Modification and Characterization of Nafion Perfluorinated

Home , Ionomer, Nafion

MODIFICATION AND CHARACTERIZATION OF NAFION PERFLUORINATED

IONOMER MEMBRANE FOR POLYMER ELECTROLYTE FUEL CELLS

A Dissertation

Presented to

The Graduate Faculty of The University of Akron

In Partial Fulfillment

of the Requirements for the Degree

Doctor of Philosophy

Nadzrinahamin Ahmad Nazir

August, 2011

MODIFICATION AND CHARACTERIZATION OF NAFION PERFLUORINATED

IONOMER MEMBRANE FOR POLYMER ELECTROLYTE FUEL CELLS

Nadzrinahamin Ahmad Nazir

Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Thein Kyu Dr. Sadhan C. Jana

______Committee Member Dean of the College Dr. Kevin Cavicchi Dr. Stephen Z. D. Cheng

______Committee Member Dean of the Graduate School Dr. Li Jia Dr. George R. Newkome

______Committee Member Date Dr. Alamgir Karim

______Committee Member Dr. Wiley Youngs

ABSTRACT

Nafion perfluorinated ionomer membrane has been regarded as a benchmark material in proton electrolyte membrane fuel cells (PEMFC) due to its excellent perm selectivity properties. However, the excessive swelling and low operating temperature window (up to 80 oC) of Nafion marked its vital shortcomings in the proton fuel cell

operations.

This dissertation is focused on the modification of hydrophobic fluorocarbon

matrix and hydrophilic ionic domains/clusters of Nafion to alleviate the aforementioned

shortcomings. The first modification is via solution blending of Nafion with copolymer

of poly(vinylidenefluoride-trifluoroethylene) or PVDF-TrFE, which modified the

fluorocarbon matrix of Nafion, while the second is via simple in–situ impregnation of

Nafion with several kinds of functionalized supramolecules, which locally alters the

properties of the ionic domains/clusters.

Nafion/PVDF-TrFE blends revealed an hourglass type phase diagram, consisting of single phase crystal (Cr1), and crystal + liquid (Cr1 + L2) and liquid + liquid (L1 + L2) coexistence regions. Blends of Nafion with PVDF-TrFE demonstrated swelling reduction upon hydration as confirmed by Fourier transform infrared (FTIR) spectroscopy and water uptake measurements. The 60/40 Nafion/PVDF-TrFE blend with bicontinuous

morphology exhibits both capacitor and proton conductivity properties. This blend was

iii

found to have lower proton conductivity with a decreasing trend upon increasing

temperature.

Supramolecules of photocurable hyperbranched polyester (HBPEAc-COOH)

were used in the first part of in-situ impregnation attempt. Of particular importance is that

the present study is the first to successfully incorporate polymer molecules/networks into

the Nafion ionic domains by means of impregnation with hyperbranched supramolecules

followed by photopolymerization in-situ. HB impregnated Nafion membranes were found to render swelling suppression as well as improved mechanical stability. This

impregnated membrane exhibited an increasing trend of proton conductivity with

increasing temperature, which eventually surpassed that of neat Nafion above 100 oC.

Two unique supramolecules terminated with hydroxyl (Noria) and tert-

butyloxycarbonyl (Noria-BOC), which resembles a waterwheel, were used in the second

impregnation attempt. We anticipated that these waterwheel supramolecules, Noria in

particular, will have potential utility as the “solid proton carrier” in proton fuel cells to

substitute the role of water at high temperatures. Noria and Noria-BOC impregnated

membranes exhibited excellent swelling suppression and the mechanical stability of both

impregnated membranes increased beyond the existing neat Nafion. Only Noria

impregnated membrane shows improved proton conductivity at elevated temperature

whereas Noria-BOC impregnated membrane revealed a plummet in the proton

conductivity at high temperatures similar to that of neat Nafion.

DEDICATION

To my parents, my family and my fiancé for your undivided support and unselfish

patience during this critical period.

ACKNOWLEDGEMENTS

This dissertation symbolizes my long standing goal to earn the PhD. scroll, for which there are many people to whom I owe thanks. Foremost, I would like to extend many thanks to my advisor, Dr. Thein Kyu for all the help, support, encouragement as well as positive criticisms throughout the years.

My heartfelt gratitude also goes to the committee members, Dr. Kevin Cavicchi,

Dr. Li Jia, Dr. Alamgir Karim and Dr. Wiley Youngs for all the suggestions given during the course of this study. Also, to the previous committee member Dr. Alexei Sokolov, thank you for your guidance and ideas during the PhD. proposal presentation.

To Dr. Matthew Espe, Dr. Antal Jákli and Dr. Robert Weiss, I really appreciate your help on the solid state NMR (SSNMR), AC electrical impedance and fuel test measurements. Thank you for the time, guidance and assistance in order to complete the study. This study would not be a success without the materials that specially synthesized by our Japanese collaborator, Dr. Takashi Nishikubo and Dr. Hiroto Kudo. I am truly grateful for their assistance and advice.

I would like to thank the United States Government, whose Fulbright Scholarship allowed me to pursue and mark my journey in the graduate studies. Special thanks also to the Malaysian Palm Oil Board (MPOB) for sponsoring a portion of my studies via the

MPOB Education Foundation Scholarship. Thank you for believing in me.

To my groupmates, especially Dr. Chandrasekaran, Dr. Kim, Dr. Huang, and Tom

Sutter, I really appreciate all the help, the constructive evening discussions and brilliant

ideas that have been poured out for me to make this study a success. To my good friend,

Emmanuel Pitia, I am lucky to know someone like you and wish a prosperous future for

you. I cherished all the memories we had together and definitely I will not forget you. To

Dr. Alyamać, I really appreciate your help.

My parents, Haji Ahmad Nazir Mohd Baki and Hajjah Haminah Mat Piah, words cannot describe the sacrifice, support and encouragement that you have given me all these years in pursuing my dreams. I am indebted to both of you and nothing can repay what you have done for me. To my siblings, Mrs. Nadzmin, Mr. Nadzham, Dr. Nadzrin and Mrs. Nadzila, I am so lucky to have you as my siblings and though we were apart for several years, distance has taught me that there is nothing in the world that can replace family. I also owe many thanks to my Malaysian friends, acquaintances at Lehigh

University and also my future in-laws, Mr. and Mrs. Lewis, Mike, Nonnie and all the

Lewis’ family who have been there for me and treating me like their own family.

Last, but not least, to my beloved fiancé, Mr. Stephen Gerard Lewis. Thank you for putting up with my crankiness, grumpiness and my ups and downs for all these years we have been together. You are my soulmate.

vii

TABLE OF CONTENTS

Page

LIST OF TABLES ...... xiv

LIST OF FIGURES ...... xv

CHAPTER

I. INTRODUCTION ...... 1

II. BACKGROUND AND LITERATURE REVIEW ...... 7

2.1. Introduction to Fuel Cells ...... 7

2.1.1. Type of Fuel Cells ...... 9

2.2. Polymer Electrolyte Membrane Fuel Cells (PEMFCs) ...... 13

2.2.1. Principle Operation of PEM Fuel Cells ...... 16

2.2.2. Assembly of PEM Fuel Cells ...... 19

2.2.2.1. Electrodes (Anode and Cathode) ...... 19

2.2.2.2. Bipolar Plates ...... 20

2.2.2.3. Proton Electrolyte Membrane ...... 21

2.3. Perfluorinated Ionomer Membrane, Nafion ...... 21

2.3.1. Morphology of Nafion ...... 23

2.3.2. Hydration Effect of Nafion ...... 26

viii

2.3.3. Proton Conductivity and Proton Transport Mechanism of Nafion ...... 30

2.3.4. Mechanical Properties of Nafion ...... 35

2.4. Recent Advances of Ionomer Blends and Nafion Blends for PEM Fuel Cell Application ...... 38

2.5. In-situ Impregnation Approach of Nafion Membrane for PEM Fuel Cell Applications ...... 41

2.6. Promising Applications of Nafion ...... 44

2.7. Homopolymer of Poly(vinylidene fluoride), PVDF ...... 45

2.7.1. Copolymer Poly(vinylidene fluoride-co- trifluoroethylene), PVDF-TrFE ...... 47

2.8. Hyperbranched Polymers/Supramolecules ...... 50

2.8.1. Photoinitiated Polymerization of Supramolecules ...... 51

2.9. Polymer Soultions and Blends ...... 53

2.9.1. Flory-Huggins Theory for Polymer Solutions ...... 53

2.9.2. Phase Field Theory for Crystal Solidification ...... 56

2.9.3. Phase Equilibrium ...... 57

2.9.4. Dynamics of Phase Separation Process ...... 61

2.9.4.1. Nucleation and Growth (NG) Mechanism ...... 61

2.9.4.2. Spinodal Decomposition (SD) Mechanism ...... 63

2.10. Fundamentals of the Impedance Spectroscopy ...... 64

2.10.1. Presentation and Interpretation of Impedance Data ...... 67

III. MATERIAL SPECIFICATIONS AND EXPERIMENTAL TECHNIQUES ...... 70

3.1. Materials ...... 70

3.1.1. Perfluorinated Ionomer Membrane, Nafion ...... 70

3.1.2. Copolymer Poly(vinylidene fluoride-trifluoroethylene), PVDF-TrFE...... 72

3.1.3. Photocurable Hyperbranched Polyester with Pendant Carboxylic Acid Functional Groups, HBPEAc-COOH...... 73

3.1.4. Waterwheel Supramolecules with Terminal Hydroxyl Function Groups, Noria and Tert-butyloxycarbonyl, Noria-BOC ...... 74

3.2. Experimental Methods ...... 81

3.2.1. Thermo Gravimetric Analysis (TGA) ...... 82

3.2.2. Differential Scanning Calorimetry (DSC) ...... 83

3.2.3. Polarized Optical Microscopy (POM) ...... 83

3.2.4. Fourier Transform Infrared (FTIR) Spectroscopy ...... 84

3.2.5. Solid State Nuclear Magnetic Resonance (SSNMR) ...... 85

3.2.6. Wide Angle X-Ray Diffraction (WAXD) ...... 85

3.2.7. Small Angle X-Ray Scattering (SAXS) ...... 86

3.2.8. Water Uptake ...... 86

3.2.9. Ion Exchange Capacity (IEC) ...... 87

3.2.10. Dynamic Mechanical Analyses (DMA) ...... 88

3.3. Conductivity Properties ...... 88

3.3.1. Proton Conductivity/Fuel Cell Measurements ...... 88

3.3.2. AC Electrical Impedance Measurements ...... 92

IV. BLENDS OF NAFION AND POLY(VINYLIDENE FLUORIDE- TRIFLUOROETHYLENE), PVDF-TrFE COPOLYMER ...... 93

4.1. Introduction ...... 93

4.2. Sample Preparation and Experimental Methods ...... 95

4.3. Theoretical Scheme for Establishment of Phase Diagram ...... 97

4.4. Results and Discussion ...... 98

4.4.1. Phase Transition ...... 98

4.4.2. Binary Phase Diagram ...... 100

4.4.3. Water Uptake and Hydration Effect on the Nafion/PVDF- TrFE Blends ...... 103

4.4.4. Electrical Properties of Nafion/PVDF-TrFE Blends ...... 106

4.4.5. Proton Conductivity Properties of Nafion/PVDF-TrFE Blends ...... 114

4.4. Conclusions ...... 120

V. IN-SITU IMPREGNATION OF NAFION WITH NOVEL PHOTOCURABLE HYPERBRANCHED POLYESTER, HBPEAc-COOH ...... 121

5.1. Introduction ...... 121

5.2. Experimental Section ...... 124

5.2.1. Materials and Sample Preparation ...... 124

5.2.2. Experimental Methods ...... 125

5.3. Results and Discussion ...... 127

5.3.1. Characterization of Neat Nafion and Pure HBPEAc- COOH ...... 127

5.3.2. Structural Characterization of HBPEAc-COOH Impregnated Membranes ...... 130 xi

5.3.3. Water Uptake and Ion Exchange Capacity (IEC) of HBPEAc-COOH Impregnated Membranes ...... 138

5.3.4. Dynamic Mechanical Properties of HBPEAc-COOH Impregnated Membranes ...... 141

5.3.5. Proton Conductivity of HBPEAc-COOH Impregnated Membranes ...... 143

5.4. Conclusions ...... 151

VI. IN-SITU IMPREGNATION OF NAFION WITH NOVEL WATERWHEEL SUPRAMOLECULES HAVING TERMINAL HYDROXYL FUNCTIONAL GROUPS, NORIA ...... 153

6.1. Introduction ...... 153

6.2. Experimental Section ...... 155

6.2.1. Materials and Sample Preparation ...... 155

6.2.2. Experimental Methods ...... 157

6.3. Results and Discussion ...... 159

6.3.1. Thermal and Mechanical Stability of Noria Impregnated Membranes ...... 159

6.3.2. Structural Characterization of Noria Impregnated Membranes ...... 162

6.3.3. Water Uptake and Ion exchange Capacity (IEC) of Noria Impreganted Membranes ...... 169

6.3.4. Proton Conductivity of Noria Impregnated Membranes ...... 171

6.4. Conclusions ...... 179

xii

VII. IN-SITU IMPREGNATION OF NAFION WITH NOVEL WATERWHEEL SUPRAMOLECULES HAVING TERMINAL TERT-BUTYLOXYCARBONYL FUNCTIONAL GROUPS, NORIA-BOC ...... 181

7.1. Introduction ...... 181

7.2. Experimental Section ...... 182

7.2.1. Materials and Sample Preparation ...... 182

7.2.2. Experimental Methods ...... 182

7.3. Results and Discussion ...... 184

7.3.1. Characterization of Neat Noria-BOC ...... 184

7.3.2. Structural Characterization of Noria-BOC Impregnated Membranes ...... 186

7.3.3. Water Uptake and Ion Exchange Capacity (IEC) of Noria Impregnated Membranes ...... 190

7.3.4. Dynamic Mechanical Properties of Noria-BOC Impregnated Membranes ...... 192

7.3.5. Proton Conductivity of Noria-BOC Impregnated Membranes ...... 195

7.4. Conclusions ...... 196

VIII. OVERALL SUMMARY AND RECOMMENDATIONS ...... 198

8.1. Overall Summary ...... 198

8.2. Recommendations ...... 201

BIBLIOGRAPHY ...... 203

xiii

LIST OF TABLES

Table Page

2.1 Summarization of major fuel cells, classified by types of electrolyte ...... 12

3.1 Physical properties of perfluorinated ionomer membrane, Nafion used in the present study ...... 77

3.2 Physical properties of poly(vinylidene fluoride-trifluoroethylene), PVDF- TrFE used in the present study ...... 78

3.3 Physical properties of hyperbranched polyester, HBPEAc-COOH used in the present study ...... 79

3.4 Physical properties of waterwheel supramolecules, Noria and Noria-BOC used in the present study ...... 80

4.1 Table of electron conductivity values of hydrated Nafion with increasing of temperature that signifies the plummet of its electrical conductivity ...... 111

xiv

LIST OF FIGURES

Figure Page

2.1 Simplistic view of PEMFC ...... 8

2.2 Schematic diagram of single PEM fuel cell ...... 15

2.3 Schematic diagram of basic principle operation of PEMFC ...... 15

2.4 Chemical structure of Nafion ...... 22

2.5 Small angle x-ray scans of hydrolyzed Nafion Na+ with different equivalent weight, swollen in water ...... 25

2.6 Yeagers's Three Phase Model and Gierke's Cluster Model ...... 26

2.7 AFM micrographs of Nafion K+ at dry state (left) and hydrated state (right)...... 28

2.8 Hydration effect on ionic clusters of Nafion manifested by Gierke ...... 28

2.9 FTIR spectrum of Nafion membrane with varies hydration level ...... 29

2.10 Conductivity of Nafion 117 as a function of hydration level ...... 32

2.11 Simplified schematic diagram of proton transport mechanism involving (a)Grotthuss mechanism or ion hopping and (b)vehicular mechanism ...... 34

2.12 Dynamic mechanical analyses (DMA) of Nafion-Cs with varying degrees of neutralization at 1 Hz ...... 36

2.13 Cloud point phase diagram of Nafion/PVDF blends at heating rate of 0.2 oC/min signifying the LCST above the melting temperature of PVDF ...... 41

2.14 Artificial muscles applications ...... 44

2.15 Chemical structure of PVDF homopolymer ...... 46 xv

2.16 Schematic representation of three conformation phases of PVDF ...... 46

2.17 Chemical structure of copolymer poly(vinylidene fluoride- trifluoroethylene), PVDF-TrFE ...... 48

2.18 Phase behavior of PVDF-TrFE as a function of VDF content ...... 49

2.19 Schematic diagram of dendrimer and dendron...... 50

2.20 Lattice model where open circles represent solvent molecules and filled represent solute molecules ...... 53

2.21 Lattice model for a polymer solution ...... 54

2.22 Single well potential...... 58

2.23 (a) Double well potential and (b) binary phase diagram for a polymer solution...... 60

2.24 Phase separation process regimes ...... 62

2.25 Nucleation and growth (NG) mechanism ...... 63

2.26 Spinodal decomposition (SD) mechanism ...... 64

2.27 Two-dimensional plane of absolute complex impedance, |Z|...... 66

2.28 Nyquist (or Cole-Cole) and Bode plot of pure resistor, pure capacitor and typical electrochemical cell containing a resistor in series with a parallel capacitor and resistor respectively...... 69

3.1 Chemical structure of Nafion...... 71

3.2 Chemical structure of poly(vinylidene fluoride-trifluoroethylene) PVDF- TrFE ...... 72

3.3 Chemical structure of hyperbranched polyester, HBPEAc-COOH ...... 74

3.4 Chemical structure of photoinitiator, Irgacure 907 ...... 74

xvi

3.5 (a) Chemical structure waterwheel supramolecules, Noria, which resembles the waterwheel. Each ring consists of alternating resorcinol and methylene units and the two rings are connected through six resorcinol units. (b) Chemical structure of resorcinol and 1,5-pentanedial and its representative symbols as shown in (a). (c) Chemical structure of waterwheel supramolecules with tert-butyloxycarbonyl, Noria-BOC ...... 76

3.6 Flow diagram of the AC impedance fuel cell setup...... 90

3.7 Polarization curves of 95/5 Noria/Nafion impregnated membranes at 74% RH obtained from the present experimental findings signifying the three regions of voltage loss; i) voltage loss due to the activation recation resistivity, ii) voltage loss attributed to the ohmic polarization and iii) voltage loss due to the concentration polarization...... 91

3.8 Typical Nyquist (or Cole-Cole) plot, the equivalent circuit and the manifestation of Rmax, Rp and fmax used to calculate the polarization capacitance, Cp...... 92

4.1 DSC thermograms for Nafion/PVDFTrFE blends demonstrating consistent Curie transition temperature (TCurie) and little or no movement of the melting temperature (Tm) of the entire blends composition range with 10 wt % increment of Nafion...... 99

4.2 Optical micrographs of Nafion/PVDFTrFE in the extreme and intermediate compositions observed under polarized mode and parallel mode suggesting presence of isotropic phase at high temperature for the extreme composition and presence of liquid–liquid phase separated structure for the intermediate composition of the blends ...... 100

4.3 The theoretical binodal (solid line) and spinodal (dash line) curves of Nafion/PVDF-TrFE blends which describe well the experimental crustal melting transition results obtained from DSC (as denoted by ‘□’) and OM (as denoted by ‘○’ for isotropic phase and ‘●’ for liquid-liquid phase separated region) ...... 102

4.4 Plot of (a) water uptake in weight increment (%) of Nafion/PVDF-TrFE blends while insets are pictures of the actual hydrated membrane demonstrating reduction of warping phenomenon as the PVDF-TrFE composition increased, (b) normalized FTIR absorbance based on the O–H stretching band as a function of Nafion. The insets show the morphology of the extreme and intermediate blend compositions which may be used to explain the hydration effect. Both plots signify the remarkable swelling suppression of the blends towards aprotic liquid, i.e., water ...... 104

xvii

4.5 (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of pure PVDF-TrFE from 43 oC to 150 oC over frequency range of 1 Hz–100 kHz...... 107

4.6 Cole-Cole plot and curve fitting plot of pure PVDF-TrFE copolymer with temperature obtained from the AC electrical impedance measurement illustrating the diminishing capacitor behavior above the Curie transition temperature...... 107

4.7 Plot of (a) capacitance and (b) electron conductivity versus temperature of pure PVDF-TrFE demonstrate the polarization capacitance, Cp and electron conductivity with increasing temperature...... 108

4.8 (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of hydrated Nafion from 23 oC to 125 oC over a frequency range of 1 Hz–100 kHz...... 110

4.9 Cole-Cole plot and curve fitting plot of hydrated Nafion with increasing of temperature obtained from the AC electrical impedance measurement illustrating the conductor-like behavior...... 111

4.10 (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of 60/40 Nafion/PVDF-TrFE from 43 oC to 125 oC over frequency range of 1 Hz–100 kHz...... 113

4.11 The Cole-Cole plot of 60/40 Nafion/PVDF-TrFE blend collected from the AC electrical impedance measurements shows the occurrence on the capacitor behavior of the blend...... 113

4.12 Plot of polarization capacitance, Cp versus temperature for 60/40 Nafion/PVDF-TrFE shows an increasing of capacitance values upon increasing of the temperature...... 114

4.13 (a) Log-log plot of storage impedance versus frequency and (b) semi-log plot of loss tangent versus frequency of neat Nafion at 100% RH...... 115

4.14 Cole-Cole plot of (a) pure Nafion and (b) 60/40 Nafion/PVDF-TrFE blend measured in the AC impedance fuel cell environment at 100 % RH. With increasing of temperature, polarization resistance decreases but exceeding the boiling temperature of water, and polarization resistance within the cell increases denoting loss of proton conductivity efficiency...... 115

xviii

4.15 Proton conductivity plot as a function of temperature of neat Nafion and 60/40 Nafion/PVDF-TrFE. It is apparent that the proton conductivity trends of the 60/40 blend resemble that of neat Nafion ...... 118

4.16 Schematic diagrams illustrating the (a) sea-and-island morphology signifying the dispersed and discontinuous structure of 90/10 blend with Nafion-rich region representing the matrix, (b) cococontinous morphology exhibited by 60/40 blend where Nafion and PVDF-TrFE percolated pathways coexist side by side, (c) inverted sea-and-island morphology which may be seen in the 10/90 blend with matrix of PVDF-TrFE and (d) an array of cylindrical channels which will favor enhancement of protonic and/or ionic conductivity within the blends ...... 119

4.17 Surface morphology of Teflon grid supported Nafion membrane developed by Gore Associate Inc ...... 119

5.1 DSC thermograms of pure HBPEAc-COOH (a) before and (b) after photocuring exhibiting an increase in Tg of the HBPEAc-COOH supramolecules upon photocuring ...... 129

5.2 FTIR spectra of (a) pure HBPEAc-COOH without initiator; (b) pure HBPEAc-COOH with initiator before curing and (c) pure HBPEAc- COOH after photocuring demonstrating the reduction of acrylate C=C stretching peak at 1630 cm-1 ...... 129

5.3 125.6 MHz 1H-13C CP/MAS SSNMR spectra of (a) HBPEAc-COOH impregnated Nafion and (b) cured neat HBPEAc-COOH. The spectra were collected using 1H decoupling so that the peaks in the 13C NMR spectrum only arise from the HBPEAc-COOH and the photoinitiator. Peaks in (b) labeled with * arise from the photoinitiator ...... 132

5.4 125.6 MHz 19F-13C CP/MAS SSNMR spectra of (a) impregnated membrane and (b) neat Nafion ...... 133

5.5 499.5 MHz 1H DP/MAS SSNMR spectra of (a) impregnated membrane, (b) pure HBPEAc-COOH and (c) pure Nafion. The single narrow line at the chemical shift of 8.6 ppm in (c) corresponds to 2 H2O per SO3H. The peak in (a) labeled with * arises from the photoinitiator ...... 133

xix

5.6 (a) 3700-2500 cm-1 region of FTIR spectra shows the slight shift of the broad O–H stretching band of impregnated membrane to lower wavenumber with increasing of HB polyesters; (b) the S-O stretching band of impregnated illustrates a systematic shift of the S-O band to a lower wavenumber (i.e., about 9 cm-1) suggesting inter-species hydrogen bonding interaction between the carboxylic acid of HBPEAc-COOH and the sulfonic acid group of Nafion ...... 134

5.7 (a) Plot of wavenumber of the S-O stretching band of FTIR as a function of HBPEAc-COOH in feed demonstrating a considerable shift of the S-O band to a lower wavenumber (i.e., about 9 cm-1) indicating inter-species hydrogen bonding interaction of the carboxylic acid of HBPEAc-COOH with the sulfonic acid terminal group of Nafion; (b) quantification of the estimated amount of HBPEAc-COOH present in Nafion membrane based on the integrated area under the curve of the FTIR spectra from C=O and C=C aromatic stretching band showing ~7% of HBPEAc-COOH incorporated in the impregnation process. The dashed line represents the actual amount of HB incorporated in the ideal situation ...... 138

5.8 (a) Plot of water uptake as a function of soaking time for both neat and impregnated membranes at room temperature. The water uptake of neat Nafion in acid form was 37wt% while the impregnated Nafion membranes show a remarkable reduction in water uptake; (b) graph of λ (moles of - H2O per mole of SO3 ) of impregnated Nafion membranes as a function of HBPEAc-COOH composition in feed (after soaking for 24 h), illustrating significant reduction of water molecules due to impregnation. The IEC curve shows an increment trend (i.e., proton storage capacity) with HBPEAc-COOH loading. Insets show photograph of neat Nafion membrane (10 mm in width and 22 mm in length) exhibiting severe warpage upon hydration and significantly less warpage in the 10% HBPEAc-COOH impregnated Nafion membrane demonstrating improved dimensional stability ...... 140

5.9 Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and 10% HBPEAc-COOH impregnated Nafion membrane (dashed line). An additional relaxation peak, attributable to the Tg of the cured HBPEAc-COOH, appears in the impregnated membrane at around ~90 oC. The inset photographs show improved thermal stability of the impregnated membrane showing light beige as compared to the dark brownish color of neat Nafion-acid after DMA experiments at 150 oC. The sample dimension was 10 mm in width x 25 mm in length ...... 143

5.10 Plots of (a) real impedance versus log frequency and (b) imaginary impedance versus log frequency of 90/10 Nafion/HBPEAc-COOH at 74% RH measured from 30 oC to 100 oC with 10 oC increment and (c) Cole- Cole plots of 90/10 Nafion/HBPEAc-COOH impregnated membranes generated at 74% RH. The diameter of semi-circular plots decreases with temperature indicates lowering of the overall cell resistance thus leading to better proton conductivity properties ...... 145

5.11 Proton conductivity plots of neat Nafion and impregnated membrane as a function of temperature at 74% RH. Proton conductivity of neat Nafion plummets beyond 80 oC while the conductivity value of the impregnated membrane continues to increase with temperature. The impregnated membrane surpassed the conductivity of that neat Nafion at 110 and 115 oC indicates the improved thermal stability and proton conduction of the impregnated membrane ...... 147

5.12 Heating and cooling cyclic measurement of neat Nafion and impregnated membrane at 100 %RH demonstrates the subtle proton conductivity hysteresis of neat Nafion after 160 hours of testing. However the HB impregnated membrane tends to show relatively consistent proton conductivity after all 5 cyclic measurements ...... 150

5.13 Schematic drawing of the ionic domain of the Nafion membrane before and after impregnation with HBPEAc-COOH showing the isolated cured HB polyester network from the fluorocarbon matrix ...... 150

6.1 Chemical structural scheme of Noria which consists the components of resorcinol and 1,5-pentanedial and resembles the waterwheel ...... 156

6.2 Impregnation process of 95/5 Nafion/Noria where after 24 hours it is evident that the supramolecules has been infused completely within Nafion membrane ...... 156

6.3 (a) TGA thermograms of neat Nafion, impregnated membranes and pure Noria signifying an excellent thermal stability exhibited by the impregnated membranes while (b) DSC scan of pure Noria illustrates the melting temperature at 373 oC which coincided with the temperature at 5% weight loss occurs ...... 159

xxi

6.4 Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and Noria impregnated Nafion membranes (dashed and dotted lines). The inset photographs show membrane samples taken after DMA experiments at 150 oC. The brownish appearance of the neat Nafion-acid is the manifestation of thermal instability of the protonated cluster whereas the Noria impregnated membrane evidently sustained its inherent color, i.e., the color of the supramolecules powder, confirming enhanced thermal stability up to 150 oC tested. The sample dimension was 10 mm in width x 25 mm in length ...... 161

6.5 FTIR spectra of neat Nafion, neat Noria, and the impregnated membranes with different feed ratios of supramolecules. These spectra were -1 normalized to the –CF2 backbone peaks of Nafion. (a) 3900–2300 cm region shows the broadness of O–H stretching present in the impregnated membranes whereas (b) 1800 cm-1 to 570 cm-1 region implying that there is little or no influence of Noria on the fluorocarbon chain of Nafion but with a shift of the S–O stretching peak to a lower wavenumber with the addition of Noria ...... 163

6.6 (a) Plot of wavenumber of the S–O stretching of FTIR as a function of Noria composition. A systematic movement of the S–O band to a lower wavenumber suggesting an inter-species through hydrogen bonding interaction might occur between the hydroxyl groups in Noria with sulfonate terminal groups of Nafion, (b) Estimated amount of Noria supramolecules incorporated within the Nafion membranes based on the integration of area under the C=C aromatic characteristic peak and physical weighing measurements. The dashed line represents the ideal condition where the amount of Noria incorporated in Nafion increases linearly with composition ...... 165

6.7 WAXD scan of (a) neat Nafion, (b) 95/5 Nafion/Noria impregnated membrane and (c) pure Noria over the 2θ range of 4o-28o at room temperature demonstrate the amorphous halo of neat Nafion and presence of crystalline peaks in pure Noria. It is evident that there is no discernible crystalline peak of Noria upon impregnation with Nafion ...... 167

6.8 (a) SAXS scans of neat Nafion and impregnated Nafion/Noria membranes with varying feed ratios. Note that the arrow indicates the peak position of Iq2 versus q plot. (b) Plot of domain spacings as a function of waterwheel supramolecules in feed in which the domain spacings increases and remains stationary up to 10 wt% thereafter. The increment of the domain size (or radius) for 2 times corresponds to the expansion of volume by 8 folds...... 169

xxii

6.9 Plot of number of mole of water per sulfonate site, λ, as a function of impregnation level. The water uptake in Noria impregnated membranes is reduced by 73% as compared to the neat Nafion. Note that λ was determined after 24 h hydration in deionized water at room temperature. On the other hand, the IEC of Noria impregnated membranes show increased of 1.39 meq g-1 at 5% feed ratio as compared to the IEC of neat Nafion which is 0.92 meq g-1 ...... 170

6.10 (a) Log-log plots of storage loss and (b) a semi-log plot of loss tangent versus frequency obtained for the 95/5 Nafion/Noria impregnated membrane at 74% RH. The real impedance (Z’) demonstrated a sigmoidal reduction of impedance values with frequency, whereas the loss tangent peaks show a consistent movement towards higher frequency as temperature is raised ...... 173

6.11 Cole-Cole plot of (a) neat Nafion and (b) 5% Noria impregnated Nafion membranes at 74% relative humidity from 40 oC to 115 oC over the frequency range of 0.1 Hz–10 kHz. It is evident that the diameter of the semi-circular plots contracted with increasing temperature, indicating that the resistance of the cell is reduced as the temperature increases ...... 174

6.12 Proton conductivity plot neat Nafion and Noria impregnated membranes from 70 to 115 oC at 74% RH. Evidently the proton conductivity of neat Nafion exerted a declining trend whereas the impregnated membranes demonstrate an increment proton conductivity trend which surpasses to that of neat Nafion at 110 oC and 115 oC. Later, it was found that Noria- BOC impregnated membranes which utilizes the hydrophobic tert- butyloxycarbonyl (–BOC) waterwheel supramolecules shows poor proton conductivity at all temperatures tested and drops its value comparable to those of the pure Nafion at higher temperatures of 110 and 115 oC which confirms that the hydroxyl group (i.e., solid water) of Noria contributes to the improved proton conductivity at elevated temperatures ...... 176

6.13 (a) Heating and cooling cyclic measurement of neat Nafion and impregnated membrane at 100% RH. Neat Nafion shows subtle proton conductivity drop after 5 cycles (i.e., 16 hrs per cycle) whereas the impregnated membrane showed sustained durability upon prolonged fuel cell tests. Note that the first, third and fourth cycle runs were omitted in the plot for clarity. (b) Plot of proton conductivity of the 5wt% Noria impregnated membrane at 100 oC and at 100% RH in comparison with that of the neat Nafion as a function of heating cycle which demonstrated the excellent durability of impregnated membrane after total of 160 hours of cyclic measurements ...... 178

xxiii

6.14 FTIR spectra of Noria impregnated membrane (a) before and (b) after subjected to 5 sets of heating and cooling fuel cell measurements ...... 179

7.1 Chemical structure of waterwheel supramolecules with tert- butyloxycarbonyl, namely Noria-BOC ...... 183

7.2 DSC and TGA thermograms of pure Noria-BOC where the melting temperature coincides with degradation temperature ...... 184

7.3 Overlay FTIR spectra of (a) neat Nafion in acid form and (b) pure Noria- BOC at 100 oC ...... 185

7.4 FTIR spectra of Noria-BOC impregnated Nafion membranes exemplifies the movement of S–O stretching, which is associated with sulfonic acid sites in Nafion from 1057 cm-1 to 1050 cm-1 ...... 186

7.5 (a) Plot of wavenumber of the S–O stretching of FTIR as a function of Noria-BOC in feed. The systematic movement of the S–O band to a lower wavenumber suggests an inter-species hydrogen bonding interaction might occur between the carbonyl groups in Noria-BOC and the sulfonic acid terminal groups of Nafion. (b) Noria-BOC supramolecules incorporated within the Nafion membranes through two different approaches, i.e., integration of area under the C=O characteristic peaks and physical weighing measurements ...... 188

7.6 SAXS studies of Noria-BOC impregnated Nafion membranes shows the changes of size in the ionic domains of the impregnated Nafion membranes with amount of waterwheel supramolecules in feed. d- spacings of ionic domains remain constant as the amount of waterwheel supramolecules reaches 3 wt% Noria-BOC in feed ...... 190

7.7 (a) Plot of water uptake as a function of soaking time for neat Nafion and Noria-BOC impregnated membranes at room temperature. The water uptake of neat Nafion in acid form is 37wt%, while the Noria-BOC impregnated Nafion membranes show a reduction to 20 wt% in water - uptake whereas (b) graph of λ (moles of H2O per mole of SO3 ) of impregnated Nafion membranes as a function of Noria-BOC composition in feed (after soaking for 24 h), illustrating significant reduction of water molecules due to impregnation. The IEC value of Noria-BOC impregnated membranes bear little to no effect with the value of 1.01 meq g-1 at 5% feed ratio compared to the IEC value of neat Nafion which is 0.92 meq g-1 ...... 192

xxiv

7.8 Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and 5wt.% Noria-BOC impregnated Nafion membrane (dotted line). The inset photographs show improved thermal stability of the impregnated membranes with unchanged color of impregnated membranes as compared to the dark brownish color of neat Nafion-acid after DMA experiments at 150 oC. The yellowish color of Noria-BOC impregnated membranes is inherited from the initial color of the waterwheel supramolecules solid. The sample dimension was 10 mm in width x 25 mm in length ...... 194

7.9 Conductivity plots of neat Nafion and 5% Noria-BOC impregnated membranes as a function of temperature at 74% RH illustrate the decreasing trend of both membranes at elevated temperatures. An error bar indicates the quadruplet readings for each temperature point ...... 196

xxv

CHAPTER I

INTRODUCTION

Over the past several decades, research on alternative energy has increasingly expanded among engineers, researchers and environmentalists. Primary concerns are global warming, the depletion of fossil fuels, and the need to reduce CO2 emissions as the

byproduct of the automotive and utility industries. As these concerns have grown to

alarming levels, clean energy technologies such as fuel cells have grown exponentially

[1].

The most important driver of new alternative energy is the possible exhaustion of

fossil fuels within the next few decades. The key component for the production of

alternative energy has been focused on energy efficiency and minimal environmental

impact [2]. As a result, fuel cell technology has made a notable and promising debut as

one of the most innovative energy options.

Among other alternative energy fuel cell types, proton electrolyte membrane fuel cells

(PEMFC) have received major attention and were developed for the transportation

industry as well as the stationary and portable fuel cell applications [3]. Operated in

lower operating temperature conditions (i.e., 60 oC–80 oC), PEMFC functions on a very

simple principle. PEMFC consists of four main components, which are cathode/anode

electrodes, gas diffusion layers, graphite bipolar plate and proton electrolyte membrane,

hence the name, proton electrolyte or exchange membrane fuel cells [4]. The heart of

PEMFC is the electrolyte membrane, which allows the conduction of only proton ions.

Perfluorinated ionomer membrane, commercially known as Nafion has been the potential

candidate for the electrolyte membrane used in the PEMFC. Thermal and chemical

resistance combined with excellent perm-selectivity has put Nafion as the benchmark of electrolyte membrane in PEMFC applications.

Structurally, Nafion is unique. This micro-phase separated perfluorinated ionomer

was developed by Dr. Walther Grot in late the 1960s. Nafion consists of chemically and

thermally stable fluorocarbon backbone, hydrophobic side chains and hydrophilic ionic

domains which are actively involved in the proton conducting process [5-8]. Even though

this ionomer has outstanding characteristics to be used in PEMFC applications, there are

several obstacles exhibited by Nafion which lower some efficiency aspects of the

PEMFC. It is known that the operational principle of PEMFC relies heavily on sufficient

hydration levels for proton conduction to take place where water is produced as the

byproduct. Due to the highly hydrophilic properties possessed by Nafion, the excessive

water absorption activity leads to dimensional instability of the ionomer and the

possibility of producing a condition wherein flooding and/or back-diffusion of water

might occur. It has been the motivation of the US Department of Energy to tailor the

solution for the water management of Nafion without hampering the membrane’s

performance [9].

Besides being prone to swelling upon contact with polar solvents, another

prevalent shortcoming of Nafion is its low and limited operating temperature. One major

drawback is that its operation temperature is limited to 60–80 oC, above which the proton conduction progressively gets worse. In the actual proton fuel cell operation, multiple

PEM fuel cells need to be stacked to obtain desirable power output [3]. An inevitable consequence is that the cell temperature rises significantly beyond the optimum operating temperature of Nafion. This presents a major challenge for researchers attempting to make the PEMFC function at higher operating temperatures.

Several efforts and attempts have been made to fabricate new proton electrolyte membranes, notably polyimides based ionomers, sulfonated thermoplastics poly(arylene ether sulfone) as well as sulfonated fluorocarbon poly(styrene) [10-14]. Even though it has been proven that these newly synthesized materials demonstrated excellent suppression of excessive swelling upon contact with solvents, these materials come short in the proton conducting performance compared to Nafion. Hitherto, Nafion remained the best ionomer for the PEMFC applications, and several approaches for modifying the

existing Nafion have been carried out to improve its limitations.

Modification strategies include physical blending of Nafion with other polymers

and also through in-situ impregnation of Nafion membranes with small molecules, such

as ionic liquids and/or polyacids [15-19]. The first mentioned modification alters the

fluoromatrix part of Nafion while the latter modification locally targets only the ionic

clusters of Nafion.

Solution blending of Nafion with homopolymer of poly(vinylidene fluoride),

PVDF, was first studied by Kyu et al. [20,21]. Blends of Nafion/PVDF were

demonstrated to exhibit a lower critical solution temperature (LCST) phase behavior

where development of small angle light scattering (SALS) halo and appearance of

interconnected domain structures observed by polarized optical microscope (POM) at

approximately 210oC revealed thermally induced liquid–liquid phase separation by

spinodal decomposition above the melting temperature of PVDF. The phase separation

behavior of Nafion/PVDF blends is highly dependent on the types of counterions

attached to sulfonate groups of Nafion [22]. Phase separation of Nafion/PVDF in sodium

(Na+) form demonstrated large-scale phase separation, while larger counterions such as tetrabutylammonium (TBA+) contributed to an isotropic phase behavior of the blends at

all temperatures.

The approach of in-situ impregnation of Nafion, which locally targets only the

ionic clusters, is increasingly being utilized to address the aforementioned limitations of

Nafion. Through the impregnation approach, small molecules such as ionic liquids have

been widely used to infuse through these ionic domains to provide better swelling

suppression as well as to increase the protonic and/or ionic conductivity properties of

Nafion. It is anticipated that the exceptional conductivity properties possessed by ionic

liquids will afford the increase of the conductivity performance of these impregnated

Nafion membranes. However, these ionic liquids are prone to leach out and serve as a

potent plasticizer to Nafion, ultimately lowering the mechanical properties of the

impregnated membranes [15].

In this study, we have characterized the structural and conductive properties of the

modified Nafion membranes which were modified using the two aforementioned

approaches. Chapter II is focused on the literature review on the basic operation of proton

electrolyte membrane fuel cell (PEMFC), structural properties of Nafion, the advances in

development of new electrolyte membranes and the description of concepts related to

thermodynamics of polymer solutions and blends on the binary phase diagrams together

with dynamics of the phase separation process. Chapter II also briefly introduces the

fundamental concepts of electrochemical impedance analyses which have been applied

throughout this dissertation in order to demonstrate the conductive properties of the

modified membranes.

Physical and chemical properties of Nafion ionomer membrane, poly(vinylidene

fluoride-trifluoroethylene) (PVDF-TrFE) copolymer, and all three unique supramolecules

are described in Chapter III together with the experimental and instrumental techniques

employed in the present study. Chapter IV discusses the phase behavior, phase

morphology and conductive properties of the Nafion/PVDF-TrFE blends. The

Nafion/PVDF-TrFE blends are sought to suppress the swelling of Nafion upon solution

blending with the hydrophobic PVDF-TrFE copolymer.

Chapter V describes the proof of concept on incorporating the large polymer molecules, i.e., photocurable hyperbranched polyester (HBPEAc-COOH), into the ionic domains of Nafion through impregnation and subsequent in-situ polymerization of photopolymerizable hyperbranched polyester. The in-situ polymerization of HB network within the ionic domains will afford potential leaching in actual fuel cell operations and the formation of functional polymer networks will provide better mechanical and dimensional stability of the impregnated membranes.

Chapter VI demonstrates the structural characterization and the conductive

behavior of the in-situ impregnated membranes with unique supramolecules namely

Noria or waterwheel supramolecules terminated with hydrophilic hydroxyl functional

groups. Presence of 24 hydroxyl groups surrounding each supramolecule are anticipated

to increase proton concentration within ionic clusters of Nafion while performing like

‘solid water’ molecules at high operating temperatures within proton electrolyte fuel

cells. Chapter VII illustrates the structure characterization as well as the conductivity

properties of in-situ impregnated Nafion membrane using waterwheel supramolecules with encapped tert-butyloxycarbonyl groups (Noria-BOC) in comparison to the previous

chapter.

The concluding Chapter VIII is devoted to the overall summary and objectives

achieved in the study together with recommendations for future work.

CHAPTER II

BACKGROUND AND LITERATURE REVIEW

2.1. Introduction to Fuel Cells

A fuel cell is an electrochemical energy converter that converts chemical energy

of fuel directly into direct current [3]. Typically, electricity generation from fuels involves several energy conversion steps such as combustion of fuels which then converts chemical energy into heat and/or mechanical energy which in turn powers a generator. A fuel cell circumvents all these processes and generates electricity in a single step without involving any moving parts [4]. Due to this simplicity, fuel cells applications hold significant promise in mainstream industries.

To some level, a fuel cell is somewhat similar to a battery. It has an electrolyte, positive and negative electrodes, and generates a direct current through electrochemical reactions. Batteries generate electricity by the electrochemical reactions involving the reactant that are contained within the assembly. As these reactants depleted, a battery may be discharged. The difference between battery and fuel cell is that a fuel cell requires a continuous and constant supply of reactants [3]. A fuel cell cannot be discharged as long as the reactants (e.g. fuel and oxidant) are supplied. Typical reactants for fuel cell are hydrogen and oxygen where neither has to be in its pure form [1, 2]. Hydrogen may

be present either in a mixture with other gasses (e.g., CO2, N2) or in hydrocarbon such as natural gas, or even in liquid hydrocarbon such as methanol. Ambient air contains enough oxygen to be used in fuel cells. Yet another difference between a fuel cell and a battery is that a fuel cell generates waste heat and water byproducts. An efficient system is needed to manage these byproducts as to ensure the efficiency of fuel cell continues.

Today research has focused on developing fuel cells for stationary, automotive, portable, and military power applications. Fuel cells, in general, are attractive because they provide an innovative alternative to current power sources with higher efficiencies, renewable fuels, and a lower environmental cost.

Figure 2.1: Simplistic view of PEMFC [redrawn from ref. 1].

A simplistic view of a polymer electrolyte fuel cell is shown in Figure 2.1. The

main components of a fuel cell (specifically polymer electrolyte membrane fuel cell;

PEMFC) are catalyst layers, gas diffusion layers and the polymer electrolyte membrane.

These three components made up to what is known as the membrane electrode assembly

(MEA). The catalyst layers act as electrodes and typically consist of platinum or platinum

alloys. Gas diffusion layers are carbon fiber cloths that act as a medium to disperse the

fuel evenly across the polymer electrolyte membrane. The polymer electrolyte membrane

acts as a proton conductor and also as the reactant separator. The proton electrolyte

membrane typically consists of strongly acidic groups which aid the proton transport

through the membrane. The fuel cell produces power via an electrochemical reaction in

which a fuel is oxidized at the anode to produce positively charged protons and

negatively charged electrons. Protons will then travel through the proton electrolyte

membrane where they react with oxygen at the cathode to produce water. Electrons, however, need to travel along an external circuit to power the load attach to the fuel cell

[23, 24].

2.1.1. Types of Fuel Cell

Fuel cells are classified primarily based on the electrolyte used. Each type of fuel cells has intended applications based on power output limitations, operating temperature, and size of the power system [25-27].

The highest electrical efficiency compared to all alternative fuel cell is the alkaline fuel cells (AFCs). AFC normally uses potassium hydroxide (KOH) as the

electrolyte and the main necessity for its reactant is it has to be an ultra pure gases. The

operating temperature for AFCs is below 100 oC, but higher temperatures are desirable for improved hydrogen oxidation kinetics. The expected power output of an AFC is in the range of tens of megawatts (MW) [25].

Simple construction, thermal and chemical stability of phosphoric acid as the electrolyte makes phosphoric acid fuel cells (PAFC) the most advanced fuel cell system and can be operated in the range of 150-200 oC. The main usage of PAFC is for

stationary power ranging from dispersed power to on-site generation plants. Its power

outputs of 0.2- 20 MW made PAFC the best alternative energy provider that is able to

supply shopping malls and hospitals with electricity, heat water and provide primary or

backup power [25].

Molten carbonate fuel cells (MCFCs) and solid oxide fuel cells (SOFCs) operate

at high temperatures, 600-800 oC and 800-1000 oC, respectively. These are useful for

stationary power applications. MCFC uses liquid lithium potassium or lithium sodium

carbonate stabilized in a matrix as the electrolyte for the system, while SOFC utilizes

ceramics as the solid electrolyte. An advantage for SOFCs over MCFCs is the usage of

solid electrolyte, which eliminates the concern over liquid electrolyte leaking out of the

system [25].

Direct methanol fuel cells (DMFCs) are similar to the polymer electrolyte

membrane fuel cells (PEMFCs) in the sense that its electrolyte is a polymer and the

charge carrier is the hydrogen ion (proton). The liquid methanol (CH3OH) is oxidized in

the presence of water at the anode to generate CO2, hydrogen ions and the electrons that

travel through the external circuit as the electric output of the fuel cell [25]. Hydrogen ions travel through the electrolyte and react with oxygen from the air and the electrons from the external circuit to form water at the anode completing the circuit. These cells have been tested in a temperature range from about 50-120 ºC. The DMFC is still in the premature stages of development, but it has been successfully demonstrated powering mobile phones and laptop computers which makes it the potential target end uses in the future [26].

The polymer electrolyte membrane fuel cells (PEMFCs) use a proton-conducting polymer membrane as the electrolyte with an operating temperature of 60-80 oC.

Commonly used reactants for this system are hydrogen and oxygen [27]. Due to its low

operating temperature, hydrogen and methanol fuel cells are popular for use in

automotive and portable electronic (consumer) applications. PEM will be discussed in

details below. The major types of fuel cells, classified by the type of electrolyte, are

outlined in Table 2.1.

Table 2.1: Summarization of major fuel cells; classified by types of electrolyte.

Fuel Cell Electrolyte Operating Electrochemical Reactions Temperature (oC)

+ - Polymer Solid organic 60-80 Anode: H2 → 2H + 2e + - Exchange polymer Cathode: ½O2 + 2H + 2e → Membrane H2O (PEMFC) Cell: H2 + ½O2 → H2O

- Aqueous 90-100 Anode: H2 + 2(OH ) → 2H2O Alkaline (AFC) solution of + 2e- - potassium Cathode: ½O2 + H2O + 2e → hydroxide 2(OH-) soaked in a Cell: H2 + ½O2 → H2O matrix

+ - Phosphoric Acid Phosphoric acid 150-200 Anode: H2 → 2H + 2e + - (PAFC) soaked in a Cathode: ½O2 + 2H + 2e → matrix H2O Cell: H2 + ½O2 → H2O

2- Molten Solution of 600-800 Anode: H2 + CO3 → H2O + - Carbonate lithium, sodium, CO2 + 2e - (MCFC) and/or Cathode: ½O2 + CO2 + 2e → 2- potassium CO3 carbonates Cell: H2 + ½O2 + CO2→ H2O soaked in a + CO2 matrix (CO2 is consumed at anode and produced at cathode, thus it is included in each side of the equation)

- - Solid Oxide Solid zirconium 800-1000 Anode: H2 + O2 → H2O + 2e - - (SOFC) oxide with a Cathode: ½O2 + 2e → O2 small amount of Cell: H2 + ½O2 → H2O yttria

2.2. Polymer Electrolyte Membrane Fuel Cells (PEMFCs)

The most commonly used electrolyte membrane for PEMFC application is

Nafion, a perfluorinated ionomer developed by E.I. DuPont Company. PEMFC operates with the use of carbon-supported platinum catalysts at the anode and cathode with a polymer electrolyte membrane sandwiched between them. Hydrogen is fed to the anode and catalytically oxidized. Electrons are created in the oxidation reaction and flow through an outer circuit to power the connected load. Protons flow through the polymer membrane and react with oxygen at the cathode in a reduction reaction to give a clean byproduct of water.

In 1968, E.I. DuPont Company commercialized a proton electrolyte membrane

based on poly(perfluorosulfonic acid) under the trade name Nafion®. The highly

fluorinated backbone structure of Nafion displays excellent resistance to degradation in a

fuel cell environment and thus increasing the fuel cell lifetimes [6].

Since then, other companies, such as Asahi Chemical in Japan and briefly Dow

Company in the U.S., have investigated membranes based on poly(perflurosulfonic acid)

structures. Flemion which having the same chemical and thermal stability as Nafion with different terminal ionic group (carboxylate terminating ionic group) were manufactured

by Asahi Chemical. Nafion has remained the industry standard proton electrolyte

membrane and almost all current PEM fuel cell research from a device standpoint focuses

on this type of electrolyte. Major applications for Nafion also include chlorine synthesis

via electrolysis (chlor-alkali processes) [28].

Moderate operating temperatures for PEM fuel cells are required because of the

need for aqueous proton transport and the polymers used have relatively low glass

transition temperatures (Tg), especially when hydrated. Polymeric electrolytes based on

sulfonic acid ion conducting sites require humidified reactant streams to hydrate the

membrane and increase its conductivity. In current poly(perfluorosulfonic acid)

copolymer membranes, hydration must be quite high to produce sufficient conductivity,

thus restricting the maximum operating temperature of PEM fuel cells to about 80 ºC in order to prevent membrane or catalyst layer from drying out. One current thrust of fuel cell research is to increase the operating temperature of PEM fuel cells to 120 ºC or higher. This may be possible by producing membranes that retain conductivity as well as more thermally and mechanically robust at high temperatures.

A single PEM fuel cell is illustrated in Figure 2.2 can operate on just a single membrane with its power output of ≤ 0.5 Watts. An increase in power output of a fuel cell is achieved by integrating single cells in series by constructing a fuel cell stack where the voltage of each single cell is additive. PEM fuel cells have shown the most promising applications in automotive and portable power microelectronics fields. Renewed interest in the commercial development of fuel cells has fostered much research into new proton electrolyte membranes. Requirements for the next generation proton electrolyte membranes includes high proton conductivity over a range of water contents, dimensional stability in hydrated condition, sustaining high proton conductivity at high temperatures, low reactant permeation, and low electrical conductivity.

Figure 2.2: Schematic diagram of single PEM fuel cell [redrawn from ref. 29].

Figure 2.3: Schematic diagram of basic principle operation of PEMFC [redrawn from ref. 30]. 15

2.2.1. Principle Operation of PEM Fuel Cells

Polymer electrolyte membrane or PEM is one of the promising types of fuel cells that have been developed which uses the concept of electrochemical devices oxidizing hydrogen to produce energy and environmental friendly byproducts. PEMFCs, also known as proton electrolyte membrane fuel cells have found a great application in energy

field and have been identified as the new resource for energy and power alternative. The

principle of PEMFCs involves the oxidation of hydrogen and splitting of hydrogen into

two primary constituents, positively charged proton and negatively charged electrons.

Figure 2.3 shows the schematic diagram of basic principle of PEM fuel cell operation.

Each hydrogen atom consists of one electron and one proton traveling through the

membrane, whereas the electrons travel through electrically conductive electrodes, and

through the outside circuit. Water is produced in the electrochemical reaction and being

pushed out of the cell with excess flow of oxygen. The net result of these simultaneous

reactions is the production of direct electrical current [1-3].

The heart of PEM fuel cells is a polymer membrane that has some unique

characteristics. This membrane must exhibit relatively high proton conductivity and

perm–selectivity in order to ensure the productivity of the electrochemical cell.

In PEMFC, the electrochemical reaction occurs at the anode and cathode of the

electrodes. Oxidation of hydrogen where it is oxidized to liberate two electrons and two

protons occur at the anode giving;

+ - H2  2 H + 2 e [2.1]

The protons are conducted from the catalyst layer through the proton electrolyte

membrane and the electrons travel through the electronic circuit. At the cathode,

reduction of oxygen will take place;

+ - ½ O2 + 2 H + 2 e  H2O [2.2]

thus giving the overall cell reaction:

H2 + ½ O2  H2O [2.3]

Both reactions can be catalyzed by nanocrystalline platinum supported with

carbon black, however, other modified catalysts are also often used to minimize carbon

monoxide poisoning at the anode. As an example, most state-of-the-art anode catalysts are alloys of platinum and ruthenium supported on carbon black. The ruthenium helps to maintain fuel cell performance even with presence of hundreds of parts per million of carbon monoxide in the anode feed stream, while the carbon black support increases the surface area of the heterogeneous catalyst to increase utilization.

DuPont’s perfluorinated ionomer membrane, Nafion has always been the most prevalent copolymer membrane used in hydrogen fuel cells [3]. Specifically, the Nafion

1135 (1100 equivalent weight, 3.5 mils thick) and Nafion 112 (1100 equivalent weight, 2

mils thick) are the best electrolyte membrane used [3]. Thinner membranes are often the best bet due to the decrease in cell resistance more than offsets any performance losses associated with the permeability of hydrogen and oxygen through the membrane. Even though these poly(perfluorosulfonic acid) copolymer membranes are expensive, they

remained the best candidate in the proton electrolyte membrane fuel cell applications.

Apart from the high cost, the greatest challenges of Nafion are the poor water management and its low thermal stability [24]. Nafion is required to be fully wet thus it will give its greatest proton conductivity. In operation of the PEM, reactant are mainly humidified and together with the water produced at the cathode resulting in accumulation of water, which is highly affecting the electrochemical performance through flooding.

This in turn will also leads to dimensional instability issues of the Nafion membrane.

Besides, presence of mildly acidic water in the usage of metallic bipolar plate will cause trace quantities of peroxide via the reduction reaction and will lead to presence of transition metal cations (Ni2+ and Fe3+) in the water. These cations will diffuse into the

membrane, or if the water from the cathode is recycled to the anode side, may lead to

reduction in the conductivity of the membrane. Additionally, the presence of the cations

with peroxide formed at the cathode can lead to an enhanced oxidation of the membrane,

which affects it durability and long–term stability. On top of that, back-diffusion of

accumulated water would definitely plunge the fuel cell performance. Therefore, it is

essential for the on-going development of the electrolyte membrane to consider the

swelling suppression of the membrane, yet still manage to accommodate the amount of

water needed for the efficient fuel cell performance.

Current membrane technology dictates that the maximum temperature for

hydrogen fuel cells remains at about 80 °C [1]. However, heat will start to generate upon

stacking of the cells. Thus, increasing of stacking will leads to higher heat generation and subsequently increasing the temperature within the cell. It is therefore, important for the electrolyte membrane to possess increased thermal stability above its ionic glass

transition. On the bright side, increasing the operating temperature would create more

high quality waste heat to be used in the system, e.g., to heat a home or radiated to the

environment. Membrane development programs for PEM fuel cells focus almost exclusively on raising the operating temperature of the cell while remaining its

mechanical stability and efficiency. Therefore, the improvement needed by Nafion is on

tailoring solution that involves its water management and pushing its thermal stability to

elevated or higher temperatures (> 80°C).

2.2.2. Assembly of PEM Fuel Cells

A fuel cell consists of electrodes (anode and cathode), electrolyte membrane and

bipolar plates. Each assemble will be briefly discussed in the next chapter.

2.2.2.1. Electrodes (Anode and Cathode)

Principle operation of PEM involved the movement of proton from anode to

cathode in order to produce an electrochemical process which will then be converted to

energy output [31]. The best catalyst for both anode and cathode is platinum. The

platinum catalyst is formed into very small particles on the surface of somewhat larger

particles of finely divided carbon powders. This will gives a very high proportion of the

surface area to be in contact with the reactants. There are two different ways of fixing these two electrodes in the system.

In the separate electrode method, the carbon–supported catalyst is fixed, using propriety techniques, to a porous and conductive material such as carbon cloth or carbon

19 paper. Poly(tetrafluoroethylene) or PTFE will often be added, due to its hydrophobic nature and thus will expel the byproduct (water) to the surface where it can evaporate. In addition to provide the basic mechanical structure for the electrode, the carbon paper or cloth also allows diffusion of the gas onto the catalyst which being referred as the gas diffusion layer. The electrode is then fixed to each side of a piece of polymer electrolyte membrane.

Another method is by building the electrode directly onto the electrolyte. The platinum on carbon catalyst is fixed directly to the electrolyte, thus manufacturing the electrode directly onto the polymer electrolyte membrane, rather than separately assembled [31]. The catalyst which often being mixed with hydrophobic PTFE, is applied to the electrolyte using rolling method [32], by spraying [33] or an adapted printing process [34]. Once the catalyst is fixed on the membrane, a gas diffusion layer must be supplied. This will be carbon cloth or paper, about 0.2 to 0.5 mm thick, as is used in the separate electrodes method. The gas diffusion layers will also form an electrical connection between the carbon–supported catalyst and the bipolar plate, or other current collector. In addition, it carries the by–product water away from the electrolyte surface and also forms a protective layer over the very thin layer of catalyst.

2.2.2.2. Bipolar Plates

The bipolar plate has to collect and conduct the current from anode of one cell to the cathode of the next, while evenly distributing the fuel gas over the surface of the anode, and the oxygen/air over the surface of the cathode. In addition to this, it often has

20 to carry a cooling fluid through the stack and keep all these reactant gases and cooling fluids apart. In addition, it also has to contain the reactant gases within the cell, so that the edges of the cell will have a sufficient area or size to allow space for sealing. Extra channel on bipolar plates are also essential to fuel cell applications as to provide cooling process throughout the cells [31].

2.2.2.3. Proton Electrolyte Membrane

The heart of PEM is the polymer membrane, which is located in between the anode and cathode. This polymer membrane is designed to have very unique capabilities in order to conduct protons as well as the need to be impermeable to gases [36-39]. This membrane acts as the electrolyte is squeezed between two porous electrodes; electrically conductive electrodes. Further details on perfluorsulfonate ionomer or Nafion will be discussed in the next section.

2.3. Perfluorinated Ionomer Membrane, Nafion

Perfluorinated ionomer or commercially produced Nafion possesses all the criteria needed to be the best electrolyte membrane candidate. These materials are produced by copolymerization of a perfluorinated vinyl ether comonomer with tetrafluoroethylene (TFE) [5, 7]. Figure 2.4 shows the chemical structure of Nafion.

Nafion is graded and classified by its equivalent weight. Equivalent weight (EW) is the number of grams of dry Nafion per mole of sulfonic acid groups when the material is in the acid form. The equivalent weight value is related to the perm-selectivity property and

21 may be expressed as the ion exchange capacity (IEC) where IEC is equivalent to

1000/EW [6].

Figure 2.4: Chemical Structure of Nafion.

Nafion perfluorinated ionomer membrane was initially used as a perm-selective membrane separator in chlor-alkaline and electrochemical cells for a large scale industrial production of NaOH, KOH and Cl2. Nafion functioned as a membrane separator separate

- Cl2 and H2 gases thus preventing the back migration of hydrated OH ions from the catholyte to the anolyte chamber. In addition to that, Nafion membrane will only allows the transport of hydrated Na+ ions in the form of aqueous sodium chloride.

DuPont was the first to manufacture this outstanding ionomer membrane until

Asahi Chemical Company came up with perfluorinated ionomers terminated with carboxylate (-COOH) terminal groups namely Flemion. Dow Chemical Company then came up with an ionomer that exhibit similar resemblance of Nafion but with an exception of shorter side chain which contains only one ether oxygen, rather than two ether oxygens, that is –O-CF2-CF2-SO3H [40]. Despite all of these newcomer perfluorinated ionomer membranes, Nafion remains the preferred electrolyte membrane

for the electrochemical cells applications and now Nafion is widely used for fuel cells

applications especially in the field of PEM fuel cells.

2.3.1. Morphology of Nafion

Unique molecular structure of Nafion is similar to poly(tetrafluoroethylene)

- + (PTFE) but with sulfonic acid (-SO3 H ) functional groups at the terminal end. The

- + Teflon backbone provides mechanical strength while sulfonic acid (-SO3 H ) chains

provide charged sites for proton transport. It is assumed that free volume of Nafion

aggregate into interconnected nanometer–sized pores whose walls are lined by sulfonic

acid groups. These aggregations are referred to as the ionic domains/clusters.

Nafion possess a very complex structure and several models have been proposed

since the early 1970s to describe aggregation of ionic groups within the Nafion polymer.

Gierke et al. [41] was the first to determine the morphology of Nafion using WAXD and

SAXS and proposed that Nafion exhibit networks of spherical cluster interconnected by

narrow channel. Based on on SAXS studies, clusters of sulfonate ended perfluoroalkyl

ether groups was probed to be ~40Å in diameter and these clusters are connected with

- ~10 Å channels. These –SO3 coated channels were taken into account for intercluster ion

hopping of positive charge species and reject the negative ions. Gierke also found that increasing of equivalent weight enhances the crystallinity of Nafion [42]. Concentration of the side chains tends to decrease with increasing of the equivalent weight.

Through SAXS, Gierke demonstrated that upon hydration, the ionomer peak tends to increase in intensity and shift to lower angles with decrease in equivalent weight [56,

57] (Figure 2.5). Furthermore, for a given equivalent weight, this peak was found to shift

to lower angles and increase in intensity with increasing of water content. These ionic

domains were found to exhibit increased linearly relationship between the numbers of

water molecules per exchange site with water content. However, higher value of EW

suppresses the water uptake since the number of exchange sites per cluster is reduced

thus making the membrane less hydrophilic. Yeager et al. [43] then came out with a

proposal of the “Three Phase Model” for Nafion. The three phases consist of (a)

microcrystalline fluorocarbon backbone, (b) pendant side chains with interfacial region

with relatively large fractional void volume containing some, some water and those sulfate or carboxylic groups and counter ions, which are not in clusters, and (c) the hydrophilic ionic exchange clustered regions.

Nafion in acid form was found to have three contrast regions in the angular range probed by SAXS and SANS [44-46]. These three contrast regions are crystalline phase, ionic clusters and inhomogeneous matrix phase. Neutralization of Nafion in the Na+ form and subsequent thermally quenched from 330 oC, the sample was found to be amorphous,

and the low angle scattering maximum, corresponding to a Bragg spacing of ~18 nm

peak diminished. Transmission electron microscopy (TEM) also has been utilized to define the morphology of Nafion. The morphology of Nafion’s cluster region in Cs+

form was found to exhibit a diameter of 25–50 Å [47, 48].

There are few other models proposing the morphology of Nafion [49, 50] and all

of these models, agreed upon one similar fact that Nafion consist of phase separated

morphology. Polar and non–polar microphase separation in Nafion are discrete by the

fact that polar (hydrophilic region) contains the ionic groups and their counter ions while

non–polar (hydrophobic region) composed of the polymer fluorocarbon backbone.

Figure 2.6 shows several models proposed in order to explain the morphology of Nafion.

Figure 2.5: Small angle x-ray scans of hydrolyzed Nafion Na+ with different equivalent weight, swollen in water [Reprinted with permission from J. Polym. Sci.: Polym. Phys., 19, T. D. Gierke, G. E. Munn, F. C. Wilson, pages 1687 – 1704, Copyright 1981, Wiley].

Figure 2.6: Yeager’s Three Phase Model and Gierke’s Cluster Model [redrawn from ref. 42].

2.3.2. Hydration Effect of Nafion

There are several other studies being done on Nafion to determine the hydration

effect and from all of these researches, it was agreed that with increasing of hydration,

the number of clusters decreased, the size of average cluster increased [51, 52]. Through

the SAXS studies, the larger structures of clusters were rationalized due to the cluster

agglomeration [53]. However, with increasing of equivalent weight, water uptake was

found to decrease. This is due to the fact that with increment of equivalent weight, less

sulfonate groups are attached at the main backbone chain, thereby leading to higher

crystallinity and lowering the hydrophilic properties. By recognizing the increment of crystallinity, stiffness of Nafion increases thus at higher equivalent weight, more energy is required to hydrate each exchange site.

Hydrophilic sulfonate groups of Nafion tend to aggregate and forming clusters that is referred to as ionic clusters and notoriously swells in water or any polar liquids

(e.g., methanol or ethanol. Upon hydration or swelling, Nafion membranes can soak up to

22 molecules of H2O per sulfonic groups [54]. Orfino and Holdcroft found that the number density of clusters decreased upon hydration while the radius of cluster size increased from 1.12nm to 2.05nm [55]. Several studies also have been conducted on the ionic clusters upon hydration [56, 57].

McLean et al. [58] have conducted the microscopy studies on hydrated Nafion K+ with the aid of AFM using low oscillation energy tapping mode. It was found that the average cluster separation increases when the cluster number density decreases. At controlled humidity environment, sample exhibited 40–100 Å in diameter clumps of multiple ionic domains. Upon exposure to hydration, the ionic features became enlarged as channels with the width of 70–150 Å. Figure 2.7 shows the low oscillation energy tapping mode phase images of Nafion K+ samples upon hydration. The expansion on the dimensional of ionic domains is due to the polar–nonpolar interfacial tension which resulting in restriction to form a high surface/volume barrow channels. Meanwhile,

Gierke et al. implies that increasing hydration leads to a smaller number of larger clusters, due to cluster coalescence as shown in Figure 2.8.

Besides SAXS, WAXD and AFM, Fourier transform infrared spectroscopy

(FTIR) also has been used in probing the hydration effect of Nafion membranes [59-61].

It has been proved that upon increasing of humidity or level of hydration, O-H stretching despite being broad, is very prominent in the vicinity of band ~3500–3670cm-1 [59].

Figure 2.9 shows the O-H stretching of Nafion upon hydration. Concrete proof also has been delivered as the FTIR spectrum as a function of relative humidity (% RH), showing an increment of the O-H stretching spectra.

Figure 2.7: AFM micrographs of Nafion K+ at dry state (left) and hydrated state (right). [Reprinted with permission from Macromolecules, 33, R. S. McLean, M. Doyle, B. B. Sauer, pages 6541 – 6550, Copyright 2000, American Chemical Society].

Figure 2.8: Hydration effect on ionic clusters of Nafion manifested by Gierke [redrawn from ref.42].

Effect of temperature upon swelling of Nafion has been studied by several

researchers [62, 63]. Degree of swelling becomes more prominent with increasing of water temperature. Recently, Alberti et al. [63] concluded that thermal treatment prior to

immersion in water significantly marked higher value of water content, λ. Water content,

λ is defined by number of moles of H2O per mole of SO3H groups [64]. This finding may

be attributed to the fact that the dried sample simply contains less bound water.

Figure 2.9: FTIR spectrum of Nafion membrane with varies hydration level. [Reprinted from Electrochimica Acta, 45, M. Ludvigsson, J. Lindgren, J. Tegenfeldt, pages 2267 – 2271, 2000 with permission from Elsevier].

Several researches have also used the solid state nuclear magnetic resonance

(SSNMR) measurement to determine the molecular assignment, molecular dynamics as

well as the quantification of hydration level in Nafion ionomer membrane. Bunce et al.

[65] demonstrated the linear relationship between water content in Nafion and the

chemical shift of the hydroxyl resonance. Batamack and Fraissard [66] found that

terminal sulfonate groups of Nafion will highly interact with water even at low hydration

level. Broad peak from the 1H NMR at 10.4 ppm signify the formation of hydronium ions

- with presence of only λ = 1 H2O/SO3 . And thus facilitate the proton conductivity

properties. Also by means of SSNMR, Ye et al. found that Nafion possessed relatively

low activation energy thus possessing excellent proton transport properties compared to

sulfonated poly(ether ether ketone) with high degree sulfonation.

Nafion membrane swells to a greater extent in a number of organic solvents than

it does in water. Yeo investigated the swelling of Nafion sulfonate membrane in a number of hydrogen bonding solvents [67] and found that Nafion swells more in methanol compared to water. This is presumably due to the larger difference in chemical potentials between the membrane and solvent.

2.3.3. Proton Conductivity and Proton Transport Mechanism of Nafion

Conductivity behavior of Nafion membrane is considered as the expression of the cation–sulfonate interaction besides the proton mobility within sulfonic groups of the membrane. Four point electrode configurations (in-plane) are generally used to measure conductivity. The measured values, however, reflect conductivity along the plane of the membrane rather than across the membrane thickness.

In the actual fuel cell measurement, proton conduction of the electrolyte

membrane was done using the through-plane measurement where protons are anticipated

to conduct perpendicular to the electrodes. Perfluorinated ionomer membrane, Nafion

117 marked an exceptional proton conductivity of ~0.1 S/cm at 80 oC with 100% of

- -4 relative humidity (i.e., ~23 H2O/SO3 ) and a value of ~1x10 S/cm for ionic conductivity

- at room temperature with ~1 H2O/SO3 . Anantaraman et al. [68] studied the effect of

humidity on the proton conductivity of Nafion and concluded that there is a tendency of

increment in conductivity value with relative humidity. Relative humidity is defined as

the amount of water vapor present in the gas compared to the amount that could be

present in the gas at the same temperature. Figure 2.10 shows the specific conductivity of

Nafion 117 as a function of moles of water molecules per sulfonate groups, λ.

Either if it is four point electrode configuration (in-plane) or through-plane

conductivity experiment, it has been proved that proton conductivity increases with

increasing water content, ion content, and temperature. Increasing of ion content leads to

an increase in the hydrophilic and ionic nature of the polymer, which results in higher

conductivities and higher water levels in PEMs [69]. However, if the water content is too

high, the membrane will highly be plasticized and will experience lowered mechanical

strength. Also, when the temperature reaches the α-relaxation temperatures (80 oC ~ 100

oC), the membrane becomes softer and dehydrates, which leads to a reduction in proton

conductivity [64]. Higher water contents will excessively swells the Nafion, thus leading

to dimensional instability. Paddison [70] suggested that to achieve a complete

dissociation of proton, thus allowing proton to conduct at its optimum efficiency, Nafion

- needs only 6 H2O/SO3 .Therefore, most investigations exhibit inconsistency in proton

transport or conductivity properties showing either increase or decrease in sulfonic acid

containing PEMs due to changes in polymer properties.

Figure 2.10: Conductivity of Nafion 117 as a function hydration level. [Reprinted with permission from J. Phys. Chem., 95, T. A. Zawodzinski, M. Neeman, L. O. Sillerud, S. Gottesfeld, pages 6040 – 6044, Copyright 1991, American Chemical Society].

According to Chen and Chou [71], the transport of proton in aqueous and organic- aqueous systems was found to be similar with a minor difference that the diffusion coefficient of proton is smaller in the organic-aqueous system than in the aqueous system.

Diffusion and proton conductivity properties for Nafion 117 were determined by Kreuer et al. [72] as a function of temperature and water content. The diffusivity and conductivity of the hydrated protonic Nafion were found to be comparable to those in the acidic aqueous solutions.

Proton transport within the ionic channels of Nafion used in PEM have been suggested to undergo two different mechanism namely proton diffusion and proton dissociation. It is known that proton is the only ion which has no electron shell and this resulting in the strong interaction with its electron density environment.

The proton dissociation mechanism, namely Grotthuss [74] mechanism or

“proton-hopping” mechanism, involves hopping of protons within the hydrogen bonds

- + - + from one hydrolyzed site (i.e., SO3 H or SO3 H5O2 ) to another across the membrane.

This mechanism undergo an additional reorganization of the proton environment which mainly comprises the reorientation of individual species or even more extended ensembles, thus resulting the formation of an uninterrupted trajectory of proton migration

[75]. At the center of the pore bulk or water bulk within the ionic channels, Grotthuss mechanism is anticipated to occur. It is assumed that dissociated protons with ~2 water molecules remain close to the surface thus designated as the surface water, while those with ≥ 2 water molecules will stratify to the pore bulk [76]. Figure 2.11 shows the simplified schematic diagram of proton transport mechanism involving (a) Grotthuss mechanism and (b) vehicular mechanism.

Protons will start to interact with the neighboring oxygen, which is well separated from its electronegative species, forming a hydrogen bonding. These protons will then tend to migrate through the translational dynamics. Known as the “vehicular

+ mechanism”, proton diffuses together with water, forming the hydronium ions (e.g. H3O ,

+ H5O2 ) [73]. In fully hydrated Nafion, the diffusion of protons is fast and mostly

exhibited the vehicular transport mechanism near the pore walls where an excess proton in bulk water existed.

Figure 2.11: Simplified schematic diagram of proton transport mechanism involving (a) Grotthuss mechanism or ion hopping and (b) vehicular mechanism where hydronium ions diffuse together with water. [Reprinted with permission from, J. Phys. Chem B, 109, M. Saito, K. Hayamizu, T. Okada, pages 3112-3119, Copyright 2005, American Chemical Society].

2.3.4. Mechanical Properties of Nafion

Dynamic mechanical properties on Nafion ionomer membrane have been studied

by several researchers [78-82]. The notable findings are three major relaxation peaks,

viz., α-, β- and γ-relaxation in the descending order of temperature. The α-relaxation

(~110oC) was assigned to the glass transition of the polar regions because it is sensitive to

ion type and degree of neutralization. Figure 2.12 illustrates the dynamic mechanical

scans of Nafion Cs+ with varying degree of neutralization. The β-relaxation was attributed to the glass transition of the Nafion matrix while γ-relaxation was assigned to the local –CF2– backbone motions [79]. Kyu and Eisenberg [80] discussed that hydration

sensitivity of both α- and β-relaxations is attributed to the close proximity of both

nonionic and ionic regions. Plasticization by water of the ionic domains would thus

influence not only the molecular motions within the ionic domains, but also motions of

the fluorocarbon backbone chains. Thus it is sensible to relate the water sensitivity to

both the α- and β-relaxations. This finding is the reversal of the early study of Yeo and

Eisenberg [78], who attributed the α-relaxation to the nonionic phase due to the minor

effect on hydration and the β-relaxation to the glass transition of the ionic phase.

Kyu and Eisenberg [81] demonstrated the effect of neutralization giving a

dominant effect on the shifting of tan δ to higher temperatures with increasing ionic

content thus leads to the reassignment of α-relaxation. Figure 2.12 shows the α-, β- and γ-

relaxations of Nafion neutralized with cesium salts as a function of neutralization.

Figure 2.12: Dynamic mechanical analysis (DMA) of Nafion–Cs with varying degrees of neutralization at 1 Hz. [Reprinted with permission from Can. J. Chem., 61, T. Kyu, M. Hashiyama, A. Eisenberg, pages 680 – 687, Copyright 1983, NRC Research Press].

In the absence of electrostatic interaction, the sulfonyl fluoride precursor displays

a single α-relaxation near 0 oC. As a result of this low temperature relaxation, the precursor may be easily melt–processed into thin films. On the other hand, upon

conversion to sodium sulfonate form, a profound shift in the α-relaxation to a temperature

near 250 oC was observed to that of ionomer. It is important to note that this form of

ionomer is no longer melt–processible due to strong Coulombic interactions that yield a

dynamic electrostatic network (i.e., a possible physically cross–linked system) which

persists to temperatures well above the melting point of the PTFE–like crystallites. In

addition to the dominant α-relaxation, a weak shoulder near 150 oC is observed and

assigned to the β-relaxation [81]. For Nafion in the acid form (H+), the α- and β-

relaxations are observed at temperature ~100 oC lower than that for the sodium sulfonate form which can be attributed to a reduction in the strength of the interactions between the

SO3H groups; comparably weak hydrogen bond interactions relative to the strong dipole–

- + dipole interactions between the SO3 Na groups [82].

Following the earlier work of Kyu and Eisenberg, Moore et al [83] investigated the effect of counterion type and size on dynamic mechanical properties of Nafion.

Attachment of alkylammonium ions to sulfonate groups of Nafion shows a systematic

alteration in the dynamic mechanical response of the ionomer with respect to the strength

of electrostatic interactions. Attachment of larger counterions i.e., tetrabutylammonium

was found to shift the α- and β-relaxations to an even lower temperature than the acid

form of the Nafion. Moreover, the magnitude of the β-relaxation is seen to increase

significantly to a level comparable to that of the α-relaxation. This behavior indicated the

weakening effect of the electrostatic interactions and plasticization effect of the large

organic counterions [84, 85].

In thermal studies, it was found that with increasing counterion radius, the

endothermic peak related to crystalline melting starts to show a decrease in temperature

which is in fair agreement with the dielectric measurement in demonstrating the effect of

different counterion [86].

Moore and Martin [87, 88] found that the solution casted Nafion film from

ethanol–water solutions at room temperature are brittle and soluble at room temperature

in a variety of polar organic solvents and exhibit low mechanical properties. These

properties are contrasting to the as–received Nafion membranes which are highly

flexible, tough and insoluble virtually in all solvents at temperatures below ~200 oC.

2.4. Recent Advances of Ionomer Blends and Nafion Blends for PEM Fuel Cell

Applications

Extensive research on replacing Nafion for the fuel cell membrane applications

have been undertaken over the past years. There have numerous attempts in designing

and synthesizing new proton electrolyte membranes, notably polyimide based ionomers, sulfonated thermoplastics poly(arylene ether sulfone), sulfonated poly(ether ketone ketone) and their blends such as poly(ether imide) [11–14].

Swier et al. [11] found that by blending sulfonated poly(ether ketone ketone)

(SPEKK) and poly(ether imide) (PEI), the mechanical stability of the sulfonated membrane was improved and reduced the degree of swelling in water. Nevertheless,

these blends had limited practical use due to the low proton conductivity of the parent

SPEKK.

Sulfonated poly(ether ketone ketone) (SPEKK) was developed by Lavorgna et al.

[12] as potential new membrane material in fuel cell applications. Formation of

cocontinuous morphology in the blends of SPEKK having two different ionic exchange

capacities (IEC) was found to affect the water swelling and sorption of the membrane.

The presence of percolative pathways in the cocontinuous structure also led to enhanced

proton transport.

Kim et al. [13] have came up with the fabrication of heteropolyacid and directly

polymerized sulfonated poly(arylene ether sulfone) to be used as membrane in fuel cell

applications. These authors incorporated heteropolyacid as the solid electrolyte in the

sulfonated membranes that showed a remarkable increase in operating temperatures

window. When being composited with heteropolyacid, Kim et al found that sulfonated

poly(arylene ether sulfone) can withstand higher operating temperatures (100oC –

130oC).

Synthesis of partially sulfonated poly(styrene) (sPS) and partially sulfonated

poly(2,6-dimethyl-1,4-phenylene oxide) (sPPO) were identified to resemble an excellent

mechanical properties of fuel cell membrane. Kim and Jung [14] developed blends of

SPS/sPPO and concluded that transport properties were fairly good, but still cannot

outperform the transport properties of Nafion. While these newly synthesized PEM

membranes show considerable improvement in suppression of swelling in water or

methanol, the proton transport properties of these materials turn out to be inferior than

Nafion when subjected to repeated durability and performance tests in the fuel cell

environment. Nafion has remained as the benchmark material for proton electrolyte

applications.

Kyu and Yang [20, 21] were among the first to use blends of Nafion and

poly(vinylidene fluoride) (PDVF) in order to study the effect of blends as a function of compositions and kinetics of phase separation. It was observed that blends of Nafion and

PVDF were partially miscible in liquid state. It was also found that the α- and β-form of

PVDF crystals develop during solvent casting in a manner dependent on the blend

39 composition and temperature. Phase diagram of Nafion and PVDF blends (Figure 2.13) established by Kyu and Yang through cloud point measurement showed a lower critical solution temperature (LCST), which can be used to acquire more controlled morphology.

In addition, incorporation of crystalline component in Nafion is expected to reduce the swelling of the membrane.

Effect of counterions in blends of Nafion and PVDF was shown to play a role in phase separation and crystal morphology. Landis and Moore [22] concluded that when radius of counterion increases, miscibility of blends and form of crystal morphology also started to change. As a conclusion, immiscibility of Nafion (Na+) and PVDF blends was

- + attributed to the strong Coulombic forces between the -SO3 Na dipoles which lead to aggregation of the ion pairs which act as strong electrostatic crosslinks. In contrast, neutralization of the Nafion with more weakly interacting cations, i.e., tetrabutylammonium (TBA+) cations, significantly reduces the strength of the dynamic electrostatic network at elevated temperatures and blends were found miscible in melt state. Smaller counterions attached to the sulfonic domains also affected the transport properties of Nafion.

Phillips and Moore [89] verified that membranes of Nafion prepared 50:50 wt% of Na+: TBA+, followed by conversion to the H+-form, showed a minimum water content yet relatively high proton conductivity. This behavior suggests that specific interactions during processing affect the organization of the ionic domains to yield consistent structures that significantly improved membrane transport properties.

Figure 2.13: Cloud point phase diagram of Nafion/PVDF blends at heating rate of 0.2 oC/min. [Reprinted with permission from Macromolecules, 23, T. Kyu, J. C. Yang, pages 176 – 182, Copyright 1990, American Chemical Society].

Cho et al [90] showed that native Nafion coated with blends of Nafion and PVDF shows lower methanol crossover compared to native Nafion. This proves that the addition

- of hydrophobic components such as PVDF to Nafion reduces the swelling of –SO3 hydrophilic domains in Nafion.

2.5. In-situ Impregnation Approach of Nafion Membrane for PEM Fuel Cell Applications

As mentioned earlier, two most important obstacles of Nafion membrane in proton electrolyte fuel cell applications are poor water management and lower operating temperature. Several attempts have been made to overcome these shortcomings.

A number of researchers have studied incorporation of polymers within Nafion,

referred to as Nafion supported membranes using several strategies, including in-situ

polymerization and sorption experiments. In-situ polymerization of poly(pyrrole) at the

surface of Nafion membrane has been demonstrated by Smit et al. [91]. This approach is

done by immersing Nafion in an acid electrolyte containing the monomer and then

polymerizing the membrane galvanostatically with the aid of an electrochemical cell. The

resultant layers and porous structures embedded onto Nafion were found to help reducing

the methanol permeability compared to nascent Nafion. In-situ polymerization of porous layers of poly(pyrrole) on the Nafion membrane through free-radical initiator has also been employed. [92]. Decrease in water sorption compared to nascent Nafion membrane was reported.

Nafion with polymerized layers of poly(2-acrylamido-2-methyl-1-

propanesulfonic acid-co-1,6-hexanediol propylate diacrylate-co-ethyl methacrylate), i.e., crosslinked PAMPS was found to reduced in its methanol permeability as compared to

Nafion. In addition to that, PAMPS polymerized Nafion exhibit relatively good conductivity properties at room temperature [93]. Bae et al. [94] composited Nafion membrane with poly(1-vinylimidazole) and showed a 25% increase in proton conductivity when compared to nascent Nafion. Impregnation and compositing of ionomer membrane with heteropolyacids in order to improve the water conservation properties at high temperature have also been conducted [12, 19]. However, these membranes only showed high proton conductivity at 100–115 oC for a short period and

tended to lose the heteropolyacids.

A simple and direct approach of impregnating Nafion with ionic liquids has been

carried out by several researchers [15-19] and specific interactions, e.g., inter-molecular hydrogen and ionic bonding have been proven to occur between these small molecules and the ionic clusters of Nafion through this act of infusion. Due to the high conductivity coefficient of the ionic liquids, the impregnated Nafion membrane with ionic liquid has been reported to exhibit good ionic conductivity at high temperature. Schmidt et al. [16] impregnated Nafion with hydrophilic ionic liquids and found that these impregnated membranes exhibited increased proton conductivity at 120 oC with tolerable swelling.

Mistry et al. [16] found that the Nafion membrane supported with ionic liquid of 1-butyl-

3-methylimidazolium bis(trifluoromethylsulfonyl)imide (BMI-BTSI) showed a remarkable increase of proton conductivity and improved thermal stability with temperature in anhydrous condition. These ionic liquids however, showing tendency to leach out as it is highly prone to dissolves back in its solvent. In addition to that, these ionic liquids also act as potent plasticizer to Nafion which ultimately lowers the mechanical strength and modulus of the impregnated membranes. Besides that, these ionic liquids are potent plasticizer thus demonstrated lower the mechanical strength of the membrane.

2.6. Promising Applications of Nafion

Besides PEM fuel cell applications, Nafion membrane has been utilized in

artificial muscles (or robotic arms), bio-medical and chemical sensors, and energy conversion devices by virtue of electrically or photochemically induced bending moment driven by ionic migration of cations inside Nafion [95-99].

Shahinpoor et al. [100, 101] found that the presence of hydrophilic domains in

Nafion is useful to disperse metal ions, which are subsequently reduced to corresponding metal atoms. In artificial muscles applications, Nafion needs to be hydrated to enhance the conductivity. The swelling of ion–exchange polymer solvate the counterions and the

- + fixed ionic groups (–SO3 ), thereby lowering the interaction between cations (e.g. H ,

Na+). Thus the solvation process increased the conductivity of the ion–exchange polymer

upon presence of electric field.

Figure 2.14: Artificial muscles applications.

Recently, Nafion also has found its niche in exploring through medical

applications. Multiuse planar amperometric biosensor modified with Nafion and/or

polyion have been investigated by Yuan et al. [102]. Nafion showed a significant effect

in eliminating the interference of ascorbic acid and uric acid. Nafion allowed the

permeation of hydrogen peroxide, but restricted the passage of anions (e.g., ascorbic acid

and uric acid) across the membrane. It was postulated that Nafion can effectively restrict

the anionic interference by adherence of anions to the surface of electrode thereby

reducing the fouling of oxidized ascorbic acid.

2.7 Homopolymer of Poly(vinylidene fluoride), PVDF

Homopolymer of poly(vinylidene fluoride) or PVDF was first introduced in 1961

as Kynar by the Pennsalt Chemical Corporation. This homopolymer, like many fluorine–

containing polymers, has superior weathering and chemical resistance. Apart from that,

PVDF also exhibits an exceptional variety of mechanical properties. By virtue of various polymorphic phases [103,104], the PVDF homopolymer possess extraordinary piezoelectric and pyroelectric properties.

Piezoelectricity may be defined as the ability of materials to generate an electric potential in response to an applied mechanical stress. Pyroelectricity on the other hand is

defined as the ability of certain materials to generate an electrical potential when

subjected to change in temperature. PVDF is known to exist in at least four different

crystal forms (α, β, γ, and δ) [103-106], depending on the processing and thermal

conditions.

Figure 2.15: Chemical structure of PVDF homopolymer.

Owing to its simple chemical structure (Figure 2.15), PVDF tends to have the most favorable torsional arrangement of with all –trans (TTTT), alternate –trans –gauche

(TGTG’) or –trans –gauche (TTTGTTTG’) conformation in one plane [107]. Figure

2.16 exhibits the schematic representation of conformations in PVDF. Modifications of sterochemical structure of PVDF by introduction of other kind of units such as a comonomer lead to improvement of physical properties of PVDF [108]. Thus a structural irregularity of –trans and –gauche linkage of monomers is introduced into the skeletal chain of PVDF. This type of irregularity affects the structural ability of the crystal modifications significantly.

Figure 2.16: Schematic representation of three conformation phases of PVDF.

The resulting material due to the copolymerization of VDF with trifluorothylene

(TrFE) shows ferroelectric phase transition between the ferro– and para–electric

crystalline phases [109]. Ferroelectricity is a physical property of a material whereby it exhibits a spontaneous electric polarization by the application of an external electric field, the direction of which can be switched between equivalent states. The existence of TrFE

monomer plays an important role in ferroelectric phase transition of the copolymer

PVDF-TrFE.

2.7.1. Copolymer Poly(vinylidene fluoride–co–trifluoroethylene), PVDF-TrFE

Due to the extensive applications of β–form crystals in ferroelectric properties, attempts have been made to induce β–form crystals from PVDF. Most common crystal form of PVDF is α–form, in which the unit cell contains two molecular chains of TGTG’ conformation, however for copolymer PVDF-TrFE, β–form of crystals is favored. The

β–form has the planar TTTT (all –trans) zigzag chain conformation packed in an orthorhombic unit cell and were found to be very polar in nature. Lando and Doll had

shown that the introduction of 17 mol% of trifluoroethylene allows the resulting

copolymer to adopt the all –trans conformation hence the β–form [103]. In recent years,

it has also been demonstrated that electron or gamma radiation can also introduce

structural disorder and defects in β–form of PVDF-TrFE, thus inducing paraelectric phase in otherwise ferroelectric polymer and eliminating the hysteresis in its polarization

response [110]. Similar effects can also be achieved by introducing a small amount of

ter–monomer; chlorofluoroethylene (CFE) into PVDF-TrFE copolymer [111].

PVDF-TrFE copolymer (Figure 2.17) was the first polymer to show ferroelectric phase transition [112-115]. For a 55 mol% of VDF and 45 mol% TrFE copolymer sample, the Curie phase transition temperature was found to be in the vicinity of ~65 oC,

which was interpreted as a result of a ferroelectric phase transition occurring within a

crystalline region. The Curie transition behavior of PVDF-TrFE copolymer was

dependent on several external factors such as VDF content, external electric field and

poling hydrostatic pressure [116-119]. Figure 2.18 shows the phase behavior of

copolymer PVDF-TrFE as a function of VDF content. At room temperature, PVDF-TrFE

exhibits ferroelectric properties and upon increasing of temperature, this copolymer tends

to undergo a ferroelectric to paraelectric transition, namely the Curie transition. As the

temperature increases, thermal motion, or entropy, competes with the ferromagnetic

tendency for dipoles to align. When the temperature rises beyond the Curie temperature,

there is a second-order phase transition and the system can no longer maintain a

spontaneous magnetization, although it still responds paramagnetically to an external

field.

Figure 2.17: Chemical structure of copolymer poly(vinylidene fluoride-trifluoroethylene), PVDF-TrFE.

Figure 2.18: Phase behavior of PVDF-TrFE as a function VDF content [redrawn from ref. 117].

Miscibility and ferroelectric transition behaviors for blends of PVDF-TrFE with poly(1,4–butylene adipate) (PBA) have been studied by Kim and Kyu [120]. PVDF-

TrFE and PBA were elucidated to be partially immiscible in amorphous state and exhibits lower critical solution temperature (LCST) which was located above the melting point of

PVDF-TrFE. Ferroelectric transition temperature of the blends increased upon the addition of PBA. A possible formation of ferroelectric phase directly from the melt at high PBA compositions was reported.

PVDF-TrFE and PMMA blends have also been clarified to be partially miscible.

Ferroelectric transitions are insensitive to the addition of amorphous components of

PMMA [121, 122]. PVDF-TrFE with poly(vinyl acetate) (PVAc) blends were miscible

49 at amorphous state and ferroelectric transitions were stabilized by presence of PVAc

[123]. Compensation between pyroelectricity and piezoelectricity along with ferroelectricity of PVDF-TrFE can also be done with inclusion of lead titanate within the

PVDF-TrFE matrix [124].

2.8. Hyperbranched Polymers/Supramolecules

Figure 2.19: Schematic diagram of dendrimer and dendron [redrawn from ref. 125].

The developments of dendrimers and hyperbranched polymers have expanded to all areas encompassing theory, synthesis, characterization of structures and properties, and investigations of potential applications. Dendrimers molecules are known to be repeatedly branched, monodisperse and usually consist of highly symmetric compounds and contains a single chemically addressable group that is called the focal point. The dendrimers are basically fractal-like branched molecules spanning out from a common

50 center, pointing to every direction (Figure 2.19).On the other hands, hyperbranched polymers are polydisperse and do not consist any initial focal point.

Dendrimers and hyperbranched polymers are one of the first nanoscaled synthetic materials that can be tailored with consistent structural characteristics producing controlled size, uniform size distribution, and surface functionality. Thus, these materials are optically isotropic as their branches are randomly oriented resulting in birefringence- free and have been synthesized for optical applications [125-128]. Moreover, these dendrimers can be tailored to create a precise composition and with alteration of the terminal functionality, surface control can be facilitated. Maruyama et al. has successfully synthesized a unique photocurable supramolecules through its pendant acrylate functional groups. As these supramolecules are still fairly new to the polymer field, their potentials for various applications including in the fuel cell applications are yet to be explored.

2.8.1. Photoinitiated Polymerization of Supramolecules

Photoiniated polymerization and/or UV–radiation curing of multifunctional monomers, or supramolecules being the most recent [129], have found a large number of applications in various industrial sectors. This technology is now commonly utilized to perform ultrafast drying, e.g., in the coatings, varnishes and adhesive industries. The same approach has also being used in producing the high–definition images that are required in the manufacturing microcircuits and printed plates.

The main advantage of using UV radiation is to initiate the chain reaction lies in the higher polymerization rates that can be reached under intense illumination, so that the

51 liquid to solid phase change takes place within a fraction of a second. Another distinct feature of light–induced reactions is that the polymerization occurs only in the illuminated areas, thus allowing complex patterns to be produced.

UV curing is typically a process that transforms a multifunctional monomer into a crosslinked polymer by a chain reaction initiated by reactive species, which are generated by UV radiation. Most monomers do not produce initiating radicals with sufficiently high yields when they are exposed to UV light, and thus a photoinitiator must be added to the formulation (i.e., vinyl group-containing mono/macromers). Once initiated, the chain reaction precedes much like in a conventional thermal polymerization, except for higher rates of initiation that can be reached by intense illumination.

The photoinitiated polymerization kinetics adapting the free radical mechanism usually occurs via three reaction steps, namely initiation, propagation and termination.

These steps may be described using these simple equations;

[ ] = I k [R ][M] Initiation step ∗ φ (2.4) d R ∗ a i [dt ] − = k [R ][M] k [P ] Propagation step ∗ (2.5) d P ∗ ∗ 2 i t [dt ] − Termination step = k [R ][M] k [M][P ] (2.6) d M ∗ ∗ dt i p where [R ] is the radical concentration− activated− from photoinitiator, [P ] is the total ∗ ∗ concentration of polymer radicals and [M] is the monomer concentration. k , k and k

i t p are the reaction kinetic constants of initiation, propagation and termination respectively.

2.9. Polymer Solutions and Blends

2.9.1. Flory–Huggins Theory for Polymer Solutions

Flory–Huggins theory [130, 131] was developed based on a lattice model and can be used to explain mixing behavior in polymer solutions. Figure 2.20 illustrated a lattice model. A solution of two components of small molecules is being considered. Only one molecule can occupy any given site in the lattice. Entropy of a system is the measure of randomness of the system. The increase in entropy due to mixing of two components can be given by Boltzmann’s relation;

∆Sm = k ln Ω (2.7)

N! ln Ω = (2.8) n1!n2!

where N = n1 + n2 , is the total number of molecules consisting of n1 molecules of

component 1 and n2 is molecules of component 2. Using Stirling’s approximation

(ln N!= N ln N − N ) Equation 2.7 can be written as;

∆Sm = k[(n1 + n2 )ln(n1 + n2 ) − n1 ln n1 − n2 ln n2 ] (2.9)

Figure 2.20: Lattice model where open circles represent solvent molecules and filled circles represent solute molecules. 53

n1 n2 Substituting x1 = and x2 = ; n1 + n2 n1 + n2

∆Sm = −k(n1 ln n1 + n2 ln n2 ) (2.10)

where x1 , x2 are mole fractions of component 1 and 2 respectively.

For a polymer solution, Equation 2.10 needs to be modified because the entropy of mixing decreases due to the long chain nature of polymer. For equal density system, mole fractions and volume fractions are equal. The entropy of mixing for a polymer solution is given by;

∆Sm = −k(n1 lnφ1 + n2 lnφ2 ) (2.11)

where φ1 and φ2 represent the volume fractions of polymer and solvent. Lattice model for

a polymer solution is shown in Figure 2.21.

Figure 2.21: Lattice model for a polymer solution.

For an ideal solution, enthalpy of mixing is zero, ∆H m = 0 . Polymers are regular

solution, in which ∆H m ≠ 0 . Flory provided following expression for enthalpy of mixing;

∆H m = zn1r1φ1∆ω12 (2.12) 54

where z, lattice coordination number, gives the number of neighboring cells to a given

cell, r1 is the number of statistical segment length of component 1 that occupies one

lattice site and ∆ω12 is internal energy of formation of an unlike 1-2 molecule pair.

ω +ω ∆ω = ω − 11 12 (2.13) 12 12 2

where ωij is the interaction energy for i-j contact. Substituting χ = zr1∆ω12 into Equation

2.12;

∆H m = χn1φ2 (2.14)

where χ is called as Flory–Huggins interaction parameter [132] and has a reciprocal dependence with absolute temperature. An empirical relationship between the interaction parameter and temperature can be given as [140];

B χ = A + (2.15) T

Most generalized form of Flory–Huggins interaction parameter is given by

Koningsveld to account for deviation from ideality [133];

 B  χ =  A + + C lnT (1+ Dφ + Eφ 2 ) (2.16)  T 

Hence a general Flory–Huggins equation for a binary system is the sum of

entropic and enthalpic terms given by;

∆Fm φ1 lnφ1 φ2 lnφ2 = + + χ12φ1φ2 (2.17) kBT r1 r2

2.9.2. Phase Field Theory for Crystal Solidification

For polymer blends consisting of a crystalline component, it is imperative to take

into account the crystal–amorphous interaction in addition to the conventional

amorphous–amorphous Flory–Huggins interaction. To achieve this goal, phase field

theory of crystal solidification may be combined with the FH theory. Phase field theory is

introduced to explain and represent the solidification of crystalline polymer blends that

undergo the solid (crystal)–liquid (melt) phase transition [134, 135]. f (ψ ) is the free

energy of crystallization of the crystalline component expressed by Landau expansion;

F(ψ ) ζ (T )ζ o (Tm ) 2 ζ (T ) +ζ o (Tm ) 3 1 4  f (ψ ) = = W  ψ − ψ + ψ  (2.18) kBT  2 3 4 

where W is the energy coefficient for the system to overcome the nucleation barrier.

ζ (T ) and ζ o (Tm ) represent the location of the nucleation hump and the free energy

minimum on the ψ -axis respectively. These solidification potentials are dependent on

temperature of crystallization.

Polymer crystallization may be explained in the framework of a phase field free

energy pertaining to a crystal order parameter, ψ . When ψ =0, crystal will tend to melt

and is assumed to have finite values close to unity in the metastable crystal phase.

However, when ψ =1 it is assume that crystal is at equilibrium limit. The crystal order parameter may be defined as the ratio of the lamellar thickness (l ) to the lamellar

thickness of a perfect polymer crystal (l o ) giving;

l ψ = (2.19) l o

The total free energy density of mixing of such crystalline–amorphous polymer

blend consists of the free energy density pertaining to crystal order parameter of the crystalline constituent weighted by its volume fraction (φ ) and the free energy of liquid–

liquid demixing,

φ (1−φ) 2 f (ψ ,φ) = φf (ψ ) + ln(φ) + ln(1−φ) +{χaa + χcaψ }φ(1−φ) (2.20) r1 r2

where χaa is the Flory–Huggins interaction parameter representing the amorphous–

amorphous interaction of the constituent chains in the isotropic melt, and χca represents

the repulsive crystal–amorphous interaction parameter. Note that the subscripts denote

the constituent 1 (crystal) and constituent 2 (amorphous). r1 and r2 correspond to the

number of statistical segment lengths of the respective components.

2.9.3. Phase Equilibrium

A stable system shows a negative free energy of mixing over the entire

composition range. From Equation 2.17, it can be seen that entropy term is always

negative due to the logarithmic terms involving volume fractions. Hence, the stability of

the system is decided by the enthalpic term, which in turn, is decided by Flory–Huggins

interaction parameter, χ aa . If χ aa is negative, the free energy becomes more negative.

This leads to the formation of a single well potential as depicted in Figure 2.22.

The free energy is concave upwards in the whole composition range. A system

always tries to minimize its free energy to become stable. In this regard if we consider a

mixture represented by two points having a composition φ1 and φ2 on the free energy

curve, the free energy of mixture is given by the point Q’, which is higher than the

minimum free energy where the system prefers to be. When χ aa is negative, this mixture

will dissolve and attain a minimum free energy represented by point Q.

Figure 2.22: Single well potential.

When χ aa is positive, the summation of enthalpic and entropic term will lead to

‘double well potential’ as illustrated in Figure 2.23(a). In this case any mixture having

compositions up to φ1 and beyond φ2 will form a single phase because in the composition

range 0 – φ1 and φ2 – 1 the lowest free energy is at φ1 and φ2 , respectively. However, in

between φ1 and φ2 the free energy curve is concave downwards instead of upwards.

Hence, the system will then phase separate into two phases the composition of which is

given by the line connecting two points B1 and B2 by a common tangent line. Any point

in this line represents the proportions for two phases that are formed. The free energy of

any single point between B1 and B2 is greater than that of a single phase system. Thus the 58

system is immiscible over this composition range. The point of contact of the double

tangent line to the free energy curve defines this composition range of immiscibility. The

slope of this line is called chemical potential which represent the first derivative of free

energy with respect to composition. Hence, the condition for phase equilibrium is the equality of chemical potential between two phases. This is mathematically represented as;

 ∂∆G   ∂∆G   m  =  m  (2.21)  ∂φ   ∂φ   φ1  φ2

The condition for phase equilibrium can be written by equating the chemical

α β potentials of each phase, μi as µi = µi where i represents the component and α and β

represents the two phases. The compositions at which the chemical potential matches are

called binodal points and their loci at different temperature forms a bimodal curve, which

separates a single phase region from a two phase region.

The free energy curve changes slope in between the two points B1 and B2 at S1

and S2. These two inflection points are called spinodal points. In between B1 and S1 and

B2 and S2, the system is stable to small changes in local composition and energy is required to drive the system to phase separate. Hence these regions are called metastable

regions. Mathematically, the condition for spinodal point is the second derivative of free

energy with respect to composition be zero (i.e slope at the inflection point = 0). This is

given by;

∂ 2∆G m = 0 (2.22) ∂φ 2

The loci of the spinodal points at different temperature form the spinodal curve. The

region between S1 and S2 form unstable region which undergoes phase separation 59

instantaneously. Figure 2.23(b) shows a binary phase diagram of a polymer solution

which is plot of temperature as a function of composition.

For a small molecule system both r1 and r2 are equal to 1, hence the entropy of mixing is very large compared to polymer solvent system. But for a polymer–solvent

system r1 is much larger whereas r2 is equal to 1 and for polymer blends both r1 and r2 are

large numbers. The entropy of mixing for a polymer–solvent and polymer–polymer system is much smaller than for a small molecule system.

(a)

(b)

Figure 2.23: (a) Double well potential and (b) binary phase diagram for a polymer solution.

2.9.4. Dynamics of Phase Separation Process

Phase separation occurs when a binary system is quenched from a single phase to

a two–phase by thermal process [136]. The dynamics of phase separation can be

classified into two regimes: nucleation and growth (NG) and spinodal decomposition

(SD). Quenching into the metastable region usually triggers the nucleation and growth if

there is a large concentration fluctuation while any small fluctuation in the unstable

region will be amplified to initiate phase separation through spinodal decomposition.

Figure 2.24 shows different regimes of phase separation.

Phase separation through nucleation and growth (NG) will occur upon thermal

quench into region I and III from single phase region. In region II, phase separation

occurs through SD. In region I, NG mechanism leads to the nucleation of polymer–poor

β α phase, φ p dispersed in a polymer–rich matrix phase, φ p , while in region III, polymer–

α β rich phase, φ p , is dispersed in a polymer–poor matrix phase, φ p . In region II, phase

separation occurs through SD thus leading to bicontinuous or cocontinuous structure.

2.9.4.1. Nucleation and Growth (NG) Mechanism

Nucleation and growth (NG) mostly occurs in the phase transition processes such

as liquid–solid transition, liquid–vapor transition and crystallization. In the metastable

region, the excess free energy gives the initial unstable embryos. In these regions, the system initially needs to overcome a certain energy barrier. Thus a finite fluctuation is necessary for the process to occur. The nucleation and growth require activation energy to originate the process and critical size has to be reached before it becomes energetically

favorable to grow. These activated sites later serve as nuclei subsequently grow on

individual nuclei, and the liquid–liquid phase separation finally leads to the droplet in the surrounding polymer matrix. Figure 2.25 illustrates the NG mechanism.

Nucleation and growth mechanism is characterized by downhill diffusion and the

two separated phases are always at equilibrium. These droplets will then grow only in size with time.

Figure 2.24: Phase separation process regimes.

Figure 2.25: Nucleation and growth (NG) mechanism.

2.9.4.2. Spinodal Decomposition (SD) Mechanism

Spinodal decomposition process occurs usually in the second–order phase transition involving the phase separation of polymer blends. The fundamental differences of SD as compared to the NG are: the concentrations do not have a step change to form an embryo, instead, the concentration fluctuations are typical causes of spinodal decomposition associated with phase separation. The SD is unstable at infinitesimal concentration fluctuations and thus it occurs spontaneously without requiring any activation energy.

The process of SD goes through three stages according to time independence on concentration fluctuations; early, intermediate, and late stages. The growth rate of concentration fluctuations of the early stage of SD can be characterized by a linearized diffusion equation, which was first solved analytically by Cahn and Hilliard.

The SD mechanism is characterized by uphill diffusion where both wavelength and amplitude grows with elapsed time as illustrated in Figure 2.26. The phase

boundaries are diffused and gradually become sharper with time. Finally, it evolves to an

interconnected morphology.

In SD, both compositions and size depend on time, while in NG, the composition

of the separated domains is constant and only the size and size distribution of the nuclei

change with time. It is worth noting that NG and SD process can only be distinguished

during initial stage where the linearized Cahn–Hillard theory is operative.

Figure 2.26: Spinodal decomposition (SD) mechanism.

2.10. Fundamentals of the Impedance Spectroscopy

The impedance spectroscopy is an experimental technique that involves imposing

a small sinusoidal (AC) voltage or current signal of known amplitude and frequency to an

electrochemical cell and monitoring the AC amplitude and phase response of the cell.

The ratio and phase–relation of the AC voltage and current signal response is the complex impedance, Z(ω) or the resistivity of a cell with frequency.

When a sinusoidal current signal of amplitude, I(amps) and frequency, ω(rad/sec)

imposed an electrochemical cell, the input of the AC signal voltage can be defined as;

V = V e (2.23) iωt o with current, I is denoted as;

I = I e ( ) (2.24) i ωt+θ o where θ is the phase angle in radians and ω is the angular frequency defined as ω = 2πf

with f is the frequency in cycles per second in Hz; (Hz = 1/sec). According to Ohm’s

Law, impedance, Z(ω) is defined as;

( ) Z( ) = (2.25) ( ) V ω ω I ω The complex impedance can be written in the rectangular form;

Z ( ) = Z ( ) iZ"( ) (2.26) ∗ ′ whereω Z’(ω) ωis the− real ωcomponent and Z”(ω) is the imaginary component with i = 1.

The impedance absolute magnitude or |Z| is defined as; √−

|Z| = ( ) + (Z" ) (2.27) 2 2 and; � Z′

Z ( ) = Re(Z) = |Z|cos (2.28) ′ Z"(ω) = Im(Z) = |Z|sin θ (2.29) withω phase angle, θ; θ

Phase angle, = tan (Z"/ ) (2.30) −1 When θexpressed in twoZ′ -dimensional plane, absolute complex impedance, |Z| can be presented as in Figure 2.27 where the vertical and horizontal axes are called real and imaginary impedance plane. 65

Figure 2.27: Two-dimensional plane of absolute complex impedance, |Z|.

Polarization capacitance in Farads can be calculated at the maximum frequency,

fmax as;

C = (2.31) 1 p 2πfmaxRp where angular frequency, ω = 2πf with f is the frequency measured in Hz and Z = Rp

where the value of Rp is read from the real impedance axis. Polarization resistance, Rp is equal to 2Rmax where Rmax is the resistance value that corresponds to the maximum

frequency of the Cole-Cole plot obtained [137].

The electrochemical reactions normally consist of electron transfer at the

electrode surface. These reactions mainly involve electrolyte resistance, adsorption of

electroactive species, charge transfer at the electrode surface, and mass transfer from the bulk solution to the electrode surface. Each process can be considered as an electric component or a simple electric circuit. The whole reaction process can be represented by

an electric circuit composed of resistance, capacitors, or constant phase elements

combined in parallel or in series.

2.10.1. Presentation and Interpretation of Impedance Data

Generally, the impedance spectrum of an electrochemical system can be presented

in Nyquist (or Cole-Cole) and Bode plots as a function of frequency. A Nyquist plot is

displayed for the experimental data set Z; where Z” versus Z’ measured at different

frequencies with lower frequency is plotted at the right side. The complex impedance

component of log Z’ or Z” versus log frequency are also an alternative graphical

presentations of the impedance data.

A Bode plot, on the other hand, is a representation of the secondary impedance

data output. There are two types of Bode diagram, log |Z| versus log ω and θ versus log

ω, describing the frequency dependencies of the modulus and phase, respectively. A

Bode plot is normally depicted logarithmically over the measured frequency range

because the same number of points is collected at each decade.

Both Nyquist (or Cole-Cole) and Bode plots usually start at a high frequency and

end at a low frequency, which enables the initial resistor to be found more quickly.

Figure 2.28 shows the typical impedance plots for pure resistor, pure capacitance and typical electrochemical cells which consist of a resistor in series with a parallel circuit containing a capacitor and a resistor respectively. Rs is the Ohmic contact resistance between the electrolyte membrane and the catalyst layer; known as the solution resistance. The diameter of the semi-circular plot, denoted by Rp is the polarization

67 resistance, compensating the Ohmic and kinetic resistance [138]. The magnitude of Rp represents the true resistance from which the conductivity value may be evaluated. Thus, determination of the conductivity value in typical electrochemical cells with semi-circular

Nyquist (or Cole-Cole) plot is given by;

Conductivity, (S cm) = (2.32) d σ ⁄ RpA where d is the thickness of the membrane used in the electrochemical cell (cm), A is the

2 contact cell area (cm ), and Rp is the cell resistance (polarization resistance), compensating the Ohmic and kinetic resistance. The value of Rp can be obtained from the real impedance, Z’ intersection point of the Nyquist (or Cole-Cole) plot or the diameter of the semi-circular plot.

Figure 2.28: Nyquist (or Cole-Cole) and Bode plot of pure resistor, pure capacitor and typical electrochemical cell containing a resistor in series with a parallel capacitor and resistor respectively [138].

CHAPTER III

MATERIAL SPECIFICATIONS AND EXPERIMENTAL TECHNIQUES

3.1. Materials

Nafion 115 was selected as the focal interest of the present studies due to its excellent chemical, physical as well as the proton transport properties. Ferrolectric

copolymer named poly(vinylidene fluoride-trifluoroethylene) or PVDF-TrFE and two different kinds of supramolecules, i.e., photocurable hyperbranched polyesters (HBPEAc-

COOH) and waterwheel supramolecules with terminal hydroxyl (Noria) and tert- butyloxycarbonyl (Noria-BOC) functional groups were selected to modify the nascent

Nafion via two different approaches: physical blending and in-situ impregnation. The

selection criteria for perfluorinated ionomer membrane, copolymer and supramolecules

along with their physico-chemical properties are highlighted in this chapter.

Summarization on the physical properties of all materials can be found in Table 3.1-3.4.

3.1.1. Perfluorinated Ionomer Membrane, Nafion

Commercialized extruded perfluorinated ionomer membrane, Nafion 115, utilized

in these studies, was received from Fuel Cell Store Inc. (San Diego, USA) in the original

protonated form (H+). The nomenclature of Nafion 115 implies that the ionomer

membrane has an equivalent weight of 1100 g/mol and thickness of 5 mils (~110–127

µm). Equivalent weight (EW) is defined as number of grams of dry Nafion per mole of sulfonic acid group. The equivalent weight is related to the property more often seen in the field of conventional ion exchange resins, which usually are described in terms of ion exchange capacity, IEC = 1000/EW [139]. For Nafion, it is not applicable to denote the molecular weight due to the variability of processing parameters and solution morphology. Some of the chemical and physical properties of Nafion are tabulated in

Table 3.1. Nafion consists of poly(tetrafluoroethylene) or TEFLON backbone polymer with pendant side chains that are terminated with sulfonic acid groups (Figure 3.1) where x = 5–13.5, y = 1, z ≥ 1 and 100

38% dry weight basis for 1100 EW membrane. Absorption of a polar solvent, i.e., water, will free the cation disassociated with terminal pendant sulfonic acid groups to move within the polymer matrix while the anions maintain a bond to the fluorocarbon backbone. It is this property of Nafion which allows the transport characteristics of cations, making Nafion perm-selective ion exchange membranes.

Figure 3.1: Chemical structure of Nafion.

3.1.2. Copolymer Poly(vinylidene fluoride-trifluoroethylene), PVDF-TrFE

Copolymer of PVDF-TrFE (Figure 3.2) with 37% trifluoroethylene monomer was used in the study of physical blending with Nafion. PVDF-TrFE possessed two endothermic peaks in the thermal analyses characterization, i.e., 73 oC and 163 oC. The first endotherm peak is denoted as the Curie temperature wherein the material starts to lose spontaneous polarization and becomes paraelectric, and the later part is the crystal melting temperature of PVDF-TrFE.

Figure 3.2: Chemical structure of poly(vinylidene fluoride-trifluoroethylene), PVFD- TrFE.

PVDF-TrFE copolymer is the resulting material of copolymerization of VDF with trifluoroethylene monomer. This copolymer exhibits interesting ferroelectric and piezoelectric properties due to the stereo-regularity of the planar –trans conformational structure. Ferroelectricity is the spontaneous electric polarization of a material that can be reversed by the application of an external electric field, while piezoelectricity is defined as the accumulation of charge in piezoelectric material, i.e., PVDF-TrFE in response to mechanical strain. Insertion of trifluoroethylene monomer promotes the planar TTTT (all

-trans) zigzag conformation rather than TGTG’ (alternate -trans-gauche) and

TTTGTTTG’ (-trans-gauche) in one plane as observed in homopolymer poly(vinylidene

fluoride) (PVDF). Thus the structure irregularity of PVDF-TrFE affects the structural ability of crystal formation and modification significantly.

The most common crystal form of PVDF is α–form, in which the unit cell contains

two molecular chains of TGTG’ conformation; however, for copolymer PVDF-TrFE, β–form

of crystals is favored. The β–form has the planar TTTT (all –trans) zigzag chain

conformation packed in an orthorhombic unit cell and was found to be very polar in nature.

Due to the interesting ferroelectric and piezoelectric properties possessed, copolymer PVDF-

TrFE was chosen in the first part of the study.

3.1.3. Photocurable Hyperbranched Polyester with Pendant Carboxylic Acid Functional

Groups, HBPEAc-COOH

The HB polyester was synthesized from polyaddition of bisphenol A diglycidyl

ether (BPGE) with trimesic acid (TMA) and methacrylic acid (MA), by adding cis-

1,2,3,6-tetrahydrophthalic anhydride (THPA), triphenylphosphine (TPP), and

hydroquinone (HQ) having pendant methacryloyl and carboxyl groups. The detailed

synthetic scheme of these HB polyesters may be found in a previous paper [140]. The

number average molecular weight and polydispersity of the HBPEAc-COOH as

measured by GPC (Model 1515, Waters) using tetrahydrofuran (THF) with polystyrene

standard were Mn=6,800 and Mw/Mn=1.42, respectively. The uniqueness of the above

HB polyester is in its photocurability through free radical photopolymerization of the

acrylate double bonds and carboxylic acid functionality that affords ionic interactions

with its counterpart. Figures 3.3 and 3.4 illustrate the chemical structure of HBPEAc-

COOH and the photoinitiator used in the study. 73

Figure 3.3: Chemical structure of hyperbranched polyester, HBPEAc-COOH.

Figure 3.4: Chemical structure of photoinitiator, Irgacure 907.

3.1.4. Waterwheel Supramolecules with Terminal Hydroxyl Functional Groups, Noria

and Tert-Butylhydroxyloxycarbonyl, Noria-BOC

Supramolecules, Noria were synthesized via condensation reaction of resorcinol as a difunctional compound with 1,5-pentanedial [(CH2)n(CHO)2] where n=3, as a

tetrafunctional compound. A similar approach to that employed for the synthesis of

calixarenes was used to give a unique double-cyclic ladder-type oligomer with a central hole which resembles a waterwheel structure as shown in Figures 3.5(a) and (b).

The reaction of resorcinol and 1,5-pentanedial was carried out ethanol at 80 oC in

the presence of concentrated hydrochloric acid (HCl). The initially homogenous reaction

mixture afforded an insoluble product after 48 h. The reaction mixture was poured in the

methanol, and the precipitated product was washed with diethyl ether several times and

dried in vacuo at 60 oC for 48 h. The supramolecule Noria is soluble in aprotic, highly polar solvents such as dimethylformamide (DMF), dimethylsulfoxide (DMSO), dimethylacetamide (DMAc) and 1-methyl-2-pyrrolidinone (NMP).

Noria derivatives were then converted into tert-butyloxycarbonyl, -COOC(CH3)3

with di-tert-butyldicarbonate to increase the solubility properties of the supramolecules.

The resultant of the conversion, Noria-BOC, yields in having different terminal functional groups as illustrated in Figure 3.5(c). These supramolecules, Noria-BOC, are highly soluble in a large number of solvents such as methanol, ethanol, acetonitrile, toluene and cyclohexane. The synthesis scheme of the supramolecule waterwheel may be found in the previous literature [141].

Both supramolecules were used in the study as to provide comparison between the two different terminal functional groups of polar hydroxyl groups and tert- butyloxycarbonyl groups.

(a)

(b)

(c)

Figure 3.5: (a) Chemical structure of Noria, which resembles awaterwheel. Each ring consists of alternating resorcinol and methylene units, and the two rings are connected through six resorcinol units, (b) Chemical structure of resorcinol and 1,5-pentanedial and its representative symbols as shown in (a), (c) Chemical structure of waterwheel supramolecules with tert-butyloxycarbonyl, Noria-BOC [141].

Table 3.1: Physical properties of perfluorinated ionomer membrane Nafion used in the present study.

a, b Manufacturer’s data. d, e, f Experimental findings obtained in the study.

Table 3.2: Physical properties of poly(vinylidene fluoride-trifluoroethylene), PVDF-TrFE used in the present study.

a Synthesis data. b,c,d Experimental findings obtained in the study.

Table 3.3: Physical properties of hyperbranched polyester, HBPEAc-COOH used in the present study.

a, b, c, d, e Experimental findings obtained in the study.

Table 3.4: Physical properties of waterwheel supramolecules, Noria and Noria-BOC, used in the present study.

a, b, c, d, e Experimental findings obtained in the study.

3.2. Experimental Methods

All Nafion membranes were thoroughly dried in vacuum oven to ensure all traces

of water were removed before physical blending and impregnation processes took place.

As-received Nafion 115 membranes were pretreated to ensure full conversion to the acid form as described elsewhere [142]. Nafion membranes were first pretreated in 2 %

o hydrogen peroxide (H2O2) solution at 80 C for 2 hours and subsequently rinsed with

deionized water. Nafion membranes were then immersed in 0.1M hydrochloric acid

(HCl) for 48 hours at room temperature and finally rinsed again with deionized water to

remove excess hydrochloric acid. Nafion membranes were then dried in a vacuum oven

and kept in dessicator prior to each sample’s preparation process.

For the membrane preparation of Nafion/PVDF-TrFE membrane blends, the

Nafion membrane was cut into tiny pieces and dried at 100 oC in vacuum oven for 48

hours in order to eliminate residual moisture prior to blending. Nafion membrane and

PVDF-TrFE were dissolved separately by rigorous stirring in dimethylacetamide

(DMAc) at 120 oC and at room temperature, respectively. The solutions were then mixed

forming a homogenous mixture at various blend ratios while maintaining polymer

concentration at 5 wt% followed by solvent removal under vacuum at 150 ºC for 24

hours.

On the other hand, in-situ impregnation process was carried out with weighed

neat Nafion membranes which were pre-swollen in methanol for 24 hours and then

impregnated in the respective supramolecules solution for 48 hours at room temperature.

Only for impregnated Nafion with HB, the curing agent (Irgacure 907) in an amount of 3

wt % with respect to that of HBPEAc-COOH was added to the above solution. The

Nafion to supramolecule ratios were maintained at 98/2, 97/3, 95/5, 90/10 and 80/20. The

ratios described above correspond to the feed compositions, which will be hereafter

referred to as the ‘feed ratio’. Impregnated membranes were then removed from the supramolecule solutions and were gently blotted with tissue paper (Kimwipes) to remove

excess solution on the surface. Only for the HBPEAc-COOH impregnated membranes,

photocuring was subsequently carried out in NuLine curing chamber (model NL22-8C,

2 o 90mW/cm ) at ~85 C for 5 minutes, which is slightly above the Tg of the HB polyester.

Blotting and curing procedures were conducted in a dark room so as to prevent accidental

curing of the membranes. These supramolecule impregnated Nafion membranes were

then dried in avacuum oven at 70 oC for 12 hours. These modified membranes were kept in a dessicator prior to use.

3.2.1. Thermo Gravimetric Analysis (TGA)

Thermal stability studies were conducted using a thermo gravimetric analyzer

(TA Instruments, Model 2950). The recommended amount of 5 mg of the samples was used for each run. As-received Nafion, various Nafion/PVDF-TrFE blends, as well as

Nafion impregnated membranes were subjected to TGA analysis at a heating rate of 10

ºC/min from 25 ºC to 650 ºC in a nitrogen atmosphere with a flow rate of 90 mL/min.

The temperature at which the weight loss was more than 5% of the initial weight was regarded as the degradation temperature. Care was taken to ensure that the temperature

employed for any further characterization of materials was selected below the

degradation temperature.

3.2.2. Differential Scanning Calorimetry (DSC)

Thermal analysis of the samples was conducted at a heating rate of 10 ºC/min

using a TA Instruments (Model 2920) differential scanning calorimeter calibrated for

temperature and enthalpy using indium standard having a melting point of 165.5 °C.

Samples weighing 7−10 mg were sealed in aluminum hermetic DSC pans using a

crimping device. A sealed empty aluminum pan was used as a reference. Nitrogen gas was purged with a flow rate of 80mL/min to the chamber as to maintain an inert atmosphere. The first DSC run was employed in order to eliminate thermal history. The data from second run DSC were used for further analyses.

3.2.3. Polarized Optical Microscopy (POM)

The samples for POM experiments were cast on glass substrates forming a thin film (~10 µm). All samples were stored in desiccators for further use. An optical microscope (BX60, Olympus) equipped with a 35 mm digital camera (EOS 400D,

Canon) and a hot stage (TMS 93, Linkam) were used for the POM study. The images were captured at 200X magnification.

3.2.4. Fourier Transformed Infrared (FTIR) Spectroscopy

FTIR spectra were collected in the attenuated total reflected mode (ATR) on a

Nicolet 380 (Thermo Scientific) spectrometer with an average of 32 scans and a spectral

resolution of 4 cm-1 at ~100 oC. Given that each of the components show characteristic

affinity towards moisture, care was exercised to minimize the effects of moisture with the

aid of a temperature cell by heating the samples to 120 °C, and spectra were recorded after equilibrating at 100 °C to assure the absence of moisture and residual solvent.

To estimate the amount of supramolecules present in the in-situ impregnated membranes, integration of the area under the FTIR curves of the corresponding characteristics bands, e.g., C=O (carbonyl) or C=C (aromatic) was done. The area under

the curve was normalized to the respective pure supramolecules. The estimated

supramolecules amount was calculated by multiplying the percentage of each

supramolecule present in membrane with the initial amount in feed and can be calculated

in what follows:

Supramolecules in membrane, x % = × 100 (3.1) AModi�ied Membrane ASupramolecules Actual weight of supramolecules = x % × initial amount in feed (3.2)

where A is the area under the curve from corresponding characteristics

Modi�ied Membrane bands of that impregnated Nafion membranes and A is the area under the

Supramolcules curve from the corresponding characteristics bands of pure supramolecules. Physical

weighing measurements with a membrane dimension of 5 cm x 5 cm were undertaken in

order to counter-check the quantified amount of incorporated suprmaolecules in Nafion

membranes obtained from the value of integrated area under the curve of the FTIR

spectra.

3.2.5. Solid State Nuclear Magnetic Resonance (SSNMR)

1H-13C and 19F-13C cross polarization (CP) and 1H direct polarization (DP) magic

angle spinning (MAS) NMR spectra were collected on a Varian (Model: NMRS 500,

11.7 T) spectrometer operated at 125.62 MHz for 13C and using a Varian narrow-bore

triple-resonance T3 MAS NMR probe. Samples were packed into 4 mm zirconia rotors

and spun at 10 kHz. The 13C and 1H chemical shifts were referenced to

hexamethylbenzene (17.3 ppm; methyl) and adamantane (1.76 ppm). The 1H-13C and 19F-

13C CP/MAS data were collected under continuous wave (CW) proton and TPPM

fluorine decoupling, respectively. A 90° pulse width of 4 μs was used for all nuclei.

Recycle delays of 2, 2.5 and 1 s were used for 1H-13C CP, 19F-13C CP and 1H DP

experiments with contact times of 3.5 and 1.5 ms for the 1H-13C and 19F-13C CP

experiments. For the 1H spectra, 32 transients were collected. Samples of neat Nafion-

acid and impregnated membrane were stored in a dessicator and only exposed to ambient

atmosphere while loading. The membranes were cut into a rectangular shape and rolled to

fit into the rotors.

3.2.6. Wide Angle X-Ray Diffraction (WAXD)

Wide angle x-ray diffraction (WAXD) measurement was conducted at room temperature using a Bruker x-ray generator (AXS D8) equipped with a copper target tube

and a two–dimensional detector. The x-ray generator was operated at 40 kV and 40 mA

using monochromatized CuKα radiation with a wavelength of 1.5418 Å. A customary

detector-to-sample distance of 15 cm was employed with 2θ covering up to 30°.

3.2.7. Small Angle X-Ray Scattering (SAXS)

2D small angle x-ray scattering (SAXS) analyses of impregnated membranes

were characterized using a 18kW rotating anode x-ray generator (MicroMax-002+)

equipped with a Cu tube operated at 45 kW and 0.88mA. The wavelength of the x-ray

beam was 0.15418 nm. The standard zero pixel of the 2D SAXS was calibrated using

silver behenate with first-order scattering vector, q=10.76 Å-1 where q=4π sin θ/λ with λ is the wavelength and 2θ is the scattering angle. For all scans, the background scattering was subtracted from the sample scans. All membranes were isothermally treated at 100

oC for 20 minutes prior to each scan as to eliminate possible moisture absorption.

3.2.8. Water Uptake

Prior to the water uptake measurement, neat Nafion and impregnated Nafion

membranes were dried in vacuum oven until constant weight was reached, which was

taken as mass in dry state. All membranes were immersed in 10 ml deionized water at

room temperature and equilibrated for 72 hours for water uptake measurement.

Membranes were then removed and blotted with dry tissue paper (Kimwipes), and mass

in wet state was recorded. Water uptake in percentage was determined as follows;

Water uptake = (3.3) Mwet −Mdry Mdry 86

The number of water molecules per mole of sulfonic group, can be calculated as

described below; λ

( )× (moles H O/SO ) = (3.4) × Mwet−Mdry Meq Na�ion 115 2 3 λ Mdry MH2O where M is mass of membrane in wet state, M is mass of membrane in dry state,

wet dry M is the equivalent weight of Nafion 115 and M is the molecular mass of

eq Na�ion 115 H2O water.

3.2.9. Ion Exchange Capacity (IEC)

Determination of ion exchange capacity of the impregnated membranes was

completed with a titration method wherein weighed membrane samples were soaked in

sodium chloride (NaCl) solution for approximately 120 hours to ensure complete

conversion of proton to sodium cations. Solutions were then titrated with standardized

sodium hydroxide (NaOH, 0.01N) using a phenolphthalein indicator to the end point. IEC

values in meq g-1 were calculated using the volume of NaOH used to neutralize the

solution in accordance to the equation below;

× Ion Exchange Capacity (IEC), meq g = (3.5) −1 Vtitrated NNaOH mmembrane where V is the volume (mL) of NaOH titrated to neutralize the solution, N is

titrated NaOH the standardized normality of NaOH and m is the mass (gram) of impregnated

membrane membranes or neat Nafion. Titration of each membrane sample was repeated five times,

and the mean IEC values were reported.

3.2.10. Dynamic Mechanical Analyses (DMA)

Dynamic mechanical properties were measured in a sinusoidal tension mode

using Pyris Diamond DMA (Perkin Elmer) equipped with a liquid nitrogen cooling

controller (Seiko Equipment). Dried pure Nafion and impregnated membranes were cooled down to -140 oC and maintained isothermally for 5 minutes, then heated up to 200

oC at a heating rate of 1 oC/min. Frequency of the dynamic mechanical measurements

was fixed at 1 Hz and measurements were carried out under nitrogen environment.

3.3. Conductivity Properties

3.3.1. Proton Conductivity/Fuel Cell Measurements

Proton conductivity characterization of neat Nafion 115 and impregnated

membranes were conducted using the AC impedance fuel cell test system (Scribner,

Model 850e) equipped with an adjustable reactant humidifier unit. Figure 3.6 shows the

flow diagram of the AC impedance fuel cell setup equipped with the adjustable reactant

humidifier unit. Prior to the conductivity measurement, a spray gun (Badger Gun, Model

No. 150) was used to spray the membranes with platinum supported carbon solution

catalyst. Sprayed membranes with total weight catalyst of 2 mg and 1 mg on the cathode

and anode side respectively were then manually compressed with 13.8 Mpa at 60 oC for

20 seconds. Hydrogen (H2) and compressed air were used as reactants with flow rate of

0.2 L/min. Contact cell dimension used was 5 cm2 and membranes were tested at

frequency range of 0.1 Hz–10 kHz. Measurements were conducted at two different

relative humidity (RH) levels: 100% RH and 74% RH ranging from 30 oC to 115 oC.

Complex impedance data obtained were analyzed using ZView program. Conductivity value can be calculated in accordance to eq (2.32) as stipulated in the previous chapter.

Proton conductivity measurement on all samples was conducted at the moderate current range = 0.5 A, which is within the Ohmic polarization or voltage loss due to the ionic resistivity within the electrolyte and electronic current occurred from the electrodes.

Polarization curves shown in Figure 3.7 of 95/5 Nafion/Noria impregnated membranes at different temperature with 74% RH obtained from the study illustrated several regions of voltage loss within the cell upon feeding with H2/air reactant.

Figure 3.6: Flow diagram of the AC impedance fuel cell setup.

Figure 3.7: Polarization curves of 95/5 Noria/Nafion impregnated membranes at 74% RH obtained from the present experimental findings signifying the three regions of voltage loss: i) voltage loss due to the activation reaction resistivity, ii) voltage loss attributed to the Ohmic polarization and iii) voltage loss due to the concentration polarization.

3.3.2. AC Electrical Impedance Measurements

Electrical impedance measurements were conducted using the Schlumberger

Impedance/Gain-Phase Analyzer over the frequency range of 1 Hz–100 KHz and

equipped with a heating stage which is connected to the Omega CN8500 temperature

controller where membranes were heated from 27 oC to 150 oC. Samples were

sandwiched between the indium tin oxide (ITO) cells with cell contact area of 1 cm2 and

thicknesses of membranes ranging between 90–110 μm. The capacitive and electron

conductivity values may be determined in agreement to the eqs. (2.31) and (3.32) as

mentioned in the preceding chapter. Figure 3.8 demonstrates the typical Nyquist (or Cole-

Cole) plot, its equivalent circuit and the manifestation of Rmax, Rp and fmax used in

determining the capacitive and conductivity values of the membranes.

Figure 3.8: Typical Nyquist (or Cole-Cole) plot, it equivalent circuit and the manifestation of Rmax, Rp and fmax used to calculate the polarization capacitance, Cp.

CHAPTER IV

BLENDS OF NAFION AND POLY(VINYLIDENE FLUORIDE-

TRIFLUOROETHYLENE), PVDF-TrFE COPOLYMER

4.1. Introduction

Perflourinated ionomer membrane, commercially known as Nafion, was first developed by DuPont for chloro-alkaline cell applications [5,6]. Nafion has been the benchmark for proton electrolyte membranes (PEMs) by virtue of its outstanding chemical and thermal stability [6]. Nafion perfluorinated ionomer also possesses an excellent perm-selectivity towards proton and immensely insulates electrons at the anode from crossing over to the cathode, i.e., one of the most crucial requirements for an electrolyte membrane in fuel cell applications.

The molecular structure of Nafion is made up of poly(tetrafluoroethylene) backbone and perfluorovinyl ether side chain terminated by sulfonic acid groups [5]. The aggregation of sulfonic acid forms the microphase separated ionic clusters, and these ionic clusters determine the ion transport properties of Nafion [7,8].

One of the major shortcomings of Nafion is the dimensional instability caused by excessive swelling in polar solvents such as water due to the hydrophilic nature of the terminating sulfonate groups of Nafion. Gore Associates, Inc. developed Teflon grid supported Nafion membranes, in which the mesh of hydrophobic Teflon gives

mechanical support to the films with better dimensional stability to suppress the

excessive swelling of Nafion.

Simple solution blending is an alternative approach to introduce the hydrophobic

fluorocarbon polymer into Nafion to improve swelling suppression. Kyu et al. [20,21]

were among the first to blend Nafion with fluorocarbon polymer such as poly(vinylidene

fluroride) (PVDF). The effect of solvent casting conditions on phase morphology and

miscibility of Nafion/PVDF blends was demonstrated, exhibiting the lower critical

solution temperature (LCST) phase behavior above 210 oC. Landis and Moore [22] found

that larger counterions of Nafion such as tetrabutylammonium (TBA+) lowered the

crystallization rate of PVDF molecules, thereby enhancing the miscibility of

Nafion/PVDF blends. However, it is desirable for the blends to possess the unremitting

percolated phase separated morphologies because the hydrophobic fluorocarbon domains can render swelling suppression, while the hydrophilic Nafion can provide interconnected

channels for ion and/or proton transports.

The main objective of this chapter is to provide a better dimensional stability of

Nafion upon hydration without projecting the proton conductivity though blending with

fluorocarbon polymers. Copolymer of poly(vinylidene fluroride)-trifluoroethylene

(PVDF-TrFE) was chosen because PVDF would be miscible with Nafion at least in the amorphous phase at fuel cell operating temperatures. It was reported that PVDF-TrFE copolymer readily developed the β-crystal with exceptionally high specific capacitance

value in the order of nano-Farads in the copolymer with 30% trifluoroethylene (TrFE)

molar ratio [110]. This polymer is known for its ferroelectric property which transforms to paraelectric with increasing temperature through a so-called Curie transition [111,112].

In fuel cell applications, conventional wisdom is to have a composite material with the highest proton transport but the lowest electron conductivity. Hence, blending of

PVDF-TrFE with Nafion has been sought because PVDF-TrFE may be a good capacitor but poor electron conductor, whereas Nafion is an excellent proton conductor especially in the hydrated state. Upon phase separation, PVDF-TrFE and Nafion may provide different pathways for electron and proton to transport. Moreover, PVDF-TrFE will act as physical networks to provide better dimensional stability to Nafion membranes upon hydration. We first established the experimental and theoretical phase diagram of the mixture to manipulate the preferred final morphology, i.e., the bicontinuous phase separated domains. The solution blended mixture may be phase separated during solvent casting or upon thermal treatment, which in turn forms various domain morphologies of

Nafion and PVDF-TrFE, including sea-and-island or comingled biphasic structures.

Electron and proton conductivities were determined for the 60/40 Nation/PVDF-TrFE blend and discussed in relation to the bicontinuous phase separated domains.

4.2. Sample Preparation and Experimental Methods

Nafion 115 membranes with equivalent weight of 1100 in its original acid form were purchased from the Fuel Cell Store, Inc. Pre-treatment for as-received Nafion 115 membranes was made in agreement to the experimental section in the preceding chapter to ensure complete conversion to the acid form. PVDF-TrFE (33 mol% of TrFE) having

the molecular weight, Mw of 6,900 and polydispersity, Mw/Mn of 1.403 was used in the

present study. Prior to blending, the Nafion membrane was cut into tiny pieces and dried

at 100 oC in a vacuum oven for 48 hours in order to eliminate residual moisture. Nafion

membrane and PVDF-TrFE were dissolved separately by thoroughly stirring in

dimethylacetamide (DMAc) at 120 oC and at room temperature, respectively. The

solutions were then mixed at various blend ratios while maintaining polymer

concentration at 5 wt%. The homogeneous solution was cast on pre-cleaned microscope slides and dried in a vacuum oven at 100 oC for 3 days. For optical microscopic analysis,

thin film with average thickness of about 20 μm was prepared. Thicker films

(approximately 90-100 μm nominal thicknesses) were used for differential scanning calorimetry (DSC), water uptake, and alternating current (AC) impedance measurements.

Phase transition temperature of the Nafion/PVDF-TrFE mixtures was determined using differential scanning calorimetry (DSC7, Perkin Elmer). The morphology of the pure and blend samples were examined by polarized optical microscopy (POM) (BX 60,

Olympus) equipped with a digital camera (EOS 300D, Cannon) and a hot stage (TMS 93,

Linkam). Experimental procedures of DSC and POM were conducted in accordance to the previous chapter.

Prior to water uptake measurement, samples were subjected to thermal treatment

and subsequently soaked in deionized water at room temperature for 24 hours. Water

uptake was determined by monitoring weight change before and after hydration, i.e.,

(Mwet–Mdry)/Mdry, where Mwet is mass of the membrane in a wet state and Mdry is mass of

the membrane in a dry state. The hydration effect on Nafion/PVDF-TrFE blends was

further investigated by Fourier transformed infrared spectroscopy (FTIR). FTIR spectra

(Nicolet 380, Thermo Scientific) with attenuated reflected mode (ATR) were acquired at

room temperature with the average of 32 scans and a spectral resolution of 4 cm-1.

AC electrical impedance measurements were conducted using the Schlumberger

Impedance/Gain-Phase Analyzer over the frequency range of 1 Hz–100 kHz and equipped with a heating stage which is connected to the Omega CN8500 temperature controller where membranes were heated from ambient temperature to 150 oC. Samples

were sandwiched between two indium tin oxide (ITO) cells with cell contact area of 1

cm2 and with thicknesses of membranes ranging between 90–110 μm. The electrical capacitance and conductivity values were calculated in agreement to eqs (2.31) and (2.32) respectively.

Proton conductivity characterization of neat Nafion and impregnated membranes was carried out using the AC impedance fuel cell device (Scribner, Model 850e Multi

Range) equipped with adjustable reactant humidifier unit. Membrane preparation and experimental procedures were described in the preceding chapter and complex impedance data obtained were analyzed using ZView program. Proton conductivity was calculated in accordance to Equation (2.32).

4.3. Theoretical Scheme for Establishment of Phase Diagram

A theoretical phase diagram containing crystalline and amorphous polymer may

be constructed by combining the free energy density of Flory–Huggins (FH) [130] , phase field (PF) [132], and their coupling interaction. The FH theory is employed to describe a

liquid-liquid demixing together with the FH interaction parameter representing

amorphous–amorphous interaction [134] and may be expressed as a function of

temperature (may be found in Chapter II, eq 2.15). The PF theory is considered for

crystal solidification of copolymer PVDF-TrFE that undergo the solid (crystal)–liquid

(melt) phase transition and thus total free energy density may be obtained through

combining the FH, PH and coupling parameter as stipulated in the eq (2.20) in Chapter II.

4.4. Results and Discussion

This section will demonstrate all experimental findings as well as discussion on

the studied Nafion/PVDF-TrFE blends.

4.4.1. Phase Transition

Figure 4.1 depicts the DSC thermograms of Nafion/PVDF-TrFE blends obtained

at a heating rate of 10 oC/min. The pure PVDF-TrFE showed two endothermic peaks. The

o first peak at 73 C corresponds to the Curie transition, TCurie, which may vary with TrFE

composition. This transition is denoted as the solid state order–disorder phase transition

in which ferroelectric property is lost as it becomes paraelectric [113]. The second peak

o observed at 163 C denotes the crystal melting transition, Tm, which is slightly depressed

by the addition of Nafion, implying partial miscibility of the mixtures.

Figure 4.2 illustrates the morphological change of Nafion/PVDF-TrFE blends with increasing temperature. The anisotropic PVDF-TrFE crystals and isotropic dark region can be seen clearly below 30oC in all compositions, but the isotropic (dark) region

gets expanded with increasing Nafion compositions, indicating the crystal + liquid

coexistence region. As temperature approaches to the melting point of PVDF-TrFE, the

PVDF-TrFE crystals start to melt and then the bright crystalline entities completely

disappear under the crossed polarization condition. However, the morphology changes

can be confirmed under the parallel mode, i.e., the blends containing 10 and 90 wt %

Nafion become homogeneous at 250oC (Figures 4.2(a) and 4.2(c)). However, the micrograph at 60 wt % Nafion reveals coarse interconnected domain morphology, characteristic of liquid + liquid phase separation (Figure 2(b)), indicating an immiscible gap above the melting point at intermediate compositions.

Figure 4.1: DSC thermograms for Nafion/PVDFTrFE blends demonstrating consistent Curie transition temperature (TCurie) and little or no movement of the melting temperature (Tm) of the entire blends composition range with 10 wt % increment of Nafion.

Figure 4.2: Optical micrographs of Nafion/PVDFTrFE in the extreme and intermediate compositions observed under polarized mode and parallel mode, suggesting presence of isotropic phase at high temperature for the extreme composition and presence of liquid– liquid phase separated structure for the intermediate composition of the blends.

4.4.2. Binary Phase Diagram

The complete phase behavior of the blends may be understood by establishing a theoretical phase diagram. Figure 4.3 shows the phase diagram of Nafion/PVDF-TrFE blends. The theoretical binodal (solid lines) curves accord well with the experimental points obtained from DSC (as denoted by ‘□’) and POM (as denoted by ‘○’ for isotropic phase and ‘●’ for liquid-liquid phase separated region) measurements. The model

100

-2 parameters used for calculation were r1 = 70, r2 = 90, A = -0.4, C = 6.1 × 10 , χ aa = 3.0 ×

-2 -3 10 , and χ ca = 7.4 × 10 at Tm along with ΔHu = 1.8 kJ/mol, respectively. Consistent

with the previous DSC results, the melting transition of PVDF-TrFE was found to be slightly depressed with increasing Nafion composition, thus giving very small χca value.

Above melting temperature (Tm), the crystalline PVDF-TrFE melts to form isotropic (I) phase at high concentration and liquid + liquid (L1 + L2) coexistence phase at

intermediate compositions, as shown in the insets of Figures 4.3(i), (ii) and (iii). It is

reasonable to infer that the mixture exhibits an hourglass type phase diagram in which the

LCST and UCST curves are presumably merged. Upon lowering the temperature below

Tm, three distinct regions are discerned, i.e., crystal + liquid (Cr1 + L2) coexistence

between the single phase crystal (Cr1) and the isotropic liquid (I) phase. Insets in Figure

4.3 denoted by (iv), (v) and (vi) illustrate the morphologies of PVDF-TrFE crystals at

three different compositions.

The present hourglass phase diagram is different from the Nafion/PVDF blends

that exhibited the LCST type phase diagram at about 210 oC. These authors identified the

occurrence of liquid-liquid phase separation via spinodal decomposition above the

melting temperature of PVDF [21,22]. From the LCST phase behavior, the blends will

revert back into the isotropic state at the operation temperatures of fuel cells (e.g., below

200 oC), which may compromise the fuel cell performance. In this regards, the

Nafion/PVDF-TrFE blends are expected to be in two phase regions at all fuel cell

operation temperatures.

101

Figure 4.3: The theoretical binodal (solid line) and spinodal (dash line) curves of Nafion/PVDF-TrFE blends which describe well the experimental crustal melting transition results obtained from DSC (as denoted by ‘□’) and OM (as denoted by ‘○’ for isotropic phase and ‘●’ for liquid-liquid phase separated region).

102

4.4.3. Water Uptake and Hydration Effect on the Nafion/PVDF-TrFE Blends

As shown in Figure 4.4(a), the hydration of Nafion/PVDF-TrFE blends was

significantly reduced with increasing amount of PVDF-TrFE. Hydrated Nafion exhibited dimensional increment of 27% in length and 20% in width. Note that the average thickness of the Nafion/PVDF-TrFE blend membrane (~90 – 110 µm) was controlled to be close to the thickness of that pure Nafion membrane (~110 µm). The excessive swelling of Nafion membrane eventually results in membrane warping (see inset right picture). The water uptake at the intermediate compositions, i.e., the 60/40 Nafion/PDVF-

TrFE blend was reduced to 15–20 wt % and the warping is visually reduced (see inset middle picture) with only slight dimensional increment of 10% in length and 7% in width. In this intermediate region, the turbidity of the membranes also increased due to phase separation. The suppression of water uptake gives better dimensional stability to the blend membranes. As expected, little to no change in length and width was observed at the 10/90 PVDF-TrFE blend because this blend primarily consists of hydrophobic

PVDF-TrFE rich region.

The composition-dependent hydration effect was further examined by FTIR. The intensity of O–H stretching at around 3700–3000 cm-1 decreased upon increase of the

PVDF-TrFE concentrations. The normalized absorbance of O–H stretching band with

respect to the O–H stretching band of pure Nafion is displayed in Figure 4.4(b). The same

characteristic bands have been used by several researchers to quantify water content in

Nafion-acid and Nafion-neutralized membranes [62,63].

103

Figure 4.4: Plot of (a) water uptake in weight increment (%) of Nafion/PVDF-TrFE blends while insets are pictures of the actual hydrated membrane demonstrating reduction of warping phenomenon as the PVDF-TrFE composition increased, (b) normalized FTIR absorbance based on the O–H stretching band as a function of Nafion. The insets show the morphology of the extreme and intermediate blend compositions which may be used to explain the hydration effect. Both plots signify the remarkable swelling suppression of the blends towards aprotic liquid, i.e., water.

104

In the intermediate Nafion/PVDF-TrFE compositions, water sorption is reduced

to about 80% relative to that of the neat Nafion, suggesting the intermediate blends

becoming less hydrophilic. When PVDF-TrFE concentration is increased above 90 wt %,

the mixture contained water less than 5%. The water content measured by FTIR and

weight increment showed a similar sigmoidal trend but with a slight discrepancy

occurring in their trend due to the crude accuracy of the weighing process.

It is well known that phase separation of polymer blend can take place via

nucleation and growth (NG) and spinodal decomposition (SD) mechanism. It can be

envisaged from the optical micrographs (see insets of Figure 4.4(b)), that the intermediate

composition, i.e., 60/40 Nafion/PVDF-TrFE, exhibits bicontinuous morphology

(interconnected spinodal structure) while in the extreme compositions, viz., 10/90 and

90/10, exhibit the sea-and-island type morphology which probably occurs through

nucleation and growth although by no means a proof. The sea-and-island type

morphology is not interesting in the evaluation of the electron or proton conductivity

performance due to the isolated and discontinuous droplet structures of Nafion in PVDF-

TrFE matrix or vice versa. The preferred morphology is the interconnected spinodal

structure that gives the percolated pathways for electron or proton transport [25].

The observed interconnected SD structure at 60/40 Nafion/PVDF-TrFE

composition is expected to contain PVDF-TrFE crystals comingled with the amorphous

Nafion region. The bicontinous pathways of Nafion upon hydration will allow protons to

conduct, while hydrophobic percolated PVDF-TrFE may restrict the excessive swelling and proton conduction. Since PVDF-TrFE primarily forms the β–form crystals, the

105

capacitor behavior are expected to dominate in this percolated PVDF-TrFE network channels. Hence, it is of interest to measure the electrical properties and proton conductivity on the intermediate blend composition, i.e., 60/40 Nafion/PVDF-TrFE blend

since this will give an insight on the interplay between the Nafion-rich channels for

proton transport and PVDF-TrFE-rich percolated network having high capacitance (or

resistivity).

4.4.4. Electrical Properties of Nafion/PVDF-TrFE Blends

Figures 4.5(a) and (b) illustrate the log-log plot of storage impedance and semi log plot of loss tangent as a function of frequency for pure PVDF-TrFE. The storage impedance reveals the sigmoidal reduction of impedance values with frequency, whereas the loss tangent peaks show a consistent movement towards higher frequency as temperature is raised. The loss tangents also exhibit a lowering of its intensity with increasing of temperature as the crystals of PVDF-TrFE loses their properties and

ultimately start to melt.

Next, the Cole-Cole plot with the curve fitting plot (solid line) of pure copolymer

PVDF-TrFE with temperature was plotted (Figure 4.6). At low temperature, the presence of the vertical line, almost parallel to the imaginary impedance axis (Z”), indicates that pure PVDF-TrFE behaves like a pure capacitor. When temperature exceeds the

o ferroelectric-paraelectric transition (TCurie) at around 73 C, the capacitor behavior

diminishes. Formation of semicircle at > 100 oC suggests that pure PVDF-TrFE acts like

a conductor but the conductivity value is still within the range of an insulator.

106

Figure 4.5: (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of pure PVDF-TrFE from 43 oC to 150 oC over frequency range of 1 Hz–100 kHz.

Figure 4.6: Cole-Cole plot and curve fitting plot of pure PVDF-TrFE copolymer with temperature obtained from the AC electrical impedance measurement illustrating the diminishing capacitor behavior above the Curie transition temperature.

107

Figure 4.7: Plot of (a) capacitance and (b) electron conductivity versus temperature of pure PVDF-TrFE demonstrate the polarization capacitance, Cp and electron conductivity with increasing temperature.

Figure 4.7 demonstrates the plot of capacitance and electron conductivity with temperature. Capacitance and conductivity values were obtained by eqs. (2.31) and (2.32) may be interpreted with great reservation because the polarized resistance (Rp) value was estimated. Nevertheless, it should be emphasized that the capacitance value of pure

PVDF-TrFE increases from 43 oC and declines above 80 oC due to the ferroelectric property which eventually transitions from the ferroelectric to the paraelectric above the

Curie transition. On the other hand, the electron conductivity shows an opposite trend with temperature. However, the electron conductivity value is so small (in the order of

10-11~10-12) and thus can be considered an insulator. This is the key requirement for the electrolyte membrane to be used in the proton electrolyte fuel cells, i.e., to possess poor electron conduction while allowing rapid proton conduction through the Nafion-rich bicontinuous network domains [3]. 108

Neat Nafion is known as an excellent proton conductor but has the ability to

immensely insulate conduction of electrons. Figures 4.8(a) and (b) show the plot of log

storage impedance and loss tangent versus log frequency of hydrated Nafion. This

hydrated Nafion membrane was equilibrated in deionized water for 24 h prior to the AC

electrical impedance measurements. Storage impedance plot of hydrated Nafion shows

the same sigmoidal reduction with frequency as observed in pure PVDF-TrFE. However, with increasing of temperature, magnitude of storage impedance increases, denoting increase of resistance within the Nafion membrane as water starts to evaporate (Figure

4.8(a)). The semi-log loss tangent plot in Figure 4.8(b) shows a prominent peak around

105 Hz at room temperature. This loss tangent peak may be ascribed as the relaxation of

the ionic clusters. As the temperature further increases, another loss tangent peak starts to

appear at 75 oC and this may be attributed to the loss of water in the membrane. Both of

the loss tangent peaks of hydrated Nafion exhibit a systematic movement to lower

frequency with increased temperature.

The Cole-Cole plot of the hydrated neat Nafion is plotted in Figure 4.9 in order to

determine the conductivity value of that of Nafion. The semi-circular occurrence in the

Cole-Cole plot suggests that neat Nafion behaves like an electrical conductor when

hydrated. This is due to the fact that Nafion is highly hydrophilic and water conducts

electrons. With increasing of temperature, resistivity exerted by neat Nafion begins to

increase due to the loss of water within the membrane. This can be seen in the increasing

of the storage impedance (Z’) value by two orders of magnitude.

109

Electron conductivity as a function of temperature of hydrated Nafion is depicted

in Table 4.1. Evidently, the conductivity of Nafion shows a drastic plummet as

temperature is raised. This can be explained due to the excessive evaporation of water

making the Nafion less conductive. These electron conductivity values are, however, very

small, thus confirming the fact that Nafion is indeed a great insulator towards electrons

even in the hydrated state.

Figure 4.8: (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of hydrated Nafion from 23 oC to 125 oC over a frequency range of 1 Hz–100 kHz.

110

Figure 4.9: Cole-Cole plot and curve fitting plot of hydrated Nafion with increasing of temperature obtained from the AC electrical impedance measurement illustrating the conductor-like behavior.

Table 4.1: Table of electron conductivity values of hydrated Nafion with increasing of temperature signifying the plummet of its electrical conductivity.

111

A log-log plot of storage impedance, Z’ versus frequency and a semi log plot of

loss tangent versus frequency of 60/40 Nafion/PVDF-TrFE are shown in Figures 4.10(a)

and (b). The storage impedance of 60/40 blends decreases with increasing of temperature.

Systematic movement of loss tangent peak to higher frequency was observed as the

temperature increases. Figure 4.11 illustrates the Cole-Cole plot of 60/40 Nafion/PVDF-

TrFE blend. The vertical plot parallel to the imaginary impedance axis signifies that the

blend exhibits a capacitor behavior. The 60/40 Nafion/PVDF-TrFE blend shows the

capacitor property inherited from the copolymer PVDF-TrFE. Polarization capacitance,

Cp as a function of temperature of 60/40 Nafion/PVDF-TrFE blend is depicted in Figure

4.12. With 40 wt% of PVDF-TrFE, the capacitance value exerted by the blends below the

Curie transition is approximately reduced by 50% of the pure PVDF-TrFE. Similar to the pure PVDF-TrFE, capacitor value tends to increase with temperature and reached the optimum capacitance behavior at 100 oC. However, above 100 oC, it is ambiguous to

determine the capacitance value of the blend due to the incomplete plot within the limited

frequency range measurement. It is anticipated that the comingled PVDF-TrFE

bicontinous networks within the membrane provide the interconnected pathways for the

electron conduction despite its insulator-like behavior. On top of that, these interphases

may have introduced an increment of the electron accumulation areas; thus, the blend

shows similar capacitor behavior to that of the PVDF-TrFE, below the Curie transition

temperature.

112

Figure 4.10: (a) Log-log plot of storage impedance, Z’ versus frequency and (b) semi log plot of loss tangent versus frequency of 60/40 Nafion/PVDF-TrFE from 43 oC to 125 oC over frequency range of 1 Hz–100 kHz.

Figure 4.11: The Cole-Cole plot of 60/40 Nafion/PVDF-TrFE blend collected from the AC electrical impedance measurements shows the occurrence on the capacitor behavior of the blend.

113

Figure 4.12: Plot of polarization capacitance, Cp versus temperature for 60/40 Nafion/PVDF-TrFE shows an increasing of capacitance values upon increasing of temperature.

4.4.5. Proton Conductivity Properties of Nafion/PVDF-TrFE Blends

Proton conductivity measurements of both neat Nafion and Nafion/PVDF-TrFE blends were carried out in the AC impedance fuel cell environment at 74% and 100%

RH. Figures 4.13(a) and (b) show the log-log storage impedance and semi log tan δ as a function of log frequency of neat Nafion at 100% RH. Storage impedance of neat Nafion under the fuel cell measurement tests showed a sigmoidal reduction in resistance with frequency. The storage impedance initially decreases with temperature but starts to increase above 80 oC. This may be attributed to the water evaporation from the

membrane. In addition, the thermal instability of the protonated clusters beyond 80 oC

also contributes to the increasing of resistance of that of neat Nafion. Loss tangent peaks

114

show a systematic movement towards lower frequency as the increasing of temperature.

Note that 60/40 Nafion/PVDF-TrFE exhibits a similar trend of storage impedance and loss tangent, except with different Z’ value (data not shown).

Figure 4.13: (a) Log-log plot of storage impedance versus frequency and (b) semi-log plot of loss tangent versus frequency of neat Nafion at 100% RH.

Figure 4.14: Cole-Cole plot of (a) pure Nafion and (b) 60/40 Nafion/PVDF-TrFE blend measured in the AC impedance fuel cell environment at 100 % RH. With increasing of temperature, polarization resistance decreases but exceeding the boiling temperature of water, and polarization resistance within the cell increases denoting loss of proton conductivity efficiency. 115

Figures 4.14(a) and (b) illustrate the Cole-Cole plot of neat Nafion and 60/40

Nafion/PVDF-TrFE samples measured at 100% RH. The first intersection of the plot at

the x-axis,denoted as Rs, indicates the resistance between the electrolyte membrane and

the catalyst layer known as the solution resistance. The diameter of the semi-circular plot,

denoted by Rp is the resistance from the electrolyte membrane, known as polarization

resistance [138]. The Rp value is used to evaluate the conductivity in accordance with

Equation (2.32). An equivalent circuit of the cell may be represented by a resistor in

series with a parallel circuit containing a capacitor and a resistor. It is clear that the

diameter of the Cole-Cole plots for both samples contract from 40 to 80 oC, suggesting

less resistance within the cell. However, upon reaching the boiling temperature of water,

both samples exhibit an increase in the polarization resistance due to loss of water leading

to decrease in proton conductivity efficiency.

Plot of proton conductivity as a function of temperature is demonstrated in Figure

4.15. Neat Nafion demonstrates an increasing proton conductivity trend with increasing temperature. As temperature increases beyond 80 oC, the proton conductivity of neat

Nafion shows a declining trend. This may be attributed to the dehydration of the

membrane and the interconnected channels starting to disrupt, thus making the ionic

domains more isolated, and hence lowering the proton conductivity. It is evident that the

proton conductivity of neat Nafion samples decline noticeably (σ = 0.043 S/cm) when

temperature is raised further up to 115 oC at 74% RH.

The 60/40 blend was found to exhibit lower proton conductivity at all

temperatures by almost one half to that of neat Nafion. At 80 oC, the 60/40 blend

116

manages to perform the utmost proton conductivity with the value of 0.048 S cm-1 at

100% RH. Based on Equation (2.32), conductivity of the membrane is highly dependent

on the bulk thickness and the macroscopic resistance of the particular membrane. This

equation may be applicable to homogeneous materials. However, the Nafion/PVDF-TrFE

blend tends to create additional microscopic interphase layers, which in turn makes the

membrane a complete heterogeneous membrane. Thus proton conductivity of the blend

membrane is now dependent on its Nafion composition/volume fraction as well as the

orientation of its percolated pathways. Therefore, considering only the bulk thickness of

the blend membranes will not reflect the actual conductivity value. The individual

percolated pathways need to be compensated thus by reducing the Nafion composition to

only 60wt% (i.e., almost one half relative to 100wt% Nafion), so that the proton

conductivity obtained is reasonable. Should the proton conductivity have been corrected

for the bulk thickness of the 60/40 blend, a similar decreasing trend to that of neat Nafion

can be anticipated.

Figures 4.16 show the schematic diagrams of the sea-and-island (nucleation and growth) morphology that is anticipated to occur in the extreme Nafion or PVDF-TrFE blend compositions while the bicontinuous percolated pathways (spinodal decomposition) morphology is likely for the 60/40 Nafion/PVDF-TrFE blend. The bright

regions denote the Nafion-rich regions, while the dark regions signify the PVD-TrFE-rich

regions. The sea-and-island type morphology is not suitable for the present electron or

proton conductivity studies; we therefore focused only on the 60/40 blend.

117

Figure 4.16(b) illustrates the bulk thickness of the membrane, d, and the

microscopic thickness, dm, of each torturous percolated Nafion or PVDF-TrFE pathway.

Presence of these interphases needs to be taken into consideration as the thickness of each percolated Nafion pathway will signify different individual microscopic conductivity behavior. It is desirable to form an array of channels perpendicular to the electrodes as this has been known to help form a continuous pathways for electron and/or proton to conduct, thus greatly improving the conductivity properties (Figure 4.16(d)). Such perpendicular alignment of domains may be fabricated under thermal gradient of pattern photopolymerization induced phase separation like in photonic crystals [143]. This is the approach employed by Gore Associates Inc. through the invention of the Teflon grid supported Nafion membrane (Figure 4.17) [144].

Figure 4.15: Proton conductivity plot as a function of temperature of neat Nafion and 60/40 Nafion/PVDF-TrFE. It is apparent that the proton conductivity trends of the 60/40 blend resemble that of neat Nafion. 118

Figure 4.16: Schematic diagrams illustrating the (a) sea-and-island morphology signifying the dispersed and discontinuous structure of 90/10 blend with Nafion-rich region representing the matrix, (b) bicontinous morphology exhibited by 60/40 blend where Nafion and PVDF-TrFE percolated pathways coexist side by side, (c) inverted sea- and-island morphology which may be seen in the 10/90 blend with matrix of PVDF- TrFE and (d) an array of cylindrical channels which will favor enhancement of protonic and/or ionic conductivity within the blends.

250 µm

20 mm

Figure 4.17: Surface morphology of Teflon grid supported Nafion membrane developed by Gore Associate Inc.

119

4.5. Conclusions

Phase diagrams of Nafion and PVDF-TrFE copolymer blends have been established experimentally and theoretically. At intermediate composition, the crystal +

liquid coexistence region, transformed to liquid + liquid upon increasing temperature

above melting point,showing an hour-glass type phase diagram. The constructed phase diagram was employed to manipulate the bicontinuous or dispersed domains via phase separation after thermal quenching. The PVDF-TrFE rendered the dimensional stability of Nafion by suppressing water uptake. The bicontinuous morphology at 60/40

Nafion/PVDF-TrFE composition provided both electrical and proton conductivity properties. This supports the hypothesis that the percolated pathways may not only act as the physical crosslink to suppress swelling but also exhibit electronic and protonic conductivity properties within the individual percolated pathways. Formation of the bicontinuous pathways within the Nafion/PVDF-TrFE has compromised the proton conductivity of the membrane. By modifying only the fluorocarbon matrix of Nafion, the swelling is suppressed but there is no improvement in the proton conductivity unless these percolated domains are aligned perpendicular to the membrane.

120

CHAPTER V

IN-SITU IMPREGNATION OF NAFION WITH NOVEL PHOTOCURABLE

HYPERBRANCHED POLYESTER, HBPEAc-COOH

5.1. Introduction

The perfluorinated ionomer membrane, commercially known as Nafion, has received widespread interest from researchers in both academia and industry. Over the past several decades, the extensive structure characterizations of Nafion have led to a general consensus that the perfluorinated ionomer has a micro-separated structure that consists of a hydrophobic backbone chain with shorter fluorinated side chains and hydrophilic ionic domains/clusters of the sulfonate terminal group. The hydrophobic fluorocarbon structure of the perfluorinated ionomer offers excellent thermal and chemical stability, whereas the hydrophilic ionic domains/clusters tethered to the perfluorinated backbone provide exceptional ion and/or proton transport properties [42-

45].

By virtue of its outstanding proton conductivity, Nafion is currently regarded as the bench marked polymer electrolyte membrane (PEM). One major drawback is that its operation temperature is limited to 60 – 80 oC, above which the proton conduction

progressively gets worse. In the actual proton fuel cell operation, a numbers of PEM fuel

121

cells have to be stacked to obtain desirable power output. An evitable consequence is that

the cell temperature rises significantly beyond the optimum operating temperature of that

of Nafion membrane, e.g., up to 130 oC for portable fuel cells and 150 oC for larger units

where water is no longer available. Hence, there is a growing demand for developing new

PEM materials capable of functioning at high temperatures under the actual proton fuel

cell operations.

There are several attempts to remedy the aforementioned short comings of Nafion

perfluorinated ionomer membranes, especially to increase proton conduction at elevated

temperatures. One notable approach is the in-situ impregnation (or doping) by means of incorporating functional small molecules such as ionic liquids within the Nafion ionic

domains/clusters [15-18]. Infusion of ionic liquids into Nafion has afforded considerable

improvement in ionic conduction. In addition, the occurrence of specific interaction has

been identified, i.e., hydrogen and/or ionic bonding between these small molecules and

the sulfonate groups of Nafion has helped retaining these functional small molecules

during use [17, 18].

The main idea of the present research is to incorporate functional

photopolymerizable supramolecules into the ionic domains and subsequently polymerize

it in-situ to form functional polymer networks in order to prevent any potential leaching

in actual fuel cell operations. Another goal is to improve proton conduction at elevated

temperatures where Nafion performs poorly.

In this chapter, we infused solid supramolecules such as photocurable hyperbranched (HB) polyester terminated with functional carboxylic acid groups, namely

122

HBPEAc-COOH [140], into the micro-phase separated ionic domains of Nafion via swelling in methanol.

It can be anticipated that the terminal carboxylic acid of HB will not only enhance ion exchange capacity, but also impart inter-species hydrogen bonding with the sulfonate groups of Nafion for retention of the modifier molecules within the ionic domains.

Subsequently, photo-crosslinking will be performed in-situ to retain the modifier supramolecules within the ionic domains and also to enhance its proton conduction above the current operating temperature of Nafion in the reduced water environment.

Suppression of the excessive swelling upon hydration is also expected through occupying available space within the ionic domains of Nafion by these supramolecules. The incorporated amount of HB supramolecules and inter-species interaction was evaluated by means of Fourier transform infrared (FTIR). The physical and structural characterization of the HB impregnated Nafion was performed using solid-state nuclear magnetic resonance (SSNMR), differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), dynamic mechanical analysis (DMA), and AC impedance measurements under the proton fuel cell environment. Of particular importance is that the present work will be the first to successfully incorporate large polymer molecules into the ionic domains of Nafion through impregnation and subsequent in-situ polymerization of photopolymerizable hyperbranched polyester [140].

More importantly, the present approach will afford better dimensional and thermal stability of the impregnated membrane with improved proton conduction during high temperature proton fuel cell operations relative to the unmodified Nafion.

123

5.2. Experimental Section

5.2.1. Materials and Sample Preparation

Perfluorinated ionomer membrane, Nafion 115 membranes with equivalent weight of 1100 in its original acid form were purchased from Fuel Cell Store. As- received Nafion membranes were first pretreated for fully conversion into protonated/acid form. Pretreatment process is analogous as described in the previous

chapter.

The HB polyester was synthesized from polyaddition of bisphenol-A-diglycidyl

ether (BPGE) with trimesic acid (TMA) and methacrylic acid (MA), by adding cis-

1,2,3,6-tetrahydrophthalic anhydride (THPA), triphenylphosphine (TPP) and hydroquinone (HQ) having pendant methacryloyl and carboxyl groups. The molecular weight and polydispersity of neat HBPEAc-COOH in tetrahydrofuran (THF) solution were measured by GPC (Model 1515, Waters) using polystyrene standard. The number average molecular weight and polydispersity were found to be Mn=6,800 and

Mw/Mn=1.42, respectively. The advantage of the above HB polyester is in its photocuring

capability of the acrylate double bonds and the carboxylic acid functionality that affords

ionic interactions with its Nafion counterpart. The detailed synthetic scheme of these HB

polyesters may be found in a previous paper [140]. ACS grade methanol was purchased

from Sigma-Aldrich and used without further purification for dissolving hyperbranched

polyester and swelling the ionic domains of Nafion.

Weighed Nafion membranes were pre-swollen in methanol for 24 h and then

immersed in the HBPEAc-COOH/methanol solution for 48 h at room temperature. The

124

curing agent (Irgacure 907) in an amount of 3wt% with respect to HBPEAc-COOH was added to the above solutions. The Nafion to HBPEAc-COOH polyester mass ratios were

98/2, 97/3, 95/5, 90/10 and 80/20. These ratios correspond to the feed compositions, which will be hereafter referred to as ‘feed ratio’. Impregnated membranes were then removed from the HBPEAc-COOH/methanol solution and were gently blotted with tissue paper (e.g., Kimwipes) to remove excess methanol solution on the surface. Subsequently, the impregnated membranes were photocured in a curing chamber (NuLine, model

NL22-8C, 90mW/cm2) at ~85 oC for 5 min, which is slightly above the glass transition

temperature (Tg) of the HB polyester. Blotting and curing procedures were conducted in a

dark room so as to prevent accidental curing of the membranes. The HBPEAc-COOH impregnated Nafion membranes will be referred to as ‘impregnated membranes’. These impregnated membranes were kept in a dessicator prior to use.

5.2.2. Experimental Methods

Thermal analyses of neat HBPEAc-COOH before and after photocuring were performed using a differential scanning calorimeter (Modulated DSC 2920, TA

Instruments). The recommended amount of 7-10 mg of each sample was encapsulated in aluminum hermatic pans. DSC heating and cooling scans were carried out from 30 oC to

120 oC at a rate of 10 oC/min unless indicated otherwise. The DSC chamber was purged

with nitrogen gas at a rate of 80mL/min.

125

1H-13C and 19F-13C cross polarization (CP) and 1H direct polarization (DP) magic

angle spinning (MAS) NMR spectra were collected on a Varian (Model: NMRS 500,

11.7 T) spectrometer operated at 125.62 MHz for 13C and using a Varian narrow-bore

triple-resonance T3 MAS NMR probe. The 1H-13C and 19F-13C CP/MAS data were

collected under continuous wave (CW) proton and TPPM fluorine decoupling,

respectively. A 90° pulse width of 4 μs was used for all nuclei. Recycle delays of 2, 2.5

and 1 s were used for 1H-13C CP, 19F-13C CP and 1H DP experiments with contact times

of 3.5 and 1.5 ms for the 1H-13C and 19F-13C CP experiments. For the 1H spectra, 32

transients were collected.

FTIR spectra were collected in the attenuated total reflected mode (ATR) on a

Nicolet 380 (Thermo Scientific) spectrometer with an average of 32 scans and a spectral

resolution of 4 cm-1 at ~100 oC. Prior to the spectra collection, the membrane samples

were equilibrated at 100 oC to eliminate moisture absorption. Quantification of the

estimated amount of HBPEAc-COOH in the impregnated membranes was done by integrating the area under the FTIR curves within the wavenumber ranges of 1670–1770

cm-1 and 1550-1450 cm-1, which correspond to the carbonyl group and C=C aromatic

stretching peaks. The estimated amount of HB polyester was calculated in accordance to

Equation (3.1) as stipulated in the previous chapter. Physical weighing measurements

were carried out in order to counter-check the quantified amount of incorporated

HBPEAc-COOH in Nafion membranes relative to those obtained from the integrated area

under the FTIR spectral peaks.

126

Dynamic mechanical properties were measured in sinusoidal tension mode using

a Pyris Diamond DMA (Perkin Elmer) equipped with a liquid nitrogen cooling controller

(Seiko Equipment). Dried neat Nafion and impregnated membranes were cooled down to

-140 oC and maintained isothermally for 5 minutes, then heated up to 200 oC at a heating

rate of 1 oC/min. The frequency of the dynamic mechanical measurements was fixed at 1

Hz and measurements were carried out under the nitrogen environment.

Unless indicated otherwise, water uptake, ion exchange and proton conductivity

measurements of neat Nafion and impregnated membranes were conducted as mentioned

in the preceding chapter. Impregnated membranes were subjected to cyclic measurements

under the fuel cell operating conditions which consist of 5 heating and cooling cycles

with a total of 160 hours.

5.3. Results and Discussion

5.3.1. Characterization of Neat Nafion and Pure HBPEAc-COOH

Prior to incorporation into the Nafion membranes, the physical and chemical

characterization of HB polyester was conducted by means of DSC, FTIR and SSNMR.

Figure 5.1 shows the DSC thermograms of pure HB polyesters before and after

photocrosslinking. The movement of the Tg (i.e., the mid-point of the transition) of pure

HB from 79 oC to 95 oC after photocuring indicates that the overall motion of the HB

polyester chains becomes restricted. The act of photocuring leads to the formation of a chemical crosslinked network within the structure, thus requiring a higher thermal energy for the chains to undergo any chain motions.

127

The overlay FTIR spectra of neat HBPEAc-COOH, HBPEAc-COOH with the initiator before photocuring and HBPEAc-COOH after photocuring are shown in Figure

5.2. Note that the amount of Irgacure 907 added was 3 wt% relative to the weight of

HBPEAc-COOH. It is evident that with the addition of photoinitiator to the neat HB

polyester, there is no discernible change in the characteristic FTIR peaks such as 1727

cm-1 (C=O carbonyl stretching) and/or 1508 cm-1 (C=C stretching of the aromatic rings).

The peak invariance can also be noticed in other characteristics bands at 1230 cm-1 and

1190 cm-1 corresponding to the C-O-C stretching of cyclic ether linkages of HBPEAc-

COOH, and 830 cm-1 and 740 cm-1 peaks attributable to C-H bending and C=C aromatic

out-of-plane bending vibrations. However, the addition of 3wt% of photoinitiator is

sufficient to achieve a high conversion (i.e., ~60%) in 4 min as pointed out in our

previous paper [140]. Upon photocuring, the C=C stretching band of acrylate near 1630

cm-1 decreases as it transformed into a C-C single bond after being attacked by free radicals of the photoinitiator. It should be emphasized that the HB polyester film is brittle as without photocuring.

128

Figure 5.1: DSC thermograms of pure HBPEAc-COOH (a) before and (b) after photocuring exhibiting an increase in Tg of the HBPEAc-COOH supramolecules upon photocuring.

Figure 5.2: FTIR spectra of (a) pure HBPEAc-COOH without initiator; (b) pure HBPEAc-COOH with initiator before curing and (c) pure HBPEAc-COOH after photocuring demonstrating the reduction of acrylate C=C stretching peak at 1630 cm-1.

129

5.3.2. Structural Characterization of HBPEAc-COOH Impregnated Membranes

The 13C CP/MAS NMR spectra acquired from cured HBPEAc-COOH and

impregnated membranes are shown in Figure 5.3. It should be emphasized that the peaks

in these spectra arise solely from the perprotio species of HBPEAc-COOH and

photoinitiator. The spectra are complex as a result of the many different types of carbons

present in the HB polyester. The most significant differences between the spectra from

neat HBPEAc-COOH and impregnated membranes are the additional peaks at 26, 52, 68

and 74 ppm that arise from the substituted morpholino portion of the photoinitiator used

to crosslink the HB supramolecules.

In order to quantify the amount of hyperbranched polyester incorporated, the total

signal intensity of the 13C CP/MAS NMR spectra of HBPEAc-COOH was evaluated.

Correcting for the total number of transients collected and sample size, the amount of

HBPEAc-COOH present in the impregnated membrane was determined to be 7 ± 1wt%.

As will be discussed later, this mass of HB polymer in the Nafion film matches the results obtained in the FTIR studies (vide infra).

The fluoropolymer backbone of the Nafion membrane was investigated to determine if there were any changes in this environment upon impregnating Nafion with

HBPEAc-COOH. The 13C NMR spectrum was collected using 19F-13C CP and 19F

decoupling, so that the peaks in the 13C NMR spectrum can only arise from the

fluorocarbon component of Nafion. The 13C spectrum from Nafion, Figure 5.4, is

consistent with those reported previously [145,146] with the peaks at 114 ppm from

backbone -CF2 groups, 120 ppm from the -OCF2 group, 122 ppm from the -CF3, 112 ppm

130

from the backbone -CF groups and 106 ppm from the -CF side chain groups. There are

no noticeable changes in the chemical shifts in the 13C NMR spectra observed for the neat

Nafion and impregnated membrane. These results suggest that there is little or no

reactivity between the Nafion fluorocarbon backbone and HBPEAc-COOH. However,

there is a slight change in linewidths, with the impregnated membrane having narrower

lines than those of the neat Nafion. This minor change presumably occurs due to changes

in the environment of the fluoro component as this portion of the membrane becoming

more mobile. As will be discussed later, the SSNMR observations accidentally

correspond to the mechanical β-relaxation temperature attributable to the glass transition temperature of fluorocarbon backbone of pure Nafion. It appears that the inclusion of HB polyester in the Nafion ionic domains not only alters the environment of the ionic region, but also exerts some influence on the motion of the fluorocarbon backbone chain and/or side groups.

Nafion membranes were characterized prior to and after the infusion of the HB polyester by determining water content, HB polyester uptake levels, and the interactions between the Nafion and HB polyester. The 1H NMR spectra of neat Nafion (Figure 5.5)

shows a single peak at 8.6 ppm, and this chemical shift corresponds to approximately 2

water molecules per SO3H group present in the Nafion membrane [65,66]. The narrow

linewidth of the peak indicates that the H2O molecules are mobile. The two broad peaks

in the 1H NMR spectrum from neat hyperbranched polyester, without the presence of

initiator, are assigned to the aromatic protons, i.e., downfield peak and the aliphatic

protons, i.e., upfield peak, Figure 5.5. After addition of the HB polymer to the Nafion

131

membrane, the 1H NMR spectrum is similar with two broad peaks arising from the HB

polyester. The presence of the additional peak in the aliphatic region of the spectrum of the impregnated membrane is attributed to the photoinitiator. The absence of any narrow

peaks in the 1H NMR spectrum from the impregnated membrane implies that there are no

mobile water molecules present. This suggests that the infusion of the HB supramolecule

into the ionic regions of the Nafion presumably has replaced, if not all, some water

molecules and thus suppressed the mobility of any remaining water molecules.

Figure 5.3: 125.6 MHz 1H-13C CP/MAS SSNMR spectra of (a) HBPEAc-COOH impregnated Nafion and (b) cured neat HBPEAc-COOH. The spectra were collected using 1H decoupling so that the peaks in the 13C NMR spectrum only arise from the HBPEAc-COOH and the photoinitiator. Peaks in (b) labeled with * arise from the photoinitiator.

132

Figure 5.4: 125.6 MHz 19F-13C CP/MAS SSNMR spectra of (a) impregnated membrane and (b) neat Nafion.

Figure 5.5: 499.5 MHz 1H DP/MAS SSNMR spectra of (a) impregnated membrane, (b) pure HBPEAc-COOH and (c) pure Nafion. The single narrow line at the chemical shift of 8.6 ppm in (c) corresponds to 2 H2O per SO3H. The peak in (a) labeled with * arises from the photoinitiator.

133

Figure 5.6: (a) 3700-2500 cm-1 region of FTIR spectra shows the slight shift of the broad O–H stretching band of impregnated membrane to lower wavenumber with increasing of HB polyesters; (b) the S-O stretching band of impregnated illustrates a systematic shift of the S-O band to a lower wavenumber (i.e., about 9 cm-1) suggesting inter-species hydrogen bonding interaction between the carboxylic acid of HBPEAc-COOH and the sulfonic acid group of Nafion.

Attenuated total reflected mode (ATR) FTIR experiments were conducted in

order to probe the interactions between HBPEAc-COOH and the ionic domains of

Nafion, as well as to determine the estimated amounts of HBPEAc-COOH incorporated into Nafion. The FTIR spectra of neat Nafion, neat HBPEAc-COOH and impregnated membranes with different HB polyester feed ratios are depicted in Figure 5.6. These spectra were collected under the minimal moisture absorption at 100 oC, e.g., bound

water. At a higher FTIR wavenumber range around 3300-3600 cm-1, the IR absorption band is extremely broad such that the peak position is not identifiable. This broad band implies the presence of bound water within pure Nafion. On the other hand, the cured

134

HBPEAc-COOH exhibits a pronounced O–H stretching band (at ~ 3500 cm-1) albeit

broad, arising from its pendant carboxylic acid functionalities [147] which may be

interacting with some bound water molecules. With impregnation of HB supramolecules

into Nafion, the broad O–H stretching band remains virtually stationary with increasing

HB contents from 2-10 wt% (see Figure 5.6(a)).

Figure 5.6(b) exhibits the FTIR spectrum of neat Nafion membrane showing two

strong bands at 1200 cm-1 and 1140 cm-1, attributable to the C-F asymmetric and

symmetric stretching vibrations of the main chain, respectively, along with the C-F

stretching vibrations from the side chains of Nafion at 980 cm-1. There is no movement of

these peaks upon incorporation of the HB polyester. Similarly, the C-F bending mode at

-1 -1 967 cm and -CF2 rocking mode at 625 cm remained stationary upon impregnation of

the HB polyester in Nafion. Together with the SSNMR results, it is reasonable to infer

that in the dried state there is little or no influence on the reactivity on the fluoropolymer

components of the Nafion backbone upon inclusion of the HB polyester into the ionic

domains.

The FTIR spectrum from neat HBPEAc-COOH illustrates a number of

characteristic bands; specifically the peak at 1727 cm-1 arising from the carbonyl stretching and the peak at 1508 cm-1 associated with the C=C stretching of the aromatic

rings. The bands at 1230 cm-1 and 1190 cm-1 correspond to the C-O-C stretching of cyclic

ether linkages of HBPEAc-COOH. The 830 cm-1 and 740 cm-1 peaks are associated with

the C-H bending and C=C aromatic out-of-plane bending vibrations, respectively. The

C=O band of the neat HB, which is located at 1727 cm-1, shows a minor shift upon

135

impregnation for about 2~3 cm-1 which is below the spectral resolution of 4 cm-1 and thus it is inconclusive.

In the case of neat Nafion, the S-O band, located at 1057 cm-1, is slightly lower

than that of the dried Nafion is 1061 cm-1, which implies that some bound water might be

present in our Nafion at 100 oC. According to Falk, this symmetric S-O stretching band

shifts to a lower wavenumber upon hydration engaging in the water-water and/or water-

ion interactions [148,149]. However, the S-O band of neat Nafion is noticeably shifted

from 1057 cm-1 to 1048 cm-1 in the impregnated Nafion. The shifting of the S-O band to a

lower wavenumber upon HB impregnation suggests the occurrence of complexation of S-

O possibly with O-H of carboxylic acid group of the HB supramolecules bridging

through some bound water molecules. A natural question is why the C=O band shows

little or no movement. It may be hypothesized that when the bound water in HB interacts

with S-O of Nafion, some C=O bonds may be free up, thereby contributing to a red

spectral shift. This may be compensated by the blue shift of C=O if inter-species complexation were to occur between the carboxylic groups of HBPEAc-COOH and the sulfonate groups of Nafion, e.g., hydrogen bonding.

In order to quantify the estimated amount of HBPEAc-COOH incorporated in the

Nafion membrane, two different FTIR approaches were employed. Note that the FTIR spectra were collected after wiping off any residual HBPEAc-COOH that might be deposited on the surface of Nafion membranes. After drying, the impregnated membrane was photocured in accordance with the procedure described in the experimental section.

The first approach focuses on the peak (position) shift of S-O band with HB polyester

136

content and the second approach utilizes the integrated area under the C=O and C=C

aromatic peak at 1727 cm-1 and 1508 cm-1 to determine the amount of HBPEAc-COOH

in Nafion.

As shown in Figure 5.7(a), the S-O band at 1057 cm-1 moves to a lower

wavenumber (i.e., about 9 cm-1) with increasing loading level of HBPEAc-COOH in

Nafion, but this movement levels off at about 5wt% of HB polyester in the feed suggesting that the impregnation of HBPEAc-COOH into the Nafion ionic domain has reached a saturation level. In the second method, the area under the curves of C=O and

C=C aromatic band were calculated as described before. The dashed line indicates an ideal situation, where the amount of HBPEAc-COOH in the impregnated membrane conforms exactly to that of the feed. As seen in Figure 5.7(b), the estimated amount of

HBPEAc-COOH that is incorporated into the Nafion conforms relatively well with the ideal situation up to 5wt% and then it saturates out after 10wt%. Hence, the true amount of incorporated HBPEAc-COOH was estimated to be about 7wt%. Based on the FTIR investigation, it may be concluded that the estimated amount of HB polyester incorporated is about 5 to 7wt% which is in good accord with the relative intensity quantification by SSNMR. The FTIR and SSNMR studies revealed that the Nafion membrane was indeed impregnated with the HB polyester network, through occupying the hydrophilic packets of the membrane and interacting with the SO3H groups.

137

Figure 5.7: (a) Plot of wavenumber of the S-O stretching band of FTIR as a function of HBPEAc-COOH in feed demonstrating a considerable shift of the S-O band to a lower wavenumber (i.e., about 9 cm-1) indicating inter-species hydrogen bonding interaction of the carboxylic acid of HBPEAc-COOH with the sulfonic acid terminal group of Nafion; (b) quantification of the estimated amount of HBPEAc-COOH present in Nafion membrane based on the integrated area under the curve of the FTIR spectra from C=O and C=C aromatic stretching band showing ~7% of HBPEAc-COOH incorporated in the impregnation process. The dashed line represents the actual amount of HB incorporated in the ideal situation.

5.3.3. Water Uptake and Ion Exchange Capacity (IEC) of HBPEAc-COOH Impregnated

Membranes

Swelling of Nafion membrane is crucial in the presence of polar aprotic and protic

liquids or solvents. Figure 5.8(a) shows the plot of water uptake as a function of soaking time for both neat and impregnated membranes at room temperature. The water uptake of neat Nafion in acid form was 37wt%. Elliot et al. [53] demonstrated the swelling of the ionic domains by water in turn plasticizes the fluorocarbon matrix. As shown in Figure

5.8(a), this swelling can be suppressed considerably by incorporating HBPEAc-COOH into the Nafion, i.e., the water uptake drops from 15wt% to 10wt% as the composition of 138

HBPEAc-COOH in feed increases from 3wt% to 10wt%. Note that the above water

uptake experiment is important to evaluate the dimensional stability, since the physical

dimensions of both impregnated membrane and neat Nafion are difficult to compare due

to the severe warping displayed especially by the pure Nafion.

In Figure 5.8(b) is shown the number of moles of water per mole of sulfonic acid

site (λ) as a function of HBPEAc-COOH composition in feed. In neat Nafion-acid, up to

22 moles of water are absorbed per mole of sulfonic acid site after 24 hours of soaking; this number increases with water temperature [62]. However, in the case of the impregnated membrane, only 7 moles of water can be absorbed by one sulfonic acid site

as the number of moles of water saturates out at ~10wt% of HBPEAc-COOH in feed.

Upon hydration, the fluorocarbon matrix undergoes reorganization, which imposes large internal stresses on the membrane resulting in warping as depicted in the inset right picture of Figure 5.8(b). Consequently lower swelling capability and minimized internal stress within the impregnated membrane leads to better dimensional stability upon hydration as seen in the inset left picture of Figure 8(b). It can be envisioned that the sulfonic acid sites have been surrounded with HBPEAc-COOH molecules and this shielding of the ionic sites lowers the hydrophilicity of the membrane. The formation of the cured HB polyester network has contributed to improved dimensional stability of the

Nafion membrane.

Furthermore, Figure 5.8(b) demonstrates that the ion exchange capacity (IEC) of impregnated membranes exceeds the IEC of neat Nafion, i.e., ~ 0.92 meq g-1, showing increasing trend of the IEC value with addition of HB.

139

The IEC for the impregnated membrane, however, saturated out at 5 wt% of HB in feed with a maximum value of ~1.41 meq g-1. The enhancement of IEC may be

attributed to the increase in the number of acidic protons, i.e., the carboxylic acid sites in

HB within the ionic clusters.

Figure 5.8: (a) Plot of water uptake as a function of soaking time for both neat and impregnated membranes at room temperature. The water uptake of neat Nafion in acid form was 37wt% while the impregnated Nafion membranes show a remarkable reduction - in water uptake; (b) graph of λ (moles of H2O per mole of SO3 ) of impregnated Nafion membranes as a function of HBPEAc-COOH composition in feed (after soaking for 24 h), illustrating significant reduction of water molecules due to impregnation. The IEC curve shows an increment trend (i.e., proton storage capacity) with HBPEAc-COOH loading. Insets show photograph of neat Nafion membrane (10 mm in width and 22 mm in length) exhibiting severe warpage upon hydration and significantly less warpage in the 10% HBPEAc-COOH impregnated Nafion membrane demonstrating improved dimensional stability.

140

5.3.4. Dynamic Mechanical Properties of HBPEAc-COOH Impregnated Membrane

DMA was performed to mimic the viscoelastic properties of the impregnated

membrane by analyzing the relaxation temperature of the ionic domain, i.e., α-relaxation.

Figure 5.9 shows the plots of storage modulus, E’ and tan δ curves versus temperature for

10% HBPEAc-COOH impregnated Nafion and neat Nafion membranes. In the neat

Nafion-acid, there are three relaxations termed γ, β and α-relaxations in ascending order

of temperature from -150 to 160 oC that were observed at -70, 15 and 100 oC

respectively. Kyu et al. [79] assigned these γ-, β- and α-relaxations to the –CF2 local motions, glass transition of Nafion matrix and glass transition of the ionic domain, respectively. Underwater stress relaxation of neat Nafion in acid and sodium forms showed that the α-relaxation temperature is profoundly affected by the presence of water in the ionic regions [80]. One drawback of the neat Nafion-acid form is that the hydrophilic ionic domains become extremely mobile above this α-relaxation temperature and the interconnected channels will eventually disrupt upon prolonged exposure to elevated temperature.

The impregnation of Nafion by HBPEAc-COOH raised the α-relaxation temperature from 100 oC to 130 oC, thereby improving the thermal stability of the

membrane. This improved α-relaxation temperature may be due to the inter-species

hydrogen bonding interaction and/or the chemical crosslinking of the supramolecules

contributing to the enhanced thermal stability as demonstrated earlier in the FTIR study.

In the inset of Figure 5.9, are shown the pictures of membrane samples taken after DMA

experiments at 150 oC. It is striking to witness the brownish color of neat Nafion, i.e., the

141

manifestation of the thermal instability of protonated ionic domains, turned into light

beige in the impregnated membrane, confirming improved thermal stability at 150 oC.

With descending temperature, an additional relaxation peak can be discerned at

o ~90 C, Figure 5.9. This peak temperature is analogous to the DSC Tg of cured neat HB

polyester and thus it is reasonable to assign it to the glass transition of cured HBPEAc-

COOH within the impregnated membrane. The original β-relaxation peak of neat Nafion,

i.e., the glass transition of the fluorocarbon backbone of Nafion, slightly decreases from

15 oC to 5 oC (Figure 5.9), but the tan δ value becomes more pronounced. However, it should be pointed out that this β-relaxation peak is considerably overlapped with the onset of the Tg peak of HB supramolecules. This overlap might have contributed to the

enhanced movement of the fluorocarbon backbone chains, thereby increasing intensity of

the β-relaxation peak. Recall the finding of the SSNMR of 19F-13C CP/MAS in Figure 5.4

that shows virtually no reactivity between the HB supramolecules and the fluorocarbon

backbone matrix in the dried state, except the narrowing of the linewidth of the –CF2

backbone of Nafion. This SSNMR narrowing indicates that the fluorocarbon chain is

becoming more mobile, which is consistent with the enhanced strength of the β-

relaxation peak caused by the HB impregnation (Figure 5.9).

Another important observation is the increase in the storage modulus values of the

impregnated membrane. As can be seen in Figure 5.9, the modulus of the HBPEAc-

COOH impregnated Nafion membrane increases over one order of magnitude at lower

temperatures, which may be attributed to the HBPEAc-COOH network. Although both

storage moduli of impregnated Nafion and neat Nafion membranes decrease with

142

temperature, the crosslinked HB network yielded a higher storage modulus at elevated temperatures (100 oC ~ 130 oC).

Figure 5.9: Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and 10% HBPEAc-COOH impregnated Nafion membrane (dashed line). An additional relaxation peak, attributable to the Tg of the cured HBPEAc-COOH, appears in the impregnated membrane at around ~90 oC. The inset photographs show improved thermal stability of the impregnated membrane showing light beige as compared to the dark brownish color of neat Nafion-acid after DMA experiments at 150 oC. The sample dimension was 10 mm in width x 25 mm in length.

5.3.5. Proton Conductivity of HBPEAc-COOH Impregnated Membrane

In order to determine whether or not there is any improvement in the proton

transport of the HBPEAc-COOH impregnated membrane over neat Nafion, proton

conductivity measurements were conducted by AC impedance fuel cell tests at humidity

levels of 100% RH and 74% RH as a function of temperature. Figures 5.10(a) and (b)

143

illustrate log-log plot of complex impedance versus frequency plots obtained for the

90/10 Nafion/HBPEAc-COOH impregnated membrane at 74% RH. The experiment at

100% RH was not feasible beyond 115 oC with the current instrument due to the

equipment limitation. The real impedance (Z’) obtained shows a sigmoidal reduction of

impedance values with frequency, whereas the imaginary component (Z”) reveals a

corresponding peak around ~20 Hz. The Z” peaks slightly shift to higher frequencies with

temperature. The reduction of storage impedance with frequency implies an increasing

ionic (proton) resistance of the impregnated electrolyte membrane. It was found that the

complex impedance plots of the impregnated membrane at 100% RH showed a similar

trend to that at 100% RH, except for slightly different AC impedance values (data not

shown).

Next, the Cole-Cole plots were constructed over the frequency range of 0.1 Hz-10

kHz. Figure 5.10(c) illustrates the Cole-Cole plot of 10wt% HBPEAc-COOH

impregnated membrane at 74%. The first intersection of the plot at the x-axis is denoted

as Rs indicating the contact resistance between the electrolyte membrane and the catalyst layer; known as the solution resistance. The diameter of the semi-circular plot, denoted

by Rp is taken as the overall cell resistance or polarization resistance. The magnitude of

Rp represents the true resistance from which the conductivity value may be evaluated.

Conductivity, σ, was then calculated in accordance with Equation (2.32) as

explained in the preceding chapter. An equivalent circuit of the cell may be represented

by a resistor in series with a parallel circuit containing a capacitor and a resistor.

144

Figure 5.10: Plots of (a) real impedance versus log frequency and (b) imaginary impedance versus log frequency of 90/10 Nafion/HBPEAc-COOH at 74% RH measured from 30 oC to 100 oC with 10 oC increment and (c) Cole-Cole plots of 90/10 Nafion/HBPEAc-COOH impregnated membranes generated at 74% RH. The diameter of semi-circular plots decreases with temperature indicates lowering of the overall cell resistance thus leading to better proton conductivity properties.

145

Figure 5.11 shows the plots of proton conductivity as a function of temperature from 70 oC to 115 oC at 74% RH of neat Nafion and impregnated membranes. The proton conductivity of neat Nafion is the highest at 80 oC, but gradually declined with increasing of temperature. Since the proton conductivity of Nafion consists of both proton (H+) and

+ + hydroniums (e.g., H3O , H5O2 ) transport, the loss of water above the boiling temperature can significantly reduce the overall proton conductivity of the hydrophilic clusters [68].

Further increment of temperature results in possible disruption of the percolated ionic channels and the ionic clusters become more isolated thereby reducing the efficiency of proton to conduct.

The impregnated membranes with 10wt% of HB, on the other hand, exhibits an increasing conductivity trend with temperature afforded by the acidic protons from the pendant carboxylic groups from the HB supramolecules, which gained higher mobility when temperature is raised. The conductivity value of the impregnated membrane increased from 0.040 S cm-1 at 80 oC to 0.056 S cm-1 at 115 oC and at 74% RH, which eventually exceeded that of the neat Nafion (σ = 0.043 S cm-1 at 115 oC and at 74% RH).

This observation is not surprising in view of the fact that a sizable portion of the Nafion ionic domain was occupied by the HB networks, which in turn reduced the water uptake.

Incorporation of hyperbranched polyesters within the ionic clusters lowers water uptake and therefore lowering conductivity in the impregnated membrane. By reducing the HB amount to 4wt%, more water molecules can be accommodated with the ionic domains and thus the proton conduction can increase slightly at low temperatures as compared with the 10 wt% sample. At elevated temperature above 100 oC, it shows the proton

146

conductivity behavior that is intermediate between the neat Nafion and the 10 wt%

impregnated membrane.

Figure 5.11: Proton conductivity plots of neat Nafion and impregnated membrane as a function of temperature at 74% RH. Proton conductivity of neat Nafion plummets beyond 80 oC while the conductivity value of the impregnated membrane continues to increase with temperature. The impregnated membrane surpassed the conductivity of that neat Nafion at 110 and 115 oC indicates the improved thermal stability and proton conduction of the impregnated membrane.

Of particular importance is that the proton conduction of the HBPEAc-COOH impregnated membrane showed an increasing trend that eventually exceeded that of neat

Nafion at 110 oC and 115 oC, even though with only slight improvement in the proton

conductivity value. It can be anticipated that the hydrogen bonded carboxylic acids

groups present in the ionic clusters/HBPEAc-COOH networks will be ionized with water

147

vapor when the temperature is raised, resulting in free hydrogen ions that ultimately

promotes the proton conduction process under the minimal water condition. Hence, it is

reasonable to infer that the proton conducting mechanism of the impregnated membrane

at elevated temperatures is becoming less dependent on the free water molecules but

more on the ionized protons of the carboxylic acid groups. Although the present proton

conductivity measurement is limited to 115 oC due to the present instrumental configuration, the impregnated membrane proved to be more thermally stable up to the experimental temperature of 150 oC as manifested in the previous DMA experiment.

There is a concern that the HB polyester in Nafion membrane tends to undergo

hydrolysis, which is prevalent in most common polyesters. Nafion itself may suffer to

some level of membrane deterioration upon prolonged or repeated usage in fuel cells due

to the formation of peroxide radical originating from the platinum/carbon supported

catalyst [150]. HB polyesters might as well be susceptible to such radical attack. Thus, neat Nafion and HB impregnated membranes were subjected to cyclic AC impedance measurements in proton fuel cell environment by heating and cooling from 30 oC to 100

oC at 100% RH for a total of 5 cycles. Figure 12 exhibits the heating and cooling cycles

of neat Nafion and impregnated membrane and there is a subtle proton conductivity

hysteresis between the heating and cooling cycles of neat Nafion (only the first and fifth

cycles are shown for clarity) starting from 70 oC up to 100 oC. However, the HB

impregnated Nafion reveals little or no change in the proton conductivity up to the 5

cycles tested, which is equivalent to the total of 160 hours of operation. Contrary to the

general perception on possible hydrolysis of conventional polyester, the present AC

148

measurements on the impregnated membranes show no sign of deterioration in

performance in the repeated cycles under the true proton fuel cell conditions up to 100

oC. It may be hypothesized that the formation of photocured HB network within the

Nafion might have improved the stability of the polyester.

Based on the observed SSNMR, FTIR, DMA and AC impedance behaviors, a schematic drawing in Figure 5.13 is presented to hypothesize a plausible mechanism on aggregated structure within the ionic domains of the HB impregnated Nafion. It seems that the inter-species hydrogen bonding between the –COOH of HBPEAc-COOH and the

–SO3H of Nafion is prevalent. Moreover, the HB polyester networks appear isolated from

the surrounding fluorocarbon matrix, having virtually little or no direct interaction with

the fluorocarbons except for contributing to a slight gain in mobility of the backbone

chains. More importantly, the proton conductivity was found to improve with increasing

temperature, which eventually exceeded that of pure Nafion above 110 oC upon

impregnation of the neat Nafion with HBPEAc-COOH networks.

149

Figure 5.12: Heating and cooling cyclic measurement of neat Nafion and impregnated membrane at 100 %RH demonstrates the subtle proton conductivity hysteresis of neat Nafion after 160 hours of testing. However the HB impregnated membrane tends to show relatively consistent proton conductivity after all 5 cyclic measurements.

COO

H2 CH2 O O C OOC OH O O COO CH2 O O CH2 O OH O O n O O OH O OH HBPEAc-COOH

CF2 CF2 CF CF2 x y n [ OCF CF ] O(CF ) SO H 2 z 2 2 3 Nafion CF3

Figure 5.13: Schematic drawing of the ionic domain of the Nafion membrane before and after impregnation with HBPEAc-COOH showing the isolated cured HB polyester network from the fluorocarbon matrix.

150

5.4. Conclusions

Nafion membranes were successfully impregnated with photocurable

hyperbranched polyester having terminal carboxylic acid functional groups, HBPEAc-

COOH. The FTIR studies suggested the occurrence of complexation, presumably inter-

species hydrogen bonding between the carboxylic group of HB supramolecules and the

sulfonate group of Nafion through bound water molecules. HBPEAc-COOH impregnated

Nafion membranes exhibited lower swelling with improved dimensional stability in the

presence of polar solvents. It was found that impregnating Nafion with HB solution

increases the IEC values and therefore the proton density within the ionic clusters

increases. The enhancement of IEC values in impregnated membranes can be attributed

to the presence of collective protons from carboxylates in HB together with the primary

protons of sulfonate groups in Nafion. Another important finding is the improved thermal stability of HBPEAc-COOH impregnated Nafion, showing the movement of the

mechanical α-relaxation to a higher temperature of 130 oC relative to 100 oC of the neat

Nafion as well as increase in the storage modulus. It may concluded that the impregnated

membrane reveals an increasing trend of proton conductivity with temperature as

opposed to the neat Nafion that shows the declining trend at elevated temperatures.

Moreover, the present AC measurements on the impregnated membranes show no sign of

deterioration in performance under the true proton fuel cell operating conditions in the repeated cycles up to 100 oC, implying the sustained performance of these impregnated membranes. With the above findings, we have established the proof of concept on the

151 incorporation of polymer molecules into the Nafion ionic domains by means of impregnation of hyperbranched supramolecules and photo-crosslinking in-situ.

152

CHAPTER VI

IN-SITU IMPREGNATION OF NAFION WITH NOVEL WATERWHEEL

SUPRAMOLECULES HAVING TERMINAL HYDROXYL FUNCTIONAL GROUPS,

NORIA

6.1. Introduction

Nafion is notoriously known for its excessive swelling upon hydration [51-58], causing dimensional instability with significant reduction in proton conductivity, especially at elevated temperature operations. Water management of Nafion has been another major problem in the proton fuel cell operation due to water flooding. One of the important criteria, yet very challenging, is to ensure that the fuel cell membrane is sufficiently hydrated so as to retain optimal protonic (or ionic) conductivity especially at elevated operating temperatures. Above the glass transition of the ionic domains of

Nafion (i.e., ~ 100 oC), there has been a major concern with the reduced proton conductivity. Therefore, it is vital to modify Nafion and raise the thermal stability above its glass transition, reduce its ion/proton conductivity dependency on water besides improving its swelling properties.

Several attempts have been made by compositing and/or impregnating Nafion with suitable materials, e.g., ionic liquids [15-18] so as to resolve issues on limited

153

operating temperature range and excessive swelling of Nafion. The occurrence of specific

interactions such as inter-molecular hydrogen and ionic bonding has been demonstrated

between these small molecules and the ionic clusters of Nafion through this act of

infusion. However, these ionic liquids shown a tendency of leaching out upon

impregnation and act as potent plasticizers to Nafion, which in turn lower the mechanical

strength and modulus of the impregnated membranes. Contradictory to the above

observations, the increased in modulus and excellent swelling suppression were observed

when Nafion membranes were impregnated with solid photocurable hyperbranched

supramolecules containing carboxylic acid functionality, which were presented and

discussed in Chapter V.

In this present chapter, we have incorporated the uniquely synthesized solid

supramolecule into the Nafion ionic domains. The as-synthesized supramolecule has a

double-cyclic ladder-type oligomer with a central hydrophobic hole with 6 cavities at the

side and terminated with 24 hydroxyl functional groups, named ‘Noria’ to mean

‘waterwheel’ in Latin [141]. Presence of 24 hydroxyl groups surrounding each supramolecule is anticipated to increase proton concentration within ionic clusters of

Nafion while performing like ‘solid water’ molecules at high operating fuel cell temperatures. The Noria supramolecules were infused into the Nafion ionic domains via swelling with mixed methanol and dimethylacetamide solutions. The incorporated amount of Noria and inter-species interaction was investigated by means of Fourier

transform infrared (FTIR). The physical and structural characterization of the Noria

impregnated Nafion was performed using differential scanning calorimetry (DSC),

154

thermogravimetric analysis (TGA), dynamic mechanical analysis (DMA), wide angle x- ray diffraction (WAXD), and small angle x-ray scattering (SAXS) and AC impedance measurements under the proton fuel cell environment. Of particular interest is that the proton conductivity was found to improve with the Noria impregnation, especially at high temperatures where Nafion loses its performance in proton fuel cell operations.

6.2. Experimental Section

6.2.1. Materials and Sample Preparation

Nafion 115 membranes having equivalent weight of 1100 in its original acid form were purchased from Fuel Cell Store Inc. As-received Nafion membranes were pretreated to ensure full conversion into acid form in accordance to the previous section.

The Noria supramolecules were synthesized based on the condensation reaction of resorcinol, as a difunctional compound, with 1,5-pentanedial, [(CH2)n(CHO)2] (n=3) as a

tetrafunctional compound, which yielded a unique double-cyclic ladder-type oligomer with a central hole, resembling a waterwheel structure [141]. Figure 6.1 illustrates the chemical structure of Noria.

Weighed Nafion membranes were pre-swollen in methanol for 24 h and then

immersed in the Noria/dimethylsulfonamide (DMAc)/methanol with 5% wt/wt solution

concentration for 48 h at room temperature. 3:1 ratio of DMAc/methanol solution was

used to dissolve Noria due to the selective solubility characteristics of these

supramolecules. The Nafion to Noria mass ratios were 98/2, 97/3, 95/5 and 90/10. These

ratios correspond to the feed compositions, which will be hereafter referred to as ‘feed

155

ratio’. Impregnated membranes were then removed from the Noria/DMAc/methanol solution.

Figure 6.1: Chemical structural scheme of Noria which consists the components of resorcinol and 1,5-pentanedial and resembles the waterwheel [141].

(a) Initial Noria (b) Impregnation (c) Impregnation (d) Remaining solution process after 12 process after 24 solution after hours hours impregnation process

Figure 6.2: Impregnation process of 95/5 Nafion/Noria where after 24 hours it is evident that the supramolecules have been infused completely within Nafion membrane.

156

All membranes were washed with methanol to remove the excess supramolecule solution on the surface, blotted with tissue paper (e.g., Kimwipes) and subsequently dried in vacuum at 100 oC for 24 h. These dried impregnated membranes were kept in a desiccator prior to use. Figure 6.2 illustrates the impregnation process of 95/5

Nafion/Noria for 24 hours and it is evident that the supramolecules have been infused completely within the Nafion membrane. Note that leaching out tests of Noria supramolecules were carried out by immersing the impregnated membranes in the respective solvent, i.e., mixed DMAc/methanol solvent for 24 h. FTIR spectra measurements show no distinctive peak diminishing or shifting after the leaching tests.

DMAc with 99% purity and ACS grade methanol were purchased from Sigma-Aldrich and used without further purification for dissolving both supramolecules and swelling the ionic domains of Nafion.

6.2.2. Experimental Methods

Thermal stability analysis of neat Noria and neat Nafion in acid form was performed using thermogravimetric analyzer (TGA 2050, TA Instruments).

Approximately 10 mg of the samples were used for each run. Neat Noria was subjected to

TGA analysis from 25 oC to 600 oC at a heating rate of 10 oC min-1 under nitrogen environment with a flow rate of 120mL min-1.

Determination of melting temperature of neat Noria was conducted using a differential scanning calorimeter (Modulated DSC 2920, TA Instruments). The recommended amount of 7-10 mg of each sample was encapsulated in aluminum

157

hermatic pans. DSC heating and cooling scans were carried out from 30 oC to 480 oC at a

rate of 10 oC min-1 unless indicated otherwise. The DSC chamber was purged with nitrogen gas at a rate of 80mL min-1. Prior to TGA and DSC runs, both neat Nafion and

pure Noria samples were annealed at 100 oC for 10 min and subsequently quenched it to

20 oC.

Dynamic mechanical analyses (DMA) were conducted in accordance to the

previous chapter. Unless indicated otherwise, Fourier transform infrared (FTIR)

measurements were conducted in agreement to the previous chapter. Quantification on

the estimated amount of Noria supramolecules in the impregnated membranes was done

by integrating the ratio of area under the FTIR curves within the wavenumber ranges of

1620-1750 cm-1 which corresponded to the C=C aromatic stretching band. The estimated

amount of Noria supramolecules were calculated in accordance to eq (3.1) and physical

weighing measurements were carried out in order to cross-check the estimated amount of

Noria in Nafion membranes relative to those obtained from the integrated area under the

FTIR spectral peaks.

Wide angle x-ray diffraction (WAXD) measurement was conducted using Bruker

x-ray generator (AXS D8) equipped with a copper target tube and a two–dimensional

detector. The x-ray generator was operated at 40 kV and 40 mA using monochromatized

CuKα radiation with a wavelength of 1.5418 Å. A customary detector-to-sample distance

of 15 cm was employed with 2θ covering up to 30°.

2D small angle x-ray scattering (SAXS) analyses of impregnated membranes

were characterized using the 18kW rotating anode x-ray generator (MicroMax-002+)

158 equipped with operating Cu tube of 45 kW and 0.88 mA. The wavelength of the x-ray beam was 1.5418 Å and the standard zero pixel of the 2D SAXS was calibrated using silver behenate. Background scattering was subtracted from the sample scans for all scans conducted. All membranes were isothermally kept at 100 oC for 10 min prior to each scan as to eliminate possible moisture absorption.

Water uptake, ion exchange capacity (IEC) and proton conductivity measurements of impregnated membranes and neat Nafion were conducted as mentioned in the Chapter III. The proton conductivity value is calculated analogous to eq (2.32) as mentioned in Chapter II.

6.3. Results and Discussion

6.3.1. Thermal and Mechanical Stability of Noria Impregnated Membranes

Figure 6.3: (a) TGA thermograms of neat Nafion, impregnated membranes and pure Noria signifying an excellent thermal stability exhibited by the impregnated membranes while (b) DSC scan of pure Noria illustrates the melting temperature at 373 oC which coincided with the temperature at 5% weight loss occurs. 159

Figure 6.3(a) illustrates the TGA scan exhibiting three distinct drops in the weight

loss curve. The initial weight drop at 140 oC may be attributed to loss of bound water

from the Nafion membrane. The second drop occurring at 320 oC may be associated with desulfonation of the sulfonate group of Nafion as reported by Surowiec and Bogoczek

[151]. The final drop in the weight loss at 430 oC may be assigned to the degradation of

the fluorocarbon backbone [152]. In the TGA curve, neat Noria reveals two

decomposition steps. The first weight drop occurring at 130 oC may be attributed to the

loss of absorbed water within the supramolecules. The second TGA weight loss begins at

about 340 oC, but the melting peak of pure Noria at 373 oC coincided with the

temperature at which 5% weight loss occurs as depicted in the DSC scan of pure Noria

(Figure 6.3b). It appears that as soon as the Noria crystal melts, thermal degradation

concurrently takes place. The impregnated membranes exhibit the improved thermal

stability relative to neat Nafion upon incorporation of 3–5 wt% of the Noria

supramolecules.

Dynamic mechanical analysis (DMA) tests were performed from -150 to 160 oC.

Figure 6.4 shows the plots of storage modulus, E’ and tan δ versus temperature for neat

Nafion membranes and Noria impregnated membranes. In the ascending order of

temperature, the neat Nafion-acid exhibits three relaxation peaks at -70, 15, and 100 oC

corresponding to γ, β and α-relaxations. These relaxations are assigned to the –CF2 local

motions, glass transition of Nafion matrix, and glass transition of the ionic domain,

respectively [79]. Exceeding the α-relaxation temperature, the ionic domain becomes

160

very mobile. Reorganization within the ionic domains due to loss of water eventually

leads to disruption of interconnected channels, and thus losing proton transport [153].

Figure 6.4: Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and Noria impregnated Nafion membranes (dashed and dotted lines). The inset photographs show membrane samples taken after DMA experiments at 150 oC. The brownish appearance of the neat Nafion-acid is the manifestation of thermal instability of the protonated cluster whereas the Noria impregnated membrane evidently sustained its inherent color, i.e., the color of the supramolecules powder, confirming enhanced thermal stability up to 150 oC tested. The sample dimension was 10 mm in width x 25 mm in length.

Of particular importance is that the 5wt% Noria impregnation of Nafion raised the

α-relaxation temperature from 100 oC to 135 oC. The inter-species hydrogen bonding

between the sulfonate groups of Nafion and the hydroxyl groups of Noria might be

161

contributing to the observed higher α-relaxation temperature. In the inset of Figure 6.4, are shown the pictures of membrane samples taken after DMA experiments at 150 oC.

The brownish appearance of the neat Nafion-acid is the manifestation of thermal

instability of the protonated cluster. On the other hand, the Noria impregnated membrane

evidently sustained its inherent color, i.e., the color of the supramolecules powder,

confirming enhanced thermal stability up to 150 oC tested. In a nut shell, the

impregnation indeed improves both mechanical and thermal stabilities of the membrane.

6.3.2. Structural Characterization of Noria Impregnated Membranes

Figure 6.5 illustrates the FTIR spectra of neat Nafion, neat Noria, and the

impregnated membranes with different feed ratios of supramolecules. Given the

molecular structure of the supramolecules, Noria is a self-associating molecule due to the presence of 24 highly polar hydroxyl groups surrounding 6 benzene cavities for each molecule [141], which are capable of forming intra-molecular hydrogen bonding as well as inter-molecular hydrogen bonding within the same species. Hence, it is reasonable to assign the broad but intense peak of O-H stretching at 3410 cm-1 to the self-associated

hydroxyl groups of neat Noria (Figure 6.5(a)). The possibility of bound water contributing to this broad O-H band cannot be ruled out, although the FTIR experiments were conducted at 100 oC with minimal absorbed moisture, i.e., bound water. Upon

impregnation of Noria into Nafion membrane the broad O–H stretching peak moves to a higher wavenumber, suggesting that some of the self-associated O-H bonds of Noria may have free up. However, some O-H groups, upon interacting with the functional site of

162

Nafion such as sulfonate groups, may lead to a shift of the O-H stretching band to a lower

wavenumber. The complexation with bound water can also contribute to such blue shift.

The competition between the above opposing mechanisms will eventually dictate

movement of O-H stretching band, if any, with Noria concentration within the ionic

domains.

Figure 6.5: FTIR spectra of neat Nafion, neat Noria, and the impregnated membranes with different feed ratios of supramolecules. These spectra were normalized to the –CF2 backbone peaks of Nafion. (a) 3900–2300 cm-1 region shows the broadness of O–H stretching present in the impregnated membranes whereas (b) 1800 cm-1 to 570 cm-1 region implying that there is little or no influence of Noria on the fluorocarbon chain of Nafion but with a shift of the S–O stretching peak to a lower wavenumber with the addition of Noria.

The neat Nafion membrane shows two strong bands at 1200 cm-1 and 1140 cm-1,

which may be attributable to the C-F asymmetric and symmetric stretching vibrations of

the main chain, respectively (Figure 6.5(b)). The C-F stretching vibration arising from the side chains of Nafion occurs at 980 cm-1. The band at 1057 cm-1 corresponding to the S-O

163

stretching band may be assigned unambiguously to the sulfonate groups of Nafion. This

S-O peak is particularly important to monitor because of its sensitivity to any

modification or environment alteration taking place within Nafion. With increasing Noria

weight in feed ratio from 2% to 10%, there is a 9 cm-1 wavenumber shift of the S-O band of the Nafion sulfonate groups. The present observation is consistent with that by

Tannenbaum et al. [154] who observed the 7 cm-1 wavenumber shift of the S-O

characteristics band upon addition of poly(ethylacrylate-co-4-pyridine) to Nafion. Their finding was attributed to the lowering of polarization of the S-O dipole resulting from the proton transfer between sulfonate groups and positively charged pyridine. However, in the present case, the shifting of the S-O band is most likely due to the inter-species hydrogen bonding between the hydroxyl groups of Noria and the sulfonate groups of

Nafion.

In addition, the characteristic peaks of C–F stretching (both backbone and side chain), C–F bending and–CF2 rocking mode remains stationary implying that there is

little or no influence of Noria on the fluorocarbon chain of Nafion. Note that these spectra

were normalized to the –CF2 backbone peaks of Nafion. Other characteristic bands such

-1 as 2930 cm corresponds to the C-H asymmetrical stretching of –CH2 groups and the

strong bands at 1623 cm-1, 1507 cm-1 and 1440 cm-1 attributable to C=C aromatic

stretching of resorcinol rings are not expected to involve in any specific interactions and

thus may be useful only in the determination of the level of impregnation in accordance

with Equation (3.1).

164

Figure 6.6: (a) Plot of wavenumber of the S–O stretching of FTIR as a function of Noria composition. A systematic movement of the S–O band to a lower wavenumber suggesting an inter-species through hydrogen bonding interaction might occur between the hydroxyl groups in Noria with sulfonate terminal groups of Nafion, (b) Estimated amount of Noria supramolecules incorporated within the Nafion membranes based on the integration of area under the C=C aromatic characteristic peak and physical weighing measurements. The dashed line represents the ideal condition where the amount of Noria incorporated in Nafion increases linearly with composition.

In order to approximately quantify the amount of incorporated supramolecules inside Nafion, three different approaches were adopted; viz., (1) the wavenumber shift of

S-O peak versus Noria feed ratios; (2) integration of the area under the curves of C=C aromatic peak of the respective impregnated membrane; and (3) physical weighing of the

impregnated membranes. Figures 6.6(a) exhibits the shift of S-O band against the Noria supramolecules feed ratios. It is clearly seen that the S-O band in the impregnated

membranes increases with increasing Noria feed ratio initially, but it levels off rapidly

around 5 wt% (Figure 6.6(a)). The shift of the S-O band to a lower wavenumber implies

165

that the sulfonate groups of Nafion formed inter-species hydrogen bonding with the

terminal hydroxyl groups of Noria.

Figures 6.6(b) depicts the amount of Noria present in Nafion membranes based on

the integration of C=C aromatic characteristic peak and the basic physical weighing

approach. Note that the FTIR spectra were collected at 100 oC, whereas the physical

weighing was done at least five times per point on the vacuum dried sample, until the

constant weight was reached. Based on the integration of the C=C aromatic peak, the estimated amount (i.e., ~5 wt%) of Noria was infused in the Nafion membrane. This C=C peak however, shifted to higher wavenumber suggesting that the C=C of resorcinol gained some flexibility as some of the self-associated terminal hydroxyl groups were relieved upon hydrogen bonding to the sulfonate groups of Nafion. The physical weighing approach, although less precise, exhibited reasonable agreement with the FTIR results of both integrated peak area as well as the S-O band movement as shown in Figure

6.6(a). Therefore, the estimated amount of infused Noria into Nafion ionic domain is roughly ~5 wt%.

The WAXD scans revealed the broad amorphous peak of the neat Nafion at 2θ =

17o (Figure 6.7(a)), whereas pure Noria supramolecules exhibited several distinct crystalline peaks at 2θ values of 6.2o, 10.4o, 12.2o and 16.1o, which correspond to the d- spacings of 1.76 Å, 2.94 Å, 3.45 Å and 4.5 Å (Figure 6.7(c)). Although the crystal structure of Noria has yet to be determined, the present WAXD observation confirms the observed crystalline nature of pure Noria by the DSC. However, upon impregnation it

into the Nafion ionic domains, there is no discernible Noria crystal peaks in the WAXD

166 scan (Figure 6.7(b)). It appears that the Noria becomes amorphous when the supramolecules are confined within the nano-sized Nafion ionic domains. Another possibility is the formation of strong inter-species hydrogen bonding between the supramolecules and Nafion protonic domains, which might have prohibited these supramolecules to form any crystals. The broad WAXD peak of impregnated Nafion membrane at 2θ = 17.6o comes from the amorphous fluorinated backbones.

Figure 6.7: WAXD scan of (a) neat Nafion, (b) 95/5 Nafion/Noria impregnated membrane and (c) pure Noria over the 2θ range of 4o-28o at room temperature demonstrate the amorphous halo of neat Nafion and presence of crystalline peaks in pure Noria. It is evident that there is no discernible crystalline peak of Noria upon impregnation with Nafion.

The incorporation of Noria into the Nafion ionic domains is expected to cause some changes in the ionic domains, which may be evaluated by small angle x-ray scattering (SAXS). Figure 6.8 illustrates the SAXS scans of neat Nafion and impregnated

167

Nafion/Noria membranes with varying feed ratios. The SAXS profile of neat Nafion in

the acid form showed a very broad SAXS peak around q = 0.08 Å-1, which corresponds

to the inter-domain spacing of 78.5 Å (Figure 6.8(a)). Note that the arrow indicates the peak position of Iq2 versus q plot (data not shown). The observed domain size is

somewhat larger than the literature value of 40 ~ 50 Å of the dried Nafion. According to

Gierke et al. [41] and Fujimura et al. [47] the average ionomer domain size will increase

significantly due to water absorption or upon neutralizing with larger counterions. There

is also a possibility that the present Nafion sample might have absorbed some moisture

from the air during handling and mounting of our sample. With addition of 2 wt% Noria

(Figure 6.8(b)), this SAXS peak moves to a smaller wavenumber of q = 0.063 Å-1.

Increasing the feed ratios of Noria to 5 wt%, the SAXS peak further moves to q = 0.041

Å-1 and thereafter, it remains stationary up to 10 wt% of Noria which corresponds to the

inter-domain spacing of impregnated Noria to be 152.8 Å. The increment of the domain size (or radius) for 2 times corresponds to the expansion of volume by 8 folds assuming that the domains are spherical and in close contact. Although the incorporation 5 wt%

Noria to the total weight of Nafion seems not large, the actual incorporated Noria by volume relative to the average Nafion ionic domains is significant.

168

Figure 6.8: (a) SAXS scans of neat Nafion and impregnated Nafion/Noria membranes with varying feed ratios. Note that the arrow indicates the peak position of Iq2 versus q plot. (b) Plot of domain spacings as a function of waterwheel supramolecules in feed in which the domain spacings increases and remains stationary up to 10 wt% thereafter. The increment of the domain size (or radius) for 2 times corresponds to the expansion of volume by 8 folds.

6.3.3. Water Uptake and Ion Exchange Capacity (IEC) of Noria Impregnated Membranes

Water uptake measurement of Noria impregnated membrane was carried out at room temperature. In Figure 6.9 is shown the plot of number of mole of water per sulfonate site, λ, as a function of impregnation level. The water uptake in Noria impregnated membranes is reduced by 73% as compared to the neat Nafion, i.e., λ of ~6 changes to ~22 moles of H2O per sulfonate group. This reduction in water uptake shows a positive contribution towards the dimensional stability of the impregnated membranes.

According to Paddison et al. [70] the ab initio calculation for the elucidation of the proton

- transport mechanism of Nafion revealed that only 6 H2O/SO3 are needed to exhibit the complete proton dissociation resulting in efficient proton conduction. Even though the 169

proton transport of Nafion is primarily governed by the amount of water, the inherent

hydroxyl functional terminal groups of the present Noria supramolecules can conduct

proton at high temperature operating conditions without the concern of water

evaporation. This will be discussed later in relation to the results of AC impedance fuel

cell experiment.

Figure 6.9: Plot of number of mole of water per sulfonate site, λ, as a function of impregnation level. The water uptake in Noria impregnated membranes is reduced by 73% as compared to the neat Nafion. Note that λ was determined after 24 h hydration in deionized water at room temperature. On the other hand, the IEC of Noria impregnated membranes show increased of 1.39 meq g-1 at 5% feed ratio as compared to the IEC of neat Nafion which is 0.92 meq g-1.

170

Determination of ion exchange capacity (IEC) of that impregnated membrane was

carried out using a titration method against NaOH (0.01N) at room temperature. Also in

Figure 6.9 is shown the variation of IEC of the impregnated membranes as a function of

Noria amount in feed. The IEC value of neat Nafion was found to be 0.92 meq g-1. The

IEC of Noria impregnated membranes increased initially with the feed ratio, but it

saturated out at 5 wt% with the IEC value of 1.39 meq g-1. The enhancement of IEC in

Noria impregnated membranes primarily arises from the hydroxyl groups of Noria and

the primary protons of sulfonate groups in Nafion.

6.3.4. Proton Conductivity of Noria Impregnated Membrane

The AC impedance fuel cell test was undertaken to determine the proton

conductivities of both neat Nafion and the Noria impregnated membrane. Figures 6.10(a)

and (b) illustrate a typical log-log plot of storage impedance versus frequency and a semi-

log plot of loss tangent versus frequency obtained for the 95/5 Nafion/Noria impregnated membrane at 74% RH. The experiment at 74% RH was not feasible beyond 115 oC with

the current instrument. The real impedance (Z’) obtained shows a sigmoidal reduction of

impedance values with frequency, whereas the loss tangent peaks show a consistent

movement towards higher frequency as temperature is raised. The reduction of storage

impedance with frequency implies lesser ionic (proton) resistance of the impregnated

electrolyte membrane with increasing temperature.

171

Figures 6.11 show the Cole-Cole plot of (a) neat Nafion and (b) 5wt% Noria impregnated membrane obtained at the frequency range of 0.1 Hz–10 KHz at 74% RH.

The semi-circular Cole-Cole plots can be represented by equivalent circuit of a resistor in series with parallel capacitor and resistor circuit, where Rp is the polarization resistance

measured within the cell representing the true resistance of the electrolyte membrane and

Rs is the solution resistance, which a measure of the contact resistance between the

electrolyte membrane and catalyst layer. Subsequently, the proton conductivity at a given

temperature was calculated in accordance with Equation (2.32) based on the observed

polarization resistance, Rp [138]. It was found that the Cole-Cole plots of both neat

Nafion and impregnated membrane at 100% RH exhibited a similar semi-circular trend, except for different Rp values (data not shown). From Figure 6.11, it is evident that the

diameter of the semi-circular plots contracted with increasing temperature, indicating that

the resistance of the cell is reduced as the temperature increases. Increasing of

temperature ultimately promotes better mobilization of water and hydronium ions

contained within the sulfonate groups, which in turn enhances the proton conducting

behavior thus less resistance within the fuel cell.

Figure 6.12 reveals the proton conductivity of neat Nafion and Noria impregnated

membranes from 70 to 115 oC at 74% RH. The proton conductivity of the neat Nafion is

the highest at the operating temperature of 80 oC, beyond which it shows a declining

trend. When temperature exceeded 100 oC, any absorbed water within the ionic domain evaporates leading to disruption of the percolated ionic channels. Besides, this temperature (i.e., 100 oC) corresponds to the α−relaxation temperature of the ionic

172

domains where the storage modulus declines, thus losing the mechanical strength. Since

+ + the transport of protons, especially that of hydronium ions (e.g., H3O , H5O2 ), strongly

depends on the interconnectivity of these percolated channels, any loss of water above its

boiling point will significantly reduce the overall proton conductivity [68].

Figure 6.10: (a) Log-log plots of storage loss and (b) a semi-log plot of loss tangent versus frequency obtained for the 95/5 Nafion/Noria impregnated membrane at 74% RH. The real impedance (Z’) demonstrated a sigmoidal reduction of impedance values with frequency, whereas the loss tangent peaks show a consistent movement towards higher frequency as temperature is raised.

173

0.4 0.4 (a) Neat Nafion 40 C (b) 95/5 Nafion/Noria 40 C with 74% RH 50 with 74% RH 50 60 60 0.3 70 0.3 70 80 80 90 90 100 100 0.2 110 0.2 110 115 115 Z” (ohm) Z” Z” (ohm) Z”

0.1 0.1

0 R Rp 0 s R R 0 0.1 0.2 0.3 0.4 0 s 0.1 0.2 0.3 p 0.4 Z’/ ohm Z’/ ohm Figure 6.11: Cole-Cole plot of (a) neat Nafion and (b) 5% Noria impregnated Nafion membranes at 74% relative humidity from 40 oC to 115 oC over the frequency range of 0.1 Hz–10 kHz. It is evident that the diameter of the semi-circular plots contracted with increasing temperature, indicating that the resistance of the cell is reduced as the temperature increases.

At lower temperatures, the 5 wt% Noria impregnated membrane shows a lower

proton conductivity (i.e., nearly one-half of that of Nafion) due to the reduced

hydrophilicity within the ionic domains. At the relative humidity of 74%RH, the proton

conductivity of Noria impregnated membrane was found to surpass that of neat Nafion

with the proton conductivity values of σ = 0.066 and 0.068 S cm-1 at 110 oC and 115 oC,

correspondingly. These values are considerably larger as compared to the neat Nafion at

the same temperatures and humidity; viz., σ = 0.053 and 0.043 S cm-1, i.e., 24 ~ 58%

improvement respectively. The AC impedance fuel cell measurements at 74% RH were

not feasible beyond 115 oC due to the limitation of the current equipment. Besides the

cumulative protons (in Noria), presence of hydrogen bonding within Noria impregnated

174

membrane readily provide the path for proton transfer [155, 156] to occur thus resulting

in an increment trend of its proton conductivity. Kreuer [156] explained that the tendency

of proton transfer to occur is highly enhanced with formation of structural and dynamic

disorder of hydrogen bonding network. As pointed earlier from SAXS study that the

5wt% of Noria impregnation corresponds to the 8 fold increment of ionic domain size by

volume assuming that the domains are spherical and in close contact. If one backs off by

reducing the Noria to 3wt%, there will be more water available for proton conduction at

low temperatures, yet sufficient ‘solid waterwheel’ will be available at elevated

temperatures above 100 oC, and thus higher proton conductivity values can be achieved at

all temperatures relative to those of the 5 wt% sample. Lo and behold, the 3wt%

impregnated Nafion indeed shows the linear increment trend of proton conduction with

the proton conductivity value of 0.056 S cm-1 at 70 oC increasing to 0.069 S cm-1 at 115

oC. It should be emphasized that even though with reduced water uptake of Noria

- impregnated membrane (i.e., 6 H2O/SO3 ), this hybrid impregnated membrane still

manage to perform an exceptional proton conductivity properties especially at higher

temperatures above 100 oC.

We later used the Noria-BOC, where the O–H groups are replaced with the hydrophobic tert-butyloxycarbonyl (–BOC) groups to confirm our hypothesis that it is the hydroxyl group (i.e., solid water) of Noria that is contributing to the improved proton conductivity at elevated temperatures. The synthesis procedure of Noria-BOC may be found elsewhere [141]. Further characterization of Noria-BOC impregnated membrane will be discussed in the Chapter VII. There is a tendency of declining trend in proton

175

conduction of the Noria-BOC impregnated membrane at 110 oC and 115 oC (data may be

found in Chapter VII) thus implying that it is the hydroxyl group of Noria that serves as

‘solid water’ or ‘solid proton carrier’ in the Noria impregnated membrane for improved

proton conductivities. Hence, there is potential for these supramolecules to function as

the water replacement during the proton conduction process in the limited water

operating condition.

Figure 6.12: Proton conductivity plot neat Nafion and Noria impregnated membranes from 70 to 115 oC at 74% RH. Evidently the proton conductivity of neat Nafion exerted a declining trend whereas the Noria impregnated membranes demonstrate an increment proton conductivity trend which surpasses to that of neat Nafion at 110 oC and 115 oC. Later, it was found that the Noria-BOC impregnated membranes which utilizes the hydrophobic tert-butyloxycarbonyl (–BOC) group waterwheel supramolecules shows poor proton conductivity at all temperatures tested and drops its value comparable to those of the pure Nafion at higher temperatures of 110 and 115 oC which confirms that the hydroxyl group (i.e., solid water) of Noria contributes to the improved proton conductivity at elevated temperatures. 176

To test the performance stability of the membranes, both neat Nafion and the

impregnated membranes were subjected to cyclic AC impedance fuel cell measurements

by heating and cooling from 30 oC to 100 oC at 100% RH for a total of 5 cycles. In Figure

6.13(a), there is a subtle proton conductivity hysteresis between the heating and cooling

cycles of neat Nafion (only the second and fifth cycles are shown for clarity) from 70 oC

to 100 oC. It is seen from the data collected, the neat Nafion showed systematic drops in

the proton conductivity values after each cycle. However, the Noria modified Nafion

shows no discernible change in the proton conductivity showing consistent increasing

trend for at least 5 cycles tested, indicating the sustained performance of the impregnated

membrane under the fuel cell environment.

Figure 6.13(b) further illustrates the proton conductivity of the 5wt% Noria impregnated membrane at 100 oC and at 100% RH in comparison with that of the neat

Nafion as a function of heating cycle. Of particular interest is that the Nafion showed a

slight reduction in proton conductivity for each cycle, whereas the Noria impregnated

membranes show excellent sustained performance during the cyclic AC impedance

measurements under the proton fuel cell conditions. More importantly, the impregnated

membrane exhibited thermal, mechanical and chemical stability exceeding those of the

neat Nafion after treatment of a total of 160 hrs of the cyclic experiments. This signifies

sustained performance of the Noria impregnated membrane during prolonged proton fuel

cell operations at elevated temperatures. Figures 6.14 show the FTIR spectra of Noria

impregnated membrane (a)before and (b)after the fuel cell cyclic measurement. Evidently

177 there is no noticeable difference in FTIR spectra before and after the proton fuel cell tests, implying the chemical stability of the Noria impregnated membranes.

Figure 6.13: (a) Heating and cooling cyclic measurement of neat Nafion and impregnated membrane at 100% RH. Neat Nafion shows subtle proton conductivity drop after 5 cycles (i.e., 16 hrs per cycle) whereas the impregnated membrane showed sustained performance upon prolonged fuel cell tests. Note that the first, third and fourth cycle runs were omitted in the plot for clarity. (b) Plot of proton conductivity of the 5wt% Noria impregnated membrane at 100 oC and at 100% RH in comparison with that of the neat Nafion as a function of heating cycle which demonstrated the excellent performance durability of impregnated membrane after total of 160 hours of cyclic measurements.

178

Figure 6.14: FTIR spectra of Noria impregnated membrane (a) before and (b) after subjected to 5 sets of heating and cooling fuel cell measurements.

6.4. Conclusions

In-situ impregnation of Nafion perfluorinated ionomer membrane was successfully carried out using novel waterwheel supramolecules containing hydroxyl functional terminal groups, namely Noria. FTIR analyses showed the occurrence of specific inter-species interaction through hydrogen bonding within the impregnated membranes. The amount of Noria supramolecules through the physical weighing and integration of area under the peak of C=C aromatic band were estimated to be approximately 5 wt%. With the aid of SAXS, the interspacing of ionic clusters of the impregnated membranes showed an increment in d-spacings from 78.5 to 152.8 Å for the

Noria impregnated membrane. This finding confirms the incorporation of waterwheel supramolecules into the ionic clusters. Modification of ionic domains of Nafion with

179 supramolecules has demonstrated reduction in hydrophilic behavior of ionic clusters, which leads to swelling suppression in aprotic liquids such as water from ~22 moles of

- - H2O/SO3 to only ~6 moles of H2O/SO3 . Contrary to the downward proton conductivity trend exhibited by the neat Nafion, the Noria impregnated membranes showed an increasing trend, which eventually exceeded the proton conductivity of Nafion above 110 oC. Of particular importance is that the Noria impregnated membrane exhibited improved thermal, mechanical, and chemical stability with enhanced proton storage capacity relative to those of the neat Nafion. It can be anticipated that these waterwheel supramolecules may have potential utility as high temperature electrolytes (or solid proton carrier) in proton fuel cells.

180

CHAPTER VII

IN-SITU IMPREGNATION OF NAFION WITH NOVEL WATERWHEEL

SUPRAMOLECULES HAVING TERMINAL TERT-BUTYLOXYCARBONYL

FUNCTIONAL GROUPS, NORIA-BOC

7.1. Introduction

Attempts were made by direct impregnation of Nafion with suitable small

molecules to comprehend the heat and water management issues of Nafion. In the

previous section, enhancement of ion exchange capacity, proton conductivity at elevated

temperature, and thermal stability of Noria impregnated Nafion membranes were

demonstrated. To seek the effect of different terminal functional groups, the unique

waterwheel supramolecules terminated with 24 tert-butyloxycarbonyl functional groups, named Noria-BOC were used. These less hydrophilic supramolecules with terminal tert- butyloxycarbonyl, -COOC(CH3)3 groups possessed broader solubility characteristics, thus the in-situ impregnation took place by swelling the ionic domains of Nafion in only methanol containing supramolecule solution. The present chapter is the extended studies

relevant to the previous chapter.

181

7.2. Experimental Section

7.2.1. Materials and Sample Preparation

Nafion 115 membranes with equivalent weight of 1100 in its original acid form

were purchased from Fuel Cell Store. As-received Nafion membranes were pretreated in sulfuric acid to ensure complete acidification of the membrane.

Noria-BOC supramolecules were synthesized using a similar scheme to that of the hydroxyl terminated waterwheel supramolecules (Noria). However, the terminal hydroxyl groups were converted to tert-butyloxycarbonyl functional groups using di-tert- butyldicarbonate [141]. Chemical structure of Noria-BOC is shown in Figure 7.1.

The in-situ impregnation process of Nafion membranes is similar to that described in Chapter VI. ACS grade methanol purchased from Sigma-Aldrich were used without further purification for dissolving both supramolecules and swelling the ionic domains of

Nafion.

7.2.2. Experimental Methods

Thermal stability of both neat Nafion and Noria-BOC was obtained using the thermogravimetric analyzer (TGA 2050, TA Instruments). Approximately 10 mg of the samples were used for each run. Each neat Noria-BOC was subjected to TGA analysis from 25 oC to 600 oC at the heating rate of 10 oC/min under nitrogen environment with a

flow rate of 120mL/min. The temperature at which 5% weight loss occurred was

regarded as the degradation temperature.

182

Fourier transform infrared (FTIR), 2D small angle x-ray scattering (SAXS), water uptake ion exchange capacity (IEC) and dynamic mechanical properties (DMA) measurements were conducted in accordance with the previous chapter.

Unless indicated otherwise, proton conductivity measurements and determination of neat Nafion 115 and Noria-BOC impregnated membranes was conducted in accordance with the preceding chapter.

Figure 7.1: Chemical structure of waterwheel supramolecules with tert-butyloxycarbonyl, namely Noria-BOC [141].

183

7.3. Results and Discussion

In this section, we characterize the Noria-BOC impregnated membranes, and the

observations are presented as a comparison to those of the neat Nafion membrane.

7.3.1. Characterization of Neat Noria-BOC

In order to provide a better understanding of the physical properties of neat Noria-

BOC, these supramolecules were characterized using DSC, TGA and FTIR prior to the

impregnation process. An overlaid plot of DSC and TGA of neat Noria-BOC are

presented in Figure 7.2. It is apparent that Noria-BOC exhibits moderate thermal stability

as the degradation starts as early as 155 oC, which is reproducible from our previous

literature [157], and proceeds to abruptly decompose at 185 oC. It is worthy to note that

crystal melting temperature of Noria-BOC coincides with its degradation temperature.

Figure 7.2: DSC and TGA thermograms of pure Noria-BOC where the melting temperature coincides with degradation temperature.

184

The overlay FTIR spectra of neat Nafion and neat Noria-BOC are depicted in

Figure 7.3. The spectrum of the neat Nafion membrane shows two strong bands at 1200

cm-1 and 1140 cm-1, attributable to the C-F asymmetric and symmetric stretching vibrations of the main chain, respectively, along with the C-F stretching vibrations from the side chains of Nafion at 980 cm-1. In the case of the neat Noria-BOC spectrum, peaks

of 2982 cm-1 and 2935 cm-1 are assigned to C-H stretching while the sharp intense peak at

1760 cm-1 is due to the C=O stretching of the tert-butyloxycarbonyl terminal groups.

C=C aromatic stretching and C-O-C linkage peaks are located at 1497 cm-1 and 1145 cm-

1 respectively.

Figure 7.3: Overlay FTIR spectra of (a) neat Nafion in acid form and (b) pure Noria-BOC at 100 oC.

185

7.3.2. Structural Characterization of Noria-BOC Impregnated Membranes

To probe the inter-species interactions of Nafion and waterwheel supramolecules,

FTIR with attenuated total reflection mode (ATR) was used. The FTIR spectra of neat

Nafion, neat Noria–BOC, and impregnated membranes with different supramolecules

feed ratios are illustrated in Figure 7.4. Spectra were normalized to the –CF2 backbone

peaks of Nafion.

Figure 7.4: FTIR spectra of Noria-BOC impregnated Nafion membranes exemplifies the movement of S–O stretching, which is associated with sulfonic acid sites in Nafion from 1057 cm-1 to 1050 cm-1.

Figures 7.4 shows the overlay spectra of impregnated Nafion membranes with

different Noria-BOC feed ratios in the higher and lower wavenumber range correspondingly. At higher wavenumber region (data not shown), there is an absence of

O–H stretching in all feed ratios of the impregnated membranes. In Figure 7.4, the S-O

186

band in the impregnated membranes starts to shift 7 cm-1 towards lower wavenumber and

remains consistent at 2% feed ratios of Noria-BOC. The movement of S-O band towards

lower wavenumber is due to the possible inter-species interaction of carbonyl groups of tert-butyloxycarbonyl in Noria-BOC with the sulfonic groups of Nafion. Analogous to

the previous studies with Noria supramolecules, the stretching and bending band

associated to the fluorocarbon of Nafion membrane showed little to no movement upon

impregnation. The C=O stretching band that initially seen in the pure Noria-BOC

spectrum showed a slight movement (~4cm-1 shift, which is within the spectral

resolution) towards lower wavenumber upon impregnation indicating the possible

interaction of Noria-BOC supramolecules with Nafion likely through the inter-species

electron acceptor-donor interaction.

In order to estimate the amount of incorporated supramolecules inside Nafion,

three different approaches were used. First, the shift of S-O band were plotted against the

supramolecule feed ratios. Second, by integrating the area under the curves of C=O,

peaks of the respective impregnated membrane were calculated as depicted in the

experimental section. Third, the impregnated membranes were physically weighed.

Figure 7.5(a) illustrates the shift of the S-O band against the feed ratios of

supramolecules. The S-O band shows 7 cm-1 low wavenumber shift with 2 wt % of

incorporated Noria-BOC (Figure 7.5(a)). The S-O band shows little to no movement beyond 3 wt% of the Noria-BOC in feed. Presence of the terminal functional groups of tert-butyloxycarbonyl in Noria-BOC allows the interaction of supramolecules with the sulfonate groups of Nafion through the possible inter-species electron acceptor-donor

187 type of interaction. Due to its bulky and less polar terminal groups, smaller amounts of

Noria-BOC supramolecules can occupy the ionic clusters of Nafion. Based on Figure

7.5(a), it is estimated that amount of Noria-BOC that can be incorporated in Nafion is about ~2-3wt%.

Figure 7.5: (a) Plot of wavenumber of the S–O stretching of FTIR as a function of Noria- BOC in feed. The systematic movement of the S–O band to a lower wavenumber suggests an inter-species hydrogen bonding interaction might occur between the carbonyl groups in Noria-BOC and the sulfonic acid terminal groups of Nafion. (b) Noria-BOC supramolecules incorporated within the Nafion membranes through two different approaches, i.e., integration of area under the C=O characteristic peaks and physical weighing measurements.

Figure 7.5(b) depicts the amount of Noria-BOC supramolecules present in Nafion membranes based on the integration of characteristic peaks and basic physical weighing.

Note that the FTIR spectra were collected at 100 oC and physical weighing was done on the vacuum dried sample at least five times per point until constant weight was reached.

In Figure 7.5(b), the plots showed the amount of Noria-BOC leveling off around 3 wt% 188

in both methods, i.e., integration of peak and physical weighing measurements. By evaluating the area under the C=O peak, the estimated amount of Noria-BOC inside

Nafion was calculated to be ~2.5 wt%. Physical weighing measurements were carried out as to countercheck the estimated amount of Noria-BOC incorporated within Nafion. It is worthy to note that physical weighing measurement is a crude experimental technique and thus leads to discrepancy on the estimated feed ratios of Noria-BOC. Therefore, we decided to use the same feed ratios of Noria-BOC supramolecules in Nafion, i.e., 5 wt% in all of the next characterization experiments for an analogous comparison to the hydroxyl terminated Noria in the previous chapter.

To probe the incorporation of supramolecules into the ionic domains, small angle x-ray scattering (SAXS) analyses were carried out for both neat Nafion and supramolecule impregnated membranes. Figure 7.6 illustrates the SAXS scans of neat

Nafion and impregnated Nafion/Noria-BOC with varying feed ratios. The SAXS profile of dry neat Nafion in acid form showed a broad ionomer peak around q = 0.08 Å-1 which

corresponds to d-spacing of 78.5. Å. Gierke et al. [41] and Fujimura et al. [47] showed

that this d-spacing corresponds to the ionomer cluster peak and increases with presence

of water or upon neutralizing with larger counterions. Therefore, it should be noted that

even though our neat Nafion was previously dried, there is a possibility that the

membrane absorbed some moisture during the handling and mounting of the sample.

The SAXS scans of impregnated Noria-BOC membranes show that the q value

are slightly larger than that of the neat Nafion. At 3 wt% of Noria-BOC, the ionomer

peak observed is at 0.049 Å-1 which corresponds to a larger d-spacing of 130.4 Å.

189

However, this q value remains constant with increasing of the Noria-BOC composition.

Presence of steric hindrance of the tert-butyloxycarbonyl terminal groups of Noria-BOC limits the number of these supramolecules that can reside within the ionic domains of

Nafion, thus leading to smaller increment in the domain spacings.

Figure 7.6: SAXS studies of Noria-BOC impregnated Nafion membranes shows the changes of size of ionic domains in the impregnated Nafion membranes with amount of waterwheel supramolecules in feed. d-spacings of ionic domains remain constant as the amount of waterwheel supramolecules reaches 3 wt% Noria-BOC in feed.

7.3.3. Water Uptake and Ion Exchange Capacity (IEC) of Noria-BOC Impregnated

Membrane

Nafion is known to be extremely susceptible to polar solvents and can swell up to twice its initial weight upon hydration in water. Extreme swelling of Nafion membrane

190

leads to unstable dimensional properties as well as lowering the efficiency of this

electrolyte membrane. We hypothesize that the incorporation of supramolecules within

the ionic domains might lead to lower water uptake and promote better dimensional

stability. Figure 7.7(a) is plotted to illustrate the water uptake in percentage of weight

increments as a function of soaking time for neat Nafion and impregnated Noria-BOC membranes (i.e., 2, 3, and 5 wt% feed ratio). All impregnated membranes show reduction of water uptake compared to that neat Nafion. Neat Nafion is highly susceptible to

swelling with 37 wt % increment under hydrated conditions. Impregnated Noria-BOC

membrane showed significantly lower water uptake, which is only 20 wt% increment

with 5 wt% Noria-BOC composition.

By plotting the graph of number of mole of water per sulfonate site, λ, (Figure

7.7b) it is clear that the reduction of water uptake in impregnated membranes with Noria-

BOC is reduced by 57% as compared to the neat Nafion. The reduction of water uptake

in impregnated membranes shows a positive improvement towards the dimensional

stability of the membranes.

Determination of ion exchange capacity (IEC) of that impregnated membrane was

completed using the titration method against 0.01N sodium hydroxide (NaOH). Figure

7.7(b) illustrates the change of IEC of the impregnated membranes with addition of

Noria-BOC. IEC of neat Nafion was calculated to be 0.92 meq g-1 while Noria-BOC

impregnated membrane shows a slight increment of IEC, i.e., 1.01 meq g-1 with 5 wt%

Noria-BOC. Impregnating Nafion with Noria-BOC solution shows little to no influence

on the changes of IEC values due to the stability of terminal –BOC groups upon

191

ionization. However, presence of inter-species interaction through the electron acceptor-

donor type provides a sense of strengthening effect upon thermal stability of this

impregnated membrane. Further discussion on thermal stability of Noria-BOC

impregnated membrane will be discussed in the DMA as well as the proton conductivity

observations and discussion.

Figure 7.7: (a) Plot of water uptake as a function of soaking time for neat Nafion and Noria-BOC impregnated membranes at room temperature. The water uptake of neat Nafion in acid form is 37wt%, while the Noria-BOC impregnated Nafion membranes show a reduction to 20 wt% in water uptake; (b) graph of λ (moles of H2O per mole of - SO3 ) of impregnated Nafion membranes as a function of Noria-BOC composition in feed (after soaking for 24 h), illustrating significant reduction of water molecules due to impregnation. The IEC value of Noria-BOC impregnated membranes bear little to no effect with the value of 1.01 meq g-1 at 5% feed ratio compared to the IEC value of neat Nafion which is 0.92 meq g-1.

7.3.4. Dynamic Mechanical Properties of Noria-BOC Impregnated Membrane

The impregnating approach of Nafion in supramolecule solutions is expected to improve the thermal stability of the present electrolyte membrane through the presence of

192

inter-species hydrogen bonding interaction. This alternative approach will locally modify

the ionic clusters in Nafion by self-directing the supramolecules into the sulfonate groups

of Nafion. Presence of carbonyl groups in Noria-BOC might lead to an inter-species

interaction which ultimately forms possible hydrogen bonding via electron acceptor- donor interaction between the supramolecules and the sulfonate groups of Nafion. DMA was used to probe the viscoelastic properties of the impregnated membrane by analyzing

the relaxation temperature of the ionic domain, i.e., α-relaxation. Figure 7.8 shows the

plots of storage modulus, E’ and tan δ curves versus temperature for neat Nafion

membranes and 5 wt% Noria-BOC impregnated Nafion. In the neat Nafion-acid, there

are three relaxations termed γ, β and α-relaxations in ascending order of temperature from

-150 to 160 oC that were observed at -70, 15 and 100 oC respectively. These γ-, β- and α-

relaxations correspond to the –CF2 local motions, glass transition of Nafion matrix and

glass transition of the ionic domain [79] respectively. Above this α-relaxation

temperature, the hydrophilic ionic domains will become extremely mobile and the

clustered structures eventually disrupt and/or collapse upon prolonged exposure to

elevated temperature.

Upon impregnation with Noria-BOC supramolecules, the α-relaxation peak shows a slight movement to 110 oC. The contribution of inter-species electron acceptor-donor interaction provides some strengthening effect on the ionic domains, thus slightly increasing the glass transition of the clusters. The insets of Figure 7.8 are the pictures of membrane samples taken after DMA experiments at 150 oC. Noria-BOC impregnated

193

membrane was found to retain its initial appearance as compared to the brownish color of

neat Nafion, thus confirming its enhanced thermal stability.

The original β-relaxation peak of neat Nafion, i.e., the glass transition of the

fluorocarbon matrix remain unchanged for Noria-BOC impregnated membrane (Figure

7.8). The γ-relaxation, however, moves to higher temperature for Noria-BOC

impregnated membranes than those of the neat Nafion, signifying that the local motion of

the –CF2 might become more constrained from the bulky ionic domain tethering the

fluorocarbon chain.

Figure 7.8: Change of storage modulus and loss tangent as a function of temperature of neat Nafion-acid form (solid line) and 5wt.% Noria-BOC impregnated Nafion membrane (dotted line). The inset photographs show improved thermal stability of the impregnated membranes with unchanged color of impregnated membranes as compared to the dark brownish color of neat Nafion-acid after DMA experiments at 150 oC. The yellowish color of Noria-BOC impregnated membranes is inherited from the initial color of the waterwheel supramolecules solid. The sample dimension was 10 mm in width x 25 mm in length. 194

7.3.5. Proton Conductivity Properties of Noria-BOC Impregnated Membrane

Proton conductivity of neat Nafion and impregnated membrane was measured

using the AC impedance fuel cell test. Figures 7.9(a) and (b) demonstrate the proton

conductivity of neat Nafion and Noria-BOC impregnated membrane. Proton conductivity

at each temperature was determined in accordance to Equation (2.32) where polarization

resistance, Rp value was taken from the diameter of the Cole-Cole plot (data not shown).

Proton conductivity of neat Nafion exhibits a maximum value of at 80 oC but plummet as

the temperature increases to 115 oC. This is due to the rapid evaporation of water and

possible disruption of the interconnected channels of the ionic clusters of neat Nafion. It

is known that the proton conductivity of Nafion is highly dependent on the protons (H+)

+ + and hydronium ions (e.g., H3O , H5O2 ) transports; thus, loss of water will significantly

reduce the overall proton conductivity. Neat Nafion in acid form was also found to turn

o into a brownish color when temperature exceeded the Tg of the ionic clusters, i.e., ~90 C.

This observation is often referred to as the thermal instability of the protonated ionic

clusters of Nafion.

Impregnating Nafion with Noria-BOC, which has tert-butyloxycarbonyl

functional groups, exerted a predictable trend in its proton conductivity as illustrated in

Figure 7.9. Below 90 oC, Noria-BOC impregnated membrane exhibits lower proton conductivity values compared to the neat Nafion. Incorporation of Noria-BOC

supramolecules results in lowering the hydrophilic properties of ionic clusters in Noria-

BOC impregnated membrane thus declining the proton conductivity. However, starting at

90 oC, Noria-BOC impregnated membrane reveals a slight increment of proton

195

conductivity. It is possible that the bounded water within the ionic clusters remains intact

with the aid of specific inter-species interaction (i.e., electron acceptor-donor interaction),

hence resulting in a slower water evaporation rate when temperature is raised. The proton

conductivity eventually drops further to 115 oC as the bounded water is completely

evaporated.

Figure 7.9: Conductivity plots of neat Nafion and 5% Noria-BOC impregnated membranes as a function of temperature at 74% RH illustrate the decreasing trend of both membranes at elevated temperatures. An error bar indicates the quadruplet readings for each temperature point.

7.4. Conclusions

Modification of Nafion membranes via an impregnation method was carried out

using novel supramolecules, namely Noria-BOC. Self-directing these supramolecules into ionic clusters of Nafion was conducted by swelling the respective Nafion membranes

196

in the methanol solutions of supramolecules. Fourier transform analyses (FTIR), small-

angle x-ray scattering scans (SAXS), atomic force microscopy-tapping mode analyses

(AFM-TM), determination of ion exchange capacity (IEC), dynamic mechanical analyses

(DMA) and AC impedance measurements in fuel cell environment were conducted as to characterize the respective impregnated membranes. Through FTIR analyses, it was demonstrated that inter-species hydrogen bonding possibly via electron acceptor-donor type occurs between these supramolecules with the sulfonate groups of Nafion. The estimation of the amount of Noria-BOC was found to be 3 wt% by measuring the area under the curves of C=O peak for Noria-BOC impregnated membranes. The size of d- spacing of ionic domains in the impregnated membranes was confirmed as 130.4 Å as compared to the neat Nafion-acid which was 78.5 Å through the SAXS analyses. Water uptake showed that the impregnation of Nafion suppresses the membrane swelling, and the IEC showed little to no increment with supramolecules in feed, indicating that there is no change in the proton density within the ionic clusters. The presence of the inter- species hydrogen interaction of Noria-BOC with Nafion led to slight improvement in the thermal stability of the ionic domain with temperature as shown by DMA analyses. In the fuel cell operating conditions, the impregnated membranes were found to exhibit lower proton conductivity due to reduced hydrophilic properties and continue to plummet when temperature is further raised.

197

CHAPTER VIII

OVERALL SUMMARY AND RECOMMENDATIONS

8.1. Overall Summary

In this dissertation, Nafion perfluorinated ionomer membrane was modified by two different approaches. The first approach was made by physical blending and solution casting of the Nafion membrane with a copolymer, poly(vinylidene fluoride- trifluoroethylene) (PVDF-TrFE) to alter the fluorocarbon matrix of Nafion membrane, while the second was done via in-situ impregnation approach, which locally targeted the ionic clusters of the Nafion membrane. The impregnation process was carried out using two different supramolecules, namely photocurable hyperbranched polyesters (HBPEAc-

COOH) and waterwheel supramolecules (Noria).

According to the US Department of Energy, it is the challenge of the present electrolyte membrane, namely Nafion, to outstand the current operating temperature of

80 oC. Additionally, the tendency of Nafion membrane to excessively absorb water must be suppressed as the pooling and overflowing of water during operation will cause dimensional instability as well as operating failure of the fuel cell. The motivation of these studies was to apprehend these requirements through modifying the present electrolyte membrane, i.e., Nafion.

198

Studies on blends of Nafion and ferroelectric copolymer PVDF-TrFE were

presented in Chapter IV. The phase diagrams of Nafion/PVDF-TrFE blends were

constructed with the aid of differential scanning calorimetry (DSC), polarized optical

microscope (POM) and also via the cloud point measurements. An hourglass type of

phase diagram was discerned for the Nafion/PVDF-TrFE blends through self-consistently solving the combined free energy density of Flory-Huggins (FH) for liquid-liquid demixing and the phase field (PF) free energy for crystal solidification. The solution blended mixture may be phase separated during solvent casting or upon thermal treatment, which in turn forms various domain morphologies of Nafion and PVDF-TrFE, including sea-and-island or comingled biphasic structures. The dispersed and bicontinuous morphology of blends was used to describe the sigmoidal trend of the water suppression. The 60/40 blend was found to exhibit both capacitor and proton conductivity properties. This supports the hypothesis that the comingled biphasic structure of PVDF-

TrFE and Nafion may provide individual pathways for electron and proton to transport.

Unfortunately, these comingled pathways compromised the proton conductivity of the intermediate blend (i.e., 60/40 Nafion/PVDF-TrFE).

The modification of ionic clusters of Nafion with photocurable hyperbranched polyesters (HPEAc-COOH) was discussed in Chapter V. These unique HB supramolecules are comprised of the acrylate double bond which offers a site to be photocured through radical polymerization, and the carboxylic acid functionality will afford an ionic interaction or hydrogen bonding with its counterpart. Through in-situ photo-curing of the HB polyester, it is anticipated that these supramolecules will form a fixed network

199

within the ionic domains preventing the supramolecules from leaching out while also

providing an enhanced modulus effect on the membrane. The solid state nuclear magnetic

resonance (SSNMR) measurements demonstrated presence of HB supramolecules inside

Nafion but there was little to no change of the chemical environment imposed on the

fluorocarbon backbone of Nafion. The HBPEAc-COOH impregnated Nafion membrane

exhibited lower swelling with improved dimensional stability in the presence of water.

Inter-species hydrogen bonding interactions between the functional carboxylic acid of

HB polyester with the sulfonate groups of Nafion, which was probed by Fourier

transform infrared (FTIR) studies, gave rise to the enhancement of glass transition

temperature of ionic domains (α-relaxation) of the impregnated membrane. It was found that impregnating Nafion with HB solution increases the ion exchange capacity (IEC).

Proton conductivity behavior of impregnated membrane showed a reasonably increasing

trend with temperature as opposed to the declining trend in the neat Nafion at elevated

temperature.

In Chapters VI and VII, in-situ impregnation of Nafion took place using the novel waterwheel supramolecules with two different terminal functionalities, Noria terminated with highly polar hydroxyl groups and Noria-BOC encapped with tert-butyloxycarbonyl functional groups. The in-situ impregnation was completed by swelling the ionic domains of Nafion in the supramolecules solutions. Infused supramolecules were likely to reside within the ionic clusters forming a specific hydrogen bonding interaction with the sulfonate groups of Nafion, illustrated through the S-O band movement by means of

FTIR. Small angle x-ray scattering (SAXS) demonstrated that the inter-domain spacing

200

of ionic clusters increases upon impregnation denoting presence of the incorporated

supramolecules within the Nafion membrane. The swelling behavior of both of the

impregnated membranes was found to be profoundly suppressed as the supramolecules

accommodated the ionic domains. Noria impregnated membrane marked an increment in

the IEC value which may primarily arise from the hydroxyl groups of Noria and protons

of sulfonic acids in Nafion. On the other hand, Noria-BOC impregnated membrane

showed little to no change in IEC value as the encapped tert-butyloxycarbonyl did not

contribute any additional protons within the ionic domains of Nafion. It was observed

that only Noria impregnated membranes demonstrated an increasing trend in the proton

conductivity at elevated temperature with minimal hydration. Proton conductivity of

Noria-BOC impregnated membrane, on the other hand, was found to plummet at the

same elevated temperature range. It is highly anticipated that the proton conduction of

Noria impregnated membrane will rely on the ionized proton of the polar hydroxyl

groups upon exceeding the evaporation temperature of free and bound water. Thus Noria

supramolecules are potential to be used as the solid electrolyte for the elevated operating temperature of fuel cell applications.

8.2. Recommendations

It is known that the conductivity is highly dependent on the orientation or percolation pathways of the ionic domains/clusters within the electrolyte. In Chapter IV we demonstrated that swelling of Nafion and PVDF-TrFE blend was greatly suppressed

but the proton conductivity of the blends was brought down possibly due to the

201

comingled and torturous pathways formed within the membrane. Previously, our group

has reported the accomplishment on the formation of holographic channels/hierarchical structures of photocurable poly(ethylene oxide) PEO with pendant diacrylate through the two- or four-wave photonics interference. These holographic arrays of channels might

enhance the transport properties of the blend membrane. Substitution of copolymer

PVDF-TrFE to photocurable hydrophobic polymers in the Nafion blends are sought to

diminish the effect of torturous percolated pathways.

202

BIBLIOGRAPHY

1. Hoogers G., in Fuel Cell Technology Handbook, CRC Press, 2002, Boca Raton, FL.

2. Sorensen B., in Hydrogen and Fuel Cells: Emerging Technologies and Applications, 2005, Elsevier Academic Press, London, UK.

3. O’Hayre R.P., Cha S.-W., Colella W., and Prinz F.B., in Fuel Cell: Fundamentals, 2005, John Wiley & Sons, New York, N.Y.

4. Barbir F., in PEM Fuel Cells: Theory and Practice, 2005, Elsevier Academic Press, Burlington, MA.

5. W. G. F. Grot, Chem. Eng. Technol. 1972, 44, 167-170.

6. W. G. F. Grot, to E.I Du Pont de Nemours and Company, Apr. 1976, Electrolysis Cell using Cation Exchange Membranes of Improved Perm-selectivity, US Patent 4, 026, 783.

7. W. G. F. Grot, Chem. Ind., 1985, 9, 647–649.

8. W. G. F. Grot, Macromol. Symp. 1994, 8, 161-167.

9. http://www1.eere.energy.gov/hydrogenandfuelcells/pdfs/htwg_rd_plan.pdf

10. K. D. Kreuer, J. Membr. Sci. 2001, 185, 29-39.

11. S. Swier, M. T. Shaw, R. A. Weiss, J. Membr. Sci., 2006, 270, 22-31.

12. Y. S. Kim, F. Wang, M. Hickner, T. A. Zawodzinski, J. E. McGrath, J. Membr. Sci., 2003, 212, 263-282.

13. M. Lavorgna, G. Mesitieri, G. Scherillo, M. T. Shaw, S. Swier, R. A. Weiss, J. Polym. Sci.: Part B: Polym. Phys., 2007, 45, 395 – 404.

14. B. Kim, B. Jung, Macromol. Rapid Commun., 2004, 25, 1263 – 1267.

15. M. Doyle, S. K. Choi, G. Proulx, J. Electrochem. Soc., 2000, 147, 34-37.

203

16. C. Schmidt, T. Glück, G. Schmidt-Naake, Chem. Eng. Technol. 2008, 31, 13-22.

17. M. K. Mistry, S. Subianto, N. R. Choudhury, N. K. Dutta, Langmuir, 2009, 25, 9240-9251.

18. A. G. Kannan, N. R. Choudhury, N. K. Dutta, J. Membr. Sci., 2009, 333, 50-58.

19. S. Malhotra, R. Datta, J. Electrochem. Soc., 1997, 144, 2, 23 - 26.

20. T. Kyu, J. C. Yang, Macromolecules, 1990, 23, 176 – 182.

21. J. C. Yang, T. Kyu, Macromolecules, 1990, 23, 182 – 186.

22. F. A. Landis, R. B. Moore, Macromolecules, 2000, 33, 6031 – 6041.

23. R. L. Busby, Hydrogen and Fuel Cells, PennWell Corp., 2005, Tulsa, OK.

24. C. S. Spiegel, Designing & Building Fuel Cells, 2007, 1st Ed., McGraw – Hill, New York, NY.

25. L. Carrette, K. A. Friedrich, U. Stimming, Chem. Phys. Chem., 2000, 1, 162-193.

26. M. Corbett, Opportunities in Advanced Fuel Cell Technologies: Stationary Power Generation 1998-2008, 1998, Kline & Company, Inc., Fairfield, NJ.

27. S. Gottesfeld, T. A. Zawodzinski, Polymer Electrolyte Fuel Cell, (Eds. R. C. Alkire), Advances in Electrochemical Science and Engineering, Wiley – VCH, 1998, New York, NY.

28. K. Kordesch, G. Simander, Fuel Cells and Their Applications, VCH Publishers, 1996, New York, N.Y.

29. http://americanhistory.si.edu/fuelcells/basics.htm

30. www.fueleconomy.gov

31. J. Larminie, A. Dicks, Fuel Cell Systems Explained, John Wiley and Sons, 2000, New York, NY.

32. D. Bevers, N. Wagner, M. VonBradke, Int. J. Hydrogen Energy, 1998, 23, 57 – 63.

33. L. Giorgi, E. Antolini, A. Pozio, E. Passalacqua, Electrochimica Acta, 1998, 43, 3675 – 3680.

34. T. R. Ralph, G. A. Hards, J. E. Keating, S. A. Campbell, D. P. Wilkinson, H. Davis, J. St. Pierre, M. C. Johnson, J. Electrochem. Soc., 1997, 11, 3845 – 3857. 204

35. Eisenberg A., King M., Ion-Containing Polymers: Physical Properties and Structure: Vol.2, Academic Press, 1977, New York, NY.

36. T. Kyu, Structure and Properties of Perfluorinated Ion – Exchange Membranes, (ed. D.R. Lloyd), in Materials Science of Synthetic Membranes, ACS Symposium Series 269, American Chemical Society, 1985,Washington DC and references therein.

37. K. A. Mauritz, in Morphological Theories, (Eds. M. R. Tant, K. A. Mauritz, G. L. Wilkes), in Ionomers: Synthesis, Structure, Properties and Applications, Blackie Academic & Professional, 1997, London, UK.

38. T. Sata, in Ion Exchange Membranes: Preparation, Characterization, Modification and Application, The Royal Society of Chemistry, 2004, Dorset, UK.

39. W. M. Risen Jr., In Chemistry in Ionomers, (Eds. M. Pineri, A. Eisenberg), Structure and Properties of Ionomers, NATO ASI Series, Kluwer Academic Publishers, 1986, Norwell, MA.

40. M. R Tant, K. P. Darst, K. D. Lee, C. W. Martin, in Multiphase Polymers: Blends and Ionomers, (Eds. L. A. Ultracki, R. A. Weiss), ACS Symposium Series No: 395, American Chemical Society, Washington DC, 1989, chapter 15, pg 370.

41. T. D. Gierke, G. E. Munn, F. C. Wilson, J. Polym. Sci.: Polym. Phys., 1981, 19, 1687 – 1704.

42. T. D. Gierke, W. Y. Hsu, The Cluster–Network Model of Ion Clustering in Perflurosulfonated Membranes, In Perflourinated Ionomer Membranes, (Eds. A. Eisenberg, H. L. Yeager), ACS Symposium Series No: 180, American Chemical Society, Washington DC, 1982, Ch. 13, 283 – 307.

43. H. L. Yeager, In Transport Properties of Pefluorinated Ionomer Membranes, (Eds. A. Eisenberg, H. L. Yeager), Perfluorinated Ionomer Membranes, ACS Symposium Series, 180, ACS, Washington DC., 1982.

44. E. J. Roche, M. Pineri, R. Duplessix, A. M. Levelut, J. Polym. Sci.; Polym. Phys. Ed., 1981, 19, 1 – 11.

45. E. J. Roche, M. Pineri, R. Duplessix, J. Polym. Sci.; Polym. Phys. Ed., 1982, 20, 107 – 116.

46. S. Kumar, M. Pineri, J. Mol. Sci., Part B; Polym. Phys., 1986, 24, 1767 – 1782.

47. M. Fujimura, R. Hashimoto, H. Kawai, Macromolecules, 1982, 15, 136 - 144.

205

48. T. Xue, J.S. Trent, K. Osseo–Assare, J. Membr. Sci., 1989, 45, 261 - 271.

49. H. W. Starkweather Jr., Macromolecules, 1982, 15, 320 – 323.

50. M. H. Litt, Polym. Prepr., (Am. Chem. Soc. Div. Polym. Chem), 1997, 38, 80–81.

51. G. Gebel, J. Lambard, Macromolecules, 1997, 30, 7914 – 7920.

52. G. Gebel, R. B. Moore, Macromolecules, 2000, 33, 13, 4850 – 4855.

53. J. A. Elliot, S. Hanna, A.M.S. Elliot, G. E. Cooley, Macromolecules, 2000, 33, 4161 - 4171.

54. T. A. Zawodzinski, T. E. Springer, J. Davey, R. Jestel, C. Lopez, J. Valerio, S. Gottesfeld, J. Electrochem. Soc., 1993, 7, 1981 – 1985.

55. F. P. Orfino, S. Holdcroft, New Mater. Electrochem. Syst., 2000, 3, 287 – 293.

56. D. J. Yarusso, S. L. Cooper, Macromolecules, 1983, 16, 1871 - 1880.

57. D. J. Yarusso, S. L. Cooper, Polymer, 1985, 26, 371 – 381.

58. R. S. McLean, M. Doyle, B. B. Sauer, Macromolecules, 2000, 33, 6541 – 6550.

59. Y. Wang, Y. Kawano, S. R. Aubuchon, R. A. Palmer, Macromolecules, 2003, 36, 1138 – 1146.

60. T. A. Zawodzinski, C. Derouin, S. Radzinski, R. J. Sherman, V. T. Smith, T. E. Springer, S. Gottesfeld, J. Electrochem Soc., 1993, 4, 1041 – 1047.

61. M. Ludvigsson, J. Lindgren, J. Tegenfeldt, Electrochimica Acta, 2000, 45, 2267 – 2271.

62. J. T. Hinatsu, M. Mizuhatta, H. Takenata, J. Electrochem. Soc., 1994, 6, 1493 - 1498.

63. G. Alberti, R. Naducci, M. Sganappa, J. Power Sources, 2007, 178, 575 – 583.

64. T. A. Zawodzinski, M. Neeman, L. O. Sillerud, S. Gottesfeld, J. Phys. Chem., 1991, 95, 6040 - 6044.

65. N. J. Bunce, S. J. Sondheimer, C. A. Fyfe, Macromolecules 1986, 19, 333-339.

66. P. Batamack, J. Fraissard, Catalysis Letters 1995, 35, 135-142.

206

67. R. S Yeo, Perflourinated Ionomer Membranes, (Eds. H. L. Yeager, A. Eisenberg), ACS Symposium Series No: 180, American Chemical Society, 1982, Chapter 5, Washington DC.

68. A. V. Anantaraman, C. L. Gardner, J. Electroanal. Chem. 1996, 414, 115-120.

69. S. J. Paddison, R. Paul, Phys. Chem. Chem. Phys., 2002, 4, 1158 – 1163.

70. S. J. Paddison, Annu. Rev. Mater. Res. 2003, 33, 289-319.

71. Y. L. Chen, T. C. Chou, Electrochem. Acta, 1993, 38, 2171 - 2175.

72. K. D. Kreuer, T. Dippel, W. Meyer, J. Maier, Mater. Res. Soc. Symp. Proc., 1993, 293, 273 - 282.

73. K. D. Kreuer, W, Weppner, A. Rabenau, Angew. Chem. Int. Ed. Eng., 1982, 21, 208-211.

74. C. J. D. van Grotthuss, Ann. Chem., 1806, 58, 54-65.

75. K. D. Kreuer, Chem. Mater., 1996, 8, 610-641.

76. N. Jalani., Development of Nanocomposite Polymer Electrolytes Membrane for Higher Temperature PEM Fuel Cells, PhD. Thesis, Worcester Polytechnic Institute, 2006.

77. M. Saito, K. Hayamizu, T. Okada., J. Phys. Chem B, 2005, 109, 3112-3119.

78. R. S. Yeo, A. Eisenberg, J. Appl. Sci., 1977, 21, 875 - 883.

79. T. Kyu, A. Eisenberg, In Mechanical Relaxations in Perfluorosulfonate Ionomer Membranes, Perflourinated Ionomer Membranes, (Eds. A. Eisenberg, H. L. Yeager), ACS Symposium Series No: 180, American Chemical Society, 1982, Washington DC.

80. T. Kyu, A. Eisenberg, J. Polym. Sci.: Polym. Symp., 1984, 71, 203-219.

81. T. Kyu, M. Hashiyama, A. Eisenberg, Can. J. Chem., 1983, 61, 680 – 687.

82. K. M. Cable, Tailoring Morphology–Property Relationships in Perfluorosulfonate Ionomers, PhD. Dissertation, University of Southern Mississippi, 1996.

83. R. B. Moore, K. M. Cable, Polym. Prepr. (Am. Chem. Soc. Div. Polym. Chem.), 1997, 38, 272 – 273.

84. R. B. Moore, K. M. Cable, T. L. Croley, J. Membr. Sci., 1992, 75, 7 -14.

207

85. D. L. Feldheim, D. R. Lawson, C. R. Martin, J. Polym. Sci.: Part B: Polymer Physics, 1993, 31, 953 – 957.

86. S.H. de Almeida, Y. Kawano, J. Thermal Anal. Calorimetry, 1999, 58, 569 – 577.

87. R. B. Moore, C. R. Martin, Anal. Chem., 1986, 58, 2569 - 2570.

88. R.B. Moore, C. R. Martin, Macromolecules, 1988, 21, 1334 – 1339.

89. A. K. Phillips, R. B. Moore, J. Polym. Sci. Polym. Phys., 2006, 44, 2267 – 2277.

90. K.-Y. Cho, H.-Y. Jung, N.-S. Choi, S.-J. Sung, J.-K. Park, J.-H. Choi, Y.-E. Sung, Solid State Ionics, 2005, 176, 3027 – 3030.

91. M. A. Smit, A. L. Ocampo, M. A. Espinosa-Medina, P. J. Sebastian, J. Power Sources, 2003, 124, 59-64.

92. A. Sungpet, J. Memb. Sci., 2003, 226, 131-134.

93. . C. W. Jr. Walker, J. Electrochem. Soc., 2004, 151, 1797 – 1803.

94. B. Bae, H. Y. Ha, D. Kim, J. Electrochem. Soc. 2005, 152, 1366-1372.

95. T. Shiga, Y. Hirose, A. Okada, T. Kurauchi, J. Appl. Polym. Sci., 1993, 47, 113 – 119.

96. M. Shahinpoor, Y. Bar–Cohen, J. O. Simpson, J. Smith, Smart Mater. Struct., 1998, 7, 15–30.

97. M. Shahinpoor, K. J. Kim, M. Mojarrad, Artificial Muscles: Applications of Advanced Polymeric Nanocomposites, Taylor and Francis, 2007, Boca Raton, FL.

98. Y. Bar–Cohen, Electroactive Polymer Actuators as Artificial Muscles, 2004, 2nd Ed., SPIE Press, Washington, DC.

99. M. D. Bennet, D. J. Leo, Sensors and Actuators, 2004, 115, 79 – 90.

100. K. J. Kim, M. Shahinpoor, Smart Mater. Struct., 2005, 14, 197 – 214.

101. K. J. Kim, M. Shahinpoor, Polymer, 2002, 43, 797–802.

102. C.-J. Yuan, C.-L. Hsu, S.-C. Wang, K.-S. Chang, Electroanal., 2005, 17, 2239– 2245.

103. W. W. Doll, J. B. Lando, J. of Macromol. Sci., Part B, 1968, 2, 205 – 218.

208

104. J. C. Salamone, in Polymeric Materials Encyclopedia, CRC Press, 1996, New York, N.Y, and references therein.

105. M. Bachmann, W. L. Gordon, S. Weinhold, J. B. Lando, J. Appl. Phys., 1980, 51, 5095–5100.

106. A. J. Lovinger, Macromolecules, 1982, 15, 1, 40 – 44.

107. M. Kobayashi, K. Tashiro, H. Tadokoro, Macromolecules, 1975, 8, 158 – 171.

108. R. R. Kolda, J. B. Lando, J. Macromol. Sci., Part B, 1975, 11, 21 – 39.

109. A. J. Lovinger, G. T. Davis, T. Furukawa, M. G. Broadhurst, Macromolecules, 1982, 15, 323–328.

110. Q. M. Zhang, V. Bharti, X. Zhao, Science, 1998, 280, 5372, 2101 – 2104.

111. B. J. Chu, Science, 2006, 313, 5785, 334 – 336.

112. T. Furukawa, M. Date, E. Fukada, Y. Tajitsu, A. Chiba, Jpn. J. Appl. Phys., 1980, 19, 109 – 112.

113. T. Yamada, T. Ueda, T. Kitayama, J. Appl. Phys., 1981, 52, 948 – 956.

114. Y. Higashihata, J. Sako, T. Yagi, Ferroelectrics, 1981, 32, 85 – 97.

115. T. Furukawa, G. E. Johnson, H. E. Bair, Ferroelectrics, 1981, 32, 61 – 73.

116. K. J. Kim, G.B. Kim, J. Appl. Polym. Sci., 1993, 47, 1781 - 1789.

117. K. J. Kim, N. M. Reynolds, S. L. Hsu, J. Polym Sci. Polym. Phys. Ed., 1993, 31, 1555 - 1566.

118. K. J. Kim, G. B. Kim, C. L. Vanlencia, J. F. Rabolt, J. Polym. Sci.: Polym. Phys. Ed., 1994, 32, 2435 - 2444.

119. K. J. Kim, G.B., Kim Polymer, 1997, 38, 4881.

120. K. J. Kim, T. Kyu, Polymer, 1999, 40, 6125 – 6134.

121. L. O. Faria, R. L. Moreira, Polymer, 1999, 40, 4465 – 4471.

122. L. O. Faria, R. L. Moreira, J. Polym. Sci.:Part B: Polym. Phys., 2000, 38, 34 – 40.

123. Y. Tang, J. Scheinbeim, J. Polym. Sci.:Part B: Polym. Phys., 2003, 41, 927 – 935. 209

124. B. Ploss, F. G. Shin, H. L. W. Chan, C. L. Choy, Appl. Phys. Lett., 2000, 76, 19, 2776 – 2778.

125. S. Yokoyama, A. Otomo, T. Nakahama, Y. Okuno, S. Mashiko, Top. Curr. Chem., 2003, 228, 205-226.

126. S. Yokoyama, T. Nakahama, A. Otomo, S. Mashiko, J. Am. Chem. Soc., 2000, 122, 3174-3181.

127. H. Ma, K.-Y. Jen, Adv. Mater., 2001, 13, 15, 1201-1205.

128. A. Adronov, J. M. J. Fréchet, G. S. He, K. S. Kim, S.J. Chung, J. Swiatkiewicz, P.N. Prasad, Chem. Mater., 2000, 12, 2838-2841.

129. C. Decker, Prog. Poly. Sci., 1996, 21, 593 – 650.

130. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, 1953, Ithaca, NY.

131. C. Tanford, Physical Chemistry of Macromolecules, Wiley, 1961, New York, NY.

132. M. L. Huggins, J. Am. Chem. Soc., 1942, 64, 1712 – 1719.

133. L.A. Kleintjens, M.H. Onclin, R. Koningsveld, EFCE Pubs, Series 11, 1980, 521 – 526.

134. R. A. Matkar, T. Kyu, J. Phys. Chem. B., 2006, 110, 12728 – 12732.

135. H. Xu, R.A. Matkar, T. Kyu, Phys. Rev. E., 2005, 72, 011804 – 011813.

136. F. F. Abraham, Homogeneous Nucleation Theory, 1974, Academic Press, New York, NY.

137. J. Ross Macdonald, Impedance Spectroscopy: Theory, Experiment and Application, Wiley-Interscience, New Jersey, 2005.

138. X. Z. Yuan, C. Song, H. Wang, J. Zhang, in Electrochemical Impedance Spectroscopy in PEM Fuel Cells, Springer, London, 2010, pp. 39-93.

139. A. Steck, H. L. Yeager, Anal. Chem., 1980, 52, 1215 – 1218.

140. K. Maruyama, H. Kudo, T. Ikehara, N. Ito, T. Nishikubo, J. Polym. Sci. Part A: Polym. Chem. 2005, 43, 4642-4653.

141. H. Kudo, R. Hayashi, K. Mitani, T. Yokozawa, N. C. Kasuga, T. Nishikubo, Angew. Chem. Int. Ed. 2006, 45, 7948-7952. 210

142. T. A. Zawodzinski, T. E. Springer, J. Davey, R. Jestel, C. Lopez, J. Valerio, S. Gottesfeld, J. Electrochem. Soc. 1993, 140, 1981-1985.

143. T. Kyu, S. Meng, H. Duran, K. Nandjundiah, G. R. Yandek, Macromol. Res., 2006, 14, 155-165.

144. B. Bahar, A. R. Hobson, J. A. Kolde, to W.L Gore and Associates Inc., U.S Patent 5635041, June 3rd, 1997.

145. S. -F. Liu, K. Schmidt-Rohr, Macromolecules, 2001, 34, 8416-8418.

146. Q. Chen, K. Schmidt-Rohr, Macromolecules, 2004, 37, 5995-6003.

147. R. M. Silverstein, F. X. Webster, D. J. Kiemle, in Spectrometric Identification of Organic Compounds, ed. S.Wolfman-Robichaud, John Wiley & Sons Inc., Hoboken, NJ, Seventh edition 2005, Ch. 2, pp. 72–126.

148. M. Falk, Mechanical Relaxations in Perfluorosulfonate Ionomer Membranes, Perflourinated Ionomer Membranes, (Eds. A. Eisenberg, H. L. Yeager), ACS Symposium Series No: 180, American Chemical Society, 1982, Washington DC.

149. M. Falk, Can. J. Chem. 1980, 58, 1495-1501.

150. E. Endoh, S. Terazono, H. Widjaja, Y. Takimoto, Electrochem. Solid State Lett., 2004, 7, A209-A211.

151. J. Surowiec, R. Bogoczek, J. Thermal Anal. 1988, 33, 1097-1102.

152. L. G. Lage, P. G. Delgado, Y. Kawano, J. Thermal Anal. 2004, 75, 521-530.

153. M. Matsuyama, E. Kokufuta, T. Kusumi, K. Harada, Macromolecules. 1980, 13, 198-200.

154. R. Tannenbaum, M. Rajagopalan, A. Eisenberg, J. Polym. Sci.: Part B., Polym. Phys., 2003, 41, 1814-1823.

155. N. Agmon, Chem. Phys. Lett. 1995, 244, 456-462.

156. K. D. Kreuer, Solid State Ionics 1997, 94, 55-62.

157. M. Tanaka, A. Rastogi, H. Kudo, D. Watanabe, T. Nishikubo, C. K. Ober, J. Mater. Chem., 2009, 19, 4622-4626.

211