Producing Melt Blown Nano-/Micro-fibers with Unique Surface Wetting Properties

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY

Zaifei Wang

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Frank S. Bates and Christopher W. Macosko

September, 2015

© Zaifei Wang 2015

ALL RIGHTS RESERVED

Acknowledgement

The past five years at the University of Minnesota have been the most marvelous experience of my growth. There are so many beautiful memories that will never ever fade in all my life. Most importantly, I met many great people who have helped and supported me throughout my doctoral career and personal life. I am truly grateful to these friends, and thanking them here is the least that I can do to express my sincere gratitude.

First of all, I would like to thank my advisors, Dr. Frank Bates and Dr. Chris

Macosko, for their great help, support, and guidance. Easy does not enter into grown-up life, especially when I came in with a metallurgical background, neither chemistry nor chemical engineering. I could not have reached this important milestone without your encouragement and support. I have been motived by your great enthusiasm to science. I learned from you how to think critically, how to solve issues, and how to have science make difference in the real world. I appreciated this excellent and joyful experience learning and exploring science under your guidance. Also, I would like to thank Dr.

Timothy Lodge, Dr. Satish Kumar, and Dr. Xiang Cheng for their willingness to review my thesis, provide comments, and be in my defense.

I am grateful to many collaborators at the University of Minnesota. I thank Dr.

Dawud Tan and Dr. Feng Zuo for getting me start quickly in the project. I thank Dr.

Satish Kumar, Dr. Changkwon Chung, Dr. Leonardo Espín for their great collaboration. I thank Dr. David Giles for his trainings on rheometers. I thank Dr. Conrado Aparicio at the School of Dentistry for offering the device to measure water contact angles. I thank

i

Wieslaw Suszynski for helping with setting up high-speed camera, Chris Frethem and

Fang Zhou for useful discussions on SEM and microtome, and Dr. Bing Luo for training and discussing XPS. I appreciated the hard work from Xiaotun Liu and helpful discussions from Huaiyu Yan. I would also like to express my gratitude to people from

Cummins Filtration, Dr. Soondeuk Jeung, Dr. Kan Wang, Dr. Bill Haberkamp, Dr. Peter

Herman for their experimental support and helpful discussions. In addition, I thank Dr.

Deborah Massouda from DuPont for offering the new perfluorinated copolyesters and useful discussions. I thank Dr. Jeff Galloway for offering the water-dispersible polymers,

Dr. Scott George for helpful discussions on the material properties.

Life would have been black and white if there were no friends. I thank the following friends for sharing the wonderful life at the University of Minnesota: Zhen Ren, Zhongda

Pan, Jie Song, Jing Han, Donglin Tang, Han Zhang, Lian Bai, Yuqiang Qian, Jun Xu,

Chris Thurber, Yogesh Dhande, Yunlong Zou, Yoshihasi Takamori, Aaron Hedegaard,

Luca Martinetti, Shigeru Aoyama, Alex Mannion, Liangliang Gu, Siddharth Chanpuriya,

Karen Haman, Sam Dalsin, Tessie Panthani, Camelo Declet-Perez, Yulong Li, Xi Chen.

I am grateful for warm and continuous support from my wife Zihan (Vera) Cheng, my parents, my younger brother, my grandparents, and Debbie Gustafson. It has been a long-distance relationship for five years. I could not think of any word to express my sincere gratitude to my wife. I thank my wife for her extreme support over the past five years. Also, I thank my parents and my younger brother. I could not have been through all the difficult times without their support. In addition, I want to thank Debbie for her great help and support in my life. She has been part of my family in my heart.

ii

Finally, I would like to thank Cummins Filtration for their financial support. And, parts of this work were carried out in the Institute of Technology Characterization

Facility at the University of Minnesota.

iii

Dedication

To my family

iv

Abstract

Melt blowing is a one-step process producing microfibers with an average fiber diameter (dav) of 1 – 5 μm. One application of melt blown fibers is filter media of water- in-diesel engine fuel filtration. In order to meet the increasingly higher requirements on fuel cleanliness, enhancing the filter efficiency has been desired. One attractive strategy is applying nanofibers (dav < 1 μm). The other is modifying the fiber surface wetting properties, such as superhydrophobic. The goal of this work is to produce melt blown nano-/micro-fibers with superhydrophilic and/or superhydrophobic surfaces for water-in- diesel filtration.

In this thesis, we advanced the new technique of producing nanofibers from melt blown fiber-in-fiber polymer blends by understanding the nanofiber formation and developing water-extractable nanofibers. Firstly, poly(styrene) (PS)/poly(butylene terephthalate) (PBT) and PS/poly(ethylene-co-chlorotrifluoroethylene) (ECTFE) binary blends were melt blown followed by tetrahydrofuran (THF) extraction. The control extrusions demonstrated that the core nanofibers were primarily generated during the fiber blowing process. We speculate that both aerodynamic and fiber-fiber drag influence the nanofiber formation. Varying the minor phase volume fraction revealed the nanofiber breakup. We proposed that both the single-droplet-elongation and dumbbell-formation contribute to the formation of nanofibers. In addition, we hypothesize that the coalescence leads to the formation of non-uniform nano-/micro-fibers.

Following the discussion on nanofiber formation, we showed nanofiber fabrication

v from water-extractable melt blown fiber-in-fiber polymer blends containing a water- dispersible sulfopolyester (SP). This new method eliminates issues associated with organic solvents. Also, it provides another route to prepare multilayer nano-/micro-fiber composites.

Surface wetting modifications of melt blown PBT fibers were achieved by alkaline hydrolysis and subsequent fluorination. Application of alkaline hydrolysis generates sponge-like PBT fibers decorated with hydroxyl and carboxyl functional groups, resulting in a superhydrophilic fiber mat surface. The subsequent fluorination creates a sticky-superhydrophobic surface.

An alternative achieving enhanced or even superhydrophobic PBT fiber surfaces is incorporation of a random perfluorinated multiblock copolyester (PFCE). Adding only 5 wt% of PFCE into PBT leads to over 20 wt% of PFCE on PBT fiber surfaces. The blooming of fluorine and the overall mat roughness increased the water contact angle from about 128° to above 150° and lowered the fiber mat surface adhesion energy by

25%.

Finally, the appendices show some preliminary works on: 1) synthesis and characterizations of poly(L-lactide) and perfluoropolyether block copolymer; 2) the effects of alkaline hydrolysis on surface topography and mechanical properties of melt blown PBT fiber mats; 3) compatibilized 6 and poly(ethene-co- tetrafluoroethene) (PETFE) blends, which, however, were not suitable for melt blowing due to high viscosity; 4) collection of individual melt blown nanofibers.

vi

Table of Contents

Acknowledgement……………………………………………………………….……….i Dedication……………………………………………………………………………...... iv Abstract…………………………………………………………………………………..v List of Tables……………………………………………………………………………xii List of Figures………………………………………………………………………….xiii

Chapter 1 Introduction...... 1

1.1 Nonwoven ...... 1 1.2 Melt blown nanofibers ...... 3 1.2.1 Spinneret design ...... 5

1.2.2 Air & Polymer flow rates (Qa, Qp) ...... 8

1.2.3 Air & Polymer temperatures (Ta, Tp) ...... 10

1.2.4 Air pressure (Pin) & Air velocity (va) ...... 11 1.2.5 Die-tip-to-collector distance (DCD) ...... 12 1.2.6 Molecular weight & MFI ...... 14 1.2.7 Viscoelasticity ...... 15 1.3 Surface wetting modifications ...... 17 1.4 Thesis overview ...... 21 Chapter 2 Nanofiber Formation from Melt Blown Fiber-in-Fiber Polymer Blends 23

2.1 Introduction ...... 23 2.2 Experimental ...... 25 2.2.1 Materials ...... 25 2.2.2 Preparation of polymer blends ...... 27 2.2.3 Characterization of blend morphology ...... 29 2.2.4 Melt blowing and solvent-extraction ...... 29 2.2.5 Fiber characterization...... 31 2.3 Results ...... 32

vii

2.3.1 Blend morphology ...... 32 2.3.2 Morphology of the melt blown fibers ...... 36 2.3.3 Morphology of the extracted nanofibers ...... 42 2.4 Discussions ...... 50 2.4.1 Effects of composition on the blend morphology ...... 50 2.4.2 Effects of composition on the melt blown fiber morphology ...... 52 2.4.3 Effects of composition on the nanofiber morphology ...... 52 2.4.4 Nanofiber formation...... 53 2.5 Conclusion ...... 64 Chapter 3 Nanofibers from Water-extractable Melt Blown Fiber-in-Fiber Polymer Blends……………………………………………………………………………………65

3.1 Introduction ...... 65 3.2 Experimental ...... 67 3.2.1 Materials ...... 67 3.2.2 Preparation of polymer blends ...... 70 3.2.3 Characterization of blend morphology ...... 70 3.2.4 Melt blowing and water-extraction ...... 71 3.2.5 Fiber characterization...... 71 3.3 Results & Discussion ...... 72 3.3.1 Polymer blends...... 72 3.3.2 Melt blown fibers ...... 73 3.3.3 Water-extracted nanofibers ...... 76 3.3.4 Multi-layer PBT nano-/micro-fiber composites...... 78 3.4 Conclusion ...... 79 Chapter 4 Tuning Surface Properties of Poly(butylene terephthalate) Melt Blown Fibers by Alkaline Hydrolysis and Fluorination...... 81

4.1 Introduction ...... 81 4.2 Experiment ...... 83 4.2.1 Fabrication of melt blown PBT fibers ...... 83

viii

4.2.2 NaOH hydrolysis of PBT fibers ...... 84 4.2.3 Fluorination of h-PBT fibers ...... 84 4.2.4 Characterization of fibers and fiber mats ...... 85 4.3 Results and Discussion ...... 88 4.3.1 Size and integrity of the fiber mats ...... 88 4.3.2 Mass loss and fiber diameter...... 91 4.3.3 Fiber surface morphology ...... 96 4.3.4 Superhydrophilicity of h-PBT fiber mats ...... 98 4.3.5 Superhydrophilic f-PBT fiber mats ...... 103 4.4 Conclusion ...... 106 Chapter 5 Fluorine Enriched Melt Blown Fibers from Polymer Blends of Poly(butylene terephthalate) and A Perfluorinated Copolyester ...... 108

5.1 Introduction ...... 108 5.2 Experimental ...... 111 5.2.1 Materials ...... 111 5.2.2 Preparation of polymer blends ...... 113 5.2.3 Melt blowing ...... 114 5.2.4 Characterization of polymer blends ...... 114 5.2.5 Characterization of fiber morphology ...... 115 5.2.6 Characterization of fiber surface properties ...... 116 5.3 Results and Discussion ...... 117 5.3.1 Polymer blends of PBT/PFCE ...... 117 5.3.2 Melt blown fibers from PBT/PFCE ...... 122 5.3.3 Quantification of PFCE on fiber surfaces ...... 127 5.3.4 Fluorine enrichment on melt blown PBT/PFCE fibers ...... 136 5.3.5 Surface wetting of melt blown PBT/PFCE fibers ...... 140 5.4 Conclusion ...... 145 Chapter 6 Exploring Melt Blown Fibers: Shaping the Future of Engineering Materials……………………………………………………………………………….147

ix

6.1 Introduction ...... 147 6.2 Bimodal distributed nano-/micro-fibers ...... 147 6.3 Self-reinforced melt blown fibers ...... 151 6.4 Superoleophobicity & Wetting robustness ...... 157 6.5 Water-in-oil Filtration ...... 162 References ...... 167

Appendix A Synthesis and Characterization of Poly(L-lactide) and Perfluoropolyether Block Copolymer ...... 192

A.1 Introduction ...... 192 A.2 Experimental ...... 193 A.2.1 Materials ...... 193 A.2.2 Synthesis of PLLA-PFPE-PLLA ...... 193 A.2.3 Characterizations of PLLA-PFPE-PLLA ...... 194 A.2.4 Melt blowing & Spin coating ...... 195 A.2.5 Fiber characterization...... 195 A.2.6 Surface characterization ...... 195 A.3 Results and Discussion ...... 196 A.3.1 Synthesis & Characterizations of PLLA-PFPE-PLLA ...... 196 A.3.2 Melt blown fibers ...... 199 A.3.3 Surface wetting properties ...... 202 A.4 Summary ...... 205 Appendix B Effects of Alkaline Hydrolysis on Morphology and Mechanical Properties of Melt Blown Poly(butylene terephthalate) Fibers ...... 206

B.1 Introduction ...... 206 B.2 Experimental ...... 206 B.2.1 Alkaline hydrolysis of melt blown PBT fiber mats ...... 206 B.2.2 Characterization of fibers and fiber mats ...... 207 B.2.3 Tensile tests ...... 208 B.3 Results and Discussion ...... 209

x

B.3.1 Physical characterizations of the mats ...... 209 B.3.2 Fiber diameter and surface topography ...... 211 B.3.3 Mechanical properties of h-PBT fiber mats ...... 214 B.4 Summary ...... 216 Appendix C Compatibilized Polyamid 6 and Poly(ethylene-alt-tetrafluoroethylene) Blends…………………………………………………………………………………..217

C.1 Introduction ...... 217 C.2 Experimental ...... 217 C.2.1 Materials ...... 217 C.2.2 Preparation of polymer blends ...... 220 C.2.3 Characterization of polymer blend morphology ...... 220 C.2.4 Surface wetting characterization ...... 220 C.3 Results & Discussion ...... 221 C.3.1 Polymer blends...... 221 C.3.2 Surface wetting ...... 224 C.4 Summary ...... 224 Appendix D Collection of Individual Melt Blown Fibers using A Rotating Frame…………………………………………………………………………………..226

xi

List of Tables

Table 1.1 Parameters influencing the diameter and diameter distribution of melt blown fibers ...... 5

Table 2.1 Melt Flow Index (MFI), thermal, and rheological properties of PBT, PS, and ECTFE...... 26

Table 2.2 Compositions of polymer blends in this study...... 28

Table 2.3 Interfacial tensions calculated from polymer surface tension ...... 51

Table 3.1 Molecular weight and thermal properties of PBT, SP1, and SP2...... 68

Table 3.2 Compositions of polymer blends in this study...... 70

Table 4.1 Effects of NaOH hydrolysis temperature (T) and duration (t) on porosity (ε), fiber volume fraction (εf), average fiber diameter (dav), and coefficient of variation (CV) ...... 93

Table 5.1 An example of calculating the solubility parameter (δ) of PBT using the group- contribution method...... 119

Table 5.2 Surface chemical compositions derived from the photoemission areas of C 1s, O 1s, and F 1s collected from XPS survey scans...... 129

Table 5.3 Line fitting results of C 1s and O 1s for PBT ...... 131

Table 5.4 The line fitting results of C 1s and O 1s for PFCE...... 133

Table 5.5 Surface chemical compositions derived from the line fitting of C 1s of melt blown PBT/PFCE blends...... 136

Table 6.1 Design parameters of the robust composite surfaces ...... 159

Table 6.2 Molecular weight of PLLA-PFPE-PLLA ...... 172

Table 6.3 Effects of the hydrolysis on the physical properties of h-PBT fiber mats...... 186

Table A.1 Molecular weight of PLLA-PFPE-PLLA ...... 197

Table B.1 Effects of the hydrolysis on the physical properties of h-PBT fiber mats. .... 211

xii

List of Figures

Figure 1.1 The schematic of the conventional melt blowing line.20 ...... 4

Figure 1.2 Schematic of the original Exxon slot melt blowing die...... 6

Figure 1.3 Schematic of the melt blowing die composed of stacked plates from Fabbricante et al.24 ...... 7

Figure 1.4 Schematic of our single-hole melt blowing die: (a) sectional and (b) end-on views of the two pieces...... 8

Figure 1.5 Effects of air flow rate on the average fiber diameter of melt blown PBT fibers...... 9

Figure 1.6 Effects of air-to-polymer mass flux ratio on (a) the average fiber diameter and (b) diameter distribution of melt blown PBT fibers...... 10

34 Figure 1.7 The effect of air velocity (va) on average fiber diameter from Milligan et al...... 12

Figure 1.8 The effect of DCD on the average diameter of melt blown fibers obtained from both the commercial and research melt blowing lines.39 ...... 13

Figure 1.9 The effect of DCD on the average diameter of melt blown thermoplastic (TPU) fibers.40 ...... 14

Figure 1.10 The elasticity (λ1) effects of the starting materials on (A) the average fiber diameter dav (B) the diameter distribution CV of melt blown PS fibers...... 17

Figure 1.11 Liquid droplet wetting behavior on (a) ideal solid surfaces, (b) Wenzel state, and (c) Cassie-Baxter state...... 18

Figure 2.1 Complex viscosity (η*), elastic modulus (G’), and viscosity modulus (G’’) of PS, PBT, and ECTFE measured by dynamic frequency sweep at 265 °C with 25mm parallel plates...... 27

Figure 2.2 A schematic of the HAAKE batch compounder used to prepare polymer blends in this study...... 28

Figure 2.3 Photograph and schematic of our lab-scale melt blowing apparatus...... 30

Figure 2.4 An exemplified quantification of fiber diameter and diameter distribution.. .. 32

Figure 2.5 Representative SEM images of PS/PBT blends with a composition of (a) ϕ = 5

xiii

%; (b) ϕ = 25 %; (c) ϕ = 30 %; (d) ϕ = 40 %...... 33

Figure 2.6 Representative SEM images of PS/ECTFE blends with a composition ratio of (a) ϕ = 5 %; (b) ϕ = 25 %; (c) ϕ = 30 %; (d) ϕ = 40 %...... 34

Figure 2.7 Effect of the blend composition on the number average diameter Dn,av of the dispersed droplets...... 35

Figure 2.8 Effects of the blend composition on (a) the geometric average droplet diameter Dav and (b) the corresponding droplet size distribution CV of PS-containing blends within the range of 0 ≤ ϕ ≤ 30 %...... 35

Figure 2.9 Comparison of the minor phase droplet distribution between (a) PS/PBT_30 (ϕ = 30 %) and (b) PS/ECTFE_30 (ϕ = 30 %)...... 36

Figure 2.10 Representative SEM images showing the morphology and the corresponding diameter distribution of melt blown fibers of (a) PS/PBT_5; (b) PS/PBT_10; (c) PS/PBT_25; (d) PS/PBT_30; (e) PS/PBT_40...... 38

Figure 2.11 Representative SEM images showing the morphology and the corresponding diameter distribution of melt blown fibers of (a) PS/ECTFE_5; (b) PS/ECTFE _10; (c) PS/ECTFE _25; (d) PS/ECTFE _30; (e) PS/ECTFE _40...... 40

Figure 2.12 The effect of blend composition ϕ on (a) melt blown fiber diameter dav and (b) fiber diameter distribution CV...... 41

Figure 2.13 Representative SEM images and the corresponding fiber diameter analysis of PBT nanofibers extracted from melt blown (a) PS/PBT_5 fibers; (b) PS/PBT_10 fibers; (c) PS/PBT_25 fibers; (d) PS/PBT_30 fibers (d) PS/PBT_40 fibers...... 44

Figure 2.14 Representative SEM images and the corresponding fiber diameter analysis of ECTFE nanofibers extracted from melt blown (a) PS/ECTFE_5 fibers; (b) PS/ECTFE _10 fibers; (c) PS/ECTFE _25 fibers; (d) PS/ECTFE _30 fibers (d) PS/ECTFE _40 fibers...... 46

Figure 2.15 Representative SEM images and the corresponding fiber diameter analysis of ECTFE nanofibers extracted from melt blown (a) PS/ECTFE_5 fibers; (b) PS/ECTFE _10 fibers; (c) PS/ECTFE _25 fibers; (d) PS/ECTFE _30 fibers (d) PS/ECTFE _40 fibers...... 48

Figure 2.16 The effect of blend composition on (a) the average fiber diameter (dN, av) and (b) fiber diameter distribution of nanofibers...... 49

Figure 2.17 Schematic of the formation of nanofibers-in-microfiber from melt blown immiscible binary polymer blends...... 54

xiv

Figure 2.18 Representative SEM images showing the morphology of the minor phase from both the cryo-fractured cross-section and THF-extracted extrudates of (a) PS/PBT_5 blend; (b) PS/PBT_25 blend; (c) PS/ECTFE_5 blend; (d) PS/ECTFE_25 blend...... 56

Figure 2.19 Schematic of the hypothesized blend morphology before blowing extension. (a) bimodal distribution, (b) side-by-side...... 61

Figure 3.1 The chemical structure of the sulfopolyester (SP)...... 68

Figure 3.2 Complex viscosity (η*), elastic modulus (G’), and viscosity modulus (G’’) of PBT, SP1, SP2 at 265 °C...... 69

Figure 3.3 Representative SEM images showing the morphology and the corresponding droplet distribution of (a) SP1/PBT_30 blend, (b) SP2/PBT_30 blend...... 72

Figure 3.4 The morphology and the corresponding diameter distribution of melt blown fibers of (a) SP1/PBT_30, (b) SP2/PBT_30...... 74

Figure 3.5 The morphology and the corresponding diameter distribution of melt blown fibers of (a) SP1/PBT_5, (b) SP1/PBT_10, (c) SP1/PBT_40...... 76

Figure 3.6 Morphology and the corresponding diameter distribution of water-extracted PBT nanofibers from melt blown (a) SP1/PBT_5, (b) SP1/PBT_10, (c) SP1/PBT_30, (d) SP1/PBT_40...... 78

Figure 3.7 A representative SEM image showing the morphology of double-layer PBT nano-/micro-fiber composite...... 79

Figure 4.1 Schematic of the experimental set-up employed to record the water droplet absorption by superhydrophilic h-PBT fiber mats...... 87

Figure 4.2 Fiber mat of (a) untreated PBT (b) h-PBT treated at 45 °C for 20 min (c) f- PBT...... 89

Figure 4.3 Thermal analysis of (a) PBT and representative h-PBT fibers (b) hydrolysis effect on crystallinity (Xc)...... 90

Figure 4.4 Mechanism of PBT alkaline hydrolysis...... 91

Figure 4.5 Mass loss (Δm %) of h-PBT fiber mats...... 92

Figure 4.6 The impact of NaOH hydrolysis on (a) dav and (b) CV...... 94

Figure 4.7 Representative distributions of fiber diameters and associated log-normal function for (a) untreated PBT fibers, and h-PBT fibers treated at 45 °C for (b) 10 min, (c) 20 min, (d) 30 min...... 95

xv

Figure 4.8 SEM images of (a) Untreated PBT fibers, and h-PBT fibers treated at 45 °C for (b) 10 min, (c) 20 min, (d) 30 min...... 97

Figure 4.9 SEM images of h-PBT fiber morphology (a) 50 °C, 10 min; (b) 50 °C, 20 min; (c) 50 °C, 30 min; (d) 55 °C, 10 min; (e) 55 °C, 20 min; (f) 55 °C, 30 min...... 98

Figure 4.10 Water contact angle of (a) PBT melt blown fiber mat (b) PBT compression molded film ...... 99

Figure 4.11 Sessile drop measurements on h-PBT fiber mats treated (a) at 45 °C for 10 min (b) at 45 °C for 20 min...... 100

Figure 4.12 Water droplet spreading and imbibition on h-PBT-4 fiber mat, captured by a 1000 fps camera...... 101

Figure 4.13 Influence of NaOH etching duration on the rate of water droplet absorption...... 102

Figure 4.14 Mechanism for reaction of carboxyl group and PFOA ...... 103

Figure 4.15 Analysis of (a) the surface chemical composition of f-PBT fibers by XPS and (b) the surface topography of f-PBT fibers by SEM...... 104

Figure 4.16 (a) Water contact angle on f-PBT fiber mat ...... 105

Figure 4.17 Schematic formation of the sticky-superhydrophobicity on f-PBT fiber mat ...... 106

Figure 5.1 Molecular structure of PFCE...... 112

Figure 5.2 Rheological properties of PBT and PFCE. (a) Complex viscosity (η*); (b) elastic modulus (G’) and viscosity modulus (G’’)...... 113

Figure 5.3 Representative SEM images of PBT/PFCE blends with a composition of (a) ϕ = 1 wt%; (b) ϕ = 3 wt%; (c) ϕ = 5 wt%; (d) ϕ = 10 wt%...... 118

Figure 5.4 DSC analysis of PBT, PFCE, and PBT/PFCE blends...... 118

Figure 5.5 Representative ATR-IR analysis of PBT, PFCE, and their blends: (a) normalized IR spectra and (b) the corresponding 2nd derivatives...... 121

Figure 5.6 Representative SEM images and the corresponding statistical fiber diameter analysis of (a) PBT/PFCE_1; (b) PBT/PFCE_3; (c) PBT/PFCE_5; (d) PBT/PFCE_10...... 123

Figure 5.7 Representative SEM images showing the internal morphology of cryo-

xvi fractured melt blown fibers of PBT/PFCE_10 blend. (a) longitudinal surface; (b) cross- sectional surface...... 124

Figure 5.8 Representative TEM image showing the internal (cross-sectional) morphology of melt blown PBT/PFCE_10 fibers...... 125

Figure 5.9 Representative SEM images showing the morphology of (a) PBT/PFCE_10 blend, (b) PBT/PFCE_10 annealed at 240 °C, (c) PBT/PFCE_10 sheared at 240 °C, (c) PBT/PFCE_10 extrudate from 240 °C...... 127

Figure 5.10 XPS survey spectra showing the photoemission areas of C 1s, O 1s, and F 1s of (a) PBT/PFCE_1; (b) PBT/PFCE_3; (c) PBT/PFCE_5; (d) PBT/PFCE_10...... 128

Figure 5.11 Line fitting analysis of (a) C 1s envelopes and (b) O 1s envelopes of pure melt blown PBT fibers...... 130

Figure 5.12 Line fitting analysis of (a) C 1s envelopes, (b) O 1s envelopes, and (c) F 1s envelope of PFCE...... 132

Figure 5.13 High-resolution scans of C 1s collected from melt blown fibers of pure PBT and PBT/PFCE blends, respectively...... 134

Figure 5.14 Line fitting analysis of C 1s from melt blown fibers of PBT/PFCE_10 (ϕ = 10 %) blend...... 135

Figure 5.15 Surface enrichment of PFCE on melt blown fibers of PBT/PFCE blends. . 137

Figure 5.16 Comparison between the effects of the loaded fluorinated component on the surface fluorine enrichment of melt blown and electrospun fibers...... 138

Figure 5.17 Typical images demonstrating the static water contact angle measurements on (a) PBT fiber mat, CA = 128 ± 4°; and (b) PBT/PFCE_5 fiber mat, CA = 155 ± 3°. 141

Figure 5.18 Impact of the PFCE concentration in the bulk on the static water contact angle of melt blown PBT fiber mats...... 142

Figure 5.19 Variation of both the advancing contact angle (θA) and receding contact angle (θR) as a function of the PFCE concentration in the bulk...... 143

Figure 5.20 Effect of the PFCE concentration on the adhesion energy of melt blown PBT fibers...... 144

Figure 6.1 A representative SEM image showing the morphology of bimodal distributed polyamide fibers.281...... 148

Figure 6.2 Schematic of the combination of melt blowing and electrospinning used to

xvii produce modal distributed nano-/micro-fibers. The figure is adopted from Erben et al.217 ...... 149

Figure 6.3 Schematic of the parallel electrospinning set-up to produce bimodal distributed nanofibers.281 ...... 150

Figure 6.4 The morphology of ECTFE nano-/micro-fibers after THF extraction of (a) the melt blown PS/ECTFE_30 fibers, (b) the melt blown PS/ECTFE_40 fibers...... 151

Figure 6.5 Schematic of the shish-kebab crystalline structure. The figure is adopted from Mackley et al.288 ...... 153

Figure 6.6 Representative SEM images showing the shish-kebab-like morphology observed on NaOH hydrolyzed melt blown PBT fibers. (a) An overview of the mat structure, (b) A zoom-in showing the shish-kebab-like fibers...... 156

Figure 6.7 Schematic of micro-hoodoos. The figure is adopted from Tuteja et al.318 .... 158

Figure 6.8 Design parameters for a robust composite surface. (A) electrospun fibers, (B) and (C) micro-hoodoos. All the parameters in (A), (B), and (C) are consistent with Tuteja et al.322 ...... 159

Figure 6.9 Design chart of robust liquid repellent surfaces. The figure is adopted from Tuteja et al.319...... 160

Figure 6.10 Flatsheet filter holder for fuel/water filtration. The picture is adopted from Yang et al.340 ...... 162

Figure 6.11 The filter holder (filtration cell) that Chase Group used for water-in-diesel filtration...... 163

Figure 6.12 Two-stage filter setup designed by Krasinski and Wierzbaa.55 (a) schematic of flow, (b) photograph of the two-stage filter system: 1-coalescence element, 2-sample point, 3-separation element...... 164

Figure 6.13 Our filtration cell (a) 3-D view, (b) the photograph of the cell...... 165

Figure 6.14 The design of our filtration cell...... 166

Figure A.1 The molecular structure of PFPE...... 193

Figure A.2 Photograph of the synthesized PLLA-PFPE-PLLA triblock copolymer...... 196

Figure A.3 (a) 1H (b) 19F NMR characterizations of PLLA-PFPE-PLLA...... 197

Figure A.4 Dynamic elastic modulus (G’) and loss modulus (G’’) of PLLA-PFPE-PLLA.

xviii

...... 198

Figure A.5 Dynamic complex viscosity of PLLA-PFPE-PLLA at 180, 190, 200, 210 °C...... 199

Figure A.6 Representative SEM images of (A) PLLA-PFPE-PLLA fibers, Tprocessing = 180 °C, (B) irregularities at Tprocessing = 180 °C, (C) and (D) PLLA-PFPE-PLLA fibers at Tprocessing = 200 °C...... 200

Figure A.7 Volumetric comparison of a piece of melt blown PLLA-PFPE-PLLA fiber mat between (A) before-soaking and (B) after-soaking...... 201

Figure A.8 Differential scanning calorimety results for PLLA-PFPE-PLLA meltblown fibers. (a) cooling at 10 °C/min; (b) cooling at 50 °C/min...... 202

Figure A.9 Water contact angle (θ) on spun-cast films of : (a) PLLA, θPLLA = 76 ± 5° ; (b) PLLA-PFPE-PLLA, θPLLA-PFPE-PLLA = 106 ± 3° ...... 203

Figure A.10 XPS analysis of PLLA-PFPE-PLLA spun-cast film...... 204

Figure A.11 Water contact angle on (a) pure PLLA fiber mat, θPLLA = 115 ± 5°, (b) PLLA-PFPE-PLLA fiber mat, θPLLA-PFPE-PLLA = 143 ± 3° ...... 204

Figure B.1 Schematic of the tensile test setup...... 209

Figure B.2 Effects of hydrolysis time on the mass loss of h-PBT fiber mats...... 210

Figure B.3 Comparison in the length and width between h-PBT-50 and PBT fiber mats...... 210

Figure B.4 Effect of NaOH hydrolysis on dav and CV...... 212

Figure B.5 Representative SEM images of surface morphologies of (a) PBT (b) h- PBT_10 (c) h-PBT_30 (d) h-PBT_60 (e) h-PBT_70 (f) h-PBT_100...... 214

Figure B.6 Representative images showing the tensile test...... 214

Figure B.7 Representative engineering stress-strain curves of the h-PBT fiber mats. ... 216

Figure C.1 TGA analysis of the mateirals...... 218

Figure C.2 (a) Complex viscosity and (b) dynamic elastic modulus (G’) and loss modulus (G’’) of PA6, PETFE, PETFE-g-MA...... 219

Figure C.3 Representative SEM images showing the morphology of (a) PA6/PETFE_90/10 wt%, (b) PA6/PETFE_80/20 wt%, (c) PA6/PETFE_70/30 wt%, (d)

xix

PA6/PETFE_50/50 wt%...... 221

Figure C.4 Schematic of the compatibilization between ...... 222

Figure C.5 Representative SEM images showing the morphology of (a) PA6/PETFE-g- MA_90/10 wt%, (b) PA6/PETFE-g-MA_80/20 wt%, (c) PA6/PETFE-g-MA _70/30 wt%, (d) PA6/PETFE-g-MA _50/50 wt%...... 223

Figure C.6 Representative SEM image showing the morphology of compatibilized PETFE-g-MA droplet in a PA6 matrix...... 223

Figure C.7 Effect of PETFE-g-MA on the surface wetting of PA6 molded films...... 224

Figure D.1 Schematic of the collection of individual melt blown fibers using a rotating frame...... 227

xx Chapter 1 Introduction

This chapter starts with a general introduction to nonwovens. Among various nonwoven production techniques, we focus on melt blowing which is currently one of the major manufacturing methods in industry. One application of melt blown fibers is water- in-diesel fuel filtration. Driven by the requirement of increasingly higher filter performance, melt blown nanofibers (dav < 1 μm) and modified fiber surfaces (e.g. superhydrophilic/superhydrophobic) are desired. Approaches of reducing the diameter of melt blown fibers and fundamentals of surface wetting are presented in this chapter.

Finally, the overview of this thesis is presented.

1.1 Nonwoven

Nonwovens are broadly defined as sheet or web structures bonded together by individual entangling fibers or filaments (and by perforating films) mechanically, thermally, or chemically.1 They are not made by weaving or knitting and do not require any downstream processing to convert the fibers to yarn.2 Compared with other conventional woven and textile fabrics, nonwovens exhibit comparable or even superior appearance and mechanical strength.3 Moreover, nonwovens are engineered to provide specific functions to ensure fitness for purpose.

The nonwoven market has been growing steadily in the past two decades. The global demand for nonwovens is forecast to rise 5.3 % annually to 9.0 million metric tons in

2017.4 Nonwoven fibers have been widely used and made an impact in a variety of

1 applications ranging from baby diapers, medical apparel, to filter media.5-9 Applications like battery separators are opening new doors to nonwovens and driving the innovations of nonwoven technology.10, 11

Nonwovens are generally classified into dry-laid, wet-laid, and polymer-laid based on the manufacturing process. In dry-laid web formation, fibers are carded (including carding and cross-lapping) or aerodynamically formed (air-laid) and then bonded by mechanical, chemical or thermal methods.2 Wet-laid nonwovens originated from papermaking refer to paper-like nonwoven fabrics which are manufactured with machinery designed to manipulate short fibers suspended in liquid. The last but the most important is polymer-laid. Polymer-laid nonwovens include melt spinning, spunbonding, melt blowing, electrospinning, and forcespinning.

Melt spinning is by far the most widespread system globally and it produces fibers typically with an average diameter larger than 10 μm.12 Additionally, conjugate spinning like islands-in-the-sea, can be accomplished by modifying the spinneret.13, 14 Both spunbonding and melt blowing are currently the two main processes in the production industries. They have been growing in parallel since their inception in the late 1950s and for some applications they are used in combination to produce multilayer laminates (e.g. spunbond/meltblown/spunbond which is also known as SMS).2 The conventional spunbonding technique fabricates microfibers (dav ~ 5 – 10 μm). In contrast, melt blowing produces finer fibers with an average diameter of 1-5 μm, and we will address this issue more in the following sections. Electrospinning has attracted much attention because of its capability to produce nanofibers as small as 1.6 nm.15 Forcespinning

2 emerged recently as a new method to produce nanofibers. It overcomes the present limitations of electrospinning such as solvent recovery.12

A variety of polymers have been used for polymer-laid nonwovens, in particular for melt blowing and spunbonding. Polypropylene is the most widely used material because of its low viscosity (high MFR), versatile molecular structure (fiber grade PP resins are mainly isotactic homopolymers), and low cost. Biodegradable polymers have also been used for nonwovens. In 2009, NatureWorks announced that they developed new grades of Ingeo poly(lactic acid) (PLA) for melt blown nonwovens.16

1.2 Melt blown nanofibers

Fabrication of nonwoven nanofibers has stirred up great interest and facilitated more rapid development of nonwovens. Although the National Science Foundation (NSF) defines nanofibers as having at least one dimension of 100 nm or less, the nonwoven industry generally considers nanofibers as having diameters of less than 1 μm.17

Nanofibers offer an array of unique features, such as small pore size, large surface area, and high level of molecular alignment, which can enhance the material performance. For example, liquid and aerosol filtration can benefit greatly from the introduction of nanofibers.18, 19

Currently electrospinning is the most popular process to produce nanofibers.

However, there are several disadvantages limiting the industrialization of this technique, such as solvent handling/recycling, high voltage, and high cost associated with equipment maintenance. Furthermore, many commercial polymers having desirable properties are difficult to dissolve in common solvents at room temperature, such as poly(butylene

3 terephthalate) (PBT) and ethylene-co-chlorotrifluoroethylenethe (ECTFE).

Forcespinning is an attractive alternative but it has been argued that the spinneret design depends on the material properties, which becomes costly.12

Melt blowing has offered economic mass production of fine fibers in the range of 1-5

μm. During melt blowing, polymer is fed into the extruder where it is heated and melted until the appropriate temperature and viscosity are reached. Then, the molten polymer is fed to the melt blowing die. Upon exiting the die the polymer melt is drawn down by hot high-velocity air jets and cooled down by the ambient air as it travels towards the fiber collector, as shown in Figure 1.1. Thus, the fibers are formed between the extrusion temperature and the solidification temperature (i.e. glass transition temperature Tg or crystallization temperature Tc).

Figure 1.1 The schematic of the conventional melt blowing line.20

If the capability of melt blowing can be extended to nanofibers (dav < 1 μm), it would provide a more attractive alternative to fabricate nanofibers than electrospinning as well

4 as forcespinning. There has been some research focusing on reducing the size of melt blown fibers below 1 μm. Factors influencing the diameter and diameter distribution of melt blown fibers can be generally categorized into process and material parameters, as summarized in Table 1.1. The impacts of these parameters are addressed as follows.

Table 1.1 Parameters influencing the diameter and diameter distribution of melt blown fibers

Process parameters Material parameters

Spinneret (die) design --

Air & polymer flow rates (Qa, Qp) Viscoelasticity

Air & polymer temperatures (Ta, Tp) Molecular weight

Air pressure in the die (Pin) Melt flow index (MFI)

Air velocity (va) --

Die-tip-to-collector distance (DCD) --

1.2.1 Spinneret design

Among the process parameters, the spinneret (die) design is the most important because it directly ties to the fiber diameter and production rate. The initial melt blowing die is a slot die developed by Van A. Wente in the 1950s.21 Later on, Exxon developed a

36-inch-wide die based on the “slot” concept, as shown in Figure 1.2.22, 23 This die design produced fibers as small as 500 nm. But more importantly, this design improved the throughput from the original 0.03 – 0.09 lb/hour/inch by Wente to commercially attractive throughput levels of 1.3 – 2.6 lb/hour/inch.22 Fabbricante et al patented a new design where the die consists of stacked plates resulting in a row of orifices as small as

5 0.0125 mm in diameter, as shown in Figure 1.3.24 They reported that fibers with an average diameter of 300 nm could be achieved. Podgorski et al. developed a new die where the polymer nozzles were surrounded by air nozzles, obtaining nanofibers with diameters ranging from 0.74 – 1.41 μm.25 Brang et al. produced nanofibers using a modified die with plate edge profile having very large length to diameter ratio and small orifice diameter.26

Figure 1.2 Schematic of the original Exxon slot melt blowing die.

6

Figure 1.3 Schematic of the melt blowing die composed of stacked plates. The picture is adopted from Fabbricante et al.24

In 2007, Ellison et al. from our group reported nanofiber fabrication with a single- hole die having an orifice with a diameter of 0.2 mm.27 Our melt blowing die is based on the industrial design of the Exxon die. As shown in Figure 1.4, an air-gap distance (dg) of

1 mm and a set-back distance (ds) of 1.5 mm were employed. Each of the v-slot channels were 10 mm by 1 mm. Poly(butylene terephthalate) (PBT) nanofibers with an average diameter of 440 nm and poly(propylene) (PP) nanofibers with an average diameter of 300 nm were successfully achieved.

7

Figure 1.4 Schematic of our single-hole melt blowing die: (a) sectional and (b) end-on views of the two pieces.

1.2.2 Air & Polymer flow rates (Qa, Qp)

Air flow rate (Qa) and polymer flow rate (Qp) play key roles in the final fiber diameter. Increasing Qp while fixing other processing conditions can increase the average fiber diameter, as confirmed by several researchers, because the same drag force from the

28-30 air jets is acting on more polymer mass. While increasing Qa reduces the average fiber diameter because of an increased drag force. However, the increase of air flow rate, usually achieved by increasing the air pressure (Pin), can break up fibers and generate loose and uncollected fibers known as flies.31 Also, Ellison et al. found droplet-

8 27 instabilities with the increase of air flow rate. Tan also studied the effects of both Qa

32 and Qp on fiber diameter and diameter distribution. PBT and PP were fabricated into melt blown fibers with a 5-hole melt blowing die (dhole = 0.2 mm). As shown in Figure

1.5, increasing Qa reduces dav, but the fibers are still larger than 1 μm.

Figure 1.5 Effects of air flow rate on the average fiber diameter of melt blown PBT fibers. The data were adopted from Tan and the figure was replotted.

It is clear from Figure 1.5 that decreasing Qp along with increasing Qa results in nanofibers, implying the importance of Qp on dav. This observation brings in the air-to- polymer mass flux ratio (Γ) which was first employed by Milligan and Haynes.33 Γ incorporates not only Qp but also the die geometry, as is seen in the definition of Γ in equation (1.1).

Qa Air Exit Area   (1.1) Qp Polymer Exit Area

9 Figure 1.6 shows the influence of Γ on both dav and the diameter distribution

(represented by the coefficient of variation, namely CV). Firstly, increasing Γ results in a

33 decrease in dav, which follows a power law decay, as reported by other researchers. On the other hand, increasing Γ broadens the fiber diameter distribution CV, which has been attributed to the increase in the turbulence of the air flow field.32

Figure 1.6 Effects of air-to-polymer mass flux ratio on (a) the average fiber diameter and (b) diameter distribution of melt blown PBT fibers. Data were adopted from Tan, and the figures were replotted.

1.2.3 Air and polymer temperatures (Ta, Tp)

It is intuitive that increasing air temperature (Ta) and polymer temperature (Tp) can reduce the fiber diameter because the fiber attenuation occurs between the processing temperature (Tpr) and solidification temperature (Tg or Tc). Milligan et al. reported that increasing melt temperature can decrease the average fiber diameter of melt blown polypropylene to about 3 μm.34 Haynes et al. patented a design to increase the length of the meltblown jet thermal core in order to increase the dwell time of the polymer filament, but the size of the resulting fibers was actually not substantially reduced.35 Lee and Wadsworth demonstrated that increasing the die temperature from roughly 210 °C to

10 250 °C led to nearly 50% reduction in the average fiber diameter, but the smallest fiber diameter they obtained was around 5 μm.36 Zhang et al. also reported that the fiber diameter can be decreased to about 2 μm by increasing either the polymer melt temperature or air temperature. To the best of our knowledge, there are no reports of achieving nanofibers by simply increasing Tp or Ta, which is possibly because of the risk that high temperature is likely to cause thermal degradation.

1.2.4 Air pressure (Pin) and velocity (va)

Air pressure (Pin) and air velocity (va), which are often associated with the air flow rate (Qa), can impact the fiber diameter. Choi et al. reported that the diameter can be

37 38 decreased to less than 5 μm by increasing the Pin at the die. Uppal et al. fabricated polypropylene nanofibers with average diameters of 500 nm by applying high air pressure (Pin > 100 kPa) at the die. However, Milligan et al. reported that the fiber diameter is insensitive to the air velocity when va is increased over about 280 mps, as demonstrated in Figure 1.7.34 This interesting observation is contrary to the results obtained by other researchers. In general, increasing the air pressure at the die (Pin) and air velocity (va) can reduce the average fiber diameter to nanometers. But it has not been favored by industries since it significantly increases the cost of production.

11

34 Figure 1.7 The effect of air velocity (va) on average fiber diameter. The picture is from Milligan et al.

1.2.5 Die-tip-to-collector distance (DCD)

The die-tip-to-collector distance (DCD) can also influence the fiber diameter since it correlates with the dwell time of the fiber attenuation resulting from both the aerodynamic drag and fiber-fiber contact/entanglement, and its impact varies with different materials because of the inherent material properties like crystallization.

Although the fiber attenuation primarily occurs within the first 5 – 6 cm from the die exit, the diameter keeps changing at a smaller rate until fiber laydown is completed.39 Bresee et al studied the effect of DCD on the diameter melt blown PP fibers with commercial and research melt blowing lines.39 As shown in Figure 1.8, the average fiber diameter decreases with the increase of DCD. However, the influence is very weak. A 10-cm-

12 increase of DCD can only reduce the average fiber diameter by 2.3 % for the research line and 4 % for the commercial line. They also found that the maximum fiber diameter increased with the increase of DCD, which they attributed to the fiber-fiber fusion.

Figure 1.8 The effect of DCD on the average diameter of melt blown polypropylene fibers obtained from both the commercial and research melt blowing lines. The figure was replotted from Bresee et al.39

Lee et al studied the influence of DCD on the average diameter of met blown PP fibers.36 They found that the average fiber diameter appears to be insensitive to the DCD after a DCD of 30 cm. Interestingly, Lee et al reported that the average diameter of melt blown thermoplastic polyurethane (TPU) increased with increasing of DCD, as shown in

Figure 1.9. Comparing this result with their previous studies on melt blown PP fibers, they attributed the fiber diameter increase with the increase of DCD to the thermal and elastic relaxations.40

13

Figure 1.9 The effect of DCD on the average diameter of melt blown thermoplastic polyurethane (TPU) fibers. Interestingly, increasing DCD leads to an increase in the fiber diameter, which is different from melt blown PP fibers. The figure is from Lee et al.40

Tan found that the change in DCD had little impact on the diameter of melt blown

PBT fibers.32 The average fiber diameter and diameter distribution appeared to be insensitive to the DCD variations. This is likely because of the fast crystallization of PBT which weakens the influence of DCD on fiber diameter.

1.2.6 Molecular weight & MFI

Melt blowing is compatible with a variety of polymers and polymer blends, which makes it competitive and attractive compared with electrospinning. However, in order to achieve nanofibers, the molecular weight associated with the melt flow index (MFI) needs to be appropriate. Bhat and Malkan suggested that the melt flow rate (MFR) suitable for melt blowing is 30 – 1500.41 In general, a low molecular weight

14 corresponding to a low viscosity and a high MFI is desired, because it can lower the processing temperature and thus decrease the manufacturing cost. Nayak et al produced melt blown PP nanofibers (average fiber diameter ~ 800 nm). The PP they used has a

MFI of 300.42 Ellison et al. used PP with a MFR of 1500 at 230 °C. The MFR of PP that

Hassan et al. used to produce nanofibers was as high as 1800.43

Along with the average molecular weight the molecular weight distribution is also important. A narrow molecular weight distribution reduces the melt elasticity and melt strength of the polymer so that the melt stream can be drawn into fine denier filaments without excessive draw force.41, 44, 45 A broad molecular weight distribution is prone to produce fiber breaks due to draw resonance (melt instability).41

1.2.7 Viscoelasticity

Viscoelasticity can influence the fiber diameter and diameter distribution.

Uyttendaele and Shambaugh first established a 1D model involving the drag effects, hear transfer, and viscoelastic effects to simulate the dynamics of melt blown fibers.46 They used the Phan-Thien and Tanner (PTT) constitutive equation and concluded that the viscoelasticity had little effect on the fiber diameter.47 Later on, Shambaugh and his coworkers improved the 1D model and developed 2D and 3D models which incorporated the lateral velocity perturbations of the fibers, allowing them to simulate the fiber whipping during melt blowing.48, 49 Both the 2D and 3D models also indicated that the viscoelasticity had little influence on the fiber diameter. Jarecki et al. studied the dynamics of stationary air drawing during the fabrication of melt blown PP fibers based on a single-filament model in a thin-filament approximation that accounts for polymer

15 viscoelasticity. They did not provide insights into the effects of viscoelasticity.50

Tan et al. systematically studied the viscoelasticity effects on fiber diameter and diameter distribution through experiments.51 Poly(styrene) (PS) was selected as the model material because it is amorphous and avoids the complexity of crystallization kinetics. Tan blended small amounts of high molecular weight PS with a very low molecular weight PS matrix (Mn = 3 kg/mol) for melt blowing. This permitted tuning of the viscoelasticity so that the viscous and elastic effects could be either isolated or combined. Through this study it was demonstrated that increasing the melt viscosity (η0) leads to an increase in average fiber diameter (dav) but has little impact on fiber diameter distribution (coefficient of variation, CV). On the other hand, increasing elasticity

(correlated with the longest melt relaxation time λ1) beyond a threshold value increased dav while reducing CV, as shown in Figure 1.10.

Subsequently, our collaborators, Kumar and coworkers, studied melt blowing of viscoelastic materials using a slender-jet model.52 The upper-convected Maxwell, Phan-

Thien and Tanner (PTT), and Giesekus constitutive equations have been used to explore the effects of viscoelasticity. These authors found that viscoelasticity tends to enlarge the fiber diameter, consistent with experimental observations. Also, the model suggested that non-uniform shear stress and thermal effects do no play important roles. In addition, they showed that viscoelasticity reduces the magnitude of disturbance amplification during melt blowing, which indicates a mechanism for the narrower fiber diameter distribution observed in Tan’s experiments.

16

Figure 1.10 The elasticity (λ1) effects of the starting materials on (A) the average fiber diameter dav (B) the diameter distribution CV of melt blown PS fibers. The figure is from Tan et al. The error bars show the spread of dav and CV values from two or three melt blowing runs.

1.3 Surface wetting modifications

The surface wetting properties of melt blown fibers play essential roles in both the academic and industrial communities, for example, in understanding and designing superhydrophobic/superhydrophilic fibrous structures for self-cleaning, anti-fogging, protective clothes, and water-oil separation.53-55 In particular, the surface wetting of melt blown fibers is of great importance to water-in-diesel fuel filtration. Water in diesel fuel can promote biological growth and reduce the lubricity of the fuel, leading to plugging of the engine components and abrasion wear.56, 57 Tuning surface wetting properties of the filter media, achieving either superhydrophilicity or superhydrophobicity, has been demonstrated to be an effective way to optimize the filter performance.58, 59

The understanding of wetting on solid surfaces was first established for ideal solid surfaces which are defined as being smooth, rigid, insoluble, nonreactive, and chemically

17 homogeneous.60 Young pioneered the research of wetting on ideal solid surfaces and developed the well-known Young’s equation (1.2),

 SG SL cosY  (1.2)  LG

where θY is the Young contact angle, γSG and γLG are the surface tensions of the solid and liquid, respectively, and γSL is the solid-liquid interfacial tension. It is seen that the surface tension (surface energy) associated with the surface chemical compositions plays an important role in wetting phenomenon. According to the measured contact angle (CA), a solid surface can be defined as being hydrophilic (CA < 90°) or hydrophobic (CA >

90°).

Figure 1.11 Liquid droplet wetting behavior on (a) ideal solid surfaces, (b) Wenzel state, and (c) Cassie- Baxter state.

However, real solids are rough, which affects their surface wetting properties along with the surface chemical compositions. Because of the roughness, there exists a range of the measured CA, which is referred to as CA hysteresis (H). It is worth addressing that not only surface roughness but also chemical heterogeneity can induce the CA

18 hysteresis.61 The most useful theories describing the wetting on rough real solid surfaces are embodied in the Wenzel and Cassie-Baxter equations. The Wenzel wetting state (as is seen in Figure 1.11 (b)) assumes that the liquid follows all the topological variations of the material so that every piece of liquid/air interface gets replaced by a solid/liquid interface of the same surface area, as expressed in equation (1.3),

coswY r cos (1.3)

where θw is the apparent CA, r (r > 1), defined as the ratio of the true surface area of the solid to its horizontal projection.62, 63 Wenzel’s equation predicts that the wetting is improved by r if θY < 90° and that anti-wetting is enhanced if θY > 90°. In contrast, the

Cassie-Baxter wetting state postulates that the liquid droplet sits on a composite surface of flat solid tops and flat air pockets where the apparent CA is given by equation (1.4),

cosCB ff1 cos  1 2 cos  2 (1.4)

where f1and f2 are liquid/solid contact area fractions of solid components 1 and 2 on the surface, respectively. θ1 and θ2 are the equilibrium contact angles of the same liquid on each of the flat surfaces of these components.64 When air is present, equation (1.4) becomes,

cosCB ff1 cos 1  1  1 (1.5)

On one hand, roughness can induce superhydrophobic surfaces (CA > 150°, H < 10°)

19 which is often explained based on both the Wenzel and Cassie-Baxter equations. The two wetting states often coexist and can transit from one to the other by external perturbations.65-69 On the other hand, roughness can lead to superhydrophilicity. Drelich defined superhydrophilic surfaces as those textured and/or structured materials (rough and/or porous) having a surface roughness factor (as defined by Wenzel equation) larger than one, r > 1, on which water (liquid) spreads completely.70

There have been several approaches developed to modify the surface wetting properties of nonwoven fibers, such as wet coating, physical vapor deposition (PVD), chemical vapor deposition (CVD), nanoparticle coating, layer-by-layer deposition (LBL), surface grating, and plasma treatment. Several researchers have reviewed these techniques for modifying the surface wetting properties of nonwoven fibers.71-73

Currently, the dominant approach in industry is wet coating. But it has been reported that wet coating is not durable and results in non-uniform surfaces.74 More importantly, it can potentially influence the pore size of the mats, which impacts the materials performance in applications like filtration.74 Another important method is blending in additives during the fiber fabrication process. Additive can migrate to fiber surfaces and thus modify the fiber surface wetting properties. This approach has been used primarily to improve the dyeability of nonwoven fibers.75

More interestingly, most publications reporting superhydrophobic/superhydrophilic fiber surfaces have focused on electrospun fibers rather than melt blown fibers.

Electrospun fibers generally have a smaller diameter (< 500 nm) than melt blown fibers, which facilitates tuning the surface wetting properties because of the greater surface-to-

20 volume ratio.

1.4 Thesis overview

Producing melt blown nano-/micro-fibers with unique surface wetting properties for water-in-diesel filtration in diesel engines is the main driving force of this project. The research presented in this dissertation has been divided into five chapters and three appendices.

Chapter 2 covers the investigation of nanofiber fabrication from melt blown fiber-in- fiber polymer blends. We describe the underlying mechanism of nanofiber formation inside single melt blown microfibers during the melt blowing process. We focus on primarily the effects of the blend composition on morphology, melt blown fiber diameter, and the nanofiber diameter. In Chapter 3, we further advance this technology by developing water-extractable nanofibers so that the problems of solvent handling and recycling associated with organic solvents are eliminated.

Both Chapter 4 and Chapter 5 show surface wetting modifications of melt blown PBT fiber mats. In Chapter 4, alkaline hydrolysis followed by the subsequent simple fluorination is used to tune the surface properties of melt blown PBT fibers. Alkaline hydrolysis imparts superhydrophilicity and the subsequent fluorination generates sticky- superhydrophobicity to the melt blown PBT fiber mats. Chapter 5 discusses fluorine enriched melt blown PBT fibers by blending in small amounts of a perfluorinated random multiblock copolyester (PFCE). We observed significant surface blooming which affected the surface wetting properties of melt blown PBT fibers are discussed.

Chapter 6 provides some future perspectives associated with both the fundamental 21 studies and potential applications of melt blown fibers. The possibility of producing bimodal-distributed and self-reinforced melt blown fibers are briefly explored. Then, we discuss superoleophobicity and the wetting robustness of the fiber mats. In addition, the design of the filtration cell for water-in-oil will be presented.

This dissertation concludes with three appendices. Appendix 1 summarizes the synthesis and characterizations of poly(L-lactide) and perfluoropolyether block copolymer. Appendix 2 presents the effects of alkaline hydrolysis on surface topography and mechanical properties of melt blown PBT fiber mats. Appendix 3 covers compatibilized polyamide 6 and poly(ethene-alt-tetrafluoroethene) (PETFE) blends.

Appendix 4 shows the collection of single melt blown fibers.

22

Chapter 2 Nanofiber Formation from Melt Blown Fiber-in-Fiber Polymer Blends

2.1 Introduction

Reducing the diameter of nonwoven fibers from microns (1 – 100 µm) to nanometers

(1 – 1000 nm) can dramatically increase the fiber surface area, improve the flexibility in surface functionalities, decrease the pore size of the fiber mats, and potentially improve the fiber mechanical properties,76 which enhances the materials performance. Nonwoven nanofibers can be used for filtration media,5 packaging materials,77 performance apparel,78, 79 biological tissue engineering,80-82 medical textiles,83, 84 and composite materials.85, 86 The global nanofiber market is forecast to rise 30.3% annually to $570.2 million by 2017.4

We have developed a new method to produce nanofibers from melt blown binary immiscible polymer blends. In this new approach, we can produce the hierarchical structure of fiber-in-fiber through one-step melt blowing immiscible polymer blends and then achieve nanofibers by dissolving the matrix polymer.87 We obtained poly(ethylene- co-chlorotrifluoroethylene) (ECTFE) nanofibers with an average diameter of 70 nm through melt blowing poly(butylene terephthalate) (PBT) and ECTFE blend followed by washing away the PBT matrix using trifluoroacetic acid (TFA). This new method appears to be a versatile technique and provides a facile and potentially high-throughput route to

23 produce nanofibers.

However, the formation of nanofibers inside single melt blown microfibers has not been well understood yet. Melt blowing is a process where molten polymer is drawn down rapidly into fibers with an average diameter of 1-5 μm by hot high-velocity air jets.

There have been some studies, both numerical simulations and experimental work, describing fiber formation during melt blowing,22, 27, 46, 51, 52, 88, 89 but none of these focuses on the formation of fiber-in-fiber. Another similar process is melt spinning of islands-in-the-sea. Some effort has been put into the analysis of the “islands” fiber formation during melt spinning.90-95 These research mainly studied the effects of draw ratios and blend compositions on fibril morphology. It was demonstrated that Taylor’s theory was not applicable with the minor phase deformation along the spin line. Reduced capillary number (Ca*), which we will discuss more in Section 2.4, was used to estimate the droplet deformation/breakup.90-95 In addition, it has been proved that fibrils are formed between die exit and solidification positions and that coalescence leads to fibril formation. Nonetheless, the prediction of the final morphology is still not clearly established because of the complexity of the temperature, the flow field, the viscoelastic effects, and the competition between deformation, breakup and coalescence of the dispersed phase.90 Compared with melt spinning, melt blowing is more complex due to both highly non-linear air drawing and 3D motion of the fibers. Understanding the formation of melt blown fiber-in-fiber would be beneficial to further development of the process.

In this study, we investigated the effects of the blend composition, which is directly

24 tied to the production rate of nanofibers, on nanofiber formation using poly(styrene) (PS) and PBT and PS and ECTFE blends as model materials. The minor phase, either PBT or

ECTFE, was blended with PS at compositions of 5 vol%, 10 vol%, 25 vol%, 30 vol%, and 40 vol%. Then, the blends were melt blown with our lab-scale melt blowing apparatus followed by solvent extraction of the majority phase PS. The morphologies of polymer blends, melt blown fibers, and the extracted nanofibers were analyzed using electron microscopy. Discussions on the morphologies of the blends, melt blown fibers, and the extracted nanofibers nanofiber formation are presented. This work targets an understanding of the nanofiber formation of melt blown fiber-in-fiber morphology.

2.2 Experimental

2.2.1 Materials

PS (Sigma Aldrich, Mw = 200 kg/mol, Mw/Mn = 2.3), PBT (Ticona Celanex 2008),

ECTFE (Solvay Halar 1400LC) were the starting materials used in this study. The thermal properties of these materials were characterized by differential scanning calorimetry (DSC) (TA Instrument Q1000). The samples were analyzed under a standard run sequence of heating-cooling-heating from -20 °C to 250 °C at a scan rate of 10

°C/min, and Tg and Tm were analyzed in the second heating run. The DSC results are summarized in Table 2.1.

Rheological properties were measured with a strain-controlled ARES rheometer (TA

Instrument) equipped with a 25 mm parallel-plate fixture. Firstly, dynamic strain sweeps were performed to determine the linear viscoelastic regions of the materials. Secondly,

25 dynamic frequency sweeps were employed to measure the complex viscosity (η*) and dynamic elastic modulus (G’) and loss modulus (G”) as a function of frequency (1 ≤ ω ≤

100 rad/s) (3 % strain) at 265 °C, as shown in Figure 2.1.

Table 2.1 Melt Flow Index (MFI), thermal, and rheological properties of PBT, PS, and ECTFE.

Melt Flow Index ρ T T η at 265 °C λ Materials g m 0 (MFI) / T (°C) (g/cm3) (°C) (°C) (Pa • s) (10-3 s) PBTb 250 / 250 1.50 43 223 60 2 PSa -- 1.05 100 -- 170 6 ECTFEb 500 / 275 1.67 75 233 40 0.8 a. The molecular weight of PS was measured by size exclusion chromatography (SEC) using

tetrahydrofuran (THF) as the solvent, Mw = 200 kDa, Ð = 2.04. b. The molecular weight of PBT and ECTFE could not be measured because of their poor solubility in THF which is employed for SEC measurement. c. The longest relaxation time λ is estimated from Maxwell model, see Morrison, F. A.96

26

Figure 2.1 Complex viscosity (η*), elastic modulus (G’), and loss modulus (G’’) of PS, PBT, and ECTFE measured by dynamic frequency sweep at 265 °C with 25mm parallel plates.

2.2.2 Preparation of polymer blends

Pellets of PBT and ECTFE were dried overnight (about 6-9 hours) under vacuum at

100 °C, while PS pellets were dried at about 70 °C under vacuum to remove the moisture absorbed by the materials. Then, the polymer pellets were weighed and mixed at room temperature with compositions that are summarized in Table 2.2. The mixed pellets were fed into a HAAKE (Thermo Scientific) batch compounder at 265 °C with roller blades rotating at 100 rpm. The schematic of the compounder is shown in Figure 2.2. The blending was maintained for 8 minutes before the resulting blend was quenched with liquid nitrogen (LN2) to freeze the blend morphology.

27

Figure 2.2 A schematic of the HAAKE batch compounder used to prepare polymer blends in this study.

Table 2.2 Compositions of polymer blends in this study

PS PBT ECTFE Sample ϕ (%) (wt% / vol%) (wt% / vol%) (wt% / vol%) PS/PBT_5 5 93 / 95 7 / 5 -- PS/PBT_10 10 86 / 90 14 / 10 -- PS/PBT_25 25 68 / 75 32 / 25 -- PS/PBT_30 30 62 / 70 38 / 30 -- PS/PBT_40 40 51 / 60 49 / 40 -- PS/ECTFE_5 5 92 / 95 -- 8 / 5 PS/ECTFE_10 10 85 / 90 -- 15 / 10 PS/ECTFE_25 25 65 / 75 -- 35 / 25 PS/ECTFE_30 30 60 / 70 -- 40 / 30 PS/ECTFE_40 40 49 / 60 -- 51 / 40

Note: composition ϕ = (vol%)minor phase

28

2.2.3 Characterization of blend morphology

Polymer blends were prepared through two approaches for scanning electron microscopy. One approach is cryo-fracture in LN2, and the other one is microtomy at -60

°C. The resulting sample surfaces were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. A scanning electron microscope (SEM) (Hitachi S-

4700 or JEOL 6500) was used to image the sample surface. Both secondary and backscattered electron detectors were employed. 20 – 25 SEM micrographs were taken for each sample. Approximately 300 – 400 droplets of the minor phase were measured with ImageJ software. Then, the measurement results were fitted with a log-normal distribution function using the software package of OriginLab to obtain the number average droplet size (Dav).

2.2.4 Melt blowing and solvent-extraction

A lab-scale melt blowing apparatus was used in this study, as shown in Figure 2.3.

The apparatus consists of a custom-built melt blowing die (the die design is described in

Chapter 1) attached onto a capillary rheometer and home-made fiber collector. The

TM capillary rheometer (Goettfert Rheo-Tester 1500) was used to heat and feed the molten polymer to the melt blowing die. Our melt blowing die mimics the configuration of an industrial die which is usually referred to as the Exxon die.97 A melt blowing die containing five 0.2-mm-diameter holes was used in this study. The continuous fiber collector (drum diameter = 32 cm, screen width = 18 cm) was used to mimic the

29 industrial process. The collector was rotated with a DC motor whose speed was controlled with a custom-built controller. A blower was placed behind the perforated fiber collector to create a weak suction for efficient fiber collection.

The materials were dried under vacuum overnight (6 – 9 hours) at 70 °C. Melt blowing experiments were performed at Qpolymer ≈ 0.2 g/(min • hole), Pinlet = 6 psi, and

Tprocessing = 265 °C. Melt blown fiber mats were collected with a continuous fiber collector located about 35 cm away from the melt blowing die tip. The control extrusion experiments were carried out under the same conditions as melt blowing except for no air blowing.

Figure 2.3 Photograph and schematic of our lab-scale melt blowing apparatus.

30

In order to expose the nanofibers, samples of melt blown fibers were cut (~ 3 x 3 cm2) from the collected continuous fiber mats with a scissors or a rotary cutter (Fiskars, 45 mm diameter). The samples of PS-containing melt blown fibers were soaked in THF without stirring for 3 times at room temperature. Each soaking lasted for 2 hours. The extracted materials were dried in the air for 24 hours. The materials were weighed both before and after soaking.

2.2.5 Fiber characterization

Both melt blown fibers and solvent-extracted nanofibers were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. 25-30 SEM micrographs were taken for each sample with a scanning electron microscope (SEM)

(Hitachi S-4700 or JEOL 6500). The diameters of 400 - 500 fibers were measured using the ImageJ software. As shown in Figure 2.4 (a), a diagonal line (from the top right to the bottom left) is first drawn on the SEM image, which helps rule out measured fibers.

Then, set up the scale in ImageJ by drawing a line along the scale bar. After that, fibers at the focus are measured by drawing short bars across the fiber diameter. The OriginLab was employed to fit the measurement results with a log-normal distribution function

(equation (2.1)). The geometric average fiber diameter (dav) was extracted from the log- normal fitting. The geometric standard deviation (GSD) and coefficient of variation (CV) were calculated based on the log-normal fitting results using equation (2.2) and equation

(2.3), respectively.98

31

Figure 2.4 An example quantification of fiber diameter and diameter distribution. (a) An SEM image showing the measurement of fiber diameter using ImageJ software. (b) The corresponding fiber diameter distribution fitted with a log-normal distribution function using OriginLab software package. The geometric average fiber diameter (dav), geometric standard deviation (GSD), and diameter distribution (CV) were extracted from the fitting. The sample is melt blown PS/PBT_5 used in this study.

11d fd( ) exp[ {ln( )}] 2 (2.1) d 2 2 dav

GSD  exp( ) (2.2)

2 CV exp( )  1  100% (2.3)

2.3 Results

2.3.1 Blend morphology

As shown in Figure 2.5 and Figure 2.6, the minor phase, either PBT or ECTFE, appears as approximately spherical droplets dispersed in a PS matrix within the range of

0 ≤ ϕ ≤ 30 %. The interfaces between the droplets and matrix show sharp gaps/cavities.

32

(a) PS/PBT_5 (ϕ = 5 %) (b) PS/PBT_25 (ϕ = 25 %)

(c) PS/PBT_30 (ϕ = 30 %) (d) PS/PBT_40 (ϕ = 40 %)

Figure 2.5 Representative SEM images of PS/PBT blends with a composition of (a) ϕ = 5 %; (b) ϕ = 25 %; (c) ϕ = 30 %; (d) ϕ = 40 %. The samples were prepared by microtoming the blends at -60 °C. SEM micrographs were captured using a secondary detector. Phase inversion occurs at ϕ = 40 %, where PS appears as approximate droplets dispersed in a PBT matrix.

33

(a) PS/ECTFE_5 (ϕ = 5 %) (b) PS/ECTFE_25 (ϕ = 25 %)

(c) PS/ECTFE_30 (ϕ = 30 %) (d) PS/ECTFE_40 (ϕ = 40 %)

Figure 2.6 Representative SEM images of PS/ECTFE blends with a composition ratio of (a) ϕ = 5 %; (b) ϕ = 25 %; (c) ϕ = 30 %; (d) ϕ = 40 %. The samples were prepared by microtoming at -60 °C. SEM micrographs were captured using a backscattered detector which is able to distinguish ECTFE and PS based on the differences in atomic number. ECTFE containing atoms with higher atomic numbers like F and Cl appears brighter than PS because it generates more back-scattered electrons.

Both the size and size distribution of the droplets increase with the minor phase concentration in the range of 0 ≤ ϕ ≤ 30 %, as shown in Figure 2.7 and Figure 2.8. The droplet size at ϕ = 40 % were not analyzed for both PS/PBT and PS/ECTFE blends, because phase inversion occurs, as evidenced by Figure 2.5 (d) and Figure 2.6 (d) where the brighter ECTFE or PBT phase encapsulates the PS matrix. Within 0 ≤ ϕ ≤ 30 %, the 34

Dav of PBT droplets is slightly smaller than that of ECTFE droplets in a PS matrix at the same composition. However, the ECTFE domains at ϕ = 30 % show approximately a bimodal distribution because of the coalescence, which deviates Dav and the corresponding CV extracted from the log-normal fitting, as demonstrated in Figure 2.9.

Figure 2.7 Effect of the blend composition on the number average diameter Dn,av of the dispersed droplets.

Figure 2.8 Effects of the blend composition on (a) the geometric average droplet diameter Dav and (b) the corresponding droplet size distribution CV of PS-containing blends within the range of 0 ≤ ϕ ≤ 30 %. At ϕ = 30 %, the distribution of ECTFE droplets in the PS matrix fails to follow a log-normal distribution.

35

Figure 2.9 Comparison of the minor phase droplet distribution between (a) PS/PBT_30 (ϕ = 30 %) and (b)

PS/ECTFE_30 (ϕ = 30 %). PS/ECTFE_30 (ϕ = 30 %) deviates from the log-normal fitting, so Dav, GSD, and CV are not accurate.

2.3.2 Morphology of the melt blown fibers

Melt blowing PS-containing polymer blends produces uniform microfibers. As shown in Figure 2.10 and Figure 2.11, entangled cylindrical fibers are randomly distributed, resulting in a porous network. But, ribbon-like fibers can be observed occasionally.

(a) PS/PBT_5 (ϕ = 5 %)

36

(b) PS/PBT_10 (ϕ = 10 %)

(c) PS/PBT_25 (ϕ = 25 %)

(d) PS/PBT_30 (ϕ = 30 %)

37

(e) PS/PBT_40 (ϕ = 40 %)

Figure 2.10 Representative SEM images showing the morphology and the corresponding diameter distribution of melt blown fibers of (a) PS/PBT_5 (ϕ = 5 %); (b) PS/PBT_10 (ϕ = 10 %); (c) PS/PBT_25 (ϕ = 25 %); (d) PS/PBT_30 (ϕ = 30 %); (e) PS/PBT_40 (ϕ = 40 %). The fiber diameter distributions were fitted with a log-normal distribution function using the OriginLab software package. The geometric average fiber diameter (dav), geometric standard deviation (GSD), and diameter distribution (CV) were extracted from the fitting.

(a) PS/ECTFE_5 (ϕ = 5 %)

38

(b) PS/ECTFE_10 (ϕ = 10 %)

(c) PS/ECTFE_25 (ϕ = 25 %)

(d) PS/ECTFE_30 (ϕ = 30 %)

39

(e) PS/ECTFE_40 (ϕ = 40 %)

Figure 2.11 Representative SEM images showing the morphology and the corresponding diameter distribution of melt blown fibers of (a) PS/ECTFE_5 (ϕ = 5 %); (b) PS/ECTFE _10 (ϕ = 10 %); (c) PS/ECTFE _25 (ϕ = 25 %); (d) PS/ECTFE _30 (ϕ = 30 %); (e) PS/ECTFE _40 (ϕ = 40 %). The fiber diameter distributions were fitted with a log-normal distribution function using the OriginLab software package. The geometric average fiber diameter (dav), geometric standard deviation (GSD), and diameter distribution (CV) were extracted from the fitting.

The average diameters (dav) of both the melt blown PS/PBT and PS/ECTFE fibers seem to be insensitive to ϕ, as shown in Figure 2.12 (a). The average fiber diameters are within the range of 1 – 1.5 μm. Also, the dav shows no dependence on the constituent

(either PS/PBT or PS/ECTFE) of the blend at a given ϕ. On the other hand, the diameter distribution (CV) of the melt blown PS/ECTFE fibers seems to narrow down from about

70 % to less than 60 %, while the CV of the melt blown PS/PBT fibers broadens from about 40 % to over 50 %.

40

Figure 2.12 The effect of blend composition ϕ on (a) melt blown fiber diameter dav and (b) fiber diameter distribution CV.

41

2.3.3 Morphology of the extracted nanofibers

The resulting melt blown microfibers possess hierarchical fiber-in-fiber structure where nanofibers are embedded inside single melt blown microfibers.87 Extracting the PS matrix of the melt blown PS/PBT (0 ≤ ϕ ≤ 30%) fibers and PS/ECTFE (0 ≤ ϕ ≤ 25%) fibers using tetrahydrofuran (THF) can expose continuous PBT and ECTFE nanofibers, respectively, as shown in Figure 2.13 and Figure 2.15. These nanofibers have large aspect ratios (length/diameter ≥ 104) under the assumption of no fiber break-up. Irregularities such as spheres, elongated ellipsoids, and wave-like sinusoidal fibers are also observed at

ϕ = 5 %, as shown in Figure 2.13 (a) and Figure 2.15 (a). Increasing ϕ tends to suppress irregularities and improve the uniformity of the nanofibers.

Also, nanofiber bundles are observed in both the PBT and ECTFE nanofibers. Each bundle contains from tens to hundreds of nanofibers sticking together. Moreover, it appears that the larger the ϕ the more significantly the nanofibers bundle.

When ϕ goes up to 40 %, where PS and PBT phase inverts in the starting blend, washing away the PS matrix generates a nonuniform fibrous structure consisting of PBT nanofibers and microfibers simultaneously, as shown in Figure 2.13 (e) and Figure 2.15

(e). Similar structures are also observed even with the melt blown PS/ECTFE_30 blend where phase inversion is absent.

42

(a) PBT nanofibers extracted from melt blown PS/PBT_5 (ϕ = 5 %)

(b) PBT nanofibers extracted from melt blown PS/PBT_10 (ϕ = 10 %)

(c) PBT nanofibers extracted from melt blown PS/PBT_25 (ϕ = 25 %)

43

(d) PBT nanofibers extracted from melt blown PS/PBT_30 (ϕ = 30 %)

(e) PBT nanofibers extracted from melt blown PS/PBT_40 (ϕ = 40 %)

Figure 2.13 Representative SEM images and the corresponding fiber diameter analysis of PBT nanofibers extracted from (a) melt blown PS/PBT_5 (ϕ = 5 %) fibers; (b) melt blown PS/PBT_10 (ϕ = 10 %) fibers; (c) melt blown PS/PBT_25 (ϕ = 25 %) fibers; (d) melt blown PS/PBT_30 (ϕ = 30 %) fibers (d) melt blown PS/PBT_40 (ϕ = 40%) fibers.

44

(a) ECTFE nanofibers extracted from melt blown PS/ECTFE_5 (ϕ = 5 %)

(b) ECTFE nanofibers extracted from melt blown PS/ECTFE_10 (ϕ = 10 %)

(c) ECTFE nanofibers extracted from melt blown PS/ECTFE_25 (ϕ = 25 %)

45

(d) ECTFE nanofibers extracted from melt blown PS/ECTFE_30 (ϕ = 30 %)

(e) ECTFE nanofibers extracted from melt blown PS/ECTFE_40 (ϕ = 40 %)

Figure 2.14 Representative SEM images and the corresponding fiber diameter analysis of ECTFE nanofibers extracted from (a) melt blown PS/ECTFE_5 (ϕ = 5 %) fibers; (b) melt blown PS/ECTFE _10 (ϕ = 10 %) fibers; (c) melt blown PS/ECTFE _25 (ϕ = 25 %) fibers; (d) melt blown PS/ECTFE _30 (ϕ = 30 %) fibers (d) melt blown PS/ECTFE _40 (ϕ = 40 %) fibers.

46

(a) ECTFE nanofibers extracted from melt blown PS/ECTFE_5 (ϕ = 5 %)

(b) ECTFE nanofibers extracted from melt blown PS/ECTFE_10 (ϕ = 10 %)

(c) ECTFE nanofibers extracted from melt blown PS/ECTFE_25 (ϕ = 25 %)

47

(d) ECTFE nanofibers extracted from melt blown PS/ECTFE_30 (ϕ = 30 %)

(e) ECTFE nanofibers extracted from melt blown PS/ECTFE_40 (ϕ = 40 %)

Figure 2.15 Representative SEM images and the corresponding fiber diameter analysis of ECTFE nanofibers extracted from (a) melt blown PS/ECTFE_5 (ϕ = 5 %) fibers; (b) melt blown PS/ECTFE _10 (ϕ = 10 %) fibers; (c) melt blown PS/ECTFE _25 (ϕ = 25 %) fibers; (d) melt blown PS/ECTFE _30 (ϕ = 30 %) fibers (d) melt blown PS/ECTFE _40 (ϕ = 40 %) fibers.

The average fiber diameter (dN,av) of the extracted PBT nanofibers slightly increases from about 100 nm to 150 nm, while the dN,av of ECTFE nanofibers appears to be constant at 90 nm with increasing ϕ, as shown in Figure 2.16 (a). The diameter distribution (CV) of the PBT nanofibers slightly increases from 40 % to 55 %, while the

48

CV of the ECTFE nanofibers plateaus around 45 %. The fiber diameters of the extracted nano-/micro-fibers, such as the extracted ECTFE nanofibers at ϕ = 30 %, were not analyzed because of the great non-uniformity.

Figure 2.16 The effect of blend composition on (a) the average fiber diameter (dN, av) and (b) fiber diameter distribution of nanofibers. The nanofibers were extracted with THF from melt blown microfibers of PS- containing polymer blends.

49

2.4 Discussion

2.4.1 Effects of composition on the blend morphology

The droplet-matrix morphology of the PS-containing blends demonstrates the immiscibility of PS/PBT and PS/ECTFE. The sharp interfaces with gaps/cavities reveal the poor interfacial adhesion. The increase of the droplet size with ϕ is because of the coalescence.99 Since PS/PBT and PS/ECTFE blends experienced the same mixing

-1 process and have similar viscosity ratios during mixing ( ̇ ≈ 130 s at 100 rpm, ηPBT/ηPS ≈

100-102 ηECTFE/ηPS ≈ 0.4), we speculate that the slightly smaller size of PBT droplets than that of the ECTFE droplets in the PS matrix is primarily because of the interfacial tension

103 (Γ12).

Interfacial tension between immiscible polymer blends can be obtained by fitting the dynamic moduli (G’ and G’’) to the Palierne model.104, 105 However, it is challenging to achieve reliable linear viscoelastic data (elastic modulus G’ and loss modulus G’’) in the terminal region (ω ≤ 1 rad/s) for our blends because G’ and G’’ are too small to be measured by the rheometer, resulting in the failure of the Palierne model fitting.

Therefore, we estimated the interfacial tensions using the harmonic mean (equation (2.4)) which has been suggested to calculate interfacial tensions of polymer blends,103, 106

d d p p 441  2  1  2 12  1   2 d d  p p (2.4) 1  2  1  2

d p where Γ12 is the interfacial tension between polymer 1 and polymer 2, and γi and γi the

dispersion and polar components of γi, respectively. As shown in Table 2.3, ΓPS/ECTFE is

50

larger than ΓPS/PBT, contributing to higher extent of coalescence resulting in a bigger size of ECTFE domains than PBT and the bimodal distribution.

Table 2.3 Interfacial tensions calculated from polymer surface tension

γ γ p γ d Γ Γ Polymer Blend i i i PS/PBT PS/ECTFE (mN/m) (mN/m) (mN/m) (mN/m) (mN/m)

PS 40.7a 0.6a 40.1a -- 2.6 PBT 46.0b 42.0b 4.0b 4.6 ECTFE 30.5c 27.0c 3.5c -- a. Reference Biresaw, G.107 b. Reference 108.108 c. Reference Halar ECTFE Design & Processing Guide.109

Phase inversions occur for both PS/PBT and PS/ECTFE blends at ϕ = 40%. There have been several models to predict phase inversions of binary polymer blends, such as the Miles-Zurek,110 Metelkin-Blekht,111 Utracki,112 and Bourry-Favis.113 We employed the Utracki model (equation (2.5)) to predict the phase inversions of our PS-containing blends because the viscosity ratios are far away from unity,114, 115

[]m mI 2    (2.5) mI 1

where λ = viscosity ratio = η1( ̇)/η2( ̇), ϕm = the maximum packing volume fraction =

0.84, [η] ≈ 1.9, and ϕxI = the volume fraction of phase x at the phase inversion point. The shear rate corresponding to 100 rpm in the HAAKE batch mixer is approximately 130 s-1

99, 100, 114 and the viscosities were estimated accordingly. Therefore, ϕPBT,I ≈ 44 % for

51

PS/PBT blends and ϕECTFE,I ≈ 42 % for PS/ECTFE, which agrees well with the experimental observation of ϕ = 40 %.

2.4.2 Effects of composition on the melt blown fiber morphology

The average fiber diameters (dav) of the melt blown PS-containing blends show little dependence on the compositions (ϕ) of the starting blends. We attribute the independence of the dav with ϕ to the relatively constant zero shear viscosity of the starting polymer blends at the melt blowing temperature (η0,blends,265 °C = 50 – 70 Pa • s), which directly

51 impacts the dav of the resulting melt blown fibers.

At low ϕ (ϕ ≤ 30%), we speculate that the stronger phase segregation in PS/ECTFE

3 0.5 3 0.5 3 0.5 116-118 blends (δPS = 9.1 (cal/cm ) , δECTFE = 20.0 (cal/cm ) , δPBT ≈ 11.0 (cal/cm ) ) produce more instabilities, like ribbon-like fibers, during fiber blowing resulting in wider fiber diameter distributions than PBT/PS blends. As ϕ increases, coalescence becomes more significant which probably suppresses the instabilities, leading to more uniform fibers.

2.4.3 Effects of composition on the nanofiber morphology

Uniform nanofibers were obtained after THF-extraction of the melt blown PS/PBT (0

≤ ϕ ≤ 30 %) and PS/ECTFE (0 ≤ ϕ ≤ 25 %) fibers. The existence of spheres and elongated threads at low ϕ (ϕ = 5 %) indicates fiber breakup during the nanofiber formation.

Increasing ϕ tends to suppress the irregularities and improve the nanofiber uniformity likely implies that the nanofibers originate from the continuous deformation of the dispersed droplets and the subsequent merging of the elongated threads (short fibers),

52 where the coalescence plays a key role.

The nanofiber bundles are intriguing. Our previous studies on melt blown

PBT/ECTFE fibers showed that the ECTFE nanofibers dispersed separately inside single

PBT microfibers.87 We hypothesize that the inter-fiber cohesion occurs among the adjacent nanofibers during the solvent extraction, leading to the formation of nanofiber bundles.119 Increasing ϕ results in more densely packed nanofibers and thus more nanofiber bundles.

At ϕ = 40 % when phase inversions occur in both the PS/PBT and PS/ECTFE blends, the extracted nanofibers coexist with microfibers. In order to determine whether there are

PS nanofibers embedded in the extracted single nano-/micro-fibers, we cryo-fractured the extracted fiber mats, but no PS nanofibers were found. This is possibly because the volume fraction of the minor phase (40 %) is not big enough to form the fiber “shells” although it hits the phase inversion point. We hypothesize that the strong nonuniform coalescence of the minor phase during the fiber blowing process contributes to the formation of such nonuniform nano-/micro-fiber networks, which we discuss more in

Session 2.4.4.

In addition, the fiber diameter analysis shows that average diameters (dN,av) of the extracted ECTFE nanofibers are a little bit smaller than those of the extracted PBT nanofibers, which is likely because of the slightly lower viscosity of ECTFE.

2.4.4 Nanofiber formation

The formation of nanofibers from melt blown fiber-in-fiber polymer blends involves

53 three processes: the dispersion of the minor phase in the matrix during compounding which prepares the starting blends, the shear inside a capillary die connected with the melt blowing spinneret, and the subsequent non-isothermal uniaxial extension when the polymer melt exits the melt blowing die. We have discussed the minor phase dispersion in Session 2.4.1. Here we consider the latter two processes which directly tie to the formation of nanofibers-in-microfiber.

Figure 2.17 Schematic of the formation of nanofibers-in-microfiber from melt blown immiscible binary polymer blends.

Firstly, the starting polymer blend is melted and extruded through a capillary die

(Dcapillary die = 3 mm, Lcapillary die ≈ 7 cm) connected with our melt blowing spinneret (Lmelt

54 blowing spinneret ≈ 7 cm), as illustrated in the schematic in Figure 2.17. The melt blowing die tip can be considered as another capillary die (ddie hole = 0.2 mm, Ldie hole ≈ 1 mm). To better understand the effect of this extrusion process on nanofiber formation, we extruded the PS-containing blends (ϕ = 5 % and ϕ = 25 %) using the same apparatus as melt blowing without applying the blowing air jets. Then, the morphology of the dispersed phase in the extrudates was analyzed by both cryo-fracture and THF-extraction.

As shown in Figure 2.18, at low ϕ (ϕ = 5 %), PBT and ECTFE exist primarily as near-spherical droplets in the extrudates, respectively. There are a large number of tiny droplets (D < 1 μm) despite some large droplets (D ~ 10 μm). Also, a small amount of ellipsoids and long threads with sharp ends are also observed. Increasing ϕ to 25 % results in more larger droplets (D ~ 20 μm) and more elongated threads (as long as 1 mm).

(a) PS/PBT_5 (ϕ = 5 %) extrudate

55

(b) PS/PBT_25 (ϕ = 25 %) extrudate

(c) PS/ECTFE_5 (ϕ = 5 %) extrudate

(d) PS/ECTFE_25 (ϕ = 25 %) extrudate

Figure 2.18 Representative SEM images showing the morphology of the minor phase from both the cryo- fractured cross-section and THF-extracted extrudates of (a) PS/PBT_5 (ϕ = 5%) blend; (b) PS/PBT_25 (ϕ = 25%) blend; (c) PS/ECTFE_5 (ϕ = 5%) blend; (d) PS/ECTFE_25 (ϕ = 25%) blend.

56

It has been demonstrated that the elongational flow at the entrance of a capillary die favors the deformation of droplets.120 The drops come together, collide and coalesce into a fiber-like morphology at the entrance region.121 During the extrusion, the dispersed droplets are first elongated at the entrance of the capillary die. Before reaching the melt blowing die hole, those deformed droplets in the center of the shear flow region will recoil due to elasticity (the normal stress differences).121-123 Those near the wall will experience relatively larger shear rate ( ̇ ≈ 5 s-1) compared with those in the center.

Huneault et al suggested that drop deformation and breakup depend on the reduced capillary number (Ca*),124

* Ca Ca  (2.6) Cac

where Ca is the capillary number and Cac the critical capillary number. Ca and Cac are estimated using equation (2.7) and (2.8),125

 R Ca  m (2.7) 

Ca 0.00056 log(c )  0.64853  0.02442(log )  0.02221(log )2  2 log  0.00645 (2.8)

where ηm is the matrix viscosity, ̇ the shear rate, R the droplet radius, Γ the interfacial tension, p the viscosity ratio (λ = ηd/ηm).

Ca* is estimated to be about 0.5 for both PS/PBT and PS/ECTFE blends at ϕ = 5 %

57

(the droplet size in the starting blends Rav,PBT ≈ 0.6 μm and Rav,ECTFE ≈ 1 μm were used) without considering the coalescence. However, the coalescence happens when the droplets enter the capillary die, which increases the droplet size. Huneault et al suggested that droplets deform without breakup when 0.1 < Ca* < 1 , while break up when Ca* >

1.124 Taking the coalescence into consideration, we speculate that the droplets near the wall will deform and some (R > 1.5 – 2 μm) will break up. Raising ϕ to 25 % increases the droplet size because of coalescence, resulting in the increase of Ca* (Ca* > 2) and thus more breakup near the wall.

When the droplets arrive at the entrance of the melt blowing die hole, we speculate that the overall droplet size is increased because of the coalescence inside the capillary

126 tube (Lcapillary die + Lmelt blowing spinneret ≈ 14 cm). These droplets are elongated and coalesce again as they enter the melt blowing die hole. The droplets are intensely stretched and subsequently break up because of the very high shear rate at the hole wall

-1 120, 123 122 ( ̇ ≈ 2880 s , Ca ≫ Cac). But, the breakup is not instantaneous under flow. Van

Puyvelde showed that the total breakup time tT, consisting of the deformation time plus the breakup time, in a simple shear flow scales as,122

 2/3 tT   Ca (2.9)

where ̇ is the shear rate. According to relation (2.9), tT ~ 0.02 s which is comparable with the material residence time in the die hole (tresidence ~ 0.02 s). Moreover, the breakup can be accelerated by the higher elastic modulus of the PS matrix which generates high

58 normal forces.127 So, we speculate that most elongated droplets break up, resulting in a large population of droplets in the THF-extracted extrudates, as shown in Figure 2.18 (a) and (c). In addition, the extrudates were collected about 5 s later when they came out of the die because of the die set-back and extrudate swell, leading to the relaxation of elastic energy which also influences the extrudate morphology.123 More coalescence resulting from the increase in ϕ can increase the tT and thus lead to more elongated (even 1-mm- long) threads.121, 128

Based on the above analysis, we conclude that the hierarchical structure of nanofibers-in-microfiber from melt blown binary immiscible polymer blends are primarily formed during the fiber blowing process. To the best of our knowledge there have been no theoretical studies associated with nanofiber formation inside single microfiber from melt blowing. Therefore, we present our discussion as follows.

Upon exiting the melt blowing die, the melt of the polymer blend is drawn down into fibers by hot high-velocity air jets. Shambaugh et al. reported that most fiber attenuation occurred rapidly due to the aerodynamic drag within the first 5 cm below the die.88, 89

Then, the fiber diameter decreases very slightly because of the fiber drag resulting from fiber-fiber entanglements/contacts until the fiber temperature reaches the solidification

129 temperature (Tg or Tc). Therefore, we postulate that the nanofibers inside single melt blown microfibers are primarily formed within the first 5 cm and the subsequent fiber- fiber drag also contributes until the matrix microfibers solidify.

The molten polymer blend is subjected to strong extensional flow (extensional strain rate ~ 107 s-1) upon exiting the melt blowing die. The matrix polymer is elongated 59 continuously with a concomitant reduction in diameter. In the meanwhile, the dispersed

130, 131 phase deform affinely with the matrix filament (Ca ≫ Cac). The strong extensional flow generates elongated threads/fibrils of the minor phase. Several researchers reported the coalescence mechanism of microfibril formation in binary polymer blends in both shear and extensional flows.132-134 They proposed that two droplets approaching each other and undergoing shear-induced coalescence, thus forming a dumbbell-shaped droplet flowed by a three-droplet dumbbell, etc. We speculate that the single-droplet-elongation and dumbbell-formation also occurs to the elongated threads dispersed in the matrix filament during the fiber blowing, which eventually results in the nanofibers with large aspect ratio (length/diameter ≈ 104 – 105).135 Moreover, the elongated threads can break up into droplets through both end-pinch and capillary instability,120, 136 as evidence by the representative SEM image in Figure 2.13 (a).

The droplet size and its size distribution when the blend melt just comes out of the melt blowing die before extension is important to the subsequent nanofiber formation and the final nanofiber morphology. The coalescence when the droplets flow through the capillary die and the melt blowing die increases the non-uniformity of the blend.

Increasing ϕ can further enhance this non-uniformity. We hypothesize that the blend may show a bimodal distribution (as is seen in Figure 2.9) or even a side-by-side morphology, as shown in the schematic in Figure 2.19. During the subsequent extension, the small domains are drawn and coalescence into nanofibers and those large domains potentially form large fibers (d ~ 1 – 2 μm), which generates the non-uniform nano-/micro-fibers after solvent extraction.

60

Figure 2.19 Schematic of the hypothesized blend morphology before blowing extension. (a) bimodal distribution, (b) side-by-side.

Interfacial slip between the deformed droplets and the matrix fiber can influence the droplet deformation. Levitt and Macosko reported that slip can reduce the deformation of a drop in shear-type flow.137 Ramachandran et al found that the shape of the deformed drop for the same drop elongation is relatively insensitive to the slip coefficient as defined by equation (2.10),138

dI d   (2.10) I R

where dI is the thickness of the diffuse interface between the liquids, ηI the viscosity of the interfacial region, ηd the viscosity of the dispersed liquid, R the drop radius. But slip

138 slows down the deformation process (Ca ≫ Cac). Also, these authors claimed that slip causes capillary instability to produce larger drops at slightly faster rates relative to the no-slip case. We postulate that interfacial slip may slow down the thread elongation but has little effect on the final nanofiber aspect ratio. It is likely that interfacial slip

61 facilitates the nanofiber breakup, but there is no evidence to support this hypothesis.

The droplet deformation has been demonstrated to be proportional to the droplet size in the case of steady shear flow and Newtonian phases, as described by equation

(2.11).139

 3 19pp 16 3R 19 16 D  Ca   m  (2.11) 2 16pp 16 2 16 16

where D is droplet deformation, ηm the matrix viscosity, ̇ the shear rate, p the viscosity ratio (ηd / ηm), Γ the interfacial tension, R the droplet radius. However, applying this equation to the droplet deformation in the highly non-linear extensional flow of fiber blowing process is inappropriate because of the viscoelastic effects and non-isothermal condition. This has also been revealed by the experimental results that the average nanofiber diameter appears to be insensitive to the droplet size in the starting polymer blends.

The extensional viscosities of the blend components are also important to the formation of nanofibers. However, it has been challenging to measure the extensional viscosity under the high extensional rate (~ 107 s-1) of melt blowing process. Both

Trouton’s law and the Cogswell model have been used to estimate extensional viscosities. Trouton’s law assumes a very low extension rate which is not applicable to melt blowing.140, 141 By measuring the pressure drop at the entrance of a capillary rheometer the Cogswell model can predict the extensional behaviors of polymers.142 But measuring the required pressure drop of both PBT and ECTFE is beyond the capability of

62 our capillary pressure transducer. Here we give general considerations associated with extensional viscosities. PBT and ECTFE would exhibit strain hardening due to crystallization, inducing higher extensional viscosities at high extension rates. Taking these into considerations we speculate that the nanofiber formation could be interrupted at those high extension rates along the fiber blowing line. However, the non-isothermal condition results in higher temperatures in the core but lower temperatures near the surface of a single melt blown fiber, which makes the local deformation associated with extensional viscosities more complicated.

Finally, we postulate that the fiber drag due to fiber-fiber entanglement/contact also influence the nanofiber morphology. Although the melt blown fiber diameter decreases sharply within the first 5 cm below the die, the fiber temperature can be higher than the

89 solidification temperature (Tg or Tc) beyond the 5 cm. Fakirov et al studied that the microfibril formation in polymer blends during cold drawing.95 They drew a PP/PET

(60/40 wt%) blend at 80 °C (slightly higher than the Tg of PET) and produced long (even more than 1 mm) PET nano-/micro-fibrils because of coalescence. We hypothesize that the fiber drag among the PS matrix fibers can further deform the nanofibers inside and cause coalescence until the core temperature of the PS matrix fibers cools down below the solidification temperature (i.e. Tg,PS = 100 °C).

63

2.5 Conclusion

We have advanced the understanding on nanofiber fabrication from melt blown fiber- in-fiber polymer blends by investigating the nanofiber formation inside single melt blown microfibers.

The formation of nanofibers from melt blown fiber-in-fiber polymer blends was investigated using PS/PBT and PS/ECTFE blends. The minor phase, either PBT or

ECTFE, was blended with PBT at controlled compositions of 5 vol%, 10 vol%, 25 vol%,

30 vol%, and 40 vol%. The blends were melt blown followed by THF-extraction of the

PS matrix. PBT and ECTFE nanofibers (dN,av ≈ 100 – 150 nm) with large aspect ratio

(length/diameter ~ 104 – 105) were produced, respectively. The average nanofiber diameter appears to be independent with the minor phase volume fraction.

The controlled extrusions demonstrated that the core nanofibers are primarily formed during fiber blowing process. Both the aerodynamic and fiber-fiber drag of the matrix microfibers influence the nanofiber formation. The single-droplet-elongation and dumbbell-formation contribute to the formation of nanofibers. Coalescence plays a key role in nanofiber formation. We hypothesize that the non-uniform coalescence can lead to a bimodal or a side-by-side blend morphology before fiber blowing. These nonuniform morphologies finally result in nano-/micro-fibers after solvent extraction.

64

Chapter 3 Nanofibers from Water-extractable Melt Blown Fiber- in-Fiber Polymer Blends

3.1 Introduction

Nonwoven nanofibers have found uses in filtration,5 performance apparel,78, 79 biological tissue engineering,80-82 medical textiles,83, 84 and composites materials.85, 86

Several techniques can be applied to fabricate nonwoven nanofibers, such as electrospinning,143 melt spinning,144 forcespinning,145 and melt blowing.27 Among these techniques, electrospinning, being able to produce fibers with an average diameter of less than 100 nm, has attracted the most attention. According to the database of Web of

Science, there are nearly 1000 research articles associated to electrospun nanofibers in

2014. However, this method is costly and has a relatively low productivity.12 In addition, issues associated with solvents have increased the concerns.146

We have reported a new approach to fabricate nanofibers from melt blown fiber-in- fiber polymer blends.87 By melt blowing immiscible polymer blends we can produce a hierarchical structure of nanofibers-in-microfiber. Then, extracting the matrix polymer with an appropriate solvent can expose the nanofibers. For example, we can produce a poly(ethylene-co-chlorotrifluoroethylene) (ECTFE) nanofibers with an average diameter of 70 nm after washing away the poly(butylene terephthalate) (PBT) matrix using trifluoroacetic acid (TFA). This new method can be applied potentially for mass

65 production of nanofibers. However, the use of selective organic solvents such as TFA leads to issues of solvent recycling and incremental manufacturing cost, challenging this technique to offer an optimal solution on cost, productivity, and environment at the same time.

In this regard, application of water-extractable polymers, either water-soluble or water-dispersible, as a sacrificial matrix of melt blown fiber-in-fiber can eliminate issues associated with organic solvents. Poly(vinyl alcohol) (PVA) is one of the most common commercial water-soluble polymer. An Tran et al. recently reported conjugate melt spinning of poly(lactic acid) (PLA) and poly(vinyl alcohol) (PVA) blends, using PVA as the “sea” and PLA the “islands”.147 After dissolving the PVA using water, they obtained

PLA nanofibrils with an average diameter of 60 nm. In order to minimize the thermal degradation of PVA, they controlled the process at a temperature lower than 200 ºC, which increased the viscosity. But, it is difficult to process PVA at a temperature lower than 200 ºC while accessing a suitable viscosity for melt blowing. Although various studies such as plasticization148-152 and blending153, 154 have been developed to stabilize

PVA for thermal processing, it is still challenging to perform melt blowing at a temperature where an optimal viscosity is attained. Moreover, these approaches are complicated for industrialization. However, Nishio et al patented generating nanofibers (5

– 500 nm but no fiber image) from melt blown PVA-containing polymer blends at 230

°C.155 The PVA they used had a polymerization degree of 400 and saponification value of

62 %. Interestingly, they did not discuss thermal degradation of PVA. Other commercial water-soluble polymers such as poly(ethylene oxide), poly(acrylic acid),

66 poly(methacrylic acid), and poly(vinyl pyrrolidone) encounter the same trade-offs between thermal degradation and optimal viscosity for melt blowing.156-159

In this work, we demonstrate nanofiber fabrication from melt blown fiber-in-fiber polymer blends containing a commercial water-dispersible sulfopolyester (SP) (Eastman

Chemical Company, US). This new method enables us to use water to expose nanofibers after melt blowing. It not only eliminates issues associated with organic solvents but also provides another approach to prepare multilayer nano-/micro-fiber composites.

3.2 Experimental

3.2.1 Materials

PBT (Ticona Celanex 2008) and two grades of sulfopolyester (SP1, SP2) (Eastman

Chemical Company, US) were used in this study. The chemical structure of SP is given in Figure 3.1. SP1 and SP2 consist of the same monomers (isophthalatic acid, diethylene glycol, 5-sodiosulfoisophthalic acid, 1,4-cyclohexanedimethanol) but in different ratios.160-162 The thermal properties of the materials were characterized by differential scanning calorimetry (DSC) (TA Instrument Q1000). The samples were analyzed under a standard run sequence of heating-cooling-heating from -20 °C to 250 °C at a scan rate of

10 °C/min, and Tg and Tm were analyzed in the second heating run. The DSC results are summarized in Table 3.1.

67

Figure 3.1 The chemical structure of the sulfopolyester (SP).

Table 3.1 Molecular weight and thermal properties of PBT, SP1, and SP2.

M T T η at 265 °C Materials n Ð g m 0 (kg/mol) (°C) (°C) (Pa • s) PBTa -- -- 43 223 60 SP1b ~ 10 ~ 2.6 – 3 45 -- 350 SP2b ~ 10 ~ 2.6 – 3 35 -- 150 a. The molecular weight of PBT could not be measured because of their poor solubility in THF which is employed for SEC measurement. The melt blow index (MFI) of PBT is 250 at 250 °C. b. The molecular weight of SP1 and SP2 were obtained from Eastman Chemical Company.

Rheological properties were measured with a strain-controlled ARES rheometer (TA

Instrument) equipped with a 25 mm parallel-plate fixture. Firstly, dynamic strain sweeps were performed to determine the linear viscoelastic regions of the materials. Secondly, dynamic frequency sweeps were employed to measure the complex viscosity (η*) and dynamic elastic modulus (G’) and loss modulus (G”) as a function of frequency (1 ≤ ω ≤

100 rad/s) (3 % strain) at 265 °C, as shown in Figure 3.2.

68

Figure 3.2 Complex viscosity (η*), elastic modulus (G’), and viscosity modulus (G’’) of PBT, SP1, SP2, measured by dynamic frequency sweep at 265 °C with 25 mm parallel plates.

69

3.2.2 Preparation of polymer blends

PBT pellets were dried overnight (about 6-9 hours) under vacuum at 100 °C. SP1 and

SP2 pellets were dried at about 40 °C under vacuum for 24 hours to remove the moisture.

Then, the polymer pellets were weighed and mixed at room temperature with compositions that are summarized in Table 3.2. The mixed pellets were fed into a

HAAKE (Thermo Scientific) batch compounder at 265 °C with roller blades rotating at

100 rpm. The blending was maintained for 8 minutes before the resulting blend was quenched with liquid nitrogen (LN2) to freeze the blend morphology.

Table 3.2 Compositions of polymer blends in this study

ϕ PBT SP1 SP2 Sample (wt%) (wt% / vol%) (wt% / vol%) (wt% / vol%) SP1/PBT_30 30 30 / 18 70 / 82 -- SP2/PBT_30 30 30 / 20 -- 70 / 80 SP1/PBT_5 5 5 / 3 95 / 97 -- SP1/PBT_10 10 10 / 5 90 / 95 -- SP1/PBT_40 40 40 / 26 60 / 74 --

3.2.3 Characterization of blend morphology

Polymer blends were cryo-fractured in LN2. The resulting sample surfaces were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. A scanning electron microscope (SEM) (Hitachi S-4700 or JEOL 6500) was used to image the sample surface. 20 – 25 SEM micrographs were taken for each sample.

Approximately 300 droplets of the minor phase were measured with the ImageJ software.

Then, the measurement results were fitted with a log-normal distribution function using 70 the software package of OriginLab to obtain the number average droplet size (Dav).

3.2.4 Melt blowing and water-extraction

A lab-scale melt blowing apparatus (see Chapter 2) was used in this study. The materials were dried under vacuum overnight (10 – 12 hours) at 40 °C. Melt blowing experiments were performed at Qpolymer ≈ 0.2 g/(min • hole), Pinlet = 6 psi, and Tprocessing =

265 °C. Melt blown fiber mats were collected with our continuous fiber collector

(rotation speed ≈ 0.5 rpm) located about 35 cm away from the melt blowing die tip.

In order to expose the nanofibers, samples of melt blown fibers were cut from the collected continuous fiber mats with a scissor or a rotary cutter (Fiskars, 45 mm- diameter). The samples were soaked in distilled water for 3-5 times without stirring at 60

°C. Each soaking used refresh distilled water and lasted for 2 hours. The extracted materials were dried in the air for 24 hours. The materials were weighed both before and after soaking.

3.2.5 Fiber characterization

Both melt blown fibers and water-extracted nanofibers were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. 25-30 SEM micrographs were taken for each sample with a scanning electron microscope (SEM)

(Hitachi S-4700 or JEOL 6500). The diameters of 400 - 500 fibers were measured using the ImageJ software. Then, the OriginLab was employed to fit the measurement results with a log-normal distribution function. The geometric average fiber diameter (dav) was extracted from the log-normal fitting. The geometric standard deviation (GSD) and

71 coefficient of variation (CV) were calculated based on the log-normal fitting results (see

Chapter 2).

3.3 Results & Discussion

3.3.1 Polymer blends

(a) SP1/PBT_30 (ϕ = 30 %)

(b) SP2/PBT_30 (ϕ = 30 %)

Figure 3.3 Representative SEM images showing the morphology and the corresponding droplet distribution of (a) SP1/PBT_30 (ϕ = 30 %) blend, (b) SP2/PBT_30 (ϕ = 30 %) blend. The samples were cryo-fractured in LN2.

In this study, we started with SP1/PBT_30 and SP2/PBT_30 polymer blends and the

72 corresponding melt blown fibers to determine the optimum material for melt blown fibers, because SP1 and SP2 have different Tgs which can influence the performance of the resulting melt blown fibers.27, 163 As shown in Figure 3.3, PBT appears as approximately spherical droplets dispersed in the SP1 and SP2 matrix, respectively, although the droplets are fractured. The droplet-matrix morphology indicates that PBT forms immiscible polymer blends with SP1 and SP2, respectively. The number average diameter of PBT droplets in SP1 is slightly larger than that in SP2 (1.55 μm vs. 1.38 μm), which we attribute to the smaller viscosity ratio of PBT/SP1 compared to PBT/SP2 (ηPBT/

103 ηSP1 ≈ 0.16, ηPBT/ ηSP2 ≈ 0.5) assuming they have comparable interfacial tension.

3.3.2 Melt blown fibers

We first melt blew SP1/PBT_30 and SP2/PBT_30 blends, respectively. Both of the blends produce uniform melt blown fibers, as shown in Figure 3.4. The average fiber diameter is about 2 – 3 μm which is typical for melt blowing processing. However, inter- fiber cohesion occurs to the melt blown fibers of SP2/PBT_30 after 2-week storage in the lab, because the Tg of SP2 is so close to the room temperature (~ 28 °C). The overall fiber mat loses the fibrous structure because of the inter-fiber cohesion. In contrast, there is no cohesion issue associated with the melt blown SP1/PBT_30 fibers due to its higher Tg.

Therefore, SP1 is selected for further study on water-extractable nanofibers, as discussed in the following Session 3.3.3.

73

(a) SP1/PBT_30 (ϕ = 30 %)

(b) SP2/PBT_30 (ϕ = 30 %)

Figure 3.4 The morphology and the corresponding diameter distribution of melt blown fibers of (a) SP1/PBT_30, (b) SP2/PBT_30.

In order to better understand the effects of the PBT weight fraction on the nanofiber morphology, in addition to melt blown fibers of SP1/PBT_30, we prepared three more sets of melt blown SP1/PBT fibers which contain 5 wt%, 10 wt%, and 40 wt% of PBT,

74 respectively. As shown in Figure 3.5, melt blown microfibers (dav ~ 1.5 – 3 μm) were obtained. Most fibers are cylindrical, but ribbon-like and fused fibers are also observed, which broadens the fiber diameter distribution. We postulate that the ribbon-like fibers are formed due to the instability during fiber blowing and fused fibers because of inter- fiber diffusion before the fibers solidify. Similar phenomena have been found in melt blown poly(propylene) fibers.42

(a) SP1/PBT_5 (ϕ = 5 %)

(b) SP1/PBT_10 (ϕ = 10 %)

75

(c) SP1/PBT_40 (ϕ = 40 %)

Figure 3.5 The morphology and the corresponding diameter distribution of melt blown fibers of (a) SP1/PBT_5, (b) SP1/PBT_10, (c) SP1/PBT_40.

3.3.3 Water-extracted nanofibers

Washing away the SP1 phase simply with distilled water at 60 °C exposes the PBT nanofibers, as shown in Figure 3.6. The average diameter of the PBT nanofibers is in the range of 70 – 100 nm, being insensitive to the PBT concentration which is consistent with our previous studies (see Chapter 2). At ϕ = 5 %, droplets, ellipsoids, and wave-like fibers are found, which are primarily because of the nanofiber breakup from both end- pinch and capillary stability.120, 136 Increasing ϕ enhances the coalescence of the elongated threads during the nanofiber formation and tends to suppress the breakup, as indicated by Figure 3.6. On the other hand, nanofiber bundles increase with the increase of ϕ, which has also been observed in our previous studies (see Chapter 2). We

76 hypothesize that the formation of nanofiber bundles is because of the cohesion between adjacent nanofibers during the solvent extraction. More discussions about the effects of the minor phase concentration on the nanofiber formation are presented in Chapter 2.

(a) PBT nanofibers extracted from SP1/PBT_5 (ϕ = 5 %)

(b) PBT nanofibers extracted from SP1/PBT_10 (ϕ = 10 %)

77

(c) PBT nanofibers extracted from SP1/PBT_30 (ϕ = 30 %)

(d) PBT nanofibers extracted from SP1/PBT_40 (ϕ = 40 %)

Figure 3.6 Morphology and the corresponding diameter distribution of water-extracted PBT nanofibers from melt blown (a) SP1/PBT_5, (b) SP1/PBT_10, (c) SP1/PBT_30, (d) SP1/PBT_40.

3.3.4 Multi-layer PBT nano-/micro-fiber composites

Extracting nanofibers from water not only solves issues associated with organic solvents but also provides another way to prepare multi-layer nano-/micro-fiber composites. Here we demonstrate the preparation of a double-layer PBT nano-/micro- fiber composite. Upon melt blowing the SP1/PBT_30 blend and depositing the fiber mat on top of a pre-made continuous PBT fiber mat we obtained a double-layer fiber mat consisting of a top layer of melt blown SP1/PBT_30 fibers and a bottom layer of melt 78 blown PBT fibers. Then, the resulting double-layer mat was soaked in distilled water to remove the SP1 component, resulting in a double-layer PBT nano-/micro-fiber composite, as shown in Figure 3.7. The thickness of both the nanofiber layer and the microfiber is expected to be controllable through tuning the blend composition and melt blowing process. Such materials are potentially useful for filtration media.

Figure 3.7 A representative SEM image showing the morphology of double-layer PBT nano-/micro-fiber composite. The PBT nanofibers (dN,av = 82 nm, CV = 36 %) were extracted from melt blown SP1/PBT_30 (ϕ = 30 %) deposited on top of continuous melt blown PBT fiber mat.

3.4 Conclusion

We have demonstrated achieving nanofibers as small as 70 nm from water- extractable melt blown fiber-in-fiber polymer blends. A commercial water-dispersible

79 sulfopolyester (SP1) serving as the sacrificial phase was blended with PBT at compositions of 95/5 wt%, 90/10 wt%, 70/30 wt%, and 60/40 wt%. Melt blowing these immiscible binary blends produced microfibers (dav ~ 1.5 – 3 μm). Washing away the

SP1 simply with distilled water expose the PBT nanofibers. A brief discussion on the nanofiber morphology was presented. Comparing with our previous studies using selective organic solvents like THF to obtain nanofibers, this new method not only eliminates issues associated with organic solvents but also provides another route to prepare multilayer nano-/micro-fiber composites.

80

Chapter 4 Tuning Surface Properties of Poly(butylene terephthalate) Melt Blown Fibers by Alkaline Hydrolysis and Fluorination

4.1 Introduction

Melt blown fibers, cost-effective engineering materials, have been used for filtration media5, particularly for water-in-diesel engine fuel filtration. Modern diesel engines, driven by increasingly stringent emission regulations, require an even greater levels of fuel cleanliness, including less particles, water, biological materials, wax crystals and asphaltines.57 Particles cause wear to engine parts and block fuel injector nozzles. Water can serve as a medium for micro-organisms and thus promote biological growth which can plug engine components.56, 57 Water can reduce the lubricity of fuel and thus increase abrasive wear. In addition, water corrodes fuel tanks and generates particles that plug filters and orifices and causes failure of fuel injection equipment.56 Both wax crystals and asphaltines can plug a filter, shortening the filter’s life.

In order to meet the higher requirements on fuel cleanliness more efficient filters must be developed. Among the various solutions to enhance filter efficiency, tuning surface wetting properties of the filter media, achieving either superhydrophilicity or

 Reproduced in parts with permission from Z. Wang, C. W. Macosko, F. S. Bates. ACS Appl. Mater. Interfaces 2014, 6, 11640-11648.  2014 American Chemical Society

81 superhydrophobicity, has been demonstrated to be an effective way to optimize the filter performance, especially in water filtration. Hydrophilic or superhydrophilic fibers are used as a depth porous medium to cause coalescence of water drops. The massive drops coalesce as they move through the filter with the fuel and fall out of the fuel flow on the downstream side of the filter due to their enlarged size and higher density compared to fuel. Another strategy is to apply hydrophobic or superhydrophobic fibers which prevent water drops from entering the filter and thus separate them on the upstream side of the media.

The current dominating approach in industrial settings is to render nonwoven fibers superhydrophobic or superhydrophilic is by wet coating. By coating a cellulose based media with a silicone based material via adsorption process Donaldson enhanced the filter performance in removing harmful water and coarse particulate contaminant.164

However, it has been reported that wet coating is nonuniform and can leach into the surroundings. More importantly, it can affect the filter pore size.74 Although other techniques such as physical vapor deposition,165, 166 chemical vapor deposition,167 addition of low-surface-energy additives through blending,168 copolymerization,169, 170 surface grafting,171, 172 layer-by-layer deposition,173 sol-gel technique,174, 175 and plasma treatment176-181 have been studied, it is challenging to industrialize these methods because of economic and public health concerns. Another approach to modify surface wetting properties of materials is alkaline hydrolysis, which has been studied with poly(ethylene terephthalate) (PET) fabrics.182-189 However, to the best of our knowledge this method has not been applied to poly(butylene terephthalate) (PBT) melt blown fibers

82 which is a major material for diesel fuel filter media, possibly because PBT woven fabrics are more resistant to aqueous sodium hydroxide solutions than PET woven fabrics.184

Here we studied surface modifications of PBT melt blown fibers by hydrolysis using sodium hydroxide (NaOH) and subsequent fluorination. PBT melt blown fibers were fabricated with our lab-scale melt blowing apparatus and subsequently soaked in NaOH solutions. After hydrolysis, a simple fluorination reaction was conducted. The effects of

NaOH hydrolysis on fiber surface morphology, average fiber diameter, mass loss, and structural integrity of the fiber mats were evaluated. Sessile drop measurements revealed that superhydrophilicity was achieved by hydrolysis and sticky superhydrophobicity was obtained by subsequent fluorination. Moreover, the influence of NaOH hydrolysis duration on the rate of water droplet absorption by superhydrophilic h-PBT fiber mats was studied experimentally using a high-speed camera.

4.2 Experiment

4.2.1 Fabrication of melt blown PBT fibers

PBT pellets (Celanex 2008, Ticona) were dried at 100 °C for 12 hours under vacuum and then melt blown at 265 °C using a previously described lab-scale apparatus.27 The melt blowing die containing five 0.2 mm diameter holes is a modified version of a commercial design.97 A stainless steel screen fitted with a blower, located 35 cm away from the melt blowing die, was used to collect the blown fibers which were generated at a polymer flow rate of 0.18 g/(min·hole) and an air volumetric flow rate of 4.5 SCFM.

83

The air pressure at die exit was kept at about 6 psi.

4.2.2 NaOH hydrolysis of PBT fibers

Melt blown PBT fiber mats were cut into small square pieces (1 x 1 cm2 and approximately 1 mm thick) and soaked in a methanol solution containing NaOH in a glass vial without stirring. We used methanol as the solvent because alcohols tend to accelerate hydrolysis of polyester.184 NaOH pellets (0.5 g) (Macron Fine Chemicals) were mixed with 2 mL of methanol (Fisher Scientific) and 2 mL deionized water (pH =

7) and preheated to 45, 50, and 55 °C before soaking the PBT fiber mats for 10, 20, and

30 min. After hydrolysis, the fiber mats were washed three times (one-hour soaking each time) with distilled water and then immersed in HCl solution (0.1 mol/L), then washed again with distilled water until neutrality (pH = 7) of the rinse was achieved. Finally, the hydrolyzed PBT (h-PBT) fiber mats were dried in air for 24 hours, followed by vacuum drying overnight at 20 °C. All experiments were conducted three times.

4.2.3 Fluorination of h-PBT fibers

A piece of h-PBT fiber mat (hydrolyzed at 45 °C for 20 min) was soaked in dimethyl sulfoxide (DMSO) (Sigma-Aldrich) at 80 °C for 24 hours. 30 mg 1H, 1H- perfluorooctylamine (PFOA) (Fisher Scientific) and 15 mg of 4-(4,6-Dimethoxy-1,3,5- triazin-2-yl)-4-methylmorpholinium chloride (DMTMM) (Sigma-Aldrich) were dissolved in a mixture of DMSO (6 mL) and methanol (2 mL) with vigorous stirring. The solution was heated to 50 °C and the h-PBT fiber mat was immersed into the solution for

2 hours. The fluorinated product, denoted f-PBT, was washed repeatedly with methanol

84 and then distilled water and dried under vacuum for 24 hours.

4.2.4 Characterization of fibers and fiber mats

The square fiber mats were photographed and the mass was determined using an electric analytical balance (Denver Instrument M–220). Before each measurement, the fiber mats were vacuum dried overnight at 20 °C. The mass loss (Δm %) was calculated using equation (4.1),

mm0  m%   100% (4.1) m0

where m0 is the mass of the unhydrolyzed fiber mat, m is the mass of the h-PBT fiber mat. The porosity (ε) and the fiber volume fraction (εf) were estimated from equation

(4.2) and (4.3),190

w  (1Az ) (4.2)  fibers

f 1 (4.3)

where ε is the porosity of the fiber mat, εf is the fiber volume fraction, w is the weight of the fiber mat, A and z are the area and thickness of the mat, and ρfibers is assumed to be the density of bulk PBT (1.4 g/cm3).

PBT and h-PBT fiber mats, dried overnight under vacuum, were analyzed by

85 differential scanning calorimetry (TA Instrument Q1000). Heating and cooling scans were controlled from 0 to 250 °C at a constant rate of 10 °C /min. Crystallinity (Xc) was estimated as,

H X  f c 0 (4.4) H f

where ΔHf represents the heat of fusion of the melting peak, ΔHf° = 140 J/g represents

100% crystalline PBT.191

Fiber morphology was determined by a scanning electron microscopy (SEM) (Hitachi

S-4700) applied to fiber mats that were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. For each fiber mat, 25 – 35 SEM micrographs were taken and 400 – 500 fiber diameter measurements were made using ImageJ software.

Origin Lab software was employed to fit a log-normal function to the distribution of fiber diameters from which the average fiber diameter (dav) and the coefficient of variation

(CV) was determined.

X-ray photoelectron spectroscopy (XPS) (Surface Science SSX-100) was employed to determine the surface chemical composition of the h-PBT and f-PBT fiber mats. A monochromatic Al Kα source with a spot size of 1 mm was applied at a take-off angle of

35°, while the pressure of the analysis chamber was maintained at 10-10 Torr. Survey spectra (6 scans/sample, 0-1100 eV binding energy) were recorded at a rate of 1 eV/step and the data were processed using Hawk Data Analysis 7 software.

The wetting properties of the fiber mats were evaluated by sessile drop

86 measurements using a FAMAS Interface Measurement & Analysis system (Kyowa, DM

– CE1). For PBT and the h-PBT fiber mats, the static contact angle (CA) was determined by placing a 5 µL water droplet onto the surface of a mat, and images were recorded with a CCD camera. For the f-PBT fiber mats, a 7 µL water droplet was used for both static and slide-off angle measurements, and images were captured photographically. Five measurements were taken with each sample.

Figure 4.1 Schematic of the experimental set-up employed to record the water droplet absorption by superhydrophilic h-PBT fiber mats (prepared from Cummins melt blown PBT mats shown in Appendix).

Finally, the influence of the time of hydrolysis (duration) on the rate of water droplet absorption by h-PBT fiber mats (prepared from Cummins melt blown PBT mats, as shown in the Appendix) were studied using a high-speed camera. Distilled water droplet absorption on the fiber mats were recorded with a 1000 fps camera (Photron FASTCAM-

87 ultima APX 120k) using a Nikon Micro-nikkor 105 mm lens and Kenko Extension tubes.

The schematic of the experimental setup is shown in Figure 4.1. In general, the fiber mats, glued onto glass slides using double sided carbon tape, were placed on a horizontal stage. A glass syringe with a 22G needle (Kyowa Interface Science Co., Ltd) was fixed by a dispensing system and was suspended above the specimen. Water droplets with a controlled volume of about 10 ± 1 µL were dispensed manually. The droplet falling, impact, and other associated phenomena were captured by the high-speed camera. 3 – 5 droplet measurements were performed on each sample. Each group, experiencing the same hydrolysis duration, contained 3 samples.

4.3 Results and Discussion

4.3.1 Size and integrity of the fiber mats

The size and integrity of the PBT fiber mats were relatively insensitive to the conditions of NaOH treatment, i.e. the temperature and duration of the reaction. As shown by the representative photographs in Figure 4.2, the dimensions of the h-PBT fiber mat treated at 45 °C for 20 min are nearly identical to those of the original material. This behavior differs from the situation when PET fabrics are treated with aqueous solutions of NaOH. PET fabrics shrink during NaOH treatment, which has been shown to depend on the crystallinity of the fibers, the duration and the temperature of the treatment, the fiber diameter, and the fabric structure.183, 189 DSC measurements shown in Figure 4.3 (a) demonstrate a relatively high degree of crystallinity in the untreated PBT fibers (Xc =

88

31%), and this value did not change significantly with hydrolysis. In contrast, Hadjizadeh

183 et al. reported that recrystallization of melt blown PET fibers (Xc = 15%) during hydrolysis leads to noticeable shrinkage. We attribute the relative stability of the h-PBT fiber mats to several factors, including the higher crystallinity due to fast crystallization,192 the mild hydrolysis conditions (i.e. the relatively low NaOH concentration, modest temperature, and short reaction times), and the associated localization of the hydrolysis reaction at the fiber surfaces resulting in the maintenance of the fiber dimensional integrity (see Figure 4.2). Subsequent fluorination did not affect this property, as shown in Figure 4.2 (c).

Figure 4.2 Fiber mat of (a) untreated PBT (b) h-PBT treated at 45 °C for 20 min (c) f-PBT.

In addition, thermal analysis indicated that hydrolysis had only a minor impact on the fiber crystallinity, as shown in Figure 4.3 (b). Within the experimental uncertainty Xc increased slightly (ca. 2 – 4 %). In contrast, single poly(lactide) (PLA) or poly(glycolic acid) (PGA) fibers are reported to display a greater increase in crystallinity with hydrolysis.193, 194 89

Figure 4.3 Thermal analysis of (a) PBT and representative h-PBT fibers (b) hydrolysis effect on crystallinity (Xc).

90

4.3.2 Mass loss and fiber diameter

Reacting the melt blown PBT fiber mat with NaOH results in the loss of mass. The most plausible mechanism of PBT alkaline hydrolysis is shown in Figure 4.4, where hydroxide ions (OH-) attack carbonyl carbons of the ester linkage, breaking PBT chains and leading to the formation of a carbonate salt of sodium, 1,4-butanediol, and fragmented polymer chains with carboxyl or hydroxyl end groups.195

Figure 4.4 Mechanism of PBT alkaline hydrolysis.

As shown in Figure 4.5, the mass loss (for 0 ≤ Δm ≤ 60%) of the h-PBT fiber mats increases linearly with increasing treatment time at a constant temperature, similar to what has been reported during NaOH hydrolysis of PET-cotton spun fabric and PET woven fabrics,184, 185, 196 but contrary to the non-linear loss of mass reported for PET melt blown fiber mats and single PLA fibers.183, 193 Alkaline hydrolysis of polyester fabrics and single fibers has been shown to be a surface reaction, where the fiber weight loss increases nonlinearly with respect to the etching time, depending on several factors, including the concentration of OH-, the fiber surface area relative to the fiber diameter, and the fabric structure.182, 193, 197 We attribute the linear dependence of the mass loss with time (0 ≤ t ≤ 30 min) to the relatively constant OH- concentration associated with the excess NaOH. A more comprehensive understanding of these kinetic issues will be investigated in the future. 91

Also shown in Figure 4.5, the rate of the mass loss increases with increasing reaction temperature, which we attribute to the underlying Arrhenius kinetics.193 As the hydrolysis proceeds, the fiber volume fraction (εf) decreases from the initial 5.9% for the untreated

PBT fiber mat to 2.6% for the h-PBT fiber mats, as shown in Table 4.1. This decrease in the fiber volume fraction is due to the mass loss resulting from the fiber surface corrosion as evidenced by no change in the fiber mat size and overall integrity.

Figure 4.5 Mass loss (Δm %) of h-PBT fiber mats.

92

Table 4.1 Effects of NaOH hydrolysis temperature (T) and duration (t) on porosity (ε), fiber volume fraction (εf), average fiber diameter (dav), and coefficient of variation (CV)

T (°C) t (min) ε (%) εf (%) dav (nm) CV (%)

-- -- 94.1 5.9 865 47 10 94.7 5.3 614 59 45 20 95.4 4.6 565 78 30 96.1 3.9 380 69 10 95 5 820 60 50 20 95.9 4.1 746 86 30 96.4 3.6 406 72 10 95.5 4.5 726 67 55 20 96.3 3.7 690 90 30 97.4 2.6 451 82

Surface hydrolysis also leads to a decrease in the fiber diameters and an increase in the distribution of fiber diameters (i.e. the coefficient of variation, CV). As shown in

Figure 4.6 and Figure 4.7, after 30 minutes the average fiber diameter (dav) has been approximately cut in half with considerably broadening of CV. This behavior supports our hypothesis that NaOH hydrolysis is a surface reaction that proceeds at a constant rate independent of the fiber diameter in the specified reaction periods (0 ≤ t ≤ 30 min). As a consequence, the fractional loss in fiber mass is greater for the smallest fibers, which explains the increase in CV with overall reaction corrosion, as shown in Figure 6 (b).

Hence, if the reaction was extended to longer time, the smallest fibers would be completely consumed, eventually degrading the integrity of the entire fiber mat.

93

Figure 4.6 The impact of NaOH hydrolysis on (a) dav and (b) CV.

94

Figure 4.7 Representative distributions of fiber diameters and associated log-normal function for (a) untreated PBT fibers (b) h-PBT fibers treated at 45 °C for 10 min (c) h-PBT treated at 45 °C for 20 min (d) h-PBT treated at 45 °C for 30 min.

Preliminary tensile experiments (see Appendix) on the etched fiber mats demonstrate that the mechanical properties are not significantly affected after 10 % mass loss (i.e. less than 30 % reduction in the tensile strength ζT and strain at break εb), which is beyond the point of full surface wetting modification discussed in the following section.

95

4.3.3 Fiber surface morphology

Hydrolysis influences the topology of the melt blown PBT fiber surfaces, as shown in the SEM images in Figure 4.8. The relatively smooth surfaces of the untreated PBT fibers are etched by the reaction resulting in a textured, sponge-like interface. This process resembles the corrosion of a metal surface, where the reaction rate depends on the crystal grain orientation.198, 199 We believe that the amorphous and low-crystallinity regions react more readily with OH- than the less accessible ordered crystals, as reported in other studies.182 Preferential fast hydrolysis produces a sponge-like surface, which becomes more dramatic with the extent of mass loss as seen in Figure 4.8. Varying the hydrolysis conditions like temperature and duration has little impact on the sponge-like surface topography, as shown in Figure 4.9, which we attribute to the mild hydrolysis conditions.

96

Figure 4.8 SEM images of (a) Untreated PBT fibers (b) h-PBT fibers treated at 45 °C for 10 min (c) h-PBT fibers treated at 45 °C for 20 min (d) h-PBT fibers treated at 45 °C for 30 min.

97

Figure 4.9 SEM images of h-PBT fiber morphology (a) 50 °C, 10 min; (b) 50 °C, 20 min; (c) 50 °C, 30 min; (d) 55 °C, 10 min; (e) 55 °C, 20 min; (f) 55 °C, 30 min.

Interestingly, this behavior has not been reported for NaOH treated melt blown fibers of PBT, which we attribute to quantitative differences in the reactivity and crystallinity of each polyester. We speculate that the detailed rate of etching of the amorphous versus crystalline domains will strongly influence the development of surface roughness and porosity by analogy with the tradeoff between fundamental reaction kinetics and morphology which contributes to the processes operating during the corrosion of metals.200-202

4.3.4 Superhydrophilicity of h-PBT fiber mats

Surface hydrolysis modifies the surface morphology and the wetting properties of the fiber mats. Based on the contact angle (CA) criterion used to define hydrophobic

203 materials, CAhydrophobic > 90°, the unmodified PBT melt blown fiber mats are hydrophobic, where CAPBT fiber mats =126 ± 4° as shown in Figure 4.10. We attribute this enhanced hydrophobicity compared to PBT molded film to the surface roughness of the 98 fiber mat which results in a meta-stable Cassie wetting state.64, 204

Figure 4.10 Water contact angle of (a) PBT melt blown fiber mat (b) PBT compression molded film

Hydrolysis of the PBT fiber mats significantly enhances the hydrophilicity, as shown in Figure 4.11, where water droplets that are deposited are spontaneously sucked into the h-PBT fiber mats. This wickability is directly tied to the presence of polar carboxyl and hydroxyl moieties generated during surface hydrolysis. Another contributing factor is the sponge-like fiber surfaces and the overall fiber mat roughness, which adds to the surface area and the hydrophilicity of the fiber mats.205, 206 In addition, the high porosity associated with the low fiber volume fraction results in capillary phenomenon that enhances the spontaneous spreading and absorption of the droplet. We therefore classify the h-PBT fiber mats as superhydrophilic.70

99

(a)

(b)

Figure 4.11 Sessile drop measurements on h-PBT fiber mats treated (a) at 45 °C for 10 min (b) at 45 °C for 20 min.

In detail, the rate at which water is drawn into the h-PBT fiber mats is correlated with the duration of NaOH treatment, i.e. water is sucked into the mat treated for 20 minutes much more rapidly than the one etched for 10 minutes, as shown in Figure 10, both at 45

°C. However, there was no significant difference between the mats treated for 20 and 30

100 minutes, and this trend was duplicated at 50 °C and 55 °C. We speculate the rate of water droplet absorption reflects the combined effects of the fiber surface roughness, porosity, pore size, and water droplet volume.207, 208 In order to better to understand the correlation between the hydrolysis duration and the rate of water droplet absorption, we recorded the water droplet absorption by h-PBT fiber mats prepared from Cummins melt blown PBT mats (the experimental details are elaborated in Appendix) using a 1000 fps camera, as exemplified in Figure 4.12. Then, we measured the droplet absorption time (tA) of each h-

PBT sample hydrolyzed for different periods of time (te). We define tA as the duration of the absorption process since the droplet touches the mat until it is completely absorbed according to the images captured by our high-speed camera.

Figure 4.12 Water droplet spreading and imbibition on h-PBT-4 fiber mat, captured by a 1000 fps camera. h-PBT-x refers to the PBT fiber mat hydrolyzed by NaOH methanol solution at 45 °C for x minute(s).

As shown in Figure 4.13, the h-PBT fiber mat starts absorbing water droplets after 2-

101 min etching (blue open square in Figure 4.13), indicating the surface wetting transformation from hydrophobic to superhydrophilic. As the etching time extends, the droplet absorption accelerates and tA decreases sharply to hundreds of milliseconds within 10-min etching. Considering there is no significant variation in both the porosity and the thickness of the fiber mats in the range of 0 ≤ te ≤ 10 min (see Appendix), we attribute such remarkable acceleration of the droplet absorption to both the increased polar moieties and roughness on fiber surfaces. Subsequently, tA increases gradually as the hydrolysis proceeds. We speculate that the decreased fiber mat thickness mainly accounts for the decreased droplet absorption. As the thickness decreases, the overall roughness r decreases, leading to the decrease of the driving force of spreading γLV(rcosθe

- cosθ).209, 210

Figure 4.13 Influence of NaOH etching duration on the rate of water droplet absorption. Red infinity ∞ represents hydrophobic surface. Blue open square □ represents the absorption time of h-PBT-2.

102

4.3.5 Superhydrophilic f-PBT fiber mats

Generation of carboxyl and hydroxyl functional groups at the fiber surfaces provides an opportunity to transform the superhydrophilic fiber mat into a low-surface-energy and textured assembly of PBT fibers with superhydrophobic properties. We implemented this strategy by reacting PFOA with h-PBT fiber mats, resulting in fluorinated PBT (f-PBT) fiber mats. Presumably the fluorinated compound reacts with the surface carboxyl groups, as illustrated in Figure 4.14.

Figure 4.14 Mechanism for reaction of carboxyl group and PFOA

As shown in Figure 4.15 (a), XPS survey spectra obtained from PBT, h-PBT, and f-

PBT fibers demonstrate the presence of fluorine in the f-PBT product. According to the spectra, the untreated PBT fiber mat yields a C/O ratio of 77/23, which is close to the chemical stoichiometry and the literature value of 75/25.211 The C/O ratio from the h-

PBT fiber mat is 74/26, which within experimental uncertainty is the same as that from the untreated PBT fiber mat, consistent with the stoichiometry of the surface chemistry.

For the f-PBT fiber mat, the C/O ratio is 69/25, and the measurement reveals about 5 atom% of fluorine. This indicates a significant amount of fluorine on the surface, considering that fluorination is limited to the outmost surface layer (approximately 10

103 nm-thick) of the fibers. SEM images revealed essentially no change in surface roughness, as illustrated in Figure 4.15 (b).

Figure 4.15 Analysis of (a) the surface chemical composition of f-PBT fibers by XPS and (b) the surface topography of f-PBT fibers by SEM.

104

To better understand the effects of the etching time and hydrolysis temperature on the subsequent fluorination, we determined the surface F atom% for the 45 °C-10 min- etched, 45 °C-20 min-etched, and 55 °C-20 min-etched f-PBT fiber mats and found that this was constant at 5 ± 0.7 atom% in all three cases. This indicates indirectly that the number and the density of the carboxyl group are relatively insensitive to the hydrolysis conditions, which has also been reported by Chen et al.212

To investigate the effect of fluorination on the wetting properties of the fiber mat, we measured the water contact angle. CAf-PBT fiber mat = 156 ± 5°, as shown in Figure 13, confirming a superhydrophobic surface character.69, 213 To evaluate the slide-off angle, we tilted, and even inverted, the substrate, yet the water droplet always remained attached to the surface. This demonstrates simultaneous superhydrophobicity and a high level of droplet adhesion for the f-PBT fiber mat.

Figure 4.16 (a) Water contact angle on f-PBT fiber mat

Such sticky superhydrophobicity has been reported on rose petal and peanut

105 leaves.213-215 We attribute the sticky-superhydrophobicity to the hierarchical micro-/nano- scale roughness of the f-PBT fiber mat. As illustrated in Figure 4.17, the sticky feature results from the Wenzel wetting regime because of water intrusion into the increased interfibrillar space caused by surface hydrolysis. On the other hand the nanoscale roughness on fiber surfaces creates localized Cassie wetting state, generating superhydrophobicity. Therefore, the three-phase contact line is quasi-continuous at the microscale but discontinuous at the nanoscale, resulting in the formation of sticky- superhydrophobicity. Such a special wetting state has been classified as a Cassie impregnating wetting state also known as the petal effect.214, 215 In addition, presumably these little “sponge bumps” on the surface of each fiber are not large enough to mimic the lotus leaf hair-like nanostructure that leads to roll-off superhydrophobicity.216

Figure 4.17 Schematic formation of the sticky-superhydrophobicity on f-PBT fiber mat

4.4 Conclusion

PBT fiber mats fabricated with our lab-scale melt blowing apparatus were exposed to

106 a NaOH methanol solution for controlled periods of time at several temperatures.

Hydrolysis leads to mass loss and increased porosity without disrupting the original overall size and integrity of the material. Etching the fiber mat for 30 minutes reduced the average fiber diameter from 865 nm to less than 400 nm while the diameter distribution

(represented by coefficient of variation, CV) significantly broadens over 80%.

At the same time, hydrolysis creates a textured, sponge-like fiber surface decorated with hydrophilic carboxyl and hydroxyl groups. This combination of chemical and physical surface modification imparts superhydrophilicity. The correlation between hydrolysis duration and the rate of water droplet absorption by h-PBT fiber mats

(prepared from Cummins melt blown PBT mats) was revealed by a high-speed camera. A

2-min hydrolysis at 45 °C would initiate the wetting transformation of the fiber mat from hydrophobic to superhydrophilic.

A subsequent reaction of the h-PBT fibers with PFOA leads to a sticky superhydrophobic surface (f-PBT). 5 atom% of fluorine was achieved through the fluorination. And, the textured, sponge-like topography was retained after fluorination, creating the overall hierarchical nano-/micro-scale roughness. Therefore, the three-phase contact line is quasi-continuous at the microscale but discontinuous at the nanoscale, resulting in the formation of sticky-superhydrophobic fiber mat surface.

107

Chapter 5 Fluorine Enriched Melt Blown Fibers from Polymer Blends of Poly(butylene terephthalate) and A Perfluorinated Copolyester

5.1 Introduction

Melt blowing is a one-step process that generates nonwoven fibers with a typical average diameter of 1-5 µm. Melt blown fibers have been widely applied in various applications, such as filtration media, biological tissue scaffolds, performance apparel, and hygiene fabrics.2, 5, 43, 217, 218 Imparting enhanced hydrophobicity or even superhydrophobicity to melt blown fibers is important to go beyond both scientific and industrial challenges like high-performance water-oil filtration, anti-fogging, and anti- corrosion.5, 219

There have been many strategies to achieve superhydrophobic nonwoven fibers. The easiest way is to fabricate low-surface-energy materials like fluoropolymers directly. Ma et al. electrospun block copolymer poly(styrene-b-dimethylsiloxane) and achieved superhydrophobic fiber mats with fiber diameter in the range of 150 – 400 nm, exhibiting a water contact angle of 163° and a contact angle hysteresis of 15°.170 Miyauchi et al. produced electrospun (PS) microfibers with nano-/micro-pores on fiber surfaces leading to a superhydrophobic fiber mat showing a water contact angle of 159.5°.220 Cengiz et al.

108 fabricated superhydrophobic nanofiber mats (water contact angle 172 ± 1°) of a terpolymer containing perfluoroethyl alkyl methacrylate (Zonyl-TM), methyl methacrylate and butyl acrylate.221 Recently Ferret et al. at Donalson Company discovered that copolymers of vinylidene fluoride (VF2) with hexafluoropropylene (HFP) having HFP content below 20 mol% can be successfully electrospun into fibers showing superhydrophobicity.222

Currently most of the superhydrophobic fibers from direct fabrication of low-surface- energy materials are obtained from electrospinning because of its great capability of reducing fiber diameter to nanoscale and tuning fiber surface topography. However, a lot of low-surface-energy materials like most fluoropolymers are poor candidates for electrospinning because of solubility issues. Moreover, electrospinning involves hazardous organic solvents which limit scale-up of the process and can cause safety concerns.

In contrast to electrospinning, melt spinning, spunbonding, and melt blowing, which all process polymer melts are compatible with a wider range of materials. But it has been challenging to produce superhydrophobic nonwovens directly from these methods because the larger fiber size (1-100 μm), greatly reduces the overall surface roughness which is critical to superhydrophobicity. Therefore, various post-surface treatments aimed at either increasing fiber surface roughness or lowering surface energy have been utilized to achieve superhydrophobicity. Shin et al. used oxygen plasma to etch ply(ethylene terephthalate) (PET) nonwoven fibers followed by plasma enhanced chemical vapor deposition (PECVD) of hexamethyldisiloxane, producing superhydrophobic PET fabrics.223 Zimmermann et al. applied the so called silicone

109 nanofilaments coating to various fabrics and produced durable superhydrophobic surfaces.224 Zhou et al. found a two-step coating process, consisting of dip-coating with silica nanoparticles followed by a second dip-coating of poly(vinylidene fluoride- hexafluoropropylene) (PVDF-HFP), which can impart durable superhydrophobicity to fabrics.225 Other techniques like surface grafting have also been demonstrated to generate superhydrophobic nonwovens.226 But, these techniques are difficult to implement on a large scale for a number of reasons, such as specialized equipment and process steps which can raise the cost.

Another potentially economical alternative approach is blending of functionalized additives which can migrate to the host polymer’s surface and thus enhance surface hydrophobicity or even achieve superhydrophobicity. Surface migration and segregation in thin films of polymer blends have been extensively studied.168, 227-235 But there have been few publications on surface migration of nonwoven fibers, especially for fibers fabricated from melt processing. Hardman et al. reported that superhydrophobic electrospun PS fibers can be obtained by adding in end-functionalized polymer additives.

They characterized the amount of fluorine on the fiber surfaces using XPS and the dependence of water contact angle on the additive concentration in the bulk.168 But they did not identify the amount of additives on the surface, which is more critical for understanding the migration phenomenon. Kaplan et al. used poly(lactic acid-co-glycerol monostearate) (PLA-PGC18) as a dopant to increase hydrophobicity of poly(lactic acid- co-glycolic acid) (PLGA) fibers without discussing surface migration.236 Several patents describe improving water repellency of nonwoven webs by blending in migrating

110 additives, but there are no detailed quantifications of surface enrichment of the additives.

237-239

In this study we report fluorine enriched melt blown fibers from polymer blends of poly(butylene terephthalate) (PBT) and a perfluorinated multiblock copolyester (PFCE).

PFCE was blended with PBT at controlled composition from 1 %, 3 %, 5 %, to 10 % by the overall blend weight, followed by melt blowing. The amount of fluorine on fiber surfaces was characterized by x-ray photoelectron microscopy (XPS), and the amount of corresponding PFCE was calculated based on a detained analysis of the spectra. Fluorine blooming and the effects of the fluorine segregation on surface wetting of melt blown

PBT/PFCE fibers will be discussed in this chapter.

5.2 Experimental

5.2.1 Materials

PBT pellets (Celanex 2008, Ticona) were dried at 100 °C for 12 hours under vacuum to remove moisture. The perfluorinated random multiblock copolyester (PFCE) (Mw =

208 kDa, Ð = 2.41), synthesized by melt polycondensation of poly(butylene terephthalate) (PBT) and 1,3-propanediol based perfluorinated isophthalic polyester

(3GF16), was kindly provided by DuPont USA. Detailed synthesis and characterizations were reported by Drysdale et al.240 The molecular structure of PFCE is illustrated in

Figure 5.1. PFCE was received as solid chunks and was dried at 25 °C for 12 hours under vacuum before use.

111

(m/n ≈ 1)

Figure 5.1 Molecular structure of PFCE.

Rheological properties were measured with a strain-controlled ARES rheometer (TA

Instrument) equipped with a 25 mm parallel-plate fixture. Firstly, dynamic strain sweeps were performed to determine the linear viscoelastic region. Secondly, dynamic frequency sweeps were employed to measure the complex viscosity (η*) and dynamic elastic modulus (G’) and loss modulus (G”) as a function of frequency (1 ≤ ω ≤ 100 rad/s) at

240 °C, as shown in Figure 5.2.

112

Figure 5.2 Rheological properties of PBT and PFCE. (a) Complex viscosity (η*); (b) elastic modulus (G’) and viscosity modulus (G’’). The measurements were carried out using 25-mm-parallel plates at 240 °C.

5.2.2 Preparation of polymer blends

PFCE was cut into small pieces with a razor blade and weighed at controlled compositions of 1 %, 3 %, 5 %, 10 % by the overall weight of the PBT/PFCE blend. The pre-dried PBT pellets were first fed into a HAAKE (Thermo Scientific) batch compounder with roller blades rotating at 100 rpm at 240 °C. The weighed PFCE pieces were added in after 3 min when PBT was melted. Then, the mixing was controlled for another 8 min before the resulting blend was quenched with liquid nitrogen (LN2) to freeze the blend morphology.

113

5.2.3 Melt blowing

The PBT/PFCE blends were dried at 25 °C for 24 hours before melt blowing. Melt blowing was performed with our lab-scale melt blowing apparatus described in Section

2.3.4 in Chapter 2. A melt blowing die with five 0.2-mm-diameter holes was used for the experiments. A stainless steel screen fitted with a blower, located 35 cm away from the melt blowing die, was used to collect the blown fibers. Melt blowing experiments were performed at controlled 240 °C (Tair = Tprocessing = 240 °C) at which PFCE is completely disordered as demonstrated by the rheological measurements in Figure 5.2. The polymer flow rate was controlled at Qpolymer ≈ 0.2 g/(min • hole), and the air pressure at the die exit

(Pinlet) was 6 psi.

Three control experiments, including quiescent annealing, shearing, and extrusion, were done as follows. The quiescent annealing and shearing were carried out at 240 °C under N2 protection with an ARES rheometer (TA Instruments). The shearing was controlled at a shear rate of 10 s-1. The extrusion was performed at 240 °C using the same apparatus as melt blowing except for no air jets blowing.

5.2.4 Characterization of polymer blends

The PBT/PFCE blends were prepared through two approaches for electron microscopy. One approach is cryo-fracture in LN2, and the other one is microtomy at -60

°C with a Leica UC6 microtome using a diamond knife. The resulting sample surface was coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. A scanning electron microscope (SEM) (Hitachi S-4700 or JEOL 6500) was used to image

114 the sample surface. Both secondary and backscattered detectors were employed. 20 – 25

SEM micrographs were taken for each sample.

The resulting polymer blends were also characterized with attenuated total reflectance infrared spectroscopy (ATR-IR) (Thermo Nicolet Avatar 370). Thin films (~ 0.3 mm) of the blends were prepared by compression-molding (230 °C, 5 min) with an ARES rheometer (TA Instruments).

In addition, the thermal properties of PBT, PFCE, and PBT/PFCE blends were characterized by differential scanning calorimetry (DSC) (TA Instrument Q1000). The samples were dried overnight under vacuum to remove moisture. The analysis followed the standard run sequence of heating-cooling-heating. Heating and cooling scans were controlled from -20 to 250 °C at a constant rate of 10 °C /min. Tg and Tm were determined from the second heating scan.

5.2.5 Characterization of fiber morphology

Melt blown fibers of PBT/PFCE blends were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. 25-30 SEM micrographs were taken for each sample with a scanning electron microscope (SEM) (Hitachi S-4700 or JEOL 6500).

The diameters of 400 - 500 fibers were measured using ImageJ software. Then, the

OriginLab software was employed to fit the measurement results with a log-normal distribution function. The geometric average fiber diameter (dav) was extracted from the log-normal fitting. The geometric standard deviation (GSD) and coefficient of variation

(CV) were calculated based on the log-normal fitting results.98

115

The internal morphology of the resulting melt blown PBT/PFCE fibers was investigated by both SEM and TEM. For SEM, the fibers were cryo-fractured in LN2. For

TEM, the fibers were first embedded in epoxy followed by air-drying at room temperature. Then, thin pieces of the epoxy-fiber composite were sectioned with a Leica

UC6 microtome using a diamond knife.

5.2.6 Characterization of fiber surface properties

X-ray photoelectron spectroscopy (XPS) (Surface Science SSX-100) was employed to determine the surface chemical composition of melt blown fibers from PBT/PFCE

2 polymer blends. A monochromatic Al Kα source (200 W) with a spot size of 1 x 1 mm was applied at a take-off angle of 35°, while the pressure of the analysis chamber was maintained at 10-10 Torr. For each sample a survey spectrum was first obtained at a rate of 1 eV/step (6 scans, 0-1100 eV binding energy). Then, high-resolution scans were carried out at a rate of 0.1 eV/step. A low-energy electron flux (5 eV) was used for charge neutralization. Overlapping peaks were deconvoluted using the ESCA 2005 software provided with the XPS system. Binding energies were referenced to C-C/C-H level at 285.0 eV. A Gaussian-Lorentzian model with Gaussian percentages of 80-100% was applied to the peak fittings.

Surface wetting properties of the fiber mats were evaluated by sessile drop measurements using a FAMAS Interface Measurement & Analysis system (Kyowa, DM

– CE1). The static contact angle (CA) was determined by placing a 5 ± 0.5 µL water droplet gently onto the surface of a mat, and images were recorded with a CCD camera.

116

The dynamic contact advancing (θA) and receding (θR) contact angles were measured by pumping liquid into and out of a water droplet, respectively. The droplet was increased from about 5 µL to about 9 µL slowly to determine θA, and it was decreased from about 6

µL to 2 µL to determine θR. Five measurements were taken for each fiber mat sample.

5.3 Results and Discussion

5.3.1 Polymer blends of PBT/PFCE

The miscibility of PBT and PFCE was first characterized by SEM. As shown in

Figure 2.5, PFCE appears as approximatly spherical droplets finely dispersed in the PBT matrix, indicating immiscible polymer blends. The DSC measurements, however, seems to be insensitive to both Tg and Tm so that they provide little valuable information about the miscibility, as shown in Figure 5.4.

(a) PBT/PFCE_1 (ϕ = 1 wt%) (b) PBT/PFCE_3 (ϕ = 3 wt%)

117

(c) PBT/PFCE_5 (ϕ = 5 wt%) (d) PBT/PFCE_10 (ϕ = 10 wt%)

Figure 5.3 Representative SEM images of PBT/PFCE blends with a composition of (a) ϕ = 1 wt%; (b) ϕ = 3 wt%; (c) ϕ = 5 wt%; (d) ϕ = 10 wt%. The samples were prepared by microtoming the blends at -60 °C. SEM micrographs were captured using a backscattered detector which is able to distinguish PFCE and PBT based on the differences in atomic number. PFCE containing atoms with higher atomic numbers like F appears brighter than PBT because it generates more back-scattered electrons.

Figure 5.4 DSC analysis of PBT, PFCE, and PBT/PFCE blends. The data shows the features of the second heating scans at 10 °C.

118

In addition, the solubility parameters of PBT and PFCE were estimated using the group-contribution theory to predict the miscibility,241 as exemplified in Table 5.1. In particular, δPFCE was calculated with equation (5.1),

12xx 1  1 2  2 (5.1)

where δ12 = solubility parameter of copolymers, xi = volume fraction of component i, δi

242 = solubility parameter of component i. Through the calculation we obtained δPBT =

3 0.5 3 0.5 11.2 (cal/cm ) , δPFCE = 9.96 (cal/cm ) . The value of δPBT agrees well with other publications.117, 243, 244 Δδ is calculated to be 1.24 (cal/cm3)0.5 which is larger than the

3 0.5 245 critical value of Δδ (Δδcritical = 0.1 (cal/cm ) ) for miscible blends. Therefore, PBT is predicted to be immiscible with PFCE based on the calculation of Δδ, which is consistent with the SEM observations.

Table 5.1 An example of calculating the solubility parameter (δ) of PBT using the group-contribution method.

unit E unit V δ Group h (J/mol) (cm3/mol) ((cal/cm3)0.5) phenylene 31940 52.4

Eh, total ester 18000 18  PBT 11.2 Vtotal methylene 4940 16.1 a. Eh = cohesive energy

241 b. The values of Eh and V are obtained from Krevelen, D. W.

However, the interfaces of the PFCE droplets are blurred, and some PFCE droplets even tend to mix, possibly forming homogeneous phase with the PBT matrix, as seen in 119

Figure 5.3. Such phenomena differentiate the PBT/PFCE blends from other immiscible binary polymer blends, such as PS/PBT blends discussed in Chapter 2, where sharp gaps/cavities indicating poor adhesion between the dispersed phase and the matrix are observed at the interfaces. It has been reported that copolyesters consisting of units of the immiscible parent homopolymers can improve the miscibility of the polymers in the amorphous phase.246, 247 We speculate that PBT and PFCE form compatibilized blends because of the PBT block in PFCE. Moreover, it is possible that there are transesterifications leading to the formation of copolymers which can serve as compatibilizers and improve the compatibility of PBT and PFCE.248-250 The average diameter of the PFCE domains is about 1 μm. Both the viscosity ratio ( ̇ ≈ 130 s-1 at 100 rpm, ηPFCE/ηPBT ≈ 0.8), which is close to unity, and the improved compatibility, contribute to the small domain size of PFCE.99, 103, 247

In addition, we postulate that there are specific interactions between PBT and PFCE, which also improves the compatibility of the blends. It has been reported that poly(vinylidene fluoride) (PVDF) can form miscible blends with ester containing polymers due to specific interactions, such as hydrogen bonding or dipole-dipole

251, 252 interactions, between the pendent carbonyl groups and the CF2 groups in PVDF.

Also, PVDF has been shown to be partially miscible (limited miscibility) with some semicrystalline polymers having carbonyl groups in their backbone chains due to dipole- dipole interactions, such as with poly(1,4-butylene adipate).253-256 Based on these previous studies, it is likely to achieve dipole-dipole interactions between the C=O groups in PBT and the CF2 groups in PFCE.

120

Figure 5.5 Representative ATR-IR analysis of PBT, PFCE, and their blends: (a) normalized IR spectra and (b) the corresponding 2nd derivatives.

121

As shown in Figure 5.5, the C=O vibration in pure PBT located at 1707 cm-1 is attributed to the crystalline carbonyl group.257 If there are dipole-dipole interactions between the C=O and CF2 groups in the blends, the C=O peak would be expected to shift to lower frequency. However, there are negligible shifts of the C=O peaks in PBT/PFCE blends, as demonstrated by the second derivatives of the spectra in Figure 5.5 (b).

Moreover, there is no noticeable broadening associating with the C=O peaks. Therefore, we conclude that there are no, or extremely weak, specific interactions.

5.3.2 Melt blown fibers from PBT/PFCE

Melt blowing PBT/PFCE blends generates uniform melt blown fibers, as shown in the SEM images in Figure 5.6. The average fiber diameter (dav) is about 1 μm and appears to be relatively insensitive to the amount of blended PFCE. It has been demonstrated that the dav of melt blown fibers primarily depends on the zero shear

51 viscosity of the materials. We attribute the constant dav to the relatively constant zero shear viscosity of the blends (η0,blends ≈ 60 Pa • s).

(a) PBT/PFCE_1 (ϕ = 1 %)

122

(b) PBT/PFCE_3 (ϕ = 3 %)

(c) PBT/PFCE_5 (ϕ = 5%)

(d) PBT/PFCE_10 (ϕ = 10 %)

Figure 5.6 Representative SEM images and the corresponding statistical fiber diameter analysis of (a) PBT/PFCE_1 (ϕ = 1 %); (b) PBT/PFCE_3 (ϕ = 3 %); (c) PBT/PFCE_5 (ϕ = 5 %); (d) PBT/PFCE_10 (ϕ = 10 %).

123

We have demonstrated that the fiber-in-fiber structure can be produced from melt blown binary immiscible polymer blends, as shown in Chapter 2. Since both PBT and

PFCE dissolve in trifluoroacetic acid (TFA), the internal structure of PFCE inside single melt blown PBT fibers cannot be studied by washing away the PBT matrix. Instead, we cryo-fractured the melt blown fibers of PBT/PFCE blends to investigate the internal structure of PFCE, as shown in Figure 5.7. Surprisingly, there are no nanofibers of PFCE embedded inside single melt blown PBT fibers. The representative TEM image in Figure

5.8 further confirms the SEM observation.

Figure 5.7 Representative SEM images showing the internal morphology of cryo-fractured melt blown fibers of PBT/PFCE_10 (ϕ = 10 wt%) blend. (a) longitudinal surface; (b) cross-sectional surface. SEM micrographs were captured using a backscattered detector

124

Figure 5.8 Representative TEM image showing the internal (cross-sectional) morphology of melt blown PBT/PFCE_10 (ϕ = 10 %) fibers. The fibers were embedded in epoxy followed by air-drying at room temperature. The section was prepared by cryo-microtome at -60 °C.

We postulate that the improved compatibility of the blends contributes to the absence of PFCE nanofibers inside PBT microfibers.95 During melt blowing, the starting blends were heated up to 240 °C and were then extruded through a capillary die connected with the melt blowing spinneret. The residence time of the molten polymer in the melt blowing apparatus is about 6 – 8 min before it was blown into fibers. It is likely that the transesterification develops which further enhances the compatibility of the blends. Also, it has been reported that extensional and shear flow fields can induce miscibility of polymer blends.258-260 The molten polymer blends experience primarily shear flow inside the capillary die and melt blowing spinneret, followed by strong extensional flow during

125 fiber blowing. It is possible that the flow fields may also have an impact to the blends compatibility.

To investigate the effects of thermal annealing and flow fields during the process on the internal morphology of melt blown PBT/PFCE fibers, we performed three control experiments where the PBT/PFCE_10 blend was (i) annealed quiescently at 240 °C for 8 min; (ii) sheared at 240 °C ( ̇ = 10 s-1) for 8 min; (iii) extruded at 240 °C using the same apparatus as melt blowing except for no applied air jets.

As shown in Figure 5.9, it is evident that the annealed PFCE domains tend to mix with the PBT matrix. Furthermore, the impact of the flow field is significant. The PFCE domains in both the sheared sample and extrudate appear to be completely mixed with the PBT matrix. All these observations demonstrated that enhanced compatibility would be expected after the blends experience annealing and strong flow field during melt blowing process, which eventually leads to the uniform internal morphology of melt blown PBT/PFCE blend fibers.

(a) PBT/PFCE_10 blend (b) Annealed at 240 °C

126

(c) Sheared at 240 °C (d) Extrudate

Figure 5.9 Representative SEM images showing the morphology of (a) PBT/PFCE_10 (ϕ = 10 %) blend, (b) PBT/PFCE_10 annealed at 240 °C for 8 min, (c) PBT/PFCE_10 sheared at 240 °C ( ̇ = 10 s-1) for 8 min, (c) PBT/PFCE_10 extrudate from 240 °C-extrusion. All samples were cryo-fractured in LN2.

5.3.3 Quantification of PFCE on fiber surfaces

It is well known that low-surface-energy components tend to segregate to the polymer-air interface. Therefore, we expect PFCE to migrate to the surface of melt blown fibers. The fluorine content on fiber surfaces was characterized by XPS and the amount of PFCE near fiber surfaces can be quantitatively determined through two approaches.

One method is based upon the areas of photoemission spectra of C 1s, O 1s, and F 1s from the survey scans of XPS. The other approach is utilizing the curve fitting of the C 1s envelopes from the high-resolution scans of XPS. We present both of the two methods below.

Firstly, calculating the surface chemical compositions through the photoemission areas of C 1s, O 1s, and F 1s is the simplest and most convenient method and is commonly reported in the literature. The photoemission spectra obtained from the survey

127 scans of melt blown PBT/PFCE fibers are shown in Figure 5.10.

(a) PBT/PFCE_1 (ϕ = 1 wt%) (b) PBT/PFCE_3 (ϕ = 3 wt%)

(c) PBT/PFCE_5 (ϕ = 5 wt%) (d) PBT/PFCE_10 (ϕ = 10 wt%)

Figure 5.10 XPS survey spectra showing the photoemission areas of C 1s, O 1s, and F 1s of (a) PBT/PFCE_1 (ϕ = 1 wt%); (b) PBT/PFCE_3 (ϕ = 3 wt%); (c) PBT/PFCE_5 (ϕ = 5 wt%); (d) PBT/PFCE_10 (ϕ = 10 wt%).

Both equation (5.2) and (5.3) were used to determine the amount of PFCE on the fiber surfaces,

128

nF W PFCE F PFCE M  PFCE C C (5.2) C nPFCE nPBT WWPFCE PBT MMPFCE PBT

WW416b PBT 1 (5.3)

where F = atomic concentration of fluorine (F), C = atomic concentration of carbon (C),

i i nPFCE = the average number of element i present in the representative unit of PFCE. nPBT

= the average number of element i present in the representative unit of PBT. For example,

F O nPFCE = 16, and nPBT = 4. M = number average molecular weight of the representative unit. W = surface weight percentage of each component. The surface atomic compositions from the survey scans of XPS and the corresponding weight percentage of PFCE on fiber surfaces are summarized in Table 5.2. Melt blown PBT fibers, PFCE, and 3GF16 were also analyzed for reference.

Table 5.2 Surface chemical compositions derived from the photoemission areas of C 1s, O 1s, and F 1s collected from XPS survey scans.

C O F F * W * W Fiber Sample F/C bulk PFCE, bulk PFCE, surface (atom%) (atom%) (atom%) (atom%) (%) (%)

PBT 76.8 23.2 0 0 0 0 0 PBT/PFCE_1 75.7 21.7 2.6 0.03 0.3 1 9.9 PBT/PFCE_3 70.0 22.8 7.2 0.10 0.7 3 27.7 PBT/PFCE_5 67.4 22.4 10.2 0.15 1.3 5 38.9 PBT/PFCE_10 61.1 20.0 18.9 0.31 2.5 10 69.7 PFCE 43.1 15.5 41.4 0.96 27.6 100 100 3GF16 40.4 15.1 44.5 1.10 38.1 -- --

* Fbulk and WPFCE,bulk were calculated according to the stoichiometry of the molecules.

129

The other approach to determine the surface chemical compositions is line fitting of C

1s envelops collected from the high-resolution scans of XPS. Firstly, the curve fitting and peak deconvolution were carried out on pure PBT melt blown fibers for reference, as shown in Figure 5.11.

Figure 5.11 Line fitting analysis of (a) C 1s envelopes and (b) O 1s envelopes of pure melt blown PBT fibers. 130

The results of the peak deconvolution for PBT is summarized in Table 5.3. The ratio of C1, C2, and C3 is about 5 : 1 : 1 , which is roughly consistent with the relative numbers of C1, C2, and C3 (8, 2, 2) in the repeating unit of PBT. The results also matches the

211 analysis by Burrell et al. Also, the ratio of O1 and O2 is close to 1 : 1.

Table 5.3 Line fitting results of C 1s and O 1s for PBT

C1 C2 C3 Shake-up O1 O2

BE (eV) 285.00 286.46 288.95 291.44 532.00 533.51 Atom % 70.20 14.35 13.89 1.56 46.06 53.93 FWHM (eV) 1.49 1.31 1.32 1.65 1.54 1.72 Sensitivity Factor 1.0 1.0 1.0 1.0 2.49 2.49

In order to quantify the amount of PFCE on surface, the pure PFCE was also analyzed for reference, as shown in Figure 5.12.

131

Figure 5.12 Line fitting analysis of (a) C 1s envelopes, (b) O 1s envelopes, and (c) F 1s envelope of PFCE.

The C 1s envelop of PFCE is more complicated and was fitted with 7 peaks, as shown

261-263 in Figure 5.12 (a). In particular C4 corresponds to CF2–CHF–O at 290.5 eV. C5

264, 265 represents CF2–CFCF3–O and CF2–CF2–CF3 at 291.52 eV. C6 corresponds to O– 132

264 266 CF2–CHF–O and O–CF2–CF2 at 292.69 eV, and C7 correlates with CF3. The results of the peak deconvolution for PFCE is summarized in Table 5.4.

Table 5.4 The line fitting results of C 1s and O 1s for PFCE.

C1 C2 C3 C4 C5 C6 C7 O1 O2 O3

BE (eV) 285.00 286.71 289.01 290.50 291.52 292.69 293.62 532.00 533.72 535.31 Atom % 43.80 8.92 10.22 4.97 10.35 12.90 9.46 33.26 34.46 32.24 FWHM 1.87 1.31 1.43 1.30 1.40 1.43 1.26 1.69 1.69 1.76 (eV) Sensitivity 1.0 1.0 1.0 1.0 1.0 1.0 1.0 2.49 2.49 2.49 Factor

The ratio of C1, C2, C3, C4, C5, C6, and C7 (43.80 : 8.92 : 10.22 : 4.97 : 10.35 : 12.90 :

9.46) is slightly lower than the relative numbers of C1, C2, C3, C4, C5, C6, and C7 (14 : 5 :

4 : 1 : 2 : 3 : 2) in the repeating unit of PFCE, which may be attributed to the surface segregation of the pendent perfluorinated segment. The ratio of O1, O2, and O1 appears to be insensitive to the segregation of the pendent segment.

The high-resolution scans of C 1s from the melt blown PBT/PFCE fibers are presented in Figure 5.13. It is observed qualitatively that the binding energy range of 290

– 295 eV associated with C-F becomes more convex with the increase of the bulk PFCE, implying more fluorine on the fiber surfaces.

133

Figure 5.13 High-resolution scans of C 1s collected from melt blown fibers of pure PBT and PBT/PFCE blends, respectively.

With the referenced C 1s peaks we are able to analyze the surface chemical compositions of the melt blown PBT/PFCE fibers. The curve fitting analysis of C 1s from melt blown PBT/PFCE_10 (ϕ = 10 %) is demonstrated as an example in Figure 5.14. The peak areas of C4, C5, C6, and C7 were used to determine the surface weight percentage of

PFCE, as shown in equation (5.4). According to the line fitting of C 1s of PFCE (Figure

5.12), the total atom% of peak is 37.7 % which is used as the reference for PFCE. The shake-up C 1s was subtracted from the total atom% of PBT C 1s. Peak C3 corresponding to C=O was used to determine the amount of PBT on fiber surfaces. Since peak C3 involves both PBT and PFCE, it is necessary to subtract the amount of C3 from PFCE, as shown in equation (5.5). Combining equation (5.4) and (5.5) we can obtain the surface weight percentage of PFCE, as summarized in 134

Table 5.5.

Figure 5.14 Line fitting analysis of C 1s from melt blown fibers of PBT/PFCE_10 (ϕ = 10 %) blend.

100 M PFCE ()CCCC4 5  6  7   C 37.7 nPFCE WPFCE  100MMPFCE (70.20 14.35 13.89) PBT (5.4) (C4 C 5  C 6  C 7 )  CC   ( Atom % of CCO in PBT )  37.7nnPFCE 13.89 PBT

10.22 Atom% of C C   ( C  C  C  C ) CO in PBT 34.97 10.35  12.90  9.46 4 5 6 7 (5.5)

135

Table 5.5 Surface chemical compositions derived from the line fitting of C 1s of melt blown PBT/PFCE blends.

W W Fiber Sample PFCE, bulk PFCE, surface (%) (%)

PBT 0 0 PBT/PFCE_1 1 4.7 PBT/PFCE_3 3 13.1 PBT/PFCE_5 5 23.5 PBT/PFCE_10 10 41.2

5.3.4 Fluorine enrichment on melt blown PBT/PFCE fibers

The survey scans of XPS shows the amount of fluorine near the fiber surfaces (~ top

10 nm), where the significant fluorine segregation is directly exposed, as is seen in Table

5.2. The surface coverage of PFCE was further determined quantitatively through both the photoemission areas and the C 1s line fitting. The results are compared in Figure 5.15.

It is clear that the WPFCE,surface calculated from the photoemission areas is about twice than that obtained from the C 1s line fitting. It has been suggested that the photoemission area based method is less accurate than the C 1s line fitting.261 Therefore, here we primarily consider the results from the C 1s fitting.

136

Figure 5.15 Surface enrichment of PFCE on melt blown fibers of PBT/PFCE blends.

It is observed that WPFCE,surface is almost 3 times larger than WPFCE,bulk. Nearly 40 vol%

3 (assuming ρpolymer = 1 g/cm ) of the fiber surfaces (~ top 10 nm) are covered by PFCE with only 10 wt% loading in the bulk, indicating the blooming of PFCE on fiber surfaces.

We attribute the blooming to the lower surface energy of PFCE compared to PBT.

Although PFCE segregates on fiber surfaces, the fiber surfaces have not been fully covered by PFCE. Because the fiber solidification happens extremely fast (< 0.8 ms),267 it is highly probable that the distribution of PFCE associated with the diffusion of PFCE near fiber surfaces has not reached the equilibrium. Therefore, to examine whether the distribution of PFCE during fiber blowing reaches the equilibrium, we annealed the melt blown PBT/PFCE_10 fibers at 90 °C for 10 hours. The annealing increases the surface

137 fluorine from 18.9 atom% to 22.4 atom%, corresponding to the increase of the surface coverage of PFCE from about 40 vol% to 45 vol%, which is consistent with our previous hypothesis.

The atomic concentration of fluorine (F atom%) on the fiber surfaces goes up to almost 20 atom% (corresponding to about 40 wt% PFCE on fiber surfaces) by blending in only 10 wt% of PFCE, as presented in Table 5.2. Comparing with the previous studies on melt blown PBT/ECTFE blends,32, 87 the segregation is much more significant. As shown in Figure 5.16, the ratio of F/C goes beyond 0.3 by blending in only 10 wt% of

PFCE while F/C just approaches 0.26 by incorporating 30 wt% of ECTFE.32

Figure 5.16 Comparison between the effects of the loaded fluorinated component on the surface fluorine enrichment of melt blown and electrospun fibers. The data of electrospun PS/PSCF fibers were adopted from Hardman, S. J.

138

The surface enrichment of fluorine in PFCE is stronger than that in ECTFE, resulting in a lower surface energy of PFCE. As shown in Table 5.2 where the F atom% on the surface of PFCE is over 40 atom% compared with the theoretical 27.6 atom% in the bulk.

In contrast, the surface F atom% (37.2 atom%) of ECTFE is roughly consistent with the F atom% in the bulk (37.5 atom%).32 A bigger difference in surface energy would contribute to more fluorine segregation on melt blown PBT/PFCE fibers compared to that on melt blown PBT/ECTFE fibers.268

Blend compatibility should impact surface segregation, although there have been some controversial results.269, 270 PBT appears to be more compatible with PFCE

3 0.5 3 0.5 compared with ECTFE (ΔδPBT/PFCE = 1.24 (cal/cm ) < ΔδPBT/ECTFE = 9.0 (cal/cm ) ), as evidenced by the blends morphology discussed previously. More fluorine segregation on melt blown PBT/ECTFE fibers than that on PBT/PFCE fibers would be expected because of strong phase segregation, which, however, conflicts with the experimental results. A similar phenomenon has also been found in the comparison of surface segregation between compatible PS/PVME and PEO/PMMA blends.269 We postulate that crystallization plays a more powerful role. Since ECTFE has a much stronger capability of crystallization than PFCE, the crystallization of ECTFE itself would weaken the chain mobility and may interrupt the diffusion of ECTFE chains towards to surface because of the formation of crystalline lamallae.231, 271, 272

In addition, it is difficult to evaluate the molecular weight effects here, because it is challenging to measure the molecular weight of ECTFE. However, it has been suggested that complex molecular structures, such as triblock or starblock architectures may

139 enhance entropy-drive surface segregation.273, 274 Compared to linear ECTFE, multiblock copolymer PFCE with a tiny perfluorinated “arm” apparently has a more complex structure, which may also play a role in the more pronounced PFCE segregation.

We also compared our results with electrospun poly(styrene) (PS) fibers loaded with fluoroalkyl end-capped polystyrene168 As shown in Figure 5.16, the fluorine segregation of our melt blown PBT/PFCE fibers is quite comparable with that of electrospun

PS/PSCF fibers with increasing the loaded fluorinated component within 5 wt%, and it is even more significant at 10 wt% loading although the molecular weight of PFCE is almost 20 times larger than that of PSCF. We attribute this more pronounced segregation of PFCE on melt blown PBT fibers to a combined result of the molecular structure, crystallization, and surface energy difference.

5.3.5 Surface wetting of melt blown PBT/PFCE fibers

Fluorine enrichment resulting from the surface segregation of PFCE enhances the hydrophobicity of melt blown PBT fibers. As shown in Figure 5.17, the contact angle

(CA) increases steadily from 128 ± 4° to 155 ± 3° by blending in only 5 wt% of PFCE in the bulk, indicating a superhydrophobic surface.

140

Figure 5.17 Typical images demonstrating the static water contact angle measurements on (a) PBT fiber mat, CA = 128 ± 4°; and (b) PBT/PFCE_5 (ϕ = 5 wt%) fiber mat, CA = 155 ± 3°.

The wetting of solid surfaces depends strongly on both the surface free energy (or chemical composition) and the surface roughness. The melt blown PBT fiber mat exhibits hydrophobicity due to primarily the fiber mat roughness leading to a metal-stable Wenzel wetting state, as discussed in Chapter 4. The overall roughness of melt blown PBT/PFCE fiber mats should be comparable with that of melt blown PBT fiber mats, considering there is little influence on the average fiber diameter (dav), the pore size, and the porosity of the mats by blending in PFCE. On the other hand, the surface energy of PBT fibers is expected to be lowered because of the fluorine segregation on fiber surfaces, which primarily contributes to the increase of CA.

Interestingly, the CA appears to plateau at 5 wt% loading of PFCE, as is seen in

Figure 5.18. A further increase in the loading of PFCE to 10 wt% shows little impact on the static water contact angle. A similar phenomenon was also observed by Hardman et al. on electrospun PS/PSCF fibers.168 They actually found that the static water contact

141 angle decreased with increasing PSCF concentration from 4 wt% to 10 wt%, and they were not able to provide a conclusive argument to explain this counterintuitive fact.

Figure 5.18 Impact of the PFCE concentration in the bulk on the static water contact angle of melt blown PBT fiber mats.

In our present work, XPS analysis has demonstrated that the surface fluorine rises from 10.2 atom% to 18.9 atom% as the PFCE loading increases from 5 wt% to 10 wt%, albeit the fiber surface has not been fully covered by fluorine. It is possible that the surface roughness dominates the surface energy at this point so that wetting is less sensitive to the chemical variations.275 Another possible explanation relies on the understanding of the measured CA. It is likely that the measured macroscopic CA is a metastable CA corresponding to the local Gibbs energy minimum instead of the most stable CA which can be theoretically related to the Young, Wenzel, as well as Cassie-

142

Baxter equations.276 This argument leads to the discussion of the contact angle hysteresis which is presented as follows.

Contact angle hysteresis, evaluated by the measurement of both the advancing and receding contact angles (θA, θR), provides a better understanding of the influence of the fluorine enrichment on fiber surface wetting. θA reveals the ascendant surface asperities, energetically favorable for the moving contact line, whereas the θR is related to the

277 descendent ones. As is shown in Figure 5.19, both θA and θR increase with the increase of WPFCE,bulk. The contact angle hysteresis (H = θA – θR) is relatively stable in the range of 15 – 20°, which may indicate the more pronounced impact of the surface roughness.

Because of the large hysteresis (H > 10°), we classify the melt blown PBT/PFCE fiber mats as sticky-superhydrophobic surfaces.

Figure 5.19 Variation of both the advancing contact angle (θA) and receding contact angle (θR) as a function of the PFCE concentration in the bulk.

143

On the other hand, some researchers have suggested that hysteresis (H) is less meaningful than the adhesion energy (E) which is calculated according to equation (5.6),

E LV(cos  R cos  A ) (5.6)

where E = adhesion energy, and rLV = liquid-vapor surface tension. So, E is proportional to (cosθR – cosθA). We calculated (cosθR – cosθA), and the results are presented in Figure

5.20. It is clear that the adhesion energy decreases with the increase of WPFCE,bulk, which is consistent with our prediction that the surface fluorine enrichment lowers the surface energy of melt blown PBT fiber mats.

Figure 5.20 Effect of the PFCE concentration on the adhesion energy of melt blown PBT fibers.

144

5.4 Conclusion

We investigated melt blown fibers of PBT and PFCE blends. PFCE was melt blended with PBT at controlled composition from 1%, 3%, 5%, to 10% by the overall weight, followed by melt blowing with our lab-scale melt blowing apparatus. SEM analysis showed that PFCE is finely dispersed in the PBT matrix as approximately spherical droplets, indicating immiscible blends. We speculated that the PBT block in PFCE and transesterification play key roles in improving the blends compatibility. Melt blowing

PBT/PFCE blends produced uniform fibers with an average diameter of about 1 μm.

Surprisingly, there are no PFCE nanofibers inside single melt blown PBT microfibers.

We attributed the absence of PFCE nanofibers to the enhanced compatibility of

PBT/PFCE blends because of the transesterification, supported by the control experiments of quiescent annealing, shear, and extrusion.

XPS revealed fluorine enrichment on the surfaces of melt blown PBT/PFCE fibers.

Two approaches, the photoemission area and C 1s line fitting based methods, were employed to quantify the amount of PFCE near the fiber surfaces. The surface fluorine rises to nearly 20 atom% corresponding to 41.2 wt% of PFCE by incorporating only 10 wt% of PFCE in the bulk. The striking fluorine segregation of PFCE was discussed in terms of migration equilibrium, surface energy, blends compatibility as well as molecular structure and was compared with both the melt blown PBT/ECTFE fibers and electrospun

PS/PSCF fibers.

Fluorine blooming results in enhanced hydrophobicity and even superhydrophobicity of melt blown PBT/PFCE fibers. The static water contact angle (155°) plateaus at 5 wt%

145 loading of PFCE, which we attributed to the more pronounced surface roughness effect and the existence of metastable apparent contact angle. Dynamic contact angle measurement further revealed the effects of fluorine enrichment on the surface wetting properties of melt blown PBT/PFCE fibers. Both θA and θR increase with increasing

WPFCE,bulk, leading to the reduction of the adhesion energy.

146

Chapter 6 Exploring Melt Blown Fibers: Shaping the Future of Engineering Materials

6.1 Introduction

The diverse nonwoven manufacturing approaches practiced in industry today will continue to become even more varied in the coming years. As one of the main industrialized methods, melt blowing is proving to be a viable process in the nonwoven ecosystem, delivering cost-effective solutions that address niche opportunities in roll goods, healthcare, industrial and consumer applications.278 Despite extensive studies in both academia and industries since the 1960s, there are still many intriguing issues associated with melt blown fibers to be explored, such as bimodal distributed nano-

/micro-fibers. These issues not only reveal the gaps in the state of art understanding of melt blown fibers but also provide great opportunities to further engineer the materials and to explore future applications. In this chapter, we present several interesting phenomena observed in our studies on melt blown fibers and suggest opportunities for advancing the field in the future. Finally, we conclude with a brief presentation of the design of a filtration cell for use in future research.

6.2 Bimodal distributed nano-/micro-fibers

The fiber diameter distribution can influence the total surface area, pore size, porosity, and even mechanical properties of the fiber mats, so it is of great importance to

147 many advanced fiber applications. One goal has been to produce uniform fibers with narrow diameter distributions.41 But some recent studies indicate that a bimodal distribution of fiber sizes offers improved material performance compared with uniformly distributed fibers. Here, we define a bimodal fiber distribution as a structure comprised of fibers with small average diameters (d) randomly mixed with those having large average diameters (D ≥ 5d), as exemplified in Figure 6.1. The ratio 5 is specified based on recent studies on aerosol filtration using bimodal distributed fibers.279, 280 The overall fiber diameter distributions of such structures exhibit bimodal distributions.

Figure 6.1 A representative SEM image showing the morphology of bimodal distributed polyamide fibers. The small fibers have an average diameter (d) of 110 nm while the larger ones 1520 nm. D/d ≈ 14. The figure is adopted from Holzmeister et al.281

Erben et al found that bimodal nano-/micro-fibers are beneficial for cell growth, adhesion, and proliferation.217 They produced the bimodal fibrous structure by combining

148 melt blowing and electrospinning, as shown in Figure 6.2. Payen et al reported that bimodal distributed fibers perform better than unimodal diameter samples in air filtration, especially when there is a great difference between the two fiber diameters in the bimodal distributed samples.280 Although the bimodal distributed fibers they used are all microfibers, it still demonstrated the significance of the bimodal distribution of fiber diameter on air filtration. But they did not clearly identify how the bimodal distributed microfibers were prepared. In another study on the efficiency of aerosol filtration by Mei et al, it was demonstrated that the bimodal distributed nanofibers can enhance the filtration quality factor compared with unimodal fiber mats.279 The method they used to prepare the bimodal distributed nanofibers is called parallel electrospinning developed by

Holzmeister et al, as shown in Figure 6.3.281 Several other studies associated with the bimodal distributed fibers have also been published.282-284

Figure 6.2 Schematic of the combination of melt blowing and electrospinning used to produce modal distributed nano-/micro-fibers. The figure is adopted from Erben et al.217

149

Figure 6.3 Schematic of the parallel electrospinning set-up to produce bimodal distributed nanofibers. The figure is adopted from Holzmeister et al.281

To the best of our knowledge, all the bimodal distributed fibrous structures have been obtained primarily from either parallel electrospinning or melt blowing/electrospinning.

Is it possible to produce bimodal distributed nano-/micro-fibers from the conventional one-step melt blowing? If melt blowing can producing bimodal distributed nano-/micro- fibers, the manufacturing cost of such structures would be substantially lowered and the applications of the structures would be greatly facilitated.

During our studies of nanofiber fabrication from melt blown fiber-in-fiber polymer blends (Chapter 2), we observed ECTFE nano-/micro-fibers after THF extraction of the melt blown PS matrix, as shown in Figure 6.4. The nanofibers have an average diameter of about 100 nm and the observed microfibers have an approximate average diameter of 1

μm. We speculate that both a bimodal distribution of the minor phase droplets contributes 150 to the formation of the nano-/micro-fibers. It is clear that the overall fiber mats are still porous although the nanofibers bundle together. In addition, the distribution of the nanofibers and microfibers appear to be very heterogeneous. However, such structures may indicate that it is possible to achieve bimodal distributed nano-/micro-fibers through the melt blowing process. Further studies of the melt blown fiber-in-fiber polymer blends, especially the nanofiber formation mechanism, would be beneficial to the exploration of melt blown bimodal distributed nano-/micro-fibers.

Figure 6.4 The morphology of ECTFE nano-/micro-fibers after THF extraction of (a) the melt blown PS/ECTFE_30 (ϕ = 30 %) fibers, (b) the melt blown PS/ECTFE_40 (ϕ = 40 %) fibers. ϕ is the volume fraction of the minor phase ECTFE in the starting PS/ECTFE blends.

6.3 Self-reinforced melt blown fibers

Melt blown fibers have been considered to be mechanically weak compared to spunbond fibers.41 One strategy to improve the tensile strength of melt blown fibers is to generate fibril-reinforced structures, such as island-in-the-sea.285 But this method requires

151 the incorporation of another add-in polymer into the host polymer, which may increase the fabrication cost. Another approach to achieve the strength enhancement is self- reinforcement. Self-reinforced materials can be considered environmentally benign, because they represent likely the best recycling option when reprocessing via remelting is targeted. In addition, these materials offer the possibility of manufacturing lightweight parts and structures because their density is well below those of traditional filled polymers.286 The basic concept of self-reinforcement is the creation of a one-, two or three-dimensional alignment (1D, 2D, or 3D, respectively) within the matrix to fulfill the role of matrix reinforcement.286 As to melt blown fibers, we focus primarily on the formation of the shish-kebab structure leading to the improvement in the mechanical properties of melt blown fibers.

The shish-kebab is a particular crystalline superstructure for semicrystalline polymers like subjected to external flow fields, in which the folded-chain lamellae serving as kebabs that grow perpendicularly to the fully extended-chain crystalline entity

(shish) via lattice/periodic matching.287, 288 The shish-kebab structure is illustrated in

Figure 6.5. Researchers have demonstrated that the shish-kebab can be pure (or classic) or hybrid.289-292 In pure shish-kebabs, both the shishes and kebabs are comprised of the same molecules; while in hybrid shish-kebabs, the shishes and kebabs can be chemically different. Since we mainly consider self-reinforcement, only the classic shish-kebab is discussed here.

152

Figure 6.5 Schematic of the shish-kebab crystalline structure. The figure is adopted from Mackley et al.288

It has been demonstrated that extension is the most effective flow field for generating the shish-kebabs.293, 294 In particular, the shish-kebab can be formed during fiber fabrication and can enhance the mechanical properties of the resulting fibers.295-297 For example, Bashir et al reported stiff and strong poly(ethylene) with a shish kebab morphology by continuous melt extrusion.298 Barham et al reviewed the high-strength fibers from solution and gel spinning where the shish-kebab structure plays a key role.299 Interestingly, there has been no literature reporting the discovery of the shish-kebab structure in melt blown fibers of polyesters like PBT and its effects on the mechanical properties of the melt blown fibers.

We have reported that the surface wetting properties of melt blown PBT fibers can be tuned by alkaline hydrolysis and the subsequent fluorination (see Chapter 4). In our studies we observed a unique surface topography which is very similar to the shish-kebab

153 structure, as shown in Figure 6.6 (b). The shish-kebab-like topography was observed only with the hydrolyzed melt blown PBT fibers produced by Cummins Filtration with diameters less than 2 μm. Fibers with diameters larger than 2 μm show a uniform sponge- like topography instead of the shish-kebab-like structures, as is seen in Figure 6.6 (a). We postulate that the amorphous and low-crystallinity regions react more readily with NaOH, exposing the internal crystalline structures. Since it is challenging to pick out these single fibers with the shish-kebab-like morphology for X-ray diffraction, it is difficult to further identify this morphology.

Beginning in the 1960s, many researchers have studied of the formation of the shish- kebab structures, mainly based on poly(ethylene) and poly(propylene) crystallized from solutions or melts.300-302 Although there have been some controversies regarding the shish nucleation,293, 303, 304 experimental works have suggested that several critical parameters exist for the formation of the shish-kebab structure. For example, the weight average

295 molecular weight (Mw) should be appropriate, and a small amount of high molecular weight chains beyond a critical molecular weight (M*) are usually considered to be essential to form precursors for shish nucleation.305 Also, some critical flow parameters like critical strain are also important.306 Generally speaking, the formation of the shish- kebab structure requires high enough strain and shear rate as well as appropriate molecular weight and its distribution, which are crucial for the formation of stable nuclei after the orientation and alignment of chains rather than relaxation and disengagement.307

Li and de Jeu studied shear-induced crystallization of PBT and found that the flow- induced row nuclei failed to generate the classic shish-kebab structures.308 They argued

154 that the relaxation of the flow-induced row nuclei broke down the nuclei into point precursors (or microshishes). The subsequent growth of lamellar crystals from these point precursors led to little preferred orientation. Therefore, no shish-kebab structure was formed. However, the extensional flow field during melt blowing is very strong with an extensional strain rate of about 10-6 s-1. Is it possible to suppress the relaxation of the flow-induced row nuclei and thus generate the shish-kebabs? In addition, Tian et al recently reported the shish-kebab formation during hot stretching of ultra-high molecular weight polyethylene precursor fibers.309 They found that the formation of the shish- kebabs originated from broken lamellae. As to PBT, is it possible that the strong extensional flow field can realign the point precursors along the flow direction and thus reform the rod-like row nuclei for the shish-kebabs?

Imagine if we were able to fabricate uniform melt blown shish-kebab fibers, both the mechanical properties of the materials and the surface wetting properties would be affected, perhaps favorably. The hierarchical shish-kebab structure increases significantly the surface roughness, providing opportunities to tune the surface wetting properties. For example, Laird et al achieved reversible lotus-to-rose transition on nanohybrid shish- kebab papers.310

Discovery of the shish-kebab-like morphology on alkaline hydrolyzed PBT fibers, shown in Figure 6.6, remains something of a mystery in the absence of a definitive identification of the structure. However, this intriguing morphology reveals that there are still intriguing and unexplained opportunities associated with melt blown fibers. It would be worth exploring the possibility of producing a complete shish-kebab-like fibers using

155 the melt blowing process, because strong and functional melt blown fibers are always desired. Such materials would broaden greatly the applications of melt blown nonwovens.

Figure 6.6 Representative SEM images showing the shish-kebab-like morphology observed on NaOH hydrolyzed melt blown PBT fibers. (a) An overview of the mat structure, (b) A zoom-in showing the shish- kebab-like fibers. The melt blown PBT fibers were received from Cummins Filtration. The fiber mat (1 x 1 2 inch ) was hydrolyzed in a NaOH (0.5 g) methanol solution (Vmethanol + Vwater = 4 + 4 mL) at 45 °C for 60 min.

156

6.4 Superoleophobicity & Wetting robustness

Superhydrophobic surfaces are useful for water-in-diesel fuel filtration. We have created sticky-superhydrophobic melt blown PBT fiber mats (f-PBT) by alkaline hydrolysis and subsequent fluorination. However, our f-PBT fiber mats do not repel oil.

An oil droplet deposited onto a piece of f-PBT fiber mat spontaneously spreads out quickly. We classify such surfaces as superoleophilic surfaces.311, 312 Therefore, our f-

PBT fiber mats are superhydrophobic and superoleophilic.

According to the database of Google Scholar, the superhydrophobic/superoleophilic surface was first reported by Feng et al in 2004.313 They prepared a superhydrophobic/superoleophilic mesh film through the fabrication of micro- and nanostructured rough surfaces from a fluorine-containing material (PTFE). Such films can efficiently separate water from diesel oil, which is similar to our goal. Since then more and more research has been focusing on creating superhydrophobic/superoleophilic surfaces by chemical and/or physical modifications, especially for fibrous structures.58,

314-317

Another very useful surface wetting property, associated with superhydrophobicity, superhydrophilicity, and superoleophilicity, is superoleophobicity. Superoleophobic surfaces refer to those structured surfaces that resist wetting of liquids with much lower surface tension, such as decane (γLV = 23.8 mN/m) or octane (γLV = 21.6 mN/m), exhibiting a contact angle larger than 150°.318, 319 Superoleophobic surfaces can be used for applications like separation of oil from water.320 Superoleophobic surfaces are more difficult to be achieved than superhydrophobic surfaces. Teflon (PTFE) has been

157 considered as the “benchmark” low-surface-energy material on which the water contact angle is about 116°, while the contact angle of decane on flat PTFE film is only 46°, indicating oleophilicity.321

Since surface wetting depends on both the surface energy (surface chemical composition) and surface roughness, researchers have been studying creating superoleophobic surfaces by tuning both the surface chemical composition and surface roughness of the materials. Tuteja et al first reported creating superoleophobic surfaces by designing the so called “re-entrant” or “micro-hoodoo” structure, as shown in Figure

6.7.318

Figure 6.7 Schematic of micro-hoodoos. The figure is adopted from Tuteja et al.318

Tuteja et al reported two types of re-entrant structures.318 One is electrospun fibers of

1H,1H,2H,2H-heptadecafluorodecyl polyhedral oligomeric silsesquioxane (referred to as fluorodecyl POSS), and the other is micro-hoodoos which were vapor phase treated with

1H,1H,2H,2H-perfluorodecyltrichlorosilane, as shown in Figure 6.8. They also demonstrated the robustness of such composite surfaces. They introduced three dimensionless parameters, the spacing ratio D*, the robustness height H*, and the robustness angle T*. The three parameters are calculated according the equations in Table 158

6.1. The most significant contribution of this research to the scientific community is the

“design chart” of liquid repellent surfaces, as shown in Figure 6.9. This design chart provides rational criteria for designing textured superhydrophobic and superoleophobic surfaces. It is observed that cylindrical fibers (wires) possess lower D* and H* compared to other fabricated surfaces with designed regular features like micro-hoodoos, implying more difficulties to achieve robust superoleophobicity.

Figure 6.8 Design parameters for a robust composite surface. (A) electrospun fibers, (B) and (C) micro- hoodoos. All the parameters in (A), (B), and (C) are consistent with Tuteja et al.322

Table 6.1 Design parameters of the robust composite surfaces

Spacing ratio Robustness height Robustness angle Sample D* H* T*

RD 2(1 cos )Rl l Electrospun fibers cap cap sin( ) R D2 2D min

WD 2[(1 cos )R H ] l (sin( min ))lcap Micro-hoodoos ()2 cap W D2 DD(1 * )

However, it is seen that Tuteja’s design parameters (D* and H*) do not take the

159 orientation of nonwoven fibers into considerations, which can introduce huge deviations to the predictions. For example, Areias et al found that a piece of electrospun PLLA fiber mat with randomly oriented fibers showed a smaller water contact angle compared to that with relatively aligned fibers (65.9° vs. 121.3°).323 Therefore, it is difficult to apply these two design parameters to predict the liquid repellency of nonwoven fibers.

Figure 6.9 Design chart of robust liquid repellent surfaces. The figure is adopted from Tuteja et al.319

In addition to the surface roughness, the surface chemical composition is also extremely important. In Tuteja’s study, fluorodecyl POSS was used for both electrospun fibers and micro-hoodoos. Currently almost all superoleophobic surfaces use

160 perfluorinated materials with long chains consisting of more than seven carbons in the backbone. Fluorinated materials repel water and other oils because of the intrinsic properties of fluorine, such as very high electronegativity and low polarizability.324 But it has been suggested that perfluorinated long chains (C ≥ 7) are more bio-accumulated than short chains (C ≤ 4).325, 326 Application of short-chain perfluorinated materials to produce superoleophobic surfaces presented challenges until Guittard et al reported the success of creating superoleophobicity with poly(3,4-ethylenedioxythiophene) (PEDOT) containing perfluorobutyl (C = 4) chains.327 In addition to short perfluorinated materials, silane- based materials have also been used to create superoleophobicity.328, 329 However, both the short-chain perfluorinated and silane-based materials are still bio-accumulated.325, 330

In the long run pure carbon materials will be desired.

Considering the f-PBT fiber mats, several factors may contribute to the absence of the superoleophobicity. First, the number and the density of PFOA which depends primarily on the number and density of the –COOH groups on fiber surfaces may be too low (5 atom% of fluorine). Chen et al reported that the number and density of the –COOH groups was insensitive to the hydrolysis conditions like temperature and time of reaction.

Another strategy to increase the number of –COOH is to graft oligomers or polymers with more –COOH like poly(acrylic acid). Second, the chain length of PFOA can also be varied to explore the impact of the chain length on the liquid repellency. In addition, the surface roughness likely does not satisfying the roughness design criterion for oil repellency. Exploring these three factors in the future may lead us to the superoleophobic

PBT fiber mats.

161

6.5 Water-in-oil Filtration

The goal of producing melt blown nanofibers and modifying the fiber surface wetting properties is to achieve functional fibrous materials serving as the filter media for water- in-oil filtration. Researchers have done numerous studies on the impact of the material properties on the filtration efficiency, such as the surface wetting properties like superhydrophobicity, the fiber diameter and its distribution, the filter depth (thickness).56,

315, 331-337 However, the fundamental influence of the fiber surface roughness on water droplet coalescence and drainage on the fiber mat surface is not yet well understood.338,

339

In order to determine the performance of the functional melt blown fibers prepared in the lab, it is necessary to build up a benchtop filtration setup. Yang et al showed a flatsheet filter holder used for water-in-diesel fuel filtration in Donaldson Company, as shown in Figure 6.10.340

Figure 6.10 Flatsheet filter holder for fuel/water filtration. The picture is adopted from Yang et al.340

162

The Chase Group at the University of Akron has been focusing on water-in-diesel filtration using electrospinning fibers.56, 332 The filter holder they used is quite similar to the Donaldson design, as shown in Figure 6.11.

Figure 6.11 The filter holder (filtration cell) that Chase Group used for water-in-diesel filtration.

However, such filter holders appear as “black boxes” where the filtration details like droplets attaching and coalescence cannot be observed and studied. Krasinski and

Wierzbaa studied the removal of emulsified water from diesel fuel using melt blown PP fiber mats modified by ionization.55 As shown in Figure 6.12, they designed a two-stage filter housing consisting of a separator and a water absorbing filter, which is much closer to the real diesel fuel filter.

163

Figure 6.12 Two-stage filter setup designed by Krasinski and Wierzbaa.55 (a) schematic of flow, (b) photograph of the two-stage filter system: 1-coalescence element, 2-sample point, 3-separation element.

In contrast to the previous filter cells, we designed a clear and transparent filtration cell, as shown in Figure 6.13. The details of the filtration cell are shown in Figure 6.14.

We used transparent (PC) to build the filtration cell so that the filtration details could be seen and even be captured by camera. Such a filtration cell would allow us to study the fundamentals of water droplet coalescence and drainage and correlate the experimental observations with numerical simulations.

164

Figure 6.13 Our filtration cell (a) 3-D view, (b) the photograph of the cell.

165

Figure 6.14 The design of our filtration cell.

166

References

(1) http://www.inda.org/about-nonwovens/ (accessed May 3, 2014).

(2) Russell, S., Development of the nonwovens industry. In Handbook of nonwovens,

Wilson, A., Ed. Woodhead Publishing: Cambridge, England, 2006; pp 1-15.

(3) http://www.technicalnonwovens.com/advantages (accessed May 3, 2015).

(4) http://www.bccresearch.com/market-research/nanotechnology/nanofibers-market- nan043c.html (accessed May 20, 2015).

(5) Hutten, I. M., Processes for Nonwoven Filter Media. In Handbook of Nonwoven

Filter Media, Elsevier: Burlington, MA, 2007; pp 195-244.

(6) Ng, F. S.-F.; Muthu, S. S.; Li, Y.; Hui, P. C.-L., Crit. Rev. Environ. Sci. Technol.

2012, 43, 1795-1822.

(7) Wilson, P. A.; Dallas, M. J., Pediatr Dermatol 1990, 7, 179-184.

(8) Czajka, R., Fibres Text. East. Eur. 2005, 13, 49.

(9) Youn, B.; Lee, E.; Ph, D.; Wadsworth, L. C.; Ph, D., Int. Nonwovens J. 2005, 2-9.

(10) Cavaliere, S.; Subianto, S.; Savych, I.; Jones, D. J.; Rozière, J., Energy Environ.

Sci. 2011, 4, 4761-4761.

(11) Arora, P.; Zhang, Z., Chem. Rev. 2004, 104, 4419-4462.

(12) Luo, C. J.; Stoyanov, S. D.; Stride, E.; Pelan, E.; Edirisinghe, M., Chem. Soc. Rev.

2012, 41, 4708-35.

(13) Keiichi, A.; Miyoshi, O.; Shinzo, T.; Koji, W. Apparatus for spinning synthetic

"islands-in-a-sea" type composite filaments. U.S. Patent 3,692,423, September 19, 1972.

(14) Kim, Y. S.; Jo, D. J.; Kim, J. S.; Yang, I. Y. Spinneret for preparing island-in-the-

167 sea yarns. U.S. Patent 2010/0205926 A1, August 19, 2010.

(15) Huang, C. B.; Chen, S. L.; Lai, C. L.; Reneker, D. H.; Qiu, H.; Ye, Y.; Hou, H.

Q., 2006, 1558.

(16) http://www.natureworksllc.com/News-and-Events/Press-Releases/2009/09-21-09-

Meltblown-Nonwovens (accessed October 9, 2014).

(17) http://www.engr.utk.edu/mse/Textiles/Nanofiber%20Nonwovens.htm (accessed

July 8, 2013).

(18) Misslitz, H.; Kreger, K.; Schmidt, H.-W., Small 2013, 9, 2053-2058.

(19) Lalagiri, M.; Bhat, G.; Singh, V.; Parameswaran, S.; Kendall, R. J.; Ramkumar,

S., Ind. Eng. Chem. Res. 2013, 52, 16513-16518.

(20) http://www.inda.org/BIO/rise2015_810_PPT.pdf (accessed April 20, 2015).

(21) Wente, V., Ind. Eng. Chem 1956, 48, 1342-1346.

(22) Shambaugh, R. L., Ind. Eng. Chem. Res. 1988, 27, 2363-2372.

(23) Butin, R.; Harding, J.; Keller, J. Nonwoven mats by melt blowing. U.S. Patent

3,849,241, November 19, 1974.

(24) Fabbricante, A. S. Micro-denier nonwoven materials made using modular die units. U.S. Patent 6,114,017, September 5, 2000.

(25) Podgórski, A.; Bałazy, A.; Gradoń, L., Chem. Eng. Sci. 2006, 61, 6804-6815.

(26) Brang, J. E.; Wilkie, A.; Haggard, J. S. Method and apparatus for production of meltblown nanofibers. U.S. Patent 2008/0023888 A1, Janauary 31, 2008.

(27) Ellison, C. J.; Phatak, A.; Giles, D. W.; Macosko, C. W.; Bates, F. S., Polymer

2007, 48, 3306-3316.

168

(28) Kayser, J. C.; Shambaugh, R. L., Polym. Eng. Sci. 1990, 30, 1237-1251.

(29) Moore, E. M.; Papavassiliou, D. V.; Shambaugh, R. L., Int. Nonwovens J. 2004,

13, 43-53.

(30) Tyagi, M. K.; Shambaugh, R. L., Ind. Eng. Chem. Res. 1995, 34, 656-660.

(31) Bresee, R. R., 2002, 11, 27-34.

(32) Dawud H, T. Tuning the Diameter and Surface Properties of Meltblown Fibers.

University of Minnesota Twin Cities, Minneapolis, November 2011.

(33) Milligan, M. W.; Haynes, B. D., J. Appl. Polym. Sci. 1995, 58, 159-163.

(34) Milligan, M. W.; Lu, F.; Buntin, R. R.; Wadsworth, L. C., J. Appl. Polym. Sci.

1992, 44, 279-288.

(35) Haynes, B. D.; McManus, J. L.; Duellman, J. M.; Clark, D. F.; Quincy, R. B.

Method of increasing the meltblown jet thermal core length via hot air entrainment. U.S.

Patent 6,613,268 B2, September 2, 2003.

(36) Lee, Y.; Wadsworth, L. C., Polym. Eng. Sci. 1990, 30, 1413-1419.

(37) Choi, K. J.; Spruiell, J. E.; Fellers, J. F.; Wadsworth, L. C., Polym. Eng. Sci. 1988,

28, 81-89.

(38) Uppal, R.; Bhat, G.; Eash, C.; Akato, K., Fibers Polym. 2013, 14, 660-668.

(39) Bresee, R. R.; Qureshi, U. a., J. Eng. Fibers Fabr. 2006, 1, 32-46.

(40) Lee, Y. E.; Wadsworth, L. C., J. Appl. Polym. Sci. 2007, 105, 3724-3727.

(41) Bhat, G.; Malkan, S., Polymer-laid web formation. In Handbook of nonwovens,

Russell, S. J., Ed. Woodhead Publishing LImited: Cambridge, England, 2007; pp 143-

200.

169

(42) Nayak, R.; Kyratzis, I. L.; Truong, Y. B.; Padhye, R.; Arnold, L., J. Text. Inst.

2015, 106, 629-640.

(43) Hassan, M. A.; Yeom, B. Y.; Wilkie, A.; Pourdeyhimi, B.; Khan, S. A., J. Membr.

Sci. 2013, 427, 336-344.

(44) Ron, B.; Zhao, R.; Corporation, B. F., Int. Nonwovens J. 2005, 19-24.

(45) Ghijsels, A.; De Clippeleir, J., Int. Polym. Process. 1994, 9, 252-257.

(46) Uyttendaele, M. A. J.; Shambaugh, R. L., AIChE J. 1990, 36, 175-186.

(47) Thien, N. P.; Tanner, R. I., J. Non-Newton. Fluid Mech 1977, 2, 353-365.

(48) Rao, R. S.; Shambaugh, R. L., Ind. Eng. Chem. Res. 1993, 32, 3100-3111.

(49) Marla, V. T.; Shambaugh, R. L., Ind. Eng. Chem. Res. 2003, 42, 6993-7005.

(50) Jarecki, L.; Ziabicki, A.; Lewandowski, Z.; Blim, A., J. Appl. Polym. Sci. 2011,

119, 53-65.

(51) Tan, D. H.; Zhou, C.; Ellison, C. J.; Kumar, S.; Macosko, C. W.; Bates, F. S., J.

Non-Newtonian Fluid Mech. 2010, 165, 892-900.

(52) Zhou, C.; Tan, D. H.; Janakiraman, A. P.; Kumar, S., Chem. Eng. Sci. 2011, 66,

4172-4183.

(53) Brochocka, A.; Majchrzycka, K., Fibres Text. East. Eur. 2009, 76, 92-98.

(54) Lee, H. J., J. Membr. Sci. 2009, 44, 4645-4652.

(55) Krasinski, A.; Wierzba, P., Sep. Purif. Technol. 2014, 50, 1541-1547.

(56) Patel, S. U.; Chase, G. G., Sep. Purif. Technol. 2014, 126, 62-68.

(57) http://www.scribd.com/doc/262368932/Design-and-Performance-Diesel-Fuel-

Filters#scribd (accessed July 8, 2012).

170

(58) Lee, M. W.; An, S.; Latthe, S. S.; Lee, C.; Hong, S.; Yoon, S. S., ACS Appl.

Mater. Interfaces 2013, 5, 10597-10604.

(59) Liu, Q.; Patel, A. A.; Liu, L., ACS Appl. Mater. Interfaces 2014, 6, 8996-9003.

(60) Marmur, A., Annu. Rev. Mater. Res. 2009, 39, 473-489.

(61) Gao, L.; McCarthy, T. J., Langmuir 2006, 22, 6234-7.

(62) Wenzel, R. N., J. Phys. Chem. 1949, 53, 1466-1467.

(63) Wenzel, R. N., Ind. Eng. Chem 1936, 28, 988-994.

(64) Cassie, A.; Baxter, S., Trans. Faraday Soc. 1944, 40, 546-551.

(65) Koishi, T.; Yasuoka, K.; Fujikawa, S.; Ebisuzaki, T.; Zeng, X. C., Proc. Natl.

Acad. Sci. U. S. A. 2009, 106, 8435-8440.

(66) Boreyko, J. B.; Chen, C. H., Phys. Rev. Lett. 2009, 103, 1-4.

(67) Bormashenko, E.; Pogreb, R.; Whyman, G.; Bormashenko, Y.; Erlich, M., Appl.

Phys. Lett. 2007, 90, 201917.

(68) Bormashenko, E., Adv. Colloid Interface Sci. 2014.

(69) Lafuma, A.; Quéré, D., Nat. Mater. 2003, 2, 457-460.

(70) Drelich, J.; Chibowski, E., Langmuir 2010, 26, 18621-18623.

(71) Wei, Q., Surface modification of textiles. CRC Press: Boca Raton, 2009.

(72) Wolfs, M.; Darmanin, T.; Guittard, F., Polym. Rev. 2013, 53, 460-505.

(73) Roach, P.; Shirtcliffe, N. J.; Newton, M. I., 2008.

(74) Barhate, R.; Ramakrishna, S., J. Membr. Sci. 2007, 296, 1-8.

(75) Marcinčin, A., Prog. Polym. Sci. 2002, 27, 853-913.

(76) Baji, A.; Mai, Y.-W.; Wong, S.-C.; Abtahi, M.; Chen, P., Compos. Sci. Technol.

171

2010, 70, 703-718.

(77) Benton, D. M.; Ferreira, R. Sterilizable and Printable Nonwoven Packaging

Materials. U.S. Patent 2013/0269294 A1, October 17, 2013.

(78) Jinka, S.; Behrens, R.; Korzeniewski, C.; Singh, V.; Arunachalam, A.;

Parameswaran, S.; Coimbatore, G.; Kendall, R.; Wolf, R.; Ramkumar, S., J. Ind. Text.

2013, 42, 501-514.

(79) Çeken, F.; Erdoğan, Ü. H.; Kayacan, O.; Uğurlu, Ş. S., J. Text. Inst. 2011, 103,

669-675.

(80) Catalani, L. H.; Collins, G.; Jaffe, M., Macromolecules 2007, 40, 1693-1697.

(81) Woodfield, T. B. F.; Malda, J.; de Wijn, J.; Péters, F.; Riesle, J.; van Blitterswijk,

C. A., Biomaterials 2004, 25, 4149-4161.

(82) Agarwal, S.; Wendorff, J. H.; Greiner, A., Adv. Mater. 2009, 21, 3343-3351.

(83) Abou-Okeil, A.; Sheta, A. M.; Amr, A.; Ali, M. A., Carbohydr. Polym. 2012, 90,

658-666.

(84) Ching Wen Lou; Lin, C.-M.; Hsing, W.-H.; Chen, A.-P.; Lin, J.-H., Text. Res. J.

2011, 81, 1331-1343.

(85) Huang, Z. M.; Zhang, Y. Z.; Kotaki, M.; Ramakrishna, S., Compos. Sci. Technol.

2003, 63, 2223-2253.

(86) Li, D.; Wang, Y.; Xia, Y., Nano Lett. 2003, 3, 1167-1171.

(87) Zuo, F.; Tan, D. H.; Wang, Z.; Jeung, S.; Macosko, C. W.; Bates, F. S., ACS

Macro Lett. 2013, 2, 301-305.

(88) Bansal, V.; Shambaugh, R. L., Ind. Eng. Chem. Res. 1998, 37, 1799-1806.

172

(89) Marla, V. T.; Shambaugh, R. L.; Papavassiliou, D. V., Ind. Eng. Chem. Res. 2009,

48, 8736-8744.

(90) Song, C. H.; Isayev, A. I., Polymer 2001, 42, 2611-2619.

(91) Xing, Q.; Zhu, M.; Wang, Y.; Chen, Y.; Zhang, Y.; Pionteck, J.; Adler, H. J.,

Polymer 2005, 46, 5406-5416.

(92) Semba, T.; Kitagawa, K.; Nakagawa, M.; Semo Ishiaku, U.; Hamada, H., J. Appl.

Polym. Sci. 2005, 98, 500-508.

(93) Afshari, M.; Kotek, R.; Haghighat Kish, M.; Nazock Dast, H.; Gupta, B. S.,

Polymer 2002, 43, 1331-1341.

(94) Si, X.; Guo, L.; Wang, Y.; Lau, K.-t., Compos. Sci. Technol. 2008, 68, 2943-

2947.

(95) Fakirov, S.; Bhattacharyya, D.; Lin, R.; Fuchs, C.; Friedrich, K., J. Macromol.

Sci., Part B: Phys. 2007, 46, 183-194.

(96) Morrison, F. A., Understanding rheology. Oxford University Press: New York,

2001.

(97) Keller, J. P.; Buntin, R. R. Melt-Blowing Die for Producing Nonwoven Mats.

U.S. Patent 3,825,380, July 23, 1974.

(98) Limpert, E.; Stahel, W. A., PLoS ONE 2011, 6.

(99) Sundararaj, U.; Macosko, C. W., Macromolecules 1995, 28, 2647-2657.

(100) Bousmina, M.; Ait-Kadi, A.; Faisant, J. B., J. Rheol. 1999, 43, 415-433.

(101) Breuer, O.; Sundararaj, U.; Toogood, R. W., Polym. Eng. Sci. 2004, 44, 868-879.

(102) Maric, M.; Macoskot, C. W., Polym. Eng. Sci. 2001, 41.

173

(103) Wu, S., Polym. Eng. Sci. 1987, 27, 335-343.

(104) Palierne, J. F., Rheol Acta 1990, 29, 204-214.

(105) Kim, H.-J.; Seo, Y., Langmuir 2003, 19, 2696-2704.

(106) Wu, S., J. Polym. Sci., Part C: Polym. Symp. 1971, 34, 19-30.

(107) Biresaw, G.; Carriere, C., J. Polym. Sci., Part B: Polym. Phys. 2002, 40, 2248-

2258.

(108) http://www.accudynetest.com/index.html (accessed August 8, 2014).

(109) http://www.solvayplastics.com/sites/solvayplastics/EN/Pages/Solvayplastics.aspx

(accessed March 5, 2012).

(110) Miles, I. S.; Zurek, A., Polym. Eng. Sci. 1988, 28, 796-805.

(111) Metelkin, V.; Blekht, V., Colloid J. USSR 1984, 46, 425-429.

(112) Utracki, L. A., J. Rheol. 1991, 35, 1615-1637.

(113) Bourry, D.; Favis, B. D., J. Polym. Sci. Part B Polym. Phys. 1998, 36, 1889-1899.

(114) Mekhilef, N.; Verhoogt, H., Polymer 1996, 37, 4069-4077.

(115) Astruc, M.; Navard, P., J. Rheol. 2000, 44, 693-712.

(116) Jang, B. N.; Wang, D.; Wilkie, C. A., Macromolecules 2005, 38, 6533-6543.

(117) Vučković, M. V.; Antić, V. V.; Govedarica, M. N.; Djonlagić, J., J. Appl. Polym.

Sci. 2010, 115, 3205-3216.

(118) Narses, S.; Sadaka, F.; Brachais, C.-H.; Couvercelle, J.-P., J. Appl. Polym. Sci.

2013, 129, 2891-2904.

(119) Oh, H. J.; Han, S. H.; Kim, S. S., J. Polym. Sci., Part B: Polym. Phys. 2014, 52,

807-814.

174

(120) Dae Han, C.; Funatsu, K., J. Rheol. (Melville, NY, U. S.) 1978, 22, 113-133.

(121) Chapleau, N.; Favis, B. D., J Mater Sci 1995, 30, 142-150.

(122) Van Puyvelde, P.; Yang, H.; Mewis, J.; Moldenaers, P., J. Rheol. (Melville, NY,

U. S.) 2000, 44, 1401-1415.

(123) Tucker III, C. L.; Moldenaers, P., Annu. Rev. Fluid Mech. 2002, 34, 177-210.

(124) Huneault, M.; Shi, Z.; Utracki, L., Polym. Eng. Sci. 1995, 35, 115-127.

(125) Utracki, L.; Shi, Z., Polym. Eng. Sci. 1992, 32, 1824-1833.

(126) Perilla, J. E.; Jana, S. C., Polym. Eng. Sci. 2004, 44, 2254-2265.

(127) Elmendorp, J. J.; Maalcke, R. J., Polym. Eng. Sci. 1985, 25, 1041-1047.

(128) Gonzalez-Nunez, R.; Arellano, M.; Moscoso, F.; Gonzalez-Romero, V.; Favis, B.,

Polymer 2001, 42, 5485-5489.

(129) Bresee, B. R. R.; Ko, W.-c.; Science, M., INJ Summer 2003.

(130) Elemans, P. H. M.; Bos, H. L.; Janssen, J. M. H.; Meijer, H. E. H., Chem. Eng.

Sci. 1993, 48, 267-276.

(131) Tran, N. H. A.; Brünig, H.; Boldt, R.; Heinrich, G., Polymer 2014, 55, 6354-6363.

(132) Rodriguez-Gonzalez, F. J.; Virgilio, N.; Ramsay, B. A.; Favis, B. D., Adv. Polym.

Technol. 2003, 22, 297-305.

(133) Padilla-Lopez, H.; Vázquez, M. O.; González-Núñez, R.; Rodrigue, D., Polym.

Eng. Sci. 2003, 43, 1646-1656.

(134) Migler, K. B., Phys. Rev. Lett. 2001, 86, 1023-1026.

(135) Li, Z.-M.; Yang, M.-B.; Xie, B.-H.; Feng, J.-M.; Huang, R., Polym. Eng. Sci.

2003, 43, 615-628.

175

(136) Stone, H. A.; Leal, L. G., J. Fluid Mech. 1989, 198, 399-427.

(137) Levitt, L.; Macosko, C. W., Macromolecules 1999, 32, 6270-6277.

(138) Ramachandran, A.; Tsigklifis, K.; Roy, A.; Leal, G., J. Rheol. (Melville, NY, U.

S.) 2012, 56, 45-97.

(139) Cox, R. G., J. Fluid Mech. 1969, 37, 601-623.

(140) Joo, Y. L.; Sun, J.; Smith, M. D.; Armstrong, R. C.; Brown, R. A.; Ross, R. A., J.

Non-Newtonian Fluid Mech. 2002, 102, 37-70.

(141) Matovich, M. A.; Pearson, J. R. A., Ind. Eng. Chem. Fundam. 1969, 8, 512-520.

(142) Cogswell, F., J. Rheol. 1972, 16, 383.

(143) Bognitzki, M.; Czado, W.; Frese, T.; Schaper, A.; Hellwig, M.; Steinhart, M.;

Greiner, A.; Wendorff, J. H., Adv. Mater. 2001, 13, 70-72.

(144) Huang, T.; Marshall, L. R.; Armantrout, J. E.; Yembrick, S.; Dunn, W. H.;

Oconnor, J. M.; Mueller, T.; Avgousti, M.; Wetzel, M. D. Production of nanofibers by melt spinning. U.S. Patent U.S. Patent 2008/0242171 A1, October 2, 2008.

(145) Padron, S.; Fuentes, A.; Caruntu, D.; Lozano, K., J. Appl. Phys. 2013, 113,

024318.

(146) Greiner, A.; Wendorff, J. H., Angew. Chem., Int. Ed. 2007, 46, 5670-5703.

(147) An Tran, N. H.; Brünig, H.; Hinüber, C.; Heinrich, G., Macromol. Mater. Eng.

2014, 299, 219-227.

(148) Jiang, X.; Jiang, T.; Zhang, X.; Zhang, X.; Dai, H., Polym. Eng. Sci. 2013, 53,

1181-1186.

(149) Jang, J.; Lee, D. K., Polymer 2003, 44, 8139-8146.

176

(150) Wu, W.; Tian, H.; Xiang, A., J. Polym. Environ. 2012, 20, 63-69.

(151) Jiang, X.; Jiang, T.; Zhang, X.; Dai, H.; Zhang, X., Polym. Eng. Sci. 2012, 52,

2245-2252.

(152) Mohsin, M.; Hossin, A.; Haik, Y., J. Appl. Polym. Sci. 2011, 122, 3102-3109.

(153) Nishino, T.; Kani, S. c.; Gotoh, K.; Nakamae, K., Polymer 2002, 43, 2869-2873.

(154) Kubo, S.; Kadla, J. F., Biomacromolecules 2003, 4, 561-567.

(155) Nishio, H. Microfibers-generating fibers and a woven or non-woven fabric microfibers - US patent 5,290,626. U.S. Patent 5,290,626, March 1, 1994.

(156) Caulfield, M. J.; Hao, X.; Qiao, G. G.; Solomon, D. H., Polymer 2003, 44, 1331-

1337.

(157) Fyfe, C. A.; McKinnon, M. S., Macromolecules 1986, 19, 1909-1912.

(158) Lau, C.; Mi, Y., Polymer 2002, 43, 823-829.

(159) Han, S.; Kim, C.; Kwon, D., Polymer 1997, 38, 317-323.

(160) George, S. E.; Sublett, B. J.; Miller, R. A. Water-dispersible sulfonated polyester.

U.S. Patent 6,162,890, December 9, 2000.

(161) George, S. E.; Sublett, B. J. Water-dispersible block copolyesters. U.S. Patent

5,709,940, January 20, 1998.

(162) George, S. E.; Lefler, E. W.; Spain, M. Water-dispersible copolyester-ether compositions. U.S. Patent 5,646,237, July 8, 1997.

(163) Shi, B.; Palfery, D., J. Polym. Environ. 2010, 18, 122-127.

(164) https://www.donaldson.com/en/engine/support/datalibrary/070616.pdf (accessed

Mar 23, 2015).

177

(165) Jiang, S. X.; Qin, W. F.; Guo, R. H.; Zhang, L., Surf. Coat. Technol. 2010, 204,

3662-3667.

(166) Huang, F.; Wei, Q.; Liu, Y.; Gao, W.; Huang, Y., J Mater Sci 2007, 42, 8025-

8028.

(167) Ma, M.; Mao, Y.; Gupta, M.; Gleason, K. K.; Rutledge, G. C., Macromolecules

2005, 38, 9742-9748.

(168) Hardman, S. J.; Muhamad-Sarih, N.; Riggs, H. J.; Thompson, R. L.; Rigby, J.;

Bergius, W. N. A.; Hutchings, L. R., Macromolecules 2011, 44, 6461-6470.

(169) Zambianchi, I. L.; It, R. E.; Bonaguidi, P.; It, P.; It, M. Filter for the Removal of

Substances from Blood Products. U.S. Patent U.S. Patent 7,736,516 B2, June 15, 2010.

(170) Ma, M.; Hill, R. M.; Lowery, J. L.; Fridrikh, S. V.; Rutledge, G. C., Langmuir

2005, 21, 5549-5554.

(171) Fareghi, A. R.; Moghaddam, P. N.; Entezami, A. A.; Avval, M. E., Iran. Polym. J.

2013, 22, 361-367.

(172) Bongiovanni, R.; Marchi, S.; Zeno, E.; Pollicino, A.; Thomas, R. R., Colloids

Surf., A 2013, 418, 52-59.

(173) Li, H.; Fu, S.; Peng, L.; Zhan, H., Cellulose 2012, 19, 533-546.

(174) Vasiljević, J.; Gorjanc, M.; Tomšič, B.; Orel, B.; Jerman, I.; Mozetič, M.; Vesel,

A.; Simončič, B., Cellulose 2013, 20, 277-289.

(175) Raghavanpillai, A.; Franco, V. A.; Meredith, W. E., J. Fluorine Chem. 2012, 135,

187-194.

(176) Fernández-Blázquez, J. P.; Fell, D.; Bonaccurso, E.; Campo, A. d., Adv. Colloid

178

Interface Sci. 2011, 357, 234-238.

(177) Wei, X.; Zhao, B.; Li, X.-M.; Wang, Z.; He, B.-Q.; He, T.; Jiang, B., J. Membr.

Sci. 2012, 407–408, 164-175.

(178) Salvagnini, C.; Roback, A.; Momtaz, M.; Pourcelle, V.; Marchand-Brynaert, J., J.

Biomater. Sci., Polym. Ed. 2007, 18, 1491-1516.

(179) Gérard, E.; Bessy, E.; Hénard, G.; Ducoroy, L.; Verpoort, T.; Marchand-Brynaert,

J., J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 5087-5099.

(180) Kim, E. J.; Yeo, G.-d.; Pai, C.-m.; Kang, I.-k., J. Biomed. Mater. Res., Part B

2009, 90B, 849-856.

(181) Liu, H.; Zheng, Y.; Gurgel, P. V.; Carbonell, R. G., J. Membr. Sci. 2013, 428,

562-575.

(182) Ng, R.; Zhang, X.; Liu, N.; Yang, S.-T., Process Biochem. 2009, 44, 992-998.

(183) Hadjizadeh, A.; Ajji, A.; Bureau, M. N., J. Mech. Behav. Biomed. Mater. 2010, 3,

574-583.

(184) Shukla, S. R.; Mathur, M. R., J. Appl. Polym. Sci. 2000, 75, 1097-1102.

(185) Dave, J.; Kumar, R.; Srivastava, H. C., J. Appl. Polym. Sci. 1987, 33, 455-477.

(186) Holmes, S.; Zeronian, S., J. Appl. Polym. Sci. 1995, 55, 1573-1581.

(187) Kotek, R.; Jung, D.-W.; Kim, J. H.; Smith, B.; Guzman, P.; Schmidt, B., J. Appl.

Polym. Sci. 2004, 92, 1724-1730.

(188) Tavanai, H., J. Text. Inst. 2009, 100, 633-639.

(189) Kim, E. S.; Lee, C. H.; Kim, S. H., J. Appl. Polym. Sci. 2009, 112, 3071-3078.

(190) Barhate, R.; Loong, C. K.; Ramakrishna, S., J. Membr. Sci. 2006, 283, 209-218.

179

(191) Illers, K.-H., Colloid. Polym. Sci. 1980, 258, 117-124.

(192) Chuah, H. H., Polym. Eng. Sci. 2001, 41, 308-313.

(193) Sun, S.-P.; Wei, M.; Olson, J. R.; Shaw, M. T., ACS Appl. Mater. Interfaces 2009,

1, 1572-1578.

(194) Ginde, R. M.; Gupta, R. K., J. Appl. Polym. Sci. 1987, 33, 2411-2429.

(195) Goje, A. S.; Chauhan, Y. P.; Mishra, S., Chem. Eng. Technol. 2004, 27, 790-799.

(196) Namboori, C.; Haith, M. S., J. Appl. Polym. Sci. 1968, 12, 1999-2005.

(197) Kallay, N.; Grancari, A. M.; Tomi, M., Text. Res. J. 1990, 60, 663-668.

(198) Park, H.; Szpunar, J., Corr. Sci. 1998, 40, 525-545.

(199) Abayarathna, D.; Hale, E.; O'Keefe, T.; Wang, Y.-M.; Radovic, D., Corr. Sci.

1991, 32, 755-768.

(200) Frankel, G. S., J. Electrochem. Soc. 1998, 145, 2186-2198.

(201) Burstein, G.; Pistorius, P., Corrosion 1995, 51, 380-385.

(202) Sasaki, K.; Burstein, G., Corr. Sci. 1996, 38, 2111-2120.

(203) Quéré, D.; Reyssat, M., Philos. Trans. R. Soc., A 2008, 366, 1539-1556.

(204) Ma, M.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.;

Rutledge, G. C., Adv. Mater. 2007, 19, 255-259.

(205) Quéré, D., Annu. Rev. Mater. Res. 2008, 38, 71-99.

(206) Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K., Langmuir 2000, 16,

5754-5760.

(207) Kissa, E., Text. Res. J. 1996, 66, 660-668.

(208) Patnaik, A.; Rengasamy, R.; Kothari, V.; Ghosh, A., Text. Prog. 2006, 38, 1-105.

180

(209) Wang, L.; Zhao, Y.; Wang, J.; Hong, X.; Zhai, J.; Jiang, L.; Wang, F., Appl. Surf.

Sci. 2009, 255, 4944-4949.

(210) McHale, G.; Shirtcliffe, N.; Aqil, S.; Perry, C.; Newton, M., Phys. Rev. Lett.

2004, 93, 036102-036102.

(211) Burrell, M. C.; Chera, J. J., Surf. Sci. Spectra 1999, 6, 5-8.

(212) Chen, W.; McCarthy, T. J., Macromolecules 1998, 31, 3648-3655.

(213) Balu, B.; Breedveld, V.; Hess, D. W., Langmuir 2008, 24, 4785-4790.

(214) Feng, L.; Zhang, Y.; Xi, J.; Zhu, Y.; Wang, N.; Xia, F.; Jiang, L., Langmuir 2008,

24, 4114-4119.

(215) Yang, S.; Ju, J.; Qiu, Y.; He, Y.; Wang, X.; Dou, S.; Liu, K.; Jiang, L., Small

2014, 10, 294-299.

(216) Bhushan, B.; Her, E. K., Langmuir 2010, 26, 8207-8217.

(217) Erben, J.; Pilarova, K.; Sanetrnik, F.; Chvojka, J.; Jencova, V.; Blazkova, L.;

Havlicek, J.; Novak, O.; Mikes, P.; Prosecka, E.; Lukas, D.; Kuzelova Kostakova, E.,

Mater. Lett. 2015, 143, 172-176.

(218) Maugans, R. A.; Allgeuer, T. A.; Martin, J. M. Method of Making A Breathable

Meltblown Nonwoven. U.S. Patent 6,176,952 B1, January 23, 2001.

(219) Ward, G. F., Filtr. Sep. 2001, 38, 42-43.

(220) Miyauchi, Y.; Ding, B.; Shiratori, S., Nanotechnology 2006, 17, 5151-5156.

(221) Cengiz, U.; Z. Avci, M.; Erbil, H. Y.; Sarac, A. S., Appl. Surf. Sci. 2012, 258,

5815-5821.

(222) Ferrer, I.; Yang, C. Fluoropolymer Fine Fiber. U.S. Patent 8,978,899 B2, March

181

17, 2015.

(223) Shin, B.; Lee, K.-R.; Moon, M.-W.; Kim, H.-Y., Soft Matter 2012, 8, 1817-1823.

(224) Zimmermann, J.; Reifler, F. A.; Fortunato, G.; Gerhardt, L.-C.; Seeger, S., Adv.

Funct. Mater. 2008, 18, 3662-3669.

(225) Zhou, H.; Wang, H.; Niu, H.; Gestos, A.; Lin, T., Adv. Funct. Mater. 2013, 23,

1664-1670.

(226) Higaki, Y.; Kabayama, H.; Tao, D.; Takahara, A., Macromol. Chem. Phys. 2015,

216, 1103-1108.

(227) Ma, J.; Hashimoto, K.; Koganezawa, T.; Tajima, K., J. Am. Chem. Soc. 2013,

135, 9644-9647.

(228) Lee, H.; Archer, L. A., Polymer 2002, 43, 2721-2728.

(229) Lee, H.; Archer, L. A., Macromolecules 2001, 34, 4572-4579.

(230) Khayet, M.; Vázquez Álvarez, M.; Khulbe, K. C.; Matsuura, T., Surf. Sci. 2007,

601, 885-895.

(231) Cheung, Z.-L.; Ng, K.-M.; Weng, L.-T.; Chan, C.-M.; Li, L., Polymer 2006, 47,

3164-3170.

(232) Cheung, Z.-L.; Weng, L.-T.; Chan, C.-M.; Hou, W. M.; Li, L., Langmuir 2005,

21, 7968-7970.

(233) Datla, V.; Shim, E.; Pourdeyhimi, B., Polym. Eng. Sci. 2012, 52, 1920-1927.

(234) Rezaei Kolahchi, A.; Ajji, A.; Carreau, P. J., Polym. Eng. Sci. 2014, 55, 349-358.

(235) Hardman, S. J.; Hutchings, L. R.; Clarke, N.; Kimani, S. M.; Mears, L. L. E.;

Smith, E. F.; Webster, J. R. P.; Thompson, R. L., Langmuir 2012, 28, 5125-5137.

182

(236) Kaplan, J. A.; Lei, H.; Liu, R.; Padera, R.; Colson, Y. L.; Grinstaff, M. W.,

Biomacromolecules 2014, 15, 2548-2554.

(237) Quincy, R. B.; Yahiaoui, A.; McManus, J. L. Nonwoven Webs Having Zoned

Migration of Internal Additives. U.S. Patent 6,203,889 B1, March 20, 2001.

(238) Powers, M. D.; Fitting, S. W. Water Repellant Meltblown Webs and Laminates.

U.S. Patent 2002/0142691 A1, October 3, 2002.

(239) Drysdale, N. E.; Madeleine, D. G.; Nederberg, F. Fabric Prepared from

Fluorinated Polyester Blend Yarns. U.S. Patent 8,889,256 B2, November 18, 2014.

(240) Drysdale, N. E.; Brun, Y.; McCord, E. F.; Nederberg, F., Macromolecules 2012,

45, 8245-8256.

(241) Krevelen, D. W.; Van Krevelen, D., Properties of polymers: their correlation with chemical structure, their numerical estimation and prediction from additive group contributions. Elsevier: Amsterdam, 1990.

(242) Paul, D.; Barlow, J., Polymer 1984, 25, 487-494.

(243) Juhl, T. B.; Bach, D.; Larson, R. G.; Christiansen, J. d.; Jensen, E. A., Polymer

2013, 54, 3891-3897.

(244) Papadopoulou, C. P.; Kalfoglou, N. K., Eur. Polym. J. 1997, 33, 191-197.

(245) Coleman, M. M.; Serman, C. J.; Bhagwagar, D. E.; Painter, P. C., Polymer 1990,

31, 1187-1203.

(246) Köppl, T.; Brehme, S.; Pospiech, D.; Fischer, O.; Wolff-Fabris, F.; Altstädt, V.;

Schartel, B.; Döring, M., J. Appl. Polym. Sci. 2013, 128, 3315-3324.

(247) Pospiech, D.; Häußler, L.; Korwitz, A.; Fischer, O.; Starke, S.; Jehnichen, D.;

183

Köppl, T.; Altstädt, V., High Perform. Polym. 2012, 24, 64-73.

(248) Hopfe, I.; Pompe, G.; Eichhorn, K. J., Polymer 1997, 38, 2321-2327.

(249) Ihm, D. W.; Park, S. Y.; Chang, C. G.; Kim, Y. S.; Lee, H. K., J. Polym. Sci., Part

A: Polym. Chem. 1996, 34, 2841-2850.

(250) Porter, R. S.; Wang, L. H., Polymer 1992, 33, 2019-2030.

(251) Saito, H.; Stuehn, B., Macromolecules 1994, 27, 216-218.

(252) Tomura, H.; Saito, H.; Inoue, T., Macromolecules 1992, 25, 1611-1614.

(253) Penning, J. P.; St. John Manley, R., Macromolecules 1996, 29, 77-83.

(254) Jo, W. H.; Park, S. J.; Kwon, I. H., Polym. Int. 1992, 29, 173-178.

(255) Fujita, K.; Kyu, T.; St. John Manley, R., Macromolecules 1996, 29, 91-96.

(256) Liu, L.-Z.; Chu, B.; Penning, J. P.; Manley, R. S. J., Macromolecules 1997, 30,

4398-4404.

(257) Di Lorenzo, M. L.; Rubino, P.; Cocca, M., Eur. Polym. J. 2013, 49, 3309-3317.

(258) Katsaros, J.; Malone, M.; Winter, H. H., Polym. Bull. 1986, 16, 83-88.

(259) Fernandez, M. L.; Higgins, J. S.; Horst, R.; Wolf, B. a., Polymer 1995, 36, 149-

154.

(260) Papathanasiou, T. D.; Higgins, J. S.; Soontaranun, W., Polym. Eng. Sci. 1999, 39,

2461-2474.

(261) Toselli, M.; Gardella Jr, J. A.; Messori, M.; Hawkridge, A. M.; Pilati, F.; Tonelli,

C., Polym. Int. 2003, 52, 1262-1274.

(262) Lei, Y.-G.; Ng, K.-M.; Weng, L.-T.; Chan, C.-M.; Li, L., Surf. Interface Anal.

2003, 35, 852-855.

184

(263) Cui, W. G., Polym. Degrad. Stab. 2008, 93, 731-738.

(264) Mori, S.; Morales, W., J. Vac. Sci. Technol., A 1990, 8, 3354-3356.

(265) Kassis, C. M.; Steehler, J. K.; Betts, D. E.; Guan, Z.; Romack, T. J.; DeSimone, J.

M.; Linton, R. W., Macromolecules 1996, 29, 3247-3254.

(266) No l, S., J. Vac. Sci. Technol., A 1991, 9, 32-32.

(267) Ko, W. C.; Bresee, R. R., Appl. Spectrosc. 2003, 57, 636-641.

(268) Affrossman, S.; Hartshorne, M.; Jerome, R.; Pethrick, R. A.; Petitjean, S.; Vilar,

M. R., Macromolecules 1993, 26, 6251-6254.

(269) Sakellariou, P., Polymer 1993, 34, 3408-3415.

(270) Bhatia, Q. S.; Pan, D. H.; Koberstein, J. T., Macromolecules 1988, 21, 2166-

2175.

(271) Cicero, J.; Dorgan, J., J. Polym. Environ. 2001, 9, 1-10.

(272) Yang, J.; Liang, Y.; Han, C. C., Langmuir 2013, 30, 394-401.

(273) Sikka, M.; Singh, N.; Karim, A.; Bates, F. S.; Satija, S. K.; Majkrzak, C. F., Phys.

Rev. Lett. 1993, 70, 307-310.

(274) Walton, D. G.; Soo, P. P.; Mayes, A. M.; Sofia Allgor, S. J.; Fujii, J. T.; Griffith,

L. G.; Ankner, J. F.; Kaiser, H.; Johansson, J.; Smith, G. D.; Barker, J. G.; Satija, S. K.,

Macromolecules 1997, 30, 6947-6956.

(275) Di Mundo, R.; Palumbo, F.; d’Agostino, R., Langmuir 2010, 26, 5196-5201.

(276) Wolansky, G.; Marmur, A., Colloids Surf., A 1999, 156, 381-388.

(277) Ruiz-Cabello, F. J. M.; Rodríguez-Valverde, M. A.; Cabrerizo-Vílchez, M. A.,

Adv. Colloid Interface Sci. 2014, 206, 320-327.

185

(278) http://www.nonwovens-industry.com/issues/2015-05-

01/view_features/meltblown-pla-nonwoven-materials-push-the-innovation-index- forward/ (accessed May 20, 2015).

(279) Mei, Y.; Wang, Z.; Li, X., J. Appl. Polym. Sci. 2013, 128, 1089-1094.

(280) Payen, J.; Vroman, P.; Lewandowski, M.; Perwuelz, A.; Callé-Chazelet, S.;

Thomas, D., Text. Res. J. 2012, 82, 1948-1959.

(281) Holzmeister, A.; Rudisile, M.; Greiner, A.; Wendorff, J. H., Eur. Polym. J. 2007,

43, 4859-4867.

(282) Abdal-hay, A.; Pant, H. R.; Lim, J. K., Eur. Polym. J. 2013, 49, 1314-1321.

(283) Pant, H. R.; Park, C. H.; Tijing, L. D.; Amarjargal, A.; Lee, D.-H.; Kim, C. S.,

Colloids Surf., A 2012, 407, 121-125.

(284) Pant, H. R.; Bajgai, M. P.; Nam, K. T.; Chu, K. H.; Park, S.-J.; Kim, H. Y.,

Mater. Lett. 2010, 64, 2087-2090.

(285) Fedorova, N.; Pourdeyhimi, B., J. Appl. Polym. Sci. 2007, 104, 3434-3442.

(286) Kmetty, Á.; Bárány, T.; Karger-Kocsis, J., Prog. Polym. Sci. 2010, 35, 1288-

1310.

(287) Wang, K.; Chen, F.; Li, Z.; Fu, Q., Prog. Polym. Sci. 2014, 39, 891-920.

(288) Mackley, M. R.; Keller, A., Philos. Trans. R. Soc., A 1975, 278, 29-66.

(289) Xu, H.; Xie, L.; Chen, J.-B.; Jiang, X.; Hsiao, B. S.; Zhong, G.-J.; Fu, Q.; Li, Z.-

M., Mater. Horiz. 2014, 1, 546-546.

(290) Xu, X.; Wang, H.; Jiang, L.; Wang, X.; Payne, S. A.; Zhu, J. Y.; Li, R.,

Macromolecules 2014, 47, 3409-3416.

186

(291) Nie, M.; Kalyon, D. M.; Fisher, F. T., ACS Appl. Mater. Interfaces 2014, 6,

14886-14893.

(292) Wang, X.; Salick, M. R.; Wang, X.; Cordie, T.; Han, W.; Peng, Y.; Li, Q.; Turng,

L.-S., Biomacromolecules 2013, 14, 3557-3569.

(293) Somani, R. H.; Yang, L.; Zhu, L.; Hsiao, B. S., Polymer 2005, 46, 8587-8623.

(294) Statton, W. O., J. Polym. Sci., Part C: Polym. Symp. 1967, 20, 117-144.

(295) Bashir, Z.; Keller, A., Colloid & Polymer Sci 1989, 267, 116-124.

(296) Nakae, M.; Uehara, H.; Kanamoto, T.; Zachariades, A. E.; Porter, R. S.,

Macromolecules 2000, 33, 2632-2641.

(297) Pennings, a. J.; Smook, J.; de Boer, J.; Gogolewski, S.; van Hutten, P. F., Pure

Appl. Chem. 1983, 55, 777-798.

(298) Bashir, Z.; Odell, J. A.; Keller, A., J Mater Sci 1986, 21, 3993-4002.

(299) Barham, P. J.; Keller, A., J Mater Sci 1985, 20, 2281-2302.

(300) Kimata, S.; Sakurai, T.; Nozue, Y.; Kasahara, T.; Yamaguchi, N.; Karino, T.;

Shibayama, M.; Kornfield, J. A., Science 2007, 316, 1014-1017.

(301) Kumaraswamy, G., J. Macromol. Sci., Part C: Polym. Rev. 2005, 45, 375-397.

(302) Mykhaylyk, O. O.; Chambon, P.; Graham, R. S.; Fairclough, J. P. A.; Olmsted, P.

D.; Ryan, A. J., Macromolecules 2008, 41, 1901-1904.

(303) Wang, K.; Chen, F.; Zhang, Q.; Fu, Q., Polymer 2008, 49, 4745-4755.

(304) Shen, B.; Liang, Y.; Zhang, C.; Han, C. C., Macromolecules 2011, 44, 6919-

6927.

(305) Somani, R. H.; Hsiao, B. S.; Nogales, A.; Srinivas, S.; Tsou, A. H.; Sics, I.; Balta-

187

Calleja, F. J.; Ezquerra, T. A., Macromolecules 2000, 33, 9385-9394.

(306) Yan, T.; Zhao, B.; Cong, Y.; Fang, Y.; Cheng, S.; Li, L.; Pan, G.; Wang, Z.; Li,

X.; Bian, F., Macromolecules 2010, 43, 602-605.

(307) Xu, H.; Zhong, G.-J.; Fu, Q.; Lei, J.; Jiang, W.; Hsiao, B. S.; Li, Z.-M., ACS Appl.

Mater. Interfaces 2012, 4, 6774-6784.

(308) Li, L.; de Jeu, W. H., Macromolecules 2004, 37, 5646-5652.

(309) Tian, Y.; Zhu, C.; Gong, J.; Yang, S.; Ma, J.; Xu, J., Polymer 2014, 55, 4299-

4306.

(310) Laird, E. D.; Bose, R. K.; Qi, H.; Lau, K. K. S.; Li, C. Y., ACS Appl. Mater.

Interfaces 2013, 5, 12089-12098.

(311) Wang, C.; Yao, T.; Wu, J.; Ma, C.; Fan, Z.; Wang, Z.; Cheng, Y.; Lin, Q.; Yang,

B., ACS Appl. Mater. Interfaces 2009, 1, 2613-2617.

(312) Drelich, J., Surf. Innovations 2013, 1, 248-254.

(313) Feng, L.; Zhang, Z.; Mai, Z.; Ma, Y.; Liu, B.; Jiang, L.; Zhu, D., Angew. Chem.,

Int. Ed. 2004, 43, 2012-2014.

(314) Li, H.; Wu, W.; Bubakir, M. M.; Chen, H.; Zhong, X.; Liu, Z.; Ding, Y.; Yang,

W., J. Appl. Polym. Sci. 2014, 131, n/a-n/a.

(315) Zhang, W.; Shi, Z.; Zhang, F.; Liu, X.; Jin, J.; Jiang, L., Adv. Mater. 2013, 25,

2071-2076.

(316) Wu, J.; Wang, N.; Wang, L.; Dong, H.; Zhao, Y.; Jiang, L., ACS Appl. Mater.

Interfaces 2012, 4, 3207-3212.

(317) Xue, C.-H.; Ji, P.-T.; Zhang, P.; Li, Y.-R.; Jia, S.-T., Appl. Surf. Sci. 2013, 284,

188

464-471.

(318) Tuteja, A.; Choi, W.; Ma, M.; Mabry, J. M.; Mazzella, S. A.; Rutledge, G. C.;

McKinley, G. H.; Cohen, R. E., Science 2007, 318, 1618-1622.

(319) Tuteja, A.; Choi, W.; McKinley, G. H.; Cohen, R. E.; Rubner, M. F., MRS Bull.

2008, 33, 752-758.

(320) Zhang, X.; Shi, F.; Niu, J.; Jiang, Y.; Wang, Z., J. Mater. Chem. 2008, 18, 621-

633.

(321) Coulson, S. R.; Woodward, I.; Badyal, J. P. S.; Brewer, S. A.; Willis, C., J. Phys.

Chem. B 2000, 104, 8836-8840.

(322) Tuteja, A.; Choi, W.; Mabry, J. M.; McKinley, G. H.; Cohen, R. E., Proc. Natl.

Acad. Sci. U. S. A. 2008, 105, 18200-18205.

(323) Areias, A. C.; Ribeiro, C.; Sencadas, V.; Garcia-Giralt, N.; Diez-Perez, A.;

Gomez Ribelles, J. L.; Lanceros-Mendez, S., Soft Matter 2012, 8, 5818-5825.

(324) Giannetti, E., Polym. Int. 2001, 50, 10-26.

(325) Buck, R. C.; Franklin, J.; Berger, U.; Conder, J. M.; Cousins, I. T.; de Voogt, P.;

Jensen, A. A.; Kannan, K.; Mabury, S. A.; van Leeuwen, S. P. J., Integr. Environ. Assess.

Manage. 2011, 7, 513-541.

(326) Guo, J.; Resnick, P.; Efimenko, K.; Genzer, J.; DeSimone, J. M., Ind. Eng. Chem.

Res. 2008, 47, 502-508.

(327) Bellanger, H.; Darmanin, T.; De Givenchy, E. T.; Guittard, F., RSC Adv. 2013, 3,

5556-5562.

(328) Im, M.; Im, H.; Lee, J.-H.; Yoon, J.-B.; Choi, Y.-K., Soft Matter 2010, 6, 1401-

189

1404.

(329) Zhang, J.; Seeger, S., Angew. Chem., Int. Ed. 2011, 50, 6652-6656.

(330) Rücker, C.; Kümmerer, K., Chem. Rev. 2014.

(331) Kulkarni, P. S.; Patel, S. U.; Patel, S. U.; Chase, G. G., Sep. Purif. Technol. 2014,

124, 1-8.

(332) Patel, S. U.; Patel, S. U.; Chase, G. G., Energy Fuels 2013, 27, 2458-2464.

(333) Viswanadam, G.; Chase, G. G., Sep. Purif. Technol. 2013, 104, 81-88.

(334) Kulkarni, P. S.; Patel, S. U.; Chase, G. G., Sep. Purif. Technol. 2012, 85, 157-164.

(335) Shin, C.; Chase, G. G., J. Dispersion Sci. Technol. 2006, 27, 517-522.

(336) Shin, C.; Chase, G. G.; Reneker, D. H., AIChE J. 2005, 51, 3109-3113.

(337) Shin, C.; Chase, G. G., AIChE J. 2004, 50, 343-350.

(338) Ma, S.; Kang, Y.; Cui, S., J. Dispersion Sci. Technol. 2013, 35, 103-110.

(339) Agarwal, S.; Arnim, V. V.; Stegmaier, T.; Planck, H.; Agarwal, A., Sep. Purif.

Technol. 2013, 107, 19-25.

(340) Yang, C.; Larsen, S., Understanding emulsified water filtration from diesel fuels.

In Proceedings of the 8th International Filtration Conference, San Antonio, TX, 2007.

(341) Muller, D.; Krobjilowski, A., Nonwovens 2001, 11.

(342) Mehta, R.; Kumar, V.; Bhunia, H.; Upadhyay, S. N., J. Macromol. Sci., Part C:

Polym. Rev. 2005, 45, 325-349.

(343) Harris, A. M.; Lee, E. C., J. Appl. Polym. Sci. 2008, 107, 2246-2255.

(344) Byun, Y.; Kim, Y. T.; Whiteside, S., J. Food Eng. 2010, 100, 239-244.

(345) Postema, A.; Pennings, A., J. Appl. Polym. Sci. 1989, 37, 2351-2369.

190

(346) Fundador, N. G. V.; Takemura, A.; Iwata, T., Macromol. Mater. Eng. 2010, 295,

865-871.

(347) Singh, A.; Naskar, A. K.; Haynes, D.; Drews, J.; Smith, D. W., Polym. Int. 2011,

60, 507-516.

(348) Wang, Z.; Macosko, C. W.; Bates, F. S., ACS Appl. Mater. Interfaces 2014, 6,

11640-11648.

(349) Arai, K.; Funaki, A.; Phongtamrug, S.; Tashiro, K., Polymer 2009, 50, 4612-

4617.

191

Appendix A Synthesis and Characterization of Poly(L-lactide) and Perfluoropolyether Block Copolymer

A.1 Introduction

Conventional nonwoven products are usually based on traditional thermoplastic resins, such as PP, PE, and PET. However, the increasing environmental consciousness and demands of legislative authorities, the manufacture, use and removal of products made of these traditional polymers are considered more critically. Application of biodegradable polymers is one remedy to solve this issue. Poly(lactic acid) (PLA) is one of the most promising biodegradable polymers. It is one of the few polymers in which the stereochemical structure can be modified by polymerizing a controlled mixture of L- or

D- isomers. It can be easily processed with standard polymer processing equipment to produce molded parts, film, and fibers. But the surface wetting properties of PLA do not meet the requirements for applications like high-performance water-in-diesel fuel filtration. A superhydrophobic surface is desired to improve the filtration efficiency.

This study aims at enhancing the surface hydrophobicity of poly(L-lactide) (PLLA) by synthesizing PLLA based block copolymer containing a perfluoropolyether (PFPE) block. The synthesis and characterizations of the PLLA-PFPE-PLLA triblock copolymer and the subsequent fabrication of melt blown fibers will be presented.

192

A.2 Experimental

A.2.1 Materials

L-latide was purchased from Purac, perfluoropolyether (PFPE) poly(tetrafluoroethylene oxide-co-difluoromethylene oxide) α,ω-diol (Fomblins Z DOL, Mn = 2 kg/mol) was purchased from Solvay-Solexis (USA). The molecular structure of PFPE is shown in Figure A.1. 1,8- diazabicycloundec-7-ene (DBU) and tin (II) octoate [(tin (II) bis(2-ethylhexanoate),

(Sn(Oct)2)] were purchased from Sigma Aldrich.

Figure A.1 The molecular structure of PFPE.

A.2.2 Synthesis of PLLA-PFPE-PLLA

Solution polymerization 6 g L-lactide and 0.42 g PFPE were fed into a round bottom flask held in a glove box. DBU was selected as the catalyst and was added dropwise. dichloromethane (CH2Cl2) and anhydrous chloroform (CHCl3) were selected as solvents, respectively. The polymerizations were performed at room temperature (~ 25 °C).

Bulk polymerization 6 g L-lactide and 0.42 g PFPE were feed into a pressure vessel held in a glove box. At the same time, 0.5 wt% Sn(Oct)2 was added as the catalyst. No solvent was added, which is different from the approach reported by D.W. Smith group.

193

The pressure vessel was then immersed in oil bath that was preheated to the reaction temperature. The polymerization was carried out at 130 °C for 12 hours. The reaction mixture was stirred until the mixture solidified. After polymerization, the pressure vessel was cooled to room temperature (~ 25 °C). The reaction product was dissolved in chloroform (CHCl3), followed by precipitation in methanol at room temperature. The precipitated product was then dried under vacuum at 40 °C for 12 hours.

A.2.3 Characterizations of PLLA-PFPE-PLLA

1 19 H NMR and F NMR spectra analysis was performed using CDCl3 as the solvent to confirm the formation of PLLA-PFPE-PLLA triblock copolymer. Size exclusion chromatography (SEC) was applied to obtain the molecular weights and polydispersity index (PDI) based on polystyrene standards. The thermal properties of the resulting block copolymers were characterized by differential scanning calorimetry (DSC) (TA

Instrument Q1000). The samples were analyzed under a standard run sequence of heating-cooling-heating from -20 °C to 250 °C at a scan rate of 10 °C/min, and Tg and Tm were analyzed in the second heating run.

Rheological properties were measured with a strain-controlled ARES rheometer (TA

Instrument) equipped with a 25 mm parallel-plate fixture. Firstly, dynamic strain sweeps were performed to determine the linear viscoelastic region. Secondly, dynamic frequency sweeps were employed to measure the complex viscosity (η*) and moduli (G’, G”) as a function of frequency (1 ≤ ω ≤ 100 rad/s) at 180 °C.

194

A.2.4 Melt blowing & Spin coating

The synthesized block copolymer was dried under vacuum overnight (6 – 9 hours) at at 50 °C. Melt blowing experiments were performed at Qpolymer ≈ 0.2 g/(min • hole), Pinlet

= 5 psi, and Tprocessing = 180 °C , 190 °C, 200 °C, respectively. Melt blown fiber mat was collected with our continuous fiber collector located about 55 cm away from the melt blowing die.

Spun-cast thin films were prepared on glass substrates using a spin-coater (1500 rpm). The solution was prepared by dissolving about 1 g PLLA-PFPE-PLLA into about 5 mL CHCl3. The spun-cast films were dried for 24 hours in vacuum.

A.2.5 Fiber characterization

Melt blown fibers were coated with gold/palladium for 30 seconds using a Denton

DV – 502 sputter coater. 25-30 SEM micrographs were taken for each sample with a scanning electron microscope (SEM) (Hitachi S-4700 or JEOL 6500).

A.2.6 Surface characterization

X-ray photoelectron spectroscopy (XPS) (Surface Science SSX-100) was employed to determine the surface chemical composition of the PLLA-PFPE-PLLA spun cast film.

A monochromatic Al Kα source with a spot size of 1 mm was applied at a take-off angle of 35°, while the pressure of the analysis chamber was maintained at 10-10 torr. Survey spectra (6 scans/sample, 0 – 1100 eV binding energy) were recorded at a rate of 1 eV/step and the data were processed using Hawk Data Analysis 7 software.

195

The wetting properties of the spun cast film and fiber mat of PLLA-PFPE-PLLA were evaluated by sessile drop measurements using a FAMAS Interface Measurement &

Analysis system (Kyowa, DM – CE1). The static contact angle (CA) was determined by placing a 5 µL water droplet onto the sample surface, and images were recorded with a

CCD camera. 3 – 5 measurements were taken with each sample.

A.3 Results and Discussion

A.3.1 Synthesis & Characterizations of PLLA-PFPE-PLLA

The solution polymerizations were not able to produce the designed block copolymer.

The solution became cloudy after about 20 minutes polymerization had occurred, which was probably due to the poor solubility of PFPE in both CHCl3 and CH2Cl2.

A white fibrous material was obtained from the bulk polymerization, as shown in

Figure A.2. 1H NMR and 19F NMR spectra analysis confirms the formation of PLLA-

PFPE-PLLA triblock copolymer, as shown in Figure A.3. The molecular weight and thermal properties are summarized in Table A.1.

Figure A.2 Photograph of the synthesized PLLA-PFPE-PLLA triblock copolymer.

196

(a)

(b)

Figure A.3 (a) 1H (b) 19F NMR characterizations of PLLA-PFPE-PLLA.

Table A.1 Molecular weight of PLLA-PFPE-PLLA

Targeted M NMR M Materials n n Ð T (°C) T (°C) (kg/mol) (kg/mol) g m PLLA-PFPE-PLLA 14 – 2 – 14 13.8 – 2 – 13.8 1.3 52 163

197

The dynamic elastic modulus (G’) and loss modulus (G’’) was first measured at

180 °C, as shown in Figure A.4. It is observed that G’ ~ ω1.7 and G’’ ~ ω, which indicates that the block copolymer is disordered at 180 °C, consistent with the optimal conditions for melt blowing. In order to investigate the viscosity of PLLA-PFPE-PLLA, the dynamic complex viscosity was measured at different temperatures, as demonstrated in Figure A.5.

Obviously the viscosity goes down as the temperature increases from 180 °C to 210 °C.

The viscosity at 210 °C is around 80 Pa·s which is in the optimal viscosity range for melt blowing. However, the block copolymer starts to degrade at 210 °C, especially after

20 minutes in the rheometer. This was confirmed by a decrease in the SEC measured molecular weight. Therefore, the optimum temperature for processing PLLA-PFPE-

PLLA is 200 °C.

Figure A.4 Dynamic elastic modulus (G’) and loss modulus (G’’) of PLLA-PFPE-PLLA.

198

Figure A.5 Dynamic complex viscosity of PLLA-PFPE-PLLA at 180, 190, 200, 210 °C.

A.3.2 Melt blown fibers

All of the melt blowing experiments produced fiber mats. However, for the cases when Tprocessing = 180 °C and Tprocessing = 190 °C, highly non-uniform fiber mats were generated containing mostly shots (i.e. a polymer globule with a diameter larger than the melt blown fibers) with a random occurrence of small diameter fibers, as seen in Figure

A.6 (A) and (B). When Tprocessing was increased to 200 °C, the fiber mat was partially non- uniform, but with relatively well formed fibers could be found although there were still some irregularities, as is shown in Figure A.6 (C) and (D); some irregular fibers are also visible. The reason why the quality of the fiber mats improved is that the dynamic viscosity of PLLA-PFPE-PLLA approaches the optimal viscosity range for melt blowing

(≤100 Pa·s)

199

Figure A.6 Representative SEM images of (A) PLLA-PFPE-PLLA fibers, Tprocessing = 180 °C , (B) irregularities at Tprocessing = 180 °C , (C) and (D) PLLA-PFPE-PLLA fibers at Tprocessing = 200 °C .

The melt blown fibers of PLLA-PFPE-PLLA are expected to be used in engine oil filtration, which requires the fiber mat to be structure-stable in engine oil, i.e. insoluble and non-shrink. In order to test the structural stability, a piece of melt blown fiber mat (2 cm × 2 cm) was soaked in ultra-low-sulfur diesel oil provided by Cummins at 100 °C for

24 hours. The results are shown below in Figure A.7.

200

Figure A.7 Volumetric comparison of a piece of melt blown PLLA-PFPE-PLLA fiber mat between (A) before-soaking and (B) after-soaking. The fiber mat was soaked in ultra-low-sulfur diesel at 100 °C for 24 hours

It was observed that the fiber mat shrank to almost half of the original size (1/4 area) after soaking. This shrinkage is attributed to cold crystallization, which decreases the volume of the fiber mat. Cold crystallization occurs because the PLLA-PFPE-PLLA fiber mat possessed only 24 % crystallinity rather than the value found in well-annealed specimens (Xc ~ 43 %), as demonstrated in Figure A.8. Therefore, the PLLA-PFPE-

PLLA is not appropriate for the application of water-in-diesel fuel filtration unless the crystallization kinetics can be accelerated.

201

Figure A.8 Differential scanning calorimety results for PLLA-PFPE-PLLA meltblown fibers. (a) cooling at 10 °C/min; (b) cooling at 50 °C/min.

A.3.3 Surface wetting properties

The spun-cast films were dried for 24 hours in vacuum. Then, water contact angle

202 measurements were performed. The results are shown in Figure A.9. The contact angle of spun-cast PLLA-PFPE-PLLA thin film is significantly greater than that of neat PLLA, which we attribute to the surface segregation of fluorine. XPS analysis demonstrates the blooming of fluorine on the film surface, as shown in Figure A.10.

Water contact angle on the PLLA-PFPE-PLLA fiber mat is about 143 ± 3°, which is larger than that on pure PLLA fiber mat (~ 115 ± 5°). The larger contact angle can be attributed to both the surface roughness and fluorine blooming.

(a) (b)

Figure A.9 Water contact angle (θ) on spun-cast films of : (a) PLLA, θPLLA = 76 ± 5° ; (b) PLLA-PFPE-

PLLA, θPLLA-PFPE-PLLA = 106 ± 3°

203

Figure A.10 XPS analysis of PLLA-PFPE-PLLA spun-cast film.

(a) (b)

Figure A.11 Water contact angle on (a) pure PLLA fiber mat, θPLLA = 115 ± 5°, (b) PLLA-PFPE-PLLA fiber mat, θPLLA-PFPE-PLLA = 143 ± 3°

204

A.4 Summary

We synthesized PLLA-PFPE-PLLA triblock copolymer through bulk polymerization.

The molecular weight, thermal properties, and rheological properties of the triblock copolymer were characterized. Rheological measurements determined the optimum processing temperature (200 °C) for the subsequent melt blowing based on the complex viscosity. The melt blown fibers have an average diameter of about 2 μm. The surface hydrophobicity of PLLA is enhanced by incorporation of PFPE, demonstrated by the contact angle measurement on spun-cast film of PLLA-PFPE-PLLA. More importantly, the surface hydrophobicity of the melt blown fiber mat is significantly increased.

However, cold-crystallization occurs to the resulting melt blown fibers, leading to the shrinkage of the fiber mat when soaking in diesel oil at 100 °C.

205

Appendix B Effects of Alkaline Hydrolysis on Morphology and Mechanical Properties of Melt Blown Poly(butylene terephthalate) Fibers

B.1 Introduction

We have studied tuning the surface properties of melt blown poly(butylene terephthalate) (PBT) fibers using alkaline hydrolysis and the subsequent fluorination. The alkaline hydrolysis can generate sponge-like surface topography, increasing the fiber surface roughness. However, we maintained the reaction time for 10 – 30 min in our previous studies, where the fiber surface topography is relatively insensitive to the reaction time. One intriguing question is: how does the fiber surface topography evolve if we extend the reaction time? Another important question is: how does the hydrolysis influence the mechanical properties of the fiber mats? This work aims at an understanding of these two questions. Cummins melt blown PBT fiber mats were hydrolyzed using NaOH at 45 for controlled reaction time. The fiber surface morphology and tensile properties of the fiber mats were studied.

B.2 Experimental

B.2.1 Alkaline hydrolysis of melt blown PBT fiber mats

Melt blown PBT fiber mats used in these experiments were obtained from Cummins

Filtration. The fiber mats were cut into small square pieces (1 x 1 inch2) and were soaked

206 in a methanol solution containing sodium hydroxide (NaOH) in a glass vial without stirring. We used methanol as the solvent because alcohols tend to accelerate hydrolysis of polyester.1 NaOH pellets (0.5 g) (Macron Fine Chemicals) were mixed with 4 mL of methanol (Fisher Scientific) and 4 mL deionized water (pH = 7) and preheated to 45 °C before soaking the PBT fiber mats. We maintained the hydrolysis reaction for different periods of time ranging from 1, 2, 3, 4, 5, 10, 20, 30, 40, 50, 60, to 70 min. After hydrolysis, the fiber mats were rinsed three times with distilled water and then immersed in HCl solution (0.1 mol/L), then washed again with distilled water until neutrality (pH =

7) of the rinse was achieved. Finally, the hydrolyzed PBT (h-PBT) fiber mats were dried in air for 24 hours, followed by vacuum drying overnight at 20 °C. All experiments were conducted three times.

B.2.2 Characterization of fibers and fiber mats

The square fiber mats were photographed and the thickness was measured using a digital caliper (Mitutoyo). The mass of each specimen was determined using an electric analytical balance (Denver Instrument M–220). Before each measurement, the fiber mats were vacuum dried overnight at 20 °C. The mass loss (Δm %) was calculated using the following equation,

mm m% 0  100% (B.1) m0

207 where m0 is the mass of the unhydrolyzed fiber mat, m is the mass of the h-PBT fiber mat.

The porosity (φ) and the fiber volume fraction (φf) were estimated from,

w  (1Az ) (B.2)  fiber

f 1 (B.3)

where φ is the porosity of the fiber mat, φf is the fiber volume fraction, w is the weight of the fiber mat, A and z are the area and thickness of the mat, and ρfiber is assumed to be the density of bulk PBT (1.4 g/cm3).

Fiber morphology was analyzed by a scanning electron microscopy (SEM) (Hitachi

S-4700). Prior to SEM analysis, the specimens were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. For each fiber mat, 25 – 35 SEM micrographs were taken and 400 – 500 fiber diameter measurements were made using

ImageJ software. Origin Lab software was employed to fit a log-normal function to the distribution of fiber diameters from which the average fiber diameter (dav) and the coefficient of variation (CV) were determined.

B.2.3 Tensile tests

The square h-PBT fiber mats were cut into strips (2 cm x 1 cm) for tensile tests. An

AR-G2 (TA Instruments) rheometer was used to characterize the tensile properties of the strips. The experimental setup is illustrated in Figure B.1. Double-side tapes were used to 208 glue the ends of the strips to prevent the sample from slipping. The initial grip-to-grip distance (gauge length) is 1 cm. The stretching speed was 0.01 mm/s (corresponding to strain rate of 1 x 10-3 s-1). 6 – 8 strips were measured for each group (every group experienced the same hydrolysis time).

Figure B.1 Schematic of the tensile test setup.

B.3 Results and Discussion

B.3.1 Physical characterizations of the mats

Details of the hydrolysis mechanism are discussed in Chapter 4. Within the range of 1

≤ t (etching time) ≤ 70 min, the mass loss (Δm %) of the h-PBT fiber mats increases linearly with the increase of the etching time, as shown in Figure B.2. Also, the integrity

209 of the PBT fiber mats seems to be insensitive to the hydrolysis, as shown in Figure B.3.

These observations are consistent with our previous studies. As the reaction time further increases, the fiber mat becomes thinner. Moreover, the dependence of the mass loss with the hydrolysis time deviates from the linear relationship, which we attribute to the sharp decrease of the fiber surface area. The mat loses the integrity at t = 140 min, as shown in

Figure B.2. The overall fiber mat properties are summarized in Table B.1.

Figure B.2 Effects of hydrolysis time on the mass loss of h-PBT fiber mats.

Figure B.3 Comparison in the length and width between h-PBT-50 and PBT fiber mats.

210

Table B.1 Effects of the hydrolysis on the physical properties of h-PBT fiber mats.

Average fiber Diameter Fiber volume Mass Loss Thickness Porosity Sample diameter distribution fraction (Δm %) (mm) φ (%) dav (µm) CV (%) φf (%) PBT - 0.217 ± 0.005 2.68 ± 0.54 57.6 88.0 12.0 h-PBT-1 - 0.218 ± 0.006 2.58 ± 0.50 53.3 88.1 11.9 h-PBT-2 - 0.216 ± 0.005 2.72 ± 0.45 47.3 88.0 12.0 h-PBT-3 0.57 ± 0.2 0.214 ± 0.004 2.80 ± 0.53 57.0 87.9 12.1 h-PBT-4 1.12 ± 0.5 0.214 ± 0.006 2.41 ± 0.47 50.0 88.0 12.0 h-PBT-5 1.5 ± 0.6 0.212 ± 0.005 2.36 ± 0.37 38.3 88.0 12.0 h-PBT-10 5.29 ± 1.5 0.210 ± 0.008 2.42± 0.46 48.0 88.0 12.0 h-PBT-20 14.33 ± 1.5 0.204 ± 0.012 2.27 ± 0.53 56.9 88.9 11.0 h-PBT-30 26.45 ± 1.1 0.185 ± 0.014 2.26 ± 0.55 59.4 90.1 9.9 h-PBT-40 30.70 ± 2.2 0.167 ± 0.009 2.00± 0.49 52.1 90.5 9.5 h-PBT-50 49.57 ± 3.0 0.136 ± 0.010 1.54 ± 0.63 70.9 90.0 10.0 h-PBT-60 54.07 ± 7.0 0.127 ± 0.007 1.30 ± 0.95 121 90.0 10.0 h-PBT-70 63.68 ± 4.1 0.122 ± 0.005 1.39 ± 0.86 105 91.3 8.7

B.3.2 Fiber diameter and surface topography

Reacting the melt blown PBT fiber mats with NaOH also leads to a decrease in fiber diameter and an increase in fiber diameter distribution (i.e. the coefficient of variation,

CV), which supports the hypothesis that the hydrolysis is a surface reaction. As shown in

Figure B.4, the average fiber diameter (dav) has been cut from about 2.8 µm to less than

1.4 µm as the hydrolysis proceeds, which is approximately 50% reduction within 60 min.

In the meanwhile, the CV considerably broadens over 100%, which is more dramatic compared with our previous studies.2 Such a broadening behavior of CV is potentially useful for filtration media to capture particles or droplets with a wide size distribution.

211

Figure B.4 Effect of NaOH hydrolysis on dav and CV.

Associated to the fiber diameter decrease, the fiber surface topography changes due to the surface hydrolysis reaction. As demonstrated in the representative SEM images in

Figure B.5, the relatively smooth surfaces of the unhydrolyzed PBT fibers are etched, resulting in textured, sponge-like surfaces. Such sponge-like surfaces, probably resulting 212 from the preferential hydrolysis of the amorphous and low-crystallinity regions, become more dramatic as the hydrolysis proceeds. We speculate that the detailed etching rate of the amorphous versus crystalline domains will strongly influence the development of the surface roughness and porosity by analogy with the trade-off between fundamental reaction kinetics and morphology, which contributes to the processes operation during the corrosion of metals.2-5 As the reaction time extends to 70 – 100 min, the fibers are more like etched ropes. More interestingly, we see shish-kebab-like fiber topography. More discussions are presented in Chapter 6.

213

Figure B.5 Representative SEM images of surface morphologies of (a) PBT (b) h-PBT_10 (c) h-PBT_30 (d) h-PBT_60 (e) h-PBT_70 (f) h-PBT_100. (h-PBT_x refers to the h-PBT fiber mat etched for x minutes)

B.3.3 Mechanical properties of h-PBT fiber mats

Figure B.6 shows a representative evolution of a fiber strip during a tensile test.

Figure B.6 Representative images showing the tensile test.

214

The representative tensile stress-strain curves of the h-PBT fiber mats are shown in

Figure B.7. Clearly the fiber mats becomes weaker as the reaction time increases. But within the range of 1 ≤ t ≤ 5 min the tensile strength (σTS) and strain at break (εb) of the h-

PBT fiber mats are approximately the same as the untreated PBT fiber mats. At t = 10 min, σTS and εb maintain about 85 % of the original value. We attribute this relatively weak dependence of the mechanical properties with the etching time to the surface reaction which does not impact the bulk properties of the fibers. Also, the slightly increased fiber surface roughness due to the hydrolysis may increase the fiber-fiber friction and thus the overall fiber mat mechanical properties. The fiber mats lose the mechanical properties sharply as the reaction time extends, which is likely because of the increase of the surface roughness which causes significant fiber surface defects.

Therefore, along with our studies on the surface wetting (Chapter 4) the mechanical studies here establish a process window to impart the melt blown PBT fiber mats superhydrophilicity without losing the mechanical properties.

215

Figure B.7 Representative engineering stress-strain curves of the h-PBT fiber mats.

B.4 Summary

We investigated the effects of the alkaline hydrolysis on the fiber surface topography and the mechanical properties of the fiber mats. The fiber mat maintains the structural integrity until the reaction extends to 140 min. The fiber surface topography evolves from sponge-like to etched-rope-like. More importantly, the overall mechanical properties of the h-PBT fiber mats maintain about 85 % of the original σTS and εb within the range of 1

≤ t ≤ 10 min. Then, the fiber mats lose σTS and εb sharply as the reaction time extends. A process window to impart the melt blown PBT fiber mats superhydrophilicity without losing the mechanical properties was established.

216

Appendix C Compatibilized Polyamide 6 and Poly(ethylene-alt- tetrafluoroethylene) Blends

C.1 Introduction

Poly(ethylene-alt-tetrafluoroethylene) (PETFE) is an alternating semi-crystalline copolymer. It is a melt-processable fluoropolymer having excellent weatherability, thermal and chemical stability. Commercially available PETFE copolymers are terpolymers consisting of ETFE unit copolymerized with the termonomer for improving the practical properties, especially the anti-stress cracking and tensile elongation at high temperature.6 More importantly, PETFE possesses low surface energy. The water contact angle on PETFE molded film is about 110°. Therefore, we studied the possibility of enhancing the surface hydrophobicity of polyamide 6 (PA6) by blending in PETFE, aiming at producing melt blown superhydrophobic PA6 fiber mats.

C.2 Experimental

C.2.1 Materials

Polyamide 6 pellets (PA6) were purchased from Sigma Aldrich. PETFE and maleic anhydride grafted PETFE (PETFE-g-MA) pellets were obtained from DuPont (US).

Thermal stability of the materials was evaluated with thermogravimetric analysis (TGA), as shown in Figure C.1. The thermal properties of the resulting block copolymers were 217 characterized by differential scanning calorimetry (DSC) (TA Instrument Q1000). The samples were analyzed under a standard run sequence of heating-cooling-heating from -

20 °C to 250 °C at a scan rate of 10 °C/min, and Tg and Tm were analyzed in the second heating run.

Figure C.1 TGA analysis of the mateirals.

Rheological properties were measured with a strain-controlled ARES rheometer (TA

Instrument) equipped with a 25 mm parallel-plate fixture. Firstly, dynamic strain sweeps were performed to determine the linear viscoelastic region. Secondly, dynamic frequency sweeps were employed to measure the complex viscosity (η*) and moduli (G’, G”) as a function of frequency (1 ≤ ω ≤ 100 rad/s) at 270 °C, as shown in Figure C.2.

218

(a)

(b)

Figure C.2 (a) Complex viscosity and (b) dynamic elastic modulus (G’) and loss modulus (G’’) of PA6, PETFE, PETFE-g-MA. The measurements were carried out at 270 °C.

219

C.2.2 Preparation of polymer blends

The materials were dried for 12 hours under vacuum at 100 °C before using. Then,

PA6 was mixed with PETFE and PETFE-g-MA at room temperature as 90/10, 80/20,

70/30, and 50/50 wt%, respectively. The mixed pellets were fed into a HAAKE (Thermo

Scientific) batch compounder at 270 °C with roller blades rotating at 100 rpm. The blending was maintained for 8 minutes before the resulting blend was quenched with liquid nitrogen (LN2) to freeze the blend morphology.

C.2.3 Characterization of polymer blend morphology

Polymer blends were cryo-fractured in LN2. The resulting sample surfaces were coated with gold/palladium for 30 seconds using a Denton DV – 502 sputter coater. A scanning electron microscope (SEM) (Hitachi S-4700 or JEOL 6500) was used to image the sample surface. Both secondary and backscattered detectors were employed. 20 – 25

SEM micrographs were taken for each sample.

C.2.4 Surface wetting characterization

The resulting PA6/PETFE-g-MA blends were compression molded into 25 mm- diameter discs (~ 1 mm) at 270 °C. The surface wetting of the films was evaluated by sessile drop measurements using a FAMAS Interface Measurement & Analysis system

(Kyowa, DM – CE1). The static contact angle (CA) was determined by placing a 5 µL water droplet onto the sample surface, and images were recorded with a CCD camera. 3 –

5 measurements were taken with each sample.

220

C.3 Results & Discussion

C.3.1 Polymer blends

PA6 forms immiscible polymer blends with PETFE. As shown in Figure C.3, PETFE appears approximately as spherical droplets dispersed in a PA6 matrix. There are sharp gaps/cavities at the interfaces, implying poor interfacial adhesion. As the weight fraction of PETFE increases, the droplet size increases from about 1.5 μm to 2.7 μm because of coalescence.

Figure C.3 Representative SEM images showing the morphology of (a) PA6/PETFE_90/10 wt%, (b) PA6/PETFE_80/20 wt%, (c) PA6/PETFE_70/30 wt%, (d) PA6/PETFE_50/50 wt%.

221

In order to improve the compatibility between PA6 and PETFE, we used PETFE-g-

MA. The amine end group of PA6 is expected to react with the functional group of maleic anhydride in PETFE-g-MA, forming compatibilized blends. The reaction is schematic in Figure C.4. As shown in Figure C.5, clearly the droplet size decreases significantly to less than 1 μm due to the compatibilization. Also, the droplets are bridged by nanoscale fibrils with the PA6 matrix, as shown in Figure C.6, leading to enhanced interfacial adhesion.

Figure C.4 Schematic of the compatibilization between

222

Figure C.5 Representative SEM images showing the morphology of (a) PA6/PETFE-g-MA_90/10 wt%, (b) PA6/PETFE-g-MA_80/20 wt%, (c) PA6/PETFE-g-MA _70/30 wt%, (d) PA6/PETFE-g-MA _50/50 wt%.

Figure C.6 Representative SEM image showing the morphology of compatibilized PETFE-g-MA droplet in a PA6 matrix. Nano fibrils bridge the interfaces.

223

C.3.2 Surface wetting

Water contact angles were measured on the molded films of PA6/PETFE-g-MA blends to characterize the surface wetting properties of the blends. As shown in Figure

C.7, the contact angle increases with the increase in the blended PETFE-g-MA, indicating enhanced surface hydrophobicity. More details of the blooming of surface fluorine in binary polymer blends are discussed in Chapter 5. Unfortunately, the blends were not able to be melt blown because of the high viscosity of PETFE and PETFE-g-

MA.

Figure C.7 Effect of PETFE-g-MA on the surface wetting of PA6 molded films.

224

C.4 Summary

We studied the binary polymer blends of PA6 and PETFE. PETFE forms immiscible blends with PA6. By replacing PETFE with PETFE-g-MA we obtained compatibilized polymer blends. SEM analysis demonstrated the compatibilization between PETFE-g-

MA and PA6, as evidenced decreased droplet size and fibril-bridged interfaces. Water contact angle measurement showed that the surface hydrophobicity of PA6 was enhanced with the increase of the blended PETFE-g-MA. However, the blends cannot be melt blown to produce uniform fibers due to the high viscosity of PETFE and PETFE-g-MA.

225

Appendix D Collection of Individual Melt Blown Fibers using A Rotating Frame

There have been many studies on melt blown fibers, ranging from the fundamental fiber formation to functional applications like filtration media. Most studies primarily focus on melt blown fiber mats. There has been no publication associated with the fundamental studies of individual melt blown fibers, which is likely because it is challenging to achieve individual melt blown fibers, especially from the fabricated fiber mats. Fibers entangle physically with each other during melt blowing.

We proposed a method to achieve individual melt blown fibers from the fiber blowing line. As shown in the schematic in Figure D.1, by using a manually or motor- driven rotating paper frame instead of the conventional drum collector we can obtain floating individual fibers on the frame. The density of the floating fibers can be adjusted by tuning the collection time and frame rotating speed. Then, individual fibers can be picked out from the floating fibers on the frame. These selected individual fibers can be readily used for fundamental studies of individual melt blown fibers, such as mechanical properties, surface wetting of individual fibers.

226

Figure D.1 Schematic of the collection of individual melt blown fibers using a rotating frame.

227