Polymer Diffusion in Latex Films

Transcription

1 Polymer Diffusion in Latex Films by YUANQIN LIU A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Chemistry University of Toronto Copyright by Yuanqin Liu 2009

2 Polymer Diffusion in Latex Films Yuanqin Liu Doctor of Philosophy Department of Chemistry University of Toronto 2009 Abstract In this thesis, I describe experiments that provide a new and deeper understanding of factors that affect polymer diffusion in acrylic latex films. This is the step that leads to the growth of mechanical properties of these films. Polymer diffusion was studied by fluorescence resonance energy transfer (FRET) in films prepared from dye-labeled latex particles. Poly(n-butyl acrylate-co-methyl methacrylate) [P(BA-MMA)] was chosen for the study of copolymer composition on the polymer diffusion rate. Four sets of P(BA-MMA) copolymers were prepared from various weight ratios of BA/MMA. Polymer diffusion was monitored as a function of annealing temperature, and apparent diffusion coefficients (D app ) were calculated from the FRET data, using a simple diffusion model. The temperature dependence of polymer dynamics (G, G ) obtained by linear rheology measurements is in good agreement with the temperature dependence of D app. Increasing the BA content of the copolymer led to an apparent increase in long-chain branching, which is reflected in both the time dependence of D app and in the dynamic moduli measurements. To study the effect of branching on polymer diffusion rates, latex particles comprised of branched poly(n-butyl methacrylate) (PBMA) were prepared. The degree of branching was controlled by adding various amounts of bisphenol A dimethacrylate as a branching agent, plus ii

3 1-dodecanethiol as a chain transfer agent to prevent gel formation and to control the polymer molecular weight. The results of rheology (G, G ) measurements are consistent with the absence of entanglement in these polymers. After correcting for the effects of T g, by comparing results at a constant T- T g, ET data show that the PBMA with the highest degree of branching had the highest diffusivity. In a separate set of experiments I tested the effect of incorporating the highly branched PBMA (HB-PBMA) into P(BA-MMA) dispersions to examine its influence on polymer diffusion in the latex films. Three different approaches were taken to combine these different polymers: latex blends, using HB-PBMA seeds in the synthesis of P(BA-MMA) by semicontinuous emulsion polymerization, and dissolving HB-PBMA in the mixture of BA and MMA for latex particles prepared by miniemulsion polymerization. ET studies indicate that HB-PBMA significantly enhances polymer diffusion rate, comparable with Texanol TM, a volatile organic coalescing agent. Tensile tests show that the films containing HB-PBMA have significant higher mechanical properties than the films containing Texanol TM. iii

4 Acknowledgments First, I would like to express my greatest appreciation to Professor Mitchell A. Winnik for his guidance, encouragement, and support during my doctoral study. His insight, kindness, humor, and knowledge truly impress and affect me. From him I have learned not only how to work, but also how to live. He is far beyond a research supervisor. Second, I want to thank my industry supervisor, Dr. Willie Lau, who gave me this great opportunity to work with Rohm and Haas Company, a world s leading company in coatings industry. Without his continuous help my research could not be carried out smoothly. Third, my appreciation will go to my research committee members, Professor Michael Georges, Professor Eugenia Kumacheva, Professor Gilbert Walker, and Professor Tim Bender, for their advices and help on my research. Fourth, I give my appreciation to my colleagues, especially Dr. Jeffrey C. Haley and Dr. Walter F. Schroeder. Working with Jeff was very enjoyable since he is smart, creative and humorous. Walter is an excellent research partner with highly reliance. I would also thank Dr. Zhihui Yin, Dr. Fugang Li, Dr. Xudong Lou, Dr. Jingshe Song, Dr. Jun Wu, Dr. Jung Kwon Oh, Mr. Kangqing Deng, Dr. Baohang Han, Dr. Sheng Dai, Mr. Mohsen Soleimani, Mrs. Neda Felorzabihi, Mr. Robert Roller, Dr. Gerald Guerin and all other group members for their kind help. I like to thank Rohm and Haas, Rohm and Haas Canada, and NSERC Canada for their support of this research. My appreciation will also go to University of Toronto, Ontario Centres of Excellence, Ontario Graduate Scholarship in Science and Technology (OGSST) for scholarships. Finally, I will express my deepest gratefulness to my wife Ying (Sunny) Sun and the other family members for their love, company, and support. iv

5 Table of Contents Chapter 1 1 Introduction Research background Water-based coatings and environment concern In-situ seeding emulsion polymerization Acrylic latex Latex film formation and polymer diffusion Research objectives Thesis outline References 6 Chapter 2 2 Experimental Materials Synthesis of latices Synthesis of poly(n-butyl acrylate-co-methyl methacrylate) P(BA-MMA) Synthesis of branched poly(n-butyl methacrylate) PBMA Preparation of P(BA-MMA)/HB-PBMA latex blends Preparation of P(BA-MMA) latices containing TexanolTM Synthesis of P(BA-MMA) latices containing hyperbranched poly(n-butyl methacrylate) (HB-PBMA) Synthesis of P(BA-MMA) latices containing hyperbranched poly(n-butyl methacrylate) (HB-PBMA) seed particles Characterizations of latex particles Characterizations of latex polymers Molecular weight and molecular weight distribution Copolymer composition Glass transition temperature Acceptor dye concentration measurements Gel content measurements Mechanical property measurements Rheology measurements Tensile testing Dynamic mechanical analysis Film preparation 17 v

6 2.7 Film annealing Fluorescence decay measurements and data analysis Calculation of apparent diffusion coefficients Dapp Fujita-Doolittle fitting References 21 Chapter 3 3 Effect of Polymer Composition on Polymer Diffusion in Poly(n-butyl acrylateco-methyl methacrylate) Latex Films Introduction Results Preparation and Characterization of the Latex Samples Energy Transfer Studies of Polymer Diffusion Polymer Diffusion in P(BA60-MMA39) Films at Different Temperatures Polymer Diffusion in Different Composition P(BA-MMA) Films Temperature Dependence of the Viscoelastic Properties of P(BA-MMA) Films Discussion Comparison between Different Experiments Effect of long chain branching on the time-dependence of Dapp Summary References 45 Chapter 4 4 Synthesis of Branched Poly(n-butyl methacrylate) via Semi-Continuous Emulsion Polymerization Introduction Experimental Section Latex Preparation Synthesis of High Molecular Weight Linear PBMA Characterization of Latex Polymers Results Synthesis of Branched PBMAs Architectures of Branched PBMAs Branched PBMA Latex Particles Rheology Measurements Discussion Control of Molecular Weight Entanglement Considerations in branched PBMA 59 vi

7 4.5 Summary Reference 62 Chapter 5 5 Effect of Branching on Polymer Diffusion in Branched Poly(n-butyl methacrylate) Latex Films Introduction Results and Discussion Synthesis of Dye-Labeled Branched PBMA Latex Particles Polymer Diffusion in Branched PBMA Latex Films at Same Temperature Polymer Diffusion in Branched PBMA Latex Films at (Tg+20) oc Summary Reference 83 Chapter 6 6 Effect of Hyper-Branched Poly(n-butyl methacrylate) on Polymer Diffusion in Poly(n-butyl acrylate-co-methyl methacrylate) Latex Films Introduction Blending Approach Effect of Blended HB-PBMA Particles on Polymer Diffusion in P(BA-MMA) Latex Films Comparing the Plasticization Effect of Blended HB-PBMA Particles and Texanol TM Effect of Blended HB-PBMA Particles on Mechanical Properties of the P(BA- MMA) Latex Films Conclusion Miniemulsion Polymerization Approach Miscibility of HB-PBMA and P(BA-MMA) Synthesis of Miniemulsion P(BA 55 -MMA 44 ) Latex Particles Kinetic Study of the Miniemulsion Polymerization Effect of the HB-PBMA on Polymer Diffusion in P(BA 55 -MMA 44 ) Latex Films Comparing the Plasticization Effect of HB-PBMA and Texanol TM Effect of HB-PBMA on Mechanical Properties of the P(BA 55 -MMA 44 ) Latex Films Conclusion Seeded Emulsion Polymerization Approach Synthesis of Seeded P(BA 55 -MMA 44 ) Latex Particles 108 vii

8 6.4.2 Effect of HB-PBMA Seed Particles on Polymer Diffusion in Seeded P(BA 55 - MMA 44 ) Latex Films Effect of HB-PBMA Seed Particles on Mechanical Properties of the Seeded P(BA 55 -MMA 44 ) Latex Films Conclusion Summary References 115 Appendix 6.1 The 1 H NMR spectrum of P(BA 55 -MMA 44 ) A 1%ME final product 117 Appendix 6.2 Tensile stress-strain curves for P(BA-MMA) latex films by miniemulsion polymerization 118 Appendix 6.3 Values of Ф ET (0) and Ф ET ( ) of the three approaches 119 Appendix 6.4 Chemical structure of Me-β-CD 120 Appendix 6.5 Chemical structure of Texanol TM 121 viii

9 List of Tables Table 2-1. Typical semi-continuous emulsion polymerization recipe for the synthesis of nonlabeled P(BA 60 -MMA 39 ) latex. Table 2-2. Typical semi-continuous emulsion polymerization recipe for the synthesis of branched PBMA latex. Table 2-3. Typical miniemulsion polymerization recipe for the synthesis of donor labeled P(BA-MMA) latex containing 1 wt% HB-PBMA. Table 2-4. Typical seeded emulsion polymerization recipe for the synthesis of A-labeled P(BA-MMA) latex containing 5 wt % HB-PBMA seed particles. Table 3-1. Characteristics of the P(BA-MMA) latex polymers and particles. Table 3-2. E a values of the latex polymers. Table 4-1. Characteristics of unlabeled latex polymers and particles Table 4-2. Estimate of branching. Table 5-1. Typical recipe for the synthesis of D-labeled branched PBMA latex. Table 5-2. Characteristics of the latex polymers and particles. Table 5-3. Estimation of the chain structure of the different branched latices. Table 5-4. Limiting values of energy transfer efficiency for the different labeled latex mixtures. Table 6-1. Characteristics of the P(BA-MMA) latex polymers and particles. Table 6-2. T g of P(BA-MMA) polymers containing HB-PBMA and Texanol TM. Table 6-3. Tensile properties of P(BA 55 -MMA 44 ) latex films. Table 6-4. Characteristics of P(BA-MMA) latex polymers made by miniemulsion polymerization. Table 6-5. Miscibility of HB-PBMA and P(BA-MMA). Table 6-6. Fitting parameters for the Fujita-Doolittle equation. Table 6-7. Tensile testing results of P(BA 55 -MMA 44 ) films. Table 6-8. Characteristics of the P(BA 55 -MMA 44 ) latex polymers and particles. Table 6-9. Tensile testing results of P(BA 55 -MMA 44 ) films. Table A1. Comparison of Ф ET (0) and Ф ET ( ) of the three approaches. ix

10 List of Figures Figure 1-1. Scheme of in-situ emulsion polymerization reaction. Figure 1-2. Mechanism of latex film formation. Figure 1-3. (A) The chemical structures of D and A are shown; (B) Polymer diffusion can only occur after the polymer in adjacent cells comes into intimate contact. Polar material trapped between cells can interfere with polymer diffusion. Figure 3-1. Plot of 1/M n against concentration of C 12 -SH of P(BA 60 -MMA 39 ) latex samples. Figure H NMR spectra of (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ). CDCl 3 was used as solvent. Peaks a and peak b correspond to protons at positions a and b respectively. Figure 3-3. Phenanthrene (donor) fluorescence decay curves [I D (t)] measured at 23 o C for Phe- P(BA 60 -MMA 39 ) latex films. (1) Phe-labeled latex only, (2) a newly formed film dried at 4 o C, consisting of a 1:1 ratio of Phe-P(BA 60 -MMA 39 ) and NBen-P(BA 60 - MMA 39 ), (3) the same film as in (2) aged for 47 min at 23 o C, and (4) a solvent-cast film from a 1:1 mixture of the two freeze-dried polymers dissolved in THF and then annealed at 120 o C for 2 h. Note that curves (1) and (2) overlap. The inset shows curves (1) and (2) at short times on a linear scale. Figure 3-4. Plot of P vs [NBenMA, mm] for fully mixed solvent-cast films prepared from Phe- P(BA 60 -MMA 39 ) plus varying amounts of free monomer MBenMA. The P values were obtained by fitting individual Phe decay curves to equation (2-5) with τ D fixed at 44.3 ns. From the slope of the plot, I calculate R 0 = 2.51 nm. Figure 3-5. Plots of Ф ET (A) and f m (B) vs annealing time for the P(BA 60 -MMA 39 ) latex films annealed at 23, 45, 70, and 90 o C. Figure 3-6. Plots of Ф ET and f m vs annealing time for the P(BA 60 -MMA 39 ) latex films annealed at 23 o C. x

11 Figure 3-7. Plots of the apparent diffusion coefficient D app as a function of f m for (A) P(BA 60 - MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ) latex films annealed at various temperatures. Figure 3-8. Plot of ln D app against 1/T over the temperature range from 23 to 60 o C at f m values of 0.59 for P(BA 60 -MMA 39 ) latex. Figure 3-9. Plot of the Φ ET for films formed from D/A labeled latex mixtures annealed for various periods of time at (A) 23 o C, (B) 70 o C and (C) 90 o C. ( )P(BA 60 -MMA 39 ), ( )P(BA 55 -MMA 44 ), ( ) P(BA 50 -MMA 49 ) and ( ) P(BA 40 -MMA 59 ). Figure The master curves of D app values for (A) P(BA 60 -MMA 39 ) at 23 o C (calculated using E a = 33.4 kcal/mol as a shift factor); (B) P(BA 55 -MMA 44 ) at 23 o C (calculated using E a = 39.1 kcal/mol as a shift factor); (C) P(BA 50 -MMA 49 ) at 23 o C (calculated using E a = 45.2 kcal/mol as a shift factor) and (D) P(BA 40 -MMA 59 ) at 70 o C (calculated using E a = 64.1 kcal/mol as a shift factor). Figure Plots of master curves of G' and G'' for (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ) latex films at T 0 = 25, 50, 80 and 90 o C respectively. Figure Plots of shifted D app and log(a T ) against the inverse of the absolute temperatures for (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 - MMA 59 ) latex films. Figure H NMR spectra of PBMAs with different branching levels. CD 2 Cl 2 was used as solvent. Peaks a and peaks b correspond to protons of BPDM and BMA respectively. Figure 4-2. UV and RI traces in the GPC analysis of branched-pbma2. Figure 4-3. Polymer architecture for (A) linear-pbma1, (B) branched-pbma2, (C) branched- PBMA5 and (D) branched-pbma7. These drawings assume a uniform distribution of branch points in the polymer molecules. xi

12 Figure 4-4. GPC traces (A) RI signal and (B) log M n vs retention volume for linear-pbma1 (L1), branched-pbma2 (B2), branched-pbma5 (B5) and branched-pbma7 (B7). The vertical line in (B) indicates that the polymer with Mn = 34,000 has a retention volume of 17.3 ml. The vertical line in (A) indicates that ca. 30% of the L1 sample has M n lower than 34,000. Figure 4-5. Plots of master curves of G' and G'' for (A) linear-pbma1, (B) branched-pbma2, (C) branched-pbma5 and (D) branched-pbma7. Figure 4-6. Plots of master curves of G' and G'' for the high molecular weight linear PBMA sample. The vertical dotted line corresponds to the minimum value of G. Figure 5-1. GPC traces (RI) for (A) A- and (B) D-labeled PBMA samples. Figure 5-2. GPC traces of (A) LB-PBMA D and (B) LB-PBMA A. The UV and RI traces overlap, which indicates a nearly random distribution of dye comonomers in the polymer chains. Figure 5-3. UV calibration curve for NBenMA in THF solution. The extinction coefficient is ε = (2.47 ± 0.03) 10 4 M -1 cm -1. Figure 5-4. Plot of the Φ ET for films formed from D/A labeled latex mixtures annealed for various periods of time at (A) 23 o C, (B) 45 o C and (C) 70 o C. (О) HB-PBMA D/A, ( ) MB-PBMA D/A, ( ) LB-PBMA D/A and ( ) LR-PBMA D/A. Figure 5-5. Plot of the f m for films formed from D/A labeled latex mixtures annealed for various periods of time at (A) 23 o C, (B) 45 o C and (C) 70 o C. (О) HB-PBMA D/A, ( ) MB- PBMA D/A, ( ) LB-PBMA D/A and ( ) LR-PBMA D/A. Figure 5-6. Plots of D app as a function of f m for (A) LR-PBMA D/A, (B) LB-PBMA D/A, (C) MB- PBMA D/A, and (D) HB-PBMA D/A. As f m 1, the ET experiment loses sensitivity because the incremental changes in Φ ET become small. Thus the calculated D app values for f m > 0.9 may not be meaningful. xii

13 Figure 5-7. Master curves of D app for (A) LR-PBMA D/A, (B) LB-PBMA D/A, (C) MB-PBMA D/A, and (D) HB-PBMA D/A. 23 o C was used as the reference temperature. The apparent precipitous decrease in D app values at values of f m 1 may not be real, since the energy transfer methodology loses its sensitivity as polymers approach the fully mixed state, and the acceptor concentrations in the films become uniform. Figure 5-8. (A) Plot of f m for films annealed at (T g +20) o C. (B) Plot of D app over f m at (T g +20) o C. (О) HB-PBMA D/A, ( ) MB-PBMA D/A, and ( ) LB-PBMA D/A. Figure 5-9. Plots of master curves of G' and G'' for (A) LR-PBMA D/A, (B) LB-PBMA D/A, (C) MB-PBMA D/A, and (D) HB-PBMA D/A. 40 o C was used as reference temperature. Figure 6-1. Plots of the f m versus annealing time for (A,B) P(BA 55 -MMA 44 ), (C,D) P(BA 50 - MMA 49 ), and (E,F) P(BA 45 -MMA 54 ) latex films containing ( )0 wt % of additives, ( )5 wt % blended HB-PBMA particles, ( )10 wt % blended HB-PBMA particles at 23 C. Figure 6-2. Comparison of the plots of the the f m versus annealing time for (A,B) P(BA 55 - MMA 44 ), (C,D) P(BA 50 -MMA 49 ), and (E,F) P(BA 45 -MMA 54 ) latex films containing HB-PBMA and Texanol TM at 23 C. ( ) 0 wt % of additives, ( ) 5 wt % blended HB-PBMA particles, ( ) 10 wt % blended HB-PBMA particles, ( ) 5 wt % Texanol TM, ( ) 10 wt % of Texanol TM. Figure 6-3. Comparison of the plots of the D app versus f m for P(BA 55 -MMA 44 ) latex films containing ( )0 wt % of additives, ( )5 wt % of blended HB-PBMA particles, and ( ) 5 wt % of Texanol TM at 23 C. Figure 6-4. Tensile stress-strain curves for P(BA 55 -MMA 44 ) latex films with 0 wt % additive, 5 wt % blended HB-PBMA particles, 10 wt % blended HB-PBMA particles, 5 wt % Texanol TM, and 10 wt % Texanol TM. Figure 6-5. Total conversion as a function of reaction time for the miniemulsion polymerization of P(BA-MMA) A 1%ME. xiii

14 Figure 6-6. Experimentally determined (points) and modeling (lines) of the partial conversion of ( ) MMA and ( ) BA vs the overall conversion for P(BA-MMA) A 1%ME. Figure 6-7. (A) Comparison of the plots of f m vs annealing time for P(BA-MMA) at 45 C with (*)0 wt %, ( )1 wt %, ( )3 wt %, ( )5 wt %, and ( )6 wt % of HB-PBMA; (B) Plot of D app as a function of f m for P(BA-MMA) at 45 C with (*)0 wt %, ( )1 wt %, ( )3 wt %, ( )5 wt %, and ( )6 wt % of HB-PBMA; (C) Master curve of D app for HB-PBMA in P(BA-MMA) films; (D) Comparison of the plots of f m vs annealing time for P(BA-MMA) at 45 C with ( )0 wt %, ( )2 wt %, ( )5 wt %, and ( )7 wt % of Texanol TM ; (E) Plots of D app vs f m for P(BA-MMA) at 45 C with ( )0 wt %, ( )2 wt %, ( )5 wt %, and ( )7 wt % of Texanol TM ; (F)Master curve of D app for Texanol TM in P(BA-MMA) films. Figure 6-8. Plot of 1/ln[D p (T,Φ a )/D p (T,0)] vs 1/Φ a at 45 C for the P(BA-MMA) films containing HB-PBMA Figure 6-9. Plot of 1/ln[D p (T,Φ a )/D p (T,0)] vs 1/Φ a at 45 C for the P(BA-MMA) films containing Texanol TM. Figure (A)Tensile stress-strain curves for P(BA-MMA) latex films with various concentration of HB-PBMA and Texanol TM. (B)Early stages of the stress-strain curves in (A). Figure Comparison of the plots of f m versus aging time for P(BA 55 -MMA 44 ) at 23 C with ( )0 wt % additives, ( )5 wt % HB-PBMA seeds, ( )10 wt % HB-PBMA seeds, ( )5 wt % Texanol TM, and ( )10 wt % Texanol TM. (C) and (D) show the early stages of the f m plots in (A) and (B), respectively. Figure (A)Comparison of the plots of the D app versus f m for P(BA 55 -MMA 44 ) latex films containing ( )0 wt % additives, ( )5 wt % HB-PBMA seeds, ( )10 wt % HB- PBMA seeds, ( )5 wt % Texanol TM, and ( )10 wt % Texanol TM at 23 C. (B)Plot of 1/ln[D p (T,Φ a )/D p (T,0)] vs 1/Φ a at 23 C for the P(BA 55 -MMA 44 ) films containing different amounts of ( )Texanol TM and ( )HB-PBMA seeds. xiv

15 Figure (A)Tensile stress-strain curves for P(BA-MMA) latex films with various concentration of HB-PBMA seed particles and Texanol TM. (B)Early stages of the stress-strain curves in (A). Figure A1. 1 H NMR spectrum of P(BA 55 -MMA 44 ) A 1%ME, which was synthesized by miniemulsion polymerization and contains 1 wt % HB-PBMA. CDCl 3 was used as solvent. Figure A2. Tensile stress-strain curves for P(BA 55 -MMA 44 ) latex films with various concentration of HB-PBMA and Texanol TM, which were synthesized by miniemulsion polymerization. xv

16 List of Abbreviation AIBN BA BMA BPDM C 12 -SH D app 2,2 -Azobis(2-methylpropionitrile) n-butyl acrylate n-butyl methacrylate bisphenol A dimethacrylate 1-dodecanethiol apparent diffusion coefficient E storage modulus E loss modulus G' shear storage modulus G shear loss modulus ET f m FRET HB-PBMA HD I D KPS MAA Me-β-CD MFT MMA NBenMA P(BA-MMA) PBMA PDI PheMMA energy transfer fraction of mixing fluorescence resonance energy transfer hyper branched poly(n-butyl methacrylate) hexadecane donor fluorescence intensity decay Potassium persulfate methacrylic acid methyl-β-cyclodextrin minimum film formation temperature Methyl methacrylate 4 -Dimethylamino-2-methacryloxy-5-methyl benzophenone poly(n-butyl acrylate-co-methyl methacrylate) poly(n-butyl methacrylate) polydispersity index Phenanthrylmethyl methacrylate xvi

17 PMMA PTFE R 0 SDS Texanol TM THF T g VOC Φ ET ω τ D poly(methyl methacrylate) polytetrafluoroethylene critical Förster radius for energy transfer sodium dodecyl sulfate 2, 2, 4-Trimethyl-1, 3-pentanediol monoisobutyrate tetrahydrofuran glass transition temperature volatile organic compounds quantum efficiency of energy transfer shear frequency fluorescent donor lifetime xvii

18 Chapter 1 1 Introduction My research work focuses on the fundamental aspects of latex film formation of acrylic latices including poly(n-butyl acrylate-methyl methacrylate) [P(BA-MMA)] copolymer and poly(n-butyl methacrylate) (PBMA) homopolymer. P(BA-MMA) copolymers of different compositions were chosen as the base latex polymer because the P(BA-MMA) latex is widely used in commercial architecture coatings. This choice makes these fundamental research experiments also of practical importance. Since P(BA-MMA) copolymers are randomly branched polymers, it is impossible to quantitatively study the effect of branching on polymer diffusion. To achieve this objective, I synthesized a series of latex particles containing branched PBMA polymer and explored their diffusion behavior. 1.1 Research background Water-based coatings and environment concern For many decades, industrial coatings have relied on solvent-based formulations. Today, increasingly strict environmental regulations are the impetus for a change in technology. The most important substitutes for solvent-based coatings are water-based coatings. Solvent-free water-based coatings should be environment friendly. However, most commercial water-based coatings still contain significant amounts of volatile organic compounds (VOC) as additives. For example, solvents added as coalescing agents promote film formation, but later evaporate and contribute to air pollution. A key challenge for the coating industry is to develop new technology that is environmentally compliant and meets current performance specifications. To develop this new technology, a deeper fundamental understanding of how small changes in formulation or latex composition affect the development of useful coating properties is necessary. 1

19 1.1.2 In-situ seeding emulsion polymerization Good model latices are the essential feature of my research. The model latices should be able to approximate the performance properties of the commercial coating products and also be sufficiently well-defined for scientific research. Previous research in our laboratory always employed latex particles synthesized using seed latex samples, which were prepared in advance. 1, 2 The advantage of this approach is the easy control of particles size. The downside is that the polymer composition of the final particles may not be unique, since the seeds normally contain high molar mass polymers. Their presence may affect the polymer diffusion studies that are a big part of my thesis research. In this research, I used an in-situ seeding process (Figure 1-1), in which the seeds and final particles are formed from the same monomer mixture as the subsequent monomer feed. In this way, the whole particle will have a homogeneous polymer Figure 1-1. Scheme of in-situ emulsion polymerization reaction. composition. The problem for in-situ seeding is that particles sizes are difficult to match from batch to batch, especially for the small scales in our reations. At the beginning of my doctoral work I tried very hard to scale down the in-situ seeding recipe from our industrial partner and eventually managed to synthesize uniform latex particles with controlled size. All emulsion polymerization reactions carried out in this thesis used this in-situ seeding process. 2

20 1.1.3 Acrylic latex In my research acrylic copolymer latices were chosen as model latices. One attractive feature of acrylic monomers is their ability to copolymerize with each other. In this way one can obtain copolymers with a wide range of compositions and physical properties. These various properties make acrylic latex copolymers interesting to both industrial and academic researchers. For example, one series of base latices for my experiments are n-butyl acrylate-methyl methacrylate (BA-MMA) copolymers, which are also the base latices of commercial architecture paints. With a weight ratio ranging from 60/40 to 40/60, their glass transition temperatures range from 3 o C to 28 o C. Another series of acrylic latices are n-butyl methacrylate-bisphenol A dimethacrylate (BMA-BPDM) copolymers, which are interesting to me because they offer control over polymer chain branching Latex film formation and polymer diffusion A number of recent publications have reviewed the mechanism of latex film formation 3, 4 Our research group has made a significant contribution to this field. 5, 6 When a latex dispersion is cast onto a substrate and allowed to dry, an initial film will form as the particles deform and pack into space-filling polyhedral cells. These newly formed latex films have poor mechanical properties because there is only weak adhesion at the boundary between adjacent cells. If the minimum film forming temperature (MFT) of the latex polymer is lower than the drying temperature, polymer molecules will diffuse across the intercellular boundary to produce mechanically coherent films. The film formation process is depicted in Figure 1-2. Figure 1-2. Mechanism of latex film formation 3

21 Due to technology limitations, polymer diffusion during latex film formation was not investigated until 1980 s. Oberthür and coworkers studied polymer diffusion in latex films using neutron scattering (SANS). 7, 8 Our research group contributed to the field by introducing the fluorescent resonance energy transfer (FRET) technique, which is based upon fluorescence decay measurements, to study polymer diffusion rates. 9 A typical approach includes the synthesis of two virtually identical samples of latex particles, one of which was labeled with a fluorescent donor dye (phenanthrene, D), while the other was labeled with an acceptor dye (anthracene or a benzophenone derivative NBen, A). The chemical structures of D and A are shown in Figure 1-3(A). Latex films were cast from a mixture of the two dispersions. As shown in Figure 1-3(B), D and A are initially separated from each other in a newly formed film. FRET does not occur at this stage and no fluorescence decay can be observed. As polymer chains diffuse, D and A are brought together. FRET take places and leads to a strong decrease in the measured fluorescence decay. In the past 20 years, extensive studies on polymer diffusion have been carried out using this technique in our laboratory. A number of factors that affect the rate of polymer diffusion in latex films were explored, which include temperature, composition of the latex, and the presence of various additives such as coalescing agents and common surfactants. Figure 1-3. (A) The chemical structures of D and A are shown; (B) Polymer diffusion can only occur after the polymer in adjacent cells comes into intimate contact. Polar material trapped between cells can interfere with polymer diffusion. 4

22 1.2 Research objectives The objective of my research is to develop molecular-level knowledge of coalescence and film formation from acrylic latex particles used for interior architectural coatings. When these coatings dry, the latex particles deform and coalesce, and the useful performance properties of the coatings arise from the diffusion of polymer molecules across the boundaries between cells formed by the latex particles. Here my focus is on the factors that connect events at the molecular level to the development of final film properties. Scientists in the coating industry anticipate that this new knowledge will help to explain a series of in-house observations about how small changes in formulation or latex composition affect the development of useful coating properties. In addition, I hope that the new knowledge developed here can be used to enhance the development of a new generation of high-performance latex coatings with substantially reduced VOC content. 1.3 Thesis outline The research work in this thesis mainly focuses on the synthesis of acrylic latices, FRET measurements of latex films, and mechanical property tests of latex polymers. I present this thesis in six chapters. In Chapter 2, I describe experimental methodologies for latex synthesis, characterizations of latex particles and latex polymers, and details of the FRET technique including data analysis. In Chapter 3, I describe polymer diffusion measurements in P(BA-MMA) copolymer latex films by FRET. The effect of copolymer composition on polymer diffusion rates is illuminated. I also show that the temperature dependence of polymer dynamics extracted by the rheology experiments is in good agreement with the temperature dependence of apparent diffusion coefficients (D app ), which were calculated from the FRET data. In Chapter 4, I describe the preparation of latex particles comprised of branched PBMA via semicontinuous emulsion polymerization. The extent of branching was controlled by adding various amounts of BPDM as a branching agent, and 1-dodecanethiol (C 12 -SH) was used as a 5

23 chain transfer agent to prevent cross-linking and to control molecular weight. The degrees of branching were determined using 1 H NMR. Rheology measurements indicated no significant entanglement contributions to the rheological properties. In Chapter 5, as a continuation from Chapter 4, I describe the FRET studies on the effect of different degree of branching on polymer diffusion rates in the PBMA latex films. I found that after correcting for the effects of T g, by comparing results at a constant T- T g, the PBMA with the highest degree of branching had the highest diffusivity. In Chapter 6, I describe the use of the HB-PBMA as a polymeric coalescing agent in P(BA-MMA) latex film formation. These approaches were taken to incorporate HB-PBMA into P(BA-MMA) latex samples: a blending approach, a miniemulsion polymerization approach, and a seeded emulsion polymerization approach. FRET studies show that the HB-PBMA enhances polymer diffusion rates in films of latex prepared using all three approaches. HB-PBMA has a similar effectiveness to a traditional coalescing agent, Texanol TM, in promoting polymer diffusion rates. Unlike Texanol TM, HB-PBMA does not lower significantly the temperature at which the latex form films upon drying. Mechanical measurements indicate that the HB-PBMA has negligible or limited influence on film mechanical properties, whereas the presence of Texanol TM causes a loss of tensile strength and toughness. 1.4 References 1 Oh, J. K.; Wu, J.; Winnik, M. A.; Craun, G. P.; Rademacher, J.; Farwaha, R. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Wu, J.; Oh, J. K.; Yang, J.; Winnik, M. A.; Farwaha, R.; Rademacher, J. Macromolecules 2003, 36,

24 3 Winnik, M. A. in Emulsion Polymerization and Emulsion Polymers; Lovell, P. A.; El-Aasser, M. S. Eds. Wiley: New York, Winnik, M. A. Curr. Opinion Coll. Polym. Sci. 1997, 2, ; Keddie, J. L. Mat. Sci. Eng. 1997, 21, Wang, Y.; Zhao, C. L.; Winnik, M. A. J. Phys. Chem. 1991, 95, Winnik, M. A.; Wang, Y.; Haley, F. J. Coatings Technol. 1992, 64, Hahn, K.; Ley, G.; Schuller, H.; Oberthur, R. Colloid & Polym Sci. 1986, 264, Hahn, K.; Ley, G.; Oberthur, R. Colloid & Polym Sci. 1988, 266, Zhao, C. L.; Wang, Y.; Hruska, Z.; Winnik, M. A. Macromolecules 1990, 23,

25 Chapter 2 2 Experimental 2.1 Materials Potassium persulfate (KPS), bisphenol A dimethacrylate (BPDM), sodium carbonate (Na 2 CO 3 ), sodium dodecyl sulfate (SDS), 1-dodecanethiol (C 12 -SH), hexadecane (HD), methacrylic acid (MAA), 2,2,4-trimethyl-1,3-pentanediol monoisobutyrate (Texanol TM ) and 2,2 - azobis(2-methylpropionitrile) (AIBN) were used as received from Aldrich. Polystep A-16 (22% solution of dodecylbenzene and tridecylbenzene sulfonates, 1 Stepan Co., Maywood NJ) and methyl-β-cyclodextrin (Me-β-CD) were kindly supplied by Rohm and Haas Co. and used as received. Methyl methacrylate (MMA, Aldrich), n-butyl acrylate (BA, Aldrich),and n-butyl methacrylate (BMA, Aldrich) were distilled at reduced pressure, and the purified monomers were stored at 4 o C until use. Water was purified by a Milli-Q ion-exchange filtration system. Phenanthrylmethyl methacrylate (PheMMA) was used as received from Toronto Research Chemicals Inc. 4 -dimethylamino-2-methacryloxy-5-methyl benzophenone (NBenMA) was synthesized as described elsewhere. 1, Synthesis of latices Synthesis of poly(n-butyl acrylate-co-methyl methacrylate) P(BA- MMA) A typical recipe for the semi-continuous emulsion polymerization of P(BA-MMA) is shown in Table 2-1. In the first stage, a dispersion of seed particles was prepared by batch emulsion polymerization with 3 wt % of a pre-emulsion of monomers, surfactant, chain transfer agent, and water. Water (3.0 g), Polystep A-16 (0.13 g) and Me-β-CD (0.13 g) were added in a 100 ml 3-neck flask equipped with a mechanical stirrer, nitrogen inlet and condenser. The flask was immersed in an oil bath. The system was thoroughly purged with nitrogen while the reaction mixture was heated to 80 o C. After the reactor temperature stabilized at 80 o C, the KPS solution (0.06 g in water 0.5 g) as an initiator and the Na 2 CO 3 solution (0.05 g in water 0.5 g) as a ph 8

26 buffer were added into the reactor followed by the addition of 3 wt % of monomer pre-emulsion (0.44 g). The mixture was stirred for 20 min at 80 o C. In the second stage of polymerization, the remaining monomer pre-emulsion was fed into the seed latex dispersion together with an initiator aqueous solution (0.01 g in water 2.0 g). The monomer feeding rate was kept identical (0.1 ml/min), controlled by Fluid Metering QG50 pumps, with a total feeding time of 3 h. After the addition was completed, the system was maintained at 80 o C for 0.5 h. Then the reaction was cooled to room temperature. A latex dispersion with ca. 50 wt % solids content was produced. The particle size is about 150 nm in diameter with a narrow size distribution. Table 2-1. Typical semi-continuous emulsion polymerization recipe for the synthesis of nonlabeled P(BA 60 -MMA 39 ) a latex ingredients (g) first stage second stage H 2 O 3.0 Polystep A-16 b 0.13 Me-β-CD 0.13 Na 2 CO KPS monomer pre-emulsion H 2 O 4.5 Polystep A BA 6.0 / 47 mmol / 60 wt % MMA 3.9 / 39 mmol / 39 wt % MAA 0.1 / 1 mmol / 1 wt % C 12 -SH c / 0.25 wt % a The subscripts refer to the wt % of each monomer. All latex samples contain 1 wt % MAA. b. 22 wt % surfactant solution, primarily sodium dodecylbenzene sulfonate. c. 1-Dodecanethiol, chain transfer agent, used at 0.25 wt % of total monomers. 9

27 Fluorescence dye labeled latex samples were synthesized in a similar fashion. For the donor (D) labeled particles, 1.0 mol % (2.4 wt %) PheMMA (based on total monomer) was added into the monomer pre-emulsion. For the acceptor (A) labeled particles, 0.3 mol % (0.8 wt %) NBenMA (based on total monomer) was added into the monomer pre-emulsion Synthesis of branched poly(n-butyl methacrylate) PBMA Branched PBMA latex samples were synthesized by semi-continuous emulsion polymerization reactions. A typical recipe for the synthesis of branched PBMA is shown in Table 2-2. A monomer pre-emulsion was prepared by shaking a mixture of monomer, branching agent, surfactant, chain transfer agent, and water for 30 min. In the first stage, a dispersion of seed particles was prepared by batch emulsion polymerization. Water (3.0 g), Me-β-CD (0.02 g) and SDS (0.03 g) were added in a 100 ml 3-neck flask equipped with a mechanical stirrer, nitrogen inlet and condenser. The flask was immersed in an oil bath. The system was thoroughly purged with nitrogen while the reaction mixture was heated to 80 o C. After the reactor temperature stabilized at 80 o C, the KPS solution (0.06 g in water 0.5 g) as an initiator and the Na 2 CO 3 solution (0.05 g in water 0.5 g) as a ph buffer were added into the reactor followed by the addition of 3 wt % of the monomer pre-emulsion. The mixture was stirred for 20 min at 80 o C. In the second stage of polymerization, the remaining monomer pre-emulsion was fed into the seed latex dispersion together with an initiator aqueous solution (0.01 g in water 2.0 g) in 4 h. The feeding rates were kept identical, controlled by Fluid Metering QG50 pumps. After the addition was completed, the system was maintained at 80 o C for 0.5 h. Then the reaction was cooled to room temperature. Fluorescence donor and acceptor labeled PBMA latices were synthesized in a similar fashion. For the donor labeled particles, 1 mol % (2.4 wt %) PheMMA (based on total monomer) was added into the monomer pre-emulsion. For the acceptor labeled particles, 0.3 mol % (0.8 wt %) NBenMA (based on total monomer) was added into the monomer pre-emulsion. 10

28 Table 2-2. Typical semi-continuous emulsion polymerization recipe for the synthesis of branched PBMA latex ingredients (g) first stage second stage H 2 O 3.0 SDS Me-β-CD 0.02 Na 2 CO KPS monomer pre-emulsion H 2 O 10.0 SDS BMA 4.60 / 32 mmol / 65 wt % BPDM 1.18 / 3.2 mmol / 17 wt % C 12 -SH 1.31 / 6.4 mmol / 18 wt % Preparation of P(BA-MMA)/HB-PBMA latex blends P(BA-MMA)/HB-PBMA latex blends were prepared by mixing HB-PBMA latex with the mixtures of donor (D)- and acceptor(a)- labeled P(BA-MMA) latices. The weight ratio of donor and acceptor labeled P(BA-MMA) is 1 : 1. And the weight ratio of D to A in this mixture is 2.9 : 1.0. The HB-PBMA content is in the range of 0 wt % to 10 wt % (based on total P(BA-MMA) polymer). 11

29 2.2.4 Preparation of P(BA-MMA) latices containing Texanol TM P(BA-MMA) latices containing Texanol TM were prepared by adding 0 wt % to 10 wt % Texanol TM (based on total P(BA-MMA) polymer) into the mixtures of D- and A-labeled P(BA- MMA) latices. The latices were stirred for 48 h at ambient condition before use Synthesis of P(BA-MMA) latices containing hyperbranched poly(nbutyl methacrylate) (HB-PBMA) P(BA-MMA) latices containing hyperbranched PBMA were synthesized using a miniemulsion polymerization method. A typical recipe for the miniemulsion polymerization reaction is shown in Table 2-3. A monomer emulsion was prepared by sonicating a mixture of all of the ingredients in an ice bath for 20 minutes using a Branson Model 450 Digital Sonifier (400w, microtip 40% maximum power, pulse: 1.0 s on/1.0 s off). The monomer emulsion was then transferred into a reactor. The system was thoroughly purged with nitrogen before the reactor was immersed in an oil bath which was pre-heated to 80 o C. The system was maintained at 80 o C for 5 h. The reaction was then cooled to room temperature. Fluorescence dye comonomers were used to label the latex polymers. To synthesize the donor and acceptor labeled particles, 1 mol % (2.4 wt %) PheMMA (based on total monomer) and 0.3 mol % (0.8 wt %) NBenMA (based on total monomer) were added into the monomer pre-emulsion, respectively. 12

30 Table 2-3. Typical miniemulsion polymerization recipe for the synthesis of donor labeled P(BA- MMA) latex containing 1 wt% HB-PBMA. ingredients (g) H 2 O 13.0 SDS 0.06 HD 0.36 / 1.6 mmol BA 5.5 / 43 mmol / 55 wt % MMA 4.4 / 44 mmol / 44 wt % MAA 0.1 / 1 mmol / 1 wt % C 12 -SH / 0.25 wt % PheMMA / 1 mol % / 2.4 wt % HB-PBMA 0.10 AIBN Synthesis of P(BA-MMA) latices containing hyperbranched poly(nbutyl methacrylate) (HB-PBMA) seed particles HB-PBMA seeded P(BA-MMA) latex samples were synthesized by seeded emulsion polymerization reactions. A typical recipe for the seeded polymerization reaction is shown in Table 2-4. In the first stage, the HB-PBMA latex dispersion (2.87 g, containing 1.0 g polymer), water (3.0 g), and methyl-β-cyclodextrin (0.12 g) were added in a 100 ml 3-neck flask equipped with a mechanical stirrer, nitrogen inlet and condenser. The flask was immersed in an oil bath. The system was thoroughly purged with nitrogen while the reaction mixture was heated to 80 o C. The KPS solution (0.14 g in water 1.0 g) as an initiator and the Na 2 CO 3 solution (0.12 g in water 1.0 g) as a ph buffer were added into the reactor In the second stage of polymerization, a pre-emulsion (39.3 g) of monomers, surfactant, chain transfer agent, and water was fed into the seed latex dispersion together with an initiator 13

31 aqueous solution (0.01 g in water 2.0 g). The monomer feeding rate was kept identical (0.1 ml/min), controlled by Fluid Metering QG50 pumps, with a total feeding time of 5 h. After the addition was completed, the system was maintained at 80 o C for 0.5 h. Then the reaction was cooled to room temperature. For the donor (D) labeled particles, 1.0 mol % (2.4 wt %) PheMMA (based on total monomer) was added into the monomer pre-emulsion. For the acceptor (A) labeled particles, 0.3 mol % (0.8 wt %) NBenMA (based on total monomer) was added into the monomer preemulsion. Table 2-4. Typical seeded emulsion polymerization recipe for the synthesis of A-labeled P(BA- MMA) latex containing 5 wt % HB-PBMA seed particles ingredients (g) first stage second stage H 2 O 3.0 HB-PBMA latex 2.87 methyl-β-cyclodextrin 0.12 Na 2 CO KPS monomer pre-emulsion H 2 O 20.0 SDS BA / 82 mmol / 55 wt % MMA 8.36 / 83 mmol / 44 wt % MAA 0.19 / 2 mmol / 1 wt % C 12 -SH a / 0.25 wt % NBenMA / 0.3 mol % a 1-Dodecanethiol, chain transfer agent, used at 0.25 wt % of total monomers. 14

32 2.3 Characterizations of latex particles The solids content of each latex dispersion was determined by gravimetry. Particle diameters were measured by dynamic light scattering at a fixed scattering angle of 90º at 23 o C with a Brookhaven Instruments model BI-90 particle sizer equipped with a 10 mw He-Ne laser. Particle sizes and size distributions were also measured by capillary hydrodynamic fractionation chromatography using a MATEC model 2000 CHDF. 2.4 Characterizations of latex polymers Molecular weight and molecular weight distribution Polymer molecular weight and polydispersity index (PDI) were measured by gel permeation chromatography (GPC) using a Viscotek liquid chromatograph equipped with a Viscotek model 2501 UV detector and a Viscotek TDA302 triple detector. Two Viscotek GMHHR Mixed Bed columns were used with tetrahydrofuran (THF) as the elution solvent at a flow rate of 0.6 ml/min. Polystyrene standards were used for calibration Copolymer composition Copolymer compositions in P(BA-MMA) and branched PBMA latices were determined by 1 H NMR spectroscopy. 1 H NMR spectra were recorded on a Varian Mercury 300 MHz NMR spectrometer using CD 2 Cl 2 or CDCl 3 as the solvent in 5 mm NMR tubes. The residual signal of the solvent was used as a reference in all the spectra (CD 2 Cl 2 δ 5.32, CDCl 3 δ 7.24) Glass transition temperature The glass transition temperature (T g ) of copolymers was measured with a TA Instruments DSC Q100 V7.3 Build 249 differential scanning calorimeter over a temperature range of -50 to 150 o C at a heating rate of 10 o C/min. Each sample was taken through two runs. T g values were calculated from the second heating run. 15

33 2.4.4 Acceptor dye concentration measurements The UV spectra of NBen-labeled latex samples were measured by a Lambda 25 UV/VIS spectrometer (PerkinElmer Instruments). A calibration curve was made based on the absorbance at 341 nm of NBenMA solutions in THF. Then the dye concentration in the NBen-labeled polymer was computed based on this calibration curve Gel content measurements Gel content was measured by the centrifugation method developed in our laboratory: 3 A latex sample (1.0 g) was dried to a constant weight W 0. The dried polymer was subsequently immersed in tetrahydrofuran (THF, 10 ml). The mixture was agitated gently at room temperature for 24 h. The resulting solution was then centrifuged at 20,000 rpm for 30 min, and the top transparent layer was poured off. When gel was present, a precipitate remained. The precipitate was washed three more times with excess THF to remove residual sols from the gel. The remaining sample (the gel fraction) was dried and weighed (W 1 ). The gel content (%) was calculated from the equation gel content (%) = (W 1 / W 0 ) 100 (2-1) 2.5 Mechanical property measurements Rheology measurements. The viscoelastic response of P(BA-MMA) and PBMA samples was studied at several temperatures above T g with a Rheometrics RAA instrument in the oscillatory shear mode. These experiments employed a pair of parallel plates (25 mm diameter). The frequency was scanned between 0.01 and 100 rad/s at a constant temperature. Strain sweeps were employed to insure that all measurements were made in the linear viscoelastic regime. The range of temperatures studied was selected to be as close as possible to the range of temperature used in the energy transfer experiments performed on these materials. However, the lowest temperatures used are limited by the sample modulus at temperatures close to T g, and by the fact that samples that are too stiff can damage the transducer of the rheometer. The following procedure was used to prepare the samples for the measurements of viscoelastic properties. First, the samples were dried under vacuum at 40 o C for 12 h to eliminate 16

34 any trace of volatiles. Then, the samples were pressed between cleaned polytetrafluoroethylene (PTFE) sheets in a Carver Press at 100 o C to eliminate air bubbles. The thickness of the samples was controlled using separators between the plates of the press. In this way, samples free of bubbles, approximately 25 mm in diameter and 1 mm thick were obtained Tensile testing Tensile testing was performed at a strain rate of 50 mm/min with an Instron 5543 tensile tester at ambient temperature. The averaged values of tensile strength, elongation, and Young's modulus were obtained from at least five specimens for each sample. The following procedure was used to prepare the specimen for the stress-strain measurements. First, the samples were dried under vacuum at 80 o C for 12 h to eliminate any trace of volatiles. Then, the samples were pressed in a stainless steel mold between cleaned polytetrafluoroethylene (PTFE) sheets with a Carver Press at 120 o C to eliminate air bubbles. The thickness of the samples was controlled by the mold (22 mm 4.8 mm 1 mm). In this way, dumbbell shape specimens free of bubbles, approximately 1 mm thick, were obtained Dynamic mechanical analysis The storage modulus (E ) and loss modulus (E ) of latex polymers were measured by dynamic mechanical tests using a TA Instruments DMA Q800 dynamic mechanical thermal analyzer in the single cantilever mode. The frequency used was 1.0 Hz, and the heating rate was 2.0 o C/min. The rectangular specimens were prepared using the procedure described above with a stainless steel mold (20 mm 5.3 mm 0.5 mm). The thickness is ca. 0.5 mm. The experiments were carried out from -80 o C to 100 o C. 2.6 Film preparation Latex films for fluorescence decay measurements were prepared from 1:1 particle mixtures of the D- and A-labeled dispersions. Several drops of a latex dispersion (about 20 wt % solids content) were spread on a small quartz plate (20 8 mm). The film was allowed to dry uncovered at 4 o C until visually transparent. It took about 2 h for a film to dry. The films prepared in this way have a thickness of ca. 60 μm. 17

35 Solvent-cast films were prepared from the same latex mixture. A latex film was allowed to dry at 23 o C overnight, and the dry film was dissolved in a minimum amount of THF. The solution was re-cast onto a small quartz plate and allowed to dry at room temperature for 24 hours. 2.7 Film annealing Latex films on quartz plates were placed directly on a high mass (2 cm thick) aluminum plate in an oven preheated to the annealing temperature and then annealed for various periods of time. Under these conditions, I estimate that it takes less than 1 minute for the film to reach the preset oven temperature. The annealed films were taken out of the oven and placed directly on another high mass aluminum plate at 4 o C for 2 minutes before carrying out fluorescence decay measurements. 2.8 Fluorescence decay measurements and data analysis Fluorescence decay profiles of the films at 23 o C were recorded using a nanosecond Time- Correlated Single Photon Counting System from IBH. 4 Each film was placed in a quartz tube and excited with a NanoLED (λ ex = 296 nm). An emission monochromator (350 ± 16 nm) was used to minimize the amount of scattered light from the sample entering the detector. Data were collected until 5000 counts were accumulated in the maximum channel. The instrumental response function was obtained by using a degassed p-terphenyl solution (0.96 ns lifetime) as a mimic standard. 5 For all the latex polymers examined in this work, the donor (Phe) decay profile in films free of acceptors was exponential with a lifetime τ D = 44.3 ns. For films containing both donor and acceptor chromophores, the fluorescence-decay profiles became non-exponential. The shape of the curve depends upon the details of the donor-acceptor pair distribution. In a system containing uniformly distributed donor and acceptors in three dimensions in the absence of diffusion, the donor fluorescence intensity decay I D ( t ) following instantaneous excitation is described by the Förster equation, 6 18

36 I D 2 [ t / τ P( t / τ ) ] 1/ ( t ) = Aexp D D (2-2) where 1/ / P = π κ N AR0 [ A] (2-3) 3 2 Here, P is proportional to the acceptor (quencher) concentration [A]. R 0 is the critical Förster radius for energy transfer, which for the Phe/NBen pair takes a value of 2.51 ± 0.04 nm. 7 N A is Avogadro s number. The orientation factor 2 κ describes the average orientation of dipoles of donor and acceptor molecules. For a random distribution of immobile chromophores in threedimensions, 2 κ is replaced by 2 κ =0.476, a situation typical of dyes in polymer matrices. 5, 8 The quantum efficiency of energy transfer Φ ET (t) is defined by the middle term in the following expression Φ ET ( t) = I D ( t, t)dt area( t) = 1 I ( t )dt τ D 0 D (2-4) where 0 ( t ) is the decay curve of donor fluorescence intensity in the donor-only film. Because I D the unquenched donor decay profile was exponential in each sample, its integral is equal to the unquenched donor lifetime τ D. Here t is the annealing time after film preparation; t is the fluorescence decay time; and area(t) refers to the normalized area under the fluorescence decay curve of a film annealed for time t. To obtain the area for each decay profile, I fitted each decay curve to the empirical equation (2-5) and then evaluate the integral analytically from the 1, 2, 9 magnitude of the fitting parameters, A 1, A 2, and p. I D 2 [ t / τ p( t / τ ) ] 1/ + A exp( t / τ ) ( t, t) = A1 exp D D 2 D (2-5) The fraction of mixing f m is an important parameter measuring the extent of growth of Ф ET due to polymer diffusion, and is defined in such a way that it corrects for the energy transfer efficiency in the nascent films. Values of f m are calculated from fluorescence decay data using the following equation f m () t Φ () t ΦET ( 0) ( ) ( 0) ET = (2-6) Φ Φ ET ET 19

37 where, in principle, the numerator represents the change in energy transfer efficiency between the freshly prepared film and that annealed for time t, and the denominator describes the difference in energy transfer efficiency between the initial and the fully mixed films. Because some polymer diffusion can occur during sample drying, I fitted data to equation (2-6) using a value of Ф ET (0) = 0.07, which corresponds to films formed from latex particles of the size employed here, which achieve intimate contact upon drying, 10, 11 but for which no polymer diffusion takes place Calculation of apparent diffusion coefficients D app If the diffusing polymer is initially distributed uniformly through a sphere of radius R, by assuming a Fickian diffusion model for spherical geometry a concentration profile C(r, t) of the polymer corresponding to an apparent diffusion coefficient D app can be calculated. A numerical method can be then used to obtain the best D app matching an experimental f m value to the theoretical expression: f m f s R 1 2 = 1 C( r, t)4π r dr (2-7) V 0 where C(r, t) is the concentration of polymer at radius r and time t, and V is the volume of a particle of radius R. In equation (2-7) it has been assumed that the quantum fraction of mixing f m is equal to the mass fraction of mixing f s. Our group has analyzed this assumption in the past finding that f m and f s values are proportional for all values of f m 0.7, except at the very end of the experiment, and that the assumption f m = f s overestimates the diffusion coefficient by a factor of Fujita-Doolittle fitting The diffusion coefficients were fitted into the Fujita-Doolittle equation (FDE), which is a one-parameter model used to compare the diffusion coefficients in the presence of a plasticizer to that in its absence. 13 The FDE is shown as following: 20

38 Dp ln D ( T, a ) ( T,0) p 1 ( T, ) ( T ) 2 φ f p 0 = f p ( T,0) + φaβ (2-8) where D p represents the polymer diffusion coefficient, Φ a refers to the volume fraction of the coalescing agent, f p (T,0) is the fractional free volume of the polymer with no added coalescing agent, and β(t) is the difference in fractional free volume between the coalescing agent and the polymer at temperature T. {ln[d p (T, Φ a )/D p (T, 0)]} -1 was plotted against 1/Φ a to give a linear line. From the intercept and the slop I calculated f p (T,0) andβ(t) References 1 Oh, J. K.; Wu, J.; Winnik, M. A.; Craun, G. P.; Rademacher, J.; Farwaha, R. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Oh, J. K.; Wu, J.; Winnik, M. A.; Craun, Gary P.; Rademacher, J.; Farwaha, R. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Tronc, F.; Liu, R.; Winnik, M. A.; Eckersley, S. T.; Rose, G. D.; Weishuhn, J. M.; Meunier, D. M. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, O Connor, D. V.; Phillips, D. Time-Correlated Single Photon Counting. Academic Press: New York, James, D. R.; Demmer, D. R. M.; Verrall, R. E.; Steer, R. P. Rev. Sci. Instrum. 1983, 54, Bartels, C. R.; Buckley, C.; Graessley, W. W. Macromolecules 1984, 17, Liu, Y.; Haley, J. C.; Deng, K.; Lau, W.; Winnik, M. A. Macromolecules 2007, 40, Baumann, J.; Fayer, M. D. J. Chem. Phys. 1986, 85, Liu, Y.; Haley, J. C.; Deng, K.; Lau, W.; Winnik, M. A. Macromolecules 2008, 41, Wool, R.P.; O Connor, K.M., J. Appl. Phys., 1981, 52, Kim, Y.H.; Wool, R.P. Macromolecules 1983, 16, Farinha, J.P.S.; Martinho, J.M.G.; Yekta, A.; Winnik, M.A. Macromolecules 1995, 28,

39 13 Juhué, D.; Wang, Y.; Winnik, M. A. Makromol. Chem. Rapid Commun. 1993, 14,

40 Chapter 3 3 Effect of Polymer Composition on Polymer Diffusion in Poly(n-butyl acrylate-co-methyl methacrylate) Latex Films 3.1 Introduction Latex paints contain significant amounts of volatile organic solvents (VOCs). 1 These solvents serve as fugitive plasticizers. They soften the particles so that the forces associated with drying are sufficient to deform the spherical latex particles into polyhedral cells that form a continuous and void-free film. They also enhance the rate of the diffusion of polymer molecules across the boundaries between these cells. This is the step that leads to the development of good mechanical properties of the latex film. At this time the polymer film is often soft and tacky. Over time, the VOCs evaporate from the film, increasing its glass transition temperature and its hardness at room temperature. The disadvantage of VOCs is that they contribute to air pollution. New knowledge is needed, however, to develop latex coatings that do not require VOCs and have similar or enhanced performance properties to current technology. My thesis research has the objective of providing some of this new knowledge. As a step in this direction, I carried out a study of polymer diffusion in films formed from a series of latex consisting of n-butyl acrylate-methyl methacrylate-methacrylic acid (BA-MMA- MAA) copolymers of different compositions. These acrylic latices are widely used in architectural coatings (house paints). Typical coatings in current use consist of BA/MMA weight ratios of 50/49 with 1 wt % MAA. Increasing the BA content in the latex reduces the glass transition temperature (T g ) of the latex polymer, however, the other effects are kept unknown. In this chapter, I described experiments that explore how the composition of the latex polymer affects the rate of polymer diffusion (and its temperature dependence), for a series of BA/MMA latex containing 1 wt % MAA, all with polymers of very similar molar mass (M n 40,000 g/mol, M w /M n 3.0). Their glass transition temperatures range from 4 o C to 28 o C. These polymer diffusion rates were studied using the fluorescence resonance energy transfer (FRET) methods developed in our laboratory. 2 To perform the FRET experiments, fluorescent donor (D)- 23

41 and acceptor (A)-labeled latex particles were synthesized using semi-continuous emulsion polymerization. Latex films were cast from a 1:1 mixture of D- and A-labeled latex samples. Polymer diffusion was monitored as a function of annealing temperature, and apparent diffusion coefficients (D app ) were calculated from the energy transfer data using a simple diffusion model. These values increased with annealing temperature and decreased with T g. Rheology measurements recorded the response of the dynamic moduli (G', G ) with respect to oscillatory shear frequency (ω) over a range of temperature close to that of the diffusion experiments. It was found that the temperature dependence of polymer dynamics extracted by the rheology experiments is in good agreement with the temperature dependence of D app. Increasing the BA copolymer content leads to an apparent increase in long-chain branching, which is reflected in both the time dependence of D app and in the dynamic moduli measurements. I can concluded that a greater degree of branching leads to a broader distribution of polymer diffusion coefficients, and a stronger time dependence of D app. 3.2 Results Preparation and Characterization of the Latex Samples Poly(n-butyl acrylate-co-methyl methacrylate) [P(BA-MMA)] was used as the base copolymer in my diffusion study. All latexes were prepared by semi-continuous emulsion polymerization. In previous work in our laboratory, in order to obtain similar size D- and A- labeled particles, unlabeled seed particles were used for the synthesis of both the donor and acceptor labeled latex, and the polymerizable dye derivative was added only in the second stage. 3, 4 By using the same unlabeled seed particles it is easy and efficient to control particle size. These seeds represent ca wt % of the final latex particles. If the seed particles are prepared by batch emulsion polymerization, they normally have a higher molar mass and broader PDI than the second stage polymer. To some extent, this may lead to a non-uniform dye distribution in the particles. Our laboratory has always treated this as a minor problem. Nevertheless, in order to overcome the disadvantage of preformed non-labeled seeds, I used an in-situ seeding process in the experiments described here. While this is common practice in industry for large-scale reactions, this is a skill-testing challenge for emulsion polymerization 24

42 reactions run on the small scale (10 g of total monomers) I employ to synthesize labeled latex. On a small scale, slight variations in particle nucleation can have large consequences for particle size and size distribution. One can also experience problems in matching the polymer molar mass and PDI between samples. The challenge for me, which in the past led to the use of common unlabeled seeds, was the need for the donor and acceptor labeled particles to have similar diameters and contain polymers of similar M n and PDI. For the latex samples described here, I employed the same monomer pre-emulsion containing: monomers, dye comonomers, surfactant, chain transfer agent and water for making the seed particles in the first stage and for particle growth in the second stage as well. These reactions worked well, and I was able to achieve reasonable control over particle size and polymer molar mass, not only for D- and A- labeled particles of a given composition, but for the entire series of latex examined here. 25

43 Table 3-1. Characteristics of the P(BA-MMA) latex polymers and particles. Latex sample a M n PDI T g ( o C) d b (nm) d n c (nm) d w /d n c solids content (%) P(BA 60 -MMA 39 ) 31, P(BA 60 -MMA 39 ) D 31, P(BA 60 -MMA 39 ) A 29, P(BA 55 -MMA 44 ) 48, P(BA 55 -MMA 44 ) D 47, P(BA 55 -MMA 44 ) A 46, P(BA 50 -MMA 49 ) 50, P(BA 50 -MMA 49 ) D 45, P(BA 50 -MMA 49 ) A 43, P(BA 40 -MMA 59 ) 51, P(BA 40 -MMA 59 ) D 44, P(BA 40 -MMA 59 ) A 41, a Superscript D and A refers to D- and A-labeled latices, respectively. b Data for the particle diameter d from the BI-90 Particle Sizer. c Number average d n and weight average d w diameter data from the CHDF 2000 (MATEC). 26

44 All emulsion polymerizations contained 1 wt % of MAA. In commercial latex, small amounts of methacrylic acid are normally employed to enhance the colloidal stability of the latex. I follow this practice here. The four pairs of samples I synthesized had monomer weight ratios of BA:MMA:MAA of 60:39:1, 55:44:1, 50:49:1 and 40:59:1. These copolymers are named Figure 3-1. Plot of 1/M n against concentration of C 12 -SH of P(BA 60 -MMA 39 ) latex samples. according to their BA:MMA compositions as P(BA 60 -MMA 39 ), P(BA 55 -MMA 44 ), P(BA 50 - MMA 49 ) and P(BA 40 -MMA 59 ). The glass transition temperatures (T g ) of the copolymers were measured by DSC, giving T g ca. 4 o C for P(BA 60 -MMA 39 ), 7 o C for P(BA 55 -MMA 44 ), 12 o C for P(BA 50 -MMA 49 ) and 27 o C for P(BA 40 -MMA 59 ). All values are similar to those estimated values from the Fox equation using T g (PBA) = -47 o C and T g (PMMA) = 105 o C. 5 Dodecyl mercaptan (C 12 -SH) was added in the emulsion polymerization reactions as a chain-transfer agent both to control the molecular weight and to limit or suppress gel formation. My target in these studies was to obtain high molar mass with very low gel content. This will serve as a baseline for future experiments with latex comprised of lower molar mass polymer. In order to optimize the reaction conditions, a series of P(BA 60 -MMA 39 ) latex samples were prepared in the presence of various amounts of C 12 -SH. The latex polymer samples obtained were analyzed by a triple detector GPC. Molecular weights were determined using polystyrene standards. As shown in Figure 3-1, the plot of 1/M n against [C 12 -SH] was linear, indicating that in the presence of β-cyclodextrin, C 12 -SH provides good control over polymer molar mass. When the amount of C 12 -SH in the reaction was 0.25 wt % based on monomer in the pre- 27

45 emulsion, the polymers obtained had less than 5% gel; with lower amounts of chain transfer agent, the latex formed had a significant gel content. For example, 0.10 wt % C 12 -SH led to a product containing 46% gel content. Thus all the latex samples used in the diffusion experiments were prepared in the presence of 0.25 wt % C 12 -SH. The characteristics of all of the latex particles synthesized are summarized in Table 3-1. The M n values were in the range of 30,000 to 50,000 with a PDI between 2 and 3.7. By comparing M n of the dye labeled and non-labeled latex polymers (see Table 3-1), I infer that the dye monomer did not significantly affect the polymerization reaction. The 1 H NMR spectra of the four different latex polymer compositions are compared in Figure 3-2. From the ratio of integrals of peaks a and b, which correspond to protons at position a and b respectively, I calculated the mole ratios of BA:MMA were 1.5:1.0 for P(BA 60 -MMA 39 ), 1.1:1.0 for P(BA 55 -MMA 44 ), 0.8:1.0 for P(BA 50 -MMA 49 ) and 0.5:1.0 for P(BA 40 -MMA 59 ), the weight ratios of BA:MMA were 65:35 for P(BA 60 -MMA 39 ), 58:42 for P(BA 55 -MMA 44 ), 51:49 for P(BA 50 -MMA 49 ) and 41:59 for P(BA 40 -MMA 59 ). These results indicate that the composition of the polymers closely resembled the monomer feed composition in my emulsion Figure H NMR spectra of (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ). CDCl 3 was used as solvent. Peaks a and peak b correspond to protons at positions a and b respectively. 28

46 polymerization reactions under monomer-starved conditions. The particle size and size distribution were characterized by both right-angle dynamic light scattering and by CHDF. As shown in Table 3-1, all samples have particle diameters of ca. 150 nm. For all latex samples the d w /d n values obtained by CHDF are less than 1.3, which indicates a narrow size distribution Energy Transfer Studies of Polymer Diffusion Films for FRET experiments were prepared from a 1:1 mixture D- and A-labeled latex particles. Several drops of the latex mixture was cast onto a small quartz plate (20 8 mm) and allowed to dry in a refrigerator at 4 o C over 1 h. The film thicknesses were ca. 60 μm. The films obtained from P(BA 60 -MMA 39 ), P(BA 55 -MMA 44 ) and P(BA 50 -MMA 49 ) D/A mixtures were transparent and free of cracks. However, the films prepared from P(BA 40 -MMA 59 ) latex were turbid and showed some cracks even if they were dried at 23 o C. I attribute this behavior to the high minimum film formation temperature (MFT) of this sample caused by the high T g of the latex polymers. Freshly formed latex films were transfer to the sample chamber of the fluorescence decay instrument in a cold quartz tube, and fluorescence decays were measured immediately. This whole process took less than 2 min. The films were then annealed in a preheated oven for various periods of time at constant temperatures. The fluorescence decays were monitored as a function of annealing time for samples annealed at different temperatures. Figure 3-3 shows representative donor fluorescence decays for a D-labeled P(BA 60 - MMA 39 ) latex film [curve (1)] and a D/A mixed P(BA 60 -MMA 39 ) latex film aged for various periods of time at room temperature (23 o C) [curves (2-3)]. The decay curve (1) in Figure 3-3 is exponential with a lifetime of 44.3 ns. Films of the D-labeled P(BA 60 -MMA 39 ) latex with the other three BA-MMA compositions gave the same lifetimes. For these polymers, I see that the donor lifetime is independent of polymer composition. Curve (2) in Figure 3-3 is the decay profile of a newly formed film from a 1:1 D/A mixture. It is not exponential, but the deviation at early times is small (shown in Figure 3-3 inset) because very little polymer diffusion has occurred. I assume that the major contribution to the curvature of this plot is energy transfer across the particle boundaries. From the decays of a series of similar films, I calculate quantum efficiencies of energy transfer (Ф ET ) of in newly formed films using equation (2-4). I take 0.06 as the value of Ф ET (0). Curve (3) in Figure 3-3 depicts the decay profile of the film in curve (2) after 47 min aging at 23 o C. The increased curvature at early decay times in an 29

47 indication that some polymer diffusion has taken place, resulting in an increase in Ф ET. Figure 3-3. Phenanthrene (donor) fluorescence decay curves [I D (t)] measured at 23 o C for Phe- P(BA 60 -MMA 39 ) latex films. (1) Phe-labeled latex only, (2) a newly formed film dried at 4 o C, consisting of a 1:1 ratio of Phe-P(BA 60 -MMA 39 ) and NBen-P(BA 60 -MMA 39 ), (3) the same film as in (2) aged for 47 min at 23 o C, and (4) a solvent-cast film from a 1:1 mixture of the two freeze-dried polymers dissolved in THF and then annealed at 120 o C for 2 h. Note that curves (1) and (2) overlap. The inset shows curves (1) and (2) at short times on a linear scale. Curve (4) in Figure 3-3 shows much more pronounced curvature. It represents the decay profile for a film cast from a THF solution. In solvent cast films, the donor and acceptor labeled polymers may be thought of as randomly mixed in solution, but can undergo some demixing upon drying, due to correlation hole effects. I refer to the Ф ET value obtained from solvent-cast films as Ф ET (lim). It represents the limiting maximum value of Ф ET that could be obtained from diffusive mixing. This value can be smaller than, but is often equal to Ф ET ( ), the value corresponding to complete randomization of the dyes in the system. For latex films formed from linear polymers, Ф ET (lim) and Ф ET ( ) are commonly very similar in magnitude, but in latex films consisting of highly branched polymers or polymer with a significant gel content, Ф ET (lim) < Ф ET ( ). Values of Ф ET ( ) were determined in a series of model experiments in which samples of each of the Phe-labeled polymers were mixed with different amounts of NBenMA as a low molar mass acceptor. Films were prepared by solvent casting and I D (t) decay profiles were measured. Individual decays were fitted to equation (2-5), and values of the fitting parameter P were plotted against [NBenMA] (Figure 3-4). The plot were linear and for each polymer led to a 30

48 value of R o = 2.51 nm, consistent with the value reported previously. 5 From this value and the composition of the 1:1 Phe/NBen latex films, I calculated values of Ф ET ( ) = I used this value in all of my calculations of f m [(equation (2-6)]. P [NBenMA, mm] Figure 3-4. Plot of P vs [NBenMA, mm] for fully mixed solvent-cast films prepared from Phe-P(BA 60 - MMA 39 ) plus varying amounts of free monomer MBenMA. The P values were obtained by fitting individual Phe decay curves to equation (2-5) with τ D fixed at 44.3 ns. From the slope of the plot, I calculate R 0 = 2.51 nm Polymer Diffusion in P(BA 60 -MMA 39 ) Films at Different Temperatures A series of Phe- and NBen-labeled P(BA 60 -MMA 39 ) latex films were cast at 4 o C and dried for 1 h. The films were annealed at various temperatures, and their fluorescence decay curves were measured at different periods of annealing time. From the newly formed films I found the quantum efficiency before annealing, Ф ET (0) = The maximum Ф ET value was obtained from fully mixed D/A films which were cast from a THF solution of 1:1 Phe- and NBen-labeled P(BA 60 -MMA 39 ) polymers. The Ф ET value for this newly formed solvent cast film was Annealing this film at 120 o C for 2 h lead to a decrease of the Ф ET value to 0.38, and there was no additional decrease with further annealing. I take Ф ET (lim) = The calculated Ф ET values are plotted against annealing time in Figure 3-5(A) for experiments at different temperatures ranging from 23 o C to 90 o C. The curves show a large increase in Ф ET values at early stages and a smaller increase at longer times. The plot shows that 31

49 Ф ET has a strong temperature dependence. From 23 o C to 90 o C, not only the plateau Ф ET values, but also the growth rate of Ф ET increased significantly. Since I know that Ф ET (0) = 0.06 and Ф ET ( ) = 0.5, fraction of mixing f m values were calculated from the corresponding areas under the donor decay profiles using equation (2-6). In Figure 3-5(B), I plot f m as a function of annealing time. The curves have a similar shape to those in Figure 3-5(A). At 90 o C, maximum Figure 3-5. Plots of Ф ET (A) and f m (B) vs annealing time for the P(BA 60 -MMA 39 ) latex films annealed at 23, 45, 70, and 90 o C. Figure 3-6. Plots of Ф ET and fm vs annealing time for the P(BA60-MMA39) latex films annealed at 23 oc. 32

50 mixing of donor and acceptor was achieved in 90 min with Ф ET (lim) = 0.38 and f m = For architectural coatings, one of the most important considerations is the time scale for polymer diffusion at room temperature. In our laboratory, this is 23 C. To emphasize that the polymer molecules in films of the P(BA 60 -MMA 39 ) latex, with T g = 4 C undergo substantial diffusion at room temperature, I plot the evolution of Ф ET and f m in Figure 3-6. To quantitatively compare polymer diffusion rates at different temperatures, one needs to be able to compute diffusion coefficients D. Because there is no proper way to calculate absolute values of D for mixtures of polymers of different lengths and extents and distribution of branches, I resort to a prescription that has served our research group well in the past: I calculate apparent diffusion coefficients D app by fitting f m data to a Fickian diffusion model. 4,6, 7 Figure 3-7. Plots of the apparent diffusion coefficient D app as a function of f m for (A) P(BA 60 - MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ) latex films annealed at various temperatures. 33

51 In Figure 3-7(A), values of D app calculated in this way for P(BA 60 -MMA 39 ) films are plotted against f m values for various annealing temperatures. At each temperature, these D app values decrease with increasing annealing time as slower diffusing species make their contribution to the growth in Ф ET. The plot also shows that the diffusion rate is faster at higher temperature for the same f m value. For example, at f m = 0.59, the annealed film sample gave a value of D app = nm 2 /s at 23 o C, 0.16 nm 2 /s at 45 o C, and 4.4 nm 2 /s at 60 o C. Arrhenius-type plots (ln D app vs 1/T) of the data in Figure 3-7(A) are linear for D app values at f m = These plots are shown in Figure 3-8. From the slopes of these plots, I obtained an apparent activation energy E a = 33.4 kcal/mol over the temperature range o C. Since temperature affects the rate of diffusion by a change in the monomeric friction factor, the magnitude of E a should be independent of f m. Therefore, I used this value of E a as a shift factor to create a master curve of D app values at 23 o C. The shifted values calculated in this way are shown in Figure 3-10 below. The success in generating the master curve serves as strong support for the validity of my analysis to obtain D app values. 2.0 E a = 33.4 kcal/mol ln D app (nm 2 /s) (1000/T) [1/K] Figure 3-8. Plot of ln D app against 1/T over the temperature range from 23 to 60 o C at f m values of 0.59 for P(BA 60 -MMA 39 ) latex. 34

52 3.2.4 Polymer Diffusion in Different Composition P(BA-MMA) Films To compare the polymer diffusion in P(BA-MMA) latex films with various polymer compositions, I monitored the increase in Ф ET for a series of latex films of each composition, each annealed at a series of temperatures. In Figure 3-9, these Ф ET values are plotted as a function of annealing time at their corresponding temperatures. Figure 3-9(A) shows that at 23 o C, three of the four latex films of different composition undergo a significant extent of diffusion on the time scale of tens of hours. The T g values of these polymer compositions range from 3 to 12 C (Table 3-1), and the rate of diffusion at room temperature increases with decreasing sample T g. Only the P(BA 40 -MMA 59 ) film showed no detectable diffusion at this temperature. It has a T g slightly above room temperature, 28 C. There are several noteworthy features of this particular set of films. When these films were cast and dried at 4 C, well below the MFT, they were turbid and cracked. Better films for polymer diffusion studies were obtained by casting and drying at room temperature although they were still not clear and crack free due to their high MFT. Upon annealing at higher temperature, the films became more transparent, but the film became fully transparent only after it was annealed for ca. 1 h at 90 o C. Thus dry sintering plays a role in particle coalescence in these films. 8 An increase in Ф ET in these films due to polymer diffusion could be measured at 70 C [Figure 3-9(B)]. This occurred over the first hour and then appeared to cease. I speculate that this increase in Ф ET is due to the contribution of diffusion of low molar mass chains in the sample. For the remainder of the experiment, polymer diffusion was very slow at this temperature, but became more pronounced at 90 C [Figure 3-9(C)]. The other samples underwent rapid polymer diffusion at 70 C. For P(BA 60 -MMA 39 ) film, Ф ET approached Ф ET (lim) (0.38) in a few minutes. The P(BA 55 -MMA 44 ) and P(BA 50 -MMA 49 ) latex film samples exhibited a behavior analogous to that of the high T g sample: rapid diffusion at early times, which appeared to level off at a Ф ET value less than Ф ET (lim) [Ф ET (lim) = 0.50 for P(BA 55 -MMA 44 ) and 0.52 for P(BA 50 -MMA 49 )]. These two films reached somewhat higher values of Ф ET values when annealed for several hours at 90 C. 35

53 Figure 3-9. Plot of the Φ ET for films formed from D/A labeled latex mixtures annealed for various periods of time at (A) 23 o C, (B) 70 o C and (C) 90 o C. ( )P(BA 60 -MMA 39 ), ( )P(BA 55 - MMA 44 ), ( ) P(BA 50 -MMA 49 ) and ( ) P(BA 40 -MMA 59 ). 36

54 It is noteworthy that Ф ET (lim) values for P(BA 55 -MMA 44 ) and for P(BA 50 -MMA 49 ) are higher than that for P(BA 60 -MMA 39 ), and indistinguishable from Ф ET ( ). This result suggests that there is a higher degree of branching or some undetected microgel in this P(BA 60 -MMA 39 ) sample that limits the extent to which donor and acceptor labeled polymers can interpenetrate. D app values for these films were calculated as a function of f m at each temperature. These plots are presented in Figure 3-7(A)-(D). From Arrhenius plots of D app values as described above, apparent activation energies were computed. The values of E a and the temperature ranges for which they were obtained are listed in Table 3-2. These were then used as shift factors to create Master Curves of the diffusion data as shown in Figure Figure The master curves of D app values for (A) P(BA 60 -MMA 39 ) at 23 o C (calculated using E a = 33.4 kcal/mol as a shift factor); (B) P(BA 55 -MMA 44 ) at 23 o C (calculated using E a = 39.1 kcal/mol as a shift factor); (C) P(BA 50 -MMA 49 ) at 23 o C (calculated using E a = 45.2 kcal/mol as a shift factor) and (D) P(BA 40 -MMA 59 ) at 70 o C (calculated using E a = 64.1 kcal/mol as a shift factor). 37

55 Table 3-2. E a values of the latex polymers. energy transfer experiments Rheology Measurements latex E a Temperature E a Temperature T 0 C 2 (kcal/mol) range (kcal/mol) range ( o C) C 1 (K) P(BA 60 -MMA 39 ) o C o C P(BA 55 -MMA 44 ) o C o C P(BA 50 -MMA 49 ) o C o C P(BA 40 -MMA 59 ) o C o C Temperature Dependence of the Viscoelastic Properties of P(BA- MMA) Films The Williams-Landel-Ferry (WLF) equation 9 is widely employed to describe the temperature dependence of polymer diffusion using parameters obtained from viscoelastic relaxation measurement. 4,7, 10 To describe polymer diffusion, the WLF equation takes the following form log ( ) a T DT C ( T T ) = log = (3-1) D0T C2 + T T0 where D 0 is the diffusion coefficient at an arbitrary chosen reference temperature T 0. C 1 and C 2 are parameters that depend on the choice of the T 0, and they are easily transferred to other reference temperatures. 38

56 For each composition, I carried out oscillatory shear measurements as a function of frequency over a range of temperature close to that of the energy transfer experiments. Nonlabeled samples, which have similar molecular weight and PDI to dye labeled samples (see Table 3-1), were used for viscoelastic measurements. I measured the storage modulus (G ) and loss modulus (G ) as a function of frequency (ω) at a series of temperatures ranging from 25 to 100 o C for P(BA 60 -MMA 39 ), from 50 to 120 o C for P(BA 55 -MMA 44 ), from 80 to 180 o C for P(BA 50 -MMA 49 ) and from 130 to 200 o C for P(BA 40 - MMA 59 ) (not shown). I used the time-temperature superposition principle (TTS) to obtain the shift factors (a T ) of the temperature dependence. Strictly speaking, the TTS principle can be only applied to a system in which the various relaxation times belonging to a given relaxation process have the same temperature dependence, such as linear amorphous polymers above T g. P(BA- MMA) copolymer is composed of polydisperse chains with various degrees of branching. 11 It is well-known that branching may affect slightly the temperature sensitivity of the viscoelastic response, but TTS basically holds. 12, 13 In Figure 3-11(A)-(D), I show the G' and G'' master curves after applying TTS, by choosing T 0 = 25 o C for P(BA 60 -MMA 39 ), T 0 = 50 o C for P(BA 55 - MMA 44 ), T 0 = 80 o C for P(BA 50 -MMA 49 ) and T 0 = 90 o C for P(BA 40 -MMA 59 ) as the reference temperatures. Only shifts in the horizontal scale were applied. Shift factors at each temperature were extracted using the generally accepted procedure of overlaying plots of tan(δ) (G''/ G') for data at different temperatures. The rheological response of all four samples is consistent with what is normally found in entangled polymer melts, in particular, that G' > G'' over the portion of the relaxation spectrum that is usually associated with the plateau regime. As observed in this figure, good matching between curves was obtained. 39

57 Figure Plots of master curves of G' and G'' for (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ) latex films at T 0 = 25, 50, 80 and 90 o C 3.3 Discussion Comparison between Different Experiments Values of the apparent activation energy in the range of temperatures studied can be obtained by plotting ln(a T ) in Arrhenius fashion against the inverse of the absolute temperature, as empty squares shown in Figure Normally the ln(a T ) vs 1/T plot is curved, but when the data are limited over a relatively narrow range of temperatures, the plot appears linear. The activation energy E a for viscoelastic relaxation can be calculated from the slope of this linear fraction of the plot. The magnitude of E a value will increase as the measurement temperature approaches T g. From Figure 3-12 average values of E a for each latex sample were calculated over the temperature range close to energy transfer experiments. I compare these E a values as well as their corresponding temperature ranges with those obtained from energy transfer experiments in Table 3-2. It can be clearly observed that the E a values from the two different methods are in good agreement when a similar temperature range was chosen. 40

58 In Figure 3-12, I make a direct comparison between data obtained from rheology measurements and from diffusion experiments. I plotted both data sets in the same graph, where shift factors obtained from rheology are shown as empty squares and those from the diffusion experiments are shown as filled squares. The diffusion data used in Figure 3-12(A) were taken from Figure 3-8, and then shifted vertically to compensate for the reference temperature. The full line represents the WLF fitting, calculated using the C 1 and C 2 listed in Table 3-2. Figure 3-12(B), 11(C), and 11(D) were plotted using the same process. The diffusion data and rheology data appear to track together with changing temperature. Figure Plots of shifted D app and log(a T ) against the inverse of the absolute temperatures for (A) P(BA 60 -MMA 39 ), (B) P(BA 55 -MMA 44 ), (C) P(BA 50 -MMA 49 ) and (D) P(BA 40 -MMA 59 ) latex Effect of long chain branching on the time-dependence of D app One of the striking features of my data is the time-dependence of D app. This effect is best appreciated by examining the master curves of the shifted D app vs. f m in Figure D app 41

59 decreases by several orders of magnitude over the course of the experiment for each of the four copolymers. In Figure 3-10(B)-(D), there is a sharp drop in D app for values of f m less than about 0.2; the high initial rate of diffusion made it impossible for me to capture the short time behavior in the P(BA 60 -MMA 39 ) sample [Figure 3-10(A)], as I could not make measurements at f m values less than 0.2 in this sample. My focus here is on the change in D app over the range in f m values from 0.2 to 0.8. For P(BA 60 -MMA 39 ), D app drops by about a factor of 100. For P(BA 55 -MMA 44 ), there is roughly a factor of reduction in D app. There is a significant amount of scatter in the data for the P(BA 50 -MMA 49 ) master curve data at f m = 0.2; taking this into account, I find that D app decreases over the relevant f m range by a factor of 10 to 50. Again, there is scatter in the data in the P(BA 40 -MMA 59 ) master curve at f m = 0.2, but I estimate an overall drop in D app of a factor of 5 to 10. I rationalize the decrease of D app with increasing time in terms of the distribution of diffusion coefficients of various species in the system. My colleague Dr. Jeffery Haley proposed a simple example to illustrate how this might occur. Imagine a system consisting of two species, one with D = 1 nm 2 /s and a second with D = 0.01 nm 2 /s. Calculations indicate the effect of a diffusing species on the time dependence Φ ET is greatly diminished once that species has diffused over some characteristic distance in the sample. The actual distance involved for a particular experiment will depend on details such as the latex particle sizes and the ratio of donor labeled particles to acceptor labeled particles. Assigning a number to this distance is not so important. The important fact is that for both species in this hypothetical sample, this distance is the same. This means that in this example, the slower moving species will take 100 times longer to diffuse over this characteristic distance. In an energy transfer experiment, one measures the change in Φ ET vs. time, and then convert this information into D app with a diffusion model. 14 Over the course of an experiment in this hypothetical system, the D app value extracted will be some sort of weighted average of the apparent diffusion coefficients of the two species in the system. Early in the experiment, D app will be weighted more heavily towards the diffusion of the faster moving species, as the motion of the faster moving species is largely what causes the change in Φ ET. Once the fast moving species has diffused over a characteristic distance, its contribution to the increase in energy transfer has reached its maximum value. The experiment is 42

60 no longer sensitive to the faster species diffusion, and the rate of Φ ET increase will drop significantly. The result of this is that the extracted D app value will now be weighted more heavily towards the diffusion coefficient of the slower moving species. Of course, this drop in D app is not expected to be sudden for the hypothetical system, but instead occurs gradually. Based on this argument, I suspect that the time-dependence of D app between f m = 0.2 and f m = 0.8 is mostly due to a broad distribution of diffusion coefficients for polymer chains present in the sample. This breadth is apparently larger in P(BA 60 -MMA 39 ) and P(BA 55 -MMA 44 ) than it is in P(BA 50 -MMA 49 ) and P(BA 40 -MMA 59 ), based on the relative magnitudes in the change of D app with f m. I cannot rationalize these differences in terms of sample PDI, as no trend in the PDI values in Table 3-1 exists. Instead, I suspect that differences in the breadth of diffusion coefficients are due to differences in the details of molecular architecture between samples. In particular, the presence of long chain branching can dramatically decrease the diffusion coefficient of a polymer; 15 a sample consisting of a range of branching architectures would be expected to have a broad distribution of diffusion coefficients, and a time-dependent D app. It is difficult to make quantitative determinations about the relative degree of long chain branching for polymers studied here, but I can make some general qualitative comments based on the rheology data in Figure The rheological responses of P(BA 60 -MMA 39 ) and P(BA 55 - MMA 44 ) are clearly quite different from P(BA 50 -MMA 49 ) and P(BA 40 -MMA 59 ). The most obvious difference is the lack of any terminal regime in the P(BA 60 -MMA 39 ) and P(BA 55 - MMA 44 ) data, despite the fact that the data is over a similar range of reduced frequencies. I believe that the differences in mechanical response between samples are due to differences in molecular architecture. Dr. Jeffery Haley, who has a lot of experience with polymer rheology, noticed that the rheological responses of P(BA 60 -MMA 39 ) and P(BA 55 -MMA 44 ) are strikingly similar to what is reported for a crosslinking polymer in the vicinity of the gel point. 16 This is not to say that P(BA 60 -MMA 39 ) and P(BA 55 -MMA 44 ) are in fact gels, but they likely share some limited structural similarities. In particular, the very broad distribution of mechanical relaxation times suggests a high degree of branching with a wide range of branch lengths. I believe that a substantial fraction of the P(BA 60 -MMA 39 ) and P(BA 55 -MMA 44 ) samples are made up of branched polymer. The data for P(BA 50 -MMA 49 ) and P(BA 40 -MMA 59 ) are quite different [Figure 3-11 (C) and (D)], and show crossovers between G and G in the interval between the 43

61 rubbery zone and the terminal zone of the master curve. This response is typical for linear polymers of unimodal molecular weight distributions of moderate PDI, and is anticipated by theory. 17 Based solely on the rheological response of P(BA 50 -MMA 49 ) and P(BA 40 -MMA 59 ), I suspect that chains with significant degrees of long chain branching make up a relatively small amount of the overall population of chains in the system. It is well established that chain transfer to polymer often produces highly branched structures in the emulsion polymerization of n-butyl acrylate. 18, 19 This is consistent with what I infer from rheology experiments, where evidence suggests that a greater degree of long chain branching is present in P(BA 60 -MMA 39 ) and P(BA 55 -MMA 44 ). This accounts for the broader distribution of diffusion coefficients in these samples that is inferred from the time dependence of D app. 3.4 Summary I synthesized donor and acceptor labeled P(BA-MMA) copolymer latex particles by semicontinuous emulsion polymerization in the presence of 0.25 wt % C 12 -SH. Four sets of copolymers were prepared from various weight ratios of BA and MMA. Weight ratios of BA:MMA:MAA for these latexes are 60:39:1, 55:44:1, 50:49:1 and 40:59:1. Their glass transition temperatures (T g ) are 4 o C, 7 o C, 12 o C and 28 o C respectively. Donor labeled latex samples were prepared in the presence of 1 mol % of PheMMA as the dye-containing comonomer. Acceptor labeled latex samples were prepared in the presence of 0.3 mol % of NBenMA. The latex particles had diameters of ca. 150 nm with narrow size distribution. FRET experiments were used to determine the apparent polymer diffusion coefficients as a function of temperature for each of the latex samples. Analysis of the diffusion data gave apparent activation energy E a values of ca. 33 kcal/mol for P(BA 60 -MMA 39 ), 39 kcal/mol for P(BA 55 -MMA 44 ), 45 kcal/mol for P(BA 50 -MMA 49 ) and 64 kcal/mol for P(BA 40 -MMA 59 ). The temperature dependence of the polymer diffusion coefficients closely matches the temperature dependence extracted from a master curve analysis of the rheology data for each latex. The rheology data lead me to conclude that latex polymers with increasing BA content have a greater degree of long-chain branching. Differences in long-chain 44

62 branching also show up in the diffusion measurements, which indicate a larger distribution of polymer diffusion coefficients in samples with higher BA content. These results will guide the development of the next generation of low VOC coatings. 3.5 References 1 Paton, T. C. Paint Flow and Pigment Technology Wiley: New York, Zhao, C. L.; Wang, Y.; Hruska; Z.; Winnik, M. A. Macromolecules 1990, 23, Oh, J. K.; Wu, J.; Winnik, M. A.; Craun, G. P.; Rademacher, J.; Farwaha, R. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Wu, J.; Tomba, J. P.; Winnik, M. A.; Farwaha, R.; Rademacher, J. Macromolecules 2004, 37, Brandrup, J.; Immergut, E. H.; Grulke, E. A. Polymer Handbook 4th ed.; John Wiley & Sons: 1999; Oh, J. K.; Wu, J.; Winnik, M. A.; Craun, Gary P.; Rademacher, J.; Farwaha, R. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Oh, J. K.; Tomba, J. P.; Ye, X.; Eley, R.; Rademacher, J.; Farwaha, R.; Winnik, M. A. Macromolecules 2003, 36, Sperry, P. R.; Snyder, B. S.; O'Dowd, M. L.; Lesko, P. M. Langmuir 1994, 10, Ferry, J. D. Viscoelastic Properties of Polymers Wiley: New York, Nemoto, N.; Landry, M. R.; Noh, I.; Yu, H. Polym. Commun. 1984, 25, Wu, J.; Oh, J. K.; Yang, J.; Winnik, M. A.; Farwaha, R.; Rademacher, J. Macromolecules 2003, 36, Bartels, C. R.; Buckley, C.; Graessley, W. W. Macromolecules 1984, 17, Carella, J. M.; Gotro, J. T.; Graessley, W. W. Macromolecules 1986, 19, Wang, Y.; Zhao, C-L.; Winnik, M. A. J. Chem. Phys. 1991, 95, McLeish, T. C. B. Advances In Physics 2002, 51, Winter, H. H.; Chambon, F. J. Rheol , Wasserman, S. H.; Graessley, W. W. J. Rheol ,

63 18 Former, C.; Castro, J.; Fellows, C. M.; Tanner, R. I.; Gilbert, R. G. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, Gonzalez, I.; Leiza, J. R.; Asua, J. M. Macromolecules 2006, 39,

64 Chapter 4 4 Synthesis of Branched Poly(n-butyl methacrylate) via Semi-Continuous Emulsion Polymerization 4.1 Introduction Latex paints, particularly house paints, are formulated with latex particles consisting of branched polymer. Branching occurs naturally and not intentionally through the choice of monomers used in the synthesis by emulsion polymerization. Coatings referred to as allacrylate are typically copolymers of n-butyl acrylate (BA) and methyl methacrylate (MMA), whereas the remainder of the market is dominated by latex consisting of copolymers of n-butyl acrylate and vinyl acetate (VAc). In the emulsion copolymerization of these monomers, the n- butyl acrylate and vinyl acetate propagating radicals have a strong propensity to abstract hydrogens from the polymer backbone, both intramolecularly and intermolecularly, leading respectively to short- and long-chain branches. Lovell s group 1-3 in particular has presented careful studies of these chain-transfer-to-polymer reactions, quantifying the mean number of branch sites per polymer. From the understanding that has been developed, one would predict that the extent of branching in P(BA-MMA) copolymers would increase as the fraction of BA in the monomer mixture was increased. Nevertheless, it remains very difficult to determine the number of long chain branches. With the addition of a chain transfer reagent to the reaction mixture, some control is possible over the number average molecular weight (M n ). Even with our current understanding of the factors that affect the formation 4-6 and properties 7, 8 of latex films, we have a little knowledge of the influence of chain branching on these properties. When grafting competes with polymerization, as in the reactions described above, even characterizing the polymer and distinguishing long-chain from short-chain branches becomes a challenge Long chain branches, particularly if they participate in entanglements, would be expected to have a significant influence on polymer diffusion in latex films and in the polymer rheology. Extensive branching should lead to structures resembling hyperbranched polymers, which have some properties similar to dendrimers. This chapter is based on the hypothesis that one can begin to learn about the importance of branching in latex polymers through the study of polymers in which there is greater control over the molecular weight and extent of branching. 47

65 Many methods have been developed for the synthesis of branched polymers. Gauthier used successive steps of anionic polymerization and chloromethylation to synthesize a series of hyperbranched polystyrenes D. M. Knauss developed a convergent living anionic polymerization method to produce polystyrene (PS) with dendritic branching. 18 A series of longbranched PS of various architectures were prepared by anionic polymerization using a multifunctional initiator. 19 Fréchet and his co-workers synthesized hyperbranched polymers using self-condensing vinyl polymerization (SCVP). 20, 21, 22 Others have used group transfer polymerization 23 and controlled radical polymerization (e.g., atom transfer radical polymerization (ATRP)) 24, 25 for the preparation of hyperbranched polymers. For example, Steven P. Armes prepared branched polymers via ATRP and reversible addition fragmentation chain transfer (RAFT) polymerization. 26, 27 Even though well-defined hyperbranched structures could be produced, none of these methods can be applied generally. Recently Sherrington s group reported a facile and broadly applicable method for producing branched polymers from vinyl monomers In this approach, they carry out conventional free radical polymerization of the vinyl monomer in the presence of both a crosslinking agent (to generate branches) and a chain transfer agent (to prevent gel formation). With the correct balance of these two reagents, soluble branched polymer can be obtained. For example, they polymerized methyl methacrylate (MMA) in the presence of but-2-ene-1,4- diacrylate (BDA, a bifunctional comonomer). To inhibit gelation, they added 1-dodecanethiol (C 12 -SH) as a chain transfer agent. They found that a high yield (77 97%) of soluble polymer was produced when relatively low levels (<2 mol %) of BDA were employed together with a similar or a higher level of C 12 -SH. 28 The effects of various branching agents on polymer architecture and properties were also studied. In this way, they synthesized a variety of soluble, branched copolymer architectures using polyacrylate branching agents containing between two and six acrylate functional groups. 31 More interestingly, the approach was applied to the synthesis of highly branched PMMA in aqueous emulsion polymerization. 28 Different ratios of cross-linker and chain transfer agent were added into a series of batch emulsion polymerization reactions. When proper amount of chain transfer agent was used, highly branched PMMAs were produced without cross-linking. The polymers obtained had relatively low molecular weights (M n < 3,300 g/mol) and broad 48

66 molecular weight distributions (M w /M n > 10). Despite these limitations, this new strategy provides a facile one-step method for making highly branched polymers confined to colloidal nanoparticles. I was interested in exploring Sherrington s emulsion polymerization approach. In this chapter I describe experiments in which I applied this method to the polymerization of n-butyl methacrylate (BMA). Instead of batch emulsion polymerization, where problems can arise because of unfavorable reactivity ratios or chain transfer constants, I used semi-continuous emulsion polymerization, which normally provides better control over molecular weight and molecular weight distribution. My primary goal in this first set of experiments was to see if semicontinuous emulsion polymerization extended the range of accessible molecular weights, and offered better control over molecular weight and its distribution. Then I was interested in how polymer properties (such as the glass transition temperature, and linear rheological properties) varied with the extent of branching. I found that the molecular weights of the branched PBMA that I obtained were dramatically higher than the PMMAs made in batch emulsion by the Sherrington group, while the molecular weight distributions were much narrower. 4.2 Experimental Section Latex Preparation Latex samples were synthesized by semi-continuous emulsion polymerization reactions. A typical recipe for the synthesis of branched PBMA is shown in Table 2-2. A monomer preemulsion was prepared by shaking a mixture of monomer, branching agent, surfactant, chain transfer agent, and water for 30 min. In the first stage, a dispersion of seed particles was prepared by batch emulsion polymerization. Water (3.0 g), Me-β-CD (0.02 g) and SDS (0.03 g) were added in a 100 ml 3-neck flask equipped with a mechanical stirrer, nitrogen inlet and condenser. The flask was immersed in an oil bath. The system was thoroughly purged with nitrogen while the reaction mixture was heated to 80 o C. After the reactor temperature stabilized at 80 o C, the KPS solution (0.06 g in water 0.5 g) as an initiator and the Na 2 CO 3 solution (0.05 g in water 0.5 g) as a ph buffer were added into the reactor followed by the addition of 3 wt % of the monomer pre-emulsion. The mixture was stirred for 20 min at 80 o C. In the second stage of 49

67 polymerization, the remaining monomer pre-emulsion was fed into the seed latex dispersion together with an initiator aqueous solution (0.01 g in water 2.0 g) in 4 h. The feeding rates were kept identical, controlled by Fluid Metering QG50 pumps. After the addition was completed, the system was maintained at 80 o C for 0.5 h. Then the reaction was cooled to room temperature Synthesis of High Molecular Weight Linear PBMA High molecular weight linear PBMA was synthesized by miniemulsion polymerization using SDS/hexadecane as surfactant/co-stabilizer and 2,2 -Azobis(2-methylpropionitrile) (AIBN) as initiator. A mixture of BMA, SDS, hexadecane, AIBN, and water was emulsified using a Branson Models250 digital sonifier (40% amplitude) at 0 o C for 30 min. The reaction was carried out at 80 C for 5 h. Water was evaporated and the polymer was dried in a vacuum oven for 24 h at 40 C. The dry polymer was fractionated using hexane/ethanol (v:v = 3:1). Then the high molecular weight fraction was collected. GPC and rheology measurement were carried out on this sample Characterization of Latex Polymers Differential Refractometer (dn/dc). The refractive index increment (dn/dc) was measured using a Brookhaven Instruments model BI-DNDC differential refractometer at 35 o C. For each polymer sample, five concentrations were used. n was plotted against concentration to give dn/dc. The obtained dn/dc values were listed in Table 4-1. Triple Detector Array Gel Permeation Chromatography (TDA/GPC). Polymer molecular weights and molecular weight distributions were measured by gel permeation chromatography (GPC) using a Viscotek liquid chromatograph equipped with a Viscotek model 2501 UV detector and a Viscotek TDA302 triple detector array (TDA). Two Viscotek GMHHR Mixed Bed columns were used with tetrahydrofuran (THF) as the elution solvent at a flow rate of 0.6 ml/min. The GPC column oven temperature was 35 o C and the injection volume was 0.1 ml. The absolute molecular weight was calculated using the appropriate dn/dc value. All molecular weight data are listed in Table

68 51

69 4.3 Results Synthesis of Branched PBMAs There were three objectives in this work. First, I wanted to prepare high molecular weight branched polymers, which here refers to a number-average molecular weight (M n ) of at least several tens of thousands g/mol. Second, I wished to control the extent of branching in these high M n branched polymers, which means the extent of branching in the final product could be tuned by adjusting the recipe for latex synthesis. Finally, I hoped to achieve the above two objectives in a straight-forward one-pot reaction. With the three goals in mind, I carried out the synthesis of n-butyl methacrylate (BMA) copolymers by semi-continuous emulsion polymerization using bisphenol A dimethacrylate (BPDM) as the branching agent and 1-dodecanethiol (C 12 -SH) as the chain transfer agent. A typical recipe is presented in Table 2-2. I chose BPDM as the branching agent. While the reactivity ratios for the reaction with BMA are not known, and these will be influenced by the relative solubility of the two monomers in the aqueous medium, I assume that I can overcome any problems with the reactivity ratio differences by running the reactions under monomerstarved conditions. The aromatic rings of BPDM provide UV and NMR signatures for analyzing its incorporation into the polymer. C 12 -SH is an efficient chain transfer agent to suppress gel formation in semi-continuous emulsion polymerizations of acrylate and methacrylate monomers. 34 To ensure good transport of these reactants, particularly C 12 -SH, through the aqueous phase during the synthesis, methyl-β-cyclodextrin (Me-β-CD) was included in the reaction mixture. 35 While many semi-continuous emulsion polymerization reactions are carried out in the presence of pre-formed seed particles, I wished to avoid the contribution of even small amounts of high molecular weight linear polymer typical of PBMA seed particles. Therefore, I generated seed particles in situ using a small fraction (typically 3 wt %) of the total feed of monomer pre-emulsion in a batch reaction. The pre-emulsion contained the monomer mixture, water and additional surfactant. The remaining pre-emulsion was then fed continuously into the reaction over 5 h. All of the potassium persulfate (KPS) initiator was added in the first stage. 36 Samples of latex produced from the above process were dried in an oven overnight. The dry polymers were then used for molecular weight measurements using a TDA/GPC. Values of 52

70 dn/dc were determined independently using a differential refractometer. As shown in Table 4-1, the dn/dc values increase as the fraction of BPDM in the polymer increases. Without BPDM, linear-pbma1 has a dn/dc value of The dn/dc value becomes for branched-pbma2 which contains 1 mol % of BPDM (relative to BMA). However, there is only a small increase of dn/dc when the BPDM fraction was increased to 10 mol % (for branched-pbma7, dn/dc = 0.084). The calculated molecular weights are presented in Table 4-1. For most samples shown in Table 4-1, M n is greater than 40,000 g/mol, and M w is greater than 100,000 g/mol. Among all the branched polymers, there is one sample with a much larger M n than all the others (for branched- PBMA1, M n = 160,000 g/mol). Thus the first goal, synthesis of high molecular weight branched polymers, was achieved. Moreover M w /M n is less than 2.8 for most cases, which indicates that the molecular weight distribution is relatively narrow. Another feature can also be observed from the molecular weight data, which is that the molecular weights are comparable for the linear PBMA1 sample and the branched-pbma2 8 samples. This feature demonstrates the feasibility of controlling the molecular weight of branched polymers by tuning the ratio of branching agent and chain transfer agent. In addition to obtaining high molecular weight branched polymers, controlling the extent of branching and preventing gel formation at the same time was also a challenge. Figure H NMR spectra of PBMAs with different branching levels. CD 2 Cl 2 was used as solvent. Peaks a and peaks b correspond to protons of BPDM and BMA respectively. 53

71 Branching control was realized by adjusting the feed ratio of BMA/BPDM, while gel formation was avoided by feeding appropriate amount of C 12 -SH. As presented in Table 4-1, four molar feed ratios of BMA/BPDM were investigated, in which the BPDM content ranged from 0 mol % to 10 mol % (relative to BMA). After polymerization, the BPDM content in each polymer was determined from the 1 H NMR spectrum (Figure 4-1) and the results are listed in Table 4-1. One can observe that all experimental ratios are similar to the corresponding feed ratios, even for samples with the highest BPDM concentration (branched-pbma6 8). Thus the extent of branching could be varied by modifying the feed ratio of BMA/BPDM. No gelation was detected in any of the latex. Gelation could be suppressed when an appropriate amount of C 12 -SH was fed together with the branching agent Architectures of Branched PBMAs I deduced the architectures of the branched PBMA chains from a combination of TDA/GPC traces and 1 H NMR spectra. From the TDA/GPC data and dn/dc values, I calculated M n values and polydispersities of the samples. A GPC trace for branched-pbma2 is shown in Figure 4-2. Here one sees that the RI trace and UV trace (due to BPDM groups) overlap very well. This result suggests that the BPDM groups are relatively uniformly distributed over the polymer chains of different molecular weight. The absence of a UV peak at low mass indicates that all of the BPDM monomer was incorporated into the polymer. I used the M n values and the BMA/BPDM ratios to calculate the average number of BMA units (N BMA ) and BPDM units (N BPDM ) per polymer chain. The BPDM units divide the polymer chain into a number of parts. The network functionality (f ) of BPDM is 4; thus (f -1)=3 parts are added to the chain for each BPDM unit. The average number of parts per chain (X) was calculated with the equation X = N BPDM (4-1) and the average number of BMA units per part (n BMA ) is given by n = N X (4-2) BMA BMA / The calculated results of four representative samples containing different BMA/BPDM ratios are shown in Table 4-2. All samples have a similar average number of BMA units per chain, whereas X values increase significantly with N BPDM. As a consequence, n BMA drops as X 54

72 increases. For example, the most highly branched sample branched-pbma7 contains an average of only 3 BMA units between branch points. In contrast, linear-pbma1 has no branch points. The information in Table 4-2 was used to generate the idealized graphical depiction of the chain architectures shown in Figure 4-3. As a linear chain, linear-pbma1 adopts a random coil structure (Figure 4-3A). Containing only four branch points per chain, branched-pbma2 maintains a loose random-coil-like as depicted in Figure 4-3B. This description of the polymer shape in solution is consistent with the hydrodynamic R h values acquired by dynamic light scattering (see Table 4-1). For the more branched PBMAs of similar molecular weight, the chain dimensions in solution become more compact. Branched-PBMA7, the most highly branched polymer chain, is divided by an average of 34 branch points per chain into more than 100 parts, which makes the chain in solution act like a dense sphere (R h = 2.1 nm, see Table 4-1 and Figure 4-3D). The decreasing R h with increasing degree of branching was also consistent the GPC elution sequence which can be observed from the GPC traces and a plot of log M n versus retention volume of the four PBMAs presented in Figure 4-4. Figure 4-2. UV and RI traces in the GPC analysis of branched-pbma2 55

73 Figure 4-3. Polymer architecture for (A) linear-pbma1, (B) branched-pbma2, (C) branched-pbma5 and (D) branched-pbma7. These drawings assume a uniform distribution of branch points in the polymer molecules. Figure 4-4. GPC traces (A) RI signal and (B) log M n vs retention volume for linear-pbma1 (L1), branched-pbma2 (B2), branched-pbma5 (B5) and branched-pbma7 (B7). The vertical line in (B) indicates that the polymer with Mn = 34,000 has a retention volume of 17.3 ml. The vertical line in (A) indicates that ca. 30% of the L1 sample has M n lowerthan