60 at 660 nm, and (b) mass spectrum of one component of Disperse Blue 60 in one sample (7.53 min).

As can be seen from Table III, Disperse Blue 60 existed in several different compositions. Peak areas are calculated from HPLC (DAD) chromatograms (absorbances). The far right column indicates the manufacturer. The dyes from each manufacturer are grouped by rows of the same color. Four different structures, indicated by retention times in blue columns, existed, all of which were uniquely identified by MS at m/z = 380, 394, 408, and 424. The composition also varied for samples from the same manufacturer. The variability in these samples was real and probably a function of the manufacturing process. As a result, the initial goal of assigning a dye from a forensic sample to a specific manufacturer may not always be possible. Note that the sample from Manufacturer 5 contained three components (Fig. 2).

OPEN IN VIEWER

Table III

Composition of Disperse Blue 60 Samples from Six Manufacturers ^a

Peak 1 RT	Peak 1 m/z	Area %	Peak 2 RT	Peak 2 m/z	Area %	Peak 3 RT	Peak 3 m/z	Area %	Peak 4 RT	Peak 4 m/z	Area %	Man.
4.11	424.151	100	4.25	380.124	85							1
4.11	424.150	82	4.24	380.123	100.0							1
4.11	424.150	100	4.24	380.124	956							1
4.11	424.150	68	4.24	380.124	100							1
4.11	424.150	51	4.24	380.124	100							1
4.11	424.150	100	4.24	380.124	70							2
4.11	424.150	84	4.24	380.124	100							2
4.11	424.150	100	4.24	380.123	84							2
4.11	424.151	100										2
			4.23	380.124	100							3
			4.24	380.123	100							3
			4.24	380.123	100	4.60	394.139	68.4				4
			4.24	380.124	100	4.60	394.140	87.6	4.89	408.155	75.1	5
			4.24	380.124	100	4.60	394.139	65.2				6

a

RT = retention time, m/z = mass to charge ratio of the ion, and Man. = manufacturer.

To illustrate the utility of HPLC-TOF MS in the identification of disperse dyes extracted from automotive fabrics, a few fibers from a sample of black fabric from a 2011 Ford Fusion ¹² were extracted.

From the TOF MS data, three of the HPLC (DAD) peaks in Fig. 3a can be assigned to Disperse Blue 60 (7.69 min), Disperse Blue 73 (8.99 min), and Disperse Blue 77 (10.25 min). The chromatogram extracted at 540 nm (see Fig. 3) is more complicated because several of the dyes in the extract absorbed at this wavelength. The peak at 7.70 min eluted at the same time as Disperse Blue 60, but was actually Disperse Red 86 as shown by the extracted ion chromatogram (EIC) and exact mass (m/z = 423.103). An EIC for m/z = 380.124 gave a peak for Disperse Blue 60. In this case, the retention times alone were inadequate for identification, but both the DAD visible spectra (red and blue) and exact masses would conclusively discriminate between these two possibilities. The peak at 8.40 min was the major yellow dye as shown by a single peak in the 410 nm extracted chromatogram. This dye gave m/z = 378.1199, which corresponded to a molecular formula of C₁₉H₁₅N₅O₄.

OPEN IN VIEWER

Fig. 3

Chromatogram of dyes extracted from a 2011 Ford Fusion fabric at (a) 660 nm, and (b) 540 nm.

TOF SIMS Analysis of Acid Dyes

The molecular ion of C.I. Acid Blue 25 was easily detected by TOF SIMS from 1% owf dyed nylon 6.6 cross sections (Fig. 4a). However, it was hardly detectable from the 1% owf dyed nylon surface (Fig. 4b). The low S/N value might be due to surface contamination on the dyed nylon surface.

OPEN IN VIEWER

Fig. 4

Negative ion TOF SIMS spectra showing the molecular ion of Acid Blue 25 obtained from (a) 1% on-weight-of-fabric (owf) dyed nylon cross section, and (b) 1% owf dyed nylon surface.

To improve the S/N value, we developed an analytical pro-tocol adapted to the removal of both surface contamination and at least partial Bi ion beam induced damaged material. ¹³ The C₆₀⁺ sputtering beam and the Bi₃⁺ analysis ion beams were used alternatively and in a cyclic manner to remove damage and to acquire data. Fig. 5 shows the negative ion TOF SIMS spectra acquired from the dyed nylon surface (a) with only Bi₃⁺ ion beam data acquisition, and (b) with C₆₀⁺ion beam sputtering followed by spectrum acquisition with Bi₃⁺ in a cyclic manner. The results clearly show the signifi-cant S/N improvement with cyclic C₆₀⁺ sputtering.

OPEN IN VIEWER

Fig. 5

Negative ion TOF SIMS spectra showing the molecular ion of Acid Blue 25 obtained from 0.5% owf dyed nylon surface (a) without C₆₀⁺ sputtering, and (b) with cyclic C₆₀⁺ sputtering.

TOF SIMS of Disperse Dyes

Both the surface and cryo-microtomed fiber cross sections dyed with Disperse Blue 60 were analyzed by TOF SIMS directly without extraction. Disperse Blue 60 molecular ions were all observed for 7% to 0.1% owf dyed polyester (PET) fabrics surfaces (Fig. 6).

OPEN IN VIEWER

Fig. 6

TOF SIMS spectra of Disperse Blue 60 dyed PET fabric surfaces with varied dyes (owf) showing the molecular ions at m/z = 379 and 423.

The molecular ions at m/z = 379 and at m/z = 423 for the two components in Disperse Blue 60 were selected to test quantification of this dye in PET. All spectra were first normalized to an internal reference at m/z = 357, a molecular fragment ion attributed to the PET polymer, to allow direct comparison of the mass spectra. The normalized molecular ion intensity of Disperse Blue 60 on fabric surfaces versus various dye concentrations (owf) is shown in Fig. 7a. Note that the molecular peaks at m/z = 379 and m/z = 423 gave a nearly linear relationship with dye concentration. It should be noted that the detection limit of TOF SIMS is highly dependent on the materials analyzed, especially for organic species. Since the ion yield varies for different elemental and molecular species, the response may not always be linear.

OPEN IN VIEWER

Fig. 7

Normalized Disperse Blue 60 molecular ion intensities at m/z = 379 and m/z = 423 dye components from fabric surfaces versus dye loading. (a) the first set of Disperse Blue 60 dyed fabrics analyzed. (b) the second set of Disperse Blue 60 dyed fabrics analyzed.

The above experiment was repeated on a second set of Disperse Blue 60 dyed PET fabrics and the results obtained from the various dye concentration shown in Fig. 7b were consistent with the first set of dyed fabrics.

Resin, PET fiber, and Disperse Blue 60 in the cross section were unambiguously differentiated by TOF SIMS using their characteristic ions. Fig. 8 shows the chemical mapping of resin, fiber, and dye in the left column. Tese TOF SIMS images are color encoded and then overlaid, which are shown in the column on the right. The overlaid image shows the relative distributions of resin (red) and PET fiber (green) over the cross section analyzed, which provided clear delineation of the PET fibers. Secondary ions attributed to the Disperse Blue 60 dye (blue encoded) clearly coincided with the position of the fibers.

OPEN IN VIEWER

Fig. 8

TOF SIMS secondary ion images of resin (red), PET (green) and Disperse Blue 60 dye (blue) are color encoded and overlaid to produce the images in the right column. The signal from Disperse Blue 60 clearly overlays with the PET.

Fig. 9 shows TOF SIMS images for Disperse Blue 60 molecular ion with varied dye concentration. Significant Disperse Blue 60 molecular ion intensity was observed for the fibers down to 0.5% owf dyed PET cross sections (Fig. 7). However, for the 0.1% owf dye PET cross section, the dye signal was difficult to distinguish from the background, similar to that found in the undyed control sample. The data suggest that the TOF SIMS detection limit for the cross-sectioned Disperse Blue 60 PET was as low as 0.5% owf.

OPEN IN VIEWER

Fig. 9

The top row contains TOF SIMS images of the m/z = 379 (molecular ion) of Disperse Blue 60 for the dye concentration given. The lower row of images has the dye (encoded in blue) overlaid onto the TOF SIMS secondary ion image of the embedding resin.