UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Electrospun nanofibrous membranes for water vapour transport applications Huizing, Ryan Nicholas 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_may_huizing_ryan.pdf [ 11.35MB ]
Metadata
JSON: 24-1.0344019.json
JSON-LD: 24-1.0344019-ld.json
RDF/XML (Pretty): 24-1.0344019-rdf.xml
RDF/JSON: 24-1.0344019-rdf.json
Turtle: 24-1.0344019-turtle.txt
N-Triples: 24-1.0344019-rdf-ntriples.txt
Original Record: 24-1.0344019-source.json
Full Text
24-1.0344019-fulltext.txt
Citation
24-1.0344019.ris

Full Text

  ELECTROSPUN NANOFIBROUS MEMBRANES FOR WATER VAPOUR TRANSPORT APPLICATIONS by Ryan Nicholas Huizing B.A.Sc, Engineering Chemistry, Queen’s University, 2005 M.A.Sc., Chemical Engineering, University of Waterloo, 2007 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Materials Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2017 © Ryan Nicholas Huizing, 2017   ii  Abstract This study discusses the development of selective water vapour permeable membranes for the separation of water vapour from air. These membranes are useful as separation media for membrane based energy recovery ventilation devices. Current generation composite membranes for these devices consist of a polymer film layer which is permeable to water vapour but selective for water vapour over gases. In these composite membranes, this film layer is attached to one surface of a microporous polymeric support substrate. However, as it is demonstrated in this work, the microporous support contributes 30 to 50% of the resistance to water vapour transport in the composite membranes. In an attempt to eliminate this microporous substrate and its associated resistance, a membrane was developed using an electrospun nanofibrous layer as a support structure for the selective coating layer. In these membranes, the electrospun nanofibers are deposited on a non-woven micro-fibrous carrier and then the nanofibers are impregnated with a selectively permeable polymer to make impregnated electrospun nanofibrous membranes (IENM). The nanofibers are found to contribute a resistance to vapour transport in these membranes and their effect on water vapour permeability is quantified through a fiber-filled-film model. An optimization study of the IENMs demonstrated that fiber diameter and fiber volume in the film layer effect the water vapour permeance of these membranes and reducing these variables improved water vapour permeance. IEMNs were demonstrated to have water vapour permeance (>10000 GPU) while still having sufficient selectivity for the application, exceeding the performance of current generation composite membranes. The membranes were demonstrated to be particularly useful in the fabrication of ‘formable membranes’ which could be thermally-formed into exchanger plates for energy recovery ventilator exchangers. Thermal and mechanical properties of the membrane components iii  were reported individually and as complete membranes. A membrane composition was demonstrated to fabricate exchanger plates from formable IENMs. This work contributes to the development of membranes for air-to-air exchangers specifically for energy recovery ventilation. A new class of membrane based on electrospun nanofibers for these devices was successfully demonstrated to have improved performance properties and a novel formable membrane was developed.   iv  Preface This dissertation is an original intellectual product of the author, Ryan Huizing.  Some contents of Chapters 3 through 7 have been published in patents US 9255744, US 8936668, patent application CA2016050610, and other international patent applications within these patent families. Various results from this dissertation have been presented as oral presentations at ICOM 2011, NAMS 2012, NAMS 2015, and EUROMEMBRANE 2015. The water vapour transport testing apparatus used throughout the dissertation was designed and built at dPoint Technologies by myself and a coop student (Tom Curran) under my supervision in 2010. DVS data was collected on samples prepared by me by an external laboratory. I completed all other sample preparation, experimentation, and analysis in Chapter 3. Some membrane fabrication and water vapour transport data Chapter 4 was collected under my supervision by Frankie Wong at dPoint Technologies. I completed all analysis and interpretation of data in Chapter 4, and also completed the experimental work. A version of Chapter 5 has been published as R. Huizing, W. Mérida, and F. Ko, “Impregnated electrospun nanofibrous membranes for water vapour transport applications,” Journal of Membrane Science, vol. 461, pp. 146–160, Jul. 2014. I completed all of the experimental and analytical work, Dr. Mérida and Dr. Ko provided direction and feedback on the work. A coop student under my supervision, Scott Dornian fabricated and tested a portion of the membranes in chapters 6 and 7. I completed a portion of the experimental work and all analysis and interpretation in these chapters. Addie Bahi at UBC completed DMA testing in Chapter 7. Taesik Chae at UBC fabricated membrane films for DMA testing in Chapter 7 at dPoint Technologies based on methods and techniques developed by me. The forming press used in Chapter 7 was designed and built to my specifications by David Kadylak and Chris Barr at dPoint Technologies. This project was sponsored by dPoint Technologies through a NSERC-CRD grant.   v  Table of Contents Abstract .......................................................................................................................................................... ii Preface ........................................................................................................................................................... iv Table of Contents ............................................................................................................................................v List of Tables ................................................................................................................................................... xi List of Figures ................................................................................................................................................ xiv List of Abbreviations ..................................................................................................................................... xxi List of Symbols ............................................................................................................................................ xxiii Acknowledgements ..................................................................................................................................... xxv Dedication .................................................................................................................................................. xxvi  Introduction and Background ................................................................................................................ 1 1.1 Membrane Science ................................................................................................................................... 1 1.2 Energy Recovery in Buildings ................................................................................................................... 2 1.3 Membranes Requirements for ERV .......................................................................................................... 5 1.4 Composite Membranes ............................................................................................................................ 7 1.5 Nanofibrous materials ............................................................................................................................. 7 1.6 Literature Review ..................................................................................................................................... 8 1.6.1 Plate and Frame ERVs ..................................................................................................................... 8 1.6.1.1 Building Energy Use Analysis and System Design ................................................................. 9 1.6.1.2 Exchanger Design ................................................................................................................ 10 1.6.1.3 Heat and Mass Transport Modeling ................................................................................... 10 1.6.1.4 Membrane Design .............................................................................................................. 11 1.6.2 Water Vapour Transport in Membranes ...................................................................................... 12 1.6.3 Water Vapour Transport in Nanofibrous Materials ..................................................................... 15 1.6.4 Nanofibrous Membranes ............................................................................................................. 16 1.7 Literature of Particular Relevance and Contributions of the Present Work ........................................... 17 vi  1.7.1 Knowledge Gaps Addressed in this Dissertation .......................................................................... 18 1.8 Research Objectives ............................................................................................................................... 19  Experimental ....................................................................................................................................... 22 2.1 Chemicals and Materials ........................................................................................................................ 22 2.2 Polymer Coating Formulations ............................................................................................................... 22 2.3 Fabrication of Nanofibrous Substrates .................................................................................................. 22 2.4 Fabrication of Selective Nanofibrous Membranes ................................................................................. 23 2.5 Polymer Film Fabrication ....................................................................................................................... 23 2.6 Composite Membrane Fabrication ........................................................................................................ 23 2.7 Basis Weight, Density, and Thickness Measurements ........................................................................... 23 2.8 Nanofiber Basis Weight Measurements ................................................................................................ 24 2.9 Swell Ratio.............................................................................................................................................. 24 2.10 Optical and Polarized Light Microscopy ............................................................................................ 24 2.11 Electron Microscopy .......................................................................................................................... 24 2.12 Laser Confocal Microscopy ................................................................................................................ 25 2.13 Mercury Intrusion Porosimetry ......................................................................................................... 25 2.14 Contact Angle Goniometry ................................................................................................................ 25 2.15 Oxygen Crossover and Permeance .................................................................................................... 25 2.16 Pressurized Air Crossover .................................................................................................................. 26 2.17 Gas Permeability ............................................................................................................................... 27 2.18 Water Vapour Flux, Permeance, and Permeability ........................................................................... 27 2.19 Permeance and Permeability Units and Selectivity ........................................................................... 29 2.20 Water Vapour Sorption (DVS) ........................................................................................................... 29 2.21 Fourier Transform Infrared Spectroscopy (FTIR) ............................................................................... 30 2.22 Nuclear Magnetic Resonance (NMR) ................................................................................................ 30 2.23 Gel Permeation Chromatography (GPC) ........................................................................................... 30 vii  2.24 Differential Scanning Calorimetry (DSC)............................................................................................ 30 2.25 Thermogravimetric Analysis (TGA) .................................................................................................... 31 2.26 Dynamic Mechanical Analysis (DMA) ................................................................................................ 31 2.27 Mechanical Analysis .......................................................................................................................... 31  Polyethylene Oxide-Polyurethane Copolymers for Water Vapour Transport Membranes ................... 32 3.1 Introduction ........................................................................................................................................... 32 3.2 PEO-PU Polymers ................................................................................................................................... 33 3.3 NMR ....................................................................................................................................................... 37 3.4 Molecular Weight of the PEO-PU Polymer ............................................................................................. 40 3.5 Crosslinking of PERMAX™ 230 ............................................................................................................... 41 3.5.1 FTIR of cross-linked films .............................................................................................................. 42 3.6 DSC and TGA Thermal Analysis .............................................................................................................. 46 3.7 Liquid Water Interactions With PEO-PU Polymers ................................................................................. 48 3.7.1 Water Uptake of Carbodiimide Cross-linked PEO-PU .................................................................. 48 3.7.2 Hydrated DSC ............................................................................................................................... 53 3.8 Water Vapour Interactions with PEO-PU films ...................................................................................... 56 3.8.1 Kinetics of Sorption in PEO-PU Films ............................................................................................ 57 3.8.2 Sorption and Desorption Curves .................................................................................................. 59 3.8.3 Solubility Coefficients ................................................................................................................... 61 3.8.4 Flory-Huggins Interaction Parameter ........................................................................................... 62 3.9 Zimm-Lundberg Clustering Analysis ....................................................................................................... 64 3.10 Permeability ...................................................................................................................................... 67 3.11 Conclusions ........................................................................................................................................ 70  Composite Membranes for Water Vapour Transport Applications ...................................................... 72 4.1 Introduction ........................................................................................................................................... 72 4.2 Water Vapour Transport Testing ........................................................................................................... 73 4.3 Resistance in Series Model ..................................................................................................................... 76 4.4 Boundary Layer Resistance .................................................................................................................... 79 viii  4.5 Substrates Structure and Performance .................................................................................................. 86 4.6 Composite Membranes .......................................................................................................................... 89 4.6.1 Silica-Polyethylene Substrate Based Membranes ........................................................................ 90 4.6.2 Polypropylene Substrate Based Membranes ............................................................................... 94 4.7 Effect of PEO Crystallization ................................................................................................................... 97 4.8 Elimination of the Thermal Switching Effect ........................................................................................ 105 4.9 Conclusions .......................................................................................................................................... 107  Impregnated Electrospun Nanofibrous Membranes ...........................................................................109 5.1 Introduction ......................................................................................................................................... 109 5.2 Nanofibrous Materials ......................................................................................................................... 109 5.3 Concept of the Impregnated Electrospun Nanofibrous Membrane for Water Vapour Transport ....... 110 5.4 Electro-spinning of Nanofibrous Layers ............................................................................................... 113 5.5 Membrane Fabrication ........................................................................................................................ 114 5.6 Membrane Structure ............................................................................................................................ 115 5.7 Permeation Testing of Nanofibrous Membranes ................................................................................. 117 5.7.1 Pressurized Air Crossover and Oxygen Permeance .................................................................... 117 5.7.2 Hydrophobicity ........................................................................................................................... 119 5.7.3 Water Vapour Transport ............................................................................................................ 119 5.7.4 Permeability and Selectivity in IENM ......................................................................................... 123 5.7.4.1 Group I: MEM-60-11 ......................................................................................................... 124 5.7.4.2 Group II: MEM-60-15 and MEM-90-11 ............................................................................. 124 5.7.4.3 Group III: MEM-60-21 and MEM-120-11 .......................................................................... 124 5.7.4.4 Group IV: MEM-120-21 and MEM-90-21 .......................................................................... 125 5.7.4.5 Group V: MEM-120-15 and MEM-90-15 ........................................................................... 125 5.7.5 Fiber Ratio in the Dense Layer.................................................................................................... 125 5.7.6 Water Vapour Activity ................................................................................................................ 127 5.8 Conclusions .......................................................................................................................................... 130  Optimization of Impregnated Electrospun Nanofibrous Membranes ..................................................132 6.1 Introduction ......................................................................................................................................... 132 ix  6.2 Solutions and Electrospinning .............................................................................................................. 134 6.3 Membrane Optimization: Fiber Diameter and Deposition ................................................................... 136 6.4 Fiber Volume Fraction .......................................................................................................................... 141 6.5 Fiber Filled Film Membrane Permeability Model ................................................................................. 148 6.6 Defects in the Nanofibrous Film Layer ................................................................................................. 151 6.7 Gravure Coating of Nanofibrous Membranes ...................................................................................... 155 6.7.1 Gravure Coating Method ............................................................................................................ 155 6.7.2 Gravure Coating Results ............................................................................................................. 157 6.8 Coating Formulation Parameters and Wetting .................................................................................... 160 6.9 Conclusions .......................................................................................................................................... 163  Formable Impregnated Electrospun Nanofibrous Membranes ...........................................................164 7.1 Introduction ......................................................................................................................................... 164 7.2 Carrier Layer ......................................................................................................................................... 167 7.2.1 Non-woven Material Properties ................................................................................................. 167 7.2.2 Thermal Analysis of Carrier Layer ............................................................................................... 168 7.2.3 Thermo-mechanical Testing of Carrier Layers ............................................................................ 169 7.3 Nanofiber + PEO-PU Film Layers .......................................................................................................... 172 7.3.1 Thermal Analysis of Fiber, Film, and Fiber-filled Film Layers ..................................................... 172 7.3.2 Thermogravimetric Analysis of Fiber-Film Layer ........................................................................ 174 7.3.3 Thermo-mechanical Testing of Film Layers ................................................................................ 176 7.4 Defects During Mechanical Deformation of Membranes .................................................................... 178 7.5 Forming of Membranes ....................................................................................................................... 183 7.6 Transport Properties of Membranes .................................................................................................... 184 7.7 Conclusions .......................................................................................................................................... 186  Conclusions and Recommendations for Future Work .........................................................................188 8.1 Overview .............................................................................................................................................. 188 8.2 Conclusions, contributions, and Future Work ...................................................................................... 189 8.2.1 Characterization of PEO-PU Polymers ........................................................................................ 189 8.2.2 Analysis of Composite Membranes ............................................................................................ 190 x  8.2.3 Impregnated Electrospun Nanofibrous Membranes: Fabrication and Optimization ................. 191 8.2.4 Formable Electrospun Nanofibrous Membranes ....................................................................... 193 References ...................................................................................................................................................196    xi  List of Tables Table 1-1: Literature of particular relevance to this dissertation. ................................................................. 17 Table 3-1: Properties of commercially available PERMAX™ dispersions from Lubrizol Advanced Materials. 37 Table 3-2: Assignment of protons in the 1H NMR of the PEO-PU polymer. .................................................... 40 Table 3-3: Molecular weight summary for two samples of PEO-PU by GPC. ................................................. 41 Table 3-4: Water uptake and density of PEO-PU polymers at various cross-linking ratios. ............................ 49 Table 3-5: Swell ratios for polymer films with increasing temperatures. ...................................................... 51 Table 3-6: Water uptake as a function of total polyethylene oxide groups in the polymer. .......................... 52 Table 3-7: Calculated bound and free water in PEO-PU polymers with different cross-linker content. ......... 54 Table 3-8: Summary of mass uptake and kinetic data for each step .............................................................. 58 Table 3-9: Kinetic parameters for sorption of water vapour for PEO-PU polymer at different levels of crosslinking at 25oC.............................................................................................................................. 58 Table 3-10: Summary of diffusivity and solubility coefficients and permeability in PEO-PU polymers at various temperature and water vapour activity. ................................................................................. 69 Table 4-1: Geometric properties of permeation module ............................................................................... 79 Table 4-2: Coefficients for Sherwood number correlations for a water vapour transport module at 50°C. ... 83 Table 4-3: Properties of substrates used to make composite membranes, errors reported are sample standard deviations for measurement of 18 samples of SiO2-PE1, 18 samples of SiO2-PE2, and 6 samples for PP. .................................................................................................................................... 86 Table 4-4: Summary of the substrate and membrane ................................................................................... 91 Table 4-5: Water vapour transport resistances in SiO2-PE1-PEO-PU membranes at 50°C, aH2O,feed=0.5. ......... 93 Table 4-6: Water vapour transport resistances in PP-PEO-PU membranes at 50°C, aH2O,feed=0.5. .................. 95 Table 4-7: Water vapour transport resistances in PP-PEO-PU membranes at 25°C, aH2O,feed=0.5. .................. 96 Table 5-1: Properties of the nanofibrous substrates generated at three electrospinning times, basis weight errors are standard deviations based on measurements from three samples, fiber diameter errors are standard deviation based on 150 measurements, 50 measurements three locations in each nanofiber mat. ....................................................................................................................................................113 Table 5-2: Properties of nanofibrous substrates and impregnated nanofibrous membranes, thickness is the average of 75 measurements, 25 measurements at each of three locations in each membrane, the error in thickness is standard deviation from the measurements. ......................................................114 Table 5-3: Oxygen crossover in impregnated nanofibrous membranes, one membrane sample was tested for each measurement. ............................................................................................................................118 xii  Table 5-4: Water vapour transport properties of impregnated nanofibrous membranes, one membrane sample was measured for each test, at standard test conditions: T=T1=T3=50°C;Pv,H2O,1=0 Pa; Pv,H2O,3=6122  Pa; υ1= υ3=4.75 m/s. ......................................................................................................121 Table 5-5: Permeance and permeability of water vapour in each impregnated nanofibrous membrane averaged over four flow rates (υ1= υ3=4.75,6.35,7.94 and 9.52 m/s) at standard test conditions: T=T1=T3=50°C;Pv,H2O,1=0 Pa; Pv,H2O,3=6122  Pa; errors are the standard deviation on the average of the four measurements for each membrane. ...........................................................................................122 Table 5-6: Membranes as catagorized for permeability-selectivity interpretation of the IENM’s. ................123 Table 6-1: Review of various reported methods of fabricating non-porous membranes based on electrospun nanofibers. .........................................................................................................................................133 Table 6-2: Electrospinning solution properties and resulting fiber diameters, errors are sample standard deviations for a total of 150 diameter measurements at 3 locations in each material. .......................135 Table 6-3: Summary of PEO-PU impregnated PAN nanofiber membranes on PET non-woven carriers, results are for one membrane sample at each set of fabrication parameters. ...............................................137 Table 6-4: Based on experimental results, minimum fiber basis weights required to achieve oxygen permeance target at a given fiber diameter. ......................................................................................145 Table 6-5: Gravure cylinders used in the study. ...........................................................................................156 Table 6-6: Formulation parameters for coating dispersions. ........................................................................160 Table 6-7: Results of wetting tests on various fiber mats, errors are sample standard deviations from the measurement of three wetting tests. .................................................................................................162 Table 7-1: Non-woven carrier layer properties. ...........................................................................................168 Table 7-2: Summary of thermal properties of the non-wovens ....................................................................169 Table 7-3: Mechanical properties of non-woven carriers in machine (MD) and cross (XD) directions at various temperatures, errors are the sample standard deviations for measurement of three materials samples. ...........................................................................................................................................................170 Table 7-4: Thermal properties of electrospun PAN fibers, PEO-PU films, and PEO-PU filled PAN fibers. ......174 Table 7-5: Summary of thermal degradation of PEO-PU and PAN nanofiber filled PEO-PU films in air and nitrogen. .............................................................................................................................................175 Table 7-6: Summary of mechanical properties for PEO-PU films and PAN NF filled PEO-PU films (average for three samples) at 150°C, error is the sample standard deviation for measurement of three samples of each material. .....................................................................................................................................177 Table 7-7: Mechanical properties of membranes at different temperatures and orientations, errors are the sample standard deviations for measurement of three materials samples. ........................................179 Table 7-8: NF membrane elongated at 95°C to different levels of elongation. .............................................180 Table 7-9: NF membrane elongated at 125°C to different levels of elongation. ...........................................182 xiii  Table 7-10: Summary of performance properties of formed membranes, data are for individual samples of each material. .....................................................................................................................................185   xiv  List of Figures Figure 1-1: Typical function of ERV system and membrane in summer conditions. ......................................... 2 Figure 1-2: Cross-flow plate-type ERV system function in cooling and heating conditions, and an example of a cross-flow ERV core. ........................................................................................................................... 4 Figure 1-3: Water vapour permeability and selectivity of water vapour over nitrogen in various polymers at 30oC including the PEO-PU polymers from the present study with data from [62] and [63] (1 Barrer = 1×10-10 cm3 (STP)•cm cm-2 s-1 cmHg-1). ................................................................................................. 13 Figure 2-1: Water vapour transport test station setup. ................................................................................. 27 Figure 2-2: Counter-flow water vapour permeance test orientation. ............................................................ 28 Figure 3-1: Structure of Tegomer® D-3404 (Evonik). ..................................................................................... 35 Figure 3-2: Diisocyanate (IPDI) reaction with polyols (PEG, Tegomer©) to generate a polyether-polyurethane as described in the PERMAX™ patent literature. ................................................................................. 36 Figure 3-3: Atom labels for the general structure of PERMAX™ polymers for NMR analysis. ........................ 38 Figure 3-4: 1H NMR of the PEO-PU polymer. ................................................................................................. 38 Figure 3-5: Predicted structure of PEO-PU polymer. ..................................................................................... 39 Figure 3-6: GPC data for the PEO-PU polymer in DMF. .................................................................................. 41 Figure 3-7: General carbo-diimide crosslinking scheme with a carboxyl functionalized polymer. ................. 42 Figure 3-8: FTIR transmission spectra for uncrosslinked and cross-linked films containg different weight percentage cross-linker (%XL). ............................................................................................................. 43 Figure 3-9: IR absorbance spectra (normalized to the largest absorption peak in all spectra) for dried films of (a.) cross-linker (XL-702), (b.) PEO-PU film with 15% cross-linker, and (c.) PEO-PU (PERMAX™ 230), ... 45 Figure 3-10: IR absorbance spectra for films with increasing cross-linker addition, less the spectra of the uncross-linked film (upward peaks are increased absorption and downward peaks are decreased absorption relative to the uncross-linked samples). ............................................................................ 46 Figure 3-11: (a.) representative DSC of a crosslinked (6.7% wt) PEO-PU polymer (2nd heat cycle at 20°C/min after cooling from 150°C at 10°C/min) and (b.) TGA/DTA in N2 (20°C/min). ......................................... 47 Figure 3-12: Deconvolution of the DTA data for the thermal degradation of PEO-PU in Nitrogen. ................ 48 Figure 3-13: Liquid water uptake in cross-linked PEO-PU polymers at 23°C, error bars are sample standard deviation about the mean for three samples of the material. ............................................................. 49 Figure 3-14: Liquid water uptake (swelling) of cross-linked and uncross-linked PEO-PU polymers at various temperatures, uncrosslinked samples gelled above 60°C and dissolved above 70°C, error bars are sample standard deviations about the mean for measurements of three samples of the material. .... 53 Figure 3-15: (a.) Heating at (left, 20°C/min) and, (b.) cooling (right, 10°C/min) DSC thermographs for water saturated polymer at different cross-linking ratios. ............................................................................. 55 xv  Figure 3-16: DSC heating of PEO-PU with 6.3% cross-linker and increasing weight percentage of water in the polymer. .............................................................................................................................................. 56 Figure 3-17: Water vapour uptake curves at various water vapour activities for PEO-PU polymer with 6.7% cross-linker at 50°C. ............................................................................................................................. 57 Figure 3-18: Water vapour sorption isotherms for a 6.7% cross-linked PEO-PU sample at various temperatures. ..................................................................................................................................... 60 Figure 3-19: Water vapour sorption isotherms for three levels of cross-linking in PEO-PU polymer at 25oC. 61 Figure 3-20: Solubility coefficients for a PEO-PU polymer cross-linked at 6.7% at different temperatures. ... 62 Figure 3-21: Interaction parameter (χ) for water vapour with PEO-PU polymer cross-linked at 6.7% wt. carbodiimide at different temperatures. ............................................................................................. 63 Figure 3-22: Interaction parameter (χ) for water vapour with PEO-PU at different levels of cross-linking at 25°C. .................................................................................................................................................... 64 Figure 3-23: Average cluster size of water molecules in the PEO-PU polymer at 6.7% wt. carbodiimide cross-linker at various temperatures and water vapour activities. ............................................................... 66 Figure 3-24: Average cluster size of water molecules in the PEO-PU polymer at two levels of cross-linking at at varying relative humidity and 25°C. ................................................................................................. 66 Figure 3-25: Water vapour permeability for a 25 micron thick film of a 6.7% crosslinked PEO-PU sample at various temperatures. ......................................................................................................................... 68 Figure 3-26: Solubility and diffusivity coefficients, and permeability in the PEO-PU polymer. ...................... 70 Figure 4-1: Membrane permeation test module. .......................................................................................... 73 Figure 4-2: Representation of the resistances in the water vapour transport process through a cross-section of a composite membrane. .................................................................................................................. 77 Figure 4-3: a. Module used for permeation experiments [102]; b. CFD analysis of flow distribution in module [101]. ................................................................................................................................................... 80 Figure 4-4: Linear regression line with 95% confidence interval for the regression for the boundary layer model based on the least squares analysis of the fitting parameters for equation [4-33] against the experimental boundary layer measured by changing substrate thickness for flows between 4 and 12 sL/min. ................................................................................................................................................ 83 Figure 4-5: Total feed and sweep boundary layer resistance predicted from model at different temperatures and flow rates (aH2O,feed = 0.5, aH2O,sweep=0.0). ....................................................................................... 84 Figure 4-6: Comparison of the predicted and observed boundary layer resistances based on the empirically predicted boundary layers for the test module for various porous materials at various conditions [4 to 15 nL/min, aH2O > 0.20, T=25 to 70°C]. .................................................................................................. 85 Figure 4-7: Pore size distributions of three substrates from mercury intrusion porosimetry. ....................... 88 xvi  Figure 4-8: Substrate surfaces at 5kx and 20kx (inset) magnification, silica-polyethylene (SiO2-PE1, left), polypropylene (PP, right). .................................................................................................................... 88 Figure 4-9: Flux, apparent permeance and boundary layer ‘corrected’ or actual permeance for two microporous substrates (SiO2-PE1, left and PP, right) at various temperatures (aH2O=0.5); data is for single material samples. ...................................................................................................................... 89 Figure 4-10: Cross-sectional images of a commercially available ERV membrane, the ‘composite membrane’ has a thin dense polymer layer on the surface of a microporous support. ........................................... 89 Figure 4-11: Example of a coating thickness measurement for a SiO2-PE1-PEO-PU membrane image. ......... 90 Figure 4-12: PEO-PU film water vapour permeability for three levels of film thickness at 50°C, data points are the sample mean of 20 data points at each activity, error bars are sample standard deviations about the mean for 20 data points at each activity for each film. .................................................................. 92 Figure 4-13: Comparison of the film layer thickness as measured by SEM and the film layer thickness as predicted by water vapour transport measure for a series of composite membranes consisting of a PEO-PU film layer coated on SiO2-PE1 substrates, error bars are the sample standard deviation based on 30 thickness measurements for each sample.................................................................................. 92 Figure 4-14: Substrate surface roughness images from laser confocal microscopy at 50× magnification. ..... 94 Figure 4-15: SEM Surface images of the PEO-PU film layer on a coated silica-PE1 substrate, showing film layer non-uniformities due to the underlying substrate (left) and a higher magnification image of a defect in the PEO-PU barrier layer due to substrate surface roughness. .............................................. 94 Figure 4-16: Temperature effect on water vapour permeance in (a.) SiO2-PE1 substrate and coated PEO-PU coated SiO2-PE1 membrane; (b.) PP substrate and PEO-PU coated PP membrane; data is for single material samples. ................................................................................................................................ 97 Figure 4-17: Polarized light optical microscopy images of cast PEO-PU films showing crystallinity and phase separated regions. ............................................................................................................................... 98 Figure 4-18: DSC first heating of PEO-PU polymers films as cast at three levels of cross-linking. .................. 98 Figure 4-19: First, second, and third heating cycles for a PEO-PU polymer with 6.7% crosslinking, heating rate is 20°C/min and cooling rate is 10°C/min. ...........................................................................................100 Figure 4-20: DSC first heat of as received of PEO-side chain polyol (Tegomer® D3403) used in PEO-PU polymer at 20°C/min heating rate. .....................................................................................................101 Figure 4-21: ATR-FTIR of SiO2-PE-PEO-PU membrane, PEO-PU film, and PP-PEO-PU membrane. .................103 Figure 4-22: Aqueous dispersion coating process with film formation. ........................................................104 Figure 4-23: Contact angles of water on SiO2-PE (left, 120°) and PP substrate (right, 102°). .........................105 Figure 4-24: FTIR of PP-PEO-PU membrane as cast and after short exposure to liquid water. .....................105 xvii  Figure 4-25: DSC heating (20°C/min) of PEO-PU polymer with 6%XL: a. as cast, dried, and cured and then with a short exposure to liquid water; b. after exposure to liquid water, with multiple increasing heating cycles of increasing temperature. ..........................................................................................106 Figure 4-26: Permeance of PP-PEO-PU membranes as cast and after water exposure, at various temperatures (aH2O=0.5). ....................................................................................................................107 Figure 5-1: Fabrication procedure for impregnated electrospun nanofibrous membranes. .........................112 Figure 5-2: (a.) Cross-sectional image of an impregnated electrospun nanofibrous membrane (MEM-90-21) with the impregnated film layer binding to the non-woven fiber at 500x magnification and (b.) the impregnated nanofibrous layer with a thickness of 9 microns at 2500x magnification. ......................112 Figure 5-3:  Histograms of fiber diameters for three deposition times (60, 90, and 120 minutes) each based on 150 measurements, with 50 measurements at each of three locations on each of the three fibrous mats, a representative image of fibers at 10000x magnification. ........................................................114 Figure 5-4: Surface images of impregnated electrospun nanofibrous membranes at various electrospinning times (depositions) and impregnation solution polymer solids at 2000x magnification, scale bars represent 20 microns. .........................................................................................................................116 Figure 5-5: Freeze fractured film layer cross-sections for all impregnated electrospun nanofibrous membranes at various electrospinning times (depositions) and impregnation solution polymer solids at 5000x magnification, scale bars represent 10 microns. ...................................................................117 Figure 5-6: A water droplet on an IENM, contact angle is 35°. .....................................................................119 Figure 5-7: Water vapour flux in IENMs and three commercially available ERV membranes as a function of velocity in the test module channels, one membrane sample was measured for each test, testing is completed isothermally at 50°C with the wet stream inlet (stream 3) at water vapour pressure of 6120 Pa and the dry stream inlet (stream 1) at water vapour pressure of 0 Pa. ..........................................120 Figure 5-8: Water vapour permeance at the standard test conditions in the IENMs fabricated as a function of impregnation coating solids at various fiber depositions, permeance results are average of measurements at four flow conditions. ..............................................................................................122 Figure 5-9:  Water vapour permeability (reported in Barrer units, where Barrer=1×10-10 cm3 (STP) cm cm-2 s-1 cmHg-1) against selectivity for a series of nanofibrous membranes averaged over a range of gas flow velocities (υ1= υ3=4.75, 6.35, 7.94, and 9.52 m/s). .............................................................................124 Figure 5-10: Water vapour permeability as a function of the PAN fiber ratio to PEO-PU ratio in a series of IENM films with low defects (water over oxygen selectivity, αH2O/O2>100), error bars are the standard deviation around the average of four measurements at different flows for each membrane sample. ...............................................................................................................................................127 Figure 5-11:Water vapour permeability in IENMs and a PEO-PU film (32 micron) as a function of vapour pressure ratio (activity, aH2O) in the wet inlet (stream 3), conditions were isothermal at 50°C, flow xviii  velocity υ1= υ3=6.35 m/s, error bars are the standard deviation for measurements of each sample at each activity. ......................................................................................................................................129 Figure 6-1: Water vapour permeance and selectivity of membranes made from nanofibers of different average fiber diameter with comparison to Chapter 5 IENMs and commercially available membranes [184]. ..................................................................................................................................................140 Figure 6-2: Response surface fit model for all data with desirability output to maximize water vapour permeance while minimizing oxygen permeance, at three levels of coating solids in the impregnation solution a. 12%, b. 15%, and c. 18%. ...................................................................................................141 Figure 6-3: Schematic for fiber filled polymer film cross-sections with different fiber diameters, constant fiber-to-fiber distance.........................................................................................................................142 Figure 6-4: (a.) Estimated fiber volume fraction based on fiber diameter and fiber-to-fiber distance for square packing (b.) reduce curve based on d/σ. .................................................................................143 Figure 6-5: Images of fibers of varying diameter produced by electrospinning (first three images from left are at 10kx magnification, scale bar = 5μm; rightmost image at 2kx magnification, scale bar = 20 μm). ...143 Figure 6-6: Cross-sections of several membranes fabricated from fibers of different diameter and basis weight with thicknesses based on the SEM imaging and based on gravimetric measurement, scale bar for all images = 10 µm; errors on δSEM are the sample standard deviations based on 75 measurements, 25 measurements at three locations in each membrane. ...................................................................146 Figure 6-7: Effective permeability over matrix permeability as a function of fiber volume in the membrane with two models based on permeability and fiber volume calculated based on gravimetric measurements of the coating deposition in the membranes, and thickness measurements based on SEM imaging. ......................................................................................................................................150 Figure 6-8: Effective permeability over matrix permeability as a function of fiber volume in the membrane with two models based on permeability and fiber volume calculated based on corrected gravimetric measurements of the coating deposition in the membranes accounting for the excess coating deposited in the non-woven carrier layer of the membranes. ............................................................151 Figure 6-9: Low Fiber Diameter - Low fiber loading – Low Coating (df=225 nm, ωf=0.185 g/m2, nf=5, 12% polymer solution: PO2/δ=140 GPU, PH2O/δ = 17700 GPU, αH2O/O2=126) ................................................153 Figure 6-10: Mid diameter – Low fiber loading – Mid coating (df=253 nm, ωf=0.214 g/m2, nf=9, 16% polymer solution: PO2/δ =136 GPU, PH2O/δ = 8712 GPU, αH2O/O2=64) .................................................................153 Figure 6-11: Mid diameter – Mid fiber loading – Low coating (df=566 nm, ωf=0.44 g/m2, nf=13, 12% polymer solution: PO2/δ =142 GPU, PH2O/δ = 11000 GPU, αH2O/O2=77) ...............................................................154 Figure 6-12: High Fiber Diameter – Low fiber loading – Mid coating (df=1250 nm, ωf=0.377 g/m2, nf= 3, 16% polymer solution, too leaky to test) ....................................................................................................154 xix  Figure 6-13: High Fiber Diameter – High fiber loading – Mid coating (df=1250 nm, ωf=1.26 g/m2, nf= 12, 16% solution: PO2/δ=361 GPU, PH2O/δ=5285 GPU, αH2O/O2=15) ....................................................................154 Figure 6-14: Benchtop hand gravure roller, gravure cell pattern, and coating method. ...............................156 Figure 6-15: Cross-sectional images comparing (a.) dip-coated and (b.) gravure coated nanofibrous layers attached to the microfibers of the non-woven carrier; arrows show excess coating deposition in the non-woven carrier layer. ....................................................................................................................157 Figure 6-16: Performance of nanofibrous membrane made from gravure coating with various gravure cell volumes, at three polymer coating concentrations and three nanofiber loadings, only membranes with >100 selectivity are shown on the right; all data is from one membrane sample at each fabrication condition. ...........................................................................................................................................158 Figure 6-17: Permeance of adequate selectivity nanofibrous membranes at various coating impregnation levels (based on gravure volume), all data is from one membrane sample at each fabrication condition. ...........................................................................................................................................158 Figure 6-18: Gravure coated nanofibrous membrane surface images,0.52 g/m2 fiber loading, 260nm fibers, 8.2 cm3/m2 200 quad gravure, 33% solids solution; PO2/δ=57 GPU, PH2O/δ=14366 GPU, αH2O/O2=248) .159 Figure 6-19: Gravure coated nanofibrous membrane cross-sectional images, 0.75 g/m2 fiber loading, 260nm fibers, 26% solids coating solution, two gravure cell volumes; 8.2 cm3/m2 (left image) and 35 cm3/m2 (right image). ......................................................................................................................................160 Figure 6-20: Droplets of various volume and coating formulation on a nanofibrous surface after drying. ...161 Figure 7-1: Cross-flow and counter-flow orientations in membrane air-to-air exchangers. ..........................164 Figure 7-2: (a.) A sensible (heat) only exchanger with multi-directional transport based on formed polymer sheets, (b.) an image of the entrance region of a formed heat exchanger plate [189]. .......................165 Figure 7-3: Formable membrane concepts with triangular and square channel cross-sections, and a concept for a triangular channel design with alternating formed and flat sheet membranes (right image) [190]. ...........................................................................................................................................................166 Figure 7-4: First heating cycle DSC for various non-woven carriers (20°C /min). ..........................................168 Figure 7-5: Average stress-strain curves for Asahi Kasei Smash™15100 Non-woven with increasing temperatures, curves are averaged from three sample measurements at each temperature. ...........171 Figure 7-6: DMA of non-wovens in tension in machine direction (left) and cross-direction (right), 1Hz frequency, 10 micron amplitude, heating at 5°C/min, data is for one sample of each material. .........171 Figure 7-7:  First heating cycle of DSC (20°C/min) of crosslinked polyether-polyurethane film (PEO-PU, solid line), electospun polyacrylonitrile nanofibers (PAN_NF, dashed line), and a film of PAN nanofibers impregnated with crosslinked polyether-polyurethane co-polymer (PEO-PU+PAN_NF, dash-dot), curves are offset for visibility..............................................................................................................173 xx  Figure 7-8: Second heating cycle (20°C/min) of crosslinked polyether-polyurethane film (PEO-PU, solid line), electospun polyacrylonitrile nanofibers (PAN_NF, dashed line), and a film of PAN nanofibers impregnated with crosslinked polyether-polyurethane co-polymer (PEO-PU+PAN_NF, dash-dot). ....174 Figure 7-9: TGA and DTA of PEO-PU copolymer films and PAN nanofiber filled PEO-PU films in air and nitrogen atmospheres. .......................................................................................................................176 Figure 7-10: Mechanical testing of PEO-PU films and PAN NF filled PEO-PU films at 0.2 mm/min and 150°C, averaged curves for three samples, only shown to minimum elongation for each sample set. ...........177 Figure 7-11: DMA of a PEO-PU polymer film and a PEO-PU film loaded with PAN nanofibers in tension, 1Hz frequency, 10 micron amplitude, heating at 5°C/min, data is for single samples of each material. ....178 Figure 7-12: Defect in elongated nanofibrous membranes, direction of elongation is across the page, isothermal at 125°C, 185% elongation. ...............................................................................................183 Figure 7-13: Heated press designed and built for forming membranes, press has temperature control on the enclosure, and heating and cooling control on the plates, the rate of compress can also be controlled. ...........................................................................................................................................................183 Figure 7-14: Plate mold geometry for membrane forming press. .................................................................184 Figure 7-15: Formed nanofibrous membrane (a.) and magnification of flow feature in formed membrane showing no defects (b.). .....................................................................................................................186 Figure 8-1: A formable membrane exchanger. .............................................................................................193 Figure 8-2: Assembly of a formable nanofibrous membrane exchanger for testing. ....................................194 Figure 8-3: (a.) Total effectiveness performance of cross-flow exchangers based on flat sheet membranes and formable nanofibrous membranes; (b.) total pressure drop in the exchanger channels; error bars include instrument and experimental errors as the standard deviation of three test measurements. 194    xxi  List of Abbreviations General Abbreviations ASHRAE American Society of Heating, Refrigeration, and Air Conditioning Engineers ASTM American Society for Testing and Materials ERV Energy Recovery Ventilator GPU Gas Permeance Unit  HRV Heat Recovery Ventilator HVAC&R Heating, Ventilation, Air Conditioning, and Refrigeration IENM Impregnated electrospun nanofibrous membrane LPM Liter per Minute STP Standard Temperature and Pressure (273.15K and 101.3 kPa) VOC Volatile Organic Compound WVT Water Vapour Transport WVTR Water Vapour Transport Rate Materials CO2 Carbon dioxide DMF Dimethylformamide DMPA Dimethylolpropanoic acid DMSO Dimethyl sulfoxide H2O Water IPDI Isophorone diisocyanate N2 Nitrogen NF Nanofiber or Nanofibrous NW Non-woven O2 Oxygen PAN Polyacrylonitrile PE Polyethylene PEBAX Polyester-block-amide PEG Polyethylene Glycol PEO Polyethyleneoxide PEO-PU Polyether-polyurethane copolymer based on polyethylene oxide PET Polyethylene terephthalate (polyester) PMMA Poly(methyl methacrylate) xxii  PP Polypropylene PTFE Polytetrafluoroethylene PU Polyurethane PUD Polyurethane dispersion PVC Polyvinyl Chloride SiO2 Silica TPU Thermoplastic polyurethane XL Cross-linker Techniques ATR-FTIR Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy DMA Dynamic Mechanical Analysis DTA Differential Thermal Analysis DSC Differential Scanning Calorimetry DVS Dynamic Vapour Sorption FTIR Fourier Transform Infrared Spectroscopy GPC Gel Permeation Chromatography NMR Nuclear Magnetic Resonance SEM Scanning Electron Microscopy TGA Thermogravimetric Analysis      xxiii  List of Symbols Symbol Unit Description  General Nomenclature  A cm2, m2 Membrane active area, or area aH2O - Water vapour activity in the feed stream (pv,H2O/psat) at temperature C mol/m3 Concentration d µm Fiber diameter D m2/s Diffusivity E MPa Modulus H J/g, J/mol Enthalpy h m height JH2O kg/m2/day, mol/m2/s Water vapour flux JO2 mol/m2/s Oxygen Flux  k1 s kinetic parameter ki m/s Mass transport coefficient  kmem m/s Membrane mass transport coefficient kobs m/s Observed/measured mass transport coefficient L m Length ?̇?H2O kg/s Mass flow, water vapour  m g, mg Mass  Pi Barrer Permeability (1 Barrer=1×10-10 cm3 (STP) cm cm-2 s-1 cmHg-1) Pi/ δ GPU Permeance (1 GPU = 1×10-6 cm3 (STP) cm-2 s-1 cmHg-1) PH2O Barrer Water Vapour Permeability PH2O/ δ GPU Water Vapour Permeance pO2 Pa Partial pressure oxygen PO2 Barrer Oxygen Permeability psat Pa Water vapour saturation pressure at test temperature pv,H2O Pa Partial pressure water vapour Q m3/s Gas flow rate R J/mol/K Gas constant Rbl s/m Total boundary layer resistance (feed+sweep) Rfeed s/m Boundary layer resistance to water vapour transport in feed (wet) stream  Rmem s/m Membrane resistance to water vapour transport in feed stream Robs s/m Observed/measured resistance to water vapour transport in feed stream  Rsweep s/m Boundary layer resistance to water vapour transport in sweep (dry) stream  S cm3 (STP)/cm3 /Pa Solubility Coefficient  T K, °C Temperature t s time Vm cm3/mol Molar Volume of Gas at standard temperature and pressure (22414) xxiv  x kg/kg Humidity ratio (kg H2O/kg air) W - Uptake w m width V - Volume fraction  Greek Letters/Symbols  αH2O/O2 - Selectivity (Water Vapour over Oxygen = PH2O/PO2) αf m3/m2 Volume areal density γ  Activity coefficient of gas (assumed 1) δ µm Film thickness  ℓf m Fiber length ε % Porosity ρ g/cm3 Density σ µm Pore size υ m/s Air flow velocity µ Pa-s Dynamic Viscosity µi J/mol Chemical potential τ - Tortousity ω g/m2 Basis weight, areal density      xxv  Acknowledgements Firstly, I would like to thank Dr. Frank Ko who acted as my primary supervisor throughout this project for inviting me to work under him in his lab. Dr. Ko has been a source wisdom, support, and guidance throughout my studies.  I would also like to thank Dr. Walter Mérida who co-supervised this project, and provided timely advice and direction throughout my studies. I would also like to thank all of the members of the AFML lab at UBC (past and present) who have helped with equipment, advice, and training over the years. I would like to thank thesis committee members Pierre Bérubé in Civil Engineering, Göran Fernlund in Materials Engineering, and Edouard Asselin in Materials Engineering, for their time and insightful comments and suggestions in the development of this dissertation. Finally, a thank you to Dr. Greg Smith in Forestry and Dr. Madjid Mohseni in Chemical Engineering for reviewing the dissertation and acting as University Examiners for my final defense.   I would like to acknowledge the Natural Sciences and Engineering Research Council (NSERC) for financial support through a CRD project sponsored by dPoint Technologies. This work would not have been possible without the ongoing support and understanding of all of the team at dPoint Technologies. James Dean, in particular has been incredibly supportive of my work and professional development over many years. Colleagues David Kadylak, Frankie Wong, Scott Dornian, and Hao Chen have provided excellent discussions leading to ongoing insight and understanding throughout this project. To my family and my friends: I regret that this generally ‘selfish’ endeavor of throwing oneself so deeply into a narrow field of study can become all-consuming at times (especially when one is ‘moonlighting’ as a doctoral student). I apologize for not being present or available at many times over the course of this project. I will always be grateful for your understanding, love, and support.    xxvi  Dedication My study of psychrometrics and the nature of water began at a young age in my Finnish grandparents’ cedarwood sauna. There, an understanding of temperature, relative humidity, and even enthalpy was implicit. Water poured on hot stone turned instantly to steam. Incipient vapours condensed onto skin or a cold window. Beads of water mingled with purifying sweat and a refreshing plunge into the lake outside provided clarity. This work is dedicated to my parents: Nancy and Nick, my grandparents: Eva and Ollie; and Myna and Gerry for being excellent teachers, for instilling a sense of wonder, inquiry, and diligence in me, and providing the foundation for the person I am today. This is also dedicated to the myriad other teachers and mentors, whom through my great fortune have entered my life and provided wisdom, guidance, time, criticism, questions, inspiration, and support along the way. For all this I am eternally grateful. As recompense, I can only hope to offer the same to others over the years to come.   from Alpha when I entered the cave through Omega still to come to all those who guided and revealed the nature of light and lightness thank you   1   Introduction and Background 1.1 Membrane Science Polymer membrane science has greatly matured over the past 30 years. This has led to the use of membranes in a wide range of applications, with membrane technology now playing a central role in industrial desalinization and the wastewater treatment [1]–[3].  A brief review of the literature indicates that polymer membrane research is active in several areas including but not limited to; water treatment, desalination, gas separation, fuel cells, solvent separation, and dialysis [4]–[11]. Polymer membranes have proven to have versatile industrial applications as they allow for highly selective separations with low energy requirements compared to incumbent technologies.  There are generally two key metrics of concern in membrane technology, permeability and selectivity. Permeability describes the rate of transport of a species through a unit area of a homogeneous material of known thickness at a known differential in chemical potential, pressure, or concentration across the material. Selectivity describes the relative permeability of two chemical species of interest, for example, that of water vapour over oxygen through a material at a given set of conditions. Other important definitions include flux, which describes the rate of transport through a membrane per unit area; and permeance, which describes the flux normalized by the concentration, pressure, vapour pressure, or chemical potential driving force. Flux and permeance are not inherent properties of a material, but rather result from the structural properties of the material, the geometry of the membrane module, and operating environment in which the membrane is utilized. Flux is best used to describe the transport performance of a membrane in a device under actual operating conditions. Permeance is best used to describe and compare functional performance of composite membranes which are non-homogeneous in structure and composition. This work focuses on gas and vapour separation applications, specifically water vapour transport applications, in which a dense, non-porous polymer layer both permeable and selective for water vapour is utilized to achieve separation. In gas separations, as in many membranes processes it is desirable to maximize both permeance and selectivity. However, these goals are contradictory: as first described by Robeson, for a given pair of gas species there will tend to be a trade-off between permeability and selectivity in polymeric materials. This is referred to as the upper bound [12]. However, it is noted that for water vapour-gas 2  separations this often does not hold true, and both high selectivity and permeability can be achieved [13]. The solution diffusion model describes permeability in dense polymers as the product of solubility and diffusivity [14]. Diffusivity is dependent on size and nature of the permeating species and the nature of the polymer, which includes the polymer chain packing and orientation, functional groups in the polymer chain, polarity, degree of crystallinity, polymer additives, crosslinking density, presence of fillers, and degree to which the polymer is plasticized in the presence the permeating species. Solubility depends on the temperature, and the thermodynamic interactions of the polymer and the species that is being dissolved.  1.2 Energy Recovery in Buildings Air-to-air heat and moisture exchange devices, also known as enthalpy exchangers, utilize membranes to selectively transport heat and water vapour between gas streams. These heat and mass exchangers are an integral part of many modern ventilation systems.  Ventilation systems are used to exhaust stale air and bring fresh air into sealed buildings. In order to maintain a comfortable indoor environment, this incoming air must be heated or cooled, which consumes energy.  Consequently, membrane based plate-type energy recovery ventilators (ERVs) are a component used in many energy efficient ventilation systems.  Figure 1-1: Typical function of ERV system and membrane in summer conditions. In these air-to-air ERV exchangers, incoming and outgoing air streams are passed over opposing sides of a membrane through which heat and moisture are transferred as shown in 3  Figure 1-1. This decreases the energy consumption of buildings by using the exhaust air to heat/cool and humidify/dehumidify the incoming air depending on the season.  Plate type membrane based energy recovery ventilators (ERVs) improve ventilated building energy efficiency by transporting heat and water vapour between incoming and outgoing air streams in buildings, effectively ‘recycling’ the energy used to condition building air. In modern buildings, ventilation systems are used to exhaust stale air, and bring in fresh air. In ‘cooling’ conditions for example (see Figure 1-2), the incoming air is hot and humid, and the exhaust air is cool and dry. Most of the energy used in air conditioning is expended condensing water vapour, so by passing the cool dry exhaust air over one side of a permeable membrane, and the hot humid incoming air over the opposing side of the membrane, the exhaust air can be used to precool and dehumidify the incoming air. In ‘heating’ conditions (North American winters, see Figure 1-2), the enthalpy exchanger operates in the opposite direction, incoming cold air is preheated with outgoing warm exhaust air, and moisture is transported through the membrane from the exhaust air to humidify the incoming air. The use of these devices can have a significant impact on building heating and cooling efficiency, while also improving indoor air quality and comfort [15]. Energy recovery devices improve indoor air quality through ventilation and improve overall building ventilation energy efficiency over a wide range of climates and seasonal conditions [16]. In cooling conditions, removal of moisture and cooling of incoming air improves energy efficiency by decreasing air conditioning loads. In heating conditions, energy efficiency is improved by pre-heating incoming air. An additional benefit of using an ERV over an HRV in heating conditions (particularly in cold climates) is that condensation (which would otherwise need to be captured with an added drain from the heat exchanger) is transferred to the dry incoming air improving on the low humidity levels which are often experienced indoors with sub-freezing outdoor winter conditions. The additional transport of moisture though the membrane in ERVs also allows less use of defrost cycles compared to HRV systems (saving energy), as the temperature at which frosting is initiated in the ERV is lowered due to latent transport and frosting/freezing in the exchanger is decreased [17]. The overall ‘effectiveness’ of these ERVs is defined in the ASHRAE-84 standard in terms of sensible (heat transport) and latent (moisture transport) effectiveness [18]. Membranes for these devices must have high water vapour permeance and should be selective for water vapour over other gases, volatile organic compounds (VOCs), and contaminants that may be present in the outgoing indoor air. 4   Figure 1-2: Cross-flow plate-type ERV system function in cooling and heating conditions, and an example of a cross-flow ERV core. High-efficiency HVAC ERV systems have the potential to save up to 65% of energy associated with the conditioning of outdoor air in buildings [15]. In 2013, residential and commercial buildings in Canada were reported to represent 27% of total energy consumption in Canada, and about 58% of this energy was used in space heating and cooling [19]. This means that 16% of all energy consumption and 8% of all greenhouse gas emissions in Canada are associated with HVAC systems. This represents 1523 PJ of energy and 70 megatonnes of CO2 equivalents in Green House Gas (GHG) emissions annually. The implementation of high-efficiency energy recovery systems in buildings could drastically decrease energy consumption in Canada. As of 2002, less than 1% of buildings utilized ERV technology [15].  However, in recent years the use of this technology is growing as greater interest in building energy efficiency and indoor air quality are driving the growth of the energy recovery ventilation market. Current market estimates place building energy recovery ventilation as 5  growing market which is currently estimated to reach $3 Billion in annual revenue world-wide by 2020 [20]. Research by NRCAN on the effectiveness of complete enthalpy recovery systems has demonstrated the improved energy savings of these devices over heat recovery devices, especially in summer condition [21]. Previously, concerns over freeze-thaw longevity in incumbent ERV paper transport media had stalled their adaptation in colder climates. Recently, polymer-based ERV devices developed by dPoint Technologies have been proven to be more effective than heat only recovery devices currently used in winter and freezing conditions [22]. A vital component of this technology is the functional membrane that interfaces with the gas streams. The membranes must be selective for water and water vapour, not allowing the gases between the streams to mix. The membranes must have extremely high water vapour permeance rates allowing for compact module design. This leads to lower pressure losses in operation, lower cost manufacturing, and decreased material consumption for manufacturing. The membrane materials must be able to function over a wide range of operating conditions from sub-zero temperatures down to less than -25°C to temperatures above 50°C. They must demonstrate long lifetimes, and have sufficient handling so they can be manufactured into modules that can be placed into systems in a cost-effective manner. With such complex requirements, careful membrane design is necessary to create effective enthalpy exchange devices.  1.3 Membranes Requirements for ERV Many commercially available planar membrane ERV devices utilize desiccant filled paper as a transport media. These paper-based membranes are associated with low performance of ERV devices, poor durability in wet and freezing conditions, and susceptibility to degradation by mold and bacteria [23]. Alternative membranes for these devices are composite membranes which are manufactured based on micro-porous substrates, coated with a thin dense layer of water vapour-transporting polymer. The substrate provides structural rigidity to the membrane while the thin functional polymer layer provides selectivity. The resistance to water transport, and thus the performance of these materials, is the result of resistance to transport both in the substrate and in the coating layer. In the present work it is demonstrated that more than 50% of water vapour transport resistance comes from the substrate itself in these composite membranes. Consequently, one of the key ways to increase enthalpy 6  exchange efficiency in these devices is by decreasing the resistance to water vapour transport in the substrate material.     Performance of high-efficiency ERV systems is currently around 65% of total enthalpy exchange effectiveness [24].  With improvements in materials design, total exchange effectiveness of greater than 75% should be achievable [25]. Current water vapour permeance performance is at least half of what is realistically achievable from water vapour transport membranes. Some level of improved transport of water vapour can be achieved with the use of high porosity, microporous substrate materials, or polymer films with high inherent water vapour permeability. Membranes with higher vapour transport performance are still required in order to develop high efficiency exchangers. At present, no singular membrane technology has met the requirements of high performance ERVs. Further materials research is required to achieve the improvements needed in permeance, selectivity, durability, and cost in membranes for these devices. Desirable properties of membranes for enthalpy exchangers generally include the following: • High water vapour permeance; • High water vapour sorption; • Low or zero air, gas, and contaminant crossover (high selectivity); • Non-flammable; • Resistance to microbial growth; • Favorable mechanical properties when dry or when wet, so that the membrane is easy to handle, does not tear easily; • Formability; • Good dimensional stability in the presence of liquid water and washable, allowing cleaning for maintenance purposes without damaging or compromising the functionality of the ERV exchanger and membrane; • Long lifetime under the required operating conditions, without detrimental leaching or loss of membrane components and without significant degradation in water vapour transport performance or increased contaminant crossover; • Tolerance to freeze/thaw cycles in the presence of liquid water condensation without significant deterioration in performance; • Low cost 7  The research in the present study focuses on materials with high water vapour permeance, high gas selectivity, and the potential for low cost. Furthermore, a membrane which can be thermal formed into exchanger plates is demonstrated which allows for advanced higher efficiency exchanger designs. 1.4 Composite Membranes Composite membranes in this study refer to a membrane with a microporous substrate layer which is coated on one side with a thin dense layer of a selective water vapour permeable polymer. At least one industrial manufacturer (dPoint Technologies) currently produces membranes of this nature for energy recovery ventilation and fuel cell humidification applications [26]. The composite membrane approach is advantageous as the substrate gives the membrane mechanical strength and robust handling properties. The selective layer on the microporous substrate surface can be relatively thin (<2 microns) leading to high permeance of water vapour. The substrate also provides dimensional stability to the selective coating layer if it swells under variable relative humidity, freezing, and condensing conditions in which the membrane is exposed to liquid water. Current generation composite membranes are studied in detail in Chapter 4 of this thesis. 1.5 Nanofibrous Materials Electrospun nanofibrous materials have been the focus of significant research interest in the past 15 years [27]–[33]. Nanofibrous materials are of particular interest in researching highly permeable membranes for building energy recovery. These materials have inherently high porosity due to small fiber diameters. For fiber diameters of less than 1000 nm, the bulk of these materials are void space (porosity 70% to > 95%) [34]. This high porosity means that the materials will have low resistance to transport. For the purposes of this research, nanofibrous materials will thus refer to materials containing fibers less than 1000 nm in diameter. Nearly all the pores in these materials will be through-pores, meaning the pores are connected between the two material surfaces, with few dead-end pores. This means that water vapour will diffuse through the nanofibrous layer via open space between the fibers, with a relatively unobstructed straight path through the thickness of the layer and little interaction with the fibers themselves. Thus, the path length of a diffusing molecule of water vapour travelling in the axis perpendicular to the plane of the nanofibrous layer is within the range of the thickness of the layer.  This leads to lower resistance to the transport of diffusing species, as molecules will have a more direct and unimpeded pathway 8  through the nanofibrous layer, and thus higher permeance. This is consistent with a number of studies highlighting the low pressure drop per filtration efficiency of these materials [35]. As such, research and application of nanofibrous materials in filtration application has grown dramatically in the past 10 years [36]. Although nanofibrous materials will have high water vapour transport rates, due to their high porosity, they will also be highly permeable to other gases and will have little selectivity, making them insufficient for ERV applications. In order to make selective gas separation membranes, a dense polymer layer both selective and highly permeable to water vapour will be required.  The focus of the present research will be to utilize nanofibrous materials as a scaffold for the fabrication of water vapour transport membranes for energy recovery ventilator applications. This study presents a new type of membrane for ERV applications based on a nanofibrous layer. These membranes use electrospun nanofibers to create a scaffold which can be filled or partially filled with a selective water vapour permeable polymer. The selective layer created is a fiber filled polymer layer attached to the surface of a non-woven carrier. The development of these membranes is summarized in Chapter 5, and the optimization of these membranes is reported in Chapter 6. Finally, membranes that can be thermally formed into exchanger plates are made from these electrospun nanofibrous membranes, as reported in Chapter 7. 1.6 Literature Review 1.6.1 Plate and Frame ERVs Plate and frame energy recovery ventilators have been utilized for energy recovery in buildings dating back to the 1970s [37]. Numerous studies have been completed to evaluate and improve these devices and systems. Recent interest in building energy efficiency has led to increased research in energy recovery ventilation devices, and systems. This work focuses on design and fabrication of membranes for building enthalpy exchangers, with particular attention to flat sheet membranes for plate and frame designs. The literature review is thus mostly limited to these designs. Published literature on these efforts can be divided into four general categories: 1. Building energy use analysis and system design 2. Exchanger design 3. Heat and mass transport modeling 4. Membrane design 9  1.6.1.1 Building Energy Use Analysis and System Design With indoor air quality becoming a greater concern some studies and building standards recommend up to 0.5 air exchanges in residential buildings per hour [38]. From an energy consumption perspective, the appeal of energy recovery technology to improve building ventilation system efficiency is apparent.  In particular, ERV systems offer a significant benefit when utilized in ventilated, air-conditioned buildings, in warm humid climates. However, in order to justify the use of energy recovery system the extent of energy savings must be quantified. To this end, studies have been completed measuring the effectiveness of different energy recovery schemes in various climates and seasons. In an early study, Drost reported the results of a study on 38 houses with air-to-air energy recovery using heat recovery ventilators (HRV) systems over one heating season in Northern Washington, demonstrating 52% overall thermal efficiency [39]. A more recent study commissioned by the US Department of Energy to study energy consumption in commercial buildings and compare various technological approaches for energy savings, reported that Energy Recovery Ventilators, although utilized in less than 1% of buildings, had the second highest potential for energy savings in commercial buildings compared to various other technologies in the study [15]. This is supported by heat and mass transfer modeling for cross-flow ERV cores in various geographic regions with the results indicating that 58% of energy for air conditioning can be saved by utilizing ERV cores, and only 10% for HRVs [40]. This was confirmed in testing at the Canadian Center for Housing Technology (CCHT), where energy consumption was measured over the summer months in twin houses, one with an HRV, the other with an ERV. The results showed that the house with an ERV had reduced overall energy consumption as well as improved indoor humidity control [21]. An analysis comparing the use of ERVs and HRVs in the different provinces of China recommended the use of ERVs in most locations [41]. Zhang et. al. presented a thermodynamic model of a coupled air conditioning and energy recovery system in which the results indicate that adding an ERV can save up to 33% of primary cooling energy [42]. Nasif, et. al. reported experimental data was used in a HPRate model to determine total energy savings in coupling a ERV with an air conditioner [43]. The model determined annual overall energy savings of 4 to 8% in various locations worldwide. In another study, ERV performance was analyzed for various regions and weather conditions throughout China and operational recommendations were made dependent on the local climate conditions [44]. ERV’s were recommended in nearly all regions, however the authors indicate that higher efficiency fans and higher performance enthalpy exchange materials are 10  still required. A study by Nui considered combined cooling and dehumidification strategies, and found that the use of an ERV as part of the system demonstrated the best energy savings in the climates modeled [45]. Another study demonstrated the benefits of using a mechanical direct expansion dehumidification system coupled with ERVs [46]. Considering these studies as a whole indicates there is strong scientific consensus recommending the use of energy recovery ventilators in building ventilation systems, and that improved enthalpy exchanger designs and materials are desirable. 1.6.1.2 Exchanger Design A number of recent reports have considered modifications of cross-flow or counter-flow building energy recovery devices. Kotcioglu et. al. used an entropy generation minimization procedure to analyze a crossflow heater exchanger with winglets as turbulence generation [47]. They found that heat transfer was improved with this design at high velocities, indicating that new plate designs have the potential to improve exchanger efficiencies in these devices. Lu et. al reported a novel thin film cross-flow heat exchanger in which the film is able to vibrate, they conclude that the vibration of the film improves heat transfer efficiency [48].   Generally flat plate enthalpy exchangers are reported in the literature, however some hollow fiber, shell and tube designs have also been reported. Kistler and Cussler reported the use of both hollow fiber and flat sheet devices from interfacially polymerized polyamide membranes, they demonstrated that both devices had high heat and moisture exchanger effectiveness, and that ease of manufacture would be the deciding factor in producing these exchangers on the commercial scale [49]. Zhang recently reported an exchanger model for a hollow fiber exchanger, demonstrating that this geometry could also be utilized in building energy recovery devices [50]. Zhang has also reported a quasi-counterflow exchanger, which demonstrated improved heat and mass transport performance over cross-flow designs for building energy recovery [25]. A high efficiency membrane heat and mass exchanger design based on thermally-formable membranes is presented in Chapter 7 of this dissertation, and membranes which can be formed into exchanger plates are demonstrated. 1.6.1.3 Heat and Mass Transport Modeling Numerous studies have reported the results of heat and mass transport modeling of energy recovery devices. This area is of interest due to the unique non-isothermal coupling of heat and moisture transport in these devices. Most studies also include some coefficients to 11  describe the properties of the membranes utilized in these devices in their models. These studies indicate the material and geometrical properties of membranes that will improve device performance.  In an early study, Zhang reported a heat and mass transport model for cross-flow core based on hydrophilic and porous membrane core [51]. This study indicated that membrane area is not effectively utilized in cross-flow devices, and that a large portion of the transport in the ERV core occurs at the inlet.  Sparrow et. al. reported a selective and water permeable core that had 50% latent (water vapour) recovery, with an NTU (number of transfer unit) model based on a heat exchanger analogy which fit the experimental data [52]. Zhang and Nui also demonstrated and NTU method for cross-flow ERV cores and included the sorption coefficients of the membranes in the model [53]. Sorption relations for materials were compared, and materials with C=1 (linear, flat) sorption isotherms demonstrate the best performance over the relevant operating conditions in energy recovery ventilators. Zhang built upon previous models to model a cross-flow exchanger in which flow is hydro-dynamically developed, but developing in thermal and vapour concentration profile, the model matches experimental results for outlet humidity predictions [54]. Min and Su recently reported a model for energy recovery devices, which includes the membrane parameters of moisture diffusivity, moisture sorption, thermal conductivity, and maximum moisture uptake, these results mostly confirm results previously reported by Zhang [55]. They report thermal effectiveness is not significantly affected by the membrane since the thermal resistance in the membrane accounts for little of the total thermal resistance. It is also noted that increasing the maximum moisture uptake of the membrane increases the humidity transport performance and changing the sorption coefficient demonstrates a maximum in humidity transport performance. They report that moisture diffusivity in the membrane has the strongest effect on overall device performance.  1.6.1.4 Membrane Design The membrane is the central functional component of the enthalpy exchanger core, and the focus of the current dissertation. The literature contains reports of the development and testing of a number of membrane materials for energy recovery cores. Zhang reports the permeation testing of porous mixed cellulose and cellulose acetate films to determine the moisture diffusivity for use in an energy recovery ventilator model [56]. The membranes are not selective and are not specifically designed for the application. In a later study, Zhang and 12  coworkers built ERV cores from these membranes and confirmed that these polymer membrane cores had much improved exchange effectiveness over paper-based cores [57]. It is not clear if they have utilized the porous films from their previous study, as they do not state whether the materials are selective for water vapour over other permeants. It is a requirement of these cores to not allow gases and contaminants other than water vapour through the membrane. If the materials utilized are porous, then the effectiveness values reported are artificially high. Later Zhang reported the development of membranes for ERV applications using a porous PES substrate with a polyvinyl alcohol coating containing lithium chloride [23]. It was reported that increasing LiCl content, increased the permeation rate of the membrane material. No selectivity or gas crossover results are reported for these materials, furthermore, LiCl being a water soluble salt is likely to leach out of the membrane on contact with water, which will decrease the useful operational regime, lifetime, and washability of the material. Hwang et.al. reported that sulphonated styrene-ethylene-butylene-styrene (sSEBS) triblock co-polymers had high water uptake and high water vapour transport, suggesting that membranes fabricated from sulphonated SEBS might be promising candidate membranes for ERV applications [58]. A study by Min et.al. analyzed the vapour transport properties of microporous polyethersulfone (PES), polyvinylidene fluoride (PVDF), and cellulose, with 0.22 micron pore size and a thickness of 100 micron [59]. They reported that the membranes with higher sorption coefficients also demonstrated higher water vapour transport. These results are for porous materials, and the membranes are not selective, however there is an indication that improved sorption in the substrate will also improve the permeation performance. Zhang et. al. most recently fabricated asymmetric cellulose acetate membranes for ERV applications by a wet phase inversion technique, demonstrating a selective membrane that was permeable to water vapour but not other gases [60]. They found that in order to decrease the CO2 transfer rate to a sufficiently low level, the dense layer had to be quite thick (>2 micron) and consequently the water permeation decreased significantly (by over 50%) in these materials.  1.6.2 Water Vapour Transport in Membranes Water vapour transport membranes are of interest for gas drying, clothing, and humidification applications. Work by Metz et.al. considered permeation properties of various dense polymers for gas drying and contains an extensive review of water vapour permeability and selectivity in various polymers, shown in Figure 1-3 [61].  13   Figure 1-3: Water vapour permeability and selectivity of water vapour over nitrogen in various polymers at 30oC including the PEO-PU polymers from the present study with data from [62] and [63] (1 Barrer = 1×10-10 cm3 (STP)•cm cm-2 s-1 cmHg-1). Metz’s work focused on polyethylene-oxide containing PEO-PBT copolymers, these block co-polymers containing PEO groups have desirable permeability properties for water vapour transport applications, and similar co-polymers such as polyether block amides, polyether-polyurethanes, and polyether-polyesters have also been studied [64]–[66].  Moisture transport in PEO containing polyurethane polymers is described in detail by Lomax [67]. In the presence of water, the PEO blocks of the polymer create channels through which water can travel, while the polyurethane segment provides mechanical integrity to the polymer. High levels of swelling in certain polymers will lead to the formation of liquid channels through the polymer structure which can increase water vapour transport. If water is the solute, this is often accomplished spontaneously due to the polymer undergoing internal phase separation, as thermodynamic forces cause the alignment of hydrophilic groups within the polymer. Generally, the more mobile a polymer chain is within the coating material, the greater the diffusive water transport rate in the material. Thus, diffusion is greatly increased in polymers in which the chains are mobile. A combination of high sorption for water vapour and rapid diffusion of water vapour in the polymer matrix leads to polymers with high permeability. 14  Chapter 3 of this dissertation, studies a specific PEO-PU based polymer containing PEO side-chains for use in water vapour transport membranes. The polymer interactions with water and water vapour are reported in detail. In a study of a composite membrane based on PVDF porous substrate coated with a PVA dense layer, Zhang reported the resistances of the layers in this membrane and recommended that the substrate resistance be decreased to improve membrane performance [68]. In another study, Zhang reported a resistance model for composite type membranes based on porous substrates with variable pore structures [69]. These two studies by Zhang are important to the literature on ERV membranes as they no longer treat the membrane as a ‘black box’ and rather numerical models are presented which consider the membrane structure as a critical component of the model. However, in both of these studies only one membrane composition is reported and the numerical model is used to predict the resistances of the membrane layers.  Most recently Koester, et. al. have completed experimentation on a range of water vapour transport membranes, and demonstrated the effects of feed and permeate side water vapour activity on water vapour permeance in these membranes [70]. In their study, a composite membrane produced by dPoint Technologies similar to those studied in Chapter 4 of this dissertation was analyzed, however without knowing the permeability and thickness of the ‘active later’ and the thickness and permeance of the substrate layers individually, the authors were unable to fully characterize their results for these materials. A further study completed by Koester et. al. presented a ‘deconvolution’ approach in an attempt to determine the individual resistances of the fluid boundary, porous support, and membrane layers in composite water vapour transport membranes [71]. However, the results for the support resistance in their study were anomalous, showing negative resistances, the substrate and film layer resistances could not be accurately decoupled without studying the permeance properties of the layers individually. In Chapter 4 of this dissertation composite membranes for ERV applications are fabricated and analyzed using a resistance in series model to determine the individual contributions of the boundary layer, the substrate layer, and film layer water vapour transport resistance in composite membranes for ERVs. By using multiple test conditions, different substrates and film layer thicknesses, and by measuring substrate resistances without the film layer, the polymer film layer resistance itself, and then assembled membrane resistances, the contribution of each resistance depending on membrane composition was demonstrated.  15  Chapter 5 also introduces a new class of membranes for water vapour transport, based on electrospun nanofibers which have improved performance and also have the benefit of potential lower material cost and longer lifetime, a requirement of the application. 1.6.3 Water Vapour Transport in Nanofibrous Materials Water vapour transport in nanofibrous materials was first reported in a series of papers by Gibson et. al., who reported high water vapour diffusion and gas convection through polybenzimidazole, polyacrylonitrile, and nylon 6,6 nanofiber mats of varying loadings on open-cell polyurethane foams [72], [73]. Dynamic water vapour transport test equipment was used to demonstrate the mats to have sufficient resistance to convective gas flow, and very low resistance to diffusive water vapour transport. It was concluded that the water vapour transport was sufficiently high for textile applications. Since the authors were more concerned with the use of nanofiber membranes in aerosol resistance in protective clothing, the effects of the nanofiber mat porosity, density, and fiber diameter on water transport were not fully analyzed.  In another series of studies, polyurethane fibrous mats were fabricated and tested for air permeability and water vapour transport [74], [75]. The authors observed that the water vapour transport for nanofibrous materials and microporous materials was statistically the same, even though the air permeation rate of the nanofibrous materials was much greater. Others have electrospun polyurethane mats and compared the water transport properties to PU coated membranes (20 micron thickness), in the study the electrospun membranes had much greater water and air permeation than the coated membrane, this as would be expected from a porous membrane compared to a coated dense film membrane [76]. In both of these studies, the authors used the cup test (ASTM E96) which has been demonstrated to be ineffective for high permeation materials [77]–[81]. Borhani et. al. report spinning PAN nanofiber mats of varying density, and observing the heat and moisture transport using the sweating hot plate method [82]. They report testing fibrous mats with density from 0.84 - 6.64 g/cm3, and bulk porosities of greater than 99%. Air permeability and water transport decreased with increasing density. This ‘sweating hot plate’ method shows some better resolution, but may not be as appropriate as the dynamic test. In a recent study on water vapour transport in nanofibrous polyurethane materials by Zhou, the author correctly utilized the dynamic test method, as reported by Gibson, and was successful at demonstrating resolution in the performance in water vapour transport in various nanofibrous materials [83]. Water vapour permeation measurements were made on shape memory polyurethane (SMPU) nanofibrous mats. A decrease in transport 16  above 50% relative humidity was demonstrated and thought to be related to the uptake of water leading to swelling of the nanofibers, and a decrease in the porosity of the mat. It is evident from the literature that nanofibrous materials have high water vapour transport rates due to high porosity. The nanofibrous materials demonstrated in the above reports will not be selective in gas separations, since virtually all species of gas including water vapour will be capable of transporting through the porous materials.  1.6.4 Nanofibrous Membranes Recent reviews report application-based research on nanofibrous membranes for air filtration, water filtration (micro, ultra, nano), sorption materials for heavy metal removal, and an antimicrobial filters [33]. However, it is noted that very few reports have been made on selective nanofibrous membrane-based separation processes [84]. Recently, the groups of Benjamin Chu and Benjamin Hsiao have led the most significant body of work in the area of selective membranes based on electrospun nanofibers. The membranes fabricated in their work are selective membranes based on multiple layer structures, in which a nanofibrous layer supports a selective polymer layer, these membranes are targeted at water filtration applications.  Their first report of a selective and permeable nanofibrous membrane involved crosslinking a layer of electrospun PVA on a non-woven support, and then coating with MWNT loaded PEBAX or PVA layers. The water flux was shown to increase with MWNT content and to be greater for PVA coatings over PEBAX [85]. The rejection of oil over water was greater than 98% for all samples, but increased slightly with decreasing MWNT content. Further results demonstrated nanofibrous membranes fabricated for water filtration applications; in which PAN fibers are spun on a PET non-woven, and then coated with a chitosan top-layer [86]. Asymmetric multi-tier membranes are also reported with increasing fiber diameter layers. Porosities over 70% are reported for the fibrous layers, and greatly improved flux and rejection are demonstrated over current nanofiltration membranes. In another study, nanofibrous PAN-polyamide membranes have been fabricated by interfacial polymerization [87].  Other recent studies consider the use of filled nanofibrous based membranes for fuel cell applications [88], [89]. These reports as a whole indicate that nanofibrous membranes are of research interest in a variety of fields and applications. However, no research has thus far been completed on selective water vapour transport membranes based on electrospun nanofibers, furthermore no reports indicate that nanofibrous membranes have been studied for building energy ventilation recovery applications. Research into the requirements, fabrication, and transport 17  properties of these nanofibrous membrane materials is the focus of Chapters 5 through 7 in the present dissertation. 1.7 Literature of Particular Relevance and Contributions of the Present Work Table 1-1: Literature of particular relevance to this dissertation.  Nanofibrous Membrane Selective Membrane Water Vapour Transport ERV Application Comments Gibson, et.al. [72], [73], [77], [80], [90]  X  X  Demonstrated water vapour flux in nanofibrous materials utilizing a dynamic water vapour transport apparatus. Also made contributions to heat and mass transport modeling in vapour transport applications. Zhang, et.al. [53], [57], [60], [91]  X X X Demonstrated water vapour flux in various materials for ERV applications. Also made significant contributions to heat and mass transport models for ERV applications. Metz and Wessling  [61], [92], [93]  X X  Demonstrated water vapour transport in various selective polymeric materials for gas drying applications using a resistance model. Koester and Wessling [70], [71]  X X X Studied the resistances in water vapour transport membranes and the effects of water vapour activity on membrane permeance. Chu et.al. [86], [87], [87], [94], [95] X X   Demonstrated first selective membranes based on fabricating thin coatings on nanofibrous substrates, used for oil/water separations and water filtration applications. This Work (2011-2016) X X X X Fabrication of selective water vapour permeable nanofibrous membranes for energy recovery ventilators. Novel materials have demonstrated performance benefits. Nanofibrous filled polymer film layer is studied as a route to membrane fabrication. Thermally formable membranes are demonstrated.  18  1.7.1 Knowledge Gaps Addressed in this Dissertation 1. PEO-PU polymer permeability and interactions with liquid water and water vapour. Although many polymers including other PEO-PU polymers have been studied, the properties and performance of the specific side chain PEO containing polymers (PERMAX™ 230) studied herein have not been reported in the literature. These polymers are of interest due to their high weight fraction of hydrophilic PEO groups, allowing them to have high water vapour permeability. Through the analysis of these polymers and cross-linking reactions for these polymers, we show how water and water vapour interact with these polymers, and determine the solubility, diffusivity, and permeability of water vapour in these polymers so that they can be used in membranes and membrane models for ERVs. 2. Water Vapour Transport in ERV Membranes Other researchers, notably Zhang and more recently Koester have studied the resistances associated with the various components of ERV membranes [68], [69], [71]. Several studies have evaluated resistance in composite membranes for VOC/gas separations [96], [97]. Others such as Metz, Zhang, and Kissler have studied the effects of boundary layers in the resistance to water vapour transport in air-to-air membrane transport [49], [56], [62]. However, no studies have addressed air-to-air water vapour transport in composite membranes in systematic method, confirming vapour transport resistance in test module boundary layer, in the substrates themselves, in the dense film layer, and then finally in assembled membranes. A transport analysis of composite membrane assemblies and an evaluation of these membranes via a resistance-in-series model is presented in this work. 3. Electrospun Nanofibrous Membranes for Water Vapour Transport Membranes made with electrospun nanofibers have been demonstrated for various applications including water filtration, most notably by Chu and Hsiao [86]. However, at the time this study began no literature reported electrospun nanofibrous membranes with a selective barrier layer for water vapour transport, and specifically no research has been reported on the use of these membranes for ERV applications. Based on our study of composite membranes, we hypothesize and then demonstrate that eliminating the substrate layer used in composite ERV membranes could improve the performance of ERV membranes. We then hypothesize that the use of an electrospun nanofibrous layer to carry the dense selective layer could be one approach to eliminating the substrate layer from ERV membranes.  19  4. Transport in Membranes Based on Impermeable Fibers in a Water Vapour Permeable Matrix No research in the literature addresses the resistances to transport associated with the nanofibrous layers within electrospun nanofibrous membranes. This work analyzes the contribution of this layer to transport resistance. Many studies present transport models for ‘mixed matrix membranes’ with particles in the dense matrix layer [98]. The literature for fiber-reinforced polymer composites also presents many models for diffusion of liquid water, and water vapour in fibrous composite materials [99], [100]. However, no studies have addressed a water vapour transport membrane containing an impermeable nanofibrous layer with a permeable polymer matrix. A transport model for this type of membrane is addressed in our studies, and presents the first step the confirmation and validation of a ‘impermeable fiber-filled-film membrane model’.  5. Thermally Formable Membranes for ERV Applications Many studies have confirmed performance of heat and moisture exchangers of various geometry for building energy recovery, however no studies have considered the use of formable membrane exchanger plates. Furthermore, no studies have reported the development of membranes for ERV’s which are thermally formable. Our work presents for the first time a method to fabricate membranes for ERVs which are thermally formable. With further material improvements and exchanger level validation building on this work, the development of advanced high efficiency ‘formable’ membrane based exchangers can be achieved. 1.8 Research Objectives This dissertation is a contribution to the field of polymeric membrane science, specifically it addresses the fabrication of novel water vapour selective membranes for energy recovery ventilation applications. The overarching purpose of this work is to improve the water vapour transport of these membranes so that building ventilation can be more efficient. The central hypothesis of this work is that through the use of electrospun nanofibers, ERV membranes structures can be made which eliminate the resistance to water vapour transport associated with the microporous substrate layer in composite membranes for energy recovery ventilation. If this is achieved, this will demonstrate a route to ERV membranes with improved water transport performance, allowing for improvements in ERV exchanger efficiency, and the improved overall efficiency of building ventilation.  20  This dissertation proposes that fabricating membranes based on electrospun nanofibers can lead to membranes with improved water vapour transport and improved forming functionality over current composite membrane approaches. In order to do so a greater understanding of the polymers and substrates used in composite membranes is required and this is the initial focus of the dissertation (Chapters 3 and 4). Once the permeability performance of the dense polymer used in the film layer of the membranes was understood and the limitations of the composite membrane approach were demonstrated, a membrane fabrication approach using electrospun nanofibers was proposed. It was hypothesized that the high resistance substrate layer used in the composite membranes could be eliminated by using an electrospun nanofibrous layer impregnated with a dense water vapour permeable polymer (Chapter 5). However, it was found that since the nanofibers used were impermeable to water vapour, an additional transport resistance was introduced which was associated with the nanofibers themselves. In order to achieve the performance target for the membranes (>10000 GPU water vapour permeance and <100 GPU oxygen permeance), this resistance had to be minimized by optimizing the nanofiber geometry and impregnated nanofibrous film layer such that a thin but defect-free selective layer was fabricated (Chapter 6). Finally, these membranes were demonstrated to be thermally formable, which allowed the membranes to be formed into exchanger plates for ERV exchangers (Chapter 7). This would allow the development of advanced higher performance heat and mass exchangers from these membranes. With this framework in mind a number of research objectives are addressed in the dissertation: 1. Analyze the chemical, thermal, and transport properties of a specific polyethylene oxide-polyurethane (PEO-PU) copolymer to be used throughout the study. The interactions of water and water vapour with this PEO-PU polymer with varying degrees of cross-linking must also be considered in this analysis (Chapter 3). Understanding the function and transport performance of these polymeric materials is necessary to interpret the performance of membranes made from these materials in Chapters 4 through 6. 2. Study current approaches to fabricating and testing composite membranes for energy recovery ventilation applications, and demonstrate how resistances to water vapour transport in these membranes can be decreased through the elimination of the microporous substrate (Chapter 4). One hypothesized method to eliminate the microporous substrate resistance in the membranes, is to use a nanofibrous layer to carry the selective polymer film layer, and proving the effectiveness of this approach is the focus of Chapters 5 and 6. 21  3. Fabricate novel nanofibrous membranes for water vapour transport applications and evaluate the performance of these materials based on water vapour and gas transport measurements (Chapter 5). 4. Optimize performance and manufacture of these nanofibrous membranes for gas-to-gas water vapour transport applications by studying the effects of nanofiber geometry and deposition, and the nanofibrous layer’s interaction with coating polymers. Present a model for the performance of these fiber filled membranes, and demonstrate the feasibility of these membranes for the application by proving performance for water vapour permeance >10000 GPU, oxygen permeance <100 GPU and selectivity of water vapour over oxygen >100 (Chapter 6). 5. Demonstrate novel formable membranes based on the electrospun nanofibrous membranes that can be thermally-formed into three dimensional structures such as exchanger plates for high performance heat and mass exchangers (Chapter 7).      22   Experimental 2.1 Chemicals and Materials  Polyacrylonitrile (PAN) was obtained from Scientific Polymer Products (#134), with an average molecular weight of 150 000 g/mol. Reagent grade solvents dimethylformamide (DMF), ethanol, acetone, hexane, and other solvents and chemicals (Triton 100X, PEG400, lithium chloride) were obtained from Sigma Aldrich. Polyether-polyurethane (PEO-PU) dispersion was obtained from Lubrizol under the trade name Permax® 230, with a polymer concentration of 33% by weight in water. Polycarbodiimide crosslinker solution was obtained from Stahl under the trade name Picassian® XL-702, with a concentration of 40% by weight in water. Deionized water was obtained from a SpectraPure Maxcap UHE 100 water filtration system. Spunbond polyester non-woven was obtained from Fiberweb under the trade name REEMAY®, with basis weights from 14 to 71 g/m2.. Spunbond polyester non-woven was also obtained from Asahi Kasei under the tradename Smash™ 15100. 2.2 Polymer Coating Formulations Polyether-polyurethane dispersions were created by mixing Permax® 230 and XL-702 crosslinker at the reported weight ratios and dispersions ranging from 5 to 33% total solids content were created by diluting the polymer dispersions with de-ionized water. 2.3 Fabrication of Nanofibrous Substrates PAN solution was created by dissolving PAN polymer in DMF at 70°C under stirring for 20 hours.  A Katotech Belt Electrospinning unit was used to create nanofibers. To create the fibers the PAN/DMF solution was placed in a 10 mL syringe with an eighteen-gauge hypodermic needle, 1.5” length without bevel. The syringe was mounted 15 cm from a rotating grounded collector onto which the non-woven support was affixed. An electrode was placed on the needle tip. A 15.7 kV potential was applied to the needle tip causing polymer to jet from the syringe towards the collector. A syringe pump containing one to three syringes was used to maintain a constant supply of polymer solution to the needle tip, at a rate of about 0.105 mL/min per syringe. Polyester non-woven support was cut to size and attached to a grounded collector target which was a 36 cm wide belt, rotated at a speed of 1.5 m/min. Fibers were electrospun and deposited as random fiber mat onto one surface of the non-woven support layer. 23  2.4 Fabrication of Selective Nanofibrous Membranes Samples of the nanofiber-covered non-wovens were cut to a known area and weighed. A coating bath of PEO-PU polymer solution was used to dip coated the electrospun nanofibrous membrane, fiber side down. The coating was readily wicked by the nanofibrous layer visibly wetting the nanofibers with coating solution. Excess impregnation solution was removed with an absorptive tissue, and the samples were dried in an oven at 50°C to remove the aqueous solvent and cure the membrane. This coating procedure led to membrane assemblies (a fiber filled film layer attached to one surface of a non-woven support), which were bound together with the selective water vapour permeable PEO-PU coating. Impregnated samples were weighed after coating and drying.  2.5 Polymer Film Fabrication Polymer films (15 to 100 microns) were cast on a clean glass surface using a gap applicator (RK PrintCoat Instruments) on a linear actuator.  The coating solution containing a known solids content was cast at a controlled wet thickness on the glass surface and then dried in an oven at 50°C. Samples of polymer were obtained by evaporating the solvent from polymer formulations in petri dishes on a level surface.   2.6 Composite Membrane Fabrication The substrate was attached to a glass or rubberized surface on a coating table (RK PrintCoat Instruments), with a metering rod or gap applicator attached on a linear actuator on the coater. A pipette of polymer solution was deposited on the substrate and the coating device was used to spread the polymeric solution over the substrate surface at a predetermined wet coating loading. Wet coating control is achieved by the gap distance between the substrate and blade coating head in the gap applicator, and by the depth of the etched grooves in the coating rods. Membranes were dried in an oven at 50oC to remove solvent. Other methods of membrane fabrication are discussed in Chapter 6. 2.7 Basis Weight, Density, and Thickness Measurements Substrate samples were weighed dry, msubstrate and measured for area, A before coating and then the membrane weight, mmembrane was measured after coating, and the basis weight (coating weight), ωcoating was determined in g/m2:  substratesubstratemembranemembranecoatingAmAm  [2-1] 24  Thickness of films and substrates was measured by micrometer with an average of at least 20 measurements over the sample area. Density of films was measured by weighing samples of a known area and thickness. Density of films was also measured by water and hexane displacement in a pycnometer.  2.8 Nanofiber Basis Weight Measurements The basis weight of the nanofibrous scaffolds were measured by cutting three 10x10 cm samples from the mats. The fibers were removed from the non-woven substrate and weighed; basis weight of the nanofibrous layer was calculated by dividing the nanofiber mass by the area. The impregnation basis weight was determined by cutting the non-woven coated with nanofiber to a known area and weighing the sample before impregnation and after drying of the impregnated nanofibrous membranes. The impregnation basis weight was determined by the difference in dry mass before and after impregnation and dividing this mass by the area of the sample. 2.9 Swell Ratio Samples were weighed dry and then placed in a 20 mL vials with de-ionized water and held at a constant temperature using a temperature controlled water bath. Samples were removed from the vials at a measured time, quickly patted dry twice with a lint-free delicate task wipe, and then weighed. The swell ratio was measured as: 100initialinitialwetUptakemmmW  [2-2]  2.10 Optical and Polarized Light Microscopy Optical and polarized light microscopy images were captured using a Nikon Eclipse LV100 Optical Microscope. 2.11 Electron Microscopy Samples were sputter coated with a thin gold layer and placed in a Hitachi Model S3000N VPSEM for surface analysis, high vacuum mode was used and a secondary electron detector was used for imaging. Scanning electron microscope images of the electrospun fiber substrates were used to determine the average fiber diameter using ImageJ, from an average of 50 measurements at each of three locations on each the fibrous mats. Membrane cross-25  sections were prepared by freeze fracture and cross-sectional images were used to measure the thickness of the samples. Thickness measurements were taken by freeze fracturing samples from three areas of each membrane, and measuring at least 25 measurements of cross-sectional images from each sample (totaling a minimum of 75 measurements per membrane). 2.12 Laser Confocal Microscopy Laser confocal microscopy images were captured using Olympus LEXT 4000 Confocal Microscope, for determining average surface roughness, and maximum peak to valley distances of microporous substrate materials.  2.13 Mercury Intrusion Porosimetry Mercury intrusion porosimetry analysis was completed using an instrument from Porous Materials Inc. (PMI), this was used to measure the porosity, pore size distribution, and average pore size of microporous substrates. 2.14 Contact Angle Goniometry Contact angles on membranes and substrate were measured by placing a controlled volume liquid water droplet on the surface of interest and capturing an image of the droplet on the surface. Images after the droplet had time to stabilize on the membrane or substrate surface. Several images were taken over one minute to ensure that the droplet had stabilized.  2.15 Oxygen Crossover and Permeance The oxygen transport provides an indication of the overall selectivity of the membrane material, specifically the level of defects in the dense coating layer. Oxygen crossover is measured in a custom module operating in counter-flow. The module is designed for testing permeation properties of membrane materials and consists of flow pathways to direct the flow of gas over the surfaces of either side of the membrane. Further details of the module can be found in Chapter 4 and in previously published works [101], [102].  In the oxygen/nitrogen test permeance test, one side of the membrane a nitrogen sweep stream (>99.999%) flows over the membrane surface. A feed stream composed of dry, clean air flows over the opposing membrane surface in counter-flow. The module outlets on both streams are open to atmosphere, such that differential pressure between the two streams is minimized and only diffusive transport and not convective transport occurs through the membrane. An oxygen sensor (Gas Badge) measures the oxygen concentration on the feed 26  outlet from which oxygen permeance and percentage crossover are calculated. Since the concentration of oxygen in dry supply air is known, and the nitrogen stream contains no oxygen at the inlet, the percentage of oxygen passing through the membrane by diffusion can be reported as:  %100)(3,2,222 OOCCOCrossover  [2-3] where C refers to the percent concentration of oxygen (O2) at points 2 and 3, in a counter-flow test module, with point 2 being at the nitrogen-side outlet (measured by the sensor), and point 3 (see Figure 2-2) being at the air-side inlet (20.95%). This test is completed at a series of flow rates and the oxygen flux is calculated from these results:  RTAVpQJmOO3,222  [2-4] Subsequently permeance is calculated using the partial pressure difference of oxygen across the membrane:  2,23,222OOOOppJP  [2-5] 2.16 Pressurized Air Crossover Membrane samples (18x5cm) were placed in a test module with an active area of 45.6 cm2. A constant static air pressure is maintained on one side of the membrane; the other side of the membrane is at ambient pressure. The air flow through the membrane is measured with a flow meter on the low pressure side of the membrane. Air flow rates through the membrane are measured at a range of the pressures, but are reported at 20.7 kPa (3 psi) upstream pressure. For porous materials this test determines the permeance of air through the material and is related to the porosity, pore size, thickness, and tortuosity of the porous material. For membrane materials with a dense selective polymer layer, this test determines the extent of ‘through-pores’ or defects in the polymer film layer of the membrane. 27  2.17 Gas Permeability Gas permeability measurements were completed in a MOCON OpTech-O2 Platinum oxygen transmission equipment for oxygen transport rate, and in a MOCON Permatran C-441 for carbon dioxide.  2.18 Water Vapour Flux, Permeance, and Permeability  A dynamic water vapour transport testing station was designed and built to test the membranes under a range of operating conditions, see Figure 2-1. The dynamic water vapour transport rate (WVTR) testing procedure was developed which was designed to test the membranes under conditions which are similar to those in which they might be utilized.     Figure 2-1: Water vapour transport test station setup. This test apparatus is similar to that described as a dynamic moisture permeation test by Gibson et. al. and also summarized in ASTM F2298 [77], [103]. A membrane sample was sealed in a test apparatus with flow field pathways on both sides of the membrane to evenly distribute gases over the both surfaces of the sample, the gas streams being separated by the membrane. Details of the test module geometry can be found in Chapter 4, section 4.2 through 4.4.   The flow rate, temperature, and relative humidity of each inlet gas stream could be controlled, and the outlet temperatures and relative humidity of each gas stream could be measured via a Labview© interface. The flow rates are controlled by mass flow controllers (MFCs) (Bronkhorst EL-FLOW), temperature is controlled using temperature controllers (OMEGA CN7800), and measured using thermocouples (K-type). Relative humidity was measured by 28  Vaisala probes (HMT 221). Humidity was controlled via a temperature controlled water bath loop passing through a plate and frame membrane humidifier (dPoint Technologies Px1) and humidity is controlled by proportioning the ratio of the humidifier outlet stream (saturated) with dry air using MFCs. The feed and sweep gases were supplied and directed in counter-flow over the opposing surfaces of the membrane (see Figure 2-2). The membrane active area in the test jig was 33 cm2, and the length of the flow pathway was 16 cm. All gas lines are heat traced into a temperature controlled oven, where the module and humidity probes are located, such that all measurements are isothermal and that no condensation may occur within the membrane module or on the humidity probes. During a ‘standard’ vapour transport test a first gas stream (S1) was supplied at 50°C and 0% relative humidity (Pv,H2O,1=0 Pa) to the inlet on one side on the membrane. A second gas stream (S3) was supplied to the inlet on the other side of the membrane at 50°C and ~50% relative humidity (Pv,H2O,3=6122 Pa), and at the same flow rate as the first gas stream (Figure 2-2). For tests where the temperature, flow, and RH were not at the ‘standard’ test conditions (feed stream at 50°C and 50% RH), testing conditions are noted and the temperature, flow, and humidity of the Stream 1 and Stream 3 inlets are reported.   Figure 2-2: Counter-flow water vapour permeance test orientation. The water content and temperature of the two streams were measured and recorded at the outlets, using calibrated Vaisala humidity probes (HMT 221).  From these values, a mass balance could be completed on the system to determine the mass flow of water vapour through the membrane active area, which can be used to determine the water vapour flux of the material. Dividing the flux by the average water vapour partial pressure difference across the membrane in the module allowed water vapour permeance to be measured.  For a film of 29  known thickness, the product of the actual membrane permeance and the thickness gives the permeability. Further details on the water vapour transport test, measurements, and analysis can be found in Chapter 4, section 4.2 through 4.4. 2.19 Permeance and Permeability Units and Selectivity Permeance is the chemical potential difference normalized flux through the membrane, in the case of water vapour transport, the flux is normalized by the water vapour pressure difference across the membrane. Permeability used to describe transport in homogeneous polymer films and is the permeance normalized by the film thickness and is the product of the permeance and the film thickness (δ). In the synthetic membrane science literature, membranologists have adopted units for permeance (GPU units) and permeability (Barrer units) that are more convenient to work with than SI units on the scale of typical membrane measurements. Using these units, a 1 micron thick film having a permeability of 1 Barrer would have a permeance of 1 GPU [104]. For permeance, the gas permeance unit (GPU) is used, where:  1 GPU = 1×10-6 cm3 (STP) cm-2 s-1 cmHg-1 = 3.348×10-10 mol m-2 s-1 Pa-1 For permeability, the Barrer unit is used, where: 1 Barrer = 1×10-10 cm3 (STP)•cm cm-2 s-1 cmHg-1 = 3.347×10-16 mol•m m-2 s-1 Pa-1. Selectivity refers to the relative permeance or permeability of two chemical species through a membrane, where the species of interest is in the numerator. Selectivity for water vapour transport membranes is usually determined by measuring the water vapour permeance and the permeance of another gas that it would be desirably to separate the water vapour from, such as water vapour over oxygen:   222222OOHOOHOOHPPPP  [2-6] For example, a membrane with 10000 GPU water vapour permeance and 100 GPU oxygen permeance has a selectivity for water vapour over oxygen of 100. If for example the oxygen permeance must not exceed a certain level (ie. 100 GPU), then the required selectivity must be increased as the water vapour permeability is increased. 2.20 Water Vapour Sorption (DVS) Dynamic water vapour sorption and sorption isotherms were completed using a Quantichrome Aquadyne DVS instrument. Film samples were preconditioned as cast at the analysis temperature under a dry air stream. Relative humidity was increased in a stepwise manner in 30  increments of 5% from 0 to 90% relativity humidity at the test temperature. Gravimetric uptake was measured over time, and the total vapour uptake at the set condition was measured when the sample came to equilibrium. Desorption measurements were then completed by ramping back down to 0% relative humidity in stepwise manner. Sorption and desorption isotherms were reported as the mass uptake of water vapour at equilibrium. 2.21 Fourier Transform Infrared Spectroscopy (FTIR) Samples were measured using an attenuated total reflectance (ATR) attachment on a Perkin Elmer Spectrum 100 FTIR spectrometer. Samples were the average of 32 scans between 3500 and 450 cm-1. FTIR is used to analyze the chemical composition of the polymers used in this work to study cross-linking reactions, and to determine the presence of PEO crystallinity in the polymer film layers in composite membranes. 2.22 Nuclear Magnetic Resonance (NMR) Polymer samples were dissolved in deuterated dimethyl sulfoxide (DMSO-d6) in Norell® Select Series™ 5 mm NMR tubes. 1H (proton) NMR spectra were measure using Bruker 400 MHz spectrometer. NMR is used to determine the chemical composition of the PEO-PU polymers used in the study. 2.23 Gel Permeation Chromatography (GPC) A TDA Model 302 Tetra Detector System by Viscotek was used for GPC. A PMMA standard with a narrow molecular weight range was used for calibration (64,868 Da) and a PMMA standard with a broader range (92,262 Da) was used as a reference standard. Samples were dissolved in HPLC grade dimethylformamide with 0.1 M LiBr, and filtered after 24 hour with a 0.2μm PTFE filter. Concentration was ~3 mg/mL, injection volume at 100 μL, flow rate was 1.0 mL/min, column temperature was 50°C, the column consisted of two Jordi Gel™ mixed bed cross-linked divinylbenzene (DVB) columns (25cm x 10mm). GPC is used to determine the average molecular weight of the PEO-PU polymers used in the study. 2.24 Differential Scanning Calorimetry (DSC) Samples of conditioned polymer film (5-15 mg) were cut into discs and placed in hermetically sealed aluminum DSC pans. All preconditioning procedures, temperature ramp rates, and thermal history are summarized for any results reported. All thermal scans were completed in a TA Instruments Q1000 Tzero™ DSC equipped with a refrigerated cooling system, under 50 mL/min nitrogen flow, using an empty hermetically sealed aluminum pan as a reference. The DSC was calibrated against a sapphire disc for heat capacity and iridium for heat flow. DSC 31  is used to measure thermal properties (glass transitions, melting) of the polymers and non-woven carriers used in the study. DSC is also used to study the interactions of water with the polymers used in the study and the state of water in these polymers through the measurement of freezing and melting of water in the polymers. 2.25 Thermogravimetric Analysis (TGA) Thermogravimetric measurements were completed with a TA Instruments Q500 TGA under either a nitrogen or air environment. Temperature ramp rates and gas flow rates are reported for individual results. Samples (~10 mg) were placed in a platinum sample pan for analysis. TGA is used to determine the thermal stability of polymers used in the study and the thermal degradation of these polymeric materials in air and nitrogen environments. 2.26 Dynamic Mechanical Analysis (DMA) Dynamic mechanical analysis was completed with a TA Instruments Q800 DMA. Temperature ramp rates, sample dimensions, frequencies, and amplitudes are reported for individual results. DMA is used to determine the thermal-mechanical properties of non-wovens, polymers, and nanofiber filled polymer films. DMA is used to determine the glass transition temperature, Tg under dynamic conditions in tension and provides the required temperatures to thermally form the materials tested. 2.27 Mechanical Analysis Samples were tested on a Katotech KES-G1 Tensile Tester. Sample dimensions, strain rates, and temperatures are noted for any reported data.  32   Polyethylene Oxide-Polyurethane Copolymers for Water Vapour Transport Membranes 3.1 Introduction Energy recovery ventilators (ERVs) or enthalpy exchangers are used to transport heat and moisture vapour between incoming (fresh) and exhaust (stale) air stream in building ventilation systems. Central to the function of these air-to-air heat and mass exchangers, is a semi-permeable membrane which selectively transports water vapour over other chemical species, vapours, and gases. Membranes for these device should have a number of desirable properties: • High water vapour permeance • Low permeance of gases (O2, CO2), selectivity of water vapour over gases • Low cross leakage under pressure • Antimicrobial function • Low flammability • Durability under humidity cycling and freeze cycling • Relatively low cost  Current generation membranes use a microporous substrate to support a thin, dense polymer film layer [26]. The present work studies a type of polyethyleneoxide-polyurethane block co-polymer for water vapour transport membranes. This polymer is commercially available as PERMAX™ and contains ‘soft’ blocks of polyethylene oxide (PEO) in the main chain polymer backbone and PEO in side chains branching from the backbone. These polymers are believed to be useful for the energy recovery ventilation application because: 1. They have high water vapour sorption and permeability due to the large number of hydrophilic PEO groups. 2. They have relativity low transport of gases (good selectivity) due to the low solubility of permeant gases in the amorphous PEO segments. 3. They can be cast as membrane dense film layers from aqueous dispersions which lowers manufacturing costs over solvent based coatings. 4. They have good hydrolytic stability under humidity cycling and freezing conditions. 33  Since aqueous dispersions are preferred from a manufacturing perspective, the polymers must be cross-linked on drying in order to make them stable in conditions where they will be exposed to high humidity and liquid water condensate.  The present work reports the formation of polycarbodiimide cross-linked PEO-PU polymers under a range of crosslinking ratios and studies the chemical nature of these polymeric materials. The interactions of liquid water and water vapour with these copolymers are reported, as well as the thermal properties of these polymeric materials. Finally, the water vapour diffusivity, sorption, and permeability properties of films made from these polymers are determined. These polymers are the functional selective water vapour permeable layer used in all the membranes in this dissertation. Through water vapour sorption and permeability measurement techniques, we determine the water vapour permeability and the solubility and diffusivity contributions to permeability in cross-linked PEO-PU polymers. This is critical to determining the membrane component resistances in Chapter 4, and for the modeling and optimization of impregnated electrospun nanofibrous membranes in Chapters 5 and 6. Knowledge of the transport properties and interactions of liquid water and water vapour with these PERMAX™ 230 based polymers have not previously been reported in the literature. 3.2 PEO-PU Polymers Thermoplastic polyurethanes (TPU) are a commercially available family of polymers, which are highly customizable to offer a wide variety of end properties. These polymers are widely in applications where toughness, durability, and broad temperature flexibility are required, for example in footwear, wire and cable, and automotive applications as well as protective coatings and films [105]. The general reactions for production of these polyurethanes involves reacting a diisocyanates with short chain diols (‘chain extenders’) and long chain polyols. The reaction of an isocyanate with an alcohol group produces an ethyl carbamate or “urethane” linkage, and through a step growth addition polymerization process, linear block copolymers can be created. These block copolymers contain ‘hard blocks’ of isocyanate groups connected by the short chain diols, and ‘soft blocks’ based on long chain diols. This chemistry supports the use of broad range of monomeric building blocks which can be incorporated in the final polymer allowing chemists to tailor the polyurethane functionality and properties. TPUs are often classified by the monomeric building blocks used in their fabrication, common categorizations include aliphatic or aromatic diisocyanates, and polyether or polyester diols.  34  Polyurethane dispersions (PUD) are a subgroup of TPUs in which polymer particles are dispersed in an aqueous system. Aqueous dispersions are attractive for polymer coatings and membrane fabrication compared to solvent-based systems which have significant environmental, economic, and health implications associated with evaporating and exhausting solvents during the drying process. In formulating PUD the diisocyanate and diols are initially reacted to create a ‘pre-polymer’ which is then mixed in water in a secondary step where it is reacted with a ‘chain extender’ to increase the polymer molecular weight, a dispersion of polymer particles in water remains at the end of the reaction [106].  Since the polymer is in a dispersion of water it may require crosslinking once cast into a film in order to improve chemical stability and decrease solubility in water. Polyurethanes can be synthesized to contain a number of active functional groups on which a cross-linking reaction can be based such as amine, hydroxyl, and carboxyl groups. Common cross-linkers for polyurethanes include isocyanates which will react with hydroxyl and amide functional groups, and carbodiimides and aziridines which will react with carboxyl functional groups. Aqueous polycarbodiimide dispersions are commercially available for crosslinking aqueous polyurethane dispersions at room temperature conditions. Crosslinking leads to a large increase in the polyurethane molecular weight, rendering the dried coating layer or polymer film insoluble.  The present work studies a specific polyurethane dispersion, PERMAX™ 230 available commercially from Lubrizol Advanced Materials. The manufacturer describes PERMAX™ 230 as an “aliphatic polyether waterborne polyurethane dispersion”. PERMAX™ dispersions are reported in the patent literature as PERMAX™ 120, 200, 202, 220, 230, 232, and 300; however with over 40 patents mentioning uses of products, minimal specific information about the polymers themselves are reported. Table 3-1 summarizes the properties of presently available formulations of PERMAX™ from the manufacturer. A comprehensive literature search for these polymers and their associated CAS numbers only shows three non-patent references, one is an earlier work of the present authors [107], another reports the use of PERMAX™ 200 as a highly non-conductive polyurethane carrier for conductive graphite and single walled carbon nanotubes (SWCNT) [108], and the final is a technical paper by the inventors of the PERMAX™ dispersions discussing the development and properties of the PERMAX™ polymers [109]. This final publication when considered with the granted US patent 6,897,281 by the same authors reveals significant details about the molecular structure of the polymer [110]. The polyurethane polymer contains between 12 and 80% poly(alkylene oxide) 35  side chains, in which the preferred side chain monomer is incorporated into the polyurethane backbone by using the polyol Tegomer® D-3404. This polyol is a polyether-1,3-diol available from Evonik Industries with an average molecular weight of 1200 g/mol, shown in Figure 3-1, where m=23.   Figure 3-1: Structure of Tegomer® D-3404 (Evonik). According to the PERMAX™ inventors, the use of these side chain units allows the incorporation of a large amount of hydrophilic PEO into the polymer while still allowing aqueous dispersions to be made. The main chain unit is also stated to contain poly(alkylene oxide) in amounts less than 25% of the total weight which are incorporated through the use of a polyether polyol. The fabrication process involves reacting a diisocyanate (the preferred is aliphatic diisocyanate appears to be IPDI) with the poly(alkylene oxide) side-chain diol and the poly(alkylene oxide) main chain diol in the presence of heat and a catalyst to generate a isocyanate terminated pre-polymer. Preferred embodiments in the invention disclosure include at least one compound having at least one cross-linkable functional group, dimethylolpropanoic acid (DMPA) is most preferred which would add carboxyl functional groups to the polyurethane. The pre-polymer is then dispersed in water and reacted with a chain extender (the preferred embodiment uses hydrazine) to create the final polyurethane dispersion.  36   Figure 3-2: Diisocyanate (IPDI) reaction with polyols (PEG, Tegomer©) to generate a polyether-polyurethane as described in the PERMAX™ patent literature. Another patent by the same inventors discusses “PERMAX™ 230 type prepolymers” and provides formulations for these pre-polymers which contain Tegomer D-3403 (side chain PEO), PEG 1450 (main chain PEO), and IPDI (diisocyanate) [111]. These formulations contain additional methoxy-PEG 750 to ‘end cap’ the pre-polymer and plasticizers for blending with PVC in the invention, however a clear picture of the block units used is PERMAX™ 230 can be deduced from this formulation as well as examples in references [109] and  [110] and Table 3-1 that the high water transport PERMAX™ 230, contain side chain PEO in the range of 30% and total soft segment in the range of 75%. On a molar basis there are ~1.7 to 2 parts diisocyanate hard groups to soft groups on a molar basis, and generally the ratio of side chain to main chain PEO is 0.75 to 1.5 on a molar basis. The general reactions and structure of these polymers as described in the patent literature can be represented by Figure 3-2. Other block copolymers with PEO soft segments have been more widely studied in the literature for water vapour transport, for example PEBAX®, PEO-PBT, polyether-polyamides, and other PEO-PU have been reported [112]–[116]. However, no reports in the literature specifically study these PERMAX™ polymers containing PEO side chains. 37  Table 3-1: Properties of commercially available PERMAX™ dispersions from Lubrizol Advanced Materials. Polymer PERMAX™ 200 PERMAX™ 230 PERMAX™ 300 CAS 797803-20-4 1082849-09-9 1370699-70-9 Type Aliphatic polyether polyurethane Aliphatic polyether polyurethane Aromatic polyurethane Viscosity (cP) 200 600 500 Solids Content (%) 44 33 42 pH 6.0 5.6 7.0 Surface Tension (mN/m) 55 55 49 Upright WVTR, ASTM E96 (g/m2/day)a 500 800 450 Permeability (Barrer)b 23400 37400 21000 a21oC, 50% RH external conditions [117], 33gsm film  bCalculated from ASTM E96 test conditions reported, (1 Barrer = 1×10-10 cm3 (STP)•cm cm-2 s-1 cmHg-1). 3.3 NMR Samples of the PEO-PU polymer were vacuum dried and dissolved in deuterated DMSO-d6 and analyzed using proton NMR. The general structure of the PERMAX™ polymers is labelled with numbers to identify the hydrogens in Figure 3-3.  38  Figure 3-3: Atom labels for the general structure of PERMAX™ polymers for NMR analysis. A proton NMR of the polymer is sample is shown in Figure 3-4, a large water peak is found in the spectrum at 3.35 ppm and could not be removed with excess drying, this makes the NMR analysis difficult due the importance of the peaks near the water peak in the analysis. Furthermore, it is unknown what excess processing additives are in the polymer, so the NMR analysis is only an estimate of the polymer structure. The methoxy group at the end of the side chain of the copolymer should have distinct singlet at 3.2 ppm (location 30 in Figure 3-3). This should integrate to 3 hydrogens for the total structure of the copolymer. Using this peak as an integration reference, the total number of protons in the polymer molecular unit can be estimated. All other protons in the spectra overlap in the 0.5-2 ppm range and the 3.4-4.0 range. A large peak at 3.5 arises from the polyethylene oxide groups (CH2-CH2-O) in the polymer.   Figure 3-4: 1H NMR of the PEO-PU polymer. Integration for the rest of the peaks in the polymer indicates that there are about 154 protons per molecular unit of polymer. By calculating the total number of protons expected from the 39  side-chain monomer (~110), and those from each urethane (IPDI) segment (20), it can be determined that there must be nearly no main chain PEO in the polymer, as nearly all of the PEO protons at ~3.5ppm are accounted for in side chain. This is supported by the lack of a peak at 4.25 ppm which would be expected from the alkyl-urethane protons where the main chain PEO would meet the urethane group. This means that the hydrophilic PEO groups are branched from the main chain of the polymer and would be expected to have significant mobility, compared to PEO in the main chain which would be more constrained by the PU hard segments which would phase separate and crystalize at application conditions. The excess protons would indicate that about 2 urethane groups are present units per every side-chain PEO group, likely from chain extension via hydrazine. This is supported by the descriptions of the polymer molar formulations in the patent literature. The polymerization reaction would be expected to be random, and segments of the polymer would contain repeating PU groups making ‘hard segments’. If the above assumptions are true, the polymer would be expected to have the structure in Figure 3-5, where x=0.6, y=0.4, and m=23. This formulation is only an estimate for the purposes of interpreting the performance of these polymers, as the total additives to the dispersion, and exact formulas for the polymer remain unknown and are not disclosed by the manufacturer. On a mass basis the total PEO in the polymer from this analysis is in the range of 66 to 72%.   Figure 3-5: Predicted structure of PEO-PU polymer.   40  Table 3-2: Assignment of protons in the 1H NMR of the PEO-PU polymer. Segment Description Location (Figure 3-3) Predicted 1H NMR Shift (ppm) Total Hydrogens in polymer unit Estimated from NMR Side Chain CH2-CH3 25 0.96 3 3 PU-ring C-CH3 5,6,17,18 1.11 12 15 PU-ring CH2 in ring 7,19 1.15 4 5 PU-ring C-CH3 8,20 1.15 6 7.5 Side Chain CH2-CH3 24 1.25 2 2 PU-ring CH-CH2-C in ring 3,4, 15, 16 1.71 8 10 PU C-CH2-NH  9,21 3.01 4 4 Side Chain -O-CH3 30 3.24 3 3 Side Chain C-CH2-O 27 3.41 2 2 Main Chain PEO O-CH2-CH2-O 11,12 3.54 (n-2)*4 0 PU-ring NH-CH-CH2 2,14 3.54 2 2.5 Side Chain O-CH2-CH2-O 28,29 3.54 m*4=96 96 Side Chain urethane-CH2 23,26 3.88 4 4 Main Chain PEO urethane-CH2-CH2- 11,12 4.25 4 0     Total 154 3.4 Molecular Weight of the PEO-PU Polymer Two samples of the PEO-PU polymer were cast from dispersion and dried before dissolving in DMF for gel permeation chromatography (GPC). The molecular weight data for the two samples run on GPC are summarized in Table 3-3, the number average molecular weight (Mn) is 228,868 Da, the weight average molecular weight (Mw) is 829,217 Da, and the Z average molecular weight (Mz) is 2,538,000, the polydispersity index is 3.623. This shows the polymer covers a wide range of molecular weights, which can be observed in Figure 3-6 where 41  the normalized weight fraction of polymer is plotted against retention volume with the corresponding molecular weight. Table 3-3: Molecular weight summary for two samples of PEO-PU by GPC.  Mn (Da) Mw (Da) Mz (Da) Mw / Mn Sample 1 229,404 818,574 2,607,000 3.568 Sample 2 228,332 839,859 2,469,000 3.678 Average 228,868 829,217 2,538,000 3.623   Figure 3-6: GPC data for the PEO-PU polymer in DMF.  3.5 Crosslinking of PERMAX™ 230 As mentioned above, the preferred embodiments of the PERMAX™ 230 polymers contain at least one compound which incorporates a cross-linkable functional group into the polymer and the preferred compound is DMPA, which would add carboxyl functional groups to the polymer chain [110].  The general carbodiimide reaction is summarized in Figure 3-7. 42  Polycarbodiimides contain numerous –N=C=N– groups within each cross-linker molecule allowing multiple crosslinking reactions to occur by the same molecule at various carboxyl groups throughout the polyurethane polymer.  Figure 3-7: General carbo-diimide crosslinking scheme with a carboxyl functionalized polymer. Carbodiimide crosslinking is a common approach for the carboxylic acid containing polyurethanes [118]. One commercially available cross-linker for aqueous polyurethane systems is an aqueous dispersion of polycarbodiimide from Stahl, Picassian® XL-702. The molecular weight and chemical structure of the cross-linker are not explicitly reported by the manufacturer. The manufacturer describes the reaction associated with this system as occurring on polymer drying as the pH drops and carboxyl groups in the TPU become activated. This reaction will greatly increase the molecular weight of the polymer, and the swell in water and solubility of the polymer in water should decrease.  A series of PEO-PU polymer-cross-linker formulations were mixed and then cast in a petri dish, they were left to dry at room temperature over several days. The films had a thickness in the range of 200 to 500 microns. 3.5.1 FTIR of Cross-linked Films Studies from Stahl International are reported in the literature and discuss the expected crosslinking reactions of the polycarbodiimides with carboxyl groups in polyurethane polymers: carbodiimide groups in the cross-linker react with a carboxyl groups in the urethane leading to an N-acyl urea bond [119]. Acidic conditions promote the reaction as the carboxyl group must be protonated for the reaction to proceed. This is beneficial as the pot life will generally be extended in the relatively neutral aqueous dispersions. After coating, during the drying process, the reaction will proceed as water evaporates from the dispersion and the carboxyl groups become more acidic and protonated.  FTIR scans were taken of each film sample long after drying and curing, spectra were collected in at least two locations on the film and the spectra were then normalized and averaged. From Figure 3-8 it is clear that only minor changes occur in the FTIR spectra with the addition of cross-linker, the reasons for this are three-fold. Firstly, relatively small amounts 43  of cross-linker were added to the films, with a maximum being 16.7 wt.% (1:5 weight equivalents of cross-linker to polyurethane). Secondly, according to US 6897281 only a 0.1 to 0.3 milli-equivalents of carboxyl groups per gram of dry polyurethane are present in the PEO-PU polymer [110]. This means that on a molar basis, there are only a few cross-linking sites per polymer chain. This combined with the low relative addition of cross-linker means that new peaks formed via the cross-linking reaction will be low in magnitude relative to other peaks in the spectra. Finally, the cross-linker, a polycarbodiimide has a fairly similar structure to the polyurethane: in the synthesis of the cross-linker, diisocynates are reacted in the presence of a polyol and a carbodiimide forming catalyst to make a polycarbodiimide cross-linker [120], [121]. The spectra for dried cross-linker and dried uncross-linked urethane are compared in Figure 3-9. Details of the IR absorbance spectra of various PEO-PU films with different polycarbodiimide cross-linker ratios are compared in Figure 3-10, in which the difference in absorption between the crosslinked film samples and the uncross-linked film is summarized at different wavelengths.  Figure 3-8: FTIR transmission spectra for uncrosslinked and cross-linked films containg different weight percentage cross-linker (%XL). 44  In Figure 3-9 a and c, the cross-linker alone and the PEO-PU alone show broad amine peaks at 3300 cm-1 associated with amine (N-H) stretching in the urethane group in the polymer backbone. This peak is present in the cross-linked samples as well, and tends to increase slightly with increasing cross-linker content as shown in Figure 3-10a. This would align with an increase in amine groups associated with the cross-linking reaction. Peaks in the 2850-2920 cm-1 are present in all samples and are associated with C-H stretching, increasing cross-linker content leads to a slight decrease in absorbance due to lower molar ratio polyethylene oxide groups in the cross-linker as shown in Figure 3-10b. The strong peak in the cross-linker spectra at 2111 is the characteristic carbodiimide peak (-N=C=N-), there is a slight peak in the 16.7% wt. cross-linker sample in Figure 3-8 and in the PU15XL in Figure 3-9c at this frequency, indicating that unreacted carbodiimide remains in this sample at this concentration, this is emphasized in the difference spectra Figure 3-10c. Samples with less cross-linker do not show this peak, indicating that ~15% wt. cross-linker is likely the upper limit for effective cross-linker addition. A peak at 1713 cm-1 is present in all samples in Figure 3-9 and is associated with stretching of the carbonyl bond (C=O) in the urethane group in the polymer backbone. This peak increases and shifts with increasing cross-linker as shown in Figure 3-10d and could be indicative of the additional carbonyls associated with the urea bond in the crosslinking reaction. A peak at 1530 cm-1 is associated with the –C-NH– group in the urethane/urea bond and the increase in the absorbance observed in Figure 3-10e may be indicative of the cross-linking reaction. A peak at 1241 cm2 in the polyurethane and all cross-linked samples in Figure 3-8 is associated with the C-N and C-O bonds in the urethane linkage in the polymers. The large peak at 1096-1098 cm-1 present in all samples is the ether stretching peak (–C-O-C–) associated with the PEO soft segments in the polymers, the intensity of this peak is 3 to 4 times that of other peaks in the polyurethane spectra for the uncross-linked sample in Figure 3-9 aligning with the ~75% soft segment content reference in the patent document for this polymer [110]. This peak is also of significantly greater intensity in the PU samples than the cross-linker in Figure 3-9b, and decreases with increased cross-linker content in the samples as observed in Figure 3-10g due to the blending of the cross-linker polyethylene oxide groups and the PU polyethylene oxide groups. There is no absolutely definitive evidence of a new peak formation associated with the N-acyl urea in the cross-linked polymers, and most changes in the spectra associated with increasing the cross-linker ratio in the polymer can be attributed to simply the blending of cross-linker with the polyurethane. However, the lack of a carbodiimide peak at 2111 cm-1 in the polymers with lower cross-linker as observed in Figure 3-10c, along with the presence of this peak arising in higher cross-linker 45  loaded samples would be indicative of a crosslinking reaction. Furthermore, the growth and broadening of peaks at 1713 and 1530 cm-1 and the shoulder growing into a peak at 1640 cm-1 (as evident in the difference in spectra in Figure 3-10d and Figure 3-10e) point towards N-acyl urea formation through the cross-linking reaction. N-acyl urea formation is associated with peaks in the ranges of 3500–3300 cm-1; 1730–1690 cm-1; 1670–1630 cm-1; and 1550–1500 cm-1 [122]. Clearly no peaks arise near the 1800 cm-1, as observed in Figure 3-10c and Figure 3-10d, indicating no anhydride formation which can be associated with side reactions of the carbodiimides. This IR analysis along with the dramatic decrease in swelling in water observed with increased cross-linker content in the polymer (as discussed in Section 3.7) and with the complete inability to dissolve the cross-linked polymer in water at greater than 70oC it is clear that a cross-linking reaction has been achieved.   Figure 3-9: IR absorbance spectra (normalized to the largest absorption peak in all spectra) for dried films of (a.) cross-linker (XL-702), (b.) PEO-PU film with 15% cross-linker, and (c.) PEO-PU (PERMAX™ 230), 46   Figure 3-10: IR absorbance spectra for films with increasing cross-linker addition, less the spectra of the uncross-linked film (upward peaks are increased absorption and downward peaks are decreased absorption relative to the uncross-linked samples). 3.6 DSC and TGA Thermal Analysis A DSC scan of the PEO-PU polymer cross-linked at 6.7 wt.% polycarbodiimide cross-linker is shown in Figure 3-11. The DSC in Figure 3-11a is a scan of the polymer as cast and dried in a vacuum and kept in a desiccator. The sample was heated to 150°C, cooled at a rate of 10°C/min to -85°C before the heating ramp which is shown at 20°C/min. Three notable features are observable in this scan, the ‘soft’ segment PEO glass transition is a -53°C, a melting peak near 50°C which is associated with the crystalline portion of the soft-segment PEO melting, and finally a sharp melting peak at 235°C associated with the melting of the crystalline portion of the ‘hard’ polyurethane segments of the copolymer. 47   Figure 3-11: (a.) representative DSC of a crosslinked (6.7% wt) PEO-PU polymer (2nd heat cycle at 20°C/min after cooling from 150°C at 10°C/min) and (b.) TGA/DTA in N2 (20°C/min).  A representative TGA of the same cross-linked polymer is shown in Figure 3-11b, it can be seen that the onset of degradation is around 250°C, indicating that no polymer degradation has occurred in the DSC scan. The differential thermal analysis (DTA) of the polymer degradation in N2 is summarized in Figure 3-12. During the thermal degradation in nitrogen, two distinct peaks are observed in the DTA plot. The lower temperature peak found at 347°C is associated with the thermal degradation of the urethane segments in the copolymer. The higher temeprature peak at 421°C is associated with the degradation of the PEO soft segments peaks [123]. In nitrogen all of the copolymer is lost at 500°C. The deconvolution of the curves indicates a relative weight ratio by integration of the deconvoluted curves of hard (PU) to soft (PEO) segments of about 31:69, based on the cumulative fitted peaks in Figure 3-12 which have R2=0.998. This aligns with FTIR and NMR data reported above on the relative ratios of the soft to hard segments in the PEO-PU polymer.  48   Figure 3-12: Deconvolution of the DTA data for the thermal degradation of PEO-PU in Nitrogen. 3.7 Liquid Water Interactions With PEO-PU Polymers 3.7.1 Water Uptake of Carbodiimide Cross-linked PEO-PU Samples of each film were cut and weighed and then immersed in water at room temperature (23oC). Measurements were taken over time to ensure that a steady state was achieved. Density measurements were completed by pycnometer dry and then wet density measurements were taken again after reaching equilibrium at room temperature. A summary of the liquid water uptake in the PEO-PU polymers with different levels of cross-linking can be found in Table 3-4 and a graphical summary in Figure 3-13. 49   Figure 3-13: Liquid water uptake in cross-linked PEO-PU polymers at 23°C, error bars are sample standard deviation about the mean for three samples of the material.  Table 3-4: Water uptake and density of PEO-PU polymers at various cross-linking ratios. Samples (XL:PU wt.) Wt. % Cross-linker Mass% Water Uptake at Steady State Volume Fraction Water  Dry Polymer Density (g/cm3) Swollen Polymer Density (g/cm3) 0 0 134 0.573 1.114 1.050 1:25 3.85 113 0.532 1.113 1.050 1:20 4.76 108 0.519 1.092 1.057 1:15 6.25 102 0.506 1.092 1.051 1:10 9.09 92 0.478 1.115 1.064 1:5 16.67 87 0.441 1.117 1.062  From Figure 3-13, increasing the crosslinking ratio clearly decreases water uptake in the polymer at room temperature conditions, above ~9% cross-linker the change in swell ratio with change in is cross-linker concentration decreases indicating that there is diminishing impact with increase cross-linker content. This agrees with the FTIR results indicating that 50  there is some unreacted carbodiimide functional groups at 16% cross-linker. Increasing the cross-linker content above 16% actually led to higher swell ratios, further indicating that the reaction stoichiometry is reached above ~16%.  The polymer films were then placed in a water bath and the temperature was increased incrementally. Samples were allowed to equilibrate for 24 hours in water at each temperature. Samples were quickly removed, placed on a lint-free task wipe to remove excess surface water and weighed. Results are reported in Figure 3-14 and Table 3-5. The impact of crosslinking is clear from these results, film samples which were not cross-linked swelled with nearly 300% total uptake at 60°C before gelling and then dissolving completely in water above 70°C. All crosslinked samples remained intact in water up to 90°C. However, the films with different cross-linker ratios responded differently to increasing temperature.  Samples with lower cross-linker content (3.9 and 4.8% in Figure 3-14), showed an increase in uptake to 60°C, a decrease in uptake at 70°C, before increasing dramatically in swell ratio at 80°C, and then decreasing again at 90°C. The final decrease may have been due to some uncross-linked polymer dissolution, however the large decrease at 70°C is of some interest and has been reported in previous studies on PEO polymers. It has been shown that the optimal conformation for chains of polyethylene oxide are assemblies of 72 helices, in which the oxygen groups available for hydrogen bonding with water are well spaced for low energy conformations of hydrogen bonded water-PEO structures [67], [124]. The optimal conformation of PEO allows about three water molecules per PEO monomer group where two water molecules are hydrogen bonded to each ether oxygen group and a third is associated via hydrogen bonding to the first two molecules [67]. These conformations allow for further ‘clustering’ of water molecule via water-water hydrogen bonds. These ‘clustered’ water molecules are referred to as ‘unbound water’ as opposed to water molecules directly hydrogen bonded to EO groups, which are referred to as ‘bound’ water. Since the bound water is in a local minimal free energy state, it is thermodynamically favorable and bound water will remain in the polymer even at elevated temperatures in air. In liquid water, as the temperature rises, the polymer molecular chains become more mobile, larger spaces open between polymer chains and clustering increases, eventually sufficient water surrounds the individual polymer chains and the polymer becomes soluble. In cross-linked samples, this final solubility is not possible as the PEO containing polymers are linked together, instead a swollen crosslinked polymer remains. With lower levels of cross-linking the polymer has sufficient mobility to swell and uptake larger amount of water. As the temperature rises, the EO groups are known to change in conformation moving away from the ideally space 72 helix, with the oxygen group 51  no longer optimally spaced for hydrogen bonding with water, the thermodynamic impetus to hold excess water is diminished, and swelling decreases. This partially explains the anomalous water uptake behavior between 50 and 90°C for polymers with less cross-linking (3.9 and 4.3% in Figure 3-14). With high amounts of crosslinking (6.3, 9.7, and 16.7% in Figure 3-14), the polymer is insufficiently mobile to accommodate much swelling, and the polymers generally demonstrate a steady decrease in water uptake with increasing temperature associated with EO group mobility as conformation diverts from the 72 helical structure. Samples with higher crosslinking ratios have maximum uptake of water before 45°C, and the highest (16%) cross-linked sample shows a steady decrease in swell with temperature increase. As water temperature increases, the chemical potential of water decreases, at lower temperatures the free volume of the polymer fills with water, and water bonds to the hydrophilic sites in the polymer. Increasing the temperature increases the mobility of the polymer chains to reconfigure, increasing the free volume, and allow uptake of more water. However, if the polymer is constrained due to heavy cross-linking such as in the 16% cross-linker sample, no excess water is absorbed at increased temperature and the polymer absorbs water in direct proportion to the chemical potential. Table 3-5: Swell ratios for polymer films with increasing temperatures. Wt. Ratio XL:PU Wt. % XL Mass% Water Uptake at Steady State (g H2O/g dry polymer) 23°C 35°C 40°C 45°C 50°C 55°C 60°C 70°C 80°C 90°C 0 0 132 160 177 196 225 261 289 g d d 1:25 3.85 109 123 127 130 140 154 165 165 294 260 1:20 4.76 102 114 118 118 122 139 145 120 204 113 1:15 6.25 101 105 107 107 103 103 101 94 87 57 1:10 9.09 84 91 90 88 85 86 82 75 78 63 1:5 16.67 75 72 71 69 66 61 59 53 52 42 g: gelled; d: dissolved  Since the total mass ratio of polyethylene oxide groups in the PEO-PU polymer is estimated as ~70% weight as supported by the FTIR, NMR, and TGA data. The total molar ratio of water 52  uptake to PEO monomer groups can be then calculated based on the water uptake, the weight of PEO-PU polymer, and cross-linker content. This data is summarized in Table 3-6. Table 3-6: Water uptake as a function of total polyethylene oxide groups in the polymer. Wt. Ratio (XL:PU) Wt. % XL mol H2O/mol polyethylene oxide monomer 23°C 35°C 40°C 45°C 50°C 55°C 60°C 70°C 80°C 90°C 0 0 4.6 5.6 6.2 6.8 7.9 9.1 10.1 g d d 1:25 3.85 4.0 4.5 4.6 4.7 5.1 5.6 6.0 6.0 10.7 9.4 1:20 4.76 3.7 4.2 4.3 4.3 4.5 5.1 5.3 4.4 7.5 4.1 1:15 6.25 3.8 3.9 4.0 4.0 3.8 3.8 3.8 3.5 3.2 2.1 1:10 9.09 3.2 3.5 3.5 3.4 3.3 3.3 3.1 2.9 3.0 2.4 1:5 16.67 3.1 3.0 3.0 2.9 2.8 2.6 2.5 2.2 2.2 1.8 g: gelled; d: dissolved  It can be observed that for polymers with higher cross-linking, that the total water uptake aligns with the optimal conformation for PEO, with ~3 water molecules per group. This would indicate that there is very little free volume created due to swelling for cross-linking levels above ~6%. For lower cross-linking ratios the swelling allows excess free water to enter the polymer.  53   Figure 3-14: Liquid water uptake (swelling) of cross-linked and uncross-linked PEO-PU polymers at various temperatures, uncrosslinked samples gelled above 60°C and dissolved above 70°C, error bars are sample standard deviations about the mean for measurements of three samples of the material. 3.7.2 Hydrated DSC Samples (~10 mg) were dried and weighed and then a controlled amount of water was added or the samples were saturated (48+ h) prior to being weighed again. The samples were then placed in a hermetically sealed pan for DSC analysis. Heating and cooling thermographs are reported in Figure 3-15 for samples of PEO-PU with 6 levels of cross-linking for samples which were saturated in liquid water at 23oC. Water can exist in three states in the polymers, bound water, bound-freezing water and freezing ‘free’ water [125], [126]. ‘Bound water’ is directly connected via hydrogen bonding to the polar PEO groups in the polymer, will not freeze, and should not be detected by any peaks in DSC measurements. ‘Freezing water’ is in the free volume of the swollen polymers and is available to freeze and melt as normal water. ‘Freezing-bound water’ is associated with water surrounding the bound water associated via hydrogen bonding in the hydration shell of the polymer, it is not directly bound to the polymer itself, it will freeze, but at much lower temperatures than free water. The heating thermographs in Figure 3-15a show broad endotherms with a peak at 0°C, these endotherms as a whole represent all freezing water in the polymers. The total available freezing water can be calculated from integration of these endotherms [127]: 54  wfHHW  [3-1] Where ΔHw is the heat of fusion of pure water (334 J/g). Which along with the total water uptake in the polymer can be used to determine the amount of bound water, and the bound water per PEO group in the polymer can be determined by:   1001441822XLPEOpolyfOHboundWWWWWEOOH [3-2] Where WH2O is the total weight fraction of water uptake in the polymer, Wpoly is the weight fraction of dry polymer with cross-linker, WPEO is the weight ratio of PEO groups in the polymer (~0.70) and WXL is the weight fraction of cross-linker in the polymer. The calculated bound and freezing water per PEO group is summarized in Table 3-7. For all levels of cross-linking, the bound water remains around 2 mole water per mole of PEO, which would be predicted as two water molecules can hydrogen bond directly with each oxygen group in the ethylene oxide monomer (-CH2-CH2-O). Table 3-7: Calculated bound and free water in PEO-PU polymers with different cross-linker content. Cross-linker, WXL (g/g polymer) ΔH (J/g) Total Water at Saturation, WH2O (g/g) Freezing water, Wf (g/g) Bound water, WH2O (g/g) Bound water/EO (mol/mol) Freezing water/EO (mol/mol) Total water/EO (mol/mol) 0.000 111.4 0.569 0.333 0.236 1.91 2.70 4.61 0.039 87.9 0.522 0.262 0.259 1.97 1.99 3.96 0.048 99.8 0.505 0.298 0.207 1.53 2.21 3.74 0.063 82.9 0.498 0.248 0.251 1.86 1.84 3.70 0.091 53.5 0.457 0.160 0.297 2.10 1.13 3.23 0.167 54.9 0.427 0.164 0.263 1.92 1.20 3.12 55  As the crosslinking ratio increases, less free and freezing water is found in the polymer due to crosslinking constraining the polymer and preventing swelling and excess water from entering the polymer. The cooling thermographs in Figure 3-15b show a separation of the different states of water in the polymers. Free water crystallizes near -20°C in all polymers as shown by sharp exothermic peaks, while the bound-freezing water crystallizes at lower temperatures in broader peaks. Interestingly the cross-linking content affects the crystallization peaks and temperatures of the bound water, and in higher cross-linking degrees (9-16%) the crystallization peaks for the bound water are extended to lower temperatures implying that the bound freezing water is more strongly associated with the polymer. This may be due to a greater portion of the available freezing water being associated only with the water molecules bound to the polymer itself, the ‘free’ water freezing peaks are of a lower magnitude for the higher cross-linked polymers, and due to the cross-linking less swelling occurs and there is less free volume for free water in the polymer.  Figure 3-15: (a.) Heating at (left, 20°C/min) and, (b.) cooling (right, 10°C/min) DSC thermographs for water saturated polymer at different cross-linking ratios. Heating DSC thermographs of a PEO-PU polymer with increasing content of water in a polymer with 6.3% cross-linker are shown in Figure 3-16, here no endothermic peak associated with melting of water are observed until a water content of greater than ~40% is present in the polymer, this would correspond to a H2O/EO molar ratio of about 1.7, meaning 56  that the water that enters the polymers is hydrogen bonded to the PEO groups until most of the available ether oxides are associated with two water molecules.  Figure 3-16: DSC heating of PEO-PU with 6.3% cross-linker and increasing weight percentage of water in the polymer. 3.8 Water Vapour Interactions with PEO-PU films In the membrane application, the polymers will rarely interact with liquid water. In most conditions the membrane will be interfacing with air streams containing varying degrees of moisture in vapour form. Polymers will interact differently with vapour than with liquids therefore it is important to study sorption of water vapour in the polymers as well as liquid water in the polymer. 57  3.8.1 Kinetics of Sorption in PEO-PU Films In dynamic vapour sorption studies the sample is held in a controlled temperature environment at a given relative humidity until an equilibrium weight has been obtained. A step change in relative humidity is applied to the gas in the sample chamber and the uptake response in the polymer can be recorded. This allows information about the kinetics of water vapour sorption. In this study the kinetics of sorption are determined for PEO-PU polymers at three levels of cross-linking and at different humidity and temperature conditions. Uptake curves normalized for the total uptake in the step for a 6.7% weight crosslinked PEO-PU polymer at 50°C at various water vapour activities are shown as an example in Figure 3-17.   Figure 3-17: Water vapour uptake curves at various water vapour activities for PEO-PU polymer with 6.7% cross-linker at 50°C. The uptake curve can be fit to the following equation to determine the uptake kinetic parameter associated with the settling time k1 (s) for the polymer at the given conditions [128].  10expktmmmmt  [3-3] The kinetic parameter is summarized for the polymers at different temperatures and relative humidity along with the total uptake in the step at the specified conditions are summarized in Table 3-8. As temperature increases, the total uptake in the step decreases, and the kinetics 58  are faster. As the relative humidity increases the total uptake in the step increases, and the kinetics of uptake are slower. As the polymer takes up more water, swelling occurs and the rate of settling in the polymer is longer. These values can be related to the diffusivity in the polymer for known samples geometry and thickness, however due to the relatively thin samples (~25 micron) and the large sample size required for the experiment (~100 mg), many layers of polymer were required in this case, and accurate diffusivity coefficients could not be extracted from the data. Instead the sorption data along with the permeability data for the films was used to determine the diffusivity, this is discussed in detail below in Section 3.10. Table 3-8: Summary of mass uptake and kinetic data for each step  Temperature  15°C 35°C 50°C 70°C RH (%) k1 (s)  (%) k1 (s)  (%) k1 (s)  (%) k1 (s)  (%) 25 1224 0.22 360 0.22 216 0.22 324 0.20 50 2376 0.52 756 0.43 252 0.39 360 0.34 75 6984 2.80 2592 1.80 648 1.12 432 0.78 85 - - 7200 5.04 1260 2.91 864 1.38 Table 3-9: Kinetic parameters for sorption of water vapour for PEO-PU polymer at different levels of crosslinking at 25oC. RH (%) Kinetic Parameter, k1 (s) 0% XL 6.7% XL 16.7% XL 5 612 864 1368 15 756 792 1296 25 828 1044 1332 50 972 1620 1692 75 2916 3816 4536 85 5544 6444 6732  The kinetic parameters and total uptakes at 25°C for three levels of cross-linking are summarized in Table 3-9. Increasing the cross-linking ratio tends to increase the kinetic 59  parameter for water vapour uptake meaning that the kinetics of sorption are slower. Since cross-linking constrains the polymer it should take longer for the polymer to come to equilibrium as water vapour absorbs in the polymer. 3.8.2 Sorption and Desorption Curves Water vapour sorption experiments were completed on 25 micron thick films of dried and cured PEO-PU polymers with 6.7% poly-carbodiimide cross-linker by weight at 15, 25, 35, 50, and 70oC. Sorption and desorption isotherms were completed; however, no significant hysteresis was observed for desorption curve. This would be expected for rubbery polymer, in which no relaxation should occur at higher humidity which would lead to additional free volume to be filled during desorption measurements.  The vapour uptake at various temperatures is reported in Figure 3-18. At lower activity (Pv/Psat < 0.5) the sorption is not significantly affected by temperature, and the change in uptake with increasing activity is relatively constant. However, at higher activities, the polymer shows an exponential increase in water vapour uptake, and the curves are convex towards to the activity axis, following a Flory-Huggins model for uptake [129]. This is typical for condensable species (ie. vapours) in rubbery polymers (ie. non-crystalline polymers above the glass transition), even those which are cross-linked [129]. The effect of temperature on sorption is much more significant above activity of 0.5, as shown in Figure 3-18 and a significant decrease in sorption is associated with higher temperatures in the polymers. This is expected as the enthalpy of sorption (ΔHs) is the key thermodynamic parameter effecting gas and vapour solubility:    RTHSS sexp0  [3-4] And the enthalpy of sorption has contributions from both the enthalpy of condensation (ΔHc) of the absorbing species and the enthalpy of mixing (ΔHm) of the species with the polymer: mcs HHH   [3-5] For water vapour, which has a large negative enthalpy of condensation, the total change in enthalpy of sorption will be negative, and an increase in temperature will lead to a decrease in solubility.  The sorption of water vapour at 25oC in the PEO-PU polymers at three levels of crosslinking are shown in Figure 3-19. The level of cross-linking has a minimal effect on the sorption of 60  water vapour at lower water vapour activity. At high water vapour activity the constraint on excess water uptake induced by the cross-linking in the polymers begins to have an effect, and the total sorption decreases slightly with increased cross-linking. However, compared to the cross-linking effect on liquid water reported in Section 3.7, the effect of cross-linking on water vapour uptake is much less significant. This means that cross-linking should only have a minor impact on the water vapour transport properties of the polymer at typical ambient operating conditions.  Figure 3-18: Water vapour sorption isotherms for a 6.7% cross-linked PEO-PU sample at various temperatures. 61   Figure 3-19: Water vapour sorption isotherms for three levels of cross-linking in PEO-PU polymer at 25oC. 3.8.3 Solubility Coefficients The solubility coefficient for the polymer is calculated by:  pCS   [3-6] where C is the concentration of water in the polymer (cm3 H2O(v) (STP)/cm3polymer) and p is the water vapour pressure. Solubility coefficients are reported in Figure 3-20. For an ideal system Henry’s law of sorption is expected and the solubility would be expected to be linear. Clearly from Figure 3-20, solubility of water vapour in the PEO-PU polymers is non-ideal, this is expected as interactions between the PEO-PU polymer and water as discussed in Section 3.7. 62   Figure 3-20: Solubility coefficients for a PEO-PU polymer cross-linked at 6.7% at different temperatures. 3.8.4 Flory-Huggins Interaction Parameter For the PEO-PU polymers studied in the presence of water vapour, penetrant-polymer interactions and penetrant-penetrant interactions are expected. Flory-Huggins solvation relationships are typically used to describe sorption isotherms for condensable penetrants in rubbery polymers. This model uses an interaction parameter, χ to describe the uptake to variable activity based on the volume fraction of the penetrant (water), ΦW [130]:        2111lnlnln WWPWWsatvWVVppa    [3-7] Since the molecular weight of the polymer is large, ~230,000 g/mol (before crosslinking), the molar volume of the polymer VP will be very large relative to the molar volume of the penetrant, VW and the center term in the right side of Equation [3-7] is insignificant for these studies. Using the sorption data for the PEO-PU polymer at different temperature conditions the interaction parameter can be determined at variable water vapour activity, as summarized in Figure 3-21. The interaction parameter can be seen to vary with concentration or the volume 63  fraction of water in the polymer, this is expected for polymer solvent systems with polar penetrants [131]. The interaction parameter is low when the polymer and penetrant are completely miscible, for 0.5<χ<2 interactions are moderate, and for χ>2 interactions are smaller (ie. in the case of water the polymer is more hydrophobic) [129]. Generally the third term on the right side of Equation [3-7] refers to the enthalpy mixing term, and the first two terms to the entropic mixing [132].  The results suggest that at lower water content, the water associates with the polymer but the mixing is less energetically favorable than at higher water contents where mixing of water into polymer is increasingly favorable as the polymer becomes increasingly hydrophilic due to the increasing water content in the polymer. Furthermore, as temperature increases, the interaction parameter is greater indicating that the polymer is more ‘hydrophobic’ at higher temperature. The interaction parameter results for three levels of crosslinking are shown in Figure 3-22. The crosslinked samples have a slightly higher interaction parameter as lower water content, likely associated with the cross-linker being less hydrophilic than the PEO-PU polymer. This has been observed in other systems such as crosslinked acrylate hydrogels [133].   Figure 3-21: Interaction parameter (χ) for water vapour with PEO-PU polymer cross-linked at 6.7% wt. carbodiimide at different temperatures. 64   Figure 3-22: Interaction parameter (χ) for water vapour with PEO-PU at different levels of cross-linking at 25°C. 3.9 Zimm-Lundberg Clustering Analysis Water cluster formation in polymers arises from hydrogen bonding interactions between the solvent molecules, this is typically studied via Zimm-Lundberg analysis [134]. Clustering arising from water-water interactions can have a significant effect on the diffusivity and permeability of polymers as the size of the penetrant increases from a single water molecule, to multiple water molecules clustered together. Clustering is generally described by:  1,11 TPWWWPW aaVG [3-8] Where ΦW and ΦP are the volume fractions of water and polymer, and G11/Vw is referred to as the clustering function. If G11/Vw = -1, the system is an ideal solution and the water is evenly distributed throughout the polymer as individual molecules excluding only their own volume. If G11/Vw > -1 then the solution becomes non-ideal as the molecules begin to cluster together. This arises from swelling associated with water molecules creating more volume for excess water molecules interacted by hydrogen bonding. This analysis describes the tendency of 65  solvent molecule to cluster, either while attached to the polymer, or in the free volume. Using sorption data, the derivative in Equation [3-8] can be integrated and the clustering function can be solved. The average number of solvent molecules in a cluster is defined as [135]: 111 WWCVGN  [3-9] For ideal mixing of water in the polymer, Nc = 1 and there is no clustering of water molecules. The results for a PEO-PU polymer at different temperatures is summarized in Figure 3-23. At low activities, there is minimal clustering (Nc~1) as the water molecules bind to the PEO groups in the polymer. However, as more water enters the polymer, the cluster number increases at higher water vapour activity. Increased temperature decreases the size of clusters, as the amount of water in the polymer is lower and clustering becomes less favorable. As shown in Figure 3-24, the cross-linking of the polymer leads to a slight decrease of the total cluster size at high activity is likely due to a decrease is polymer mobility.  The cluster number represents a statistical thermodynamic calculation based on deviation from ideal mixing, so the clustering integral captures an estimation of how water is aggregating in the polymer. This may include changes in polymer morphology associated with movement of the PEO side chains to further phase separate from the PU backbone to accommodate for excess water. From the liquid uptake and DSC analysis the absolute number of water molecules per PEO group was determined to be in the range of ~3 for complete saturation. The clustering analysis shows that there is further aggregation of these water molecules into clusters, even at conditions below saturation. This clustering may somewhat impact the diffusivity of water through the membrane due to water-water interactions, decreasing the overall permeability, however significant clustering only occurs at very high humidity conditions, which are not typical in ERV applications. 66   Figure 3-23: Average cluster size of water molecules in the PEO-PU polymer at 6.7% wt. carbodiimide cross-linker at various temperatures and water vapour activities.  Figure 3-24: Average cluster size of water molecules in the PEO-PU polymer at two levels of cross-linking at at varying relative humidity and 25°C. 67  3.10 Permeability Films of PEO-PU with 6.3% cross-linker were cast, dried, and tested for water vapour transport in a counter-flow dynamic water vapour transport apparatus, at various feed stream activity and temperature conditions. All tests were isothermal, and the feed stream vapour pressure was held constant until a steady state was achieved for the permeability measurement. The details of this apparatus are described in Chapters 2.20 and Chapter Sections 4.2 to 4.4. The permeability of the films are reported here in Barrer units to make the data easier to read and to compare with results in the membrane science literature. The conversion of the Barrer unit from SI is as follows [1]: cmHgscmcmSTPcmBarrerPasmmmol 2310152)(101110989.2  [3-10] Permeability is normalized by the average vapour pressure differential across the film in the test module, which is corrected for the boundary layer resistances. The boundary layer corrections are discussed in the Chapter 4. The permeability of a 25 micron thick 6.3% crosslinked PEO-PU film measured at various temperatures and water vapour activities are reported in Figure 3-25. It can be seen that water vapour permeability increases slightly with temperature (up to 50°C). Permeability also increases with increasing activity, even though the results are normalized for concentration differential. In order to further understand these effects, the diffusivity was determined using the solubility coefficients (S) determined from the sorption isotherms and permeability (P) measurements [1]: SDP OH 2  [3-11] Which is modified to account for the average concentration of water on both sides of the membrane in the permeation module: sfsfOHppCCDP 2  [3-12] Where Cf and Cs are the concentration of water in the polymer at the feed and sweep sides of the film interface as determined from the boundary layer corrected average vapour activity at the membrane surface and the sorption data at the given temperature and humidity. The 68  vapour pressures pf and ps are the boundary layer corrected vapour pressures at the membrane surfaces.  Figure 3-25: Water vapour permeability for a 25 micron thick film of a 6.7% crosslinked PEO-PU sample at various temperatures. Diffusivity, solubility, and permeability as determined from the dynamic sorption data and the film permeability data are reported in Table 3-10 and summarized in Figure 3-26. Diffusivity increases with temperature at a given water vapour activity as expected due to increases molecular mobility, whereas sorption decreases with temperature. At constant temperature the diffusivity tends to decrease at higher water vapour activity as the polymers begin to swell and clustering becomes significant. The effect of increasing solubility with activity on permeability can also be observed in Figure 3-26. At higher water activity, the diffusivity decreases, but the solubility coefficient increases to a greater extent. Since the permeability is the product of diffusivity and solubility (Equation [3-11]), the overall effect is an increase in permeability with increase water vapour activity. Similarly, at lower temperatures the diffusivity is decreased significantly, and even though solubility increases at lower temperatures the permeability is generally lower at lower temperatures. The effect of crystallinity in the PEO segments of the polymer may also be a factor effecting the overall permeability at temperatures below the PEO melting temperature ~50oC, and at lower humidity conditions. At higher temperatures (70°C) the permeability is affected by the decrease in the solubility coefficient at higher temperature, but since less water is in the polymer the clustering number 69  is decreased and so the diffusivity is less affected by higher water vapour activity than at lower temperatures, and the overall permeability is greatest at higher temperatures.  Table 3-10: Summary of diffusivity and solubility coefficients and permeability in PEO-PU polymers at various temperature and water vapour activity. Temperature (°C) Pv/Psat D (cm2/s) S (cm3(STP)/cm3/Pa) P (cm3(STP)-cm/cm2/s/Pa) P (Barrer)a 25 0.21 1.06E-07 1.56E-02 1.91E-09 25425 25 0.44 1.91E-07 1.89E-02 3.56E-09 47415 25 0.50 1.79E-07 2.12E-02 3.87E-09 51565 25 0.72 1.31E-07 3.85E-02 4.81E-09 64133 35 0.16 1.81E-07 8.18E-03 1.68E-09 22336 35 0.26 3.04E-07 9.17E-03 2.84E-09 37920 35 0.48 3.28E-07 1.22E-02 4.10E-09 54702 35 0.50 3.23E-07 1.28E-02 4.25E-09 56596 35 0.73 2.58E-07 2.27E-02 5.64E-09 75256 35 0.86 1.59E-07 5.27E-02 7.02E-09 93600 50 0.14 2.87E-07 3.79E-03 1.30E-09 17345 50 0.29 7.16E-07 4.33E-03 3.15E-09 41933 50 0.50 7.88E-07 5.50E-03 4.47E-09 59575 50 0.57 7.20E-07 6.25E-03 4.78E-09 63721 50 0.67 7.09E-07 7.34E-03 5.44E-09 72522 50 0.80 6.87E-07 1.05E-02 6.70E-09 89310 50 0.85 5.77E-07 1.56E-02 7.55E-09 100707 70 0.15 1.26E-06 1.63E-03 2.18E-09 29083 70 0.25 1.65E-06 1.75E-03 2.97E-09 39581 70 0.49 2.17E-06 2.10E-03 4.48E-09 59754 70  70 0.68 2.07E-06 2.75E-03 5.81E-09 77436 a 1 Barrer = 1×10-10 cm3 (STP)•cm cm-2 s-1 cmHg-1.  Figure 3-26: Solubility and diffusivity coefficients, and permeability in the PEO-PU polymer. 3.11 Conclusions An estimate for the chemical structure for the polyether-polyurethane copolymer (PERMAX™ 230) was determined, with the polymer containing about 70% by weight of PEO side chains. This is supported by NMR, TGA, FTIR, and water uptake results. The PEO side chains cause the polymer to have high uptakes of liquid water and water vapour. In order to be useful in the application, the PEO-PU polymers must not be dissolved in liquid water, and so the polymers must be cross-linked. A polycarbodiiimide cross-linker was used to cross-link the PEO-PU polymers, crosslinking was confirmed by FTIR and water uptake measurements.  The interactions of liquid water and water vapour with the cross-linked PEO-PU polymers were then studied. Water uptake measurements and hydrated DSC measurements were used to determine the interaction of the PEO-PU polymers with liquid water at different levels of cross-linking. Liquid water exists in several states in the polymers depending on the level of hydration and the level of cross-linking. At <40% weight water, only water which is directly hydrogen bonded to the PEO groups in the polymer is present. Additional water can enter the polymer and be associated with the bound water, or as free water in the polymer void space.  Water vapour sorption studies were carried out to determine the sorption and solubility coefficients for the cross-linked polymer at various conditions. Furthermore, an analysis of the Flory-Huggins interaction parameters for the polymers was conducted at different temperatures and cross-linking ratios. Cross-linking had a minimal overall effect on water vapour sorption and the interaction parameters for the polymers. A Zimm-Lundberg clustering 71  analysis was also carried out to study the effects of clustering of absorbed water vapour in the polymers. The analysis showed that especially at high humidity the polymer morphology allows excess clustering of water in the polymer, which impacts diffusivity of water vapour through the polymer.  Polymer films were tested for water vapour permeability, and the solubility and diffusivity coefficients of the polymers were determined. Although the diffusivity decreases slightly at higher humidity (likely due to clustering), the solubility increases greatly at these conditions which off-sets the effect of lower diffusivity. Permeability of the polymer thus increases with increasing water vapour activity, and is not strongly impacted by temperature between 25 and 70°C.    72   Composite Membranes for Water Vapour Transport Applications 4.1 Introduction The cross-linked polyethylene oxide-polyurethane copolymers studied in Chapter 3 have appropriate gas selectivity properties for the application, and as a result of their high weight fraction of PEO groups they have high water vapour uptake and permeability. However, in order to use these polymers in a humidity exchanger they must be able to be handled, processed, and built into exchanger devices. Films of the polymer alone would need to be greater than 15 microns in thickness to manufacture a free standing film which has sufficient mechanical properties for further processing. However, the permeance or water flux properties of a dense polymer film decrease proportional to thickness, and so due to their thickness these films would have relatively poor water vapour transport performance. In order to overcome this, the selective polymer film layer can be cast on the surface of a much thicker microporous carrier substrate. The substrate should ideally have high porosity and low resistance to vapour transport, it must be thick and strong enough to handle, and have small enough pore size that a dense film layer of selective polymer can be cast on the surface of the substrate. By using an appropriate microporous substrate as the carrier for the selective film layer, the film layer thickness can be easily reduced to less than 5 micron or even less than 1 micron in thickness. These ‘composite membranes’ made by combining a microporous substrate with a thin dense selective polymer layer on one surface are easy to handle and process, and since the selective layer is thin, the water vapour permeance of the membrane can be significantly higher than a free standing film of the same polymer.  In this chapter the component resistances in composite membranes are determined, which demonstrates the individual resistances associated with boundary layers, the membrane film layer, and the membrane substrate layer in vapour transport membranes. Other studies have evaluated these resistances in composite membranes for ERVs, but have not studied how each resistance independently contributes to the total transport. The fabrication and testing of several of these composite membranes is completed in this chapter and a discussion addressing some the challenges and deficiencies of this approach are presented. The microporous substrates used in these composite membranes are shown to contribute up to 50% of the total membrane resistance to water vapour transport. This chapter is critical to the central hypothesis of this dissertation which posits that the substrate contributes a significant resistance to transport and that eliminating this substrate and its associated resistance will lead to improved water vapour transport membranes. 73  4.2 Water Vapour Transport Testing Details of the dynamic water vapour testing apparatus are summarized in Chapter 2 Section 2.21. In a water vapour transport test, the membrane is placed in a counter-flow test module and air at controlled humidity is supplied to the ‘feed’ side of the membrane and dry air supplied to the ‘sweep’ side of the membrane in counter-flow as shown in Figure 4-1. The module geometry for testing the membrane is summarized in more detail in the sections below.  Figure 4-1: Membrane permeation test module. During a ‘standard’ vapour transport test a first ‘sweep’ gas stream was supplied at 50°C and 0% relative humidity (Pv,H2O,1=0 Pa) to the inlet on one side on the membrane. A second ‘feed’ gas stream was supplied to the inlet on the other side of the membrane at 50°C and ~50% relative humidity (Pv,H2O,3=6122 Pa), and at the same flow rate as the first gas. The water content and temperature of the two streams were measured and recorded at the outlets, utilizing calibrated Vaisalla humidity probes (HMT 221). For tests where the temperature, flow, and RH were not at the ‘standard’ test conditions (feed stream at 50°C and 50% RH), testing conditions are noted and the temperature, flow, and humidity of the Stream 1 and Stream 3 inlets are reported. From the absolute humidity (x, kg H2O/kg air) values, a mass balance could be completed on the system to determine the mass flow of water vapour through the membrane (kg/min):  44223311 QxQxQxQx   [4-1] And in the case where x1=0 (dry sweep stream), 74   2244332QxQxQxm OH   [4-2] The results may also be reported as a water flux by dividing the mass flow by the membrane area over which the transport has occurred in units of mass per area per time (kg/m2/h):  AmJOHOH22  [4-3] Alternatively, the flux can be calculated by assuming an ideal gas based on the sweep outlet conditions using the following equation:  RTAVpQJmsOHvOH,,2 22  [4-4] By dividing the water flux by the calculated mean water vapour pressure differential within the test module, the apparent water vapour permeance value [PH2O/δ] can be determined in units of mass per area per time per vapour pressure differential and is typically reported in gas permeance units (1 GPU = 1×10-6 cm3 (STP) cm-2 s-1 cmHg-1):   sOHvfOHvOHapparentOHppJP,,,, 2222 [4-5] Where pv,H2O is the bulk vapour pressure in the feed and sweep streams.  Permeance units are useful when the membrane is asymmetrical in nature, is not homogeneous, or is a multiple layer material. When a polymeric film of known thickness is tested, the vapour permeability (PH2O) can be determined for the film material. Permeability is typically reported in Barrer units (1 Barrer = 1×10-10 cm3 (STP) cm cm-2 s-1 cmHg-1). General Fickian diffusion in one-dimension states:  dxdRTCDJ iiii  [4-6] Where μi is the chemical potential of the diffusing substance, x is diffusion distance, Di is the diffusivity coefficient of the substance in the polymer matrix (in the present case this is H2O), and C is the concentration of the penetrant in the substance. For activity coefficients close to unity : 75   dxdCDJ OHOH 22   [4-7] When diffusivity is constant and independent of concentration:  memsfOHOHCCDJ 22  [4-8] To account for the actual concentration in the membrane surfaces the sorption coefficient of the material must be measured and can be related to the vapour pressure at the membrane-gas stream interphase (See Chapter 3, Section 3.8): OHmemfvOHmf pSC 2,,,2,   [4-9] And; OHmemsvOHms pSC 2,,,2,   [4-10] Substituting Equations [4-9] and [4-10] into [4-8]: memOHmemsvOHmemfvOHOHOHppSDJ2,,,2,,,222  [4-11] And with the permeability as defined by the solution diffusion model: OHOHOH SDP 222   [4-12] Equation [4-11] becomes: memOHmemsvOHmemfvOHOHppPJ2,,,2,,,22  [4-13] Furthermore, it is often convenient to define permeance as the permeability divided by the thickness of the membrane:  OHmemsvOHmemfvmemOHOH ppPJ 2,,,2,,,22  [4-14] 76  This equation is similar to [4-5], but the permeance here is the ‘actual’ permeance since the vapour pressure (pv,f,mem,H2O and pv,s,mem,H2O) values used are those at the membrane-air stream interphase. In most testing conditions, only the bulk vapour pressures in the feed and sweep gas streams can be measured directly, and due to boundary layers near the membrane surface, the observed mass transport coefficient is not the actual mass transport coefficient for the membrane. In order to understand how changing membrane properties and operating condition affects the material performance, the unknown concentration or vapour pressure at membrane-air stream interphases must be determined through what is known as a resistance in series model. 4.3 Resistance in Series Model When testing water vapour transport through a membrane between a feed air stream and a sweep air stream, a steady state non-equilibrium mass transport process is being examined. In this process multiple steps occur: a. Water vapour diffuses through the boundary layer from the bulk feed air stream to the membrane surface. b. Water vapour crosses the interface between the air on the feed side and enters membrane surface. c. Water vapour transports through the membrane, in the case of a coated microporous membrane, the water vapour must transport through a dense polymer film layer and a microporous substrate. d. Water vapour crosses the interface between the membrane surface and the air on the sweep side, desorbing from the membrane. e. Water vapour diffuses through the boundary from the membrane surface to the bulk sweep stream. This process is summarized in Figure 4-2, and the resistances associated with each step of the transport are additive. As explored by Metz et. al., real permeance of the membrane must be considered in terms of the module geometry, and the boundary layers present in the module under the given operating and flow conditions [93]. This is particularly the case for membranes with high levels of water vapour transport, where the boundary layers can represent a significant portion of the total resistance to mass transport. This means that the permeance should be calculated based on the actual driving force at the membrane surfaces. This is achieved by using the so called “resistance in series” model, which proposes that the 77  resistances to vapour transport in the test module and through the membrane are additive, the resistance is the inverse of conductivity:  iikR1  [4-15] And generally for the present system (see Figure 4-2):  sblmemfblobs kkkk .,1111  [4-16] Where kbl,f and kbl.s are the mass transfer coefficients associated with the boundary layers on the feed and sweep sides of the membrane respectively. For a microporous substrate with a dense selective polymeric film layer on one surface, the total resistance to water vapour transport in the membrane is a sum of the resistance of the substrate and the coating:  subfilmmem RRR   [4-17] For the total system:  sblsubfilmfblmemblobs RRRRRRR ,,   [4-18]  Figure 4-2: Representation of the resistances in the water vapour transport process through a cross-section of a composite membrane. The mass transfer coefficient in the coating is defined using the solution-diffusion model and mass transport coefficient is the permeance or the thickness normalized permeability:  78   filmfilmPk  [4-19] Where P is the permeability of the polymer film, and δfilm is the film layer thickness. The mass transfer coefficient of the porous substrate is a function of the substrate thickness (δsub), porosity (ε), pore radius (σ), and tortuosity (τ):  subsubk2  [4-20] If the Hagen-Poiseuille equation for simple porous media is used:  8822 subkandxPJ  [4-21] Where η is the fluid viscocity, and Δx can be interpreted as the substrate thickness δ. The pore structures of substrates and porous membranes are generally more complex than the straight cylindrical pores assumed in the Hagen-Poiseuille model, and relations such as the Kozeny-Carmen equation or Knudson flow for small pore size is typically used to describe transport through porous membranes. Discussions of flow in porous media are beyond the scope of the present work.  Returning to mass transport of vapour through membranes in the present system, the observed mass transport coefficient (kobs) can be determined from the flux measured through the experiment (Equation [4-4]) and the bulk vapour pressures on the feed and sweep sides of the membrane, which can be converted to concentrations assuming the ideal gas law applies (the tests are at near atmospheric conditions):   bulksweepbulkfeedobsOH CCkJ ,,2   [4-22] The boundary layer resistance (Rfeed + Rsweep) can be determined from mass transfer correlations for known geometries or experimentally by measuring the observed transport resistance at a set of operating conditions and flows for a series of similar membranes of varying thickness. This is discussed in greater detail in Section 4.4 below. With the boundary layer resistance determined and the observed total resistance measured experimentally, the membrane resistance and membrane mass transport coefficient can be calculated from Equation [4-16] and [4-18]. Finally, the effective average vapour concentration difference at the membrane surfaces can be calculated from: 79    mOHmemmemsweepmemfeedkJCCC 2,,   [4-23] Which can be converted to a surface vapour pressure difference to determine the ‘actual’ membrane water vapour permeance:    memOHvOHmemsOHvmemfOHvOHactualOHpJppJP,2,,,,,,,22222 [4-24] 4.4 Boundary Layer Resistance The vapour transport module consists of seven channels with rectangular cross-sections as summarized in Table 4-1. Table 4-1: Geometric properties of permeation module Parameter Value Unit Number of channels 7 - Channel length 16 cm Channel width 3 mm Channel depth 1 mm Total Area 33 cm2  The test module is a counter-flow permeation module designed for heat and moisture transport measurements during the author’s master’s thesis [101]. This permeation module is used for air, oxygen permeation, carbon dioxide, and water permeation measurements. An isometric view of the module is shown in Figure 4-3a [102]. 80    Figure 4-3: a. Module used for permeation experiments [102]; b. CFD analysis of flow distribution in module [101]. The module comprises of two half shells with machined flow field pathways, the two sides are placed on either side of a membrane sample and sealed by compression. Figure 4-3b is a computational fluid dynamics (CFD) analysis of air-flow through the module under standard conditions, this CFD analysis was not validated, but it provides an indication that the flow distribution in the module channels was well distributed. The module contains an entrance region in which fluid flow is distributed into each of the channels. Fluid flow in channel ducts as well as heat and mass transfer in these ducts cans be described by a series of dimensionless correlations. The Reynolds number (Re) describes the ratio of momentum to viscous forces of air flowing in the channel:   hdRe  [4-25] where ρ is fluid density, μ is the fluid dynamic viscosity, v is the fluid average velocity, and dh is the hydraulic diameter of the duct:   hwwhdh24 [4-26] where w is the channel width, and h is the channel height. The average fluid velocity in the channel is determined by:  whQv   [4-27]  where Q is the volumetric flow in the channel. 81  The Schidmt number (Sc) describes the ratio of convective and molecular diffusive forces in the fluid:  abDSc  [4-28] Where Dab is the diffusivity of the component of interest in the carrier fluid, in this case the diffusivity of water vapour in air. The diffusivity, viscosity, and density of air-water vapour systems will be dependent on the temperature and vapour pressure in the stream [136]. All testing is completed isothermally, so these values can be determined for the system through the following equations for air containing water vapour. This is important to consider as viscosity and density change significantly with vapour pressure and temperature and so the fluid velocity of air in the channel will change with the operating conditions. As well diffusivity changes significantly with temperature. For diffusivity of water in air (m2/s) the following empirical equation can be used [66]:   81.15215.27315.2731019.2   TTD AirOH  [4-29] For density of moist in air (kg/m3) the following empirical equation can be used:  1013253783.015.2732929.1 vpBT  [4-30] Where B is the atmospheric pressure, and pv is the vapour pressure. For viscosity of dry air (Pa s) the following empirical equation can be used up to 54°C:    62 100004.0067.01.17  TT  [4-31] However, for the calculations look-up tables were used for diffusivity, density, and viscosity for moist air at the test temperatures and humidity. The Sherwood number (Sh) describes the ratio of total mass transport in the channel to the diffusive mass transport:  abhDkdSh   [4-32] Where k is the mass transfer coefficient in the channel associated with the fluid boundary layers at the membrane surface. The inverse of these mass transport coefficients can be used to determine the feed and sweep side transport resistances in equation [4-18]. Sherwood 82  numbers can be described by correlations involving the Reynolds and Schmidt numbers and geometric properties of the system, the channel average Sherwood number generally can be described by:  chbaLdScSh  Re  [4-33] where β, a, b, and c are constants that must be empirically fit to the geometric system from transport experiments, and L is the channel length. For developing concentration in rectangular ducts in laminar flow where the channel length is greater than the length of the developing region the channel average Sherwood number described by the following equation has been reported to be accurate [2].:  33.033.033.0Re988.1 LdScSh h  [4-34] Alternatively the Wilson plot methodology has been shown to be effective for determining the membrane mass transport coefficients from experimental transport measurements in a module at different flow velocities [63]. Here the total mass transport resistance in the gas boundary layers at the membrane surface:  cmobs vakk111  [4-35] Linear regression is used to determine the coefficients a and c which best fit the experimental data, where the intercept is 1/km.  The boundary layer resistance in a module can also be determined through a similar method by using membranes of the same composition with different thicknesses [93], where a plot of the observed resistance (Robs) against the thickness is fit by linear regression, where the intercept is the total boundary layer resistance:  mblobs Pkk11 [4-36] A microporous substrate consisting of silica and polyethylene was used as a reference material to determine the boundary layer resistances in the module. In this analysis up to three layers of membrane were stacked together to estimate the boundary resistances at a series of flow rates. 83  An empirical fit of performance data was done to determine the parameters (β, a, b, c) in equation [4-33] by minimization of the sum of the squared residuals, the results are compared against the mass transport resistances in the boundary layers as predicted by equation [4-34], and the results are summarized in Table 4-2. Table 4-2: Coefficients for Sherwood number correlations for a water vapour transport module at 50°C. Flow (sL/min) Feed /Sweep Channel Velocity (m/s) Experimental Result (s/m) Least Squares Analysis of Eqn [4-33]b (s/m) Eqn. [4-34]c (s/m) Total, Rbl Feed Rbl,f Sweep Rbl,s Total Rbl Feed, Rbl,f Sweep, Rbl,s Total Rbl 6 5.8/5.6 37.7 19.0 19.2 38.1 18.1 18.2 36.3 8 7.7/7.5 33.3 16.4 16.5 32.9 16.4 16.6 33.0 10 9.6/9.4 30.4 14.6 14.7 29.3 15.3 15.4 30.7 12 11.5/11.2 25.6 13.3 13.4 26.7 14.4 14.5 28.8 a. sL/min (controlled via mass flow controller, standard conditions 0°C, 101.325 kPa) b. β=1.907, a=0.513, b=0.357, c=0.567; SSR=2.7 c. β=1.988, a=0.333, b=0.333, c=0.333; SSR=12.2  Figure 4-4: Linear regression line with 95% confidence interval for the regression for the boundary layer model based on the least squares analysis of the fitting parameters for 84  equation [4-33] against the experimental boundary layer measured by changing substrate thickness for flows between 4 and 12 sL/min. The fitting parameters along with equation [4-33] have a better overall fit for the module used compared to assumptions for rectangular ducts in equation [4-34] since the SSR values for the experimental results against the model results are lower for equation [4-33]. The model can be used to predict the boundary layer resistances at various flow, temperatures, and feed and sweep relative humidity set points, so that estimated boundary layer resistances can be determined for specific testing conditions used for both feed and sweep streams. From Table 4-2, it can be observed that flow has a strong effect on the boundary layer resistance, whereas the difference between the feed the sweep stream is relatively insignificant. The difference between the feed and sweep is related to the average relative humidity in the stream and its effect on the fluid properties of the air in the stream, the results show that this has little overall impact on the resistance, and the maximum error in the boundary layer calculation associated with ignoring the water vapour activity in the feed and sweep streams is less than <3% at 70°C, and approximately 1% at 50°C. However, the average stream water vapour activity is included in the model for completeness. The boundary layer resistances determined by the model at different flows and temperatures and are summarized in Figure 4-5, temperature has an effect on the boundary layer resistance as temperature not only effects the fluid properties of the air streams, but also the diffusivity of water vapour in the air, errors of about 22% would be expected by neglecting to account for temperature between 25 and 70°C.  Figure 4-5: Total feed and sweep boundary layer resistance predicted from model at different temperatures and flow rates (aH2O,feed = 0.5, aH2O,sweep=0.0). 85  With the boundary layers accounted for, a porous membrane should have equivalent permeance at all temperature, flow, and feed stream activity conditions. A series of tests were completed on various porous membranes at different conditions to determine the accuracy of the modelled boundary layer predictions. This was done by comparing the modelled boundary layer against the calculated boundary layer resistance based on the expected permeance of each porous material. It was found that at the lowest flow rate (2.5 sL/min) and at very low feed activities (aH2O=0.0 to 0.20) the modelled boundary layer results differed significantly from the model, possibly due to measurement accuracy limits on the test apparatus at low flow and activity. These results were not included in the analysis and all further membrane testing was completed at stable conditions (4 to 15 nL/min, aH2O > 0.20, T=25 to 70°C). The results for several materials at a number of conditions in this range are shown in Figure 4-6, where the model boundary layer resistance is plotted against the measured boundary layer resistance. It was determined that the empirical model predicted the boundary layer resistance sufficiently for the purposes of this study as the regression line of the observed data plotted against the model data is linear, with a slope near one and an intercept near zero. Since this model can be used, it allows membranes to be measured at a single test condition without having to complete a full flow and temperature analysis for each material.  Figure 4-6: Comparison of the predicted and observed boundary layer resistances based on the empirically predicted boundary layers for the test module for various porous materials at various conditions [4 to 15 nL/min, aH2O > 0.20, T=25 to 70°C]. 86  4.5 Substrates Structure and Performance In this study, three substrates were used to make composite membranes. Mercury intrusion porosimetry was used to analyze the pore structure of the substrates, the porosity of the and the pore size distribution of the substrates are shown in Figure 4-7. The structural and composition data for each substrate is summarized in Table 4-3. The silica loaded polyethylene substrates are similar in structure, except that SiO2-PE2 is a thicker material, with a slightly smaller average pore size. The polypropylene based substrate has a different microstructure, is much thinner, and has a narrow pore size distribution averaging around a slightly larger average pore size than the silica-polyethylene materials. Surface images of the SiO2-PE1 and polypropylene (PP) materials are shown in Figure 4-8, which clearly shows the microstructural difference between the substrates. The silica polyethylene (SiO2-PE) materials are comprised of a high loading of precipitated silica (~60% from TGA) held together by polyethylene, creating a microporous structure. Whereas the polypropylene material has a pore structure created via a dry stretching process [137].  Table 4-3: Properties of substrates used to make composite membranes, errors reported are sample standard deviations for measurement of 18 samples of SiO2-PE1, 18 samples of SiO2-PE2, and 6 samples for PP. Substrate SiO2-PE1 SiO2-PE2 PP Porosity (cm3/cm3) 0.49 0.48 0.44 Average Pore Diameter (μm) 0.038 0.023 0.095 Thickness (μm) 105 145 25 Basis Weight (g/m2) 53.8±1.6 93.6±1.5 10.1±0.1 Material Composition Silica Polyethylene Silica Polyethylene Polypropylene Area Surface Roughness, Sq (μm) 2.72 2.24 0.16 Average Peak to Valley Distance, Sz (μm) 15.0 15.5 1.9 Performance Properties  Air Permeability (cm3/cm2/s/Pa) 1.36×10-5 ±3.36×10-7 2.93×10-6 ±1.24×10-7 2.39×10-5 ±1.94×10-6 Water Vapour Flux (mol/m2/s)a 1.71×10-2 ±4.56×10-4 1.03×10-2 ±2.12×10-4 2.09×10-2 ±2.89×10-4 Apparent Water Vapour Permeance (GPU) 11697 ± 432 6068 ± 152 15640 ± 335 Corrected Vapour Permeance (GPU) 19558 ± 1164 7664 ± 243 33757 ± 1576    a Water vapour transport is measured at T=50°C, af=0.5, as=0.0, v1=v3=5.8 m/s. 87  The average performance properties of the substrates are also summarized in Table 4-3. The silica polyethylene substrates are thicker and have smaller average pore size and so the pressurized air permeability is lower relative to the polypropylene substrates. Similarly, the water vapour flux and permeance are lower for the silica polyethylene materials. The apparent permeance is the vapour flux normalized by the average pressure differential across the membrane samples in the test module. The corrected permeance accounts for the boundary layer vapour transport resistance associated with the crossflow test module at the given conditions. As the vapour transport performance of the materials increase, the effect of the boundary layer becomes a more significant portion of the total resistance to transport. As such, the difference between the apparent and corrected permeance increases with increasing water vapour permeance as observed when comparing the various substrates in Table 4-3. This is import to note, as often in testing of water vapour transport in materials, the boundary layers are not accounted for, and permeance can reported as artificially low. The error associated with neglecting the boundary layer is much greater for higher permeance materials. It has been noted in numerous studies the importance of considering boundary layers for water vapour transport measurements, in particular errors associated with stagnant boundary layers in the ASTM E96 “cup test” are common and dynamic vapour transport testing with air flow on both sides of the membrane is recommended [77], [81]. All further permeance measurements reported in this study are the ‘corrected’ water vapour permeance, unless stated otherwise. Water vapour flux results for SiO2-PE1 porous substrate and the PP porous substrate tested at various temperatures and constant feed stream relative humidity (aH2O=0.5) are summarized in Figure 4-9. The flux increases with increasing temperatures with both porous materials as the actual vapour pressure vapour pressure difference between the feed and sweep streams is increasing with temperature. Normalizing for this differential, the permeance remains constant at all temperatures. This is expected as the materials do not interact significantly with water vapour and the pore structure remains unchanged over the temperature range. However, these results again illustrate that the boundary layers at the membrane surfaces must be accounted for and the corrected permeance should be reported especially when working with materials with high water vapour transport. 88   Figure 4-7: Pore size distributions of three substrates from mercury intrusion porosimetry.  Figure 4-8: Substrate surfaces at 5kx and 20kx (inset) magnification, silica-polyethylene (SiO2-PE1, left), polypropylene (PP, right). 89   Figure 4-9: Flux, apparent permeance and boundary layer ‘corrected’ or actual permeance for two microporous substrates (SiO2-PE1, left and PP, right) at various temperatures (aH2O=0.5); data is for single material samples. 4.6 Composite Membranes Composite membranes are made by coating a thin dense polymer layer on one surface of a microporous substrate, the substrate acts as a carrier for the thin film barrier layer made of cross-linked PEO-PU polymer. The dense film layer provides selectivity for water vapour over other gases in the membrane. Example images of cross-sections this type of composite membrane are shown in Figure 4-10, these types of membranes are in commercial use for ERV applications and have been reported previously in patents by the present author [26].  Figure 4-10: Cross-sectional images of a commercially available ERV membrane, the ‘composite membrane’ has a thin dense polymer layer on the surface of a microporous support. 90  4.6.1 Silica-Polyethylene Substrate Based Membranes Only the thinner Silica-polyethylene substrate (SiO2-PE1) was used for further study. For the silica polyethylene materials, the variation of the basis weight is over 3% (+/- 1.6 g/m2), this is greater than the dense polymer coating areal weight in many cases (0.5 to 5 g/m2). This means that the coating thickness cannot be accurately determined by gravimetric methods. Instead the film thickness must be determined either by microscopy images or by water vapour transport measurements along with a transport model, and coating and substrate permeability measurements. In order for confirm that water vapour transport models and PEO-PU film permeability measurements could be used, several membranes were produced with different loadings of the PEO-PU polymer (1 to 6 g/m2). Thickness measurements for the silica-polyethlyene were taken by measuring the coating thickness from freeze-fractured cross-sections using SEM. A minimum of three cross-sections from various locations in a sample were analyzed, and a minimum of three images per section were used, with a total of at least 10 measurements per image. An example image measurement is shown in Figure 4-11.   Figure 4-11: Example of a coating thickness measurement for a SiO2-PE1-PEO-PU membrane image. The membranes that were imaged had been tested for water vapour transport, and so coating thickness measurements could be used to determine the correlation between SEM measured thickness and the predicted thickness from water vapour transport measurements. In order to do so the substrate permeance had to be measured for each material prior to applying the film layer. Also the permeability of the PEO-PU film had to be known as well. The substrate permeance as measured for each substrate, is reported in Table 4-4 along with the membrane permeance after the PEO-PU coating layer was applied. A summary of the permeability for 91  three PEO-PU films of three thicknesses (32, 60, and 95 micron) in the range of aH2O = 0.4 to 0.95 at 50°C is summarized in Figure 4-12. The substrates and then the coated membranes were tested isothermally at 50°C and feed activity of aH2O=0.5. From Figure 4-12 the film permeability should be in the range of 45000-55000 Barrer at these conditions, and a regression analysis was used to find the film thickness that minimized the sum of squared residuals for the measured and predicted thicknesses from the vapour transport model, a SSR value of 1.7 was determined for a permeability of 47250 Barrer, and a linear plot of the predicted versus measured thicknesses is shown in Figure 4-13. The SEM measured film thickness and film thickness measured by water vapour transport are also summarized in Table 4-4. The results indicate that the film layer thickness can be estimated by the water vapour transport model since the linear regression line for the measured to predicted has slope near one and an intercept near zero. The SEM thickness measurements tend to have a high level of variability due to only measuring thickness at only a few discrete points, representing only a small portion of the membrane surface. The thickness as determined from water vapour transport measurements represents the effective thickness for the entire area of the membrane being tested, and is the preferred method for film thickness determination. Table 4-4: Summary of the substrate and membrane  Membrane Substrate Permeance (GPU) Membrane Permeance (GPU) SEM Film Thickness Film Thickness from WVT SiO2-PE1-PEO-PU-1 23420 6156 5.5 5.7 SiO2-PE1-PEO-PU-2 23420 7479 3.8 4.3 SiO2-PE1-PEO-PU-3 23420 8345 4.2 3.6 SiO2-PE1-PEO-PU-4 23420 7675 3.5 4.1 SiO2-PE1-PEO-PU-5 23420 9268 2.9 3.1 SiO2-PE1-PEO-PU-6 21300 10270 3.1 2.4 SiO2-PE1-PEO-PU-7 22545 14243 1.1 1.2 SiO2-PE1-PEO-PU-8 19800 8915 2.8 2.9 SiO2-PE1-PEO-PU-9 19980 10870 2.6 2.3 SiO2-PE1-PEO-PU-10 22590 14147 1.3 1.0  92   Figure 4-12: PEO-PU film water vapour permeability for three levels of film thickness at 50°C, data points are the sample mean of 20 data points at each activity, error bars are sample standard deviations about the mean for 20 data points at each activity for each film.  Figure 4-13: Comparison of the film layer thickness as measured by SEM and the film layer thickness as predicted by water vapour transport measure for a series of composite membranes consisting of a PEO-PU film layer coated on SiO2-PE1 substrates, error bars are the sample standard deviation based on 30 thickness measurements for each sample. 93  The boundary layer, substrate, and film layer water transport resistances for the SiO2-PE1-PEO-PU membranes are summarized in Table 4-5. As a percentage of resistance, the boundary layers are significant when testing these high permeance membranes. As the water vapour permeance increases the boundary layer becomes more significant, representing up to greater than 30% of the resistance in the water transport measurement. As the film layer thickness decreases, the percentage of resistance of the substrate also becomes more significant, representing nearly 50% of the membrane resistance in membranes with ~10000 GPU performance.  Table 4-5: Water vapour transport resistances in SiO2-PE1-PEO-PU membranes at 50°C, aH2O,feed=0.5. Film Layer (µm) Membrane Water Vapour Permeance (GPU) Water Vapour Transport Resistance (s/m) Percentage of Resistance in WVT Test (%) Percentage Resistance in Membrane (%) BL Sub Film BL Sub Film  Sub Film 1.0 14147 38.1 55.6 22.9 32.7 47.7 19.7 70.8 29.2 1.2 14243 38.1 49.3 28.7 32.8 42.4 24.7 63.2 36.8 2.3 10870 38.1 49.2 53.1 27.2 35.1 37.8 48.1 51.9 2.4 10270 38.1 52.2 56.1 26.0 35.7 38.3 48.2 51.8 2.9 8915 38.1 56.2 68.6 23.4 34.5 42.1 45.0 55.0 3.1 9268 38.1 47.5 72.5 24.1 30.0 45.9 39.6 60.4 3.6 8345 38.1 47.5 85.7 22.2 27.7 50.0 35.6 64.4 4.1 7675 38.1 47.5 97.4 20.8 25.9 53.2 32.8 67.2 4.3 7479 38.1 47.5 101.2 20.4 25.4 54.2 31.9 68.1 5.7 6156 38.1 47.5 133.1 17.4 21.7 60.9 26.3 73.7  In order to improve the membrane performance, either the film layer thickness must be further decreased, or else the SiO2-PE1 substrate must be replaced with a higher performing substrate material. Decreasing the film thickness to below 2 microns becomes challenging with the SiO2-PE1 substrate due to its high level of surface roughness. As shown in Figure 4-14, and summarized in Table 4-3, the structure of the SiO2-PE substrate has a greater area surface roughness (2.72 µm) when compared with the PP substrate (0.16 µm) and  the SiO2-PE substrate has a maximum peak to valley distance of 15 microns compared to 1.9 microns for the PP substrate. This leads to defects in the coating layer especially for coatings much thinner than about 2.5 microns, which will compromise the gas and contaminant barrier 94  properties and the selectivity of the membrane. A coating defect in the PEO-PU film on the silica-PE substrate is shown in Figure 4-15.  Figure 4-14: Substrate surface roughness images from laser confocal microscopy at 50× magnification.  Figure 4-15: SEM Surface images of the PEO-PU film layer on a coated silica-PE1 substrate, showing film layer non-uniformities due to the underlying substrate (left) and a higher magnification image of a defect in the PEO-PU barrier layer due to substrate surface roughness. 4.6.2 Polypropylene Substrate Based Membranes The polypropylene substrate has a higher water vapour permeance due its lower thickness, so it is a promising candidate for a higher performance membrane material. The substrate also has relatively low surface roughness which should allow thinner defect-free barrier film layers to be coated on the substrate surface. Furthermore, since the substrate basis weight is low, and variability of the basis weight is also lower, the thin layer thickness can be accurately estimated by gravimetric methods. A series of coating thickness were applied to polypropylene 95  substrates the water vapour permeance of each membrane was measured, the film and substrate resistances were determined for each membrane, and the results are summarized in Table 4-6. However, all these tests are at elevated temperature (50°C), when the temperature is reduced to 25°C the performance changes significantly as shown in Table 4-7. At 25°C, the membrane permeance is over an order of magnitude lower than at 50°C, the film layer resistance to transport has increased dramatically, and the film layer now accounts for nearly all of the membrane resistance.  Table 4-6: Water vapour transport resistances in PP-PEO-PU membranes at 50°C, aH2O,feed=0.5. Film Layer (µm) Membrane Water Vapour Permeance (GPU) Water Vapour Transport Resistance (s/m) Percentage of Resistance in WVT Test (%) Percentage Resistance in Membrane (%) BL Sub Film BL Sub Film  Sub Film 0.7 20921 38.1 36.3 16.8 41.8 39.8 18.5 68.3 31.7 1.5 16198 38.1 36.3 32.3 35.7 34.0 30.3 52.9 47.1 2.7 9635 38.1 36.3 79.1 24.8 23.6 51.5 31.5 68.5 2.9 9210 38.1 36.3 84.4 24.0 22.9 53.1 30.1 69.9 3.6 7556 38.1 36.3 110.8 20.6 19.6 59.8 24.7 75.3  With the higher performance substrate, the resistance associated with the substrate is now only ~30% for a 10000 GPU membrane, as opposed to nearly 50% for the Si-PE1 based membrane at the same performance. Furthermore, since the coatings can be thinner on the PP substrate higher performance membranes can be produced. However, all these tests are at elevated temperature (50°C), when the temperature is reduced to 25°C the performance changes significantly as shown in Table 4-7. At 25°C, the membrane permeance is over an order of magnitude lower than at 50°C, the film layer resistance to transport has increased dramatically, and the film layer now accounts for nearly all of the membrane resistance.  96  Table 4-7: Water vapour transport resistances in PP-PEO-PU membranes at 25°C, aH2O,feed=0.5. Film Layer (µm) Membrane Water Vapour Permeance (GPU) Water Vapour Transport Resistance (s/m) Percentage of Resistance in WVT Test (%) Percentage Resistance in Membrane (%) BL Sub Film BL Sub Film  Sub Film 0.68 2090 42.8 37.6 539 6.9 6.1 87.0 6.5 93.5 1.47 658 42.8 37.6 1795 2.3 2.0 95.7 2.1 97.9 2.73 344 42.8 37.6 3465 1.2 1.1 97.7 1.1 98.9 2.9 234 42.8 37.6 5104 0.8 0.7 98.4 0.7 99.3 3.58 1137 42.8 37.6 1022 3.9 3.4 92.7 3.6 96.4  Interestingly, this temperature effect does not appear to be present in the SiO2-PE1-PEO-PU membranes. The water permeance performance of the SiO2-PE1 and PP substrates and coated membranes made from both substrates are reported at various temperatures in Figure 4-16. The substrates both demonstrate no change in performance with temperature, and the PEO-PU coated SiO2-PE1 substrate in shown in Figure 4-16a shows no change in permeance with temperature. However, the membranes made with a PP substrate coated with the same PEO-PU polymer show a significant temperature ‘switch’ effect in Figure 4-16b. In Below 50°C the permeance of the membrane is low, but between 45 and 50°C, the water vapour permeance of the material increases by an order of magnitude in Figure 4-16b. Polymer materials which respond to external stimuli have been well reported in the literature [138]. Furthermore, polyurethane based polymer materials with temperature switching permeance have been reported with soft segments of polycaprolactone (PCL), polypropylene glycol, and polyethylene glycol [139]–[141]. For example, ‘smart’ materials, based on polyurethanes containing soft-segments of PCL demonstrated ‘switch’ temperatures in the range of 50 to 60°C going from 250 g/m2/day to 3000 g/m2/day water flux [142]. In the referenced study, when the flux results reported are normalized by the vapour pressure in at different temperatures in the test, the actual ‘switch’ in permeance is only from about 300 to 500 GPU (or 20000 to about 28000 Barrer when normalized for thickness to give a permeability number). In the case of PEO-PU used in the present study on the polypropylene substrate, the difference in permeance above and below the switch temperature is much greater (~1800 GPU to about 15500 GPU, or 2700 to 45000 Barrer when normalized for film layer thickness permeability). Although these phenomena are of interest for further study for various 97  applications, in the application of the present study, high water vapour transport is generally desirable across all temperature conditions, and the low permeance values found in the PP based membrane at ambient conditions would not be desirable. However, the fact that the same PEO-PU polymer shows no switch with temperature when coated on the silica-polyethylene deserves further discussion.  Figure 4-16: Temperature effect on water vapour permeance in (a.) SiO2-PE1 substrate and coated PEO-PU coated SiO2-PE1 membrane; (b.) PP substrate and PEO-PU coated PP membrane; data is for single material samples. 4.7 Effect of PEO Crystallization The switch in permeability observed in the polyurethanes can be associated with glass transitions in the polymer as well as melting of crystalline portions in the polymer [141]. In Figure 4-16b it is likely that the increase in permeance above 45°C is caused by the melting of crystalline PEO groups in the polymer. Crystalline PEO soft segments in the polymer will have very low water vapour permeability, once the PEO melts to the completely amorphous state, water can readily transport through the mobile PEO groups. Due to a large portion of the total PEO in the polymer being in side chains, and cross-linking only occurring in the hard segments of the polymer, the PEO groups can phase separate into PEO rich regions, where crystallization may occur. Samples of the cast of films of the PEO-PU polymer alone show visible crystalline structure when left at ambient conditions for extended periods of time. Polarized light optical microscopy images of cast PEO-PU films (200 to 500 micron thickness) left at ambient conditions for several weeks are shown in Figure 4-17. 98   Figure 4-17: Polarized light optical microscopy images of cast PEO-PU films showing crystallinity and phase separated regions. DSC heating scans of the cast PEO-PU polymers at three levels of cross-linking are shown in Figure 4-18 of the polymer as cast. All samples show a melting events centering near 50°C, although at the highest level of cross-linking the melting peak is more complex, this may be due to the cross-linking constraining polymer microstructural rearrangement.   Figure 4-18: DSC first heating of PEO-PU polymers films as cast at three levels of cross-linking.  Heat Flow (W/g)25 30 35 40 45 50 55 60Temperature (°C)                 PEO-PU-0%XL–––––––                 PEO-PU-6.7%XL– – – –                 PEO-PU-16.7%XL––––– ·Exo Down99  DSC of the PEO-PU polymer with 6.7% weight of polycarbodiimide cross-linker over three heating cycles (20°C/min) is shown in Figure 4-19. In the first cycle the polymer as cast and dried is cooled to -20°C before heating to 150°C, a melting peak for the PEO soft segments is found with a peak at 51.7°C, a long endotherm between 70 and 150°C represents the evaporation of excess water held in the polymer. The polymer was then cooled (10°C/min) to -85°C before the second heating cycle (20°C/min). In the second heating cycle, the PEO soft segment glass transition is observed centered at -52.5°C, a melting peak associated with the recrystallized PEO groups is now found as a broad peak at 33.6°C, this has likely shifted due to restructuring of the PEO soft segments after the previous melting. The ‘as cast’ samples tested in the first heating cycle have essentially been annealed at ambient conditions near but below the PEO melting temperature which may explain why the initial PEO melting peak is higher [143]. The ‘hard’ PU region of the polymer melts with a sharp peak at 232.6°C. The polymer was then cooled again (10°C/min) to -85°C and a third heating cycle is completed at 20°C/min. The PEO segment glass transition is centered at -55°C, and a complex multiple melting peak for the PEO groups is found with peaks at 18.2, 32.3, and 39.7°C. No significant hard segment melting peak is observed up to 250°C. After the hard segments melt at 232°C in the second heating, the polymer is free to restructure and recrystallize on cooling leading to the complex peak in the range of 18 to 40°C. For a reference, the ‘soft’ side-chain PEO polyol (Tegomer® D3403) used as a monomer in the PEO-PU manufacture was also tested using DSC, this material is a ‘waxy’ like solid at ambient conditions, with a molecular weight of 1200 g/mol, and ~23 PEO groups in the polymer, such that 90% of the weight of the material is PEO groups. The first heating of a sample of this material conditioned at room temperature (dry) after cooling to -20°C is shown in Figure 4-20. The structure of the peaks shows some similarity to the PEO segment melting in the third heat of the PEO-PU. The enthalpy of melting (ΔHm) of the Tegomer polyol is 157.7 J/g, which considering the heat of fusion of perfect crystalline PEO (ΔHf°=188.9 J/g ), and that 90% of the weight is PEO, the degree of crystallinity in the polyol at atmospheric temperature is about 93%. Accordingly, the presence of crystallinity in the PEO-PU copolymer based on this material is not surprising. The complexity of the thermal analysis of the PEO-PU copolymer arises from the ability of these segmented polyurethanes to undergo microphase separation into hard and soft domains, in which the soft PEO domains can contain both amorphous and crystalline domains, the hard PU segment domains may also be crystalline or non-crystalline, and also contain intermediate regions containing intermixed PU and PEO segments can also be found 100  [144]. The morphology of segmented polyurethane can obviously be complex, and numerous studies have analyzed the effects of polymer compositions and thermal treatment on the microphase separation in polyurethanes [145]–[150]. A study mentioning crystallization in ‘side chain’ PEG-PU showed that longer chain lengths were shown to have a significant impact on the degree of crystallinity in the side chains [151].  Figure 4-19: First, second, and third heating cycles for a PEO-PU polymer with 6.7% crosslinking, heating rate is 20°C/min and cooling rate is 10°C/min.  Heat Flow (W/g)-70 -50 -30 -10 10 30 50 70 90 110 130 150 170 190 210 230 250Temperature (°C)                  PEO-PU-1st Heating–––––––                  PEO-PU-2nd Heating– – – –                  PEO-PU-3rd Heating––––– ·Exo Down101   Figure 4-20: DSC first heat of as received of PEO-side chain polyol (Tegomer® D3403) used in PEO-PU polymer at 20°C/min heating rate. For the purposes of this study what is important is that these polymers can crystallize at ambient conditions. The PEO melting peak in the ‘as cast’ polymer from the first heating cycle in Figure 4-19 has a total integration enthalpy of only 1.8 J/g, considering the PEO-PU polymer is about 70% wt. PEO with 6.7% crosslinking, the degree of crystallinity is only 1.5% in the (300-500 micron) films that were left at ambient conditions for several weeks after drying and curing at 50°C. It would not be expected that such a small degree of crystallinity in the PEO segments would have a great impact on water vapour permeability. However, more rapid crystal growth would be expected in thinner films cast on the microporous substrates in this study, or a crystalline region at the polymer-substrate interface could become a barrier to transport. This accounts for the presence of the PEO crystallinity on the PP substrate. However, if the PEO-PU polymer has a tendency to crystallize, an explanation for why the crystallization of PEO segments in the polymer coating does not occur on the silica-PE based substrates is still required. DSC measurements for crystallinity work well for bulk polymers, but when the polymer is cast in thin layer (0.5 to 3 g/m2) on a substrate (10 to 60 g/m2), the polymer film layer only represents a small fraction of the total weight of the membrane, and DSC analysis of the film layer is not readily achievable. In order to study the presence of crystalline PEO in the 46.71°C43.12°C157.7J/gHeat Flow (W/g)0 20 40 60 80Temperature (°C)                 Tegomer D3403–––––––Exo Down102  membrane, ATR-FTIR was used to evaluate the film layer in the membranes. Various studies have indicated that the amorphous and crystalline phases of PEO have distinctly different absorbance peaks in infrared spectroscopy studies. Specifically, in IR studies, crystalline PEO has two distinct peaks appearing as a sharp doublet at 1359 and 1343 cm-1 associated with wagging motions of CH2 in PEO crystal regions, whereas in amorphous PEO, only one broad peak is found centered at 1349 cm-1 [152], [153]. FTIR of the PEO-PU polymer in the coated membrane surfaces are shown in Figure 4-21. The SiO2-PEO-PU membrane shows a broad peak centered at 1349 cm-1 indicating that the PEO is in the amorphous state, the PEO-PU polymer alone shows peaks at 1359 and 1343 cm-1 indicting the presence of crystalline PEO, and in the PP-PEO-PU membrane peaks are found near 1359 and 1343 cm-1, with some portion of the absorption in the middle range, indicating both crystalline and amorphous PEO. This demonstrates that crystalline PEO is prominent on the PP based membrane, whereas amorphous PEO is prominent on the SiO2-PE based membrane. This agrees with the permeance data in Figure 4-16 where the crystalline PEO in the PP based membrane inhibits permeability in the PEO-PU at temperatures below the PEO segment melting range.  The presence of silica in the substrate creates a rough surface, which should provide more sites to initiate crystallization than the smooth PP surface. However several studies on silica-PEO/PEG systems (2000 to 20000 g/mol) have shown significant inhibition of crystallization in which no melting enthalpy is present below 50% PEG, proposing that interactions between the PEG and silica cause the PEG to be confined via adsorption on the silica surfaces [154], [155]. No studies were found which report this effect on coated membrane materials, where the silica is only found at the interface, however this effect may be at play in the present systems.  103   Figure 4-21: ATR-FTIR of SiO2-PE-PEO-PU membrane, PEO-PU film, and PP-PEO-PU membrane. It should be noted that the low performance of the PP-PEO-PU membranes was time dependent, and initial water vapour permeance in the membranes at 25°C was as expected, but within a few hours to a few days, the performance had decreased to the levels observed in Figure 4-16. This indicates that the crystallization in the film layers of the membranes is a time dependent process, which would be expected for crystallization just below the PEO melting temperature. It was found that the level of cross-linker (0, 6.7, 16.7%) did not affect the occurrence of PEO crystallization, indicating a morphological time dependent process was responsible for the performance loss over time instead of a chemical reaction time dependent process dependent on the cross-linking reaction. Since membranes are produced by casting polymer dispersions onto the substrate surface, the film formation process applies to the membrane fabrication process. The film formation process for the present system is summarized in Figure 4-22. In film formation the polymer dispersion initially wets the substrate. As water is evaporated from the dispersion, polymer particles begin to pack together and interact. As more water evaporates the particles deform and begin to inter-diffuse as particle boundaries break down. After further evaporation hard and soft segment phase separation occurs, crosslinking is initiated after sufficient drying, and the hard segments crystallize below their melting temperature. Soft segment crystallization can occur after extended periods of temperature below the soft segment melting temperature.  104   Figure 4-22: Aqueous dispersion coating process with film formation. During the film forming process, the PEO-PU polymers are free to orient themselves into microphase morphologies. At the interface of the polymer coating and the substrate, preferential ordering of the PU and PEO groups in the polymer may occur depending on the affinity of each group for the substrate. Both substrates are considered hydrophobic from contact angle measurements as shown in Figure 4-23, where the contact angle for water on the SiO2-PE substrate is 120° and for the PP substrate 102°, so substrate hydrophilicity cannot be a significant factor. However, wettability of PEG-400 on the substrates differs: on the SiO2-PE substrate, PEG-400 readily wets the substrate imbibing into the pores, whereas on the PP substrate PEG-400 remains on the surface as a droplet. This would indicate that during microstructural re-organization during the film forming process that the soft PEO groups in the PEO-PU polymer may preferentially interact with the SiO2 substrate and not the PP substrate. This may prevent the initialization of crystallization at the substrate coating boundary during film formation.    105   Figure 4-23: Contact angles of water on SiO2-PE (left, 120°) and PP substrate (right, 102°). 4.8 Elimination of the Thermal Switching Effect Interestingly, the application of liquid water to the PEO-PU layer on the PP substrate based membranes, appeared to eliminate the crystallinity. Exposures of even a few seconds to liquid water appeared adequate to disrupt the crystallinity, and return the PEO membrane to the amorphous state. The FTIR results in Figure 4-24 shows that the PEO groups in the PEO-PU polymer on the PP substrate return to the amorphous state after a one second exposure to liquid water.  Figure 4-24: FTIR of PP-PEO-PU membrane as cast and after short exposure to liquid water. This indicates that the crystallinity is disrupted by liquid water interacting with the PEO groups in the polymer, similar effects are observed in PEO-PU polymers films in DSC studies. In 106  Figure 4-25a, PEO-PU polymer is shown as cast, cured, and dried with crystalline PEO melting present, and a sample from the same film after exposure to liquid water for a short time (6 s) to show that exposure to liquid water eliminates the crystallization. In Figure 4-25b a PEO-PU polymer film which has been exposed to liquid water is heated to increasing temperatures, up to 150°C to show that the water remains bound to the PEO groups in the polymer preventing recrystallization. This further indicates that the thermal switching phenomena was related to the membrane fabrication/film formation process, in that the effect can be permanently eliminated and that polymer likely reorders into a state that favors PEO crystallization as it is dried from dispersion.   Figure 4-25: DSC heating (20°C/min) of PEO-PU polymer with 6%XL: a. as cast, dried, and cured and then with a short exposure to liquid water; b. after exposure to liquid water, with multiple increasing heating cycles of increasing temperature. A PP-PEO-PU membrane was tested similar to Figure 4-16b, as cast showing the ‘temperature switch’ associated with PEO crystallization. The material was then exposed to liquid water and then tested again at the same conditions, the results can be observed in Figure 4-26. The results show that after drying and curing of the membrane the temperature switch is observed. However, after a short exposure to liquid water, the PEO crystallinity is disrupted as water molecules bond to the PEO groups, and they remain amorphous, and the permeability of the PEO-PU polymer is recovered. The permeance of the membrane exposed to liquid water is returned to the expected level, and is constant at all temperatures.  107   Figure 4-26: Permeance of PP-PEO-PU membranes as cast and after water exposure, at various temperatures (aH2O=0.5). 4.9 Conclusions In this chapter, a resistance in series model for water vapour transport was presented for the test module used for water vapour transport testing. The importance of the transport boundary layers at the membrane surface was demonstrated, especially for higher performance membranes, where the boundary layers can be a significant portion of the resistance to transport. Composite membranes consisting of a microporous substrate layer coated with a thin selective polymer film layer were fabricated and the substrate and film layer resistances were confirmed for a range of film layer thicknesses. The contribution of the substrate resistance to transport was shown to be significant (50% of the total membrane resistance) for composite membranes based on a SiO2-PE based microporous substrate. A higher water vapour transport substrate based on polypropylene was also used to make composite membranes. However, the substrate resistance to transport was still nearly 30% of the total membrane resistance. An unexpected thermal switching of permeability was observed when using polypropylene substrates that was not observed for the SiO2-PE substrate based membranes. The thermal switching occurred above 50°C and was shown to be associated with crystallization and melting of the soft PEO segments of the PEO-PU polymer. It was found that this switching effect could be eliminated by disrupting the crystallinity with liquid water after curing the membrane. Crystallization was also prevented by casting on a silica loaded polyethylene substrate. Testing below 50°C was found to lead to crystallization of the PEO 108  segments of the PEO-PU polymers in some membranes, which decreased the water vapour transport performance. It is recommended that future testing be completed at 50°C in order to ensure that the PEO segments are in the amorphous state and that the degree of crystallinity does not have to be known in order to compare membrane materials based on PEO-PU polymers. Overall, the composite membranes were shown to have acceptable water vapour transport, but the resistance to transport associated with the substrates used is undesirable and the substrates add significant cost to the membrane materials. The following Chapters will discuss a new approach to fabricating water vapour transport membranes which eliminates the use of a microporous substrate layer.   109   Impregnated Electrospun Nanofibrous Membranes 5.1 Introduction Chapter 4 presented work involving in the fabrication and testing of composite membranes for water vapour transport applications. The microporous substrate layers of these membranes were found to add significant resistance to vapour transport in the materials. The central hypothesis of this dissertation is that this resistance can be eliminated (thereby improving water vapour transport performance) by using an electrospun nanofibrous layer to fabricate membranes. This chapter presents a novel type of water vapour transport membrane based on electrospun nanofibers that eliminates the use of the costly and high resistance substrate layer. Although other studies in the literature have used electrospun nanofibers to fabricate membranes for liquid water filtration and other membrane application, this work is the first report of selective membranes for water vapour transport based on electrospun nanofibers.  In Chapter 4, the testing PEO-PU based membranes below 50°C could be affected by the crystallization of the PEO segments of the PEO-PU polymers, so further testing was done at 50°C to ensure that polymer crystallinity did not bias results when comparing membrane test results. Portions of this Chapter have been reported in a peer-reviewed journal article in the Journal of Membrane Science [156]. 5.2 Nanofibrous Materials Nanofibrous materials can be fabricated via the electrospinning process which can generate fibrous materials in the 100-1000 nm diameter range. Electrospinning involves drawing polymer solution from a spinneret needle via an electrostatic potential field. As the polymer jet discharges from the needle tip, the solvent in the solution begins to evaporate; the jet destabilizes and undergoes a whipping motion. This whipping further draws the fiber to thinner diameter before the dry fibers are deposited on the grounded collector. This process can be used to spin fibers from a wide variety of polymeric materials [32], [157], [158]. Due to high porosity and high surface area, electrospun nanofibrous materials have demonstrated use in a broad range of applications [28]. In those end uses involving permeation through the fiber layer, nanofibrous materials have been used in “breathable” performance clothing, air filtration, and water filtration. In breathable clothing it is desirable to rapidly transport sweat produced by the body due to exertion, while preventing wind and external moisture from penetrating the fabric. Studies on nanofibrous materials have demonstrated high water vapour permeation rates for performance clothing [159]. Work on 110  nanofiber based air filtration materials has also established that increased air permeability is related to decreased fiber diameter [36], [160]. Research on electrospun nanofiber ultrafiltration membranes has demonstrated that transport performance may be increased by over a full order of magnitude from previous membranes by using a nanofibrous layer in the membranes [86], [161]. Other recent work has identified electrospun nanofibrous membranes as promising candidates for membrane based osmosis processes [162], [163]. However, no reports demonstrate nanofibrous membranes which contain a dense selective polymer layer for water vapour transport applications. In this study we present a novel method of fabricating enthalpy exchanger membranes by using electrospun nanofibrous mats as scaffolds to be impregnated by water vapour permeable polymers. To the best of our knowledge, this is the first published report of an electrospun nanofibrous membrane (ENM) which contains a dense film selective for water vapour over other gases and of using this type of membrane in water vapour transport or ‘gas-to-gas’ membrane applications [164]. More specifically, this is the first report of an electrospun nanofibrous based membrane for air-to-air plate type enthalpy exchanger applications. We demonstrate a method to combine a support layer, electrospun nanofibers, and water vapour permeable polymers to create impregnated electrospun nanofibrous membranes (IENM) with water vapour permeance and selectivity that is acceptable for energy recovery ventilation applications. In this report, we study how controllable factors in the membrane fabrication affect the final structural and transport properties of the membranes. 5.3 Concept of the Impregnated Electrospun Nanofibrous Membrane for Water Vapour Transport Current high performance membrane materials for ERV devices are based microporous substrates coated with a thin dense water vapour permeable film layer [26]. Improved water vapour permeation performance in these materials is desirable, both to improve recovery efficiency and decrease size and cost of exchangers. Our analysis of composite type membrane materials in Chapter 4, indicated that the microporous substrate materials used contributes significant resistance to vapour transport in the membrane (over 50%) [165]. Minimizing or eliminating the resistance in the substrate is of key importance to increasing the water vapour transport of these materials. We found that casting continuous thin films from PEO-PU polymer dispersions was challenging, and defect free films could not be cast below 15 microns in thickness with a gap applicator. Furthermore, these freestanding films were fragile and could not be easily handled. 111  Casting or impregnating a film layer into the non-woven support was also unsuccessful. Coating a defect free film using the non-woven alone required multiple impregnation steps and resulted in a very thick film layer (>100 microns in thickness).  By impregnating electrospun nanofibrous layers with a water vapour permeable polymer dispersion of PEO-PU, we found that nanofiber filled films between 1 and 15 microns in thickness could be readily made on the surface of a non-woven support. This allows the elimination of low pore size, high resistance, and high cost substrates layer used in commercially available composite membrane materials. Nano-fibrous layers are promising scaffolds for permeable films since they can have high porosity (>80%), essentially no dead-ended pores (material pores that do not lead through the material), and their pore size can be controlled to facilitate application of a coating layer.   These impregnated electrospun nanofibrous membranes (IENM) are fabricated by filling a nanofibrous layer with a selective, water permeable polymer solution; and then evaporating the solvent to create a continuous nanofiber filled film layer. The fabrication procedure is summarized in Figure 5-1. The controlled deposition of the nanofibrous layer on an open non-woven support layer provides a scaffold for carrying the selective coating layer. By controlling the deposition time and fiber diameter of the nanofibrous layer, nanofibrous scaffolds of low thickness and high porosity can be generated on the surfaces of non-woven supports. By changing the concentration of the impregnating (coating) solution the extent of impregnation of the nanofibrous layer can be controlled. The goal is to create continuous defect-free impregnation of the nanofibrous layer, at a minimum thickness bound to the top of a non-woven mechanical support. This will lead to nanofibrous reinforced films with high water vapour flux, which maintain adequate selectivity of water vapour over other gaseous molecules. The target water vapour permeance for these membranes is greater than 10000 GPU with an oxygen permeance of less than 100 GPU, these criteria are explained in greater detail in Section 5.7.  112   Figure 5-1: Fabrication procedure for impregnated electrospun nanofibrous membranes. An additional benefit of this approach is that if the impregnation solution wets both the nanofibrous layer and the non-woven support layer the polymer will bond the filled nanofibrous film to the support. A cross-sectional image of MEM-90-21 in Figure 5-2 shows that impregnating the nanofibrous layer with a water vapour permeable polymer creates a continuous film layer that is also bound to the non-woven layer. Binding the impregnated nanofibrous layer to the non-woven fibers increases the mechanical integrity of the membrane, and allows for handling for manufacturability and long lifetime in the application.  Figure 5-2: (a.) Cross-sectional image of an impregnated electrospun nanofibrous membrane (MEM-90-21) with the impregnated film layer binding to the non-woven fiber at 113  500x magnification and (b.) the impregnated nanofibrous layer with a thickness of 9 microns at 2500x magnification. 5.4 Electro-spinning of Nanofibrous Layers Fibers were fabricated via electrospinning and collected on a non-woven support layer. Nanofibrous layers with three depositions of fiber were created by spinning for 60, 90, and 120 minutes. The fiber mats were examined via scanning electron microscopy and image analysis software (ImageJ) was used to determine fiber diameter as summarized in Figure 5-3. The areal densities measured via gravimetric methods are reported in Table 5-1 along with the average fiber diameters in the mats. The average fiber diameter is fairly uniform since all the mats were electrospun using the same polymer solution and spinning conditions. The thickness and areal density of the mats increased with increasing deposition time, as more layers of fiber were electrospun onto the support layers.  Table 5-1: Properties of the nanofibrous substrates generated at three electrospinning times, basis weight errors are standard deviations based on measurements from three samples, fiber diameter errors are standard deviation based on 150 measurements, 50 measurements three locations in each nanofiber mat. Sample Electrospinning time (min) Nanofiber Basis Weight (g/m2) Fiber Diameter (nm) PAN-60 60 0.63 +/- 0.06 483 +/- 64 PAN-90 90 0.91 +/- 0.04 456 +/- 71 PAN-120 120 1.45 +/- 0.04 458 +/- 65 114   Figure 5-3:  Histograms of fiber diameters for three deposition times (60, 90, and 120 minutes) each based on 150 measurements, with 50 measurements at each of three locations on each of the three fibrous mats, a representative image of fibers at 10000x magnification. 5.5 Membrane Fabrication In order to study the effect of the impregnating solution on the water vapour transport and selectivity of the nanofibrous membranes, nanofiber mats at three levels of fiber deposition were dip coated with each of three PEO-PU solutions containing 11, 15, and 21% polymer solids. The PAN nanofibrous layers were therby impregnated with the cross-linked PEO-PU coating and dried. The coating areal weight in the nanofibrous layer was measured and is reported in Table 5-2.  Table 5-2: Properties of nanofibrous substrates and impregnated nanofibrous membranes, thickness is the average of 75 measurements, 25 measurements at each of three locations in each membrane, the error in thickness is standard deviation from the measurements. Sample Electro-spinning time (min) Nanofiber Basis Weight, ωfiber (g/m2) Impregnation Solution Solids Concentration (%) Impregnation Dry Weight, ωimpreg (g/m2) Gravimetric Thickness (µm) Thickness as Measured by SEM (µm) MEM-60-11 60 0.63 11 2.08 2.32 2.02+/-0.39 MEM-60-15 60 0.63 15 4.29 4.22 4.52+/-0.89 MEM-60-21 60 0.63 21 9.54 8.74 7.41+/-1.24 115  MEM-90-11 90 0.91 11 3.42 3.71 4.18+/-0.50 MEM-90-15 90 0.91 15 6.54 6.40 6.31+/-1.12 MEM-90-21 90 0.91 21 14.22 13.00 10.11+/-4.67 MEM-120-11 120 1.45 11 3.99 4.66 4.51+/-0.94 MEM-120-15 120 1.45 15 7.64 7.80 6.67+/-0.61 MEM-120-21 120 1.45 21 13.24 12.62 12.06+/-1.21  The amount of coating impregnated into the nanofibrous layer increased with increasing fiber deposition. The basis weight or amount of dry impregnated coating increased with the percent solids in the impregnation solution. The membrane film thickness (δfilm) is estimated by the dry density of impregnation polymer (1.162 g/cm3) the dry density of the PAN polymer (1.184 g/cm3) and the basis weights (areal densities) of the fibers and dry impregnation polymer measured in g/m2 and is reported as the gravimetric thickness in Table 5-2:   PUPEOimpregPANfiberfilm m  [5-1] The average film thickness measured by SEM correlates well (R2=0.94) with the reported gravimetric thickness with the known polymer density as calculated by Equation [5-1]. Using the gravimetric thickness allows an average effective thickness over the entire sample to be used for permeability calculations which should be more representative of the entire sample tested than using discrete points measured by SEM.  5.6 Membrane Structure A pictorial summary of representative images of each membrane surface is shown in Figure 5-4. At low impregnation solution polymer concentration and lower fiber deposition, the electrospun layers are not completely impregnated and large pores can be observed in the membranes. As the impregnation solution solids concentration increases it can be observed that there are no visible pores in the impregnated membranes and that continuous dense films were produced. Membranes with large pores, which are not completely impregnated such as image MEM-60-11 in Figure 5-4, will have increased contaminant and gas crossover compared to fully impregnated materials. However, increasing the degree of impregnation and the amount of total polymer in the nanofibrous matrix will increase the film thickness, 116  increasing the path length of diffusing water molecules though the membrane. This leads to increased resistance to vapour transport and decreased water vapour flux in the membrane. Representative cross-section images of each membrane film layer are summarized in Figure 5-5, it can be observed that increasing the impregnation solution concentration and increasing the fiber loading tend to increase the membrane film layer thickness.   Figure 5-4: Surface images of impregnated electrospun nanofibrous membranes at various electrospinning times (depositions) and impregnation solution polymer solids at 2000x magnification, scale bars represent 20 microns. 117   Figure 5-5: Freeze fractured film layer cross-sections for all impregnated electrospun nanofibrous membranes at various electrospinning times (depositions) and impregnation solution polymer solids at 5000x magnification, scale bars represent 10 microns. 5.7 Permeation Testing of Nanofibrous Membranes 5.7.1 Pressurized Air Crossover and Oxygen Permeance The oxygen crossover rates expressed as a percentage of supplied oxygen at two flow rates in a cross flow oxygen permeance test are reported in Table 5-3. The calculated oxygen permeance is also reported. The PEO-PU coating material has a low inherent oxygen/nitrogen permeability (~1-5 Barrer), which is typical for these types of block copolymers [166]. Almost any oxygen crossover observed through the membrane must be associated with defects or pores in the film layer. The pressured air crossover is a direct indication of the level of defects 118  in the membranes, as it measures air flow through pores in the membranes. Defect free films will be air barriers at the pressures tested and should have zero cross-over.  Table 5-3: Oxygen crossover in impregnated nanofibrous membranes, one membrane sample was tested for each measurement. Sample Pressurized Air Crossover at 1 PSI (cm3/min) Oxygen Crossover 2000 cm3/min flow (%) Oxygen Crossover 500 cm3/min flow (%) Oxygen Permeance (GPU) MEM-60-11 14000 41.3 73.6 3450 MEM-60-15 1640 1.9 11.5 236 MEM-60-21 374 0.5 1.9 46 MEM-90-11 1130 5.8 21.2 503 MEM-90-15 39 0.0 0.5 9 MEM-90-21 20 0.0 0.5 8 MEM-120-11 474 0.5 4.3 81 MEM-120-15 26 0.0 0.5 8 MEM-120-21 15 0.0 0.5 8 Generally, membranes that were fabricated from lower fiber deposition and low impregnation solution polymer solids had higher oxygen permeance and air crossover. For example, membrane surface images for MEM-60-11 and MEM-90-11 in Figure 5-4 confirm that these membranes have relatively open pore structures due to low impregnation and the high oxygen permeance measured in these materials is explained by this open pore structure. Membranes with more complete impregnation as observed in the surface images have improved (lower) gas permeance and lower pressurized air crossover. Due to the level of oxygen crossover and pressurized air crossover observed, it can be concluded that all membranes fabricated have some small defects. However, in the application the static pressure differential across the membrane is minimal (<250 Pa), and a small level of crossover is acceptable (usually less than 1% of rated flow at the highest operating pressure). In North America, certification for ERVs involves an ‘exhaust air transfer ratio’ (EATR) test which is completed using sulfur hexafluoride (SF6) tracer gas [18]. An acceptable crossover rate is less than 1% in an ERV 119  core. The equivalent flow in the test module used in this study in comparison to an ERV core is at 2000 cm3/min in the module, so the target is less at 1% transport at this flow, or less than 100 GPU of oxygen permeance. 5.7.2 Hydrophobicity The membranes fabricated were hydrophilic with a contact angle of ~35° as shown in Figure 5-6, this is due to the hydrophilic nature of the PEO-PU polymer coating used.   Figure 5-6: A water droplet on an IENM, contact angle is 35°. 5.7.3 Water Vapour Transport Water vapour transport measurements were completed on each membrane at a range of flow velocities as reported in Figure 5-7 along with flux results for three commercially available membrane materials. It can be observed that the flux rates of the IENM’s are equivalent or better than currently available commercial materials. All permeation measurements were taken isothermally at 50°C with Stream 3 inlet conditions of T3=50°C, Pv,H2O,3= 6120 Pa (aH2O=0.5) and Stream 1 inlet conditions T1=50°C, Pv,H2O,3= 0 Pa (aH2O=0.0). The testing was completed in counter-flow, and there is a continuous vapour pressure differential across the membrane which drives vapour flux. The outlets of the module are open to atmosphere. All measurements are taken at equal flows on both sides of the membrane, and there is no average differential static pressure across the membrane.  Increasing the flow velocity in the test module led to increased flux rates due to decreasing the boundary layer resistance at the membrane surfaces. As the flow increases the flux rate tends to plateau, this indicates that the resistance in the boundary layer has less impact on the observed performance as the flow rate increases. The effect of the boundary layer becomes more prominent as the vapour transport rate of the membrane increases; this is evident in Figure 5-7 as higher performance membranes demonstrate a greater change in flux with increasing flow velocity. This is further evident in Table 5-4, where the percentage contribution of the boundary layers is considerably 120  less in membrane materials with lower vapour flux than in higher vapour transport materials. The apparent and actual permeance values determined as discussed in Chapter 4, Section 4.3 are also reported in Table 5-4. It is evident that as the membrane water vapour transport performance increases, accounting for the boundary layer resistance associated with the feed and sweep streams in the test module gains greater importance in the permeance analysis. All further reported water vapour permeability and permeance values are ‘actual’ permeance values, are reported with the boundary layer resistance removed, and are based on the effective vapour pressure differential across the membrane surface.  It is desirable to maximize the water vapour permeance in the IENMs while maintaining an oxygen permeance of less than 100 GPU, the target water vapour permeance is 10000 GPU or greater, this corresponds to a higher performing membrane than current generation commercially available materials. In the nanofibrous membranes fabricated, both higher loading of fibers and higher percent solids in the impregnation solution tended to lead to lower vapour transport rates, this corresponds with a general increase in coating loading (and thickness) as summarized in Table 5-2.    Figure 5-7: Water vapour flux in IENMs and three commercially available ERV membranes as a function of velocity in the test module channels, one membrane sample was measured for each test, testing is completed isothermally at 50°C with the wet stream inlet (stream 3) at water vapour pressure of 6120 Pa and the dry stream inlet (stream 1) at water vapour pressure of 0 Pa. 121   Table 5-4: Water vapour transport properties of impregnated nanofibrous membranes, one membrane sample was measured for each test, at standard test conditions: T=T1=T3=50°C;Pv,H2O,1=0 Pa; Pv,H2O,3=6122  Pa; υ1= υ3=4.75 m/s. Sample Water Vapour Flux (kg/m2/day) Rmem (s/m) % Boundary Layer Resistance Apparent Permeance (GPU) Corrected Water Vapour Permeance (GPU) MEM-60-11 29.2 38 53.8 13415 29044 MEM-60-15 17.4 121 27.0 6727 9213 MEM-60-21 11.3 231 16.2 4033 4812 MEM-90-11 20.8 87 33.8 8425 12726 MEM-90-15 11.4 228 16.4 4078 4875 MEM-90-21 7.8 371 10.7 2676 2998 MEM-120-11 13.8 173 20.5 5107 6422 MEM-120-15 9.2 302 12.9 3212 3687 MEM-120-21 7.8 369 10.8 2685 3009 In Figure 5-8 water vapour permeance of the impregnated nanofibrous membranes is plotted as a function of impregnation solution concentration at three nanofiber loadings. Decreases in water vapour permeance and flux are observed as polymer solids in the impregnation solids increases from 11 to 21%. Permeance and flux also decrease with the nanofiber loading in the membranes for a given impregnation solution concentration. Membranes with many surface pores or defects as observed in Figure 5-4, such as MEM-60-11 and MEM-90-11 demonstrate considerably higher vapour transport rates. However, the gas permeance through these materials due to defects and pores in the dense selective layer was also too great for these membranes to be effectively used in an ERV application.  122    Figure 5-8: Water vapour permeance at the standard test conditions in the IENMs fabricated as a function of impregnation coating solids at various fiber depositions, permeance results are average of measurements at four flow conditions. Table 5-5: Permeance and permeability of water vapour in each impregnated nanofibrous membrane averaged over four flow rates (υ1= υ3=4.75,6.35,7.94 and 9.52 m/s) at standard test conditions: T=T1=T3=50°C;Pv,H2O,1=0 Pa; Pv,H2O,3=6122  Pa; errors are the standard deviation on the average of the four measurements for each membrane. Group Sample Thickness from Basis Weights (µm) Ratio of PAN/PU Volume  Water Vapour Permeance (GPU) Water Vapour Permeability (Barrer) Selectivity (H2O/O2) I MEM-60-11 2.32 0.229 31629 ± 2707 73499 ± 6290 9 ± 1 II MEM-60-15 4.22 0.126 9248 ± 174  39074 ± 736 39 ± 1 III MEM-60-21 8.74 0.054 4800 ± 178 41984 ± 1559 105 ± 4 II MEM-90-11 3.71 0.207 13498 ± 717 50089 ± 2659 27 ± 1 V MEM-90-15 6.40 0.120 5282 ± 308 33810 ± 1973 569 ± 33 IV MEM-90-21 13.00 0.059 3176 ± 127 41328 ± 1650 391 ± 16 III MEM-120-11 4.66 0.263 6849 ± 399 31920 ± 1858 85 ± 5 V MEM-120-15 7.80 0.157 3831 ± 157  29916 ± 1229 472 ± 19 IV MEM-120-21 12.62 0.082 3143 ± 106 39698 ± 1333 387 ± 13 123    5.7.4 Permeability and Selectivity in IENM The membranes are fiber filled films bonded to one surface of a non-woven support. The gravimetric thickness (Equation [5-1]) is used as an estimate for the membrane film thickness. This gravimetric thickness agrees well with measurements taken by SEM, and gives an average effective thickness over the entire membrane area. This thickness is used to calculate the water vapour permeability in Barrer units. The electrospun PAN nanofibers do not contribute to water vapour transport, as PAN has very low water vapour permeability (300 Barrer) relative to the PEO-PU polymer (~50000 Barrer) [167]. It is useful to consider the volume fraction of PAN fibers to PEO-PU polymer in the final membranes in the interpretation of permeability. The fiber volume percentage is reported in Table 5-5 along with the thickness, water vapour permeance and permeability, and selectivity of the nanofibrous membranes. The transport data is the average of measurements taken with each membrane at four flow rates. In Figure 5-9 it can be observed that in normalizing permeance for thickness accounts for a large portion of the variation in the membrane flux and permeance; and permeability performance of the membranes falls into a narrower range. However, selectivity varies significantly between the membranes. Permeability and selectivity are associated with the structural properties of the membranes. Considering Figure 5-4, Figure 5-5, and Figure 5-9 the membranes are categorized into five groupings (I through V). These loosely defined categories are summarized in Table 5-6 and are used to describe the permeability and selectivity observed as a function of the membrane fabrication variables.  Table 5-6: Membranes as catagorized for permeability-selectivity interpretation of the IENM’s. Nanofiber Deposition Impregnation Solution Concentration 11% 15% 21% 60 min (0.45 g/m2) I II III 90 min (0.91 g/m2) II V IV 120 min (1.45 g/m2) III V IV  124   Figure 5-9:  Water vapour permeability (reported in Barrer units, where Barrer=1×10-10 cm3 (STP) cm cm-2 s-1 cmHg-1) against selectivity for a series of nanofibrous membranes averaged over a range of gas flow velocities (υ1= υ3=4.75, 6.35, 7.94, and 9.52 m/s). 5.7.4.1 Group I: MEM-60-11  Low nanofiber loading and low solids in the impregnation solution led to membranes with an open pore structure, such as MEM-60-11 as observed in the surface image in Figure 5-4. These membranes had high water vapour permeance and low selectivity, due to most transport occurring through the open pores in the material. Due to the high leakage and low selectivity, these membranes would be insufficient for the application. 5.7.4.2 Group II: MEM-60-15 and MEM-90-11 Increasing the fiber loading or the impregnation solids over Group I membranes led to improved films, however the surface images of MEM-90-11 and MEM-60-15 (Figure 5-4), crossover rates, and lower selectivity indicate that there are still significant defects in the film layers. Higher permeability in MEM-90-11 is related to water vapour transport through defects in the film layer. 5.7.4.3 Group III: MEM-60-21 and MEM-120-11 These membranes had moderate permeance and higher selectivity than Group II membranes. One of the membranes (MEM-60-21) was fabricated with low fiber loading (0.63 g/m2) and high impregnation solution concentration (21%) and the other membrane (MEM-120-11) was 125  fabricated with high fiber loading (1.45 g/m2) and low coating impregnation solids (11%). These membranes had oxygen permeance in the target range (<100 GPU). MEM-120-11 had lower water vapour permeability (~30000 Barrer) which demonstrates the significance of higher PAN to PEO-PU ratio in thinner membranes with higher fiber loadings. Here the ratio of PAN fiber in the film layer is large (0.263, Table 5-5) and causes a decrease in the permeability of the membrane, this finding is discussed in further detail in the Section 3.5.5. 5.7.4.4 Group IV: MEM-120-21 and MEM-90-21 This group of membranes were fabricated with the high polymer solids solution (21% wt.) and had higher nanofiber basis weights (0.91 g/m2 and 1.45 g/m2), leading to high coating loading and thickness (>10 micron). This caused the membranes to have the lowest water vapour permeance of all membranes fabricated. The membranes also had low oxygen permeance, low air crossover, and high selectivity indicating that there were low defects in the film layers. The water vapour permeability of these membranes was generally aligned with other membranes (~40000 Barrer). MEM-120-21 has a slightly lower permeability than MEM-90-21 which is likely associated with the increase in fiber ratio in the film layer in MEM-120-21 (0.082) over MEM-90-21 (0.059). 5.7.4.5 Group V: MEM-120-15 and MEM-90-15 Group V membranes had moderate (0.91 g/m2) and high (1.45 g/m2) loading of nanofiber and were fabricated from moderate impregnation solution solids concentration (15%). These membranes had low defects showing low air crossover and oxygen permeance (Table 5-3). The film thickness was lower than Group IV membranes and the Group V membranes have higher selectivity due to increased water vapour permeance. However the permeability of the membranes in comparison to Group IV membranes is low. This phenomenon can be explained by considering the proportion of PAN nanofiber volume to PEO-PU polymer volume in the membrane. In Group IV membranes, the fiber ratio is 0.059 (MEM-90-21) and 0.082 (MEM-120-21). In Group V membranes the ratio of PAN fibers increases from the Group IV membranes at the same fiber deposition to 0.12 (MEM-90-15) and 0.157 (MEM-120-15). This impacts the permeability of these membranes, as the ratio of relatively impermeable PAN throughout the membrane increases with decreasing film thickness and increased fiber loading.  5.7.5 Fiber Ratio in the Dense Layer Figure 5-5 is a pictorial summary of the membrane cross-sections. At a constant fiber loading (for example, Figure 5-5 at ωfiber=1.45 g/m2) as the impregnation solution concentration used 126  to fabricate the membranes decreases, the impregnation basis weight and film layer thickness decreases, leading to increased permeance. Concurrently though, decreased permeability is observed due to the increase in PAN fiber to PEO-PU coating ratio. This effect is most notable in membranes with high selectivity, low defects, and low oxygen permeance and is readily evident when comparing membranes of Group IV and V. The ratio of fiber volume in the film layer is expected to become more significant with attempts to increase the permeance/flux of electrospun nanofibrous membranes by decreasing the membrane thickness. Comparing membranes with water vapour over oxygen selectivity greater than 100 in Figure 5-10, a clear decreasing trend in permeability is demonstrated with increasing fiber ratio. The permeability of a pure PEO-PU film (zero fiber ratio) is shown in Figure 5-10 as well. All the membranes in this study were fabricated with a constant fiber diameter (~450 nm). At this fiber diameter it is evident that the PAN fibers occupy a significant volume of the IENM film layers and can have a detrimental effect on water vapour permeability. It is notable that although the IENM created demonstrated moderate water vapour permeance and selectivity, and vapour flux rates greater than some commercially available membrane materials, none of the membranes met both the water vapour permeance criteria of greater than 10000 GPU and the oxygen permeance criteria of less than 100 GPU. Decreasing the fiber diameter may decrease the membrane thickness and decrease the ratio of fiber in the impregnated nanofibrous film layer. This should allow for significantly increased water vapour permeance. Chapter 6 will focus on the effect of fiber diameter and fiber loading on permeance and permeability.  127   Figure 5-10: Water vapour permeability as a function of the PAN fiber ratio to PEO-PU ratio in a series of IENM films with low defects (water over oxygen selectivity, αH2O/O2>100), error bars are the standard deviation around the average of four measurements at different flows for each membrane sample. 5.7.6 Water Vapour Activity The membranes were tested for water vapour transport under a range of humidity conditions and are reported in terms of water vapour permeability (Figure 5-11) as a function of water vapour activity (aH2O) in the feed stream.  These results are compared to a pure film of the PEO-PU polymer with a thickness of 32 microns. Activity is defined here as the vapour pressure at the Stream 3 inlet (feed) normalized by the saturation vapour pressure of air in the feed stream at the test temperature (in this case 50°C). This allows membranes to be compared at a partial pressure reference which represents the maximum vapour pressure at any temperature.  Flux in the membranes increased with increasing water vapour partial pressure in the feed stream. This is expected as the driving force across the membrane in solution-diffusion processes is proportional to the chemical potential gradient, which in this case is determined by the vapour pressure at the feed inlet [14]. Permeability (flux normalized for partial pressure differential and thickness) tends to only show a slight increase with changing feed activity for most membranes at aH2O < 0.5 and the permeability of the membranes is between 20000 and 50000 Barrer in this 128  activity range. At higher activity in the feed (above aH2O=0.5) the permeability curves upward. This effect is reported in many water vapour permeable-hydrophilic polymer systems and is related to increased water sorption in the dense PEO-PU film layer as discussed in Chapter 3 [92]. As the feed vapour pressure increases, more vapour is absorbed into the film layer and the local concentration of water vapour in the polymer increases, see Figure 3-18 also. This leads excess free volume in the polymer allowing a further increase in sorption into the film layer. Increased sorption in the dense layer at higher activity leads to higher permeability. As shown in Figure 5-11, the permeability of all nanofibrous membranes except MEM-60-11 (which had a high level of defects) is lower than the permeability of a pure film of the polymer. This demonstrates that the presence of embedded PAN fibers decreases the permeability of the films. Each membrane group has been discussed to explain this in detail above, but it is can be observed in Figure 5-11 that at high fiber loading (120 min, 1.45 g/m2) membranes made with the highest impregnation solids (21%) have significantly higher permeability than those with lower solids (15%). This is due to the lower PAN to PEO-PU volume ratio in MEM-120-21 over MEM-120-15. Similarly, MEM-90-21 has higher permeability than MEM-90-15. This trend only holds for membranes with sufficiently low defects (low oxygen permeance) such that the transport is largely through the dense polymer where PAN fibers can inhibit motion of water molecules through the membrane film. As such MEM-90-11, MEM-60-15, and MEM-60-11 with higher oxygen permeance deviate from the trend of decreasing permeability with increasing fiber ratio as a significant portion of transport in these materials is through the pore structure created by defects in the film layer. 129   Figure 5-11:Water vapour permeability in IENMs and a PEO-PU film (32 micron) as a function of vapour pressure ratio (activity, aH2O) in the wet inlet (stream 3), conditions were isothermal at 50°C, flow velocity υ1= υ3=6.35 m/s, error bars are the standard deviation for measurements of each sample at each activity. 130  5.8 Conclusions  The central hypothesis of this dissertation is that the substrate used in current ERV membranes can be eliminated by the use of an electrospun nanofibrous layer. This chapter presents our first attempts to validate this hypothesis, and demonstrates a proof of concept for these impregnated electrospun nanofibrous membranes (IENM). Although the performance target (>10000 GPU water vapour permeance and less than 100 GPU oxygen permeance) was not achieved, the fiber geometry was held constant in this Chapter, and further optimization by modifying the fiber diameter must be completed. We find also that this approach, using a relatively impermeable fiber layer, contributes an added transport resistance to the membranes, which must be addressed. Impregnated electrospun nanofibrous membranes (IENM) were fabricated using a process that involved the electrospinning of a PAN nanofibrous layer onto a porous support, impregnating the nanofibrous layer with an aqueous polyether-polyurethane solution, and then drying and crosslinking the impregnation solution. This procedure allowed for the successful fabrication of low defect membrane film layers which were bonded to the support layer, making them materials with good handling properties for further manufacturing. Membranes were created using a range of nanofiber depositions and polymer impregnation solution concentrations in order to study how the fabrication process variables affected the structural properties of the membrane. In order to determine how structure affected function, the water vapour and oxygen permeance properties of the membranes were measured. Overall the membranes demonstrated moderate water vapour permeance and acceptable selectivity for enthalpy exchanger applications, comparable or better than some commercially available membrane materials. However, improved water vapour transport is still desirable and the target permeance of 10000 GPU with oxygen permeance of less than 100 GPU was not achieved in this study. Low fiber loading and low polymer solids in the impregnation solution led to thinner film layers and higher water vapour permeance; but these membranes tended to have higher levels of defects and insufficient oxygen permeance. Generally, increasing the fiber deposition and the coating polymer solids in the impregnation solution led to thicker film layers which decreased the water vapour permeance.  It was shown that increased fiber ratio in the IENM led to decreased permeability in defect free film layers. This effect was more pronounced as the volume ratio of low permeability PAN fibers increased, especially as thickness of the film layers was decreased. Chapter 6 will focus on optimization of the 131  performance of IENM’s by modifying the structural and geometrical properties of the fiber layer.  132   Optimization of Impregnated Electrospun Nanofibrous Membranes 6.1 Introduction In the previous chapter, impregnated electrospun nanofibrous membranes (IENM) were fabricated by electrospinning a layer of sub-micron polymer fibers onto the surface of a non-woven carrier, the electrospun layer was then impregnated with an aqueous polymer dispersion and then dried, creating a dense, cross-linked nanofiber filled layer bonded to the surface of a non-woven support. In Chapter 5, the fiber diameter was held constant, while evaluating the effects of fiber deposition and impregnation solution polymer concentration. Water vapour transport performance and selectivity of water vapour over oxygen was affected by the nanofiber deposition and the polymer concentration in the impregnation solution used to fabricate the membranes. In this chapter, approaches to optimizing the performance and manufacture of these IENM’s are discussed. The initial work in Chapter 5 is expanded upon by analyzing the effect of fiber diameter on the membrane performance.  First, it was hypothesized that smaller fiber diameters would lead to thinner fiber-filled-film layers, containing lower impermeable fiber volume, thereby leading to improved water vapour transport performance. This is validated through optimization involving factors of fiber diameter, fiber deposition, and coating deposition, leading to membranes which achieve the target of greater than 10000 GPU water vapour permeance with an oxygen permeance of less than 100 GPU. Building on mixed matrix membrane models and models for diffusion in fiber-reinforced polymer composites in the literature, a fiber-filled-film model is presented to account for the contribution of the water vapour transport resistance of the impermeable nanofibers in the membranes produced in this study. Secondly, an alternative membrane fabrication process was evaluated. In Chapter 5 a dip coating process was used to fabricate membranes, this method was a straightforward method for membrane fabrication. The membranes made were largely defect free and had moderate water vapour permeance. However, the dip coating method had a major drawback in that there was little control over the deposition of coating polymer into the nanofibrous layer. Furthermore, excess polymer was deposited into the carrier layer in the dip coating process, and it would not be feasible to scale this dip-coating process beyond the fabrication of laboratory samples. Greater control over how much and where the dense coating layer was deposited into the nanofibrous layer is desirable. This led us to consider the gravure coating 133  process as a potential route to impregnated electrospun nanofiber membrane (IENM) fabrication and scale-up [168].  A survey of the literature on electrospun nanofibrous membranes shows a number of approaches to fabricating nanofibrous membranes for various applications. Coating methods for membranes with a non-porous layer, included dip and saturation coating, rod and gap coating, spin coating, slot die coating and interfacial polymerization. These methods are summarized in Table 6-1, along with a qualitative assessment of scalability to a high volume roll-to-roll manufacturing process. This assessment is based on whether the fabrication process can be used to make continuous rolls of material; whether method required to fabricate the membrane successfully is used in industrial coating processes presently; and whether the nanofibrous layer would significantly complicate membrane manufacture by the coating method. In this review, we only found one mention of the use of a gravure coating process in the literature [169]. In this reference only one pattern of gravure cylinder was reported with one coating formulation, and in this case the membrane was still a porous material, a continuous film layer to make a selective membrane was not the goal. In the second portion of this chapter, we contribute a study on selective (non-porous) membranes based on electrospun nanofibers produced via the gravure process. This process is scalable to high volume roll-to-roll manufacturing processes, and allows a controlled or ‘metered’ amount of coating to be deposited into or on the nanofibrous layer. Table 6-1: Review of various reported methods of fabricating non-porous membranes based on electrospun nanofibers. Method Description Scalable REF Dip/ Immersion Nanofibrous layer is dipped or immersed in coating solution to wet and fill the fibers. Maybe [89], [170]–[173] Gap Applicator  Excess coating is placed on the sample and an applicator with a micrometer controlled gap above the substrate surface draws away the excess coating while allowing a metered amount to remain. No  [174] Gravure Coating solution is metered into nanofibrous layer using a roller with an etched cell pattern, the cells hold a controlled liquid volume which is transferred to the nanofibrous layer. Yes [169] 134  Method Description Scalable REF Interfacial Polymerization Nanofibrous layer is wet with a monomer solution and then a reacting monomer solution is cast on the surface, a film layer is synthesized at the solution interface. Casting can be achieved via various methods (blade, rod, etc). Excess reactant is typically washed from the membrane. Yes [87], [95], [162], [175], [176] Melting  Vapour or liquid melting of the nanofibrous layer to create a continuous film layer Maybe [177], [178] Phase inversion Nanofibrous layer is filled with a solvent system and then immersed in a non-solvent to induce phase separation, creating a porous structure, a non-porous film layer can be cast on the surface. Yes [163] Rod or blade coating Excess coating solution is deposited on nanofibrous surface and then spread over the nanofibrous layer via a rod or casting blades, rods can be etched to allow metered amounts of coating to remain on the substrate. Maybe [94], [179], [180] Slot Die Solution is metered as a bead onto the nanofibrous layer via a die with a narrow slot. Yes [181] Spin coating Nanofibrous material is attached to a plate, and polymer solution is deposited onto the nanofiber while rapidly spinning the plate, excess solution is removed via angular momentum. No [182] Spray Coating A solution of polymer is sprayed onto the nanofibrous material, layer by layer, as droplets collide with the nanofibrous layer, a coating is layer formed bridging the gaps between the nanofiber, creating a dense film. Maybe [183] 6.2 Solutions and Electrospinning PAN solutions were formulated over a range of polymer concentrations and viscosity and surface tensions measurement were taken for the solutions, the results are summarized in Table 6-2. The results show that the solution viscosity increases exponentially with increasing 135  polymer concentration, surface tension of the solutions increases slightly in a linear fashion over the range of polymer concentrations measured. A viscosity increase was observed for the solutions after aging for 7 days. Due to this, all further solutions used were made new prior to electrospinning.  Fiber diameters were determined by taking three samples from the electrospun mat and generating SEM images of the surfaces. Diameters were measured using ImageJ software based on fiber images at 10000x magnification, a minimum of 150 measurements were taken from a minimum of three locations within the electrospun mat. PAN solutions at less than 8% PAN polymer concentration in solution showed fibers with beads and no fiber diameter could be determined. No continuous fibers were formed below 5% PAN polymer concentration and just a polymer spray was emitted from the needle during electrospinning. Above 16% PAN concentration, no fibers were electrospun. In the range of 8 to 15% PAN concentration in solution smooth fibers were collected, the fiber diameter increased with increasing concentration and increasing polymer solution viscosity. Table 6-2: Electrospinning solution properties and resulting fiber diameters, errors are sample standard deviations for a total of 150 diameter measurements at 3 locations in each material. PAN Electrospinning Solution wt.% Viscosity t=0(cPs) Viscosity t=7 days (cPs) Surface Tension (mN/m) Fiber Diameter (nm) 4 35.5 - 35.5 Spray 5 63.3 - 35.6 Beads 6 83.9 - 36.0 Beads 7 121 127 36.0 Beads 8 160 176 36.3 198 +/- 47 9 227 244 36.4 221 +/- 42 10 314 332 36.6 260 +/- 37 11 385 415 36.6 295 +/- 45 12 539 556 36.8 492 +/- 84 13 665 691 - 577 +/- 94 14 878 - - - 15 1004 - - 1250 +/- 239 16 1305 - - no jet  136  6.3 Membrane Optimization: Fiber Diameter and Deposition In the previous study, PAN nanofibers in the range of 450 nm diameter were deposited at 0.6, 0.9, 1.5 g/m2 and impregnated with PEO-PU polymer from dispersions with solids contents of 11, 15, and 21%. The target of a 10000 GPU water vapour permeance membrane with less than 100 GPU oxygen permeance was not achieved. However, it was evident that middle to low fiber loadings (0.63 and 1 g/m2) and middle PEO-PU coating formulation solids (15%) had the best overall performance. The fiber ratio in the film layer represented a significant resistance to water vapour transport, and thinner effective film layers with low defects would be necessary to achieve the performance target. It was hypothesized that lower fiber diameter would help achieve thinner films with lower fiber loadings. The total experimental design space for this new round of experiments was PAN fiber spinning solution concentration= [8,10,12,13%]; PEO-PU impregnation solution concentration= [12, 15, 18%]; Nanofiber deposition= [low (0.25 g/m2), mid-low (0.35 g/m2), medium (0.5 g/m2), mid-high (~0.75 g/m2), high (~1 g/m2)]. About 70 membranes were made and tested covering this design space. First, electrospun nanofibers were deposited on a PET based nonwoven (Reemay) with an average basis weight of 14.1 +/- 0.9 g/m2 (PET-NW). Due to the variability of the non-woven, the fiber deposition could not be determined directly gravimetrically on the non-wovens, instead sections of sample were cut from the fiber-nonwoven, and the fibers were removed and weighed to determine the areal loading of fiber on each sample, these are reported in Table 6-3. All samples were spun by attaching a 25 x 80 cm non-woven sample to a grounded belt collector rotating at 3 m/min, three 18G needles were traversed in an alternating movement back and forth across the width of the sample at a rate of 10 cm/min, the electrospinning needles containing PAN solution were at a distance of 15 cm from needle tip to collector, a potential of 20 kV was applied across the gap between the needle and collector and the pump rate was adjusted to maintain a slight bead at the needle tip.     137  Table 6-3: Summary of PEO-PU impregnated PAN nanofiber membranes on PET non-woven carriers, results are for one membrane sample at each set of fabrication parameters. ID Fiber Diameter (nm) ωfiber (g/m2) Coating Solids (%) ωPEO-PU (g/m2) Fiber Volume (%) [gravimetric] Film Layer Thickness (μm) [gravimetric] Permeance (GPU) Selectivity H2O O2  H2O /O2 397 198 0.13 12 2.6 4.4 2.50 54162 1005 54 398 198 0.13 15 3.5 3.3 3.32 42690 1141 37 399 198 0.13 18 6.2 1.9 5.80 14621 283 52 403 198 0.17 12 3.5 4.2 3.36 39398 920 43 404 198 0.17 15 4.0 3.7 3.76 38795 741 52 405 198 0.17 18 6.5 2.3 6.07 15639 428 37 475 198 0.26 12 3.7 6.1 3.61 29876 192 156 476 198 0.26 15 5.1 4.5 4.88 21632 127 171 477 198 0.26 18 7.6 3.0 7.20 12679 71 177 400 198 0.35 12 3.1 9.4 3.18 17817 99 181 401 198 0.35 15 5.5 5.6 5.29 13593 42 326 402 198 0.35 18 5.8 5.3 5.58 8643 22 384 359 198 0.45 12 5.0 7.6 4.99 9758 10 1015 360 198 0.45 15 5.8 6.7 5.65 12012 35 344 361 198 0.45 18 8.4 4.7 8.12 7133 7 1010 478 198 0.59 12 4.0 12.1 4.12 16029 21 755 479 198 0.59 15 5.7 8.8 5.69 13390 97 138 480 198 0.59 18 7.9 6.5 7.73 8774 26 333 362 198 0.62 12 4.9 10.5 5.00 8587 11 788 363 198 0.62 15 7.1 7.4 7.05 6919 7 980 364 198 0.62 18 8.8 6.1 8.56 5602 2 2829 365 198 0.84 12 4.6 14.3 4.93 7711 2 3894 366 198 0.84 15 7.1 9.8 7.25 6829 7 973 367 198 0.84 18 9.6 7.5 9.46 4147 6 721 406 260 0.22 12 5.2 3.7 4.95 34484 1040 33 374 260 0.22 15 4.6 4.1 4.39 23026 738 31 408 260 0.22 18 6.5 2.9 6.16 12831 441 29 409 260 0.34 12 3.7 7.8 3.69 20404 543 38 410 260 0.34 15 4.8 6.2 4.68 16728 270 62 411 260 0.34 18 7.8 3.9 7.42 9307 47 198 368 260 0.36 12 3.7 8.2 3.71 17918 48 371 138  ID Fiber Diameter (nm) ωfiber (g/m2) Coating Solids (%) ωPEO-PU (g/m2) Fiber Volume (%) [gravimetric] Film Layer Thickness (μm) [gravimetric] Permeance (GPU) Selectivity H2O O2  H2O /O2 369 260 0.36 15 4.8 6.5 4.70 15066 35 434 370 260 0.36 18 7.5 4.2 7.17 8440 19 454 472 260 0.41 12 4.4 7.8 4.39 15175 102 148 473 260 0.41 15 5.3 6.6 5.22 13358 7 1891 474 260 0.41 18 6.9 5.1 6.69 10791 14 798 469 260 0.46 12 3.6 10.6 3.67 14111 56 252 470 260 0.46 15 5.3 7.4 5.23 14297 258 55 471 260 0.46 18 7.3 5.5 7.07 8344 25 333 466 260 0.59 12 3.8 12.6 3.97 14337 179 80 467 260 0.59 15 5.0 9.8 5.11 10447 26 397 468 260 0.59 18 7.6 6.7 7.48 6718 10 699 371 260 0.66 12 4.4 12.3 4.55 9780 8 1180 372 260 0.66 15 5.8 9.6 5.82 9178 10 960 373 260 0.66 18 9.2 6.2 8.95 5563 6 968 374 260 0.95 12 4.1 17.7 4.53 8316 10 870 375 260 0.95 15 5.8 13.2 6.09 6852 22 306 376 260 0.95 18 9.9 8.2 9.84 3790 2 1926 377 492 0.33 12 4.8 5.9 4.68 19243 1141 17 378 492 0.33 15 5.5 5.2 5.30 20043 772 26 379 492 0.33 18 7.9 3.7 7.54 8742 98 90 380 492 0.54 12 4.2 10.6 4.28 11540 48 240 381 492 0.54 15 5.9 7.7 5.88 9789 25 389 382 492 0.54 18 7.9 5.9 7.73 6603 10 682 383 492 0.90 12 5.0 14.4 5.31 6826 7 966 384 492 0.90 15 6.7 11.1 6.87 6090 2 3075 385 492 0.90 18 10.2 7.6 10.09 4600 2 2300 386 492 1.54 12 4.6 23.5 5.52 5474 2 2737 387 492 1.54 15 7.3 16.3 8.00 4489 2 2244 388 492 1.54 18 10.2 12.2 10.64 3512 2 1756 412 527 0.36 12 3.5 8.8 3.48 - >2500 - 413 527 0.36 15 5.3 5.9 5.20 - >2500 - 414 527 0.36 18 6.8 4.7 6.53 - >2500 - 415 527 1.02 12 4.6 16.8 5.10 14006 671 21 139  ID Fiber Diameter (nm) ωfiber (g/m2) Coating Solids (%) ωPEO-PU (g/m2) Fiber Volume (%) [gravimetric] Film Layer Thickness (μm) [gravimetric] Permeance (GPU) Selectivity H2O O2  H2O /O2 416 527 1.02 15 6.3 12.9 6.66 - >2500 - 417 527 1.02 18 8.1 10.3 8.31 8399 196 43 418 527 1.28 12 4.9 19.3 5.58 9216 190 49 419 527 1.28 15 6.8 14.7 7.33 7501 88 85 420 527 1.28 18 8.3 12.4 8.68 4206 49 87  Samples were cut from the PET NW+PAN NF samples (25 cm x 15 cm) and weighed directly prior to coating. Samples were dip coated in a bath of PEO-PU based coating dispersion (12%, 15%, and 18% solids), excess coating was removed from the samples, and the samples were then dried and cured. Dry sample weights were taken to determine the average PEO-PU polymer which was added to the NF layer in each membrane. Each membrane was tested for oxygen permeance at ambient conditions and water vapour permeance isothermally at 50°C with a feed stream activity (aH2O=0.5). All results are reported in Table 6-3. 140   Figure 6-1: Water vapour permeance and selectivity of membranes made from nanofibers of different average fiber diameter with comparison to Chapter 5 IENMs and commercially available membranes [184]. The overall water vapour permeance and H2O/O2 selectivity is plotted in Figure 6-1 along with results from Chapter 5 and from commercially available ERV membrane materials [156], [184]. The selectivity represents the permeance of water vapour over oxygen permeance in the samples, it is essentially a measure of defects in the membranes, as membranes with absolutely zero defects will have a <5 GPU permeance to oxygen (O2 permeability in the PEO-PU polymer is in the range of 1 to 5 Barrer). As mentioned previously, the upper limit of acceptable oxygen permeance (PO2/δ) is about 100 GPU, for the application at no static pressure difference. The water vapour permeance (PH2O/δ) target for the application is >10000 GPU. This corresponds to a selectivity (αH2O/O2=PH2O/PO2=10000/100) of 100. As the permeance of water vapour permeance increases, the selectivity must go up to meet to 100 GPU oxygen permeance target, the low limit line is shown in Figure 6-1, and the target region is bound by dashed lines, with the target in the upper right region of Figure 6-1. From the data, it is evident that decreasing the fiber diameter allows for membranes with improved water 141  vapour permeance and selectivity by allowing thinner defect free film layers in the impregnated nanofibrous membranes. From Table 6-3 the highest performing membranes in terms of selectivity and water vapour permeance are produced from 200 to 260 diameter fibers, in the range of 0.3 to 0.75 g/m2 fiber areal density.  A response surface analysis was used to analyze the membrane performance based on factors of fiber diameter and fiber deposition at three levels of polymer coating solids concentration in the impregnation solution. An optimization procedure was run on the response surface fit model based on maximizing water vapour permeance while minimizing oxygen permeance. The results are shown in Figure 6-2, for the three levels of coating solids. This optimization study indicates that lower fiber diameters will generally lead to optimal membrane performance. It also confirms that fiber depositions in the range of 0.3 to 0.75 g/m2 tend to have the optimal performance for all levels of coating impregnation solids concentrations.   Figure 6-2: Response surface fit model for all data with desirability output to maximize water vapour permeance while minimizing oxygen permeance, at three levels of coating solids in the impregnation solution a. 12%, b. 15%, and c. 18%. 6.4 Fiber Volume Fraction Using smaller diameter fibers in the fiber filled polymer film layer should have two benefits. Firstly, if some critical number of fibers is required to make a defect free impregnated film layer, thinner film layers can be produced by using thinner fibers since the total fiber layer thickness will be lower for the same critical number of fibers for smaller diameter fibers. Secondly thinner fibers allow defect free films to be fabricated with lower fiber volume ratios. If a critical number of fibers layers is necessary to support a continuous defect free film, and the fibers have similar spacing then it can be argued on a geometrical basis that thinner fibers   142  will lead to both a thinner film layer (fiber+polymer coating) and a lower fiber ratio in the film as summarized pictorially in Figure 6-3.  Figure 6-3: Schematic for fiber filled polymer film cross-sections with different fiber diameters, constant fiber-to-fiber distance. Based on a geometric model with square packing as shown in Figure 6-4a, the fiber volume fraction, εf depends on the fiber to fiber distance (average pore diameter, σ) and the fiber diameter, d:   [6-1] In the range of 100 to 1000 nm, which is typical for electrospun fibers, the theoretical fiber volume increases significantly with fiber dimeter. Assuming an average fiber-to-fiber distance of 2.5 microns, the fiber volume fraction increases from 0.04 at 100 nm fiber diameter to 0.26 at 1000 nm fiber diameter. This would mean that the fiber volume fraction in a nanofiber filled film layer would be expected to decrease with decreasing fiber diameter. Figure 6-4a can be reduced to a master curve by plotting the fiber volumes against the ratio of the d/σ in common units as shown in Figure 6-4b.   ddf4143   Figure 6-4: (a.) Estimated fiber volume fraction based on fiber diameter and fiber-to-fiber distance for square packing (b.) reduce curve based on d/σ. Image analysis was completed to measure average fiber to fiber distances in the top fiber layer of the mats with various fiber diameters. Representative images of the fiber mats measured are shown in Figure 6-5. For fiber mats with average fiber diameters in the range of 200 to 600 nm, the average fiber to fiber distance in the top layer did not vary significantly and averaged around 2.6 micron. Fiber at larger diameter (1250 nm average) appeared to be somewhat fused (possibly due the large diameter leading to incomplete drying/solidification of the fibers during electrospinning), the average fiber to fiber distance is 10.2 μm for these larger fibers. These values align with results for nanofiber pore diameters reported in the literature as a function of porosity and fiber diameter [34]. The results indicate that typical inter-fiber pore diameters are in the range of 0.1 to 10 microns, which is in the range where the fiber-to-fiber distance is expected to have a significant impact on fiber volume in the fiber-polymer film layer in Figure 6-4a.  Figure 6-5: Images of fibers of varying diameter produced by electrospinning (first three images from left are at 10kx magnification, scale bar = 5μm; rightmost image at 2kx magnification, scale bar = 20 μm). 144  Various stochastic models have been reported based on randomly oriented fibers to predict fiber pore size distributions and the number of fibers in a unit cell based on the porosity, fiber diameter, and mat thickness [185]–[187]. In the present study a geometric model based on square fiber packing is used, in which the pore size is constant so that porosity, number of fibers, and total fiber height can be estimated based on fiber diameter and fiber areal density. The square fiber packing unit cell is shown in Figure 6-4a, with a unit length L=(σ+d), and unit area of A=(σ+d)2. With a given the basis weight (areal density) of fibers in a mat, ω (g/m2), the fiber to fiber distance σ, the fiber polymer density ρf (g/m3), and the fiber diameter d (m), equations can be derived for the fiber volume density per unit area αf (m3/m2), the length fiber length per unit area ℓ𝑓 (m), the number of fibers stacked per unit area nf, and the thickness of the fiber mat δf (m):  Volume areal density   [6-2] Fiber volume per unit area   [6-3] Fiber length    [6-4] Fibers per unit area   [6-5] Mat thickness  [6-6] From these equations, the expected fiber mat thickness can be estimated for a known fiber deposition (areal fiber density), fiber diameter, and average fiber to fiber distance. The fiber volume fraction is determined by equation [6-1] and the porosity is:   [6-7] The above equations imply that if a minimum number of fibers per unit area are required to achieve a defect free fiber-filled polymer coating layer, that increasing the fiber diameter requires an increase in basis weight to achieve a selective membrane and that the mat thickness and correspondingly the film layer thickness must also increase. Furthermore, the volume fraction of fibers in the mat will also increase.  Experimental data for all membranes ff  2dV ff    222ddAV ffff    22dddLnffffdnff  f  1145  at each fiber diameter was evaluated to determine the minimal fiber areal density (ω) at which membranes at a given fiber diameter had sufficiently low defects to achieve an oxygen permeance target. The minimum basis weight at which the oxygen permeance target was achieved based on a large series of fabricated membranes at different fiber diameters are summarized in Table 6-4 for oxygen permeance of less than 100 and oxygen permeance of less than 10. As expected based on equations [6-2] through [6-7], as the fiber diameter increases, the basis weight required to achieve sufficiently low oxygen permeance increases as well. Table 6-4: Based on experimental results, minimum fiber basis weights required to achieve oxygen permeance target at a given fiber diameter. d (nm) 225 253 566 1250 min ω (g/m2) for PO2/δ <100 0.12 0.24 0.54 >1.26 min ω (g/m2) for PO2/δ <10 0.25 0.39 0.83 -   146   Figure 6-6: Cross-sections of several membranes fabricated from fibers of different diameter and basis weight with thicknesses based on the SEM imaging and based on gravimetric measurement, scale bar for all images = 10 µm; errors on δSEM are the sample standard deviations based on 75 measurements, 25 measurements at three locations in each membrane. 147  Cross-sectional images of the fiber-filled film layer in the nanofibrous membranes with fiber diameters of 200, 260, 492, and 816 nm in the range of 0.35 to 1.275 g/m2 after impregnation with a coating solution at 18% solids are shown in Figure 6-6. It is evident from these images that as the fiber diameter increases, the film layer thickness increases. Figure 6-6 also reports the thickness of each membrane based on several SEM cross-sectional images from the membrane samples (δSEM). The thickness as estimated by gravimetric methods (δgrav) is also reported, this is calculated based on the measured PEO-PU dry coating and fiber deposition areal weights, the matrix PEO-PU polymer density, and the PAN fiber polymer density. The thickness of the film as measured by SEM images (δSEM) was significantly lower that the thickness estimated by weight measurements (δgrav) in all cases. In the gravimetric coating weight measurement, the combine non-woven and nanofiber material is weighed after spinning the fibers and then again after the membrane is dried following impregnation with the PEO-PU coating solution. This means that only the total polymer coating weight deposited into the whole material (non-woven carrier and nanofiber) during the impregnation step can be measured. However, a significant portion of PEO-PU polymer is also deposited in the non-woven carrier layer during dip coating impregnation, and this leads to an artificially high thickness estimate for the film layer. The deposition of the coating polymer into the non-woven carrier layer shown in the cross-sectional image in Figure 6-15a. The fiber volume fractions (Vf) are also reported in Figure 6-6, calculated based on the measured areal weights of the fiber in the membrane and total weight of polymer in the dried membrane for the gravimetric fiber volume, Vf,grav:  PANfiberPUPEOPUPEOPANfibergravfV, [6-8] For the actual fiber volume in the film layer, Vf,SEM, the fiber basis weight is used in the calculation (as this is measured prior to coating), but the coating basis weight in the film layer (ωPEO-PU,FILM) is determined instead from the film layer thickness as measured by SEM images:   PANfiberSEMPUPEOfilmPUPEO  ,  [6-9] And then Vf,SEM can be calculated: 148   PANfiberPUPEOfilmPUPEOPANfiberSEMfV ,,  [6-10] The discrepancies associated with excess coating remaining in the non-woven carrier layer after impregnation have implications when considering the effects of the fiber volume in the film on membrane permeability. Also, the film thickness must be used to determine permeability of the membrane film layer, and it was not feasible to verify the thickness of all membranes by SEM imaging, this is discussed further in terms of predicting the performance data based on a fiber volume model the Section 6.5. 6.5 Fiber Filled Film Membrane Permeability Model Various models have been presented to predict the effective permeability (Peff) of membranes with a disperse phase of particulate fillers within a polymer matrix. Based on the Maxwell model for electrical conductance in composite materials, a model for ‘mixed matrix membranes’ (MMM) with particle fillers of aspect ratio=1 has been adopted to describe molecular permeability in these types of membranes [98]:   [6-11] Where PC is the permeability of the continuous phase, PD is the permeability of the disperse phase, and φ D is the volume fraction of the dispersed phase. These mixed matrix membranes utilize a highly permeable (zeolite) particle phase to achieve improved permeability and selectivity performance. In our system, the matrix phase is PEO-PU polymer (PC=PPEO-PU,H2O = 45000 to 60000 Barrer) containing a lower permeability fibrous phase (PD = PPAN,H2O = 300 Barrer), the fibers are used as a carrier for the permeable polymer phase.  The MMM model applies for particle fillers, however a fiber filled film layer would be expected to behave differently than a particle filled layer. Fibers have significantly different aspect ratios than particles, the fibers interact by overlapping in the layer, and fibers are not ideally dispersed as particles might be. A random fiber mat tends to have fibers with their length dimension distributed parallel to the x-y plane of the film more so than in the z-plane (ie. the fibers length tends to be normal to direction of permeation). It would be more desirable to use a fiber-reinforced polymer composite models to describe the system.  For fiber-reinforced polymer composites, models have been presented to predict the effective diffusivity (Deff) of the composites for a permeable matrix (Dm) and impermeable fibers (Df). In )(22)(22DCDCDDCDCDCeffPPPPPPPPPP149  the present case the permeability to water vapour of the PAN polymer used in the fibers is several orders of magnitude lower than that of the PEO-PU used in the matrix. For a fiber filled composite with impermeable fibers in a permeable matrix (where Df << Dm), with diffusion occurring normal to the direction of the fibers Shen and Springer reported that effective diffusivity of the composite was [100]:   [6-12] The various models describing effective diffusivity in fiber filled composites in the literature are focused on materials that are significantly thicker than the materials in our study. The models are concerned with slow diffusion of water or water vapour into the matrix without a concentration differential maintained across the thickness of the material. In our case, the matrix polymer has high permeability to water vapour, and is in a thin film layer. The material is being used as a membrane and water vapour is absorbing at one interface, diffusing through the film layer, and desorbing at the other interface, based on a vapour pressure differential which is being maintained across the membrane surface at steady state. As discussed in Chapter 4, the solubility coefficient of the film layer must also be accounted for by factoring the non-absorbing fiber volume in the polymer:   [6-13] And from the solution diffusion model:  SDP   [6-14] Our proposed ‘fiber-filled-film’ membrane model for the effective permeability of a permeable matrix filled with impermeable random fibers for water vapour permeability as a function of fiber volume fraction is determine by combining equations [6-12],[6-13], and [6-14]:   [6-15] The calculated permeability of the film layer for all membranes based on gravimetric measurements of the film thickness based on the measured total PEO-PU areal coating weight and the measured fiber areal weight is summarized in Figure 6-7 along with both the fmeff VDD21 fmeffVSS 1 ffmeffmeffmeffVVSSDDPP 121150  present model based on the Shen and Springer and for the mixed matrix membrane model. Only membranes with less than 50 GPU oxygen permeance are plotted to eliminate any materials that might have significant defects or voids in the fiber filled film layer, which would affect the modeled permeability. Here the permeability appears to be much greater than expected in nearly all cases. However, the Vf and permeability data based on the SEM measured thicknesses of the films (such as those summarized in Figure 6-6) are added to the graph in Figure 6-7 and they appear to align better with the fiber-filled membrane model.  Figure 6-7: Effective permeability over matrix permeability as a function of fiber volume in the membrane with two models based on permeability and fiber volume calculated based on gravimetric measurements of the coating deposition in the membranes, and thickness measurements based on SEM imaging. From the results in Figure 6-6, along with a number of other membranes in which the film thickness was measured by SEM, the actual SEM measured film layer thickness is on average about 40% of that measured gravimetrically by the coated weight. This is due to excess coating entering the non-woven carrier layer, which is shown in Figure 6-15a and discussed further in the Section 6.7.1 below. The excess coating in the non-woven increases the calculated weight of PEO-PU polymer in the membrane, and decreases the calculated fiber volume fraction, but since this excess coating is not in the nanofibrous film layer, it does not 151  contribute to the actual resistance to transport of the membrane, the Vf, or the thickness of the film layer. Using a 40% factor for the amount of the total PEO-PU coating weight deposited in the film layer, the average film layer thickness, permeability, and fiber volume fractions of the membranes are recalculated and the model has an improved fit with the data as shown in Figure 6-8.  Figure 6-8: Effective permeability over matrix permeability as a function of fiber volume in the membrane with two models based on permeability and fiber volume calculated based on corrected gravimetric measurements of the coating deposition in the membranes accounting for the excess coating deposited in the non-woven carrier layer of the membranes. Future studies should further validate this model based on the fabrication and testing of precisely controlled fiber filled films without a non-woven carrier to eliminate the uncertainty associated with polymer in the non-woven layer. Furthermore, sorption and permeability studies should be carried out on these films to validate the effect of fiber volume on the solubility and diffusivity coefficients in defect free films of exact geometry and composition.  6.6 Defects in the Nanofibrous Film Layer Leakage, cross-over, and low selectivity are undesirable for membranes for water vapour separations. For a continuous defect-free film layer, the oxygen permeability of the membrane will be low, and high selectivity is achieved. However, any defects in the film layer allow both 152  convective transport of air and contaminants through the pores via differential pressure across the membrane, and diffusive transport of contaminants and gases through the defect pores. In the fabrication of impregnated electrospun nanofibrous membranes, a number of defects are observed and are associated with material system and fabrication conditions. Figure 6-9 through Figure 6-13 show surfaces of impregnated electrospun nanofiber membranes with defects at various fiber diameters, fiber loadings, and polymer solids in the impregnation solution. Figure 6-9 shows defects in membranes made with low fiber loading (0.185 g/m2) and low coating solids (12%), the estimated number of fiber layers for this membrane based on the diameter and fiber areal density is five layers (estimated nf=5), and the defects are the result of insufficient polymer in the coating solution to create a dense film across the pores, with the limited density of fibers available. This membrane is close to achieving the oxygen permeance target (<100 GPU) and has high water vapour permeance well above the performance target (>10000 GPU), but is just on the edge of achieving the selectivity target, increasing the fiber loading or coating solids slightly, eliminates the defects in similar materials. Similarly, the membrane in Figure 6-10 is close to the selectivity target, but defects are present near interface with the underlying non-woven fibers, here there are nearly sufficient fiber layers to create a dense film (estimated nf=9) and the nanofibers are observed to be intact at the defect, but the fibers were of an insufficient density near the underlying non-woven fiber for a film to be created. A slight increase in fiber loading eliminates the defects in similar membrane materials. The membrane shown in Figure 6-11 is very close to the performance target but has very minor defects near the underlying non-woven, increasing the solids concentration in the impregnation solution allows defect free membranes to be fabricated. Alternatively, a very thin second PEO-PU coating layer could be applied as a ‘defect’ filling layer in membranes with only minor defects such as those shown in Figure 6-11. The previous three examples of defects were of membranes that were close to the performance target, the membrane in Figure 6-12 had too many defects to be tested for performance. This membrane was made from larger fibers (average 1250nm) with only an estimated 3 fiber layers, there is insufficient nanofiber structure to support the casting of a continuous film in the fiber layer. A much higher density of fibers is required support a continuous coating. Increasing the fiber loading as shown in Figure 6-13, leads to a much lower level of defects in the film, but the nanofibrous layer is under-filled for this system due 153  to the use of a 16% polymer solids coating solution, leading to defects. Increasing the concentration in the casting solution and/or slightly increasing the fiber loading should eliminate the defects, but the vapour transport performance is already low for this membrane (only 5300 GPU), the benefit of using lower fiber diameter nanofibers is evident from this example.   Figure 6-9: Low Fiber Diameter - Low fiber loading – Low Coating (df=225 nm, ωf=0.185 g/m2, nf=5, 12% polymer solution: PO2/δ=140 GPU, PH2O/δ = 17700 GPU, αH2O/O2=126)  Figure 6-10: Mid diameter – Low fiber loading – Mid coating (df=253 nm, ωf=0.214 g/m2, nf=9, 16% polymer solution: PO2/δ =136 GPU, PH2O/δ = 8712 GPU, αH2O/O2=64)  154   Figure 6-11: Mid diameter – Mid fiber loading – Low coating (df=566 nm, ωf=0.44 g/m2, nf=13, 12% polymer solution: PO2/δ =142 GPU, PH2O/δ = 11000 GPU, αH2O/O2=77)   Figure 6-12: High Fiber Diameter – Low fiber loading – Mid coating (df=1250 nm, ωf=0.377 g/m2, nf= 3, 16% polymer solution, too leaky to test)  Figure 6-13: High Fiber Diameter – High fiber loading – Mid coating (df=1250 nm, ωf=1.26 g/m2, nf= 12, 16% solution: PO2/δ=361 GPU, PH2O/δ=5285 GPU, αH2O/O2=15) 155  6.7 Gravure Coating of Nanofibrous Membranes For electrospun nanofibrous membranes the gravure coating process would have the benefit of allowing a controlled deposition of polymer into the nanofibrous layer. Since we found previously that PAN nanofiber layers readily wick aqueous polymeric dispersions, and wetting of the nanofibers with the coating solutions was rapid, it was reasonable to assume that coating uniformity with the gravure method would not be a challenge. Theoretically, any level of polymer coating could be deposited into the nanofibrous layer by selecting an appropriate gravure pattern and cell size, and controlling the coating solution polymer concentration. In the fabrication process the fiber deposition and impregnation solution concentration would still be expected to affect the final membrane structural and performance properties, but using this gravure coating method should allow much greater control over the membrane fabrication process. Furthermore, gravure coating is a commonly used in industrial roll-to-roll coating processes, so this method would be desirable from a manufacturability perspective. 6.7.1 Gravure Coating Method Samples of electrospun fiber attached to one surface of the nonwoven were cut into sheets (30x 20 cm). The dry nanofiber coated nonwovens were weighed dry and the pre-coated basis weight was determined by dividing the weight of the sample by the area of the sample. The samples were affixed to a rubberized bed for coating. A hand gravure proofing device (Pamarco, Evenflo) was used to deposit a coating formulation onto the nanofibrous layer in the lab trials. The gravure has a known cell volume and defined pattern as summarized in Table 6-5. The hand gravure has a fluid reservoir into which a coating formulation of a known solids concentration is deposited. The gravure roller is fed from this reservoir, filling the etched surface of the gravure roller, excess coating solution is scraped from the gravure using a ‘doctor blade’, leaving only the ‘cells’ of the gravure filled with coating solution. The ‘primed’ gravure then interfaces with the nanofibrous layer. Pressure is applied to the roller during the coating process, and the filled cells contact the nanofibrous layer transferring the coating to the nanofibrous layer. The hand gravure and gravure coating process is shown in Figure 6-14. Solution in the gravure cells are deposited on the nanofibrous layer and wick into the layer, wetting the nanofibrous layer. Deposition efficiency is affected by the solution fluid properties and impression pressure.  156   Figure 6-14: Benchtop hand gravure roller, gravure cell pattern, and coating method. Table 6-5: Gravure cylinders used in the study. Gravure Type Gravure Cell Geometry Cell Volume (cm3/m2) 300Q3 Quad 7.0 200Q174 Quad 14.7 200Q63 Quad 8.2 140Q18 Quad 21.2 120Q21 Quad 35.0 85Q177 Quad 50.8  An advantage of the gravure method over the dip coating method is that there is no excess coating deposited into the membranes. During the dip coating process, the non-woven layer is also filled with polymer solution, and excess polymer is deposited into the non-woven fibers, and at the nanofiber-nonwoven fiber interface. This can lead to added resistance to vapour transport, due to greater film layer thickness at the film-non-woven interface, as well as by blocking pores in the non-woven. A small amount of excess is desirable, to help bind the impregnated nanofibrous layer to the non-woven, however it should be minimized. Since the 157  gravure method allows a controlled volume of coating solution to be deposited into the nanofibrous layer, better performance membranes can be fabricated.  Figure 6-15: Cross-sectional images comparing (a.) dip-coated and (b.) gravure coated nanofibrous layers attached to the microfibers of the non-woven carrier; arrows show excess coating deposition in the non-woven carrier layer. 6.7.2 Gravure Coating Results Three levels of fiber deposition (250 nm average diameter) were tested in the design matrix (0.25, 0.50, 0.75 g/m2), with three levels coating formulation solids (18, 26, 33%), with 6 gravure types consisting of a range of cell volumes (7 to 50 cm3/m2). Based on the solids and cell volumes, the theoretical deposition for the dry PEO-PU in the membrane ranges from 1.4 to 18 g/m2.  Figure 6-16 shows the water vapour permeance and water vapour/oxygen selectivity of the gravure coated membranes. Clearly, high solids coating formulations (33%) do not lead to adequately selective membranes, this is discussed further below in the section on wetting of nanofibrous layers and the viscosity of the coating formulation. When the nanofibrous later deposition is low (0.25 g/m2), there are also a large number of defects in the selective layer, and the membranes have poor selectivity. Samples which met the permeance and selectivity criteria where at medium to high fiber loadings (0.5 to 0.75 g/m2) with medium to low formulation solids (18 to 26%). Higher fiber loading (0.75 g/m2) leads to better selectivity (less defects) due to better wetting of the coating solution, but lower permeance due to the increased fiber ratio and a thicker fiber layer. The permeance of water vapour against the expected dry coating deposition based on the gravure cell volume and the formulation solids is plotted in Figure 6-17 for all samples with >100 selectivity for water vapour over oxygen. From Figure 6-17, increasing the coating weight as expected decreases the permeance as 158  there is a thicker selective film layer. The effect of the higher loading fibers can be observed in Figure 6-17 as well, with lower fiber loading (0.5 g/m2) leading to increase overall permeance of water vapour.   Figure 6-16: Performance of nanofibrous membrane made from gravure coating with various gravure cell volumes, at three polymer coating concentrations and three nanofiber loadings, only membranes with >100 selectivity are shown on the right; all data is from one membrane sample at each fabrication condition.  Figure 6-17: Permeance of adequate selectivity nanofibrous membranes at various coating impregnation levels (based on gravure volume), all data is from one membrane sample at each fabrication condition. 159  Surface images of a gravure coated nanofibrous membrane are shown in Figure 6-18, here the gravure pattern can be observed in the image on the left, this was due to the high solution viscosity of the 33% solids coating solution, and may be improved with better wetting or lower solids/viscosity as discussed in the following section. The membrane is defect free as observed on the right image in Figure 6-18.   Figure 6-18: Gravure coated nanofibrous membrane surface images,0.52 g/m2 fiber loading, 260nm fibers, 8.2 cm3/m2 200 quad gravure, 33% solids solution; PO2/δ=57 GPU, PH2O/δ=14366 GPU, αH2O/O2=248) Cross-sectional images of two nanofibrous membranes coated with low (8.2 cm3/m2) and high (35 cm3/m2) cell volume gravure etch patterns are shown in Figure 6-19. Both membranes achieve the <100 GPU oxygen permeance target with permeance of 20 and 2 GPU respectively, but the lower cell volume gravure leads to a structured film layer, with a lower effective thickness and so the water vapour permeance is significantly higher (15000 GPU versus 8000 GPU).  160  Figure 6-19: Gravure coated nanofibrous membrane cross-sectional images, 0.75 g/m2 fiber loading, 260nm fibers, 26% solids coating solution, two gravure cell volumes; 8.2 cm3/m2 (left image) and 35 cm3/m2 (right image). 6.8 Coating Formulation Parameters and Wetting The gravure coating method relies on the coating within the gravure cells to be deposited as discrete droplets on the nanofiber surface and then wet the nanofibrous layer spreading to create a continuous wet film layer within the nanofiber matrix before drying and solidifying into a continuous defect free polymer film. During the wetting process the surface tension of the coating solution will determine how well the fibers are wet, and the viscosity of the coating solution will determine the resistance to the spread of the coating. In order to study this phenomena, nanofibrous layers of constant fiber diameter were created at different deposition levels, and coating solutions with different polymer solids content (variable viscosity) and with different levels of surfactant (variable surface tension) were deposited as controlled droplets on the nanofiber surface. Triton 100X was used to modify the surface tension of the coating solutions. The properties of the solutions are summarized in Table 6-6. Increasing the polymer solids content in the dispersion increases the viscosity, the addition of Triton 100X decreases the surface tension, the addition of Triton also modifies the viscosity to a greater extent for solutions with higher polymer solids. Table 6-6: Formulation parameters for coating dispersions. PEO-PU Polymer Solids in Solution Concentration of Triton 100X Solution Viscosity (cP) Solution Surface Tension (mN/m) 10 0 <10 46.3 10 1 <10 33.2 10 4 <10 32.8 17 0 11.6 45.4 17 1 12.2 36.5 17 4 14.8 32.5 24 0 27.2 45.2 24 1 32.6 37.4 24 2 42.3 34.6 24 4 160 32.7 33 0 241 46.8  161  A controlled droplet of solution (5 µL) was placed on the surface of the nanofibrous web using a micropipette and allowed to dry at room conditions, the diameter of spherical half droplet as it touches the surface for a 5 µL droplet is 1.7mm. The diameter of the wetted area was measured to determine the extent of wetting, an example of the spreading of various droplets on a nanofibrous layer is shown in Figure 6-20. The results are summarized in Table 6-7.  .  Figure 6-20: Droplets of various volume and coating formulation on a nanofibrous surface after drying. The surface energy of PAN is reported to be near 44 mN/m, since the coating solutions are all near or below this surface tension, the coating solutions should all wet the PAN nanofibers and not remain as droplets on the surface [188]. Furthermore, capillary phenomenon should cause wicking of the polymer dispersion into the fibrous matrix. As shown in Figure 6-20, the various formulations all wet and wicked into the nanofibrous layers. In Table 6-7, the lowest viscosity formulations (10% solids) had the greatest overall spreading, and for equivalent formulation surface tension, the viscosity dominated the degree of wicking. Without any modification to the surface tension of the formulations, the extent of wetting was improved for fiber densities greater than 0.25 g/m2, however at higher viscosity (24% solution), the wetting appears to decrease somewhat at 1.09 g/m2. This aligns with the results of the gravure coated samples reported in Figure 6-16, where lower fiber loadings (0.25 g/m2) did not lead to any 162  defect free membranes, and also high solids, higher viscosity formulations (33%, 241 cP) were not successful at 0.75 g/m2 fiber loading and only one coating level at 0.5 g/m2 and 33% solids was successful in fabricating a membrane which achieved the performance target. This membrane, whose surface was shown in images in Figure 6-18, had a gravure cell pattern on the membrane surface, which indicated poor spreading of the coating into the nanofibrous layer. Overall, from the gravure results reported in Figure 6-18, a nanofibrous loading in the range of 0.5 to 0.75 g/m2 and a coating solution in the range of 18% polymer concentration appears lead to the best performing membranes in terms of selectivity and water vapour permeance. The wetting results Table 6-7 confirm these results, and indicate that decreasing the surface tension of the coating solution may allow coating with a broader range of polymer concentrations and lower overall coating weights due to better spreading. Table 6-7: Results of wetting tests on various fiber mats, errors are sample standard deviations from the measurement of three wetting tests. Fiber Areal Density (g/m2) Polymer Solids (%) 10 17 24 Wetting Diameter (mm) for PEO-PU (aqueous dispersion) 0.25 6.2 +/- 0.3 2.8 +/- 0.2 3.1 +/- 0.3 0.52 11.1 +/- 0.5 9.4 +/- 0.9 5.2 +/- 0.5 1.09 11.9 +/- 1.9 8.2 +/- 0.2 4.7 +/- 0.3 Wetting Diameter (mm) for PEO-PU with 1% Triton 0.25 12.5 +/- 1.7 12.6 +/- 0.5 5.7 +/- 0.4 0.52 15.5 +/- 1.4 13.5 +/- 0.5 7.5 +/- 0.3 1.09 15.3 +/- 0.5 14.0 +/- 1.0 4.9 +/- 0.5 Wetting Diameter (mm) for PEO-PU with 2% Triton 0.25 - - 7.0 +/- 0.4 0.52 - - 6.9 +/- 0.7 1.09 - - 4.2 +/- 0.6 Wetting Diameter (mm) for PEO-PU with 4% Triton 0.25 14.9 +/- 0.4 12.7 +/- 0.5 - 0.52 17.4 +/- 0.3 13.8 +/- 0.6 4.0 +/- 0.5 1.09 17.2 +/- 0.3 14.6 +/- 0.4 4.1 +/- 0.2  163  6.9 Conclusions  In this Chapter electrospun nanofibrous membranes for water vapour transport have been fabricated which exceed the 10000 GPU target for water vapour permeance while achieving the 100 GPU oxygen permeance target for low defect membranes. This was achieved by building on the work in Chapter 5 and reducing the fiber size. Fabricating impregnated electrospun nanofibrous membranes based on 200 to 260 nm fibers at areal densities in the range of 0.3 to 0.75 g/m2 produced the best performing membranes in terms of water vapour permeance and selectivity of water vapour over gases. A model to account for the permeability observed in these fiber-filled-film membranes was presented and validated against a range of membranes, this model should be further validated and refined in future work. An alternative method of membrane fabrication was also presented using a gravure coating method. By using this gravure coating method, the fabrication of these membranes should be scalable to roll-to-roll manufacturing processes. The method also allows the metering of the coating polymer into the nanofibrous layer, which improved the membrane performance. Finally, coating solution modification experiments were completed to gain insight into the findings of the gravure coating experiments, and suggest opportunities for future fabrication improvements through the modification of the properties of the impregnation solution.    164   Formable Impregnated Electrospun Nanofibrous Membranes 7.1 Introduction In the previous chapters high performing membranes based on electrospun nanofibers have been introduced. These membranes are based on depositing electrospun nanofibers on the surface of a non-woven carrier based on microfibers (10 to 50 μm fibers). The nanofiber layer was then impregnated with a coating solution containing a polymer which is permeable to water vapour and blocks the transport of gases. The coating solution was dried to remove the solvent (water) and a dense polymer film incorporating the nanofibers remains attached to the non-woven surface (refer to Figure 5-1 for a pictorial summary of the process). The membranes demonstrated were flat-sheet materials that could be used in cross-flow or counter-flow energy recovery exchanger designs as shown in Figure 7-1.  Figure 7-1: Cross-flow and counter-flow orientations in membrane air-to-air exchangers. These exchangers are used in energy recovery systems which use fans to move air through the exchanger channel and supply/extract air from locations inside a building. It is desirable to minimize the pressure drop (or losses) associated with air flow through the exchanger channels, which will lead to lower fan power and maximize the overall system efficiency. Membranes with higher transport of moisture lead to exchangers with higher latent transport effectiveness at an equivalent fan power. Counter-flow exchanger designs allow for effectiveness improvements over cross-flow designs due to the increased average temperature or concentration differential generated within the exchanger channels when air is flowing in counter-current directions over the membrane surfaces. In flat plate devices, the  165  membrane is separated by spacers to create channels through which air flow is directed over the opposing surfaces of the membrane. Transport in these devices only occurs through the plane of the flat membranes, as shown in Figure 7-1. The highest efficiency heat exchanger designs incorporate multi-surface heat exchange, in which every surface in the heat exchanger is used for heat transport. This is accomplished by ‘forming’ a plastic or metal sheet into a heat exchanger plate with channels formed into the sheet. These structures when assembled into stacks can direct air flow over the exchange surface in such a manner as to allow multi-directional heat exchange, with heat transport occurring through all surfaces in the exchanger channels. An example of a plastic sensible (heat) only exchanger with square channel cross-sections (a.) and plate (b.) is shown in Figure 7-2 to demonstrate the concept.  Figure 7-2: (a.) A sensible (heat) only exchanger with multi-directional transport based on formed polymer sheets, (b.) an image of the entrance region of a formed heat exchanger plate [189]. A ‘formable membrane’ which could be formed similar to these heat exchanger plates would allow the creation of exchangers with multi-directional heat and mass (moisture) transport, as well as improved overall efficiency. A membrane which can be formed in such a manner would allow significant improvements in energy recovery ventilator design through improved heat and moisture transport performance, decreased exchanger pressure drop per unit area of membrane, the elimination of costly spacer plates of flow channels, and improved manufacturing assembly through-put. Designs based on rectangular, square or triangular channel cross-section can be used as shown in Figure 7-3.   166   Figure 7-3: Formable membrane concepts with triangular and square channel cross-sections, and a concept for a triangular channel design with alternating formed and flat sheet membranes (right image) [190]. One advantage of using electrospun nanofibers to make membranes is the capability of spinning fibers on a wide variety of support layers. The fibrous layer can be selected from the wide variety of polymers which can be electrospun, and the fiber layer structure can be controlled through fiber diameter, morphology, and deposition. The nanofibrous layer can then be filled or coated with a non-porous or ‘dense’ selective polymer layer, and through this processing step may be simultaneously bound to the support layer. In this chapter it is proposed that selecting and combining supports, fibers, and dense polymer layers which can be thermally formed will lead to membranes which can be thermally formed into ERV exchanger plates. Several studies have considered the forming of fibrous, woven, and composite structures [191]–[193]. Cai and Ko defined the required dimensionless energy for forming (UF*) of a fibrous composite pre-form defined by:   3*FFFFhVEUU   [7-1]  167  Where UF is average forming energy from experiment, EF is the fiber modulus, h is the preform thickness, and VF is the fiber volume fraction. From this equation, increasing the fiber modulus, thickness, and fiber density will increase the energy required to form a preform. In the forming of impregnated electrospun nanofibrous membranes on non-woven carrier layers, the non-woven carrier will comprise most of the weight fraction of the total membrane. This means that the physical and thermo-mechanical properties of the non-woven will define the overall formability, forming energy, and structural characteristics of the formed membranes. However, the selective fiber-filled polymer film layer is thin, and its structure must remain intact through the forming process to make formed membranes with adequate selectivity. This means that the film layer will define the limits of forming and the allowable elongation in the formed membranes. This elongation limit will define the geometry of the mold design for membrane forming, and the total draw in the mold cannot exceed the elongation at which defects are created in the membrane film layer. In this chapter it is hypothesized that the IENM approach could be used to fabricate membranes which could be thermally formed into exchanger plates. This chapter presents thermal-mechanical studies on the carrier support layer and the fiber-filled-film layers in IENM’s to select materials and compositions which allow the membranes to be formed into exchanger plates without defects. This is the first report of a designed thermally formable membrane for ERV exchangers in the literature. This chapter demonstrates nanofibrous membrane structures which can be formed into exchanger plates, and discusses some of the current challenges and limitations of fabricating and forming these membranes.  7.2 Carrier Layer 7.2.1 Non-woven Material Properties  A number of non-woven carrier layers based on microfibers were evaluated as the carrier layer for the formable nanofibrous membranes. Since the non-woven carrier comprises most of the weight of the membrane, the non-woven properties will largely define the mechanical properties of the membrane. Polyester (PET) based non-wovens were selected since they can be thermally formed at temperatures above their glass transition near 80°C and then cooled to maintain their shape. The usage temperature range of the exchanger will be between -25 to 50°. This is well below the glass transition of the carrier layer, and the formed structure should be stable during the operating lifetime of the device. A summary of the basic properties for the non-woven carriers tested in this study are summarized in Table 7-1. 168  Table 7-1: Non-woven carrier layer properties. Material Name Supplier Material Type Basis Weight (g/m2) Thickness (µm) Fiber Profile Average Fiber Diameter (µm) Smash™ 15100 Asahi Kasei PET spunbond 105 226 Round 21 Reemay® 2016 FiberWeb PET spunbond 46 254 Triobial 21 Reemay® 2214 FiberWeb PET spunbond 46 229 Round 16 Reemay® 2024 FiberWeb PET spunbond 71 305 Triobial 21 7.2.2 Thermal Analysis of Carrier Layer Thermograms for 3 types of non-woven polyester carrier layers are shown in Figure 7-4, and the transitions and endotherms of interest are summarized in Table 7-2.   Figure 7-4: First heating cycle DSC for various non-woven carriers (20°C /min). All of the carrier layers show a glass transition in the range of 76-81°C and do not begin melting until above 235°C with peak melting is at 250°C. These values are typical for poly(ethylene terephthalate) (PET) based materials [194]. The forming temperature range for  Heat Flow (W/g)0 50 100 150 200 250Temperature (°C)                 SMASH-15100–––––––                 R2016–– –– –                 R2214– – – –Exo Down169  membranes based on these non-woven carrier layers will be between 85 and 200°C, between the non-woven glass transition and melting onset. Table 7-2: Summary of thermal properties of the non-wovens Material Tg (°C) Melt Onset (°C) Tm (°C) ΔHm (J/g) % Crystallinity Smash™ 15100 80.5 235.4 252.2 50.7 36.2 Reemay® 2016 80.5 238.2 251.0 42.8 30.5 Reemay® 2214 80.4 243.3 253.6 47.6 33.9 Reemay® 2024 78.4 239.1 252.0 45.5 32.5  7.2.3 Thermo-mechanical Testing of Carrier Layers The non-woven mechanical properties at various temperatures are summarized in Table 7-3. It can be seen that increasing the temperature leads to a decrease in modulus, yield strength, and an increase in elongation for all non-wovens. A significant change is observed going from 60 to 95°C as the PET non-woven passes through the glass transition. Further increase in temperature leads to an increase in elongation. The stiffness and strength tend to be lower in the cross-roll direction (XD) compared to the machine direction (roll direction or MD) for the non-wovens. This means that forming the samples should require less force to create features when elongating the samples in the cross-direction. The Reemay® non-woven samples show significantly less elongation before break at elevated temperature than the Smash™ non-wovens. For this reason, only the Smash™ 15100 non-woven was selected for fabricating membranes. Figure 7-5 shows a summary of the averaged isothermal stress-strain curves for the Smash™ nonwoven at different temperatures in the roll direction (MD) of non-woven.  Increasing the temperature decreases the modulus, decreases the ultimate strength, and increases the elongation. A notable increase in elongation is observed by increasing the temperature from 125 to 150°C. This is further confirmed in the dynamic mechanical analysis results summarized in Figure 7-6, where the storage modulus for the Smash™ non-woven is still decreasing until >125°C, this mean that better draw would be expected above this temperature. From Table 7-3 the elongation is not significantly different between the roll (MD) and cross-web (XD) directions at all temperatures measured, so the orientation of forming should not have a great impact on the formability of the materials. 170  Table 7-3: Mechanical properties of non-woven carriers in machine (MD) and cross (XD) directions at various temperatures, errors are the sample standard deviations for measurement of three materials samples. Material Tiso (°C) Direction  Modulus (MPa) UTS (MPa) Yield, 2% (MPa) Elongation (%) Smash™ 15100 25 MD XD 387±81 234±6 11.0±1.0 7.0±0.5 9.92±0.88 5.46±0.87 30±11 31±14 Smash™ 15100 60 MD 285±49 9.8±1.3 5.5±0.9 116±30 Smash™ 15100 80 MD 68±18 9.2±0.9 2.81±0.18 152±13 Smash™ 15100 95 MD XD 21±4 10±1 9.1±0.6 5.0±0.4 1.42±0.06 0.85±0.24 183±7 172±14 Smash™ 15100 110 MD 19±2 8.6±0.6 0.98±0.27 195±17 Smash™ 15100 125 MD 17±2 7.9±0.6 0.90±0.31 198±8 Smash™ 15100 150 MD XD 10±2 5±2 6.4±1.4 4.4±0.4 0.15±0.09 0.12±0.04 236±57 245±36 Reemay® 2016 25 MD XD 285±36 81±5 11.2±1.1 6.1±1.9 4.90±0.58 2.31±0.59 31±7 43±10 Reemay® 2214 25 MD XD 220±5 185±34 10.2±0.6 9.8±1.6 4.07±0.36 3.71±0.62 54±7 56±3 Reemay® 2024 25 MD XD 292±40 184±19 15.1±4.1 9.7±1.7 5.25±0.90 3.96±0.35 37±8 40±8 Reemay® 2024 150 MD XD 15+1 7±1 4.3±0.4 2.3±0.6 0.33±0.03 0.14±0.04 60±6 64±10  171   Figure 7-5: Average stress-strain curves for Asahi Kasei Smash™15100 Non-woven with increasing temperatures, curves are averaged from three sample measurements at each temperature.  Figure 7-6: DMA of non-wovens in tension in machine direction (left) and cross-direction (right), 1Hz frequency, 10 micron amplitude, heating at 5°C/min, data is for one sample of each material. 172  7.3 Nanofiber + PEO-PU Film Layers 7.3.1 Thermal Analysis of Fiber, Film, and Fiber-filled Film Layers Differential thermograms were taken of the electrospun polyacrylonitrile (PAN) fibers (10% PAN, average fiber diameter = 250 nm), films of the crosslinked polyether-polyurethane (PEO-PU) copolymer, and a cross-linked PEO-PU copolymer filled PAN fiber filled film (5.9 g/m2 of spun PAN filled with 13.4 g/m2 of PEO-PU, Vf=0.29 fiber volume). The first heating cycle thermograms for each material are shown in Figure 7-7, samples were brought to equilibrium at -20°C before ramping to 150°C at 20°C/min on the first heating cycle. Samples were then cooled at 10°C /min to -85°C before reheating to 150°C at 20°C/min to produce the second heating cycle thermograms shown in Figure 7-8. All notable transitions are reported in Table 7-4. The cross-linked co-polymer film shows a prominent transition at 50.9°C in the first heating cycle, this transition has a small endothermic peak, associated with the melting of the soft block in the polyether-oxide side chains of the copolymer, this has been discussed in Chapter 4. In the first heating cycle of the copolymer a gradual endotherm in the PEO-PU copolymer film begins immediately following the the melting endotherm  and peaks around 130°C. This endotherm is not present on a second heating cycle, and is believed to be associated with loss of residual water from the polymer. In the second heating of the copolymer film, shown in Figure 7-8 a melting peak is observed at 235°C, with a enthalpy of melting of 4.37 J/g, which is associated with the melting of crystaline ‘hard segments’ in the copolymer. This melting begins around 200°C and defines an upper forming temperature limit for the polymer membrane. The transition at -52.5°C in the copolymer film is associated with the glass transition of the PEO soft segments of the block-copolymer. In Figure 7-7, The electrospun polyacrylonitrile fibers (PAN) show two merged transitions at 67.6°C and 86.5°C in the first heating cycle. These peaks are absent from the second heating cycle, shown in Figure 7-8 which only has one transiton across the entire temperature range at 101.6°C. This transition is in the typical range reported for PAN polymer and PAN nanofibers of 85 to 140°C [143], [195].  The polyether-polyurethane copolymer filled PAN fiber film shown in Figure 7-7, has four transitions in the first heating cycle, between 0 and 150°C which align with the transitions observed in the first heating of the copolymer film and the fibers individually. The transition associated with the PEO-PU polymer soft segment melting is shifted down from ~51°C to 173  43.1°C, which indicates that the PAN nanofibers likely interact with the soft segments of the copolymer. Additionally, as shown in Figure 7-8, the soft segment glass transition is shifted in the fiber filled film from -55.3°C, down about 3°C from the film alone, which further suggests some interaction of the nanofibers with the copolymer soft-segments. Modification of thermal behaviour and shifting of thermal transitions are commonly observed in polymer blends, and the extended thermal analysis of the PAN nanofiber-PEO-PU system as a polymer blend could be the focus of future studies [196], [197]. The second heating cycle for the fiber filled polymer film, as shown in Figure 7-8 has only three transitions, two associated with the copolymer, and one at 100.0°C associated with the PAN fibers. From a thermal-forming perspective, increasing the temperature of the membrane above 100°C prior to mechanical forming, will ensure that all polymers in the membrane (the PET in carrier layer non-woven, the nanofibers in the fiber-polymer layer, and the copolymer film) are all above their glass transition, meaning that all polymer chains will have good mobility under the forces of forming. From the DMA data for the non-wovens shown in Figure 7-6, greater than 125°C was recommneded for the improved elongation. Subsequent cooling of the membrnae below at least the Tg of the PET layer (70oC) will allow the formed membrane to maintain its shape after forming, cooling below 20°C would be recommended to ensure that the material is cooled below the melting of the PEO soft segments of the coating copolymer.  Figure 7-7:  First heating cycle of DSC (20°C/min) of crosslinked polyether-polyurethane film (PEO-PU, solid line), electospun polyacrylonitrile nanofibers (PAN_NF, dashed line), and a  Heat Flow (W/g)0 20 40 60 80 100 120 140Temperature (°C)                 PEO_PU_FILM–––––––                 PAN_NF– – – –                 PEO_PU+PAN_NF––––– ·Exo Down174  film of PAN nanofibers impregnated with crosslinked polyether-polyurethane co-polymer (PEO-PU+PAN_NF, dash-dot), curves are offset for visibility.  Figure 7-8: Second heating cycle (20°C/min) of crosslinked polyether-polyurethane film (PEO-PU, solid line), electospun polyacrylonitrile nanofibers (PAN_NF, dashed line), and a film of PAN nanofibers impregnated with crosslinked polyether-polyurethane co-polymer (PEO-PU+PAN_NF, dash-dot). Table 7-4: Thermal properties of electrospun PAN fibers, PEO-PU films, and PEO-PU filled PAN fibers. Material Cycle Tg,s T1 T2 T3 T4 PEO-PU film 1st heat N/A 50.9 - - - PAN Fibers 1st heat N/A - 67.6 86.5 - PEO-PU filled PAN fibers 1st heat N/A 43.1 66.9 86.9 103.7 PEO-PU film 2nd heat -52.5 28.7 - - - PAN Fibers 2nd heat - - - - 101.6 PEO-PU filled PAN fibers 2nd heat -55.3 43.1 - - 100.0  7.3.2 Thermogravimetric Analysis of Fiber-Film Layer Samples of the copolymer film and the copolymer filled nanofiber film were tested using TGA in air to determine the upper forming temperature for the membranes, testing was completed  Heat Flow (W/g)-70 -20 30 80 130Temperature (°C)                 PEO_PU_FILM–––––––                 PAN_NF– – – –                 PAN_NF+PEO_PU––––– ·Exo Down175  in nitrogen as well to compare the thermal degradation of the films. The melting of the hard segments in the copolymer peaks at 235°C with onset at 200°C, as shown by DSC in Figure 7-8. The TGA and DTG plots of both materials in Air and Nitrogen are shown in Figure 7-9 and summarized in Table 7-5. In both air and nitrogen, the copolymer film does not demonstrate significant degradation until above 250°C, which is slightly above the melting peak of the copolymer at 235°C. As previously discussed in Chapter 3, during  the thermal degradation of PEO-PU in nitrogen, two distinct peaks are observed in the DTA plot, the earlier peak at T1 found at 347°C is associated with the thermal degradation of the urethane segments in the copolymer, the second peak at 421°C is associated with the degradation of the PEO soft segments peaks [123]. In nitrogen all of the copolymer is lost by 500°C. For films containing the PAN nanofibers in nitrogen, the urethane degradation peak occurs at a lower temperature of 326°C, while the polyalkylene oxide degradition peak occurs at higher temperature 434°C. This may indicate some interaction of the PEO segments with the PAN fibers which was also suggested by the movement of the soft phase melting transition in the DSC analysis. Residue associated with the PAN in the fiberous matrix remains at 600°C. In air, oxidative degradation occurs, and char remains in both the PEO-PU and PEO-PU+PAN nanofiber above 500°C. Table 7-5: Summary of thermal degradation of PEO-PU and PAN nanofiber filled PEO-PU films in air and nitrogen. Material Air  Nitrogen  2.5% loss T1 T2 T3 Residue 2.5% loss T1 T2 Residue PEO-PU film 275 354 408 - 8% (500°C) 264 347 421 0% (500°C) PEO-PU filled PAN fibers 249 340 427 566 14% (590°C) 265 326 434 13% (600°C)  176   Figure 7-9: TGA and DTA of PEO-PU copolymer films and PAN nanofiber filled PEO-PU films in air and nitrogen atmospheres.  7.3.3 Thermo-mechanical Testing of Film Layers Samples were placed in a temperature controlled oven fitted around a tensile testing apparatus and the mechanical properties were determined at different temperatures.  Figure 7-10 shows the isothermal stress-strain curve for films of the PEO-PU copolymer as well as PAN nanofiber filled films of the copolymer with different fibers loadings at 150°C. The average results are summarized in Table 7-6, the results in show that the modulus and the ultimate tensile strength of the films increase with the fiber loading. The elongation is greater than 150% for all samples. A DMA of the polymer film and a fiber loaded film are shown in Figure 7-11, the storage of modulus of the fiber loaded sample is greater at all temperatures,  0.00.51.01.5Deriv. Weight (%/°C)0 100 200 300 400 500 600 700Temperature (°C)                 PEO-PU-N2–––––––                 PAN NF+PEO-PU-N2–– –– –0.00.51.01.5Deriv. Weight (%/°C)0 100 200 300 400 500 600Temperature (°C)                 PEO-PU-Air–––––––                 PAN NF+PEO-PU-Air–– –– –020406080100Weight (%)0 100 200 300 400 500 600Temperature (°C)                 PEO_PU in Air–––––––                 PAN_NF+PEO_PU in Air– – – –020406080100Weight (%)0 100 200 300 400 500 600 700Temperature (°C)                 PEO_PU in N2–––––––                 PAN_NF+PEO_PU in N2– – – –177  and a transition in the DMA plot is observable associated with the PAN fibers above 100°C, ending above 125°C. This further confirms the recommended forming temperature of >125°C.  Figure 7-10: Mechanical testing of PEO-PU films and PAN NF filled PEO-PU films at 0.2 mm/min and 150°C, averaged curves for three samples, only shown to minimum elongation for each sample set. Table 7-6: Summary of mechanical properties for PEO-PU films and PAN NF filled PEO-PU films (average for three samples) at 150°C, error is the sample standard deviation for measurement of three samples of each material. Material Coating (g/m2) Fiber (g/m2) % wt. Fiber UTS (MPa) Modulus (MPa) Elongation (%) Crosslinked (XL) PEO-PU 17.6 - - 0.9±0.3 1.0±0.4 349±247 XL-PEO-PU in PAN NF 6.1 13.0 31.9 1.2±0.03 4.7±1.1 290±69 XL-PEO-PU in PAN NF 5.0 3.0 37.5 1.7±0.3 7.2±5.0 167±22 XL-PEO-PU in PAN NF 9.1 10.0 47.8 5.1±1.2 12.4±2.6 320±68  178   Figure 7-11: DMA of a PEO-PU polymer film and a PEO-PU film loaded with PAN nanofibers in tension, 1Hz frequency, 10 micron amplitude, heating at 5°C/min, data is for single samples of each material. 7.4 Defects During Mechanical Deformation of Membranes A membrane was fabricated with 0.975 g/m2 fibers (10% PAN, average diameter = 250 nm) on the Asahi Kasei substrate. It was dip coated with a 12 wt.% PEO-PU copolymer dispersion with a 10:1 polymer to polycarbodiimide crosslinker solids ratio. The coating dispersion filled the nanofibrous matrix and wet the non-woven, excess polymer was removed with a lint-free critical task wipe and the sample was dried and cured at 50°C. The final sample had 10.4 g/m2 dry polymer in the 0.975 g/m2 nanofiber matrix, total fiber volume of 0.08. Leakage through the membrane at 3psi was zero, oxygen permeance was <14 GPU, water vapour permeance was 4850 GPU, H2O/O2 selectivity was 347.  Samples were then elongated at a rate of 0.2 mm/s to varying degrees [50%, 100%, 150%, 175%, break] at two temperature set-points [95°C, 125°C] and in the machine and cross direction for the non-woven [MD, XD].  Mechanical properties of the membranes in the various test are summarized in Table 7-7, each result is the average of three samples. The mechanical properties of the membranes are very similar to the non-woven alone at each temperature, since the polymer filled nanofibrous layer contributes very little to the final weight of the 010002000300040005000Storage Modulus (MPa)-100 -50 0 50 100 150 200Temperature (°C)                 PEO-PU Film–––––––                 PEO-PU - 42% PAN Nanofiber–––––––179  membrane (<2% wt). The modulus and yield decrease with temperature indicating that forming is energy will be lower at higher temperatures. The maximum elongation was slightly increased at higher temperatures.  Table 7-7: Mechanical properties of membranes at different temperatures and orientations, errors are the sample standard deviations for measurement of three materials samples.  Temperature (°C) Strain Limit Direction  Modulus (MPa) UTS (MPa) Yield, 2% (MPa) Elongation (%) 25 None MD 463±39 13.2±1.4 11.4±0.8 61±18 25 None XD 238±14 6.5±0.7 4.9±0.7 28±4 95 50 MD 45±8 4.6±0.2 1.6±0.5 42±3 95 100 MD 28±4 6.4±0.9 1.2±0.3 100±7 95 150 MD 35±12 9.9±0.3 1.6±0.2 148±7 95 175 MD 43±13 10.5±0.8 1.7±0.1 176±6 95 None MD 21±4 9.3±1.1 2.2±0.3 174±23 95 None XD 17±6 4.9±0.8 0.7±0.2 173±15 125 100 MD 16±3 4.7±0.8 0.8±0.4 96±5 125 150 MD 12±2 6.6±0.7 1.3±0.5 147±3 125 175 MD 12±1 7.4±0.6 1.1±0.6 193±26 125 None MD 16±1 7.7±0.9 0.7±0.6 179±12 125 None XD 8±1 4.3±0.5 0.4±0.1 188±18  Representative images of the membrane surfaces after elongation to each strain limit are shown in Table 7-8 for 95°C and Table 7-9 for 125°C at 100 and 500× magnification. All images are for elongation in the machine direction, defect formation begins between 150 and 175% elongation at 125°C, some deformation is observed by 150% at 95°C. The final image in Table 7-8 shows the defect formation in ‘cross-machine’ direction. It is clear that defect formation is due to tearing of the nanofiber filled film layer as can be observed in the images of the ‘no limit’ elongation samples. From the images the tearing appears to occur near the interface of the non-woven with the nanofibrous filled layer. Images of these defects in more detail are shown in Figure 7-12 where nanofiber breakage can be observed. The fiber-filled film layer is attached to the underlying non-woven fibers, as the non-woven is stretched, the film layer is stretched locally at points of contact between the film layer 180  and the non-woven fibers. Tear and breakage of the film results from excessive local stretching of the film in these regions. From these results, the elongation limit before defect formation for these membranes between 150% and 170% elongation, this corresponds to a maximum draw ratio of 2.5 to 1, where a flat sheet of the material would be formed into a channel with a perimeter 2.5 times the length of the flat sheet. Table 7-8: NF membrane elongated at 95°C to different levels of elongation. E (%) Magnification x100 X500 0   50   100   181  E (%) Magnification x100 X500 150   175   no limit (>175) Cross-web (XD)      182  Table 7-9: NF membrane elongated at 125°C to different levels of elongation. E (%) Magnification 100× 500× 150   175   no limit (190)    183   Figure 7-12: Defect in elongated nanofibrous membranes, direction of elongation is across the page, isothermal at 125°C, 185% elongation. 7.5 Forming of Membranes A press was built to test the formability of membranes, the press is shown in Figure 7-13, and one half of the forming mold is shown in Figure 7-14. In the press, the membrane was suspended in tension between the plates in the oven, and the oven, membrane, and plates were heated to a controlled temperature. The male and female plates were then pressed together at a controlled rate, compressing the membrane. A cooling loop was run through the plates, which was used to quench-cool the membrane (<10°C). The plates as shown in Figure 7-14 have a triangular channel with a depth of 2.5 mm and a span of 5.5mm, corresponding to a draw ratio of 1.6 to 1, meaning that a flat sheet of membrane must be elongated to about 30% without defect formation for the forming to be successful. The plates were coated with a PTFE layer to minimize any adhesion of the membrane to the plates.  Figure 7-13: Heated press designed and built for forming membranes, press has temperature control on the enclosure, and heating and cooling control on the plates, the rate of compress can also be controlled. 184   Figure 7-14: Plate mold geometry for membrane forming press. 7.6 Transport Properties of Membranes Impregnated electrospun nanofibrous membranes fabricated on a PET non-woven carrier (Smash™ 15100) layer using 250nm diameter electrospun PAN nanofibers of varying fiber deposition and varying coating solids in the impregnation solution. Samples were placed in the test mold and heated to the forming temperature stated in Table 7-10, the mold was then used to compress the membranes. The membranes were then cooled to below room temperature prior to removing from the mold. An example of a formed membrane is shown in Figure 7-15. Samples were cut from the triangular channel region of the mold for leak and performance testing. The membrane performance before and after forming at 100 and 125°C for a range of samples at different coating and fiber depositions are summarized in Table 7-10.   The oxygen permeance of the membranes after forming is an indication of any defects created on forming. It can be observed in Table 7-10 that increasing the forming temperature from 100°C to 125°C generally improves the formability of the membranes as the oxygen permeance of similar materials formed at the higher temperature is lower than at lower temperature. The membranes with higher fiber deposition and higher solids in the coating (an indication of the polymer coating loading) lead to formed membranes with lower oxygen permeance. Since forming elongates the membrane thereby making the fiber-filled film layer thinner, the water vapour permeance increases for the formed materials. Forming the membranes in an orientation such that the machine (roll) direction of the non-woven is perpendicular to the exchanger plate channel length leads to lower oxygen permeance and thus less defects. Formed membranes which had <100 GPU oxygen permeance and >10000 185  GPU water vapour permeance were achieved for membranes with higher fiber loadings (>0.65 g/m2) and lower coating solids (~14%). Table 7-10: Summary of performance properties of formed membranes, data are for individual samples of each material. Membrane Fabrication Properties Forming Condition Transport Properties PU Solids (%) Fiber Deposition (g/m2) O2 Permeance (GPU) H2O Permeance (GPU) Selectivity (H2O/O2) 14% 0.461 Unformed 100 °C 125 °C 44 209 60 9884 11497 10399 226 55 174 14% 0.663 Unformed 125 °C 6 28 6837 10334 1121 372 14% 0.71 Unformed 100 °C 125 °C 5 2227 16 7334 9511 9089 1548 4 562 17% 0.457 Unformed 100 °C 125 °C 29 2275 16 6223 8344 7802 217 4 482 17% 0.554 Unformed 100 °C 125 °C 6 20 5 5471 6837 7183 897 337 1516 17% 0.691 Unformed 125 °C 6 40 4691 7437 769 186 20% 0.451 Unformed 100 °C 125 °C 145 23 21 5302 6362 5907 36 277 283 20% 0.457 Unformed 100 °C 125 °C 15 13 28 4495 6550 6837 291 505 246 20% 0.714 Unformed 125 °C 6 8 3679 6223 603 733 20% 0.714 Unformed 100 °C 125 °C 2 - 8 3503 5386 4931 1682 - 581 14% 0.799 Unformed 150 °C - MD 150 °C - XD 10 90 109 10859 9030 11251 1143 100 103 186  Membrane Fabrication Properties Forming Condition Transport Properties PU Solids (%) Fiber Deposition (g/m2) PU Solids (%) Fiber Deposition (g/m2)  17% 0.799 Unformed 150 °C - MD 150 °C - XD 7 56 100 6685 8008 10264 948 143 103 17% 0.799 Unformed 150 °C - MD 150 °C - XD 11 93 211 7302 8340 8667 674 90 41 17% 1.240 Unformed 150 °C - MD 150 °C - XD 2 24 29 5232 6636 8743 2616 279 303 20% 1.240 Unformed 150 °C - MD 150 °C - XD 2 10 33 3417 5393 7241 1709 564 222   Figure 7-15: Formed nanofibrous membrane (a.) and magnification of flow feature in formed membrane showing no defects (b.). 7.7 Conclusions Membranes which could be formed into exchanger plates were fabricated based on electrospun nanofibrous membranes. A number of non-woven carrier layers based on PET microfibers were evaluated for use in these membranes and the thermal and thermo-187  mechanical properties of the nonwovens were evaluated. From this testing, a non-woven PET carrier layer was selected for further membrane fabrication and evaluation. The thermal and thermo-mechanical properties of the film and fiber filled film layers of the nanofibrous membranes were also evaluated to confirm maximum elongation and forming temperatures. From the materials evaluation forming temperatures of over 125°C were recommended, and membrane with higher fiber loadings appeared to show better elongation before break at elevated temperatures. Elongation tests on impregnated electrospun nanofibrous membranes over a range of temperatures were completed, and defects in the film layer were observed in the roll (MD) and cross-roll (XD) of the non-woven carriers at elongations over 175% at 125°C. A heated press was assembled to test the formability of these membranes at elevated temperatures, the exchanger plate molds for the press required a maximum membrane elongation of about 170% or a draw ratio of about 1.6 to 1. A series of nanofibrous membranes were fabricated and compression molded at various temperatures. Higher molding temperature leads to less defect formation in the formed membranes. Forming the membranes also leads to increased water vapour permeance due to stretching and thereby thinning the membrane selective film layer. Future work will involve membrane scale up and making prototype ‘formable membrane’ exchangers from these membranes. These exchangers are expected to have higher latent (water vapour transport) and higher sensible (heat transport) performance to flat sheet membrane exchangers at similar exchanger volume, exchanger face velocity, flow rate and pressure drop to the flat plate exchangers.   188   Conclusions and Recommendations for Future Work 8.1 Overview As outlined in Chapter 1, the focus of this thesis is on membranes for selective separations of water vapour from air. These membranes are the central functional component of heat and mass exchangers, used in the transport of heat and moisture between building ventilation supply and exhaust air streams. In modern sealed buildings mechanical ventilation is used to exhaust stale indoor air and supply fresh outdoor air to occupied spaces. ‘Energy recovery ventilation’ (ERV) exchangers are a critical component in modern high efficiency building ventilation systems. In these air-to-air exchangers outgoing air passes over one surface of a membrane, and incoming air passes over the opposing membrane surface. This allows building exhaust air to be used to pre-cool and dehumidify incoming fresh air in cooling conditions, and preheat and humidify incoming air in heating conditions. This improves the efficiency of building ventilation and allows engineers to meet building indoor air change-over requirements without having to significantly increase the size and energy load on heating and cooling equipment.  Since the exhaust air being removed from buildings contains indoor air contaminants (carbon dioxide, gases, VOCs) the membranes in these exchangers must be moderately selective and preferentially transport heat and moisture without transporting other gases and vapours from the stale exhaust air to the fresh supply air. Although membrane based exchangers are presently commercially available, there is ever-increasing focus on improving building energy efficiency, which drives a need to improve the heat and moisture transport effectiveness of these devices. By improving the membrane performance and functionality, higher efficiency exchangers can be created, leading to reduced energy consumption associated with ventilation, and improving overall building energy efficiency.  In this work, it was proposed that the vapour transport of ERV membranes could be improved by eliminating the need for a microporous substrate layer which is used in current generation composite ERV membranes. Evidence for this claim was presented in an analysis of composite membranes for ERVs in Chapter 4 of the thesis. It was hypothesized that one way to make ERV membranes without a substrate could be through the use of electrospun nanofibers. Our initial validation of this concept is presented in Chapter 5. However, the membranes did not exceed the water transport performance of composite membranes while maintaining adequate selectivity. We proposed that this is due to the use of polymers for the nanofibrous layer which are relatively impermeable to water vapour, and that these impermeable fibers create a new resistance to transport in the membranes. 189  The resistance associated with these impermeable fibers was quantified through a systematic approach to fabricating nanofibrous membranes with different fiber volume factions, and validating a model for the water vapour permeability of the fiber-filled-film layer in the membranes in Chapter 6 of the dissertation. Through the optimization of these nanofibrous membranes, materials which exceed the performance of composite membranes for ERVs were demonstrated. This validated the claim that nanofibrous membranes can be used to eliminate the substrate layer from ERV membranes and thereby improve their water vapour transport performance. Finally, we proposed that nanofibrous membranes can be made into thermally formable membranes for advanced exchanger designs. This is validated in Chapter 7. 8.2 Conclusions, contributions, and Future Work The purpose of this thesis work was to generate a knowledge base for the design and fabrication of improved performance membranes for air-to-air membrane based exchangers. This was to be achieved firstly through the chemical, thermal, and permeability analysis of selective polymers for these membranes, and then through material mass transport modelling and validation of the performance of current generation composite membranes. From this analysis, methods to improve the performance of these membranes were proposed and developed through systematic experimental optimization, modelling, and results directed refinement of material design. This approach led to membranes which as flat sheet membranes exceeded the performance of comparable current generation membranes, and had an additional ‘thermal-formability’ function allowing for the assembly of advanced exchanger designs. 8.2.1 Characterization of PEO-PU Polymers The first portion of this project, was to complete an analysis of the current generation composite membranes for ERV applications. These membranes consist of a dense selective water vapour permeable film layer on the surface of a micro-porous polymer substrate layer. Polymers for the dense layer must have high water vapour permeability and moderate selectivity of water vapour over other gases and vapours. A large number of polymers have been studied in the literature for water vapour permeability. In this study, we focused on a commercially available polyether-polyurethane copolymer containing a large weight fraction of polyethyleneoxide (PEO) incorporated as ‘side chains’ from the polymer backbone. These types of ‘side chain’ PEO containing polymers and their interactions with water and water 190  vapour have not been widely studied in the literature, and the specific polymer (PERMAX™230) used in the present study has not previously been analyzed and reported in the literature. The work in this dissertation presents previously unreported details on this polymer’s chemical, thermal, and transport properties. Furthermore, cross-linking reactions of this PEO-PU polymer with poly-carbodiimides were studied and liquid water and water vapour interactions with this polymer were analyzed. Cross-linking was shown to reduce swell and dissolution in liquid water up to 90°C, while not adversely affecting the uptake and transport of water vapour. Interactions of water vapour with these polymers was analyzed through sorption and permeability studies. The interaction of water with the polymers was further analyzed through DSC studies of hydrated polymer to determine the various states of water in these polymers at various levels of hydration. The solubility coefficients of the polymers were determined at a range of temperature and water vapour activity conditions. Permeability, solubility, and diffusivity of water vapour in the polymers was determined and the effects of temperature and water vapour activity on the transport properties of the polymers was reported. The Flory-Huggins interaction parameters for the polymers were reported at these conditions as well. A Zimm-Lundberg clustering analysis was completed to understand the tendency of water to cluster in the polymers elucidating the micro-phase morphology of these polymers in the presence of water vapour and providing a further understanding of how solubility and diffusivity of water vapour in the polymers was impacted by changes in water vapour activity and temperature.  8.2.2 Analysis of Composite Membranes After the polymer permeability properties of the PEO-PU polymers were determined in Chapter 3, an analysis of two types of composite membranes for water vapour separations was completed in Chapter 4. In order to study the contributions of resistance to water vapour transport in the film and substrate layers of these membranes, a resistance in series model was presented and validated. A method to determine the boundary layer resistance in air-to-air transport measurement device was also presented and validated. This boundary layer mass transport resistance was shown to be significant in air-to-air membrane based water vapour transport, and must be considered not only in membrane testing, but also in heat and mass exchanger design. Also from the resistance in series model, the substrate resistance in the composite membranes was demonstrated to be between 30 to 50% of the total transport resistance in the membranes. Although other reports in the literature have evaluated composite membranes and presented models for their performance based on the resistances 191  of each component, this study confirmed the resistances associated with each layer individually, and then validated the resistances in each component in assembled membranes. This allowed us to hypothesize that creating membrane structures which eliminated this substrate and its resistance to transport would allow for higher performance membranes. We proposed a novel type of membrane for water vapour transport applications based on a nanofibrous layer impregnated with a water vapour permeable polymer. This membrane fabrication approach allowed selective membranes to be generated without the use of a microporous substrate layer. Fabricating and optimizing these so called ‘impregnated electrospun nanofibrous membranes’ (IEMNs) was the focus of Chapters 5 and 6 of this study. However, through the analysis of the composite membranes based on two substrates in Chapter 4 of this study, some interesting results were observed, demonstrating unexpected performance properties depending on the substrate which was used to fabricate the composite membranes. The tendency of the side-chain PEO portion of the PEO-PU polymers to crystalize at ambient conditions led to composite membranes with ‘thermal switching’ permeability properties when the PEO-PU polymers were cast on microporous polypropylene substrates. Surprisingly, this thermal switching property was not present when membranes were synthesized on microporous substrates containing high weight fractions of silica. It was proposed that during film formation of the PEO-PU polymers on the two substrates, preferential micro-phase separation of the copolymer structure must occur dependent on the substrate, leading to a greater tendency towards PEO crystallization when coating the film layer on the microporous polypropylene substrates. Interestingly, treating cured membranes containing crystalline PEO with liquid water disrupted the PEO crystallinity via hydrogen bonding of water to the PEO segments of the polymer, and eliminated the thermal switching behavior of the membranes. Future work studying the interfacial phenomena at play during membrane fabrication could be of interest for membranes where polymer crystallization is an issue, or for applications where thermally controlled switching of permeability over several orders of magnitude with temperature is desirable. 8.2.3 Impregnated Electrospun Nanofibrous Membranes: Fabrication and Optimization Through optimization in Chapters 5 and 6, IENMs were demonstrated to achieve the water vapour permeance target of >10000 GPU with gas permeance of <100 GPU. The results presented in this study are the first report in the literature on electrospun nanofibrous membranes for gas selective water vapour separations and the first report of the use of 192  electrospun nanofibrous membranes for enthalpy exchanger applications. This work on nanofibrous membranes has led to a patent family which has patents granted in Canada and the United States, pending in Europe and Asia, with a continuation patent also pending in the United States. These membranes exceed the performance of current generation composite membrane materials. The approach of eliminating the microporous substrate layer through the use of an electrospun nanofibrous layer was proven to be successful. However, since a PAN polymer which is relatively impermeable to water vapour was used for the nanofiber layer, a new resistance to transport associated with the nanofiber in the membranes was identified. Decreasing the fiber diameter to less than 250nm allowed thinner defect free film layers to be fabricated. Furthermore, the fiber volume in the membranes could be decreased by using smaller fiber diameters, and less fiber areal density, allowing higher film layer permeability. A model for the permeability of the impermeable fiber filled polymer film layer was presented where the effective permeability of the film is a function of the impermeable fiber volume, which effects both the diffusivity and solubility of the fiber-filled film layer. Since the materials studied were actual membranes with a non-woven microfiber carrier layer, the carrier layer complicated a precise analysis of the permeability of fiber-filled film layer. Future work should further validate the fiber-filled film permeability model by fabricating free standing fiber-filled films (without the non-woven carrier layer used in this study) of exact fiber volume and thickness, and confirming the effect of fiber volume and diameter on the diffusivity, sorption, and permeability of these films. The use of nanofibers which were impermeable to water and water vapour in this study produced membranes in which the permeable polymer matrix was supported within the void space between the nanofibers. The use of impermeable fibers may be beneficial from a dimensional stability and longevity perspective in the presence of liquid water or high relative humidity. Nonetheless, future studies could also consider the use of water vapour permeable fibers to improve the performance of the membranes.  Also in Chapter 6, a gravure coating method was proposed to improve the fabrication of the impregnated electrospun nanofibrous membranes by allowing control of the polymer coating deposition into the nanofiber layer. This coating method should allow scale up of membrane production to roll-to-roll coating processes. Recommendations for further modifications to the coating impregnation solution were made to improve wetting of the fibrous layer by optimizing the solution surface tension and viscosity. Membrane manufacturing scale up in a roll-to-roll process is the focus of current efforts.   193  8.2.4 Formable Electrospun Nanofibrous Membranes One further benefit of the IEMNs which was recognized during the study was that by using a nanofibrous scaffold, defect-free water vapour permeable film layers could be produced on a wide variety of porous support layers. As reported in Chapter 7, this led us to identify carrier layers (non-wovens based on microfibers) which could be thermally formed into exchanger plates. These non-wovens were used as a carrier layer for assembling membranes based on electrospun nanofibers. Through measuring the thermo-mechanical properties of the membrane components and the membranes themselves, thermally formable membranes were fabricated which could be elongated up to 150% (draw ratio of 2.5 to 1) before defect formation. To our knowledge the formable membranes demonstrated in this study are the first thermally formable membranes for energy recovery ventilation exchangers to be reported or studied in the literature. Prototype exchanger plates were fabricated based on molds with a draw ratio of 1.6 to 1 and showed minimal defects after forming exchanger plates with triangular channels. These membranes have excellent potential as membranes for ‘next-generation’ high efficiency exchangers which have multi-directional heat and mass transfer in the exchanger channels. Further work has been completed to fabricate prototype formable membrane exchangers from these materials and to quantify the performance benefits of these types of exchangers. Some initial examples of these exchangers and their assembly are shown in Figure 8-1 and Figure 8-2.  Figure 8-1: A formable membrane exchanger. 194   Figure 8-2: Assembly of a formable nanofibrous membrane exchanger for testing. Initial performance results of the nanofibrous membrane based exchangers in Figure 8-3a shows that the formed nanofibrous membrane exchanger has significantly higher ‘total effectiveness’ than flat sheet exchangers of similar geometry. The total effectiveness is a measure of the combined heat and moisture transport in the exchanger. Figure 8-3b shows that the two exchangers have similar total pressure drop in the channels.   Figure 8-3: (a.) Total effectiveness performance of cross-flow exchangers based on flat sheet membranes and formable nanofibrous membranes; (b.) total pressure drop in the exchanger channels; error bars include instrument and experimental errors as the standard deviation of three test measurements. Unfortunately, these first formable membrane exchanger prototypes had leakage exceeding 10% of the rated flow due to membrane defects and challenges with assembly, so absolute conclusions cannot be drawn based on the exchanger performance. Ongoing work should 195  focus on scale up improvement of nanofibrous membranes to make continuous membrane rolls with better uniformity. Furthermore, exchanger plate forming methods must be improved to minimize defect formation in larger formed sheets. Improved edge and perimeter sealing techniques should improve the exchanger assembly and decrease leakage. Improvements in membrane design should also focus on achieving greater draw ratios which would allow for features with greater pitch (channel depth), allowing for lower pressure drop, and higher channel geometries with improved transport. Exchanger designs where square channel cross-sections could be formed (formable membranes with draw ratios of 3 to 1) would be desirable from an exchanger design perspective.   196  References [1] R. W. Baker, Membrane technology and applications. John Wiley and Sons, 2004. [2] P. T. Cardew and M. S. Le, Membrane processes: a technology guide. Royal Society of Chemistry, 1999. [3] N. N. Li, Advanced membrane technology and applications. Wiley-Interscience, 2008. [4] M. A. Aroon, A. F. Ismail, T. Matsuura, and M. M. Montazer-Rahmati, “Performance studies of mixed matrix membranes for gas separation: A review,” Sep. Purif. Technol., vol. 75, no. 3, pp. 229–242, Nov. 2010. [5] E. Klein, “Affinity membranes: a 10-year review,” J. Membr. Sci., vol. 179, no. 1–2, pp. 1–27, Nov. 2000. [6] K. P. Lee, T. C. Arnot, and D. Mattia, “A review of reverse osmosis membrane materials for desalination--Development to date and future potential,” J. Membr. Sci., vol. In Press, Corrected Proof. [7] T. Matsuura, “Progress in membrane science and technology for seawater desalination -- a review,” Desalination, vol. 134, no. 1–3, pp. 47–54, Apr. 2001. [8] P. Shao and R. Y. M. Huang, “Polymeric membrane pervaporation,” J. Membr. Sci., vol. 287, no. 2, pp. 162–179, Jan. 2007. [9] B. Smitha, S. Sridhar, and A. A. Khan, “Solid polymer electrolyte membranes for fuel cell applications--a review,” J. Membr. Sci., vol. 259, no. 1–2, pp. 10–26, Aug. 2005. [10] M. Takht Ravanchi, T. Kaghazchi, and A. Kargari, “Application of membrane separation processes in petrochemical industry: a review,” Desalination, vol. 235, no. 1–3, pp. 199–244, Jan. 2009. [11] B. D. Freeman and I. Pinnau, Polymer membranes for gas and vapor separation. ACS Publications, 1999. [12] L. M. Robeson, “Correlation of separation factor versus permeability for polymeric membranes,” J. Membr. Sci., vol. 62, no. 2, pp. 165–185, Oct. 1991. [13] S. P. Nunes and K.-V. Peinemann, Membrane technology in the chemical industry. Wiley-VCH, 2006. [14] J. G. Wijmans and R. W. Baker, “The solution-diffusion model: a review,” J. Membr. Sci., vol. 107, no. 1–2, pp. 1–21, Nov. 1995. [15] K. W. Roth, D. Westphalen, J. Kieckmann, S. D. Hamilton, and W. Goetzler, “Volume III: Energy Savings Potential,” DOE Rep. Energy Consum. Charact. Commer. Build. HVAC Syst., 2002. [16] A. Mardiana-Idayu and S. B. Riffat, “Review on heat recovery technologies for building applications,” Renew. Sustain. Energy Rev., vol. 16, no. 2, pp. 1241–1255, Feb. 2012. [17] M. Rafati Nasr, M. Fauchoux, R. W. Besant, and C. J. Simonson, “A review of frosting in air-to-air energy exchangers,” Renew. Sustain. Energy Rev., vol. 30, pp. 538–554, Feb. 2014. [18] ASHRAE, “Standard 84-Method of Testing Air-to-Air Heat/Energy Exchangers,” Am. Soc. Heat. Refrig. Air-Cond. Eng. Inc Atlanta GA, 2013. [19] N. R. C. Governement of Canada, “Energy Use Data Handbook 1990 to 2013.” [Online]. Available: http://oee.nrcan.gc.ca/corporate/statistics/neud/dpa/data_e/publications.cfm. [Accessed: 02-Feb-2015]. [20] “Energy Recovery Ventilation Systems Will Reach Nearly $3 Billion in Annual Revenue by 2020,” Navigant Research, 12-Aug-2014. . [21] B. K. Ouazia, M. C. Swinton, M. Julien, and M. Manning, “Assessment of the enthalpy performance of houses using energy recovery technology,” Trans.-Am. Soc. Heat. Refrig. AIR Cond. Eng., vol. 112, no. 1, p. 26, 2006. 197  [22] M. R. Nasr, M. Fauchoux, D. Kadylak, R. Huizing, and C. Simonson, “Frosting limit in Air-to-Air Membrane Energy Exchangers,” in Proceedings of the 10th Nordic Symposium on Building Physics, Lund, Sweden, 2014. [23] L.-Z. Zhang, Y.-Y. Wang, C.-L. Wang, and H. Xiang, “Synthesis and characterization of a PVA/LiCl blend membrane for air dehumidification,” J. Membr. Sci., vol. 308, no. 1–2, pp. 198–206, Feb. 2008. [24] Y. P. Zhou, J. Y. Wu, and R. Z. Wang, “Performance of energy recovery ventilator with various weathers and temperature set-points,” Energy Build., vol. 39, no. 12, pp. 1202–1210, Dec. 2007. [25] L.-Z. Zhang, “Heat and mass transfer in a quasi-counter flow membrane-based total heat exchanger,” Int. J. Heat Mass Transf., vol. 53, no. 23–24, pp. 5478–5486, Nov. 2010. [26] R. N. Huizing, “United States Patent: 9255744 - Coated membranes for enthalpy exchange and other applications,” 9255744, 09-Feb-2016. [27] C. Burger, B. S. Hsiao, and B. Chu, “NANOFIBROUS MATERIALS AND THEIR APPLICATIONS,” Annu. Rev. Mater. Res., vol. 36, no. 1, pp. 333–368, Aug. 2006. [28] Z.-M. Huang, Y.-Z. Zhang, M. Kotaki, and S. Ramakrishna, “A review on polymer nanofibers by electrospinning and their applications in nanocomposites,” Compos. Sci. Technol., vol. 63, no. 15, pp. 2223–2253, Nov. 2003. [29] Krishnan Jayaraman, M. Kotaki, Yanzhong Zhang, Xiumei Mo, and S. Ramakrishna, “Recent Advances in Polymer Nanofibers,” J. Nanosci. Nanotechnol., vol. 4, pp. 52–65, Jan. 2004. [30] V. Thavasi, G. Singh, and S. Ramakrishna, “Electrospun nanofibers in energy and environmental applications,” Energy Environ. Sci., vol. 1, no. 2, pp. 205–221, 2008. [31] W. E. Teo and S. Ramakrishna, “A review on electrospinning design and nanofibre assemblies,” Nanotechnology, vol. 17, no. 14, pp. R89–R106, 2006. [32] A. L. Andrady, Science and technology of polymer nanofibers. John Wiley & Sons, 2008. [33] K. Yoon, B. S. Hsiao, and B. Chu, “Functional nanofibers for environmental applications,” J. Mater. Chem., vol. 18, no. 44, pp. 5326–5334, 2008. [34] A. Szentivanyi, T. Chakradeo, H. Zernetsch, and B. Glasmacher, “Electrospun cellular microenvironments: Understanding controlled release and scaffold structure,” Adv. Drug Deliv. Rev., vol. 63, no. 4–5, pp. 209–220, Apr. 2011. [35] K. Graham, “Polymeric nanofibers in air filter applications,” presented at the Fifteenth Annual Technical Conference & Expo of the American Filtration & Separations Society, Galveston, Texas, 09-Apr-2002. [36] R. S. Barhate and S. Ramakrishna, “Nanofibrous filtering media: Filtration problems and solutions from tiny materials,” J. Membr. Sci., vol. 296, no. 1–2, pp. 1–8, Jun. 2007. [37] M. Yoshino, A. Kamiya, N. Ogawa, A. Oguri, and S. Muramatsu, “United States Patent: 3666007 - APPARATUS FOR EFFECTING CONTINUOUS AND SIMULTANEOUS TRANSFER OF HEAT AND MOISTURE BETWEEN TWO AIR STREAMS,” 3666007, 30-May-1972. [38] M. H. Sherman and A. T. Hodgson, “Formaldehyde as a basis for residential ventilation rates,” Indoor Air, vol. 14, no. 1, pp. 2–8, Feb. 2004. [39] M. K. Drost, “Air-to-air heat exchanger performance,” Energy Build., vol. 19, no. 3, pp. 215–220, 1993. [40] L. Z. Zhang and J. L. Niu, “Energy requirements for conditioning fresh air and the long-term savings with a membrane-based energy recovery ventilator in Hong Kong,” Energy, vol. 26, no. 2, pp. 119–135, Feb. 2001. [41] K. Zhong and Y. Kang, “Applicability of air-to-air heat recovery ventilators in China,” Appl. Therm. Eng., vol. 29, no. 5–6, pp. 830–840, Apr. 2009. [42] L. Z. Zhang, D. S. Zhu, X. H. Deng, and B. Hua, “Thermodynamic modeling of a novel air dehumidification system,” Energy Build., vol. 37, no. 3, pp. 279–286, Mar. 2005. 198  [43] M. Nasif, R. AL-Waked, G. Morrison, and M. Behnia, “Membrane heat exchanger in HVAC energy recovery systems, systems energy analysis,” Energy Build., vol. 42, no. 10, pp. 1833–1840, Oct. 2010. [44] J. Liu, W. Li, J. Liu, and B. Wang, “Efficiency of energy recovery ventilator with various weathers and its energy saving performance in a residential apartment,” Energy Build., vol. 42, no. 1, pp. 43–49, Jan. 2010. [45] J. L. Niu, L. Z. Zhang, and H. G. Zuo, “Energy savings potential of chilled-ceiling combined with desiccant cooling in hot and humid climates,” Energy Build., vol. 34, no. 5, pp. 487–495, Jun. 2002. [46] C.-H. Liang, L.-Z. Zhang, and L.-X. Pei, “Performance analysis of a direct expansion air dehumidification system combined with membrane-based total heat recovery,” Energy, vol. In Press, Corrected Proof. [47] I. Kotcioglu, S. Caliskan, A. Cansiz, and S. Baskaya, “Second law analysis and heat transfer in a cross-flow heat exchanger with a new winglet-type vortex generator,” Energy, vol. In Press, Corrected Proof. [48] Y. Lu, Y. Wang, L. Zhu, and Q. Wang, “Enhanced performance of heat recovery ventilator by airflow-induced film vibration (HRV performance enhanced by FIV),” Int. J. Therm. Sci., vol. In Press, Corrected Proof. [49] K. R. Kistler and E. L. Cussler, “Membrane Modules for Building Ventilation,” Chem. Eng. Res. Des., vol. 80, no. 1, pp. 53–64, Jan. 2002. [50] L.-Z. Zhang, “An analytical solution for heat mass transfer in a hollow fiber membrane based air-to-air heat mass exchanger,” J. Membr. Sci., vol. 360, no. 1–2, pp. 217–225, Sep. 2010. [51] L. Z. Zhang and Y. Jiang, “Heat and mass transfer in a membrane-based energy recovery ventilator,” J. Membr. Sci., vol. 163, no. 1, pp. 29–38, 1999. [52] E. M. Sparrow, J. P. Abraham, G. P. Martin, and J. C. Y. Tong, “An experimental investigation of a mass exchanger for transferring water vapor and inhibiting the transfer of other gases,” Int. J. Heat Mass Transf., vol. 44, no. 22, pp. 4313–4321, Nov. 2001. [53] L. Z. Zhang and J. L. Niu, “Effectiveness Correlations for Heat and Moisture Transfer Processes in an Enthalpy Exchanger With Membrane Cores,” J. Heat Transf., vol. 124, no. 5, pp. 922–929, Oct. 2002. [54] L.-Z. Zhang, “Heat and mass transfer in a cross-flow membrane-based enthalpy exchanger under naturally formed boundary conditions,” Int. J. Heat Mass Transf., vol. 50, no. 1–2, pp. 151–162, Jan. 2007. [55] J. Min and M. Su, “Performance analysis of a membrane-based enthalpy exchanger: Effects of the membrane properties on the exchanger performance,” J. Membr. Sci., vol. 348, no. 1–2, pp. 376–382, Feb. 2010. [56] L.-Z. Zhang, “Evaluation of moisture diffusivity in hydrophilic polymer membranes: A new approach,” J. Membr. Sci., vol. 269, no. 1–2, pp. 75–83, Feb. 2006. [57] L.-Z. Zhang, C.-H. Liang, and L.-X. Pei, “Heat and moisture transfer in application scale parallel-plates enthalpy exchangers with novel membrane materials,” J. Membr. Sci., vol. 325, no. 2, pp. 672–682, Dec. 2008. [58] H. Y. Hwang, H. C. Koh, J. W. Rhim, and S. Y. Nam, “Preparation of sulfonated SEBS block copolymer membranes and their permeation properties,” Desalination, vol. 233, no. 1–3, pp. 173–182, Dec. 2008. [59] J. Min, T. Hu, and X. Liu, “Evaluation of moisture diffusivities in various membranes,” J. Membr. Sci., vol. 357, no. 1–2, pp. 185–191, Jul. 2010. [60] X.-R. Zhang, L.-Z. Zhang, H.-M. Liu, and L.-X. Pei, “One-step fabrication and analysis of an asymmetric cellulose acetate membrane for heat and moisture recovery,” J. Membr. Sci., vol. 366, no. 1–2, pp. 158–165, Jan. 2011. 199  [61] S. J. Metz, W. J. C. van de Ven, J. Potreck, M. H. V. Mulder, and M. Wessling, “Transport of water vapor and inert gas mixtures through highly selective and highly permeable polymer membranes,” J. Membr. Sci., vol. 251, no. 1–2, pp. 29–41, 2005. [62] S. J. Metz, W. J. C. van de Ven, J. Potreck, M. H. V. Mulder, and M. Wessling, “Transport of water vapor and inert gas mixtures through highly selective and highly permeable polymer membranes,” J. Membr. Sci., vol. 251, no. 1–2, pp. 29–41, Apr. 2005. [63] H. Lin, S. M. Thompson, A. Serbanescu-Martin, J. G. Wijmans, K. D. Amo, K. A. Lokhandwala, and T. C. Merkel, “Dehydration of natural gas using membranes. Part I: Composite membranes,” J. Membr. Sci., vol. 413–414, pp. 70–81, Sep. 2012. [64] J. Potreck, K. Nijmeijer, T. Kosinski, and M. Wessling, “Mixed water vapor/gas transport through the rubbery polymer PEBAX® 1074,” J. Membr. Sci., vol. 338, no. 1–2, pp. 11–16, Aug. 2009. [65] L. Di Landro, M. Pegoraro, and L. Bordogma, “Interactions of polyether--polyurethanes with water vapour and water--methane separation selectivity,” J. Membr. Sci., vol. 64, no. 3, pp. 229–236, Dec. 1991. [66] P. W. Gibson, “Effect of temperature on water vapor transport through polymer membrane laminates,” Polym. Test., vol. 19, no. 6, pp. 673–691, 2000. [67] G. R. Lomax, “Breathable polyurethane membranes for textile and related industries,” J. Mater. Chem., vol. 17, no. 27, pp. 2775–2784, 2007. [68] L.-Z. Zhang, “Numerical study of heat and mass transfer in an enthalpy exchanger with a hydrophobic-hydrophilic composite membrane core,” Numer. Heat Transf. Part Appl., vol. 51, no. 7, pp. 697–714, Mar. 2007. [69] L.-Z. Zhang, “Coupled heat and mass transfer through asymmetric porous membranes with finger-like macrovoids structure,” Int. J. Heat Mass Transf., vol. 52, no. 3–4, pp. 751–759, Jan. 2009. [70] S. Koester, F. Roghmans, and M. Wessling, “Water vapor permeance: The interplay of feed and permeate activity,” J. Membr. Sci., vol. 485, pp. 69–78, Jul. 2015. [71] S. Koester, J. Lölsberg, L. Lutz, D. Marten, and M. Wessling, “On individual resistances of selective skin, porous support and diffusion boundary layer in water vapor permeation,” J. Membr. Sci., vol. 507, pp. 179–187, Jun. 2016. [72] P. W. Gibson, H. L. Schreuder-Gibson, and D. Rivin, “Electrospun fiber mats: Transport properties,” AIChE J., vol. 45, no. 1, pp. 190–195, 1999. [73] P. Gibson, H. Schreuder-Gibson, and D. Rivin, “Transport properties of porous membranes based on electrospun nanofibers,” Colloids Surf. Physicochem. Eng. Asp., vol. 187–188, pp. 469–481, Aug. 2001. [74] S. Lee and S. K. Obendorf, “Use of Electrospun Nanofiber Web for Protective Textile Materials as Barriers to Liquid Penetration,” Text. Res. J., vol. 77, no. 9, pp. 696–702, Sep. 2007. [75] S. Lee and S. Obendorf, “Transport properties of layered fabric systems based on electrospun nanofibers,” Fibers Polym., vol. 8, no. 5, pp. 501–506, Dec. 2007. [76] Y. Kang, C. Park, J. Kim, and T. Kang, “Application of electrospun polyurethane web to breathable water-proof fabrics,” Fibers Polym., vol. 8, no. 5, pp. 564–570, Dec. 2007. [77] P. Gibson, C. Kendrick, D. Rivin, L. Sicuranza, and M. Charmchi, “An Automated Water Vapor Diffusion Test Method for Fabrics, Laminates, and Films,” J. Ind. Text., vol. 24, no. 4, pp. 322–345, Apr. 1995. [78] P. Gibson and M. Charmchi, “The use of volume-averaging techniques to predict temperature transients due to water vapor sorption in hygroscopic porous polymer materials,” J. Appl. Polym. Sci., vol. 64, no. 3, pp. 493–505, 1997. [79] P. W. Gibson and M. Charmchi, “Modeling convection/diffusion processes in porous textiles with inclusion of humidity-dependent air permeability,” Int. Commun. Heat Mass Transf., vol. 24, no. 5, pp. 709–724, Sep. 1997. 200  [80] P. Gibson, D. Rivin, and C. Kendrick, “Convection/diffusion test method for porous textiles,” Int. J. Cloth. Sci. Technol., vol. 12, no. 2, pp. 96–113, 2000. [81] E. A. McCullough, M. Kwon, and H. Shim, “A comparison of standard methods for measuring water vapour permeability of fabrics,” Meas. Sci. Technol., vol. 14, no. 8, pp. 1402–1408, 2003. [82] S. Borhani, S. G. Etemad, and S. A. H. Ravandi, “Dynamic heat and moisture transfer in bulky PAN nanofiber mats,” Heat Mass Transf., Feb. 2011. [83] H. Zhuo, J. Hu, and S. Chen, “Study of Water Vapor Permeability of Shape Memory Polyurethane Nanofibrous Nonwoven,” Text. Res. J. [84] C. Feng, K. C. Khulbe, and T. Matsuura, “Recent progress in the preparation, characterization, and applications of nanofibers and nanofiber membranes via electrospinning/interfacial polymerization,” J. Appl. Polym. Sci., vol. 115, no. 2, pp. 756–776, 2010. [85] X. Wang, X. Chen, K. Yoon, D. Fang, B. S. Hsiao, and B. Chu, “High Flux Filtration Medium Based on Nanofibrous Substrate with Hydrophilic Nanocomposite Coating,” Environ. Sci. Technol., vol. 39, no. 19, pp. 7684–7691, Oct. 2005. [86] K. Yoon, K. Kim, X. Wang, D. Fang, B. S. Hsiao, and B. Chu, “High flux ultrafiltration membranes based on electrospun nanofibrous PAN scaffolds and chitosan coating,” Polymer, vol. 47, no. 7, pp. 2434–2441, Mar. 2006. [87] K. Yoon, B. S. Hsiao, and B. Chu, “High flux nanofiltration membranes based on interfacially polymerized polyamide barrier layer on polyacrylonitrile nanofibrous scaffolds,” J. Membr. Sci., vol. 326, no. 2, pp. 484–492, Jan. 2009. [88] J. Choi, K. M. Lee, R. Wycisk, P. N. Pintauro, and P. T. Mather, “Sulfonated Polysulfone/POSS Nanofiber Composite Membranes for PEM Fuel Cells,” J. Electrochem. Soc., vol. 157, no. 6, pp. B914–B919, Jun. 2010. [89] H.-L. Lin, S.-H. Wang, C.-K. Chiu, T. L. Yu, L.-C. Chen, C.-C. Huang, T.-H. Cheng, and J.-M. Lin, “Preparation of Nafion/poly(vinyl alcohol) electro-spun fiber composite membranes for direct methanol fuel cells,” J. Membr. Sci., vol. 365, no. 1–2, pp. 114–122, Dec. 2010. [90] P. Gibson, “Governing Equations for Multiphase Heat and Mass Transfer in Hygroscopic Porous Media with Applications to Clothing Materials.” 2001. [91] J. L. Niu and L. Z. Zhang, “Membrane-based Enthalpy Exchanger: material considerations and clarification of moisture resistance,” J. Membr. Sci., vol. 189, no. 2, pp. 179–191, Aug. 2001. [92] S. J. Metz, W. J. C. van de Ven, M. H. V. Mulder, and M. Wessling, “Mixed gas water vapor/N2 transport in poly(ethylene oxide) poly(butylene terephthalate) block copolymers,” J. Membr. Sci., vol. 266, no. 1–2, pp. 51–61, Dec. 2005. [93] S. J. Metz, “Water vapor and gas transport through polymeric membranes,” Universiteit Twente (The Netherlands), Netherlands, 2003. [94] K. Yoon, B. S. Hsiao, and B. Chu, “High flux ultrafiltration nanofibrous membranes based on polyacrylonitrile electrospun scaffolds and crosslinked polyvinyl alcohol coating,” J. Membr. Sci., vol. 338, no. 1–2, pp. 145–152, Aug. 2009. [95] L. Yung, H. Ma, X. Wang, K. Yoon, R. Wang, B. S. Hsiao, and B. Chu, “Fabrication of thin-film nanofibrous composite membranes by interfacial polymerization using ionic liquids as additives,” J. Membr. Sci., vol. 365, no. 1–2, pp. 52–58, Dec. 2010. [96] U. Beuscher and C. H. Gooding, “The influence of the porous support layer of composite membranes on the separation of binary gas mixtures,” J. Membr. Sci., vol. 152, no. 1, pp. 99–116, Jan. 1999. [97] L. Liu, N. Jiang, C. M. Burns, A. Chakma, and X. Feng, “Substrate resistance in composite membranes for organic vapour/gas separations,” J. Membr. Sci., vol. 338, no. 1–2, pp. 153–160, Aug. 2009. [98] R. D. Noble, “Perspectives on mixed matrix membranes,” J. Membr. Sci., vol. 378, no. 1–2, pp. 393–397, Aug. 2011. 201  [99] L. L. Marsh, R. Lasky, D. P. Seraphim, and G. S. Springer, “Moisture solubility and diffusion in epoxy and epoxy-glass composites,” IBM J. Res. Dev., vol. 28, no. 6, pp. 655–661, Nov. 1984. [100] C.-H. Shen and G. S. Springer, “Moisture Absorption and Desorption of Composite Materials,” J. Compos. Mater., vol. 10, no. 1, pp. 2–20, Jan. 1976. [101] R. Huizing, “Design and membrane selection for gas to gas humidifiers for fuel cell applications,” University of Waterloo, Canada, 2007. [102] D. Kadylak and W. Mérida, “Experimental verification of a membrane humidifier model based on the effectiveness method,” J. Power Sources, vol. 195, no. 10, pp. 3166–3175, May 2010. [103] ASTM Standard F2298, “Standard Test Methods for Water Vapor Diffusion Resistance and Air Flow Resistance of Clothing Materials Using the Dynamic Moisture Permeation Cell,” ASTM International, 2009. [104] R. W. Baker, “Appendix,” in Membrane Technology and Applications, John Wiley & Sons, Ltd, 2012, pp. 559–570. [105] L. Johnson and J. Samms, “Thermoplastic Polyurethane Technologies for the Textile Industry,” J. Ind. Text., vol. 27, no. 1, pp. 48–62, Jul. 1997. [106] S. A. Madbouly and J. U. Otaigbe, “Recent advances in synthesis, characterization and rheological properties of polyurethanes and POSS/polyurethane nanocomposites dispersions and films.,” Prog. Polym. Sci., vol. 34, no. Copyright (C) 2015 American Chemical Society (ACS). All Rights Reserved., pp. 1283–1332, 2009. [107] R. Huizing, W. Merida, and F. Ko, “Impregnated electrospun nanofibrous membranes for water vapour transport applications.,” J. Membr. Sci., vol. 461, no. Copyright (C) 2015 American Chemical Society (ACS). All Rights Reserved., pp. 146–160, 2014. [108] G. K. Kannarpady, B. Mohan, and A. Bhattacharyya, “Nonlinear surface electrical resistivity of graphite-polyurethane composite films.,” J. Appl. Polym. Sci., vol. 106, no. Copyright (C) 2015 American Chemical Society (ACS). All Rights Reserved., pp. 293–298, 2007. [109] A. Lubnin, G. Anderle, G. Snow, R. Varn, and S. Lenhard, “Novel ‘breathable’ polyurethane dispersions.,” Proc. Int. Waterborne High-Solids Powder Coat. Symp., vol. 31st, no. Copyright (C) 2015 American Chemical Society (ACS). All Rights Reserved., pp. 299–316, 2004. [110] A. V. Lubnin, G. E. Snow, R. D. Varn, and G. A. Anderle, “Breathable polyurethanes, blends, and articles,” US6897281 B2, 24-May-2005. [111] G. A. Anderle, A. V. Lubnin, G. E. Snow, and R. D. Varn, “Vinyl chloride, acrylate, and urethane polymers with increased moisture vapor permeability and static dissipative properties.,” 28-May-2009. [112] H. Sijbesma, K. Nymeijer, R. van Marwijk, R. Heijboer, J. Potreck, and M. Wessling, “Flue gas dehydration using polymer membranes,” J. Membr. Sci., vol. 313, no. 1–2, pp. 263–276, Apr. 2008. [113] J. Potreck, K. Nijmeijer, T. Kosinski, and M. Wessling, “Mixed water vapor/gas transport through the rubbery polymer PEBAX® 1074,” J. Membr. Sci., vol. 338, no. 1–2, pp. 11–16, Aug. 2009. [114] S. J. Metz, W. J. C. van de Ven, M. H. V. Mulder, and M. Wessling, “Mixed gas water vapor/N2 transport in poly(ethylene oxide) poly(butylene terephthalate) block copolymers,” J. Membr. Sci., vol. 266, no. 1–2, pp. 51–61, Dec. 2005. [115] M. S. Yen and K. L. Cheng, “The effects of soft segments on the physical properties and water vapor permeability of H12MDI-PU cast films,” J. Appl. Polym. Sci., vol. 52, no. 12, pp. 1707–1717, 1994. [116] L. Di Landro, M. Pegoraro, and L. Bordogma, “Interactions of polyether—polyurethanes with water vapour and water—methane separation selectivity,” J. Membr. Sci., vol. 64, no. 3, pp. 229–236, Dec. 1991. [117] ASTM Standard E96, “Standard Test Methods for Water Vapor Transmission of Materials,” ASTM International, 2015. 202  [118] R. G. Coogan, “Post-crosslinking of waterborne urethanes.,” Prog. Org. Coat., vol. 32, no. Copyright (C) 2015 American Chemical Society (ACS). All Rights Reserved., pp. 51–63, 1997. [119] L. C. J. Hesselmans, A. J. Derksen, and J. A. M. van den Goorbergh, “Polycarbodiimide crosslinkers,” Prog. Org. Coat., vol. 55, no. 2, pp. 142–148, Feb. 2006. [120] L. C. J. Hesselmans and J. A. M. van den Goorbergh, “Containing carbodiimide functions; for polyurethanes, poly(meth)acrylates,” US5258481 A, 02-Nov-1993. [121] L. C. J. Hesselmans, A. J. Derksen, and J. C. Munneke, “Process for preparation of stable polycarbodiimide dispersions in water, which are free of organic solvents and may be used as crosslinking agent,” US7439316 B2, 21-Oct-2008. [122] W. Posthumus, A. J. Derksen, J. A. M. van den Goorbergh, and L. C. J. Hesselmans, “Crosslinking by polycarbodiimides,” Prog. Org. Coat., vol. 58, no. 2–3, pp. 231–236, Feb. 2007. [123] M. G. Lu, J. Y. Lee, M. J. Shim, and S. W. Kim, “Thermal degradation of film cast from aqueous polyurethane dispersions,” J. Appl. Polym. Sci., vol. 85, no. 12, pp. 2552–2558, Sep. 2002. [124] Y. Aray, M. Marquez, J. Rodríguez, D. Vega, Y. Simón-Manso, S. Coll, C. Gonzalez, and D. A. Weitz, “Electrostatics for Exploring the Nature of the Hydrogen Bonding in Polyethylene Oxide Hydration,” J. Phys. Chem. B, vol. 108, no. 7, pp. 2418–2424, Feb. 2004. [125] H. Hatakeyama and T. Hatakeyama, “Interaction between water and hydrophilic polymers,” Thermochim. Acta, vol. 308, no. 1–2, pp. 3–22, Jan. 1998. [126] N. S. Schneider, J. L. Illinger, and F. E. Karasz, “The interaction of water with polyurethanes containing block copolymer soft segments,” J. Appl. Polym. Sci., vol. 47, no. 8, pp. 1419–1425, Feb. 1993. [127] D. Husken and R. J. Gaymans, “The Structure of Water in PEO-Based Segmented Block Copolymers and its Effect on Transition Temperatures,” Macromol. Chem. Phys., vol. 209, no. 9, pp. 967–979, May 2008. [128] A. Cova, A. J. Sandoval, V. Balsamo, and A. J. Müller, “The effect of hydrophobic modifications on the adsorption isotherms of cassava starch,” Carbohydr. Polym., vol. 81, no. 3, pp. 660–667, Jul. 2010. [129] Y. Yampolskii, I. Pinnau, and B. D. Freeman, Materials science of membranes for gas and vapor separation. John Wiley and Sons, 2006. [130] G. S. Park, “Transport Principles—Solution, Diffusion and Permeation in Polymer Membranes,” in Synthetic Membranes: Science, Engineering and Applications, P. M. Bungay, H. K. Lonsdale, and M. N. de Pinho, Eds. Springer Netherlands, 1986, pp. 57–107. [131] S. J. Metz, N. F. A. van der Vegt, M. H. V. Mulder, and M. Wessling, “Thermodynamics of Water Vapor Sorption in Poly(ethylene oxide) Poly(butylene terephthalate) Block Copolymers,” J. Phys. Chem. B, vol. 107, no. 49, pp. 13629–13635, Dec. 2003. [132] J. A. Emerson, D. T. W. Toolan, J. R. Howse, E. M. Furst, and T. H. Epps, “Determination of Solvent–Polymer and Polymer–Polymer Flory–Huggins Interaction Parameters for Poly(3-hexylthiophene) via Solvent Vapor Swelling,” Macromolecules, vol. 46, no. 16, pp. 6533–6540, Aug. 2013. [133] K. Bergfeldt, L. Piculell, and P. Linse, “Segregation and Association in Mixed Polymer Solutions from Flory−Huggins Model Calculations,” J. Phys. Chem., vol. 100, no. 9, pp. 3680–3687, Jan. 1996. [134] B. H. Zimm and J. L. Lundberg, “Sorption of Vapors by High Polymers,” J. Phys. Chem., vol. 60, no. 4, pp. 425–428, Apr. 1956. [135] N. Dolmaire, E. Espuche, F. Méchin, and J.-P. Pascault, “Water transport properties of thermoplastic polyurethane films,” J. Polym. Sci. Part B Polym. Phys., vol. 42, no. 3, pp. 473–492, Feb. 2004. 203  [136] P. T. Tsilingiris, “Thermophysical and transport properties of humid air at temperature range between 0 and 100 °C,” Energy Convers. Manag., vol. 49, no. 5, pp. 1098–1110, May 2008. [137] P. Arora and Z. (John) Zhang, “Battery Separators,” Chem. Rev., vol. 104, no. 10, pp. 4419–4462, Oct. 2004. [138] H. Meng and G. Li, “A review of stimuli-responsive shape memory polymer composites,” Polymer, vol. 54, no. 9, pp. 2199–2221, Apr. 2013. [139] Q. Meng and J. Hu, “A poly(ethylene glycol)-based smart phase change material,” Sol. Energy Mater. Sol. Cells, vol. 92, no. 10, pp. 1260–1268, Oct. 2008. [140] H. Zhou, Y. Chen, H. Fan, H. Shi, Z. Luo, and B. Shi, “Water vapor permeability of the polyurethane/TiO2 nanohybrid membrane with temperature sensitivity,” J. Appl. Polym. Sci., vol. 109, no. 5, pp. 3002–3007, Sep. 2008. [141] Y. Chen, Y. Liu, H. Fan, H. Li, B. Shi, H. Zhou, and B. Peng, “The polyurethane membranes with temperature sensitivity for water vapor permeation,” J. Membr. Sci., vol. 287, no. 2, pp. 192–197, Jan. 2007. [142] H. Zhou, J. Zeng, H. Fan, Y. Liu, and J. Zhou, “Thermal sensitive polyurethane membranes with desirable switch temperatures,” Macromol. Res., vol. 18, no. 11, pp. 1053–1059, Nov. 2010. [143] J. D. Menczel and R. B. Prime, Thermal Analysis of Polymers: Fundamentals and Applications. John Wiley & Sons, 2014. [144] S. R. Reijerkerk, R. Jordana, K. Nijmeijer, and M. Wessling, “Highly hydrophilic, rubbery membranes for CO2 capture and dehydration of flue gas,” Int. J. Greenh. Gas Control, vol. 5, no. 1, pp. 26–36, Jan. 2011. [145] J. T. Koberstein and T. P. Russell, “Simultaneous SAXS-DSC study of multiple endothermic behavior in polyether-based polyurethane block copolymers,” Macromolecules, vol. 19, no. 3, pp. 714–720, 1986. [146] J. T. Koberstein and A. F. Galambos, “Multiple melting in segmented polyurethane block copolymers,” Macromolecules, vol. 25, no. 21, pp. 5618–5624, Oct. 1992. [147] L. M. Leung and J. T. Koberstein, “DSC annealing study of microphase separation and multiple endothermic behavior in polyether-based polyurethane block copolymers,” Macromolecules, vol. 19, no. 3, pp. 706–713, 1986. [148] D. J. Martin, G. F. Meijs, G. M. Renwick, S. J. Mccarthy, and P. A. Gunatillake, “The effect of average soft segment length on morphology and properties of a series of polyurethane elastomers. I. Characterization of the series,” J. Appl. Polym. Sci., vol. 62, no. 9, pp. 1377–1386, Nov. 1996. [149] D. J. Martin, G. F. Meijs, P. A. Gunatillake, S. J. McCarthy, and G. M. Renwick, “The effect of average soft segment length on morphology and properties of a series of polyurethane elastomers. II. SAXS-DSC annealing study,” J. Appl. Polym. Sci., vol. 64, no. 4, pp. 803–817, Apr. 1997. [150] Y. Yanagihara, N. Osaka, S. Murayama, and H. Saito, “Thermal annealing behavior and structure development of crystalline hard segment domain in a melt-quenched thermoplastic polyurethane,” Polymer, vol. 54, no. 8, pp. 2183–2189, Apr. 2013. [151] M.-S. Yen, H.-C. Tsai, and P.-D. Hong, “Effect of soft segment composition on the physical properties of nonionic aqueous polyurethane containing side chain PEGME,” J. Appl. Polym. Sci., vol. 105, no. 3, pp. 1391–1399, Aug. 2007. [152] Li, Xuan, Wang, Yingkang, and Yang, Yang, “STUDIES OF THE CRYSTALLIZATION BEHAVIOR IN THE CRYSTALLINE/AMORPHOUS POLYMER BLENDS: POLY (ETHYLENE OXIDE)/POLY (METHYL METHACRYLATE) AND POLY (ETHYLENE OXIDE)/POLY (VINYL ACETATE),” Chin. J. Polym. Sci., vol. 3, p. 11, 1985. [153] H. Schönherr and C. W. Frank, “Ultrathin Films of Poly(ethylene oxides) on Oxidized Silicon. 1. Spectroscopic Characterization of Film Structure and Crystallization Kinetics,” Macromolecules, vol. 36, no. 4, pp. 1188–1198, Feb. 2003. 204  [154] Q. Guo and T. Wang, “Influence of SiO2 pore structure on phase change enthalpy of shape-stabilized polyethylene glycol/silica composites,” J. Mater. Sci., vol. 48, no. 10, pp. 3716–3721, Jan. 2013. [155] S. Jiang, D. Yu, X. Ji, L. An, and B. Jiang, “Confined crystallization behavior of PEO in silica networks,” Polymer, vol. 41, no. 6, pp. 2041–2046, Mar. 2000. [156] R. Huizing, W. Mérida, and F. Ko, “Impregnated electrospun nanofibrous membranes for water vapour transport applications,” J. Membr. Sci., vol. 461, pp. 146–160, Jul. 2014. [157] S. Ramakrishna, An Introduction to Electrospinning And Nanofibers. World Scientific, 2005. [158] Y. Gogotsi, Nanotubes and nanofibers. CRC Press, 2006. [159] P. Gibson, H. Schreuder-Gibson, and D. Rivin, “Transport properties of porous membranes based on electrospun nanofibers,” Colloids Surf. Physicochem. Eng. Asp., vol. 187–188, pp. 469–481, Aug. 2001. [160] T. Grafe and K. Graham, “Polymeric nanofibers and nanofiber webs: a new class of nonwovens,” Nonwoven Technol Rev, pp. 51–5, 2003. [161] X. Wang, X. Chen, K. Yoon, D. Fang, B. S. Hsiao, and B. Chu, “High Flux Filtration Medium Based on Nanofibrous Substrate with Hydrophilic Nanocomposite Coating,” Env. Sci Technol, vol. 39, no. 19, pp. 7684–7691, 2005. [162] N.-N. Bui, M. L. Lind, E. M. V. Hoek, and J. R. McCutcheon, “Electrospun nanofiber supported thin film composite membranes for engineered osmosis,” J. Membr. Sci., vol. 385–386, pp. 10–19, Dec. 2011. [163] L. A. Hoover, J. D. Schiffman, and M. Elimelech, “Nanofibers in thin-film composite membrane support layers: Enabling expanded application of forward and pressure retarded osmosis,” Desalination, vol. 308, pp. 73–81, Jan. 2013. [164] R. Huizing, W. Merida, and F. Ko, “Selective Nano-fibrous Membranes for Water Vapour Transport Applications,” presented at the International Congress on Membranes (ICOM 2011), ICOM 2011, Amsterdam, NL, 26-Jul-2011. [165] R. Huizing, W. Merida, and F. Ko, “Selective Electrospun Nanofibrous Membranes for Water Vapour Transport Applications.,” presented at the North American Membrane Society Meeting (NAMS 2012), NAMS 2012, New Orleans, USA, 13-Jun-2012. [166] V. I. Bondar, B. D. Freeman, and I. Pinnau, “Gas transport properties of poly(ether‐b‐amide) segmented block copolymers,” J. Polym. Sci. Part B Polym. Phys., vol. 38, no. 15, pp. 2051–2062, Aug. 2000. [167] S. M. Allen, M. Fujii, V. Stannett, H. B. Hopfenberg, and J. L. Williams, “The barrier properties of polyacrylonitrile,” J. Membr. Sci., vol. 2, pp. 153–163, 1977. [168] R. N. Huizing, CA, F. K. Ko, and CA, “United States Patent: 8936668 - Selective water vapour transport membranes comprising a nanofibrous layer and methods for making the same,” 8936668, 20-Jan-2015. [169] K. A. Hong, H. S. Yoo, and E. Kim, “Effect of waterborne polyurethane coating on the durability and breathable waterproofing of electrospun nanofiber web-laminated fabrics,” Text. Res. J., p. 40517514542141, Jul. 2014. [170] S. Mollá and V. Compañ, “Polyvinyl alcohol nanofiber reinforced Nafion membranes for fuel cell applications,” J. Membr. Sci., vol. 372, no. 1–2, pp. 191–200, Apr. 2011. [171] S.-H. Wang and H.-L. Lin, “Nafion/ poly(vinyl alcohol) nano-fiber composite and Nafion/ poly(vinyl alcohol) blend membranes for direct methanol fuel cells,” J. Membr. Sci. [172] J. Choi, K. M. Lee, R. Wycisk, P. N. Pintauro, and P. T. Mather, “Nanofiber network ion-exchange membranes,” Macromolecules, vol. 41, no. 13, pp. 4569–4572, 2008. [173] J. Choi, R. Wycisk, W. Zhang, P. N. Pintauro, K. M. Lee, and P. T. Mather, “High Conductivity Perfluorosulfonic Acid Nanofiber Composite Fuel‐Cell Membranes,” ChemSusChem, vol. 3, no. 11, pp. 1245–1248, Nov. 2010. 205  [174] Z. Zhao, J. Zheng, M. Wang, H. Zhang, and C. C. Han, “High performance ultrafiltration membrane based on modified chitosan coating and electrospun nanofibrous PVDF scaffolds,” J. Membr. Sci., vol. 394–395, pp. 209–217, Mar. 2012. [175] S. Kaur, R. Barhate, S. Sundarrajan, T. Matsuura, and S. Ramakrishna, “Hot pressing of electrospun membrane composite and its influence on separation performance on thin film composite nanofiltration membrane,” Desalination, vol. 279, no. 1–3, pp. 201–209, Sep. 2011. [176] H. Zhang, J. Zheng, Z. Zhao, and C. C. Han, “Role of wettability in interfacial polymerization based on PVDF electrospun nanofibrous scaffolds,” J. Membr. Sci., vol. 442, pp. 124–130, Sep. 2013. [177] H. You, Y. Yang, X. Li, K. Zhang, X. Wang, M. Zhu, and B. S. Hsiao, “Low pressure high flux thin film nanofibrous composite membranes prepared by electrospraying technique combined with solution treatment,” J. Membr. Sci., vol. 394–395, pp. 241–247, Mar. 2012. [178] X. Wang, K. Zhang, Y. Yang, L. Wang, Z. Zhou, M. Zhu, B. S. Hsiao, and B. Chu, “Development of hydrophilic barrier layer on nanofibrous substrate as composite membrane via a facile route,” J. Membr. Sci., vol. 356, no. 1–2, pp. 110–116, Jul. 2010. [179] D. M. Yu, S. Yoon, T.-H. Kim, J. Y. Lee, J. Lee, and Y. T. Hong, “Properties of sulfonated poly(arylene ether sulfone)/electrospun nonwoven polyacrylonitrile composite membrane for proton exchange membrane fuel cells,” J. Membr. Sci., vol. 446, pp. 212–219, Nov. 2013. [180] H. Ma, B. S. Hsiao, and B. Chu, “Thin-film nanofibrous composite membranes containing cellulose or chitin barrier layers fabricated by ionic liquids,” Polymer, vol. 52, no. 12, pp. 2594–2599, May 2011. [181] X. Wang, D. Fang, B. S. Hsiao, and B. Chu, “Nanofiltration membranes based on thin-film nanofibrous composites,” J. Membr. Sci., vol. 469, pp. 188–197, Nov. 2014. [182] W.-S. Hung, C.-L. Lai, Q. An, M. De Guzman, T.-J. Shen, Y.-H. Huang, K.-C. Chang, C.-H. Tsou, C.-C. Hu, and K.-R. Lee, “A study on high-performance composite membranes comprising heterogeneous polyamide layers on an electrospun substrate for ethanol dehydration,” J. Membr. Sci., vol. 470, pp. 513–523, Nov. 2014. [183] K. C. Krogman, J. L. Lowery, N. S. Zacharia, G. C. Rutledge, and P. T. Hammond, “Spraying asymmetry into functional membranes layer-by-layer,” Nat. Mater., vol. 8, no. 6, pp. 512–518, 2009. [184] R. Huizing, H. Chen, and F. Wong, “Contaminant transport in membrane based energy recovery ventilators,” Sci. Technol. Built Environ., vol. 21, no. 1, pp. 54–66, Jan. 2015. [185] N. Lifshutz, “On the ‘Mean Flow’ Pore Size Distribution of Microfiber and Nanofiber Webs,” Int. Nonwovens J., vol. 14, pp. 18–24, 2005. [186] L. Y. Wan, H. Wang, W. Gao, and F. Ko, “An analysis of the tensile properties of nanofiber mats,” Polymer, vol. 73, pp. 62–67, Sep. 2015. [187] W. W. Sampson, “A multiplanar model for the pore radius distribution in isotropic near-planar stochastic fibre networks,” J. Mater. Sci., vol. 38, no. 8, pp. 1617–1622. [188] L.-H. Lee, “Relationships between surface wettability and glass temperatures of high polymers,” J. Appl. Polym. Sci., vol. 12, no. 4, pp. 719–730, Apr. 1968. [189] “PAUL Wärmerückgewinnung GmbH.” [Online]. Available: http://paul-lueftung.de/. [Accessed: 01-Jan-2016]. [190] J. F. Dean, D. E. Kadylak, R. N. Huizing, J. B. Balanko, and C. W. Mullen, “Counter-flow energy recovery ventilator (erv) core,” US20140326432 A1, 06-Nov-2014. [191] R. K. Okine, “Analysis of Forming Parts from Advanced Thermoplastic Composite Sheet Materials,” J. Thermoplast. Compos. Mater., vol. 2, no. 1, pp. 50–76, Jan. 1989. [192] J. Z. Yu, Z. Cai, and F. K. Ko, “Formability of textile preforms for composite applications. Part 1: Characterization experiments,” Compos. Manuf., vol. 5, no. 2, pp. 113–122, Jun. 1994. [193] Z. Cai, J. Z. Yu, and F. K. Ko, “Formability of textile preforms for composite applications. Part 2: Evaluation experiments and modelling,” Compos. Manuf., vol. 5, no. 2, pp. 123–132, Jun. 1994. 206  [194] D. Bloch, Polymer Handbook, 4th Edition, 4th ed. Wiley-Interscience, 1999. [195] L. Ji and X. Zhang, “Ultrafine polyacrylonitrile/silica composite fibers via electrospinning,” Mater. Lett., vol. 62, no. 14, pp. 2161–2164, May 2008. [196] P. P. Lizymol and S. Thomas, “Thermal behaviour of polymer blends: a comparison of the thermal properties of miscible and immiscible systems,” Polym. Degrad. Stab., vol. 41, no. 1, pp. 59–64, Jan. 1993. [197] P. Basak and S. V. Manorama, “PEO-PU/PAN semi-interpenetrating polymer networks for SPEs: influence of physical properties on the electrical characteristics,” Solid State Ion., vol. 167, no. 1–2, pp. 113–121, Feb. 2004.     

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0344019/manifest

Comment

Related Items