Analysis of trace fibers by IR-MALDESI imaging coupled with high resolving power MS

Abstract

Trace evidence is a significant portion of forensic cases. Textile fibers are a common form of trace evidence that are gaining importance in criminal cases. Currently, qualitative techniques that do not yield structural information are primarily used for fiber analysis, but mass spectrometry is gaining an increasing role in this field. Mass spectrometry yields more quantitative structural information about the dye and polymer that can be used for more conclusive comparisons. Matrix-assisted laser desorption electrospray ionization (MALDESI) is a hybrid ambient ionization source being investigated for use in mass spectrometric fiber analysis. In this manuscript, IR-MALDESI was used as a source for mass spectrometry imaging (MSI) of a dyed nylon fiber cluster and single fiber. Information about the fiber polymer as well as the dye were obtained from a single fiber which was on the order of 10 μm in diameter. These experiments were performed directly from the surface of a tape lift of the fiber with a background of extraneous fibers.

Introduction

Forensic evidence plays a vital role in many criminal cases. Trace evidence, especially, can be used to establish contact between two objects, people, or locations. Textile fibers are among the more commonly observed forms of trace evidence. A variety of analytical methodologies are currently used for fiber analysis and are primarily directed towards identifying the fiber type. Methods such as FT-IR spectroscopy, microscopy, and UV-visible microspectrophotometry are used to assess optical properties of the fiber [1]. In addition to identifying the fiber, the analysis of the dye from these fibers can provide further discrimination. Separation methods such as thin layer chromatography, liquid chromatography, and capillary electrophoresis are commonly combined with some form of spectrophotometric detection to help support dye identification from dye extracts from the fiber. However, the amount of dye that is typically extracted from trace fibers is on the nanogram of dye per millimeter of fiber scale which requires that these methods be reasonably sensitive [2]. While these analyses are good for supporting identification, the use of retention time or retention factor even when combined with microspectrophotometry is typically not specific enough to make a unique identification of the dye(s) [3]. For this reason, mass spectrometry (MS) is playing an increasing role in trace fiber analysis due to its exceptional selectivity and sensitivity [4].

A typical mass spectrometric fiber analysis involves extracting the dye from the fiber using an appropriate extraction solvent followed by separation and MS detection [2, 5–8]. These extraction and separation steps, however, can be time consuming, thus there have been investigations into the use of direct analysis techniques to allow for obtaining MS information with minimal sample preparation. Laser desorption photoionization (LD-PI) [9], direct analysis in real time (DART) [10], and matrix-assisted laser desorption electrospray ionization (MALDESI) [11] have each been used for direct sampling of dyestuffs from textile fabric/fibers. The unique attributes of direct analysis MS techniques has also found application in other areas of forensic science [12–14].

MALDESI is a hybrid ambient ionization source that combines features from matrix-assisted laser desorption ionization (MALDI) and electrospray ionization (ESI) [15]. In general, MALDESI utilizes a pulsed laser to resonantly excite a matrix (endogenous or exogenous) to facilitate sample desorption. The desorbed sample partitions into the charged solvent droplets of an electrospray plume and is then post-ionized by an ESI-like process. The wavelength of laser light that is used for desorption can be tuned or chosen to correlate with a strong absorber in the sample. The use of an infrared (IR) laser allows for the use of water as a matrix which comes with the inherent benefit of having little to no matrix background interference in the mass spectra. The IR-MALDESI source was recently configured for performing imaging experiments and the use of ice as a laser energy absorbing matrix was explored [16, 17]. Also, this source has been coupled with the Thermo Q Exactive allowing for fast, high resolving power imaging data to be collected through a variety of different acquisition modes [18].

Mass spectrometry imaging (MSI) is a technique whereby a surface is analyzed as a two-dimensional array of points. At each point, a mass spectrum is generated. By relating the intensity of any observed ion to its position, a heatmap can be generated that represents the ions distribution on the analyzed surface. Early accounts of using MALDI for spatial imaging have been presented by Spengler [19] and Caprioli [20]. This unique combination of molecular specificity with spatial information has permitted MSI to be practical for a number of different applications in biological and clinical research [21–26]. In addition, a few forensic applications have also been presented in the literature. Desorption electrospray ionization (DESI) has been used to investigate inks and dyes for forensic document analysis [27], as well as latent fingerprint imaging [28]. A recent publication demonstrated the use of secondary ion mass spectrometry (SIMS) for imaging the distribution of dyes within single nylon fibers [29].

One of the most commonly used technique for the recovery of textile fiber evidence is through the use of tape lifts [30–33]. Other techniques, such as hand retrieval with tweezers, often involve transferring any fibers picked up with tweezers to tape for storage and later analysis [34]. A good tape lifting tape has good uptake, low saturation, a good recovery rate, and offers ease of analysis [35]. Given that adhesive tape is commonly used for recovery or storage of fiber evidence, it was the goal of the authors to directly analyze textile fibers from a tape lift to simulate a “real-world” sample matrix. Herein, we describe the application of IR-MALDESI MSI to the analysis of textile fibers directly from adhesive tape. The use of polarity switching for this type of analysis is also presented. Polarity switching is useful for this type of analysis because it allows for observation of particles with different ionization characteristics. Some dyes, such as acid dyes, are anionic and thus observable in negative ionization mode, while basic dyes are cationic and observable in positive ionization mode [7]. Previous experiments demonstrated proof of principle for analyzing the dyes directly from fabric [11]; however, this manuscript pushes towards a more relevant single-fiber analysis.

Experimental

Materials

Acid Black 58 was purchased from Ciba-Geigy (Dover Township, NJ) and Basic Violet 16 was purchased from Clariant (Charlotte, NC). HPLC-grade water and methanol were purchased from Burdick and Jackson (Muskegon, MI, USA) and ammonium hydroxide was purchased from Sigma-Aldrich (St. Louis, MO). 3M double-sided Scotch tape was purchased from a local office supply store.

Dyeing procedure

Aqueous stock solutions (1 mg/mL) of the commercial Acid Black 58 dye powder were diluted to produce a 300-mL dyebath for dyeing at 1 % on weight of fiber (owf) on 3.00 g nylon 6.6 (Dupont, Spun Nylon 6.6, Type 200) then acidified to pH 4 using glacial acetic acid. The dye and fabric were placed into beakers and inserted into the Pyrotec MB2 (Roaches, UK) dyeing machine. The dyebath was heated to 100 °C at a ramp rate of 4 °C/min and was held at this temperature for 60 min. Dyed fabric was then rinsed with copious amounts of cold water and dried.

For dyeing acrylic fiber with Basic Violet 16, 10 g of 100 % woven acrylic fabric (Testfabrics, Inc.) was dyed to 0.5 % (owf) Sevron Brilliant Red 2B (C.I. Basic Violet 16) at pH 4.0–4.5 by addition of acetic acid in a total solution of 1,200 mL. The dyebath temperature was raised to 95 °C at a rate of approximately 1 °C/min and maintained at this temperature for 20 min. The dyed fabric was then rinsed with copious hot and cold water, and then dried.

IR-MALDESI source

The IR-MALDESI source has been described in greater detail in previous publications [16, 17, 36]. In short, a thermoelectric sample stage is used to cool the sample to −10 °C while open to the ambient environment in order to deposit a layer of ice over the sample surface. This ice/frost layer resonantly absorbs the laser energy from a tunable wavelength IR laser (IR Opolette, Opotek, Carlsbad, CA) which facilitates sample desorption. The desorbed sample then partitions into an electrospray plume and becomes ionized by an ESI type process. The geometric parameters of the source have recently been optimized for imaging applications [17] and these conditions were used for all of the experiments in this manuscript unless stated otherwise. The composition of the electrospray solvent was 50 % aqueous methanol with 0.1 % ammonium hydroxide which admittedly favors the negative ion mode. The IR-MALDESI source was synchronized with the Q Exactive mass spectrometer such that a single acquisition mass spectrum was acquired at each pixel. This was achieved through coordinating the triggering of the laser with the trapping and acquisition of the ions. All imaging experiments were acquired with two pulses per pixel (at 20 Hz) and a step size of 100 μm in both the X and Y directions. For single polarity experiments, this resulted in pixel dimensions of 100 μm by 100 μm. For polarity switching experiments, adjacent pixels were acquired with different polarities so the pixel dimensions for a single polarity are shown as 200 μm in X and 100 μm in Y. A workflow of the polarity switching experiment is demonstrated in Fig. 1. As shown, once the dataset is filtered by polarity, the dimensions of the ion image pixels end up being twice as large in the X dimension compared to the Y dimension.

Fig. 1.

Outline of the polarity switching imaging experiment which demonstrates that adjacent pixels are collected in either positive or negative ion mode. Because of this, the actual resolution in the X (horizontal) dimension is two times larger than in the Y (vertical) dimension once the individual polarity datasets are separated. The size of the pixels is shown larger than actual to demonstrate the concept

Q Exactive mass spectrometer

The automatic gain control (AGC) feature of the Q Exactive measures ion flux and varies ion injection time to accumulate an optimal number of ions for the orbitrap acquisition. Due to the pulsed nature of the IR-MALDESI source, this feature could not be used and thus the AGC was turned off. The possible consequences of turning off the AGC include issues with ion suppression and shifts in mass accuracy due to the differing number of ions injected into the orbitrap from one acquisition to the next which would affect the external mass calibration. By using lock mass, the mass accuracy appears to be maintained within 2 ppm for these experiments. When the AGC is turned off, ions are accumulated for the maximum ion injection time. The maximum injection time for all experiments was set at 150 ms in order to accumulate ions generated from the two laser pulses at each pixel (two pulses at 20 Hz ~50–100 ms).

For the single polarity experiments, the polarity was set as either positive or negative. The resolving power of the orbitrap acquisition was set to 140,000 at m/z 200 with a mass range of 250 to 1,000. With the polarity switching experiments, an instrument method was created to acquire spectra with alternating polarities. The resolving power for these experiments was also set at 140,000 at m/z 200 with a mass range of 250–1,000m/z.

Sample preparation and data analysis

For the polarity switching experiment, a single fiber and fiber cluster (a bundle of roughly 100 single fibers that are wound together to make the yarn/thread of the fabric) were removed from both of the dyed fabrics (acid black 58 dyed nylon 6,6 and basic violet 16 dyed acetate) and placed on the double-sided tape. The double-sided tape was then adhered to a standard glass microscope slide. Optical images were obtained before and after deposition of the ice matrix. For the single polarity experiment, several single acid black 58 dyed nylon 6,6 fibers were randomly placed onto a section of the double-sided tape. An optical image was obtained to point out the location of these target fibers. The double-sided tape was pressed against the carpet to extract extraneous fibers to the tape lift. This tape lift was then adhered to a glass slide prior to analysis. These single target fibers were determined to be roughly 10 μm in diameter.

Once placed on the IR-MALDESI sample stage, a thin layer of ice was deposited over the surface of the tape lift to act as the laser energy absorbing matrix. A region of interest was then defined to incorporate a portion of the tape lift. The resulting raw MS data (.raw format) was converted to mzXML with the MSConvert from Proteowizard [37] using a polarity filter to separate the positive and negative acquisitions for the polarity switching experiment. The mzXML file was then directly loaded into the new standalone version of MSiReader V0.04 for data analysis and ion map generation [38, 39]. Ion maps were generated using an m/z bin width of 5 ppm and no normalization or interpolation was used in order to demonstrate the true quality of the raw data.

Results and discussion

IR-MALDESI MSI with polarity switching

In a previous publication, Cochran et al. demonstrated that mass information about the fiber polymer and dye could be obtained by directly analyzing dyed fabric using the IR-MALDESI source coupled to an LTQ-FT mass spectrometer [11]. These experiments used relatively large samples of fabric (1 cm by 1 cm) to demonstrate proof of principle. However, such large samples are not commonplace in forensic cases, therefore the goal of this manuscript is to determine how small of a sample can be used to still obtain specific m/z information about the dye and fiber polymer down to the single fiber level. To test this hypothesis, a sample was prepared on double-sided tape with progressively less material starting with a single thread/yarn, a bundle of roughly 100 single fibers, and ending with a single fiber. Two different dye classes and fiber types were initially investigated including a basic dye (basic violet 16) in acetate fiber and an acid dye in nylon 6,6 fiber (acid black 58). Since the basic dye is more apt to be ionized in positive ion mode and the acid dye would be more likely observed in negative ion mode, a polarity switching imaging experiment was performed in order to collect information from both polarities. For this polarity switching imaging experiment, adjacent pixels are analyzed with opposing polarities as outlined in Fig. 1.

The results of the polarity switching imaging experiment are outlined in Fig. 2. The optical images before and after deposition of the ice matrix are shown in Fig. 2a. Ion maps of acid black 58 (Fig. 2b, c) and the nylon 6,6 polymer (Fig. 2d–g) are also presented. Acid black 58 belongs to a class of 1:2 metal complex dyes that is typically coordinated to chromium. This coordination typically improves the dyeing characteristics of the dye and increases the affinity for polyamide fibers (like nylon). While the ion map for this 1:2 chromium complex was observed to have a distribution (data not shown) that is similar to the one shown in Fig. 2c, the monomer of the dye was much more abundant. Given that these fabrics were dyed over 3 years ago, it is difficult to determine whether this observation is a result of in-source fragmentation of the metal complex or if it is merely due to degradation [40]. While it may be proposed that absorption of the laser energy that is used for sample desorption may increase the internal energy of the ions and thereby induce fragmentation, recent evidence suggests that ion internal energies of IR-MALDESI are nearly identical to those of soft ionization in ESI [18, 41]. The monomer of acid black 58 was observed in both positive and negative ion mode as the protonated (Fig. 2b) or deprotonated (Fig. 2c) form, respectively. In addition to observing ions relating to the dye, several ions were also observed with similar distributions that related to the nylon 6,6 polymer. Figure 2d–g represent ion maps of the protonated or deprotonated forms of the nylon 6,6 dimer or trimer (n=2 and n=3, respectively). One important observation from this experiment is that dye and polymer information could be obtained from a single fiber that is on the order of 10 μm in diameter.

Fig. 2.

Results from the polarity switching IR-MALDESI MSI experiment. The numbers above each image are m/z values. a Optical images of the thread, fiber bundle, and single fibers for basic violet 16 in acetate and acid black 58 in nylon 6,6 both before and after deposition of the ice matrix. b, c Ion images of the protonated and deprotonated (respectively) monomer of the acid black 58 dye. d–g Ion images of the protonated and deprotonated form of the nylon 6,6 polymer (d–e are n=2 and f–g are n=3)

While dye and polymer information were observed in both polarities for acid black 58 in nylon 6,6, there were no observed ions that had unique distribution correlating to the basic violet 16 dye or the acetate fiber polymer. A potential drawback or limitation that was observed in these experiments is that even when performing polarity switching experiments, the electrospray solvent composition would be optimized for one polarity or the other. Perhaps by using a more acidic solvent, the protonation would be more favored in the positive polarity experiment.

IR-MALDESI MSI of single fibers on tape lift

After demonstrating that samples as small as single fibers are capable of providing reasonable signal for the dye and fiber polymer through direct analysis with IR-MALDESI, the next goal was to simulate the analysis of single fibers from a “real-world” sample matrix. To simulate analysis from a tape lift, single acid black 58 nylon 6,6 fibers were randomly placed onto double-sided tape on a glass microscope slide and their location was noted in the optical image in Fig. 3a. This sample was then pressed onto a section of a carpet to extract extraneous fibers to simulate a “real-world” sample matrix (Fig. 3a (after tape lift)). The sample matrix consisted of the extraneous rug fibers, hair, and dirt. The tape lift was taken from a rug in the mass spectrometry facility. This tape lift was then analyzed by IR-MALDESI imaging in the negative ion mode.

Fig. 3.

Results from the analysis of single acid black 58 fibers from a tape lift. The numbers above each image are m/z values. a Optical images of the nylon fibers before and after performing a tape lift from a carpet. b Ion image of the deprotonated monomer of the acid black 58 dye along with an overlay of the ion image with the optical image to show the correlation of the dye signal with the presence of the fiber. c, d Ion images of the deprotonated nylon 6,6 polymer (c is n=2 and d is n= 3)

The results of the single fiber tape lift are shown in Fig. 3. The optical images of the sample before and after performing the tape lift are provided in Fig. 3a. The monomer of acid black 58 that was observed in the polarity switching experiment was also detected to relate to the position of the single nylon fibers from this tape lift experiment (Fig. 3b). An overlay of this ion image and the optical image is also provided to demonstrate the correlation of the dye distribution with the location of the single fibers. The fiber polymer peaks were found to have the same distribution as the ion map of the dye which is expected (Fig. 3c). There are a couple of instances where a portion of the desorbed fiber stuck to the electrospray emitter resulting in a continuous ionization of the dye and fiber polymer. This is demonstrated in the ion maps where there are horizontal lines of slowly diminishing ion intensity. While the geometric parameters of the source were not specifically optimized for this application, the results from optimization for direct analysis of droplets [36] and for biological tissue [17] indicate that the small working distance is optimal as it is likely the smaller particles which are ultimately ionized. This observation precludes the ability to increase the distance between the ESI emitter and sample to reduce occasional carry-over from desorbed fiber attaching to the emitter. This type of carry-over could be reduced or even eliminated through the use of remote sampling techniques [42].

Conclusion

Proof of principle has been previously established for IR-MALDESI as a technique for MSI and for the direct analysis of samples of textile fabric. By performing MSI on much smaller samples, this technology was found to be capable of determining the dye m/z and spatial distribution and fiber type from samples as small as single fibers. The characterization obtained from MSI does not identify the dye directly because no database has yet been developed. This information, however, can be used for exclusion purposes. A polarity switching experiment was used to image samples in both positive and negative polarity in a single experiment. This technique allowed for the observation of species that may preferentially ionize in either mode. It is also a less destructive technique than extraction followed by LC-MS as well as yielding spatial characteristics of the dye in the fiber. In order to simulate a “real-world” sample containing a mixture of fibers, single fibers were collected on a tape lift along with other extraneous fibers. It was determined that the tape sample matrix did not interfere with the direct analysis of the tape lift, as dye and fiber mass information were observed to have identical distributions that correlated with the position of the target fibers. The results of these experiments are quite exciting. Because dye and fiber polymer signal was observed from individual pixels, this implies that this ionization technique coupled with a high-resolution, accurate-mass instrument (Q Exactive) is sensitive enough to detect dye and fiber polymer from a 10-μm-diameter fiber that is as small as 100 μm in length. The fiber incurs some damage during the laser pulsing, but the damage is minimized with the use of ice as an exogenous matrix.