New mass spectrometry-based methods for organic dyes analysis for forensic trace evidence and conservation of cultural heritage objects

(1)

MSc Chemistry

Analytical Science

Literature Thesis

New mass spectrometry-based methods for organic dyes

analysis for forensic trace evidence and conservation of cultural

heritage objects

by

Roselina Medico

11118539

July 2020

12 EC

Supervisor/Examiner:

2

nd

_Examiner:

dr. Alina Astefanei

dr. Maarten R. van Bommel

Van ‘t Hoff Institute for Molecular

Science (HIMS)/

UvA

(2)

Abstract

The analysis of trace level dyes (e.g., µg or ng levels) found in textiles is a crucial step for the proper identification and characterization of the samples. Nowhere is this more common than in the study of cultural heritage objects and forensic trace evidence. To obtain the accurate identification of dyes at trace levels requires new instruments and methods. Specifically, methods that can deal with the shortcomings that are associated with these samples, which include limited sample size, the variance between samples, and unknown state of degradation or highly degraded samples.

This review focuses on the use of mass spectrometry-based techniques such as liquid chromatography-mass spectrometry (LC-MS), liquid microjunction surface sampling probe (LMJ-SSP), atmospheric solids analysis probe (ASAP), direct analysis in real-time (DART), laser desorption ionization (LDI), and surface acoustic wave nebulization-MS (SAWN-MS). These are considered minimally invasive techniques that can analyze fiber samples between 1 to 5 mm in length or 0.1 to 2 mg in weight. They have shown to be suitable for the analysis of trace-level dyes when previously extracted (e.g., LC and SAWN), as well as when the extraction of the dye occurs a couple of seconds before the MS analysis (e.g., LMJ-SSP, ASAP, DART, and LDI). LC-MS LOD ranging from 0.4 to 90 ng/ml have been obtained for several classes of dyes (e.g., anthraquinoids, flavonoids, tannins, and indigoids) and resolutions of 2-5 mDA for fiber on dyes for ambient MS techniques such as DART. Moreover, all the methods were deemed suitable for the identification of known and unknown components while providing information regarding degradation products when a MS/MS system was used. Hence, demonstrating their suitability for the analysis of complex samples such as those obtained from cultural heritage objects and forensic cases.

(3)

List of Abbreviations

2DLC Two-dimensional liquid chromatography

ACN Acetonitrile

APCI Atmospheric pressure chemical ionization

API Atmospheric pressure ionization

APPI Atmospheric pressure photoionization

ASAP Atmospheric solid analysis probe

ATR Attenuated total reflection

CID Collision induced dissociation

CZE Capillary zone electrophoresis

DAD Diode array detector

DART Direct analysis in real-time

DE Direct exposure

DMF Dimethylformamide

DMSO Dimethyl sulfoxide

DNA Deoxyribonucleic acid

EDTA Edetate disodium

ESI Electrospray ionization

FA Formic acid

FS Forensic science

FTIR Fourier transform infrared spectroscopy

HPLC High-pressure liquid chromatography

HRMS High-resolution mass spectrometer

IDT Interdigit transducer

IT-MS Ion trap mass spectrometer

LC Liquid chromatography

LDI Laser desorption ionization

LMJ-SSP Liquid microjunction surface sample probe

LOD Limit of detection

LOQ Limit of quantitation

MALDI Matrix-assisted laser desorption ionization

MeOH Methanol

MRM Multiple reaction monitoring

MS Mass spectrometer

MS/MS Tandem mass spectrometer

m/z Mass to charge ratio

NIRs Near-infrared spectroscopy

PDA Photodiode array

PGC Pyrolysis gas chromatography

PRM Product reaction monitoring

(5)

S/N Signal to noise ratio

SAWN Surface acoustic wave nebulization

SERS Surface enhance raman spectroscopy

SIM Selected ion monitoring

TLC Thin layer chromatography

TFA Trifluoroacetic acid

TOF Time of flight

UHPLC Ultra-high-pressure liquid chromatography

UNESCO United Nations Educational, Scientific and Cultural Organization

(6)

iii

List of Figures

Figure 1. Chemical structure of alizarin a compound commonly found anthraquinoid dyes. ... 4

Figure 2. Chemical structure of luteolin a compound commonly found in flavonoid dyes. ... 4

Figure 3. General chemical structure of an indigotin dye compound. ... 5

Figure 4. Chemical structure of a) ellagic acid and b) juglone. Two compounds commonly found in tannin dyes. ... 5

Figure 5. Classification of fibers. Reproduced from ref. [6]. ... 6

Figure 6. Textile fibers with their compatible dyes. Reproduced from ref. [3]. ... 7

Figure 7. Dye classes detected by different ionization techniques and ionization modes. Reproduced from ref. [8]. ... 14

Figure 8. Comparison of a) DAD chromatogram recorded at 275 nm and b-e) MS chromatograms using SIM mode. The co-eluting compounds purpurin and xanthopurpurin are not visible on the DAD chromatogram. It is only by the use of the MS in SIM mode that they can be separated and identified. Reproduced from ref.[14]. ... 15

Figure 9. Classification of ambient techniques based on their method of ionization. ... 17

Figure 10. Schematic of the original LMJ-SSP device. Reproduced from ref. [73]... 18

Figure 11. Schematic of a Flowprobe (LMJ-SSP) device. Reproduced from ref. [68] ... 19

Figure 12. The spectrum of a woolen reference fiber dyed with an extract of the insect Dactylopius coccus Costa (anthraquinone) by LMJ-SSP with carminic acid (C22H20O13) peak at m/z 491.0834. Reproduced from ref. [41] ... 20

Figure 13. Schematic of ASAP device. Reproduced from ref. [35] ... 21

Figure 14. Spectra of a) standard compound indigo (C16H10N2O2) and b) wool fibers dyed with synthetic indigo (C16H10N2O2) showing the presence of m/z 263.2. Reproduced from ref. [35].22 Figure 15. Spectra of a) undyed wool fibers and b) historic blue wool fiber showing the presence or absence of indigo (C16H10N2O2) at m/z 263.2, respectively. Reproduced from ref. [35]. ... 22

Figure 16. Schematic of a generic DART probe. Reproduced from ref. [82] ... 23

Figure 17. Schematic of a DART probe in front of MS inlet. Reproduced from ref. [83] ... 23

Figure 18. DART mass spectra of a 128-year-old nitroalizarin (C14H7NO6) orange-dyed cotton sample showing the nitroalizarin peak at m/z 286.0348 and an alizarin (C14H8O4) peak at m/z 241.0493. Reproduced from ref. [37]. ... 24

Figure 19. DART mass spectra of an onion-dyed cotton sample with quercetin (C15H10O7) at m/z 303.0543. Reproduced from ref. [37]. ... 24

Figure 20. DART mass spectra of a 128-year-old indigo-dyed (C16H10N2O2) cotton sample with an m/z at 263.0823. Reproduced from ref. [37]. ... 25

Figure 21. The spectrum of the offset image of the red colorant from the balloon-printed tole with an insert detail of the alizarin (C14H8O4) relate spectra at m/z 241.0495. Reproduced from ref. [36]. ... 25

Figure 22. Spectra of alizarin (C14H8O4 and m/z 241.0501), purpurin (C14H8O5 and m/z 257.0450) rubiadin (C15H10O4 and m/z 255.0657) and lucidin (C15H10O5 and m/z 271.0606) in a) the red colorant on the ballon-printed toile by transmission mode and b) offset image of the red colorant on the ballon-printed toile by reflective mode. Reproduced from ref. [36]. ... 26

(7)

Figure 23. Spectra of chemical standards a) alizarin (C14H8O4) and b) xanthopurpurin (C14H8O4) with and without the use of CID. Reproduced from ref. [38] ... 27 Figure 24. Spectra of Relbunium root samples a) Relbunium hypocarpium, b) Relbunium richardianum, and c) unknown species of Relbunium. Anthraquinone dyes: X = xanthopurpurin (C14H8O4), A= alizarin (C14H8O4), P = purpurin (C14H8O5), and R = rubiadin (C15H10O4). Reproduced from ref. [38] ... 28 Figure 25. a) Schematic of the commercially available SAWN chip (LiNbO3), b) the nebulization process, and c) set up the SAWN device in front of the MS inlet. Reproduced from ref. [97]. . 30 Figure 26. Other possible IDTs configurations of SAW chips. Reproduced from ref. [91]. ... 31 Figure 27. SAWN-MS spectrum of wool fiber dyed with Acid Violet 7 (C20H16N4Na2O9S2). Reproduced from ref. [42]. ... 31 Figure 28. The degradation products obtained from a wool fiber dyed with Basic Violet 3 (C25H30ClN3). a) SAWN-MS spectrum and b) SAWN-MS/MS spectrum. Reproduced from ref. [42]. ... 32 Figure 29. LDI schematic configured to a) IT-MS and b) TOF-MS. Reproduced from ref. [39].33 Figure 30. LDI-TOF-MS spectrum of a) an alum mordant quercetin (C15H10O7) dyed wool at m/z 303 and b) a red fiber from an ancient Peruvian civilization. Reproduced from ref. [39]. ... 34 Figure 31. LDI-TOF-MS spectrum of a) a wool fiber dyed with indigo from an unspecified origin (C16H10N2O2) and b) wool fiber dyed with natural indigo (uncertain biological origin) from ancient Peruvian civilization. Reproduced from ref. [39]. ... 34 Figure 32. Spectra of the standard 1,2-dihydroxyanthraquinone (alizarin, C14H8O4) for P (m/z 241) a) ESI-MS/MS and b) LDI-MS/MS. Reproduced from ref. [40]. ... 35 Figure 33. Daughter ion spectra obtained by LDI-MS/MS of 1,2-dihydroxyanthraquinone (alizarin, C14H8O4) for P+1 m/z 242 a) dye standard b) from a painting's cross-section and c) a madder dyed silk fiber. Reproduced from ref. [40]. ... 36 Figure 34. Indigo or Vat Blue I (MW: 262.27 g/mol, C.I. 7300, C16H10N2O2). ... 37 Figure 35. Absorption spectra of three different brands of denim jeans: a) Easy Care (USA), b) Males Only (Indonesia), and c) Gitano (USA). Reproduced from ref. [101]. ... 38 Figure 36. Spectral variation of three samples taken from the same cotton cloth dyed with Paramine Black GF. Reproduced from ref. [102]. ... 39 Figure 37. Denim fiber composition in the year 2015 obtained from the analysis of 108 denim jeans originating from different countries and from different brands available in the USA. Reproduced from ref. [104]. ... 40 Figure 38. The redox process for the reduction of indigo. Reproduced from ref. [107]. ... 41 Figure 39. The proposed mechanism for the production of Indigo from E. coli for the dyeing of cotton fibers. Reproduced from ref. [108]. ... 42 Figure 40. Comparison of cotton fibers dyed with a different type of Indigo dyes. Reproduced from ref. [108]. ... 43 Figure 41. The proposed mechanism for the reduction of indigo with reducing sugars. Reproduced from ref. [110]. ... 43 Figure 42. Analysis of synthetic indigo a-b) LDI-TOF-MS in the positive mode at low laser power c) LDI-TOF-MS in negative mode. Reproduced from ref. [39]. ... 46 Figure 43. The spectrum of natural indigo by LDI-TOF-MS positive mode. Reproduced from ref.

(8)

v [39]. ... 46 Figure 44. The spectra of synthetic indigo by LDI-ITMS a) MS2 of the protonated ions at m/z 263, b) MS3 of the fragment ions of m/z 263 at m/z 234 and c) MS4 of the fragment ions of m/z 234 at m/z 205. Reproduced from ref. [39]. ... 47 Figure 45. Comparison of S/N of a mixture of standards (e.g., anthraquinoid, flavanoids, and their derivatives) dissolved in an extract of the 4th-3rd B.C. woolen thread obtained by HPLC-MS in positive and negative ionization for scan, SIM, and MRM mode. Reproduced from ref. [61] .. 49 Figure 46. Flow diagram for fiber examination. Reproduced from ref. [117] ... 54

(9)

List of Tables

Table 1. Comparison of LOD’s of several dye classes in a fixed HPLC system for MS and DAD. ... 16 Table 2. Comparison of the different MS-based techniques found in the literature for the analysis of dyes in fibers. ... 53

(10)

1

1 Introduction

Identification of the dye components, in particular at trace levels (e.g., µg), can provide a plethora of information for the identification and characterization of textile fibers. This includes the source of the textile dyestuff (e.g., animal, plant, or mineral origins) and, to a deeper extent, even the geographical origin (e.g., country or region), leading to a deeper understanding of the object. These are all common attributes studied in both cultural heritage objects (e.g., textiles, paintings, and embroidery) and forensic trace evidence (e.g., crime scene). This information can, in turn, provide the knowledge needed for the conservation, restoration, and authentication of cultural heritage objects, as well as a better understanding of ancient cultures [1]. Likewise, the analysis of dyes can be of crucial importance for the evaluation of forensic evidence for the court in criminal prosecution or defense.

According to UNESCO, cultural heritage is defined as “the legacy of physical artifacts

and intangible attributes of a group or society that are inherited from past generations, maintained in the present and bestowed for the benefit of future generations” [2]. As we seek

to preserve and understand cultural heritage objects, an abundant amount of information can be gained by studying the dyestuff employed. For instance, the chemical composition of the dyeing material, the application of mordants, the dyeing process, and the rate of decay are valuable sources of information for the maintenance and restoration process of this prized object [3]. Additionally, the authentication of cultural heritage objects, in scenarios such as fraud cases, require information to prove the samples age and country of origin, which can be done through the analysis of the dye given that dyes of the same chemical class can be produced in different regions and from different dye sources (e.g., different plant species).

In forensic cases, the analysis of dyes can be exemplified with Locard’s exchange principle, which states that “every contact leaves a trace” [4], [5]. In legal terms, trace

evidence is defined as minute amounts of materials (e.g., fibers) that are inevitably

transferred through contact between individuals or between an individual and a physical location [6]. Common occurrences are cases involving violent contact (e.g., assault, rape, and murder), but it can also be as unassuming as a person sitting or resting on a couch. As long as some physical contact takes place between two objects, there will always be a chance of leaving a trace behind, making fibers the most common type of evidence recovered from the scene of a crime. While this type of evidence can be small and provides less evidential power than DNA evidence, proper analysis of any trace can make or break a forensic case. Primarily, trace evidence of this type could be used to match evidence with reference material, which can then potentially lead to the perpetrator or to even more evidence. This information could also be used for the reconstruction of a crime scene. To a more significant extent, fiber evidence could also be used to corroborate or refute a proposed link between an individual and a crime scene, especially when there is no DNA evidence [7].

However, to obtain accurate identification of dyes at trace levels, new instruments and methods are required. Specifically, methods that can deal with the shortcomings that are associated with these samples.

(11)

1.1 Dyes and Fibers

A dye is a soluble colored substance that can absorb and reflect specific visible wavelengths of light. It can fuse to fiber by forming physical and or chemical bonds without the use of other compounds. In contrast, pigments are insoluble colored compounds, mostly inorganic, that have no affinity to fibers. They are fused to the fibers by the usage of bonding agents or added in the melt [3], [8]–[10]. Throughout the literature, the classification of dyes varies upon the targeted audience. In general, dyes can be classified according to their color, source of origin, and chemical structure [8]. The simplest classification system is the assessment of dyes based on a visual perspective, or color theory, mainly used by artists. The principles color utilized for this classification is based on the primary colors (i.e., red, blue, and yellow) and the secondary colors obtained when they are mixed (i.e., green, orange, and violet). In this system, different shades or hues of a specific color are considered a tertiary group [11]. However, identifying a color based on appearance can be very user subjective. This is not only due to the differences between a person’s visual capacity but also relates to environmental factors. Elements such as lighting conditions, matrix/background, fading of dye, can also affect how color is visualized and hence labeled.

1.1.1 Classification by application

A more scientific approach involves the classification of dyes based on how their chemical structure affects their method of application [3], [9], [12]. This method is more commonly seen in the literature for forensic applications. This type of classification system provides further in-depth knowledge of the methods and procedures utilized in the dyeing process:

1) Acid dyes

Acid dyes occur under acidic conditions resulting mainly in ionic bonds between the dye molecule and the fiber. In addition to ionic bonds, van der Waals’ force and hydrogen bonds are also formed. Acid dyes can be applied to fibers such as polyamide, wool, and silk [3], [9], [12].

2) Azoic dyes

Azoic dyes are applied using a coupling component. The dyeing process starts with the fiber impregnation with a naphthol solution followed by impregnation of a ‘fast salt’ solution (e.g., base or diazo component) to form an insoluble molecule. This dye is commonly applied to cotton and man-made polymers such as ester and rayon [3], [9], [12].

3) Basic dyes

Basic or cationic dyes are applied under acidic conditions. The dyes are usually ammonium, sulphonium or oxonium salts. They ionize in solution with the colored component of the dye being a cation. The negatively charged fiber surface attracts the dye cation, which results in the fiber being neutralized. Raising the temperature permits the dye to penetrate the fiber.

(12)

3 Basic dyes can be applied to synthetic polymers such as polyamide, polyester, and polyacrylic fibers [3], [9], [12].

4) Direct dyes

Direct dyes are applied directly from an aqueous medium containing an electrolyte (e.g., NaCl). The positively charged ion is attracted to the negatively charged surface of the fiber, which neutralized the surface, enabling the dye anion to enter the fiber. This dye can be used with both natural (e.g., cotton, wool, and silk) and some man-made fibers (e.g., rayon and polyamide fibers) [3], [9], [12].

5) Disperse dyes

Disperse dyes are applied with an aqueous dispersion consisting of the dye, water, and a surface-active agent. The dye molecule is held in the fiber through hydrogen and the van der Waals’s bonds. The dye in this category can be applied to most of the man-made fibers (e.g., polyamide, polyester, and polyacrylic) [3], [9], [12].

6) Metalized (mordant) dyes

Metallized or mordant dyes are applied through the formation of metal complexes with fibers. Three types of processes are used: chrome mordant method, metachrome method, and after chrome method. In chrome mordant, the fiber is submerged in a chrome dye in an acid medium which is heated to fix the dye molecule. In the metachrome method, dye and mordant are added at the same time to the fiber. After the chrome method, acid dyes are applied to the fiber, and the temperature is slowly raised to reach the boiling temperature. After one hour, the mordant is added, and after another hour, the complex is formed. Metalized dyes are mainly applied to wool and sometimes to polypropylene fibers [3], [9], [12].

7) Reactive dyes

Reactive dyes are applied in an alkaline medium or acidic medium, depending on the fiber. The dyeing process takes place through covalent bonds with the functional groups. It can be applied to natural fibers and polyamide fibers [3], [9], [12].

8) Sulphur dyes

Sulphur dyes are initially reduced in an alkaline medium to produce the leuco form of the dye, which will then penetrate the fiber. Afterward, the leuco form has to be oxidized to its original insoluble form. Sulphur dyes are commonly used for cotton and rayon fibers [3], [9], [12].

9) Vat dyes

Vat dyes are applied through an extensive process that starts with dispersion, production of leuco form, application to the fiber, oxidation of the leuco form, and removal of the insoluble dye. It is most commonly applied to cotton, but it can also be used for wool and rayon [3], [9], [12].

(13)

1.1.2 Classification by chromophores

Lastly, and from another perspective, dyes can also be classified based on their chemical class to which their chromophore belongs. Chromophores are the part of the molecule that absorbs light [9]. This classification system is widely used in the literature of cultural heritage for the classification of natural organic dyes. Natural dyes are composed of mordant, vat and direct dyes and can be further subdivided between animal and plant origin [3], [8], [10], [13], [14]. The chemical class can be divided into four main groups:

1) Anthraquinoid dyes

They are naturally abundant in nature in both plants and animals. They have in general red color but can range from orange to pink to crimson. They are mostly found as mordant dyes or as red lakes, a pigment created by absorption or complexation of an organic dyestuff on an insoluble inorganic substance [14]. Anthraquinones derived from plants have one substituted aromatic ring, while those derived from animals (e.g., insects) contain substituted aromatic rings. Among other things, these types of dyes are found to be stable in the photooxidation process[8], [13], [14].

Figure 1. Chemical structure of alizarin a compound commonly found anthraquinoid dyes. 2) Flavonoid dyes

This class of dye belongs to the plant kingdom and mainly produces yellow hues but can also be found as red and dark dyes. This group is mainly used as a mordant dye but can also be found as yellow lakes. The flavonoid group of dyes is more prone to fading due to the photooxidative reactions. As a result, the identification of these dyes is made through the analysis of their glycosides or secondary components. [8], [13], [14].

Figure 2. Chemical structure of luteolin a compound commonly found in flavonoid dyes. O O OH OH HO OH O O OH OH

(14)

5 3) Indigoid dyes

This group of dyes can be found in plants and animals and produce blue and purple hues. Indigoid dyes have can also be found mixed with yellow dyes to yield green hues. As opposed to the previous two categories, this dye can be found mainly as vat dyes with some uses as a lake. The blue indigoid dyes are also known for their high resistance to photooxidative reactions [8], [13], [14].

Figure 3. General chemical structure of an indigotin dye compound. 4) Tannins

This class of dyes are derived from plants and include hydrolyzable tannins and condensed tannins. They produce brown and black hues. Tannins have a wider variety of chemical structures than the previous three groups. They are quite stable to photooxidative reactions [8], [13], [14].

Figure 4. Chemical structure of a) ellagic acid and b) juglone. Two compounds commonly found in tannin dyes.

1.1.3 Types of fibers

To obtain a specific color in the dyeing process, one must first consider the type of fiber that will be utilized. Fibers are defined as “any solid object that is thin, flexible, and

elongated, with a high length to transverse cross-sectional area ratio” [6]. Currently, fibers are classified as either being produced from natural resources or man-made products. Natural fibers can be further subdivided into three categories: animal, vegetable, and mineral origin, while man-made products can be subdivided into two categories: organic and inorganic. Both the natural and man-made fibers can be further subdivided, as presented in Figure 5 [6].

Once the fiber has been chosen, the next step is to select a suitable dye that will effectively chemically bond with it. Figure 6 displays the possible dyes that can be utilized on different fiber types. Some fibers, such as cotton, are capable of absorbing dyes from various categories (e.g., azoic, direct, oxidative, and reactive), yet other fibers have a more limited

H N N H O O HO O HO O O O OH OH O O OH (a) (b)

(15)

range of dyes. Therefore, in the analysis of unknown dyes, knowing the type of fiber narrows downs and simplifies the extraction procedure.

Figure 5. Classification of fibers. Reproduced from ref. [6]. Fibers Man-made Organic Natural polymers Cellulose ester Rayon Regenerated protein Rubber Synthethic polymers Polyolefin Polyvinyl derivatives Polyure-thane Polyamide Polyester Inorganic Natural Animal Hair Silk Mineral Vegetable Bast Leaf Seed

(16)

7 Figure 6. Textile fibers with their compatible dyes. Reproduced from ref. [3].

1.2 Analytical challenges

Unsurprisingly, the analysis of dyed textiles in both cultural heritage objects and for forensic trace evidence is not a simple straight forward path. Special considerations need to be taken into account in the evaluation and characterization of these types of samples. Even though each category contains their own unique set of challenges, both share the following characteristics: 1) limited sample size (e.g., bunches of fibers or single fibers), 2) variance between samples and 3) unknown state of degradation or highly degraded samples (over-all low dye concentrations). The first issue for both cases is related to the lack of samples and hence the general apprehension for destructive analytical techniques. In forensic cases, this is highly dependent on the level of contact between the objects, the fibers type, and other environmental factors. The number of samples found may not be enough for triplicate analysis. In the context of cultural heritage objects, samples taken for analysis should be small enough to maintain the object as pristine as possible. Already limited by the lack of readily available samples, this is further perpetuated by limiting the dimension, which usually ranges in the millimeter to micrometer category. Additionally, color variation in between samples due to the dyeing process and irregular shapes of fibers can contribute to lower dye concentration and poor analytical results [15]. Lastly, given that the goal is to learn more about the dye itself, this poses the most significant challenge to both cases. Exposure of the samples to environmental factors such as a variance of weather conditions, light exposure (e.g., photooxidation), or in general improper storage of the sample can lead to severe degradation of the dye [1].

(17)

Among other limitations, one drawback to the analysis of dyes, in some cases, is related to the lack of readily available dye standards. In general, reference materials and recipes for dyes are available through the use of the Color Index (CI), a reference work that catalogs the chemical class of all dyes [9]. In other cases, standards are prepared, when needed, by using the recipes or dyeing information obtained from historical knowledge and recipes books. However, the procurement of raw dyestuff reference samples can sometimes be challenging [16]. Additionally, these types of samples also suffer from the interference of other materials such as mordents, fibers, binders, and other pigments, which makes the sample treatment and extraction of the dye it-self more arduous [17]. Moreover, there is always a risk of losses occurring during the dye extraction process, and when the sample size is an issue, this poses a more significant challenge.

Due to this nature, the overall analysis and comparison process for both the cultural heritage field and crime scene samples are usually laborious and time-consuming. All these limitations must be taken into account when selecting the best analytical techniques for the analysis of dyes. Keeping in mind that in addition to these limitations, there are also individual limitations attributed to each technique.

1.3 Analytical techniques: the golden standards

For an analytical technique to be suitablefor the study of dyes in fibers for cultural heritage or forensic research, it must adhere to the following parameters: minimally invasive or non-destructive, fast, multi-elemental, sensitive, versatile, and universal. Primarily, the proposed analysis technique must be non-invasive, or at least micro-destructive (e.g., mm of fiber or below mg range), which would ideally entail leaving the material as is or with no visible changes that can be observed by the naked eye (especially crucial for art objects). Secondly, the sample preparation must be efficient but kept to a minimum and with an over-all short analysis time. Additionover-ally, it must be a multi-elemental technique that provides information not only about the known but also any unknown elements that might bring new information or evidence. Moreover, the techniques must have the adequate sensitivity required to analyze and reliably identify different classes of dyes in different types of substrates. This should include the ability to deal with micro samples and degraded samples. Furthermore, it needs to be versatile, allowing within a single run a wide range of information (e.g., m/z composition, retention time, and UV spectra) to be gathered to increase the level of confidence for the identified species and method validation. Finally, the technique must be universal to allow the study of different materials (e.g., dye classes) employing a single instrument [1], [18]. There is currently no specific technique that can adequately apply all the different parameters set above. Therefore, it is sometimes necessary to combine several analytical techniques to obtain a robust set of results.

(18)

9

1.4 Aim

Therefore, the aim of this literature thesis to give an overview of the current methods employed for the analysis of organic dyes in (single) fibers of relevance for both cultural heritage objects and in forensic cases. The application of liquid chromatography with mass spectrometry (LC-MS) and ambient mass spectrometry (e.g., DART-MS, SAWN-MS) techniques will be discussed in greater detail. Additionally, these techniques will be compared to other routine techniques used for the characterization of organic dyes in fibers. Furthermore, a critical discussion will be made in terms of the benefits and drawbacks of the reported techniques. Finally, a brief overview of the future perspectives for organic dye analysis in single fibers will be included.

(19)

2 Analytical workflow

The world of cultural heritage, as compared to the forensic world, is more open to utilizing new techniques for the analysis of dyes. Hence, most of the articles found currently on the subject of dyes and fibers center around cultural heritage objects. Within the reviewed literature, the majority revolved around the use of spectroscopy as the primary technique for the screening of dyes, followed by liquid chromatography (LC) and, lastly, ambient mass spectrometry techniques. Each technique has its own advantages and disadvantages for which they may be better suited for different scenarios.

2.1 Analytical techniques

As mentioned previously, the primary method for the analysis of dyes on fiber was found to be spectroscopic methods. Articles found in this area included the use of Raman spectroscopy [14], [19], surface-enhanced Raman spectroscopy (SERS) [20]–[23], near-infrared spectroscopy (NIRs) [14], [19], attenuated total reflection Fourier transform infrared spectroscopy (ATF-FTIR) [24]–[26], fluorescence spectroscopy [14], [19] and UV/vis spectroscopy [27]. Spectroscopy methods are regularly utilized since they are fast, non-destructive, and require little sample preparation. Notably, it is the ability to perform in-situ analysis of the objects or samples that make these methods highly in demand. These methods allow us to obtain a spectral fingerprint of the dye, which can later be confirmed or identified through the use of a spectral database. That being said, these techniques cannot identify samples unknown to the database. Furthermore, absorption and fluorescence bands are broad, which causes emissions of chromophores to overlap [14]. Moreover, the quality of the sample (e.g., degraded samples) and the type of fiber (e.g., matrix effect) can interfere with the results obtained [14]. Hence, to obtain further information about the dye components, other analytical techniques are required.

The second most popular technique for the analysis of organic dyes is liquid chromatography (LC). This category included high pressure liquid chromatography (HPLC), ultra-high-pressure liquid chromatography (UHPLC) [28] and two-dimension liquid chromatography (2DLC) [29]. In addition to the variety of the LC classes different detectors were also reviewed, including diode array detector (DAD) [21], [26], [30], [31] mass spectrometers (MS) [15], [22], [26], [32], high resolution mass spectrometers (HRMS) [33], and tandem mass spectrometers (MS/MS) [34]. The advantage of LC over spectroscopic methods is the ability to separate dye mixtures and hence provide more detailed information of the individual dye components, degradation products, side products, and the dyeing recipe. On the other hand, LC methods have, by far, the longest analysis time. This includes the time need to extract the dye from the fiber (e.g., 10-120 minutes), as well as the time needed to separate the dye components (e.g., 10-90 minutes). In addition, this method is considered a micro-destructive technique. Meaning that small samples (e.g., 1-5 mm of yarn or fiber, depending on the detection used) must be acquired from the original object so they can be analyzed. Nevertheless, LC techniques are more sensitive and more versatile than the spectroscopic method. This is particularly true in the cases where the LC is coupled with more than one detector which in addition to information regarding the dye’s components retention time, a full UV/Vis spectrum (i.e., DAD) and chemical fingerprint (i.e., MS and MS/MS) of its major components and degradation products are obtained.

(20)

11 Lastly, the third group of techniques referenced was the direct MS category.This consists of atmospheric solids analysis probe (ASAP) [35], direct analysis in real-time (DART) [36]–[38], laser desorption ionization (LDI) [39], [40], liquid microjunction surface sampling probe (LMJ-SSP) [41] and surface acoustic wave nebulization (SAWN) [42]. The quality of the obtained data is highly dependent on the chosen technique. Nonetheless, they are highlighted for their capabilities of simplifying the overall analysis process (e.g., analysis time and sample preparation) while obtaining similar structural information to those of LC-MS methods. Even though there is no initial separation of compounds or a UV spectrum, these methods have been capable of detecting compounds in trace samples (e.g., 100-700 µg or 1-5 mm) in as little as 1 minute. Moreover, the use of HRMS and MS/MS detection has allowed these methods to identify a dye’s molecular ion accurately. Additionally, the use of collision-induced dissociation (CID) fragmentation patterns has allowed for degradation studies of the samples and even isomer identification (e.g., DART and LDI). These techniques require little (e.g., SAWN) to no sample preparation (e.g., ASAP, DART, LMJ-SPP, and LDI) as the ionization process occurs directly on the sample’s surface. Consequently, these techniques are considered micro-destructive techniques with sample sizes in the same order of magnitude as those used for LC-MS.

(21)

2.2 Liquid chromatography-mass spectrometry, LC-MS

Currently, the most widely used technique for the analysis of dyes on fibers is liquid chromatography (LC). LC can separate complex mixtures with the added benefit of being compatible with different detectors (e.g., MS and DAD), which in turn provides more valuable and detailed information.

2.2.1 Extraction procedures

The most crucial part of the analysis of dyes by LC is the dye extraction procedure. The current literature on this matter is mainly focused on improving the quality of the dye extraction technique before the LC analysis. A suitable extraction method is chosen given the known composition of the fiber and the compatibility of the solvents with the HPLC column. In addition, the chosen solvent should not interfere with MS detection [34]. Care must be taken when selecting the type of extraction method as the quality and quantity of the isolated colorant is highly dependent upon it as unwanted results such as degradation of the dye sample can occur.

The main methods found in the literature involve the use of acid hydrolysis, extraction with an organic solvent, and extraction by applying complexation agents [8], [14]. The general set-up for all techniques begins with the submersion of the fiber in the chosen solvent, which is subsequently heated within the temperature range of 40 °C to 135 °C and a time range of 10 to 120 minutes. After heating the sample for the specified amount of time, a cool-down period (e.g., room temperature) follows, after which the solvent is being evaporated, and finally, the re-solubilization of the sample. The choice of solvent is based on both the type of fiber and the dye or suspected dye.

In acid hydrolysis, the extraction of the dyes from the fiber occurs through the use of a strong acid (e.g., HCl) heated to 100 °C for 10 minutes followed by cooling and re-solubilization [20], [26], [31], [32], [43], [44]. Hence, acid extraction is typically considered a ‘harsh’ extraction technique as it breaks the bonds of compounds containing glycosides and turns them into their aglycones [32]. This is the case of many yellow dyes from the flavonoid family. This technique is better suited for mordant dyes such as anthraquinones and tannins [14]. Therefore, to prevent the formation of aglycones, the use of ‘soft’ or ‘mild’ techniques are recommended [14].

The extraction with organic solvent includes the use of dimethyl sulfoxide (DMSO) [15], [26], [44], pyridine, [22], [45], dimethylformamide (DMF) [46], [47], dimethyl sulfoxide [48] and acetonitrile [15], [34]. In addition, it is also possible to combines several organic solvents with complexation agents. Common complexation agents utilized were EDTA [32], [47], FA [15], [20], [32], [49], sulphuric acid (H2SO4), sodium hydroxide (NaOH) [29], [49], oxalic acid [21],and acetic acid [15], [34]. Contrary to acid hydrolysis, for the extraction of dyes from the fiber, these techniques usually require a lower temperature (e.g., 40-135 °C) for a longer period (e.g., 10-120 min) depending on the type of solvent used. For instance, when DMSO is applied for the extraction of direct, disperse, and acidic dyes found in different fibers (e.g., cotton, regenerated cellulose, polyester, polyamide, and wool), the sample is heated in the solvent at 100 °C for a maximum of 2 hours [15]. In a separate case, DMSO was utilized for the extraction of indigo (e.g., vat dye) by first sonicating the sample at 65 °C for 20 minutes, followed by heating the sample for 10 minutes at 135 °C [52]. In specific cases, such as for flavonoid dyes, more than one technique can be successfully applied. Flavonoids found in fibers (e.g., silk and wool) can be either extracted by EDTA

(22)

13 (at 60°C for 30 minutes) or with the use of FA (at 40 °C for 30 minutes) [32]. Both solvents were capable of extracting the dye with little to no decomposition, yet EDTA proved to be more efficient for silk and FA for wool. Although it is important to avoid using harsh extraction procedures, it is good to realize that that softer extraction techniques may be less effective in the dye extraction process.

More recently, in a paper by Carey et al. [15], an improved technique for the ‘mild’ extraction of reactive dye was developed. This technique called ‘cellulose digestion procedure’ combines the original idea of Home et al. [51] with the proposed improvements suggested by Wiggins [9] but utilizing an alternate cellulase due to a lack of commercial availability. Cellulase digestion works by dissolving the covalent bond between the dye and the fiber. After altering the method to meet their needs, the author came up with the following protocol; Initially, the fiber is submerged in NaOH and cooled down to 4 °C for 4 hours. After which the fiber is rinsed in acetic acid solution and twice in the cellulase solution. The cellulase solution contained 10 mL of acetic acid solution in water (pH5) with 0.01 g of cellulase. The fiber is then submerged in the cellulase solution, which is mixed for 20 hours, followed by centrifugation for 5 minutes. Lastly, methanol is added to the solution. This method is utilized in the extraction of reactive dyes, which can be found on fibers such as cotton, wool, silk, and polyamides [3], [15].

2.2.2 Chromatographic separation

The separation in LC is based on the component’s difference in their affinity or retention strength for the stationary phase and the mobile phase. More specifically, the affinity of the components to the liquid mobile phase will dictate which components will elute first and which will elute last. Molecules of the same compound will migrate together through the column to produce distinct bands. In the analysis of dyes, the separation of components is based on the principle of reverse-phase (RP) chromatography by employing a C18 column [8], [14], [19]. In this method, compounds are separated based on their hydrophobic interaction between the solute molecules in the mobile phase (i.e., polar) and the ligands attached to the stationary phase (i.e., non-polar) [53].

To this date, there are no set guidelines for the experimental set-up when it comes to the composition of the mobile phase and the analysis time, nonetheless gradient elution is the most common method found throughout the literature. Typical mobile phases found in the literature include a combination of solvents (e.g., acetonitrile (ACN), methanol (MeOH), and water) and complexing agents (e.g., FA, acetic acid and ammonium acetate) [8], [14], [19]. The procurement of a suitable elution method is crucial for a satisfactory elution and resolution of structurally related dyes [19].

2.2.3 Detection of the compounds

After separation of the dye mixture, the obtained fractions are then sent to the detector of choice (e.g., UV-Vis, DAD, MS). The most common coupling of LC found in literature is DAD, LC-MS, or LC-DAD-MS. Each type of detector has its unique advantage and disadvantages, but the choice of which to use depends immensely on what is the aim of the research.

The primary advantage of DAD is its ability to detect multiple wavelengths simultaneously. This allows for the detection of dye components and degradation products that would otherwise not be detected due to their absorbance at different wavelengths [19]. DAD can also be used for

(23)

quantitation and confirmation of a dye identity by the use of spectral matching with the reference database [12]. This makes this type of detector better suited for the analysis of known dyes. In the case of unknown dyes, this becomes problematic if no such standard exists or they are not commercially available [54]. Even more problematic is the analysis of degraded samples as the presence of secondary degradation products will produce compounds unknown to the spectral database [19]. Moreover, components with similar UV-vis absorption are not easily differentiated by this method [28]. In addition to sample by-products, DAD can also be influenced by the mobile phase as well as the sample treatment used to extract the dye mixture [46].

Advancements in MS technology resolves many of the problems faced in the analysis of dyes. MS allows the identification of dyes through the mass-to-charge (m/z) ratio. After the extraction and separation of the dye by LC, further, optimization can be obtained through the use of different ionization methods and MS acquisition modes. Samples must be introduced into a mass spectrometer as ions. Processes that convert neutral molecules to charged ions may be classified as ‘soft’ (primarily produce molecular ions with little fragmentation) or ‘hard’ (the signal for the molecular ion is weak or maybe even absent, while there are many peaks due to ions produced by fragmentation). Depending on the analysis needed and sample type, both methods of ion production have advantages and disadvantages. Electron spray ionization (ESI), atmospheric pressure chemical ionization (APCI), and atmospheric pressure photoionization (APPI) are three ‘soft’ ionization techniques. ESI is commonly utilized in the study of dyes due to its ability to ionize polar compounds (i.e., chromophores) efficiently. In contrast, APCI is less commonly applied, yet it is mainly used for less polar and more volatile compounds [55]–[58]. APPI, even less common than the previous two ionization technique, can be applied to very apolar compounds and has been applied successfully to the analysis of indigoids [8], [59]. Furthermore, the addition of complexing or ion-pairing components (e.g., FA, acetic acid, ammonium acetate) can enhance ionization efficiency [34]. Ionization can be performed in both positive and negative modes. In general, both ionization modes should be used to identify dyes, given that some compounds will ionize more efficiently in either one or the other [60]. This is attributed to the quantity of acidic groups found in the dye and their degree of dissociation [34]. Figure 7 gives an overview of the different dyes with the preferred ionization techniques. Dyes in parentheses produced less intense signals. For instance, in the case of anthraquinoids and flavonoids, negative ionization gave better selectivity and sensitivity expect in the case of alizarin, an anthraquinoid [61].

Figure 7. Dye classes detected by different ionization techniques and ionization modes. Reproduced from ref. [8].

(24)

15 The selectivity of the MS can be further improved by selecting the appropriate MS data acquisition mode: 1) full-scan mode, 2) selected ion monitoring, SIM, and 3) multiple reaction monitoring, MRM, or product reaction monitoring (PRM) [56]. In the full-scan mode, the entire mass range from the lowest m/z value to the highest m/z value is analyzed and provides a full view or a fingerprint of the sample. Further selectivity and sensitivity can be obtained in the SIM acquisition mode in which a specific m/z value is monitored. In MRM, particular ions are selected from the full-scan, which are then subsequently further fragmented through collision-induced dissociation (CID) to obtain and detect the fragments [34]. Through the use of the CID fragmentation patterns derived from the use of MS/MS, it becomes possible to elucidate the structure of the dye as well as to resolve its isomeric compounds that are separated by chromatography prior to MS detection. MRM has been used in the analysis of degradation products to identify natural dyes [62]. At the same time, co-eluting compounds not detected by LC-DAD can be further elucidated by the use of MS in selected ion monitoring (SIM) mode (Fig. 8) [14]. Moreover, the analysis of the fragmentation patterns obtained has led to the detection and a better comprehension of the study of dyes degradation mechanisms [62]. It is through this process that art conservators can select an adequate restoration procedure, even when most of the color has faded. Overall, the results form MS detection has shown to produce better selectivity and lower limit of detection (LOD) compared to DAD (Table 1). Hence, MS detectors are better suited than DAD for the analysis of unknown samples.

Figure 8. Comparison of a) DAD chromatogram recorded at 275 nm and b-e) MS chromatograms using SIM mode. The co-eluting compounds purpurin and xanthopurpurin are not visible on the DAD chromatogram. It is only by the use of the MS in SIM mode that they can be separated and identified. Reproduced from ref.[14].

(25)

The identification of dyes and their source of origin can be a very complicated process. This is not only related to the size and the quality of the sample but also the relatively similar chemical structures of dyes originating from the same family. By coupling LC to both DAD and MS, a plethora of information is gained for the confirmation and identification of known and unknown dyes. Table 1. Comparison of LOD’s of several dye classes in a fixed HPLC system for MS and DAD.

Dyes MS DAD Ref.

Anthraquinoids 30—90 ng/mL 0.18—1.71 µg/mL [63]

Anthraquinoids,

flavonoids, benzoic acids, and quinizarin*

0.5—12 ng/mL;

32 ng/mL* 2—31 ng/mL [61]

Anthraquinoids,

flavonoids, benzoic acids, tannins, and indigotin

0.4—20 ng/mL 10—70 ng/mL [64]

(26)

17

2.3 Ambient MS-based techniques

In the past decade, there has been a growing trend in the use of direct MS techniques. These techniques were designed to circumvent steps in the procedures involving sample treatment and chromatographic separation that are a necessity for MS analysis. The benefits of these methods include minimizing sample consumption, sample contamination, and reducing chemical waste [14], [65]. In the case of dyes in fibers, direct ionization techniques can provide information about samples as small as 1 mm [42], [66] with an analysis time of as short as 1 minute [37]. In addition, the drawbacks of sample preparation by extraction, hydrolysis, or derivatization can be avoided [19].

Mogen et al. define ambient ionization techniques by the following parameters: 1) ionization must occur in the open air; 2) sample analysis should be possible via direct surface analysis; 3) ionization of the samples must produce ions with little fragmentation; and 4) the ambient ion source should be interchangeable in mass spectrometers with an atmospheric pressure interface [67]. Ambient ionization can be attained via a one-step process “desorption ionization” or via a two-step process of “desorption and followed by ionization” [67]. These techniques can be further sub-classified according to Figure 9 based on their primary desorption/ ionization mechanism.

Figure 9. Classification of ambient techniques based on their method of ionization. Ambient Techniques One-step mechanism Solid-liquid extraction Plasma Two-step mechanism Thermal/ mechanical desportion/ ablation

(non-laser)

Laser

Acoustic

(27)

The techniques utilized in the analysis of dye on fibers are further described below. In the category of ‘one-step mechanism,’ the first subclass entails the desorption of a sample by solid-liquid extraction and then subsequently ionized. Liquid micro junction surface sampling probe (LMJ-SSP) is an ionization method based on this technique. Also, under the ‘one-step mechanism’ category is the sub-class of the plasma-based methods, which utilizes thermal or chemical sputtering neutral desorption followed by gas-phase chemical ionization. Examples of plasma-based techniques are direct analysis in real-time (DART) and atmospheric solids analysis probe (ASAP). In the category of ‘two-step mechanism,’ the most prevalent techniques for the analysis of dyes are surface acoustic wave nebulization (SAWN), an acoustic desorption technique, and the

laser desorption ionization (LDI), a multimode technique. A summary of the different ambient

ionization techniques reported for the analysis of dyes in fiber is given in Appendix I. The table describes the ionization source, the types of dyes analyzed, the types of fibers used, the sample size (e.g., mm or µg), the sample preparation technique, the overall analysis time (i.e., minutes), and the MS mode utilized.

2.3.1 Liquid microjunction surface sampling probe

The liquid microjunction surface sampling probe (LMJ-SSP) ionization technique uses a semi-static liquid junction to perform surface sampling (Fig. 10-11). The current application of LMJ-SSP includes analysis of dried blood spots [68], analysis of body tissues (e.g., brain, liver, and kidney) [69], animal tissue [70], and plant tissue [71]. This is done through the use of coaxial capillaries, which are situated close to the surface of the sample. The outer capillary, with the largest internal diameter, supplies a continuous solvent flow that produces the liquid microjunction between the probe and sample surface. At the same time, nebulizing nitrogen gas is applied to the tip of the ESI nozzle. The vacuum created by the pneumatic nebulization extracts the liquid material from the sample surface via the inner capillary. At the ESI nozzle, a high voltage is applied that converts the liquid into electrospray droplets and are subsequently introduced to the MS [72], [73].

(28)

19 Figure 11. Schematic of a Flowprobe (LMJ-SSP) device. Reproduced from ref. [68].

In the article by Kramell et al. [41], LMJ-SSP was utilized for the characterization of indigo and anthraquinone dyes as liquid standards, dyed fibers (i.e., cotton, linen, silk, and wool) and historical samples (i.e., cotton, silk, and wool). The aim of this work was to present for the first time the use of LMJ-SSP as a rapid and minimally invasive technique for the analysis of dyestuff in objects of archeological interest. This method required no previous sample preparation for the fibers. Dyed fibers, 100-700 µg, were affixed on to a glass slide by the addition of a solvent (e.g., MeOH/H2O with FA or ACN with FA). Standards were suspended or dissolved in methanol (MeOH) and spotted on a glass slide where they were left to dry. For efficiency, the extraction monitoring period was increased for those samples with poor extractability, but in general, the total analysis time was less than 5 minutes. Confirmation of the dyes was made possible by the use of an HRMS in both positive and negative modes. The results demonstrate a clear base peak corresponding to the anthraquinone dyes at m/z 491.0834 (Fig. 12) and indigo dyes at m/z 263.0815, with mass errors less than or equal to 1.5 ppm. Most notably was the ability to differentiate between plant and animal origin irrespective of the fiber type, age of the sample, or the sample’s workmanship.

(29)

Figure 12. The spectrum of a woolen reference fiber dyed with an extract of the insect Dactylopius coccus Costa (anthraquinone) by LMJ-SSP with carminic acid (C22H20O13) peak at m/z 491.0834. Reproduced from ref. [41].

(30)

21

2.3.2 Atmospheric solids analysis probe

Atmospheric solids analysis probe (ASAP) is a minimally destructive technique without the need for sample preparation. It can be used for the rapid analysis of liquids (i.e., volatile and semi-volatile) and solid samples. Furthermore, this method can ionize samples with low polarities that are not easily ionized by other methods (e.g., ESI, APCI, and APPI). It is considered a soft ionization technique that produces radicals [M] •+_{and protonated molecules (e.g., [M+H]}+_{or [M-H]}-_{) with low} fragmentation patterns [74]. ASAP has been applied for the study of drugs [75], pesticides in apples [76], analysis of fatty acids on seed surfaces [77], among others. In ASAP, ionization occurs in a standard API housing fitted with a window for the removable ASAP probe. Samples are placed at the tip of the ASAP probe, which consists of a melting point capillary, positioned between the MS inlet and a standard APCI or ESI probe (Fig. 13). The APCI or ESI probe then provides the hot gas (i.e., nitrogen) required to nebulize the samples. Immediately after nebulization, the sample comes into contact with the Corona discharge produced at the Corona discharge pin, where it becomes ionized and is subsequently introduced into the MS inlet [58].

Figure 13. Schematic of ASAP device. Reproduced from ref. [35].

The article by Kramell et al. [35] describes the advantages of using ASAP for the analysis of indigo dyes on different types of fiber. The study involved the measurements of dye standards (e.g., indigo and indirubin), undyed reference fibers (e.g., cotton, linen, silk, and wool), and dyed historical fibers (i.e., cotton, silk, and wool). No sample preparation was required, and typical results were obtained within one minute with fiber fragments between 200-500 µg in size. The results for all the samples containing indigo displayed a clear base peak at m/z 263 (Fig. 14-15). At the same time, with the use of a single quadrupole analysis in combination with ASAP, daughter ions were present at lower intensities. Although no further separation (e.g., isomeric compounds) could be obtained nor quantification was possible, this method could potentially be used as a screening tool.

(31)

Figure 14. Spectra of a) standard compound indigo (C16H10N2O2) and b) wool fibers dyed with synthetic indigo (C16H10N2O2) showing the presence of m/z 263.2. Reproduced from ref. [35].

Figure 15. Spectra of a) undyed wool fibers and b) historic blue wool fiber showing the presence or absence of indigo (C16H10N2O2) at m/z 263.2, respectively. Reproduced from ref. [35].

(a)

(b) (b)

(32)

23

2.3.3 Direct analysis in real-time

Direct analysis in real-time (DART) is a technique that utilizes a heated helium or nitrogen gas stream to desorb and ionize molecules from a sample surface (Fig. 16-17). It can be used to analyze liquid and solid samples, as well as vapor phase samples. It has been commonly applied for the analysis of drugs [78], explosives [79], and food safety [80]. A typical DART probe comprises a tube with several chambers through which the gas flows. As the gas flows into the discharge chamber, an electrical potential is added, which creates ions, electrons, and excited-state species. In the next chamber, a perforated electrode can be used to separate ions from the gas stream. Afterward, the gas flow passes through a heated chamber and exits through a grid electrode. The grid electrode works to repel ions and to discard ions of the opposite polarity. This technique has the capability of being used in transmission mode or reflectance mode. In transmission mode, the DART gas flow is aimed directly at the MS inlet. In contrast, in reflectance mode the gas flow first hits the sample surface and then it is reflected in the MS [81].

Figure 16. Schematic of a generic DART probe. Reproduced from ref. [82]

Figure 17. Schematic of a DART probe in front of MS inlet. Reproduced from ref. [83]

DeRoo et al. [37] demonstrated that through the use of DART in combination with TOF-MS for the identification of natural dyes in textiles and cultural heritage objects. DART allowed the identification of dyes from several classes (e.g., anthraquinone, flavonoid, and indigoid) as well as from different biological sources (e.g., madder root and onion skin). In addition, this technique proved to be useful for both aqueous solutions and dyes on fibers (e.g., cotton, linen, silk, and wool) with previous sample pre-treatment. Within a minute, they were able to identify alizarin at m/z 241.050 Da, quercetin at m/z 303.049 Da, nitroalizarin at m/z 286.035, and indigo at m/z 263.083 Da in pure dye standards with a mass accuracy of less than 1 mDa. The results obtained from single

(33)

fiber samples (5 mm) from both historical samples and dyes on cotton fibers had a mass accuracy of 3-5 mDa (Fig. 18-20). The lower accuracy for dyes on fibers was attributed to the textile matrix.

Figure 18. DART mass spectra of a 128-year-old nitroalizarin (C14H7NO6) orange-dyed cotton sample showing the nitroalizarin peak at m/z 286.0348 and an alizarin (C14H8O4) peak at m/z 241.0493. Reproduced from ref. [37].

Figure 19. DART mass spectra of an onion-dyed cotton sample with quercetin (C15H10O7) at m/z 303.0543. Reproduced from ref. [37].

(34)

25 Figure 20. DART mass spectra of a 128-year-old indigo-dyed (C16H10N2O2) cotton sample with an m/z at 263.0823. Reproduced from ref. [37].

Similarly, in the article by Baglia et al. [36], the capabilities of DART for the analysis for alizarin dye and its derivatives in textiles (e.g., cotton and wool) and paper were demonstrated. The authors were able to obtain the alizarin base peak at m/z of 241.0504 and mass accuracy of less than 1 mDa for powdered standards, as well as historical samples (m/z 241.0495) (Fig. 21). Furthermore, the use of high-temperature helium settings was highlighted for its ability to increase signal abundance in transmission mode, more specifically for alizarin and its analogs. In contrast, reflectance mode was better suited for the analysis of out-gassing or leaching at lower temperatures. Figure 22 shows the spectra of alizarin, purpurin, rubiadin, and lucidin when DART is configured in transmission mode (Fig. 22 a) and when it is configured in reflectance mode (Fig. 22 b). In addition, this work showed that by varying the temperature and configurations (e.g., transmission or reflectance mode), the sensitivity of the instrument could be significantly improved.

Figure 21. The spectrum of the offset image of the red colorant from the balloon-printed tole with an insert detail of the alizarin (C14H8O4) relate spectra at m/z 241.0495. Reproduced from ref. [36].

(35)

Figure 22. Spectra of alizarin (C14H8O4 and m/z 241.0501), purpurin (C14H8O5 and m/z 257.0450) rubiadin (C15H10O4 and m/z 255.0657) and lucidin (C15H10O5 and m/z 271.0606) in a) the red colorant on the ballon-printed toile by transmission mode and b) offset image of the red colorant on the ballon-printed toile by reflective mode. Reproduced from ref. [36].

(a)

(36)

27 The work of Armitage et al. [38] showed that through the use of DART and tandem mass spectrometry, the separation of isomers becomes possible. Insect and plant-derived anthraquinone dyes (i.e., alizarin, xanthopurpurin, purpurin, and rubiadin) were analyzed as reference samples. In addition, loose wool fibers and yarns from historical samples were also investigated. The results obtained showed that by increasing the negative voltage (i.e., -30 V to -90 V), they were able to produce a collision-induced dissociation (CID), which lead to a higher quantity of fragmentations. The fragmentation patterns obtained showed a significant difference between alizarin and xanthopurpurin, its isomeric structural compound (Fig. 23). Similar results were also obtained for purpurin and rubidian. Additionally, three samples originating from the Relbunium root family were separated and characterized based on their fragmentation pattern (Fig. 24). The identity of the three samples was based on the ratio of the four main anthraquinone compounds previously mentioned. Even though this method did not work for aglycones of carminic acid, DART can still be used as a confirmation tool for red colorants for both standards and dyes on fibers.

Figure 23. Spectra of chemical standards a) alizarin (C14H8O4) and b) xanthopurpurin (C14H8O4) with and without the use of CID. Reproduced from ref. [38]

(b) (a)

(37)

Figure 24. Spectra of Relbunium root samples a) Relbunium hypocarpium, b) Relbunium richardianum, and c) unknown species of Relbunium. Anthraquinone dyes: X = xanthopurpurin (C14H8O4), A= alizarin (C14H8O4), P = purpurin (C14H8O5), and R = rubiadin (C15H10O4). Reproduced from ref. [38]

(38)

29

2.3.4 Surface acoustic wave nebulization

In Surface acoustic wave nebulization (SAWN), a sample droplet is placed in the delay region of a lithium niobate (LiNbO3) chip (Fig. 25a). A sinusoidal electrical signal is then applied to the chip through an RF cable. This signal is transformed through the piezoelectric effect produced by the interdigital transducer (IDT), made of a chrome adhesion layer and gold, into an acoustic wave that propagates through the chip. The IDTs are located on both extremities of the chip, which allows it to generate recirculation that induces the capillary waves on the surface of the droplet. The capillary waves subsequently destabilize, which produces a vertical nebulization (Fig. 25b). By this process, the sample droplet is introduced to the inlet of the MS (Fig. 25c), where it is further desolvated and analyzed [84]–[87].

Surface acoustic wave nebulization (SAWN) utilizes a nanometer-order amplitude acoustic wave that propagates along and near the surface of a single-crystal piezoelectric substrate actuation to produce multiply charged ions [84], [87]. The chips or the single-crystal piezoelectric substrate, where the liquid samples are nebulized, can be found as either quartz, lithium tantalite (LiTaO3), zinc oxide (ZnO) or lithium niobate (LiNbO3), the latter being the commercially available version, with each providing different qualities of nebulization power [88]. Along with the different types of substrates, the placement and shape of the IDT provide different ion intensities. The original SAWN device used a single IDT but had inferior capabilities when compared to ESI, with ESI producing 102_-103_{higher ion intensities. After this, dual IDT devices were created to overcomes} these issues. The new design (Fig. 26) provided the same level of ion intensity as ESI but with a lower internal ion energy of 2.25 eV when compared to ESI (i.e., 2.44 eV) [89]–[91]. It is considered a soft ionization technique, and therefore, there is less fragmentation, as was exemplified in the work of Yoon et al. [92] for the analysis of lipids (e.g., lipid A and phospholipids). SAWN has also been proven to work efficiently in the study of other complex samples such as phospholipids, blood [93], fermented food products [94], and explosives [95].

(39)

Figure 25. a) Schematic of the commercially available SAWN chip (LiNbO3), b) the nebulization process, and c) set up the SAWN device in front of the MS inlet. Reproduced from ref. [96].

Delay region IDT 1 IDT 2 Incoming electrical signal (a) (b) (c)

(40)

31 Figure 26. Other possible IDTs configurations of SAW chips. Reproduced from ref. [90].

Astefanei et al. [42] reported the use of SAWN for the analysis of synthetic dyes (e.g., basic violet 3, acid violet 7, and basic blue 26, acid orange 6, acid red 88, acid red 44, and acid blue 74) on wool fibers (e.g., 1-2 mm) and as standards. This technique, unlike some of the previously mentioned methods, does require sample preparation since samples must be in liquid form. In this case, the wool samples (few mm of fiber) were dissolved in 100 µl MeOH-H2O and vortexed for 2 minutes. A total of 10 µl aliquot was placed on the SAWN chip, where it was subsequently nebulized and introduced into the MS to produce a spectrum in up to 1 minute. This work demonstrated that with the use of this ionization method, complex mixtures, such as that of wool dyed in acid violet 7 (C20H16N4Na2O9S2) (Fig. 27), can be elucidated, and the dye mixtures identified. In addition, SAWN, in combination with HRMS, identified trace level peaks of the demethylated products with a 2-ppm accuracy demonstrating its capabilities of detecting dye degradation products (Fig. 28). Moreover, LOQ’s as low as 0.001 pg were obtained for basic dyes and between 0.0005 and 0.5 pg for acidic dyes.

Figure 27. SAWN-MS spectrum of wool fiber dyed with Acid Violet 7 (C20H16N4Na2O9S2). Reproduced from ref. [42].

(41)

Figure 28. The degradation products obtained from a wool fiber dyed with Basic Violet 3 (C25H30ClN3). a) SAWN-MS spectrum and b) SAWN-MS/MS spectrum. Reproduced from ref. [42].