The Development of Nanomaterials and “Green” Methods for Separation Science

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Michael C. Beilke

Graduate Program in Chemistry

The Ohio State University

2015

Dissertation Committee:

Dr. Susan V. Olesik, Advisor

Dr. Heather Allen

Dr. Vicki Wysocki

Dr. Abigail Shoben

Michael C. Beilke

2015

Abstract

A primary focus of current separations research is directed toward the reduction of both the diameter and particle size distribution of the material utilized as a stationary phase. The work reported herein follows a common theme. Research is focused on novel approaches for the application of electrospun nanomaterials or the search for improved efficiency in separation science.

Electrospinning is a cost-effective and simple technique that relies on repulsive electrostatic forces to generate nanofibers from a conductive polymeric solution.

Electrospun nanofibers have proven to be an effective stationary phase in ultra-thin layer chromatography (UTLC), giving more efficient separations in shorter analysis time than traditional particle-based stationary phases. This technology was further enhanced by aligning the nanofibrous mats in a single direction. Aligned electrospun UTLC (AE-

UTLC) devices showed improved performance relative to non-aligned electrospun (E-

UTLC) phases, demonstrating higher separation efficiency and reduced time of analysis.

A major disadvantage of conventional TLC analysis is that the mobile phase velocity decreases with increasing separation distance. Here, the chromatographic performance of electrospun (UTLC) stationary phases were explored with induced forced-flow of mobile phase across the stationary phase with applied potential. This type of forced-flow is used in planar electrochromatography (PEC). Compared to UTLC, improved efficiency resulted from analytes with greater migration distance. ii

Utilization of nanofibers to provide a co-reactant electrochemiluminescent determination for nucleobases was examined. Nafion, a cation-exchange polymer becomes electrospun with the aid of a second polymer, poly(acrylic) acid (PAA). Good linear agreement between concentration and the evolution of electrochemiluminescent signal for guanine solutions are demonstrated.

A “green” hydrophilic interaction chromatography (HILIC), a liquid-liquid partition mechanism, method for separating mixtures with broad ranges in polarities is explored using enhanced-fluidity mobile phases. Under HILIC conditions, analytes elute with increasing polarity. Enhanced-fluidity liquid chromatography (EFLC) involves the addition of liquefied gas to conventional liquid mobile phases. The liquefied gas provides greater diffusivity and lower viscosity character to the mobile phase. The impact of carbon dioxide addition to a methanol:water mobile phase was studied to optimize HILIC conditions. Additionally, the buffer type, pH, and ionic strength were adjusted to achieve optimal chromatographic performance. For the first time a separation of 16 ribonucleic acid (RNA) nucleosides/nucleotides was achieved in 16 minutes with greater than 1.3 resolution for all analyte pairs. An optimized separation using carbon dioxide:methanol:water mobile phase was compared to methanol:water and acetonitrile:water mobile phases. Based on chromatographic performance parameters

(efficiency, resolution, and speed of analysis) and the environmental impact of the mobile phase mixtures, carbon dioxide:methanol:water mixtures are preferred over acetonitrile:water or methanol:water mobile phases for the separation of mixtures of nucleosides and nucleotides. The separation of 16 nucleosides and nucleotides,

iii representing a large group of compounds with wide ranging polarities, is taken as an example to assess the usefulness of EFL-HILIC. Addition of gradient elution conditions were also explored to provide reduced analysis time for the wide ranging polar mixture.

Dedication

This document is dedicated to my beautiful wife and our sugar bear along with family lost along my graduate school journey. Grandpa, I miss you and can’t wait to see you

again. Aliyah, I will always love you.

Acknowledgments

I would like to acknowledge all members of the Olesik research group both past and present for their support both professionally and personally. I would like to especially recognize Cherie (Owens) Pomeranz, Toni Newsome, Martin “The Mick”

Beres, and Joe Zewe for helping me get through graduate school. Chemistry talk with

Hui Wang during “tea time” helped to not only mold me into a better chemist but also a better person.

I wish to extend a special thanks to my research advisor; Dr. Susan Olesik, for her guidance. Thank you for exemplifying what is takes to always forge ahead in your pursuits despite any perceived obstacles encountered along the way.

Lastly, I could never have made it this far in school without my family. Mom you are my rock, I am me because of you, thank you. Christina, you my dear sister are special. You have always been there, no matter what, no questions asked, thank you.

Granny, I love watching baseball and chatting sports with you (yes he did play for Green

Bay 5 year ago), thank you. Dad, your stories are ridiculous. The burger was worth it.

The screen was tiny. My cribbage board was finer, thank you. Gary, you have showed me, unwillingly as it may be, how to be responsible and to do things right the first time, thank you.

Vita

2005...... B.S. Chemistry, Missouri Southern State

University, Joplin, MO

2007...... M.S. Chemistry, Saint Louis University, St.

Louis, MO

2008...... Analytical Chemist, Aerotek, Chesterfield,

2009-2010 ...... Graduate Teaching Assistant, Department of

Chemistry and Biochemistry, The Ohio

State University, Columbus, OH

2010-2014 ...... Nanoscale Science and Engineering Fellow,

The Ohio State University, Columbus, OH

2014-Present ...... Graduate Research Assistant, The Ohio

State University, Columbus, OH

Publications

M.C. Beilke, M.J. Beres, S.V. Olesik, Gradient enhanced-fluidity liquid hydrophilic interaction chromatography of RNA nucleosides and nucleotides: a “green” technique, J. Chromatogr. A, (submitted 2015).

M.C. Beilke, J.W. Zewe, J.E. Clark, S.V. Olesik, Aligned electrospun nanofibers for ultra-thin layer chromatography, Anal. Chim. Acta, (2013), 761, 201-208. vii

M.C. Beilke, T.L. Klotzbach, B.L. Treu, D. Sokic-Lazic, R.L. Arechedarra, J. Wildrick, M.J. Moehlenbrock, M. Germain, S.D. Minteer, “Enzymatic biofuel cells,” in Micro Fuel Cells, Burlington, Mass., Elsevier Inc., 2009, pp. 179-241.

M.C. Beilke, S.D. Minteer, Immobilization of glycolysis enzymes in hydrophobically modified Nafion, PMSE Preprints, (2006), 94, 556-557.

Presentations

M.C. Beilke, S.V. Olesik, Electrospun silica nanoparticle/PVP nanofiber mat as a planar electrochromatography stationary phase, The Pittsburgh Conference 2015, New Orleans, LA.

M.C. Beilke, S.V. Olesik, Development of electrochemiluminescent electrospun nanofibers, The Pittsburgh Conference 2014, Chicago, IL.

M.C. Beilke, S.V. Olesik, Characterization and implementation of ion exchange electrospun nanofibers for nucleic acid detection, The Pittsburgh Conference 2012, Orlando, FL.

M.C. Beilke, S.V. Olesik, DNA detection using electrochemiluminescence from electropun nanofibers, The Pittsburgh Conference 2011, Atlanta, GA.

M.C. Beilke, S.D. Minteer, Glycolysis biomimic in hydrophobically modified Nafion, The Pittsburgh Conference 2007, Chicago, IL.

M.C. Beilke, D. Sokic-Lazic, S.D. Minteer, Enzymatic biomimics for biofuel cell application, The Fuel Cell Seminar 2006, Honolulu, HI.

M.C. Beilke, S.D. Minteer, Immobilization of glycolysis enzymes in hydrophobically modified Nafion, ACS National Meeting 2006, Atlanta, GA.

Fields of Study

Major Field: Chemistry

viii

Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments...... vi

Vita ...... vii

Publications ...... vii

Presentations ...... viii

Table of Contents ...... ix

List of Tables ...... xiv

List of Figures ...... xvi

CHAPTER 1: INTRODUCTION ...... 1

1.1 Overview ...... 1

1.2 Electrospinning ...... 12

1.3 Aligned Ultra-Thin Layer Chromatography ...... 15

1.4 Planar Electrochromatography ...... 15

1.5 Sensing ...... 16

1.6 “Green” Separations with Enhanced-fluidity Liquid Gradients ...... 17

1.7 Research Focus ...... 18

1.8 References ...... 19

CHAPTER 2: ALIGNED ELECTRSPUN NANOFIBERS FOR ULTRATHIN-LAYER

CHROMATOGRAPHY ...... 23

2.1 Introduction ...... 23

2.2 Experimental ...... 28

2.2.1 Reagents...... 28

2.2.2 Instrumentation ...... 28

2.2.3 Aligned Electrospinning ...... 29

2.2.4 Thin-layer Chromatography ...... 31

2.3 Results and Discussion ...... 32

2.3.1 Optimization of Aligned Electrospun PAN Nanofibers ...... 32

2.3.2 Initial Separation: Laser Dye Analysis ...... 42

2.3.3 Separation of β-Blockers and Steroidal Compounds ...... 44

2.3.4 Mobile Phase Velocity...... 54

2.4 Conclusions ...... 59

CHAPTER 3: PLANAR ELECTROCHEMISTRY (PEC) SEPARATION OF LASER

DYES WITH ELECTROSPUN NANOFIBERS ...... 64

3.1 Introduction ...... 64

3.2 Experimental ...... 67

3.2.1 Reagents...... 67

3.2.2 Instrumentation ...... 70

3.2.3 Electrospinning ...... 72

3.2.4 Planar Electrochromatography ...... 72

3.3 Results and Discussion ...... 74

3.3.1 Mobile Phase Optimization ...... 74

3.3.2 Efficiency ...... 82

3.3.3 UTLC Comparison ...... 90

3.4 Conclusions ...... 93

3.5 References ...... 93

CHAPTER 4: CHEMILUMINESCENT NANOFIBER CHARACTERIZATION AND

DETERMINATION OF NUCELOBASE CONCENTRATIONS ...... 98

4.1 Introduction ...... 98

4.2 Experimental ...... 100

4.2.1 Reagents and Instrumentation ...... 100

4.2.2 Light Generating Pathway ...... 102

4.2.3 Electrospinning ...... 104

4.3 Results and Discussion ...... 104

4.3.1 Optimization of Electrospun Nanofibers ...... 105

4.3.2 Modified GCE Surface ...... 114

4.3.3 Guanine Determination...... 118

4.4 Conclusions ...... 120

CHAPTER 5: “GREEN” SEPARATIONS THROUGH ENHANCED-FLUIDITY

LIQUID CHROMATOGRAPHY GRADIENTS ...... 124

5.1 Introduction ...... 124

5.2 Experimental ...... 127

5.2.1 Instrumental ...... 127

5.2.2 Chemicals ...... 128

5.2.3 Mobile Phase Preparation ...... 132

5.2.4 Data Analysis ...... 132

5.3 Results and Discussion ...... 132

5.3.1 Traditional HILIC Mobile Phases ...... 132

5.3.2 Mobile Phase Optimization ...... 136

5.3.3 Effect of CO2 on Retention ...... 143

5.3.4 Gradient Optimization ...... 156

5.3.5 Optimized Chromatographic Comparison ...... 159

5.3.6 Efficiency and Resolution ...... 160

xii

5.3.7 Evaluation of Method “Greenness” ...... 172

5.4 Conclusions ...... 175

5.5 References ...... 175

CHAPTER 6: SUMMARY AND FUTURE WORK ...... 181

6.1 Summary of Research ...... 181

6.2 Future Work ...... 183

6.3 References ...... 184

Bibliography ...... 185

xiii

List of Tables

Table 2.1. Average diameter of aligned electrospun PAN nanofibers as a function of collector rotational speed...... 34

Table 2.2. Retardation factors Rf, of laser dyes for PAN AE-UTLC and E-UTLC plates.

...... 43

Table 2.3. Efficienceis, N, and retardation factors Rf, for the separation of acebutolol, cortisone, and propranolol on PAN AE-UTLC and E-UTLC plates using mobile phases of 40:60 and 30:70 (v:v%) chloroform:heptane with 1.0 wt% TBABr...... 51

Table 3.1. Analyte structure and pKa values for laser dyes...... 69

Table 3.2. Efficiency, N of PEC and UTLC comparison with 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA mobile phase with a 120 sec analysis time...... 89

Table 4.1. Nanofiber diameters of the observed mat containing a blend of PAA with

Nafion polymers...... 108

Table 4.2. Electrodes generated from electrospinning polymer solution at 0.50, 0.75, and

2+ 1.00 minute. Peak current, Ip determined for a 50 mM solution of Ru(bpy)3 ...... 117

Table 5.1. Name structure and log P values for analytes in their neutral unionized form calculated from ACD/ChemSketch at www.acdlabs.com/resources/freeware/index.php

...... 130

Table 5.2. Retention factors, k, during isocratic conditions for nucleosides (A,U,C,G), monophosphate nucleotides (AMP,UMP,CMP,GMP), diphosphate nucleotides xiv

(ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) under varying ACN:100 mM sodium phosphate (pH=2.65) and MeOH:400 mM sodium phosphate (pH=2.65) mobile phases...... 135

Table 5.3. Retention factors, k calculated for isocratic conditions for nucleosides

(A,U,C,G), monophosphate nucleotides (AMP,UMP,CMP,GMP), diphosphate nucleotides (ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) with 80:20 (v:v%) MeOH:H2O (40 mM sodium phosphate, pH=2.65) with CO2 addition.

...... 147

Table 5.4. S-values ranges for nucleosides (A,U,C,G), monophosphate nucleotides

(AMP,UMP,CMP,GMP), diphosphate nucleotides (ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) with varying mobile phases...... 151

Table 5.5. Mobile Phase composition and HPLC-EAT scores for optimized ACN:H2O,

MeOH:H2O, and CO2:MeOH:H2O separations. Optimized conditions are same as listed in Figure 5.9...... 174

List of Figures

Figure 1.1. Depiction of band broadening associated with the multiflow path term (A- term)...... 6

Figure 1.2. Depiction of band broadening due to the longitudinal diffusion term (B-term).

Two horizontal cross-sections with the corresponding solute concentration versus distance shown with an (1) early time and a (2) later time...... 7

Figure 1.3. Depiction of band broadening caused from mass transfer (C-term). (1) Ideal

Gaussian profiles for the solute in the mobile phase and in the stationary phase. (2,3) As the solute band moves a small distance down the column, an equilibrium between the two phases no longer exists. The red arrows show the movement of solute from the stationary phase to the mobile phase, and from the mobile phase to the stationary phase. (4) As equilibrium is reestablished, the solute band is now broader...... 8

Figure 1.4. Typical plot of the van Deemter equation showing the relationship between plate height (H) and mobile phase velocity (u)...... 9

Figure 1.5. Mobile phase flow profile with (1) pressure driven and (2) electroosmotic flow...... 10

Figure 1.6. Depiction of how resolution can be obtained from a chromatogram...... 11

Figure 1.7. An illustration of an electrospinning apparatus, consisting of (1) a syringe pump, (2) a syringe containing a conductive polymer solution, (3) a high voltage power supply, and a grounded conductive collector...... 13 xvi

Figure 1.8. Representative electrospun mats with (A) aligned and (B) random oriented nanofibers. (Scale bar = 10 micron) ...... 14

Figure 2.1. An illustration of the (A) parallel electrode and (B) rotating drum aligned electrospinning apparatuses...... 27

Figure 2.2. Rotating drum utilized in the generation of aligned electrospun PAN nanofibers. The rotating collector is contained within a Plexiglass box, which serves to isolate the electrospun nanofibers from the outside environment while acting as a safe guard against mandrel failure...... 30

Figure 2.3. SEM images illustrating aligned electrospun PAN nanofibers generated on the rotating collector at rotational speeds of (A) 500, (B) 750, (C) 1000, (D) 1250, and

(E) 1500 rpm. The white bar demonstrates the direction in which the drum was rotated; this also serves as the direction of nanofiber alignment...... 33

Figure 2.4. Percentage of electrospun PAN nanofibers positioned within a 10°, 20°, and

30° angle from the direction of nanofiber alignment for nanofibrous mats generated at

500 (■), 750 (■), 1000 (■), 1250 (■), and 1500 rpm (■)...... 36

Figure 2.5. Mat thickness as a function of electrospinning time for the generation of aligned electrospun PAN nanofibers...... 40

Figure 2.6. Mat thickness (A) and fiber morphology (B) of the aligned electrospun PAN nanofibers collected for 120 min at 1250 rpm...... 41

Figure 2.7. Retardation factors Rf, of acebutolol (♦), cortisone (■), and propranolol (▲) for PAN (A) AE-UTLC plates and (B) E-UTLC plates (n=5)...... 46

xvii

Figure 2.8. Time of analysis as a function of mobile phase composition for PAN AE-

UTLC (■) and E-UTLC (♦). The migration distance was 2.5 cm...... 47

Figure 2.9. Chromatograms for the separation of (1) acebutolol, (2) cortisone, and (3) propranolol on a PAN (A) AE-UTLC plate using a 40:60 (v:v%) chloroform:heptane mobile phase with 1.0 wt% TBABr and on a (B) E-UTLC plate using 30:70 (v:v%) chloroform:heptane mobile phase with 1.0 wt% TBABr...... 50

Figure 2.10. Change in efficiency, N (♦), of propranolol with increasing migration distance on a PAN AE-UTLC plate using a 40:60 (v:v%) chloroform:heptane mobile phase with 1.0 wt% TBABr...... 52

Figure 2.11. Comparison of mobile phase velocities of PAN AE-UTLC (■) plates, PAN

E-UTLC (♦) plates, and commercial cyano TLC (▲) plates using heptane as the mobile phase. (Error bars are contained within data points when not visible)...... 53

Figure 2.12. Comparison of mobile phase velocities of PAN AE-UTLC (■) plates, PAN

E-UTLC (♦) plates, and commercial cyano TLC (▲) plates using 50:50 (v:v%) acetone:water (A) and acetone (B) as the mobile phase. (Error bars are contained within data points when not visible)...... 58

Figure 3.1. PEC device. Components are listed (1) Anode, (2) Whatman wicks, (3)

Cathode, (4) Mobile phase reservoir, and (5) electrospun SiO2/PVP UTLC plate. 71

Figure 3.2. Initial spot widths of KR, R101, R590, R610, and SR deposited onto the

SiO2/PVP UTLC stationary phase with 100 µm Chemyx Nanojet (■), 250 µm hand- spotted (■), and 500 µm hand-spotted (■) method...... 77

xviii

Figure 3.3. Migration distance of laser dyes versus ACN and IPA concentrations: KR

(♦), R590 (▲), R101 (■), R610 (+), and SR (●). Conditions are 60 second analysis time, applied potential of 1000 V, 50% by volume 35 mM citrate buffer (pH 5.6) and 50% of combination of ACN and IPA...... 78

Figure 3.4. Migration distance of KR (♦), R590 (▲), R101 (■), R610 (+), and SR (●) for

60 second analysis time with 250, 500, and 750 V of applied potential. Mobile phase consists of 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA...... 79

Figure 3.5. Migration distance of KR (♦), R590 (▲), R101(■), R610 (+), and SR (●) with 750 V of applied potential for 60, 120, and 180 seconds. Mobile phase consists of

50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA...... 80

Figure 3.6. Migration distance of KR (♦), R590 (▲), R101 (■), R610 (+), and SR (●) with 35 mM citrate buffer (pH 4.4, 5.0, 5.6, and 6.2) with 750 V applied potential for 120 seconds. Mobile phase consists of 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer:ACN:IPA...... 81

Figure 3.7. Surface plot for KR reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■)

2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis...... 84

Figure 3.8. Surface plot for R101 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■)

2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z-

xix axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis...... 85

Figure 3.9. Surface plot for R590 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■)

Figure 3.10. Surface plot for R610 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■)

Figure 3.11. Surface plot for SR reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■)

Figure 3.12. Representation of (A) PEC (750 V applied potential) and (B) UTLC analysis on an SiO2/PVP plate with KR (●), R101 (●), R590 (●), R610 (●), and SR (●) using 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA as the mobile phase, under 120 second analysis time. Lower dotted line shows initial spotting area of

xx analytes. Upper dashed line shows solvent from migration of mobile phase for UTLC analysis...... 91

Figure 3.13. PEC separation of KR, R101, and R590, respectively. Chromatogram obtained with an applied voltage of 750, analysis time of 120 seconds, and a mobile phase of 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA. Note: analytes with lower retention travel further along the stationary phase...... 92

Figure 4.1. Four reaction co-reactant sequence for the generation of light. Step 1.

2+ Ru(bpy)3 becomes oxidized. Step 2. Guanine becomes oxidized. Step 3. Excited state

2+ formation. Step 4. Emission of light and regeneration of Ru(bpy)3 species...... 103

Figure 4.2. Representative SEM image of an electrospun nanofiber mat generated with

70:30 (w:w%) Nafion:PAA under 12.5 cm and 8 kV (Scale bar = 10 µm)...... 109

Figure 4.3. Representative SEM image of an electrospun nanofiber mat generated with

70:30 (w:w%) Nafion:PAA under 12.5 cm and 8 kV that was soaked in H2O for a period of three days (Scale bar = 10 µm)...... 110

Figure 4.4. Mat thickness plotted against electrospinning time (n=3) (R2 = 0.9850). ... 111

Figure 4.5. Mat thickness plotted against electrospinning time. Magnified view of lower electrospinning times (0.50, 0.75, and 1.00 minute)...... 112

Figure 4.6. Representative SEM image of optimized electrospun nanofibers with mat thickness generated with 180 minute electrospinning time (Scale bar = 25 µm)...... 113

Figure 4.7. Cyclic voltammogram collected with 100 mV/sec scan rate for a bare CGE electrode (■), TF Nafion:PAA (■), electrospun Nafion:PAA (■) with oxidation of 1.0 x

10-4 M guanine...... 116

xxi

Figure 4.8. Electrochemiluminescent signal obtained from (1.0 x 10-5 M, 1.0 x 10-6 M, and 1.0 x 10-7 M) guanine solutions...... 119

Figure 5.1. Effect of (A) MeOH and (B) ACN content on retention factor of A (♦), AMP

(■), ADP (▲), and ATP (+)...... 134

Figure 5.2. Chromatogram of 16 nucleoside/nucleotide analyte mixture (1.25 x 10-4 M) with 80:20 (v:v%) MeOH:H2O mobile phase. Gradient program used is (0-1.50 min)

70% B, (1.50-10.00 min) 70-90% B with 1.00 mL/min flow rate. The analyte key is (1)-

A, (2)-U, (3)-G, (4)-C, (5)-AMP, (6)-UMP, (7)-CMP, (8)-GMP, (9)-ADP, (10)-UDP,

(11)-CDP, (12)-GDP, (13)-ATP, (14)-UTP, (15)-CTP, (16)-GTP...... 137

Figure 5.3. Isocratic separation of A, AMP, ADP, and ATP analyte mixture (100 ppm) with 80:20 (v:v%) MeOH:aqueous buffer mobile phase. 10 mM ammonium phosphate buffer (pH=2.00)...... 138

Figure 5.4. Peak width for 16 nucleoside/nucleotide analyte mixture (100 ppm) with

80:20 (v:v%) MeOH:aqueous phosphate buffer at pH 2.00 (■), pH 2.65 (■), and pH 3.15

(■) in the mobile phase. Gradient program used is (0-1.00 min) 70% B, (1.00-6.67 min)

70-90% B with 1.50 mL/min flow rate...... 141

Figure 5.5. Peak width for 16 nucleoside/nucleotide analyte mixture (100 ppm) with

80:20 (v:v%) MeOH:aqueous phosphate buffer at a pH of 2.65 with an ionic strength of

10 mM (■), 25 mM (■), and 40 mM (■) in the mobile phase. Gradient program used is

(0-1.00 min) 70% B, (1.00-6.67 min) 70-90% B with 1.50 mL/min flow rate...... 142

xxii

Figure 5.6. Effect of CO2 content on retention factor of A (♦), AMP (■), ADP(▲), and

ATP (+). Mobile phase consisted of varying proportions of solvent A (CO2) and solvent

B (80:20 (v:v%) MeOH:H2O with 40 mM sodium phosphate buffer, pH=2.65)...... 146

Figure 5.7. Log k vs. φ plots for injection of 100 ppm of A (♦), AMP (■), ADP (▲), and

ATP (+) for ACN:H2O (A), MeOH:H2O (B), and CO2:MeOH:H2O (C) optimized mobile phases. Lines represent linear regressions, and S-values obtained from slopes of linear regression lines...... 148

Figure 5.8. Isoeleutropic nomogram comparing percentage of weak eluent needed to obtain identical solvent strength for (A) nucleosides, (B) monophosphate nucleotides, (C) diphosphate nucleotides, and (D) triphosphate nucleotides. The strong eluent (solvent B) is H2O for ACN and MeOH and 80:20 (v:v%) MeOH:H2O for CO2...... 152

Figure 5.9. Optimized separation of the 16 nucleoside/nucleotide analyte mixture (1.25 x

10-4 M) with (A) ACN:100 mM sodium phosphate (pH=2.65) (0-7.5 min, 70% ACN, 7.5-

15 min, 70-60% ACN), (B) MeOH:400 mM sodium phosphate (pH=2.65) (0-4 min, 90%

MeOH, 4-6 min, 90-80% MeOH), and (C) CO2:MeOH:H2O with gradient described in

Section 5.3.4. The analyte key is (1)-A, (2)-U, (3)-G, (4)-C, (5)-AMP, (6)-UMP, (7)-

CMP, (8)-GMP, (9)-ADP, (10)-UDP, (11)-CDP, (12)-GDP, (13)-ATP, (14)-UTP, (15)-

CTP, (16)-GTP. Mobile phase gradients are shown with dashed lines with addition of the instrument dwell time for each separation...... 158

Figure 5.10. Order of elution for each of the mobile phase system from least retained on top to most retained on bottom...... 162

xxiii

Figure 5.11. Pseudo van Deemter plots for 100 ppm of (A) A, (B) AMP, (C) ADP, and

(D) ATP for the CO2:MeOH:H2O (♦), MeOH:H2O (■), and ACN:H2O (▲) optimized mobile phases. Measured at various flow-rates under optimized mobile phase conditions in Figure 2.10. Lines represent linear regressions...... 163

Figure 5.12. Resolution, Rs of analyte pair at collected with 80:20 (v:v%) MeOH:H2O

(40 mM sodium phosphate, pH=2.65) with 0.50 mL/min (■), 1.00 mL/min (■), 1.50 mL/min (■), and 2.00 mL/min (■) flow-rates. The dashed line represent Rs=1.0...... 167

UMP, (7)-CMP, (8)-GMP, (9)-ADP, (10)-UDP, (11)-CDP, (12)-GDP, (13)-ATP, (14)-

UTP, (15)-CTP, (16)-GTP...... 168

xxiv

CHAPTER 1: INTRODUCTION

1.1 Overview

The drive to develop methods to increase efficiency in separation science is continual. Recently, the application of nanomaterials for use in chromatographic separations have been explored [1]. Nanomaterials offer ideal physical characteristics for use as a stationary phase in separation techniques. The potential for increases efficiency in chromatography is gained due to the nanomaterials size and relatively large surface area. Nevertheless, application of nanomaterials has proven difficult. The work reported herein describes the development of novel polymeric nanomaterials and their usage in chromatographic and sensing devices as applied to ultra-thin layer chromatography

(UTLC), planar electrochromatography (PEC), and modified glassy carbon electrode.

Advancements in enhanced-fluidity liquid chromatography (EFLC) is also detailed herein.

1.2 Basic Chromatography Fundamentals

M.S. Tswett is credited with the discovery and coining the term chromatography

[2]. However, chromatography did not become an established practice until the 1930s.

Today, chromatography takes many forms; though the essential governing characteristic remains the need for varying amounts of retention to occur with analyte mixtures.

Analyte separation is governed by differences in the relative distribution of analytes between the stationary phase and the mobile phase. This phenomenon is detailed in the partition coefficient (K) equation (Equation 1.1).

퐶 퐾 = 푠 (Eq. 1.1) 퐶푚 where Cs is the concentration of solute in the stationary phase and Cm is the concentration of solute in the mobile phase.

The relative amount of time that each analyte interacts with the stationary phase can be expressed with the individual retention factor (k) in Equation 1.2.

푡 ′ (푡 −푡 ) 푘 = 푅 = 푅 푀 (Eq. 1.2) 푡푀 푡푀 where tR’ is the adjusted retention time, tM is the time the analyte spends in the mobile phase, and tR is the time the analyte spend in the stationary phase.

The width of a chromatographic band (TLC) or peak is indicative of how much band broadening is present during separation. This band dispersion is typically described using plate height (H). From Equation 1.3, plate height is equal to the variance of the chromatographic band (σ2), divided by the length of the column.

휎2 퐻 = (Eq. 1.3) 퐿

The performance of a column is also often described by the number of theoretical plates which is described by Eqation 1.4; this equation also shows the relationship between plate height and plate number.

퐿 푁 = (Eq. 1.4) 퐻

2 where L is the distance the mobile phase travels during the separation. Ultimately, lower

H values result in a more efficient separation. Equation 1.5 (The van Deemter equation) can be used to determine the most efficient system.

퐵 퐻 = 퐴 + + (퐶 + 퐶 )푢 (Eq. 1.5) 푢 푠 푀 where u is linear velocity of mobile phase, A is the multipath flow term, B is the longitudinal diffusion term, and both Cs and CM are part of the mass transfer term.

A representation of band broadening effects from the multipath flow term is shown in Figure 1.1. The multipath flow term is proportional to particle diameter, because band broadening is determined from the distance all analyte molecules travel.

From the diagram the solute’s initial profile is rectangular. As the band travels, individual solute molecules travel along different paths. Three examples are shown, most molecules follow paths with similar lengths to the middle while few molecules travel shorter paths (bottom) and a few others follow a longer path (top) all of which end as a broader Gaussian shape.

The longitudinal diffusion term arises due to random molecular diffusion of the analytes. Solute molecules are constantly in motion, diffusing from regions of higher solute concentration to regions of lower solute concentration thus making the longitudinal diffusion term proportional to the rate of diffusion of the analyte in the mobile phase.

The longitudinal diffusion term is also inversely proportional to linear velocity of the mobile phase since this would be associated with a lessened amount of time that analytes have to diffuse before elution. A representative depiction of band broadening due to longitudinal diffusion is shown in Figure 1.2. 3

Figure 1.3 is a depiction of the mass transfer term. The mass transfer term is associated with an equilibrium between the mobile phase and stationary phase. Band broadening occurs when the movement of the analyte molecule within the mobile phase or within the stationary phase is not fast enough to maintain an equilibrium between the two phases. Band broadening here is proportional to the thickness of the stationary phase as well as the particle diameter. The mass transfer term is also proportional to the mobile phase velocity. It is also inversely proportional to the diffusion of the analyte in both the mobile phase and stationary phase.

Of course these three terms (A, B, and C) are additive and are typically shown in a van Deemter plot of plate height (H) against linear velocity (u) of the mobile phase. A typical van Deemter plot is shown is Figure 1.4. The optimal flow rate for the system can be garnered from this system at the local minimum for the HTotal curve. The individual terms are also shown in Figure 1.4. The A-term contribution is constant while the B-term decreases with flow rate. The C-term contribution is low at low flow rates and increases linearly with flow rate.

For band broadening, the linear velocity of the mobile phase is more important than the flow rate. The mobile phase flow profile and changes in local velocity depend on the driving forces used to maintain flow through the chromatographic system. These driving forces are capillary, pneumatic and electroosmotic forces. Capillary forces are responsible for the movement of mobile phase in planar chromatography. These forces are too weak to provide either an optimum or constant mobile phase velocity for separations using small particle stationary phases. Pressure driven transport of the

4 mobile phase is used in column chromatography. The mobile phase is externally pressurized and then driven through the column by the pressure gradient between the inlet and column exit. A parabolic profile (Figure 1.5-top) is generated because of the compressible mobile phase. From Figure 1.5 (top), a local velocity that varies with position is shown with decreasing flow resistance with migration along the column.

Electrosmotic flow is the source of flow in an electric field. At the column wall an electrical double layer forms from the adsorption of ions from the mobile phase. An excess of counter-ions is present in the double layer compared to the bulk liquid. In the presence of an electric field shearing of the solution occurs only within the thin diffuse part of the double layer transporting the mobile phase through the column with a plug profile (Figure 1.5-bottom). An advantage is gained with plug flow over parabolic flow because of minimizing contributions of the multiflow path term.

Finally, while the van Deemter curves describes the optimal conditions for band dispersion, another measurement must be used to better understand if a separation of an analyte mixture occurred. The resolution (Rs) of an elution is a quantitative measure of how well two neighboring anayte peaks can be differentiated in a chromatographic separation. Complete baseline resolution is achieved at 1.5. The resolution equation is shown in Equation 1.6.

2[(푡푅)퐵−(푡푅)퐴] 푅푠 = (Eq. 1.6) 푊퐵−푊퐴 where tR is the retention time for neighboring analyte peaks A and B. WA and WB are the peak widths measured at the baseline of the chromatogram. Figure 1.6 depicts a sample chromatogram with tR, WA, and WB clearly marked. 5

Figure 1.1. Depiction of band broadening associated with the multiflow path term (A- term).

1 2

Figure 1.2. Depiction of band broadening due to the longitudinal diffusion term (B- term). Two horizontal cross-sections with the corresponding solute concentration versus distance shown with an (1) early time and a (2) later time.

Figure 1.3. Depiction of band broadening caused from mass transfer (C-term). (1) Ideal Gaussian profiles for the solute in the mobile phase and in the stationary phase. (2, 3) As the solute band moves a small distance down the column, an equilibrium between the two phases no longer exists. The red arrows show the movement of solute from the stationary phase to the mobile phase, and from the mobile phase to the stationary phase. (4) As equilibrium is reestablished, the solute band is now broader.

0.7

0.6

0.5

0.4 H

Total

0.3

Cu H (µm) H 0.2 A 0.1 B/u 0

-0.1 0 2 4 6 8 u (mm/sec)

Figure 1.4. Typical plot of the van Deemter equation showing the relationship between plate height (H) and mobile phase velocity (u).

Figure 1.5. Mobile phase flow profile with (1) pressure driven and (2) electroosmotic flow.

Figure 1.6. Depiction of how resolution can be obtained from a chromatogram.

1.3 Electrospinning

Electrospinning is a strategy to produce a high surface area material from a conductive polymeric solution. Electrospinning is similar to electrospraying, though long semi-continuous fibers are generated instead of particles. Typically the generated fibers possess small diameters (100-1000 nm). A typical electrospinning apparatus is shown in

Figure 1.7. The spinneret solution, containing dissolved polymer, is forced out the end of a syringe. Nanofibers are produced as the applied voltage is increased passed a critical point (≥ 8 kV) whereby surface tension of the spinneret solution is overcome and becomes charged to form a Taylor cone. A charged jet of the material is emitted from the

Taylor cone and travels through space while whipping and splaying occurs toward the direction of a grounded collector [3,4]. Evaporation of the solvent, used to dissolve the polymer material, and formation of nanofibers occurs while traversing the void space [5].

Nanofibers are collected on a grounded conductive surface. Figure 1.8 shows representative electrospun material with aligned and randomly placed nanofibers.

Randomly placed nanofibers are typically generated with a stationary conductive collector. Aligned electrospun nanofibers are discussed in greater detail in Section 2.2.3.

4. Grounded Collector 2. Syringe 1. Syringe Pump

3. Power Supply

Figure 1.8. Representative electrospun mats with (A) aligned and (B) random oriented nanofibers. (Scale bar = 10 micron)

1.4 Aligned Ultra-Thin Layer Chromatography

Previous UTLC studies focused on electrospun mats with randomly oriented nanofibers. Aligned electrospun fibers have been utilized in a variety of applications, often demonstrating a profound impact due to the ordered structuring of the nanofibers

[4]. Aligned nanofibers have been utilized as cell scaffolds. With this fiber configuration, cells propagate in the direction of the nanofibers’ orientation [6]. It was hypothesized that UTLC with aligned electrospun nanofibers may show an increase in chromatographic performance relative to non-aligned electrospun UTLC plates.

A rotating drum apparatus is capable of giving large surface areas of aligned nanofibers [4,7,8]. Thick nanofibrous mats are also possible with the rotating drum apparatus. The rotational speed must be optimized to produce the highest alignment possible while preventing fiber breakage [4]. Due to the increased mat areas and thicknesses possible with the rotating drum, the aligned UTLC plates utilized in this study were fabricated using this method. Chapter 2 describes utilization of aligned electrospun nanomaterial as a UTLC device.

1.5 Planar Electrochromatography

Planar electrochromatography is a forced-flow technique that was developed to help overcome the poor flow profile associated with TLC analysis. Forced-flow occurs as the mobile phase is driven by an applied electric field. A major limitation of the technique is that the sorbent layer dries when there is excessive Joule heating. Joule heating will always occur during PEC, and will cause evaporation of mobile phase unless

15 the system is cooled or the environment in contact with the sorbent layer is fully saturated with vapor.

So far, in PEC analysis, commercial TLC plates have been utilized; however, the relatively large thicknesses of these plates make them ineffective at dissipating the Joule heat produced. Further development of PEC requires plates with thinner layers than commercial plates, which are more appropriate for heat dissipation [9,10]. Electrospun nanofiber stationary phases were recently introduced for UTLC [11]. The variable thickness of the electrospun stationary phase offers a promising solution to problems associated with Joule heating found in thicker layers currently in use. Chapter 3 described the utilization of a SiO2/PVP electrospun nanomaterial for use in PEC analysis.

1.6 Sensing

2+ The utilization of tris(2,2’bipyridyl)ruthenium(II) Ru(bpy)3 in an electrode platform is cost effective and sensitive [12,13]. Several efforts have been made to

2+ immobilize Ru(bpy)3 onto electrode surfaces [14,15]. However, these approaches tend to be either time consuming or involve complicated fabrication processes. So it is desirable to develop a facile and inexpensive approach for fabrication of an electrochemiluminescent sensor that possesses high surface area and can lead to large

2+ quantities of immobilized Ru(bpy)3 .

The electrospinning technique has proven to be a facile, versatile and a cost effective strategy for developing polymeric fibrous membranes with large surface areas

[16,17]. Examples of fibrous membranes that show great potential for sensors with high

A sensitivity are plentiful [18,19].

2+ A cation-exchanger, Nafion, is the immobilization material for Ru(bpy)3 . The

2+ sensing electrode is fabricated from Ru(bpy)3 /Nafion nanofibers by the electrospinning

technique. Nafion has been widely used in electrochemistry [20]. However, due to pure

Nafion’s intensive aggregates, its solution usually can only be electrosprayed into beads

rather than by electrospinning into nanofibers. The addition of a second polymer can

promote proper chain entanglement and electrospinning can occur [21]. Together the

2+ nanofibers with the combination of immobilized Ru(bpy)3 will create an

electrochemiluminescent sensing platform that would probably show a strong

electrochemiluminescent signal, and thus a high sensitivity in molecular detection. So a

great potential for electrospun Nafion nanofiber based electrochemiluminescent sensors

is offered.

1.7 “Green” Separations with Enhanced-fluidity Liquid Gradients

A current concern regarding high performance liquid chromatography (HPLC) is

its heavy use of toxic, organic solvents. The most direct method of “greening”

chromatography is simply to develop methods that utilize “green” solvents without

compromising chromatographic performance.

Hydrophilic interaction chromatography (HILIC) is a type of chromatography that

could benefit greatly from “greener” methods. Unfortunately, HILIC methods typically

rely heavily on acetonitrile (ACN) as a mobile phase component. ACN is one of the most

17 common solvents used throughout HPLC because of its highly desirable chromatographic properties. However, when compared to other commonly used chromatographic solvents, ACN is less environmentally friendly. Some attempts have been made to minimize/eliminate ACN from HILIC mobile phases for the purpose of “greener” chromatography. Recently, use of enhanced-fluidity liquids (EFLs) as “green” mobile phases in HILIC separations have been demonstrated [22,23]. These mobile phases are liquid mixtures (usually a primary alcohol and H2O) to which high proportions of a liquefied gas (usually CO2) are added [24]. By adding nonpolar CO2 as a modifier in

HILIC, the eluent strength of alcohol:H2O mobile phases can be decreased significantly, yielding retention similar to traditional ACN:H2O mobile phases. With addition of CO2, these mobile phases offer the chromatographic advantages similar to supercritical fluid mobile phases, including enhanced solute mass transfer and reduced system back pressure [25]. This often leads to improved chromatographic performance in terms of efficiency, resolution, and separation time as well as providing the analysis of polar compounds [22,26].

While the benefits of enhanced-fluidity liquid chromatography (EFLC) have been demonstrated for HILIC, they have only been demonstrated in isocratic separations.

However, mixtures containing analytes with a wide range in polarity often require gradient elution in order to resolve all peaks with a reasonable time of analysis. For the first time the optimized HILIC method uses a gradient program with EFL mobile phases.

1.8 Research Focus

The use of electrospun nanofibers as a stationary phase in UTLC and PEC analysis is extensively discussed in Chapters 2 and 3. Chapter 2 describes the first use of an aligned nanofibrous mat as a UTLC device. Chapter 3 details the advancement of

PEC separations with laser dyes as analytes. Chapter 4 details the study of an electrospun chemiluminescent nanofiber modified glassy carbon electrode for measuring nucleobase concentration in solution. Chapter 5 describes the advancements in a separation of widely polar analytes with gradients applied to enhanced-fluidity liquid hydrophilic interaction chromatography. Finally, Chapter 6 details future work for all reported research.

1.9 References

[1] Z. Zhang, Z. Wang, Z. Liao, H. Liu, Applications of nanomaterials in liquid

chromatography: opportunities for separation with high efficiency and selectivity,

J. Sep. Sci., (2006), 29, 1872-1878.

[2] L.S. Ettre, K.I Sakodynskii, M.S. Tswett and the discovery of chromatography I:

early work (1899-1903), Chromatographia, (1993), 35, 223-231.

[3] S. Ramakrishna, K. Fujihara, W.E. Teo, T.C. Lim, Z. Ma, An Introduction to

Electrospinning and Nanofibers, World Scientific, New York, NY, 2005.

[4] W.E. Teo, S. Ramakrishna, A review on electrospinning design and nanofiber

assemblies, Nanotechnology, (2006), 17, R89-R106.

[5] N. Bhardwaj, S.C. Kundu, Electrospinning: a fascinating fiber fabrication

technique, Biotechnol. Adv., (2010), 28. 325-347. 19

[6] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Aligned biodegradable nanofibrous

sturcture: a potential scaffold for blood vessel engineering, Biomater., (2004), 25,

877-886.

[7] S.F. Fennessey, R.J. Farris, Fabrication of aligned and molecularly oriented

electrospun polyacrylonitrile nanofibers and the mechanical behavior of their

twisted yarns, Polymer, (2004), 45, 4217-4225.

[8] W. Hu, Z.M. Huang, S.Y. Meng, C. He, Fabrication and characterization of

chitosan coated braided PLA wire using aligned electrospun fibers, J. Mater. Sci.:

Mater. Med., (2009), 20, 2275-2284.

[9] Sz. Nyiredy, in: Sz. Nyiredy (Ed.), Planar Chromatography: A Retrospective View

for the Third Millennium, Springer, Budapest, 2001, pp 177-199.

[10] L. Botz, S. Nagy, B. Kocsis, in: Sz. Nyiredy (Ed.), Planar Chromatography: A

Retrospective View for the Third Millennium, Springer, Budapest, 2001, pp 489-

516.

[11] J.E. Clark, S.V. Olesik, Technique for ultrathin layer chromatography using an

electrospun, nanofibrous stationary phase, Anal. Chem., (2009), 81, 4121-4229.

[12] Z.F. Ding, B.M. Quinn, S.K. Haram, L.E. Pell, B.A. Korgel, A.J. Bard,

Electrochemistry and electrogenerated chemiluminescence from silicon nanocrystal

quantum dots, Science, (2002), 296, 1293-1297.

[13] M.M. Richter, Electrochemiluminescence (ECL), Chem. Rev., (2004), 104, 3003-

3036.

[14] Z.H. Guo, Y. Shen, M.K. Wang, F. Zhao, S.J. Dong, Electrochemistry and

electrogenerated chemiluminescence of SiO2

nanoparticles/tris(2,2’bipyridyl)ruthenium(II) multilayer films on indium tin oxide

electrodes, Anal. Chem., (2004), 76, 184-191.

[15] S. Zanarini, E. Rampazzo, L. Della Ciana, M. Marcaccio, E. Marzocchi, M.

Montalti, F. Paolucci, L. Prodi, Ru(bpy)3 covalently doped silica nanoparticles as

multicenter tunable structures for electrochemiluminescence amplification, J. Am.

Chem. Soc., (2009), 131, 2260-2267.

[16] J. Yoon, S.K. Chae, J.M. Kim, Colorimetric sensors for volatile organic compounds

(VOCs) based on conjugated polymer-embedded electrospun fibers, J. Am. Chem.

Soc., (2007), 129, 3038-3039.

[17] X. Wang, C. Drew, S.H. Lee, K.J. Senecal, J. Kumar, L.A. Samuelson, Electrospun

nanofibrous membranes for highly sensitive optical sensors, Nano Lett., (2002), 2,

1273-1275.

[18] X.B. Yin, S.J. Don, E.I. Wang, Analytical applications of the

electrochemiluminescence of tris(2,2’-bipyridyl) ruthenium and its derivatives,

Trac-Trend Anal. Chem., (2004), 23, 432-441.

[19] F. Di Benedetto, A. Camposeo, S. Pagliara, E. Mele, L. Persano, R. Stabile, R.

Clingolani, D. Pilignano, Patterning of light-emitting conjugated polymer

nanofibers, Nat. Nanotiechnol., (2008), 3, 614-619.

[20] D.J. Kim, Y.K. Lyu, H.N. Choi, I.H. Min, W.Y. Lee, nafion-stabilized magnetic

2+ nanoparticles (Fe3O4) for [Ru(bpy)] (bpy=bipyridine) electrogenerated

chemiluminescence sensor, Chem. Commun., (2005), 23, 2966-2968.

[21] H. Chen, J.D. Snyder, Y.A. Elabd, Electrospinning and solution properties of

Nafion and poly(acrylic acid), Macromolecules, (2008), 41, 128-135.

[22] M.J. Beres, S.V. Olesik, Enhanced-fluidity liquid chromatography using mixed-

mode hydrophilic interaction chromatography/strong cation-exchange retention

mechanisms, J. Sep Sci., (2015), 38, 3119-3129.

[23] A. dos Santos Pereira, A.J. Giron, E. Admasu, P. Sandra, Green hydrophilic

interaction chromatography using ethanol-water-carbon dioxide mixtures, J. Sep.

Sci., (2010), 33, 834-837.

[24] J.C. Valette, C. Demesmay, J.L. Rocca, E. Verdon, Separation of tetracycline

antibiotics by hydrophilic interaction chromatography using an amino-propyl

stationary phase, Chromatographia, (2004), 59, 55-60.

[25] S.T. Lee, S.V. Olesik, Normal-phase high-performance liquid chromatography

using enhanced-liquid mobile phases, J. Chromatogr. A, (1995), 707, 217-224.

[26] G.S. Philibert, S.V. Olesik, Characterization of enhanced-fluidity liquid hydrophilic

interaction chromatography for the separation of nucleosides and nucleotides, J.

Chromatogr. A, (2011), 45, 8222-8230.

CHAPTER 2: ALIGNED ELECTRSPUN NANOFIBERS FOR ULTRATHIN- LAYER CHROMATOGRAPHY

2.1 Introduction

Originally developed in the 1950s, thin-layer chromatography (TLC), also called planar chromatography, is widely used today in environmental analysis, food, clinical, and pharmaceutical industries [1,2]. The stationary phase in TLC consists of sorbent particles that are attached to a solid support, typically an aluminum or glass plate. The analytes of interest are spotted directly onto the stationary phase, the edge of which is brought into contact with the mobile phase. The mobile phase proceeds up the plate via capillary action. Separation of analytes occurs due to interactions with both the mobile phase and the stationary phase. A number of different materials have been applied as the stationary phase in TLC, including alumina, cellulose, ion-exchange resins and, perhaps most commonly, silica gel [3,4,5,6]. More recently, nanostructured surfaces have been developed for TLC applications [7,8].

In 2001, ultra-thin layer chromatography (UTLC) was developed. This technology, which typically utilizes a stationary phase with a 5-20 µm thickness

(compared to 100-400 µm thick stationary phases for commercial TLC devices), represented a significant improvement in analysis time and sensitivity over traditional

TLC devices [4,9]. Our lab previously demonstrated that electrospinning can be utilized to generate a mat of nanofibers that can be used as a UTLC sorbent [10,11]. 23

Electrospinning is a simple and cost effective method of generating nanofibers by placing a high electric field between a syringe containing a polymeric solution and a conductive surface. At a critical voltage, the surface tension of the polymer solution at the syringe tip is overcome, and polymeric nanofibers are splayed from the droplet and are collected on the conductive surface [12]. To date, electrospinning has been used in a variety of applications including use in sensors [13], tissue scaffolds [14], and in solid phase microextraction [15,16]. Electrospun UTLC (E-UTLC) plates are particularly attractive for a number of reasons: no binder material is required in fabrication [2]; the solid support and mat thickness are readily variable; the small dimensions of the fibers provide a support with high surface area, and the chemical functionalities in the stationary phase can be easily modified [10,11]. Electrospun polyacrylonitrile (PAN)

UTLC plates not only demonstrated a decreased time of analysis, but also vastly superior separation efficiencies for steroidal compounds relative to a commercially available cyano phase TLC plate [10]. Cyano-modified phases are frequently utilized as a stationary phase in the separation of steroidal compounds, as well as alkaloids and derivitized amino acids [17]. The PAN E-UTLC plate demonstrated 500 times greater separation efficiency and a decreased time of analysis by up to 50% when compared to the conventional cyano TLC phase [10].

The previous E-UTLC studies used electrospun mats with randomly oriented nanofibers. The work reported herein marks the first use of aligned electrospun UTLC

(AE-UTLC). Aligned electrospun fibers have been utilized in a variety of applications, often demonstrating a profound impact due to the ordered structuring of the nanofibers

[18]. For example, aligned nanofibers have been utilized as cell scaffolds. With this fiber configuration, cells propagate in the direction of the nanofibers’ orientation [19]. It was hypothesized that UTLC with aligned electrospun nanofibers may show an increase in chromatographic performance relative to non-aligned electrospun UTLC plates.

While there are many different techniques that have been applied to generate aligned electrospun fibers [20], the two most commonly used models are: the parallel electrode and rotating drum models (Figure 2.1) [18]. While the parallel electrode model is both simple to use and capable of generating highly aligned nanofibers, the electrodes can only be placed about 1-5 cm apart, thus limiting the length of the nanofibrous mat that can be generated [18,21]. Furthermore, it is very difficult to obtain electrospun mats that are sufficiently thick to use as a chromatographic support; the wider the distance between the two electrodes the thinner the electrospun mat [18,22].

A rotating drum apparatus is capable of giving much larger surface areas of aligned nanofibers than the parallel electrode configuration [18,23,24]. Thicker nanofibrous mats are also possible with the rotating drum. However, it is more challenging to generate mats that are as highly aligned as those generated with parallel electrodes. Additionally, the rotational speed must be optimized to produce the highest alignment possible while preventing fiber breakage [18]. Due to the increased mat areas and thicknesses possible with the rotating drum, the AE-UTLC plates utilized in this study were fabricated using this method.

Herein, the use of a newly configured rotating device to generate PAN AE-UTLC plates is described. These plates were applied to the analysis of both laser dyes and a

25 mixture of β-blockers and steroidal compounds. The performance of the AE-UTLC plates was compared to that of the previously published PAN E-UTLC devices [11].

A B

Figure 2.1. An illustration of the (A) parallel electrode and (B) rotating drum aligned electrospinning apparatuses.

2.2 Experimental

2.2.1 Reagents

Polyacrylonitrile (PAN), molecular weight ~150,000 g mol-1, was purchased from

Sigma Aldrich (Atlanta, GA). N,N-dimethylformamide, utilized as the solvent for PAN, was also purchased from Sigma Aldrich. The laser dyes studied were purchased from

Exciton Inc. (Dayton, OH) and included kiton red, sulforhodamine 640, rhodamine 610 chloride, and rhodamine 610 perchlorate. Acebutolol, propranolol, and cortisone were purchased from Sigma Aldrich. Tetrabutylammonium bromide (≥ 99%) was purchased from Arcos Organics (New Jersey). Acetone, heptane and 2-propanol were acquired from Fisher Scientific (Fair Lawn, NJ). Chloroform (EMD Chemicals, Gibbstown, NJ), ethanol (Decon Labs, Inc., King of Prussia, PA), and methanol (Avantor Performance

Materials) were also utilized. Commercially available thin-layer chromatography plates were purchased from Macherey-Nagel (Bethlehem, PA) catalog number 818184. The plates contain a 0.15 mm layer on an aluminum sheet. The layer is composed of cyanopropyl modified silica particles ranging from 2-10 μm in size.

2.2.2 Instrumentation

A Hitachi S-4300 (Hitachi High Technologies America, Inc., Pleasanton, CA) scanning electron microscope was utilized to obtain all the SEM images of the electrospun nanofibers. ImageJ (Available from the National Institute of Health at http://www.rsbweb.nih.gov/ij/index.html) software was employed for all measurements

28 of SEM images, including those utilized to determine the degree of alignment of the electrospun nanofibers.

2.2.3 Aligned Electrospinning

The spinning drum apparatus used in this experiment is shown in Figure 2.2. The drum itself consisted of a 33.02 cm diameter x 3.18 cm wide piece of aluminum, which was drilled out to reduce weight. The drum is supported by a bearing-containing support on each end via 27.94 cm axles. The supports are bolted to rails at the base; the rails were slotted to provide easy adjustment to meet electrospinning conditions. On one side, the mandrel was attached to a ¾ horsepower, air-actuated stirrer with a chuck (Mixer

Direct, Inc., Jeffersonville, IN). This was used to rotate the drum relative to the electrospinning syringe. The motor was powered by house air, at a pressure of approximately 90 psi. The motor was capable of providing speeds of 0-3000 rpm; the rotational speed of the drum was determined using a NIST-certified, Monarch Instrument

PLT200 pocket laser tachometer (Cole-Parmer, Vernon Hills, IL). During electrospinning experiments, the surface of the drum was wrapped with 0.003” stainless steel shim stock (McMaster-Carr, Robbinsville, NJ) to serve as a solid support for the nanofibers. The rest of the apparatus consisted of the equipment necessary to electrospin the polymer: the syringe pump, a Harvard Model 33 dual syringe pump (Holliston, MA), and the power supply, a Spellman CZE 1000R high voltage source (Hauppauge, NY).

2.2.4 Thin-layer Chromatography

AE-UTLC plates were fabricated by removing the shim stock, now covered by aligned electrospun nanofibers, from the drum and cutting the material into plates that were appropriately sized for UTLC experiments (~2 x 5 cm). Analytes were spotted on the plates utilizing a fused silica capillary with an internal diameter of 250 µm; the volume of analyte spotted was ~50 nL. All UTLC experiments were conducted in a cylindrical glass development chamber (volume = 250 mL) containing 5 mL of mobile phase. The equilibration time was 10 minutes. The development of the laser dyes was carried out with a 90:10 (v:v%) 2-propanol:methanol mobile phase; the mixture of β- blockers and steroidal compounds were developed using a mobile phase composed of a mixture of chloroform and heptane with 1.0 wt% tetrabutylammonium bromide. All plates were developed until the mobile phase had reached a migration distance of 2.5 cm.

Following development, analysis was conducted utilizing a digital documentation system (Spectroline, Westbury, NY). The system consists of a CC-81 cabinet fitted with an ENF-280C 365 nm/254 nm, 8 W UV lamp and a GL-1301 universal camera adapter with a 58 mm adapter ring. The camera that was utilized was a Canon A650IS 12.1 MP digital camera. Digital photographs were subsequently analyzed with ImageJ as well as

TLC Analyzer (available at http://www.sciencebuddies.org/science-research- papers/tlc_analyzer.shtml) [25] and PeakFit. All images were darkened to enhance contrast prior to analysis. Analytes were visualized via exposure to UV radiation at

λ=254 nm.

2.3 Results and Discussion

2.3.1 Optimization of Aligned Electrospun PAN Nanofibers

An initial study was performed to determine the degree of alignment of the electrospun PAN nanofibers that could be achieved with the rotating drum. Utilizing previously published parameters for generating electrospun PAN nanofibers [10]; 30 minute electrospinning trials were performed at varying drum rotational speeds to determine optimum rotational speed of the drum for the production of highly aligned fibers. Experiments were performed at rotational speeds of 500, 750, 1000, 1250, and

1500 rpm in triplicate. Figure 2.3 shows representative SEM images of the electrospun

PAN nanofiber mats collected at each of these rotational speeds. The average nanofiber diameter was determined at each speed by measuring 45 individual nanofibers on each of three images for each run that was performed. The data in Table 2.1 illustrate that while the variance in nanofiber diameter from the lowest to the highest speed is less than 60 nm, there is a general trend toward decreasing nanofiber diameter with increasing rotational speed.

A B

C D

Figure 2.3. SEM images illustrating aligned electrospun PAN nanofibers generated on the rotating collector at rotational speeds of (A) 500, (B) 750, (C) 1000, (D) 1250, and (E) 1500 rpm. The white bar demonstrates the direction in which the drum was rotated; this also serves as the direction of nanofiber alignment.

Table 2.1. Average diameter of aligned electrospun PAN nanofibers as a function of collector rotational speed.

Rotational Speed (rpm) Nanofiber Diameter (nm) 500 520 ± 70 750 480 ± 80 1000 480 ± 120 1250 450 ± 70 1500 470 ± 80

The alignment of the nanofibers at each rotational speed was also assessed using

SEM images and ImageJ software. Using ImageJ, a virtual line was placed in the direction of alignment; the extent of the alignment was then assessed by determining the angle between the virtual line and individual nanofibers. The median of forty-five replicate measurements from each of three trials was calculated for the nanofibrous mats collected at each rotational speed. The angle from which the electrospun nanofibers deviated from the direction of the alignment was quantified; the population of nanofibers that deviated by 10°, 20°, and 30° from the direction of alignment was then determined

[26]. These measurements were performed for each of three images for three separate electrospun mats at each rotational speed. The results of these measurements are shown in Figure 2.4. This same method was also applied to non-aligned electrospun PAN nanofibers collected on a stationary collector. In this case the virtual line was placed in the same direction as it had been for the aligned nanofibers. The virtual line could be overlaid in any direction on the SEM images of the non-aligned electrospun nanofibers and the same degree of alignment would be calculated as these nanofibers were randomly deposited upon the collector.

100 90 80

60 50 40

30 Alignment (%) Alignment 20 10 0 10 20 30

Variation of Alignment (°) Figure 2.4. Percentage of electrospun PAN nanofibers positioned within a 10°, 20°, and 30° angle from the direction of nanofiber alignment for nanofibrous mats generated at 500 (■), 750 (■), 1000 (■), 1250 (■), and 1500 rpm (■).

It is clear from Figure 2.3 and Figure 2.4 that the alignment of the electrospun nanofibers collected at 500 and 1500 rpm is inferior relative to the alignment demonstrated for the nanofibers collected at 750, 1000, and 1250 rpm. Among the nanofibers collected at 750, 1000, and 1250 rpm, there is very little difference in their average alignment; for each of these three rotational speeds ~60% of the nanofibers were within 10° of the direction of alignment, ~80% of the nanofibers were within 20°, and over 90% of the electrospun nanofibers collected were within 30° of the direction of alignment. All nanofibrous mats collected on the rotating drum demonstrated a much higher degree of alignment than the non-aligned electrospun PAN nanofibers which were collected on a stationary collector. The non-aligned nanofibers had ~20% of the nanofibers within 10° of each other, ~30% of the nanofibers within 20°, and ~45% of the electrospun nanofibers collected were within 30° of the same direction.

Ultimately, samples collected from 750-1250 rpms yielded good alignment; all chromatographic experiments were conducted utilizing electrospun nanofibrous mats collected at 1250 rpm. The 750 and 1000 rpm samples probably would have resulted in similar chromatographic performance to the 1250 rpm sample, however, the 1250 rpm was chosen as a slightly smaller standard deviation of the average fiber alignment between mats was observed.

The morphologies of the aligned electrospun PAN nanofibers and the non-aligned

PAN nanofibers were further compared by assessing the degree to which the individual aligned and non-aligned nanofibers deviated from linearity via curving or twisting. In

37 order to quantitatively compare both types of nanofibers, a tortuosity factor (TF) was calculated according to Equation 2.1 [27]:

퐿푒푛푔푡ℎ 표푓 푡ℎ푒 푐푢푟푣푒 푙𝑖푛푒 푏푒푡푤푒푒푛 푡푤표 푝표𝑖푛푡푠 푇퐹 = (Eq. 2.1) 퐷𝑖푠푡푎푛푐푒 푏푒푡푤푒푒푛 푡푤표 푝표𝑖푛푡푠

All TF calculations were performed on SEM images of the same magnification (4000x) in order to ensure the scale at which the tortuosities of the nanofibers were determined was constant. The TF for the aligned nanofibers was determined to be 1.03 ± 0.02; TF of the non-aligned nanofibers was 1.28 ± 0.21. This indicates that the tortuosity of the non- aligned nanofibers was approximately 25% greater than that of the aligned nanofibers; it is clear that the alignment process not only orders the nanofibers by increasing the degree to which the nanofibers are collected in a single direction, but it also serves to straighten the individual nanofibers.

The effect of electrospinning time on mat thickness was also studied utilizing the same electrospinning parameters and maintaining the optimized collector rotational speed at ~1250 rpm. Electrospun mats were generated at times of 5, 15, 30, 45, 60, and 120 minutes; the thicknesses of the mats were then determined using SEM and ImageJ software. Figure 2.5 illustrates the relationship between electrospinning time and the thickness of the electrospun mat. Mat thickness increases with an increase in electrospinning time. However, it should be noted that after some time this increase deviates from linearity. While the mat thickness increases very little, remaining at ~25

µm thick, the mass of the electrospun mat increases by nearly 100% when the electrospinning time is increased from 60 to 120 minutes. Similar behavior has also been observed when electrospinning to a stationary collector [10]. This suggests that while the 38 thickness of the nanofibrous mat remains relatively unchanged, the nanofiber density of the mat increases. Initial experiments indicated that the aligned electrospun mats collected at lesser times, including the mat collected at 60 minutes, were of insufficient nanofiber density and demonstrated band broadening and poor chromatographic performance. As a result it was decided that the mat generated at an electrospinning time of 120 minutes would be utilized in this study. A representative SEM image of this mat is shown in Figure 2.6. While the mass of the aligned electrospun mat collected at 60 minutes correlated strongly with the mass of the non-aligned electrospun mat generated with an electrospinning time of 20 minutes, no adequate separation could be obtained, necessitating the use of the nanofibrous mat collected at 120 minutes. As a consequence, while the mat thickness remains the same, the volume of the stationary phase, VS, is greater for the AE-PAN plates than the E-PAN UTLC devices, demonstrating that a higher nanofiber density is required for the AE-PAN UTLC stationary phases. This suggests that the packing of the nanofibers is different for the aligned and non-aligned electrospun nanofibers.

In order to ensure that the alignment of the nanofibers had not been adversely affected by the increased electrospinning time, the alignment factor was recalculated using SEM images of the 120 minute, 1250 rpm mats. The alignment factor did not decrease with the increased electrospinning time (Figure 2.6).

Mat Thickness (µm) Thickness Mat 10

0 15 30 45 60 120 Electrospinning Time (min)

Figure 2.5. Mat thickness as a function of electrospinning time for the generation of aligned electrospun PAN nanofibers.

A B

Figure 2.6. Mat thickness (A) and fiber morphology (B) of the aligned electrospun PAN nanofibers collected for 120 min at 1250 rpm.

2.3.2 Initial Separation: Laser Dye Analysis

In order to demonstrate the efficacy of using aligned electrospun nanofibers as a chromatographic stationary phase, a mixture of four laser dyes, rhodamine 610 chloride, rhodamine 610 perchlorate, sulforhodamine 640, and kiton red, was applied to the PAN

AE-UTLC plates. The purpose of this separation was to establish the suitability of the

AE-PAN nanofibers for separations. The retardation factor, Rf (Equation 2.2), for each of these compounds was calculated and is listed in Table 2.2 where the results are also compared to those for a PAN E-UTLC plate. Zs is the distance the analyte travels and Zf is the distance the mobile phase travels during development. As previously noted, there is a wide range of chemical functionalities present within the selected laser dyes; it may not be possible to prepare a single mobile phase composition that is optimal for the separation of all of the analytes [10].

푍푠 푅푓 = (Eq. 2.2) 푍푓

It is interesting to note that while there is no significant difference between the retardation factors observed for the laser dyes on the AE-UTLC plate versus the E-UTLC plate, the %RSD is generally lower (~50-80% lower for the four analytes) on the AE-

UTLC plate.

Table 2.2. Retardation factors Rf, of laser dyes for PAN AE-UTLC and E-UTLC plates.

Analyte AE-UTLC plate Rf E-UTLC plate Rf (%RSD) (%RSD) Kiton red 0.50 (6.0) 0.42 (44) Rhodamine 610 chloride 0.30 (19) 0.51 (39) Rhodamine 610 perchlorate 0.34 (20) 0.54 (44) Sulforhodamine 640 0.46 (5.6) 0.33 (45)

2.3.3 Separation of β-Blockers and Steroidal Compounds

A mixture of acebutolol, propranolol, and cortisone was studied using both the

PAN AE-UTLC and E-UTLC plates. The concentrations of the analytes were 2.5 x 10-3

M, 2.0 x 10-2 M, and 5.0 x 10-3 M, respectively. These compounds represented the lowest concentrations that could be utilized while allowing for post-development visualization.

The effects of changing the mobile phase composition on retention behavior were compared for both the PAN AE-UTLC (Figure 2.7A) and E-UTLC (Figure 2.7B) plates.

A mobile phase comprised of a 40:60 (v:v%) mixture of chloroform and heptane with 1.0 wt% tetrabutylammonium bromide (TBABr) was found to give the greatest resolution for the AE-UTLC plates, whereas a mobile phase consisting of a 30:70 (v:v%) mixture of chloroform and heptane with 1.0 wt% TBABr gave the best resolution for the non- aligned E-UTLC plates. TBABr has been previously shown to decrease retention time while minimizing band broadening when applied to separations of analytes similar to acebutolol and propranolol [17,28]. When comparing Figure 2.7A and Figure 2.7B, it is interesting to note that the AE-UTLC Rf plot is shifted to the right relative to the E-UTLC plot; the analytes begin to migrate sooner on the non-aligned UTLC plate than the AE-

UTLC plate, despite the fact that the mobile phase velocity is faster for the AE-UTLC plate, as shown in Figure 2.8. The increased retention of the AE-UTLC plate can be explained by considering the increase in the volume of the stationary phase,VS, for the

AE-UTLC plates discussed earlier. Consider Equation 2.3 [1]:

푉 푘 = 퐾 푆 (Eq. 2.3) 푉푀

Where k is the retention factor, K is the equilibrium constant, VM is the mobile phase volume and VS is the volume of the stationary phase. Since the stationary phase in the

AE-UTLC and E-UTLC systems are comprised of PAN, K is the same for both plates.

However, VS is significantly greater for the AE-UTLC plate, which should give a corresponding increase in k. The retention factor, k, is further related to the retardation factor, Rf, according to Equation 2.4 [1].

1−푅 푘 = 푓 (Eq. 2.4) 푅푓

It is clear from Equation 2.4 that any increase in k would result in a decrease in Rf; this is observed in the case of the AE-UTLC plate. The increase in VS for the AE-UTLC plate leads directly to the observed lower Rf values relative to the E-UTLC stationary phase.

1.0 A 0.8

0.6

f R 0.4

0.2

0.0 0 20 40 60 80 100 % Chloroform (Chloroform:Heptane)

1.0 B 0.8

0.6

f R 0.4

0.2

0.0 0 20 40 60 80 100 % Chloroform (Chloroform:Heptane)

Figure 2.7. Retardation factors Rf, of acebutolol (♦), cortisone (■), and propranolol (▲) for PAN (A) AE-UTLC plates and (B) E-UTLC plates (n=5).

400 350

300 250 200 150

Time (seconds) Time 100 50 0 0 20 40 60 80 100 % Chloroform (Chloroform:Heptane)

Figure 2.8. Time of analysis as a function of mobile phase composition for PAN AE- UTLC (■) and E-UTLC (♦). The migration distance was 2.5 cm.

Sample chromatograms for the separation of the β-blocker/steroid mixture with

PAN AE-UTLC and E-UTLC plates using chloroform and heptane mobile phases with

1.0 wt% TBABr is displayed in Figure 2.9. The resolution of the AE-UTLC plate was observed at 1.20 between acebutolol and cortisone as compared to 0.692 for the E-UTLC plate. Resolution was 2.02 and 1.01 between cortisone and propranolol for AE-UTLC and E-UTLC, respectively. Table 2.3 shows the optimized mobile phase conditions with respect to resolution for the AE-UTLC and E-UTLC chromatograms. The run-to-run reproducibility of the separation was enhanced with the AE-UTLC device, as noted by the smaller relative standard deviations of the Rf values for the analytes separated on the

AE-UTLC plate compared to the E-UTLC plate. The AE-UTLC plates showed higher plate numbers for all three analytes when compared to the E-UTLC stationary phase at the conditions which gave the greatest resolution. Equation 2.5 was used to calculate plate number, N, which is proportional to the square of the migration distance of the analyte, ZS, divided by the spot width of the developed analyte, w [1].

푍 2 푁 = 16 ( 푆) (Eq. 2.5) 푤

A single analyte, propranolol, was chosen to demonstrate the effect of longer migration distances on plate number calculations, as shown in Figure 2.10. At longer development distances, N for propranolol was very high. The comparatively low plate numbers

(Figure 2.10) are a consequence of performing the separation at a relatively short migration distance of 2.5 cm. Increasing efficiency with increasing development distance was also observed for glassy carbon E-UTLC stationary phases [11]; to our knowledge the only other reported occurrence of increased efficiency at increased migration distance 48 involved the use of pressurized planar electrochromatography (PPEC) [29]. As discussed in greater detail below, the AE-UTLC plates demonstrate a very rapid mobile phase velocity (Figure 2.8) which does not appear to slow significantly over distance and time

(Figure 2.11). Decreasing mobile phase velocity contributes greatly to band dispersion in

TLC [1]; as the mobile phase velocity rapidly decreases (typically at or before 7 cm in

HPTLC) [30] the band broadening of the analyte exceeds its rate of migration [1]. This causes a decrease in the quality of the separation and a decrease in N [1,28]. This effect could be mitigated by the higher mobile phase velocities of the AE-UTLC phase, resulting in higher N at increased migration distances. More detailed studies are being performed to further elucidate this chromatographic behavior.

1.2 A 3 1.0

0.8

0.6 1

Intensity 0.4 2 0.2

0.0 0.0 0.5 1.0 1.5 2.0 2.5 Distance (cm)

1.2 B 1.0

0.8

0.6 3

Intensity 0.4

0.2 1 2 0.0 0.0 0.5 1.0 1.5 2.0 2.5 Distance (cm)

Table 2.3. Efficienceis, N, and retardation factors Rf, for the separation of acebutolol, cortisone, and propranolol on PAN AE-UTLC and E-UTLC plates using mobile phases of 40:60 (v:v%) and 30:70 (v:v%) chloroform:heptane with 1.0 wt% TBABr.

PAN AE-UTLC PAN E-UTLC Analyte (40:60 chloroform:heptane) (30:70 chloroform:heptane) N Rf (%RSD) N Rf (%RSD) Acebutolol 470 0.27 (11) 3 0.06 (24) Cortisone 130 0.44 (6.7) 36 0.30 (6.6) Propranolol 1700 0.56 (4.5) 870 0.53 (6.1)

16000

12000

8000 Efficiency, N Efficiency, 4000

0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Distance (cm)

Figure 2.10. Change in efficiency, N (♦), of propranolol with increasing migration distance on a PAN AE-UTLC plate using a 40:60 (v:v%) chloroform:heptane mobile phase with 1.0 wt% TBABr.

50.0

40.0

) 30.0

(cm

2 f

Z 20.0

10.0

0.0 0.0 1.0 2.0 3.0 4.0 5.0 Time (min)

Figure 2.11. Comparison of mobile phase velocities of PAN AE-UTLC (■) plates, PAN E-UTLC (♦) plates, and commercial cyano TLC (▲) plates using heptane as the mobile phase. (Error bars are contained within data points when not visible).

As previously stated, Figure 2.8 shows the times of analysis for both the PAN

AE-UTLC and E-UTLC for a final migration distance of 2.5 cm as a function of mobile phase composition. It is clear that the PAN AE-UTLC plates not only provide a more rapid time of analysis, approximately 2-2.5 times faster than E-UTLC plates, but also a more reproducible time of analysis. This enhanced reproducibility in the time of analysis for the AE-UTLC plate causes enhanced reproducibility of the separation. Furthermore, these results demonstrate that AE-UTLC phases can be used to generate separations over a very small development distance in a short time frame. All AE-UTLC plates were developed between 1-1.5 minutes, compared to the nearly 3 minute analysis time for E-

UTLC plates.

2.3.4 Mobile Phase Velocity

As illustrated in Section 4.3.3, the AE-UTLC plates demonstrated significantly faster times of analysis relative to the E-UTLC plates showing that through alignment, the mobile phase velocity could be significantly increased. Mobile phase transport through electrospun nanofibrous mats in UTLC has been previously examined using the

Lucas-Washburn equation (Equation 2.6) [31].

훾푅푡푐표푠휃 푍2 = (Eq. 2.6) 푓 2휂

In this equation, Zf is the migration distance of the mobile phase, γ is the surface tension of the mobile phase, R is the effective capillary (or pore) radius, t is time, θ is the contact angle of the mobile phase with the stationary phase, and η is the mobile phase viscosity

[1,31]. The Lucas-Washburn equation is frequently used to describe capillary flow 54 through porous substances, including traditional silica and cyano TLC stationary phases

[10,11,31]. The Lucas-Washburn equation can be applied assuming the following conditions: an adsorbed film of penetrating liquid must be present (accomplished by exposure to vapors of mobile phase and achievement of equilibrium prior to use); prevalence of laminar flow conditions (Reynolds number <1200); inertial influences must be negligible; the contact angle of the mobile face upon the stationary phase must be 0° [1,11,31]. Nonpolar heptane has a 0° contact angle with the PAN and cyano phase surfaces and was employed as the mobile phase in the determination of R described below [11]. The R of electrospun nanofibrous mats and its relationship on the rate of capillary rise has been previously studied; it was determined that as R increases the rate of capillary rise also increases [11,32].

The R of AE-UTLC plates was calculated using the Lucas-Washburn equation

(Equation 2.6) [31] and compared with R for both the E-UTLC and the commercial cyano

TLC phases. The aligned PAN nanofibers had a calculated R of 1200 ± 70 nm; this value is significantly larger than R for the non-aligned E-PAN nanofibers (280 ± 10 nm) as well as the commercial cyano TLC phase (430 ± 21 nm) [11]. This is to be expected as the

AE-UTLC phase demonstrates the highest mobile phase velocity (Figure 2.11). It has been previously noted that the pore structure of E-UTLC plates greatly differs from that of traditional particle-based TLC stationary phases; R for the commercial cyano TLC phase is < 1/10 of the particle radius, whereas R for the non-aligned E-UTLC plate is

~1.5 times greater than the nanofiber radius as measured by SEM, while the AE-PAN R is ~5 times greater than the radius of the aligned PAN nanofibers. These results agree

55 with the established trend of increasing mobile phase velocities with increasing values of

When the Lucas-Washburn equation is applied to mobile phase migration in TLC, the velocity constant, κ, replaces the term γR/2. This factor has been traditionally used to describe mobile phase velocity in typical precoated TLC phases; κ relates Zf and t according to Equation 2.7 [1]:

2 푍푓 = 휅푡 (Eq. 2.7)

2 Plotting Zf versus t should give a straight line where κ is equal to the slope; deviations from a straight line are indicative of abnormal flow behavior resulting from experimental error or stationary phase irregularities [33]. As shown in Figure 2.11 these plots were compiled for the AE-UTLC, E-UTLC, and commercial cyano phases; κ was determined to be 0.173 cm2 second-1 for the AE-UTLC plates; this value is over 200% greater than that of the non-aligned E-UTLC plates (0.055 cm2 second-1). The commercial cyano phase, with average particle sizes of 2-10 μm, has a κ of 0.107 cm2 second-1.This corresponds directly to the faster mobile phase velocity observed for the PAN AE-UTLC

2 plates. Additionally, Figure 2.11 shows a linear relationship between Zf and t, demonstrating stationary phase homogeneity for the electrospun plates as well as verifying the applicability of the Lucas-Washburn equation for describing flow through electrospun media [33]. The AE-UTLC plates demonstrated a κ value which is nearly

45% higher than that of the cyano plate.

While the mobile phase velocity for the cyano plate is more rapid than that of the

E-UTLC plate when heptane is used as the mobile phase (~90% faster) (Figure 2.11),

56 comparable mobile phase velocities are observed for the cyano and E-UTLC plates for acetone (E-UTLC ~10% faster) and 50:50 (v:v%) acetone:water (cyano phase ~5% faster) (Figure 2.12A-B). The AE-UTLC plate showed faster mobile phase velocities than all other tested stationary phases for all mobile phase conditions. While it has been previously demonstrated that among electrospun UTLC plates, larger nanofiber diameters correspond with faster mobile phase velocity as well as increased values for R and κ, this explanation cannot be used to describe the increase of the mobile phase velocity of the

AE-UTLC plates compared to the E-UTLC plates [11]. The nanofiber diameters of the

AE-UTLC plates and E-UTLC plates (453 ± 70 nm vs. 400 ± 50 nm) are not statistically different at the 95% confidence level. This suggests that the more highly aligned structure of the AE-UTLC nanofibrous mat, as well as the decreased tortuosity of the nanofibers, are primarily responsible for the observed enhancement in mobile phase velocity relative to the E-UTLC stationary phases. These results strongly corroborate previous studies which demonstrate that increased nanofiber tortuosity decreases R and thus also decreases the rate of capillary rise of liquids penetrating the nanofibrous mat

[32].

50.0 A

40.0

)

2 30.0

(cm

f 20.0 Z

10.0

0.0 0.0 1.0 2.0 3.0 4.0 5.0 Time (min)

50.0 B

40.0

)

2 30.0

(cm

f 20.0 Z

10.0

0.0 0.0 1.0 2.0 3.0 4.0 5.0 Time (min)

Figure 2.12. Comparison of mobile phase velocities of PAN AE-UTLC (■) plates, PAN E-UTLC (♦) plates, and commercial cyano TLC (▲) plates using 50:50 (v:v%) acetone:water (A) and acetone (B) as the mobile phase. (Error bars are contained within data points when not visible).

2.4 Conclusions

This study demonstrates the applicability of aligned electrospun nanofibers to

UTLC analysis. Aligning the electrospun nanofibers with a rotating drum collector not only increases the directional orientation of the nanofibers but also increases the straightness of the individual nanofibers. Furthermore, AE-UTLC devices represent a further enhancement in separation efficiency relative to E-UTLC stationary phases while decreasing time of analysis by over 50%. When compared with E-UTLC, the AE-UTLC devices provide a more rapid and consistent time of analysis while also showing increased efficiency and reproducibility; these enhanced separations occur over a very short development distance and time. These enhancements are obtained while maintaining the significant gains afforded by the cost effectiveness and versatility of the original E-UTLC plates [10].

2.5 References

[1] C. F. Poole, The Essence of Chromatography, Elsevier Science, Amsterdam, The

Netherlands, 2003.

[2] J. Sherma, Planar chromatography, Anal. Chem., (2010), 82, 4895-4910.

[3] J.S. Bernard Fried, Thin-layer Chromatography: Techniques and Applications,

Marcel Dekker, New York, NY, 1994.

[4] L.W. Bezuidenhout, M.J. Brett, Ultrathin layer chromatography on nanostructured

thin films, J. Chromatogr. A., (2008), 1183, 179-185.

[5] S.A. Nabi, M.A. Khan, Selective TLC separation of lysine and threonine in

pharmaceutical preparations, Acta Chromatogr., (2003), 13, 161-171.

[6] M. Kastelan-Macan, S. Cerjan-Stefanovic, M. Petrovic, Phenol adsorption on

active carbon by means of thin-layer chromatography, Chromatographia, (1989),

27, 297-300.

[7] J. Song, D.S. Jensen, D.N. Hutchison, B. Turner, T. Wood, A. Dadson, M.A. Vail,

M.R. Linford, R.R. Vanfleet, R.C. Davis, Carbon nanotube-templated

microfabrication of porous silicon-carbon materials with application to chemical

separations, Adv. Funct. Mater., (2011), 21, 1132-1139.

[8] S.R. Jim, A.J. Oka, M.T. Taschuk, M.J. Brett, Morphological modification of

nanostructured ultrathin-layer chromatography stationary phases, J. Chromatogr.

A., (2011), 1218, 7203-7210.

[9] H.E. Hauck, M. Schulz, Ultra thin-layer chromatography, Chromatographia,

(2003), Supp. 57, S-313-S-315.

[10] J.E. Clark, S.V. Olesik, Technique for ultrathin layer chromatography using an

electrospun, nanofibrous stationary phase, Anal. Chem., (2009), 81 4121-4130.

[11] J.E. Clark, S.V. Olesik, Electrospun glassy carbon ultra-thin layer chromatography,

J. Chromatogr. A., (2010), 1217, 4655-4662.

[12] S. Ramakrishna, K. Fujihara, W.E. Teo, T.C. Lim, Z. Ma, An Introduction to

Electrospinning and Nanofibers, World Scientific, River Edge, NJ, 2005.

[13] W.G. Shim, C. Kim, J.W. Lee, J.J. Yun, Y. Jeong, H. Moon, K.S. Yang,

Adsorption characteristics of benzene on electrospun-derived porous carbon

nanofibers, J. Appl. Polym. Sci., (2006), 102, 2454-2462.

[14] U. Boudriot, R. Dersch, A. Greiner, J.H. Wendorff, Electrospinning approaches

toward scaffold engineering - a brief overview, Artif. Organs., (2006), 30, 785-792.

[15] J.W. Zewe, J.K. Steach, S.V. Olesik, Electrospun fibers for solid-phase

microextraction, Anal. Chem., (2010), 82, 5341-5348.

[16] T.E. Newsome, J.W. Zewe, S.V. Olesik, Electrospun nanofibrous solid-phase

microextraction coatings for preconcentration of pharmaceuticals prior to liquid

chromatographic separations, J. Chromatogr. A, 1262, 1-7.

[17] W. Jost, H.E. Hauck, W. Fischer, Cyano-modified pre-coated plates - a medium

polar modification for high performance thin-layer chromatography,

Chromatographia, (1986), 21, 375-378.

[18] W.E. Teo, S. Ramakrishna, A review on electrospinning design and nanofiber

assemblies, Nanotechnology, (2006), 17, R89-R106.

[19] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Aligned biodegradable nanofibrous

structure: a potential scaffold for blood vessel engineering, Biomater., (2004), 25,

877-886.

[20] Z. Song, X. Hou, L. Zhang, S. Wu, Enhanced crystallinity and orientation by hot-

stretching to improve the mechanical properties of electrospun partially aligned

polyacrylonitrile (PAN) nanocomposites, Materials, (2011), 4, 621-632.

[21] R. Jalali, M. Morshed, S.A. H. Ravandi, Fundamental parameters affecting

electrospinning of PAN nanofibers as uniaxially aligned fibers, J. Appl. Polym.

Sci., (2006), 101, 4350-4357.

[22] D. Li, Y. Wang, Y. Xia, Electrospinning of polymeric and ceramic nanofibers as

uniaxially aligned arrays, Nano Lett., (2003), 3, 1167-1171.

[23] S.F. Fennessey, R.J. Farris, Fabrication of aligned and molecularly oriented

electrospun polyacrylonitrile nanofibers and the mechanical behavior of their

twisted yarns, Polymer, (2004), 45, 4217-4225.

[24] W. Hu, Z.M. Huang, S.Y. Meng, C. He, Fabrication and characterization of

chitosan coated braided PLA wire using aligned electrospun fibers, J. Mater. Sci.:

Mater. Med., (2009), 20, 2275-2284.

[25] A.V.I. Hess, Digitally enhanced thin-layer chromatography: an inexpensive, new

technique for qualitative and quantitative analysis, J. Chem. Educ., (2007), 84, 842-

847.

[26] M.B. Bazbouz, G.K. Stylios, Alignment and optimization of nylon 6 nanofibers by

electrospinning, J. Appl. Polym. Sci., (2007), 107, 3023-3032.

[27] W. Zhou, Y. Wu, F. Wei, G. Luo, W. Qian, Elastic deformation of multiwalled

carbon nanotubes in electrospun MWCTs-PEO and MWCNTs-PVA nanofibers,

Polymer, (2005), 46, 12689-12695.

[28] L.H. Bluhm, T. Li, Effect of analogue ions in normal-phase ion-pair chromatograpy

of quaternary ammonium compounds, J. Chromatogr. Sci., (1999), 37, 273-276.

[29] A.L. Novotny, D. Nurok, R.W. Replogle, G.L. Hawkins, R.E. Santini, Results with

an apparatus for pressurized planar electrochromatography, Anal. Chem., (2006),

78, 2823-2831.

[30] C.F. Poole, S.K. Poole, Multidimensionality in planar chromatography, J.

Chromatogr. A, (1995), 703, 573-612.

[31] E.W. Washburn, Dynamics of capillary flow, Phys. Rev., (1921), 17, 374-375.

[32] M.S. M. Jad, S.A. H. Ravandi, H. Tavanai, R.H. Sanatgar, Wicking phenomenon in

polyacrylonitrile nanofiber yarn, Fiber Polym., (2011), 12, 801-807.

[33] F. Geiss, Fundamentals of Thin-layer Chromatography (Planar Chromatography),

Huethig, Heidelberg, Germany, Basel, Switzerland, New York, United States of

America, 1987.

CHAPTER 3: PLANAR ELECTROCHEMISTRY (PEC) SEPARATION OF LASER DYES WITH ELECTROSPUN NANOFIBERS

3.1 Introduction

Thin-layer chromatography (TLC), also called planar chromatography, is a widely used separation technique that is over 75 years old [1]. The advantages of TLC are its simplicity and it also allows for multiple separations to occur simultaneously on the same chromatographic device. Additionally, TLC can be evaluated without the need to transfer the separated analytes of interest to a detector. Despite the ease of use; TLC comes with detrimental aspects. The main problem is that analysis with TLC suffers from a poor mobile phase flow profile over the sorbent layer. The mobile phase flow in TLC is primarily provided by capillary flow. The poor flow profile associated with capillary mediated flow is due to the inverse relationship between the linear velocity of the solvent front and its migration distance. The decrease in mobile phase velocity with distance traveled by the solvent front often results in long analysis times. Furthermore, it is impossible to perform a separation at an optimum flow rate. As a result, this yields poor chromatographic efficiency. These effects are magnified when using layers consisting of small particles, as the solvent velocity constant is proportional to the particle diameter of the stationary phase [2].

Several forced-flow techniques have been developed to help overcome the poor flow profile associated with TLC analysis. Of these, some have been made commercially 64 available [3,4]. These methods achieved an improved mobile phase flow profile through introducing pressure forced-flow, with moderate chromatographic success. Other forced- flow techniques are those where the mobile phase is driven by an applied electric field.

Forced-flow techniques with an applied electric field are called planar electrochromatography (PEC) [5,6,7,8,9,10,11,12,13]. The velocity, veo, of electroosmotic flow (EOF), created by the applied electric field, is given by the

Smoluchowski equation [14]:

휀∗휉∗퐸 휈 = (Eq. 3.1) 4∗휋∗휂 where ε is the dielectric constant, ζ is the zeta potential, E the electric field strength, and η is the viscosity. In theory, EOF should have a flat flow profile in contrast to the profiles in either capillary mediated or pressure driven flow. A substantial advantage to using the electroosmotic force rather than capillary forces to drive the mobile phase in planar chromatography is gained. Flow velocity should be independent of migration distance and sorbent particle diameter. EOF can be controlled by adjusting the magnitude of the electric field, and this should allow PEC to be performed at an optimum flow velocity.

The first work on PEC was by Pretorius et al. [5]. A separation of four steroids, as the separation, on pre-wetted plates, was 15 times faster than an equivalent separation by conventional TLC. Shafik et al. later demonstrated that the increase in migration velocity was attributed to the increased evaporation of mobile phase due to Joule heating [15]. A major limitation of the technique is that the sorbent layer dries when there is excessive

Joule heating. Joule heating will always occur during PEC, and will cause evaporation of

65 mobile phase unless the system is cooled or the environment in contact with the sorbent layer is fully saturated with vapor.

PEC requires plates with thinner layers than commercial plates, which are more appropriate for heat dissipation [3,16]. Electrospun nanofiber stationary phases were recently introduced for ultra-thin layer chromatography (UTLC) [17]. Electrospinning is a technique that relies on repulsive electrostatic forces to produce nanofibers from a polymer solution [18]. Electrospinning is used to fabricate a mat of randomly distributed nanofibers (∼ 25 μm thick) which is used as a stationary phase for UTLC. Electrospun

UTLC plates offer many advantages, including easily variable mat thicknesses and nanofiber diameters [17,19]. Rapid separations and enhanced efficiencies have been reported for UTLC analysis using short separation distances (20-30 mm).

Electrospinning is capable of producing nanofibers with diverse functionality [18]. This method is adaptive and has been used to create nanofibers with polymers [17,20,21,22], carbon [19], and silica [23]. Multiple polymer stationary phases have been produced from polyacrylonitrile (PAN) [17,20,21,22], poly(vinylalcohol) (PVA) [24], and cellulose acetate [25].

Although faster separations and enhanced efficiencies are possible with electrospun UTLC, the mobile phase velocity nevertheless reduces as a function of distance traveled since the flow is capillary driven [17,19,20,21,22,23,24]. Therefore, the

66 performance of electrospun UTLC plates may be further improved by further development of forced-flow driven mobile phases. Because of the contribution of electrophoresis to the separation mechanism, it is expected that PEC can offer additional selectivity compared to UTLC, as the analytes are composed of charged species [26].

The variable thickness of the electrospun stationary phase offers a promising solution to problems associated with Joule heating found in thicker layers currently in use.

Accordingly, the performance of electrospun nanofiber stationary phases is examined and it is reasonable to expect that these systems will be applicable to PEC analysis.

3.2 Experimental

3.2.1 Reagents

Silica (SiO2) nanoparticles, Angstrom Sphere monodispersed SiO2 powder, 250 nm with a particle size standard deviation of <10%, were purchased from Fiber Optic

Center Inc. (New Bedford,MA). Polyvinylpyrrolidone (PVP), average Mw=1,300,000, was purchased from Acros Organics (New Jersey). Reagent alcohol (90% ethanol, 5% methanol, and 5% 2-propanol), (HPLC grade, 99.8% purity) was purchased from Fisher

Scientific (Fair Lawn, NJ). Methanol (MeOH) (ACS grade) was purchased from Macron

Chemicals (St. Louis, MO). Acetonitrile (ACN) (HPLC grade), citric acid (ACS grade), and 2-propanol (IPA) (HPLC grade) were purchased from Fisher Scientific. Sodium citrate (ACS grade) was purchased from Jenneile Chemical Co. (Cincinnati, OH). The laser dyes studied were purchased from Exciton Inc. (Dayton, OH) and included kiton red

(KR), rhodamine 590 chloride (R590), rhodamine 101 chloride (R101), rhodamine 610

67 chloride (R610), and sulforhodamine 640 (SR). Analyte structure and pKa values shown in Table 3.1.

Table 3.1. Analyte structure and pKa values for laser dyes.

Analyte Structure pKa

KR <1.5, <2.0

R101 <2.0, 3.3

R590 <2.0

R610 <2.0, 3.6

SR <1.5, <2.0

3.2.2 Instrumentation

The PEC device used can be seen in Figure 3.1. The PEC device was prepared from an electrophoresis chamber (Mini-Sub Cell GT, Bio-Rad, Hercules, CA). A Fisher

Biotech high voltage power supply (Pittsburgh, PA) was used to apply an appropriate voltage across the anode (1) and cathode (3). The Whatman wicks (2) were essential to provide pre-wetting of the electrospun SiO2/PVP UTLC stationary phase (5) as well as maintaining contact between stationary phase and mobile phase during experimentation.

Each wick overlapped the stationary phase by 0.5 cm on each side of the SiO2/PVP

UTLC plate. A glass cover plate (not shown) was placed on top of the stationary phase, to reduce evaporation of mobile phase while also keeping the stationary phase in good contact with the pre-wetted wicks during experimentation. The reservoirs (4) simply housed mixed mobile phase. Wicks were placed in the mobile phase while the electrodes were also submerged in the liquid.

(2) Whatman wick

(1) Anode (3) Cathode

(4) MP reservoir (5) ES UTLC plate

Figure 3.1. PEC device. Components are listed (1) Anode, (2) Whatman wicks, (3) Cathode, (4) Mobile phase reservoir, and (5) electrospun SiO2/PVP UTLC plate.

3.2.3 Electrospinning

Electrospinning SiO2/PVP nanofiber mats have previously been reported [27].

During electrospinning experiments, SiO2/PVP nanofibers were deposited onto the surface of 0.003” stainless steel shim stock (McMaster-Carr, Robbinsville, NJ) to serve as a solid support for the nanofibers. The electrospinning apparatus consisted of: the syringe pump, a Harvard Model 33 dual syringe pump (Holliston, MA), and the power supply, a Spellman CZE 1000R high voltage source (Hauppauge, NY). It was imperative to keep the environment around where electrospinning occurred to 30-50% relative humidity. An applied potential of 11 kV was used for electrospinning as well as a syringe tip to collector distance of 10 cm. A SiO2/PVP UTLC plate has previously been reported [28]. Newsome et al. report that the as-spun stationary phase is soluble in common mobile phase solvents. To utilize the SiO2/PVP nanofiber mat as a UTLC plate crosslinking of the PVP was performed between 150-200° C. SiO2/PVP UTLC plates generated for use in PEC were thermally crosslinked at a final temperature of 175°C without any detrimental effects from contact with MeOH, ACN, IPA, H2O or heptane were observed. Average nanofiber diameters are 300 ± 90 nm while the mat thickness was measured to 25 ± 2 µm for the SiO2/PVP UTLC device [27]. The electrospun nanofiber mat is later transferred from the stainless steel shim stock to a glass slide for

PEC analysis.

3.2.4 Planar Electrochromatography

Prior to analysis, all analytes (3.0 x 10-4 M) are deposited (~ 50 nL) onto the

SiO2/PVP stationary phase approximately 0.75 cm from one edge of the UTLC plate.

The solvent used to make the analyte solution is allowed to evaporate for a period of one hour before analysis. The desired mobile phase was used to wet the stationary phase and

Whatman wicks followed by adhering the wetted wicks to both sides of the plate (~ 0.5 cm from edge). The plate is positioned with the analytes nearer the anode side of the chamber with the wick placed in the reservoir housing the mobile phase. Voltage is then applied across the anode and cathode. An electroosmotic flow is produced when the stationary phase becomes ionized (pH>3). An electrical double layer is formed as cations in the mobile phase flow toward the cathode while anions flow toward the anion. The net electroosmotic flow between anion and cation movement is toward the cathode. PEC analysis is then carried out for the prescribed time period. The mobile phase is then allowed to evaporate from the SiO2/PVP UTLC plate prior to visualization. Visualization of analytes was conducted utilizing a digital documentation system (Spectroline,

Westbury, NY). The system consists of a CC-81 cabinet fitted with an ENF-280C 365 nm/254 nm, 8 W UV lamp and a GL-1301 universal camera adapter with a 58 mm adapter ring. The camera was a Canon A650IS 12.1 MP digital camera. Digital photographs were subsequently analyzed with ImageJ as well as TLC Analyzer (available at http://www.sciencebuddies.org/science-research-papers/tlc_analyzer.shtml) [29] and

PeakFit. All images were darkened to enhance contrast prior to analysis. Analytes were visualized via excitation with UV radiation at λ=365 nm.

3.3 Results and Discussion

3.3.1 Mobile Phase Optimization

Laser dyes were chosen as the analytes of interest because they share similar chemical structures as well as exhibiting an ease of detection with UV radiation at 365 nm [8,17,20,30]. Analyte deposition onto the stationary phase was evaluated with measuring initial spot widths from hand-spotting with a capillary tube (250 and 500 µm) and delivery with a Chemyx Nanojet device (100 µm). Ideal initial spot widths are as narrow and reproducible as possible. All analytes were deposited onto the SiO2/PVP

UTLC stationary phase and the resulting analyte spot widths were measured. Figure 3.2 shows the measured spot widths for KR, R101, R590, R610, and SR for the three spotting methods. The 500 µm (I.D.) capillary produced the largest spots and was eliminated from consideration. The automated Chemyx Nanojet device with 100 µm (I.D.) syringe produced similar spot sizes to that of hand-spotting with a 250 µm (I.D.) piece of capillary tubing. Furthermore, the 250 µm (I.D.) capillary produced a lower standard deviation and deposition occurred quicker for the hand-spotting method. Hand-spotting with a 250 µm (I.D.) capillary was chosen for all subsequent analysis.

To begin mobile phase optimization, mixtures containing ACN, IPA, and 35 mM citrate buffer were explored because previous studies indicate that laser dye separations occurred with good selectivity for these mobile phases [30,31]. Low conductivity buffers were also explored (Homopipes and MES). The low conductivity buffers did not successfully provide enough current for analysis to occur and were eliminated as options.

Figure 3.3 shows PEC performed for 60 second analysis time with 1000 V of applied

74 potential. The migration distance of laser dyes were monitored under different mobile phase conditions. Mobile phases were composed of 50% (v:v%) 35 mM citrate buffer

(pH 5.6) with the remaining mobile phase made up of ACN and IPA. Similar migration distances were observed across all mobile phase solvent mixtures under the conditions used. The proliferation of heat due to Joule heating was observed during analysis. It was noticed that the selectivity of the analytes change across the mobile phase composition.

The analyte migration distance of R101 is largest at 50:0:50 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA while R590 is largest for 50:50:0 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA.

To better understand the generation of Joule heating during analysis, the applied voltage was explored with variation, with 250, 500, 750 and 1000 V. Figure 3.4 shows migration distances for analytes with a 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH

5.6):ACN:IPA mobile phase with 60 second analysis time. Heat was observed in proportion to the amount of applied voltage, with 1000 V producing the most heat.

Analyte selectivity did not change over all applied voltages. The migration distance was greatest for the 1000 V analysis. This corresponded to producing the fastest analyte migration velocities (0.11 mm/sec). With this in mind, various analysis times (60, 120, and 180 sec) were explored for the four chosen voltages (250, 500, 750, and 1000V).

Figure 3.5 shows the associated analyte migration distances for the 750 V with analysis times of 60, 120, and 180 seconds as analysis could not be conducted with greater than 60 seconds for 1000 V. The migration distance of R590 changed from 6.69

± 0.27 mm for 60 seconds to 14.33 ± 0.75 mm for 120 seconds. Analyte migration

75 distance increased drastically for the 120 second analysis over the 60 second sample period, while 180 seconds provided a small increase in migration distance over 120 seconds. This is further exemplified in the analyte velocity for 120 seconds (0.12 mm/sec) and 180 seconds (0.07 mm/sec). Analyte migration slowed for the period from

120 to 180 seconds.

The effects of pH were evaluated while measuring migration distance in Figure

3.6. Little information was garnered from pH changes within the explored region of 4.4 to 6.2 pH units. Citrate buffer with pH 5.6 was chosen as the optimal pH because of the increased migration distance with reduced final spot width. Analyte migration distance increases from significantly with pH at 5.0 (10.61 ± 0.84 mm) to pH 5.6 (11.79 ± 0.61 mm).

5.0

4.0

3.0

2.0

1.0 Initial Spot Width (mm) Width Spot Initial

0.0 KR R101 R590 R610 SR Analyte

Figure 3.2. Initial spot widths of KR, R101, R590, R610, and SR deposited onto the SiO2/PVP UTLC stationary phase with 100 µm Chemyx Nanojet (■), 250 µm hand- spotted (■), and 500 µm hand-spotted (■) method.

% IPA 50 40 30 20 10 0

8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 Migration Distance (mm) Distance Migration 0.0 0 10 20 30 40 50 % ACN

Figure 3.3. Migration distance of laser dyes versus ACN and IPA concentrations: KR (♦), R590 (▲), R101 (■), R610 (+), and SR (●). Conditions are 60 second analysis time, applied potential of 1000 V, 50% by volume 35 mM citrate buffer (pH 5.6) and 50% of combination of ACN and IPA.

8.0

7.0 6.0 5.0 4.0 3.0 2.0

Migration Distance (mm) Distance Migration 1.0 0.0 0 250 500 750 1000 Applied Potential (V)

Figure 3.4. Migration distance of KR (♦), R590 (▲), R101 (■), R610 (+), and SR (●) for 60 second analysis time with 250, 500, and 750 V of applied potential. Mobile phase consists of 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA.

16.0

14.0 12.0 10.0 8.0 6.0 4.0

Migration Distance (mm) Distance Migration 2.0 0.0 0 50 100 150 200 Time (seconds)

Figure 3.5. Migration distance of KR (♦), R590 (▲), R101(■), R610 (+), and SR (●) with 750 V of applied potential for 60, 120, and 180 seconds. Mobile phase consists of 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA.

16.0

14.0 12.0 10.0 8.0 6.0 4.0

Migration Distance (mm) Distance Migration 2.0 0.0 4.0 4.5 5.0 5.5 6.0 6.5 Citrate Buffer pH

3.3.2 Efficiency

The laser dye separation was explored in terms of efficiency; in order to further determine the effects that the experimental conditions place on the separation. Figure

3.11-Figure 3.11 shows the efficiency for all five analytes tested depicted in a surface plot diagram under various applied voltages, with differing mobile phases, and analysis times. Better chromatographic conditions are determined with the elevated position surface position of the plot across all analytes. As the migration distance of the analytes increases, efficiency also increased. Efficiency is shown in Equation 3.2:

푍 2 푁 = 16 ( 푠) (Eq. 3.2) 푤 where Zs is the migration distance of the analyte and w is the spot width of the developed analyte. Larger efficiencies were gained for the analytes that traveled farther along the plate. Analysis time of 60 seconds resulted in lower calculated efficiency for all analytes.

While 750 V provided larger efficiencies than at other voltages. This is primarily attributed to lower voltages producing lower migration distance while with higher applied voltages spot widths are increased. Efficiencies as large as 4200 were generated for 750

V, 120 sec, and 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA for R590.

A peltier device was explored to limit the production of Joule heat associated mobile phase evaporation in the PEC setup. The peltier device was held at 27.5 °C during analysis. The addition of the peltier device resulted in recording much lower observed efficiencies for all analytes. Comparing the same parameters for the efficiencies produced without a peltier device to with a peltier device results in 4200 to 100 for R590 with 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA, 750 V, 120 82 seconds. The utilization of the peltier device allowed for a longer time of analysis. The most likely explanation stems from more mobile phase staying in the system because of the reduced Joule heating and therefore reduced evaporation rate with the added peltier device. The largest calculated efficiency for the peltier device was 1000 for R101, R590, and R610 for 300 second sample time with 500 V applied potential (50:37.5:12.5

(v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA). The pelier device was eliminated because the addition of the peltier device did not produce superior efficiency data compared to that without a peltier device.

Efficiency for the optimized PEC conditions was compared to UTLC analysis from Table 3.2. Efficiencies were greater in UTLC compared to PEC analysis for analytes with lower migration distances (KR and SR). However, the relative standard deviation was significantly lower for the analytes with larger migration distances (R101,

R590, and R610).

Figure 3.7. Surface plot for KR reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■) 2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis.

Figure 3.8. Surface plot for R101 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■) 2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis.

Figure 3.9. Surface plot for R590 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■) 2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis.

Figure 3.10. Surface plot for R610 reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■) 2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis.

Figure 3.11. Surface plot for SR reported in terms of calculated efficiency: Y-axis is efficiency (■) 0-500, (■) 500-1000, (■) 1000-1500, (■) 1500-2000, (■) 2000-2500, (■) 2500-3000, (■) 3000-3500, (■) 3500-4000, (■) 4000-4500. X-axis is applied voltage. Z- axis is mobile phase composition with 50 vol% citrate buffer and the remaining composition of ACN:IPA shown with analysis time in parenthesis.

Table 3.2. Efficiency, N of PEC and UTLC comparison with 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA mobile phase with a 120 sec analysis time. PEC UTLC Analyte NAVE NSTDEV NRSD (%) NAVE NSTDEV NRSD (%) SR 197 162 82.0 873 354 40.5 KR 43 34 79.8 558 443 79.4 R101 1311 329 25.1 919 1036 112.7 R590 4205 642 15.3 1064 897 84.3 R610 1278 269 21.1 276 198 71.8

3.3.3 UTLC Comparison

Figure 3.12 shows individually hand-spotted analytes on a SiO2/PVP UTLC plate after analysis. Figure 3.12A and Figure 3.12B shows PEC and UTLC with 50:37.5:12.5

(v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA for 120 second analysis. The decrease in migration distance for KR and SR in PEC compared to UTLC is observed. This is due to the more negatively charged KR and SR (pka <1.5 and <2.0) analytes progressing more slowly toward the cathode, with EOF. It was not possible to successfully separate all five analytes from one another in a mixture with UTLC or PEC analysis. Two pairs of analytes (KR, SR and R101, R610) migrate similarly, and coelute with PEC analysis, shown in Figure 3.12A. A mixture of KR, R101, and R590 was prepared for PEC analysis. A successful separation was performed with KR, R101, and R590 with 750 V of applied potential and 50:37.5:12.5 (v:v:v%) 35 mM citrate buffer (pH 5.6):ACN:IPA for 120 seconds. The chromatogram of the separation is shown in Figure 3.13.

Resolution of 4.0 was calculated for the KR and R101 analyte pair while 1.0 for R101 and R590.

A B

R101

Intensity R590 KR

0 10 20 30 40 50 60 Separation Distance (mm)

3.4 Conclusions

Optimal mobile phase conditions were determined with five laser dyes as analytes. PEC analysis was conducted on an electrospun SiO2/PVP stationary phase and compared to UTLC analysis with migration distance and efficiency calculations. PEC applied voltage and time of analysis were also studied and optimized for the five laser dyes. The presence of EOF with PEC analysis was shown for the KR and SR laser dyes.

Analytes migrating further, under EOF conditions, in the PEC device produced a much lower relative standard deviation for efficiency compared to UTLC analysis. A separation of three laser dyes utilizing EOF was acheived. Three of the laser dyes analytes were shown separated with a resolution of 4.0 and 1.0 between analyte pairs.

3.5 References

[1] N.A. Ismailov, M.S. Schraiber, Farmatsiya (Sofia) (1938) 1, in : J.G. Kirchner

(Ed.), Thin-Layer Chromatography, second edition, Wiley, New York, 1978.

[2] G. Guiochon, A. Siouffi, Study of the performances of thin layer chromatography

III. flow velocity of the mobile phase, J. Chromatogr. Sci,. (1978), 16, 598-609.

[3] Sz. Nyiredy, in: Sz. Nyiredy (Ed.), Planar Chromatography: A Retrospective View

for the Third Millennium, Springer, Budapest, 2001, pp 177-199.

[4] E. Tyihak, E. Mincsovics, in: Sz. Nyiredy (Ed.), Planar Chromatography: A

Retrospective View for the Third Millennium, Springer, Budapest, 2001, p. 137-

176.

[5] V. Pretorius, B.J. Hopkins, J.D. Schieke, Electroosmosis. New concept for high

speed liquid chromatography, J. Chromatogr., (1974), 99, 23-30.

[6] D. Nurok, M.C. Frost, C.L. Pritchard, D.M. Chenoweth, The performance of planar

chromatography using electroosmotic flow, J. Planar Chromatogr. - Mod. TLC,

(1998), 11, 244-246.

[7] A.G. Howard, T. Shafik, F. Moffatt, I.D. Wilson, Electroosmotically driven thin-

layer electrochromatograpy on silica media, J. Chromatogr. A, (1999), 844, 333-

340.

[8] D. Nurok, M.C. Frost, D.M. Chenoweth, Separation using planar chromatography

with electroosmotic flow, J. Chromatogr. A, (2000), 903, 211-217.

[9] D. Nurok, J.M. Koers, D.A. Nyman, W.-M. Liao, Variables that affect performance

in planar chromatography with electroosmotic flow, J. Planar Chromatogr. Mod. -

TLC, (2001), 14, 409-414.

[10] D. Nurok, J.M. Koers, M.A. Carmichael, T.A. Dzido, The performance of planar

electrochromatography in a horizontal chamber, J. Planar Chromatogr. - Mod.

TLC, (2002), 15, 320-323.

[11] D. Nurok, J.M. Koers, M.A. Carmichael, Role of buffer concentration and applied

voltage in obtaining a good separation in planar electrochromatography, J.

Chromatogr. A, (2003), 983, 247-253.

[12] T.H. Dzido, R. Majewski, B. Polak, W. Golkiewicz, E. Soczewinski, Application

of a horizontal DS chamber to planar electrochromatography, J. Planar.

Chromatogr. - Mod. TLC, (2003), 16, 176-182. 94

[13] T.H. Dzido, R. Majewski, B. Polak, W. Golkiewicz, E. Soczewinski, Application

of a horizontal DS chamber to planar electrochrmotography, J. Planar Chromatogr.

- Mod. TLC, (2003), 16, 176-182.

[14] A. Sze, D. Erickson, L. Ren, D. Li, Zeta-potential measurement using the

Smoluchowski equation and the slope of the current-time relationship in

electroosmotic flow, J. Colloid Interface Sci., (2003), 261, 402-410.

[15] T. Shafik, A.G. Howard, F. Moffatt, I.D. Wilson, Evaporation-induced solvent

migration in electrically-driven thin layer chromatography, J. Chromatogr. A,

(1999), 841, 127-132.

[16] L. Botz, S. Nagy, B. Kocsis, in: Sz. Nyiredy (Ed.), Planar Chromatography: A

Retrospective View for the Third Millennium, Springer, Budapest, 2001, pp 489-

516.

[17] J.E. Clark, S.V. Olesik, Technique for ultrathin layer chromatography using an

electrospun, nanofibrous stationary phase, Anal. Chem., (2009), 81, 4121-4129.

[18] A.L. Andrady, Science and Technology of Polymer Nanofibers; John Wiley &

Sons, Inc.: Hoboken, NJ, 2008.

[19] J.E. Clark, S.V. Olesik, Electrospun glassy carbon ultra-thin layer chromatography

devices, J. Chromatogr. A, (2010), 1217, 4655-4662.

[20] M.C. Beilke, J.W. Zewe, S.V. Olesik, Aligned electrospun nanofibers for ultra-thin

layer chromatography, Anal. Chim. Acta, (2013), 761, 201-208.

[21] P. Kampalanonwat, P. Supaphol, G.E. Morlock, Electrospun nanofiber layers with

incorporated photoluminescence indicator for chromatography and detection of

ultraviolet-active compounds, J. Chromatogr. A, (2013), 1299, 110-117.

[22] X. Fang, S.V. Olesik, Carbon nanotube and carbon nanorod-filled polyacrylonitrile

electrospun stationary phase for ultrathin layer chromatography, Anal. Chim. Acta,

(2014), 830, 1-10.

[23] T.E. Newsome, S.V. Olesik, Silica-based nanofibers for electrospun ultra-thin layer

chromatography, J. Chromatogr. A, (2014), 1364, 261-270.

[24] T. Lu, S.V. Olesik, Electrospun polyvinyl alcohol ultra-thin layer chromatography

of amino acids, J. Chromatogr. B, (2013), 912, 98-104.

[25] T. Rojanarata, S. Plianwong, K. Su-uta, P. Opanasopit, T. Ngawhitunpat,

Electrospun cellulose acetate nanofibers as thin layer chromatographic media for

eco-friendly screening of steroids adulterated in traditional medicine and

nutraceutical products, Talanta, (2013), 15, 208-213.

[26] T.H. Dzido, P.W. Plocharz, P. Slazak, A. Halka, Progress in planar

electrochromatography, Anal. Bioanal. Chem., (2008), 391, 2111-2118.

[27] T.E. Newsome, S.V. Olesik, Electrospinning silica/polyvinylpyrrolidone composite

nanofibers, J. Appl. Polym. Sci., (2014), 131, 40966(1)-40966(9).

[28] T.E. Newsome, S.V. Olesik, Silica-based nanofibers for electrospun ultra-thin layer

chromatography, J. Chrom. A, (2014), 1364, 261-270.

[29] A.V.I. Hess, Digitally enhanced thin-layer chromatography: An inexpensive, new

technique for qualitative and quantitative analysis, J. Chem. Educ., (2007), 84, 842-

847.

[30] T.E. Newsome, S.V. Olesik, Planar electrochromaography using an electrospun

polymer nanofiber layer, Anal. Chem., (2014), 86, 10961-10969.

[31] W.F. Patton, V. Panchagnula, E. Rockney, and I.S. Krull, Taking a walk on the

wild side with planar electrochromatography and thin-layer electrophoresis: of

peptides, proteins, and proteomics, J. Liq. Chromatogr. Relat. Technol., (2006), 29,

1177-1218.

CHAPTER 4: CHEMILUMINESCENT NANOFIBER CHARACTERIZATION AND DETERMINATION OF NUCELOBASE CONCENTRATIONS

4.1 Introduction

2+ The use of tris(2,2’bipyridyl)ruthenium(II) (Ru(bpy)3 ) on an electrode surface has become a promising method and is widely used in electrochemiluminescenct sensors.

2+ The utilization of Ru(bpy)3 in an electrode platform is cost effective, sensitive, and often able to be reused [1,2,3]. Perhaps the main advantage gained from using a

2+ Ru(bpy)3 electrochemiluminescent system is its adaptability and suitability because of

2+ its high solubility in aqueous environments [4]. Ru(bpy)3 has been used to analyze compounds that are H2O soluble and has shown to be effective for analysis of a wide range of compounds such as oxalate, alkylamines, and amino acids [5,6,7]. Several

2+ efforts have been made to immobilize Ru(bpy)3 onto electrode surfaces using different methods. A few of these methods include self-assembled monolayers, layer-by-layer assembly, and covalent immobilization [8,9,10,11]. However, these approaches tend to be either time consuming or involve complicated fabrication processes. So, it is desirable to develop a facile and inexpensive approach for fabrication of an electrochemiluminescent sensor that possesses high surface area and can lead to large

2+ quantities of immobilized Ru(bpy)3 .

The electrospinning technique has proven to be a facile, versatile and a cost

effective strategy for developing polymeric fibrous membranes with large surface areas

[12,13,14,15,16,17]. Examples of fibrous membranes that show great potential for

sensors with high sensitivity are plentiful. Polydiacetylene embedded silica fibers have

been electrospun and used as a sensor for volatile organic compounds [14]. Electrospun

fluorescent polymer nanofibers have been designed and fabricated as a highly sensitive

A optical sensor for the analysis of metal ions and 2,4-dinitrotoluene. The Stern–Volmer

constant values of these sensors are two to three orders of magnitude larger than those by

thin film methods with the same sensing material [15]. A fluorescent conjugated polymer

has been synthesized and fabricated into electrospun fibers that can sense low ppb

concentration of cytochrome c and methyl viologen in aqueous solutions [16]. Not only

2+ has electrospun nanofibers been utilized in sensors, but the use of Ru(bpy)3 has

previously been reported by Moran-Mirabal et al. [17]. Moran-Mirabal created light

2+ generating nanofibers from a spinneret solution containing Ru(bpy)3 and polyethylene

oxide.

2+ A cation-exchanger, Nafion, is the immobilization material for Ru(bpy)3

through electrostatic interactions. The sensing electrode is fabricated from

2+ Ru(bpy)3 /Nafion nanofibers by the electrospinning technique. As an interesting and

2+ useful ion-exchange material for Ru(bpy)3 , Nafion has been widely used in

electrochemistry [18]. Nafion has been shown to possess a large ion-exchange capacity

2+ (1.00 ± 0.02 meq/g) [19]. Its ion-exchange selectivity coefficient (vs. Na+) for Ru(bpy)3

is very large (5.7 x 106) because the sulfonic acid side chains can strongly bind to

2+ Ru(bpy)3 [20]. However, pure Nafion can only be electrosprayed with bead formation rather than electrospun into long semi-continuous nanofibers. With the addition of a second polymer leading to proper chain entanglement electrospinning occurs. Nafion and a second polymer, polyacrylic acid (PAA), blended solution can be electrospun [21].

2+ Chen et al. demonstrated that with only a little PAA, the Ru(bpy)3 complex, and Nafion polymer solution can be electrospun into nanofibers.

The surface area to volume ratio is increased with electrospinning compared to a

2+ thin film layer. Together, the combination of immobilized Ru(bpy)3 onto the ion- exhange nanofibers will create an electrochemiluminescent sensing platform that shows a strong electrochemiluminescent signal, and thus a high sensitivity in molecular detection.

So a great potential for electrospun Nafion nanofiber based electrochemiluminescent sensors is described for determination of guanine in solution.

4.2 Experimental

4.2.1 Reagents and Instrumentation

Polyacrylic acid (PAA), (molecular weight = 450,000 g/mol), Nafion (molecular weight = 1,100 g/mol) perfluorinated resin solution (5 wt%) in lower aliphatic alcohols and H2O, contains 15-20% H2O was purchased from Sigma Aldrich (St. Louis, MO).

N,N-dimethylformamide, guanine, 2-propanol (IPA), and tris(2-2’bypyridylruthenium(II) hexahydrate was also purchased from Sigma Aldrich. Methanol from Avantor

Performance Materials (Center Valley, PA) was also utilized. Potassium chloride, potassium monobasic, and potassium dibasic was purchased from Fisher Scientific (Fair

100

Lawn, NJ). Use of nanopure H2O was made possible with purification through a

Barnstead Nanopure Infinity system (Thermo Scientific, Asheville, NC).

An Epsilon potentiostat manufactured by BASi (West Lafayette, IN) was used. A glassy carbon electrode (GCE) was procured from CH Instruments (Bee Cave, TX). A

Ag/AgCl reference electrode and Pt counter electrode was also purchased from BASi. A power supply (Orion Industries, Windham, NH), was used to power the PMT during experimentation (1250 V). A Hamamatsu Photonics (Hamamatsu, Japan) photomultiplier tube (PMT) was used to measure the generation of light. A Keithley

Instruments (Solon, OH) current amplifier was used to increase the signal produced from the PMT (108 V/A).

The electrospinning apparatus shown in Figure 1.7 consists of the equipment necessary to electrospin the Nafion:PAA composite: (1) Syringe pump, a Harvard Model

33 dual syringe pump (Holliston, MA), (2) Power supply, a Spellman CZE 1000R high voltage source (Hauppauge, NY), (3) Syringe, (10 mL Fisher Scientific), with a one inch twenty-three gauge internal diameter metallic tip from Nordson EFD (East Providence,

RI) and (4) Grounded collector, (aluminum foil or CGE). Heavy Duty aluminum foil

101 from Reynolds Wrap (Richmond, VA) was used for nanofiber characterization with the

SEM instrument.

4.2.2 Light Generating Pathway

Figure 4.1 shows the necessary steps for the generation of light during a four step

2+ co-reactant sequence. The co-reactant sequence begins as Ru(bpy)3 and the analyte of interest become simultaneously oxidized at the electrode surface in the first and second step, respectively. The two oxidized products interact to form an excited state species in

2+ the third step. The formation of the *Ru(bpy)3 from the electron transfer reaction is a very fast and has been measured with cyclic voltammetry with guanine (k ~ 106 M-1 second-1) [22]. The fourth step results in the formation of light (610 nm) as the excited

2+ electron drops back down to the ground state and the Ru(bpy)3 molecule becomes

2+ regenerated. The Ru(bpy)3 is then free to further undergo the reaction sequence again.

2+ With these reactions in mind, it is obvious that the amount of Ru(bpy)3 is critically

2+ important. As long as the Ru(bpy)3 concentration is held constant experimentally then the amount of light that is produced is wholly dependent on the concentration of the analyte of interest.

102

2+ 3+ - 1. Ru(bpy)3  Ru(bpy)3 + e

2. Guanine  [Guanine•]+ e-  Guanine• + H+

• 3+ 2+ 3. Guanine + Ru(bpy)3  *Ru(bpy)3 + products

2+ 2+ 4. *Ru(bpy)3  Ru(bpy)3 + hν

Figure 4.1. Four reaction co-reactant sequence for the generation of light. Step 1. 2+ Ru(bpy)3 becomes oxidized. Step 2. Guanine becomes oxidized. Step 3. Excited state 2+ formation. Step 4. Emission of light and regeneration of Ru(bpy)3 species.

103

4.2.3 Electrospinning

Electrospinning is discussed in further detail in Section 1.3. There are several parameters to consider in the electrospinning process to achieve nanofiberous mats. The variables to consider include: distance from the metallic tip to the grounded collector, the applied voltage, the syringe flow-rate, the humidity of the electrospinning environment, and the spinneret composition. Variables must be optimized with the goal of obtaining nanofibers with the have the smallest, most homogenous nanofiber diameters. Generally, the applied voltage and the distance from the tip of the syringe to the collector provide the largest influence on nanofiber dimensions. Ideally, the solvent is fully evaporated prior to becoming deposited onto the grounded collector in the form of a nanofibrous mat. When the applied voltage becomes too high a decrease in the time the polymer solution has to travel to the collector is achieved. This can have a detrimental effect as not all solvent may be evaporated in the electrospinning process. Similarly, the distance between the spinneret and the collection plate is also an important parameter in that enough time must pass for solvent evaporation to occur before reaching the collector. To determine nanofiber diameters and mat thicknesses at least three electrospinning nanofiber mats were collected and mounted on an SEM sample stage for each of the electrospinning conditions studied. SEM images were generated and at least 15 measurements were made on the three collected samples for a total of 45 measurements for each reported nanofiber diameter value.

4.3 Results and Discussion

104

4.3.1 Optimization of Electrospun Nanofibers

Nafion was used as the ionic exchange polymer. Nafion has a molecular weight of 1,100 g/mol. Polymers with relatively low molecular weights may not provide enough chain entanglement to electrospin individually. A second polymer, with a larger molecular weight, can be incorporated into the spinneret solution to help provide enough chain entanglement for electrospinning to occur [23]. Commercially available Nafion comes dissolved in a combination of lower aliphatic alcohols and H2O. With this in mind, it was decided to focus on utilization of IPA and H2O as the solvent system for electrospinning. Table 4.1 is a compilation of average fiber diameters obtained for various solvent compositions, voltages, and distances studied for the Nafion and PAA composite. As can be seen in the first row of Table 4.1, Nafion did not produce any nanofibers when electrospun individually. Nanofibers are generated as a second polymer, PAA, is incorporated into the spinneret solution. PAA has a molecular weight of 450,000 g/mol to enhance chain entanglement. Nanofiber diameters of 300 ± 100 nm were observed for the Nafion:PAA composite with a 15 cm distance and an electric field of 15 kV. A variation of fiber diameter with solvent composition was observed as the portion of H2O in the solvent was increased. Fiber diameters increase as more H2O was incorporated into the sample. This is most likely due to H2O evaporating less quickly than IPA. Because the evaporation rate of H2O is less than IPA, it appears that larger distances are required to generate fibers with solvent compositions possessing larger H2O portions. While the spinneret solution containing 25:75 (w:w%) Nafion:PAA provided the smallest nanofiber diameter (300 ± 100 nm) ultimately, the spinneret solution

105 containing 70:30 (w:w%) Nafion:PAA (500 ± 100 nm) was chosen as most desirable because of the larger proportion of mat weight due to Nafion. The 70:30 (w:w%)

Nafion:PAA diameters with 500 ± 100 nm were observed with an optimized distance of

12.5 cm and 8 kV of applied voltage and can be seen in Figure 4.2.

The solubility of the nanofiber mat must be studied because the nanofiber mat will ultimately be used in an aqueous environment for nucleobase sensing and the spinneret solution contains H2O as part of the solvent. Figure 4.3 is a representative SEM image for a 70:30 (w:w%) Nafion:PAA nanofiber mat that has been soaked in H2O for a period of three days. It is obvious that the morphology of the nanofibers remained largely intact and did not dissolve in the H2O after soaking for three days. However, the nanofibers were swollen after soaking. The average nanofiber diameters increased from

500 ± 100 nm to 650 ± 40 nm. The swelling of the nanofiber diameter is most likely attributed to PAA. Swelling of PAA nanofibers in H2O was noted previously [21]. Fiber swelling was minimized in all further experiments by using the electrospinning mat the same day that it was generated.

The effect of electrospinning time on mat thickness was also studied using the same optimized electrospinning parameters determined above. Electrospun mats were generated at times of 0.50, 0.75, 1.00, 60, 120, and 180 minutes; the thicknesses of the mats were then determined using SEM analysis and subsequent measurements with

ImageJ software. Figure 4.4 illustrates the relationship between electrospinning time and the thickness of the electrospun mat. Mat thicknesses range from 1.7 ± 0.8 µm to 31.1 ±

1.1 µm across all electrospinning times studied. Mat thickness samples were collected

106 through cutting the aluminum collector and adhered electrospun nanofiber mat with

Westcott 5” Titanium scissors that were previously cleaned with IPA. Cleaning the scissors with IPA prior to cutting was necessary because uncleaned scissors resulted in removal of the nanofiber mat from the aluminum collector before the mat could be completely cut. Additionally, cutting mat thickness samples with razor blades, X-acto knifes, or utility knifes were eliminated because pulling and tearing of the nanofiber mat occurred during the cutting process. While it is very difficult to determine the mat thickness without compromising the nanofiber morphology due to compression with the cutting device the mat thickness increases linearly with an increase in electrospinning time (R2=0.985). This suggests that the final mat thickness obtained can be altered by simply changing the length of time spent electrospinning. Figure 4.5 shows a magnified portion detailing the lower electrospinning times (0.50, 0.75, and 1.00 minute). Figure

4.6 contains a representative SEM image of optimized electrospun nanofibers displaying mat thickness with 180 minute electrospinning time (Scale bar = 25 µm).

107

Table 4.1. Nanofiber diameters of the observed mat containing a blend of PAA with Nafion polymers. Mat-Polymer Solvent Distance Electric Field Average Fiber (w:w%) (w:w%) (cm) (kV) Diameter (nm)

Pure Nafion 67:33 IPA:H2O 12.5 8 None

25:75 Nafion:PAA 67:33 IPA:H2O 15.0 15 300 ± 100

50:50 Nafion:PAA 67:33 IPA:H2O 17.5 15 500 ± 200

50:50 Nafion:PAA 50:50 IPA:H2O 18.0 20 800 ± 200

50:50 Nafion:PAA 50:50 IPA:H2O 25.0 20 1000 ± 200

70:30 Nafion:PAA 67:33 IPA:H2O 12.5 8 500 ± 100

108

Figure 4.2. Representative SEM image of an electrospun nanofiber mat generated with 70:30 (w:w%) Nafion:PAA under 12.5 cm and 8 kV (Scale bar = 10 µm).

109

Figure 4.3. Representative SEM image of an electrospun nanofiber mat generated with 70:30 (w:w%) Nafion:PAA under 12.5 cm and 8 kV that was soaked in H2O for a period of three days (Scale bar = 10 µm).

110

40 35

30 25 20 15

Thickness (µm) Thickness 10 5 0 0 50 100 150 200 Time (min)

Figure 4.4. Mat thickness plotted against electrospinning time (n=3) (R2 = 0.985).

111

4.0 3.5

3.0 2.5 2.0 1.5

Thickness (µm) Thickness 1.0 0.5 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Time (min)

Figure 4.5. Mat thickness plotted against electrospinning time. Magnified view of lower electrospinning times (0.50, 0.75, and 1.00 minute).

112

Figure 4.6. Representative SEM image of optimized electrospun nanofibers with mat thickness generated with 180 minute electrospinning time (Scale bar = 25 µm).

113

4.3.2 Modified GCE Surface

The co-reactant sequence shown in Figure 4.1 illustrates the necessity for

2+ Ru(bpy)3 to be in close proximity to the electrode surface. The spinneret solution

2+ containing Nafion, PAA, and Ru(bpy)3 is electrospun directly onto the GCE tip.

2+ Ru(bpy)3 remains present on the electrode through ion-exchange interactions with the sulfonic acid side chains present on Nafion. After electrospinning, to ensure complete

2+ 2+ coverage of Ru(bpy)3 , the electrode is then soaked in a solution containing Ru(bpy)3

2+ prior to analysis. Figure 4.7 demonstrates the necessity for the addition of Ru(bpy)3 to the electrode surface for guanine oxidation as measured by cyclic voltammetry. Plotted in Figure 4.7 is a bare GCE, thin film of Nafion:PAA, and electrospun Nafion:PAA electrodes measuring the oxidation signal of a 1.0 x 10-4 M guanine solution. The

2+ oxidation of both guanine and Ru(bpy)3 is observed in the voltammogram at the peak at

1.1 V. Cyclic voltammograms were collected by varying the potential from 0.80 V to

1.35 V then back to 0.80 V while measuring the resulting current. It is important to notice the general size of the peaks in relation to one another. It is clear that the bare

GCE does not produce as much current as either the thin film or electrospun Nafion:PAA electrode. The electrospun electrode performance was superior to that of the thin film electrode. This would suggest that an improvement is garnered by electrospinning the polymer mixture over that of simply using a thin film for sensing. Improvements may be provided by generating a higher surface area electrospun material on the electrode and/or obtaining an increase in diffusion through the electrospun material.

114

Studies were conducted to further optimize the modified GCE surface by determining the thickness of the nanofibrous polymer mat on the electrode. Polymer mixtures were electrospun directly to the GCE. Table 4.2 shows results for modified electrodes that were electrospun for 0.50, 0.75, and 1.00 minute. The data was collected

2+ using an analyte solution of 50 mM Ru(bpy)3 to ensure enough analyte is present to the electrode for oxidation. Peak currents were measured using cyclic voltammetry across multiple scan rates (50, 300, 400, 800, 2000 mV/second) for each of the thicknesses tested. Peak currents should increase linearly with the square root of the scan rate used to yield an electrode that is only limited by diffusion of the analyte. It appears that the peak current increases with electrospinning time until a critical point where peak current then decreases because the analyte does not diffuse effectively with too much polymer material applied to the electrode. Electrospinning for 0.75 minute garnered the highest peak current with a scan rate of 2000 mV/second over the others tested while also

2 providing the best R values for linearity for peak current, Ip values plotted against the square root of scan rates.

115

0.10 0.08 0.06 0.04 0.02 0.00

Current (mA) Current -0.02 -0.04 -0.06 -0.08 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Potential (V) vs Ag/AgCl

Figure 4.7. Cyclic voltammogram collected with 100 mV/sec scan rate for a bare CGE electrode (■), TF Nafion:PAA (■), electrospun Nafion:PAA (■) with oxidation of 1.0 x 10-4 M guanine.

116

Table 4.2. Electrodes generated from electrospinning polymer solution at 0.50, 0.75, and 2+ 1.00 minute. Peak current, Ip determined for a 50 mM solution of Ru(bpy)3 . Electrospinning Mat thickness I (mA) R2 time (minute) p (µm) 0.50 0.27 0.960 1.72 ± 0.81 0.75 0.34 0.990 2.79 ± 0.34 1.00 0.17 0.966 3.22 ± 0.34

117

4.3.3 Guanine Determination

Guanine, one of the four main nucleobases, contained in DNA’s nucleic acids.

The electrochemiluminescent determination of guanine with a Nafion:PAA modified

GCE surface is shown in Figure 4.8. The electrochemistry is performed with a standard three electrode system of a working electrode (modified GCE), counter electrode (Pt),

2+ and a reference electrode. Oxidation of Ru(bpy)3 and guanine was provided by the electrochemical technique of differential pulse voltammetry where a potential scan was used beginning at 800 mV and ending at 1500 mV with a step of 4 mV. The pulse width used was 70 msec, pulse period of 200 msec, and pulse amplitude of 50 mV. The blank

(phosphate buffer pH 7.0) electrochemiluminescent signal is subtracted from the signal associated with each of the studied guanine concentrations (0.01 mM, 0.001 mM, and

0.0001 mM) and plotted against the analyte concentration. The electrochemiluminescent signal was collected by a PMT that was placed in close proximity to the reaction vessel containing the three electrodes and analyte solution. The electrochemiluminescent signal increases linearly (R2 = 0.999) as the guanine concentration is increased. More concentrated guanine samples could not be tested because the solubility of guanine would not allow for samples greater than 0.01 mM. An electrochemiluminescent signal for samples more dilute than 1.0 x 10-7 mM guanine could not be quantified over the blank signal.

118

1.5E+06

1.0E+06

Intensity 5.0E+05

0.0E+00 0 0.002 0.004 0.006 0.008 0.01 0.012 Concentration (mM)

Figure 4.8. Electrochemiluminescent signal obtained from (1.0 x 10-5 M, 1.0 x 10-6 M, and 1.0 x 10-7 M) guanine solutions.

119

4.4 Conclusions

Electrospinning conditions for the generation of 70:30 (w:w%) Nafion:PAA polymer composite nanofibers were determined to be 8 kV of applied voltage and a tip to collector distance of 12.5 cm. The electrospinning conditions correspond to producing fibers with 500 ± 100 nm diameters. Further optimization of a modified GCE surface was determined through varying electrospinning time while measuring mat thicknesses of a composite of 70:30 (w:w%) Nafion:PAA nanofiber mat. An optimized nanofiber mat thickness of 2.79 ± 0.34 µm resulted in an electrospinning time of 0.75 minute. Use of the electrospun modified GCEs were then employed to determine the amount of guanine in solution with the generation of electrochemiluminescent signal monitored with a PMT.

This research demonstrates the successful utilization of electrochemiluminescent nanofibers for the determination of guanine concentrations in solution. Good linearity was observed between 1.0 x 10-5 M to 1.0 x 10-7 M guanine concentrations and measuring electrochemiluminescent signal.

4.5 References

[1] Z. Ding, B.M. Quinn, S.K. Haram, L.E. Pell, B.A. Korgel, A.J. Bard,

Electrochemistry and electrogenerated chemiluminescence from silicon nanocrystal

quantum dots, Science, (2002), 296, 1293-1297.

[2] W.J. Miao, Electrogenerated chemiluminescence and its biorelated applications,

Chem. Rev., (2008), 108, 2506-2553.

120

[3] M.M. Richter, Electrochemiluminescence (ECL), Chem. Rev., (2004), 104, 3003-

3036.

[4] I. Rubinstein, A.J. Bard, Electrogenerated chemiluminescence. 37. Aqueous ecl

2+ systems based on Ru(2,2’-bipyridine)3 and oxalate or organic acids, J. Am Chem.

Soc., (1981), 103, 512-516.

[5] I. Rubinstein, A.J. Bard, Electrogenerated chemiluminescent determination of

oxalate, Anal. Chem., (1983), 55, 1580-1582.

[6] J.B. Noffsinger, N.D. Danielson, Generation of chemiluminescence upon reaction

of aliphatic amines with tris(2,2’-bipyridine)ruthenium(III), Anal. Chem., (1987),

59, 865-868.

[7] S.N. Brune, D.R. Bobbitt, Effect of pH on the reaction of tris(2,2’-

bipyridyl)ruthenium(III) with amino-acids: implications for their detection,

Talanta, (1991), 38, 419-424.

[8] C.J. Miller, P. McCord, A.J. Bard, Study of Langmuir monolayers of ruthenium

complexes and their aggregation by electrogenerated chemiluminescence,

Langmuir, (1991), 7, 2781-2787.

[9] Z.H. Guo, Y. Shen, M.K. Wang, F. Zhao, S.J. Dong, Electrochemistry and

electrogenerated chemiluminescence of SiO2

nanoparticles/tris(2,2’bipyridyl)ruthenium(II) multilayer films on indium tin oxide

electrodes, Anal. Chem., (2004), 76, 184-191.

[10] S. Zanarini, E. Rampazzo, L. Della Ciana, M. Marcaccio, E. Marzocchi, M.

Montalti, F. Paolucci, L. Prodi, Ru(bpy)3 covalently doped silica nanoparticles as 121

multicenter tunable structures for electrochemiluminescence amplification, J. Am.

Chem. Soc., (2009), 131, 2260-2267.

[11] X.B. Yin, S.J. Dong, E.I. Wang, Analytical applications of the

electrochemiluminescence of tris(2,2’-bipyridyl) ruthenium and its derivatives,

Trac-Trend Anal. Chem., (2004), 23, 432-441.

[12] F. Di Benedetto, A. Camposeo, S. Pagliara, E. Mele, L. Persano, R. Stabile, R.

Clingolani, D. Pilignano, Patterning of light-emitting conjugated polymer

nanofibers, Nat. Nanotechnol., (2008), 3, 614-619.

[13] D. Yang, B. Lu, Y. Zhao, X. Jiang, Fabrication of aligned fibrous arrays by

magnetic electrospinning, Adv. Mater., (2007), 19, 3702-3706.

[14] J. Yoon, S.K. Chae, J.M. Kim, Colorimetric sensors for volatile organic compounds

(VOCs) based on conjugated polymer-embedded electrospun fibers, J. Am. Chem.

Soc., (2007), 129, 3038-3039.

[15] X. Wang, C. Drew, S.H. Lee, K.J. Senecal, J. Kumar, L.A. Samuelson, Electrospun

nanofibrous membranes for highly sensitive optical sensors, Nano Lett., (2002), 2,

1273-1275.

[16] X. Wang, Y.G. Kim, C. Drew, B.C. Ku, J. Kumar, L.A. Samuelson, Electrostatic

assembly of conjugated polymer thin layers on electrospun nanofibrous membranes

for biosensors, Nano Lett., (2004), 4, 331-334.

[17] J.M. Moran-Mirabal, J.D. Slinker, J.A. DeFranco, S.S. Verbridge, R. Llic, S.

Flores-Torres, H. Abruna, G.G. Malliaras, H.G. Craighead, Electrospun light-

emitting nanofibers, Nano Lett., (2007), 7, 458-463. 122

[18] D.J. Kim, Y.K. Lyu, H.N. Choi, I.H. Min, W.Y. Lee, Nafion-stabilized magnetic

2+ nanoparticles (Fe3O4) for [Ru(bpy)3] (bpy = bipyridine) electrogenerated

chemiluminescence sensor, Chem. Commun., (2005), 23, 2966-2968.

[19] T.Y. Chen, J. Leddy, Ion exchange capacity of Nafion and Nafion composites,

Langmuir, (2000), 16, 2866-2871.

[20] M.N. Szentirmay, C.R. Martin, Ion-exchange selectivity of Nafion films on

electrode surfaces, Anal. Chem., (1984), 56, 1898-1902.

[21] H. Chen, J.D. Snyder, Y.A. Elabd, Electrospinning and solution properties of

Nafion and poly(acrylic acid), Macromolecules, (2008), 41, 128-135.

[22] D.H. Johnston, H.H. Thorp, Cyclic voltammetry studies of polynucleotide binding

and oxidation by metal complexes: homogeneous electron-transfer kinetics, J.

Phys. Chem., (1996), 100, 13837-13843.

[23] A. Laforgue, L. Robitaille, A. Mokrini, A. Ajji, Fabrication and characterization of

ionic conducting nanofibers, Macromol. Mater. Eng., (2007), 292, 1229-1236.

123

CHAPTER 5: “GREEN” SEPARATIONS THROUGH ENHANCED-FLUIDITY LIQUID CHROMATOGRAPHY GRADIENTS

5.1 Introduction

In the past 20 years, there has been a significant push towards the development and implementation of environmentally sustainable chemistry practices. Often referred to as “green chemistry,” is the design of chemical products and processes which reduce or eliminate the use and generation of hazardous substances [1] and has been defined by a set of 12 principles [2]. These principles serve as a guideline to minimize the environmental risk of any chemical process and can be applied across all disciplines, including high-performance liquid chromatography (HPLC).

The largest environmental concern in HPLC is its heavy use of toxic, organic solvents. A single analytical chromatograph with a conventional column (4.6 mm x 150 mm) produces upwards of 500 liters of solvent waste per year [3]. To offset this, “green” chromatographic methods can be designed to minimize the generation of hazardous solvent waste while maintaining chromatographic performance. Smaller scale systems with shorter column dimensions and reduced particle diameters can lessen the required solvent volume [4], however conventional systems are generally limited to columns at least 2.1 mm in diameter. Portions of solvents can also be recovered and reused; however, this is generally only feasible for isocratic separations of clean samples. While these techniques can be successful, the most direct method of “greening” chromatography 124 is simply to develop methods that utilize “green” solvents without compromising chromatographic performance. Solvent selection guides have even been developed to aid chromatographers in making “green” solvent choices for their separations [5,6].

Hydrophilic interaction chromatography (HILIC) is an area of chromatography that could benefit greatly from “greener” methods. HILIC has long been established as the preferred chromatographic mode for the separation of highly polar and ionizable compounds. First proposed in 1990, HILIC separations are achieved via a partitioning mechanism in which analytes partition between a H2O-enriched layer on the surface of a polar stationary phase and a highly organic mobile phase [7]. As a result, polar compounds preferentially partition into the H2O-rich layer and are well-retained.

Additionally, the presence of large amounts of volatile organic solvents in the mobile phase make HILIC well suited for liquid chromatography-mass spectrometry (LC-MS) analysis. Together, these advantages have made HILIC methods extremely popular for difficult separations of polar biological compounds [8].

Unfortunately, HILIC methods typically rely heavily on acetonitrile (ACN) as a mobile phase component. ACN is already one of the most common solvents used throughout HPLC because of its highly desirable chromatographic properties, including low viscosity, high solvating power, and low UV cutoff. Additionally, its aprotic character makes it a very weak eluent in HILIC separations, yielding excellent retention of polar compounds. As a result, the vast majority of HILIC separations are performed using ACN as the organic portion of the mobile phase. However, when compared to other commonly used chromatographic solvents, ACN is much more expensive and much

125 less environmentally friendly. Given the recent drive towards sustainable chemistry, development of quality HILIC methods using “greener” solvents would be highly beneficial.

Few attempts have been made to minimize/eliminate ACN from HILIC mobile phases for the purpose of “greener” chromatography. Typically primary alcohols are used as a substitute, however the polar protic nature of these solvents results in disruption of the H2O layer [9] and drastically decreased retention relative to ACN [10,11,12].

Others have performed “reversed HILIC” separations, which use small portions of ACN with large amounts of aqueous mobile phase [13]. However, this is really just aqueous reversed-phase chromatography on a HILIC column. Additionally, given the recent revival of supercritical fluid chromatography (SFC) due to its “green” character, supercritical fluids have been used as mobile phases with HILIC columns [14,15,16,17].

However, unlike traditional HILIC methods, H2O is not typically a component of the mobile phase.

Recently, our group [18,19,20] and others [21] have demonstrated the use of enhanced-fluidity liquids (EFLs) as “green” mobile phases in HILIC separations. These mobile phases are liquid mixtures (usually a primary alcohol and H2O) to which high proportions of a liquefied gas (usually CO2) have been added [22]. This makes EFLs similar to SFC mobile phases, but CO2 is used as the modifier rather than the primary solvent. By adding nonpolar CO2 as a modifier in HILIC, the eluent strength of alcohol:H2O mobile phases can be decreased significantly, yielding retention similar to traditional ACN:H2O mobile phases. Even with lower CO2 proportions, these mobile

126 phases offer the chromatographic advantages similar to SFC mobile phases, including enhanced solute mass transfer and reduced system backpressure [23]. This often leads to improved chromatographic performance in terms of efficiency, resolution, and separation time. Additionally, allows for analysis of highly polar compounds. Furthermore, by limiting the mobile phase components to CO2, H2O, and organic solvents like methanol

(MeOH) or ethanol (EtOH), the overall “greenness” of the method is greater than traditional ACN-based HILIC separations.

While these benefits of enhanced-fluidity liquid chromatography (EFLC) have been demonstrated for HILIC, they have only been demonstrated in isocratic separations.

However, samples with analytes possessing wide retention factor ranges often require gradient elution in order to resolve all peaks with a reasonable time of analysis. By performing a reverse CO2 gradient with EFL mobile phases, the HILIC eluent strength of the mobile phase can be varied drastically, allowing for analysis of compounds with a wide range of polarities. Using sixteen RNA nucleosides and nucleotides with a wide range of polarity, this study marks the first time that “green” EFLC mobile phases have been used in gradient separations using commercial SFC instrumentation.

5.2 Experimental

5.2.1 Instrumental

The instrument used for the EFLC analysis was a 1260 Infinity Analytical SFC system (Agilent Technologies, Santa Clara, CA). The system was comprised of a Fusion

A5 SFC control module (600 bar maximum), 1260 HiP degasser, 1260 SFC bin pump,

127

1260 SFC ALS auto-sampler, 1290 thermostatted column compartment (-10°-100° C), and a 1200 DAD SL detector (190-950 nm). The wavelength for analyte detection in the

DAD was set to 270 nm. During all data collection the pressure of the back pressure regulator (BPR) was set to 120 bar and the temperature to 60.0° C. The temperature of the right column mobile phase inlet was set to 40.0° C while the left column mobile phase outlet was set to 37.5° C. A Shimadzu Scientific Instruments (Kyoto, Japan)

HPLC instrument was used for ACN:H2O and MeOH:H2O mobile phase analysis. The

HPLC was composed of two LC-20AT pumps, an SIL-20A auto-sampler, a CTO-20A column oven, an SPD-20A UV/vis detector (270 nm wavelength selection), and a CBM-

20A communications bus module. The column used for all analysis was a 4.6 mm x 150 mm XBridge™ Amide column packed with 3.5 µm particles (Waters, Milford, MA).

5.2.2 Chemicals

Supercritical fluid extraction grade CO2 (99.999% purity) was purchased and used from Praxair, Inc (Danbury, CT). HPLC grade MeOH, HPLC grade ACN, and HPLC grade o-phosphoric acid 85% was purchased from Fisher Scientific (Fair Lawn, NJ). (≥

99%) Sodium phosphate monobasic dihydrate was purchased from Sigma-Aldrich (St.

Louis, MO). Deionized H2O was purified to 18.2 MΩ on a Barnstead Nanopure Infinity system (Thermo Scientific, Asheville, NC). All nucleoside and 5’-nucleotide analytes were purchased as a kit from Sigma-Aldrich (St. Louis, MO). For clarification, the kit is composed of (≥ 99%) adenosine (A), (≥ 99%) cytidine (C), (≥ 98%) guanosine (G), (≥

99%) uridine (U), (≥ 99%) adenosine 5’-monophosphate sodium salt (AMP), (≥ 99%)

128 cytidine 5’-monophosphate disodium salt (CMP), (≥ 99%) guanosine 5’-monophosphate disodium salt hydrate (GMP), (≥ 99%) uridine 5’-monophosphate disodium salt (UMP),

(≥ 95%) adenosine 5’-diphosphate sodium salt (ADP), (≥ 95%) cytidine 5’-diphosphate sodium salt hydrate (CDP), (≥ 96%) guanosine 5’-diphosphate sodium salt (GDP), (95-

100%) uridine 5’-(trihydrogen diphosphate sodium salt (UDP), (≥ 99%) adenosine 5’- triphosphate disodium salt hydrate (ATP), (~ 95%) cytidine 5’-triphosphate disodium salt

(CTP), (≥ 95%) guanosine 5’-triphosphate sodium salt hydrate (GTP), (≥ 96%) uridine

5’-triphosphate trisodium salt hydrate (UTP). All analyte structures are shown in Table

5.1 along with associated log P values calculated using ACD/ChemSketch Freeware from www.acdlabs.com/resources/freeware/index.php [24]. Analytes were dissolved in 80:20

(v:v%) MeOH:H2O. All analyte solutions were prepared to a final concentration of 1.25 x 10-4 M unless otherwise indicated and placed into a Misonix ultrasonic cleaner

(QSonica LLC, Newtown, CT) to aid in the solvation process. When not in use analyte solutions were stored at 4° C to prolong stability.

129

Table 5.1. Name structure and log P values for analytes in their neutral unionized form calculated from ACD/ChemSketch at www.acdlabs.com/resources/freeware/index.php

Name Structure Log P

A -1.02

AMP -2.07

ADP -3.35

ATP -4.62

C -1.94

CMP -1.56

CDP -4.25

CTP -5.52

Continued

130

Table 5.1 continued

Name Structure Log P

U -1.61

UMP -1.58

UDP -4.27

UTP -5.54

G -1.72

GMP -2.28

GDP -3.55

GTP -4.82

131

5.2.3 Mobile Phase Preparation

The Agilent Technologies SFC system is capable of mixing liquefied CO2 from pump A and organic liquid mixtures from pump B at various (v:v%) concentrations. The liquid contents delivered through pump B must be prepared prior to experimentation by mixing the desired amounts of organic solvent and aqueous buffer. The ionic strength of any buffer housed in reservoir B is reported with respect to total volume (prior to CO2 addition from pump A). For example, in order to prepare 300 mL of an 80:20 (v:v%)

MeOH:H2O solution (containing 40 mM sodium phosphate buffer at pH=2.65) , 24 mL of 500 mM sodium phosphate (pH=2.65) stock solution was mixed with an additional 36 mL of H2O and then combined with 240 mL of MeOH. Regular flushing of the system with 80:20 (v:v%) MeOH:H2O was performed to aid in the removal of buffer salts from the instrument.

5.2.4 Data Analysis

Data analysis for the ACN:H2O and MeOH:H2O systems was performed using

PeakFit Version 4 software (SPSS Inc. Chicago, IL). The CO2:MeOH:H2O analysis was conducted with OpenLab software (Agilent Technologies). All chromatographic data analysis was calculated using half-height for peak measurements.

5.3 Results and Discussion

5.3.1 Traditional HILIC Mobile Phases

132

ACN is by far the most commonly used organic modifier in HILIC separations.

In the framework of “green” chromatography, “greener” solvents like alcohols have also been used, but these alcohol/H2O mobile phases suffer from significantly reduced retention. As expected, Figure 5.1 illustrates the weak eluent strength of ACN in comparison to MeOH for nucleosides/nucleotides. As ACN content is increased past

65%, retention of the nucleosides/nucleotides increases drastically. This retention increase is not observed as MeOH content is increased. In fact, 90% MeOH content is required to achieve the same retention as 65% ACN, which is not even adequate to separate A, AMP, ADP, and ATP, let alone 16 nucleosides and nucleotides. Table 5.2 illustrates the ranges in retention factors for varying ACN and MeOH content for the 16 nucleoside and nucleotides grouped according to degree of phosphorylation: nucleosides, monophosphate nucleotides, diphosphate nucleotides, and triphosphate nucleotides.

Clearly, ACN content can be increased to achieve significant retention (k >2) for compounds within each class. On the other hand, increasing MeOH content only results in adequate retention of the most polar compounds, the diphosphate and triphosphate nucleotides. This demonstrates that while it is a “greener” option than ACN, MeOH is not a suitable replacement in this particular case.

133

7 A

2 Retention Factor, k Factor, Retention 1

0 65 70 75 80 85 90 % MeOH

Retention Factor, k Factor, Retention 5

0 55 60 65 70 75 % ACN

Figure 5.1. Effect of (A) MeOH and (B) ACN content on retention factor of A (♦), AMP (■), ADP (▲), and ATP (+).

134

Table 5.2. Retention factors, k, during isocratic conditions for nucleosides (A,U,C,G), monophosphate nucleotides (AMP,UMP,CMP,GMP), diphosphate nucleotides (ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) under varying ACN:100 mM sodium phosphate (pH=2.65) and MeOH:400 mM sodium phosphate (pH=2.65) mobile phases.

Retention Factor (k) Mobile Phase A,U, AMP,UMP, ADP,UDP, ATP,UTP, v:v% Mixture C,G CMP,GMP CDP,GDP CTP,GTP 55:45 0.4-0.8 0.7-1.0 0.9-1.3 1.2-1.6 60:40 0.5-1.0 1.0-1.6 1.5-2.2 2.1-2.8 ACN:buffer 65:35 0.6-1.5 1.5-2.8 2.7-4.1 4.1-5.8 70:30 0.8-2.6 2.8-5.8 5.9-10.5 10.3-16.5 75:25 1.3-6.8 8.2-20.3 19.4-22.2 22.2-46.2 65:35 0.3-0.5 0.3-0.5 0.5-0.6 0.6-0.9 70:30 0.4-0.5 0.4-0.6 0.6-0.9 0.9-1.3 75:25 0.4-0.6 0.4-0.7 0.7-1.2 1.3-2.0 MeOH:buffer 80:20 0.4-0.6 0.5-0.8 1.0-1.7 1.9-3.5 85:15 0.5-0.8 0.6-1.0 1.4-2.7 3.3-6.5 90:10 0.5-0.9 0.6-1.3 2.0-4.2 6.0-13.7

135

5.3.2 Mobile Phase Optimization

Phosphorylated compounds often suffer from strong chromatographic tailing caused by hydrogen-bond interactions with free silanols from the stationary phase [11] or chelation of metal ions from the connecting stainless steel tubing or column [25]. Figure

5.2 shows a chromatogram of the 16 nucleosides/nucleotides obtained using a

CO2:MeOH:H2O mobile phase without the addition of any buffer. It is apparent from the separation that heavy peak tailing necessitates incorporation of buffer into the mobile phase for our system as well. The easiest way to remedy the resulting poor chromatographic peak shape is typically through the addition of phosphoric acid or phosphate buffers to the mobile phase [19,25], which compete with and minimize phosphate analyte interactions. Unfortunately, phosphate buffers have limited solubility in organic solvents, which make up a high portion of HILIC mobile phases. Several phosphate buffer salts were studied such as potassium phosphate, ammonium phosphate, and sodium phosphate salts. The potassium phosphate salt was almost insoluble in

MeOH:H2O mobile phases, and was eliminated as an option. Ammonium phosphate showed increased solubility over that of potassium phosphate, however, heavy tailing still existed (Figure 5.3). Sodium phosphate exhibited similar solubility to ammonium phosphate but yielded better overall peak shape, and thus was used for this study.

136

1,5,10,14

2,6

9,13

11,15

12,16 Intensity 3,8 4,7

0 2 4 6 8 10 Time (min)

Figure 5.2. Chromatogram of 16 nucleoside/nucleotide analyte mixture (1.25 x 10-4 M) with 80:20 (v:v%) MeOH:H2O mobile phase. Gradient program used is (0-1.50 min) 70% B, (1.50-10.00 min) 70-90% B with 1.00 mL/min flow rate. The analyte key is (1)- A, (2)-U, (3)-G, (4)-C, (5)-AMP, (6)-UMP, (7)-CMP, (8)-GMP, (9)-ADP, (10)-UDP, (11)-CDP, (12)-GDP, (13)-ATP, (14)-UTP, (15)-CTP, (16)-GTP.

137

Intensity AMP

ADP ATP

0 2 4 6 8 10 Time (min)

Figure 5.3. Isocratic separation of A, AMP, ADP, and ATP analyte mixture (100 ppm) with 80:20 (v:v%) MeOH:aqueous buffer mobile phase. 10 mM ammonium phosphate buffer (pH=2.00).

138

Because the solubility of phosphate buffers in organic solvents decreases as pH increases, only a small range of pH (2.15-3.15) could be tested. As a result, limited retention changes were seen as pH was altered, with the exception of the triphosphate nucleotides. However, drastic peak width changes were seen as a result of both buffer concentration and pH. To determine the optimized pH for the sodium phosphate buffer, peak widths of the analytes were compared at several different pH values for the 16 analyte mixture. As is observed in Figure 5.4, the peak widths of the diphosphate and triphosphate nucleotides decreased as the pH increased. This was attributed to the increased monophosphate content present in higher pH’s at the same total buffer concentration (pKa1 = 2.15), which competes with the phosphate containing analytes for interactions with free silanols or the metal column. While it appears that pH=3.15 produces the lowest peak widths, it should be noted here that the amount of CO2 that can be added to the mobile phase not only depends on the MeOH:H2O ratio, but also upon the pH or ionic strength of the buffer used. At pH=3.15, the desired amount of CO2 could not be added before precipitation of buffer would occur, so pH=2.65 was chosen due to the fact that more CO2 could be added with minimal loss in peak efficiency.

The ionic strength of the buffer system plays an important role in the separation mechanism. As increased amounts of phosphate buffer become dissolved in the H2O layer a lessened amount of peak tailing and peak broadening is observed. Similarly to the pH study there is a finite amount of sodium phosphate buffer that can be dissolved in solution. Not only must buffer salts stay in solution while in the mobile pump reservoir, but the salts must also stay in solution upon mixing with CO2 in the mobile phase. Peak

139 widths are measured for mobile phases implementing 10 mM, 25 mM, and 40 mM sodium phosphate buffers during a separation of all 16 nucleoside/nucleotide analytes with 100 ppm concentration. Figure 5.5 displays the resulting peak width measurements.

Peak widths decrease as the ionic strength of the buffer increases. Once again there is a finite limit to how much buffer salts can be added to the mobile phase. Ultimately, the 40 mM sodium phosphate buffer system yielded narrower peak widths over the 10 mM and

25 mM concentrations tested, but concentrations higher than 40 mM resulted in solubility issues.

140

2.5

2.0

1.5

1.0

Peak width (min) width Peak 0.5

0.0

U C

ATP

UTP CTP

GTP

ADP

UDP CDP

AMP GDP

UMP CMP GMP Analyte

Figure 5.4. Peak width for 16 nucleoside/nucleotide analyte mixture (100 ppm) with 80:20 (v:v%) MeOH:aqueous phosphate buffer at pH 2.00 (■), pH 2.65 (■), and pH 3.15 (■) in the mobile phase. Gradient program used is (0-1.00 min) 70% B, (1.00-6.67 min) 70-90% B with 1.50 mL/min flow rate.

141

2.5

2.0

1.5

1.0

Peak width (min) width Peak 0.5

0.0

U C

ATP

UTP CTP

GTP

ADP

UDP CDP

AMP GDP

UMP CMP GMP Analyte

Figure 5.5. Peak width for 16 nucleoside/nucleotide analyte mixture (100 ppm) with 80:20 (v:v%) MeOH:aqueous phosphate buffer at a pH of 2.65 with an ionic strength of 10 mM (■), 25 mM (■), and 40 mM (■) in the mobile phase. Gradient program used is (0-1.00 min) 70% B, (1.00-6.67 min) 70-90% B with 1.50 mL/min flow rate.

142

5.3.3 Effect of CO2 on Retention

By adding CO2 to MeOH:H2O mixtures, the eluent strength of the mobile phase is decreased. This type of retention behavior has been observed previously by others

[26,27] and within our group [18,19,20]. For a HILIC mechanism, this results in increased retention, especially of the most polar compounds. This change in analyte retention as a function of added CO2 is demonstrated in Figure 5.6 for adenosine and its phosphate nucleotides. As the CO2 content is increased, the retention factor of all compounds increases. The most drastic increases are observed for the most polar compounds (ADP and ATP). Table 5.3 summarizes the changes in retention factors for each class of compounds. Retention factors within each class show an increasing trend as an increasing proportion of CO2 is added to the mobile phase system. While the most drastic change is for the most polar compounds (the diphosphate and triphosphate nucleotides), a significant change is observed even in the nucleosides and monophosphate nucleotides.

The above behavior suggests that when CO2 is used as the weak solvent, retention behavior is similar to that seen when ACN is added to H2O. In fact, an evaluation of the solvent strengths of the mobile phases (in which ACN, MeOH, and CO2 are considered as weak eluents), illustrates exactly that. For a HILIC separation, the relationship between the retention factor, k, and the volume fraction, φ, of the stronger eluent (H2O) in the mobile phase can be estimated by Equation 5.1.

log 푘 = log 푘표푟푔 − 푆휑 (Eq. 5.1)

143 where korg is the hypothetical retention factor of the analyte when using the weak solvent as the eluent and S is the slope of the plot of log k vs φ when fitted to a linear regression model [28]. The S-value can be used as a measure of eluent strength, with greater S- values indicating greater change in retention per change in volume fraction of strong solvent. Figure 5.7 shows plots of log k vs φ for ACN:H2O, MeOH:H2O, and

CO2:MeOH:H2O mixtures. H2O was treated as the strong eluent for the ACN:H2O and

MeOH:H2O separations, and 80:20 (v:v%) MeOH:H2O was treated as the strong eluent in the CO2:MeOH:H2O case. This allowed examination of the effect of ACN, MeOH, and

CO2 on retention individually.

The S-values obtained from the slopes in Figure 5.7 are summarized in Table 5.4.

From these values, it is evident that the eluent strength of CO2 is somewhere between that of ACN and MeOH. As a result, by adding CO2 to MeOH:H2O mixtures, the solvent strength of the overall mixture begins to approach that of high ACN content mobile phases.

The impacet of the mobile phase solvent strength depends significantly on the polarity of the analytes. By examining the averarage retention factors of analytes under isocratic solvent conditions for these mobile phases, isoeluotropic nomograms (Figure

5.8) were developed for the least polar (nucleosides) and most polar (triphosphate nucleotides) compounds. These nomograms indicate the relative percentages of weak eluent (ACN, MeOH, or CO2) needed to obtain similar solvent strength. Decrease in solvent strength that are simply not possible even using 90:10 (v:v%) MeOH:H2O mixtures, become possible by adding CO2 to 80:20 (v:v%) MeOH:H2O mobile phase.

144

This effect is increased as the polarity of the analytes is increased. For example, 30%

CO2 added to an 80:20 (v:v%) ratio MeOH:H2O mobile phase results in solvent strength equivalent to 65:35 (v:v%) ACN:H2O mixtures for nucleosides, but a 77:23 (v:v%)

ACN:H2O mixture for triphosphate nucleotides. Therefore, CO2:MeOH:H2O mixtures are potential candidates for “green” HILIC separations in which MeOH:H2O mobile phases fail to achieve adequate retention.

145

5 Retention Factor, k Factor, Retention

0 0 10 20 30

CO2 content (vol%)

Figure 5.6. Effect of CO2 content on retention factor of A (♦), AMP (■), ADP (▲), and ATP (+). Mobile phase consisted of varying proportions of solvent A (CO2) and solvent B (80:20 (v:v%) MeOH:H2O with 40 mM sodium phosphate buffer, pH=2.65).

146

Table 5.3. Retention factors, k calculated for isocratic conditions for nucleosides (A,U,C,G), monophosphate nucleotides (AMP,UMP,CMP,GMP), diphosphate nucleotides (ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) with 80:20 (v:v%) MeOH:H2O (40 mM sodium phosphate, pH=2.65) with CO2 addition. Retention Factor (k) CO 2 A,U, AMP,UMP, ADP,UDP, ATP,UTP, addition C,G CMP,GMP CDP,GDP CTP,GTP (vol%) 5% 0.2-0.3 0.3-0.6 0.8-1.7 2.0-4.1 10% 0.2-0.5 0.4-0.9 1.2-2.6 3.0-6.7 15% 0.3-0.6 0.6-1.2 1.8-3.9 4.7-10.6 MeOH:buffer 20% 0.3-0.8 0.7-1.7 2.3-5.7 6.6-15.8 25% 0.4-1.0 1.0-2.5 3.7-9.0 10.4-26.1 30% 0.5-1.3 1.6-4.1 6.3-17.1 19.1-54.2

147

A 1.2 1.0 y = -0.06x + 2.98 R² = 0.98 0.8 y = -0.06x + 2.53 0.6 R² = 0.98 0.4 y = -0.04x + 1.85 log k log 0.2 R² = 0.99 0.0 -0.2 y = -0.02x + 0.71 -0.4 R² = 0.99 -0.6 20 25 30 35 40 45 50

H2O Content (vol%)

Figure 5.7. Log k vs. φ plots for injection of 100 ppm of A (♦), AMP (■), ADP (▲), and ATP (+) for ACN:H2O (A), MeOH:H2O (B), and CO2:MeOH:H2O (C) optimized mobile phases. Lines represent linear regressions, and S-values obtained from slopes of linear regression lines. Continued

148

Figure 5.7 continued

B 1.0 0.8 y = -0.04x + 1.11 0.6 R² = 0.99 y = -0.02x + 0.51 0.4

R² = 0.99

0.2 log k log 0.0 y = -0.01x - 0.10 R² = 0.99 -0.2

-0.4 y = -0.01x - 0.25 R² = 0.96 -0.6 0 10 20 30 40

H2O Content (vol%)

Continued

149

Figure 5.7 continued

C 1.5 y = -0.04x + 3.92 1.0 R² = 0.99

0.5 y = -0.03x + 3.19

R² = 0.99

0.0 y = -0.03x + 2.14 log k log R² = 0.99 -0.5 y = -0.01x + 0.67 R² = 0.94 -1.0 60 70 80 90 100

H2O Content (vol%)

150

Table 5.4. S-values ranges for nucleosides (A,U,C,G), monophosphate nucleotides (AMP,UMP,CMP,GMP), diphosphate nucleotides (ADP,UDP,CDP,GDP), and triphosphate nucleotides (ATP,UTP,CTP,GTP) with varying mobile phases.

S-values

Mobile Phase A,U, AMP,UMP, ADP,UDP, ATP,UTP, Composition C,G CMP,GMP CDP,GDP CTP,GTP

ACN:H2O 2.1-3.5 4.1-5.0 5.3-6.1 6.1-6.8

MeOH:H2O 0.63-1.2 1.0-1.8 2.5-3.5 3.9-4.8

CO2:MeOH:H2O 1.8-2.5 2.5-3.1 3.2-4.0 3.6-4.4

151

A 40 50 60 70 % ACN (in H2O)

65 75 80 90 % MeOH (in H2O)

0 10 20 25 30 % CO2 (in 80:20 (v:v%) MeOH:H2O)

Decreasing Eluent Strength

152

Figure 5.8 continued

B 45 55 65 75 % ACN (in H2O)

75 80 90 % MeOH (in H2O)

0 5 10 15 20 25 30 % CO2 (in 80:20 (v:v%) MeOH:H2O)

Decreasing Eluent Strength

Continued

153

Figure 5.8 continued

C 50 60 70 80 % ACN (in H2O)

75 80 90

% MeOH (in H2O)

0 5 10 15 20 25 30 % CO2 (in 80:20 (v:v%) MeOH:H2O)

Decreasing Eluent Strength

Continued

154

Figure 5.8 continued

% ACN (in H2O) 50 60 70 80

70 75 80 85 90 %MeOH (in H2O)

0 5 10 15 20 25 30 % CO2 (in 80:20 (v:v%) MeOH:H2O)

Decreasing Eluent Strength

155

5.3.4 Gradient Optimization

The separation of the mixture of 16 nucleosides/nucleotides was optimized using

EFLC mobile phases. The basis of this optimization was to achieve the highest resolution for all of the analytes in the shortest period of time. Due to the large range in polarity of the nucleosides and nucleotides, a reverse CO2 gradient program was implemented. High portions of CO2 were required initially to obtain adequate retention and resolution of the less polar nucleosides and monophosphate nucleotides, followed by a decrease in CO2 content to elute the more polar diphosphate and triphosphate nucleotides.

The optimized separation of the mixture using an EFLC mobile phase is shown in

Figure 5.9. The gradient program contains two isocratic holds and two linear gradients.

In this work, the isocratic hold was essential to resolve the nucleosides prior to beginning a gradient to separate the remaining nucelotides. The gradient program for the optimized

EFLC separation is described below, using CO2 as solvent A and 80:20 (v:v%)

MeOH:H2O with 40 mM sodium phosphate (pH=2.65) as solvent B with a total flow of

1.00 mL/min. Initially, 71% B is held (0-1.50 min) in order to adequately retain and resolve adenosine and uridine. B is then increased from 71% to 82% (1.50-2.75 min) to begin eluting cytidine, guanine, AMP, and UMP. B is then held at 82% (2.75-4.50 min) until these compounds elute before it is further increased to 90% (4.50-7.50 min) to begin elution of the highly polar diphosphate and triphosphate nucleotides. Finally, B is held at

90% until all analytes are eluted from the column. Isocratic holds and gradients were used previously for the analysis of nucleotides with several different types of columns

156 under HILIC conditions [29]. To the best of our knowledge, this is the first time that all

16 RNA nucleosides/nucleotides were spearted with Rs ≥ in under 20 minutes.

157

400 1 100 A 2 3 11,14 13 350 4 6 5 8 16 90 7,10 9 12 15 300 80 2,5 1 70 250 B 4,6 3,7 60 200 14,15 8 9 13 10,11 12 16 50 150

Intensity 1 40 100 C 2 3 4 30 5 50 6 7 8 9 10 11 12 13 20 14 15 16 0 10 0 5 10 15 20 -50 Time (min) 0

Figure 5.9. Optimized separation of the 16 nucleoside/nucleotide analyte mixture (1.25 x 10-4 M) with (A) ACN:100 mM sodium phosphate (pH=2.65) (0-7.5 min, 70% ACN, 7.5-15 min, 70-60% ACN), (B) MeOH:400 mM sodium phosphate (pH=2.65) (0-4 min, 90% MeOH, 4-6 min, 90-80% MeOH), and (C) CO2:MeOH:H2O with gradient described in Section 5.3.4. The analyte key is (1)-A, (2)-U, (3)-G, (4)-C, (5)-AMP, (6)-UMP, (7)- CMP, (8)-GMP, (9)-ADP, (10)-UDP, (11)-CDP, (12)-GDP, (13)-ATP, (14)-UTP, (15)- CTP, (16)-GTP. Mobile phase gradients are shown with dashed lines with addition of the instrument dwell time for each separation.

158

5.3.5 Optimized Chromatographic Comparison

The performance of traditional MeOH:H2O and ACN:H2O HILIC mobile phases using the amide stationary phase, and similar pH and buffer concentrations was also analyzed. Figure 5.9A and Figure 5.9B optimized ACN:H2O and MeOH:H2O systems, respectively. The optimized MeOH:H2O separation is less than 15 minutes; however, the quality of the separation is hindered and multiple analytes coelute. No amount of added

MeOH content could provide adequate retention of the nucleosides and monophosphate nucleotides, and peak shape of the diphosphate nucleotides and triphosphate nucleotides is very poor. The ACN:H2O system produces much better retention and peak shape, but is unable to resolve two analyte pairs (CMP and UDP, CDP and UTP). It may be possible to resolve all of the compounds by delaying or lengthening the gradient, but this would obviously come at the expense of increased analysis time. Interestingly, both the ACN and MeOH methods had different retention order than the EFL mobile phase. The different selectivity of the three mobile phase systems is shown in Figure 5.10. In short, the EFL mobile phase provided the best separation of the three mobile phases in the shortest amount of time, with different selectivity than either of the traditional HILIC mobile phases.

Others have studied the separation of RNA nucleosides/nucleotides using traditional HILIC mobile phases, either with a subset of the RNA nucleoside/nucleotide mixture or with additional biologically similar analytes, such as DNA nucleosides/nucleotides [30,31]. Zhou et al. accomplished a separation of 11 RNA nucleoside/nucleotides with four intermediates in 26 minutes (Rs ≥1.3) using a titania

159 column with ACN:H2O mobile phases. However, their study did not include the four

RNA nucleosides and CDP. Yang et al. separated monophosphate- and diphosphate nucleotides using HILIC with an amide column and ACN:buffer mixtures combined with mass spectrometric detection. The analysis time for Yang’s method was approximately

20 minutes. Complete chromatographic resolution of paired peaks was not achieved nor was it necessary because the mass spectrometer could separate the overlapping chromatographic peaks. However, nucleosides and triphosphate nucleotides were not included in their study. Previously in our group, an isocratic separation of 15 RNA nucleoside/nucleotide analytes was accomplished in just under an hour with greater than

1.3 resolution, but with broad peak widths [32]. By using gradient elution, a substantial decrease in analysis time from approximately one hour to 16 minutes was achieved.

5.3.6 Efficiency and Resolution

The advantages gained by using EFLC are not limited to the changed polarity of the mobile phase, but also include higher analyte diffusivity and a lower viscosity solution in comparison to MeOH:H2O mobile phases. This combination of improved mass transport properties allows for increased flow rates without significant loss in chromatographic efficiency. While it is not fair to directly compare the effective plate heights between these conditions (since each analyte was eluted using a gradient

2 program), Neue previously illustrated that the square of peak volume (푝푣 ) vs mobile phase flow rate (pseudo van Deemter plots) are useful for comparison of gradient mobile phase conditions, provided that the gradient volume is held constant. Figure 5.11 shows

160 pseudo van Deemter curves [33] for the adenosine containing analytes (A, AMP, ADP, and ATP) with the ACN:H2O, MeOH:H2O, and CO2:MeOH:H2O optimized gradient

2 separations. As the flow rate is increased, 푝푣 (which is proportional to peak width) increases as well for all analytes with all mobile phase gradients. However, the rate at

2 which 푝푣 (and peak width) increases is smaller for the EFLC mobile phase than for either the MeOH:H2O or ACN:H2O mobile phase.

Because this rate of decrease in efficiency is smaller, the flow rate of the optimized EFLC separation can be increased to 2.0 mL/min before the resolution of peaks AMP and UMP falls under Rs=1.0 (Figure 5.12). This effectively means that separation of these 16 RNA analytes can be completed in under 10 minutes of analysis time (Figure 5.13). All other pairs of compounds still have resolution greater than 1.3 at this point so if resolution of AMP and UMP is not important, the flow rate could be increased further to 3.0 mL/min before instrument back pressure limits (600 bar maximum) would be reached.

161

Figure 5.10. Order of elution for each of the mobile phase systems tested (MeOH:H2O, CO2:MeOH:H2O, and ACN:H2O) from least retained on top to most retained on bottom for nucleoside/nucleotides.

162

0.14 slope = 0.028 A R² = 0.95 0.12 slope = 0.025

0.10 R² = 0.98

) 2

0.08

(mL

2 0.06 pv slope = 0.012 0.04 R² = 0.96

0.02

0.00 0 1 2 3 Flow Rate (mL/min)

Figure 5.11. Pseudo van Deemter plots for 100 ppm of (A) A, (B) AMP, (C) ADP, and (D) ATP for the CO2:MeOH:H2O (♦), MeOH:H2O (■), and ACN:H2O (▲) optimized mobile phases. Measured at various flow-rates under optimized mobile phase conditions in Figure 2.10. Lines represent linear regressions.

Continued

163

Figure 5.11 continued

0.08 B 0.07 slope = 0.019 R² = 0.99

0.06

) 2

0.05

(mL

2 0.04 slope = 0.019 pv R² = 0.96 0.03 0.02 slope = 0.0041 0.01 R² = 0.99 0 0 1 2 3 Flow Rate (mL/min)

Continued

164

Figure 5.11 continued

0.14 C slope = 0.037 0.12 R² = 0.93

0.10

) 2 slope = 0.036

0.08 R² = 0.98

(mL

0.06 pv

0.04 slope = 0.0088 0.02 R² = 0.98

0.00 0 1 2 3 Flow Rate (mL/min)

Continued

165

Figure 5.11 continued

0.8 D slope = 0.35 0.7 R² = 0.99

0.6

)

2 0.5 (mL

0.4 2

pv 0.3

0.2 slope = 0.024 R² = 0.99 0.1 slope = 0.018 0 R² = 0.98 0 1 2 3 Flow Rate (mL/min)

166

s 6

4 Resolution, R Resolution, 2

G-C

A-U

U-G

C-AMP

CTP-GTP

UTP-CTP

ATP-UTP

GDP-ATP

CDP-GDP

UDP-CDP

ADP-UDP

GMP-ADP

CMP-GMP

UMP-CMP AMP-UMP Analyte Pair

Figure 5.12. Resolution, Rs of analyte pair at collected with 80:20 (v:v%) MeOH:H2O (40 mM sodium phosphate, pH=2.65) with 0.50 mL/min (■), 1.00 mL/min (■), 1.50 mL/min (■), and 2.00 mL/min (■) flow-rates. The dashed line represent Rs=1.0.

167

A 1

2 3 4 5 Intensity 6 8 9 7 10 12 11 13 14 15 16

0 5 10 15 20 25 30 35 Time (min)

Figure 5.13. Separation of the 16 nucleoside/nucleotide analyte mixture (1.25 x 10-4 M) with 0.50 mL/min (A), 1.00 mL/min (B), 1.50 mL/min (C), and 2.00 mL/min (D) total mobile phase flow rates. The analyte key is (1)-A, (2)-U, (3)-G, (4)-C, (5)-AMP, (6)- UMP, (7)-CMP, (8)-GMP, (9)-ADP, (10)-UDP, (11)-CDP, (12)-GDP, (13)-ATP, (14)- UTP, (15)-CTP, (16)-GTP. Continued

168

Figure 5.13 continued

B 1

3 4 5

Intensity 6 8 7 9 10 11 1213 14 15 16

0 5 10 15 20 Time (min)

Continued

169

Figure 5.13 continued

C 1

3 4 5 Intensity 6 7 8 9 10 11 1213 14 15 16

0 5 10 15 Time (min)

Continued

170

Figure 5.13 continued

D 1

3 4

Intensity 5 6 7 8 9 10 11 12 13 14 15 16

0 5 10 Time (min)

171

5.3.7 Evaluation of Method “Greenness”

Each method was evaluated from an environmental impact standpoint. While

MeOH and CO2 are considered “greener” alternatives to ACN, the amount of each solvent used must also be taken into consideration. In order to assess the “greenness” of the optimized LC and EFLC methods, HPLC environmental assessment tool (HPLC-

EAT) scores [34] were calculated for a single run under each set of optimized mobile phase conditions: ACN:H2O, MeOH:H2O, and CO2:MeOH:H2O. These scores take into consideration the environmental (E), health (H), and safety issues (SI) for all solvents within an HPLC method. The HPLC-EAT score for a given method is calculated using

Equation 5.2.

퐻푃퐿퐶 − 퐸퐴푇 = (푆퐼)1푚1 + 퐻1푚1 + 퐸1푚1+… + (푆퐼)푛푚푛 + 퐻푛푚푛 + 퐸푛푚푛 (Eq. 5.2) where SI, H, and E are safety, health, and environmental factors, respectively (calculated for each solvent according to Koller et al. [35]), m is the mass of solvent used, and n is the number of solvents. Lower scores indicate more environmentally friendly methods.

The scores were tabulated for a single run (at 1.0 mL/min) using each method, and are reported in Table 5.5. Not only does the optimized CO2:MeOH:H2O EFLC method use less total organic solvent per run, it also has a total HPLC-EAT score (25.4) that is less than half that of the ACN:H2O method (61.2). Additionally, the individual safety, health, and environmental scores are all lower for the EFLC method compared to the ACN:H2O method. As expected, the EFLC method has comparable HPLC-EAT scores to that of the optimized MeOH:H2O separation, but the EFLC separation offers vastly superior chromatographic performance. When examining all three methods from both a

172 chromatographic and environmental standpoint, it is clear that the CO2:MeOH:H2O method is by far the best option.

173

Table 5.5. Mobile Phase composition and HPLC-EAT scores for optimized ACN:H2O, MeOH:H2O, and CO2:MeOH:H2O separations. Optimized conditions are same as listed in Figure 5.9. Total Safety Health HPLC- Mobile Phase Environ. Solvent Impact Impact EAT Composition Impact (E) (g) (SI) (H) score ACN:H2O 13.5 36.6 14.3 10.4 61.2 MeOH:H2O 9.53 17.8 3.98 2.97 24.7 CO2:MeOH:H2O 9.27 18.3 4.10 3.05 25.4

174

5.4 Conclusions

EFLC using a gradient allowed for the first time a separation of 16 RNA nucleosides/nucleotides with greater than 1.3 resolution for all analyte pairs in less than

20 minutes. The use of “green” solvents performed better or at least comparable in

HILIC to more costly conventional solvents as mobile phases. Furthermore, addition of

CO2 to mobile phases allows for fine tuning of mobile phase polarity for the separation of analytes with wide polarity ranges when paired with gradient elution programming.

5.5 References

[1] P.T. Anastas, L.G. Heine, T. C. Williamson, Green Chemical Syntheses and

Processes: Introduction, in: P.T. Anastas, L.G. Heine, T. C. Williamson, Ed.; Green

Chemical Syntheses and Processes, American Chemical Society, Washington, DC,

2000, Chapter 1, pp 1-6.

[2] P.T. Anastas, J.C. Warner, Green Chemistry: Theory and Practice, Oxford

University Press, New York, 1998.

[3] J. Płotka, M. Tobiszewski, A.M. Sulej, M. Kupska, T. Górecki, J. Namiesnik,

Green Chromatography, Green chromatography, J. Chromatogr. A (2013), 1307, 1-

20.

[4] P. Sandra, G. Vanhoenacker, F. David, K. Sandra, A. Pereira, Green

chromatography. Part 1: Introduction and liquid chromatography, LC GC Eur

(2010), 23, 242-259.

175

[5] R.K. Henderson, C. Jimenez-Gonzalez, D.J.C. Constable, S.R. Alston, G.G.A.

Inglis, G. Fisher, J. Sherwood, S.P. Binks, A.D. Curzons, Expanding GSK’s

solvent selection guide - embedding sustainability into solvent selection starting at

medicinal chemistry, Green Chem., (2011), 13, 854-862.

[6] K. Alfonsi, J. Colberg, P.J. Dunn, T. Fevig, S. Jennings, T.A. Johnson, H.P. Kleine,

C. Knight, M.A. Nagy, D.A. Perry, M. Stefaniak, Green chemistry tools to

influence a medicinal chemistry and research chemistry based organization, Green

Chem., (2008), 10, 31-36.

[7] A.J. Alpert, Hydrophilic-interaction chromatography for the separation of peptides,

nucleic acids and other polar compounds, J. Chromatogr., (1990), 499, 177-196.

[8] B. Buszewski, S. Noga, Hydrophilic interaction liquid chromatography (HILIC)-a

powerful separation technique, Anal. Bioanal. Chem., (2012), 402, 231-247.

[9] Z. Hao, B. Xiao, N. Weng, Impact of column temperature and mobile phase

components on selectivity of hydrophilic interaction chromatography (HILIC), J.

Sep. Sci., (2008), 31, 1449-1464.

[10] R.P. Li, J.X. Huang, Chromatographic behavior of epirubicin and its analogues on

high-purity silica in hydrophilic interaction chromatography, J. Chromatogr. A,

(2004), 1041, 163–169.

[11] B.A. Olsen, Hydrophilic interaction chromatography using amino and silica

columns for the determination of polar pharmaceuticals and impurities, J.

Chromatogr. A, (2001), 913, 113 –122.

176

[12] J.C. Valette, C. Demesmay, J.L. Rocca, E. Verdon, Separation of tetracycline

antibiotics by hydrophilic interaction chromatography using an amino-propyl

stationary phase, Chromatographia, (2004), 59, 55–60.

[13] A. dos Santos Pereira, F. David, G. Vanhoenacker, P. Sandra, The acetonitrile

shortage: is reversed HILIC with water an alternative for the analysis of highly

polar ionizable solutes?, J. Sep. Sci., (2009), 32, 2001–2007.

[14] T.A. Berger, Characterization of a 2.6 µm Kinetex porous shell hydrophilic

interaction liquid chromatography column in supercritical fluid chromatography

with a comparison to 3 µm totally porous silica, J. Chromatogr. A, (2011), 1218,

4559-4568.

[15] L. Subbarao, J. Cole, R. Chen, An empirical evaluation of HILIC and monolithic

columns for SFC applications, LC GC N. Am., (2010), Feb. Suppl. 28, 62-63.

[16] M.R. Emmett, S. Kazazic, A.G. Marshall, W. Chen, S.D.-H. Shi, B. Bolaños, M.J.

Greig, Supercritical fluid chromatography reduction of hydrogen/deuterium back

exchange in solution-phase hydrogen/deuterium exchange with mass spectrometric

analysis, Anal. Chem., (2010), 78, 7058-7060.

[17] C. West, S. Khater, E. Lesellier, Characterization and use of hydrophilic interaction

liquid chromatography type stationary phases in supercritical fluid

chromatography, J. Chromatogr. A, (2012), 1250, 182-195.

[18] J.W. Treadway, G.S. Philibert, S.V. Olesik, Enhanced fluidity liquid

chromatography for hydrophilic interaction separation of nucleosides, J.

Chromatogr. A, (2011), 1218, 5897-5902. 177

[19] G.S. Philibert, S.V. Olesik, Characterization of enhanced-fluidity liquid hydrophilic

interaction chromatography for the separation of nucleosides and nucleotides, J.

Chromatogr. A, (2011), 45, 8222-8230.

[20] M.J. Beres, S.V. Olesik, Enhanced-fluidity liquid chromatography using mixed-

mode hydrophilic interaction liquid chromatography/strong cation-exchange

retention mechanisms, J. Sep. Sci., (2015), 402, doi: 10.1002/jssc.201500454.

[21] A. dos Santos Pereira, A.J. Girón, E. Admasu, P. Sandra, Green hydrophilic

interaction chromatography using ethanol-water-carbon dioxide mixtures, J Sep.

Sci., (2010), 33, 834-837.

[22] Y. Cui, S.V. Olesik, Reversed-phase high-performance liquid chromatography

using enhanced-fluidity mobile phases, J. Chromatogr. A, (1995), 691, 151-162.

[23] S.T. Lee, S.V. Olesik, Normal-phase high-performance liquid chromatography

using enhanced-fluidity liquid mobile phases, J. Chromatogr. A, (1995), 707, 217-

224.

[24] ACD/Chemsketch Freeware, Advanced Chemistry Development, Inc., Toronto,

ON, Canada, 2015.

[25] A. Wakamatsu, K. Morimoto, M. Shimizu, S. Kudoh, A severe peak tailing of

phosphate compounds caused by interaction with stainless steel used for liquid

chromatography and electrospray mass spectrometry, J. Sep. Sci., (2005), 28, 1823-

1830.

178

[26] T.A. Berger, J.F. Deye, Composition and density effects using methanol/carbon

dioxide in packed column supercritical fluid chromatography, Anal. Chem., (1990),

62, 1181-1185.

[27] T.A. Berger, J.F. Deye, “Use of solvatochromic dyes to correlate mobile phase

solvent strength to chromatographic retention in supercritical fluid

chromatography.” Supercritical Fluid Technology. Ed. F. Bright, Washington D.C.:

American Chemical Society, 1992. 132-142.

[28] P. Hemstrom, K. Irgum, Hydrophilic interaction chromatography, J. Sep. Sci.,

(2006), 29, 1784–1821.

[29] S.H. Yang, H.P. Nguyen, K.A. Schug, Method development and analysis of mono-

and dephosphorylated nucleotides by HILIC-ESI-MS, Hydrophic Interaction

Liquid Chromatography and Advanced Applications, Ed. P.G. Wang, W. He, Boca

Raton, FL: CRC Press, 2011, 445-450.

[30] A.J. Alpert, Electrostatic repulsion hydrophilic interaction chromatography for

isocratic separation of charged solutes and selective isolation of phosphopeptides,

Anal. Chem., (2008), 80, 62-76.

[31] T. Zhou, C.A. Lucy, Hydrophilic interaction chromatography of nucleotides and

their pathway intermediates on titania, J. Chromatogr. A, (2008), 1187, 89-93.

[32] G.S. Philibert, Improved separation of biological compounds using enhanced-

fluidity liquid chromatography, Dissertation, The Ohio State University. Ann

Arbor: UMI, 2011. (Publication No. ATT AAI3493528.)

179

[33] M. Gilar, U. Neue, Peak capacity in gradient reversed-phase liquid chromatography

of biopolymers theoretical and practical implications for the separation of

oligonucleotides, J. Chromatogr. A, 1169, 139-150.

[34] Y. Gaber, U. Tornvall, M.A. Kumar, M.A. Amin, R. Hatti-Kaul, HPLC-EAT

(environmental assessment too): a tool for profiling safety, health, and

environmental impacts of liquid chromatography methods, Green Chem., (2011),

13, 2021-2025.

[35] G. Koller, U. Fischer, K. Hungerbuhler, Assessing safety, health and environmental

impact during early process development, Ind. Eng. Chem. Res., (2000), 39, 960-

972.

180

CHAPTER 6: SUMMARY AND FUTURE WORK

6.1 Summary of Research

The research detailed in this document describes generation of nanomaterials and subsequent application to separation and electrochemical science. The nanomaterials were produced via the electrospinning method that has been demonstrated as a cheap and facile way to generate high surface area, tunable nanostructures suitable for yielding highly efficient separations. Chapter 2 detailed the applicability of aligned electrospun nanofibers to UTLC analysis. Aligning the electrospun nanofibers with a rotating drum collector not only increased the directional orientation of the nanofibers but also increased the straightness of the individual nanofibers. Furthermore, AE-UTLC devices represented a further enhancement in separation efficiency relative to E-UTLC stationary phases with a decreased time of analysis by over 50%. When compared with E-UTLC, the AE-UTLC devices provided a more rapid and consistent time of analysis while also showing increased efficiency and reproducibility; these enhanced separations occurred over a very short development distance and time.

Chapter 3 described the determination of optimal mobile phase conditions for five laser dyes as analytes. PEC analysis was conducted on an electrospun SiO2/PVP stationary phase and compared to UTLC analysis with migration distance and efficiency calculations. PEC applied voltage and time of analysis were also studied and optimized 181 for the five laser dyes. The presence of EOF with PEC analysis was shown for the KR and SR laser dyes. Analytes migrating further, under EOF conditions, in the PEC device produced a much lower relative standard deviation for efficiency compared to UTLC analysis. A separation of three laser dyes utilizing EOF was successfully achieved. The three laser dyes analytes were shown separated with a resolution of 4.0 and 1.0 between analyte pairs.

Chapter 4 divulges electrospinning conditions for the generation of 70:30 (w:w%)

Nafion:PAA polymer composite nanofibers with 8 kV of applied voltage and a tip to collector distance of 12.5 cm. The electrospinning conditions correspond to producing fibers with 500 ± 100 nm diameters. Further optimization of a modified GCE surface is determined through varying electrospinning time while measuring mat thicknesses of a composite of 70:30 (w:w%) Nafion:PAA nanofiber mat. An optimized nanofiber mat thickness of 2.79 ± 0.34 µm resulted in an electrospinning time of 0.75 minute. Use of the electrospun modified GCEs were then employed to determine the amount of guanine in solution with the generation of electrochemiluminescent signal monitored with a PMT.

Good linearity was observed between 1.0 x 10-5 M to 1.0 x 10-7 M guanine concentrations while measuring electrochemiluminescent signal.

In Chapter 5 the use of EFLs allowed for the first time a separation of 16 RNA nucleosides/nucleotides with greater than 1.3 resolution for all analyte pairs in less than

20 minutes. The use of “green” solvents performed better or at least comparable in

HILIC to more costly conventional solvents as mobile phases. Additionally, addition of

CO2 to mobile phases allow for fine tuning of mobile phase polarity for the separation of

182 analytes with wide polarity ranges when paired with gradient elution programming.

Solvent strength of the mobile phase played a key role in analyte retention thusly, the use of CO2 allows for differences in retention order across mobile phase systems.

6.2 Future Work

Electrospun nanofiber mats have successfully been applied to UTLC analysis.

The progression toward a packed electrospun nanofiber liquid chromatography (LC) stationary phase will be difficult. Larger channeled polymeric fibers (50 µm I.D.) have already been packed into an LC column [1]. The packed polymeric fibers provided shorter analysis time while yielding comparable performance to a commercial packed LC column [2]. Utilization of much narrower electrospun nanofibers to construct an LC stationary phase may give more efficient chromatographic separations in shorter analysis times by comparison. Presently, the nanofibers must be removed from their substrate if they are to be forced into a stainless steel column. While some nanofibers such as PAN can be easily peeled away from the surface of the conductive collector following electrospinning, other electrospun nanofibers, such as SU-8, are difficult to remove.

Electrospinning directly to material that could be packed into the column with the stationary phase would be beneficial to forcing the nanomaterial into the column.

Joule heating is detrimental to any separation with applied potential. Further research may also be pursued related to the removal of Joule heating from the PEC setup during analysis used in Chapter 3. Further, Nurok et al. describes a non-commercial PEC device that operates under pressure (PPEC) [3]. PPEC helps by reducing the amount of

183 mobile phase accumulating on the surface of the stationary phase, which will lead to an improvement in chromatographic efficiency for the system because any potential analyte streaking will become minimized.

6.3 References

[1] R.K. Marcus, W.C. Davis, B.C. Knippel, L. LaMotte, T.A. Hill, D. Perahia, J.D.

Jenkins, Capillary-channeled polymer fibers as stationary phases in liquid

chromatography separations, J. Chromogr. A, (2003), 986, 17-31.

[2] D.M. Nelson, R.K. Marcus, Characterization of capillary-channeled polymer fiber

stationary phases for high-performance liquid chromatography protein separations:

comparative analysis with a packed-bed column, Anal. Chem., (2006), 76, 8462-

8471.

[3] D. Nurok, J.M. Koers, A.L. Novotny, M.A. Carmichael, J.J. Kosiba, Apparatus and

initial results for pressurized planar electrochromatography, Anal. Chem., (2004),

76, 1690-1695.

184

Bibliography

[1] Z. Zhang, Z. Wang, Z. Liao, H. Liu, Applications of nanomaterials in liquid

chromatography: opportunities for separation with high efficiency and selectivity,

J. Sep. Sci., (2006), 29, 1872-1878.

[2] L.S. Ettre, K.I Sakodynskii, M.S. Tswett and the discovery of chromatography I:

early work (1899-1903), Chromatographia, (1993), 35, 223-231.

[3] S. Ramakrishna, K. Fujihara, W.E. Teo, T.C. Lim, Z. Ma, An Introduction to

Electrospinning and Nanofibers, World Scientific, New York, NY, 2005.

[4] W.E. Teo, S. Ramakrishna, A review on electrospinning design and nanofiber

assemblies, Nanotechnology, (2006), 17, R89-R106.

[5] N. Bhardwaj, S.C. Kundu, Electrospinning: a fascinating fiber fabrication

technique, Biotechnol. Adv., (2010), 28. 325-347.

[6] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Aligned biodegradable nanofibrous

sturcture: a potential scaffold for blood vessel engineering, Biomater., (2004), 25,

877-886.

[7] S.F. Fennessey, R.J. Farris, Fabrication of aligned and molecularly oriented

electrospun polyacrylonitrile nanofibers and the mechanical behavior of their

twisted yarns, Polymer, (2004), 45, 4217-4225.

185

[8] W. Hu, Z.M. Huang, S.Y. Meng, C. He, Fabrication and characterization of

chitosan coated braided PLA wire using aligned electrospun fibers, J. Mater. Sci.:

Mater. Med., (2009), 20, 2275-2284.

[9] Sz. Nyiredy, in: Sz. Nyiredy (Ed.), Planar Chromatography: A Retrospective View

for the Third Millennium, Springer, Budapest, 2001, pp 177-199.

[10] L. Botz, S. Nagy, B. Kocsis, in: Sz. Nyiredy (Ed.), Planar Chromatography: A

Retrospective View for the Third Millennium, Springer, Budapest, 2001, pp 489-

516.

[11] J.E. Clark, S.V. Olesik, Technique for ultrathin layer chromatography using an

electrospun, nanofibrous stationary phase, Anal. Chem., (2009), 81, 4121-4229.

[12] Z.F. Ding, B.M. Quinn, S.K. Haram, L.E. Pell, B.A. Korgel, A.J. Bard,

Electrochemistry and electrogenerated chemiluminescence from silicon nanocrystal

quantum dots, Science, (2002), 296, 1293-1297.

[12] M.M. Richter, Electrochemiluminescence (ECL), Chem. Rev., (2004), 104, 3003-

3036.

[14] Z.H. Guo, Y. Shen, M.K. Wang, F. Zhao, S.J. Dong, Electrochemistry and

electrogenerated chemiluminescence of SiO2

nanoparticles/tris(2,2’bipyridyl)ruthenium(II) multilayer films on indium tin oxide

electrodes, Anal. Chem., (2004), 76, 184-191.

[15] S. Zanarini, E. Rampazzo, L. Della Ciana, M. Marcaccio, E. Marzocchi, M.

Montalti, F. Paolucci, L. Prodi, Ru(bpy)3 covalently doped silica nanoparticles as

186

multicenter tunable structures for electrochemiluminescence amplification, J. Am.

Chem. Soc., (2009), 131, 2260-2267.

[16] J. Yoon, S.K. Chae, J.M. Kim, Colorimetric sensors for volatile organic compounds

(VOCs) based on conjugated polymer-embedded electrospun fibers, J. Am. Chem.

Soc., (2007), 129, 3038-3039.

[17] X. Wang, C. Drew, S.H. Lee, K.J. Senecal, J. Kumar, L.A. Samuelson, Electrospun

nanofibrous membranes for highly sensitive optical sensors, Nano Lett., (2002), 2,

1273-1275.

[18] X.B. Yin, S.J. Don, E.I. Wang, Analytical applications of the

electrochemiluminescence of tris(2,2’-bipyridyl) ruthenium and its derivatives,

Trac-Trend Anal. Chem., (2004), 23, 432-441.

[19] F. Di Benedetto, A. Camposeo, S. Pagliara, E. Mele, L. Persano, R. Stabile, R.

Clingolani, D. Pilignano, Patterning of light-emitting conjugated polymer

nanofibers, Nat. Nanotiechnol., (2008), 3, 614-619.

[20] D.J. Kim, Y.K. Lyu, H.N. Choi, I.H. Min, W.Y. Lee, nafion-stabilized magnetic

2+ nanoparticles (Fe3O4) for [Ru(bpy)] (bpy=bipyridine) electrogenerated

chemiluminescence sensor, Chem. Commun., (2005), 23, 2966-2968.

[21] H. Chen, J.D. Snyder, Y.A. Elabd, Electrospinning and solution properties of

Nafion and poly(acrylic acid), Macromolecules, (2008), 41, 128-135.

[22] M.J. Beres, S.V. Olesik, Enhanced-fluidity liquid chromatography using mixed-

mode hydrophilic interaction chromatography/strong cation-exchange retention

mechanisms, J. Sep Sci., (2015), 38, 3119-3129.

187

[23] A. dos Santos Pereira, A.J. Giron, E. Admasu, P. Sandra, Green hydrophilic

interaction chromatography using ethanol-water-carbon dioxide mixtures, J. Sep.

Sci., (2010), 33, 834-837.

[24] J.C. Valette, C. Demesmay, J.L. Rocca, E. Verdon, Separation of tetracycline

antibiotics by hydrophilic interaction chromatography using an amino-propyl

stationary phase, Chromatographia, (2004), 59, 55-60.

[25] S.T. Lee, S.V. Olesik, Normal-phase high-performance liquid chromatography

using enhanced-liquid mobile phases, J. Chromatogr. A, (1995), 707, 217-224.

[26] G.S. Philibert, S.V. Olesik, Characterization of enhanced-fluidity liquid hydrophilic

interaction chromatography for the separation of nucleosides and nucleotides, J.

Chromatogr. A, (2011), 45, 8222-8230.

[27] C. F. Poole, The Essence of Chromatography, Elsevier Science, Amsterdam, The

Netherlands, 2003.

[28] J. Sherma, Planar chromatography, Anal. Chem., (2010), 82, 4895-4910.

[29] J.S. Bernard Fried, Thin-layer Chromatography: Techniques and Applications,

Marcel Dekker, New York, NY, 1994.

[30] L.W. Bezuidenhout, M.J. Brett, Ultrathin layer chromatography on nanostructured

thin films, J. Chromatogr. A., (2008), 1183, 179-185.

[31] S.A. Nabi, M.A. Khan, Selective TLC separation of lysine and threonine in

pharmaceutical preparations, Acta Chromatogr., (2003), 13, 161-171.

188

[32] M. Kastelan-Macan, S. Cerjan-Stefanovic, M. Petrovic, Phenol adsorption on

active carbon by means of thin-layer chromatography, Chromatographia, (1989),

27, 297-300.

[33] J. Song, D.S. Jensen, D.N. Hutchison, B. Turner, T. Wood, A. Dadson, M.A. Vail,

M.R. Linford, R.R. Vanfleet, R.C. Davis, Carbon nanotube-templated

microfabrication of porous silicon-carbon materials with application to chemical

separations, Adv. Funct. Mater., (2011), 21, 1132-1139.

[34] S.R. Jim, A.J. Oka, M.T. Taschuk, M.J. Brett, Morphological modification of

nanostructured ultrathin-layer chromatography stationary phases, J. Chromatogr.

A., (2011), 1218, 7203-7210.

[35] H.E. Hauck, M. Schulz, Ultra thin-layer chromatography, Chromatographia,

(2003), Supp. 57, S-313-S-315.

[36] W.G. Shim, C. Kim, J.W. Lee, J.J. Yun, Y. Jeong, H. Moon, K.S. Yang,

Adsorption characteristics of benzene on electrospun-derived porous carbon

nanofibers, J. Appl. Polym. Sci., (2006), 102, 2454-2462.

[37] U. Boudriot, R. Dersch, A. Greiner, J.H. Wendorff, Electrospinning approaches

toward scaffold engineering - a brief overview, Artif. Organs., (2006), 30, 785-792.

[38] J.W. Zewe, J.K. Steach, S.V. Olesik, Electrospun fibers for solid-phase

microextraction, Anal. Chem., (2010), 82, 5341-5348.

[39] T.E. Newsome, J.W. Zewe, S.V. Olesik, Electrospun nanofibrous solid-phase

microextraction coatings for preconcentration of pharmaceuticals prior to liquid

chromatographic separations, J. Chromatogr. A, 1262, 1-7.

189

[40] W. Jost, H.E. Hauck, W. Fischer, Cyano-modified pre-coated plates - a medium

polar modification for high performance thin-layer chromatography,

Chromatographia, (1986), 21, 375-378.

[41] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Aligned biodegradable nanofibrous

structure: a potential scaffold for blood vessel engineering, Biomater., (2004), 25,

877-886.

[42] Z. Song, X. Hou, L. Zhang, S. Wu, Enhanced crystallinity and orientation by hot-

stretching to improve the mechanical properties of electrospun partially aligned

polyacrylonitrile (PAN) nanocomposites, Materials, (2011), 4, 621-632.

[43] R. Jalali, M. Morshed, S.A. H. Ravandi, Fundamental parameters affecting

electrospinning of PAN nanofibers as uniaxially aligned fibers, J. Appl. Polym.

Sci., (2006), 101, 4350-4357.

[44] D. Li, Y. Wang, Y. Xia, Electrospinning of polymeric and ceramic nanofibers as

uniaxially aligned arrays, Nano Lett., (2003), 3, 1167-1171.

[45] A.V.I. Hess, Digitally enhanced thin-layer chromatography: an inexpensive, new

technique for qualitative and quantitative analysis, J. Chem. Educ., (2007), 84, 842-

847.

[46] M.B. Bazbouz, G.K. Stylios, Alignment and optimization of nylon 6 nanofibers by

electrospinning, J. Appl. Polym. Sci., (2007), 107, 3023-3032.

[47] W. Zhou, Y. Wu, F. Wei, G. Luo, W. Qian, Elastic deformation of multiwalled

carbon nanotubes in electrospun MWCTs-PEO and MWCNTs-PVA nanofibers,

Polymer, (2005), 46, 12689-12695.

190

[48] L.H. Bluhm, T. Li, Effect of analogue ions in normal-phase ion-pair chromatograpy

of quaternary ammonium compounds, J. Chromatogr. Sci., (1999), 37, 273-276.

[49] A.L. Novotny, D. Nurok, R.W. Replogle, G.L. Hawkins, R.E. Santini, Results with

an apparatus for pressurized planar electrochromatography, Anal. Chem., (2006),

78, 2823-2831.

[50] C.F. Poole, S.K. Poole, Multidimensionality in planar chromatography, J.

Chromatogr. A, (1995), 703, 573-612.

[51] E.W. Washburn, Dynamics of capillary flow, Phys. Rev., (1921), 17, 374-375.

[52] M.S. M. Jad, S.A. H. Ravandi, H. Tavanai, R.H. Sanatgar, Wicking phenomenon in

polyacrylonitrile nanofiber yarn, Fiber Polym., (2011), 12, 801-807.

[53] F. Geiss, Fundamentals of Thin-layer Chromatography (Planar Chromatography),

Huethig, Heidelberg, Germany, Basel, Switzerland, New York, United States of

America, 1987.

[54] N.A. Ismailov, M.S. Schraiber, Farmatsiya (Sofia) (1938) 1, in : J.G. Kirchner

(Ed.), Thin-Layer Chromatography, second edition, Wiley, New York, 1978.

[55] G. Guiochon, A. Siouffi, Study of the performances of thin layer chromatography

III. flow velocity of the mobile phase, J. Chromatogr. Sci,. (1978), 16, 598-609.

[56] Sz. Nyiredy, in: Sz. Nyiredy (Ed.), Planar Chromatography: A Retrospective View

for the Third Millennium, Springer, Budapest, 2001, pp 177-199.

[57] E. Tyihak, E. Mincsovics, in: Sz. Nyiredy (Ed.), Planar Chromatography: A

Retrospective View for the Third Millennium, Springer, Budapest, 2001, p. 137-

176.

191

[58] V. Pretorius, B.J. Hopkins, J.D. Schieke, Electroosmosis. New concept for high

speed liquid chromatography, J. Chromatogr., (1974), 99, 23-30.

[59] D. Nurok, M.C. Frost, C.L. Pritchard, D.M. Chenoweth, The performance of planar

chromatography using electroosmotic flow, J. Planar Chromatogr. - Mod. TLC,

(1998), 11, 244-246.

[60] A.G. Howard, T. Shafik, F. Moffatt, I.D. Wilson, Electroosmotically driven thin-

layer electrochromatograpy on silica media, J. Chromatogr. A, (1999), 844, 333-

340.

[61] D. Nurok, M.C. Frost, D.M. Chenoweth, Separation using planar chromatography

with electroosmotic flow, J. Chromatogr. A, (2000), 903, 211-217.

[62] D. Nurok, J.M. Koers, D.A. Nyman, W.-M. Liao, Variables that affect performance

in planar chromatography with electroosmotic flow, J. Planar Chromatogr. Mod. -

TLC, (2001), 14, 409-414.

[63] D. Nurok, J.M. Koers, M.A. Carmichael, T.A. Dzido, The performance of planar

electrochromatography in a horizontal chamber, J. Planar Chromatogr. - Mod.

TLC, (2002), 15, 320-323.

[64] D. Nurok, J.M. Koers, M.A. Carmichael, Role of buffer concentration and applied

voltage in obtaining a good separation in planar electrochromatography, J.

Chromatogr. A, (2003), 983, 247-253.

[65] T.H. Dzido, R. Majewski, B. Polak, W. Golkiewicz, E. Soczewinski, Application

of a horizontal DS chamber to planar electrochromatography, J. Planar

Chromatogr. - Mod. TLC, (2003), 16, 176-182.

192

[66] A. Sze, D. Erickson, L. Ren, D. Li, Zeta-potential measurement using the

Smoluchowski equation and the slope of the current-time relationship in

electroosmotic flow, J. Colloid Interface Sci., (2003), 261, 402-410.

[67] T. Shafik, A.G. Howard, F. Moffatt, I.D. Wilson, Evaporation-induced solvent

migration in electrically-driven thin layer chromatography, J. Chromatogr. A,

(1999), 841, 127-132.

[68] A.L. Andrady, Science and Technology of Polymer Nanofibers; John Wiley &

Sons, Inc.: Hoboken, NJ, 2008.

[69] J.E. Clark, S.V. Olesik, Electrospun glassy carbon ultra-thin layer chromatography

devices, J. Chromatogr. A, (2010), 1217, 4655-4662.

[70] M.C. Beilke, J.W. Zewe, S.V. Olesik, Aligned electrospun nanofibers for ultra-thin

layer chromatography, Anal. Chim. Acta, (2013), 761, 201-208.

[71] P. Kampalanonwat, P. Supaphol, G.E. Morlock, Electrospun nanofiber layers with

incorporated photoluminescence indicator for chromatography and detection of

ultraviolet-active compounds, J. Chromatogr. A, (2013), 1299, 110-117.

[72] X. Fang, S.V. Olesik, Carbon nanotube and carbon nanorod-filled polyacrylonitrile

electrospun stationary phase for ultrathin layer chromatography, Anal. Chim. Acta,

(2014), 830, 1-10.

[73] T.E. Newsome, S.V. Olesik, Silica-based nanofibers for electrospun ultra-thin layer

chromatography, J. Chromatogr. A, (2014), 1364, 261-270.

[74] T. Lu, S.V. Olesik, Electrospun polyvinyl alcohol ultra-thin layer chromatography

of amino acids, J. Chromatogr. B, (2013), 912, 98-104.

193

[75] T. Rojanarata, S. Plianwong, K. Su-uta, P. Opanasopit, T. Ngawhitunpat,

Electrospun cellulose acetate nanofibers as thin layer chromatographic media for

eco-friendly screening of steroids adulterated in traditional medicine and

nutraceutical products, Talanta, (2013), 15, 208-213.

[76] T.H. Dzido, P.W. Plocharz, P. Slazak, A. Halka, Progress in planar

electrochromatography, Anal. Bioanal. Chem., (2008), 391, 2111-2118.

[77] T.E. Newsome, S.V. Olesik, Electrospinning silica/polyvinylpyrrolidone composite

nanofibers, J. Appl. Polym. Sci., (2014), 131, 40966(1)-40966(9).

[78] T.E. Newsome, S.V. Olesik, Planar electrochromaography using an electrospun

polymer nanofiber layer, Anal. Chem., (2014), 86, 10961-10969.

[79] W.F. Patton, V. Panchagnula, E. Rockney, and I.S. Krull, Taking a walk on the

wild side with planar electrochromatography and thin-layer electrophoresis: of

peptides, proteins, and proteomics, J. Liq. Chromatogr. Relat. Technol., (2006), 29,

1177-1218.

[80] W.J. Miao, Electrogenerated chemiluminescence and its biorelated applications,

Chem. Rev., (2008), 108, 2506-2553.

[81] I. Rubinstein, A.J. Bard, Electrogenerated chemiluminescence. 37. Aqueous ecl

2+ systems based on Ru(2,2’-bipyridine)3 and oxalate or organic acids, J. Am Chem.

Soc., (1981), 103, 512-516.

[82] I. Rubinstein, A.J. Bard, Electrogenerated chemiluminescent determination of

oxalate, Anal. Chem., (1983), 55, 1580-1582.

194

[83] J.B. Noffsinger, N.D. Danielson, Generation of chemiluminescence upon reaction

of aliphatic amines with tris(2,2’-bipyridine)ruthenium(III), Anal. Chem., (1987),

59, 865-868.

[84] S.N. Brune, D.R. Bobbitt, Effect of pH on the reaction of tris(2,2’-

bipyridyl)ruthenium(III) with amino-acids: implications for their detection,

Talanta, (1991), 38, 419-424.

[85] C.J. Miller, P. McCord, A.J. Bard, Study of Langmuir monolayers of ruthenium

complexes and their aggregation by electrogenerated chemiluminescence,

Langmuir, (1991), 7, 2781-2787.

[86] D. Yang, B. Lu, Y. Zhao, X. Jiang, Fabrication of aligned fibrous arrays by

magnetic electrospinning, Adv. Mater., (2007), 19, 3702-3706.

[87] X. Wang, Y.G. Kim, C. Drew, B.C. Ku, J. Kumar, L.A. Samuelson, Electrostatic

assembly of conjugated polymer thin layers on electrospun nanofibrous membranes

for biosensors, Nano Lett., (2004), 4, 331-334.

[88] J.M. Moran-Mirabal, J.D. Slinker, J.A. DeFranco, S.S. Verbridge, R. Llic, S.

Flores-Torres, H. Abruna, G.G. Malliaras, H.G. Craighead, Electrospun light-

emitting nanofibers, Nano Lett., (2007), 7, 458-463.

[89] T.Y. Chen, J. Leddy, Ion exchange capacity of Nafion and Nafion composites,

Langmuir, (2000), 16, 2866-2871.

[90] M.N. Szentirmay, C.R. Martin, Ion-exchange selectivity of Nafion films on

electrode surfaces, Anal. Chem., (1984), 56, 1898-1902.

195

[91] D.H. Johnston, H.H. Thorp, Cyclic voltammetry studies of polynucleotide binding

and oxidation by metal complexes: homogeneous electron-transfer kinetics, J.

Phys. Chem., (1996), 100, 13837-13843.

[92] A. Laforgue, L. Robitaille, A. Mokrini, A. Ajji, Fabrication and characterization of

ionic conducting nanofibers, Macromol. Mater. Eng., (2007), 292, 1229-1236.

[93] P.T. Anastas, L.G. Heine, T. C. Williamson, Green Chemical Syntheses and

Processes: Introduction, in: P.T. Anastas, L.G. Heine, T. C. Williamson, Ed.; Green

Chemical Syntheses and Processes, American Chemical Society, Washington, DC,

2000, Chapter 1, pp 1-6.

[94] P.T. Anastas, J.C. Warner, Green Chemistry: Theory and Practice, Oxford

University Press, New York, 1998.

[95] J. Płotka, M. Tobiszewski, A.M. Sulej, M. Kupska, T. Górecki, J. Namiesnik,

Green Chromatography, Green chromatography, J. Chromatogr. A (2013), 1307, 1-

20.

[96] P. Sandra, G. Vanhoenacker, F. David, K. Sandra, A. Pereira, Green

chromatography. Part 1: Introduction and liquid chromatography, LC GC Eur

(2010), 23, 242-259.

[97] R.K. Henderson, C. Jimenez-Gonzalez, D.J.C. Constable, S.R. Alston, G.G.A.

Inglis, G. Fisher, J. Sherwood, S.P. Binks, A.D. Curzons, Expanding GSK’s

solvent selection guide - embedding sustainability into solvent selection starting at

medicinal chemistry, Green Chem., (2011), 13, 854-862.

196

[98] K. Alfonsi, J. Colberg, P.J. Dunn, T. Fevig, S. Jennings, T.A. Johnson, H.P. Kleine,

C. Knight, M.A. Nagy, D.A. Perry, M. Stefaniak, Green chemistry tools to

influence a medicinal chemistry and research chemistry based organization, Green

Chem., (2008), 10, 31-36.

[99] A.J. Alpert, Hydrophilic-interaction chromatography for the separation of peptides,

nucleic acids and other polar compounds, J. Chromatogr., (1990), 499, 177-196.

[100] B. Buszewski, S. Noga, Hydrophilic interaction liquid chromatography (HILIC)-a

powerful separation technique, Anal. Bioanal. Chem., (2012), 402, 231-247.

[101] Z. Hao, B. Xiao, N. Weng, Impact of column temperature and mobile phase

components on selectivity of hydrophilic interaction chromatography (HILIC), J.

Sep. Sci., (2008), 31, 1449-1464.

[102] R.P. Li, J.X. Huang, Chromatographic behavior of epirubicin and its analogues on

high-purity silica in hydrophilic interaction chromatography, J. Chromatogr. A,

(2004), 1041, 163–169.

[103] B.A. Olsen, Hydrophilic interaction chromatography using amino and silica

columns for the determination of polar pharmaceuticals and impurities, J.

Chromatogr. A, (2001), 913, 113 –122.

[104] A. dos Santos Pereira, F. David, G. Vanhoenacker, P. Sandra, The acetonitrile

shortage: is reversed HILIC with water an alternative for the analysis of highly

polar ionizable solutes? J. Sep. Sci., (2009), 32, 2001–2007.

[105] T.A. Berger, Characterization of a 2.6 µm Kinetex porous shell hydrophilic

interaction liquid chromatography column in supercritical fluid chromatography

197

with a comparison to 3 µm totally porous silica, J. Chromatogr. A, (2011), 1218,

4559-4568.

[106] L. Subbarao, J. Cole, R. Chen, An empirical evaluation of HILIC and monolithic

columns for SFC applications, LC GC N. Am., (2010), Feb. Suppl. 28, 62-63.

[107] M.R. Emmett, S. Kazazic, A.G. Marshall, W. Chen, S.D.-H. Shi, B. Bolaños, M.J.

Greig, Supercritical fluid chromatography reduction of hydrogen/deuterium back

exchange in solution-phase hydrogen/deuterium exchange with mass spectrometric

analysis, Anal. Chem., (2010), 78, 7058-7060.

[108] C. West, S. Khater, E. Lesellier, Characterization and use of hydrophilic interaction

liquid chromatography type stationary phases in supercritical fluid

chromatography, J. Chromatogr. A, (2012), 1250, 182-195.

[109] J.W. Treadway, G.S. Philibert, S.V. Olesik, Enhanced fluidity liquid

chromatography for hydrophilic interaction separation of nucleosides, J.

Chromatogr. A, (2011), 1218, 5897-5902.

[110] M.J. Beres, S.V. Olesik, Enhanced-fluidity liquid chromatography using mixed-

mode hydrophilic interaction liquid chromatography/strong cation-exchange

retention mechanisms, J. Sep. Sci., (2015), 402, doi: 10.1002/jssc.201500454.

[111] A. dos Santos Pereira, A.J. Girón, E. Admasu, P. Sandra, Green hydrophilic

interaction chromatography using ethanol-water-carbon dioxide mixtures, J Sep.

Sci., (2010), 33, 834-837.

[112] Y. Cui, S.V. Olesik, Reversed-phase high-performance liquid chromatography

using enhanced-fluidity mobile phases, J. Chromatogr. A, (1995), 691, 151-162.

198

[113] ACD/Chemsketch Freeware, Advanced Chemistry Development, Inc., Toronto,

ON, Canada, 2015.

[114] A. Wakamatsu, K. Morimoto, M. Shimizu, S. Kudoh, A severe peak tailing of

phosphate compounds caused by interaction with stainless steel used for liquid

chromatography and electrospray mass spectrometry, J. Sep. Sci., (2005), 28, 1823-

1830.

[115] T.A. Berger, J.F. Deye, Composition and density effects using methanol/carbon

dioxide in packed column supercritical fluid chromatography, Anal. Chem., (1990),

62, 1181-1185.

[116] T.A. Berger, J.F. Deye, “Use of solvatochromic dyes to correlate mobile phase

solvent strength to chromatographic retention in supercritical fluid

chromatography.” Supercritical Fluid Technology. Ed. F. Bright, Washington D.C.:

American Chemical Society, 1992. 132-142.

[117] P. Hemstrom, K. Irgum, Hydrophilic interaction chromatography, J. Sep. Sci.,

(2006), 29, 1784–1821.

[118] S.H. Yang, H.P. Nguyen, K.A. Schug, Method development and analysis of mono-

and dephosphorylated nucleotides by HILIC-ESI-MS, Hydrophic Interaction

Liquid Chromatography and Advanced Applications, Ed. P.G. Wang, W. He, Boca

Raton, FL: CRC Press, 2011, 445-450.

[119] A.J. Alpert, Electrostatic repulsion hydrophilic interaction chromatography for

isocratic separation of charged solutes and selective isolation of phosphopeptides,

Anal. Chem., (2008), 80, 62-76.

199

[120] T. Zhou, C.A. Lucy, Hydrophilic interaction chromatography of nucleotides and

their pathway intermediates on titania, J. Chromatogr. A, (2008), 1187, 89-93.

[121] G.S. Philibert, Improved separation of biological compounds using enhanced-

fluidity liquid chromatography, Dissertation, The Ohio State University. Ann

Arbor: UMI, 2011. (Publication No. ATT AAI3493528.)

[122] M. Gilar, U. Neue, Peak capacity in gradient reversed-phase liquid chromatography

of biopolymers theoretical and practical implications for the separation of

oligonucleotides, J. Chromatogr. A, 1169, 139-150.

[123] Y. Gaber, U. Tornvall, M.A. Kumar, M.A. Amin, R. Hatti-Kaul, HPLC-EAT

(environmental assessment too): a tool for profiling safety, health, and

environmental impacts of liquid chromatography methods, Green Chem., (2011),

13, 2021-2025.

[124] G. Koller, U. Fischer, K. Hungerbuhler, Assessing safety, health and environmental

impact during early process development, Ind. Eng. Chem. Res., (2000), 39, 960-

972.

[125] R.K. Marcus, W.C. Davis, B.C. Knippel, L. LaMotte, T.A. Hill, D. Perahia, J.D.

Jenkins, Capillary-channeled polymer fibers as stationary phases in liquid

chromatography separations, J. Chromogr. A, (2003), 986, 17-31.

[126] D.M. Nelson, R.K. Marcus, Characterization of capillary-channeled polymer fiber

stationary phases for high-performance liquid chromatography protein separations:

comparative analysis with a packed-bed column, Anal. Chem., (2006), 76, 8462-

8471.

200

[127] D. Nurok, J.M. Koers, A.L. Novotny, M.A. Carmichael, J.J. Kosiba, Apparatus and

initial results for pressurized planar electrochromatography, Anal. Chem., (2004),

76, 1690-1695.

201