Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effect of non-solvent on viscoelastic and microstructural properties of cellulose acetate in a ternary… Korehei, Reza 2007

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

Item Metadata

Download

Media
831-ubc_2007-0460.pdf [ 14.24MB ]
Metadata
JSON: 831-1.0074946.json
JSON-LD: 831-1.0074946-ld.json
RDF/XML (Pretty): 831-1.0074946-rdf.xml
RDF/JSON: 831-1.0074946-rdf.json
Turtle: 831-1.0074946-turtle.txt
N-Triples: 831-1.0074946-rdf-ntriples.txt
Original Record: 831-1.0074946-source.json
Full Text
831-1.0074946-fulltext.txt
Citation
831-1.0074946.ris

Full Text

EFFECT OF NON-SOLVENT ON VISCOELASTIC A N D MICROSTRUCTURAL PROPERTIES OF C E L L U L O S E A C E T A T E IN A TERNARY SYSTEM by REZA KOREHEI B.Sc, The University of Tehran, 1998 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA September 2007 © Reza Korehei, 2007 Abstract In this study, the effects of component composition on rheological and microstructural behaviour of a ternary of cellulose acetate (CA), A^ A'-dimethylacetamide (DMAc) and non-solvent (1-Propanol, 2-Propanol, 1-Hexanol, 1-Octanol, 1-Decanol, 1,2-Ethanediol, 1,2-Propanediol, 1,3-Propanediol, 1,4-Butanediol, 1,6-Hexanediol, Glycerol) system was examined. In this ternary system, physical gel formation can arise as a result of phase separation, which is characterized by the observation of a gradual to extreme cloudiness in the system. Depending on the non-solvent concentration, structure and polymer concentration, phase separation leads to CA aggregation and the formation of large macromolecular assemblies. Sol-gel transition is observed at a critical non-solvent concentration, which is dependent on the non-solvent structure and CA concentration. Increasing CA and non-solvent concentration resulted in enhanced steady shear viscosity and dynamic viscoelastic properties. Enhanced dynamic viscoelastic property and gelation are due to the intensification of intermolecular hydrogen bonding and hydrophobic interactions. Increasing the available hydrogen-bonding groups within the non-solvent leads to the formation of gels with larger elastic and viscous modulus (G' and G"), and a lower concentration sol-gel transition. Likewise, increasing the hydrophobic component of the non-solvent also enhanced the gel properties and accelerated the sol-gel transition. Although hydrophobic interactions play a role in the gelation process, it appears that gel properties are greatly influenced by competitive hydrogen bonding between system components. Competitive hydrogen bonding interactions between components in the various stages of the phase separation and gel formation was used to explain the weak and strong polymer-network structures observed by rheology. ii Through the use of Fourier transfer infrared (FTIR) spectroscopy the effect of hydrogen bonding between CA, DMAc and non-solvent were probed. Shifting the hydroxyl (OH) band to a lower wavenumber in the FTIR spectra suggests the intensification of the intermolecular hydrogen bonds in the ternary system. This shift accounts for the phase separation, and development of microstructure in the sample. Increasing the non-solvent content shifts the yield strain of the gels to a lower strain value, suggesting that they are made of floes. These floes consist of aggregated macromolecules with strong-links, and the links between floes are stronger than the links within the floes. The power-law dependence of elastic modulus (G1), together with similar values of fractal dimension for gels observed through confocal microscopy, suggests that the gels are fractal in nature and that they are made through an aggregation mechanism. Scanning electron microscopy (SEM) and laser scanning confocal microscopy (LSCM) revealed differences in the gel microstructure, depending on the constituent composition. Microscopic images showed better uniform packing in the polymer network structure as the CA concentration increases. The LSCM images (fluorescence and reflective mode) confirm the rheological results, and show different texture and aggregated structure for the gel as the structure of non-solvents are varied in the ternary system. iii Acknowledgements There are numerous people who I would like to thank for the help they provided during the completion of this thesis. Firstly, I would like to express my deepest appreciation to my supervisor Dr. John Kadla for his support, guidance and the immense knowledge he has imparted on me in the preparation of this study. I would also like to thank my committee, Dr. Shawn Mansfield and Dr. Phil Evans for their constant encouragement, advice and constructive criticism. My heartfelt thanks to Dr. Qizhou Dai, for his assistance and guidance throughout my thesis, the members of Paprican Vancouver Laboratory, Mr. Andrew Goodison and Mrs. Val Lawrence for their great help and expertise in Confocal Microscopy, and all my colleagues in Wood Science Department and Biomaterial Laboratory for their contributions wonderful discussions and above all their friendship. Finally, I would like to express my gratitude to my mother and my sister for their unconditional love, patience, sacrifice and support they have given me. Lastly, this thesis is dedicated to the memory of my late father and aunt who passed on while I was studying abroad. iv T a b l e o f C o n t e n t A b s t r a c t i i A c k n o w l e d g e m e n t s i v L i s t o f f igures v i i L i s t o f T a b l e s x i C h a p t e r 1. I n t r o d u c t i o n 1 1.1 Cellulose 1 1.2 Cellulose Modification 7 1.3 Cellulose Acetate and its Properties 10 1.4 Cellulose Acetate - Applications and Emerging Uses 16 1.5 Cellulose Acetate - Gelation Mechanism 18 1.6 Rheology and the Phase Behaviour of Polymers 19 1.6.1 Liquid-solid Transition and the Solidification Process 19 1.6.2 Rheological Characterization of Polymeric Solutions and Gels 23 1.7 Goal of the Project 28 1.9 References 31 C h a p t e r 2. E f f e c t o f D i h y d r i c A l c o h o l S t r u c t u r e o n the V i s c o e l a s t i c P r o p e r t i e s a n d M i c r o s t r u c t u r e o f C e l l u l o s e A c e t a t e S o l u t i o n s 37 2.1 Abstract 37 2.2 Introduction 37 2.3 Experimental Materials and Methods 39 2.3.1 Materials 39 2.3.2 Sample Preparation 40 2.3.3 Rheological Measurements 40 2.3.4 FTIR Spectroscopy 40 2.3.5 Cloud Point 40 2.3.6 Scanning Electron Microscopy (SEM) 41 2.3.7 Laser Scanning Confocal Microscopy (LSCM) 41 2.4 Results and Discussion 41 2.4.1 Solution Behaviour 41 2.4.2. Phase Separation and the Sol-gel Transition 44 2.3.3. Network Formation and the Gelation Process 50 2.3.4 Gel Rheology and Microstructure 63 2.4 Conclusion 73 2.5 References 76 v Chapter 3. Effect of Hydrophobic and Hydrophilic Interactions on the Rheological Behaviour and Microstructure of a Ternary Cellulose Acetate System 79 3.1 Abstract 79 3.2 Introduction 80 3.3 Experimental Materials and Methods 84 3.3.1 Materials 84 3.3.2 Sample Preparation 84 3.3.3 Analysis 84 3.4 Results and Discussion 84 3.4.1 Solution Viscosity 84 3.4.2 Hydrophilic and Hydrophobic Interactions and the Sol-Gel Transition 93 3.4.3 Non-solvent Structure and Effect on Bonding Mechanism 100 3.4.4 Rheology at the Gel Point 105 3.4.5 Rheological Characterization of Network Structure 110 3.4.6 Microscopic Analysis of Gel Microstructure 117 3.5 Conclusion 120 3.6 References 123 Chapter 5. Conclusion 127 5.1 Recommendation 132 Appendix 134 Part A: Figures and Tables 135 Part B: Experimental Materials and Methods 168 vi List of figures Figure 1.1: Constitutional formula of the cellulose macromolecule illustrating cellobiose as the repeat unit. The P-D-anhydroglucopyranose units are in the 4C| chair conformation with all hydroxyl substituents orientated equatorially (Krassig 1990) 2 Figure 1.2: Intra- and intermolecular hydrogen bonding in cellulose I: intramolecular hydrogen bonding between O3-H—CV and O2-H—C>6, and intermolecular hydrogen bonding between GvH—O3 within the same lattice plane (002) (Krassig 1990) 3 Figure 1.3: Fringed fibril model of cellulose supramolecular structure (Hearle 1958) 4 Figure 1.4: Hydrogen-bonding network in cellulose II. a) centre down chains (002 plane); b) centre up chains (002 plane); c) between antiparallel chains (101 plane) (Krassig 1990) 6 Figure 1.5: Schematic representation of the coordination binding of copper in cuoxam with the C2-O" and C3-O" groups in a cellulose molecule (Seger and Burchard 1994) 9 Figure 1.6: Structural representation of Cellulose acetate 11 Figure 1.7: Possible conversion of different polymorphs of cellulose and cellulose triacetate (CTA) (Sprague 1958) 13 Figure 1.8: Dependence of Tgj Tm and Tj on the average degree of substitution DS (Pizzoli et al. 1985) .' 16 Figure 1.9: Schematic representation of cluster growth during phase separation and gel formation (Winter and Mours 1997) 21 Figure 1.10: Viscosity as function of shear rate '. 25 Figure 1.11: Elastic (G1) and viscous (G") moduli as a function of angular frequency... 27 Figure 2.1: Effect of BD content on the steady shear viscosity for a CA/DMAc/BD system at CA 15wt% and a angular frequency of 1 rad s"1 42 Figure 2.2: Viscosity as a function of BD content obtained at a shear rate of Is"1 for five CA concentrations 43 Figure 2.3: Elastic (G') and viscous (G") moduli for 15 wt% CA solution at several BD concentrations. Shown is a clear solution, cloudy system, and self-supporting gel 45 vii Figure 2.4: Effect of non-solvent a) ED, b) BD and c) HD content on elastic moduli (G' at 1 rad s"1) at various CA concentrations 47 Figure 2.5: Effect of non-solvent (BD) content on transmission intensity and elastic moduli (G') for 10, 15 and 20 wt% CA (1 rad s"1) 48 Figure 2.6: Stress sweep experiments conducted for a) 10 wt.% CA, b) 15 wt.% CA, and c) 20 wt% CA samples with varying BD contents 51 Figure 2.7: Stress sweep experiments conducted for 15 wt% CA samples at varying non-solvent contents; a) ED b) BD, and c) HD 52 Figure 2.8: FTIR analysis of the various components in the CA/DMAc/BD ternary system. Included in the figure are the FTIR spectrum of viscous solution and gel samples. 54 Figure 2.9: FTIR analysis of the CA/DMAc/BD ternary system at 10, 15 and 20 wt% CA concentration and constant BD content (26 wt%). Also included is the FTIR spectrum of the 15 wt% CA/DMAc/BD (41.6 wt%) gel sample 56 Figure 2.10: Viscous (G") and elastic (G') moduli at or near the gel point for 10, 15 and 20 wt% CA concentration at various BD contents (numbers in parentheses are weight percentage) 57 Figure 2.11: G' and G" frequency sweep spectra for ED, BD and HD as non-solvents at an intermediate state at 15 wt% CA 58 Figure 2.12: Influence of the individual solubility parameter indexes on G' at a fixed frequency of 1 rad s"1 (15 wt% CA and BD as non-solvent) 61 Figure 2.13: Effect of hydrogen bonding solubility parameters on the elastic modulus G' of ED, BD and HD ternary systems at 15 wt% CA concentration 62 Figure 2.14: FTIR spectra of the hydroxyl stretching region of a) CA/DMAc/non-solvent, and b) DMAc/non-solvent solutions at 15 wt.% CA concentration and 26.6 wt.% non-solvent content 63 Figure 2.15: Elastic and viscous moduli of CA/DMAc/BD gels as a function of frequency at 15 wt % CA concentration 64 Figure 2.16: Elastic modulus of CA/DMAc/non-solvent gels as a function of non-solvent content (G' obtained at frequency of 1 rad s"1) 65 Figure 2.17: Elastic modulus of CA gels at different a) BD and b) CA concentrations as a function of strain. The limit of linearity shifts to lower strain as concentration increases. 67 viii Figure 2.18: Change in onset point of nonlinearity of elastic modulus for ED, BD and HD systems (CA15 wt%) 68 Figure 2.19: Micrograph images of gels at the same elastic modulus (G') and at the same CA concentrations (10 wt.%) for ED, BD and HD systems. From top to bottom are LSCM in fluorescent mode (top), LSCM images in reflective mode (middle), and SEM images at 3000x magnification (bottom) 70 Figure 2.20: Micrograph images of CA/DMAc/BD gels at the same elastic modulus (G') and three different CA concentrations (10, 15 and 20wt.%), From top to bottom are LSCM in fluorescent mode (top), LSCM images in reflective mode (middle), and SEM images at lOOOx magnification (bottom) 72 Figure 3.1: Steady state viscosity as a function of shear rate (s"1) for varying concentrations of 1-Propanol at a constant CA 15 wt% concentration 85 Figure 3.2: Effect of a) hydrophilic and b) hydrophobic interactions on the viscosity enhancement of the CA/DMAc/non-solvent solutions (values obtained at a shear rate of 1 s"1) 88 Figure 3.3: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/l-Pro system 90 Figure 3.4: FTIR spectra of the hydroxyl stretching region of a) CA/DMAc/l-Pro, CA/DMAc/l,3-PD, and CA/DMAc/Gly solutions (33.3 wt% non-solvent content and 15 wt% CA concentration), and b) DMAc/l-Pro, DMAc/l,3-PD, and DMAc/Gly solutions (33.3 wt% non-solvent content) 92 Figure 3.5: FTIR spectra of the hydroxyl stretching region of CA/DMAc/l-Pro, CA/DMAc/l-Hex, CA/DMAc/l-Oct, and CA/DMAc/l-Dec solutions (33.3 wt% non-solvent content and 15 wt% CA concentration) 93 Figure 3.6: Stress sweep spectra for ternary systems consisting of 1-Pro, 1,3-PD and Gly at 33.3 wt% non-solvent content and 15 wt% CA concentration 94 Figure 3.7: Stress sweep spectra for ternary systems consisting of 1-Pro, 1,3-PD and Gly at 33.3, 21.4, Gly 16.7 wt.% non-solvent content, respectively and 15 wt% CA 95 Figure 3.8: Stress sweep spectra for ternary systems consisting of 1-Hex, 1-Oct and 1-Dec at 35wt% non-sol vent content and 15 wt% CA concentration 96 Figure 3.9: Elastic (G') and viscous (G") moduli for the CA/DMAc/l-Pro ternary system at 15 wt% CA concentration and several 1-Pro contents. Shown are a clear solution, cloudy system, and self-supporting gel 97 ix Figure 3.10: Effect of increasing non-solvent content on the elastic modulus (G' at 1 rad s"1) for the a) hydrophilic and b) hydrophobic ternary systems 99 Figure 3.11: Digital images showing the effect of a) hexane, b) dibutyl ether, and c) 1-decanol on the phase behaviour of the ternary systems in DMAc at 15wt% CA 101 Figure 3.12: Frequency sweep spectra of 1,2-propandiol 1,3-propandiol and 1-propanol at the same concentration of non-solvent (36.3 wt.%) and CA (15 wt.%) 103 Figure 3.13: Effect of increasing non-solvent content on the elastic modulus (G' at 1 rad s"1) for the 1-Pro, 1,2-PD and 1,3-PD ternary systems. (15 wt.% CA concentration) 104 Figure 3.14: Viscous (G") and elastic (G1) moduli at or near the gel point for 1-Pro, 1,3-PD and Gly non-solvent ternary systems at 15 wt% CA concentration (numbers in parentheses are weight percentage of non-solvent) 106 Figure 3.15: Viscous (G") and elastic (G1) moduli at or near the gel point for 1-Pro, 1-Hex, 1-Oct and 1-Dec non-solvent ternary systems at 15 wt% CA concentration (numbers in parentheses are weight percentage of non-solvent) 107 Figure 3.16: FTIR spectra of the hydroxyl stretching region of C A/DM Ac/1-Pro (48.3 wt%), CA/DMAc/l,3-PD (34.6 wt%), and CA/DMAc/Gly (30 wt%) systems at the sol-gel transition point (15 wt% CA concentration) 108 Figure 3.17: FTIR spectra of the hydroxyl stretching region of CA/DMAc/l-Pro (48.3 wt%), CA/DMAc/l-Hex (40 wt%), CA/DMAc/l-Oct (37.3 wt%)and CA/DMAc/l-Dec (35 wt%) systems at the sol-gel transition point (15 wt% CA concentration) 109 Figure 3.18: Elastic modulus of CA gels for the different non-solvent, 1-Pro, 1,3-PD and Gly concentrations as a function of strain 111 Figure 3.19: Elastic modulus of CA gels for the different non-solvent, 1-Hex, 1-Oct and 1-Dec concentrations as a function of strain. The limit of linearity shifts to lower strain as concentration increases 112 Figure 3.20: Elastic modulus of CA/DMAc/non-solvent gels as a function of non-solvent content for a) hydrophilic and b) hydrophobic non-solvent systems 114 Figure 3.21: LSCM images of gels at the same elastic modulus (G') and at the same CA concentrations (10 wt.%) for the 1-Pro, 1,3-PD, Gly non-solvent ternary systems. The top images are in fluorescent mode and the bottom images are in reflective mode 118 Figure 3.22: LSCM images of gels at the same elastic modulus (G') and at the same CA concentrations (15 wt%) for the 1-Hex, 1-Oct, 1-Dec non-solvent ternary systems. The top images are in fluorescent mode and the bottom images are in reflective mode 119 x List of Tables Table 1.1: Frequently observed 20 values for CTA I and II (Kono et al. 1999) 14 Table 2.1: Effect of dihydric alcohol content on the rate of change in viscosity for ED, BD and HD systems 44 Table 2.2: Concentration (wt%) of non-solvent at which a 50% reduction in transmission intensity was observed 49 Table 2.3: Hansen solubility parameters for the components of a ternary system (8d -dispersive, 8P - permanent dipole-dipole, 8h - hydrogen-bonding) 60 Table 2.4: Fractal dimensions of CA/DMAc/non-sol vent gels at 15 wt% CA concentration, as obtained by rheology and confocal microscopy 69 Table 3.1: Concentration of Non-solvent at which non-Newtonian behaviour is observed 86 Table 3.2: Fractal dimensions of CA/DMAc/non-solvent gels at 15 wt% CA concentration, obtained by rheology and confocal microscopy 116 xi Chapter 1. Introduction 1.1 Cellulose Polymeric materials, natural and synthetic, have been increasingly replacing many traditional materials such as wood and metals in many applications (Nobles and Brown 2004). As a result, extensive research into the preparation and modification of polymers, and the resulting polymer properties has been conducted in order to investigate the applicability of polymeric materials to everyday products.(Nobles and Brown 2004) Of particular interest are polymers that are derived from renewable resources. Increasing global concerns over the sustainability and environmental impact of petroleum derived synthetic polymers are forcing the search for renewable alternatives (Lu et al. 2004). Cellulose is the most abundant biopolymer produced on earth. It is produced by nature at 11 12 an annual rate of 10 -10 tons per year. It is the main constituent of higher plants, such as wood, cotton and ramie; composing up to 40%-45% of their dry weight (Delmer and Haigler 2002). Cellulose is also produced by bacteria including Acetobacter xylinum, algae, Valonia, Chladophora, Rhizoclonium and Microdictyon and by several other sources such as in tunicin, a cell wall component of ascidians. Industrially, the principal sources of cellulose are wood, cotton fibre and cotton linters. In 1838, Anselm Payen first proposed the elemental composition of cellulose to be C6H10O5 (Payen 1838), classifying it as a carbohydrate. Cellulose is a polydispersed linear threodisyndiotactic homopolymer consisting of /3-D-anhydroglucopyranose units linked via /?(l-4) glycosidic bonds (trans or diequitorial linkages). The anhydroglucopyranose units (AGU) exist in the lowest energy 4Ci-chair conformation (Kennedy et al. 1987), Taking the dimer cellobiose as the basic repeating unit, cellulose 1 can be classified as an isotactic homopolymer of cellobiose (Figure 1.1). Based on the 4Ci-chair conformation each AGU possesses hydroxyl groups at the C2, C3, and Ce positions orientated in the ring plane, equatorially. The hydroxyl groups at both ends of the cellulose molecule behave differently. The Ci end has reducing properties, while the end with a free C4 hydroxyl group is non-reducing. The molecular size of native cellulose, which can be defined by its degree of polymerization (DP), differs widely depending on the origin and method of isolation. Native cotton has a DP range on the order of 15,000, while that of native wood cellulose is 10,000. This corresponds to a molecular mass of 2.4 and 1.6 million Da, respectively, or to a molecular length of 7.7 and 5.2 pm, respectively. non-reducing reducing end-group end-group Figure 1.1: Constitutional formula of the cellulose macromolecule illustrating cellobiose as the repeat unit. The P-D-anhydroglucopyranose units are in the 4C\ chair conformation with all hydroxyl substituents orientated equatorially (Krassig 1990). The backbone conformation of the cellulose chain is determined by the bond angles, bond length and torsion angles of the glycosidic bond. Generally, a bent backbone conformation is assumed (Figure 1.2), which together with the equatorial orientation of the hydroxyl groups creates a strong tendency to form intra- and intermolecular hydrogen bonds. The intramolecular hydrogen bonds are responsible for the stiff and rigid nature of 2 the cellulose molecule, as well as its "two-fold screw axis". The special chain conformation of native crystalline cellulose (cellulose I) is adequately represented by a 1,2 helix (Gardner and Blackwell 1974; Kolpak and Blackwell 1975; Blackwell et al. 1978; Kroon-Batenburg and Kroon 1997), although deviations have been observed and attributed to changes in hydrogen bonding or degree of order (Horii et al. 1987). From IR, NMR spectroscopy and X-ray diffraction studies (Liang and Marchessault 1959; Blackwell and Marchessault 1971), it has been demonstrated that intramolecular hydrogen bonds are formed along both sides of the cellulose chain (Figure 1.2). One exists between the C3-hydroxyl of one anhydroglucopyranose unit and the pyranose ring oxygen (O5) of an adjacent unit ( O 3 -H— O 5 ) , and the other is between the C2-hydroxyl and the adjacent C6-hydrogen (O2-H—Of,) which has a tg (trans-guache) orientation. The O3-H—0 5 ' hydrogen bond length is 2.15k, and that of the O2-H—06 hydrogen bond is 2.87A (Marrinan and Mann 1954). Figure 1.2: Intra- and intermolecular hydrogen bonding in cellulose I: intramolecular hydrogen bonding between O3 -H—O5' and O2-H—06, and intermolecular hydrogen bonding between 06-H—O3 within the same lattice plane (002) (Krassig 1990). 3 The intermolecular hydrogen bonding in native cellulose is between the hydroxyl groups of the Cs and C 3 ' positions of adjacent cellulose molecules located in the same lattice-plane (Figure 1.2). Thus, the hydroxyl group, which is in a tg position is involved in two secondary valence interactions, one intramolecular and one intermolecular. Thus, it is precluded from interacting with molecules in neighbouring 002-planes, above or below. This clearly demonstrates that native cellulose has a sheet-like structure with weak van der Waals forces holding the sheets together. Cellulose chains have a strong tendency to aggregate into highly ordered structural entities. This arises from the extensive intra- and intermolecular interactions, specifically intermolecular hydrogen bonding. The inter-chain cohesion is favoured by the high spatial regularity of the hydrogen bonding sites, which involves all three of the free hydroxyl groups within the AGU. The resulting cellulose aggregates are highly crystalline, although not uniform throughout the entire structure (Figure 1.3). A two-phase model, assuming low ordered (non-crystalline) and highly ordered (crystalline) regions, is generally accepted for cellulose and cellulose fibres - fringed fibril model (Hearle 1958). Non-crystalline (disordered) Regions Crystalline Regions (highly ordered) Figure 1.3: Fringed fibril model of cellulose supramolecular structure (Hearle 1958). 4 Cellulose exists in several crystal structures, which differ in unit-cell dimensions and possible chain polarity (Kadla and Gilbert 2000). Cellulose I, or native cellulose, is the predominate crystalline structure of algal, bacterial, some animal and most higher plants, and can be converted into the other polymorphs through a variety of treatments (Hayashi et al. 1987). Two forms of cellulose I exist, cellulose Ia and cellulose Ip. Cellulose Ia is the primary polymorph in bacterial and algal celluloses, while cellulose Ip predominates in higher plants such as cotton and wood (Atalla and VanderHart 1984; VanderHart and Atalla 1984). Cellulose Ia is metastable, and can be irreversibly converted to cellulose Ip by annealing in dilute alkali at high temperature (Sugiyama et al. 1991). Electron diffraction studies have shown that cellulose Ia has a 1-chain, triclinic (PI) unit cell structure (a = 6.74 A, b = 5.93 A, and c (fibre axis) = 10.36 A; a = 117°, (3 = 113°, y = 81°), while cellulose Ip has a 2 chain monoclinic (P21) unit cell (a = 7.85 A, b = 8.17A, c (fibre axis) = 10.36 A; y = 97.3°) (Sugiyama et al. 1991). Native cellulose has a parallel chain orientation (Hieta et al. 1984; Chanzy and Henrissat 1985). In the presence of alkali (mercerization), or through regeneration from solution, both cellulose Ia and Ip are transformed into an anti-parallel cellulose II (Sarko and Muggli 1974; Takai and Colvin 1978; Hayashi et al. 1987; Nishimura and Sarko 1991; Yamada et al. 1992; Raymond et al. 1995; Kono et al. 2004). The basic unit-cell structure of cellulose II is monoclinic, consisting of two cellulose chains with a P21 space group: a = 9.08 A, b = 7.92A, c = 10.36A (fibre axis), y = 117.1°. Although quite similar to Ip in unit cell dimensions, cellulose II has a slightly larger b-lattice plane distance, 9.04A vs. 8.17 A, respectively. The hydrogen bonding in cellulose II is more complicated than cellulose I, with more intermolecular cross-linking. Unlike cellulose I, the hydrogen bonding of the center and 5 corner chains of the unit cell of cellulose I I are not equivalent. The centre chain of cellulose I I forms sheets very similar to cellulose I, with two intramolecular hydrogen bonds and an intermolecular hydrogen bond between the O v H — O 3 to the next chain within the same lattice plane (002). However, in the corner chain of the cellulose I I , the C 6 hydroxymethyl group is in the gt position and forms two intermolecular hydrogen bonds, one within the same 002 plane (06—H-O2) and the other between O2-H—Or along the diagonal in the 101 plane. a) b) c) Figure 1.4: Hydrogen-bonding network in cellulose I I . a) centre down chains (002 plane); b) centre up chains (002 plane); c) between antiparallel chains (101 plane) (Krassig 1990). Native cellulose can also be converted to other polymorphs, such as cellulose I I I and I V , through various treatments. Crystalline modification to cellulose I I I is obtained by treating either cellulose I or cellulose I I with liquid ammonia at a low temperature, then recrystallizing the preparation by evaporating the ammonia. Cellulose I I I has a 6 monoclinic unit cell (a = 9.9 A, b = 7.74 A, c = 10.3A (fibre axis), y - 122°), with small differences in lattice dimension between cellulose IIIi and IIIn. Cellulose IV on the other hand, is obtained by treating various polymorphs in a suitable liquid at high temperature under tension, resulting in an average lattice spacing of 7.9 A, b = 8.11 A, c = 10.3A (fibre axis), y = 90°. 1.2 Cellulose Modification The extensive intra- and intermolecular interactions present in native cellulose are responsible for the relative stiffness, rigidity and poor processability of the cellulose polymer. In fact these interactions are so strong that cellulose, a semicrystalline polymer, decomposes prior to melting and is insoluble in most conventional solvents (Bochek 2003). As a result, complex solvent systems and/or chemical modification of cellulose is required to facilitate its dissolution and processing, the nature and extent of which is dependent on the intended industrial or laboratory applications (Kobayashi et al. 1999). The complex supramolecular structure and strong hydrogen bonds of cellulose has led to the development of special solvent systems for dissolving it and carrying out the controlled synthesis of cellulose derivatives. The creation of several solvent-systems for cellulose has opened up new approaches to cellulose derivatization under homogeneous conditions and to the analytical characterization of cellulose macromolecules in the dissolved state. The mixture of 7V,7V-dimethylacetamide (DMAc) combined with LiCl has shown substantial potential for the analysis of cellulose and the preparation of a wide variety of derivatives (McCormick and Lichatowich 1979). Another powerful solvent-system for cellulose is a combination of dimethyl sulfoxide and tetrabutylammonium fluoride trihydrate (DMSO/TBAF) (Ass et al. 2004). An advantage of DMSO/TBAF is 7 that cellulose with a degree of polymerization as high as 650, dissolves without any pre-treatment. The acetylation of cellulose linters in DMSO/TBAF results in good control of the degree of substitution (DS), and distribution of acetyl groups during derivatization. Aqueous solutions of a number of metal complexes have also been found to dissolve cellulose. The first success was reported by Schweizer (1857), who discovered that a solution of cupric hydroxide in aqueous ammonia (cuoxam) dissolves cellulose (Schweizer 1857). Recently, a number of new metal complexes have been developed that completely dissolve cellulose through deprotonation and coordinative binding. These include cadmium complexes, Cd-tren (tren = tris(2-aminoethyl)amine, N[-CH2-CH2-NH2]3(OH)2), nickel complexes (Ni-tren) and copper complexes (Cu-tren). These solvents are capable of dissolving the high DP cellulose found in cotton linters and bacterium (DPW = 9700) (Saalwachter et al. 2000). The water-containing metal complexes bind via coordination to the deprotonated OH groups at the C2 and C3 position of the AGU (Figure 1.5). These solvents were designed such that only two of the coordination sites remained free to bind to the deprotonated hydroxyl groups within cellulose. Four of the six coordination sites of the Ni2 + and Cd2+ cations are bound to the tren, leaving two sites in the cis configuration free for binding to the AGU. A representation of the coordination binding of the Cu complex with a cellulose molecule is shown in Figure 1.5. In addition to good molecular dispersity and common solution behaviour, the metal complexes posses a remarkable chain stiffness which may be due to strengthened intrachain hydrogen bonding (Figure 1.5) (Saalwachter et al. 2000, Saalwachter, 2001, Seger, 1994). 8 Figure 1.5: Schematic representation of the coordination binding of copper in cuoxam with the C2-O" and C3-O" groups in a cellulose molecule (Seger and Burchard 1994). Very recently, several ionic liquids, termed "green solvents" were also found to dissolve cellulose (Heinze et al. 2005). By definition, an ionic liquid is a liquid that contains essentially only ions. In the broad sense, the term includes all molten salts, however, the term "ionic liquid" is commonly used for salts whose melting point is relatively low (below 100 °C). Ionic liquids such as l-N-butyl-3- methylimidazolium ([C4mim]+), with different anions like chloride are used as a reaction media for the acetylation and the carboxymethylation of cellulose in presence of DMSO. It has been demonstrated that ionic liquids can dissolve the biopolymer independent of DP up to 1200. For example, using [C4mim]+ Cf as the reaction medium, it was shown a good yield of cellulose acetate with a high DS can be rapidly synthesized (Heinze et al. 2005). The preparation of cellulose derivatives under different conditions has been investigated for many years. In general, two classes of cellulose derivatives are of commercial significance: cellulose esters (e.g. cellulose acetate) and cellulose ethers (e.g. methylcellulose). Of these cellulose esters, cellulose xanthates, acetates, acetatebutyrates, nitrates and phosphates are the most industrially significant, comprising more than 90% 9 of the production capacity in the chemical processing of cellulose. Industrial cellulose esters such as cellulose acetates are usually made by reacting of cellulose with a corresponding organic/inorganic acid and a strong acid catalyst (Steinmeier 2004; Zugenmaier 2004). Depending on the functional groups introduced and the DS, cellulose esters can be used for the production of textile fibres, coatings, cigarette filters, and membranes. In contrast, cellulose ethers are manufactured through alkali reactions with alkyl halides (Williamson ether synthesis) or alkylene oxides (ring-opening reactions). Industrial etherification is performed under heterogeneous conditions beginning with alkali cellulose. The most economically important cellulose ethers are carboxymethylcellulose (CMC), methylcellulose (MC) and hydroxyethylcellulose (HEC); which are typically water soluble and widely used to modify rheological properties in paint industry, oil recovery, foods and cosmetics (Hirrien et al. 1998; Steinmeier 2004; Zugenmaier 2004). 1.3 Cellulose Acetate and its Properties Cellulose acetates are among the most important cellulose esters in commercial production. They have a wide range of properties and applications depending on their degree of acetylation. Cellulose acetate can be fully acetylated with all three hydroxyl groups of the monomelic anhydroglucopyranose unit being derivatized or only a portion thereof in a regioselectively or statistically distributed manner (Figure 1.6). Of the various acetates, the so-called secondary acetate or cellulose diacetate is of commercial significance. Despite its name, cellulose diacetate has an average degree of 2.5 acetate groups per anhydroglucopyranose unit. Likewise, commercial cellulose triacetate (CTA) refers to cellulose acetates with a DS above 2.7. 10 Figure 1.6: Structural representation of Cellulose acetate Cellulose acetate was first discovered by Paul Schutzenberger in 1865 (Heinze and Liebert 2004). It was synthesized by heating cellulose in a sealed glass tube with acetic anhydride. The resulting CTA exhibited properties close to that of collodium and nitrocellulose, however, its solubility and elastic properties were significantly different, requiring new processing methods to be developed. Furthermore, the cost of the raw materials required to synthesize CTA rendered its utilization uneconomic. It was not until the discovery of cellulose diacetate (a mixture of di- and triacetate) in 1904 that cellulose acetates became commercially important. Cellulose diacetate is soluble in acetone and other readily available solvents, such as methyl acetate and ethyl acetate. Additionally, its mechanical properties are close to collodion (a solution of nitrocellulose in ether and acetone). The first important application of CA was their use in coatings for airplanes.(Heinze and Liebert 2004) Today, cellulose acetates are used in filter tow, textile fibres, films, mouldings and coating applications. Today, cotton linters or softwood sulphite / prehydrolyzed sulphite pulps are typically used as the raw material for cellulose acetates. Acetylation is performed in the presence of acetic acid and in combination with methylene chloride and sulphuric acid. The 11 primary product is cellulose triacetate (acetyl content of 44.8%) (Heinze and Liebert 2004; Saka 2004). Partially substituted cellulose acetate or secondary cellulose acetate (DS of 1.8 to 2.5) is then made by the hydrolysis of CTA through the addition of water, dilute acetic acid or NaOH (Heinze and Liebert 2004; Saka 2004; Steinmeier 2004). The solution properties of cellulose acetates are strongly influenced by the average degree of substitution and distribution of acetate groups along the cellulose chain. Cellulose acetates with a degree of substitution ranging from 0.5 to 1 are soluble in water. However, as the number of acetyl groups increases the solubility of CA in water decreases, but at DS>1 cellulose acetates exhibit good solubility in a variety of organic solvents, such as acetic acid, tetrahydrofuran, acetone, A^ TV-dimethylacetamide, dioxane and dimethyl sulfoxide (Gomez-Bujedo et al. 2004). This phenomenon is attributed to both the disruption of the intra- and intermolecular hydrogen bonding within the CA chains and the formation of specific interactions between functional groups in CA and the solvent. For example, basic solvents like acetone interact primarily with the hydroxyl groups on the CA chain, while acidic solvents such as formic acid primarily solvate the acetyl groups (Pintaric et al. 2000; Gomez-Bujedo et al. 2004). Therefore, the use of specific solvents can induce structural changes in solution depending on the amount of acetyl and hydroxyl groups along the partially substituted CA chains. The thermal behaviour of cellulose acetates is also influenced by their DS, the regiochemistry of acetylation, and the method of preparation and analysis. Cellulose acetates are polymers with high melting temperatures and low melting entropies. Cellulose acetate with DS of 2.5, derived from the deacetylation of CTA, exhibits a partially crystalline structure. In this secondary cellulose acetate, crystalline domains are 12 only a few molecules wide and both crystalline and non-crystalline sections are formed in an antiparallel shape which is adjacent along a fibril (Kamide and Saito 1985; Gomez -Bujedo et al. 2004). Highly substituted cellulose acetates (CTA) are also semicrystalline polymers, which can exist in two allomorphs, CTA-I and CTA-II, depending on the crystal structure of the starting cellulose and the acetylation conditions. CTA I is typically produced by the heterogeneous acetylation of native cellulose (cellulose I), while CTA II is made from regenerated or mercerized cellulose (cellulose II). Figure 1.7 outlines the interconversion of the various cellulose and CTA allomorphs (Sprague 1958, Wolf, 1992 #124). Figure 1.7: Possible conversion of different polymorphs of cellulose and cellulose triacetate (CTA) (Sprague 1958). In the structure of native cellulose I, the chain molecules are arranged in a parallel fashion. Conversely, in cellulose II they exist in an antiparallel manner. Therefore, CTA I can be obtained only from cellulose I, and only if the acetylation is performed in a heterogeneous fashion; i.e. the polymer is never dissolved or strongly swollen. 13 Accordingly, CTA I retains the parallel chain arrangement of the original cellulose I, and in the case of Ramie cellulose the process is reversible (Rustemeyer 2004). Homogeneous acetylation of cellulose I results in the production of the antiparallel CTA II allomorph, and CTA I can be irreversibly converted to CTA II by annealing (Figure 1.7). As expected, saponification of CTA II leads to cellulose II (Wolf et al. 1992). The crystal structures of CTA I and CTA II can be distinguished by X-ray diffraction, CPMAS NMR and DSC (differential scanning calorimetry). Of these, X-ray diffraction is most commonly used. The diffractogram of CTA I is relatively simple as compared to CTA II, Table 1.1 summarizes the 20 values for frequently observed reflections. Table 1.1: Frequently observed 20 values for CTA I and II (Kono et al. 1999). Polymorph 20 Value Intensity CTA I 7.64 14.58 15.90 17.80 20.30 22.37 26.50 Very strong Weak Very strong Weak Strong Weak Weak 8.43 Very strong 10.42 Very strong 13.14 Very strong 16.28 Strong CTA II 16.74 Strong 18.59 Strong 21.39 Weak 23.39 Strong 26.50 Weak 14 DSC thermograms also demonstrate the difference between CTA I and CTA II. However, as CTA I can be converted to CTA II by annealing, the first heating run must be used. A sharp melting peak at 580 K is observed in the first heating cycle, and can be assigned to CTA I. The second and subsequent heating cycles reveal a broad melting peak, an indication of CTA II formation, which upon further heating cycles drops the observed melting peak by 30-40K. The broad endotherm suggests the crystallization is not completed and/or only a small amount of crystallites have been formed, and is an indication of the conversion of CTA I to CTA II.(Zugenmaier and Vogt 1983; Kamide and Saito 1985) As mentioned above, cellulose acetates are described as polymers with a high melting temperature and low melting entropy. Thermal data for CTA shows that the polymer has a glass transition temperature (Tg) around 190 °C, melting temperature (Tm) of 307 °C and decomposition temperature (Td) of 356 °C. As expected the Tg, which is related to segmental chain motion, is affected by the degree of acetylation, varying between 190 and 220 °C for CA samples with 61 - 52.6% acetic acid content (Tang et al. 1996). In addition to the Tg and Tm, the DSC thermogram of CTA has an endothermic peak between 50 and 80°C related to the evaporation of water, and a small exothermic peak at 190 to 206 °C which has been assigned to the crystallization temperature (Tc) (Coyle et al. 1996). However, the Tc peak is not always present in commercial samples and the reason for this is the subject of ongoing investigations. For secondary CA, the DSC thermograms show a drop in melting and crystallization temperature, due in part to the smaller and less perfect crystallite structure. An extensive investigation of the thermal properties of cellulose acetates of various DS (0.49, 1.75, 2.46, 2.92) has been 15 investigated (Kamide and Saito 1985). The effect of the average degree of substitution on the glass transition temperature, melting temperature and decomposition temperature is shown in Figure 1.8 (Kamide and Saito 1985; Pizzoli et al. 1985). 6 5.3 T/K 4.7 4 0 1 2 3 DS Figure 1.8: Dependence of Tg; Tm and Tj on the average degree of substitution DS (Pizzoli etal. 1985). 1.4 Cellulose Acetate - Applications and Emerging Uses The use of CA dates back to the early 1900s, when the first photographic film was produced by the Dreyfus brothers. The properties of CA including good toughness, deep gloss and a "natural" feel have made CA attractive for a number of other applications. As well, CA has the advantage of having low toxicity. This has led to the use of CA in a variety of industrial applications such as the production of textiles, tool handles, speciality papers, and cigarette filter tow. Additionally, properties such as transparency, 16 smoothness and color have allowed CA to be used for handles of toothbrushes, umbrellas and screwdrivers. Spectacle frames have been made from sheets of cellulose acetate, but for this application it is being replaced to some extent by synthetic thermoplastic materials (Rustemeyer 2004). One draw back to cellulose acetate utilization is its limited processability. As a result, binary blends of CA with synthetic polymers have been the subject of increasing interest. Polymer blending is widely employed to improve and tailor material performance and processability. For example, cellulose acetate is used extensively for the production of safety face shields because of its excellent optical quality and durability. However, due to its fairly high Tg and Tm, CA can not be easily moulded. Therefore, vinyl polymers such as polyfvinyl acetate) (PVAc), poly(N-vinyl pyrrolidone) (PVP), and poly(N-vinyl pyrrolidone-co-vinyl acetate) [PVP-co-PVAc)] are used as flexible synthetic polymers to plasticize CA and facilitate mould processing (Miyashita et al. 2002). Another area of particular interest is in the blending of low DS water soluble CA with PVP for use as controlled release materials with unique swelling properties (Qin et al. 2003). Cellulose acetate as filters has been introduced for the manufacturing of separation and filtration media. Due to the hydrophilic nature of the cellulose acetate membrane and low binding, they are suitable for maximum sample recovery for tissue culture media preparation, sterile filtration of biological fluids, protein and enzyme filtrations and filtration of other aqueous solutions. As well, CA has shown good promise in the area of drug delivery and research is underway to determine its effectiveness and applicability. 17 1.5 Cellulose Acetate - Gelation Mechanism In polymeric systems, physical gel formation can arise as a result of phase separation. Depending on the system, phase separation leads to polymer aggregation and the formation of large macromolecular assemblies. Gelation is initiated by large macromolecular associations forming clusters, which leads to an infinite homogenous cluster extending through the entire volume of the system (Nielsen 1977; Ilyina et al. 1993; Winter and Mours 1997; Ross-Murphy 1998). Previous work on cellulose acetate gels has shown that they exhibit reversible thermal properties, depending on factors such as CA concentration, acetyl content and the type of solvent (Cai and Zhang 2006). However, unlike other cellulose esters where solidification occurs when the polymer system is heated; in most CA/solvent systems, gelation takes place after the CA solution is heated to a specific temperature and subsequently cooled (Ryskina and Averyanol971). CA/benzyl alcohol solutions have been shown to form gels when heated to 80°C, then cooled to 25°C. It has been proposed that gel formation is induced by the existence of strong intermolecular associations in the system (Ryskina and Averyano 1971; Ryskina and Averyanova 1975). In some CA/solvent systems, the addition of non-solvents such as water or alcohol can induce gelation. It was recently demonstrated that the addition of water to a CA/DMAc solution results in gelation (Appaw et al. 2007). This system exhibited a liquid-solid transition and phase separation prior to gel formation. It was shown that the sol-gel transition and gel properties were dependent on both water and polymer concentrations. Phase separation was characterized by the development of cloudiness in the system, the result of polymer aggregation and formation of large macromolecular clusters. It was 18 reported that gelation occurred from the manipulation of inter- and intramolecular hydrogen bonding interactions, as well as hydrophobic non-bonding interactions between components. In another study, water, methanol, ethanol and isopropanol were used as coagulants for CA (in acetone) membrane formation (Hao and Wang 2001). This study demonstrated that the alcohols are relatively weak coagulants as compared to water; water > methanol > ethanol > isopropanol. The phase behaviour and coagulated CA membrane morphology was shown to vary depending on the coagulant or non-solvent. For the mono-hydroxyl alcohol non-solvents, a correlation between the calculated non-solvent/polymer interaction parameter and phase behaviour was observed. It was found that the water-coagulated membranes underwent a liquid-liquid phase separation and the formation of large macrovoids, while those coagulated from methanol underwent a spinodal microphase separation leading to a honeycomb-like membrane structure. In the case of ethanol and isopropanol, there was a delay before phase separation occurred resulting in a top layer with thicker density (Hao and Wang 2001). 1.6 Rheology and the Phase Behaviour of Polymers 1.6.1 Liquid-solid Transition and the Solidification Process Rheology, which is the study of the deformation and flow of matter under the influence of an applied stress, can be a reliable method to investigate the mechanical properties of polymers in solution (Nielsen 1977; Barnes et al. 1989). In practice, rheology is primarily concerned with applying the disciplines of elasticity and (Newtonian) fluid mechanics to materials whose mechanical behaviour cannot be described with the classical theories. 19 Predictions for mechanical behaviour (on the continuum mechanical scale) can be established based on the micro- or nanostructure of the material, such as the molecular size and architecture of polymers in solution or the particle size distribution in a solid suspension. The universality of the rheological behaviour of polymers has been extensively explored, and can be used to study the solution, liquid to solid transition (LST) and gel behaviour of various systems (Winter and Mours 1997). The LST of polymers is technically important as it occurs in nearly all common fabrication processes, such as the injection moulding of semi-crystalline polymers and the processing of cross-linked polymers (Idris et al. 2003). Knowing the exact point of the liquid to solid transition of polymer is of critical importance for the design and operation of polymer processing. In chemically cross-linked polymers, the instant of LST is defined as the chemical gel point (Vilgis and Winter 1988). Prior to gelation, the system consists of a distribution of finite clusters, while after it contains infinitely large macromolecules and particulate clusters which do not flow (Dumitras and Friedrich 2004). At that gel point, there is a substantial change in the molecular mobility of the polymers, resulting in a large change in their rheology. Processing at the gel point potentially produces interesting polymer texture development, which can be critical to material properties (Winter and Mours 1997). Solidification, or gelation, is a measure of connectivity and its process differs from material to material. Many parameters effect and accelerate this process, including polymer concentration, molecular weight, type and number of functional groups and solvent quality (Chiou et al. 2001). Gelation can occur as a result of covalent linkages (chemical gels) or entanglement (physical gels) of polymer constituents. Typically, 20 chemical gel networks are characterized by junction points, while physical gel networks by junction zones. Most physical gels dissolve upon addition of a suitable solvent, whereas chemical gel networks swell but do not dissolve. Figure 1.9 illustrates cluster growth during the sol-gel transition in a typical polymeric system. At p=0 (p is the bond probability), only individual monomers are present. As the bond probability, or index of cross-linking increases, interconnectivity increases and molecular clusters grow in size. During phase separation and before the network spans the entire sample, the ratio of connected bonds or cross-linking to the total number of possible bonds is 0<p<l. At the critical extent of bond formation, p—>pc, the molecular weight distribution spreads to infinity. Molecular sizes range from small unreacted oligomers to infinite clusters, defining the gel point (Winter and Mours 1997). p = 0 0<p<l p-+l Figure 1.9: Schematic representation of cluster growth during phase separation and gel formation (Winter and Mours 1997). Chemical gels (or covalently cross-linked materials) are formed by a variety of mechanisms including cross-linking of high molecular weight linear chains (either chemically or by radiation), end-linking of reactant chains with a branching unit, or the step growth polymerization of oligmeric multi-functional precursors (Nielsen 1977; 21 Ross-Murphy 1997; Winter and Mours 1997; Ross-Murphy 1998; Eissa and Khan 2005). Chemical gels are true macromolecules in that their molecular weight is nominally infinite and they posses an infinite relaxation time and an equilibrium modulus (Ross-Murphy 1998). In contrast, physical gels are formed by the topological interaction of polymer chains, either in the melt or in solution when a certain critical molecular weight or concentration is reached. The presence of non-covalent bonds as cross-links complicates any physical description of the network properties, since the numbers and positions of physical bonds can fluctuate with time and temperature. In many cases bonding in physically cross-linked gel often involve forces such as, Coulombic, dipole-dipole, hydrophobic and hydrogen-bonding interactions (Ross-Murphy 1998; Li et al. 2001). In biopolymer gels, non-covalent cross-links are generally formed by one or more of the above listed interactions and these are combined with more specific and complex mechanisms for gelation. The gelation process in physical gels is reversible and they posses a potential advantage in that junctions can formed and broken by altering the environment (temperature, pressure and pH) (Guenet 1991; Ross-Murphy 1998; Eissa and Khan 2005). These characteristics distinguish physical gel from chemical gel network structures. Their limited life span makes physical gels fluid-like in long term applications, but also allows them to heal and repair if they are broken or disrupted (Guenet 1991). Many parameters and mechanisms can influence polymeric system and connect polymers into large-scale structures, enhancing their mechanical properties. For example, when a polymer melt is mixed (or loaded) with filler that does not interact strongly with the polymer, the viscoelastic properties increase due to partial volume occupancy by rigid and immobile masses of fillers. However, when the filler is able to react with the polymer, 22 e.g. rubber loaded with finely ground carbon black, the behaviour is somewhat different. A chemical cross-linked network structure results, with the formation of an infinite molecular weight system (Castellani and Lomellini 1991). Similarly, the addition of filler compounds to a polymeric solution can result in polymer aggregation and the development of polymer network structures (Tweddle and Sourirajan 1978; Gildert et al. 1979). The addition of a non-solvent such as water or alcohol to a biopolymer solution has been shown to induce gelation, resulting in an entangled polymer system with enhanced mechanical and rheological properties (Tweddle and Sourirajan 1978; Gildert et al. 1979; Hao and Wang 2001; Kadla et al. 2005; Appaw et al. 2007). The addition of water to a CA/DMAc solution induced gelation and the formation of an entangled polymer system (Appaw et al. 2007). Phase separation was observed to occur prior to gel formation. Depending on polymer concentration, the addition of water led to a gradual to extreme cloudiness in the system and ultimately gel formation. The resulting gels were formed by polymer aggregation and the creation of large macromolecular clusters as a result of hydrophobic (molecular entanglement) and hydrophilic (hydrogen bonding) interactions between constituents. 1.6.2 Rheological Characterization of Polymeric Solutions and Gels Steady shear measurements of polymeric solutions are performed by subjecting a sample to a constant shear rate (j) which results in the generation of a shear stress (x). For polymeric liquids, emulsions and concentrated suspensions the rate of deformation is dependent on the viscosity (n) of the system and is measured as a function of the steady shear rate, defined as (Barnes et al. 1989): 23 r|= x/y Most materials exhibit Newtonian behaviour under normal circumstances, and r\ is independent of shear rate (Barnes et al. 1989). A unique characteristic of polymer solutions and melts is the non-Newtonian behaviour, in that the apparent viscosity changes with an increase in shear rate. The flow behaviour of non-Newtonian materials helps to specify the degree of dispersion and/or flocculation, as well as shear thinning or thickening (Nielsen 1977; Ferry 1980; Barnes et al. 1989; Goodwin et al. 2000). Nearly all non-Newtonian materials exhibit shear thinning (pseudoplasticity), which is a reduction in viscosity that extends over several orders in magnitude in shear rate (Barnes et al. 1989). Following the removal of the stress, a gradual recovery of structure generally occurs, and is called thixotropy. There are some cases wherein non-Newtonian materials exhibit viscosity increases with shear rate. This unusual flow behaviour is called shear thickening (dilatancy) and is illustrated in Figure 1.10. The shear thickening region extends only about one order of magnitude of shear rate. In shear thickening, deforming the material causes rearrangement of its microstructure such that resistance to flow increases with shear rate (Nielsen 1977; Ferry 1980; Barnes et al. 1989). In this case, recovery of viscosity upon stress removal is referred to as anti-thixotropy. Most cellulosic systems typically exhibit a zero shear viscosity, n0, which is typified by a constant viscosity (Newtonian) region at low shear rates and shear thinning at high shear rates. 24 10° 1 1 1 Shear thickening I Vi 03 n , •9 Newtonian H H B - B B B B B — B Viscos lO"1 io- 2 Shear thinning ^ " ^ ^ i i i 1 10"1 10° 101 102 103 104 Shear rate (1/s) Figure 1.10: Viscosity as function of shear rate Parameters such as temperature, polymer molecular weight, polymer concentration, nature of solvent or plasticizer and fillers, determine the viscosity and rheological properties of non-Newtonian polymers. An increase in the concentration or molecular weight of the polymer, leads to an increase in the number of polymer entanglements in solution, thereby raising the viscosity values. Generally, polymer solutions at rest or at low temperature have a higher degree of entanglement or physical bonding than a polymer, which is flowing. Therefore, the polymer solution should be at an equilibrium state for shear viscosity measurements. By contrast, in a solid or gel-like state, the extent of cross-linking and overall rigidity of the sample is much higher, rendering the measurement of flow viscosity inappropriate (Nielsen 1977; Barnes et al. 1989; Macosko 1994). 25 Another method of exploring the structural properties of complex fluids is by applying a small amplitude oscillatory shear so that the fluid's microstructure is not significantly deformed. If the strain amplitude is small enough and the materials structure is not disturbed, the resulting shear stress is controlled by the rates of spontaneous rearrangements or relaxations present in the system. In dynamic, or oscillatory, experiments a sinusoidally varying strain y = y 0 sin(cot) in the linear viscoelastic (LVE) regime is applied; co is the frequency of oscillations and yo is the strain amplitude. The shear stress produced by a small-amplitude deformation is proportional to the amplitude of the applied strain. This stress is sinusoidally varying in time and can be represented by: The G' (co) term is in phase with the strain and is called the "storage modulus", while the G" (co) term is in phase with the rate of strain y and is termed the "loss modulus". The storage modulus represents the storage of elastic energy, while the loss modulus represents the viscous dissipation of energy. Most materials, including polymer solutions, exhibit both elastic and viscous properties and are referred to as viscoelastic materials. As such, both G' and G" will have values greater than zero. The plots of G' and G" as a function of frequency provide information on the structure of the material being studied. Figure 1.11 illustrates the typical behaviour of G' and G" in a viscoelastic material at three different states; liquid, transition state and gel. 26 104 103 101 10° G' G" Elastic gel Transition state ^ ^y - G" -G' — s S / Polymer 1 1 solution —I  1 I I 0.1 1 10 100 1000 Angular Frequency (rad s"1) Figure 1.11: Elastic (G1) and viscous (G") moduli as a function of angular frequency. In a liquid sample both G' and G" are frequency dependent and have slopes close to 2. At the intermediate state or LST transition G' ~ G" or a crossover of G' and G" is observed. At this point, the moduli are relatively frequency independent, indicating that the structure of the material is tending to solidify and approaching its gelation point. In the case of a complete elastic gel, both moduli become frequency independent with G' being larger than G*'(Tung and Dynes 1982; Li 2002; Avanza et al. 2005; da Silva and Areas 2005). The linear viscoelastic (LVE) is defined as the region corresponding to the stress varying linearly with strain for the sample being investigated (Macosko 1994). For dynamic frequency analysis to be valid, the experiments have to be conducted in the LVE region. 27 Therefore, dynamic strain/stress sweep experiments have to be performed to determine the LVE. This test involves the variation of strain amplitude while maintaining a constant frequency. One of the main advantages of dynamic oscillatory experimentation is that it probes material without disrupting the microstructure. Dynamic rheology has been used to characterize cellulose derivative systems and has provided important information such as the sol-gel transition (Wang and Fried 1992; Li 2002). 1.7 Goal of the Project Recent advances in biotechnology are fuelling the development of separation and purification techniques of biological molecules. These compounds are often very sensitive to traditional methods of separation, such as solvent precipitation, crystallization, or solvent extraction. As a result, interest in membrane separation techniques that are simple and reliable, and easily scaled-up are of great interest to the pharmaceutical industry. Membrane filters or "membranes" are polymer films with specific pore ratings. Membranes retain particles and/or micro-organisms that exceed their pore sizes (ratings) by acting as a physical barrier and capturing such particles on the surface of the membrane (Qin et al. 2003; Appaw et al. 2007). Asymmetric ultra-filters based on the hydrophilic properties of cellulose acetates are used for biological separation due to the compatibility of cellulose acetate with low protein adsorption, and the adjustable variation in the porosity of CA from coarse to fine. In such applications, phase separation induced gelation of CA is widely used and is of considerable interest for filtration and micro encapsulation applications (Kadla et al. 2005; Appaw et al. 2007). It has been shown that increasing CA and non-solvent concentration in a CA/DMAc/water ternary system accelerates sol-gel transition and formation of 3-D solid 28 structure. It is believed that the sol-gel transition and gelation are due to intensification of inter- and intramolecular hydrogen bonding interactions, as well as hydrophobic polymer-polymer interactions. The addition of alcohols, specifically mono-hydroxyl alcohols to a CA/acetone system also leads to phase separation. Depending on the non-solvent, water, methanol, ethanol, or isopropanol, the phase behaviour and coagulated CA membrane morphological structure was shown to vary (Hao and Wang 2001). The water coagulated membranes had large macrovoids from liquid-liquid phase separation, while those from methanol had a honeycomb-like structure from spinodal microphase separation. In the case of ethanol or isopropanol, a thicker dense top layer was observed due to the delay time phase separation. From the work reviewed above, it is anticipate that changes in non-solvent structure and properties, i.e. molecular size and propensity to form hydrogen bonds, will affect the sol-gel process and 3D network structure of the resulting gels and coagulated membranes. To assess this we have chosen a series of alcohols that differ in size and number of hydroxyl groups as non-solvents. We hypothesize that: • Increasing the alkyl chain length of a series of the dihydric alcohols (non-solvent) will influence the microstructure of the resulting gel. As the alkyl chain length increases, more open gel-networks will form. • Increasing the hydrophobic component and hydrogen bonding ability within the non-solvent (mono-, di-, trihydric alcohols) accelerates sol-gel transition and will also alter the microstructure of the resulting gel network. Since intermolecular hydrogen bonding between components plays a major role in the sol-gel transition of semi-concentrated CA ternary systems (Appaw et al. 2007), increasing the number of hydroxyl groups on the non-solvent should increase the ability 29 of the system to form hydrogen bonds between components and accelerate gel formation as well as alter gel properties. Increasing the hydrophobic characteristic of the non-solvent will assist in phase separation and as such may directly impact the gel-induced structures. Therefore, the effect of non-solvent addition on viscosity and the sol-gel transition will be studied by means of steady-state viscosity (r\) and dynamic rheology (G' & G"). Direct observation of the changes in hydrogen bonding will be studied using infrared (IR) spectroscopy. In addition, scanning electron microscopy (SEM) and laser scanning confocal microscopy (LSCM) will be used to visualize and characterize gel networks and provide insight into their formation. 30 1.9 References Appaw, C, Gilbert, R. D., Khan, S. A. and Kadla, J. F. (2007). "Viscoelastic Behaviour of Cellulose Acetate in a Mixed Solvent System." Biomacromolecules 8(5): 1541-1547. Ass, B. A. P., Frollini, E. and Heinze, T. (2004). "Studies on the homogeneous acetylation of cellulose in the novel solvent dimethyl sulfoxide/tetrabutylammonium fluoride trihydrate." Macromolecular Bioscience 4(11): 1008-1013. Atalla, R. H. and VanderHart, D. L. (1984). "Native cellulose: a composite of two distinct crystalline forms." Science fWashington, DC. United States) 223(4633): 283-5. Avanza, M., Puppo, M. C. and Anon, M. C. (2005). "Rheological characterization of amaranth protein gels." Food Hydrocolloids 19(5): 889-898. Barnes, H. A., Hutton, J. F. and Walters, K. (1989). An introduction to rheology. Amsterdam, Elsevier. Blackwell, J., Kolpak, F. J. and Gardner, K. H. (1978). "Structures of Celluloses-I and Celluloses-2." Tappi 61(1): 71-72. Blackwell, J. and Marchessault, R. H. (1971). Cellulose and Cellulose Derivatives. Bikales, N. M. New York, Wiley: 1-23. Bochek, A. M. (2003). "Effect of hydrogen bonding on cellulose solubility in aqueous and nonaqueous solvents." Russian Journal of Applied Chemistry 76(11): 1711-1719. Cai, J. and Zhang, L. (2006). "Unique gelation behaviour of cellulose in NaOH/Urea aqueous solution." Biomacromolecules 7(1): 183-189. Castellani, L. and Lomellini, P. (1991). "Dynamic Viscoelastic Properties of Molten Acrylonitrile Epdm Styrene (AES) at Different Rubber Contents." Plastics Rubber and Composites Processing and Applications 16(1): 25-29. Chanzy, H. and Henrissat, B. (1985). "Undirectional Degradation of Valonia Cellulose Microcrystals Subjected to Cellulase Action." Febs Letters 184(2): 285-288. Chiou, B. S., Raghavan, S. R. and Khan, S. A. (2001). "Effect of colloidal fillers on the cross-linking of a UV-curable polymer: Gel point rheology and the Winter-Chambon criterion." Macromolecules 34(13): 4526-4533. Coyle, F. M., Martin, S. J. and McBrierty, V. J. (1996). "Dynamics of water molecules in polymers." Journal of Molecular Liquids 69: 95-116. da Silva, M. A. and Areas, E. P. G. (2005). "Solvent-induced lysozyme gels: Rheology, fractal analysis, and sol-gel kinetics." Journal of Colloid and Interface Science 289(2): 394-401. Delmer, D. P. and Haigler, C. H. (2002). "The regulation of metabolic flux to cellulose, a major sink for carbon in plants." Metabolic Engineering 4(1): 22-28. 31 Dumitras, M. and Friedrich, C. (2004). "Network formation and elasticity evolution in dibenzylidene sorbitol/poly(propylene oxide) physical gels." Journal of Rheology 48(5): 1135-1146. Eissa, A. S. and Khan, S. A. (2005). "Acid-induced gelation of enzymatically modified, preheated whey proteins." Journal of Agricultural and Food Chemistry 53(12): 5010-5017. Ferry, J. D. (1980). Viscoelastic Properties of Polymers. New York ; Toronto Wiley, 1980. Gardner, K. H. and Blackwell, J. (1974). "Structure of Native Cellulose." Biopolymers 13(10): 1975-2001. Gildert, G. R., Matsuura, T. and Sourirajan, S. (1979). "Effect of Different Alcohol-Water Mixtures as Gelation Media During Formation of Cellulose-Acetate Reverse-Osmosis Membranes." Journal of Applied Polymer Science 24(1): 305-310. Gomez-Bujedo, S., Fleury, E. and Vignon, M. R. (2004). "Preparation of cellouronic acids and partially acetylated cellouronic acids by TEMPO/NaCIO oxidation of water-soluble cellulose acetate." Biomacromolecules 5(2): 565-571. Goodwin, J. W., Hughes, R. W. and Royal Society of Chemistry (Great Britain) (2000). Rheology for chemists : an introduction. Cambridge, England}, Royal Society of Chemistry. Guenet, J. M. (1991). "Thermoreversible Gelation of Isotactic Polystyrene." Abstracts of Papers of the American Chemical Society 202: 149. Hao, J. H. and Wang, S. C. (2001). "Calculation of alcohol-acetone-cellulose acetate ternary phase diagram and their relevance to membrane formation." Journal of Applied Polymer Science 80(10): 1650-1657. Hayashi, J., Kon, H., Takai, M., Hatano, M. and Nozawa, T. (1987). "Possible Cause of Structural Irreversibility between Cellulose-I and Cellulose-Ii Families." ACS Symposium Series 340: 135-150. Hearle, J. W. S. (1958). "Fringed-fibril theory of structure in crystalline polymers." Journal of Polymer Science 28: 432-5. Heinze, T. and Liebert, T. (2004). "Chemical characteristics of cellulose acetate." Macromolecular Symposia 208: 167-237. Heinze, T., Schwikal, K. and Barthel, S. (2005). "Ionic liquids as reaction medium in cellulose functionalization." Macromolecular Bioscience 5(6): 520-525. Hieta, K., Kuga, S. and Usuda, M. (1984). "Electron Staining of Reducing Ends Evidences a Parallel-Chain Structure in Valonia Cellulose." Biopolymers 23(10): 1807-1810. Hirrien, M., Chevillard, C, Desbrieres, J., Axelos, M. A. V. and Rinaudo, M. (1998). "Thermogelation of methylcelluloses: new evidence for understanding the gelation mechanism." Polymer 39(25): 6251-6259. 32 Horii, F., Yamamoto, H., Kitamaru, R., Tanahashi, M. and Higuchi, T. (1987). "Transformation of native cellulose crystals induced by saturated steam at high temperatures." Macromolecules 20(11): 2946-9. Idris, A., Ismail, A. F., Gordeyev, S. A. and Shilton, S. J. (2003). "Rheology assessment of cellulose acetate spinning solution and its influence on reverse osmosis hollow fiber membrane performance." Polymer Testing 22(3): 319-325. Ilyina, E. E., Daragan, V. A. and Prisment, A. E. (1993). "Features of Complexing and the Solution Structure from the Data on Rotational and Translational Mobility of Molecular Probes in Cellulose-Acetate Solutions." Macromolecules 26(13): 3319-3323. Kadla, J. F., Dai, Q. Z. and Khan, S. A. (2005). "Viscoelastic and microstructural changes in sol-gel systems of cellulose acetate in a mixed solvent system." Abstracts of Papers of the American Chemical Society 229: U292-U292. Kadla, J. F. and Gilbert, R. D. (2000). "Cellulose structure: A review." Cellulose Chemistry and Technology 34(3-4): 197-216. Kamide, K. and Saito, M. (1985). "Thermal-Analysis of Cellulose-Acetate Solids with Total Degrees of Substitution of 0.49, 1.75, 2.46, and 2.92." Polymer Journal 17(8): 919-928. Kennedy, J. F., Phillips, G. O. and Williams, P. A. (1987). Wood and cellulosics : industrial utilisation, biotechnology, structure, and properties. New York, E. Horwood. Kobayashi, K., Huang, C. I. and Lodge, T. P. (1999). "Thermoreversible gelation of aqueous methylcellulose solutions." Macromolecules 32(21): 7070-7077. Kolpak, F. J. and Blackwell, J. (1975). "Deformation of Cotton and Bacterial Cellulose Microfibrils." Textile Research Journal 45(7): 568-572. Kono, H., Numata, Y., Erata, T. and Takai, M. (2004). "C-13 and H-l resonance assignment of mercerized cellulose II by two-dimensional MAS NMR spectroscopies." Macromolecules 37(14): 5310-5316. Kono, H., Numata, Y., Nagai, N., Erata, T. and Takai, M. (1999). "CPMAS C-13 NMR and X-ray studies of cellooligosaccharide acetates as a model for cellulose triacetate." Journal of Polymer Science Part a-Polymer Chemistry 37(22): 4100-4107. Krassig, H. (1990). "Cellulose - Morphology, Structure, Accessibility and Reactivity." Papier 44(12): 617-623. Kroon-Batenburg, L. M. J. and Kroon, J. (1997). "The crystal and molecular structures of cellulose I and II." Glycoconiugate Journal 14(5): 677-690. Li, L. (2002). "Thermal gelation of methylcellulose in water: Scaling and thermoreversibility." Macromolecules 35(15): 5990-5998. Li, L., Thangamathesvaran, P. M., Yue, C. Y., Tam, K. C, Hu, X. and Lam, Y. C. (2001). "Gel network structure of methylcellulose in water." Langmuir 17(26): 8062-8068. 33 Liang, C. Y. and Marchessault, R. H. (1959). "Infrared spectra of crystalline polysaccharides. I. Hydrogen bonds in native celluloses." Journal of Polymer Science 37: 385-95. Lu, Y. S., Zhang, L. N. and Pu, X. (2004). "Structure, properties and biodegradability of water resistant regenerated cellulose films coated with polyurethane/benzyl konjac glucomannan semi-IPN coating." Polymer Degradation and Stability 86(1): 51-57. Macosko, C. W. (1994). Rheology : principles, measurements, and applications. New York, NY, VCH. Marrinan, H. J. and Mann, J. (1954). "A study by infrared spectroscopy of hydrogen bonding in cellulose." Journal of Applied Chemistry 4: 204-11. Mccormick, C. L. and Lichatowich, D. K. (1979). "Homogeneous Solution Reactions of Cellulose, Chitin, and Other Polysaccharides to Produce Controlled-Activity Pesticide Systems." Journal of Polymer Science Part C-Polymer Letters 17(8): 479-484. Miyashita, Y., Suzuki, T. and Nishio, Y. (2002). "Miscibility of cellulose acetate with vinyl polymers." Cellulose 9(3-4): 215-223. Nielsen, L. E. (1977). Polymer rheology. New York, M. Dekker. Nishimura, H. and Sarko, A. (1991). "Mercerization of Cellulose .6. Crystal and Molecular-Structure of Na-Cellulose .4." Macromolecules 24(3): 771-778. Nobles, D. R. and Brown, R. M. (2004). "The pivotal role of cyanobacteria in the evolution of cellulose synthases and cellulose synthase-like proteins." Cellulose 11(3-4): 437-448. Payen, A. (1838). Comptes Rendus De L Academie Pes Sciences Serie Iii-Sciences De La Vie-Life Sciences 7: 1052. Pintaric, B., Rogosic, M. and Mencer, H. J. (2000). "Dilute solution properties of cellulose diacetate in mixed solvents." Journal of Molecular Liquids 85(3): 331-350. Pizzoli, M., Scandola, M., Ceccorulli, G. and Pezzin, G. (1985). "Rate of Absorption of Di-N-Butylphthalate in Glassy Poly( Vinyl chloride)." Polymer Communications 26(4): 107-109. Qin, J. J., Li, Y., Lee, L. S. and Lee, H. (2003). "Cellulose acetate hollow fiber ultrafiltration membranes made from CA/PVP 360 K/NMP/water." Journal of Membrane Science 218(1-2): 173-183. Raymond, S., Kvick, A. and Chanzy, H. (1995). "The Structure of Cellulose-Ii - a Revisit." Macromolecules 28(24): 8422-8425. Ross-Murphy, S. B. (1997). "Structure and rheology of gelatin gels." Imaging Science Journal 45(3-4): 205-209. 34 Ross-Murphy, S. B. (1998). Formation, structure and properties of physical networks. Polymer networks: principles of their formation, structure and properties. Stepto, R. F. T. New York, Chapman and Hall: 288-315. Ross-Murphy, S. B. (1998). "Reversible and irreversible biopolymer gels - Structure and mechanical properties." Berichte Der Bunsen-Gesellschaft-Physical Chemistry Chemical Physics 102(11): 1534-1539. Rustemeyer, P. (2004). "Cellulose Acetates: Properties and Applications, Heidelberg (Germany), 2003 - Preface." Macromolecular Symposia 208. Rustemeyer, P. (2004). "History of CA and evolution of the markets." Macromolecular Symposia 208: 1-6. Ryskina, I. I. and Averyano.Vm (1971). "Study of Gelatin and Structure of Cellulose Acetate Cels in Benzyl Alcohol." Vysokomolekulyarnye Soedineniya Section A 13(10): 2189. Ryskina, I. I. and Averyanova, V. M. (1975). "Study of Morphology of 3-Dimensional Network of Cellulose Acetate-Benzyl Alcohol Gels." Vysokomolekulyarnye Soedineniya Seriva A 17(4): 919-922. Saalwachter, K., Burchard, W., Klufers, P., Kettenbach, G., Mayer, P., Klemm, D. and Dugarmaa, S. (2000). "Cellulose solutions in water containing metal complexes." Macromolecules 33(11): 4094-4107. Saka, S. (2004). "Wood as natural raw materials for cellulose acetate production." Macromolecular Symposia 208: 7-28. Sarko, A. and Muggli, R. (1974). "Packing Analysis of Carbohydrates and Polysaccharides .3. Valonia Cellulose and Cellulose-Ii." Macromolecules 7(4): 486-494. Schweizer, E. and Prakt, J. J. (1857). Chemistry 72: 109. Seger, B. and Burchard, W. (1994). "Structure of Cellulose in Cuoxam." Macromolecular Symposia 83: 291-310. Sprague, B. S., Riley., J. L. and Noether, H. D. (1958). Text. Res. J. 28: 275-281. Steinmeier, H. (2004). "Chemistry of cellulose acetylation." Macromolecular Symposia 208: 49-60. Sugiyama, J., Vuong, R. and Chanzy, H. (1991). "Electron diffraction study on the two crystalline phases occurring in native cellulose from an algal cell wall." Macromolecules 24(14): 4168-75. Takai, M. and Colvin, J. R. (1978). "Mechanism of Transition between Cellulose-I and Cellulose-Ii during Mercerization." Journal of Polymer Science Part a-Polymer Chemistry 16(6): 1335-1342. Tang, L. G., Hon, D. N. S. and Zhu, Y. Q. (1996). "Polymorphic trans formations of cellulose acetates prepared by solution acetylation at an elevated temperature." Journal of Macromolecular Science-Pure and Applied Chemistry A33(2): 203-208. 35 Tung, C. Y. M. and Dynes, P. J. (1982). "Relationship between Viscoelastic Properties and Gelation in Thermosetting Systems." Journal of Applied Polymer Science 27(2): 569-574. Tweddle, T. A. and Sourirajan, S. (1978). "Effect of Ethanol-Water Mixture as Gelation Medium During Formation of Cellulose-Acetate Reverse-Osmosis Membranes." Journal of Applied Polymer Science 22(8): 2265-2274. VanderHart, D. L. and Atalla, R. H. (1984). "Studies of microstructure in native celluloses using solid-state carbon-13 NMR." Macromolecules 17(8): 1465-72. Vilgis, T. A. and Winter, H. H. (1988). "Mechanical Selfsimilarity of Polymers During Chemical Gelation." Colloid and Polymer Science 266(6): 494-500. Wang, C. S. and Fried, J. R. (1992). "Viscoelastic Properties of Concentrated Cellulose-Acetate Solutions." Journal of Rheology 36(5): 929-945. Winter, H. H. and Mours, M. (1997). "Rheology of polymers near liquid-solid transitions." Neutron Spin Echo Spectroscopy Viscoelasticity Rheology 134: 165-234. Wolf, R. M., Francotte, E., Glasser, L., Simon, I. and Scheraga, H. A. (1992). "Computation of Low-Energy Crystalline Arrangements of Cellulose Triacetate." Macromolecules 25(2): 709-720. Yamada, H., Kowsaka, K., Matsui, T., Okajima, K. and Kamide, K. (1992). "Nuclear magnetic study on the dissolution of natural and regenerated celluloses into aqueous alkali solution." Cellulose Chemistry and Technology 26(2): 141-50. Zugenmaier, P. (2004). "Characterization and physical properties of cellulose." Macromolecular Symposia 208: 81-166. Zugenmaier, P. and Vogt, U. (1983). "Structural Investigations on Cellulose Tricarbanilate - Conformation and Liquid-Crystalline Behavior." Makromolekulare Chemie-Macromolecular Chemistry and Physics 184(8): 1749-1760. 36 Chapter 2. Effect of Dihydric Alcohol Structure on the Viscoelastic Properties and Microstructure of Cellulose Acetate Solutions 2.1 Abstract The effect of dihydric alcohol (non-solvent) addition on the rheological and micro structural behaviour of a cellulose acetate (CA), 7V,./V-dimethylacetarnide (DMAc) system was investigated. At low non-solvent concentration a homogenous solution formed. At the critical non-solvent concentration, a uniform turbid/cloudy and semi-solid system formed and exhibited gel-like characteristics. Increasing the CA and dihydric alcohol concentration led to enhanced steady shear viscosity and dynamic viscoelastic properties. Changing the dihydric alcohol structure from a 1,2-ethanediol to a 1,4-butanediol to a 1,6-hexanediol increases the moduli and decreases the concentration at which the sol-gel transition occurs. The effect of dihydric alcohol structure on viscoelastic properties can be explained based by Hansen solubility parameters (5d-dispersive, 5p-permanent dipole-dipole, 5h-hydrogen bonding); hydrogen bonding solubility parameter being the main key for the liquid-solid transition. FT-IR spectroscopy clearly showed the formation of strong hydrogen bonds between CA and the dihydric alcohols, where 1,6-hexanediol > 1,4-butanediol > 1,2-ethanediol. 2.2 Introduction Cellulose acetate is a linear homopolymer partially substituted at the carbon-2, 3 and 6 positions of the anhydroglucopyranose units of cellulose (Rustemeyer 2004). The solution properties of CA are influenced by the degree of substitution and distribution of functional groups along the chain (Tsunashima and Hattori 2000; Gomez-Bujedo et al. 2004). CA solution behaviour can be manipulated, and the formation of a physical gel 37 can occur through aggregation induced phase separation (Ryskina and Averyano.Vm 1971; Ryskina and Averyanova 1975). Many parameters effect phase separation, including solvent evaporation, temperature change, non-solvent addition and polymer concentration (Matsuyama et al. 2000; Appaw 2004). In a ternary system, which consists of polymer/solvent/non-solvent, phase separation is influenced by concentration and structure of the non-solvent and polymer (Appaw 2004; Appaw et al. 2007). Depending on the component concentration and structure, phase separation can lead to physical gel formation. Non-solvent induced gelation occurs as a result of physical cross-linking/polymer aggregation and the formation of large macromolecular clusters (Gildert et al. 1979; Chiou et al. 2001; Appaw et al. 2007), which significantly influence the rheological properties and potential end-uses of the polymers (Winter and Mours 1997; Vogrin et al. 2002). Previous work on the mechanism of gel formation for cellulose acetate has demonstrated that CA gels exhibit thermal reversible properties (Ryskina and Averyano.Vm 1971; Ryskina and Averyanova 1975; Kadla et al. 2005). In most CA/solvent systems, gelation occurs after the CA solution is heated to a specific temperature and subsequently cooled. For example, heating solutions of CA/benzyl alcohol to 60-80°C led to gelation when the system was cooled to 25°C (Ryskina and Averyanova 1975). In this system, gel formation is induced by the establishment of strong intermolecular associations. Likewise, intermolecular interactions between CA solutions and coagulation media have been shown to affect membrane formation and microstructure (Hao and Wang 2001). Membrane formation occurs by solvent/non-solvent exchange, and the resulting CA 38 membrane microstructure shown to be dependent on the coagulant strength of the non-solvent (water > methanol > ethanol > isopropanol) (Hao and Wang 2001). While many studies have examined the behaviour of CA in various solvent systems (Pintaric et al. 2000; Bochek and Kalyuzhnaya 2002; Hattori et al. 2002), limited systematic research has been performed to understand the rheological behaviour of CA in mixed solvents. Recently, it was demonstrated that the addition of water to a CA/DMAc solution led to enhanced steady shear viscosity and dynamic viscoelastic properties, and ultimately to phase-separated gel formation (Appaw et al. 2007). The changes in dynamic rheological behaviour of the system during gelation were found to correlate well with the Hansen hydrogen bonding solubility parameter index (8|,) of the solvent system, suggesting hydrogen-bonding interactions as the major factor initiating the sol-gel process. In this chapter, the rheological and microstructural behaviour of a CA, DMAc, dihydric alcohol (non-solvent) system is investigated. Changing non-solvent from 1,2-ethanediol (ED) to 1,4-butanediol (BD) to 1,6-hexanediol (HD) affected the hydrogen bonding characteristics of the system and resulting gel microstructure. The effect on intermolecular interactions and resulting polymer solution/gel network properties were examined using steady shear and dynamic rheology, FTIR and LSCM. 2.3 Experimental Materials and Methods 2.3.1 Materials Cellulose acetate (CA - number average molecular weight (Mn) = 30,000 g/mol and degree of acetylation = 2.5), N,N - dimethylacetamide (DMAc - HPLC grade), 1,2-ethanediol (ED), 1,4-butanediol (BD), 1,6-hexanediol (HD) and calcofluor white were 39 purchased from Sigma-Aldrich and used as received. Sodium acetate buffer (0.1M) was purchased from Fisher Scientific. 2.3.2 Sample Preparation All CA/DMAc/Non-solvents mixtures were prepared from bulk solution as out lined in experimental materials and methods section in Appendix B. 2.3.3 Rheological Measurements Steady and dynamic rheological experiments were conducted with an AR 2000 rheometer (TA Instruments, New Castle, DE). All experiments were performed at 25 °C. Solution viscosities were measured using steady shear experiments with shear rates ranging from 0.1 to 250 s"1, using cone (60mm 2° angle) geometry. Dynamic stress sweep experiments (0.1 to 2000 Pa) were performed to determine the linear viscoelastic (LVE) regime prior to the frequency sweep experiments. The elastic (G') and viscous (G") moduli were determined over the frequency range of 0.1 to 250 rad s"', using cone (liquid) and Parallel plate (gel) geometries. 2.3.4 FTIR Spectroscopy FTIR spectra were recorded on a Perkin Elmer 16PG FTIR spectrometer, using 10 mg sample pressed between ZnSe plates and 16 scans was collected at a resolution of 4 cm"1 over the range of 400 to 4000 cm"1, 2.3.5 Cloud Point Cloud point measurements were performed using a spectrophotometer UV light source operating at 340nm as outlined in experimental materials and methods section in Appendix B. 40 2.3.6 Scanning Electron Microscopy (SEM) SEM analysis of the rigid gels was performed using a Hitachi S-2600 VPSEM with an accelerating voltage of 20kV. Samples were prepared and measured as outlined in experimental materials and method section in Appendix B. 2.3.7 Laser Scanning Confocal Microscopy (LSCM) LSCM experiments were performed with a Chameleon compact ultra fast Ti laser scanning confocal system connected to an inverted microscope (Zeiss Axiovert). Calcofluor white (0.01 wt%) was used as the fluorescent dye and all samples were prepared and measured as outlined in experimental materials and method section in Appendix B. The image analysis was performed using image J analysis soft ware and a LSM image browser. 2.4 Results and Discussion 2.4.1 Solution Behaviour The solution behaviour of the CA ternary system was examined by steady shear rheology. The addition of dihydric alcohol (ED, BD and HD) to the bulk of the CA/DMAc solution led to an enhancement in viscosity. Figure 2.1 shows the effect of increasing BD concentration on the steady state viscosity as a function of shear rate for a 15 wt% CA solution. At low non-solvent (BD) content, a large Newtonian plateau is observed, followed by shear thinning. At a critical non-solvent content (BD > 35 wt%) the system became cloudy, indicating phase separation. Beyond this concentration, the sample exhibited high viscosity at low shear rates and the disappearance of the zero-shear viscosity plateau. In the case of ED and HD, the critical non-solvent content at which the 41 system became cloudy was > 37 wt% and > 33 wt%, respectively (Figure Al and Figure A2 in Appendix). Vi 03 OH, -Vi O o Vi > 10z 10' 10u 10 -2 • BD 20.0 wt.% -o BD 26.6 wt.% • BD 3 1.6 wt.% BD 35.0 wt.% • BD 36.6 wt.% -'oo. ° 0 < > o o o o o o o o o o o o o o * i 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ^ 0 $ 10"' 10u 10' Shear rate (1/s) 10 10 Figure 2.1: Effect of BD content on the steady shear viscosity for a CA/DMAc/BD system at CA 15wt% and a angular frequency of 1 rad s"1. A progressive increase in shear viscosity occurred for all dihydric alcohol non-solvents when the CA content was increased from 10 to 20 wt% (Figure 2.2). The viscosity increased by over an order of magnitude over this range of CA concentrations. The low-shear viscosity increased exponentially with increasing non-solvent content; the exponent being dependent on CA concentration and non-solvent. The same phenomenon was observed for ED and HD (Figure A3 and Figure A4 in Appendix). 42 Table 2.1 lists the exponents, rate of change in viscosity as a function of non-solvent content for the three dihydric alcohols as a function of CA concentration (ranging from 10-20 wt%). Increasing CA concentration increased the rate of change in viscosity disproportionately between the three non-solvents; 1,4-butanediol (BD) exhibited the largest change in the rate of change in viscosity (0.045 - 0.076 Pa s wt.%"1), followed by HD (0.044 - 0.064 Pa s wt%"'), and ED (0.038 - 0.057 Pa s wt%"'). The rate of change in viscosity increases linearly with increasing CA content; the slope increased from 0.18 (ED) to 0.22 (BD) to 0.26 (HD). 43 Table 2.1: Effect of dihydric alcohol content on the rate of change in viscosity for ED, BD and HD systems. Rate of Change in Viscosity (10'2, Pa-s-wt%~') CA (wt%) ED BD HD 10 3.8 4.5 4.4 12.5 4.5 6.0 5.1 ' 15 4.6 6.4 5.7 17.5 4.8 7.2 6.4 20 5.7 7.6 Slope 0.18 0.22 0.26 The increasing viscosity with CA concentration is an indication of increased molecular entanglement in the polymer solution. The entanglements act as temporary cross-links, decreasing the relative motion of the molecules and their ability to flow, and a system that exhibits many of the characteristics of a physically cross-linked polymer network (Grassi et al. 1996; Ross-Murphy 1998; Ikeda and Nishinari 2001). 2.4.2. Phase Separation and the Sol-gel Transition The loss of the Newtonian plateau at low shear rate suggests the development of microstructure within the sample and the intensification of intermolecular interactions (Steiner et al. 1987; Fang et al. 2004). To quantify the sol-gel transition, the elastic (G') and viscous (G") modules were plotted as a function of frequency (Figure 2.3) at three different BD contents for 15 wt% of CA solution. At low BD content (26.6 wt%) G" is larger than G', and the samples are transparent. Both G' and G" exhibit strong frequency dependence, reminiscent of a polymer solution. Increasing the BD content to 36.6 wt% led to a sol-gel transition with both moduli increasing and relatively frequency independent. This dynamic rheological behaviour is an indication of gel formation, in 44 which G' ~ G" (Winter and Mours 1997; da Silva and Areas 2005; Cai and Zhang 2006), and the system is cloudy. At higher BD content G' and G" increased by several orders of magnitude and the material resembled an opaque self-supporting material. The G' was larger than G" over the entire frequency range evaluated, and both are relatively frequency independent, characteristic of a 3-dimensional elastic gel (Nielsen 1977; Ferry 1980; Macosko 1994). The effect of non-solvent content on G' for ED, BD and HD are shown in Figures A5 - A7 (Appendix). c3 OH o b i o 3 i o J 10' 10_ 10 -3 10" .-•88 • G' (26.6 wt.%) o G" (26.6 wt.%) • G1 (36.6 wt.%) • G" (36.6 wt.%) • G' (43.3 wt.%) 0 G" (43.3 wt.%) 10"' 10° 101 10z Angular Frequency (rad/s) 10" Figure 2.3: Elastic (G') and viscous (G") moduli for 15 wt% CA solution at several BD concentrations. Shown is a clear solution, cloudy system, and self-supporting gel. 45 Figure 2.4 demonstrates the effect of increasing the non-solvent and polymer concentrations on G'. As the non-solvent content increased, three distinct phases were observed; an initial region in which there was a slow increase in G' with increasing non-solvent content, an intermediate phase with a sharp increase in G' and concomitant gel formation followed by a slow increase in G'. All CA concentrations showed the same sigmoidal shape, with G' enhancement occurring faster with increasing CA concentration. At the same non-solvent content, the elastic modulus of the high CA concentration was substantially higher than at the lower CA concentration. Moreover, G' increases substantially more in the low CA concentrations upon increasing non-solvent addition, as compared to high CA concentration. 46 -©— C A 1 0 . 0 w t % •B— CA12.5 wt% ^ — C A 1 5 . 0 w l % ^ — CA17.5 wt% - t — CA20.0 wt% k - 1 I I | I I M - f - l - l I I | - t — t - l - t - l I I 1 1 | I H I | M l I 15 20 25 30 35 40 45 50 Non-solvent content (wt%) Figure 2.4: Effect of non-solvent a) ED, b) BD and c) HD content on elastic moduli (G' at 1 rad s"1) at various CA concentrations. 47 The sharp increase in G' corresponds well to the sharp drop in transmission intensity at the lower CA concentration (Figure 2.5). However, as the CA concentration was increased to 15 wt% and beyond, the system appeared cloudy and the drop in transmission intensity occurred well below the sharp increase in G'. Analogous to G', increasing the CA concentration from 10 to 20 wt% resulted in the cloud point shifting to lower non-solvent content. Increasing the non-solvent content caused phase separation and gel formation, as polymer-polymer and polymer-non-solvent interactions were more pronounced (see Figures A8-A10 in Appendix for complete cloud point data for all 3 non-solvent systems). 10 104 103 102 fe 101 b 10' 10" 10": o 10 -3 - e —Tran. 10% - B — Tran. 15% -e—Tran . 20% - • — G' 10% - • — G' 15% - • — G' 20% 10 20 30 BD content (wt.%) 40 100 80 60 H s 40 $ 20 50 Figure 2.5: Effect of non-solvent (BD) content on transmission intensity and elastic moduli (G') for 10, 15 and 20 wt% CA (1 rad s"1). 48 The effect of non-solvent addition on phase separation was also dependent on the type of non-solvent used. Table 2.2 lists the non-solvent content (wt%) at which a 50 % reduction in transmission intensity was observed for the various CA concentrations studied. As with G' (Figure 2.4) phase separation occurred at lower non-solvent contents for HD was compared to BD and ED. Table 2.2: Concentration (wt%) of non-solvent at which a 50% reduction in transmission intensity was observed. Non-solvent (wt%) CA (wt%) ED BD HD 10 38.3 37.5 36.6 12.5 36.6 36.0 35.0 15 35.5 35.0 33.3 17.5 33.3 31.6 29.0 20 31.6 30.0 28.3 It is interesting to note that even in the non-solvent content range in which transmission remained constant (~ 0 - 25 wt%), G' continued to increase with the addition of non-solvent indicating the sensitivity of rheological measurements. The increase in G' may be due to enhanced intermolecular interactions (e.g. hydrogen-bonding) between CA-DMAc-non-solvent. In solution, the hydroxyl (OH) groups on CA probably interact with the carbonyl group in DMAc; analogous to the CA-acetone system (Griswold and Cuculo 1974). With the addition of non-solvent new hydrogen bonds are established between non-solvent and DMAc, and to a much lesser extent CA. The result is an intensification of the hydrogen bonding interactions in solution resulting in the observed increase in 49 viscosity and G'. This complex hydrogen bonding network system was further magnified by increasing the CA concentration from 10 to 20 wt%. 2.3.3. Network Formation and the Gelation Process Figure 2.6 shows the effect of increasing stress amplitude on elastic modulus (G1) having different CA concentrations and non-solvent content (BD). At low stress, G' was relatively independent of the applied stress, which indicates the materials are in the LVE regime. However, with increasing stress, there is a breakdown of interactive bonds in the material and a gradual decrease in G' was observed. At low CA concentration (10 wt%) a step reduction in G' with increasing applied stress is observed over the BD content range of 30 -40 wt% (Figure 2.6a). This phenomenon is characteristic of a weakly cross-linked polymer structure (Winter and Mours 1997), in which there are a minute number of cross-link connections, and a system of finite distribution of macromolecular clusters. Rheologically, the system is a solution or "sol", with G" being larger than G'. At this stage of network formation, the applied stress is sufficient to disrupt these cross-links, but not rupture the system. As a result, the elastic modulus exhibits a temporary drop in the linear viscoelastic region. Beyond the critical stress (xc) the structure ruptures and/or yields to a sharp drop in G' (Appaw et al. 2007). Increasing CA concentration leads to increase polymer-polymer interactions and the creation of a strong network structure, and formation of infinite macromolecular clusters. As a result, the stepwise drop in G' disappeared (Figure 2.6c) (Winter and Mours 1997; Kobayashi et al. 1999; Weng et al. 2004), and the material exhibits an increase in both yield stress, and xc. 50 PH 104 i o J 10/ 10' b i o u 10" 10" i o 4 10J c3 9 10 a 101 i o u cd PH 10" 10 ° 102 10' 10 T T | I I ' ' ' " '"I A A A A A A A A A A A A A A A . A A A A A . A A A A . • l - m - l - l - H i - t - H - t - H - h H ^ , JOOOOOoo^ o BD 28.3 wt.% • BD 31.6 wt.% o BD 36.6 wt.% x BD 40.0 wt.% + BD 41.6 wt.% ^ BD 45.0 wt.% OOcps&XXXnaxixcaxa I I H - I - H - W j I 1 I I I I i l ' l ' , ', I I P,I ' l l A A A A A ^ \ A A A A £ A A A A A A A A A A A A A A A A A I ^ ' A A , ' A A „ t -H - l - l - t -t .,-l-^M-r-H-H-l-H-X>S<XXOOO<>OOOo<x. :x*xs<. o BD 26.6 wt.% D BD 28.3 wt.% » BD 31.6 wt.% x BD 35.0 wt.% BD 36.6 wt.% * BD 40.0 wt.% CCCC«XXXXCGXCCO< A A A A A A A A A / \ A A A A A A A A A A A A A A A A A A / N A A A J ; . "4* W i - r - H - r - H + H - l - H - t - H I I I I I I I I 1-1+ > 0 0 » « < > ! X X A 5 0 < ? C O O O < X > « ^ < > O C < ^ < ^ C ^ ^ •xxx ° BD 23.3 wt.% a BD 26.6 wt.% o BD 30.0 wt.% x BD 31.6 wt.% + BD 333 wt.% A BD 35.0 wt.% iiill I I I I I I I I L L L C O C X O C X X C O C r X X C O ^ O D O C C X X D C C a 10"z 10" 10u 10' 102 Osc. stress (Pa.) 10J 10* Figure 2.6: Stress sweep experiments conducted for a) 10 wt.% CA, b) 15 wt.% CA, and c) 20 wt% CA samples with varying BD contents. 51 03 OH 03 PH 1(T icy i ( r b IO 1 i o u 10" 10 10" ^ 10 b io 10' i o u 103 i o 4 10J b io 10' i o u <kKK<N<)««<fc^<<M<)03<Kra«^ <! r i I I 1-1 U T T T O C O O O c c c a r ' o ED 26.6 (wt%) n ED 28.3 (wt%) o ED 30.0 (wt%) X ED 31.6 (wt%) i - ED 33.3 (wt%) A ED 35.0 (wt%) V ED 36.6 (wt%) ED 38.3 (wt%) < ED 40.0 (wt%) ED 41.6 (wt%) A ED 43.3 (wt%) «KKF3<KKKm=rag<KKl< 1 ' 'H-|-H'-|-H-4.tAaA v^,. BD BD BD BD BD BD BD BD BD BD BD 26.6% 28.3% 30% 31.6% 33.3% 35% 36.6% 38.3% 40% 41.6% 43.3% ( X X C C T X ) O D C C Q X C C O < I I I I I I I I I I -H-H-n-), HD 26.6 (wt%) HD 30.0 (wt%) HD31.6 (wt%) HD 33.3 (wt%) HD35.0 (wt%) HD 36.6 (wt%) HD 38.3 (wt%) HD 40.0 (wt%) 10 10 10 10 10 Osc. stress (Pa) 10J i o 4 Figure 2.7: Stress sweep experiments conducted for 15 wt% CA samples at varying non-solvent contents; a) ED b) BD, and c) HD. 52 The same phenomenon was observed for the ED and HD systems (Figures A11-A13 in Appendix). Figure 2.7 shows the stress sweep results obtained for varying ED (Figure 2.7a), BD (Figure 2.7b) and HD (Figure 2.7c) content at 15 wt% CA concentration. The extent of reduction in G' with increasing applied stress is greatest in ED; the LVE regime reducing by almost an order of magnitude at the lower ED contents. The HD system had the smallest change in modulus in the LVE region. Similarly, the range of non-solvent content at which the step-wise reduction in LVE was observed was also greatest for ED, -10 wt% (-26-36 wt% ED) > BD, -8 wt% (-28-36 wt% BD) > HD, -4 wt% (-31-25 wt% HD). Physical gels consist of extensive macromolecular clusters, held together by complex molecular interaction mechanisms (Winter and Mours 1997; Ross-Murphy 1998; Eissa and Khan 2005). In the ternary system examined here, CA is initially dispersed in DMAc and specific intra- and intermolecular interactions form. The addition of non-solvent leads to the formation of new interactions between the non-solvent and DMAc, as well as the non-solvent and CA. Sample preparation involved heating the samples at 100 °C, which as in a CA/DMAc/water system (Appaw 2004), disrupts the CA/DMAc and CA/CA interactions, enabling the formation of the new interactions with the non-solvent. The non-solvents (dihydric alcohols) used here can act as both hydrogen bond donors and acceptors, interacting with both the hydroxyl groups and acetyl carbonyl groups in CA. Figure 2.8 shows the hydroxyl stretching region (V0H) of the FTIR spectra for the various components in one of the ternary systems employed in this study. 53 C A / D M A c / B D 26wt% viscous solution) 3700 3500 3300 3100 Wave number (cm1) Figure 2.8: FTIR analysis of the various components in the CA/DMAc/BD ternary system. Included in the figure are the FTIR spectrum of viscous solution and gel samples. FTIR spectroscopy has proven to be a highly effective means of investigating specific interactions within and between molecules (Joesten and Drago 1962; Drago and Vogel 1992). FTIR can be used to qualitatively and quantitatively study the mechanism of intermolecular interaction through hydrogen-bonding. Spectral differences existed between the CA/DMAc and DMAc/BD band envelopes associated with the hydroxyl stretching region (VOH ~3700 - 3000 cm"1). A clear difference in band envelope shape and wavenumber position can be seen in Figure 2.8. The CA/DMAc hydroxyl stretching region was broader with a clear bimodal distribution. In both, a broad band centre and a shoulder were observed. The broad centre band at -3478 cm"1 (CA/DMAc) and -3402 cm"1 (DMAc/BD) and the shoulder at -3290 cm"1 can be assigned to the average 54 stretching of intermolecular hydrogen bonding (Kubo and Kadla 2005). Based on the band envelope shape and wavenumber of the major peak/shoulder, it is evident that stronger hydrogen bonds exist between DMAc and BD than with CA. Moreover, the FTIR spectrum of the CA/DMAc/BD viscous solution (26 wt% BD) clearly showed the presence of very strong hydrogen bonds between CA and BD, significantly stronger than those between CA and DMAc, and DMAc and BD. Furthermore, increasing the non-solvent concentration from 26 to 41 wt% further increased the intermolecular hydrogen bonding, and the hydroxyl stretching band shifted to lower wavenumber. The intensification of intermolecular hydrogen bonding in the system after addition of BD supports the increased viscosity and modulus discussed above. The same phenomenon was observed for both ED and HD. (Figures A14 and A15 in Appendix) and will be discussed further in the next section. As illustrated in Figure 2.7, polymer concentration also plays a key role in network formation and inducing the sol-gel transition (Ross-Murphy 1998; Fang et al. 2004; Cai and Zhang 2006). Increasing CA concentration increase the intermolecular interactions between CA chains as well as the hydrogen-bonding between CA/non-solvent. Figure 2.9 shows the change in the hydroxyl stretching region for CA/DMAc/BD solutions at 10, 15 and 20 wt% CA concentrations. Although not as pronounced as that resulting from increasing non-solvent content (Figure 2.8), increasing CA concentration did shift the V Q H band envelope to a slightly lower wavenumber. 55 3700 3500 3300 3100 Wavenumber (cm-1) Figure 2.9: FTIR analysis of the CA/DMAc/BD ternary system at 10, 15 and 20 wt% CA concentration and constant BD content (26 wt%). Also included is the FTIR spectrum of the 15 wt% CA/DMAc/BD (41.6 wt%) gel sample. CA concentration also affected the amount of non-solvent required to induce gel formation. Figure 2.10 shows the concentration of non-solvent (BD) at which G' crosses G" or G' is very close to G" for CA concentrations of 10, 15 and 20 wt%. The crossover of G'-G" during the liquid-solid transition is an indication of the development of a cross-linked structure and formation of a weak gel (Winter and Mours 1997; Chauvelon et al. 2003; Avanza et al. 2005; da Silva and Areas 2005; Cai and Zhang 2006). These cross-linked weak gels were made by controlling the non-solvent content and measured using plate geometry. As anticipated, the higher CA content (20 wt%) sample had a higher elastic/viscous moduli than at lower polymer contents, increasing an order of magnitude over the 10 wt% CA concentration sample. The 20 wt% CA concentration system also required less BD content to obtain the transitional state; 41.3% wt% BD at 10 wf% CA 56 versus 31.3 wt% BD at 20 wt% CA concentration. Thus, in addition to hydrogen bonding with the non-solvent, the gelation process was strongly influenced by polymer/polymer interactions with increasing CA concentration increasing the elastic modulus and microstructure development. 104 PH o b 10 10z • G ' CA(10)BD(41.3) O G"CA(10)BD(41 .3 ) • G ' CA(15)BD(36.3) • G"CA(15) BD(36.3) • G ' CA(20)BD(31.3) O G"CA(20) BD(31.3) - r r q i | <>•* 10' 10 -2 10" 10u 10' 10z 10J Angular frequency (rad s" ) Figure 2.10: Viscous (G") and elastic (G') moduli at or near the gel point for 10, 15 and 20 wt% CA concentration at various BD contents (numbers in parentheses are weight percentage). A similar effect on the sol-gel transitions was observed for the HD and ED systems. However, the concentration of non-solvent required for gel formation and the modulus of the resulting gels were very much dependent on non-solvent structure. Figure 2.11 shows 57 the difference in non-solvent content between ED, BD and HD samples at 15 wt% CA concentration where G' was very close to G". As expected, based on the rheological results discussed above, the modulus of the HD system at this intermediate state is the largest followed by BD and ED. Likewise, the amount of HD required to attain this state is slightly less than that for BD, which was again less than that required in ED system. PH o b 10" 10 : • G' ED(37.3) - o G" ED( 37.3) • G' BD(36.3) • G" BD( 36.3) • G' HD(35.6) 0 G" HD(35.6) 10z 10' 10 -2 (35.  %tt 0 o ° « H l " o ° I C « 0 0 • • o i # o o 10 10 10 10 Angular frequency (rad s'1) 10J Figure 2.11: G' and G" frequency sweep spectra for ED, BD and HD as non-solvents at an intermediate state at 15 wt% CA. The differences in solution properties and phase behaviour observed between HD, BD and ED are probably very much dependent on intermolecular interactions of the solvent/non-solvent and CA. As the strength of intermolecular forces is equal to the cohesive energy density (CED), CED values can be used to predict solubility and solution 58 behaviour. Typically, Hildebrand solubility parameters (5), which are the square root of CED are used. For mixtures, the solubility parameter is often taken as the sum of the products of the component solubility parameters and their volume fractions ($): 8 • = Y#A mix / i l TI i In polar systems, the free energy change of mixing is dominated by hydrogen bonding forces between the various groups in the polymer and solvent (Hansen 1967; Burrell 1968). Generally, complete miscibility is expected if the degree of hydrogen-bonding and the solubility parameters are similar between components. In a completely thermodynamic solution, the strength of the intermolecular forces must be equal the cohesive energy density. Hansen (1967) assumed that the cohesive energy arises from dispersive (8d), permanent dipole-dipole (8P) and hydrogen bonding (8|,) forces where: This three-term set represents different contributions to the free energy of mixing and presents improved agreement between 8 and the experimental data. Table 2.3 lists the values of these terms for the various components of the ternary system used here. 59 Table 2.3: Hansen solubility parameters for the components of a ternary system (8d -dispersive, 5P - permanent dipole-dipole, 8h - hydrogen-bonding) Solubility parameters (MPa1/2) 8 5h CA 25.1 18.6 12.7 11 DMAc 22.7 16.8 11.5 10.2 ED 34.2 17.2 12.6 26.7 BD 28.3 16.9 7.9 21.2 HD 25.1 16.5 5.7 18 The impact of the different solubility parameter components on the modulus for BD at 15wt% CA concentration is shown in Figure 2.12. The same trend of individual solubility parameters was observed for the ED and HD systems (Figures A16 and A17 in Appendix). The hydrogen bonding solubility parameter index (8h) was significantly affected, while both 8d and 8P remained relatively constant. This implies that the hydrogen-bonding interactions are the major route for initiating the sol-gel process in this system (Appaw et al. 2007). Accordingly, the addition of a dihydric alcohol to DMAc led to an increase in the solvent/non-solvent solubility parameter (Table 2.3). 60 17 03 OH 03 o3 o cn 4 - * c <u j> "3 00 16 15 14 13 12 10 10u 10 10 10 G' (Pa) •—• _J i i 10" 10 Figure 2.12: Influence of the individual solubility parameter indexes on G' at a fixed frequency of 1 rad s_1 (15 wt% CA and BD as non-solvent). Figure 2.13 illustrates G' as a function of hydrogen bonding solubility parameter (8h) for all three non-solvents. As expected (Table 2.3), the calculated hydrogen bonding solubility values are ranked such that 8h ED > 8h BD > 8h HD, and suggest a role for hydrogen bonding in initiating the sol-gel process is more significant in ED, than BD, which is more significant than HD. 61 16.5 13.5 1 1 10"1 10° 101 102 103 104 105 G' (Pa) Figure 2.13: Effect of hydrogen bonding solubility parameters on the elastic modulus G' of ED, BD and HD ternary systems at 15 wt% CA concentration. This is supported by FTIR analysis, where the ED ternary system exhibited stronger hydrogen bonding with CA than BD or HD (Figure 2.14). The three non-solvents had identical hydroxyl stretching region absorbance profiles for the DMAc/non-solvent solutions. However, the V 0 H band envelopes changed dramatically upon inclusion of CA. As illustrated in Figure 2.14, FTIR analysis revealed a stronger hydrogen-bonding system existed in the ED ternary system, followed by BD and HD. 62 3700 3500 3300 3100 Wavenumber (cm1) Figure 2.14: FTIR spectra of the hydroxyl stretching region of a) CA/DMAc/non-solvent, and b) DMAc/non-solvent solutions at 15 wt.% CA concentration and 26.6 wt.% non-solvent content. 2.3.4 Gel Rheology and Microstructure Figure 2.15 shows the effect of non-solvent (BD) content on G' as a function of CA concentration (Figures A5-A7 in Appendix). At the higher BD contents, all samples exhibited gel-like features with G' > G", and both moduli exhibiting very weak frequency dependence. 63 ca PH o C b Angular frequency (rad s~ ) Figure 2.15: Elastic and viscous moduli of CA/DMAc/BD gels as a function of frequency at 15 wt % CA concentration. Figure 2.16 shows the effect of non-solvent content (or volume fraction) on the elastic modulus of the gels. It can be seen that G' exhibited a power-law behaviour with non-solvent content (<£), G'~ 0 ", suggesting the gels were fractal in nature (Eissa and Khan 2005). The power-law dependence of G' was different between non-solvents, with the power-law exponent increasing from 23 for ED to 30 for BD and to 75 for HD. This suggests. differences in microstructure between the gels containing the different non-solvents. 64 y = 1 . 8 x l ( r 3 5 - x 2 3 R2=0.97 (ED) -y = 5 . 7 x l O - 4 5 - x 3 0 R2=0.94 (BD) -y = 6.8x10" 1 1 6 • x 7 5 R2=0.99 (HD) 36 38 40 42 44 Non-solvent (wt%) Figure 2.16: Elastic modulus of CA/DMAc/non-solvent gels as a function of non-solvent content (G' obtained at frequency of 1 rad s"1). The power-law or scaling behaviour of polymeric gels has been well studied, and has been proven to be very successful in describing the elastic properties of polymeric gels. The basic concept is to relate the elastic properties of the gels to its network structure. According to scaling theory, the structure of the polymeric gel network is considered as a collection of close-packed blobs or floes of polymers. These floes are considered to be fractal objects that they are at least partially self-similar. This means that a part of the fractal is identical to the entire fractal itself except smaller. Fractality is a powerful means of quantifying the structure of non-Euclidean objects, and the fractal dimension is a simple means of capturing the complexity of a structure's geometry in a single number. In the case of polymer aggregation-induced gelation, the fractal dimension can provide 65 information on the mechanism of aggregation and gelation (Shih et al. 1990). That is, whether the system involves fast aggregation as per the cluster-cluster aggregation model (CLCLA), or slow aggregation as per the reaction-limited aggregation model (RLA). Rheology has been used to determine the fractal dimensions of gels (Shih et al. 1990). In this approach, the interaction within and between floes of aggregated macromolecules, which interconnect to form a 3-dimensional gel network, and can be classified as either strong-linked or weak-linked systems. In strong linked systems, the links between floes are stronger than those within the floes, and as a result, failure under deformation occurs through breaking intra-floc interactions. As such, this leads to a decrease in the LVE regime with increasing sample concentration. By contrast, weak-linked gels possess strong intra-floc links, more rigid than the inter-floc linkages, and an increase in the limit of linearity occurs with increasing concentration. Figure 2.17 shows the effect of increasing non-solvent and CA concentration on the elastic modulus (G') as a function of strain. It is evident that the onset of nonlinearity shifted to lower value with increasing concentration, indicating the CA/DMAc/dihydric alcohol gels were strongly-linked. The same trend was observed for ED and HD (Figures A18 and A19 in Appendix). 66 03 PH 10D 104 10J i o z O lO1 10u 10" c3 PH I O H 10 10z O 101 10u 10" 10" ~i r rrrfmTTtitiiiiaDiiiiiDiinTn-o BD(41.6wt%) • BD (40.0wt%) o BD (36.6wt%) A BD (35.0wt%) H 1 h XX<«<«>!»«xxXXXX»«^^ •XX o CA(10.0wt%) a CA(12.5wt%) <> CA(15.0wt%) X CA(17.5wt%) + CA(20.0wt%) 10"6 10"5 10"4 10"3 10"2 10"' 10° 101 102 Strain % Figure 2.17: Elastic modulus of CA gels at different a) BD and b) CA concentrations as a function of strain. The limit of linearity shifts to lower strain as concentration increases. Figure 2.18 depicts the effect of non-solvent concentration on the on-set point of nonlinearity in G' for ED, BD and HD (the on-set point is defined as intersection between the tangent of the drop in G' and the LVE regime). All three were classified as strongly-linked gels (as discussed above), but the rate of change in nonlinearity differs between 67 the non-solvents. The ED system showed the largest change in non-linearity, i.e. smallest slope, followed by BD, then HD (ED = -0.80; BD = -1.09; HD = -1.26). This implies that the ED system may contain relatively stronger inter-floe link bonds than those in BD, which are stronger than in HD. This is consistent with the FTIR results presented in Figure 2.14. Figure 2.18: Change in onset point of nonlinearity of elastic modulus for ED, BD and HD systems (CA15 wt%). In strong-linked gels G' scales as G' ~ O ( x ) ( \ where x is the back bone fractal dimension of the floes, d is the Euclidean dimension, and D is the gel fractal dimension. Typically x ranges between 1 to 1.3 for colloidal gels (Wu and Morbidelli 2001). We arbitrarily chose d=1.15, as the range of x only differs by 0.3 (1 - 1.3) and changes the 68 values of D -2%. The corresponding fractal dimensions D, for the system here ranged from 2.90 in ED to 2.78 for HD (Table 2.4). Table 2.4: Fractal dimensions of CA/DMAc/non-solvent gels at 15 wt% CA concentration, as obtained by rheology and confocal microscopy. ED BD HD by confocal microscopy 1.92±0.02 1.92±0.03 1.92±0.02 by rheology 2.90±0.06 2.86±0.04 2.78±0.03 An alternative approach to determining fractal dimension, as well as directly visualizing gel microstructure is through confocal microscopy. Confocal images are used to extract information on the microstructure of gels, membrane pore dimensions and their distribution. LSCM provides 2D-imaging at various depths without sample processing, enabling the direct study of the actual sample (Charcosset and Bernengo 2000; Pugnaloni et al. 2005). Optical sectioning eliminates artefacts that might occur through physical sectioning and sample preparation. Figure 2.19 shows the LSCM images obtained for the various non-solvents at the same elastic modulus and CA concentration (10 wt%). 69 E D B D H D Figure 2.19: Micrograph images of gels at the same elastic modulus (G') and at the same CA concentrations (10 wt.%) for ED, BD and HD systems. From top to bottom are LSCM in fluorescent mode (top), LSCM images in reflective mode (middle), and SEM images at 3000x magnification (bottom). The LSCM images show the presence of a denser network structure as the non-solvent in the ternary system is varied from ED to BD and to HD, respectively. The texture of the HD gel appeared to be a denser with a more uniform structure. The fractal dimensions of the gels were calculated from the images using the box counting technique using Image J 70 software (Table 2.4) (Shih et al. 1990; Hagiwara et al. 1997; Eissa and Khan 2005). Using this technique fractal dimensions of approximately 1.92 were obtained from the scaling relation N QC r ~ D, where N is the number of boxes filled with CA floes, r is the size of the square box, and D is the fractal dimension of the CA network. Regardless of the non-solvent, or CA concentration the fractal dimension of the gels appeared to be the same. It has been reported that the value of the fractal dimension can vary enormously if there is a systematic departure from linear behaviour (Marangoni 2002). According to Alejandro Marangoni (2002), mass fractal dimensions determined by confocal microscopy technique belong to 2-dimensional Euclidean space, while that determined by rheological experiments translate to a 3-dimensional Euclidean space. Therefore, these two different techniques provide different values for the fractal dimension. However, for fractal scaling theories to be theoretically valid, there must be agreement between the fractal dimensions determined by rheological methods and microscopy methods. Also shown in Figure 2.19 (bottom) are SEM images of the CA materials formed after solvent/non-solvent removal. Significant differences in microstructure between the three non-solvent systems are evident. The HD system appeared to have much larger CA aggregated clusters as compared to that of BD and ED. This is consistent with the LSCM reflection images, wherein the HD samples had larger CA domains as indicated in red (in the fluorescent images the colour portions represent the calcofluor-tagged CA domains), while in the reflective images the red areas corresponds to the actual structure of the CA chains (Charcosset and Bernengo 2000; Charcosset et al. 2000). 71 CA 10wt.% CA 15wt.% CA 20wt.% Figure 2.20: Micrograph images of CA/DMAc/BD gels at the same elastic modulus (G') and three different CA concentrations (10, 15 and 20wt.%), From top to bottom are LSCM in fluorescent mode (top), LSCM images in reflective mode (middle), and SEM images at lOOOx magnification (bottom). Figure 2.20 illustrate the LSCM and SEM micrographs of CA/DMAc/BD gels obtained at the same elastic modulus (G') and different CA concentrations (10, 15 and 20 wt%). With increasing CA concentration, gel uniformity increased and a more homogenous 72 microstructure with smaller voids was formed. This is similar to the changes in microstructure observed in changing from ED to BD to HD (Figure 2.19). The increased density and uniformity is consistent with the enhanced moduli observed in the rheological studies, and accounts for the stronger gels. Again, SEM images of the isolated CA networks showed the same enhancement in the uniformity of microstructure with increasing CA concentration. The more tightly packed voids with higher CA content provided a more rigid and compact gel network, stemming from additional polymer/polymer interactions. These images reveal better uniform packing in polymer network structure as CA concentration increased. The agreement between the LSCM reflective images and SEM images is remarkable. 2.4 Conclusion The effect of dihydric alcohol (ED, BD and HD) addition to bulk CA/DMAc solutions was investigated using steady-state and dynamic rheology. The addition of these non-solvents to DMAc/CA solutions resulted in a change in the inter-molecular interactions in the system. Increasing non-solvent addition intensifies the inter-molecular interactions, and the viscosity of the system increases dramatically. At low non-solvent content, a large Newtonian plateau was observed followed by shear thinning. Increasing non-solvent addition increased the viscosity at low shear rates, and non-Newtonian behaviour was observed. At a critical non-solvent concentration, which differed between non-solvents, i.e. HD<BD<ED, a sol-gel transition occurred. Increasing the CA concentration accelerated the sol-gel transition and gel formation. The elastic modulus of the ternary system increased several orders of magnitude, and a more uniform structure was obtained with the higher polymer content. This is likely the result of increasing inter-molecular 73 interactions (physical entanglement and hydrogen bonding) as result of higher CA aggregation. Increasing the non-solvent and CA content intensified the intermolecular hydrogen bonding in the system, and led to phase separation. FTIR analysis showed that the hydroxyl stretching band envelope shifted to a lower wave number with increasing non-solvent content, and to a lesser extent polymer concentration. This is an indication of the development of stronger intermolecular hydrogen-bonding interactions between components. The stronger intermolecular interaction between non-solvent and CA led to CA aggregation and the formation of a strong physical gel. The changes in the elastic modulus during gelation correlated well with the combined solubility parameter (8) and in particular, the Hansen hydrogen bonding solubility parameter index (8h) of the solvent system, indicating hydrogen-bonding interactions are likely the major factor initiating the sol-gel process. The calculated hydrogen bonding solubility value (Sh) showed that hydrogen-bonding interactions are strongest in ED, such that ED>BD>HD. This was confirmed by FTIR, wherein the shift in wavenumber associated with the hydroxyl band envelope was greatest in the ED system and followed the same trend, ED>BD>HD. Rheological measurement revealed that the system was fractal in nature, and the gels were "strong-linked", where interfloc linkages are stronger than the intrafloc linkages. This means that failure under deformation occurs through breaking intrafloc linkages. Again, the ED system clearly showed that it contained stronger interfloc linkages than BD, which were stronger than HD. Interestingly, the concentration of non-solvent required to induce network formation and gelation was greatest for ED and lowest for HD. This suggests that one factor affecting network formation may be physical size of 74 the non-solvent. The HD molecules are larger than BD and ED, and as such, fewer molecules may be required to "bridge" or link CA chains. Moreover, based on solubility parameters, HD likely has a stronger affinity for CA, while ED does not, inferring the self-association of ED molecules and the formation of larger ED assemblies within the system. This may explain the lower content of HD, which is required to form a gel. In addition, the hydrogen bonds between non-solvents and with DMAc are weaker than those with CA. Therefore, it is likely that upon deformation, failure would occur between non-solvent/DMAc and non-solvent/non-solvent, before non-solvent/CA. Thus, as more ED/ED and ED/DMAc linkages are involved in forming the interfloc linkages than in the case of HD, they serve as the weak point in the gel network. Support for the concept that larger ED clusters were involved in linking the CA chains can be obtained from LSCM images. Imaging of gels from different non-solvent systems revealed a denser and more uniform network formed in the HD system as compared to BD and ED. In fact, LSCM analysis of the ED system revealed very large "voids" within the network. Increasing non-solvent content reduced the size of the "voids" and increased the overall uniformity of the network. Likewise, there was a positive correlation between CA concentration and the density and uniformity of polymer network structure. 75 2.5 References Appaw, C. (2004). Rheology and Microstructure of Cellulose Acetate in Mixed Solvent Systems. PhD Thesis; Chemical Engineering. Raleigh, NC, North Carolina State University: 224. Appaw, C, Gilbert, R. D., Khan, S. A. and Kadla, J. F. (2007). "Viscoelastic Behavior of Cellulose Acetate in a Mixed Solvent System." Biomacromolecules 8(5): 1541-1547. Avanza, M., Puppo, M. C. and Anon, M. C. (2005). "Rheological characterization of amaranth protein gels." Food Hvdrocolloids 19(5): 889-898. Bochek, A. M. and Kalyuzhnaya, L. M. (2002). "Interaction of water with cellulose and cellulose acetates as influenced by the hydrogen bond system and hydrophilic-hydrophobic balance of the macromolecules." Russian Journal of Applied Chemistry 75(6): 989-993. Burrell, H. (1968). "Challenge of Solubility Parameter Concept." Journal of Paint Technology 40(520): 197. Cai, J. and Zhang, L. (2006). "Unique gelation behavior of cellulose in NaOH/Urea aqueous solution." Biomacromolecules 7(1): 183-189. Charcosset, C. and Bernengo, J. C. (2000). "Comparison of microporous membrane morphologies using confocal scanning laser microscopy." Journal of Membrane Science 168(1-2): 53-62. Charcosset, C, Cherfi, A. and Bernengo, J. C. (2000). "Characterization of microporous membrane morphology using confocal scanning laser microscopy." Chemical Engineering Science 55(22): 5351-5358. Chauvelon, G., Buleon, A., Thibault, J. F. and Saulnier, L. (2003). "Preparation of sulfoacetate derivatives of cellulose by direct esterification." Carbohydrate Research 338(8): 743-750. Chiou, B. S., Raghavan, S. R. and Khan, S. A. (2001). "Effect of colloidal fillers on the cross-linking of a UV-curable polymer: Gel point rheology and the Winter-Chambon criterion." Macromolecules 34(13): 4526-4533. da Silva, M. A. and Areas, E. P. G. (2005). "Solvent-induced lysozyme gels: Rheology, fractal analysis, and sol-gel kinetics." Journal of Colloid and Interface Science 289(2): 394-401. Drago, R. S. and Vogel, G. C. (1992). "Interpretation of Spectroscopic Changes Upon Adduct Formation and Their Use to Determine E-Parameter and C-Parameter." Journal of the American Chemical Society 114(24): 9527-9532. Eissa, A. S. and Khan, S. A. (2005). "Acid-induced gelation of enzymatically modified, preheated whey proteins." Journal of Agricultural and Food Chemistry 53(12): 5010-5017. Fang, Y. P., Takahashi, R. and Nishinari, K. (2004). "Rheological characterization of schizophyllan aqueous solutions after denaturation-renaturation treatment." Biopolymers 74(4): 302-315. Ferry, J. D. (1980). Viscoelastic Properties of Polymers. New York ; Toronto Wiley, 1980. 76 Gildert, G. R., Matsuura, T. and Sourirajan, S. (1979). "Effect of Different Alcohol-Water Mixtures as Gelation Media During Formation of Cellulose-Acetate Reverse-Osmosis Membranes." Journal of Applied Polymer Science 24(1): 305-310. Gomez-Bujedo, S., Fleury, E. and Vignon, M. R. (2004). "Preparation of cellouronic acids and partially acetylated cellouronic acids by TEMPO/NaCIO oxidation of water-soluble cellulose acetate." Biomacromolecules 5(2): 565-571. Grassi, M., Lapasin, R. and Pricl, S. (1996). "A study of the rheological behavior of scleroglucan weak gel systems." Carbohydrate Polymers 29(2): 169-181. Griswold, P. D. and Cuculo, J. A. (1974). "Experimental Study of Relationship between Rheological Properties and Spinnability in Dry Spinning of Cellulose Acetate-Acetone Solutions." Journal of Applied Polymer Science 18(10): 2887-2902. Hagiwara, T., Kumagai, H., Matsunaga, T. and Nakamura, K. (1997). "Analysis of aggregate structure in food protein gels with the concept of fractal." Bioscience Biotechnology and Biochemistry 61(10): 1663-1667. Hansen, C. M. (1967). "Solubility Relationships in More Recent Polymers." Svensk Kemisk Tidskrift 79(11): 627. Hao, J. H. and Wang, S. C. (2001). "Calculation of alcohol-acetone-cellulose acetate ternary phase diagram and their relevance to membrane formation." Journal of Applied Polymer Science 80(10): 1650-1657. Hattori, K., Cuculo, J. A. and Hudson, S. M. (2002). "New solvents for cellulose: Hydrazine/thiocyanate salt system." Journal of Polymer Science Part a-Polymer Chemistry 40(4): 601-611. Ikeda, S. and Nishinari, K. (2001). ""Weak gel"-type rheological properties of aqueous dispersions of nonaggregated kappa-carrageenan helices." Journal of Agricultural and Food Chemistry 49(9): 4436-4441. Joesten, M. D. and Drago, R. S. (1962). "Validity of Frequency Shift-Enthalpy Correlations .1. Adducts of Phenol with Nitrogen and Oxygen Donors." Journal of the American Chemical Society 84(20): 3817-&. Kadla, J. F., Dai, Q. Z. and Khan, S. A. (2005). "Viscoelastic and microstructural changes in sol-gel systems of cellulose acetate in a mixed solvent system." Abstracts of Papers of the American Chemical Society 229: U292-U292. Kobayashi, K., Huang, C. I. and Lodge, T. P. (1999). "Thermoreversible gelation of aqueous methylcellulose solutions." Macromolecules 32(21): 7070-7077. Kubo, S. and Kadla, J. F. (2005). "Hydrogen Bonding in Lignin: A Fourier Transform Infrared Model Compound Study." Biomacromolecules 6(5): 2815-2821. Macosko, C. W. (1994). Rheology : principles, measurements, and applications. New York, NY, VCH. Marangoni, A. G. (2002). "The nature of fractality in fat crystal networks." Trends in Food Science & Technology 13(2): 37-47. Matsuyama, H., Nishiguchi, M. and Kitamura, Y. (2000). "Phase separation mechanism during membrane formation by dry-cast process." Journal of Applied Polymer Science 77(4): 776-783. Nielsen, L. E. (1977). Polymer rheology. New York, M. Dekker. 77 Pintaric, B., Rogosic, M. and Mencer, H. J. (2000). "Dilute solution properties of cellulose diacetate in mixed solvents." Journal of Molecular Liquids 85(3): 331-350. Pugnaloni, L. A., Matia-Merino, L. and Dickinson, E. (2005). "Microstructure of acid-induced caseinate gels containing sucrose: Quantification from confocal microscopy and image analysis." Colloids and Surfaces B-Biointerfaces 42(3-4): 211-217. Ross-Murphy, S. B. (1998). "Reversible and irreversible biopolymer gels - Structure and mechanical properties." Berichte Der Bunsen-Gesellschaft-Physical Chemistry Chemical Physics 102(11): 1534-1539. Rustemeyer, P. (2004). "Cellulose Acetates: Properties and Applications, Heidelberg (Germany), 2003 - Preface." Macromolecular Symposia 208. Ryskina, I. I. and Averyano.Vm (1971). "Study of Gelatin and Structure of Cellulose Acetate Cels in Benzyl Alcohol." Vysokomolekulyarnye Soedineniya Section A 13(10): 2189-&. Ryskina, I. I. and Averyanova, V. M. (1975). "Study of Morphology of 3-Dimensional Network of Cellulose Acetate-Benzyl Alcohol Gels." Vysokomolekulyarnye Soedineniya Seriva A 17(4): 919-922. Shih, W. H., Shih, W. Y., Kim, S. I., Liu, J. and Aksay, I. A. (1990). "Scaling Behavior of the Elastic Properties of Colloidal Gels." Physical Review A 42(8): 4772-4779. Steiner, W., Lafferty, R. M., Gomes, I. and Esterbauer, H. (1987). "Studies on a Wild Strain of Schizophyllum-Commune - Cellulase and Xylanase Production and Formation of the Extracellular Polysaccharide Schizophyllan." Biotechnology and Bioengineering 30(2): 169-178. Tsunashima, Y. and Hattori, K. (2000). "Substituent distribution in cellulose acetates: Its control and the effect on structure formation in solution." Journal of Colloid and Interface Science 228(2): 279-286. Vogrin, N., Stropnik, C, Musil, V. and Brumen, M. (2002). "The wet phase separation: the effect of cast solution thickness on the appearance of macrovoids in the membrane forming ternary cellulose acetate/acetone/water system." Journal of Membrane Science 207(1): 139-141. Weng, L. H., Zhang, L. N., Ruan, D., Shi, L. H. and Xu, J. (2004). "Thermal gelation of cellulose in a NaOH/thiourea aqueous solution." Langmuir 20(6): 2086-2093. Winter, H. H. and Mours, M. (1997). "Rheology of polymers near liquid-solid transitions." Neutron Spin Echo Spectroscopy Viscoelasticity Rheology 134: 165-234. Wu, H. and Morbidelli, M. (2001). "A model relating structure of colloidal gels to their elastic properties." Langmuir 17(4): 1030-1036. 78 Chapter 3. Effect of Hydrophobic and Hydrophilic Interactions on the Rheological Behaviour and Microstructure of a Ternary Cellulose Acetate System 3.1 Abstract The effect of component composition on the rheological and microstructural behaviour of cellulose acetate (CA) in a ternary CA, 7V,A/-dimethylacetamide (DMAc), non-solvent (alcohol) system was examined. Increasing the CA and non-solvent concentration increased the viscosity and dynamic viscoelastic properties of the system. A sol-gel transition was observed at a critical non-solvent concentration, which appeared to be dependent on the structure of the non-solvent. Increasing the available hydrogen-bonding groups within the non-solvent led to higher modulus (stronger gels) and a lower concentration sol-gel transition. Increasing the hydrophobic component of the non-solvent also enhanced gel properties, although not to the same extent. Using dynamic rheology, FTIR spectroscopy and laser scanning confocal microscopy the effect of hydrogen-bonding and hydrophobic interactions between CA, DMAc and non-solvents were probed. Although hydrophobic interactions play a role in gelation, it appeared that gel properties are greatly influenced by competitive hydrogen bonding between system components. The presence of intramolecular hydrogen bonds within the non-solvent delays the sol-gel transition and decreases the extent of intermolecular hydrogen bonding. Furthermore, the presence of proton donor-acceptor groups in the non-solvent is necessary for gel formation. 79 3.2 Introduction Cellulose is the most abundant natural biopolymer, and along with its derivatives is extensively used in textiles, cosmetics, packaging materials, films, membranes, and engineered thermoplastics (Gross and Scholz 2001; Saalwachter and Burchard 2001). However, the use of cellulose is limited by its poor solubility in most solvents, and the fact that cellulose decomposes prior to melting (Saalwachter and Burchard 2001). As a result, cellulose is derivatized, typically esterified to facilitate processing (Tsunashima and Hattori 2000). Among the many cellulose esters, cellulose acetates are most widely and commercially used. Cellulose acetates exhibit different solubility patterns depending on their degree of acetylation (DS). For example, CA with a DS between 0.5-1 is soluble in aqueous solutions, while those with DS>1 tend to be insoluble in aqueous medium, but soluble in many organic solvent systems (Bochek and Kalyuzhnaya 2002; Gomez-Bujedo et al. 2004). Cellulose acetate dissolution and solution properties are dependent on interactions between specific functional groups within the solvent and polymer. For example, it is postulated that specific intermolecular interactions, such as hydrogen bonding are responsible for the observed differences in the rheological behaviour of CA in different solvents (Bochek and Kalyuzhnaya 2002) and solvent/non-solvent systems (Appaw 2004). It has previously been shown that increasing CA and non-solvent concentration in a CA/solvent/non-solvent ternary system results in enhanced steady shear viscosity and dynamic viscoelastic properties (Pilon, 1971; Khalil, 1973; Appaw 2004). The observed rheological properties are believed to be due to the intensification of intermolecular hydrogen bond between polymer, solvent and non-solvent. At low non-solvent 80 concentrations the system exists as a clear homogenous solution. At a relatively high non-solvent content a sol-gel transition is observed with the formation of semi-solid systems that exhibit gel-like characteristics (Appaw et al. 2007). The sol-gel transition and gelation are dependent on CA concentration, and they decrease with increasing CA concentration. Phase separation and gelation are the most important phenomenon and of considerable interest for many applications including filtration and micro-encapsulation applications (Ilyina et al. 1993; Hao and Wang 2001). Phase separation and gelation are closely related, cellulosic systems typically phase separate prior to gelation (Ryskina and Averyano.Vm 1971; Ryskina and Averyanova 1975; Hao and Wang 2001; Appaw et al. 2007). Phase separation is dependent on the concentration of the specific cellulose derivative and the solvent used (Kjoniksen et al. 1998). Changes in intra- and intermolecular interactions in the solution has been proposed to account for gel formation in such systems (Cai and Zhang 2006). Gelation is initiated by large macromolecular associations that merge to form a three dimensional network solid structure (Winter and Mours 1997; Cai and Zhang 2006; Appaw et al. 2007). In CA/DMAc solutions the addition of water can induce phase separation, ultimately leading to gel formation (Appaw et al. 2007). In this system, the sol-gel transition is primarily dependent on the non-solvent (water) content and to a lesser extent CA concentration. Gelation is believed to arise from the manipulation of inter- and intramolecular hydrogen bonding interactions and hydrophobic polymer-polymer interactions. The addition of alcohols, specifically mono-hydroxyl alcohols, to a CA/acetone system also leads to phase separation (Gildert et al. 1979; Hao and Wang 2001). Depending on the non-solvent structure (methanol, ethanol, or isopropanol) the 81 phase behaviour and morphology of coagulated CA membrane varies. The water coagulated membranes had large macrovoids arising from liquid-liquid phase separation, while those from methanol showed a honeycomb-like structure due to spinodal microphase separation. In the case of ethanol or isopropanol a thick dense top layer was observed, and attributed to a delay in phase separation (Hao and Wang 2001). In thermoreversible physical gels, the thermal energy required to create or destroy the network junctions is of the same magnitude as that of polymer thermal motion (Williams et al. 1992). Physical gels do not involve covalent bonding. The cross-linked network structure is produced through ionic bonding, hydrogen bonding, and/or molecular entanglement (Williams et al. 1992; Winter and Mours 1997; Weng et al. 2004). Therefore, phase separation and physical gel formation is dependent on the specific interactions between components in the system. Consequently, increasing the number of hydrophilic groups, e.g. hydroxyl groups along the polymer backbone or on the other system components, can lead to a system in which more intermolecular hydrophilic interactions can occur (Li et al. 2001; Cai and Zhang 2006). The addition of hydrogen bonding non-solvents, such as water and alcohols to a CA solution induces gelation via the manipulation of hydrogen bonding. The addition of hydrophobic components to a polymer solution has also been shown to have a remarkable effect on its rheological properties and gel formation (Tsunashima et al. 2002). However, it is extremely difficult to control the gel network structure that is formed via the hydrophobic interactions. Hydrophobic interactions or connections can be affected by many factors, such as the number of aggregates per unit volume within the system, the average size of hydrophobic 82 aggregates, the structure of junctions, and/or the association strength (Winter and Chambon 1986; Tanaka 2000; Li et al. 2001). In the case of polymeric physical gel systems, there is no principle, or general rule for predicting the gel point in the solidification process. As a result, several techniques have been used to determine the point of gelation. Molecular weight divergence has been used, which involves the visual observation of transition from flow to non-flow behaviour (Mours and Winter 1996). Quantitative rheological measurements have also been used (Avanza et al. 2005; da Silva and Areas 2005), employing power laws or scaling laws to predict the point of gelation. The most common are zero-shear viscosity r\Q, dynamic moduli (G' and G"), and equilibrium modulus Ge before, at, and beyond the sol-gel transition (Castro et al. 1984; Mours and Winter 1996; Li et al. 2001). Although, the scaling laws, sol-gel transition and gel formation can be easily determined, there is still no universal value for determining the gel point. Furthermore, the scaling laws are not always applicable for all types of gels (Castro et al. 1984; Chambon et al. 1986; Mours and Winter 1996; Tanaka 2000). Finally, other techniques have been proposed for determining the gel point, such as a the maximum in tan 5 plot or in G", as well as the point at which G' is relatively close to, or crosses over with, G" (Chambon et al. 1986; Winter and Mours 1997; Avanza et al. 2005; da Silva and Areas 2005). In this study, the viscoelastic properties of a ternary system consisting of CA, N,N-dimethylacetamide and non-solvents (alcohols) were examined. The modulus of the ternary system and gel point formation are controlled by the addition of non-solvent to the CA solution. Different mechanisms of intermolecular interaction (hydrophobic and hydrophilic) were examined by changing the non-solvent alcohol structure, e.g. number 83 of hydroxyl groups (mono-, di-, trihydric alcohols) and alkyl chain length. The influence of the different non-solvent structures on the sol-gel transition and CA membrane (polymer conformation) properties were also examined. 3.3 Experimental Materials and Methods 3.3.1 Materials Cellulose acetate (CA - number average molecular weight (Mn) = 30,000 g/mol and degree of acetylation = 2.5), N,N - dimethylacetamide (DMAc - HPLC grade), 1-Propanol (1-Pro), 2-Propanol (2-Pro) 1-Hexanol (1-Hex), 1-Octanol (1-Oct), 1-Decanol (1-Dec), 1,2-Propandiol (1,2-PD), 1,3-Propandiol (1,3-PD) and glycerol (Gly) were purchased from Sigma-Aldrich and used as received. 3.3.2 Sample Preparation All CA/DMAc/Non-solvents mixtures were prepared from bulk solution as out lined in experimental materials and methods section in Appendix B. 3.3.3 Analysis FTIR, rheology and confocal scanning laser microscopy analysis of the various ternary systems was performed as out lined in experimental materials and methods section in Appendix B. 3.4 Results and Discussion 3.4.1 Solution Viscosity In this chapter, the effect of hydrophilic and hydrophobic interactions on the solution behaviour of a CA/DMAc/non-solvent ternary system was investigated by changing the 84 non-solvent structure. Increasing the number of hydroxyl groups on the propane backbone, i.e. 1-propanol, 1,3-propanediol, glycerol (1,2,3-propanetriol) increased the hydrophilicity of the non-solvent, and the propensity to interact via hydrogen bonding. Likewise, increasing the alkyl chain length of the monohydric alcohols from 1-hexanol to 1-octanol to 1-decanol increased the hydrophobicity of the non-solvent, and the extent to which hydrophobic interactions occurred in the system. Figure 3.1 shows the steady state viscosity of CA/DMAc/1-propanol solutions as a function of shear rate (s"1) for varying 1-propanol contents at a 15 wt% CA concentration. Similar viscosity profiles were observed for all of the alcohol non-solvents investigated (see Figures A23-A29 in Appendix). oo PH O O on > 10z 10' 10u o 1-Pro0% • 1-pro 10% © 1 -pro 20% X 1-Pro 30% + 1-Pro 40% A 1-Pro 45% V 1-Pro 48% AAAAAAAAAAAAAAAAAAAAAAAAAZ X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X x OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOo • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • n n ooooooooooooooooooooooooooooooooooooo 10" 10" 10" 10u 10' 10z 10J Shear rate (s") Figure 3.1: Steady state viscosity as a function of shear rate (s"1) for varying concentrations of 1-Propanol at a constant CA 15 wt% concentration. 85 At a low non-solvent content Newtonian behaviour was observed, followed by shear thinning at high shear rates. Increasing non-solvent content increased the viscosity to higher values, and shear thinning occurred at lower shear rates. At critical non-solvent content (Table 3.1) higher viscosities were observed at low shear rates, and the zero-shear viscosity plateau disappeared. The non-Newtonian behaviour and disappearance of the Newtonian plateau implies the development of microstructure in the sample. Microstructure development is a result of enhanced intermolecular interactions has been observed in increasing non-solvent content (Steiner et al. 1987; Fang et al. 2004). Table 3.1: Concentration of Non-solvent at which non-Newtonian behaviour is observed Non-solvent 1-Pro 2-Pro 1,2-PD 1,3-PD Gly 1-Hex 1-Oct 1-Dec Critical Cone. (wt.%) 48.3 46.6 43.3 35 30 40 34.6 31.6 Viscosity (Pa-s)* Critical Cone.a 4.4 4.3 14.3 21.6 24.2 3.8 8.9 32.3 20 wt.%b 1.4 - 2.4 3.0 5.3 1.1 1.4 1.9 a Viscosity at highest concentration exhibiting Newtonian behaviour. b Viscosity a 20 wt.% non-solvent content. * Viscosity values are within an error of 10%. Table 3.1 shows the concentrations at which non-Newtonian behaviour and microstructure development occurred for the various non-solvents, as well as the viscosities obtained at a constant non-solvent content (20 wt.%). Increasing the number of hydroxyl groups on the non-solvent significantly increased the viscosity of the system. 86 At 20 wt.% non-solvent content, the viscosity of the CA/DMAc/Gly system was almost four times the viscosity measured for the CA/DMAc/1-Pro system, 5.3 Pa-s and 1.4 Pa-s respectively. Similarly, increasing the alkyl chain length of the non-solvent alcohol from 1-Hex to 1-Dec, changed the viscosity from 1.1 Pa-s and 1.9 Pa-s, respectively. Interestingly, in both the 1-Pro and Gly systems and the 1-Hex and 1-Dec systems, at the highest concentration of non-solvents at which Newtonian behaviour is observed, the viscosity change almost an order of magnitude. Moreover, the change was slightly greater for the monohydric alcohol system, 1-Dec > Gly. Figure 3.2 illustrates the progressive increase in shear viscosity (data obtained at a shear rate of 1 s"1) with increasing non-solvent content, and changes in non-solvent structure. Increasing the number of hydroxyl groups (OH) on the non-solvent from 1 to 3 (1-Pro, 1,3-PD and Gly) led to an increase in viscosity (Figure 3.2a). At a relatively high non-solvent concentration (30 wt%), the samples, which are still clear solutions, vary in viscosity by almost an order of magnitude; 1-Pro ~ 2 Pa-s, 1,3-PD ~ 9 Pa-s, and Gly ~ 36 Pa-s. As non-solvent contents were increased, the solution viscosities increased dramatically. A change in the slope of the system with higher non-solvent contents was clearly observed for all of the non-solvent systems. Moreover, the change in slope was greatest for the Gly, such that Gly > 1,3-PD > 1-Pro. This may represent microstructure development as a result of increased intermolecular interactions between components. The same phenomenon was observed when the results were normalized to equivalent hydroxyl groups (Figure A30 in Appendix), indicating that the change in viscosity and intermolecular interactions were not solely the result of more hydrogen bonding groups. Again, the CA/DMAc/Gly showed the most rapidly change in viscosity. 87 i(r i o z OH GO O o Q 10 > 102 b-c3 OH £ 10 1/3 o o CO • I—I > 1 0.1 o 1-Pro • o 1,3-PD • Gly I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I o 1-Hex o 1-Oct • 1-Dec ' -10 0 10 20 30 40 50 Non-solvent (wt.%) Figure 3.2: Effect of a) hydrophilic and b) hydrophobic interactions on the viscosity enhancement of the CA/DMAc/non-solvent solutions (values obtained at a shear rate of 1 88 At low non-solvent contents, increasing the alkyl length of the non-solvent, or hydrophobicity also increased the solution viscosity, albeit not as dramatically as the hydrophilic non-solvents (Figure 3.2b). At 30 wt% non-solvent concentration the viscosities are ~1.5, ~2, and -14 Pa-s, for 1-hexanol, 1-octanol, and 1-decanol, respectively. Therefore, it appears that in dilute solutions hydrophilic interactions, such as hydrogen bonding have a greater effect on viscosity than hydrophobic interactions. The presence of more hydroxyl groups on the non-solvent enables increased hydrogen bonding between constituents, which leads to an increase in viscosity. However, at higher non-solvent contents, viscosity enhancement occurred very fast in the hydrophobic non-solvent system. From Figure 3.2 it is clear that the change in viscosity (slope) with increasing non-solvent content was greater for the longer alkyl chain monohydric alcohols as compared to glycerol, 1,3-propanediol or 1-propanol at high non-solvent contents (>~30 wt.%). This is probably due to hydrophobic interactions leading to phase separation and eventually solvent phase partitioning of the system (discussed below). Intermolecular interactions (physical entanglement and hydrogen bonding) are important parameters for viscosity enhancement in CA solutions. Changes in intermolecular interactions, specifically hydrogen bonding, as a function of increasing non-solvent content has been investigated using FTIR spectroscopy. In CA/DMAc solutions, CA dissolution involves the interaction of the hydroxyl groups in CA with the carbonyl groups in DMAc (Griswold and Cuculo 1974). The addition of a hydrogen bond accepting/donating non-solvent to the CA/DMAc solution led to the formation of new hydrogen bonds between the non-solvent/DMAc and the non-solvent/CA. The effect of this complex hydrogen bonding network system can be seen in the absorption bands of 89 the hydroxyl stretching region of the ternary system. Figure 3.3 shows the hydroxyl stretching region of the FTIR spectra of the various system components in the CA/DMAc/1-Pro ternary system. 3700 3500 3300 3100 Wavenumber (cm-1) Figure 3.3: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/1-Pro system. As with the dihydric alcohol system examined in Chapter 3, the FTIR spectrum of the CA/DMAc/l-Pro system revealed stronger hydrogen bond formation between 1-Pro and CA as compared to those between CA and DMAc, or DMAc and 1-Pro (the FTIR absorption band envelope shifted to lower wavenumber, implying the formation of stronger hydrogen bonds). The intensity of the lower wavenumber shoulder at 3280 cm"1 was greatest in the CA/DMAc/l-Pro system. Increasing non-solvent concentration led to 90 enhancement in the strong hydrogen bonding interactions; the CA/DMAc/1-Pro (gel) system having the highest intensity of lower wavenumber shoulder. Similar profiles were observed for all of the non-solvent systems (Figures A31-A35 in Appendix). The FTIR spectra of the hydroxyl stretching region of the hydrophilic (1-Pro, 1,3-PD and Gly) and hydrophobic (1-Pro, 1-Hex, 1-Oct and 1-Dec) CA/DMAc/non-solvent ternary systems at the same non-solvent contents are shown in Figure 3.4 and Figure 3.5, respectively. The FTIR spectra of the 1-Pro, 1,3-PD and Gly (Figure 3.4) showed a dramatic difference between the non-solvents. Increasing the number of the hydroxyl groups on the non-solvent led not only to more hydrogen bonding, but also to the formation of stronger intermolecular hydrogen bonds between components. The FTIR hydroxyl stretching band of the CA/DMAc/Gly ternary system increased not only in intensity (increased size of the band envelope), but also in hydrogen bond strength (hydroxyl stretching band shifts to lower wavenumber) as compared to that involving 1,3-PD, which is more than that involving 1-Pro. In addition, the change in the hydroxyl stretching band envelope between the DMAc/non-solvent and CA/DMAc/non-solvent is greatest for Gly. This is consistent with the observed change in viscosity (Figure 3.2a), and suggests that the hydrogen bonds formed between CA and Gly were stronger than those between CA and either 1,3-PD or 1-Pro. Interestingly, the difference between the spectra for DMAc/l,3-PD and CA/DMAc/l,3-PD and that of DMAc/l-Pro and CA/DMAc/l-Pro are quite similar, and the change in wavenumber was quite small. 91 3700 3500 3300 Wavenumber (cm"1) 3100 Figure 3.4: FTIR spectra of the hydroxyl stretching region of a) CA/DMAc/1-Pro, CA/DMAc/l,3-PD, and CA/DMAc/Gly solutions (33.3 wt% non-solvent content and 15 wt% CA concentration), and b) DMAc/l-Pro, DMAc/l,3-PD, and DMAc/Gly solutions (33.3 wt% non-solvent content). By contrast, the FTIR spectra of the hydrophobic (1-Pro, 1-Hex, 1-Oct and 1-Dec) CA/DMAc/non-solvent ternary systems showed very little change in the hydroxyl stretching band envelopes (Figure 3.5). Increasing the alkyl chain length of the non-solvent monohydric alcohol led to a slight shift of the hydroxyl stretching band envelope to higher wavenumber, indicating weaker hydrogen bonding interactions. However, this change in band profile was very slight. Thus, the change in viscosity (Figure 3.2) 92 observed in changing the non-solvent structure from 1-Hex to 1-Oct to 1-Dec, implies hydrophobic, non-bonding interactions also played a role in the development of microstructure and viscoelastic properties. 3700 3500 3300 3100 Wavenumber (cm-1) Figure 3.5: FTIR spectra of the hydroxyl stretching region of CA/DMAc/1-Pro, CA/DMAc/l-Hex, CA/DMAc/1-Oct, and CA/DMAc/l-Dec solutions (33.3 wt% non-solvent content and 15 wt% CA concentration). 3.4.2 Hydrophilic and Hydrophobic Interactions and the Sol-Gel Transition Figure 3.6 shows the stress sweep spectra of the hydrophilic (1-Pro, 1,3-PD, Gly) ternary systems at the same non-solvent content (33.3 wt%) and CA concentration (15 wt%). At 33.3 wt% non-solvent content, the CA/DMAc/l-Pro system (clear solution) displaying the lowest elastic modulus of the three non-solvent systems. The CA/DMAc/l,3-PD 93 system exhibited a substantially higher modulus, and the solutions appeared cloudy/turbid. At the same non-solvent content (33.3 wt%) the CA/DMAc/Gly system appeared as a self-supporting network, with the highest modulus of the three, G' ~ 80,000 Pa. Under these conditions, constant non-solvent content, the number of hydroxyl groups on the non-solvent increases 1 - 2 - 3 times. As a result, the elastic modulus (G') increased by nearly six orders of magnitude between 1-Pro and Gly.. o b 10c 105 104 i o 3 102 101 10 10_ 0 10" OOOOOOCXDOOOOOOOOOCXXX)OoOOOOOOOOOB§o *8§§c IAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA •^OOOCXXXXXOOOOOO 10"z 10"' 10u io ' i o 2 Osc. Stress (Pa.) 10J • G' Gly 0 G" Gly • G1 1,3-PD A G" 1,3-PD • G' 1-Pro 0 ( 1-Pro 10" Figure 3.6: Stress sweep spectra for ternary systems consisting of 1-Pro, 1,3-PD and Gly at 33.3 wt% non-solvent content and 15 wt% CA concentration. 94 However, comparing the stress sweep spectra at approximately the same hydroxyl equivalents (1-Pro 33.3 wt.% ; 1,3-PD 21.4 wt.%; Gly 16.7 wt.%) (Figure 3.7) produced very little difference in modulus between the respective non-solvents. It can be seen that the behaviour of the elastic moduli with increasing stress was slightly different between the three systems. The CA/DMAc/Gly and CA/DMAc/1,3-PD exhibited a step change in the elastic modulus with increasing applied stress. The step-wise reduction in the LVE regime was greatest in the 1,3-PD system. Interestingly, the Gly and 1,3-PD systems had identical G' at low stress, but as stress was increased the 1,3-PD system G' reduced to that of the 1-Pro system. 03 PH a i ( r i o z 10' i o u 10" OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A oooooooooooooooooooooooooooo ° o o o O O r • G' (1-pro) O G" (1-pro) A G' (1,3-PD) A G" (1,3-PD) • G' (Gly) O G" (Gly) • • • • * A A A A A A A xKl •• • • • • • • • . H A A A A • • • A . 10 -2 10 10u 101 Osc. stress (Pa) 10z 10J Figure 3.7: Stress sweep spectra for ternary systems consisting of 1-Pro, 1,3-PD and Gly at 33.3 wt%, 21.4 wt.%; Gly 16.7 wt.% non-solvent content, respectively and 15 wt% CA concentration. 95 Figure 3.8 shows the stress sweep spectra (G' & G") of the hydrophobic (1-Pro, 1-Hex, 1-Oct, 1-Dec) ternary systems at the same non-solvent content (35 wt%) and CA concentration (15 wt%). Increasing the alkyl chain length of the non-solvent from 1-Hex to 1-Oct and finally to 1-Dec appeared to enhance the elastic modulus by almost four orders of magnitude. Increasing the hydrophobic component of the non-solvent and the resulted in an increase in modulus, which further illustrates the complexity of the CA/DMAc/non-solvent system. OH a <% a i o 4 10 10z 10' 10u 10" 10 -2 - --• A A A A AAA. A. A, A, A. A, A. A. A. A. A, A, A. A. A. A. A. A, A. A . * 1 oooooooooooooooooooooooooooooooooo A AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAA AAAAAAAAAAA. 10"' 10u 10' Osc. Stress (Pa) 10 A G' 1-Hex A G" 1-Hex • G' 1-Oct O G" 1-Oct A G' 1-Dec A G" 1-Dec 10J Figure 3.8: Stress sweep spectra for ternary systems consisting of 1-Hex, 1-Oct and 1-Dec at 35wt% non-solvent content and 15 wt% CA concentration. Elastic (G') and viscous (G") moduli, as a function of frequency, were used to quantify the sol-gel transition (Figures A36-A43 in Appendix). At a 1-Pro content of 38.3 wt% 96 the system formed a clear homogenous solution, and G" was larger than G'. Increasing the 1-Pro content to 48.3 wt% led to phase separation and both moduli were approximately equivalent, G ~G". At this point, which is referred to as the gelation point, both moduli have increased and exhibited only a slight frequency dependence (Kobayashi et al. 1999; Avanza et al. 2005; da Silva and Areas 2005). Further addition of 1-Pro (53.3 wt%) led to the formation of a self-supporting gel-like material where G' was larger than G"; a characteristic feature of a three dimensional elastic gel. At this stage, both moduli are relatively frequency independent (Figure 3.9) (Nielsen 1977; Mours and Winter 1996). c3 w b b 10° io H 10z 10u 10" 4 10" 10 -2 oooo oo<>* o<>ovo , H H B ' • • • • .CP _ l -1 10u Angular frequency (rad/s) • G' (1-Pro 38.3 wt.%) o G" (1-Pro 38.3 wt.%) • G 1 (1-Pro 48.3 wt.%) • G" (1-Pro 48.3 wt.%) • G' (1-Pro 53.3 wt.%) © G" (1-Pro 53.3 wt.%) i o 1 i o 2 10J Figure 3.9: Elastic (G') and viscous (G") moduli for the CA/DMAc/l-Pro ternary system at 15 wt% CA concentration and several 1-Pro contents. Shown are a clear solution, cloudy system, and self-supporting gel 97 The effect of non-solvent content on the elastic modulus (G', selected at the frequency 1 rad s"1) of the hydrophilic and hydrophobic ternary systems is shown in Figure 3.10. As was observed for the dihydric alcohols (Chapter 2), a sigmoidal shape with three distinct phases was observed in terms of G' enhancement with increasing non-solvent content. The elastic modulus initially increases slowly with increasing non-solvent content, then an intermediate phase of sharp increase in G' and gel formation, followed once again by a slow increase in G'. Although all of the ternary systems exhibited the same sigmoidal shape, the phase of sharp increase in G' happened faster, or at a lower non-solvent content with increasing number of OH groups on the non-solvent structure (hydrophilic system), or increasing length of alkyl chain (hydrophobic system). Addition of OH group to the structure of non-solvent led to faster phase separation since hydrophilic intermolecular interactions, i.e. hydrogen bonding were more pronounced between components. The increase in hydrogen bonding led to increased intermolecular interactions between CA/DMAc/non-solvent, and ultimately phase separation and gel formation. A comparison of G' values as a function of equivalent OH groups revealed the similar trends, with the Gly system G' increased faster with increasing OH content (Figure A44 in Appendix). Increasing the alkyl chain length or hydrophobicity of the non-solvent increased the hydrophobicity of the ternary system; this affects solvent quality and led to phase separation. Unlike the hydrophilic system, the hydrophobic system gels at higher non-solvent contents underwent solution partitioning wherein a very strong gel network formed under a clear liquid solution (Figure 3.11). This "gel-liquid" phase separation 98 occurred in all of the hydrophobic non-solvents, and was most pronounced for the longer alkyl chain non-solvents. O 10 OH O 10 10 20 30 40 Non-solvent (wt%) Figure 3.10: Effect of increasing non-solvent content on the elastic modulus (G ' at 1 rad s"1) for the a) hydrophilic and b) hydrophobic ternary systems.(15 wt% C A concentration) 99 The presence of a "two-phase" system, liquid layer on top of a stiff gel, with increasing non-solvent may be related to the solution behaviour of these monohydric alcohols. It has been reported that in organic solution long chain alcohols form micelles and reverse micelles.(Tikhonov et al. 1982; Bakunin et al. 1997) Thus, hydrophobic monohydric alcohols probably interact with CA via hydrogen bonding (see below), as well as with DMAc, and with increasing hydrophobic non-solvent addition; the CA must compete against non-solvent self-association and micelle/reverse micelle formation. As a result, at a critical non-solvent content, which decreases with increasing non-solvent alkyl chain length, non-solvent self-association (micelle/reverse micelle) occurs and the system phase separates. The top liquid layer was likely comprised primarily of non-solvent and DMAc, and the bottom a highly concentrated ternary system. The net result is less non-solvent and DMAc available to interact with CA and stabilize the gel network, and a more condensed system forms with enhanced CA-CA interactions (Tsunashima and Hattori 2000; Tsunashima et al. 2002). 3.4.3 Non-solvent Structure and Effect on Bonding Mechanism Although the impact of hydrophobic interaction on G' development is an important factor in phase separation, it cannot be considered the main factor in the solidification process in this CA ternary system. For example, the ternary systems with hexane and/or polyethylene glycol (600 Mw) as non-solvent did not exhibit phase separated gel formation (Figure 3.11a). With increasing hexane or PEG concentration, the system phase separated into two clear liquid layers. Likewise, the use of dibutyl ether as a non-solvent did not led to gel formation (Figure 3.11b) Again, phase separation occurred, but the system did not form a self-supporting gel network. By contrast, the 1-decanol system 100 (Figure 3.12c) did form a self-supporting gel network at high non-solvent contents, and ultimately a "two-phase" liquid layer on top of a stiff gel system. Figure 3.11: Digital images showing the effect of a) hexane, b) dibutyl ether, and c) 1-decanol on the phase behaviour of the ternary systems in D M A c at 15wt% C A concentration. 101 These results are quite significant. Hydrogen bonding interactions in the CA ternary system examined here are essential for obtaining a physical gel, but intermolecular interactions based on proton donor-acceptor groups appeared critical. The effect of O—H intermolecular hydrogen bonding on gel formation can be seen by comparing the results obtained for the conformational isomers 1-octanol and dibutyl ether (molecular formula CsHigO). The 1-octanol ternary system produced a uniform cross-linked gel at 37.3 wt%. The single hydroxyl group in 1-octanol can act as a proton donor and acceptor group. By contrast, dibutyl ether is aprotic, and can only act as a proton acceptor (ether oxygen group). As such, dibutyl ether can only interact via hydrogen bonding with the free hydroxyl groups in the CA molecule. 1 -octanol can interact with both the free hydroxyl groups as well as the acetyl carbonyl groups, which are much more abundant (DA = 2.45). Therefore, it seems that i) hydrophobic interactions alone are not responsible for gel formation, and ii) hydrogen bond donating groups on the non-solvent are required for gel formation. Further evidence of the complexity of the interactions involved in the ternary system was seen using 1,2-PD as the non-solvent. Figure 3.12 shows the frequency sweep spectra of 1,2-PD and 1,3-PD at the same concentration of non-solvent (36.3 wt.%) and CA (15 wt.%). It can be seen that the 1,2-PD system is a clear solution with G'<G", whereas 1,3-PD at same concentration forms a 3-dimensional network gel with a much larger elastic modulus and G'>G". Also included in Figure 3.12 is the frequency sweep data obtained for 1-Pro at the same 38.3 wt% concentration. It appears that at the same concentration 1,2-PD and 1-Pro have similar moduli and frequency dependence relative to 1,3-PD. 102 104 _ 10 fe b * 10° b 10" 10 -4 10" ***** © • G' 1-pro o G" 1-pro • G' 1,2-PD • G" 1,2-PD • G' 1,3-PD © G" 1,3-PD 10"' 10u 101 102 Angular frequency (rad/s) 10J Figure 3.12: Frequency sweep spectra of 1,2-propandiol 1,3-propandiol and 1-propanol at the same concentration of non-solvent (36.3 wt.%) and CA (15 wt.%). Similarly, the behaviour of elastic moduli with increasing non-solvent concentration was substantially different for 1,3-PD and 1,2-PD systems. Figure 3.13 shows increasing non-solvent content leads to a substantial increase in G' and gel formation. The concentration of 1,3-PD being much lower than that for 1,2-PD. Again, similar trends are also observed between 1,2-PD and 1-Pro over the entire range of moduli/concentration. 103 b 10 10 ,-3 -10 0 10 20 30 40 50 60 Non-solvent (wt%) Figure 3.13: Effect of increasing non-solvent content on the elastic modulus (G' at 1 rad s"1) for the 1-Pro, 1,2-PD and 1,3-PD ternary systems. (15 wt% CA concentration) The observed differences in behaviour between the 1,2-PD and 1,3-PD ternary systems may be rationalized by the differences in molecular interactions present in the two systems. Both 1,2-PD and 1,3-PD can potentially form intramolecular hydrogen bonds. 1,3-PD (and Gly) can form syndiaxial intramolecular hydrogen bonds, while 1,2-PD can form vicinal intramolecular hydrogen bonds. Vicinal intramolecular hydrogen bonding is rare as the angular geometry, i.e. five-member ring, is unfavorable. On the other hand, syndiaxial intramolecular hydrogen bonding has more favorable six-member ring geometry. Although no data is reported on the intramolecular hydrogen bond energy of 1,3-PD, based on computational estimations for Gly and other polyols (Rozas et al. 2001; 104 Deshmukh et al. 2006), syndiaxial intramolecular hydrogen bonds are generally classified as weak hydrogen bonds. Moreover, in concentrated solutions with polar solvents, intramolecular hydrogen bonding is less favoured than intermolecular hydrogen bonding, and generally does not occur. Therefore, the difference in viscoelastic behaviour must arise due differences in intermolecular interactions in the systems. 1,2-PD is a conformational isomer of 1,3-PD, and structurally is a more compact molecule, i.e. it likely has a more spherical shape than 1,3-PD, which is likely more elongated. As a result, 1,3-PD has more surface area, or is more accessible for intermolecular interaction with the various component within the ternary system. The stronger intermolecular interactions can be seen from the higher boiling point (214°C vs. 187°C) and density (1.053 g ml"1 vs. 1.036 g ml"1) for 1,3-PD as compared to 1,2-PD. 3.4.4 Rheology at the Gel Point Figure 3.14 illustrates G' and G" for the ternary systems comprised of 1-Pro, 1,3-PD and Gly as the non-solvent at the gelation point or intermediate state, where G' crosses G" or is very close to G". The crossover of G'-G" during the liquid to solid transition is an indication of development in of a cross-linking structure and formation of a weak gel (Mours and Winter 1996; Kobayashi et al. 1999; Chauvelon et al. 2003; Avanza et al. 2005; da Silva and Areas 2005; Cai and Zhang 2006). 105 10* OH b i o J 10z A ^ < , 0 ° A * • O ° A ^ A A A ^ •o • G' (Gly 30.0wt%) o G" (Gly 30.0wt%) A G'(1,3-PD 34.6wt%) A G" (1,3-PD 34.6wt%) • G'(l-Pro48.3wt%) 0 G" (1-Pro 48.3wt%) 10' 10" 10" 10u 10' 10z 10 Angular frequency (rad s"1) Figure 3.14: Viscous (G") and elastic (G') moduli at or near the gel point for 1-Pro, 1,3-PD and Gly non-solvent ternary systems at 15 wt% CA concentration (numbers in parentheses are weight percentage of non-solvent). These cross-linked weak gels were made by controlling the amount of non-solvent, and were measured by plate geometry. As expected, based on the viscosity data, the Gly ternary system demonstrated the largest elastic modulus at the lowest non-solvent content (30 wt%). As the number of OH groups on the non-solvent increased, the ternary system exhibited a larger elastic modulus and required less non-solvent content to reach the gel point: Gly (30wt%) < 1,3-PD (34.6wt%) < 1-Pro (48.3wt%). This again shows the significance of hydrophilic intermolecular interactions between the non-solvent and CA on developing the viscoelastic properties of the ternary system. 106 A similar trend was observed for the 1-Hex, 1-Oct and 1-Dec non-solvent ternary systems. Figure 3.15 shows that the sol-gel point (G'~G") for four different non-solvents (1-Pro, 1-Hex, 1-Oct and 1-Dec) occurred at the highest modulus and lowest non-solvent concentration for the 1-Dec (35.0 wt%), with the lowest belonging to 1-Hex (40.0 wt%). Therefore, as discussed with respect to the effect of increasing alkyl chain length on viscosity, increasing hydrophobic interaction and phase behaviour of the system the elastic modulus of the ternary system at a transitional state increases accordingly. PH o b i o 3 10" i o 3 h 10' 10" ©2°-o • © # ,A AS' • v T G' (1-Pro 48.3 wt%) V G" (1-Pro 48.3 wt%) A G' (1-Hex 40.0 wt%) A G" (1-Hex 40.0 wt%) • G' (1-Oct 37.3 wt%) G" (1-Oct 37.3 wt%) • G' (1-Dec 35.0 wt%) o G" (1-Dec 35.0 wt%) 10 10u 10 Angular frequency (rad/s) 10z 10J Figure 3.15: Viscous (G") and elastic (G') moduli at or near the gel point for 1-Pro, 1-Hex, 1-Oct and 1-Dec non-solvent ternary systems at 15 wt% CA concentration (numbers in parentheses are weight percentage of non-solvent). 107 FTIR spectroscopy further supported the observation of the rheological experiments (Figures 3.14 and 3.15) showing clear absorption bands in the OH stretching regions (Figures 3.16 and 3.17). Similar to the solution systems (Figure 3.4), increasing the number of hydroxyl groups able to participate in hydrogen bonding increased the size of the band envelopes, i.e. area under the curve as well as the strength of the hydrogen bonds present in the system (shift to lower wavenumber). 3700 3500 3300 3100 Wavenumber (cm"1) Figure 3.16: FTIR spectra of the hydroxyl stretching region of CA/DMAc/ 1-Pro (48.3 wt%), CA/DMAc/l,3-PD (34.6 wt%), and CA/DMAc/Gly (30 wt%) systems at the sol-gel transition point (15 wt% CA concentration). The hydroxyl stretching band envelope of the CA/DMAc/Gly ternary system at the gel-point was very broad, extending over the widest wavenumber range. There was evidence 108 of intramolecular hydrogen bonding as the band envelope extends above ~3600cm (Deshmukh et al. 2006), as well as very strong intermolecular hydrogen bonding, the band envelope extends below -3200 cm"1. The CA/DMAc/Gly ternary system had the strongest hydrogen bonds between the three hydrophilic non-solvents investigated. Again, comparing the three non-solvents showed the same trend as observed for the viscous solutions, wherein the hydrogen bonds formed between CA and Gly were greater than those between CA and 1-PD, which were greater than those between CA and 1-Pro. 3600 3400 3200 3110 Wavenumber (cm"1) Figure 3.17: FTIR spectra of the hydroxyl stretching region of CA/DMAc/l-Pro (48.3 wt%), CA/DMAc/l-Hex (40 wt%), CA/DMAc/l-Oct (37.3 wt%)and CA/DMAc/l-Dec (35 wt%) systems at the sol-gel transition point (15 wt% CA concentration). Not surprisingly, for the hydrophobic systems, the FTIR spectra of the transition state samples were quite different. Analogous to Figure 3.5, Figure 3.17 shows that the 109 hydroxyl stretching band for the various hydrophobic non-solvent ternary systems shifts to lower wavenumber with decreasing alkyl chain length of the non-solvent. In fact, the 1-Pro non-solvent ternary system had the strongest hydrogen bonding interactions followed by 1 -Hex, 1 -Oct and 1 -Dec, respectively. This trend is opposite to that observed for the elastic moduli behaviour (Figure 3.16), and illustrates that although hydrogen bonding is essential for gel formation, non-bonding hydrophobic interactions also play a role in determining the modulus of the resulting gels (see below). 3.4.5 Rheological Characterization of Network Structure Figures 3.18 and 3.19 show the elastic modulus (G') in a strain sweep plot for the ternary systems of the different non-solvents (Figure 3.18 - 1-Pro, 1,3-PD and Gly; Figure 3.19 - 1-Hex, 1-Oct, 1-Dec) at a constant CA concentration (15 wt%). Increasing the non-solvent content enhanced the elastic modulus by several orders of magnitude as the solution changed from a clear homogenous solution to a turbid/cloudy system and ultimately formed a gel. As shown in Figure 3.18b and 3.18c, at low non-solvent concentration (cloudy solutions) G' decreased stepwise with increasing applied stress. This phenomenon is characteristic of a weak interconnected polymer network (see Chapter 2) and a semi-developed microstructure. Applying stress to these samples breaks the weak cross-links, and as a result the elastic modulus shows a temporary drop in the linear viscoelastic region (Khalil 1973; Winter and Mours 1997; Chauvelon et al. 2003; Weng et al. 2004). 110 c3 PH, O 03 PH, b 03 PH, b 10" 10 10z io 1 10u 10" 10' 10" 10J 10z 10' 10u 10 10" 10J 10z 10' 10u EHSUBJB 10" 1-Pro (26.6wt%) 1-Pro (40.0wt%) 1-Pro (45.0wt%) 1-Pro (46.6wt%) 1-Pro (48.3wt%) 1-Pro (50.0wt%) B G'1-Pro (51.6wt%) o G'1,3-PD (26.6wt%) • G' 1,3-PD (31.6wt%) o G'1,3-PD (35.0wt%) A G' 1,3-PD (36.6wt%) v G'1,3-PD (40.0wt%) * G' 1,3-PD (43.3wt%) ' A A A o G' Gly (23.3wt%) • G' Gly (25.0wt%) o G' Gly (26.6wt%) A G' Gly (28.3wt%) V G' Gly (30.0wt%) Q G' Gly (31.6wt%) a G' Gly (33.3wt%) 10 -5 10"3 Strain % 10" 10' Figure 3.18: Elastic modulus of CA gels for the different non-solvent, 1-Pro, 1,3-PD and Gly concentrations as a function of strain. The limit of linearity shifts to lower strain as concentration increases. I l l 03 fe b 03 fe b 03 fe b 10* 103 102 101 10° 105 104 103 102 101 10° lO"1 105 104 103 102 10' 10° lO"' lO"2 10" I I I I _ I -^^ +++^ -H-H-H-^ ++ o Hex (35.0 wt%) • Hex (38.3 wt%) o Hex (40.0 wt%) X Hex (40.6 wt%) + Hex (41.0 wt%) A Hex (41.6 wt%) V Hex (43.3 wt%) oocoocoooooccwc*^ I I I I I 111 11 111111111 11 111 111111111111 | - H - + ° Oct (33.3 wt%) n Oct (34.6 wt%) o Oct (36.0 wt%) x Oct (37.3 wt%) + Oct (41.6 wt%) CHmEBooBBtBoaasHBa^  «++++ o Dec (26.6 wt%) • Dec (30.0 wt%) 0 Dec (33.3 wt%) X Dec (35.0 wt%) + Dec (38.3 wt%) i r w i n n r n n i i i r x o ^  10" 10"3 10" Strain % 10' Figure 3.19: Elastic modulus of CA gels for the different non-solvent, 1-Hex, 1-Oct and 1-Dec concentrations as a function of strain. The limit of linearity shifts to lower strain as concentration increases. 112 In the case of 1-Pro, the temporary drop in the elastic modulus was not apparent. As network formation is the result of intermolecular interactions, i.e. hydrogen bonding, the structure of the non-solvent and number of OH groups will affect the cross-linked structure and behaviour under stress. In fact, the same step-wise drop in modulus was observed for the various dihydric alcohols studied in Chapter 2. As with the 1-Pro system, the other hydrophobic non-solvent ternary systems (1-Hex, 1-Oct and 1-Dec) did not exhibit a stepwise reduction in G' (Figure 3.19). Rheology can also be used to determine the mechanism of gelation, and whether or not the gels are fractal in nature (Eissa and Khan 2005). This is done by examining the relationship between G' and volume fraction or content of the polymer, or in our case non-solvent. If the G' exhibited a power-law behaviour with non-solvent content (<£>), i.e. G'~ 0 n, the gels are fractal in nature. As was found in Chapter 2, both the hydrophilic and hydrophobic non-solvent systems exhibit a power-law behaviour between elastic modulus and non-solvent content. The power-law exponents varied between the various systems, ranging from -134 - 28 (Figure 3.20). In both systems, the weaker modulus gels (1-Pro and 1-Hex, respectively) exhibited larger power-law exponents, both significantly higher than the other non-solvents. The similarity in structure of 1-Pro and 1-Hex and their larger power-law exponent values suggests that their gelation mechanism is different than the other non-solvents. Increasing both hydrophobicity and hydrophilicity in non-solvent caused significant decreases in the power-law exponent values. 113 03 O H 03 P H 10° 10 10* O 3 103 10z 10° 103 10* O 3 10 3 10z 10' 1-Pro - H — 1,3-PD -0 — Gly V = 3.4xl0"w *x&a Rl =0.98 (Pro) j = 1.5xl0"43 *x29 R2=0.91 (PD) v = 4.6xl0"56 *x38 i?2=0.98 (Gly) 1 1 1 1-Hex - E — 1-Oct -0 — 1-Dec v = 6 . 9 x l 0 - 2 , 4 * x 1 3 4 >; = 2 . 7 x l 0 - 4 1 * x 2 8 y = 7.9xlQ 1 1 6 **3 2 R2=0.94 (Hex) iJ2 =1.0 (PD) R2=\.0 (Gly) 20 30 40 50 Non-solvent (wt%) 60 Figure 3.20: Elastic modulus of CA/DMAc/non-solvent gels as a function of non-solvent content for a) hydrophilic and b) hydrophobic non-solvent systems (C obtained at frequency of 1 rad s"1). For various polymeric gel systems rheology has been employed to determine the fractal dimensions of gels (Shih et al. 1990). As discussed in Chapter 2, this approach involves 114 gels being viewed as being made up of floes consisting of aggregated macromolecules, which interconnect to form a 3-dimensional gel network. Gels can be classified as being either strong-linked or weak-linked systems. In strong linked systems, the links between floes are stronger than those within the floes, and as a result, failure under deformation occurs via breaking of intra-floc interactions. This leads to a decrease in the LVE regime with increasing sample concentration. By contrast, weak-linked gels possess strong intra-floc links, more rigid than the inter-floe linkages, and an increase in the limit of linearity occurs with increasing concentration. Figures 3.18 and 3.19 show that the onset of nonlinearity, no matter how one defines it, shifted to a lower value with increasing concentration, indicating the CA/DMAc/dihydric alcohol gels are strong-linked. In the case of the hydrophobic non-solvent systems, the onset of non-linearity for the gel samples occurred over a small range of % strain. As with the dihydric alcohol systems discussed in Chapter 2, this may imply that the 1-Dec system contained relatively weaker inter-floe link bonds than those in the 1-Oct system, which again weaker than those in the 1-Hex system. The strain spectra showed that the addition of the non-solvent to the ternary system shifted the limit of the linear elastic modulus plateau to lower strain values. This suggests that ternary systems with a large elastic modulus (gels) are made of floes consisting of aggregated macromolecules (Shih et al. 1990; Lowe et al. 2003; Eissa and Khan 2005; Kadla et al. 2005). These floes are interconnected to form a three dimensional gel network. The solid network CA ternary system is a strong-linked gel system in which the links between floes are stronger than the links within the floes. As a result, failure under deformation occurs the via the breaking of intraflocs linkage. 115 In strong-linked gels G' scales as G'~0Ka+ma~u>, where x is the backbone fractal dimension of the floes, d is the Euclidean dimension, and D is the gel fractal dimension (Shih et al. 1990; Marangoni 2002; Eissa and Khan 2005). Using theoretical and experiment power-law exponents, and substituting x=1.15 and d=3, it was found that£) = 2.9±5 (Table 3.1) for of the various gels with differing non-solvent structure and concentration. Table 3.2: Fractal dimensions of CA/DMAc/non-solvent gels at 15 wt% CA concentration, obtained by rheology and confocal microscopy. 1-Pro 1,3-PD Gly 1-Hex 1-Oct 1-Dec by confocal microscopy by rheology 1.90±0.05 2.95±0.10 1.90±0.05 2.85±0.08 1.92±0.04 2.89±0.99 1.98±0.05 2.95±0.13 1.95±0.03 2.85±0.05 1.85±0.06 2.87±0.15 An alternative approach to determining fractal dimension, as well as directly visualizing gel microstructure is through confocal microscopy (see below). The fractal dimensions of the gels can be calculated through confocal images by the box counting technique using Image J software.(Eissa and Khan 2005) In this method, the scaling relation is N oc r-D, where JV is the number of boxes filled with CA, r is the size of the square box, and D is the fractal dimension of the CA network. The fractal dimension of the gels obtained via confocal microscopy also showed similar values between the various systems, D ~ 1.90 ±0.05, regardless of CA concentration and type of non-solvent. According to this method (Table 3.2), gels are also made of floes, consisting of aggregated macromolecules that 116 are interconnected to form a three-dimensional gel network (Shih et al. 1990; Marangoni 2002; Eissa and Khan 2005; Kadla et al. 2005). It has been reported that the fractal dimension can vary enormously if there is a systematic departure from linear behaviour (Marangoni 2002). According to Alejandro Marangoni (2002), mass fractal dimensions determined by confocal microscopy technique belong to 2-dimensional Euclidean space, while that determined by rheological experiments translate to a 3-dimensional Euclidean space. Therefore, these two different techniques provide different values for the fractal dimension. However, for fractal scaling theories to be theoretically valid, there must be agreement between the fractal dimensions determined by rheological methods and microscopy methods, that is D~3 by rheology and D~2 by image analysis. 3.4.6 Microscopic Analysis of Gel Microstructure In this study, laser confocal fluorescence microscopy was employed to characterize the microporous and network structure of the various ternary system gels. Figure 3.21 shows the micrographs obtained for the gels obtained for the hydrophilic non-solvent systems (1-Pro, 1,3-PD, Gly) at the same G' and same CA concentration (15 wt.%). The dark blue images represent fluorescent mode images of the samples stained with calcofluor white. Calcofluor white is a cellulose selective dye that labels the surface of the CA network. In the fluorescent images, the CA microstructure is shown by the dark segments. By contrast, in the reflection images (seen in red) to the dark areas correspond to the non-CA domains (Ljunglof et al. 1999; Charcosset and Bernengo 2000; Charcosset et al. 2000; Fang et al. 2003). 117 Figure 3.21: LSCM images of gels at the same elastic modulus (G') and at the same CA concentrations (10 wt.%) for the 1-Pro, 1,3-PD, Gly non-solvent ternary systems. The top images are in fluorescent mode and the bottom images are in reflective mode. A denser network with a fine and uniform structure is observed as the non-solvent is changed from 1-Pro to 1,3-PD and finally to Gly (Figure 3.21). Increasing the number of hydroxyl groups on the non-solvent structure led to a more uniform gel microstructure. This increased network connectivity is consistent with the enhanced hydrophilic interactions in the ternary system. Comparison of the three non-solvent systems shows a substantially different microstructure between 1 -Pro and the other two alcohols. However, the calculated fractal dimensions are approximately the same -1.90. A similar discrepancy was reported for a series of caseinate gels. These gels exhibit visually different structures, ranging from open to dense and more uniform, yet all having the 118 same calculated fractal dimension from image analysis. The authors suggested that the fractal dimension alone may not always be sensitive enough to capture the differences observed in gel microstructure (Pugnaloni et al. 2005). Differences in microstructure were also observed between the hydrophobic non-solvent systems. Figure 3.22 shows that increasing the length of the alkyl chains in the monohydric alcohols (1-Hex, 1-Oct, 1-Dec) created a bigger and less uniform micro-porous network structure (Figure 3.22 reflective mode). This is in sharp contrast to the results obtained for the hydrophilic non-solvents (1,3-PD and Gly), where modulus enhancement was accompanied by enhanced microstructure uniformity (Figure 3.22). Oct, FD=1.95 20 (Jm Hex 20 \lm Dec, FD=1.85 20 Oct |g 20 |Jm Dec 20 Mm Figure 3.22: L S C M images of gels at the same elastic modulus (G') and at the same C A concentrations (15 wt%) for the 1-Hex, 1-Oct, 1-Dec non-solvent ternary systems. The top images are in fluorescent mode and the bottom images are in reflective mode. 119 In the hydrophobic non-solvent systems, as the hydrophobicity of the non-solvent increased the void-volume or pore size of the gel network as well as the non-uniformity. However, according to the rheological results, the moduli of the gels are greater for the 1-Dec system as compared to the 1-Oct and 1-Hex systems. Thus, in the hydrophobic non-solvent systems, the decreased uniformity and more open structure led to increased modulus. One possible explanation for this observation may lie in the deference in phase behaviour and gelation mechanism. Both systems are considered strong-linked gels, wherein failure under deformation occurs between floes or aggregated molecules. However, at high non-solvent contents, the hydrophobic non-solvent systems under go a phase separation and the formation of a dense gel network phase below a clear solution phase. In these systems, it may be that phase separation and gelations are being driven by drop in solvent "quality" and CA aggregation. The fact that the system exhibits a sol-gel transition, rather than just CA precipitation is due to the fact that monohydric alcohols were used and forming intermolecular hydrogen bonds with the CA. 3.5 Conclusion The effect of non-solvent structure on the phase behaviour and viscoelastic properties of a CA/DMAc/non-solvent system was investigated using steady state and dynamic rheology. At low non-solvent concentration, Newtonian behaviour was observed, followed by shear thinning at high shear rates. Increasing addition of the non-solvent enhanced the intermolecular interactions and the system exhibited non-Newtonian behaviour at a low shear rate; representing the improvement in the microstructure. Increasing the number of hydroxyl groups (OH) on the non-solvent from 1 to 3 (1-Pro, 1,3-PD and Gly) increased the solution viscosity by increasing the number of hydrophilic 120 interactions. Likewise, increasing the alkyl chain length of the non-solvent (1-Hex, 1-Oct, 1-Dec) increased solution viscosity, possibly due to increasing non-bonding interactions. At low non-solvent concentrations, hydrophilic interactions were found to have a greater effect on viscosity enhancement than hydrophobic interactions. Using dynamic rheology (frequency sweep and stress sweep) the effects of non-solvent addition on the elastic (G') and viscous (G") moduli was studied. Increasing the non-solvent concentration led to a cloudy/turbid phase separated polymer system, and eventually the formation of a self-supporting gel-like material. As the non-solvent content increased, three distinct phases in the elastic modulus were observed; an initial region of slow G' increase with increasing non-solvent content, an intermediate phase of sharp increase in G' and concomitant gel formation followed once again by a slow increase in G'. It was found that G' enhancement happened faster, i.e. at lower non-solvent content with i) increasing number of OH groups on the non-solvent structure (Gly>l,3-PD>1-Pro), and ii) increasing alkyl chain length of the non-solvent structure (l-Hex>l-Oct>l-Dec). Similarly, at the gelation point the moduli was higher and the concentration of non-solvent required lower for i) Gly as compared to 1,3-PD andl-Pro, respectively, and ii) 1-Dec as compared to 1-Oct and 1-Hex, respectively. These results indicate that the structure of the non-solvent plays an important role in gel formation, and that both hydrophilic and hydrophobic interactions increase the modulus at the gel point. FTIR analysis confirmed the effect of hydrophilic interactions on the sol-gel transition. In the hydrophilic non-solvent ternary systems (1-Pro, 1,3-PD, Gly), increasing non-solvent concentration increased the size of the hydroxyl stretching band envelope and extended it to a lower wavenumber. This shift in the hydroxyl band at an intermediate state (G ~G") 121 to a lower wavenumber implies the formation of stronger hydrogen bonds, and accounts for the modulus or viscosity enhancement through hydrophilic interactions. In the hydrophobic non-solvents ternary systems (1-Hex, 1-Oct, 1-Dec) no effect was observed in hydroxyl stretching at low non-solvent content, and a shift to higher wavenumber was observed at high non-solvent levels. Furthermore, at high non-solvent contents, the hydrophobic ternary systems all underwent phase separation and the formation of a rigid gel network phase underneath a clear liquid phase was observed. It was found that protic non-solvents were required for gel network development; non-solvents that only acted as hydrogen bond acceptors (e.g. dibutyl ether) did not form gels, nor did non-polar aprotic non-solvents (e.g. hexane). Both rheological and microscopic analyses of the various CA/DMAc/non-solvent gels revealed they were fractal in nature, and can be classified as strong-linked gel networks. Microstructural analysis revealed distinctly different microstructures between the hydrophilic and hydrophobic non-solvent ternary systems. The hydrophilic ternary system gels showed increased uniformity and density with increasing hydrophilicity of the system, consistent with their increased moduli. By contrast, the hydrophobic system showed a decrease in uniformity and a more open network structure with increasing hydrophobicity, but increased moduli. In this system, increasing addition of non-solvent resulted in increased CA agglomeration/association arising from the dramatic change in solvent properties. Therefore, based on these results, it seems that although hydrogen bonding is critical for gel formation, the viscoelastic properties and microstructure can be manipulated via both hydrophilic and hydrophobic interactions. 122 3.6 References Appaw, C. (2004). Rheology and Microstructure of Cellulose Acetate in Mixed Solvent Systems. PhD Thesis; Chemical Engineering. Raleigh, NC, North Carolina State University: 224. Appaw, C, Gilbert, R. D., Khan, S. A. and Kadla, J. F. (2007). "Viscoelastic Behaviour of Cellulose Acetate in a Mixed Solvent System." Biomacromolecules 8(5): 1541-1547. Avanza, M., Puppo, M. C. and Anon, M. C. (2005). "Rheological characterization of amaranth protein gels." Food Hydrocolloids 19(5): 889-898. Bakunin, V. N., Parenago, O. P. and Kuz'mina, G. N. (1997). "Molecular association in liquid-phase oxidation of hydrocarbons." Rossiiskii Khimicheskii Zhurnal 41(3): 69-75. Bochek, A. M. and Kalyuzhnaya, L. M. (2002). "Interaction of water with cellulose and cellulose acetates as influenced by the hydrogen bond system and hydrophilic-hydrophobic balance of the macromolecules." Russian Journal of Applied Chemistry 75(6): 989-993. Cai, J. and Zhang, L. (2006). "Unique gelation behavior of cellulose in NaOH/Urea aqueous solution." Biomacromolecules 7(1): 183-189. Castro, J. M., Macosko, C. W. and Critchfield, F. E. (1984). "Effect of Processing Conditions on Premature Gelling, Knit Line Strength, and Physical-Properties for the Rim Process." Journal of Applied Polymer Science 29(6): 1959-1969. Chambon, F., Petrovic, Z. S., Macknight, W. J. and Winter, H. H. (1986). "Rheology of Model Polyurethanes at the Gel Point." Macromolecules 19(8): 2146-2149. Charcosset, C. and Bernengo, J. C. (2000). "Comparison of microporous membrane morphologies using confocal scanning laser microscopy." Journal of Membrane Science 168(1-2): 53-62. Charcosset, C, Cherfi, A. and Bernengo, J. C. (2000). "Characterization of microporous membrane morphology using confocal scanning laser microscopy." Chemical Engineering Science 55(22): 5351-5358. Chauvelon, G., Buleon, A., Thibault, J. F. and Saulnier, L. (2003). "Preparation of sulfoacetate derivatives of cellulose by direct esterification." Carbohydrate Research 338(8): 743-750. da Silva, M. A. and Areas, E. P. G. (2005). "Solvent-induced lysozyme gels: Rheology, fractal analysis, and sol-gel kinetics." Journal of Colloid and Interface Science 289(2): 394-401. Deshmukh, M. M., Gadre, S. R. and Bartolotti, L. J. (2006). "Estimation of intramolecular hydrogen bond energy via molecular tailoring approach." Journal of Physical Chemistry A 110(45): 12519-12523. Eissa, A. S. and Khan, S. A. (2005). "Acid-induced gelation of enzymatically modified, preheated whey proteins." Journal of Agricultural and Food Chemistry 53(12): 5010-5017. 123 Fang, N., Lai, A. C. K., Wan, K. T. and Chan, V. (2003). "Effect of acyl chain mismatch on the contact mechanics of two-component phospholipid vesicle during main phase transition." Biophysical Chemistry 104(1): 141-153. Fang, Y. P., Takahashi, R. and Nishinari, K. (2004). "Rheological characterization of schizophyllan aqueous solutions after denaturation-renaturation treatment." Biopolymers 74(4): 302-315. Gildert, G. R., Matsuura, T. and Sourirajan, S. (1979). "Effect of Different Alcohol-Water Mixtures as Gelation Media During Formation of Cellulose-Acetate Reverse-Osmosis Membranes." Journal of Applied Polymer Science 24(1): 305-310. Gomez-Bujedo, S., Fleury, E. and Vignon, M. R. (2004). "Preparation of cellouronic acids and partially acetylated cellouronic acids by TEMPO/NaCIO oxidation of water-soluble cellulose acetate." Biomacromolecules 5(2): 565-571. Griswold, P. D. and Cuculo, J. A. (1974). "Experimental Study of Relationship between Rheological Properties and Spinnability in Dry Spinning of Cellulose Acetate-Acetone Solutions." Journal of Applied Polymer Science 18(10): 2887-2902. Gross, R. A. and Scholz, C. (2001). Biopolymers from polysaccharides and agroproteins. Washington, DC, American Chemical Society. Hao, J. H. and Wang, S. C. (2001). "Calculation of alcohol-acetone-cellulose acetate ternary phase diagram and their relevance to membrane formation." Journal of Applied Polymer Science 80110V. 1650-1657. Ilyina, E. E., Daragan, V. A. and Prisment, A. E. (1993). "Features of Complexing and the Solution Structure from the Data on Rotational and Translational Mobility of Molecular Probes in Cellulose-Acetate Solutions." Macromolecules 26(13): 3319-3323. Kadla, J. F., Dai, Q. Z. and Khan, S. A. (2005). "Viscoelastic and microstructural changes in sol-gel systems of cellulose acetate in a mixed solvent system." Abstracts of Papers of the American Chemical Society 229: U292-U292. Khalil, S. A. H. (1973). "Phase Separation of Cellulose Derivatives - Effects of Polymer Viscosity and Dielectric-Constant of Nonsolvent." Journal of Pharmaceutical Sciences 62(11): 1883-1884. Kjoniksen, A. L., Nystrom, B. and Lindman, B. (1998). "Dynamic viscoelasticity of gelling and nongelling aqueous mixtures of ethyl(hydroxyethyl)cellulose and an ionic surfactant." Macromolecules 31(6): 1852-1858. Kobayashi, K., Huang, C. I. and Lodge, T. P. (1999). "Thermoreversible gelation of aqueous methylcellulose solutions." Macromolecules 32(21): 7070-7077. Li, L., Thangamathesvaran, P. M., Yue, C. Y., Tam, K. C, Hu, X. and Lam, Y. C. (2001). "Gel network structure of methylcellulose in water." Langmuir 17(26): 8062-8068. Ljunglof, A., Bergvall, P., Bhikhabhai, R. and Hjorth, R. (1999). "Direct visualisation of plasmid DNA in individual chromatography adsorbent particles by confocal scanning laser microscopy." Journal of Chromatography A 844(1-2): 129-135. Lowe, L. L., Foegeding, E. A. and Daubert, C. R. (2003). "Rheological properties of fine-stranded whey protein isolate gels." Food Hydrocolloids 17(4): 515-522. 124 Marangoni, A. G. (2002). "The nature of fractality in fat crystal networks." Trends in Food Science & Technology 13(2): 37-47. Mours, M. and Winter, H. H. (1996). "Relaxation patterns of nearly critical gels." Macromolecules 29(22V. 7221-7229. Nielsen, L. E. (1977). Polymer rheology. New York, M. Dekker. Pugnaloni, L. A., Matia-Merino, L. and Dickinson, E. (2005). "Microstructure of acid-induced caseinate gels containing sucrose: Quantification from confocal microscopy and image analysis." Colloids and Surfaces B-Biointerfaces 42(3-4): 211-217. Rozas, I., Alkorta, I. and Elguero, J. (2001). "Intramolecular hydrogen bonds in ortho-substituted hydroxybenzenes and in 8-susbtituted 1-hydroxynaphthalenes: Can a methyl group be an acceptor of hydrogen bonds?" Journal of Physical Chemistry A 105(45): 10462-10467. Ryskina, I. I. and Averyano.Vm (1971). "Study of Gelatin and Structure of Cellulose Acetate Cels in Benzyl Alcohol." Vysokomolekulyarnye Soedineniya Section A 13(10): 2189. Ryskina, I. I. and Averyanova, V. M. (1975). "Study of Morphology of 3-Dimensional Network of Cellulose Acetate-Benzyl Alcohol Gels." Vysokomolekulyarnye Soedineniya Seriva A 17(4): 919-922. Saalwachter, K. and Burchard, W. (2001). "Cellulose in new metal-complexing solvents. 2. Semidilute behavior in Cd-tren." Macromolecules 34(16): 5587-5598. Shih, W. FL, Shih, W. Y., Kim, S. I., Liu, J. and Aksay, I. A. (1990). "Scaling Behavior of the Elastic Properties of Colloidal Gels." Physical Review A 42(8): 4772-4779. Steiner, W., Lafferty, R. M., Gomes, I. and Esterbauer, H. (1987). "Studies on a Wild Strain of Schizophyllum-Commune - Cellulase and Xylanase Production and Formation of the Extracellular Polysaccharide Schizophyllan." Biotechnology and Bioengineering 30(21: 169-178. Tanaka, F. (2000). "Thermoreversible gelation strongly coupled to polymer conformational transition." Macromolecules 33(11): 4249-4263. Tikhonov, V. P., Lebedev, R. A. and Fuks, G. I. (1982). "Study of the micelle formation of surfactants in nonpolar liquids by infrared spectroscopy. 2. Solutions of fatty alcohols in carbon tetrachloride." Kolloidnyi Zhurnal 44(1): 57-61. Tsunashima, Y. and Hattori, K. (2000). "Substituent distribution in cellulose acetates: Its control and the effect on structure formation in solution." Journal of Colloid and Interface Science 228(2): 279-286. Tsunashima, Y., Kawanishi, H. and Horii, F. M. (2002). "Reorganization of dynamic self-assemblies of cellulose diacetate in solution: Dynamical critical-like fluctuations in the lower critical solution temperature system." Biomacromolecules 3(6V. 1276-1285. Weng, L. H., Zhang, L. N., Ruan, D., Shi, L. H. and Xu, J. (2004). "Thermal gelation of cellulose in a NaOH/thiourea aqueous solution." Langmuir 20(6): 2086-2093. 125 Williams, P. A., Meadows, J., Phillips, G. O. and Tanaka, R. (1992). "Solution Properties of a Hydrophobically Associating Cellulosic Polymer - Electron-Spin-Resonance Spectroscopy." Acs Symposium Series 489: 341-356. Winter, H. H. and Chambon, F. (1986). "Analysis of Linear Viscoelasticity of a Cross-Linking Polymer at the Gel Point." Journal of Rheology 30(2): 367-382. Winter, H. H. and Mours, M. (1997). "Rheology of polymers near liquid-solid transitions." Neutron Spin Echo Spectroscopy Viscoelasticity Rheology 134: 165-234. 126 Chapter 5. Conclusion The low unspecific absorption capacity and biocompatibility of CA makes it an ideal candidate for biological filtration applications, where maximum recovery of proteins is critical. CA filters / membranes have good dimensional stability after autoclaving or steam sterilizing, and are unaffected by temperatures and pressures up to 135°C (275°F) and 130 psi respectively. The performance characteristics of CA membranes depend on the size and uniformity of the pore structure. Commercially, CA micro-filtration membranes are manufactured with pore sizes ranging from 0.1-50 pm. Increasing the pore size of CA membranes from 0.1 pm up to 20 pm can lead to an increase in eluent flow rates from 8.1 ml/min to 1441 ml/min. Therefore, by controlling CA membrane pore dimensions various medical and pharmaceutical applications can be captured ranging from pre-filtration (20 pm) to final or sterile filtration (0.65 pm). The primary objective of this research to examine controlling/tailoring the pore dimension of CA membranes through phase-separation induced gelation. Using controlled solvent/non-solvents systems we were able to alter gel microstructure, and ultimately membrane porosity. In this thesis, we studied the rheological and microstructural behaviour of ternary systems of cellulose acetate, N, /V-dimethylacetamide and various non-solvents (1-Propanol, 2-Propanol, 1-Hexanol, 1-Octanol, 1-Decanol, 1,2-Ethanediol, 1,2-Propanediol, 1,3-Propanediol, 1,4-Butanediol, 1,6-Hexanediol, Glycerol). In Chapter 2, the influence of CA and non-solvent concentration on steady shear viscosity and dynamic viscoelastic properties of various ternary systems was investigated. In the Chapter 3, the effect of hydrophilic and hydrophobic interactions on gelation and the viscoelastic properties of the ternary systems was investigated using different non-127 solvent structures at a constant CA concentration (15 wt%). In view of these findings, the following conclusions can be inferred. Addition of non-solvent to the bulk of the CA/DMAc solutions leads to phase separation in the system. At low non-solvent content, the system appears as a homogenous solution, while at a critical concentration of the non-solvent the system forms a uniform turbid/cloudy and semi-solid system, which exhibits gel-like characteristic. The sol-gel transition is primarily dependent on the content and structure of non-solvents, and to some extent on the CA concentration. Increasing the CA concentration accelerates the sol-gel transition and results in gel formation at a low concentration of the non-solvent. The steady state rheological behavior demonstrated that the addition of a non-solvent to CA solution increased viscosity. At a low concentration of non-solvent, Newtonian behavior was observed followed by shear thinning. The addition of a non-solvent enhanced intermolecular interactions (hydrogen bonding and molecular entanglement). At a critical non-solvent content, the viscosity curve exhibited high viscosity at low shear rates and the disappearance of the zero-shear viscosity plateau represented an improvement in the microstructure of the ternary system. Dynamic rheology spectra shows enhancement in the elastic (G') and viscous (G") modules as the non-solvent content increased. In the frequency sweep mode and at low concentration of the non-solvent, G" is larger than G' and both modules are frequency dependent. The influence of increasing the non-solvent concentration causes clean and neat solutions of polymers to phase separate and eventually form self-supporting gel-like materials. During liquid solid transition and at a defined level of non-solvent content, G' crosses G" or G ~G". The crossover of G'-G" during liquid-solid transition is indicative 128 of the development of a cross-link structure and the formation of critical gels. The addition of non-solvents increased both the modulus by several orders of magnitude, with G' > G" and formation of stronger gels. The Hansen equation and Hildebrand solubility parameters (8) for ED, BD and HD demonstrated that hydrogen bonding forces dominate other forces such as dispersive (Sd) and permanent dipole-dipole (8P). It was shown that the hydrogen bonding solubility parameter index (8h) is significantly affected, while both Sd and 8 P remain relatively constant. This indicates that the hydrogen-bonding interactions may be the major route for initiating the sol-gel process in this system The calculated hydrogen bonding solubility value (8h) showed that 8|, ED > 8h BD > Sh HD. The calculated solubility values also confirmed FTIR spectroscopy, showing that hydroxyl stretching band is shifted to lower wavenumber, indicating of stronger hydrogen bonding interaction in system. In order to better understand the role of hydrophilic interaction in ternary system, the availability of the hydrogen-bonding groups within the non-solvent was varied. It was found that increasing the number of OH groups on the non-solvent from 1 to 3 (1-Pro, 1,3-PD and Gly) increased the viscosity and the modulus through hydrophilic interactions and a lower concentration sol-gel transition was also observed. Likewise, increasing the hydrophobicity of the non-solvents also enhanced gel properties. Also, extension of alkyl chain length within the non-solvent (1-Hex, 1-Oct, 1-Dec) resulted in more molecular entanglement in the system and consequently a faster phase separation. However, the ternary systems with hexane and/or polyethylene glycol as non-solvent did not exhibit gel formation, it was revealed that phase separation and gel formation were greatly induced by competitive hydrogen-bonding between components and hydrophobic 129 interaction by itself is not the main key for gel formation. The effect of O—H intermolecular hydrogen bonding on gel formation was further investigated by using the conformational isomers 1-octanol and dibutyl ether (molecular formula CgHigO) as non-solvents. The 1-octanol ternary system produces a uniform cross-linked gel at 37.3 wt%. By contrast, dibutyl ether did not show phase separation or gel formation. Therefore, it is concluded that the hydrogen bonding interactions in the CA ternary system are essential for obtaining a physical gel, and intermolecular interactions based on proton donor-acceptor groups appear critical. FTIR spectra of the ternary systems showed shifts of the hydroxyl stretching band to lower wavenumber as the non-solvent content increased. Shifting the hydroxyl band to lower wavenumber effects hydrogen bond intensification in the ternary system, resulting in an increase in viscosity and G'. At the intermediate state where G' ~ G", shifting of the hydroxyl band to lower frequency for 1-Pro, 1,3-PD and Gly accounts for modulus and viscosity enhancements through hydrophilic interactions. The opposite shifting of the hydroxyl stretching band in FTIR spectra for ternary systems consisting of 1-Pro, 1-Hex, 1-Oct and 1-Dec as the non-solvent demonstrated the impact of molecular entanglement on modulus enhancement and gelation. The strain spectra demonstrated that further addition of the non-solvents to the ternary system shifted the limits of the linear elastic modulus plateau to lower strain values. This suggested that ternary systems with a large elastic modulus (gels) are made of floes consisting of aggregated macromolecules. A three dimensional gel network structure is formed as a result of the interconnected floes. The solid network CA ternary system is a strong-linked gel in which the links between floes are stronger than the links within the 130 floes. As a result, failure under deformation occurs through the breaking of the intra-floc linkage. This explains the shifts of yield strain to lower strain value with increasing non-solvent concentration. The effect of non-solvent concentration on the on-set point of non-linearity in G' (from strain sweep spectra) showed different rates of change in non-linearity between the non-solvents. The ED system has the largest change in non-linearity, i.e. smallest slope, followed by BD, then HD (ED = -0.80; BD = -1.09; HD = -1.26). This implies that the ED system may contain relatively stronger inter-floc linkage than those in BD, which are stronger than in HD. This was further confirmed with FTIR spectra of the ternary systems. In the case of the hydrophobic non-solvent systems, the onset of non-linearity for the gel samples occurred over a small range of % strain. This may imply that the ternary systems with longer alkyl chain of mono-hydroxyl alcohol contained relatively weaker inter-floc linkage than those of shorter chains. Using rheological technique, in all CA ternary systems, G' exhibited a power-law behaviour with non-solvent content (<Z>), G'~ 0 11. The power-law dependence of G', together with the similar values for n observed of gels (Z)=2.9±0.05) suggests that the gels are fractal in nature, and that they are made through an aggregation mechanisms. However, the power-law dependence of G' is different between non-solvents and increasing both hydrophobicity and hydrophilicity within non-solvents decreased the power-law exponent values. It was found that the fractal dimension of the gels obtained through confocal microscopy have similar values (1.90 ±0.05), regardless of CA concentration and type of non-solvent. The different values of fractal dimensions based on the two above methods are related to two different dimensional Euclidean spaces for confocal (2) and 3-D for rheology. A good agreement between fractal dimensions 131 determined by rheological methods and microscopy methods for all CA ternary systems showed that the fractal scaling theories are theoretically valid. Microscopy images (LSCM and SEM) exhibited uniform packing in the polymer network structures as CA concentration increased. The agreement between the LSCM reflective images and SEM images was good. The LSCM images confirmed the rheological results and a denser network structure with more rigid gel was observed as the non-solvent content varied in the ternary system from ED to BD and to HD, respectively. Increasing the hydrophilicity among non-solvent structure in CA ternary system showed more aggregated and denser network microstructure with changing from 1-Pro to 1,3-PD to Gly. Alternatively ternary system at high CA concentration (20 wt%) showed the smaller micro-porous membrane microstructures compare to lower CA concentration (10 wt%). It was also found from LSCM images (reflective mode) that the presence of mono-hydroxyl non-solvent (1-Pro) forms an open network microstructure (pore dimension >5 pm), and increasing the length of alkyl chain (1-Hex < 1-Oct < 1-Dec) led to formation of larger micro-porous membrane microstructures (up to ~60 pm). 5.1 Recommendation This research has examined the various aspects of changing the CA and non-solvent compositions in ternary system. Many different non-solvent structures were used to understand the molecular interactions, which account for phase separation and gel formation. However, there are several other parameters that could be considered to improve our understanding of molecular interactions in the ternary systems. It was assumed that phase separation is influenced by the number of hydroxyl groups available in polymers and non-solvents. Therefore, studying the sol-gel transition at 132 differing degrees of substitution of CA would be a good approach to understand the competitive hydrogen bonding in the ternary system. The modification of polymers and the use of regioselective CA (acetyl group on C-2,3 and hydroxyl on C-6) would be an interesting approach, which could reveal the effect of intra- and intermolecular hydrogen bonding within polymers and other components. Another possible approach would be using CA with different molecular weights as well as other cellulose derivatives with different functional groups, such as cellulose ester, cellulose ether and nitrocellulose to study their possible phase separation/gel formation and their mechanism of interactions for each polymer in the ternary system. 133 Appendix 134 Part A: Figures and Tables 101 PH oo O o GO • r H 10" v v V r -A A A , A A A A , A A A A A A A A A A A A A A A , F + + + + + + + - H - - H - + A A A A A A , •+++++. A V •+++. V X X X X X ' X X X X X X X X X X X X X X X X X X X X X X X ; X X X x o • o X -t-A V ED 0 (wt%) ED 5 (wt%) ED 10 (wt%) ED 26.6 (wt%) ED 33.3 (wt%) ED 35.0 (wt%) ED 36.6 (wt%) m x r x r x r x n n x D i r x r x r x a ^ ^ 10" 10"' 10" 10' 10' 10J 10' Shear rate (1/s) Figure Al: Effect of different ED contents on steady shear viscosity in CA/DMAc/ED system at CA 15wt% concentration io-cs PH O o o o > 10' 10" v ^ v W v w v v v w v v v v v v \ ^ < < 4 A A A A A A A A A A A A A A A A A A A A A A A A A ^ ^ ' g j HD 0 (wt%) HD 5 (wt%) HD 10 (wt%) HD 20 (wt%) HD 26.6 (wt%) HD 30.0 (wt%) HD31.6(wt%) HD 33.3 (wt%) HD 35.0 (wt%) ++++++ -| -H -H~ I~ I - +++++ -H -++ -| -+++ +++++ J P T I I I I I I I I II M I I I I I I I I ITTTrm OCODOOOCOOOCOOOOCOOOOCOOO^ 1 I I I 10" lO"' 10" 10' io- 10J 10* Shear rate (1/s) Figure A2: Effect of different HD contents on steady shear viscosity in CA/DMAc/HD system at CA 15wt% concentration 135 10"' 1 1 i i i i i i I -5 0 5 10 15 20 25 30 35 H D content (%) Figure A 4 : Viscosity as a function of HD content obtained at shear rate of Is"1 for five different CA concentrations 136 03 10° 10" 10-10u 10" 10 IO" ^^^^TATATAATAAAAAAAAA ^ ^ ^ ^ AA "IK A + + O _J •go 9BE 10"' 10" 10' I O ' o ED 26.6 (wt%) • ED 28.3 (wt%) 0 ED 30.0 (wt%) X ED 31.6 (wt%) + ED 33.3 (wt%) A ED35.0(wt%) V ED 36.6 (wt%) ED 38.3 (wt%) A ED 40.0 (wt%) < ED41.6(wt%) t> ED 43.3 (wt%) 10J Angular frequency (rad/s) Figure A5: Dynamic frequency sweep (G') experiments at CA concentration of 15wt% having different ED content 03 b i o 5 10J 10' 10 10" 10" AAAA AAA' AAA' AAA' A A AA A A A A - i i AAA' AT / / S " O r y 10 10u 10 10" Angular frequency (rad/s) o • © X + A A V < t> BD 26.6% BD28.3% BD 31.6% BD 30% BD33.3% BD 35% BD 36.6% BD 38.3% BD 40% BD41.6% BD 43.3% 10' 10" Figure A6: Dynamic frequency sweep (G') experiments at CA concentration of 15wt% having different BD content 137 03 fe b •10° 10* 10" 10" 10" 10" 10" AAAAAAAAAAAAAAAAAAAAAAAAAA^ \AAA AA o HD 26.6 (wt%) • HD 30.0 (wt%) o HD31.6(wt%) X HD 33.3 (wt%) + HD 35.0 (wt%) A HD 36.6 (wt%) V HD 38.3 (wt%) HD 40.0 (wt%) 10 10" 10' 10" 10J Angular frequency (rad/s) Figure A7: Dynamic frequency sweep (G') experiments at CA concentration of 15wt% having different HD content § 03 fl 100 80 60 40 20 -20 20 CA10% CAI 2.5% CAI 5% CAI 7.5% CA20% 25 45 50 30 35 40 E D content (wt%) Figure A8: Effect of ED content on transmission intensity at five CA concentrations 138 _20 1 1 1 1 i I 20 25 30 35 40 45 BD content (wt%) Figure A9: Effect of BD content on transmission intensity at five CA concentrations. _20 1 — 1 1 i i I 20 25 30 35 40 45 HD content (wt%) Figure A10: Effect of HD content on transmission intensity at five CA concentrations 139 I0"4 1 1 1 1 1 1 1 lO"2 I 1 1 1 ! 1 I 10"2 10"' 10° 10' 10! 10' I04 I0"2 10"' 10° 10' 102 I0J 104 Osc. stress (Pa) Osc. stress (Pa) 10 10' 10' 10' . Osc. stress (Pa.) CA 17.5 wt% AAAAAAAAAAAAA, ED 23.3% I ED 25% ED 26.6% I ED 28.3% | ED 30% ED 31.6% I ED 33.3% | ED 35% ED 36.6% I ED 38.3% | ED 40% ED 41.6% I ED 43.3% 10"' 10° 10' I02 10' Osc. stress (Pa) I04 C A 20 wt% AAAAAAAAAAAAAAAAAAAAAAAAAtftA j^j^  AAAAAAAAAAAAAAA^. ED 20% ED 21.6% I ED 23.3% | ED 25% ED 26.6% I ED 28.3% I ED 30% ED 31.6% ED 33.3% I ED 33.3% [ ED 35% ED 36.6% I ED 38.3% | ED 40% 10"' 10° 10' 102 10"' Osc. stress (Pa) Figure All: Stress sweep experiments conducted for a) 10 wt.% CA, b) 12.5 wt.% CA, c) 15 wt%, d) 17.5 wt%, and e) 20 wt% CA samples with varying ED contents. 140 10 ioJ io2 io1 10° 10"' IO"2 I0'J C A 10 wt% 10"' 10 10 10 Osc. stress (Pa) 10! BD 26.6% BD 28.3% BD 30% BD 31.6% BD 33.3% BD 35% BD 36.6% BD 38.3% BD 40%, BD41.6% BD 43.3%, BD 45%, BD 46.6% 10" ~ 10" O CA 12.5 wt% 10 10' 10' Osc. stress (Pa) o BD26.6% • BD28.3% 0 BD 30% X BD31.6% + BD33.3% BD 35% V BD 36.6% BD 38.3% A BD 40% < BD41.6% BD 43.3% * BD 45% 10" 10" ioJ >0O<*5OOOOO0000<i<3 o BD 26.6%, • BD 28.3% o BD 30% X BD31.6% -t- BD 33.3% A BD 35% t\ BD 36.6% A BD 38.3% < BD 40% > BD41.6% * BD 43.3% 10-' 10° 10' 10' Osc. stress (Pa.) CA 17.5 wt% BD 26.6% BD 28.3% BD 30% BD31.6% BD 33.3% BD 35% BD 36.6% BD 38.3% 40% . 10° 10' 102 Osc. stress (Pa) 10" IO4 b NNNMS ^ ^^ ^^ ^^ ^^  NNNNNNNKMNJSNKNNN N t.\ C A 20 wt% yvW77WVWVVVWVVVVVy^vwvvvv7vVv?v v AAAAAAAAAAAAAAAAAAAAA^rXA^  10° 10' 10' ' 10"' Osc. stress (Pa) o BD 23.3% • BD 25% o BD 26.6% X BD 28.3% -t- BD 30% A BD 31.6% V BD 33.3% BD 35% A BD 36.6% Figure A12: Stress sweep experiments conducted for a) 10 wt.% CA, b) 12.5 wt.% CA, c) 15 wt%, d) 17.5 wt%, and e) 20 wt% CA samples with varying BD contents. 141 CA 10 wt% +++++ + * X x * x x x 0 HD 30% D HD33.3% O HD35% X HD36.6% -t- HD38.3% A HD40% V HD4l% C HD43.3% A HD45% x x oooooc ° ° o o ' o o x x x fiea o o 0 0 ( IO'2 10'' 10° io' Osc. stress (Pa) IO'1 10° 10' IO2 Osc. stress (Pa) 1 1 1 1 CA12.5wt% d t, c C Ci t i c cC L ii d ti d ti e. d Ci d1. d c d d d t. ti d A A A A A A A A A A A A A A A A A A A A A A A A A A A A w v w v y v v v v v v v v v v v v v v v v v v v v v o HD 30% • HD3I.6% o HD33.3% X HD35% + HD36.6% A HD40% " HD 38.3% d HD43.3% " HD4I.6% x x x x x x x x x x x x x ^ + + ^ + + + + 4 * ^ ' l > x x x x x x x x x x x x x x o o o o o o o o o o o 0 0 o *x>< DDODDanDBFJBgggggggggggggggg^ -— Iff ' ° IO2 10' h C A 15 wt% W W V VW VWV W W V W W V V V W W W V ^ wwv^ AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAftA^ HD26. HD30, HD3I. HD 33. HD35. HD 36. HD38. HD 40. 6 (wt%) 0 (wt%) 6 (wt%) 3 (wt%) 0 (wt%) 6 (\vt%) .3 (wt%) 0 (wt%I CX^OCOCCOCOCOOOOOCOCOCOCOCCXDCCCOO^^tb 10° 10' IO2 Osc. stress (Pa) 10 10' I O 4 '° I02 10' CA 17.5 wt% 7 W J T O W W W T O W W W W 7 7 J V 7 W 7 7 W 1 •WVvvv^j AAAAAAAAAAAAAAAAAAAA^  ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ COCCCOOCCXXXXTCCOOOCKOOCXXOCOCCCOCO-^)^  IO'2 10"' 10° 10' 102 10' Osc. stress (Pa) HD HD HD HD HD HD HD HD HD 23.3% | 25% 26.6% I 28.3% | 30% 31.6% I 33.3% I 35% 36.6% I 10" CA 20 wt% LNJvNNNNKKN\NNNsNNNNr.NNNN^ r.t^ r^.|r: L U U J ^ U ^ . W W V V W v Y W V V W W V W V W W W AAAAAAAAAAAAAAAAAAAAAAAAAAAAA^AAAAAAAJ^^  10'' 10 10 10 Osc. stress (Pa) o HD 23.3% • HD 25% o HD 26.6% X HD 28.3% + HD 30% A HD 30% V HD31.6% HD33.3% A HD 35% I04 Figure A13: Stress sweep experiments conducted for a) 10 wt.% CA, b) 12.5 wt.% CA, c) 15 wt%, d) 17.5 wt%, and e) 20 wt% CA samples with varying HD contents. 142 C A / D M A c 15wt% D M A c / E D 8.25/6.75g C A / D M A c / E D 2.25/8.75/4 g C A / D M A c / E D 2.25/6/6.75 g 3700 3500 3300 3100 cm Figure A14: FTIR analysis of the various components in the CA/DMAc/ED ternary system. Included in the figure are the FTIR spectrum of viscous solution and gel samples. C A / D M A c 15wt% D M A c / H D 8.25/5.75g C A / D M A c / H D 2.25/8.75/4 g C A / D M A c / H D 2.25/7/5.75 g 3700 3500 3300 3100 cm Figure A15: FTIR analysis of the various components in the CA/DMAc/HD ternary system. Included in the figure are the FTIR spectrum of viscous solution and gel samples. 143 s c3 • i fl <u oo 16.5 16 15.5 15 14.5 14 13.5 h 13 IT| I I I | •• •• • 10'' 10° 10' 102 103 104 10s G' (Pa) Figure A16: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, ED as non-solvent) 2 fl fl j> "o 00 17 16 15 14 13 12 11 IT] | 5d • • • • 8h 5p ul I i , i i , n i l 10"1 10° 10' 102 103 104 10s 106 G' (Pa) Figure A17: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, HD as non-solvent). 144 03 fe b io 3 10* 10* 10" io-A A A A A / S A A A A A A A A A A A A A A A A A A A A A A A A A A A ^ • 3 i i r n 3 i i i i i i i i x n i i x i i T O 0 ED(38.3wt%) o • ED (40.0wt%) 0 ED(41.6wt%) A ED (43.3wt%) io" 10 10" 10" 10" 10u 10' Strain % Figure A18: Elastic modulus of CA 15 wt% gels at different concentrations of ED as a function of strain. The limit of linearity shifts to lower strain values as concentration increases. 03 fe b 10° io 3 10* 10' 10z 10' 10" A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A ^ ^ >A A 'Oo, 1 H I T T T T I I I i m i 0 HD (35.0wt%) • HD (36.6wt%) O HD (38.3wt%) A HD (40.0wt%) o o o D o c a x x D o r a x Q ^ ^ 10'6 10"5 10"4 10"3 10'2 10"' 10° 101 Strain % Figure A19: Elastic modulus of CA 15 wt% gels at different concentrations of HD as a function of strain. The limit of linearity shifts to lower strain values as concentration increases. 145 Figure A20: SEM micrograph images of gels at same elastic modulus (G') for ED, BD and HD as non-solvent in three different CA concentrations (10, 15 and 20wt.%), all SEM images are at same magnification (xl.Ok) 146 Figure A21: SEM micrograph images of gels at same elastic modulus (G') for ED, BD and HD system in 10wt.% CA concentration at three different magnifications (xl.Ok, x2.0k and x3.0k) Figure A22: SEM micrograph images of gels at same elastic modulus (G') for ED, BD and HD system in 15wt.% CA concentration at three different magnifications (xl.Ok, x2.0kand x3.0k) 148 cd fe O o > 10" 10' o 1,3-PD (26.6wt%) • 1,3-PD (28.3wt%) o 1,3-PD (30.0wt%) X 1,3-PD (31.6wt%) + 1,3-PD (33.3wt%) A 1,3-PD (35.0wt%) o~ x x o o o o o o o o o o o o o o o o o o o o o o o o o o o § o ^ B + 10"2 10"' 10° 10' 102 103 Shear rate (s"1) Figure A23: Effect of different 1,3-PD contents on steady shear viscosity for ternary system of CA/DMAc/l,3-PD system at constant CA 15wt% concentration 10" cd PH O o CO > 10z 10' h ~+++ + + 0 Gly (23.3wt%) • Gly (25.0wt%) 0 Gly (26.6wt%) X Gly (28.3wt%) + Gly (30.0wt%) oooo8888888888BBBr! 10"2 Iff' 10° 101 IO2 103 Shear rate (s"1) Figure A24: Effect of different Gly contents on steady shear viscosity for ternary system of CA/DMAc/Gly system at constant CA 15wt% concentration 149 c3 PH O o 10" 10' 10" 10"' o Hex (35.0wt%) • Hex (36.6wt%) o Hex (38.3wt%) X Hex (40.0wt%) 10" 10" 10' 10" 10J Shear rate (s" ) Figure A25: Effect of different 1-Hex content on steady shear viscosity of CA/DMAc/1-Hex ternary system at CA 15wt% cd PH O o > 10" 10' 10" 10" AAAAAAAA A A A AAAAAAAAA A A . O • o X + A Oct(5.00wt%) Oct(10.0wt%) Oct (20.0wt%) Oct (26.6wt%) Oct (33.3wt%) Oct (34.6wt%) A A A A A A A A X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 10" 10" 10" 10' 10" 10J Shear rate (s" ) Figure A26: Effect of different 1-Oct content on steady shear viscosity of CA/DMAc/1-Oct ternary system at CA 15wt% 150 ca PH co O o co > 10J 10" 10' 10" 10" 10" 10" 0 00 XX x x x u a D D D D D D D 0 Dec (28.3wt%) • Dec (30wt%) o Dec(31.6wt%) X Dec(33.3wt%) X oooooooooooooooooeooooooooooooooo® X o • o X o o x • o x ° "x x x ^ o x x X x x x n 10" 10" 10' -1\ 10" 10J Shear rate (s" ) Figure A27: Effect of different 1-Dec content on steady shear viscosity of CA/DMAc/l-Dec ternary system at CA 15wt% ca PH O o 10' io-10' V V V v V v v ^ K AAAAAAAAAAAAAAAASBE +++++++++++++++++++ o 1,2-PD (33.3wt%) • 1,2-PD (35wt%) o 1,2-PD (36.6wt%) X 1,2-PD (38.3wt%) + 1,2-PD (40.0wt%) A 1,2-PD (41.6wt%) V 1,2-PD (43.3wt%) 1,2-PD (45.0wt%) XXXXXXXXXXXXXXXXXXXXXXX t^SKv. ' A oooooooooooooooooooooooooSS xx$ ooooooooooooo AO 10"2 10"' 10° 10' 102 103 Shear rate (s"1) Figure A28: Effect of different 1,2-PD content on steady shear viscosity of CA/DMAc/1,2-PD ternary system at CA 15wt% cci PH cfl O o 10-10' 10u 10" 10" + + 4 o 2-Pro (40.0wt%) • 2-Pro (41.6wt%) o 2-Pro (43.3wt%) X 2-Pro (45.0wt%) + 2-Pro (46.6wt%) +++, X X X X X X X X X X X X X X X X X X X X X X X * * * y v • • • • • • • • • • • • • • • • • • D D D D D Q D Q H 8 S 8 8 X x X O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O Q Q S ^ *6 • 10" 10u 10' io- 10 Shear rate (s" ) Figure A29: Effect of different 2-Pro content on steady shear viscosity of CA/DMAc/2-Pro ternary system at CA 15wt% CS PH cn O O cfl •,—i > -0.1 0.1 0.2 Mole (equivalent OH) 0.3 0.4 Figure A30: Effect of hydrophilic interactions on the viscosity enhancement of the CA/DMAc/non-solvent solutions at calculated equivalent hydroxyl group (values obtained at a shear rate of 1 s"1). 152 DMAc/l ,3-PD 8.75/6.25g CA/DMAc/ l , 3PD 2.25/6.5/6.25g 3800 3600 3400 3200 3000 cm" 1 Figure A31: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/1,3-PD system. A 3800 3600 3400 3200 3000 cm"1 Figure A32: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/Gly system. 153 3800 3600 3400 3200 3000 cm"' Figure A33: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/Hex system. 3800 3600 3400 3200 3000 cm"1 Figure A34: FTIR spectra of the Hydroxyl stretching region of the CA/DMAc/Oct system. 154 A 3800 3600 3400 3200 3000 c m Figure A35: FTIR spectra of the OH stretching region of the CA/DMAc/Dec system. b 10" 10 10" 10"-10 7& 7 W 7 W ^ X o 1-Pro (38.3 wt%) D l-Pro(41.6wt%) O 1-Pro (40.0wt%) X 1-Pro (43.3wt%) + 1-Pro (45.0wt%) A 1-Pro (46.6wt%) V 1-Pro (48.3wt%) 1-Pro (50.0wt%) A 1-Pro (51.6wt%) < 1-Pro (53.3wt%) 10" 10 10 Angualr frequency (rad/s) 10z 10J Figure A36: Dynamic frequency sweep (G1) experiments for CA concentration of 15wt% having different 1-Pro content 155 T T T fc O 10° 10z 10" 10" 10" « « « < ' A A 3 8 F ^ 0 ° «gP o° • o° -42 1 10" 10" 10' 10' o 1,3-PD (26.6wt%) • 1,3-PD (28.3wt%) o 1,3-PD (30.0wt%) X 1,3-PD (33.3wt%) + 1,3-PD (31.6wt%) A 1,3-PD (35.0wt%) V 1,3-PD (36.6wt%) 1,3-PD (38.3wt%) A 1,3-PD (40.0wt%) < 1,3-PD (41.6wt%) t> 1,3-PD (43.3wt%) o 1,3-PD (45.0wt%) 1 103 104 Angular frequency (rad/s) Figure A37: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 1,3-PD content fc b 10 10" 10" 10" +++++ x o ^ k ° ° x©° n D o ° xgo „nno o O ° ° 10" 10' 10 o Gly (23.3wt%) • Gly (25.0wt%) o Gly (26.6wt%) X Gly (28.3wt%) + Gly(30.0wt%) A Gly (31.6wt%) 0 Gly (35.0wt%) Gly (36.6wt%) 10J Angular frequency (rad/s) Figure A38: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different Gly content 156 10° 10 4 1 0 " 10" 10" 10"4 10" x x x x : c x x x x x x x x x x x x x x x x x x x xx> o Hex (35.0wt%) • Hex (36.6wt%) o Hex (38.3wt%) X Hex (40.0wt%) + Hex(41.6wt%) A Hex (43.3 wt%) V Hex (45.0 wt%) 10"1 10" 10' 10" 10 J Angular frequency (rad/s) Figure A39: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 1 -Hex content 0. b 10" 10" 10" lO"4 AAAAA . ^ ^ W ^ W ^ - - * • • A . A A A A A A A A A A A A AAAAAAAAAAAAAAAAAAAA^^ xx: x x x x : x x x x x x : x x x + ^ x x x x x x x x x p o Oct(33.3wt%) • Oct (34.6wt%) o Oct (36.0wt%) X Oct (37.3wt%) + Oct (38.6wt%) A Oct (40.0wt%) V Oct (41.6wt%) Oct (43.3wt%) A Oct (45.0wt%) 10" 10" 10' 10z 10J Angular frequency (rad/s) Figure A40: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 1-Oct content 157 fe b 10° 10 10z 10" 10" IO"4 10" AJ AAAAA AAAAA A A A AAAAAA, AAAAAA' v v v v v v v v v v v v ^ >.++ x F o Dec (26.6wt%) • Dec (28.3wt%) 0 Dec (30.0wt%) X Dec(31.6wt%) + Dec (33.3wt%) A Dec (36.6wt%) V Dec (35.0wt%) L\ Dec (38.3wt%) A Dec (40.0wt%) io- 10" 10' 10z 103 Osc. stress (Pa.) Figure A41: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 1-Dec content b 10" 10" 10/ 10" 10" 10" 10" [X>[\>l\>|\.l\kL\.L\. AAAAA +++++ W W W v v v v v v v v v w v v v v v w , AA AA + + + + + . . A A A 4 ' AAAAA" + + + +++ ++++ X ' y X X < , X X ^oi O n S O . o n D ° o o o o _J io- 10" 10' o • o X + A V ts 2-Pro(41.6wt%) 2-Pro (43.3wt%) 2-Pro (45.0wt%) 2-Pro (46.6wt%) 2-Pro (48.3wt%) 2-Pro (50.0wt%) 2-Pro (51.6wt%) 2-Pro (53.3wt%) 10z 10J Osc. stress (Pa.) Figure A42: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 2-Pro content 158 10° fe b 10' 10 10" 10' 10" vv vvvvvvvvvvvvvvv vvvvvvvvvvvvv + +- + X X X C<>r x x x Ao§g8 C x x 5 Qo o o 10" 10' 10' o • o X + A V [\ A o 10' 1,2-PD (33. 1,2-PD (35. 1,2-PD (38. 1,2-PD (41. 1,2-PD (43. 1,2-PD (45. 1,2-PD (46. 1,2-PD (50. 1,2-PD (51. 1,2-PD (53. 3wt%) 0wt%) 3wt%) 6wt%) 3wt%) 0wt%) 6wt%) 0wt%) 6wt%) 3wt%) 104 Angular frequency (rad/s) Figure A43: Dynamic frequency sweep (G') experiments for CA concentration of 15wt% having different 1,2-PD content 03 fe b to 3 10' 10' 10"' 10"' -0.1 0.1 0.2 0.3 Mole (equivalent OH) 0.4 0.5 Figure A44: Effect of increasing non-solvent content on the elastic modulus (G' at 1 rad s"1) for the CA/DMAc/non-solvents ternary systems at calculated equivalent OH group of non-solvents. (CAI5 wt%) 159 cd b 10° 10s 104 io 3 io2 io1 10° IO"1 IO"2 10"' ^A < NsISMSNNMSMsNNtsMstSMStsNSNSMSt W V v A A A A 7 W V V W W W V W W s A 7 V v A A A A A A A A A A A A A A A A A A A A A A A ^ I I I I I I I I I I I I I I H+H-++++ AA o • o X + A V ts. A < Pro. (38.3 wt.%) Pro. (40.0 wt.%) Pro. (41.6 wt.%) Pro. (43.3 wt.%) Pro. (45.0 wt.%) Pro. (46.6 wt.%) Pro. (48.3 wt.%) Pro. (50.0 wt.%) Pro. (51.6 wt.%) Pro. (53.3 wt.%) 10" 10" 10' 10" 10J 10" i o ' Osc. stress (Pa.) Figure A45: Stress sweep spectra (C) conducted at constant CA concentration of 15wt% with varying 1 -Pro content cd PH 107 io6 io5 io 4 io 3 io2 io' 10° 10"1 10" i r i r <w<KKm<i<»ci<Ki<mci<KKmo^ IsNlstsNslslsl /]/| /] /|/|/l/1/l/1/l/1/l/l/l/MAV^VVW/1/l/14j^ AAAAAAAAAAA o • o X - t -A V ts A < t> 1,3-PD (26. 1,3-PD (28. 1,3-PD (30. 1,3-PD (31. 1,3-PD (33. 1,3-PD (35. 1,3-PD (36. 1,3-PD (40. 1,3-PD (38. 1,3-PD (41. 1,3-PD (43. 1,3-PD (45. 1,3-PD (46. 6 wt%) 3 wt%) 0 wt%) 6 wt%) 3 wt%) 0 wt%) 6 wt%) 0 wt%) 3 wt%) 6 wt%) 3 wt%) 0 wt%) 6 wt%) 10" 10" 10' 10" 10' 10" 103 Osc. stress (Pa.) Figure A46: Stress sweep spectra (G') conducted at constant CA concentration of 15wt% with varying 1,3-PD content 160 03 fe b 10° 10° 10* 10' io-10' 10u W V V W W W W V W W W W s ' V W V W v V , A A A A A A A A A A A A A A A A A A A A A A A A ^ ^ J ^ ^ V ^ AA A V "+-K o Gly (23.3 wt.%) • Gly (25.0 wt.%) o Gly (26.6 wt.%) X Gly (28.3 wt.%) + Gly (30.0 wt.%) A Gly (31.6 wt.%) Gly (33.3 wt.%) Gly (36.6 wt.%) 10"2 10"' 10° 101 102 103 104 105 Osc. stress (Pa.) Figure A47: Stress sweep spectra (G') conducted at constant CA concentration of 15wt% with varying Gly content 03 10° 105 10* 10' 10" 10' 10° 10" 10" -1 I I I I I I I I I I I I I I I I I I I I I V A A A •+++ o Hex (35.0wt%) • Hex (36.6wt%) o Hex (38.3wt%) X Hex (40.0wt%) + Hex(41.6wt%) A Hex (43.3wt%) v Hex (45.0wt%) *xx. •xxx. •>s< 10" 10" 10' 10" 10' 10* Osc. stress (Pa.) Figure A48: Stress sweep spectra (G') conducted as constant CA concentration of 15wt% with varying 1-Hex content 161 cS PH, O 10° 105 104 103 102 10' 10° lO"1 lO"2 i o -A A A ^ A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A -H-+-H-H~r~H-++++-r++++++- "+++++++++ r r r r r r r m i i i i i r t n o Oct(33.3wt%) • Oct (34.6wt%) o Oct (36.0wt%) x Oct(37.3wt%) + Oct (38.6wt%) A Oct (40.0wt%) v Oct(41.6wt%) ^ Oct (43.3wt%) A Oct (45.0wt%) 10" 10u 10' 10" 10* Osc.stress (Pa.) Figure A49: Stress sweep spectra (G') conducted as constant CA concentration of 15wt% with varying 1 -Oct content cS PH_ b 10° i o 5 104 103 102 101 10° IO"1 10'2 A^A^^AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA^ ts V V V V V W V V V V V V V W W W W W W W W V W V V y AAAAAAAAAAAAAAAAAAAAAAAAAAAA ++1H-++++-l++++++++-H-+-H-+-H-++++ X X X X X X X X X X X X X X X X X X v °0 0 o • o X + A V 1\ A Dec (26. Dec (28. Dec (30. Dec (31. Dec (33. Dec (35. Dec (36. Dec (38. Dec (40. 6wt%) 3wt%) 0wt%) 6wt%) 3wt%) 0wt%) 6wt%) 3wt%) 0wt%) IO'" 10" 10u 10' 102 Osc. stress (Pa.) 10 10* Figure A50: Stress sweep spectra (G') conducted as constant CA concentration of 15wt% with varying 1 -Dec content 162 1 1 1 1 1 WVVWVvWVWVWVWVVVVVVVVVVVVVv AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA V ,+++^++^+4+++++-H-+-r++-H-++++++ + V + + + xxxxxxxxxxxxxxxxxxxxxxxx X x x X X OOOOOOOOOOOOOOOOOOOOOOOoo x x o 2-Pro (41.6wt%) • 2-Pro (43.3wt%) o 2-Pro (45.0wt%) x 2-Pro (46.6wt%) + 2-Pro (48.3wt%) A 2-Pro (50.0wt%) v 2-Pro (51.6wt%) ^ 2-Pro (53.3wt%) - . 1 1 1 1 1 I O - 2 10_l 10° 101 102 103 104 Osc. Stress (Pa.) Figure A51: Stress sweep spectra conducted as constant CA concentration of 15wt% with varying 2-Pro content 03 PH 10° 105 104 103 102 101 10° 10"' IO"" AA VNAAAAA/VWW 10" 10" 10' 10" 10' o 1,2 -PD (33.3wt%) • 1,2 -PD (35.0wt%) © 1,2 -PD (36.6wt%) X 1,2 -PD (38.3wt%) + 1,2 -PD (40.0wt%) A 1,2 PD(41.6wt%) V 1,2 PD (43.3wt%) 1,2 PD (45.0wt%) A 1,2 PD (46.6wt%) <l 1,2 PD (48.3wt%) I> 1,2 PD (50.0wt%) * 1,2 PD (51.6wt%) ffl 1,2 PD (53.3wt%) 10* 10' Osc. Stress (Pa.) Figure A52: Stress sweep spectra (G') conducted as constant CA concentration of 15wt% with varying 1,2-PD content 163 PH ,0 3 "o 00 17 16 15 14 13 12 11 TTf I I I I I 111| 1 I I I I I11| 1 rTTTTTTj 1 111 1-Pro lil I I -e s e 10"2 10"' 10° 10' 102 I O 3 I O 4 105 106 G' (Pa) Figure A53: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, 1-Pro as non-solvent) o a u cs C H d w j> 0 0 17 16 15 14 13 "I '"i 1 10" Q B-B-1,3-PD -a—B—B B-B B-B JII—1 1 11 mi! 1 1 1 1 1 10u 10" 10" 10° G' (Pa) Figure A54: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, 1,3-PD as non-solvent) 164 B 2 c3 OH 3 fl CD "o 00 16.5 16 15.5 15 14.5 14 13.5 13 e — e — e — e -•—a—• • 10"' 10"' Gly - e — H H 10° 101 102 103 104 10s G' (Pa) Figure A55: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, Gly as non-solvent) s-<D •*-» <D a 2 cd OH X> 3 C « "3 0 0 17 16 15 14 13 12 11 10 1 1 1 111I1 '1 1 1 1 1 Mill! 1 '1 1 1 1 1 - —- ©—© - 1-Hex " - 5h -% 1 1 i )-3 IO"1 IO1 103 105 G' (Pa) Figure A56: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, 1-Hex as non-solvent) 165 2 OH fl fl 17 16 15 14 13 12 11 10 m r m — i - m i m r - i i n m TTTTFT]—r -TTTrmr 10"; 10" 10' 1-Oct 10J 10' G' (Pa) Figure A57: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, 1 -Oct as non-solvent) U< 2 OH o CO j> "o cvo 17 16 15 14 13 12 11 10 TT] 1 I I I I 1 TTTT TTT| 1—i i 111111 1 — r r i T r n p 1-Dec 10"3 10"2 10"' 10° io ' 102 103 104 10s G' (Pa) Figure A58: The influence of the individual solubility parameter index on G' at a fixed frequency 1 rad/s (CA 15 wt%, 1-Dec as non-solvent) 166 1-06 002844 HD10.9mm 16.0kV x5Q0 lOOum 19-Oct-06 002865 WD 9.5mm 16.0kV xSOO lOOum 19-Oct-06 002900 HD 9.0mm 16.0kV x500 lCOum Figure A59: SEM micrograph images of gels at same elastic modulus (G') for 1-Pro, 1,3-PD, Gly, 1-Hex, 1-Oct and 1-Dec as non-solvents at fixed CA concentrations (15wt.%), all SEM images are at same magnification (500) 167 Part B: Experimental Materials and Methods Materials Cellulose acetate (CA - M„ ca. 30,000 g/mol, degree of acetylation (DA) = 2.45 (39.7 wt% acetyl content)), jV,7V-dimethylacetamide (DMAc - HPLC grade), 1-Propanol (1-Pro), 2-Propanol (2-Pro), 1-Hexanol (1-Hex), 1-Octanol (1-Oct), 1-Decanol (1-Dec), 1,2-Ethanediol (1,2-ED), 1,2-Propanediol (1,2-PD), 1,3-Propanediol (1,3-PD), 1,4-Butanediol (1,4-BD), 1,6-Hexanediol (1,6-HD) and Glycerol (Gly) were purchased from Sigma-Aldrich and used as received. Sodium acetate buffer (0.1M) was purchased from Fisher Scientific. Calcofluor white (M2R) and 25% glutaraldehyde were obtained from Sigma-Aldrich. Aluminum SEM mounts, Pin Type (12.7 mm diameter slotted head, 3.2 mm pin) and carbon mounting tape (12.7 mm diameter) were purchased from Canemco & Marivac. Sample Preparation The first step in the preparation of the various CA systems was to prepare a bulk solution of CA (28.12 wt%) in DMAc. To facilitate mixing and produce a homogeneous solution, the initial mechanical mixture of CA/DMAc was heated to 100°C in an oven for 20 min. The mixed solvent systems were then prepared by adding the CA/DMAc bulk solution to mixed fractions of DMAc and non-solvent (mono-, di-, trihydric alcohol). The DMAc/non-solvent solutions were prepared with non-solvent concentrations varying from 26.6 to 53.3 wt%; the entire liquid to solid transition range. Five CA concentrations were prepared; 10, 12.5, 15, 17.5 and 20 wt%. To ensure complete miscibility and the elimination of air bubbles, the samples were manually mixed and heated for 10 to 15 min at 100°C. All samples were prepared in 15 mL glass scintillation vials (Fischer Scientific, fisher brand). For gels, the heated samples were immediately transferred to a wide mouth 168 plastic container (30mL, 30mm diameter, Nalgene straight-side jar (Fisher Scientific), an appropriate size for plate geometry rheology measurement). All samples were equilibrated for exactly 1 week in a desiccator at room temperature prior to analysis. All solutions and dilutions were prepared on a weight basis. Rheological Characterization A TA Instruments Advanced Rheometer (AR 2000) was used to measure the rheological. properties of the samples. All analyses were performed at ambient temperature (25 °C). In general, the viscoelastic properties of materials were measured using either parallel plate or cone geometries depending on nature of the samples being studied. In this research, dilute CA solutions were analysed using a 60 mm diameter aluminium 2 degree cone angle geometry with a fixed zero gap (automatically determined by the rheometer). For high viscosity CA solutions (close to the gel-point) a smaller surface 40 mm diameter aluminium 2 degree cone angle geometry was used. On average 3 to 4 mL of the CA solution were used per analysis. For gel samples, a stainless steel parallel plate geometry (20 mm diameter) with a variable gap (1000 -2000 microns) was used. Gap size was determined based on the thickness of the gel and the applied force (normal force) which was kept below 0.2 N cm"2. In this study rheological measurements were performed in both steady shear and dynamic modes. Steady Shear Rheology Steady shear measurements were performed by subjecting a sample to a steady shear at a constant shear rate (y), resulting in the generation of a shear stress (T). The corresponding shear stress (x) on the sample was measured using a torque transducer. The viscosity (n) was measured as a function of the steady shear rate (y). In a typical experiment 3 mL (40 169 mm cone) to 4 mL (60 mm cone) of the CA solution (excess sample) was loaded onto the centre of the bottom plate. The upper geometry was lowered to the adjusted zero gap and the excess solution around geometry was removed. The solutions were then relaxed for 1 minute to reach an equilibrium state prior to viscosity measurement. Shear rates ranged from 0.05 s"1 to 500 s"1. All viscosity measurements were conducted at 25 °C. Dynamic Rheology In the dynamic rheology experiments an oscillating shear was applied to the sample and the corresponding elastic (G') and viscous (G") moduli measured as a function of either frequency or stress amplitude. For dynamic frequency analysis to be valid, the experiments have to be conducted in the linear viscoelastic region (LVE). Therefore, initial stress sweep experiments were performed to determine the LVE. This test involved the variation of stress amplitude from 0.1 (Pa) to 2000 (Pa), while maintaining a constant frequency (1 Hz). The log sweep was set at 10 points per decade. The linear viscoelastic region was defined as the region in which G' varied linearly with stress as shown in Figure A In addition to determining the LVE, the stress sweep experiments also provided information on the yield stress of the gels. 170 IO5 104 IO3 G' (Pa) 102 IO1 IO"1 10° IO1 102 IO3 Osc. Stress (Pa) Figure A: Schematic stress sweep spectra showing LVE region and yield stress. The elastic G' and viscous G" moduli of the samples were determined using frequency sweep experiments ranging from 0.01 to 250 rad s"1 at a constant stress of 1 Pa. (identified from the LVE). In a typical experiment, the stress sweep samples, which did not reach their yield stress, were equilibrated for 2 minutes in the absence of applied stress, then a sinusoidal deformation as a function of frequency was applied. All experiments were performed at 25 °C. The shape and magnitude of the elastic and viscous moduli, as a function of frequency, provides a signature of the state of the system (e.g., solution, gel). Cone geometries were used for solution samples and parallel plate geometry (Standard steel parallel plate, 20 mm) were used for gels. Duplicate experiments were run to ensure consistency and all results were within an error of 10%. Cloud Point Measurement 111 Cloud point measurements were performed using 340 nm UV light in a spectrophotometer. Measurements were made using a cuvette having an optical path length of 1 cm. All measurements were standardized against a dilute CA/DMAc solution, which was adjusted to 100% transmission. Samples (~5 mL) were transferred into the cuvette and placed in the path of the UV light and the output readings recorded. For high viscosity samples, such as gels, a spatula was used to carefully transfer the sample into the cuvette. All cloud point measurement was performed at ambient temperature. FTIR Spectroscopy FTIR spectra were measured on a Perkin Elmer 16PG FTIR spectrometer. A total of 16 scans were collected at resolution of 4 cm"', over the range of 400 to 4000 cm"1. FTIR spectra were collected using 10 mg samples pressed between ZnSe salt plates. Care was taken not to disrupt sample microstructure during the transfer and pressing between plates. Microscopy Scanning Electron Microscopy (SEM) The microstructure of the gels was investigated using a Hitachi S-2600 variable pressure scanning electron microscope operating at an accelerating voltage of 20 kV and a working distance of 10mm. Images were taken over a magnification range of 250x to 4000x. Duplicate samples were prepared, and SEM images were obtained from fractured surfaces to minimize any artifices introduced during sample preparation. Prior SEM, the solvents had to be removed from the samples without any significant changes in microstructure of the gels. This was accomplished by a fixation/solvent exchange/freeze drying process. Small pieces of gels (~2 x ~2 x ~2 mm) were out from prepared gels and floated on a (10 mL) solution of 25% glutaraldehyde and 0.1M sodium 172 acetate buffer at pH 6.6. Fixation was accomplished by subsequently treating the samples in a microwave at 100 Watts for 40 seconds and 24 °C. The gels were then rinsed with 10 mL of 0.1 M sodium acetate buffer solution, followed by six washes with distilled water (10 mL). Each rinse was performed in a microwave under a vacuum (500 mtorr) for 40 seconds at 200 Watts and 24 °C. The samples were then quickly frozen and lypholized, gently fractured and sputter coated Coating was performed using a cressington high resolution sputter coater under vacuum between 0.04 to 0.03 torr for 3 minutes until the samples were covered with a thin layer of Au/Pd. Laser Scanning Confocal Microscopy (LSCM) A confocal laser scanning system (Chameleon, compact ultra fast Ti) connected to an inverted microscope (Zeiss Axiovert) was used to scan the gels. The LSCM images were obtained at high resolution with 3-D reconstruction. An advantage of confocal microscopy is its ability to produce blur-free images of thick specimens at various depths, and to generate clean slices through optical sections out of thick fluorescent specimens without physical sectioning. Calcofluor white (Sigma-Aldrich), a cellulose selective fluorescent dye was used to tag the CA for LSCM. Calcofluor white (0.01 wt%) was added to the CA system, mixed and heated at 100 °C as per rheological analysis. A few drop of the hot bubble free solution was then placed onto a single concavity microscopy slide (76 x 26 mm; 1.4 mm thick and 0.5 mm deep - Fisher Scientific), over laid with a cover slip (0.5 mm thickness) and sealed with clear nail polish. The sample was then conditioned for 1 week in a desiccator at ambient temperature as per the other analyses. LSCM measurements were performed with a 63 x numerical aperture 1.4 oil immersion objective lens. Optical sectioning was 173 performed as a function of the depth along the z-axis. The IR laser excitation source was set to 705 nm with two channel spectra detection (Chi filter: 544-704nm, Ch2 filter: 435-485nm). Scans were collected as stack images with a stack size of X: 142.86 pm, Y: 142.86 pm, and Z: 27.87pm. Data Analysis All rheological analyses were within an error of less than 10%. The error was determined by running multiple samples (minimum of 4 samples) for several non-solvent systems (i.e. ED/BD/HD). The multiple samples were averaged and the difference between the highest and lowest values was less than 10%. This was consistently the case for all of the non-solvents studied. Therefore, rheological analyses were run in duplicate and the results were within the 10% error. In this thesis, all of the raw data presented is from one of the replicate runs. In the CSLM experiments, image analysis was done using Adobe Photoshop 7.0 (Adobe Systems Inc.). In the determination of the fractal dimension from CSLM, five layers through the sample thickness were analyzed by Image J software using the fractal box count method. Average fractal dimension and standard deviation were obtained. The spatial resolution of the images was 512 x 512 pixels. Image analysis was performed using Image J software (NIST). 174 

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.831.1-0074946/manifest

Comment

Related Items