ABSTRACT LI, MIN. Forensic Analysis of Organic Dyes on Trace Fibers

Home , Acid Blue 25, Anthraquinone, Disperse dye

ABSTRACT

LI, MIN. Forensic Analysis of Organic Dyes on Trace Fibers and Their Photodegradation. (Under the direction of David Hinks and Nelson. R. Vinueza.)

Trace evidence collected from difference sources (victims, suspects and the crime scene) can establish connections between the crime scene, suspects and victims if they are identical. In forensic analysis, textile fiber evidence is compared in terms of both physical properties and chemical composition, among which the organic dyes are important components. Therefore unambiguous identification of organic dyes is of significance in forensic fiber analysis. The dye concentration present in trace fibers is extremely low.

Therefore, highly sensitive, efficient and accurate analytical techniques should be developed for forensic trace fiber analysis. Additionally, degradation associated with environmental exposures could influence comparative analysis. For example, a long time exposure to outdoor weathering could cause photodegradation of dyed fibers. Hence, the objective of this study is to develop analytical methods that can be used to detect and characterize organic dyes present in trace textile fibers with minimum sample consumption, and to understand the photodegradation process of a widely used disperse dye (C.I. Disperse Red 1) for polyester during outdoor weathering.

Time-of-flight secondary ion mass spectrometry (TOF-SIMS) was successfully employed to detect trace amount of C.I. Acid Blue 25 and C.I. Disperse Red 1, by analyzing fiber cross sections of thickness 500 nm~700 nm. High resolution images demonstrating dye distribution and concentration were obtained.

Additionally, operation conditions for an innovative sample preparation platform, automated microfluidic extraction device, were optimized to extract organic dyes with high efficiency to avoid the risk of contamination during sample preparation. With automated extraction, trace fibers (~2-3 mm) were successfully extracted within less than 10 min using around 9 µL solvent. Concentrated dye extracts from millimeter-length-threads and single fibers were collected and analyzed by subsequent high performance liquid chromatography separation followed by mass spectrometry. Single fibers (<5 mm length) containing 1% on- weight-of fiber (owf) were successfully detected.

For comparative analysis, the most efficient and rapid method is to compare the unknown dyes to reference standards. In this aspect of the project, dye standards collected from different manufacturers were analyzed using a high performance liquid chromatography-photodiode array detection-electrospray ionization-quadrupole time of flight mass spectrometer (HPLC-DAD-ESI-QTOF-MS). Information regarding retention time, UV- vis absorption and mass spectra of the standards was recorded in a dye database for forensic purposes. For example, C.I. Direct Blue 106 contained two main components having retention times at 4.020 min and 4.353 min in HPLC-DAD chromatograms, and both components had an absorption maximum at 635 nm, and the doubly charged ions were identified as characteristic ions with a mass-to charge ratio (m/z) at 346.9890.

Finally, a preilimanry photodegradation study of C.I. Disperse Red 1 on polyester by using tandem mass spectrometric analysis is presented. Three major degradation products were characterized by targeted collision-activated dissociation and the photofading pathway was established. In this regard, the predominant product (E)-N-ethyl-4-((4-nitrophenyl)

diacetyl) aniline was produced rapidly and then was converted to (E)-4-((4-nitrophenyl) diazenyl) aniline. The photodegradation process was compared with that obtained in solution.

Forensic Analysis of Organic Dyes on Trace Fibers and Their Photodegradation

by Min Li

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the Degree of Doctor of Philosophy

Fiber and Polymer Science

Raleigh, North Carolina

2016

APPROVED BY:

______David Hinks Nelson R. Vinueza Co-Chair of Advisory Committee Co-Chair of Advisory Committee

______Harold S. Freeman Ahmed El-Shafei

______Peter. Bloomfield

BIOGRAPHY

Min Li was born in Taiyuan, Shanxi, China. After completion of her high school education in

2006, she attended Beijing Institute of Fashion Technology, where she received two Bachelor

Degrees of Engineering, one in Textile Chemistry and the other one in Fashion Design. Then she joined Novozymes. Inc. Beijing, China, where she worked as a lab assistant for textile industry in the department of research development. In 2012, she got her Master’s Degree in

Textile Chemistry from North Carolina State University, and in the same year, she was awarded first place honors in the Herman and Myrtle Goldstein Student Paper Competition at the American Association of Textile Chemists and Colorists International Conference held in

Charlotte, North Carolina for her paper entitled “An Environmentally Benign Approach to

Cotton Preparation: One Bath Enzymatic Desizing, Scouring and Activated Bleaching.”

In 2012 she continued her studies by pursuing a Ph.D. Degree in Fiber and Polymer Science

(FPS). Her research focused on forensic analysis of dyed fibers, as well as the photodegradation of an important disperse dye under the guidance of Dr. David Hinks, Dr.

Nelson R. Vinueza and Dr. Harold S. Freeman.

ACKNOWLEDGMENTS

My deep gratitude goes first to my advisors Dr. David Hinks and Dr. Nelson R. Vinueza, who expertly guided me through my graduate education and who shared the excitement of research and discovery. I would like to thank both of you for offering me the opportunity to pursue the doctoral degree in your research team and guiding me to grow as a research scientist. Your enthusiasm toward research has deeply inspired me to learn and grow.

I would also like to thank my committee members, Dr. Harold S. Freeman, Dr. Ahmed El-

Shafei, and Dr. Peter Bloomfield, for being my committee members and guiding my research. All of you provided your brilliant comments and suggestions during my research.

Dr. Freeman offered me tremendous help in the research about photodegradation, and his wisdom, kindness, and enthusiasm toward research deeply influenced me. Dr. Peter

Bloomfield, who is always nice, patient and ready to help, helped me in statistical data analysis and modeling.

Grateful thanks are expressed to Elaine (Chuanzhen) Zhou, and everyone in AIF at NC State

University. Elaine taught me about the concept and operation of Time-of-flight Secondary ion mass spectrometry, and offered me great suggestions on my research and personal life. I would thank Mr. Roberto Garcia, who offered training on cryomicrotoming and XRD. I thank Ms. Judy Elson, for her tremendous help with dye collection, microscopy training, weathering test and her patience and kindness. I would also thank Dr. Malgorzata Szymczyk for her guidance and help with the dye synthesis.

I also thank everyone in my two research teams: Guan Wang, Nanshan Zhang, Sha Fu and

Xiuzhu Fei in Dr. Hinks’ group, who are always supportive and helpful with my research. I iii

thank Yufei Chen, Kelsey Boes, Emily Lichtenberger, and Nadia Sultana, who consistently offered help with my research and study. I really enjoyed working with them and learning from them.

Finally, and most importantly, I am grateful to my family and friends, for their love, support and encouragement throughout my life.

TABLE OF CONTENTS

LIST OF TABLES ...... x

LIST OF FIGURES ...... xii

LIST OF ABBREVATIONS ...... xxiii

Chapter 1 Introduction...... 1

1.1. Motivation ...... 1

1.2. Literature Review ...... 4

1.2.1. Introduction to organic dyes ...... 4

1.2.2. Photodegradation of dyes ...... 16

1.2.3. Forensic fiber examination ...... 26

1.2.4. Analytical techniques for trace dye analysis ...... 28

1.3. Instrumentation...... 44

1.4. Reference ...... 46

Chapter 2 Microfluidic System for Automated Dye Molecule Extraction and Detection and Characterization ...... 57

2.1. Preliminary study on dye extraction at conventional scale ...... 58

2.1.1. Dyes and chemicals...... 58

2.1.2. Sample preparations ...... 61

2.1.3. Extraction ...... 62

2.1.4. Measurement ...... 64

2.1.5. Results and discussions ...... 67

2.2. Macro-level Optimization of Direct Dye Extraction ...... 69

2.2.1. Full factorial design ...... 69

2.2.2. Statistical analysis ...... 71

2.2.3. Optimization of extraction ...... 73

2.2.4. Modeling of solvent evaporation ...... 74

2.3. Optimization of C.I. Acid Blue 25 Extraction from Nylon Fibers by HPLC-DAD-

MS 77

2.3.1. Acid dye extraction ...... 77

2.3.2. Central composite design (CCD) of acid dye extraction ...... 78

2.3.3. Statistical analysis ...... 80

2.3.4. Summary ...... 86

2.4. Automated Microfluidic Extraction –HPLC-DAD-MS Analysis of Dyes on Trace

Fibers 87

2.4.1. Sample collection ...... 87

2.4.2. Design and fabrication of microfluidic apparatus -1st Generation ...... 87

2.4.3. Quantitative analysis of dye extraction ...... 91

2.4.4. LC-QTOF-MS Mass detection ...... 91

2.4.5. Quantification and Method Detection Limit (MDL) ...... 96 vi

2.4.6. Microfluidic Extraction-MS analysis of single fiber ...... 97

2.5. Summary ...... 98

Chapter 3 High Performance Liquid Chromatography-Photodiode Array-Mass

Spectrometric Analysis of Direct Dyes ...... 100

3.1. HPLC-DAD-MS analysis of reference direct dyes ...... 100

3.1.1. Method development of chromatography ...... 101

3.1.2. Modification of gradient ...... 103

3.2. Method validation ...... 104

3.2.1. Limit of detection/quantification (LOD/LOQ) ...... 104

3.2.2. Linearity ...... 105

3.2.3. Specificity ...... 106

3.3. HPLC-DAD-MS analysis of direct dye standards ...... 108

3.3.1. LC-MS analysis of oxazine dyes ...... 109

3.3.2. LC-MS analysis of phthalocyanine dyes ...... 113

3.3.3. LC-MS analysis of azo direct dyes...... 114

3.4. References ...... 120

Chapter 4 Chromatographic Optimization of Azo Disperse Dyes ...... 121

4.1. Introduction ...... 121

4.2. Experimental ...... 121

4.3. Results and Discussions ...... 123 vii

4.4. Conclusions ...... 134

Chapter 5 ESI QTOF Tandem Mass Spectrometric Analysis of Sulfonated

Anthraquinone Acid Dyes ...... 135

5.1. Introduction ...... 136

5.2. Experimental ...... 138

5.2.1. Chemicals and materials ...... 138

5.2.2. Instrumentation ...... 139

5.3. Results and Discussions ...... 140

5.3.1. ESI mass spectra of individual dyes ...... 140

5.3.2. Tandem mass spectrometry analysis of single dye ...... 141

5.3.3. Proposed mechanisms of Collision-Induced Dissociation ...... 147

5.4. Conclusions ...... 152

Chapter 6 Photodegradation of C.I. Disperse Red 1 ...... 153

6.1. Introduction ...... 154

6.2. Experimental ...... 155

6.2.1. Dyes and solvent ...... 155

6.2.2. Accelerated weathering of dyed polyester ...... 156

6.2.3. Dye extraction ...... 157

6.2.4. UV irradiation of dye solvent under regular lamp light and UV light (330 nm)

157 viii

6.2.5. HPLC-DAD-MS analysis of dye extract ...... 158

6.2.6. LC-Auto MS/MS analysis of degradation products...... 159

6.2.7. Synthesis of standard compounds for degradation products ...... 159

6.3. Results and Discussion ...... 160

6.3.1. Spectrophotometric Test ...... 160

6.3.2. Characterization of synthesized compounds ...... 160

6.3.3. Screening degradation product candidates with LC-auto MS/MS analysis .... 164

6.3.4. Confirmation of degradation structure via LC-MS/MS analysis ...... 167

6.3.5. LC-MS analysis of the degradation products in ethyl acetate ...... 169

6.4. Conclusions ...... 174

Chapter 7 TOF-SIMS Method Development And Identification of Acid Dyes in Nylon

Fibers ...... 176

7.1. Introduction ...... 178

7.2. Experimental Section ...... 179

7.2.1. Nylon dyeing procedure ...... 179

7.2.2. TOF SIMS Sample Preparation Procedure ...... 180

7.3. ToF-SIMS Analysis ...... 181

7.4. Results and Discussion ...... 182

7.5. Conclusions ...... 194

7.6. References ...... 195 ix

LIST OF TABLES

Table 1-1 Summary of important dye classes and their applications ...... 4

Table 1-2. Summary of dye classes and the typical fiber types to which they are applied ...... 9

Table 1-3. Complementary colors...... 17

Table 1-4 Ionization techniques applied for synthetic dyes ...... 32

Table 2-1 Preliminary Screening of Extracting Solvent ...... 64

Table 2-2 Coded and uncoded values of independent variables in extraction experiment ..... 70

Table 2-3 Design of Screening experiment for direct dye extraction ...... 70

Table 2-4 Design of experiment of solvent evaporation ...... 71

Table 2-5 DOE of evaporation and results ...... 74

Table 2-6 Coded and uncoded values of independent variables in central composited design for acid dye extraction ...... 79

Table 2-7 Central composite design (CCD) of acid dye extraction at conventional lab scale 80

Table 2-8 ANOVA analysis of acid dye extraction with dad peak area as the response ...... 81

Table 2-9 ANOVA Analysis of acid dye extraction by peak areas of extracted ion chromatograms ...... 84

Table 2-10 Information of anionic dyes...... 87

Table 2-11 Mass spectra of dye extract from microfluidic extraction vs. conventional extraction...... 92

Table 3-1 Time table of B% in Original Method ...... 101

Table 3-2 HPLC gradient of method 1 ...... 101 x

Table 3-3 HPLC gradient of method 2 ...... 103

Table 3-4 Comparison of ion abundance of direct yellow 106 in liquid chromatograms .... 104

Table 3-5 Information of direct dye standards for LC-MS analysis ...... 108

Table 3-6 Analysis of optimized chromatography for direct dyes ...... 120

Table 4-1 Azo disperse dyes for LC-MS analysis...... 121

Table 4-2 Comparison of APCI and ESI mass ionization ...... 124

Table 4-3 HPLC gradient for disperse dyes ...... 125

Table 4-4 Peak information of LC-MS of disperse azo dyes...... 128

Table 4-5 Anthraquinone dyes for HPLC-DAD-MS separation and identification ...... 128

Table 5-1 Information of C.I. Acid Dyes ...... 138

Table 5-2 MSMS Fragmentation under CID 40 V ...... 145

Table 6-1 HPLC gradient for C.I. Disperse Red 1 ...... 158

Table 6-2 HPLC-MS/MS analysis of synthesized standard compounds and degraded sample

(after 20 hours) ...... 162

Table 7-1 Formulation used to embed nylon fibers ...... 180

LIST OF FIGURES

Figure 1-1. Examples of direct dyes. (a) C.I. Direct Yellow 86; (b) C.I. Direct Blue 71. (c)

C.I. Direct Blue 86; (d). Triphenodioxazine direct dyes. (Sirius Light Violet FRL)...... 6

Figure 1-2. Examples of acid dyes. (a) C.I. Acid Red 114; (b) C.I. Acid Green 16; (c) C.I.

Acid Red 186 (d) Acid Blue 25...... 7

Figure 1-3 Typical structures of disperse dyes. (a) C.I. Disperse Yellow 58; (b) C.I. Disperse

Yellow 77; (c) C.I. Disperse Blue 27; (d) C.I. Disperse Yellow 3...... 8

Figure 1-4. Example of Azo Dyes. (a) C.I. Disperse Red 1 (Monoazo); (b) C.I. Acid Black 1.

(Diazo); (c) C.I. Direct Red 81...... 10

Figure 1-5. Tautomerism of azo dyes...... 11

Figure 1-6. Acid Black 52 (2:3 Cr Complex, Mixture of 1:1 Cr Complex and 1:2 Cr

Complex, and L represents ligands)...... 11

Figure 1-7. Examples of anthraquinone dyes. (a) Basic structure of anthraquinone dyes; (b)

C.I. Disperse Red 15; (c) C.I. Acid Blue 25...... 12

Figure 1-8 Example of typical triarylmethane dyes. (a) Malachite green; (b) Ethyl violet; (c)

Methyl violet; (d) C.I. Acid Green 16)...... 13

Figure 1-9. Examples of other dye structures. (a) Xanthene C.I. Acid Orange 11; (b) Acridine

C.I. Acid Yellow 118; (c) Anthrone, Vat Orange 9; (d) Thioindiogid; (e) Indigoid Dyes. .... 14

Figure 1-10. Examples of dyes in NCSU-Max. A. Weaver Dye Library...... 15

Figure 1-11. Electromagnetic spectrum...... 16

Figure 1-12 Electron transitions between electron orbitals...... 18

Figure 1-13. Energy absorption and release by organic compounds...... 19 xii

Figure 1-14. The five classes of fading rate curve. I- First or second order fading; dyes present as single molecules or very small aggregates. II- Initially as I followed by zero-order fading. III. Zero order fading; dye in large aggregates or firmly enmeshed in substrate molecules. IV and V-dye in relatively large aggregates, affected by heat...... 20

Figure 1-15. Color shift of colorant...... 21

Figure 1-16. Tautomerism of anthraquinone compounds...... 22

Figure 1-17. Photodegradation of C.I. Disperse Red 1 on polyester and nylon substrates. ... 24

Figure 1-18. Photofading of triphenylmethane dye...... 26

Figure 1-19. Flow chart of forensic dye examination...... 27

Figure 1-20. Negative mode electrospray ionization source...... 33

Figure 1-21. Schematic of secondary ionization mass spectrometry...... 35

Figure 1-22. Isolation of parent ions using quadrupole mass analyzer...... 37

Figure 1-23. Viable band pass mass filter: By adjusting the slope of the mass scan line, the band pass region of the mass filter can be varied. The slope of the mass scan line or operating line is given by a/q=2U/V...... 38

Figure 1-24. Time-of-flight analysis of ions of various masses. For clarity and simplicity, this shown in a linear time-of-flight mass spectrometer that does not have an ion mirror...... 40

Figure 1-25. Time-of-flight mass spectrometer schematic...... 42

Figure 1-26 Schematic of tandem mass spectrometry...... 43

Figure 1-27. Schematic of Agilent 6520 QTOF mass analyzer in negative mode...... 45

Figure 2-1. Structures of direct dyes for preliminary extraction...... 59

Figure 2-2. Data Color Ahiba dyeing machine...... 62

Figure 2-3. Preliminary extraction of fabrics for UV-vis spectrometry...... 63 xiii

Figure 2-4. Pierce Reative-therm heating module and nitrogen gas delivery system...... 64

Figure 2-5. Agilent Cary 300 UV/Vis Spectrophotometer...... 65

Figure 2-6. Color differences of C.I. Disperse Red 1 on polyester before and after photodegradation...... 67

Figure 2-7. Extraction percentage with extraction solvents (organic solvent: H2O 4:3)...... 68

Figure 2-8. Interactions between extraction time and temperature on extraction performance.

...... 72

Figure 2-9. Statistical model of direct dye extraction percentage vs. time at room temperature

(25 °c) (a) Extraction percentage with 200 µl extraction solvent; (b) Extraction percentage with 800 µL extraction solvent...... 73

Figure 2-10. Cube analysis of the direct dye extraction...... 74

Figure 2-11. Statistic model for solvent evaporation by JMP...... 75

Figure 2-12. Prediction profiler of evaporation time...... 76

Figure 2-13 Conventional extraction procedures of fibers for LC/MS analysis...... 78

Figure 2-14. Contour plots of independent variables on the peak areas of HPLC-DAD peaks at 660 nm...... 82

Figure 2-15. Estimated response surfaces with related contours by plotting DAD-Peak areas versus time and temperature...... 83

Figure 2-16. Contour plots showing the effects of interactions between variables on peak area of extracted ion chromatograms...... 85

Figure 2-17. Microfluidic extraction apparatus ...... 88

Figure 2-18. Microfluidic chip and elastomeric FFKM extraction chamber ...... 88

Figure 2-19. Control window displaying operation of components and times ...... 90 xiv

Figure 2-20. Extraction parameter inputs (Note: the unit of time is second) ...... 90

Figure 2-21. Mass spectrum of C.I. Acid blue 25 isolated from nylon fibers and C.I. Direct

Red 81 from cotton fiber by pyridine extraction in conventional condition. (a) The shown mass spectrum of C.I. Acid Blue 25 was acquired at a retention time of 5.240 min. The inserts show a detail of the spectrum at m/z 393.0549; (b) The mass spectrum of C.I. Direct

Red 81 was acquired at a retention time of 4.501 min, with the detailed spectrum at

314.5337...... 93

Figure 2-22. Discoloration of C.I. Direct Red 81 dyed fibers before and after automated microfluidic extraction...... 94

Figure 2-23. LC-DAD-QTOF-MS data for C.I. Direct Red 81 extracted from a cotton fiber in microfluidic extraction system. (a) Total ion chromatogram; (b) DAD chromatogram at

540 nm; (c) Extracted ion current for m/z 314.53; (d) mass spectrum corresponding to the peak observed in the m/z 314.5333 extracted ion current...... 95

Figure 2-24. LC-QTOF-MS data for C.I. Acid Red 114 out of microfluidic extraction system.

(a) DAD chromatogram at 540 nm; (b) Extracted ion current for m/z 718.1019; (c) Mass spectrum corresponding to the peak observed in the m/z 718.1019 extracted ion current. .... 96

Figure 2-25. Quantification of C.I. Acid Blue 25 extract from textile threads (Note: the blue dot represents the dye extract from nylon fibers containing 0.0625%owf C.I. Acid Blue 25, 3 mm in length, and 0.2 mm in diameter)...... 97

Figure 2-26. MS Spectra of dye extract from single fibers (C.I. Acid Blue 25 and C.I. Acid

Yellow 49) (~5 mm) with Automatic Microfluidic Extraction-MS. (a) Mass spectra of dye extract of C.I. Acid Blue 25 and C.I. Acid Yellow 24; (b) Mass spectra of C.I. Acid Blue

25;(c) Mass Spectra of C.I. Acid Yellow 49)...... 98 xv

Figure 3-1. Chromatograms of C.I. Direct Yellow 106 at 410 nm with HPLC gradient of method 1...... 102

Figure 3-2. Mass spectra of method 1 (Green) and the original method (Red)...... 102

Figure 3-3. Comparisons of chromatograms of HPLC-DAD-MS at 660 nm. Original method

(blue), Method 2 (green), and method 3(red)...... 103

Figure 3-4. Limit of quantification of C.I. Direct Yellow 106...... 105

Figure 3-5. Linear range of peak areas vs. concentration C.I. Direct Yellow 106...... 106

Figure 3-6. Optimized HPLC-DAD chromatography and gradient B%...... 107

Figure 3-7. DAD (660 nm) Chromatogram of Direct Blue 106. UV-Visible spectrum of

Direct Blue 106 at 4.020 min and 4.353 min...... 109

Figure 3-8. MS Peaks of C.I. Direct Blue 106. (a) Doubly charged dye molecule (M-2H)-2;

(b) Doubly charged dimer (2M-2H)-2...... 110

Figure 3-9. DAD Chromatograms of Direct Blue 108. (a) UV-Visible spectrum of Direct

Blue 106 at 660nm; (b) Red substance detected at chromatograms at 540 nm; (c) the UV-Vis spectra at multiple retention times...... 111

Figure 3-10. LC-MS Mass spectra of Direct Blue 108. (a) UV-Visible spectrum of C.I. Direct

Blue 108 at 660nm; (b) Doubly charged dye molecule (M-2H)-2; (c) Doubly charged dimer

(2M-2H)-2...... 112

Figure 3-11. LC-MS Mass spectra of C.I. Direct Blue 86. (a) UV-Visible spectrum of

C.I.Direct Blue 86 at 660nm; (b) Doubly charged dye molecule (M-2H)-2; (c) Doubly charged dimer (2M-2H)-2...... 113

xvi

Figure 3-12. (a) DAD chromatogram of C.I. Direct Green 26 at 4.303 min; (b) LC-MS Mass spectra of ionized dye molecules of C.I. Direct Green 26; (c)Isotopic pattern of the dominant ion peak at (M-3H)-3 of C.I. Direct Green 26 at 4.247 -4.440 min...... 115

Figure 3-13. LC-DAD-MS analysis of Direct Green 6 at 660 nm at 3.997 min. (a)

Chromatograms obtained from C.I. Direct Blue 90 at 660 nm; (b) Chromatograms obtained from C.I. Direct Blue 90 at 254 nm...... 116

Figure 3-14. LC-DAD-MS analysis of Direct Green 6 at 660 nm at 3.997 min. (a)

Chromatograms obtained from C.I. Direct Blue 90 at 660 nm; (b) Chromatograms obtained from C.I. Direct Blue 90 at 254 nm...... 117

Figure 3-15. Chromatograms of C.I. Direct Red 2 at 3.800 min...... 118

Figure 3-16. The HPLC chromatogram of C.I. Direct Blue 15...... 119

Figure 4-1. Color and substituent groups of disperse Azo dyes. a. UV-Vis Absorption of

MDL3, MDL4 and MDL 5. (b) UV absorption of MDL 10, and MDL5. (c) UV Vis of

Disperse Red 1 and MDL3...... 126

Figure 4-2. LC-DAD detection of Azo dye mixtures: a 254 nm, b: 410 nm, c: 540 nm, d: 660 nm. (Compounds: 1,3, 5 are blue substances in disperse blue 3...... 127

Figure 4-3. LC-DAD-MS analyses of mixture of disperse anthraquinone dyes...... 131

Figure 4-4 (a) Chromatograms of C.I. Disperse Blue 56 with old chromatograms;(b)

Expected structure observed at 9.467 min; (c) Mass spectrum of the main component of C.I.

Disperse Blue 56 detected at retention time of 6.220 min...... 132

Figure 4-5. (a) Chromatograms of C.I. Disperse Blue 3 with HPLC-DAD with a newly developed chromatographic method; (b) Chromatograms with a previous developed method;

Figure 4-6. Mass spectrometry of components in C.I. Disperse Blue 3. (a) Structure proposed in Colour Index; (b) Unknown blue substance at retention time of 1.753 min; (c) component

2 retained at 3.544 min; (d) component 4 that were detected at 4.874 min...... 134

Figure 5-1. Basic structure of the 1-Amino-2 sulfonated anthraquinone dyes. R: Alkyl or aromatic groups...... 137

Figure 5-2. Structure of sulfonated anthraquinone acid dyes...... 139

Figure 5-3. Negative-Ion ESI mass spectra of C.I. Acid Blue 62...... 141

Figure 5-4. LC-MS/MS mass spectra of C.I. Acid Blue 25 CID Voltage = 40 V...... 142

Figure 5-5. Fragmentation scheme and Mass Spectra of LC-MS/MS analysis of C.I. Acid

Blue 62 CID Voltage = 40 V...... 143

Figure 5-6. LC-MS/MS Mass Spectra of C.I. Acid Blue 277 at CID Voltage = 40 V...... 144

Figure 5-7. LC-MS/MS Spectra of C.I. Acid Blue 45...... 144

Figure 5-8. Fragmentation pathway of acid blue dyes (the number indicates the orders of peak intensity of fragments in LC-MS/MS mass spectra) all dyes lost SO2 as a characteristic neutral loss. The order of fragmentation was determined by the energy required to cleave off substituents...... 146

Figure 5-9. Proposed mechanism of fragmentation of sulfonated anthraquinone CID, where the ortho-NH3 favors the rearrangement of SO2...... 147

Figure 5-10. Model Compounds Compound A and Compound B ...... 148

Figure 5-11. Tandem mass spectrum of (1 amino anthraquinone sulfonic acid) containing 32S

(a), and 34S (b)...... 148

Figure 5-12. Comparisons of survival yields of compounds A and B. Compound A with NH3 has higher survival yield under CID Voltage = 40 V...... 150 xviii

Figure 5-13. Fragmentation pathways of C.I. Acid Blue 129 under CID Voltage = 40 V. .. 151

Figure 5-14. LC-MS/MS mass spectra of C.I. Acid Blue 129 under CID Voltage = 40 V. . 151

Figure 6-1. Structure of C.I. Disperse Red 1 (Compound 1) ...... 155

Figure 6-2. Irradiance of light source of the simulated outdoor weathering condition ...... 156

Figure 6-3 Irradiance of regular lamp light spectra...... 157

Figure 6-4 Synthesis of Standard Compounds...... 159

Figure 6-5. K/S value of Photodegradation of Dye 1 ...... 160

Figure 6-6 Thin layer chromatograms of synthesized products...... 161

Figure 6-7 HPLC-DAD-MS analysis of synthesized compound 2 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours...... 162

Figure 6-8 HPLC-DAD-MS analysis of synthesized compound 3 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours...... 163

Figure 6-9 HPLC-DAD-MS analysis of synthesized compound 4 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours...... 163

Figure 6-10. LC-MS/MS analysis of Disperse Red 1 extract out of unexposed fabrics ...... 164

Figure 6-11. Proposed structure of three degradation products ...... 165

Figure 6-12. Compound 2 at m/z=287.1139 out of photodegradation of Disperse Red 1 in ethyl acetate ...... 165

xix

Figure 6-14. Fragmentation pattern of compound 4 (m/z=243.0877) out of photodegradation of Disperse Red 1 in ethyl acetate. LC-MS/MS Mass spectra of compound 1 after 5 hours (a),

10 hours (b), 40 hours (c) and 80 hours (d)...... 167

Figure 6-15. MS/MS Mass spectra of authentic samples and degradation products, (a)

Compound 2 in dye solution after 20 hour-regular lamp irradiation, (b) Authentic sample for compound 2, (c) Compound 3 in dye solution after 20 hour-regular lamp irradiation, (d)

Authentic sample for compound 3. (e) Compound 4 in dye solution after 20 hour-regular lamp irradiation, (f) Authentic sample for compound 4...... 168

Figure 6-16 LC-MS/MS mass spectra of Compound 5 (top)and Compound 6 (bottom) found in dye solution after UV irradiation at 330 nm...... 169

Figure 6-18. Extracted Ion Chromatogram (EIC) of degradation products after 80 hours: compound 1 (a) at 2.760 min, compound 2 at 5.36 min (b) and compound 3 at 2.005 min (c).

...... 170

Figure 6-19. Relative ion abundances of Cps 2-4 out of the photodegradation of Disperse Red

1, (a) Cps 2-4detected in ethyl acetate in lamp light, (b) compounds in ethyl acetate after irradiation with Ragular lamp light, (c) Cp2-4 major compounds in UV wavelength of 330 nm, (d) Cp5-6 in UV irradiation at 330 nm...... 171

Figure 6-20. Proposed photodegradation pathways of disperse red 1 (Compound 1) ...... 172

Figure 6-21. Relative Ion abundance of degradation products in polyester and ethyl acetate:

The production curve of compound 2(b), compound 3(b) compound 4(c) and a comparison of ion production in fabrics and solvent (d) ...... 174

Figure 7-1. (a) Chemical structure and molecular weight of Acid Blue 25 (AB25). Negative ion ToF-SIMS spectra showing the molecular ion of AB25 obtained from (b) dye powder, (c)

1% on-weight-of-fabric (owf) dyed nylon surface, and (d) 1% owf dyed nylon cross section.

...... 183

Figure 7-2. Reconstructed negative ion ToF-SIMS spectra obtained from 1% owf Acid Blue

+ 25 dyed nylon surface with C60 ion beam sputtering at varied time. Spectra (a – e) are reconstructed from 20 frames of 128 x 128 pixels, 1 shot/pixel acquisition. (a) at surface, 0s

14 -2 C60 sputtering, (b) at 14.6s C60 sputtering (PIDD 1 x 10 cm ), (c) at 219s C60 sputtering

15 -2 15 -2 (PIDD 1.5 x 10 cm ), (d) at 438s C60 sputtering (PIDD 3.0 x 10 cm ) and (e) at 1022s C60 sputtering (PIDD 7.0 x 1015 cm-2). (f) Reconstructed accumulated spectrum from 1400 frames of 128 x 128 pixels, 1 shot/pixel acquisition obtained from Acid Blue 25 dyed nylon surface with 1022s C60 sputtering...... 185

Figure 7-3. Negative ion ToF-SIMS spectrum obtained from 0.5% owf Acid Blue 25 dyed nylon surface (a) without C60 sputtering, and (b) with C60 sputtering for 14.6s at 1 nA

+ followed by spectrum acquisition with Bi3 at 1 shot/pixel for 20 frames and this cycle was repeated for 30 times. Both spectra (a) and (b) are reconstructed from 600 frames of 128 x

128 pixels, 1 shot/pixel acquisition...... 187

xxi

Figure 7-4. Negative ion ToF-SIMS spectrum obtained from 0.1% owf Acid Blue 25 dyed

+ nylon surface with C60 ion beam sputtering acquired under the same conditions as in Figure

3b...... 188

Figure 7-5. ToF-SIMS images (100 µm x 100 µm) of a 1% owf Acid Blue 25 dyed nylon fiber cross section showing the spatial distribution of (a) CN-, the characteristic ion of nylon,

- (b) molecular ion of Acid Blue 25, (c) C3H3O2 , the characteristic ion of the embedding resin, and (d) overlaid image of resin (in red) and the Acid Blue 25 molecular ion (in blue). The images are reconstructed from 200 frames of 256 x 256 pixels, 1 shot/pixel acquisition. ... 189

Figure 7-6. ToF-SIMS images (100 µm X 100 µm, 256 x 256 pixels, 1 shot/pixel) of 0.1%

+ owf Acid Blue 25 dyed nylon cross section acquired with Bi3 beam only. The total acquisition was 600 frames...... 190

Figure 7-7. ToF-SIMS images (100 µm x 100 µm) of 1% owf Acid Blue 25 dyed nylon cross section acquired on the same spot from a fresh cross section surface, after 200 frames of acquisition, and after various C60 sputtering times. All images are reconstructed from 50 frames of 256 x 256 pixels, 1 shot/pixel acquisition except the stack image. Image reconstruction via summation of all images (Stack image lower right above) provides significant improvement in signal to noise...... 192

Figure 7-8. ToF-SIMS images (100 µm X 100 µm, 256 x 256 pixels, 1 shot/pixel) of 0.1% owf Acid Blue 25 dyed nylon cross section acquired with C60 sputtering. The total acquisition was 600 frames...... 193

xxii

LIST OF ABBREVATIONS

εmax Molar extinction coefficient

λmax: Absorption maximum wavelength

TOF Time-of-flight

HPLC High performance liquid chromatography

MS Mass spectrometry

DAD Photodiode array detector

ESI Electrospray ionization

APCI Atmospheric pressure chemical ionization

SIMS Secondary ion mass spectrometry

CID Collision induced disassociation

MS/MS Tandem mass spectrometry

xxiii

Chapter 1 Introduction

1.1. Motivation

Trace evidence plays important roles in criminal investigations and forensic analysis. Crime investigator collect crucial trace evidence to determine what had happened in a crime scene.

Crimes often involve direct physical contact between the suspects and the victims.1 Trace evidence, such as fibers, hairs, dust, pollen, and soil, are highly likely to be transferred between objects, and are persistent in objects for a time.2 Trace evidence is of significance due to the transferable property suggesting what has happened and who was involved.

Textile fibers are regarded as one of the most important types of trace evidence, owing to their wide availability and specificity.2 Textiles are widely available in clothing, carpet, automobile fabrics and ropes. Fibers collected from different places compared in both physical and chemical properties. Chemical specificity of colorants (pigments and dyes) and additives provide an extra level of discriminatory value of fiber evidence. Therefore examination of trace fibers is of significance to improve probative value of textile evidence.

Examination of trace fiber evidence is achieved by a variety of analytical tools including microscopy, chromatography and mass spectrometry. UV-visible microspectrophotometry

(MSP), infrared microspectroscopy and Raman spectroscopy are commonly used technologies at regular forensic labs.3 Non-destructive analytical tools are highly preferable to preserve fragile trace evidence; other techniques such as pyrolysis coupled to gas chromatography, high performance liquid chromatography (HPLC) and mass spectrometry

(MS) analysis are becoming increasingly important for forensic use to achieve high efficiency and accuracy.

The first challenge associated with trace amount of dye analysis is limited sample size. In many cases textile fibers less than 2 mm are all that are found at a crime scene. Short fiber fragments (under 2.5 mm) are regarded to have higher value than longer fibers, because they are stuck to objects for much longer time. It was estimated that fibers in length of 2 mm contain around 2-200 ng dye content. Highly sensitive separation techniques such as ultra- performance liquid chromatography equipped with photodiode array detector separate and detect color substance with a detection limit of as low as 0.13~1.44 ppb, it can hardly detect fibers less than 0.05 mm, due to sample loss during sample preparation. Conventional micro extraction is labor intensive and has a risk of contamination during sample preparation and sample handling. Therefore, a highly efficient sample-handle free extraction platform is necessary for trace-level fiber analysis.

In addition, the most rapid and convenient method for comparative analysis is refereeing to standard dye references. Therefore, constructing a dye database for forensic purpose were proposed by collection, analysis and registration of dyes from various dye manufacturers are necessary.

The last challenge associated with comparative analysis is sample degradation associated with environmental changes. For example, dyed fibers exposed to outdoor weathering could cause complication to comparative analysis. Understanding the photodegradation process helps to know the forensic significance of textile fibers. Dr. Morgan’s group also investigated photodegradation process of commonly used acid dyes, and new compounds were detected

after certain time periods of UV exposure. However, the photodegradation process is not fully understood.4

In order to solve above problems involved in trace fiber analysis, the objective of this research are listed below: (1) develop highly sensitive and reliable methods to detect dyes with minimal destruction of materials using time-of-flight secondary ion mass spectrometry

(TOF-SIMS); (2) optimize extraction conditions of common dyes in an automated microfluidic extraction system for sub-micro fiber dye extraction; (3) understand the photofading pathways of a widely used disperse dye (C.I. Disperse Red 1) on polyester during outdoor weathering using electrospray ionization tandem mass spectrometry analysis.

1.2. Literature Review

1.2.1. Introduction to organic dyes

1.2.1.1. Classification by application

Dyes are classified by either application methods (i.e. direct, acid, disperse) or by chemical structure (i.e. azo, anthraquinone, triarylmethane).5 Forensic analysis identification of dyes is typically based on the application since common organic dyes are associated with fiber types.

Information about dye application and typical dye classes are summarized in Table 1-1.6

Table 1-1 Summary of important dye classes and their applications

Class Typical substrates Method of application Chemical classes Acid Nylon (polyamide), wool, From neutral to acidic acid baths: ionic bond Azo (including silk, paper, inks, and between dye molecule and polymer premetalized), leather anthraquinone, triphenylmethane, azine, xanthene, nitro and nitroso Basic Paper, polyacrylonitrile, Applied from acidic dye baths Cryanime, hemicyanine, modified nylon, polyester diazahemicranine, and inks diphenylmethane, triacrylmethane, azo, azine, xanthene, acridine, oxazine and anthrquninone Direct Cotton, rayon, paper, Applied from neutral or slightly alkaline baths Azo, phthalocyanine, leather and nylon containing additional electrolyte stilbene, and oxazine. Disperse Polyester, polyamide, Fine aqueous dispersions often applied by high Azo, anthraquinone, styryl, acetate, acrylic and plastics temperature/pressure or lower temperature carrier nitro, and benzeodifuranone. methods; dyes may be padded on cloth and baked on or thermo fixed. Hydrogen bonds and weak van der Waals forces hold the dye in the fiber Fluorescent Soaps and detergents, all Stilbene, pyrazoles, brighteners fibers, oils, paints and coumarin, and plastics naphthalimides Reactive Cotton, wool, silk, and Reactive cite on dye react with functional groups Azo, anthraquinone, polyamide of fiber to bind covalently under influence of phthalocyanine,formazan, heat and pH and benzeodifuranone Solvent Plastics, gasoline, stains, Dissolution in the substrate Azo, triphenylmethane, inks, fats, oils, and waxes anthraquinone, and phthalocyanine Vat Cellulosics Water-insoluble dyes solubilized by reducing Anthraquinone (including with sodium hydrogen sulfite, then exhausted on polycyclic quinones) and fiber and reoxidized. indigoids

Direct dyes obtain their name due to inherent substantivity to substrates. They are directly applied into cellulosic fibers (i.e. cotton, rayon, paper) in the presence of heat and an electrolyte.7 Substantivity of direct dyes are achieved by hydrogen bonding and van der

Waals, and the strength of interaction increases with the size of molecules. Therefore direct dyes are designed to have planar and conjugated double-bond systems to increase substantivity toward fabrics. Direct dyes have many applications in industry and life science.8

Direct dyes are available for a full range of hue with structures belonging to azo, oxazine, copper phthalocyanine and other classes, among which azo dyes constitute the major proportion of the direct dyes. Three typical structures are frequently seen in direct dye structures: polyazo (Figure 1-1a-b), copper phthalocyanine (Figure 1-1c) and oxazine (Figure

1-1d).7-8

Figure 1-1. Examples of direct dyes. (a) C.I. Direct Yellow 86; (b) C.I. Direct Blue 71. (c) C.I. Direct Blue 86; (d). Triphenodioxazine direct dyes. (Sirius Light Violet FRL).

(a)

(b)

Acid dyes are applied to protein fibers, such as nylon, wool, and silk in acidic aqueous solutions (PH 2.0-6.0), with typical structures include azo, anthraquinone, and

triphenylmethane, and less common structures are based on pyrazolone, azine, nitrodiphenylamine and phthalocyanine. Typical structures of acid dyes are listed in Figure

1-2.7

(a) (b)

Figure 1-2. Examples of acid dyes. (a) C.I. Acid Red 114; (b) C.I. Acid Green 16; (c) C.I. Acid Red 186 (d) Acid Blue 25.

Disperse dyes are water-insoluble that are typically applied to synthetic fibers including polyester, acrylic fiber, polytrimethyl fibers, and polyacitic acid (PLA) under high pressure or temperature. Figure 1-3 demonstrates the general structures of disperse dyes, which are small, planar and with a non-ionic chromophore, attached with polar functional groups like

hydroyalkyl, -NO2, and –CN. The interactions between dyes and the polyesters are Van der

Waals Forces and hydrogen bonding.9

(a) (b)

Figure 1-3 Typical structures of disperse dyes. (a) C.I. Disperse Yellow 58; (b) C.I. Disperse Yellow 77; (c) C.I. Disperse Blue 27; (d) C.I. Disperse Yellow 3.

Other dye classes

Fiber dyes also includes basic dyes, azoic dyes, metallized dyes, sulfur dyes and vat dyes.

Their application and characteristics are summarized in Table 1-2.9 Basic dyes are applied under acidic conditions, but they process positively charged functional groups which form ionic bonds with negatively charged functional groups in substrates. Reactive dyes form covalent bonds with substrates. Sulfur dyes and vat dyes require reducing agents to make them water soluble, then undergo oxidation within the fiber in insoluble form.10

Table 1-2. Summary of dye classes and the typical fiber types to which they are applied

Dye Class Description Typical fiber substrates Anionic compounds, water soluble, Acid form ionic bond between dye Wool, silk, polyamide, polypropylene molecule and polymer Applied in weakly acid dye bath, Polyacrylonitrile, acrylic, occasionally Basic negatively charged fiber polyester and polypropylene Water soluble, anionic compounds, Direct applied directly to fiber at aqueous Cotton, rayon and other cellulosics medium Water soluble form covalent bond Reactive Cotton, wool and other celluolosics with functional group of fiber Water insoluble; aqueous dispersion; Disperse form hydrogen bonds and van der Polyester, acetate Waals forces Compounds containing sulfur or sodium sulfide; reduced using sodium Sulfur sulfide or sodium hydrosulfite, dye cellulosics enters fiber and is oxidized to original from Vat Water-insoluble; cellulosics

1.2.1.2. Classification by structure

Azo dyes

Synthetic azo dyes are among the largest and most versatile classes of synthetic dyes covering wide spectrum. Dyes possess one or more azo groups and are described as mono, bis or dis, tris, and terakis azo dyes. Azo dyes can be conventionally categorized into subgroups to indicate methods of synthesis using letters A, D, E, Z, and M.11

A: Primary aromatic amine (normal diazo component)

D: Primary aromatic diamine (terazo component)

E: Coupling component capable of reaction with one diazonium salt ion.

Z: Coupling component capable of reacting with more than one diazo component.

M: Middle component capable of coupling and diazotization.

Examples of azo dyes are shown in Table 1-4.

(a)

(b)

(c)

Figure 1-4. Example of Azo Dyes. (a) C.I. Disperse Red 1 (Monoazo); (b) C.I. Acid Black 1. (Diazo); (c) C.I. Direct Red 81.

The azo dyes have attracted much attention in high tech areas such as thermal transfer printing, liquid crystal displays, and solar energy conversion.12 Some azo dyes undergo tautomerism depending on the properties and locations of substituents. As shown in Figure

1-5, 4-phenylazo-1-naphthols exist in solution as hydroxyl azo or as a hydrazone tautomer.12

Tautomers exhibit varied properties such as hue and photostability.

Figure 1-5. Tautomerism of azo dyes. Azo dyes are capable of producing metal-complex dyes with metal ions. Metal complexes allow the small azo dyes to have a significant increase in molecular size, and better wash fastness and light stability. Metal complexes of azo dyes ligand have extensive application in biological, pharmaceutical and food industries.13

Figure 1-6. Acid Black 52 (2:3 Cr Complex, Mixture of 1:1 Cr Complex and 1:2 Cr Complex, and L represents ligands). 11

Anthraquinone dye

Anthraquinone dyes are the second most important dye class in organic dyes. The basic structure of anthraquinone dyes is shown in Figure 1-7 (a). Anthraquinone dyes provide good light fastness, wash fastness and leveling properties and have a wide application as acid, reactive, and disperse dyes.7

(a) (b) (c)

Figure 1-7. Examples of anthraquinone dyes. (a) Basic structure of anthraquinone dyes; (b) C.I. Disperse Red 15; (c) C.I. Acid Blue 25.

Triarylmethane dye

The third important dyes are triarylmethane, which can produce brilliant green or blue hue that have wide application including ball point pen inks, photoconductors, and even therapeutic agents.14 Triarylmethane dyes have less light fastness than azo dyes and anthraquinone dyes. Photodegradation of triarylmethane dyes have been analyzed by in forensic research, because the degree of photodegradation indicates the age of the entry of inks in suspect documents. Examples of triarylmethane dyes are shown in Figure 1-8.13

(a) (b)

Figure 1-8 Example of typical triarylmethane dyes. (a) Malachite green; (b) Ethyl violet; (c) Methyl violet; (d) C.I. Acid Green 16).

Other dye structures

Other than the structures stated above, other structure including xanthene, acridine, quinonline, methine, thiazole, indigoid, and thiazine dyes have wide applications. Examples of typical structures are shown in Figure 1-9.7

(a) (b)

Figure 1-9. Examples of other dye structures. (a) Xanthene C.I. Acid Orange 11; (b) Acridine C.I. Acid Yellow 118; (c) Anthrone, Vat Orange 9; (d) Thioindiogid; (e) Indigoid Dyes.

A variety of dyes that are commercially available are registered in Colour index (C.I.) published by the Society of Dyers and Colourists (SDC) and American Association of

Textile Chemists and Colourists (AATCC). More than 27000 dyes and pigments under

13,000 C.I. Generic Names (direct, acid, disperse dyes) were recorded, complemented with detailed chemical structure, physical form, principal usage and comments supplied by the manufacturers.15 However, some latest dyes are usually not included. Commercial dyes produced by various manufacturers contain various components depending on batches, including byproducts, intermediates, and inorganic salts. For example, C.I. Acid orange 20 manufactured by a company was found to yield six color bands in gel chromatography in

addition to the expected dye, and C.I. Direct Green 6 was found to have five different λmax.

C.I. Disperse Blue 60 observed at least three components in one commercial product.16

Max A. Weaver Dye library

Max A. Weaver Dye library in North Carolina State University provides a source of more than 100,000 synthetic dyes and dyed fabric samples, complemented with information on light fastness, wash fastness, and chemical structure. Examples of the samples in the dye library are shown in Figure 1-10.

(a)NCSU-MWDL-AZ-B-303A2-C (b) NCSU-MWDL-S-X-26647-138

Figure 1-10. Examples of dyes in NCSU-Max. A. Weaver Dye Library.

1.2.2. Photodegradation of dyes

1.2.2.1. Color and dye chemistry

Colors are generated by absorption and reflection of visible light of illumination sources, such as sunlight, which has properties of both a particle and a wave with a continuous range of frequency. People sense visible light with wavelength of falling in the range of around 380 nm to 780 nm, following sequence of the violet, blue, green, yellow, orange as is illustrated in Figure 1-11.17

Figure 1-11. Electromagnetic spectrum.

Color perceived or the reflected color is complementary to the absorbed light. Light with a wavelength of 550 nm is perceived as red, and 550 nm as green, and 410 nm light as purple.

For example, if a colorant absorbs light in the region of 491-570 nm (green), the reflected light is the complementary light of green, which is red. An empirical list of absorbed and observed color is shown in Table 1-3.18

Table 1-3. Complementary colors

Color Absorbed Wavelength absorbed (nm) Color Perceived

Red 647-700 Green

Orange 585-647 Cyan (green-blue)

Yellow 570-585 Blue

Green 491-570 Red

Blue 424-491 Yellow

Violet 400-424 Greenish-Yellow

Color generation can be explained by molecular electronic transitions that when a photon of light is absorbed by a molecule, an electron is promoted from one energy level to higher energy level. The molecules at different energy level processing different ways of electron orbitals, namely non-bonding (n), pi-bonding (π), sigma bonding (σ), and two orbitals at higher energy orbitals defined as π* and σ* orbitals. The energy requirement for different electronic transitions is as follows, n→π*<π→π*

Figure 1-12 Electron transitions between electron orbitals.

Excited molecules have a tendency to go back to ground state and release energy (∆E) between the two states, and ∆E can be calculated by

hc E , (1)  where h represents Planck’s constant (6.625 ×10-34, Js), c is the speed of light (2.2998 ×10-

17, nm/s), and λ is the associated wavelength (nm). When the energy gap falls in visible light absorption spectrum, the energy is released in form of visible light.19

Figure 1-13. Energy absorption and release by organic compounds.

Dyes containing conjugated structure having alternating single and multiple bonds, which in general tend to yield narrower energy gap between the excited status and the ground status, thereby leading to low-energy, or longer wavelength spectrum absorption.19

1.2.2.2. Photodegradation of dyes

Photodegradation and light stability of dyes have being fascinating people for over hundreds of years.20 The light absorbed, as a form of energy, can cause damaging effect on both fibers and colorants. Degradation of colorants caused by light exposure is so called photodegradation, or photofading, which is one of the most important factors influencing the fading of colorants and natural aging of textiles.21

The most common technique in previous research for measuring degree of photodegradation was by the K/S value of fabrics after degradation.22 The K/S value is roughly proportional to the dye concentration as is defined in Kubelka-Munk’s equation:

ACCKSKSftft/(/)/00 (/) (2)

where A represents the relative dye concentration on fibers, Cft and Cf0 refer to the concentration before and after exposure. The decomposition rate of dyes in fibers can be described by the general equation

L n [ A ] = - a t (3)

Where a stands for a photo-destruction constant, and t represents the UV exposure time (s).

Five main classes according to the rate of loss of dye pictured in Figure 1-14.23

During weathering or photodegradation of dyes, color changes take place toward different directions of light absorption due to photochemical reactions. Color shift toward shorter wavelength is termed as hypsochromic shift (blue shift), and color change toward longer wavelength is termed as bathochromic shift (red shift). It may be accompanied with increase

(hypsochromic) or decrease (hypochromic) in molar absorptivity, which is quantified by molar extinction coefficients ().24

Figure 1-15. Color shift of colorant.

1.2.2.3. Factors affecting light stability

Since color is produced by electron transition upon light absorption, the light stability of dye molecules is closely related to the status of electron transitions associated dye structure. Dyes containing extended conjugation system tend to have higher light stability, because the extended conjugation systems facilitate redistribution of the electron density transmitted through interconnected π-bonds, so the dye structure can be protected from bonding breakage or to help dissipate the absorbed energy. The distribution of electron within the dye molecule can also be achieved by electron-drawing/pushing effect from substituent groups. For example, azo dyes having hydroxyl groups at the ortho- position to azo bond and anthraquinone dyes having a α-carbonyl are observed to have enhanced light stability, due to intermolecular hydrogen between the hydroxyl formed with nitro or hydrogen with the carbonyl group. As is shown in Figure 1-16, the transitions between keto-form and enol-form

of an anthraquinone dyes help to improve light stability due to spread of energy within the molecules without breaking bonds. Other steric factors including rotation of bulky substituents are reported to improve the stability of dyes by rotation to dissipate the absorbed

UV energy.25

Figure 1-16. Tautomerism of anthraquinone compounds.

The overall photofading rate is influenced by dye concentration. Such influence is also called layer effects, which depends upon other physical factors such as the average size of dye particle and their distribution.26 Other than the factors stated above, the light stability of dyes are determined by the environment (humidity, temperature, pollutions, auxiliary agents, etc.)27 28 and the spectral characteristic of UV radiation.29

1.2.2.4. Photofading mechanisms

To preserve colorants and to minimize the damaging effect light can achieve, researchers have never stopped investigating the photofading mechanisms in the past two hundreds of years. As early as 1966, H.C.A van Beek and P.M. Heertjes reviewed four possibilities during photodegradation of dyes: (1) photodecomposition, (2) photoreduction of dyes by H- atom absorption, (3) photoreaction with photo-excited substances, and (4) photo catalytic process.30 These theories were further testified and developed by other researches in the past

decades.31 As the advances of modern analytical technology, people have much better understanding of the chemistry and reactive species involved in photofading of different dye classes.

Photodegradation of azo dyes

In 1992, Ping Yueh Wang and Ing Jin Wang proposed that reductive cleavage of azo bonding was the major degradation pathways of azo dyes present in substrates (i.e. nylon, polyester).32 This theory was supported by other researchers.33 For example, the photo fading process of C.I. Disperse Red 1 was proposed to have a reduction of azo bonding as is shown in Figure 1-17.33 However, Hashizume and N.S. Allen etc. proposed oxidative photo fading were existing as a competing reaction with the photoreduction, and a boundary exists between the two different reactions.26

Figure 1-17. Photodegradation of C.I. Disperse Red 1 on polyester and nylon substrates.

Photodegradation of anthraquinone dyes

Anthraquinone dyes were also found to undergo either oxidation or reduction depending on the nature of the polymer and dyes.22 Compared to azo dyes, anthraquinone dyes have conjugated systems which tend to have higher stability. Singlet oxygen plays an important role in the oxidation route.27,34 Free radicals were detected in the oxidation route with dealkylation is the primary step. Reductive photodegradation, on the other hand, relies on the hydrogen atom abstraction or electron abstraction.23

Photodegradation of triarylmethane dyes

The triarylmethane dyes are regarded as less light-fast than azo dyes and anthraquinone dyes.

The carbon in the middle will isolate the double bond leads to a less stable conjugated system. An example illustrating the mechanisms of photo oxidative degradation in Figure

1-18 showing that singlet oxygen are playing an important role in photo fading.20 Other researchers proposed that photoreduction is involved by absorbing electron or radicals from the environment.21,28

Figure 1-18. Photofading of triphenylmethane dye.

1.2.3. Forensic fiber examination

Current forensic analysis of dyes on textiles involves various analytical techniques to maximize the usefulness of the fiber evidence.35 As is shown in Figure 1-19,36 known and questioned fibers are properly mounted and compared in terms of the physical and optical features, including apparent cross sections, relative diameters, and luster, and fiber types.

Microscopy allows a preliminary screening of colorants once the fiber types are determined.

Conventional fiber testing such as solvent test, or flame test sacrifice fibers more than forensic scientist want to sacrifice. Therefore, for trace fiber examination high sensitivity chemical analysis technique are employed.

Vis-microspectrophotometry (MSP) of Fibers

Figure 1-19. Flow chart of forensic dye examination.

1.2.4. Analytical techniques for trace dye analysis

1.2.4.1. Spectroscopy

Microscopic exam is the first step of forensic fiber examination. Infrared (IR) spectroscopy and UV/visible or fluorescence micro spectrophotometry may show differences and similarities between questioned and known fibers.37 UV-Vis spectroscopy is the most important analytical instrument in modern day laboratory for its simplicity, versatility, speed, accuracy and cost-effectiveness.38 UV-vis spectroscopy measures dye concentrations according to Beer’s Law:

Ac   . (4) where A represents absorption, ε is the absorption coefficient, and c is the dye concentration.

A good example of UV-Vis spectroscopy employed in forensic fiber analysis is UV-Vis micro spectrophotometer, which can be configured to measure the transmittance, absorbance, reflectance, polarization, and fluorescence and luminescence micro spectra of sub- micrometer samples.39 Microspectrophotometer allows to acquire spectra of extremely small sample non-destructively. Raman spectroscopy has also been used for forensic analysis of inks in questioned documents.40 UV-Vis spectroscopy can be used for structural comparisons between two compounds that they are having excitation status.36

1.2.4.2. Chromatography

Chromatography separates components in a mixture by their different affinities to a mobile phase and stationary phase.9 Polarity of the mobile phase are adjusted by the proportion of organic solvents.41 Chromatography techniques applied in forensic analysis of dyes include thin layer chromatography (TLC), high performance liquid chromatography (HPLC),42 gas

chromatography (GC), capillary electrophoresis (CE),43 paper chromatography and other techniques.44

TLC is the simplest technique that is used for rapid determination of homogeneity of extracted dye. The extraction and TLC analysis of acid, disperse, and basic dyes from nylon, polyester and polyacrylonitrile fibers have been successfully developed. At standardized

TLC conditions, dye compounds can be identified via Rf value, which is measured by the ratio of travel distance of spot to that of the solvent. Samples are detected visually if they are colored or under UV light.45 Alternatively, spots are visualized by iodine, which has a high affinity for both unsaturated and aromatic compounds. Recently, TLC plates are used as a medium for direct mass spectrometric analysis like MALDI-TOF,46 DESI with minimum sample preparation.47

High performance liquid chromatography (HPLC) and capillary electrophoresis (CE)

HPLC is a quantifiable method that has been successfully applied to separate and identify dyes in the past.48 Analytes are separated in mobile phase when they pass through a column filled with particles under high pressure.49 Sample components travel with mobile phase and are detected by detectors such as UV light, diode array detectors (DAD),50 and mass spectrometers.51 In order to improve separation efficiency, ion pair agents are used in separating multi-sulphonated dyes, and complex azo dyes.52

Capillary electrophoresis (CE) separates ionic species in a capillary filled with buffer solution, ionic species are separated based on their charge to mass ratio. Morgan and 29

coworkers developed CE methods with different solvent conditions have been developed that are suitable for direct, reactive, and vat dye classes. Sample components are detected by detectors such as UV light, diode array detectors (DAD),53 and mass spectrometers.54

1.2.4.3. Mass spectrometry

Mass spectrometry is a powerful analytical technique to identify compounds in a sample by mass-to-charge ratio, further it can be used to elucidate the structure by subsequent mass spectrometry analysis of ions of interest.55 56 The mass spectrometry process includes conversion of compounds to gaseous ions, with or without fragmentation, followed by characterization of compounds according to their mass-to-charge ratio (m/z) and relative abundances.57 Mass spectrometry provides advantages over other techniques because of its high sensitivity and reproducibility as well as the molecular level of information it provides.58,59

Basic components of the mass spectrometer includes an inlet system (i.e. Liquid chromatography, gas chromatography, direct probe), ionization source (i.e. Electrospray

Ionization, Chemical Ionization, Electrospray Ionization,60 Secondary Ions Mass

Spectrometry (SIMS),61 Atmospheric-pressure chemical ionization (APCI),62 Matrix-assisted laser desorption/ionization (MALDI), mass analyzer (i.e. quadrupole, Time-of-flight, Ion-

Trap, and Magnetic Sector) and detectors (Electron Multiplier, Micro-channel Plates

(MCPS)).63 In the following sections, commonly used ionization sources, and mass analyzers that have been successfully applied for analysis of dyes will be discussed.

Ionization Source

Ionization is a key element for mass spectrometric analysis because it provides gas phase ions that are analyzable. A wide range of ionization sources essentially allows for analysis of polar, ionic, non-volatile and thermal labile dyes.64 Mass spectrometry with atmospheric ionization has become a standard method for identification of dyes, metal-complex dyes, polysulphonated dyes and in general for dye mixtures. Selection of ionization source is determined by the nature of analytes. Ionic or very polar dyes can be analyzed with electrospray ionization (ESI), and for dyes and pigments with low to medium polarity the atmospheric chemical ionization (APCI) gives much better results.

Direct ionization techniques, such as secondary ionization, matrix-assisted laser desorption/ionization (MALDI),65 Desorption electrospray ionization (DESI), facilitated direct analysis with minimum sample separation.66 Soft ionization techniques ESI, and

MALDI have been employed for metal complex azo dyes to provide intact molecules for mass analysis.67 However, none of them can be universally applied for all sorts of dyes because of the wide variety of dyes.64 Table 1-4 provides a summary of common ionization methods for dye analysis.

Table 1-4 Ionization techniques applied for synthetic dyes

Dye Class Ionization Method* Unsulfonated azo dyes EI, LD, TS, FAB, ESI, APCI APCI,EI,ESI/IS,FAB,FD,TS,MA Sulfonated azo dyes LDI Other sulphonated Dyes FAB68 Cationic Dyes MALDI Cyanine Dyes ESI, FAB Anthraquinone Dyes ESI Disperse dyes EI, FAB,TS,ESI Indigoid dyes EI, LDI APCI Xanthene dyes FAB,ES, ESI Dispersants and other additives LD,FAB, ESI Benzyl indole dyes EIS Anionic Dyes FAB69, LSI, MALDI, ESI *EI: Electron Impact, Filed Ionization: FI, FAB: Fast Atom Bombardment, LDI: Laser desorption/ionization70, TS: Thermospray, ESI: Electrospray Ionization, APCI: Atmospheric-pressure chemical ionization. fast atom bombardment (FAB), liquid secondary ion mass spectrometry, atmospheric pressure ionization, electrospray, thermospray, and matrix assisted laser desorption/ionization (MALDI).

Electrospray Ionization

Electrospray ionization (ESI) was first developed in Fenn’s laboratories by Whitehouse et al.

ESI utilizes an electric field to send droplets into an atmospheric pressure region towards a counter electrode.36 ESI utilizes an electric field to send droplets into an atmospheric pressure region towards a counter electrode.17 The process from ions in solution to gas phase include three major steps: production of charged droplets; shrinkage of the charged droplets and production of gas phase ions from charged droplets. (Figure 1-20).71

The biggest advantage of ESI is that it transfer aqua phase analyte ions to gas phase ions without breaking the analytes, therefore it has been used to characterize a wide range of molecules, including biomolecules, inorganic and organic compounds and metal complex.72

Figure 1-20. Negative mode electrospray ionization source.

Electrospray ionization has three stages: the first stage is formation of charged droplets, when the penetration of high voltage field into the liquid in the capillary leads to solution polarization. At negative mode, positive ions move away from the capillary tips, but negative ions move toward the capillarity tips when the opposite electrolyte is applied with positive voltage. Solution polarization leads to destabilization of the menisus and a Talyor cone is formed to produce a jet of charged negative ions. Because droplets are negatively charged, jet further breaks up into small charged droplets due to repulsion. At the capillary tip, excess positive charge migrates toward the power, so that an electrolytic cell is formed. The ion transport does not occur through uninterrupted solution, but as charged droplets and later ions in the gas phase. Therefore, a conventional electrochemical oxidation reaction should be occurring at the liquid-metal interface of the capillary to provide negative ions to the solution by converting metal ions to positive ions and electron by converting positive ions from the solution to neutral molecules and positive ions.

In the second stage, droplets shrink as solvent evaporates and reach to a given radius when the increasing repulsion between the charges overcomes the surface tension of the droplet.

The last stage is production of ionized analytes.

The ESI process was explained by two theoretical models: the ion evaporation and the charge residue model. The charge residue model was proposed by Dole in the study of protein molecules, that when very small droplets are formed by droplet evolution, some of these droplets would contain only one analyte molecule.73 Charges at the surface of the droplets will stay on the analyte after successive coulomb fissions. The charge on the droplet can be expressed by Rayleigh equation73:

3 1/2 푄푅푦- = 8휋(휀0훾푟 ) where QRy is the charge on the droplet, 훾 is the surface tension of the solvent, R the radius of the droplet, and 휀0 the electrical permittivity. When droplets shrink at constant charge fission at or near Raleigh limit, a jet of smaller charged progeny droplets will be released.

Ion evaporates model, on the other hand, states that when the droplet radii blow 10 nm, direct ion emission occurs rather than form smaller droplets. Ion are directly emitted from the highly charged droplets by repulsion of charges. Therefore, in this model ions are not required to generate droplets that contain only one analyte, ion evaporation can occur even when the droplet contains more than analytes.74

Secondary ion mass spectrometry (SIMS)

SIMS utilize secondary mass ionization by sputtering the surface of the specimen with a focused primary ion beam (Ions Cs+, Ar+, Ga+), causing desorption of secondary ions that can 34

be analyzed by mass spectrometer. The solid sample should be stable in vacuum, the secondary particles can be analyzed by Quadrupole or time-of-flight mass analyzer, and the ions are detected by an ion detection system (i.e. Photographic plate, electron multiplier).

Secondary ion current yield can be expressed by

I Isp Y  , (5)

where Ip is primary ion current, Y refers to the sputter yield (number of atoms released per primary ion), α represents ionization probability, θ indicates fractional concentration of the species in the analyzed volume, and η is the instrumental response factor.75 The schematic of

SIMS is demonstrated in Figure 1-21.76

Figure 1-21. Schematic of secondary ionization mass spectrometry. SIMS can be classified to dynamic SIMS and static SIMS based on the influence of primary ion. At Static SIMS mode, less than 1% of surface molecules are perturbed with a depth of a few angstroms.77 SIMS allows high sensitivity for the detection of molecules at the

nanometer scale and at attomolar concentration. In contrast, the dynamic SIMS is highly destructive that yield elemental and isotopic information of samples.78

Time-of-flight secondary ion mass spectrometry (TOF-SIMS) is a well-suited platform for surface analysis that is capable of producing surface spectroscopy, high resolution surface images with minimal sample consumption.77 In forensic analysis, TOF-SIMS has been used to analyze questioned documents, finger prints, human hair and other forensic analysis to determine entry age. Sample surface is bombarded with a focused energy primary ion beam, producing desorption of secondary ions. High resolution images are obtained based on mass spectrometry that is produced by rastering the ion beam across the sample surface. In addition, TOF-SIMS provides image for sample viewing with spatial resolution of less than

0.1 um, it has been used for mapping distribution of dyes in human hair.

1.2.4.4.Mass analyzer

Quadrupole mass analyzer

Ions that produced in the ionization sources are further separated or isolated by mass analyzer. One of the most important mass analyzer is the quadrupole mass analyzer (Figure

1-22), that consists of four parallel cylindrical rods that have applied direct current (DC) and radio-frequency (RF) fields.79 Each pair of opposite rods is electrically connected electrically. One set of paired rods has a positive DC and the other two have a negative DC.

RF amplitude is super imposed with the DC potential to control the travelling direction of the ions. When ions enter into the combinational electric field, their trajectories are affected by the magnetic field. The ions begin to oscillate to a degree depending on their mass to charge ration (m/z). Ions with a range m/z ratio reach the detector while other ions out of the m/z 36

range will be neutralized on the rods’ surfaces. Therefore, the quadruple mass analyzer acts as a filter that only allows one ion or a set of individual ions of interest to reach the directors.

Figure 1-22. Isolation of parent ions using quadrupole mass analyzer. The applied AC and DC voltage to the four rods of a quadrupole mass filter can be graphically presented by the Mathieu stability diagram.79 The shaded region shown in Figure

1-23 represents a collection of points in a-q space that corresponds to the stable solutions to the equation of motion in two planes, namely x-y and x-z plane. The stability diagram is powerful because it can easily be used to rationalize the behavior of quadrupole operation.

The trajectory of any ions in terms of each ions’ initial conditions can be described by the following equations

2 2 푎푥 = −푎푦 = 4푧푒푈/푚 푟0

2 2 푞푦 = −푞푦 = 2푧푒푉/푚 휔0 where x and y are the distance along the given coordinate axes, r0 is the distance from the center axis (the z axis) to the surface of any electrode, 휔 is the angular frequency (2휋푓) of the applied ac waveform, U is the magnitude of the applied dc potential. Parameter q is proportional to DC potential U, but the x axes are proportional to RC potential. For example,

the RF-only mode is equivalent to the operating the quadrupole where the Mathieu parameter q is equal to zero.79

Reflectron Time-of-flight mass analyzer

Time-of-flight mass analyzer separate ions by flight times, when ions has dispersive mass to charge ratios are accelerated under high voltage (Uex), to achieve a velocity that is proportional to the square root of the mass to charge ratio of the charged analytes. Analyte further travels through a known distance, and the total time consumed during the traveling is essentially reversely proportional to the initial travel speed of ions, and also with the mass to charge ratio of charged analytes. Ions are detected by their flight times which are 38

proportional to m/z, where t is ion flight time, m is mass, z is charge state of the ion, e is

79 electron charge, and Uex is the extraction voltage.

As is shown in Figure 1-24, ions were focused by ion optics and enter into a vacuum region where high voltage is applied to the back plate of the ion pulser. Ions are accelerated and obtain kinetic energy prior enter into the flight tube. The measured flight time (t) starts when ion is pulsed by ion pulser and end when the ion reaches the detector. The flight time is determined by the kinetic energy of ion (E), the flight distance (d), and mass to charge ratio

(m/z). The following equations were applied to time-of-flight analysis:

Kinetic energy obtained by high voltage potential Uex and

1 퐸 = 푚푣2 = 푈 푞 2 푒푥

Flight time

푑 푡 = 푣 where

푞 = 푧푒

2푞푈 푣 = √ 푒푥 푚

푚 푚 푡 = 푑√ = 푑√ 2푞푈푒푥 2푧푒푈푒푥

Figure 1-24. Time-of-flight analysis of ions of various masses. For clarity and simplicity, this shown in a linear time-of-flight mass spectrometer that does not have an ion mirror. (Ref: TOF-MS Agilent Technical Overview)

TOF measurement does not rely on the arrival times of ions coming from a single pulse applied to the ion pulser, but instead are summations of the signals resulting from many pulses. Each time when the high voltage pulser is applied in the plate of ion pulser, the data acquisition system will record a single spectrum called a transient. The frequency of ion pulser is limited by the mass range. Because once ion pulser is triggered, it is necessary to wait until the last ion of interest arrives at the detector before the next pulse. The total flight time should not exceed the time intervals between two pulses. For example, if an ion with m/z 3200 has a flight time of about 0.1 milliseconds; it means 10000 transients can be recorded within 1 second if there is no delay between transients. For a mass of 800 m/z, the

flight time will be reduced to 0.1 √4 seconds, allowing for 20,000 transients per second, and more transients are recorded to make higher intensity of ions of interest. Therefore, in case of fast separation, less number of transients is recorded, and the number of ions of a specific mass in any given transient is quite small. At the same time, if more ions are detected to improve the sensitivity, the resolving power may be degraded. This is due to small differences of flight times for ions with the same mass, when they have distributions of initial kinetic energy, position and the time of formation of ions prior to acceleration. For orthogonal acceleration TOF instrument, the resolving power or resolution can be reached to

3000-5000 (FWHM).79

Reflectron Time-of-flight mass analyzer was designed to achieve better ion focusing and high resolution, by correcting the distribution of kinetic energy of ions with the same mass.

When two ions with the same m/z ions with different kinetic energy entered the flight tube, they will reach the detector at different times. As is shown in Figure 1-25, ions of higher kinetic energy penetrate farther with longer flight path than the ion with lower kinetic energy, but return at the same speed as they entered. It catches up with the other ion and reaches to the detector at the same time. In this way, ions are better focused and resolving power is improved.

Figure 1-25. Time-of-flight mass spectrometer schematic.

Tandem mass spectrometry

The advantage of mass spectrometry relies on its capability of structure elucidation by tandem mass spectrometry, which uses two or more m/z analyzers in series connected a chamber that can break ions into pieces.80 Ions of interest are selected and subject to fragmentation, usually by collision with neutral gas (e.g. nitrogen, argon) in the process called collision-induced disassociation (CID) or other dissociation ways.81 As is in Figure

1-26, ions of interest are selected followed by collision induced dissociation to yield product ions and neutral fragments. The fragmentation can be expressed as

 mmmpfn, (6)

+ where parent ion 푚푝 collides with the neutral molecule and subsequently breaks apart to

+ form fragment ion mf and neutral fragment (neutral loss) mn . Based on the fragment ions detected by the MS/MS mass spectra, the structure information can be obtained.81

Figure 1-26 Schematic of tandem mass spectrometry. Featured fragments act like puzzles indicate the compounds prior to fragmentation.101

Compounds containing functional groups yield featured fragments under standard conditions to yield featured neutral losses and fragments.102 For example, polysulfonated acids observe

SO3 loss or SO2 loss during negative ionization mass spectrometry; azo dyes subject to loss

82 of N2; dyes containing –COOH observe a loss of –CO2 when SO3H and OSO3H groups are absent.83

Tandem mass spectrometry can be used to identify structural isomers,60 by examining featured fragments as well as their relative abundances are affected by the molecule structure and the CID conditions.84 For example, the two isomers of rhodamine 6G can be differentiated by different fragmentation patterns under the same CID conditions.85

1.3. Instrumentation

Agilent infinity 1200 HPLC coupled with Agilent 6520 QTOF mass spectrometer was used in this research, and the schematic of which is illustrated in Figure 1-27.86 The mass spectrometer consisting of two quadruple analyzers in sequence, and a reflectron time-of- flight mass analyzer. The first quadrupole (Q1) acts as a mass filter allowing ions at certain m/z values or ranges to enter into the collision cell ions, where ions are thermalized with neutral gas molecules (nitrogen gas) and their energy spread are reduced to achieve a higher transmission efficiency. At tandem mass spectrometry mode, ions undergo collision induced dissociation (CID) in the collision cell before they enter into the second quadrupole (Q2), where the resulting fragment ions together with residue parent ions are collisional cooled and focused into a parallel beam that continuously enters the ion modulator of the TOF analyzer.

The mass resolution of this instrument can be reached to sub 5 ppm mass accuracy.87

HPLC Inlet

Nebulizer Capillary Ion mirror

Solvent Spray

Quad Mass Filter Octopole 2 Octopole 1 Detector

Lens 1 and 2 Collision Cell DC Quad Ion Pulser

Rough Pump Turbo Turbo Turbo

Negative ESI (TOP) and QTOF Mass Analyzer Schematic (Bottom)

Figure 1-27. Schematic of Agilent 6520 QTOF mass analyzer in negative mode.

1.4. Reference

1. Kumar, P.; Jaggi, C.; Sharma, V.; Raina, K. K., Advancements of vertically aligned liquid crystal displays. Micron 2016, 81, 34-47.

2. Nicholas, D. J. Forensic fiber and hair analysis. http://www.indy.gov/egov/county/fsa/documents/hair.pdf.

3. Massonnet, G.; Buzzini, P.; Monard, F.; Jochem, G.; Fido, L.; Bell, S.; Stauber, M.;

Coyle, T.; Roux, C.; Hemmings, J.; Leijenhorst, H.; Van Zanten, Z.; Wiggins, K.; Smith, C.;

Chabli, S.; Sauneuf, T.; Rosengarten, A.; Meile, C.; Ketterer, S.; Blumer, A., Raman spectroscopy and microspectrophotometry of reactive dyes on cotton fibres: Analysis and detection limits. Forensic Science International 2012, 222 (1–3), 200-207.

4. Stephen L.Morgan, P. D. Validation of Forensic Characterization and Chemical

Identification of Dyes Extracted from Millimeter-length Fibers; U.S. Department of Justice,

2014.

5. J.Shore, Colorants and auxiliaries: organic chemistry and application properties.

2002.

6. Stefan, A.; Dockery, C.; Baguley, B.; Vann, B.; Nieuwland, A.; Hendrix, J.; Morgan,

S., Microextraction, capillary electrophoresis, and mass spectrometry for forensic analysis of azo and methine basic dyes from acrylic fibers. Anal Bioanal Chem 2009, 394 (8), 2087-

2094.

7. Clark, M., Handbook of Textile and Industrial Dyeing, Volume 1 - Principles,

Processes andTypes of Dyes. Woodhead Publishing: 2011; pp 490-492.

8. Gregary, P., Industrial applications of phthalocyanines. Journal of Porphyrins and

Phthalocyanines 2000, 04 (04), 432-437. 46

9. Stefan, A.; Dockery, C.; Nieuwland, A.; Roberson, S.; Baguley, B.; Hendrix, J.;

Morgan, S., Forensic analysis of anthraquinone, azo, and metal complex acid dyes from nylon fibers by micro-extraction and capillary electrophoresis. Anal Bioanal Chem 2009, 394

(8), 2077-2085.

10. Abrahart, E. N., Dyes and their intermediates. 2nd edition ed.; Hodder Arnold: 1977.

11. El-Shafei, A. A. semi-empirical molecular orbital methods and ab initio calculations in dye chemistry: computational studies towards the design and synthesis of organic pigments. . North Carolina State University, 2002.

12. Cheon, K.-S.; Park, Y. S.; Kazmaier, P. M.; Buncel, E., Studies of azo–hydrazone tautomerism and H-bonding in azo-functionalized dendrimers and model compounds. Dyes and Pigments 2002, 53 (1), 3-14.

13. Hunger, K.; Gregory, P.; Miederer, P.; Berneth, H.; Heid, C.; Mennicke, W., Important

Chemical Chromophores of Dye Classes. In Industrial Dyes, Wiley-VCH Verlag GmbH &

Co. KGaA: 2004; 13-112.

14. Indig, G. L.; Anderson, G. S.; Nichols, M. G.; Bartlett, J. A.; Mellon, W. S.; Sieber, F.,

Effect of molecular structure on the performance of triarylmethane dyes as therapeutic agents for photochemical purging of autologous bone marrow grafts from residual tumor cells.

Journal of Pharmaceutical Sciences 2000, 89 (1), 88-99.

15. Colour Index.

16. Beck, K. R.; Hinks, D.; Crawford, A.; Weisner, N., Liquid Chromatographic and

Mass Spectrometric Analysis of Dyes for Forensic Purposes. AATCC Review 2012, 12 (1),

60-65.

17. Wilm, M., Principles of Electrospray Ionization. Molecular & Cellular Proteomics : 47

MCP 2011, 10 (7), M111.009407.

18. Pavia, D.; Lampman, G.; Kriz, G.; Vyvyan, J., Introduction to spectroscopy. Cengage

Learning: 2008.

19. Dean, J. A., Lange's handbook of chemistry. Materials and Manufacturing Processes

1990, 5 (4), 687-688.

20. Weyermann, C.; Kirsch, D.; Costa Vera, C.; Spengler, B., Evaluation of the

Photodegradation of Crystal Violet upon Light Exposure by Mass Spectrometric and

Spectroscopic Methods. Journal of Forensic Sciences 2009, 54 (2), 339-345.

21. Galvan-Gonzalez, A.; Belfield, K. D.; Stegeman, G. I.; Canva, M.; Marder, S. R.;

Staub, K.; Levina, G.; Twieg, R. J., Photodegradation of selected π-conjugated electro-optic chromophores. Journal of Applied Physics 2003, 94 (1), 756-763.

22. Katsuda, N.; Omura, T.; Takagishi, T., Photodegradation of an anthraquinone type disperse dye on polyester, diacetate, and triacetate fibers, and in solution. Dyes and Pigments

1997, 34 (2), 147-157.

23. Katsuda, N.; Yabushita, S.; Otake, K.; Omura, T.; Takagishi, T., Photodegradation of a disperse dye on polyester fiber and in solution. Dyes and Pigments 1996, 31 (4), 291-300.

24. Nassau, K., The physics and chemistry of color: the fifteen causes of color. The

Physics and Chemistry of Color: The Fifteen Causes of Color, 2nd Edition. 496. ISBN 0-471-

39106-9. Wiley-VCH, 2001, 1.

25. Daria, C.; Han, N.; Marina, B.; Lorenzo, F.; Albert, J. K.; Renè, M. W.; Maarten, R. v.

B., Crystal violet: Study of the photo-fading of an early synthetic dye in aqueous solution and on paper with HPLC-PDA, LC-MS and FORS. Journal of Physics: Conference Series 2010,

231 (1), 012011. 48

26. Allen, N. S., Photofading and light stability of dyed and pigmented polymers.

Polymer Degradation and Stability 1994, 44 (3), 357-374.

27. Himeno, K.; Okada, Y.; Morita, Z., Photofading of monoazo disperse dyes on polyester and polyamide substrates. Dyes and Pigments 2000, 45 (2), 109-123.

28. Allen, N. S.; McKellar, J. F.; Mohajerani, B., Lightfastness and spectroscopic properties of basic triphenylmethane dyes: Effect of the substrate. Dyes and Pigments 1980, 1

(1), 49-57.

29. Blus, K., Photo-stability of Acid Dyes, Derivatives of 1-phenyl-3-methylpyrazol-5- one in Polyamide Fibers. Fibres & Textiles in Eastern Europe 2005, 13 (6), 54.

30. Van Beek, H. C. A.; Heertjes, P. M., Fading by Light of Organic Dyes on Textiles and

Other Materials. Studies in Conservation 1966, 11 (3), 123-132.

31. Petrick, L. M.; Wilson, T. A.; Ronald Fawcett, W., High-Performance Liquid

Chromatography–Ultraviolet–Visible Spectroscopy–Electrospray Ionization Mass

Spectrometry Method for Acrylic and Polyester Forensic Fiber Dye Analysis*. Journal of

Forensic Sciences 2006, 51 (4), 771-779.

32. Wang, P. Y.; Wang, I. J., Photolytic Behavior of Some Azo Pyridone Disperse Dyes on Polyester Substrates. Textile Research Journal 1992, 62 (1), 15-20.

33. Freeman, H. S.; Hsu, W. N., Photolytic Behavior of Some Popular Disperse Dyes on

Polyester and Nylon Substrates. Textile Research Journal 1987, 57 (4), 223-234.

34. Kuramoto, N.; Kitao, T., The contribution of singlet oxygen to the photofading of triphenylmethane and related dyes. Dyes and Pigments 1982, 3 (1), 49-58.

35. Houck, M. M., Professional Issues in Forensic Science. 2015.

36. Blake, S. L. Dyed Fiber Analysis towards the Development of a LC-MS Comparative 49

Finished Fiber

Database for Forensic Purposes. North Carolina State University, Raleigh, 2013.

37. Appalaneni, K.; Heider, E. C.; Moore, A. F. T.; Campiglia, A. D., Single Fiber

Identification with Nondestructive Excitation–Emission Spectral Cluster Analysis. Analytical

Chemistry 2014.

38. Bell, S., Forensic Chemistry. 2006, 212-213

39. Abbott, L. C.; Batchelor, S. N.; Smith, J. R. L.; Moore, J. N., Resonance Raman and

UV–visible spectroscopy of black dyes on textiles. Forensic Science International 2010, 202

(1–3), 54-63.

40. Braz, A.; López-López, M.; García-Ruiz, C., Raman spectroscopy for forensic analysis of inks in questioned documents. Forensic Science International 2013, 232 (1–3),

206-212.

41. M.Huang, R. R., B.G. Fookes, ME. Sigma, Analysis of fiber dyes by liquid chromatography mass spectrometry (LC-MS) with electrospray ionization: Discriminating between dyes with indistinguishable UV-visible absorption spectra. Journal of Forensic

Sciences 2005, 50 (3), 526-534.

42. Rafaëlly, L.; Héron, S.; Nowik, W.; Tchapla, A., Optimisation of ESI-MS detection for the HPLC of anthraquinone dyes. Dyes and Pigments 2008, 77 (1), 191-203.

43. Lee, E. D.; Mück, W.; Henion, J. D.; Covey, T. R., Capillary zone electrophoresis/tandem mass spectrometry for the determination of sulfonated azo dyes.

Biological Mass Spectrometry 1989, 18 (4), 253-257.

44. Goodpaster, J.; Liszewski, E., Forensic analysis of dyed textile fibers. Anal Bioanal

Chem 2009, 394 (8), 2009-2018. 50

45. Titford, M., Comparison of historic Grübler dyes with modern counterparts using thin layer chromatography. Biotechnic & Histochemistry 2007, 82 (4-5), 227-234.

46. Cochran, K. H.; Barry, J. A.; Muddiman, D. C.; Hinks, D., Direct Analysis of Textile

Fabrics and Dyes Using Infrared Matrix-Assisted Laser Desorption Electrospray Ionization

Mass Spectrometry. Analytical Chemistry 2013, 85 (2), 831-836.

47. Zenobi, R.; Knochenmuss, R., Ion formation in MALDI mass spectrometry. Mass

Spectrometry Reviews 1998, 17 (5), 337-366.

48. Anne fraser, C. Liquid Chromatography – Quadrupole Time-of-Flight Mass

Spectrometry Analysis of Disperse Dyes for the Development of a Forensic Dye Database.

North Carolina State University, Raleigh, 2012.

49. Snyder, L. R., Introduction to modern liquid chromatography. Third ed.; 2010.

50. Ilisz, I.; Aranyi, A.; Pataj, Z.; Péter, A., Enantiomeric separation of nonproteinogenic amino acids by high-performance liquid chromatography. Journal of Chromatography A

2012, 1269 (0), 94-121.

51. Oberacher, H.; Schubert, B.; Libiseller, K.; Schweissgut, A., Detection and identification of drugs and toxicants in human body fluids by liquid chromatography–tandem mass spectrometry under data-dependent acquisition control and automated database search.

Analytica Chimica Acta 2013, 770 (0), 121-131.

52. Sun, H.-w.; Wang, F.-c.; Ai, L.-f., Determination of banned 10 azo-dyes in hot chili products by gel permeation chromatography–liquid chromatography–electrospray ionization- tandem mass spectrometry. Journal of Chromatography A 2007, 1164 (1–2), 120-128.

53. Petroviciu, I.; Vanden Berghe, I.; Cretu, I.; Albu, F.; Medvedovici, A., Identification of natural dyes in historical textiles from Romanian collections by LC-DAD and LC-MS 51

(single stage and tandem MS). Journal of Cultural Heritage 2012, 13 (1), 89-97.

54. Hufsky, F.; Scheubert, K.; Böcker, S., Computational mass spectrometry for small- molecule fragmentation. TrAC Trends in Analytical Chemistry 2014, 53 (0), 41-48.

55. Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S., Fourier transform ion cyclotron resonance mass spectrometry: A primer. Mass Spectrometry Reviews 1998, 17 (1), 1-35.

56. Souto, C. Analysis of Early Synthetic Dyes with HPLC-DAD-MS. 2010.

57. Vincenti, M.; Salomone, A.; Gerace, E.; Pirro, V., Application of mass spectrometry to hair analysis for forensic toxicological investigations. Mass Spectrometry Reviews 2013,

32 (4), 312-332.

58. Blake, S.; Walker, S. H.; Muddiman, D.; Hinks, D.; Beck, K., Spectral Accuracy and

Sulfur Counting Capabilities of the LTQ-FT-ICR and the LTQ-Orbitrap XL for Small

Molecule Analysis. J. Am. Soc. Mass Spectrom. 2011, 22 (12), 2269-2275.

59. Petroviciu, I.; Albu, F.; Medvedovici, A., LC/MS and LC/MS/MS based protocol for identification of dyes in historic textiles. Microchemical Journal 2010, 95 (2), 247-254.

60. Sullivan, A. G.; Garner, R.; Gaskell, S. J., Structural analysis of sulfonated monoazo dyestuff intermediates by electrospray tandem mass spectrometry and matrix-assisted laser desorption/ionization post-source decay mass spectrometry. Rapid Communications in Mass

Spectrometry 1998, 12 (17), 1207-1215.

61. Zhou, C.; Li, M.; Garcia, R.; Crawford, A.; Beck, K.; Hinks, D.; Griffis, D. P., Time- of-Flight-Secondary Ion Mass Spectrometry Method Development for High-Sensitivity

Analysis of Acid Dyes in Nylon Fibers. Analytical Chemistry 2012, 84 (22), 10085-10090.

62. Reemtsma, T., Liquid chromatography–mass spectrometry and strategies for trace- level analysis of polar organic pollutants. Journal of Chromatography A 2003, 1000 (1–2), 52

477-501.

63. Monge, M. E.; Harris, G. A.; Dwivedi, P.; Fernández, F. M., Mass Spectrometry:

Recent Advances in Direct Open Air Surface Sampling/Ionization. Chemical Reviews 2013,

113 (4), 2269-2308.

64. Betowski, L. D.; Ballard, J. M., Identification of dyes by thermospray ionization and mass spectrometry/mass spectrometry. Analytical Chemistry 1984, 56 (13), 2604-2607.

65. Knochenmuss, R.; Zenobi, R., MALDI ionization: The role of in-plume processes.

Chemical Reviews 2003, 103 (2), 441-452.

66. Weyermann, C.; Kirsch, D.; Costa-Vera, C.; Spengler, B., Photofading of Ballpoint

Dyes Studied on Paper by LDI and MALDI MS. J. Am. Soc. Mass Spectrom. 2006, 17 (3),

297-306.

67. Bich, C.; Baer, S.; Jecklin, M. C.; Zenobi, R., Probing the Hydrophobic Effect of

Noncovalent Complexes by Mass Spectrometry. J. Am. Soc. Mass Spectrom. 2010, 21 (2),

286-289.

68. J.J. Monaghan, M. B., R.S. Bordoli, R.D. Sedgwick, A.N. Tyler Fast Atom

Bombardment mass spectra of involatile sulphonated and phosphonated azo dyestuffs.

International Journal of Mass Spectrometry and Ion Physics 1983, 46, 447-450.

69. R. Borsdorf, R. H., Fast atom bombardment mass spectra of disulphonated oxonole dyes. Dyes and Pigments 1986, 7 (6), 467-474.

70. Sobott, F.; Wattenberg, A.; Kleinekofort, W.; Pfenninger, A.; Brutschy, B., Laser desorption mass spectrometry on thin liquid jets. Fresenius J. Anal. Chem. 1998, 360 (7-8),

745-749.

71. Olumee, Z.; Callahan, J. H.; Vertes, A., Droplet Dynamics Changes in Electrostatic 53

Sprays of Methanol−Water Mixtures. The Journal of Physical Chemistry A 1998, 102 (46),

9154-9160.

72. Smith, D. L., Mass Spectrometry Applications in Forensic Science. In Encyclopedia of Analytical Chemistry, John Wiley & Sons, Ltd: 2006.

73. Holčapek, M.; Volná, K.; Jandera, P.; Kolářová, L.; Lemr, K.; Exner, M.; Církva, A.,

Effects of ion-pairing reagents on the electrospray signal suppression of sulphonated dyes and intermediates. Journal of Mass Spectrometry 2004, 39 (1), 43-50.

74. Loo, J. A., Electrospray ionization mass spectrometry: a technology for studying noncovalent macromolecular complexes. International Journal of Mass Spectrometry 2000,

200 (1–3), 175-186.

75. Bedrich, W.; Koch, B.; Mai, H.; Seidenkranz, U.; Syhre, H.; Voigtmann, R.,

Distortion of Secondary Ion Extraction Due to Sample Surface Irregularities. In Secondary

Ion Mass Spectrometry SIMS III, Benninghoven, A.; Giber, J.; László, J.; Riedel, M.; Werner,

H. W., Eds. Springer Berlin Heidelberg: 1982,19, 81-87.

76. David W. Mogk. Schematic diagram of the CAMECA IonTOF ToF-SIMS instrument. http://serc.carleton.edu/download/images/8384/evans_trift_system_and_camecas_117276701

7.v2.jpg.

77. Briggs, J. C. V. a. D., TOF-SIMS surface analysis by mass spectrometry. IM

Publications and SurfaceSpectra Limited, 2011.

78. Denman, J. A.; Skinner, W. M.; Kirkbride, K. P.; Kempson, I. M., Organic and inorganic discrimination of ballpoint pen inks by ToF-SIMS and multivariate statistics.

Applied Surface Science 2010, 256 (7), 2155-2163.

79. Chernushevich, I. V.; Loboda, A. V.; Thomson, B. A., An introduction to quadrupole– 54

time-of-flight mass spectrometry. Journal of Mass Spectrometry 2001, 36 (8), 849-865.

80. Melnyk, M. C.; Schey, K. L.; Bartlett, M. G.; Busch, K. L., Charge-remote fragmentation in dialkylaminostyryl dyes. Journal of Mass Spectrometry 1998, 33 (9), 850-

857.

81. Aucella, F.; Lauriola, V.; Vecchione, G.; Tiscia, G. L.; Grandone, E., Liquid chromatography–tandem mass spectrometry method as the golden standard for therapeutic drug monitoring in renal transplant. Journal of Pharmaceutical and Biomedical Analysis

2013, 86 (0), 123-126.

82. Lech, K.; Wilicka, E.; Witowska-Jarosz, J.; Jarosz, M., Early synthetic dyes – a challenge for tandem mass spectrometry. Journal of Mass Spectrometry 2013, 48 (2), 141-

147.

83. Holčapek, M.; Volná, K.; Vaněrková, D., Effects of functional groups on the fragmentation of dyes in electrospray and atmospheric pressure chemical ionization mass spectra. Dyes and Pigments 2007, 75 (1), 156-165.

84. Holčapek, M.; Jirásko, R.; Lísa, M., Basic rules for the interpretation of atmospheric pressure ionization mass spectra of small molecules. Journal of Chromatography A 2010,

1217 (25), 3908-3921.

85. Weisz, A.; Andrzejewski, D.; Fales, H. M.; Mandelbaum, A., Structural assignment of isomeric 2-(2-quinolinyl)-1H-indene-1,3(2H)-dione mono- and disulfonic acids by liquid chromatography electrospray and atmospheric pressure chemical ionization tandem mass spectrometry. Journal of Mass Spectrometry 2001, 36 (9), 1024-1030.

86. Technologies, A. Agilent 6200 Series TOF and 6500 Series Q-TOF LC/MS System

2012. http://www.chem.agilent.com/Library/usermanuals/Public/G3335-90142_TOF_Q- 55

TOF_Concepts.pdf.

87. Fjeldsted, J. Time-of-Flight Mass spectrometry Technical Overview. https://www.agilent.com/cs/library/technicaloverviews/Public/5990-9207EN.pdf.

Chapter 2 Microfluidic System for Automated Dye Molecule Extraction

and Detection and Characterization

The work that has been done in this chapter was funded by the National Institute of

Justice, for a project entitled Microfluidic System for Automated Dye Molecule Extraction and Detection for Forensic Fiber Identification. The overall goal of this study was to develop and validate a universal methodology suitable to extract, separate, and analyze dyes in different classifications (direct, acid and disperse dyes) using a microfluidic extraction apparatus coupled to mass spectrometric analysis.

A extraction solvent consisting of pyridine/H2O 4:3 (v/v) was found to be the most efficient solvent for most of important dye classes, by quantitatively measuring the exaction percentage. Optimization of extraction of direct, acid dyes was further conducted based on statistic methodologies, by fitting a statistic model involving extraction time, temperature, sample size, and volume of extraction solvent. Optimization of extraction was firstly conducted at conventional laboratory scale, followed by extraction at the microfluidic device scale. A universal method for high performance liquid chromatography-photodiode array detection-mass spectrometry analysis (HPLC-DAD-MS) was developed and validated for direct dyes to help establishing dye database.

2.1. Preliminary study on dye extraction at conventional scale

2.1.1. Dyes and chemicals

Representative commercial direct dyes (Figure 2-1) were provided by Classic Dyestuffs Inc.

Sodium sulfate was purchased from Fisher Scientific. Water and methanol, acetone, ethanol and pyridine were purchased from Fisher Scientific (ACS grade). Water was purified using

Millipore equipment (PureLab-ELGA).

Figure 2-1. Structures of direct dyes for preliminary extraction.

Direct Red 81

Direct Yellow 106

Direct Yellow 86

Direct Black 22

Direct Violet 51

Acid Blue 25

2.1.2. Sample preparations

Direct dyeing Samples of cotton fabrics (5.00 g, Testfabrics, #400) were dyed at a liquor-to- goods ratio of 20:1 in a dye bath containing varied concentrations (0.5%, 1%, 2% and 3%) of direct dyes (weight/weight) in a Datacolor AHIBA dyeing machine (Figure 2-2) with the following heating program: ramp rate of 2 °C from room temperature to 90 °C, then the temperature was held for 90 min. Samples were rinsed using tap water, air dried and stored in darkness before use.

Figure 2-2. Data Color Ahiba dyeing machine.

2.1.3. Extraction

The extraction solvent was prepared by adding 40 mL organic solvent in certain volume of

Milli-Q filtered water according to pre-determined ratios. Dyes in dyed fabrics (2 x 2 cm, ~2 mg) were isolated using 1000 µL extracting solvent in vials with lids on in a Pierce Reacti- therm module (Figure 2-4), followed by evaporation under 10 psi nitrogen gas. The dye residual was dissolved again by adding 4 mL deionized water, filtered with 0.22 PVDF filter, and transferred to disposable vials for UV/Vis spectroscopy of Figure 2-3.

Fisher Reacti-Vials N2 Gas Flow Take out 1 mL Deionized Water UV/Vis Spectrometry Glass Vials Fabrics

Extraction Solvent Dilution by 4 times with DI water

Extraction Evaporation Recovery Vis-Spectrometry analysis

Figure 2-3. Preliminary extraction of fabrics for UV-vis spectrometry.

The heating module with an apparatus constructed to deliver nitrogen gas to the vials is shown in Figure 2-4. Polyvinyl chloride (PVC, Fisherband) tubing (0.312 inch internal diameter (I.D.) x 0.625 inch wall thickness) was connected to the nitrogen tank regulator.

Smaller diameter clear PVC tubing was connected to the 0.312 inch I.D x 0.0625 inch wall thickness tubing via screw band metal clamps. At the end of five tubes, metal needles were attached. The needles, which were delivered nitrogen gas to the vials.

Figure 2-4. Pierce Reative-therm heating module and nitrogen gas delivery system.

Extraction solvents were selected based on previous study, and the recipes as well as the proposed extraction conditions for extraction are shown in Table 2-1.

Table 2-1 Preliminary Screening of Extracting Solvent

Extracting solvent (2ml) Time (min) Temperature (°C)

Methanol/H2O 4:3 or 3:2 20 25-50

Acetone/H2O 4:3 or 3:2 20 25-50

Ethanol/H2O 4:3 or 3:2 20 25-50

Pyridine/H2O 4:3 or 3:2 30 25-90

Pyridine/H2O 2:1 15 80

Pyridine/H2O 35:65 15 80

Formic acid/ H2O 88:12 15 80

2.1.4. Measurement

2.1.4.1. UV-Vis spectroscopy

The extraction performance of extraction solution was evaluated in two ways: the first was the analysis of the percentage of dyes that have been extracted from the fabrics, and the second was to measure the color changes of the fabrics upon extraction. Dye extract solution

was scanned by Aglient Cary 300 UV-vis-spectrophotometer (Figure 2-5), by which the maximum absorption wavelengh are obtained, and quantification test can be performbed by using standard calibration curves according tothe Beer’s law shown in the Equation (7), light absorbance (A) is proportional to the dye concentration.

I Acllog. 0  (7) 10 I

where A is the measured absorbance of the dye solution to light; I0 is the intensity of incident light at a given wavelength, I is the transmitted intensity, the ratio of I0 to I is the reflectance.

Absorbance data were averaged every 10 nm. l is the path length through the vials which is

1 cm, and c is the the concentration of the solution. ε is a constant known as molar abosorptivity coefficient, which is a measure of how strong a colored species absorbs light at a given wavelength that should be determined before testing. The scan range of the spectrophotometer was set from 360 nm to 660 nm.

Figure 2-5. Agilent Cary 300 UV/Vis Spectrophotometer.

Dyeing uptake percentage was considered for quantitative measurement of direct dye extraction. It is defined as the ratio of the amount of actual dye present on the fabric to the total amount of dye during dyeing procedures, as is shown in Equation(8).

Cafter Puptake (1)100% (8) Cbefore

Where the Cafter can be measured using UV/Vis spectrophotometer. The dyeing uptake rate gives a theoretical total weight of dyes on the fabrics Mdye:

MMowfPdyefabricuptake %%. (9)

The extraction percentage is defined as the porportion of dye extracted from the fabric before extraction, as is defined in Equation (9)

VC Extraction%100%ExtractExtract (10) M dye

where Vextract is the volume of extraction solvent, and CExtract is the concentration of dye extract solution, and Mdye is the total weight of dyes present on the fibers after dyeing procedures. The recovery percentage obtained from different extraction solution will be compared.

2.1.4.2. Color differences of fabrics

The color of the fabrics can be quantitatively measured by means of the chromatic coordinates of the subject: L*, c*, h°, which correspond to lightness, chroma, and hue angle defined by CIE color model 13. The ΔL, ΔC and Δh are the differences in lightness chroma and hue angle of test samples with respect to the references (fabrics before extraction).13

Their overall color differences (ΔEcmc) were measured by Datacolor Check Pro (Datacolor,

USA) with the settings below: D65 standard illuminant, 10° observer, UV included, and specular component included. Samples were measured at three different locations and the results were averaged for comparisons using software Datacolor I Control. A snapshot of color measurement of three polyester samples containing C.I. Disperse Red 1 is shown in

Figure 2-6. The top table shows the detailed color differences of the test samples with references, and the bottom (left) shows the color locations defined in the color space, where the yellow, red, blue and green area represents the changes of hues of test samples compared with the standard; the bottom (right) compares the overlapped spectra curves of reference and samples.

Figure 2-6. Color differences of C.I. Disperse Red 1 on polyester before and after photodegradation.

2.1.5. Results and discussions

Preliminary selection of the solvent found that pyridine/water 4:3 extracted more efficiently than other solvents for disperse dyes and acid dyes at the same temperature and volume. At 67

comparatively low temperatures (at around 50 °C), most of direct dyes were extracted within10 min, but the acid dyes need temperature at around 80 °C for around 15 min, and the disperse dyes can be extracted when the temperature were set at around 100 °C for around 20 min. However, the acetate fibers containing disperse dyes were partially dissolved which may interfere the dye analysis in subsequent analysis.

Color differences between the fabrics before and after preliminary extraction of five typical direct dyes are shown in Figure 2-7. At room temperature, the pyridine/water 4:3 has significantly higher extraction yield than other extraction solvents, followed by the acetone/H2O, methanol/H2O, and ethanol/H2O. Therefore, the system pyridine/water 4:3 was treated as the optimum extraction solvent for the future research.

Extraction Performance of Solvent Systems 40

30 Direct Red 81 25 Direct Yellow 44 20 Direct Voilet 51 15 Direct Black 22 10 Direct Yellow 86

Extraction Percentage (%) Percentage Extraction 5

0 Methanol/H2O Acetone/H2O Ethanol/H2O Pyridine/H2O

Figure 2-7. Extraction percentage with extraction solvents (organic solvent: H2O 4:3).

The extraction solvent consisting of pyridine/H2O 4:3 (v/v) was selected in the preliminary screening study, by quantitatively measuring extraction percentage using UV-Vis spectroscopy. The extraction performance of pyridine/H2O was affected by the nature of dyes, extraction time, temperature and their affinity to the cellulosic fabrics, but less affected by the dye loading. Extraction conditions should be further optimized.

2.2. Macro-level Optimization of Direct Dye Extraction

Macro-level extraction in this study refers to extraction of threads samples (C.I. Direct

Yellow 106) that were conducted at conventional lab scale, in regular glass vials with a capacity of 200 µL. In contrast, microfluidic level extraction refers to extraction that was performed in a microfluidic cavity with a volume of less than 10 µL. Macro-level extraction for dye classes was preliminary optimized to obtain reasonable ranges for microfluidic extraction. In this chapter, statistic methodologies were employed in design of experiments.

2.2.1. Full factorial design

A full factorial design of experiment (DOE) allows for systematic examination significance of multiple extraction variables and interactions between variables. Extraction temperature, extraction time and the volume of extraction solvent were designated as independent variables, and the extraction percentage determined by UV-Vis spectroscopy was designated as the response. Each variable has two levels, a maximum (+1) and minimum (-1), to define its practical range: temperature ranging from 25 °C to 80 °C, time ranging from 2 min to 11 min, and the volume of extraction solvent ranging from 250 µL to 800 µL. The coded and

non-coded values of the investigated variables are listed in Table 2-2. Here center points (0) are included to account for middle values.

Table 2-2 Coded and uncoded values of independent variables in extraction experiment

Symbol Coded Level Variable Uncoded Coded -1 0 1 Extraction temperature (°C) X1 x1 25 52.5 80 Extraction time (min) X2 x2 2 11 20 Volume of extraction solvent ( µL) X3 x3 250 525 800

The formula of full factorial design of direct dye extraction is listed in Table 2-3.

Table 2-3 Design of Screening experiment for direct dye extraction

Temperature/ °C Time (min) Volume (µL) Extraction Percentage 1 25 20 250 3.876 2 25 2 250 9.342 3 25 20 800 6.373 4 25 2 800 5.833 5 80 20 800 8.833 6 80 2 250 5.147 7 80 20 250 9.089 8 80 2 800 9.377

To estimate total time that are needed for evaporation, a statistic model was built based on a full factorial design, evaluating the impact of temperatures, nitrogen gas pressure, and volume of extraction solvent on evaporation time. Practical ranges for each factor are shown in Table 2-4.

Table 2-4 Design of experiment of solvent evaporation

Coded Level Variable -1 0 1 Extraction temperature (°C) 60 90 120 Nitrogen gas pressure (Psi) 5 10 15 Volume of extraction solvent ( µL) 20 60 100

2.2.2. Statistical analysis

All of the three factors: extraction time, temperature, and solvent volume positively affected the extraction percentage (Figure 2-8), and their interactions presented as significant effect on the extraction percentage. Extraction percentage increased in both cases when extraction time extended and temperature increased. Interestingly, at higher extraction temperature, the extraction percentage increased faster than that at low concentration. It was due to the interactions between temperature and extraction time that significantly affect extraction performance. The effect of interactions of two variables can be explained by how variables affect the responses is depending on the setting of the others. For instance, at room temperature (25 °C) it took around 9 min (from 8 min to 17 min) to increase by 10% (from

40% to 50%), but at a temperature of 80 °C, it took only 3 min to increase from 70% to 80%, it is owing to higher migrating rates of dye molecules from the fabrics to extraction solvent at higher temperatures.

Extraction Percentage Design-Expert® Software 20.00

Factor Coding: Actual 80 Extraction Percentage 17.00 90.89 70 14.00 9.34 60 X1 = A: Temp 11.00

X2 = B: Time B : T im e 50 8.00 Actual Factor

C: Volume = 525.00 5.00 40

2.00 25.00 36.00 47.00 58.00 69.00 80.00

A: Temp

Figure 2-8. Interactions between extraction time and temperature on extraction performance.

Interactions between three variables were further examined by 3D surface plot (Figure 2-9), suggesting that the extraction percentage increased with temperature and extraction time. At around 80 °C the extraction percentage reached to the maximum of around 85% within 20 min (Figure 2-9).

Design-Expert® Software Factor Coding: Actual Extraction Percentage 90.89 100

75 9.34 50 X1 = A: Temp X2 = B: Time 25

Actual Factor 0 C: Volume = 200.00

Extraction20.00 Percentage 80.00 17.00 69.00 14.00 58.00 11.00 8.00 47.00 B: Time 5.00 36.00 A: Temp 2.00 25.00

Figure 2-9. Statistical model of direct dye extraction percentage vs. time at room temperature (25 °c) (a) Extraction percentage with 200 µl extraction solvent; (b) Extraction percentage with 800 µL extraction solvent.

2.2.3. Optimization of extraction

The optimized conditions were obtained using Cube analysis (Figure 2-10), showing how extraction percentages are maximized to 90.89% when the variables were set as: 80 °C, 20 min and 250 µL.

Cube Design-Expert® Software Extraction Percentage Factor Coding: Actual Extraction Percentage 63.73 88.33 X1 = A: Temp X2 = B: Time X3 = C: Volume B+: 20.00 38.76 90.89

58.83 51.47 C+: 800.00

B : T im e C: Volume

B-: 2.00 9.34 51.47 C-: 250.00 A-: 25.00 A: Temp A+: 80.00

Figure 2-10. Cube analysis of the direct dye extraction. 2.2.4. Modeling of solvent evaporation

The minimum evaporation time was predicted by the profiler, design of experiments by

JMP.11 Pro. are listed in Table 2-5. In this design two blocks were designed to evaluate the variance of experiment operation during different days.

Table 2-5 DOE of evaporation and results

Pattern Block Temperature Extracting Volume Nitrogen Time (℃) (흁풍) (Psi) (second) +−+ 1 120 20 15 30.64 000 1 90 60 10 144.74 −−− 1 60 20 5 162.64 ++− 1 120 100 5 145.82 −++ 1 60 100 15 264.34 000 2 90 60 10 79.66 −−+ 2 60 20 15 76.24 +++ 2 120 100 15 60.94 −+− 2 60 100 5 213.69 +−− 2 120 20 5 46.22

Analysis of variance (ANOVA) shows that at the 5% significance level the extraction temperature and volume having p-value less than 0.05, suggesting they are significant factors 74

affecting evaporation time (Figure 2-11). It is reasonable that higher temperature significantly accelerated the evaporation rate.

Figure 2-11. Statistic model for solvent evaporation by JMP.

The purpose of this modeling is to predict an estimated time for evaporation at certain conditions. The function of profiler displays profile traces for each variable in Figure 2-12 demonstrates at the optimized condition (100 ℃, extracting solvent volume 20 µL, 5 psi

Nitrogen gas flow rate predicted the total time for complete evaporation should be around

118.1 ± 43.9 seconds.

Figure 2-12. Prediction profiler of evaporation time.

In the present research, full factorial design was employed to discuss the extraction time, temperature, and extraction volume were proved to be important factors for direct dye extraction at the marcro-level. The extraction process of direct dye was optimized to increase its extraction efficiency within less than 10 min. The optimal extraction condition of direct dyes obtained by the experiment and factorial screening analysis were set as follows: extraction temperature as 80 ˚C min, extraction solvent is 192.6 μL, and extraction time as 10 min. Plus, the evaporation model was built to predict total time that is required for certain extraction conditions. The present research provides theoretical basis for further investigation on the microfluidic level extraction of direct dyes.

2.3. Optimization of C.I. Acid Blue 25 Extraction from Nylon Fibers by HPLC-DAD-MS

In this section, Acid Blue 25 was used as a model acid dye to optimize the acid dye extraction at the macro-level. Different than the direct dye extraction study, the study on acid dye extraction was investigated based on a surface response methodology (SRM). Surface response methodology explores factors including extraction temperature, time, and dye loading were independent variables. The concentration of C.I. Acid Blue 25 was measured by the peak areas of chromatograms during analysis HPLC-DAD-MS via photodiode array detection and extracted ion chromatograms of the expected protonated ions in mass spectrometry.

2.3.1. Acid dye extraction

Solvents that were applied during the extraction and LC-MS analysis are listed below:

Pyridine (Fisher, ACS grade), methanol (Burdick&Jackson, LC/MS grade), acetonitrile

(Burdick&Jackson, LC/MS grade), ammonium formate (Sigma-Aldrich), formic acid (~98%,

Sigma-Aldrich). Water was purified using Millipore equipment(PureLab-ELGA).

Extraction solvent was prepared by adding 40 mL pyridine in 30 mL Milli-Q filtered water.

HPLC buffer (Milli-Q H2O containing 3% acetonitrile) was used to recover dye extract for

HPLC analsis, and was obtained by dissolving 3 mL of acetonitrile in 97 mL of Milli-Q water. The dye standard solution were made by commercial dyes of 0.03 mg/mL in water to avoid saturation in the electrospray ionization (ESI) mass specrometric analysis.

Samples of nylon threads (3~10 mm in length, and ~0.2 mm in diameter) (unless stated otherwise) were submerged in 200 µL extraction solvent (pyridine/water 4:3) in a sealed vial 77

with cone bottom and heated to 80 °C with lids in a Pierce Reacti-Therm Heating Module until the threads were discolored within around 20 min. The extraction solvent was evaporated under 10 psi nitrogen gas flow until the solvent is completely evaporated. Dye residue was further reconstructed with 200 µL HPLC buffer (3% acetonitrile in Milli-Q

H2O). The dye solution was filtered by Millex-GV 13 mm, 0.22 µm PVDF filters (Milli-Q water) (Figure 2-13).

Figure 2-13 Conventional extraction procedures of fibers for LC/MS analysis.

2.3.2. Central composite design (CCD) of acid dye extraction

Central composite design can efficiently estimate first and second-order terms in a quadratic model. Extraction conditions including extraction time, temperature, volume of extraction solvent, and dye loading were selected as independent variables.

Table 2-6 Coded and uncoded values of independent variables in central composited design for acid dye extraction

Symbol Coded Level Variable Uncoded Coded -1 0 1 Extraction time (min) X1 x1 30 165 300 Extraction temperature (°C) X2 x2 60 90 120 Length of thread (µm) X3 x3 200 600 1000 Dye loading (owf%) X4 X4 0.5 1.75 3

Two responses peak areas of DAD chromatograms (660 nm) and extracted ion chromatograms were used as responses in the statistic models (Table 2-7). Samples collected from run #22,#23 and #26 were not successfully collected due to sample loss during manual extraction. However, the rest data are adequate for a statistical model.

Table 2-7 Central composite design (CCD) of acid dye extraction at conventional lab scale

Thread DAD EIC Sample# Time Temp. Dye loading (owf %) length PKA PKA 1 30 60 200 0.5 3.68 409994.9 2 300 60 200 0.5 13.53 775841.6 3 30 120 200 0.5 28.96 2027189 4 300 120 200 0.5 30.42 16896932 5 30 60 1000 0.5 10.65 829448.5 6 300 60 1000 0.5 7.4 501005.6 7 30 120 1000 0.5 30.87 1970098 8 300 120 1000 0.5 224.85 10728393 9 30 60 200 3 5.71 38806.54 10 300 60 200 3 5.29 359924.8 11 30 120 200 3 29.3 1639611 12 300 120 200 3 63.44 2922195 13 30 60 1000 3 26.9 1546808 14 300 60 1000 3 0.74 67568.94 15 30 120 1000 3 55.48 16816140 16 300 120 1000 3 63.44 2922195 17 30 90 600 1.75 26.58 1740416 18 300 90 600 1.75 198.45 9101939 19 165 60 600 1.75 52.03 25266764 20 165 120 600 1.75 225.29 10644521 21 165 90 200 1.75 46.18 2988882 22 165 90 1000 1.75 N/A N/A 23 165 90 600 0.5 N/A N/A 24 165 90 600 3 196.78 8368871 25 165 90 600 1.75 52.03 25266764 26 165 90 600 1.75 N/A N/A

2.3.3. Statistical analysis

2.3.3.1. Integrated peak areas

Among the statistic model provided by the software, transformation of natural log of the response was found to fit the extraction with a lack of fit P-value of 0.0031 (Table 2-8).

Similar to direct dye extraction, acid dye extraction temperature was the most important factor (p-value = 0.004), followed by interactions between time and temperature. The dye loading was found to be not significant, therefore it was excluded in the statistic estimates.

Table 2-8 ANOVA analysis of acid dye extraction with dad peak area as the response

Sum of Squares Df Mean Square F Value p-value (prob>F) Model 291.64 9 32.40 4.17 0.0102 A-Time 18.76 1 18.76 2.41 0.1444 B-Temperature 116.12 1 116.12 14.94 0.0020 C-Thread Length 10.65 1 10.65 1.37 0.2628 AB 15.68 1 15.68 2.02 0.1791 AC 0.033 1 0.033 0.004 0.9488 BC 6.18 1 6.18 0.80 0.3888 A^2 6.47 1 6.47 0.83 0.3784 B^2 0.11 1 0.11 0.014 0.9068 C^2 29.65 1 29.65 3.81 0.0727 Residual 101.07 13 7.77 Cor Total 392.71 22

The model F-Value of 4.16 indicates the model is significant. There is only 1.02% chance that an F-value this large could occur due to noise. Contour plots of DAD-peak areas reflect how independent variables were affecting the response, as well as the interactions between independent variables. The concentrations of dye solution are proportional to the peak areas of the chromatograms, therefore total areas of the DAD-peak areas reflect the extraction performance under the conditions. Higher DAD-Peak corresponds to a higher extraction performance under certain conditions. Based on the ANOVA analysis, the peak areas increased with the increase of temperature (Figure 2-14 (b)). At fixed temperature, extraction time of 160 seconds at a temperature of 85 °C the concentration of dye extract reached to the maximum (Figure 2-14(bottom)).

DAD-PKA Design-Expert® Software 120.00 Factor Coding: Actual 114.00 250 Original Scale 108.00 DAD-PKA 200 Design Points 102.00 225.29 96.00 150

0.74 90.00 100 84.00

X1 = A: Time 78.00 B: Temperature (°C) X2 = B: Temperature 72.00 50 66.00 Actual Factors C: Thread Length = 600.00 60.00 D: dye concentration = 1.75 30.00 84.00 138.00 192.00 246.00 300.00

A: Time (Seconds)

DAD-PKA Design-Expert® Software 1000.00 Factor Coding: Actual 920.00 Original Scale DAD-PKA 840.00 Design Points 760.00 225.29 680.00 150 200 250

0.74 600.00 100 50 520.00

440.00 X1 = B: Temperature C: Thread Length X2 = C: Thread Length 360.00

280.00 Actual Factors A: Time = 165.00 200.00 D: dye concentration = 1.75 60.00 66.00 72.00 78.00 84.00 90.00 96.00 102.00 108.00 114.00 120.00 B: Temperature (°C)

DAD-PKA 1000.00 Design-Expert® Software 40 60 40 Factor Coding: Actual 920.00 Original Scale DAD-PKA 840.00 Design Points 760.00

225.29 680.00

0.74 600.00 120 520.00 100 440.00 X1 = A: Time C: Thread Length 80 360.00 X2 = C: Thread Length 60 280.00 Actual Factors 20 40 B: Temperature = 90.00 200.00 D: dye concentration = 1.75 30.00 84.00 138.00 192.00 246.00 300.00 A: Time (Seconds)

Figure 2-14. Contour plots of independent variables on the peak areas of HPLC-DAD peaks at 660 nm.

Three dimensional (3D) plots of the model were used for graphical interpretation of interactions between variables. When extraction temperature was set at the low level (60 °C), extraction efficiency increased slightly as extraction time increased; at the high level

(120 °C), the extraction percentage reached to the maximum at around 160 seconds. This behavior was attributed to evaporation of pyridine/H2O solution at higher temperature, leading to less contact between dyed fibers and the solvent.

Design-Expert® Software Factor Coding: Actual Original Scale DAD-PKA 1146.05 Design points above predicted955.526 value Design points below predicted765.003 value 225.29 574.481 383.958 0.74 193.436 2.91335

X1 = A: Time DAD-PKA X2 = B: Temperature 120.00 114.00 108.00 300.00 Actual Factors 102.00 96.00 246.00 C: Thread Length = 600.00 90.00 192.00 84.00 D: dye concentration = 1.75 138.00 B: Temperature (°C)78.00 72.00 84.00 66.00 60.00 30.00 A: Time (Seconds)

Figure 2-15. Estimated response surfaces with related contours by plotting DAD-Peak areas versus time and temperature.

Statistical analysis of Peak areas of Extracted ion chromatograms (EIC)

ANOVA statisitic analysis of the estimated response surface of peak areas of EIC shows that dye loading in the thread was not a significant factor in constructing a statistic model, because it did not give a p-value less than 0.05. The analysis of variance showing in Table

2-9 suggests extraction temperature is the most important factor, followed by the interactions between temperature and time.

Table 2-9 ANOVA Analysis of acid dye extraction by peak areas of extracted ion chromatograms

Source Sum of Squares df Mean Square F Value p-value (prob>F) Model 51.56 10 5.16 3.69 0.0179 A-Time 0.69 1 0.69 0.50 0.4950 B-Temperature 21.64 1 21.64 15.50 0.0020 C-Thread Length 1.09 1 1.09 0.78 0.3952 AB 0.73 1 0.73 0.52 0.4840 AC 5.35 1 5.35 3.83 0.0739 BC 0.012 1 0.012 8.468E-003 0.9282 A^2 3.61 1 3.61 2.59 0.1338 B^2 0.040 1 0.040 0.029 0.8680 C^2 3.20 1 3.20 2.29 0.1559 Residual 16.75 12 1.40 Cor Total 68.31 22

Similar to surface response of DAD chromatograms, the peak areas of EIC of C.I. Acid blue

25 demonstrates the interactions between extraction temperature and thread length are significant factors. Interactions between extraction temperature and thread length was found to be significant. This can be explained by that the extraction solvent was boiling to circulate around the glass vials at temperature higher than 100 °C, which bring less contact between the dyed threads and extraction solvents. Contour plots suggest at around 160-180 seconds the peak area of extracted ion chromatograms reached to the maximum.

MS-PKA Design-Expert® Software 120.00 4E+007 Factor Coding: Actual 114.00 Original Scale 3E+007 MS-PKA 108.00 Design Points 102.00 2E+007 2.52668E+007 96.00

38806.5 90.00

84.00 1E+007

X1 = A: Time 78.00 B: Tem perature (°C) X2 = B: Temperature 72.00

66.00 Actual Factors C: Thread Length = 600.00 60.00 D: dye concentration = 1.75 30.00 84.00 138.00 192.00 246.00 300.00 A: Time (Seconds)

MS-PKA 1000.00 Design-Expert® Software 2E+006 Factor Coding: Actual 920.00 Original Scale 4E+006 MS-PKA 840.00 Design Points 760.00

2.52668E+007 680.00

600.00 38806.5 1.4E+007 520.00 1.2E+007 440.00 1E+007 X1 = A: Time C: Thread Length 4E+006 X2 = C: Thread Length 360.00 8E+006 2E+006 6E+006 280.00 Actual Factors B: Temperature = 90.00 200.00 D: dye concentration = 1.75 30.00 84.00 138.00 192.00 246.00 300.00 A: Time (Seconds)

MS-PKA Design-Expert® Software 1000.00 Factor Coding: Actual 920.00 Original Scale MS-PKA 840.00 Design Points 760.00

2.52668E+007 680.00 2E+007 3E+007 4E+007 600.00 38806.5 1E+007 520.00

440.00 X1 = B: Temperature C: Thread Length X2 = C: Thread Length 360.00

280.00 Actual Factors A: Time = 165.00 200.00 D: dye concentration = 1.75 60.00 66.00 72.00 78.00 84.00 90.00 96.00 102.00 108.00 114.00 120.00 B: Temperature (°C)

Figure 2-16. Contour plots showing the effects of interactions between variables on peak area of extracted ion chromatograms.

2.3.4. Summary

In this study, central composite design was successfully used to optimize the extraction conditions of C.I. Acid blue 25 during conventional extraction procedure. The independent variables, extraction temperature, extraction time and their interactions had significant impact on the yield of extraction. ANOVA, contour plots and surface plots were applied to predict an optimal working condition to evaluate the performance of extraction procedure for C.I.

Acid Blue 25: Extraction time 160 seconds, temperature 100 ˚C, 200 µL pyridine/H2O as the extraction solvent.

2.4. Automated Microfluidic Extraction –HPLC-DAD-MS Analysis of Dyes on Trace

Fibers

In this section, trace amount textile fibers (i.e. millimeter-length fibers or single fiber) were analyzed on a microfluidic device. Dye extraction in microfluidic device saves multiple stages of sample preparation that is required by manual extraction: dye isolation, solvent evaporation, and reconstruction, and filtration were performed in an integrated platform.

2.4.1. Sample collection

Textiles with seven different dyes as well as corresponding standard references were kindly provided by M. Dohmen, BASF, and Huntsman Co. and Classic Dyestuff (Table 2-10).

Table 2-10 Information of anionic dyes

C.I. *Name Chemical Classes Manufacturer Trade Name Chemical Formula

Acid Blue 25 Anthraquinone M.Dohmen Dorasyn Blue AG C20H14N2O5S

Acid Yellow 49 Monoazo Chemical Telon Yellow FGL 200 C16 H13 N5 O3 S Cl2

Acid Yellow 151 Metalized Monoazo M.Dohmen Dorolan Yellow RTU C32 H28 Co N8 O10 S2 Acid Orange 67 Monoazo Huntsman Erionyl Yellow A-R-01 C26 H22 N4 O8 S2 Acid Red 114 Diazo Ciba Geigy Erionyl Red RS C37H30N4O10S3

Direct Red 81 Disazo Ciba-Geigy Pergasol Red 2BA C29H19N5Na2O8S2

Direct Yellow 106 Disazo Huntsman Solophenyl Yellow ARL C48H18N8Na6O18S6 *C.I. Number: Color Index Number defined by Colour Index International.

2.4.2. Design and fabrication of microfluidic apparatus -1st Generation

The first generation of microfluidic dye extraction system is shown in Figure 2-17. It is an automatic system that only requires a lab examiner to place the fiber sample in the extraction chamber, close the lid, slide it into the system and set the extraction parameters on the user interface. The extraction process is controlled by a single board micro-processor in the box with analog and digital I/O and a user interface that is running on the PC shown the background. A) shows the removable elastomeric extraction chamber (purchased from 87

Kalrez® perfluoroelastomer (FFKM)) being loaded into the plastic (PEEK) microfluidic chip. The fiber is placed on a holding feature (B) by the inspector using a microscope. The glass cover is put into place (C) to seal the extraction chamber and fiber location is verified

(D) using the microscope.

Figure 2-17. Microfluidic extraction apparatus

Figure 2-18. Microfluidic chip and elastomeric FFKM extraction chamber 88

2.4.2.1. Design of control panel

The user interface for controlling the microfluidic device is shown in Figure 2-19. The user has the ability to define the dye extraction parameters (i.e. extraction time and temperature, evaporation time, cool temperature for the buffer steps, and the number of buffer fills and settle times for dye pickup). Air pressures can be read from the pressure transducers and used to manually set the pressures with the regulators. Limits can also be set for each pressure so that the extraction program will not run when a pressure is out of a specified range. Manual control of individual components is also available and is mainly used for troubleshooting with the system. The user interface gives the user control over extraction conditions (these options are shown in the “Parameters” box). Due to the multitude of dye and fiber types, optimal extraction conditions may vary. By giving the user control over parameters such as temperature, extraction time, the number of buffer fills, and buffer settling times, optimization can be carried out with the device for each dye and fiber type. The automated programs are selected from the “Auto” box. The extraction sequence is initiated by clicking

“Run”. A cleaning sequence is also available. The cleaning sequence is performed between extractions to cleanse the valves, tubing, and manifolds that come into contact with the extracted dye.

Figure 2-19. Control window displaying operation of components and times

Figure 2-20. Extraction parameter inputs (Note: the unit of time is second)

The extraction programming was set as below: solvent volume: 10 µL, extraction time and temperature: 80 ºC, 5 min. The evaporation was performed with 10 psi Nitrogen gas for 20 seconds upon preliminary optimization test. As the micro-chamber cannot be moved during

extraction, HPLC buffer (3% acetonitrile) was allowed to stay in the chamber for another 20 second, to completely dissolve dye extract. Then additional buffers was directed to the microfluidic chamber to completely dissolve dyes, each of buffer rinses was around 20 µL to flush dye extract to a vial or directly to mass spectrometry analysis.

2.4.3. Quantitative analysis of dye extraction

Filtered dye extracts out of microfluidic extraction was analyzed by LC-MS analysis described in the last section. Standard dye solutions with known concentration (1.25-11.25

µg/L) were analyzed with the same LC-MS method with the sample). The expected ion peaks of analyte in the extract ion chromatogram (EIC) were integrated to make into standard calibration curve for quantification. The method detection limit (MDL) in this study is defined as the minimum concentrations of the dye on the substrates that could be detected.

Dye extractions from nylon threads (3 mm in length, ~0.2 mm in diameter, 0.625% to 3% owf) and single fibers (5 mm in length, dye concentration of 1% owf) were collected by microfluidic apparatus, and then analyzed with LC-MS or directly connect to the mass spectrometer for mass spectrometric (MS) analysis. The lowest dye concentration that can be detected (signal to charge ratio (S/N) that is higher than 10) is regarded to be the method detection limit.

2.4.4. LC-QTOF-MS Mass detection

Using optimized LC-MS chromatography, the studied commercial dyes together with the extract from microfluidic extraction were detectable in the mass spectrometer under the studied conditions demonstrate the mass spectra of the dye extract from C.I. Acid Blue 25 and C.I. Direct Red 81 obtained in conventional conditions. 91

Table 2-11 Mass spectra of dye extract from microfluidic extraction vs. conventional extraction

Exp. Actual Error Lowest Con. Chemical RT Deprotonated Dye Source mass Mass (ppm) (%owf) Formula (min) molecule (m/z) (m/z) Acid Conventional 5.270 (M-H)- 339.0545 -2.1 N/A C20H14N2O5S 339.0552 Blue 25 Microfluidic 5.266 339.0548 -1.2 0.0625 Acid Conventional 4.560 424.0035 0.125 C16 H13 N5 O3 S -0.71 Yellow (M-H)- 424.0038 Cl2 49 Microfluidic 4.560 424.0047 2.12 0.0625 Acid Conventional 6.140 581.0802 0.2 0.0625 Orange C26 H22 N4 O8 S2 (M-H)- 581.0801 67 Microfluidic 6.140 581.0804 0.5 0.125 Acid Conventional 6.113 785.1011 -4.5 0.125 Red C37H30N4O10S3 (M-H)- 785.1046 114 Microfluidic 6.133 785.1019 -3.4 0.125 Direct Conventional 4.420 (M-2H)2- 314.5349 1.9 N/A C29H19N5Na2O8S2 314.5343 Red 81 Microfluidic 4.469 (M-H)- 314.5333 -3.2 0.125 Direct Conventional 4.210 406.3703 1.5 0.125 3- Green C50H38N12O18S4 (M-3H) 406.3697 26 Microfluidic 4.207 406.3701 1.0 0.125

Figure 2-21. Mass spectrum of C.I. Acid blue 25 isolated from nylon fibers and C.I. Direct Red 81 from cotton fiber by pyridine extraction in conventional condition. (a) The shown mass spectrum of C.I. Acid Blue 25 was acquired at a retention time of 5.240 min. The inserts show a detail of the spectrum at m/z 393.0549; (b) The mass spectrum of C.I. Direct Red 81 was acquired at a retention time of 4.501 min, with the detailed spectrum at 314.5337. The mass spectrum of LC-MS mass spectra associated with the structure of the analyte is shown in Figure 2-22.

Figure 2-22. Discoloration of C.I. Direct Red 81 dyed fibers before and after automated microfluidic extraction.

Figure 2-23. LC-DAD-QTOF-MS data for C.I. Direct Red 81 extracted from a cotton fiber in microfluidic extraction system. (a) Total ion chromatogram; (b) DAD chromatogram at 540 nm; (c) Extracted ion current for m/z 314.53; (d) mass spectrum corresponding to the peak observed in the m/z 314.5333 extracted ion current.

LC-QTOF-MS analysis of C.I. Acid Red 114 is shown in Figure 2-24. It is clearly shown that the C.I. Acid Red 114 can be successfully detected with the proposed method.

Figure 2-24. LC-QTOF-MS data for C.I. Acid Red 114 out of microfluidic extraction system. (a) DAD chromatogram at 540 nm; (b) Extracted ion current for m/z 718.1019; (c) Mass spectrum corresponding to the peak observed in the m/z 718.1019 extracted ion current.

2.4.5. Quantification and Method Detection Limit (MDL)

Dye extracts from the textile threads collected with the microfluidic extraction apparatus were quantified by Agilent Mass Hunter Quantitative Analysis in LC-MS analysis. It is demonstrated that the lowest concentration of C.I. Acid Blue 25 at 0.0625% owf was successfully detected by mass spectrometry analysis, and other dyes can be detected at

0.125% owf. The expected ions in the Extracted-ion chromatogram (EIC) at m/z=393.0570 96

has an integrated peak area of 230396.73, which corresponds to 11.015 μg/L according to the equation given by the standard calibration curve (Figure 2-25). However, the dyes present in the single fiber were failed to be detected by the LC-MS analysis, due to the extremely low concentration of dye, the sample dilution by the mobile phases and sample loss during filtration.

2.4.6. Microfluidic Extraction-MS analysis of single fiber

The dye extract out of single fibers was directly analyzed with electrospray ionization mass spectrometric analysis (ESI-MS), so that sample dilution and filtration was avoided. The extraction procedures took 700 seconds (~11.6 min) in total before the molecular information was obtained by mass analyzer. Comparing to conventional extraction procedure, the microfluidic extraction has significanly reduced the time for sample preparation. The mass spectra of dye extract out of single fiber are shown in Figure 2-26(a) and Figure 2-26 (b), 97

respectively. Mixed nylon fibers with C.I. Acid Blue 25 and C.I. Acid Yellow 49 in the microfludic context can be both sucessfuly detected with high sensitivity and ion intensity, as are shown in Figure 2-26 (c).

Figure 2-26. MS Spectra of dye extract from single fibers (C.I. Acid Blue 25 and C.I. Acid Yellow 49) (~5 mm) with Automatic Microfluidic Extraction-MS. (a) Mass spectra of dye extract of C.I. Acid Blue 25 and C.I. Acid Yellow 24; (b) Mass spectra of C.I. Acid Blue 25;(c) Mass Spectra of C.I. Acid Yellow 49).

2.5. Summary

An automated microfluidic extraction device coupled to HPLC-DAD-MS has been developed for the extraction, separation and mass spectrometry analysis of dyes from trace-

level fibers. Dye extracts from millimeter-length fibers were successfully detected and quantified. In particularly, microfluidic extraction coupled to mass spectrometry successfully detected dyes in single fiber. The microfluidic extraction apparatus has shortened the time required for sample analysis from around 60 min in the conventional condition to less than

12 min. Combination of the automated microfluidic extraction to LC-MS or MS analysis presents a highly efficient, sensitive, and free-contamination methodology that can be applied for trace-level fiber analysis.

Chapter 3 High Performance Liquid Chromatography-Photodiode

Array-Mass Spectrometric Analysis of Direct Dyes

Abstract

Unambiguous identification and characterization of chemical composition of dyed fibers with high efficiency are important for fiber examination. However, in many cases organic dyes present in the fibers are unknown. Since the most rapid and efficient way to characterize unknown dyes is to compare a standard, it is necessary to establish a dye database in terms of retention time, color (UV-Vis absorption), molecular weight and other information. Chemical composition of unknown dyes can be preliminary determined according to fiber types, followed by chemical separation and mass spectrometry detection. Previous research disperse dyes and acid dyes have been d. However, not many research have been done to characterize direct dyes. In this chapter, HPLC-DAD-MS analysis for the identification of direct dyes was demonstrated in this section. A photodiode array detector placed in series after HPLC facilitates monitoring the elution of dyes in the presence of other non-dye components from colored fibers. Electrospray ionization mass spectrometry was used to detect the ionization of nonvolatile and thermally-labile molecules for mass spectrometry analysis. In this chapter commercial dyes collected from different manufactures were analyzed by a standard high performance liquid chromatography equipped with photodiode array detection, followed by mass spectrometry analysis.

3.1.HPLC-DAD-MS analysis of reference direct dyes

Mobile phase A: HPLC grade water (Burdick and Jackson),

Mobile phase B: 7: 3 Methanol (LCMS grade), acetonitrile (LCMS grade)

100

In this study, different variables in the gradient mobile phase B% were studied, and they were modified based on a gradient method developed in previous research. A programmed solvent gradient is listed in Table 3-1, which is named as Original Method.

Table 3-1 Time table of B% in Original Method

Time Flow Pressure Mobile Phase B% 0 0.5 400 3 1 0.5 400 3 1.5 0.5 400 60 7 0.5 400 90 9 0.5 400 90 9.5 0.5 400 3

3.1.1. Method development of chromatography

The constitution of mobile phase can significantly affect the ionization efficiency and retention behavior of analytes, therefore, the first step of optimization was to adjust the starting concentration of the chromatography from 3% to 10% of mobile phase B at the starting point.

Table 3-2 HPLC gradient of method 1

Time Flow Pressure Mobile Phase B% 0 0.5 400 10 9 0.5 400 90 9.5 0.5 400 10

Original Method

Method 1

101

3.953 min 5.693 min

Original Method Method 1

Figure 3-1. Chromatograms of C.I. Direct Yellow 106 at 410 nm with HPLC gradient of method 1.

In the chromatogram obtained by the Method 1 suggests the UV spectra in the DAD peaks at

410 nm decreased at longer retention time (5.693 min) than original method (3.953 min). The volume-percent of organic solvent (B %) reached to around 70% when the compound was detected. Fourthly charged molecule at m/z = 298.99 (M-4H)-4 was observed to process the highest abundance, suggesting the original method is slightly higher ionization efficiency than the new method 1(Figure 3-2).

Original Method

Method 1

Figure 3-2. Mass spectra of method 1 (Green) and the original method (Red).

102

3.1.2. Modification of gradient

The method 2 and method 3 process different gradients that were modified to be 10% or 20% less than the original method at 1.5 min. Method 2: (50% at 1.5 min) and Method 3 (40% at

1.5 min).

Table 3-3 HPLC gradient of method 2

Time Flow Pressure Mobile phase B (%) 0 0.5 400 10 40 (Method 2) 1.5 0.5 400 50 (Method 3) 7 0.5 400 90 11 0.5 400 90 11.5 0.5 400 10

Comparison of DAD rays of original method (blue), method 2 (green), and method 3 (red). It demonstrates that with steeper gradient at 1.5 min, chromatograms yielded sharper peaks.

Figure 3-3. Comparisons of chromatograms of HPLC-DAD-MS at 660 nm. Original method (blue), Method 2 (green), and method 3(red).

103

The ionization efficiency of C.I. Direct yellow 106 by four gradient chromatography are compared by the ion abundance of dye molecules at m/z = 298.9981 in Table 3-4. Method 3 was observed to have the highest peak area and relatively high ion abundance in mass spectrometry detection.

Table 3-4 Comparison of ion abundance of direct yellow 106 in liquid chromatograms

Method m/z DAD Peak Area Ion abundance Retention Time (min) Original 298.9981 30.78 1591.4 3.953 Method 1 298.9974 37.63 1237.9 5.693 Method 2 298.9981 5.34 1708.2 3.804 Method 3 298.9965 50.03 1632.8 3.507

3.2.Method validation

3.2.1. Limit of detection/quantification (LOD/LOQ)

Limit of detection was determined as the sample concentration that produces a peak that is ten times the level of the baseline noise. The limit of quantification (LOQ) was determined as the lowest concentration can be linearly quantified by the peak area of extraction ion chromatogram (EIC). Linearity determined by the ratio of the peak area to the dye concentration (Figure 3-4). Demonstrates from 5 µg/L to 10 µg/L a good linearity was achieved. Therefore the limit of quantification of this LC-MS method for C.I. Direct Yellow

106 was determined as 1.25 µg/L.

104

PKA of Extracted Ion Chromatogram 8.00E+03 7.00E+03 6.00E+03 5.00E+03 4.00E+03 3.00E+03 PKA/Con.

2.00E+03 PKA/Concentration 1.00E+03 0.00E+00 0 2 4 6 8 10 12 Concentration (ug/L)

Figure 3-4. Limit of quantification of C.I. Direct Yellow 106.

3.2.2. Linearity

The linearity range was preliminarily measured by plotting the peak areas of the extracted ion chromatogram (EIC) at m/z = 298.99 versus the analyte concentration 1.25 – 125 µg/L purified C.I. Direct Yellow 106 dissolved in HPLC-grade H2O. Triplicate tests were demonstrated in Figure 3-5, showing that linearity range from 1.25 – 125 µg/L was well established for future quantification of direct dyes obtained from microfluidic extraction.

105

PKA of Extracted Ion Chromatogram 6.00E+04 y = 4967.4x + 1184.7 5.00E+04 R² = 0.9936 4.00E+04

3.00E+04 PKA PKA 2.00E+04 Linear (PKA) 1.00E+04

0.00E+00 0 2 4 6 8 10 12 Concentration (µg/L)

Figure 3-5. Linear range of peak areas vs. concentration C.I. Direct Yellow 106.

3.2.3. Specificity

The purpose of specificity is to prove the LC-MS method is highly specific, three sets of dyes were analyzed using the optimized method. According to the DAD chromatogram, dyes absorbing different visible spectra have different wavelength. Four direct dyes with similar yellow color (were mixed and analyzed in Figure 3-6. The optimized liquid chromatography provided better separation and higher DAD peak intensity. The registered maximum wavelength fulfill the ideal retention factor for gradient chromatography 2

106

(11)

ktFBVS*  /()( %GM ) (11)

(12) SMW 0.25   (12)

Figure 3-6. Optimized HPLC-DAD chromatography and gradient B%.

107

3.3.HPLC-DAD-MS analysis of direct dye standards

The optimized HPLC-DAD-MS method was used as a standard protocol for dye extracts and commercial direct dye standards. Selected direct dyes and their structures are listed in Table

3-5.

Table 3-5 Information of direct dye standards for LC-MS analysis

1 2 No. C. I. Name Dye Class Source Molecular formula Molecular weight SO3 Group

1 Direct Blue 106 Oxazine Classic Dyestuffs. C30H16Cl2N4O8S2 694.9870 2

2 Direct Blue 108 Oxazine Classic Dyestuffs C34 H22 Cl2 N4 O11 S3 827.9824 3

3 Direct Blue 86 Phthalocyanine Classic Dyestuffs C32H14CuN8Na2O6S2 734.993 2

4 Direct Green 26 Polyazo Classic Dyestuffs C50H38N12O18S4 1222.131 4

5 Direct Green 6 Polyazo Ciba Geigy C34H22N8Na2O10S2 768.1057 2

6 Direct Yellow 106 Polyazo Huntsman C30 H18 Cl2 N4 O8 S2 1325.0279 2

7 Direct Orange 34 Polyazo Ciba Geigy C15H15NO3S 291.0678 1

9 Direct Yellow 27 Polyazo Ciba Geigy C25H22N4O9S3 618.6586 3

10 Direct Red 83 Polyazo Classic Dyestuffs C33 H24 N6 O17 S4 904.0081 4

11 Direct Red 80 Polyazo Classic Dyestuffs C45H26N10Na6O21S6 1371.8984 6

14 Direct Blue 15 Polyazo Classic Dyestuffs C34H28N6O16S4 904.8640 4

15 Direct Blue 90 Polyazo Classic Dyestuffs C38H20Na6N6O16S4 1082.7826 4

16 Direct Red 2 Polyazo Huntsman C34H26N6Na2O6S2 724.7175 2

1 C.I.: Color Index

2 Molecular weight of protonated ions

108

3.3.1. LC-MS analysis of oxazine dyes

Direct Blue 106: C30 H18 Cl2 N4 O8 S2

The chromatograms at DAD= 660 nm obtained from sample #1 had two major peaks at 4.020 min and 4.353 min (Figure 3-7), corresponding to two major blue substances are present in the sample. The UV absorption of the two peaks demonstrates a strong maximum at 630 nm and 300 nm with different peak intensities. The two major components detected were possibly structural isomers having modifications at different positions of the main compound, which had identical UV spectrum at different HPLC retention times.

4.353 min 4.020 min

Figure 3-7. DAD (660 nm) Chromatogram of Direct Blue 106. UV-Visible spectrum of Direct Blue 106 at 4.020 min and 4.353 min. In order to identify the two compounds at the different retention times, their mass to charge ratios were analyzed in Figure 3-8. At 4.412 min two major compounds were observed: doubly charged dye molecules the dimer. The peak of Extracted Ion Chromatogram (EIC) at m/z=346.9899, corresponding to(푀 − 2퐻)2−, and dimer (2M-2H)2- at 694.9898.

109

(A) (M-2H)-2H

(B)

(2M-2H)-2H

(M-2H)-2

Figure 3-8. MS Peaks of C.I. Direct Blue 106. (a) Doubly charged dye molecule (M-2H)-2; (b) Doubly charged dimer (2M-2H)-2.

Compound 2: Direct Blue 108 (C34 H22 Cl2 N4 O11 S3)

The HPLC-chromatogram of C.I. Direct Blue 108 observed multiple peaks at 660 nm (Figure

3-9), suggesting blue dyes were present in a mixture. Two of the major peaks observed at

3.88 min and 4.820 min at the 540 nm suggests the presence of reddish substances.

110

RT=3.74 min c RT=3.88 min RT=4.32 min RT=4.848 min RT=4.57 min RT=5.62 min

Figure 3-9. DAD Chromatograms of Direct Blue 108. (a) UV-Visible spectrum of Direct Blue 106 at 660nm; (b) Red substance detected at chromatograms at 540 nm; (c) the UV-Vis spectra at multiple retention times.

Further mass identification doubly charged dye molecules at m/z=412.9812, corresponding to

2− -2 (푀 − 2퐻) (Figure 3-10 (a)) and dimer (2M-2H) at 827.9801 (Figure 3-10 (b)).

111

Figure 3-10. LC-MS Mass spectra of Direct Blue 108. (a) UV-Visible spectrum of C.I. Direct Blue 108 at 660nm; (b) Doubly charged dye molecule (M-2H)-2; (c) Doubly charged dimer (2M-2H)-2.

112

3.3.2. LC-MS analysis of phthalocyanine dyes

Direct Blue 86(C32H14CuN8Na2O6S2) The HPLC-chromatogram of C.I. Direct Blue 86 observed multiple peaks at 660 nm (Figure 3-11), suggesting multiple blue substances were present the sample. Two of the major peaks observed at 3.74 min and 4.78 min had the same

UV-Vis absorption spectra as demonstrated at the top right corner. The characteristic molecule ions associated are deprotonated dye molecules at m/z=733.9866, corresponding to

(M − 2H)2− (Figure 3-11 (a)) and dimer (2M-2H)2- at m/z 366.4906 (Figure 3-11 (b)).

Figure 3-11. LC-MS Mass spectra of C.I. Direct Blue 86. (a) UV-Visible spectrum of C.I.Direct Blue 86 at 660nm; (b) Doubly charged dye molecule (M-2H)-2; (c) Doubly charged dimer (2M-2H)-2. a

113

3.3.3. LC-MS analysis of azo direct dyes.

Compound 3: Direct Green 26 (C50H38N12O18S4) was detected at retention time = 4.303 min

(Figure 3-12) as a single peak. Dye molecules were detected as he dominant peaks of which were the triply charged ions in the mass spectrometry analysis (Figure 3-12(a)). The highest number of charge molecules detected was not the total number of sulfonic groups, which differs with previous study. Dyes molecules were detected as a fourthly charged and triply charged ions as are shown in Figure 3-12(b) Dominant mass ion at 406.3703 has a featured isotopic pattern as is shown in Figure 3-12(c). It is worth noticing that the ion abundance of the isotopic peaks decreases as the isotopic molecule weight increases.

114

115

Compound 3: Direct Green 6 (C34H22N8Na2O10S2)

Analysis of Direct Green 6 was found to have a single peak at 254 nm, which corresponds to expected structure in Colour Index. The dominant mass peaks were detected at m/z= 383.04, corresponding to (M-2H)-2 (Figure 3-13).

Figure 3-13. LC-DAD-MS analysis of Direct Green 6 at 660 nm at 3.997 min. (a) Chromatograms obtained from C.I. Direct Blue 90 at 660 nm; (b) Chromatograms obtained from C.I. Direct Blue 90 at 254 nm.

116

Compound 4: C.I. Direct Blue 90 (C38H20N6Na4O16S4)

C.I. Direct Blue 90 was detected by DAD chromatograms at 660 nm at 3.747 min. However, differences existed between the expected and detected. Possible compounds in this commercial samples were at m/z = 533.9127 (Figure 5-8).

Figure 3-14. LC-DAD-MS analysis of Direct Green 6 at 660 nm at 3.997 min. (a) Chromatograms obtained from C.I. Direct Blue 90 at 660 nm; (b) Chromatograms obtained from C.I. Direct Blue 90 at 254 nm.

117

Compound 5: C.I. Direct Red 2 C34H26N6Na2O6S2

Figure 3-15. Chromatograms of C.I. Direct Red 2 at 3.800 min.

Compound 6 Direct Blue 15 C34H28N6O16S4

The HPLC chromatogram of C.I. Direct blue 15 is shown in Figure 3-16. Multiple peaks were detected at chromatograms at 254 nm. UV spectra of these peaks were overlaid. It is shown that they have different absorption coefficients, but the absorption of the wavelength were the same.

118

Figure 3-16. The HPLC chromatogram of C.I. Direct Blue 15.

119

Part of the direct dyes are demonstrated in the Figure 3-7.

Table 3-6 Analysis of optimized chromatography for direct dyes

Dye name Number of SO3H MF DAD λmax (nm) MS (RT) MI* -2 Direct Blue 106 2 C30 H18 Cl2 N4 O8 S2 4.353 630 4.437 (M-2H)

-2 Direct Blue 108 2 C34 H22 Cl2 N4 O11 S3 3.880 610 3.968 (M-2H) (M-H)- -2 Direct Blue 86 2 C32H16CuN8O6S2 3.996 660 4.780 (M-2H) (M-H)- -2 Direct Green 26 5 C50H38N12O18S4 4.207 627 4.295 (M-2H) (2M-4H)-2 (M-3H)-3 (2M-3H)-3 (M-4H)-4 -2 Direct Green 6 C34H22N8Na2O10S2 3.987 680 4.087 (M-2H)

-2 Direct Yellow 106 6 C30 H18 Cl2 N4 O8 S2 5.955 410 6.055 (M-2H) (2M-4H)-2 (M-3H)-3 (2M-3H)-3 (M-4H)-4 - Direct Blue15 4 C34H28N6O16S4 3.780 630 3.791 (M-H) (M-2H)-2 (M-3H)-3 Direct Blue 90 4 C38H24N6O16S4 3.747 615 3.816 N/A

Direct Red 81 2 C29H21N5O8S2 4.313 520 4.387 M-2H *Highlighted items are the dominant peaks observed with generalized LC-MS method.

A generalized LC-MS method was used to detect direct dyes in different chemical classes.

Oxazine dyes and metal complex dyes tend to have multiple peaks in DAD chromatograms.

Multiple charged molecules were detected for polysulphonated direct dyes. Oxazine

molecules yield doubly charged dimers. Other than the essential components defined in the

Colour Index, other components were detected to differentiate samples from different

manufactures and batches.

3.4.References

1. Page, J. S.; Kelly, R. T.; Tang, K.; Smith, R. D., Ionization and Transmission

Efficiency in an Electrospray Ionization–Mass Spectrometry Interface. Journal of the

American Society for Mass Spectrometry 2007, 18 (9), 1582-1590.

120

Chapter 4 Chromatographic Optimization of Azo Disperse Dyes

4.1.Introduction

The purpose of this study was to optimize a simple and reproducible chromatography method for disperse dye characterization. Azo disperse dyes with similar structure were studied, with the purpose of investigating the effects of substituents on the retention behavior. A series of anthraquinone dyes with different substitute groups were selected.

4.2.Experimental

Dyes and chemicals

Commercial disperse dyes in this study are summarized in Table 4-1.

Table 4-1 Azo disperse dyes for LC-MS analysis.

Sample # Name Structure

1 Disperse Yellow 3

2 Disperse Red 1

3 Disperse Red 167

4 MDL3

121

Table 4-1 Continued

5 MDL4

6 MDL5

7 MDL8

8 MDL10

9 Disperse Yellow 42

10 Disperse Yellow 86

11 Disperse Yellow 54

12 DO-30-95

13 MDL7

122

Accurate mass analysis of commercial dyes/dye extracts was performed using Agilent

Technologies 6520 Accurate-Mass Quadrupole Time-of-Flight mass spectrometer (QTOF-

MS). Parameters were set as follows: dual electrospray ionization (ESI) source voltage 4 kv under positive mode, nebulizer pressure (35 psig), capillary voltage 9 (4000 V), drying gas flow (12 L/min at 350 °C), and fragmentor voltage 110 V. Data collection and analysis were performed using Agilent Qualitative Analysis B.05.00.

Direct Injection with ESI and APCI MS analysis of samples in NCSU dye library

Azo dyes collected from dye library were preliminary examined using direct injection. 0.5 mg/mL dye stock was diluted in organic solvent methanol/acetonitrile 7:3. Mobile phase B contains 1% formic acid in methanol, and Mobile phase A is composed of 1% formic acid in

Milli-Q H2O. Flow rate 0.2 ml/min, injection volume 1 μL and run for 2 min. Fragmentor voltage 100 V.

4.3.Results and Discussions

Direct infusion of dye solution collected from dye library were analyzed by different ionization sources. It is demonstrated that both ionization sources successfully ionized samples to form protonated molecule ions.

123

Table 4-2 Comparison of APCI and ESI mass ionization

ESI Expected Found Name Structure APCI Domiant APCI Ions (Y/N) Ions Y Y (M+H)+: -0.0027 -0.0027 282.27, 329.1608 (5.56 (6.72ppm) 304.25, 205.08, MDL3 ppm) 336.24, 282.27, 536.16, (M+Na)+: C17H20N4O3 N N 610.17 351.1428 Y Y (M+H)+: 0.0028 (-0.0033) 357.1557 7.73 282.27, 304.25, MDL4 ppm 282.27 536.16, C18H20N4O4 (M+Na)+: 610.17. N N 379.1377 Y Y (M+H)+: (-0.004) (-0.0054) 333.1113, (5.51 No MFG ppm) 282.27, 282.27, M+Na: 304.25, 304.25, MDL5 336.24, 336.24, 355.0902 N N 319.23, 393.20, C16H17ClN4O2 437.22. 563.54 (M+Na)+: N N 369.1089

(M+H)+: Y Y 282.29, 336.25, 282.29, MDL8 393.14, 563.54 (M+Na)+: C18H18N6O4 N N 563.55. 360.1431

Y 282.29, Y (- (M+H)+: 336.25, MDL10 (0.0034) 0.0019) 347.1269 347.13, 282.29 (-9.12 pm) (4.72 563.25 C17H19ClN4O2 ppm)

Based on direct mass spectrometry analysis, the dominant ions at m/z 282.29 were observed in both ions, but it was not the main component, the expected ions were identified.

124

HPLC-DAD-MS analysis of reference standard dyes

0.3 mg/mL dye standard solutions were separated through Agilent Technologies 1260 SL high performance liquid chromatography at 45 °C. The chromatographic separation was performed with an Agilent ZORBAX 2.1mm x 150 mm narrow bore column C18 column.

The mobile phase consists of (A) Milli-Q H2O containing 0.1% formic acid and mobile phase

B consisting of acetonitrile 0.1% Formic Acid at a flow rate of 0.5 mL/min.

Table 4-3 HPLC gradient for disperse dyes

Time (min) Mobile phase B% (Acetonitrile) Flow rate (ml/min) 0 30 0.5 2 60 0.5 10 70 0.5 18 70 0.5 18.5 30 0.5

Accurate mass analysis of commercial dyes/dye extracts was performed using Agilent

Technologies 6520 Accurate-Mass Quadrupole Time-of-Flight mass spectrometer (QTOF-

Color and structure

Colors of azo dye samples containing different substation were studied. Comparison of

MDL3, MDL4 MDL5 (Figure 4-1 (a)) demonstrates a functional group that has higher electronegativity tends to have a bathochromic shift. Hypsochromic shifts were observed

125

when subsititutents were replaced by COOH and OH, because the electronegativity order of these substituents are Cl >COOH> OH. However, addition of CH3 group to the aromatic rings demonstrated different directions of color shifts. Hypsochromic shift was observed with addition of a methyl group in MDL 10 (Figure 4-1 (b)), but opposite directions were observed with addition of a methyl group to disperse red 1 (Figure 4-1 (c)). Cl group present in structure has a stronger electron withdrawing effect than the OH groups.

MDL3: R= C2H5OH MDL4: R= C2H4COOH MDL5: R= C2H4Cl

b +CH3 c +CH3

MDL10: R1= H Dis Red 1: R=H, 500 nm MDL5: R = CH 1 3 MDL3: R=CH3, 506 nm

Figure 4-1. Color and substituent groups of disperse Azo dyes. a. UV-Vis Absorption of MDL3, MDL4 and MDL 5. (b) UV absorption of MDL 10, and MDL5. (c) UV Vis of Disperse Red 1 and MDL3.

126

LC-MS analysis of azo dye mixtures

Mixtures of disperse azo dyes in this study were successfully separated by LC-MS assay. The corresponding compounds label in the chromatograms in Figure 4-2 demonstrates all the compounds were successfully detected by DAD detection.

7 8 9 13 a 2 4 6 10 11 12 1 3 5

4 9 b 2 7 8 10 11 6 12 13

7 c 8 10 11 12 6

3 5 d 1

Figure 4-2. LC-DAD detection of Azo dye mixtures: a 254 nm, b: 410 nm, c: 540 nm, d: 660 nm. (Compounds: 1,3, 5 are blue substances in disperse blue 3. LC-DAD chromatograms at 254 nm detected thirteen compounds. Three blue substances were detected at 660 nm was found in Disperse Blue 3. Six red substances detected at 540 nm were disperse red 1 (Peak 6), MDL3 (Peak 7) and MDL 4 (Peak 8), MDL 5 (Peak 11),

MDL8 (Peak 10), and MDL 10 (Peak 12). They are also detected at DAD = 410 nm. It should be noted that MDL 8 and Direct Red 167 eluted at the same retention time at around

7.193 min, and they could be differentiated by mass spectrometry only with the current LC-

MS assay.

127

Table 4-4 Peak information of LC-MS of disperse azo dyes

Retention Times λ Number DYE m/z Peak max (EIC) (nm) 1 MDL3 329.1608 4.619 7 521 2 MDL5 333.1113 7.984 11 478 3 MDL4 357.1557 5.306 8 500 4 MDL7 428.1881 5.756 13 520 5 MDL8 384.1444 7.193 10 505 6 MDL10 347.1269 9.015 12 483 7 DY3-70 270.1250 3.457 4 360 8 DR-1-72 315.1452 4.192 6 503 9 DO-30-95 450.0730 6.395 9 421 DR-167- 11(overlaid with 10 506.1470 7.296 509 102 MDL8) 11 DB3-71 297.1233 2.095 1,3,5 616

LC-MS analysis of anthraquinone dye mixtures

The optimized method was used to separate a series of disperse dyes as summarized in

Table 4-5.

Table 4-5 Anthraquinone dyes for HPLC-DAD-MS separation and identification

Sample # Dye Structure

DR 60 1

DB3 2

DR 91 3

128

Table 4-6 continued

DB 56 4

5 DB60

6 DB27

7 DR86

8 DY54

9 DY42

129

Table 4-7 continued

10 DB73

11 DB77

Preliminary developed methods were found to be hard to separate anthraquinone dyes very well.

It is shown that the disperse Yellow 42 and direct red 91 are overlapped, but they have completely different UV-Vis absorption can be separated completely. The figures showing here found anthraquinone dyes did not separated well. Two peaks were overlaid that Disperse

Red 91 and Disperse Yellow 42, and Disperse 60 and Disperse 54 were eluted at around the same time. Therefore, new chromatographic method is required to achieve a better separation for anthraquinone dyes.

130

DR 91 DY42(370.0856, 4.033min)

DB 3 DY 54 (4.768 min)

DB 77(6.048 min) DR 86(5.526 min) DB 73

DB 60(4.8710 min)

Figure 4-3. LC-DAD-MS analyses of mixture of disperse anthraquinone dyes.

LC-MS analysis of C.I. Disperse Blue 56

The chromatographic method developed previous enables to detect multiple components in commercial samples. Take C.I. Disperse Blue 56 as an example, four blue substances were detected by DAD chromatograms at 660 nm (Figure 4-4). Mass spectrometry analysis suggests that the major components present in this sample was 1,5-diamino-4,8- dihydroxyanthracene-9,10-dione (Figure 4-4 c) instead of 4-amino-2-bromo-1,5- dihydroxyanthracene-9,10-dione (Figure 4-4 b).

131

b c

Figure 4-4 (a) Chromatograms of C.I. Disperse Blue 56 with old chromatograms;(b) Expected structure observed at 9.467 min; (c) Mass spectrum of the main component of C.I. Disperse Blue 56 detected at retention time of 6.220 min.

LC-MS analysis of C.I. Disperse Blue 3

The LC-MS analysis with previous method did not separate all components present in the commercial samples of C.I. Disperse Blue 3, instead, four components were separated by newly developed chromatographic methods (Figure 4-5c), that the corresponding components are proposed in Figure 4-6.

132

old chromatogram New chromatogram c

Figure 4-5. (a) Chromatograms of C.I. Disperse Blue 3 with HPLC-DAD with a newly developed chromatographic method; (b) Chromatograms with a previous developed method; (c) Comparison of newly developed chromatographic method and the old method.

133

a b Disperse Blue 3

c d

Figure 4-6. Mass spectrometry of components in C.I. Disperse Blue 3. (a) Structure proposed in Colour Index; (b) Unknown blue substance at retention time of 1.753 min; (c) component 2 retained at 3.544 min; (d) component 4 that were detected at 4.874 min.

4.4.Conclusions

It has been shown that the optimized chromatographic method enables successful separation of the dye components of commercial dye samples, including disperse azo, anthraquinone dyes, structure isomers present in each commercial samples. Disperse azo dyes were found to have relative pure peaks, but certain anthraquinone dyes samples contain more components than expected. Further optimization of chromatographic methods is needed to completely separate these components present in commercial dyes.

134

Chapter 5 ESI QTOF Tandem Mass Spectrometric Analysis of Sulfonated

Anthraquinone Acid Dyes

Abstract

Sulfonated dyes account for a large proportion in textile dyes. In this study a series of sulfonated dyes with the structure of 1-amino anthraquinone-2-sulfonate were studied using liquid chromatography, electrospray ionization and quadrupole-time-of-flight tandem mass spectrometry (HPLC-ESI-QTOF-MS/MS), with the purpose of understanding their fragmentation behavior during collision-induced-disassociation (CID) in tandem mass spectrometry. The fragmentation mechanisms were investigated and validated by tracking isotope sulfur atom of the sulfonic acid group. A featured ion loss of SO2∙ at 64 amu was observed in the CID, which was proposed to be caused by unimolecular rearrangement and fragmentation. Furthermore, we identified the fragmentation pattern and correlate them with particular substituent groups. The proposed mechanisms were validated and used for predicting featured fragmentation pattern of C.I. Acid Blue 129 with known structure at the same CID condition. The detailed understanding of fragmentation mechanisms enables to establish correlation between the fragmentation patterns with the major structure and particular substituent groups. The present approach is applicable for structural analysis of unknown dyes in mixtures using HPLC-MS/MS techniques.

Key words: Sulfonated anthraquinone, Mass spectrometry, Mechanisms, Tandem mass spectrometry, Structure elucidation.

135

5.1.Introduction

Detection and characterization of unknown dyes are becoming increasingly important in the areas of forensics,37 food,24 environment38 and human health.39 Combination of liquid phase separation techniques with mass spectrometry and tandem mass spectrometry37b, 37e, 40 are powerful tools to characterize unknown dyes in mixtures.39 For example, high performance liquid chromatography (HPLC) coupled to electrospray ionization (ESI) mass spectrometry has been frequently used to separate and characterize unknown dyes in waste water, food, and fibers.41 The HPLC coupled to tandem mass spectrometry (MS/MS) is a powerful tool for structural analysis.

The structure information provided by the tandem mass spectrometry relies on the interpretation of featured fragments of the ions of interest under certain fragmentation conditions. A typical way of fragmentation is collision-induced-disassociation (CID), during which the ions of interest (parent ions) are isolated and collide with neutral gas molecules to generate featured fragment ions (product ions) and neutral loss that can be analyzed in the mass spectra.42 Compounds with certain functional groups have their characteristic neutral loss during fragmentation. For example, Michal H and his coworkers found the sulfonated dyes have a featured loss of SO2 or SO3 in the electrospray ionization tandem mass spectrometry 43. This finding agreed with other research performed with different ionization sources. For example, a loss of SO3 has been observed in the tandem mass spectrometry analysis of sulfonated azo dyes using fast atom bombardment (FAB),44 atmospheric pressure ionization,45 and liquid secondary ionization46 for the analysis of sulfonated azo dyes. The

136

loss of SO2 was also reported in the fragmentation of sulfonated aromatic compounds, including benzene sulfonate 47, sulfonyl-sulfinate 48 as well as other sulfonated compounds.49

Sulfonated dyes have been extensively studied with tandem mass spectrometry and characteristic ions were observed. However, much less research has focused on the mechanism of the fragmentation, especially on the interactions between the functional groups and the correlation with the main structures. In this study, a series of sulfonated anthraquinone acid dyes containing the structure of 1-amino anthraquinone-2-sulfonate acid were analyzed with electrospray ionization tandem mass spectrometry. Such type of dyes have wide applications for their good light stability and leveling properties in textile industry

2. The purpose of this work is to have a detailed interpretation of the fragmentation pattern of sulfonated anthraquinone dyes, and to propose mechanisms that can be used for unknown dye analysis.

Figure 5-1. Basic structure of the 1-Amino-2 sulfonated anthraquinone dyes. R: Alkyl or aromatic groups.

137

5.2. Experimental

5.2.1. Chemicals and materials

Selected acid dyes and their structures are listed in Table 5-1. Methanol (high purity for LC-

MS grade), acetonitrile (high purity for LC-MS grade) was purchased from Honeywell &

Burdick Jackson (Muskegon, MI, USA). The solvents were filtered through 0.22 µm

Millipore filters (Whatman, GE Healthcare, UK) prior to use. Ammonium formate (>99%%,

HPLC grade, Fluka, Switzerland) and formic acid (~98%, MS grade) were purchased from

Sigma Aldrich (St Louis, MO, USA). The acid dyes (Table 5-1) were supplied by Ciba

Specialty Chemicals, M.Dohmen, and Classic Dyestuffs.

Table 5-1 Information of C.I. Acid Dyes

No. C. I.3 Name Manufacturer Trade Name Molecular formula4 Molecular weight Acid Group

A1 Acid Blue 25 M.Dohmen Dorasyn Blue AG C20H14N2O5S 393.0529 1

2 Acid Blue 45 Ciba Geigy Erio Cyanine S 150% C14H10N2O10S2 428.9781 2

3 Acid Blue 40 M.Dohmen Dorasyn Blue A2G 200% 450.0763 1

4 Acid Blue 62 Classic Dyestuffs Bernalizarine Blue SRA C20H20 N2O5S 399.1009 1

5 Acid Blue 277 Ciba Geigy Tectilon Blue 46 C24H23N3O8S2 544.0846 1

3 C.I.: Color Index

4 Molecular weight of protonated ions

138

C.I. Acid Blue 25 C.I. Acid Blue 62 C.I. Acid Blue 40

C.I. Acid Blue 277 C.I. Acid Blue 45

Figure 5-2. Structure of sulfonated anthraquinone acid dyes.

5.2.2. Instrumentation

Accurate MS analysis of commercial dyes were performed by using a Aglient Technologies

1260 SL liquid chromatograph coupled to an Agilent Techologies 6520 Acurrate-Mass Q-

TOF mass spectrometer equipped with an electrospray ionization (ESI) source operating in negative ion mode. Each analyte was prepared for analysis by mixing 20 uL of 1 mg/mL dye in Milli-Q water. As the commercial dyes may contain impurities includes auxiliaries, intermediates or by products during manufacture, standard single dye solution were separated on a Agilent Zorbax Eclipse Plus C18 (2.1 × 50 mm, 3.5 μm) with a Zorbax Eclipse Plus C18 narrow bore guard column (2.1×12.5 μm, 5 μm). The mobile phase used for separation 139

consists of a mixture of 20 mM ammonium formate and 0.01% formic acid in water (A) and

7:3 methanol/acetonitrile (B). Negative ESI mass spectra were acquired by setting the nebulizer pressure as 35 psi, capillary voltage as 4000 V, drying gas flow as 12 L/min at

350 °C, fragemntor voltage as 110 V, and collision voltage as 25-40 V.

Electrospray tandem mass spectrometry was performed using a quadrupole time-of-flight mass analyzer. The target mass for each dye was set according to expected masses of dye molecule at the expected retention time; the isolation width for tandem mass spectrometric experiments was set as m/z = 1. Data collection and analysis were performed using Agilent

MassHunter Workstation Acquisition and Agilent MassHunter Quanlitative Analysis

B.05.00, respectively. The capillary was held at 4 kv. The collision potential offset was 40 V with nitrogen gas as the collision gas.

5.3.Results and Discussions

5.3.1. ESI mass spectra of individual dyes

Anionic acid dyes were measured in the negative mode, where the dyes present as deprotonated molecules [M-H]−. Figure 5-3 shows an example of the mass spectrum of C.I.

Acid Blue 62, which demonstrates the deprotonated molecule [M-H]− at m/z = 399.1014.

140

[M-H]-

Figure 5-3. Negative-Ion ESI mass spectra of C.I. Acid Blue 62.

5.3.2. Tandem mass spectrometry analysis of single dye

C.I. Acid Blue 25

The LC-MS/MS mass spectrum of C.I. Acid Blue 25 is illustrated in Figure 5-4. The predominant peak at m/z 329.0926 represents the main product ions out of activated collision by losing a neutral loss at m/z –63.96 amu from the precursor ion at m/z 393.0529. The

38b, 43, 48, 50 corresponding neutral loss may correspond to SO2, according to previous research .

The second highest product ion is at m/z 301.0944 amu after a loss of NC6H5. Since a higher ion abundance of product ions indicates a more favorable fragmentation path, product ions with higher abundance are generated more easily than other fragments. Based on this theory, the relatively high peak intensity at m/z 329.0926 indicates the loss of –SO2 loss dominated the fragmentation, while only small part of the compounds have the cleavage of C-N bonding of the secondary amine connected to the anthraquinone aromatic system (R is aromatic system). In this case, the benzene acted as an electron donor system, so that the resonance

141

occurs in the molecule tend to stabilize the system. Similar fragmentation pattern were observed in both C.I. Acid Blue 40 and C.I. Acid Blue 45.

Figure 5-4. LC-MS/MS mass spectra of C.I. Acid Blue 25 CID Voltage = 40 V.

C.I. Acid Blue 62

The product-ion spectra of the [M-H]− ion of C.I. Acid Blue 62, generated by LC-MS/MS is shown in Figure 5-5. Comparing to C.I. Acid Blue 25, C.I. Acid Blue 62 underwent fragmentation in a different pathway: the predominant peak in the LC-MS/MS mass spectrum is at m/z 316.0155, corresponding to a cleavage of the N-C bonding between the secondary amine and the carbon on anthraquinone. The production ion is further fragmented by losing of a SO2 at m/z 252.0452.

142

Figure 5-5. Fragmentation scheme and Mass Spectra of LC-MS/MS analysis of C.I. Acid Blue 62 CID Voltage = 40 V. Comparison of the mass spectra of C.I. Acid blue 45 and C.I. Acid blue 25 indicates that the ion abundance of the fragments is closely related to the nature of the substituent groups attached to the secondary amine: A loss of alkyl group has much higher ion abundance than that of SO2 loss, as is demonstrated in Figure 5-5. However, when the substituent group is an aromatic ring, as in the case of C.I. Acid Blue 25 showing in Figure 5-4, SO2 loss is more favorable, suggesting the aromatic group attached to the secondary amine is more difficult to be fragmented than SO2. In this case, the aromatic ring is more stable than the alky ring.

C.I. Acid Blue 277

In order to examine the fragmentation when the aromatic rings contains substituting group,

C.I. Acid Blue 277 containing a side chain of SO2NHC2H4OH attached to the aromatic ring was analyzed. As is shown in Figure 5-6, the peak [M-H- SO2NHC2H4OH]− at m/z 420.0780 is most abundant, followed by fragment ions [M-H- SO2NHC2H4OH-SO2]− that at m/z

356.1165, and [M-H-SO2]− at m/z 480.1213. Therefore, the peak with a loss of side chain of the aromatic ring is the dominant fragment ion. This is because the precursor ion may lose 143

easily the SO2NHC2H4OH than losing a SO2. These observations indicate the energy required by the molecule to lose SO2 is higher than that required by the C-N bond between the benzene and the secondary amine when it is connected to the aromatic system 42.

Figure 5-6. LC-MS/MS Mass Spectra of C.I. Acid Blue 277 at CID Voltage = 40 V.

C.I. Acid Blue 45

Unlike other compounds, C.I. Acid Blue 45 is a disulfonated anthraquinone dyes containing two sulfonic groups. C.I. Acid Blue 45 shown in Figure 5-7 demonstrates a successive cleavage of SO2 produce two dominated peaks, corresponding successive cleavage of SO2.

Figure 5-7. LC-MS/MS Spectra of C.I. Acid Blue 45.

144

A summary of the fragmentation pattern of all the dyes studied is shown in Table 5-2, and

Figure 5-8. It is clear that the neutral loss of SO2 was observed in all dyes studied, but their peak abundances vary with the substituents attached to the amine. The product ion corresponding to a loss of SO2 has highest peak abundances when the substituents attached to the secondary amine are aromatic systems (C.I. Acid Blue 25, C.I. Acid Blue 40, and C.I.

Acid Blue 45). The ion abundances of the mass spectra of LC-MS/MS follow the orders:

Side chain>alkyl substitution> SO2> aromatic substitution.

Table 5-2 MSMS Fragmentation under CID 40 V

No. m/z Structure Abundance % - 329.0926 [M-H -SO2 ] 100 C.I. AB25 393.0529 [M-H]- 10.66 - 301.0944 [M-H -NHC6H5] 13.58 - 252.0506 [M-H-SO2-NHC6H5] 4.3 - 386.1151 [M-H -SO2 ] 100 450.0763 [M-H]- ]- 63.32 C.I. AB40 343.0961 [M-H -SO -C H O]- 43.19 2 2 3 - 237.0429 [M-H -SO2-C8H9N2O] 4.17 - 252.0540 [M-H -SO2-C8H8NO] 3.41 - 365.0052 [M-H -SO2] 100 C.I. AB45 285.0527 [M-H -SO -SO ]- 68.1 2 2 428.9781 [M-H]- 56.39 - 316.0161 [M-H -C6H11] 100 C.I.AB62 399.1006 [M-H]- 14.09 - 289.0060 [M-H -SO2] 8.59 - 252.0533 [M-H -SO2-C6H11N] 7.79 - 301.0178 [M-H -C6H12N] 1.41 544.0846 [M-H]- 100 - 420.0780 [M-H –SO2NHC2H4OH] 96.54 - C.I. AB277 356.1165 [M-H -SO2NHC2H4OH-SO2] 24.49 - 480.1213 [M-H -SO2] 18.78 - 301.0045 [M-H -C10H15N2O3S] 7.36 - 237.0427 [M-H -C10H15N2O3S-SO2] 3.22

145

C.I. Acid Blue 25, C.I. Acid Blue 40, C.I. Acid Blue 62 are sharing common fragmentation ions at m/z 252.05, and C.I. Acid Blue 45, C.I. Acid Blue 277 are sharing m/z at 237.04 amu, the shared fragmentation ions is worth studying in the future work.

Deprotonated C.I. Acid Blue 25 Deprotonated C.I. Acid Blue 40

Deprotonated C.I. Acid Blue 45

Deprotonated C.I. Acid Blue 62 Deprotonated C.I. Acid Blue 277

Figure 5-8. Fragmentation pathway of acid blue dyes (all dyes lost SO2 as a characteristic neutral loss). The order of fragmentation was determined by the energy required to cleave off substituents (Red: 1st step, Blue: 2nd step, Green: 3rd step).

146

5.3.3. Proposed mechanisms of Collision-Induced Dissociation

Mechanisms of loss of SO2

According to the detailed interoperation of mass spectra, loss of SO2 rather than SO3 was observed by all the sulfonated anthraquinone dyes. Here we proposed fragmentation pathway shown in Figure 5-9. Take C.I. Acid Blue 25 as an example, during collision induced disassociation, the electron of oxygen atom in the sulfonic acid group is abstracted by the electron deficient carbon atom, and at the same time the hydrogen in the ortho-amino group is capable to form intermolecular hydrogen bonding with one of the oxygen atoms to form a six-member ring, which acts as an anchor, favoring the rearrangement that allows the loss of

SO2 during CID. Alternatively, during rearrangement an epoxide formed from oxygen to carbon, formation a C-O bond requires less energy than C-S bond, and the atoms tends to form a stable bond between carbon and oxygen rather instead of C-S bond 51.

Figure 5-9. Proposed mechanism of fragmentation of sulfonated anthraquinone CID, where the ortho-NH3 favors the rearrangement of SO2.

147

Confirmation of loss of SO2

34 In order to prove the loss of 64 amu is SO2 loss rather than other fragments, isotopic S and

32S was traced during the fragmentation of two model compounds (Figure 5-10). Observation

34 of loss of 65.91 amu in the mass spectra shown in Figure 5-11 further proves SO2 loss in

32 34 CID. Loss of both SO2 and SO2 were observed in both of the fragments of the

deprotonated molecules. Therefore, the loss of SO2 during CID is a characteristic cleavage of

the sulfonated anthraquinone dyes.

(A) (B)

Figure 5-10. Model Compounds Compound A and Compound B

As MS_131022170309 #1 RT: 0.00 AV: 1 NL: 1.30E5 As MSMS_304 #1 RT: 0.00 AV: 1 NL: 6.05E3 T: ITMS - p ESI Full ms2 [email protected] [80.00-600.00] T: ITMS - p ESI Full ms2 [email protected] [80.00-600.00] 238.08 238.17 100 100

90 90

80 80

70 70

60 60

50 50

40 40

RelativeAbundance 30 RelativeAbundance 30 304.08

20 302.00 20

10 10 131.00 187.25 283.33 141.08 173.17 224.17 286.17 478.33 532.92 573.42 0 0 100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600 m/z m/z

(a) (b)

Figure 5-11. Tandem mass spectrum of (1 amino anthraquinone sulfonic acid) containing 32S (a), and 34S (b).

148

Function of ortho-amino group in the CID.

In order to confirm the ortho-amino group is playing a role during collision-induced disassociation, two model compounds (Figure 5-10) were investigated. They were analyzed under a CID energy ranging from 10 V to 35 V in 5V intervals. Survival yield of tandem mass spectrometry was employed to compare the energy required during fragmentation 52.

The survival yield is defined as the ratio of the precursor (parent) ion intensity to the sum of parent and fragment ion intensities in the mass spectrum, as is shown in the Equation (1).

Survival Yield = IM/(IM + IF) (1)

Where IM represents the peak intensity of parent ions left in the mixture; IF represents the intensity of fragments in mass spectra. For two compounds that are fragmented at the same

CID energy, a higher survival yield suggests the compound requires higher energy to be fragmented 52.

Comparison of the survival yield of compound A and B shown in the Figure 5-12 suggests that at the same CID voltage the model compound A with a ortho-amino group has a higher survival yield, demonstrating the presence of an ortho-amino group give a higher yield of parent ions, which is possibly caused by the amino group which is stabilizing the molecules from being fragmented 51.

149

Compound A Compound B

1.0 0.96 0.91 0.9 0.8 0.74 0.74 0.6

0.4 0.42 0.45

Survival Yield (100%) Survival 0.2 0.16 0.19

0.0 0.04

10 15 20 25 30 CID Voltage (V)

Figure 5-12. Comparisons of survival yields of compounds A and B. Compound A with NH3 has higher survival yield under CID Voltage = 40 V.

Application of fragmentation mechanisms

As the proposed mechanism has been validated and the correlation between the substituents and ion abundance has been established, they could be used to predict the fragmentation pattern (types and relative abundance of fragment ions) of dyes with the similar structures.

Here we use C.I. Acid Blue 129 (Figure 5-13) as a tester to undergo the same fragmentation condition. If the mechanism proposed above is correct, the dye fragments should have predominant fragment ions corresponding to loss of SO2, and the second highest peak should be the fragments with a loss of benzene rings, because the aromatic groups attached to the meta-position amine requires higher energy to be fragmented than loss of SO2. Furthermore, a characteristic fragmentation peak at m/z 252.05 should be found in the LC-MS/MS mass spectra as is shown in Figure 5-13.

150

1) Rearrangement: -SO2(-63.9609)

2）M-H-SO2-C9H12 (-183.0474 )

C.I. Acid Blue 129

Figure 5-13. Fragmentation pathways of C.I. Acid Blue 129 under CID Voltage = 40 V. According to the LC-MS/MS analysis of C.I. Acid Blue 129, the fragmentation pattern agrees with the prediction. As is shown in Figure 5-14, the fragmentation of the Acid Blue

129 proved that a loss of SO2 is the predominant peak at m/z 371.1409 with the presence of aromatic ring attached to the secondary amine, and the second highest peak at m/z 252.0548 was formed by losing the aromatic substituent. Therefore, the proposed mechanisms and rule of fragmentation pattern are applicable to predict fragmentation and structure elucidation of the sulfonated anthraquinone dyes.

Figure 5-14. LC-MS/MS mass spectra of C.I. Acid Blue 129 under CID Voltage = 40 V.

151

5.4.Conclusions

In this study, a series of 1-amino anthraquinone-2-sulfonate acid dyes were studied with

HPLC-ESI-QTOF-MS/MS. A loss of SO2 was observed as a featured neutral loss in all the dyes studied. Fragmentation mechanisms were proposed and validated to predict fragmentation pattern of a dye with known structure. The Ortho-amino enables to form intramolecular hydrogen bonding to form a stable six-member ring to stabilize the molecule ions and facilitate a SO2 loss. The fragmentation pattern of sulfonated anthraquinone dyes

(relative peak abundances, m/z differences between peaks) is correlated to the substituents attached to the secondary amino groups: the side-chain groups attached to aromatic rings are fragmented the most easily, followed by alkyl substitution, loss of SO2 and aromatic substitution. The detailed understanding of fragmentation pattern and mechanisms are applicable for structural analysis of other dyes with featured structures.

152

Chapter 6 Photodegradation of C.I. Disperse Red 1

Abstract

Azo dyes account for the largest proportion and remain the most important class of synthetic dyes. A concern about azo dyes arising from the use of genotoxic compounds such as aromatic amines as intermediates. The main objective of this study is to quantitatively measuring the photodegradation products of a representative disperse azo dye, C.I. Disperse

Red 1, ((E)-2-(ethyl (4-((4-nitrophenyl) diazenyl) phenyl) amino) ethan-1-ol), in both dyed polyester fabrics and in the model solvent ethyl acetate, further to explore the mechanisms of photodegradation of disperse azo dyes with high resolution mass spectrometry analysis. The dye extracts from the degraded fabrics after weathering test were quantitatively measured by high performance liquid chromatography coupled to DAD and electrospray ionization- quadrupole time-of-flight mass spectrometry (HPLC-DAD-ESI-QTOF-MS) and auto tandem mass spectrometry (HPLC-DAD-ESI-QTOF-MS/MS) analysis. Three degradation products were structurally characterized and the photofading pathways were confirmed: the predominant product (E)-N-ethyl-4-((4-nitrophenyl) diacetyl) aniline was produced rapidly and then was decomposed to (E)-4-((4-nitrophenyl) diazenyl) aniline. The kinetics of ion generation and transformation depends on the environment.

153

6.1.Introduction Azo dyes accounts for largest proportion in textile manufacturing. An increasing concern about application of azo dyes focused the genotoxicity of aromatic amines that were used during manufacturing. Certain azo dyes produce genotoxic aromatic amines during metabolisms in different cell systems. Hydrophobic azo dyes such as disperse red 1 posed a carcinogenic risk after chemical oxidation and reduction. Natural faded dyes after long time exposure to UV irradiation or weathering conditions has the potential to produces products that is mutagenic or genotoxic compounds.

Photodegradation of dyes happens when photon of the light delivers energy from ultraviolet or visible light. Although extensive research has been done, not a definitive conclusion has been made so far on the mechanisms of photofading pathways. Two major theories explained the mechanisms as competing reactions of reductive and oxidative photofading: Early research 53 experimentally proved that photo reduction during photodegradation of azo disperse dyes in substrates (i.e. nylon, polyester), by absorbing hydrogen or losing certain number of methyl groups. 53 Later this photofading pathways were supported by Hashizume, but he also proposed reductive photo fading is existing as a competing reaction with the photooxidation, and a boundary exists between the two different reactions16 17 21a, 54. Later research using LC/MS and MALDI MS found that N-Demethylation and hydroxyization are the major reaction in oxidative photodegradation.19 However, no structural information has been confirmed regarding photo fading under different light sources.

High performance liquid chromatography equipped with mass spectrometry is a powerful tool that can be used for chemical separation, spectral detection, and allows structure elucidation. In this study, degradation products were detected in the photodegraded fabrics at

154

accelerated simulated natural fading weathering condition, as well as those from dye solvent in UV irradiation and regular lamp light sources, with the purpose of investigating photodegradation mechanisms of azo disperse dyes.

6.2.Experimental 6.2.1. Dyes and solvent

All the solvents (high purity for LC-MS grade) were purchased from Honeywell & Burdick

Jackson (Muskegon, MI, USA), and were filtered through 0.22 µm Millipore filters

(Whatman, GE Healthcare, UK) prior to use. Ammonium formate (>99%%, HPLC grade,

Fluka, Switzerland) and formic acid (~98%, MS grade) were purchased from Sigma Aldrich

(St Louis, MO, USA). Disperse red 1, namely (E)-2-(4-(ethyl (2-hydroxyethyl) amino) phenyl)-1-(4-nitrophenyl) diazen-1-ium was purchased from Sigma Aldrich (Figure 6-1).

Figure 6-1. Structure of C.I. Disperse Red 1 (Compound 1)

Swathes of polyester (Testfabrics. Polyester, #730) were used in the dyeing of samples. The dye bath consisted of deionized water stock solution and glacial acetate acid. The polyester samples were dyed in a Datacolor AHIBA Nuance Top Speed II dyeing machine (Figure 2-2) in a 200-mL dye bath. The dye bath was increased from room temperature to 130°C at a ramp rate of 4°C/min for 30 minutes.

155

6.2.2. Accelerated weathering of dyed polyester

Accelerated weathering of dyed polyester

Polyester fabrics (Testfabrics 700-13) containing 1% dye (weight/weight) together with blank samples were irradiated using Atlas to a simulated weathering condition (Phoenix,

Arizona (Latitude: 31 °N average high temperature in summer 39 °C. Average relative humidity (summer) 32%, with annual total radiation 8004 MJ/m2). The light spectra covering a wide range of spectrum (Figure 6-2). The weathering meter setting was: Rack panel temperature: ~43 °C, chamber temperature: 43 °C, and the relative humidity was 30%.

Exposure times up to 200 hours were employed.

Outdoor light 4

*(nm))

Irradiance (W/(m

0 300 400 500 600 700 800 Wavelength (nm)

Figure 6-2. Irradiance of light source of the simulated outdoor weathering condition

Photodegradation of C.I. Disperse Red 1 in ethyl acetate

About 1 mg/mL disperse red 1 solution was prepared in ethyl acetate (Fisher ACS grade) in

50 mL glass tubes. Dye solutions were exposed to a lamp light irradiation in S.N.E

Ultraviolet light box equipped with lamps PRBPR-5750 A, (Brafod. Ct. 06405, USA, light

156

spectra is shown in Figure 6-3). The other dye solution was put in a monochrome UV light box at 330 nm for up to 40 hours. Dye solutions exposed for different periods of irradiation were collected in HPLC-vials.

Irradiance 0.8

0.7

0.6

0.5

0.4

Irradiance 0.3

0.2

0.1

0.0 300 350 400 450 500 550 600 650 700 750 W avelength

Figure 6-3 Irradiance of regular lamp light spectra.

6.2.3. Dye extraction

Dyes together with the degradation products were isolated by 200 µL extraction solvent

(pyridine/H2O 4:3) in a sealed vial with a cone bottom and heated to 80 °C in a Pierce Reacti-

Therm Heating Module until they were discolored at approximately 15 min, followed by evaporation under 10 psi nitrogen gas. The dye residue was dissolved in 200 µL HPLC buffer (3% acetonitrile in H2O) and was filtered into Agilent HPLC vials using Millex-GV 13 mm, 0.22 µm PVDF filters prior to LC/MS analysis.

6.2.4. UV irradiation of dye solvent under regular lamp light and UV light (330 nm)

1 mg/mL disperse red 1 was diluted in ethyl acetate (Fisher ACS grade) in 50 mL glass tubes, followed by lamp light irradiation in S.N.E Ultraviolet light box equipped with UV lamps

157

PRBPR-5750 A, (Brafod. Ct. 06405, USA), and UV light at 330 nm. 1 mL dye solutions were collected into HPLC vials after UV irradiation from 1 hour to 160 hours.

6.2.5. HPLC-DAD-MS analysis of dye extract

Fabrics exposed to UV Irradiation were extracted using pyridine/water 4:3 at 90 °C for 20 min. The dye extract were reconstructed using acetonitrile and filtered using 0.22 µm PVDF

Filter. Then each dye extract were separated through Agilent Technologies 1260 SL high performance liquid chromatography at 45 °C. The chromatographic separation was performed with an Agilent ZORBAX 2.1mm x 50 mm narrow bore column C18 column. The mobile phase consists of (A) Milli-Q H2O containing 0.1% formic acid and mobile phase B consisting of acetonitrile 0.1% Formic Acid at a flow rate of 0.5 mL/min.

Table 6-1 HPLC gradient for C.I. Disperse Red 1

Time (min) Mobile phase B% (Acetonitrile) Flow rate (ml/min) 0 30 0.5 2 60 0.5 10 70 0.5 18 70 0.5 18.5 30 0.5

Accurate mass analysis of commercial dyes/dye extracts was performed using Agilent

Technologies 6520 Accurate-Mass Quadrupole Time-of-Flight mass spectrometer (QTOF-

158

6.2.6. LC-Auto MS/MS analysis of degradation products

With the aim to obtain preliminary information of the degradation products, the dye extracts where scanned by LC-MS/MS with a collision-induced-disassociation (CID) voltage at 20V and 30V with the following setting: Ion abundance >1000 count will be selected as parent ions for fragmentation, isolation width : Narrow (~1.3 amu), Delta Mass 100 ppm.

6.2.7. Synthesis of standard compounds for degradation products

1.21 gv(5 m mol) purified Disperse Orange 3 was added to 0.687g (5.5m mol) 2-bromoetha-

1-ol and 20 mL pyridine in a 50 mL beaker. The reaction was kept in room temperature and last for 24 hours. Products were filtered and dried in heating oven for 24 hours.

2.67g 4-nitroaniline was added into 1 mol NaNO2 and 8mL HCl in ice bath for 2 hours. The products were further mixed with 1.22 g (10 m mol) N-ethylaniline for a coupling reaction that lasts for 24 hours. The resultant products were further filtered and dried overnight.

Figure 6-4 Synthesis of Standard Compounds

159

6.3.Results and Discussion 6.3.1. Spectrophotometric Test

Relative Dye concentration

1.0

0.8

0.6

0.4

0.2

Relativeconcentration Dye 0.0 0 50 100 150 200 Time

Figure 6-5. K/S value of Photodegradation of Dye 1 K/S value is proportional to dye concentration present in dyed fabrics. To measure the fading kinetics so C.I. Direct red 1, K/S value of degraded fabrics after UV irradiation was measured by spectrophotometer. Results indicate the relative dye concentration on polyester drops rapidly to less than 10% within 200 hours at the outdoor conditions. These results suggest that the C.I. Disperse red 1 was decomposed upon weathering test.

6.3.2. Characterization of synthesized compounds

Thin layer chromatography analysis of synthesized compounds Products out of organic synthesis were firstly analyzed via thin layer chromatography using a elution solvent (Methanol/Ethyl acetate 1:1). The distance traveled by the products relative to the distance traveled by the solvent front can be used to identify relative polarities between components. The retention factor is expressed as Rf value.

160

TLC # Expected structure

1 Disperse Red 1 (Rf = 0.33)

2,3 standard compound 4 (Rf = 0.53)

4 standard compound 2 (Rf = 0.087)

5 standard compound 3 (Rf = 0.65) 1 2 3 4 5

Figure 6-6 Thin layer chromatograms of synthesized products.

HPLC analysis of synthesized compounds Synthesized compounds as well as dye solution sample (dye solution that was exposed to ramp light for 20 hours) were analyzed by HPLC-MS anlaysis. The expected ions were examined in extracted ion chromatograms. It shows in Figure 6-7,Figure 6-8 and Figure 6-9 that the standard compounds and degradation products have almost identical retention behavior and featured fragmentation patterns (Table 6-2). Therefore, the structure of degradation products are identical to the synthesized compounds.

161

Table 6-2. HPLC-MS/MS analysis of synthesized standard compounds and degraded sample (after 20 hours)

RT Parent ion Fragment Fragment Fragment (min) (m/z) 1(m/z) 2(m/z) 3(m/z) Compound 2 2.416 287.1127 122.0212 106.0659 241.1180 Sample 2.415 287.1134 122.0213 106.0633 241.1172 Error (ppm) N/A 1.04 0.82 24.51 3.32 Compound 3 4.676 271.1191 122.0218 106.0638 225.1148 Sample 4.704 287.1181 122.0223 106.0630 225.1180 Error (ppm) N/A 1.11 4.10 1.89 1.42 Compound 4 2.973 243.0869 122.0224 197.0938 106.0511

Sample 2.972 287.0866 122.0217 197.0930 106.0510

Difference 0.001 0.0003 0.0007 0.0008 0.0001

Error (ppm) N/A 1.23 5.74 4.06 0.94

Standard Impurity

EIC of Standard

EIC of Degradation product

Figure 6-7 HPLC-DAD-MS analysis of synthesized compound 2 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours.

162

DAD-Standard

EIC of Standard Structural isomers

Structure isomers EIC of Degradation product

Figure 6-8 HPLC-DAD-MS analysis of synthesized compound 3 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours.

DAD-Standard

EIC of Standard

Degradation product

Figure 6-9 HPLC-DAD-MS analysis of synthesized compound 4 as well as the degradation products out of degraded C.I. Disperse Red 1 after 20 hours.

163

6.3.3. Screening degradation product candidates with LC-auto MS/MS analysis

ESI-Auto MS/MS automatically isolate ions that have ion abundance higher than certain value to be fragmented. In our study, any compounds present in the dye extract solution were analyzed via LC-ESI Auto MS/MS when their ion abundances are higher than 1000. The featured fragments out of the analysis were compared with that of Disperse Red 1, whose featured fragmentation ions were observed at m/z= 106.0651, 122.0225, 134.0957, and

255.0865 (Figure 6-10). Compounds sharing fragmentation pattern or the original dye were considered as potential degradation products.

Figure 6-10. LC-MS/MS analysis of Disperse Red 1 extract out of unexposed fabrics The photodegradation products detected by the auto-MS/MS analysis differs were different from that in GC-MS reported previously. Three major compounds were detected in both polyester and ethyl acetate at m/z= 243.0877, m/z= 271.1190, and m/z= 287.1139, as are demonstrated in Figure 6-11.

164

Compound 2 Compound 3

Compound 4

Figure 6-11. Proposed structure of three degradation products

Compound 2 at m/z = 287.11 was observed in both polyester and ethyl acetate despite of different light sources. They are sharing the fragments at m/z 106.0619 and 122.0237 with

Disperse Red 1 (Figure 6-12).

Figure 6-12. Compound 2 at m/z=287.1139 out of photodegradation of Disperse Red 1 in ethyl acetate

165

Compound 3 at m/z= 271.1191 was observed to be the most abundant ion is sharing part of the featured fragment ions at m/z 106.0651 and m/z 122.0237 were the dominant fragment ions.

a b

c d

Figure 6-13. Compound 3 at m/z=271.1191 out of photodegradation of Disperse Red 1 in ethyl acetate. LC-MS/MS Mass spectra of compound 2 after 5 hours (a), 10 hours (b), 40 hours (c) and 80 hours (d). The featured fragment ions of Compound 4 at m/z=243.0877 were sharing fragment ions with other two compounds and the Disperse Red 1 at m/z= 106.0651, 122.0237. We propose that the compound 4 may originate from loss of ethyl group from compound 2 or loss of ethyl hydroxyl group from compound 3.

166

a b

c d

Figure 6-14. Fragmentation pattern of compound 4 (m/z=243.0877) out of photodegradation of Disperse Red 1 in ethyl acetate. LC-MS/MS Mass spectra of compound 1 after 5 hours (a), 10 hours (b), 40 hours (c) and 80 hours (d).

The proposed degradation products differ with that in previous study: compounds out of loss of ethyl hydroxyl group or ethyl group attached to the amine were the dominant products, rather than azo-cleavage or reductive hydrogen absorption 16-17, 53. With the aim to monitor the generation process of each compounds during UV irradiation time, the peak areas of extracted chromatogram of the above three compounds were compared in the degraded fabrics with LC-MS analysis.

6.3.4. Confirmation of degradation structure via LC-MS/MS analysis

In order to confirm the identity of Compound 2-4, synthesis of their authentic compounds were performed. Its retention time (Rt) registered on chromatograms with DAD detection, as well as MS/MS spectrum, were identical with those of compound 2-4.

167

a b

c d

e f

Figure 6-15. MS/MS Mass spectra of authentic samples and degradation products, (a) Compound 2 in dye solution after 20 hour-regular lamp irradiation, (b) Authentic sample for compound 2, (c) Compound 3 in dye solution after 20 hour-regular lamp irradiation, (d) Authentic sample for compound 3. (e) Compound 4 in dye solution after 20 hour-regular lamp irradiation, (f) Authentic sample for compound 4.

Other than the three major compounds discussed above, small amount of two new products were found in dye solution after UV light irradiation at 330 nm. Two compounds together with their proposed structure were demonstrated in Figure 6-16. Relative ion abundances of compound 5 and compound 6 increased as UV irradiation time.

168

Figure 6-16 LC-MS/MS mass spectra of Compound 5 (top)and Compound 6 (bottom) found in dye solution after UV irradiation at 330 nm.

6.3.5. LC-MS analysis of the degradation products in ethyl acetate

HPLC equipped with photodiode array detection provide the second dimensional discriminable value by providing full UV-visible spectra of each analyte. In Figure 6-17 it

169

was observed that the UV-vis spectra of the three compounds did not shift significantly compared to the original dye in dye extract (Figure 6-17 a) and ethyl acetate (Figure 6-17 b).

λmax= 500 nm a λmax= 500 nm b

Figure 6-17. Comparison of the UV-vis spectra of Disperse Red 1 and their degradation products in the solvent (a) and the dye extract (b). The highest peak absorption shows disperse red 1 at RT= 3.601 min. The low abundance shows the UV-Vis spectra of degradation products at RT= 1.897 min, and RT= 2.674 min. The integrated peak area of Extracted Ion Chromatogram (EIC) of ions of interest during liquid chromatography is proportional to its concentration. The EIC peak area of the three compounds after UV irradiation compared upon integration (Figure 6-18).

Figure 6-18. Extracted Ion Chromatogram (EIC) of degradation products after 80 hours: compound 1 (a) at 2.760 min, compound 2 at 5.36 min (b) and compound 3 at 2.005 min (c).

170

Cp2(F) Cp2(S) 1.0E8 Cp3(F) Cp3(S) Cp4(F) 1.00E+008 Cp4(S)

8.0E7 8.00E+007

6.0E7 6.00E+007

PKA 4.0E7 PKA 4.00E+007

2.00E+007 2.0E7

0.00E+000 0.0 0 50 100 150 200 0 20 40 60 80 100 120 140 160 180 Time (Hour) Time (Hour)

(a) (b)

cp2 m/z=285.10 1E8 cp3 m/z=299.10 cp4 7E7 9E7

8E7 6E7

7E7 5E7 6E7 4E7 5E7

PKA 4E7 3E7 PKA of EIC 3E7 2E7 2E7 1E7 1E7

0 0 0 10 20 30 40 0 10 20 30 40 UV 330 nm irradiation time (hours) UV Irradiation time (hours)

I (d)

Figure 6-19. Relative ion abundances of Cps 2-4 out of the photodegradation of Disperse Red 1, (a) Cps 2-4detected in ethyl acetate in lamp light, (b) compounds in ethyl acetate after irradiation with Regular lamp light, (c) Cp2-4 major compounds in UV wavelength of 330 nm, (d) Cp5-6 in UV irradiation at 330 nm. The degradation products of three compounds in different environments: polyesters after simulated outdoor weathering (Figure 6-19 (a)), solution in in lamp light (Figure 6-19(b) and

UV lamp at 330 nm (Figure 6-19 (c)). In the polyester upon simulated outdoor conditions

171

(Figure 6-19 (a)), the compound 3 were the dominant products which decreases linearly as

UV irradiation, the compound 4 increases at a low concentration and reach to maximum at

200 hours, and Compound 2 did not show significant change.

For dyes dissolved in ethyl acetate that exposed to lamp light (Figure 6-19(b)), both compound 3 and compound 4 increased with the same rate from 0 to 40 hours. The compound 3 concentration decreased after 40 hours, rather the compound 4 increased continuously. Results indicate that compound 4 are produced from loss of ethyl group from compound 3.

However, the generation rate of compound 3 in the ethyl acetate is significantly higher than compound 4 in UV lamp (Figure 6-19 (c)). The concentration of Compound 4 increased rapidly until 80 hours and decomposed. Based on the generation kinetics of the three compounds, the photodegradation routes were proposed (Figure 6-20).

Route 2(a)

cp2

Route 3 (a) Route 3 (b) Route 1 (a) cp5 cp6

Route 1 (b)

cp3 cp4

Figure 6-20. Proposed photodegradation pathways of C.I. Disperse Red 1

172

Two major pathways during the UV induced photodegradation: Disperse red 1 was decomposed to compound 3 by losing an ethyl hydroxyl group (Figure 6-20 (route 1 (a)).

Compound 3 further lost an ethyl group to generate compound 4 (Figure 6-20 (Route 1(b)).

Competing route 2(a) are formation of compound 2 by losing ethyl group, followed by losing an ethyl hydroxyl group to form compound 4 (Figure 6-20 (Route 2(b)). Our findings differ with previous study.

Relative abundance of the compound 2, 3, 4 to the total ions (sum of peak areas of compound

1, 2 and 3) were compared in Figure 6-21. It is shown that the relative abundance of compound 2 increases dramatically in the ethyl acetate, which means that the photodegradation rate in the ethyl acetate in the route 1(b) or route 2 (b) is faster than the fabrics. This is because of Disperse Red 1 in the ethyl acetate is evenly distributed in the solution and absorbs more light.

Compound 3 increased as the same rate with the decreases of Compound 2, suggesting the compound 2 was generated from decomposition of compound 1. However, the ratio of compound 1 in the fabrics was significantly higher than that in the ethyl acetate, suggesting the simulated outdoor condition facilitate formation of compound 3. The ethyl acetate, on the other hand, facilitates compound 4 formation. The difference could be caused by 30% humidity and comparatively higher ratio of oxygen that can facilitate loss of ethyl group.

173

Cp3(F) a Cp2(F) 1.0 b Cp3(S) 0.09 Cp 2(S) 0.9

0.8 0.08 0.7

0.6 0.07 0.5

0.4 0.06 0.3

Relativeionabundance 0.2

Relative ion abundance ion Relative 0.05

0.1

0.04 0.0 0 50 100 150 200 0 50 100 150 200 Time (Hours) Time (Hours)

Cp2(F) Cp4(F) c 1.0 d Cp3(F) 0.5 Cp 4(S) Cp4(F) Cp 2(S) Cp3(S) 0.4 Cp 4(S)

0.3

0.5

0.2

0.1

Relative ion abundance ion Relative Relativeionabundance

0.0 0.0

0 50 100 150 200 0 50 100 150 200 Time (Hours) Time (Hours)

Figure 6-21. Relative Ion abundance of degradation products in polyester and ethyl acetate: The production curve of compound 2(b), compound 3(b) compound 4(c) and a comparison of ion production in fabrics and solvent (d).

6.4. Conclusions In this study, the photodegradation products were quantitatively analyzed using LC-ESI Auto

MS/MS analysis, and the UV induced photodegradation mechanism was proposed. (E)-N- ethyl-4-((4-nitrophenyl) diazenyl) aniline (Compound 3) and (E)-4-((4-nitrophenyl) diazenyl) aniline (compound 4) were produced as a major photodegradation products. UV light at 330 nm provides higher energy to accelerate the photodegradation process. The different

174

photodegradation trends of disperse red 1 in the polyester and the ethyl acetate are caused by the environment.

175

Chapter 7 TOF-SIMS Method Development And Identification of Acid

Dyes in Nylon Fibers

Chuanzhen Zhou,1 Min Li,2 Roberto Garcia,1 Anne, Fraser,2 Keith Beck,2 David, Hinks2*,

Dieter P. Griffis1,3 *

1Analytical Instrumentation Facility, North Carolina State University, Campus Box 7531

Room 318 MRC, 2410 Campus Shore Dr., Raleigh, NC 27695

2Department of Textile Engineering, Chemistry and Science, College of Textiles, North

Carolina State University, Raleigh, NC 27695

3Department of Materials Science and Engineering, North Carolina State University, Campus

Box 7907, Raleigh, NC 27695

ABSTRACT:

A minimally destructive technique for determination of dyes in finished fibers provides an important tool for crime scene and other forensic investigations. The analytical power and the minimal sample consumption of Time-of-Flight Secondary Ion Mass Spectrometric (ToF-

SIMS) analysis provide the ability to obtain definitive molecular and elemental information relevant to fiber identification, including identification of dyes, from a very small volume of

* Corresponding Author Dieter Griffis: Email: [email protected]; Ph: 919 515 2128; Fax:

919 515 6965

* Corresponding Author David Hinks: Email: [email protected]; Ph: 919 2448582; Fax: 919

515 6965

176

sample. For both fiber surface analysis and, with the aid of cryomicrotomy, fiber cross section analysis, ToF-SIMS was used to identify various dyes in finished textile fibers. The analysis of Acid Blue 25 in nylon is presented as a representative example. The molecular ion of Acid Blue 25 with higher than 3% on weight-of-fiber (owf) dye loading can be easily

+ identified on dyed nylon surfaces by ToF-SIMS using Bi3 primary ion beam. With the use of

C60 to remove surface contamination, it is possible to identify the Acid Blue 25 with only

0.1% owf dye loadings on nylon surfaces. For cross sections of dyed nylon fibers, the molecular information of the dye disappears from spectra acquired after 200 frames of imaging acquisition at 1 shot/pixel for 256 x 256 pixels in a 100 µm x 100 µm rastered area

+ presumably as a result of Bi ion beam damage. When C60 sputtering is employed to remove the Bi ion beam induced damaged material, it is possible to partially recover the molecular

+ information of the dye. If C60 is used for damage removal in a cyclic manner to remove damage and acquire data and the spectra acquired are subsequently summed, the signal to noise ratio is greatly improved and provides the ability to unambiguously identify Acid Blue

25 via its molecular ion at a concentration of 0.1% owf.

177

7.1.Introduction

Analysis for finished fibers that essentially has an insignificant negative impact on the preservation of trace evidence is significant in crime scene and other forensic investigations.[reference] However, it is presently limited by the availability of robust techniques that can bear with only nano- or micro-level destruction of the fiber. One of the most important characteristics for fiber comparisons, fiber color, reflects the dyes and pigments used on the fabric and so provides dye identity. Commonly used forensic examination of dyed fibers includes thin layer chromatography (TLC),55 polarized light microscopy (PLM),56 UV-VIS microspectrophotometry (MSP),57 and FTIR58. Nevertheless, these methods have problems of destruction of fiber evidence, low resolution and lack of reproducibility. ToF-SIMS has several advantages for such dyed fiber analysis. ToF-SIMS can analyze the dyed fiber surface directly, which avoids the need for dye extraction and therefore maximally preserve the fiber evidence.59 With the aid of cryomicrotomy, ToF-

SIMS can be used to examine the cross section of the dyed fibers with only micro-level destruction of the fiber.60 Moreover, high resolution ToF-SIMS images showing the spatial distribution of the dye and the fibers can be obtained simultaneously.59b Combined LC TOF61 and ToF-SIMS results will be used to generate a mass spectrometric dye data base for use by law enforcement and other agencies.

ToF-SIMS has been successfully employed in forensic science such as in differentiating various types of gunpowder,62 analyzing hair samples for the presence of key drugs,63 fingerprints,64 a series of colored inks65 and sealing-inks.66 However, identification of dyes in a very low concentration such as 0.1 % dye on weight of fabric (owf) has been hampered by contamination on the fabric surfaces and the limited number of dye molecules under static 178

SIMS regime. The contamination on fabric surfaces may result from the dying process, samples handling and transferring and thus masks dye information. During ToF-SIMS spectral and imaging data acquisition, the finite lifetime of the dye molecular ions resulted from their destruction by the high energy Bi ion bombardment limits the number of obtainable molecular ions of the dyes and thus limits signal to noise.

+ Previous work in our lab developed an analytical protocol based on alternating C60

+ sputtering and Bi3 spectral imaging to improve the signal to noise required to spatially resolve different type of lignins within a cell wall.67 In this study, this protocol is revised and then employed to remove both contamination on fabric surfaces and Bi ion beam induced

+ + damaged material. C60 sputtering beam and Bi3 analysis ion beam are used in a cyclic manner to remove damage and acquire data and the spectra are subsequently summed. The

+ C60 sputtering dose is optimized to maximally recover the molecular information of the dye.

The signal to noise ratio is greatly improved and provides the ability to unambiguously identify Acid Blue 25 via its molecular ion at a concentration of 0.1% owf.

7.2.Experimental Section

7.2.1. Nylon dyeing procedure

A stock solution (1 g/L) of Acid Blue 25 was prepared. A dyebath was prepared (300 mL) that included stock solution, water, and acetic acid (20 µL) where the dye concentration in the dyebath depends on the desired percent on-weight-of-fiber (% owf). 0.1%, 0.5% and 1% owf

Acid Blue 25 dyed nylon were prepared for this study. nylon 6.6 (~3 g) was dyed in a

Roaches Pyrotec machine at 2°C/min from room temperature to 100 °C and the temperature was kept at 100 °C for 60 minutes.

179

7.2.2. TOF SIMS Sample Preparation Procedure

Dye powders and swatches of dyed nylon fabrics were pressed using a hydraulic press to achieve relatively flat surfaces for efficient secondary ion extraction and better mass resolution. The cross-sections of dyed nylon fibers were produced by cryomicrotomy and were 500-700 nm in thickness. After several attempts of embedding in Spurr’s and

Eponate™ 12 -Araldite 502 (Ted Pella), it was found that the fibers sectioned better in

Eponate™ 12 -Araldite 502 due to better surface adhesion and similar cutting properties to the nylon fibers. The formulation used is listed in Table 7-1.

Table 7-1 Formulation used to embed nylon fibers

Constituent Weight (g) Eponate 3.05 Araldite 502 1.76 Dodecenyl Succinic Anhydride (DDSA) 5.5 N,N-Benzyldimethylamine (BDMA) 0.3

Samples were embedded in polyethylene molds that were rinsed with hexane to remove any chemicals, especially PDMS, from the surface. The curing routine used was 12hrs at

45°C, 24hrs at 60°C then 8hrs at 70°C followed by an oven cool down. Sections were then cut on a Leica UC7 with cryo-attachment using a 45° cryo-diamond knife. The cryo temperatures prevented smearing of chemical species that are more mobile at room temperature. Cutting temperatures was optimized to be at -40°C where the sections are with little curling and no smearing of the surface. Colder temperatures resulted in excessive curling of the sections while warmer temperatures produced smearing of chemicals on the surface. Sections were produced in the range of 500 nm-700 nm thick depending on the 180

cutting characteristic of the sample section. In many cases a section that has many fibers needed to be thicker to prevent pull out of the fibers. Sections were excised from the diamond knife with an eyelash brush and placed onto a drop of water on a clean Si section. The surface tension of the water stretches the sections to flatten them as the water droplet evaporates. The water was allowed to evaporate and the sections were stored in Fluoroware® containers until needed.

7.3.ToF-SIMS Analysis

ToF-SIMS analyses in this study were conducted using a ToF-SIMS V (ION TOF, Inc.

m+ Chestnut Ridge, NY) instrument equipped with a Bin (n = 1 - 5, m = 1, 2) liquid metal ion gun and a C60 ion gun. The Bi ion optics column is oriented 45º from the surface normal. The instrument vacuum system consists of a load lock for rapid sample loading and an analysis chamber separated by the gate valve. The analysis chamber pressure is maintained below 5.0 x 10-9 mbar to avoid contamination of the surfaces to be analyzed. For ToF-SIMS spectrum and image acquisition, a finely focused, pulsed primary ion beam is rastered across the surface of the sample and the secondary ions emitted at each irradiated point or pixel are extracted into a time of flight mass spectrometer, mass filtered, and counted. In this way, an image with sub micrometer (< 0.3µm) spatial resolution can be acquired with a full mass spectrum for each pixel (spectrum image).

+ High mass resolution spectra were acquired using a 25 keV Bi3 liquid metal ion source at a current of 0.25 pA, with a pulse width of less than 1.0 ns under high-current bunched conditions. Acquisition times and areas varied. Secondary ions were extracted into a ToF mass spectrometer with 10 keV post acceleration to improve detection sensitivity. The combination of primary ion pulse width used and the ToF analyzer tuning provides a mass 181

+ resolution of approximately 4000 – 5000 m/m at m/z 29 (C2H5 ). 100 µm by 100 µm area,

+ 256 by 256 pixel spectrum images were acquired using a 25 keV Bi3 liquid metal ion source at a current of 0.15 pA, with a pulse width of 100 ns. A low energy electron flood gun (20 eV) was used to prevent charge buildup on the insulting sample surfaces. The improvement

+ + in signal-to-noise was achieved using dual beam analysis with C60 sputtering and Bi3

+ acquisition. The C60 sputtering current was kept at 1.0 nA. The sputtering area is 300 µm by

+ + 300 µm. The C60 sputtering time and Bi3 acquisition time vary on dyed nylon fiber surface and dyed nylon cross sections.

7.4.Results and Discussion

7.4.1. ToF-SIMS analysis of C.I. Acid Blue 25 dyed nylon fabric surface

Figure 7-1 presents the chemical structure of Acid Blue 25 and negative ion TOF SIMS spectra showing this dye’s molecular ion from Acid Blue 25 dye powder, from a 1% owf dyed nylon fiber surface and from a 1% owf dyed nylon fiber cross section, respectively.

While Acid Blue 25 can be easily detected by ToF-SIMS from the dye powder and dyed nylon cross sections (Figure 7-1b and Figure 7-1d), this dye is at best barely detectable (very poor signal to noise) from the dyed nylon surface (Figure 7-1c). This lack of sensitivity and very low molecular ion intensity for Acid Blue 25 may be due to surface contamination on the dyed nylon surface resulting from the dying process, subsequent handling or other causes.

182

a) O NH2 SO3H

Mw: 394.063 O HN

24000 b) 393, [M-H] Dye Powder 18000

12000

6000

Intensity (a.u.)

0 390 392 394 396 398 m/z, Negative Ion

c) 80 Dyed Nylon Surface 60

Intensity (a.u.)

0 390 392 394 396 398 m/z, Negative Ion

6000 d) Dyed Nylon Cross Section 4500

3000

1500

Intensity (a.u.)

0 390 392 394 396 398 m/z, Negative Ion

Figure 7-1. (a) Chemical structure and molecular weight of Acid Blue 25 (AB25). Negative ion ToF-SIMS spectra showing the molecular ion of AB25 obtained from (b) dye powder, (c) 1% on-weight-of-fabric (owf) dyed nylon surface, and (d) 1% owf dyed nylon cross section.

183

Dye molecular ion signal to noise and thus detection limit on dyed nylon surface was

+ greatly improved via the use of a C60 ion beam to remove any surface contamination. The

+ analysis protocol involves sputtering the dyed nylon surface for 14.6s with 1 nA C60 ion

+ beam followed by spectrum acquisition with Bi3 at 1 shot/pixel for 20 frames. The sputtering area and analysis area is 300 µm x 300 µm and 100 µm x 100 µm, respectively. This sequence is then repeated until a sufficient signal-to-noise ratio is achieved. Figure 7-2 presents the negative ion ToF-SIMS spectra obtained from 1% owf Acid Blue 25 dyed nylon

+ surface prior to (Figure 7-2a) and after various C60 sputtering times (Figure 7-2through 2e).

+ Note that prior to C60 ion beam sputtering (Figure 7-2a) there is no molecular information

+ indicating the presence of Acid Blue 25, but that after 14.6s C60 ion beam sputtering (Figure

7-2b), which equals 1 x 1014 ions/cm2, the molecular ion of Acid Blue 25 is apparent. After

+ further C60 ion beam sputtering times (Figure 7-2c through Figure 7-2e), the molecular ion

+ intensity of Acid Blue 25 is maintained, indicating that C60 ion beam sputtering does not

+ destroy the molecular information of the dye, but rather that C60 ion beam sputtering removes surface contamination. The spectrum in Figure 2f is the sum of all spectra acquired during the entire data acquisition sequence. This summation spectrum illustrates the improvement in signal-to-noise obtained using this process for analysis of Acid Blue 25 from a dyed nylon surface. The implication of this ToF-SIMS method in forensic science is significant because it provides the ability to analyze small areas of fabric surface even if surfaces are contaminated as is likely the case for most forensic evidence.

184

a) b) c)

0 0 0 x10 x10 x10 at surface at 14.6s C60 sputtering at 219s C60 sputtering 3.0 100 counts 5.0 291 counts 6.0 497 counts 5.0 2.5 4.0 4.0

Intensity (counts)Intensity Intensity (counts)Intensity 2.0 3.0 (counts)Intensity 3.0 1.5 2.0 2.0

392.0 394.0 392.0 394.0 392.0 394.0 Mass (u) Mass (u) Mass (u)

d) e) f)

0 2 Accumulated spectrum at 438s C60 sputtering x10 at 1022s C60 sputtering x10 527 counts 5.0 245 counts 26060 counts 1.5 4.0

1.0 Intensity (counts)Intensity 3.0 (counts)Intensity

2.0 0.5

392.0 394.0 392.0 394.0 Mass (u) Mass (u)

Figure 7-2. Reconstructed negative ion ToF-SIMS spectra obtained from 1% owf Acid Blue + 25 dyed nylon surface with C60 ion beam sputtering at varied time. Spectra (a – e) are reconstructed from 20 frames of 128 x 128 pixels, 1 shot/pixel acquisition. (a) at surface, 0s 14 -2 C60 sputtering, (b) at 14.6s C60 sputtering (PIDD 1 x 10 cm ), (c) at 219s C60 sputtering 15 -2 15 -2 (PIDD 1.5 x 10 cm ), (d) at 438s C60 sputtering (PIDD 3.0 x 10 cm ) and (e) at 1022s C60 sputtering (PIDD 7.0 x 1015 cm-2). (f) Reconstructed accumulated spectrum from 1400 frames of 128 x 128 pixels, 1 shot/pixel acquisition obtained from Acid Blue 25 dyed nylon surface with 1022s C60 sputtering.

+ To further demonstrate the surface contamination removal ability of C60 ion beam,

two experiments were performed on a 0.5% owf dyed nylon surface. The negative ion ToF-

+ SIMS spectra were acquired from the dyed nylon surface with (1) only Bi3 ion beam and (2)

+ + 14.6s C60 ion beam sputtering at 1 nA followed by spectrum acquisition with Bi3 at 1

185

+ shot/pixel for 20 frames in a cyclic manner. The accumulated Bi3 acquisition time is 600 frames for both experiments. The resulted accumulated spectra were compared in Figure 7-3a

+ and 3b. In the spectrum acquired without C60 ion beam sputtering, it is difficult to unambiguously identify the presence of the dye due to the poor signal to noise (around 4) and various mass interferences originated from the surface contamination. It is apparently that the

+ spectrum obtained with C60 ion beam sputtering is notably clean: free of mass interferences in the molecular ion mass region, suggesting the removal of surface contamination. The background noise level is greatly reduced and the signal-to-noise ratio is considerably

+ improved to be around 20, 5 times higher than that acquired without C60 sputtering.

186

a) 400 0.5% wof

without C sputtering 300 60

200

100

Intensity (a.u.)

0 392 393 394 395 396 m/z, Negative Ion

b) 600 0.5% wof

with C sputtering 450 60

300

150

Intensity (a.u.)

0 392 393 394 395 396 m/z, Negative Ion

Figure 7-3. Negative ion ToF-SIMS spectrum obtained from 0.5% owf Acid Blue 25 dyed nylon surface (a) without C60 sputtering, and (b) with C60 sputtering for 14.6s at 1 nA followed + by spectrum acquisition with Bi3 at 1 shot/pixel for 20 frames and this cycle was repeated for 30 times. Both spectra (a) and (b) are reconstructed from 600 frames of 128 x 128 pixels, 1 shot/pixel acquisition. Addition tests of this ToF- SIMS analytical approach have been performed on 0.1% owf Acid Blue 25 dyed nylon surfaces. Figure 7-4 presents the negative ion ToF-SIMS

+ spectrum obtained from 0.1% owf dyed nylon surface acquired utilizing C60 ion beam sputtering. It can be clearly seen that Acid Blue 25 in nylon can be detected with a detection limit of as low as 0.1% owf. As the dye molecule is barely observable from 1% owf dyed

187

+ nylon surface when C60 ion beam is not applied, the detection limit of Acid Blue 25 in nylon

+ surfaces has been improved at least 10 times with the aid of C60 ion beam sputtering.

60 0.1% wof With C sputtering 45 60

Intensity (a.u.)

0 392 393 394 395 396 m/z, Negative Ion

Figure 7-4. Negative ion ToF-SIMS spectrum obtained from 0.1% owf Acid Blue 25 dyed + nylon surface with C60 ion beam sputtering acquired under the same conditions as in Figure 3b.

High Spatial Resolution Imaging of the Dyed Nylon Cross Section

ToF-SIMS has been employed to map the spatial distribution of acid dyes in the dyed nylon fiber cross sections. As can be seen in Figure 7-5, the nylon fiber, the embedding resin

(required for microtoming of cross sections), and the Acid Blue 25 dye in the cross sections can be unambiguously differentiated by ToF-SIMS. Secondary ions attributed to the Acid

Blue 25 dye clearly coincide with the position of the nylon fibers. ToF-SIMS secondary ion images of the resin and the molecular ion of Acid Blue 25 dye in Figure 7-5 are color encoded and then overlaid shown in Figure 7-5d. The blue encoded molecular ion of Acid

Blue 25 overlaid on the red encoded resin provides clear delineation of the dye in the nylon fibers. 188

120 a) b) 10 100 8 80 6 60 4 40

20 2

0 0 m/z 26 CN– (Nylon) m/z 393 [M-H]– (Dye)

60 c) d) 50

10 20 µm 0 – Overlaid image m/z 71 C3H3O2 Resin Resin (red), Dye (blue)

Figure 7-5. ToF-SIMS images (100 µm x 100 µm) of a 1% owf Acid Blue 25 dyed nylon fiber cross section showing the spatial distribution of (a) CN-, the characteristic ion of nylon, (b) - molecular ion of Acid Blue 25, (c) C3H3O2 , the characteristic ion of the embedding resin, and (d) overlaid image of resin (in red) and the Acid Blue 25 molecular ion (in blue). The images are reconstructed from 200 frames of 256 x 256 pixels, 1 shot/pixel acquisition. While it has been demonstrated that ToF-SIMS can identify and detect 1% owf acid dye from nylon fiber cross sections, the goal of this study is to increase the detection limit of acid dyes. Figure 7-6 presents the ToF-SIMS images of nylon and Acid Blue 25 dye of 0.1%

+ owf dyed nylon cross sections acquired with Bi3 for 600 frames of 1 shot/pixel at 256 x 256 pixels. Note that the molecular ion of Acid Blue 25 has very low intensity and thus can barely be differentiated from background signals. The intensity of the dye molecular ion decreased 60% when reaching the static SIMS limit (around 200 frames of acquisition),

+ indicating the destruction of the molecular structure of the dye by Bi3 bombardment resulting in low signal-to-noise.

189

m/z 26 CN–

20 µm m/z 393 [M-H]–

Figure 7-6. ToF-SIMS images (100 µm X 100 µm, 256 x 256 pixels, 1 shot/pixel) of 0.1% owf + Acid Blue 25 dyed nylon cross section acquired with Bi3 beam only. The total acquisition was 600 frames. To improve signal to noise and thus detection limit in these cross sections, the

+ analytical approach using C60 ion beam sputtering on dyed nylon surface has been applied as follows. A 100 µm x 100 µm area, 256 x 256 pixels ToF-SIMS image was acquired on a 1%

+ owf Acid Blue 25 dyed nylon cross section with one Bi3 primary ion pulse per pixel for 200 frames. The reconstructed image of the molecular ion of Acid Blue 25 from the first 50 image frames is shown in Figure 7-7a. The ToF-SIMS image in Figure 7-7b was acquired in

+ the same position with 50 frames of one Bi3 primary ion pulse per pixel, but shows the resulting data after a total of 200 frames. It can be clearly seen that, the molecular

190

+ information of Acid Blue 25 disappeared after 200 frames of Bi3 acquisition, suggesting the

+ + dye molecule was significantly destroyed by Bi3 bombardment. The C60 ion beam was then

+ employed to sputter the previously analyzed area for 60s followed by Bi3 image acquisition using the same data acquisition conditions as in Figure 7-7b and the reconstructed image is

+ + shown in Figure 7-7c. Clearly, the C60 sputtering reduced the Bi3 ion beam induced molecular damage and thus it was possible to partially recover Acid Blue 25 dye molecular information.

More cycles of sputtering and image acquisition were conducted to determine the

+ optimum C60 sputtering dose needed to provide maximum signal recovery. Each cycle

+ includes 60s C60 ion beam sputtered on the previously analyzed area and 50 frames of one

+ Bi3 primary ion pulse per pixel acquisition. Figure 7-7c to 7h presents the ToF-SIMS images of the molecular ion of Acid Blue 25 acquired on the same spot from each cycle. The

+ + accumulated C60 sputtering time was shown in the figures. It appears that 180s C60 sputtering (1.25 x 1015 ions/cm2), is needed to recover the maximum molecular ion signal.

+ + Note that C60 sputtering was interfered by two 50 frames of Bi3 image acquisition, which

+ may induce further damages on the acid dye. The actual optimum C60 sputtering dose may be higher. Since all images were acquired on the same spot, they can be reconstructed and summed. Clearly, the resulted summed image provides a significantly higher signal-to-noise ratio. It has also been determined that the nylon fiber cross section was sputtered away after

+ around 400s C60 sputtering. Given the fact that the thickness of the cross sections are around

+ 600 nm, the sputtering rate of nylon fiber with C60 is around 1.5 nm/s at 1 nA current in a

300 µm x 300 µm sputtered area.

191

192

+ + surface using a cyclic Bi3 data acquisition/C60 damage removal protocol: a ToF-SIMS

+ image was acquired with Bi3 for 100 frames at 1 pulse/pixel, the surface of the dyed nylon cross section was sputtered for 73 seconds using C60, and this sequence was then repeated 6

+ times. The optimum C60 sputtering time was not used because the nylon fiber will be

+ sputtered away after 2 cycles of sputtering. The C60 sputtering time in this cyclic protocol was carefully selected to produce maximum molecular ion signals. Figure 7-8 presents the

ToF-SIMS images acquired with the above described method. The improvement in signal to

193

noise provided by this protocol is sufficient to allow unambiguous identification of Acid

Blue 25 via its molecular ion at a concentration of 0.1% owf.

The greatly improved detection limit of the acid dye in nylon using ToF-SIMS was made

+ possible by the use of cyclic cleaning and damage removal protocol employing C60 ion beam sputtering. In the identification of acid dye from the nylon surface, C60 was used to sputter clean the nylon fabric surface to remove surface contamination and to expose the dye. In the identification of acid dye from the nylon cross section, C60 was used to sputter away at least a portion of the surface layer damaged by the Bi ion beam during image acquisition. This sputtering was sufficient to expose relatively undamaged dye molecules below the Bi induced damaged layer, making adequate numbers of molecular ion of the acid dye to provide a meaningful improvement in signal to noise. This combination of damage removal and signal to noise improvement makes it possible to unambiguous identify acid blue 25 from both nylon surface and cross section at a concentration as low as 0.1% owf.

7.5.Conclusions

+ C60 ion beam was successfully employed to remove both the surface contamination as well as to partially remove Bi induced damaged materials resulting from imaging acquisition, leading to significant improvement on the detection limit of acid dye in nylon fabrics. With

+ the use of C60 , the detection limit of the Acid Blue 25 from the nylon surface is improved 30 times from 3% owf to 0.1% owf. For the dyed nylon cross section, the detection limit is

+ improved to 0.1% owf as well. The capability of C60 ion source to at least partially remove

Bi induced damage provided the ability to clearly identify the Acid Blue 25 via the molecular

15 2 ion ToF-SIMS image. Optimized C60 sputtering dose (1.25 x 10 ions/cm ) to maximally 194

recover the molecular ion signal of Acid Blue 25 has been determined using dyed nylon cross section.

ACKNOWLEDGMENT: This research was sponsored by the National Institute of Justice under Grant # 147528 and made possible in part with support by the North Carolina State

University Analytical Instrumentation Facility.

7.6.References

1. Wiggins, K. G.; Holness, J. A., Sci. Justice 2005, 45, 93-96.

2. Palenik, S., Microscopical examination of fibres. In: Robertoson, J., Grieve, M. (eds).

Forensic examination of fibres. 2nd ed.; CRC: New York, 1999.

3. Grieve, M. C.; Biermann, T. W.; Davingnon, M., Sci. Justice 2003, 43, 5-22.

4. Biermann, T. W., Sci. Justice 2007, 47, 68-87.

5. Kirkbride, K.; Tungol, M., Infrared microspectroscopy of fibres. Forensic examination of fibres. 2nd ed.; CRC: New York, 1999.

6. Cho, L. L.; Reffner, J. A.; Gatewood, B. M.; Wetzel, D. L., Journal. Forensci

Science. 2001, 46, 1309-1314.

7. Lee, Y.; Lee, J.; Kim, Y.; Choi, S.; Ham, S.; Kim, K., Applied. Surface. Science.

2008, 255, 1033-1036.

8. Belu, A.; Graham, D.; Castner, D., Biomaterials 2003, 24, 3635-3653.

9. Mitchell, R.; Carr, C.; Parfitt, M.; Vickerman, J.; Jones, C., Cellulose 2005, 12, 629-

639.

10. Soltzberg, L. J., J. Am. Soc. Mass Spec. 2007, 18, 2001-2006.

195

11. Hellsing, M.; Fokine, M.; Claesson, A.; Nilsson, L.-E.; Margulis, W., Appl. Surf. Sci.

2003, 203-204, 648-651.

12. Petrick, L. M.; Wilson, T. A.; Fawcett, W. R., J. of Forensic Sci. 2006, 51, (4), 771-

779.

13. Huang, M.; Yinon, J.; Sigman, M. E., J. of Forensic Sci. 2004, 49, (2), 238-249.

14. Mahoney, C. M.; Gillen, G.; Fahey, A. J., Forensic Sci. Int. 2006, 158, 39-51.

15. Audinot, J.-N.; Yegles, M.; Labarthe, A.; Ruch, D.; Wennig, R.; Migeon, H.-N., Appl.

Surf. Sci. 2003, 203-204, 1718-1721.

16. Chen, B.-J.; Lee, P.-L.; Chen, W.-Y.; Mai, F.-D.; Ling, Y.-C., Applied. Surface.

Science. 2006, 252, 6786-6788.

17. Szynkowska, M. I.; Czerski, K.; Rogowski, J.; Paryjczak, T.; Parczewski, A.,

Forensic Sci. Int. 2009, 184, e24-e26.

18. He, A.; Karpuzov, D.; Xu, S., Surface. Interface Analysis. 2006, 38, 854-858.

19. Lee, J.; Lee, C.; Lee, K.; Lee, Y., Appl. Surf. Science. 2008, 255, 1523-1526.

20. Zhou, C.; Li, Q.; Chiang, V. L.; Lucia, L. A.; Griffis, D. P., Analytical. Chemistry.

2011, 83, 7020-7026.

196