Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers

Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Dissertation presented at Faculdade de Engenharia da Universidade do Porto to obtain the degree of Doctor in Chemical and Biological Engineering by Miguel Ângelo Diz Gonçalves Supervisor: Prof. Mário Rui P.F.N. Costa Co-supervisor: Prof. Rolando C.P.S. Dias Laboratory of Separation and Reaction Engineering, Associate Laboratory LSRE/LCM Department of Chemical Engineering, Faculty of Engineering, University of Porto April 2014 Agradecimentos Acknowledgements Aos meus orientadores, Professor Rolando Dias e Professor Mário Rui Costa, pela oportunidade concedida, acompanhamento científico, dedicação, sabedoria e amizade demonstrada ao longo deste trabalho. Ao LSRE (Laboratory of Separation Process and Engineering), nas pessoas do Professor Alírio Rodrigues e Professora Madalena Dias e à direcção da Escola Superior de Tecnologia e Gestão do Instituto Politécnico de Bragança, pelas condições técnicas disponibilizadas à realização deste trabalho. Aos meus Pais, João e Irene, pelo Amor, força e encorajamento e principalmente por estarem sempre por mim. Agradeço à minha irmã Sandra, aos meus amigos e familiares, que estiveram sempre ao meu lado. Aos meus colegas do LSRE pela amizade e espírito de camaradagem dados ao longo destes anos. Um agradecimento especial à Isabel Patrícia, ao Patrick, ao Marco, à Virgínia e à Patrícia. À Maria João e à Paula, pela ajuda e apoio dados através dos laboratórios de química (LPQ e LQA) da Escola Superior de Tecnologia e Gestão, e pela amizade demonstrada. À Fundação para a Ciência e Tecnologia (FCT) e ao FSE (Programa Operacional Potencial e Humano/POPH), pelo apoio financeiro através da bolsa de doutoramento SFRH/BD/76587/2011. FCT, Ministry of Science and Technology of Portugal (Program COMPETE-QCAIII) and European Community through FEDER are gratefully acknowledged for supporting this research through the projects POCI/EQU/44784/2002, POCI-PPCDT/EQU/60483/2004 and POCI/EQU-EQU/098150/2008. The financial support of BASF AG through the projects FEUP/IPB/BASF/2007 and FEUP/IPB/BASF/2008 is gratefully acknowledged. EU is also gratefully acknowledged for supporting this research through the Marie Curie Initial Training Network “Nanopoly” ITN-GE-2009-238700. À minha mulher Anabela Ao meu filho André Abstract Many materials used in coating and adhesive industries (paints, glues), sanitary industries and agriculture (superabsorbent polymers), biomedicine and pharmaceutics (smart hydrogels for drug delivery) or separation processes in different industries or environment (such as ion exchange or chelating resins) are produced through the polymerisation of multivinyl monomers. The performance of these materials is dependent on their production process and therefore it is important to find the link between their synthesis conditions and properties. The structure of the networks has a strong influence on their properties. The materials obtained by conventional radical polymerisation (FRP) lead to networks with high amounts of relatively short loops which contribute little to the elasticity modulus. The aim of this work was to carry out a study on the synthesis and characterization of hyperbranched polymers, gels and hydrogels arising from the radical polymerisation of multivinyl monomers to provide data for testing current knowledge on chemical and process modelling. Here the synthesis and characterization of materials obtained by free radical polymerisation (FRP) and controlled radical polymerisation techniques (CRP), namely ATRP, NMRP and RAFT, using organic (styrenics, acrylates and methacrylates) and water soluble monomers (acrylic and methacrylic acids copolymers, acrylamides) in different reaction systems (solution/bulk, suspension and inverse-suspension) using semi-batch and batch reactors in atmospheric and pressurized operation is reported. Temperature, monomers concentration, crosslinker concentration and functionalization, initiation system and concentration/kind of polymerisation controlling agents were the parameters tested in these polymerisations. A secondary goal of this work was to use these experimental results to test simulation tools for predicting molecular masses and radius of gyration before and after gelation in the absence/presence of intramolecular cyclizations. Molecular architecture of the polymerisation products was experimentally measured using size exclusion chromatography with multiple detectors, namely refractive index, multi-angle laser light scattering, intrinsic viscosity and ultraviolet extinction. In-line and off-line FTIRATR measurements were also performed in order to obtain information concerning the kinetics of formation and final structures of gels. In batch mode, FTIR yielded information on the presence and amounts of pendant double bonds. Gel fractions and swelling ratios were quantified and the performance of the synthesized water compatible polymer networks (hydrogels) was studied through drug delivery tests. The morphology of the gel beads was observed by scanning electron microscopy. The concentration of pendant double bonds was also measured by the iodine chloride addition method. It was observed that the used polymerisation mechanisms have a huge effect on the network structure and this feature can be exploited to design tailored materials. Hyperbranched polymers with different architectures result from stopping molecular growth before gel point and were obtained by changing the feed policy of the reactants, using semibatch reactors and chain transfer agents. Controlled radical polymerisation techniques (ATRP, NMRP and RAFT) were used along the experimental program. These polymerisation mechanisms allowed an improvement on the final properties of the polymers and networks, by decreasing the relative rate of intramolecular cyclizations. Water compatible monomers were used to produce superabsorbent polymers and smart hydrogels. Such networks can be used in controlled drug delivery by exploiting the change of active substance retention with pH, temperature or other environmental variables. Resumo Muitos materiais usados nas indústrias de revestimentos e adesivos (tintas, colas), indústrias sanitárias e agricultura (polímeros superabsorventes), biomedicina e farmácia (hidrogéis inteligentes para libertação controlada de fármacos) ou em processos de separação industriais e ambientais (tais como resinas de permuta iónica ou resinas quelatantes) são produzidos através da polimerização de monómeros multivinílicos. O desempenho e a utilidade destes materiais dependem do processo de produção e portanto é importante encontrar a relação entre as condições de síntese e as suas propriedades. A estrutura das redes poliméricas tem uma forte influência nas suas propriedades. Os materiais obtidos por polimerização radicalar convencional (FRP) são em geral redes de polímero contendo grandes quantidades de pequenos ciclos que contribuem pouco para a sua elasticidade. O objectivo deste trabalho consistiu na realização de estudos de síntese e caracterização de polímeros híper-ramificados, géis e hidrogéis, obtidos através de polimerizações radicalares de monómeros multivinílicos, de forma a obter dados experimentais úteis para a avaliação do conhecimento actual sobre a química e modelização desses processos. É aqui relatada a síntese e a caracterização de materiais obtidos por polimerização radicalar “livre” (FRP) e técnicas de polimerização radicalar controladas (CRP), nomeadamente ATRP, NMRP e RAFT, usando monómeros orgânicos (estirenicos, acrilatos e metacrilatos) e solúveis em água (acrílicos, acrilamidas). Foram também experimentados diversos sistemas de reacção (solução/massa, suspensão e suspensão-inversa) e considerados diferentes reactores fechados ou semi-fechados, operando à pressão atmosférica ou sob pressurização. Nestas polimerizações, a temperatura, concentração de monómeros, concentração e funcionalidade dos reticulantes, sistema de iniciação e tipo/concentração de agentes controladores da polimerização foram testados como parâmetros de operação. Um objectivo secundário deste trabalho foi a utilização destes resultados experimentais para teste de ferramentas de simulação úteis na previsão de massas moleculares e raios de giração antes e depois da gelificação e na presença (ou ausência) de ciclizações intramoleculares. A arquitectura molecular dos produtos foi medida experimentalmente usando cromatografia por exclusão de tamanhos com detecção múltipla, nomeadamente índice de refracção, dispersão de luz multi-angular, índice de viscosidade e absorção no ultravioleta. Medições in- line e off-line por FTIR-ATR foram também realizadas com o propósito de obter informação acerca da cinética de formação e das estruturas finais dos géis. Em modo fechado, a análise FTIR revelou informação sobre a presença e quantidades de ligações duplas pendentes. As fracções de gel e as razões de inchamento foram quantificadas e o desempenho das redes poliméricas compatíveis com água (hidrogéis) foram estudados através de testes de libertação controlada. A morfologia dos produtos foi observada por microscopia electrónica de varrimento e a concentração de ligações duplas pendentes foi também medida pelo método da adição de cloreto de iodo. Foi observado que os mecanismos de polimerização usados têm um enorme efeito na estrutura das redes e este aspecto pode ser explorado para produzir materiais com propriedades por medida. Polímeros hiper-ramificados com diferentes arquitecturas, resultando da paragem do crescimento molecular antes do ponto de gel, foram também obtidos mudando o modo da alimentação dos reagentes, usando reactores semi-fechados ou considerando agentes de transferência de cadeia na composição inicial do sistema químico. Diferentes técnicas de polimerização radicalar controladas (nomeadamente ATRP, NMRP e RAFT) foram usadas ao longo do programa experimental. Estes mecanismos de polimerização permitiram o melhoramento das propriedades finais dos polímeros e das correspondentes redes, através da diminuição da velocidade relativa dos mecanismos cinéticos de ciclização intramolecular. Monómeros compatíveis com a água foram usados na produção polímeros superabsorventes a hidrogéis inteligentes (sensíveis a estímulos resultantes de variações nas condições envolventes). Foi demonstrado que estas redes podem ser usadas na libertação controlada de fármacos, explorando mudanças na retenção da substância activa em função de alterações no pH, temperatura ou em outras variáveis das soluções aquosas circundantes (ex. força iónica). vi Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7.2.2 Polymerisation Set-up 7-5 7.2.3 Product Analysis by SEC with a Tetra Detector Array 7-7 7.2.4 In-line FTIR-ATR Measurements 7-8 7.2.5 Swelling Ratio Sensitivity Measurements 7-9 7.2.6. Drug Release Testing 7-12 7.3 Results and Discussion 7-17 7.4 Conclusions 7-25 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK 8-1 REFERENCES 9-1 LIST OF PUBLICATIONS A APPENDICES C Appendix A D vii Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers List of Figures INTRODUCTION 1-1 Figure 1.1. Schematic representation of (a) linear polymer chain (b) branched polymer chain and (c) crosslinked polymer chain. 1-4 Figure 1.2. Depiction of main polymer equilibration step in a RAFT polymerisation. 1-9 Figure 1.3. Schematic of a general primary cyclization reaction. 1-11 Figure 1.4. Schematic of a general secondary cyclization reaction. 1-11 CONVENTIONAL FREE-RADICAL COPOLYMERISATION OF MULTIVINYL MONOMERS 2-1 Figure 2.1. Experimental set-up used in the reaction polymerisations. 2-5 Figure 2.2. Schematic representations of (a) styrene (b) m-divinylbenzene and (c) pdivinylbenzene. 2-10 Figure 2.3. Schematic representations of (a) p-pendant double bond and (b) m-pendant double bond. 2-12 Figure 2.4. Schematic representation of (a) toluene (b) AIBN and (c) carbon tetrabromide. 2-12 Figure 2.5. Schematic representations of primary radicals from (a) initiator (b) solvent and (c) chain transfer agent. 2-12 Figure 2.6. Schematic representation of the radical from styrene. 2-12 Figure 2.7. Schematic representations of radical from (a) p-divinylbenzene and (b) radical from m-divinylbenzene. 2-12 Figure 2.8. Schematic representation of the radicals arising from (a) m-PDB and (b) pPDB. 2-13 Figure 2.9. AIBN decomposition leading to primary radicals. 2-14 Figure 2.10. Initiation reaction of the styrene monomer by primary radicals. 2-15 Figure 2.11. Initiation reaction of the m-divinylbenzene monomer by primary radicals. 2-15 Figure 2.12. Initiation reaction of the p-divinylbenzene monomer by primary radicals. 2-15 Figure 2.13. Initiation reaction of the m-divinylbenzene pendant double bond by primary radicals. 2-16 List of Figures viii Figure 2.14. Initiation reaction of the p-divinylbenzene pendant double bond by primary radicals. 2-16 Figure 2.15. Propagation of styrene monomer with a radical formed by styrene monomer. 2-16 Figure 2.16. Propagation of m-DVB monomer with a radical formed by styrene monomer. 2-17 Figure 2.17. Propagation of p-DVB monomer with a radical formed by styrene monomer. 2-17 Figure 2.18. Propagation of m-DVB pendant double bond with a radical formed by styrene monomer. 2-17 Figure 2.19. Propagation of p-DVB pendant double bond with a radical formed by styrene monomer. 2-18 Figure 2.20. Propagation of styrene monomer with a radical formed by m-DVB monomer. 2-18 Figure 2.21. Propagation of styrene monomer with a radical formed by p-DVB monomer. 2-18 Figure 2.22. Propagation of m-DVB monomer with a radical formed by m-DVB monomer. 2-19 Figure 2.23. Propagation of p-DVB monomer with a radical formed by m-DVB monomer. 2-19 Figure 2.24. Propagation of m-DVB with a radical formed by p-DVB monomer. 2-19 Figure 2.25. Propagation of p-DVB with a radical formed by p-DVB monomer. 2-19 Figure 2.26. Propagation of m-DVB pendant double bond with a radical formed by m-DVB monomer. 2-20 Figure 2.27. Propagation of p-DVB pendant double bond with a radical formed by m-DVB monomer. 2-20 Figure 2.28. Propagation of m-DVB pendant double bond with a radical formed by p-DVB monomer. 2-20 Figure 2.29. Propagation of p-DVB pendant double bond with a radical formed by p-DVB monomer. 2-21 Figure 2.30. Propagation of styrene monomer with a radical formed by m-DVB pendant double bond. 2-21 Figure 2.31. Propagation of styrene monomer with a radical formed by p-DVB pendant double bond. 2-21 Figure 2.32. Propagation of m-DVB monomer with a radical formed by m-DVB pendant double bond. 2-22 Figure 2.33. Propagation of p-DVB monomer with a radical formed by m-DVB pendant double bond. 2-22 ix Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 2.34. Propagation of m-DVB monomer with a radical formed by p-DVB pendant double bond. 2-22 Figure 2.35. Propagation of p-DVB monomer with a radical formed by p-DVB pendant double bond. 2-23 Figure 2.36. Propagation of m-DVB pendant double bond with a radical formed by m-DVB pendant double bond. 2-23 Figure 2.37. Propagation of p-DVB pendant double bond with a radical formed by m-DVB pendant double bond. 2-23 Figure 2.38. Propagation of m-DVB pendant double bond with a radical formed by p-DVB pendant double bond. 2-23 Figure 2.39. Propagation of p-DVB pendant double bond with a radical formed by p-DVB pendant double bond. 2-24 Figure 2.40. Termination by combination of two radicals derived from styrene monomer. 2-24 Figure 2.41. Termination by combination of two radicals derived from p-divinylbenzene monomer. 2-24 Figure 2.42. Termination by combination of two radicals derived from m-divinylbenzene pendant double bonds. 2-25 Figure 2.43. Transfer to solvent reaction between a radical from styrene and toluene. 2-26 Figure 2.44. Transfer to chain transfer agent reaction between a radical from styrene and CBr4. 2-26 Figure 2.45. Schematic representation of a primary intramolecular cyclization reaction in the copolymerisation of styrene and divinylbenzene. 2-26 Figure 2.46. Schematic representation of a secondary intramolecular cyclization reaction. 2-26 Figure 2.47. Time evolution of the measured overall conversion for linear and non-linear polymerisation systems considering different operation conditions. 2-30 Figure 2.48. The influence of the reactivity of the pendant double bonds on the predicted (run 2) and its comparison with the correspondent measured values. 2-30 Figure 2.49. Predicted and measured for linear and non-linear polymerisation systems in batch and semi-batch reactor. 2-31 Figure 2.50. Time evolution of , and during the radical copolymerisation of STY/DVB in semi-batch reactor using different feed policies. 2-31 Figure 2.51. Time evolution of in the presence of CTA or using different feed policies in semi-batch reactor. 2-33 List of Figures x Figure 2.52. Time evolution of in THF solution during the radical copolymerisation of STY/DVB in batch or semi-batch reactor. 2-33 Figure 2.53. Measured refractive index signal in the SEC chromatograms of samples of STY/DVB corresponding to different polymerisation times in a semi-batch reactor. 2-34 Figure 2.54. Measured 90° light scattering signal in the SEC chromatograms of samples of STY/DVB corresponding to different polymerisation times in a semi-batch reactor. 2-34 Figure 2.55. Molecular weight along the SEC chromatogram for samples of STY/DVB synthesized using different operation conditions in polymerisation times very near to gel point. 2-35 Figure 2.56. Molecular weight along the SEC chromatogram for samples of STY/DVB synthesized using different operation conditions in polymerisations times very near to gel point. 2-35 Figure 2.57. Observed relations molecular weight vs. elution volume for STY/DVB. 2-36 Figure 2.58. Time evolution of the molecular weight distribution of STY/DVB synthesized in semi-batch reactor showing a formation of a long end tail as gelation is approached. 2-36 Figure 2.59. Schematic representation of (a) methyl methacrylate and (b) ethylene glycol dimethacrylate. 2-37 Figure 2.60. Schematic representation of benzoyl peroxide (BPO) thermal initiator. 2-37 Figure 2.61. Transfer to monomer reaction between a radical from MMA and MMA. 2-39 Figure 2.62. Transfer to initiator reaction of a radical from MMA with BPO. 2-39 Figure 2.63. Time evolution of the measured and predicted overall monomer conversion for different polymerisation runs performed with AIBN and BPO as initiators. 2-43 Figure 2.64. Time evolution of the measured and predicted for polymerisations runs performed with AIBN. 2-44 Figure 2.65. Time evolution of the measured and predicted for polymerisations runs performed with BPO. 2-44 Figure 2.66. Predicted and measured time evolution of in the copolymerisation system MMA/EGDMA initiated by AIBN. 2-46 Figure 2.67. Predicted and measured time evolution of in the copolymerisation system MMA/EGDMA initiated by BPO. 2-46 xi Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 2.68. The predicted effect of the reactivity of radicals of pendant double bonds on the time evolution of . 2-48 Figure 2.69. Predicted and measured gel conversion for the system MMA/EGDMA in the presence of CTA. Experimental data collect from Li et al. (1989a and 1989b) and predictions from the present kinetic model. 2-48 Figure 2.70. Relation between average radius of gyration and average molecular weight for linear MMA and several non-linear samples of MMA/EGDMA. 2-49 Figure 2.71. Observed 90° light scattering signal in the SEC chromatograms of samples of MMA/EGDMA corresponding to different polymerisation times. 2-49 Figure 2.72. Molecular weight along the SEC chromatogram for a sample of MMA/EGDMA. 2-50 Figure 2.73. Observed relations molecular weight vs. elution volume for MMA/EGDMA copolymers. 2-50 Figure 2.74. Time evolution of the molecular weight distribution of MMA/EGDMA copolymerisation showing the formation of a long end tail as gelation is approached. 2-51 Figure 2.75. Predicted and measured time evolution of the weight average molecular weight . The effect of the mole fraction of DVB in the initial monomer mixture over the dynamics of in batch reactor is observed. 2-53 Figure 2.76. Predicted and measured time evolution of the weight fraction of gel . The effect of the mole fraction of DVB in the initial monomer mixture over the dynamics of is observed. 2-53 Figure 2.77. (a) SEM of STY/DVB gel beds obtained in run 3 with magnification of 250 . (b) Gel beads for run 6 with magnification of 250 . (c) Run 1 with magnification of 50000 . (d) Run 5 with magnification of 50000 . (e) Run 2 with magnification of 2500 . (f) Run 3 with magnification of 20000 . 2-55 Figure 2.78. Measured swelling ratio for STY/DVB gel beads synthesized using different proportions of n-heptane/toluene. 2-56 Figure 2.79. FTIR-ATR spectra observed in the in-line monitoring of the suspension copolymerisation of STY/DVB using n-heptane and toluene as diluents of the organic phase. 2-57 List of Figures xii ATOM TRANSFER RADICAL POLYMERISATION OF ACRYLATES Figure 3.1. Schematic of an ATRP equilibrium reaction of a polymer radical from methyl methacrylate. 3-3 Figure 3.2. Depiction of a chain transfer reaction involving an acrylate monomer (n-butyl acrylate used for illustration) with creation of a terminal double bond that can lead to long chain branching. 3-20 Figure 3.3. Depiction of intermolecular chain transfer to polymer in acrylates due to the Hatom abstraction of a methine hydrogen. 3-21 Figure 3.4. Depiction of intermolecular chain transfer to polymer in acrylates due to the Hatom abstraction of a –CH3 group. 3-21 Figure 3.5. Depiction of pendant double bonds propagation (crosslinking) in acrylate/diacrylate copolymerisation. 3-21 Figure 3.6. Observed refractive index signal in the SEC traces of ATRP synthesised acrylate/diacrylate copolymers corresponding to different polymerisation times (monomer conversion). 3-24 Figure 3.7. Measured refractive index and light scattering (90°) signals in the SEC trace of an ATRP synthesized acrylate/diacrylate copolymer. 3-25 Figure 3.8. Measured relation of molecular weight versus elution volume in the SEC trace of a FRP synthesized PnBuA sample. 3-25 Figure 3.9. Measured relation of root mean square radius of gyration versus elution volume for the same sample of Figure 3.8. 3-26 Figure 3.10. Scaling law for the radius of gyration versus molecular weight observed for PnBuA in a good solvent and its comparison with scaling laws measured for the dimensions of other polymers. 3-26 Figure 3.11. Estimated unperturbed dimensions of PnBuA considering Fixman extrapolation. 3-28 Figure 3.12. Estimated Θ dimensions of PnBuA with Stockmayer-Fixman extrapolation. 3-28 Figure 3.13. Estimated Θ dimensions of PnBuA with Kurata-Stockmayer-Roig extrapolation. 3-29 Figure 3.14. Off-line FTIR-ATR spectra observed for nBuA, HDDA and BEDA monomers. 3-32 Figure 3.15. FTIR-ATR spectra observed in the in-line monitoring of the nBuA polymerisation (Run 6 of Table 3.2). 3-32 Figure 3.16. FTIR-ATR spectra observed in the in-line monitoring of nBuA/HDDA copolymerisation (Run 7 in Table 3.2). 3-33 Figure 3.17. Polymer formation (Run 6 in Table 3.2) detected by FTIR-ATR monitoring of the region with wave numbers in the range of 3400 to 4000 cm-1 (aliphatic C-H bonds). 3-34 xiii Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 3.18. Similar analysis as in Figure 17 for the copolymerisation of nBuA/HDDA (Run 7 in Table 3.2). 3-34 Figure 3.19. FTIR-ATR in-line monitoring of MA homopolymerisation (run 1 in Table 3.2) using the reference background of the initial mixture. 3-35 Figure 3.20. Measured and predicted evolution of in nBuA/HDDA FRP copolymerisation with different initial amounts of diacrylate ( ). 3-36 Figure 3.21. Effect of the dilution in the evolution of for FRP copolymerisation of nBuA/HDDA. 3-38 Figure 3.22. Predicted effect of the reactivity of PDB on for FRP of nBuA/HDDA (run 7 in Table 3.2. 3-38 Figure 3.23. Effect of the initial amount of crosslinker in the evolution of for ATRP copolymerisation of nBuA/HDDA. 3-49 Figure 3.24. Predicted effect of the reactivity of pendant double bonds on for ATRP of nBuA/HDDA (run 4 in Table 3.1). 3-40 Figure 3.25. Effect of the dilution in ATRP copolymerisation of nBuA/BEDA. 3-40 Figure 3.26. Comparison of FRP and ATRP copolymerisations of nBuA/HDDA with the same initial dilution and amount of crosslinker. 3-41 Figure 3.27. Predicted effect of the reactivity of the pendant double bonds ( on the time evolution of in the ATRP of nBuA/HDDA copolymerisation (run 3 in Table 3.1). 3-42 Figure 3.28. Predicted and observed evolution of during the ATRP of nBuA/diacrylate considering different initial mole of crosslinker (runs 5 and 6 in Table 3.1). 3-42 Figure 3.29. Predicted effect of the reactivity of PDBs on the dynamics of the (Θ state and normalized by the bond length, ) of ATRP synthesized polyacrylates. Operating conditions: , rM/CuBr/PMDETA=50/0.45/0.5. Kinetic parameters: . 3-43 Figure 3.30. Predicted effect of the initial mole fraction of crosslinker on the dynamics of the of ATRP synthesized polyacrylates (same conditions of Figure 3.41). 3-44 Figure 3.31. Measured relation versus time of polymerisation in different ATRP experiments of nBuA. 3-44 Figure 3.32. Observed relation of versus monomer conversion in linear (nBuA) and non-linear (nBuA/HDDA) experiments considering different initial mole ratios monomer/initiator ( ). 3-45 Figure 3.33. Effect of the initial ratio monomer/initiator in ATRP copolymerisation of nBuA/BEDA. 3.46 List of Figures xiv Figure 3.34. Measured and predicted evolution of in FRP homopolymerisation on nBuA and nBuA/HDDA copolymerisation with different amounts of diacrylate. 3-46 Figure 3.35. Measured and predicted evolution of and in nBuA/BEDA ATRP copolymerisation (run 9 in Table 3.1). 3-47 Figure 3.36. Typical SEC traces observed in the characterization by SEC/RI/MALLS of ATRP synthesized polymer samples (only RI signal). 3-48 Figure 3.37. Predicted influence of intermolecular chain transfer to polymer on for FRP of nBuA (run 6 in Table 3.2). 3-49 Figure 3.38. Predicted influence of terminal double bonds on for FRP on nBuA at different dilutions (runs 1 and 5 in Table 3.2). 3-49 Figure 3.39. Predicted influence of long chain branching for the ATRP of nBuA (run 2 in Table 3.1). 3-50 Figure 3.40. Relation versus for FRP (run 7 in Table 3.2) and ATRP (run 9 in Table 3.1) hyperbranched polyacrylates. 3-51 Figure 3.41. Predict effect of the reactivity of radicals from transfer to monomer on of FRP synthesized polyacrylates (operating conditions: . Kinetic parameters: , , , ). 3-52 Figure 3.42. Predicted effect of the initial amount of crosslinker on the dynamics of crosslinking and branching densities of ATRP synthesized polyacrylates (operating conditions: T = 60 °C, , rM/CuBr/PMDETA=50/0.45/0.5. kinetic parameters: , , = 1, ). 3-53 Figure 3.43. Measured refractive index and light scattering (90°) signals in the SEC trace of ATRP synthesized acrylate/diacrylate sample. Change of absolute molecular weight along the elution volume is also shown. 3-54 Figure 3.44. Examples of Debye plots along the SEC trace of a branched polyacrylate sample. Three different elution volumes were considered and the measured signals with the eight MALLS detectors are showed. 3-54 Figure 3.45. Typical differential weight fraction of the molar mass distributions measured for branched polyacrylate samples. Time evolution of the crosslink process can be observed. 3-55 Figure 3.46. Typical differential weight fraction of the radius of gyration distributions measured for branched polyacrylate samples. Change of the distribution with monomer conversion is depicted. 3-55 Figure 3.47. Observed SEC/RI/MALLS traces (only RI signal) of ATRP MMA/EGDMA (run 5 in Table 3.3) polymer samples collected at different polymerisation times. 3-58 xv Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 3.48. Observed SEC/RI/MALLS traces (only LS signal) of ATRP MMA/EGDMA (run 5 in Table 3.3) polymer samples collected at different polymerisation times. 3-58 Figure 3.49. Molecular weight along the SEC trace for an ATRP synthesized sample of MMA/EGDMA with a polymerisation time close to gel point. 3-59 Figure 3.50. Predicted and observed evolution of during the ATRP of MMA/EGDMA considering different initial mole fraction of crosslinker (see runs in Table 3.3). 3-60 Figure 3.51 Predicted and observed evolution of during the ATRP of MMA/EGDMA considering different initial mole fraction of crosslinker (see runs in Table 3.3). 3-61 Figure 3.52. Predictions and experimental observations for monomer conversion in the ATRP of MMA/EGDMA at different temperatures. Initial mole ratio MBPA/CuBr/HMTETA=1/1/1. 3-61 Figure 3.53. Predicted and experimentally observed influence of the temperature on in ATRP copolymerisations of MMA/EGDMA with constant initial mole fraction of crosslinker. 3-62 Figure 3.54. Predicted and observed time evolution of , , and in the ATRP of MMA/EGDMA. 3-63 NITROXIDE MEDIATED RADICAL POLYMERISATION OF STYRENE/DIVINYLBENZENE Figure 4.1. Mechanism of activation/deactivation involving a polymer radical formed by styrene and the mediator nitroxide (TEMPO). 4-10 Figure 4.2. Schematic representation of the Mayo dimerization. 4-10 Figure 4.3. Mechanism showing the reaction of auto-initiation of styrene. 4-10 Figure 4.4. Predictions and experimental observations for monomer conversion in conventional and NMRP runs. 4-16 Figure 4.5. Predictions and experimental observations of and in FRP and NMRP linear runs. Simulations include the thermal initiation of styrene during the heating period. 4-17 Figure 4.6. Figure 4.6. Measured and predicted time evolution of in FRP and NMRP STY/DVB runs. 4-18 Figure 4.7. Measured and predicted time evolution of in FRP and NMRP STY/DVB runs with different amounts of crosslinker. 4-19 Figure 4.8. Measured and predicted time evolution of in FRP and NMRP STY/DVB runs with different amounts of crosslinker. 4-19 List of Figures xxii Figure 5.12. Main results for RAFT polystyrene synthesis in aqueous suspension at 150 °C using DDMAT CTA (run 9 in Table 5.1). (a) Measured reaction time evolution of monomer conversion. (b) Measured reaction time evolution of molecular weight dispersity (DM) and average molecular weights ( and ). (c) Normalized RI signal of polystyrene samples with different polymerisation time. (d) Normalized MALLS signal (90° detector) for the same samples described in (c). (e) Comparison of the normalized RI and MALLS signals of a polystyrene sample with polymerisation time t = 12 h. (f) Change of the RI and LS signals along the molecular weight for the same sample described in (e). 5-25 Figure 5.13. Effect of the polymerisation temperature (in the range 70 to 150 °C) on the observed RI signals of the final samples (t=12 hr). 5-26 Figure 5.14. Similar comparison described in Figure 5.13 considering the MALLS signal. 5-27 Figure 5.15. Measured dynamics of monomer conversion for RAFT styrene polymerisation in aqueous suspension at different temperatures and using DDMAT as CTA agent. 5-28 Figure 5.16. Measured dynamics of monomer conversion for RAFT styrene polymerisation in aqueous suspension at 130 °C and using DDMAT as CTA agent. 5-28 Figure 5.17. Comparison of the observed time evolution of monomer conversion for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 5-29 Figure 5.18. Comparison of the observed time evolution of for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 5-29 Figure 5.19. Comparison of the observed time evolution of for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 5-30 Figure 5.20. Comparison of the observed time evolution of DM for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 5-30 Figure 5.21. Measured dynamics of monomer conversion for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 5-31 Figure 5.22. Measured dynamics of for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 5-31 Figure 5.23. Observed RI SEC traces for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 5-33 xxiii Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 5.24. Observed MALLS SEC traces for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 5-33 Figure 5.25. Observed DM for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 5-34 Figure 5.26. Observed off-line FTIR spectra for styrene and divinylbenzene monomers and polystyrene (run 7 in Table 5.1) and poly(STY/DVB) with 5% DVB (run 1 in Table 5.2). 5-34 Figure 5.27. Observed off-line FTIR spectra for DVB networks correspondent to different polymerisation times for run 2 in Table 5.2. 5-35 Figure 5.28. Observed off-line FTIR spectra for DVB networks correspondent to different polymerisation times for run 3 in Table 5.2. 5-35 Figure 5.29. Observed off-line FTIR spectra for final samples correspondent to runs 1, 2 and 3 in Table 5.2 and for run 7 in Table 5.1. 5-36 Figure 5.30. FTIR-ATR spectra observed during the in-line monitoring for the aqueous suspension RAFT polymerisation of divinylbenzene at 70 °C with 100% DVB in the organic phase (run 2 in Table 5.2). 5-38 Figure 5.31. Estimated pendant double bond conversion using in-line FTIR-ATR monitoring and estimated monomer conversion using SEC for run 2 in Table 5.2. 5-38 Figure 5.32. FTIR-ATR spectra observed during the in-line monitoring for the aqueous suspension RAFT polymerisation of divinylbenzene at 70 °C with 50% DVB in the organic phase (run 3 in Table 5.2). 5-40 Figure 5.33. Estimated pendant double bond conversion using in-line FTIR-ATR monitoring and estimated monomer conversion using SEC for run 3 in Table 5.2. 5-40 Figure 5.34. Normalized intensity of the FTIR 1630 cm-1 peak correspondent to C=C bonds considering different internal references (842, 909, 1494 and 1600 cm-1). The values presented are correspondent to off-line FTIR analysis of samples collected at different reaction times for run 2 in Table 5.2. 5-41 Figure 5.35. Normalized intensity of the FTIR 1630 cm-1 peak correspondent to C=C bonds considering different internal references (842, 909, 1494 and 1600 cm-1). The values presented are correspondent to off-line FTIR analysis of samples collected at different reaction times for run 3 in Table 5.2. 5-41 Figure 5.36. Observed ratios between network and DVB considering the normalized FTIR 1630 cm-1 peak (using the 1494 cm-1 as internal reference). Measurements are correspondent to samples with different polymerisation times in runs 2 and 3 in Table 5.2. (see Eq. (5.26)). 5-42 List of Figures xxiv Figure 5.37. Dynamics of the PDB concentration (mol/g polymer) measured by ICL titration (run 1 in Table 5.2) and considering also the calibration between the normalized FTIR 1630 cm-1 peak intensity ( ) and PDB concentration. 5-43 Figure 5.38. Measured dynamics of , and in aqueous suspension RAFT copolymerisation of STY/DVB (run 1 in Table 5.2) 5-44 Figure 5.39. Comparison of the observed dynamics of in aqueous RAFT suspension polymerisation of styrene (run 7 in Table 5.1) and RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). 5-44 Figure 5.40. Measured dynamics of monomer conversion in aqueous suspension RAFT polymerisation of styrene (run in Table 5.1) and RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). 5-45 Figure 5.41. Measured dynamics of PDB concentration in RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). 5-46 Figure 5.42. Comparison of the measured dynamics of monomer conversion in aqueous suspension RAFT polymerisation of styrene (run 2 in table 5.1) and DVB (runs 2 and 3 in Table 5.2) at 70 °C. 5-46 Figure 5.43. Normalized RI signal of polystyrene samples (run 3 in Table 5.1) with different polymerisation times. 5-47 Figure 5.44. Comparison of the normalized RI and MALLS signals of a polystyrene sample (run 3 in Table 5.1) with polymerisation time t= 12 h. 5-48 Figure 5.45. Measured dynamics of for RAFT styrene polymerisation (run 2 in Table 5.1) and RAFT DVB (runs 2 and 3 in Table 5.2). 5-48 Figure 5.46. Normalized RI and LS signals observed for polystyrene (run 7 in Table 5.1) and soluble poly (STY/DVB) network (run 1 in Table 5.2) synthesized RAFT polymerisation at 130 °C. 5-49 Figure 5.47. Comparison of the observed RI signal for RAFT synthesized polystyrene (run 2 in Table 5.1) and soluble poly(DVB) (runs 2 and 3 in Table 5.2). 5-50 Figure 5.48. Observed RI signal for RAFT synthesized soluble poly(DVB) samples. Different stages of the crosslinking process (reaction times 1, 2 and 3 hr) are compared. 5-50 Figure 5.49. SEM micrographs of different polystyrene samples synthesized by RAFT. 5-51 xxv Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers INVERSE-SUSPENSION FREE-RADICAL POLYMERISATION LEADING TO HYDROGELS FORMATION Figure 6.1. A model depicting a stimuli-responsive drug-delivery system (adapted from Bajpai et al., 2010). 6-4 Figure 6.2. Structural representation of some anionic hydrogels. 6-6 Figure 6.3. Structural representation of some cationic hydrogels. 6-6 Figure 6.4. (a) Schematic of the response of a cationic hydrogel in basic medium. (b) Schematic of the response of a cationic hydrogel in acid medium. 6-6 Figure 6.5. (a)Schematic of the response of an anionic hydrogel in acid medium. (b) Schematic of the response of an anionic hydrogel in basic medium. 6-6 Figure 6.6. Schematic of a pH-sensitive hydrogel in acidic and basic medium. (A) Anionic (B) cationic. 6-7 Figure 6.7. Schematic representation of a polyamphoteric hydrogel of MAA/DMAEMA. 6-7 Figure 6.8. Schematic representation of some neutral hydrogels. 6-8 Figure 6.9 Structures of some temperature-sensitive polymers. 6-8 Figure 6.10. Illustration of the physical response of a negative temperature-sensitive hydrogel when submitted to heat or cool processes. 6-9 Figure 6.11. Chromatographic traces (RI signal) observed for AA/TMPTA samples synthesized through the inverse suspension process. 6-12 Figure 6.12. Chromatographic traces (RI signal) of the soluble polymer of the samples of AA/MBAm hydrogel collected at different reaction times. 6-13 Figure 6.13. MALLS chromatographic traces (only the 90° signal is shown) of the soluble polymer samples of AA/MBAm hydrogel collected at different reaction times. 6-13 Figure 6.14. Comparison of RI and LS signals observed in the SEC/RI/MALLS analysis of the soluble fraction of hydrogels or their linear counterparts. In this particularly case is an acrylamide based sample. 6-14 Figure 6.15. Chromatographic traces of two monomers (AA and AAM) present in inverse suspension polymerisation samples collected at different reactions times. 6-14 Figure 6.16. Comparison of observed and predicted reaction time evolution of monomer conversion during inverse suspension synthesis of hydrogels. 6-15 List of Figures xxvi Figure 6.17. Predicted dynamics of the weight fraction of gel and monomer conversion in batch SAP production using a trifunctional crosslinker with . 6-24 Figure 6.18. Predicted dynamics of the weight average molecular weight in batch SAP production using a trifunctional crosslinker with . 6-24 Figure 6.19. Prediction dynamics of the weight fraction of gel and monomer conversion in batch SAP production using a trifunctional crosslinker with . 6-25 Figure 6.20. Predicted dynamics of the weight fraction of gel in batch SAP production considering different initial mole fraction of trifunctional crosslinker ( . 6-25 Figure 6.21. Predicted time evolution of monomer conversion and weight gel fraction during acrylic acid/triacrylate copolymerisation with . Other parameters considered in the simulations: , (T=50 °C for V50), = 15%, , , 30000 L mol−1 s−1, = 5 × 107 L mol−1s−1, . 6-26 Figure 6.22. Numerical solution of the characteristics (vector s) correspondent to the SAP synthesis using the conditions described for Figure 6.21. For illustration purposes the chosen polymerisation time was t = 203.6 s. 6-27 Figure 6.23. Numerical solution of the characteristics (vector G) correspondent to the SAP synthesis using the conditions described for Figure 6.21. For illustration purposes the chosen polymerisation time was t = 203.6 s. 6-28 Figure 6.24. Predicted dynamics of the weight fraction of gel (wg) in batch SAP production considering three different synthesis processes correspondent to the use of three crosslinkers with different functionalities (bi-, triand tetrafunctional, considering N,N′-methylenebisacrylamide (α = 2), trimethylolpropane triacrylate (α = 3) and tetraallyloxyethane (α = 4) as case studies). 6-29 Figure 6.25. Comparison between the predictions of the Theory of Branching Process (TBP) and the proposed kinetic approach for the weight fraction of gel in the copolymerisation of a vinyl monomer with a bifunctional crosslinker (two initial compositions were considered using AA + MBAm as case study). 6-29 Figure 6.26. FTIR-ATR spectra observed during in-line monitoring of aqueous solution polymerisation of AA with 40% concentration (v/v), 80% neutralization, T=50 °C, initiation by V50 (mole ratio initiator/monomer= 0.2%). 6-31 Figure 6.27. FTIR-ATR spectra observed during in-line monitoring of aqueous solution polymerisation of AAM polymerisation at 9% (w/w), T= 20 °C, initiation by APS/TEMED (mole ratio initiator/monomer = 0.2%). 6-31 Figure 6.28. In-line FTIR-ATR estimated monomer conversion for aqueous monomer solution polymerisation of AA at 15 % concentration (v/v), 80% neutralization, T= 20 °C, initiation by 0.2% APS/TEMED. 6-32 xxvii Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 6.29. In-line FTIR-ATR estimated monomer conversion for aqueous monomer solution polymerisation of AAM at 9% concentration (w/w), T=20 °C, initiation by 0.2% APS/TEMED. 6-32 Figure 6.30. Comparison between experimentally observed and predicted dynamics of monomer conversion and weight fraction of gel during the homopolymeriation of AA and its crosslinking with TMPTA at T = 50 °C. Homopolymerisation of AA was performed in water solution with and . Crosslinking copolymerisation was performed in inverse suspension considering similar composition of the aqueous phase and . 6-33 Figure 6.31. Predicted and measured time evolution of the molecular weight of the soluble phase during inverse suspension hydrogels formation. Synthesis of AA/MBAm superabsorbent hydrogel was here considered for illustration purposes. 6-34 Figure 6.32. Predicted and measured time evolution of the weight fraction of gel for the same system described in Figure 6.31. 6-35 Figure 6.33. Repeated swelling (pH=7.5 and collapsing pH=1.2) of inverse suspension synthesized AA/AAM/MBAm hydrogels. 6-36 Figure 6.34. Observed volume change of NIPA/MBAm and DMA/MBAm inverse suspension synthesized hydrogels stimulated by temperature. 6-37 Figure 6.35. Measured dynamics of swelling of an inverse suspension synthesized pHresposnive hydrogel (AA/AAM/MBAm) in buffer solutions of different values of pH. 6-37 Figure 6.36. Examples of the measured swelling kinetics of different superabsorbent hydrogels showing the effect of synthesis conditions on the materials performance. 6-38 Figure 6.37. Gel micrographs of inverse suspension synthesized hydrogel beads. (a) AA/MBAm hydrogel. (b) DMA/MBAm hydrogel. 6-38 INVERSE-SUSPENSION REVERSIBLE ADDITIONFRAGMENTATION RADICAL POLYMERISATION LEADING TO HYDROGELS FORMATION Figure 7.1. Chemical structure of CPA RAFT agent. In chapter 5 are the chemical structures of DDMAT and CDT). 7-4 Figure 7.2. SEM micrographs of some hydrogel beads synthesized in this chapter using the inverse-suspension technique. (a) AA/MBAm hydrogel. (b) NIPA/MBAm hydrogel. (c) and (d) NIPA/AA/MBAm hydrogel. In some cases, formation of fused material was observed due to the post-treatment of the products (precipitation/drying). 7-7 List of Figures xxviii Figure 7.3. Refractive index (RI), right angle light scattering (RALS) and intrinsic viscosity-differential pressure (IVDP) signals simultaneously observed in the SEC analysis of a water soluble PAA sample. 7-9 Figure 7.4. IVDP signals observed in the SEC analysis of different water soluble polymers synthesized, highlighting the influence of operation conditions (FRP/RAFT) on the products molecular structure and properties. 7-9 Figure 7.5. In-line FTIR-ATR spectra observed during DMAEMA/EGDMA FRP polymerisation (run 1 in Table 7.3). Absorption peak at around 935 cm-1 was considered to estimate the double bonds conversion, using also the peak at 1720 cm-1 as internal reference. 7-10 Figure 7.6. FTIR-ATR estimated dynamics of monomer conversion for DMAEMA/EGDMA polymerisation (run 1 in Table 7.3). Similar measurements were performed with runs 2 and 3 in Table 7.3 but even lower monomer conversions were obtained in these experiments (almost negligible after 8 hours of polymerisation in run 3). 7-11 Figure 7.7. Measured equilibrium swelling ratio of NIPA/MBAm hydrogels in aqueous solutions at different temperatures illustrating networks sensitivity to changes in this parameter. 7-11 Figure 7.8. Dynamics of caffeine release from a pH/temperature sensitive synthesized hydrogel (NIPA/AA/MBAm) measured by UV detection at 270 nm. Two different surrounding water solutions were considered: pH=1/T=37 °C (collapsed particles). And pH=7/T= 22 °C (swollen particles). Drug loading was performed by swelling the hydrogel beads in caffeine aqueous solution during 48 hours. 7-12 Figure 7.9 Measured equilibrium weight swelling ration of anionic (AA based) and cationic (DMAEMA based) hydrogels in aqueous solutions at different pH values illustrating networks sensitivity to changes in this parameter. Inverse effect of the pH on the swelling ratio of these hydrogels can be exploited to trigger different macroscopic effects (e.g. shrunk to swollen networks by changing the pH from 1 to 8 with AA hydrogels and the opposite with DMAEMA hydrogels). 7-13 Figure 7.10. Comparison of the change of the equilibrium weight swelling ratio with pH for FRP and RAFT synthesized AA hydrogels. High effect of the synthesis technique used on this parameter is observed. The primary chain length of the networks is strongly affected when FRP is replaced by RAFt which can eventually be used to tune the swelling properties of the hydrogels. Note that results presented in Figures 7.9 and 7.10 were obtained using buffer aqueous solutions at different pH values. The swelling ratio of hydrogels is also strongly dependent on the ionic strength and size of the ions/counterions present in the solutions. A different dependence of SR can be observed if other aqueous solutions at the same pH values are considered (e.g. changing the used salts). 7-14 xxix Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Figure 7.11. Dynamics of release of 5-fluoruracil from a pH sensitive hydrogel (cationic hydrogel based on DMAEMA) measured by UV detection at 270 nm. 7-15 Figure 7.12. Dynamics of release of 5-fluoruracil from a pH sensitive hydrogel (anionic hydrogel based on AA) measured by UV detection at 270 nm. 7-15 Figure 7.13. Comparison of the dynamics release of ibuprofen from cationic (DMAEMA based) and anionic (AA based) hydrogels, both placed in aqueous solution at pH=10 (release measured by UV detection at 223 nm). 7-16 Figure 7.14. Dynamics of release of 5-fluorouracil from FRP and RAFT synthesized pH sensitive hydrogels illustrating the effect of the molecular architecture of the networks on their performance. Amount of drug released is here expressed as the fraction of drug loaded in the hydrogel that is transferred to the aqueous solution (release measured by UV detection at 270 nm). 7-16 Figure 7.15. Observed dynamics of monomer conversion (run 1 in Table 7.2) concerning the inverse-suspension of NIPA RAFT polymerisation with . 7-19 Figure 7.16. Observed dynamics of monomer conversion (run 2 in Table 7.2) concerning the inverse-suspension of AA RAFT polymerisation with . 7-20 Figure 7.17. Observed dynamics of monomer conversion (run 3 in Table 7.2) concerning the inverse-suspension of DMA RAFT polymerisation with . 7-20 Figure 7.18. Observed dynamics of monomer conversion (run 4 in Table 7.2) concerning the inverse-suspension of MAA RAFT polymerisation with . 7-21 Figure 7.19. Observed dynamics of monomer conversion (run 5 in Table 7.2) concerning the inverse-suspension of AA RAFT polymerisation with . 7-21 Figure 7.20. Observed dynamics of monomer conversion (run 6 in Table 7.2) concerning the inverse-suspension of AA/MBAm RAFT polymerisation with . 7-22 Figure 7.21. SEC chromatographic traces showing the presence of soluble material in AA/MBAm RAFT polymerisation in DMF (run 6 in Table 7.2) and the formation of a secondary population with high molecular size and very low concentration. 7-22 Figure 7.22. Dynamics of gel formation during the RAFT synthesis of stimuli-responsive hydrogels. DMAEMA/MAA/MBAm RAFT polymerisation at 50 °C using CPA agent (run 4 in Table 7.3) is here considered for illustration purposes. 7-23 Figure 7.23. SEC chromatographic traces showing the formation of a secondary population with high molecular size and very low concentration in RAFT polymerisation (run 4 in Table 7.2). 7-24 List of Figures xxx Figure 7.24. Dynamics of formation of the secondary population observed by light scattering in the same run of Figure 7.23. 7-24 APPENDICES Figure A.1. General scheme of the characterization procedure in all the copolymerisation reactions (depending on the system this scheme may be adapted). D Figure A.2. Schematic of a separation process of a polymer sample based on the molecules size (hydrodynamic volume). Note that A represents the injection step, B the separation, C the elution of large molecules and D the elution of small molecules. E Figure A.3. Photographical record of the Polymer Laboratories PL-GPC-50 containing the SEC system coupled with RI and MALLS detectors. G Figure A.4. A view of the MALLS measurement with 18-angle detector. The incident laser beam is polarized vertically with respect to the scattering plane shown (Wyatt, 1997). G Figure A.5. Photographic record of SEC system with the tetra-detector array (RI-LSIVDP-UV). H Figure A.6. Schematic of the on-line GPC viscometer with a series of four capillary bridges. I Figure A.7. Photographic record of the in-line FTIR/ATR system set-up used in this work. I Figure A.8. Schematic of the gravimetric technique used to obtain the monomer conversion in this work. K Figure A.9. Experimental procedure for determination of pendant double bonds by iodine method. N Figure A.10. Diagram RI versus time used in the determination of the dn/dc of a polystyrene sample. O xxxi Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers List of Tables CONVENTIONAL FREE-RADICAL COPOLYMERISATION OF MULTIVINYL MONOMERS Table 2.1. Description of the set of experiments performed in the study of the radical copolymerisation of STY with DVB in toluene solution at 60 °C in a semibatch reactor. 2-6 Table 2.2. Set of experiments performed in the radical copolymerisation of MMA with EGDMA in toluene solution at 60 °C. 2-7 Table 2.3. Description of the set of experiments performed in the study of the suspension copolymerisation of styrene/divinylbenzene at 60 °C. 2-7 Table 2.4. Chemical groups for the modeling of radical copolymerisation of styrene/ divinylbenzene. 2-11 Table 2.5. Kinetic scheme of radical copolymerisation of styrene/divinylbenzene. 2-14 Table 2.6. Basic set of kinetic parameters considered in the modeling of the radical copolymerisation of styrene with divinylbenzene at 60 °C. 2-27 Table 2.7. Assumptions used in the present system for some kinetic parameters considered in the radical copolymerisation of styrene with divinylbenzene at 60 °C. 2-28 Table 2.8. Chemical groups for the radical copolymerisation of MMA with EGDMA. 2-38 Table 2.9. Kinetic scheme considered in the radical copolymerisation of MMA/EGDMA at 60 °C. 2-39 Table 2.10. Assumptions used for some kinetic parameters considered in the modeling of radical copolymerisation of MMA with EGDMA at 60 °C. 2-40 Table 2.11. Basic set of kinetic parameters considered in the modeling of radical copolymerisation of MMA with EGDMA at 60 °C. 2-41 Table 2.12. Propagation rate coefficients ( ) considered in the radical copolymerisation of MMA with EGDMA at 60 °C. 2-42 List of Symbols xxxviii Rate coefficient of radical termination by disproportionation. Rate coefficient of radical termination by disproportionation of a polymer radical of kind i with a polymer radical with kind j. Rate coefficient of the irreversible chain transfer to specie x. Factor for calculating termination rate constants of acrylate monomers. Factor for calculating termination rate constants of acrylate monomers. Rate coefficient of a inhibitor Z with a primary radical of kind i. Rate coefficient of a inhibitor Z with a polymer radical of kind i. Initial mass fraction of monomer. l Bond length. Weight of the gel in equilibrium. Weight of gel. Mi Monomer, macromonomer or PDB of kind i. Weight of polymer. n M Number-average relative molecular mass. weight of sample collected from the reactor. Mw Polymer chain molecular weight. Mw0 molecular weight of a repetitive unit. z M z-average relative molecular mass. N Number of skeletal bonds of the polymer chain. nx Refractive index of the specie x. P Monomer conversion. Pk,m,n Polymer molecules bearing k radicals, m PDB and n repeating units. Reactivity ratio of PDB as compared to the double bonds of the monovinyl monomer. Reactivity ratio for radicals derived from chain transfer to monomer. Square of the random-flight end-to-end distance. xxxix Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Root mean square radius of gyration. z-average mean square radius of gyration. Unperturbed root mean square radius of gyration. Unperturbed root z-average mean square radius of gyration. Ri radical in polymer molecule of the kind i. Apparent values of reactivity ratio between a radical of kind i with a monomer or macromonomer of kind j. Polymerization reactivity ratio between a radical of kind i with a monomer or macromonomer of kind j. Reactivity ratio for internal radicals. Initial mole ratio between monomer and initiator. Reactivity ratio of PDB comparatively to the double bonds of the crosslinker monomer. Reactivity ratio for pendant double bond radicals. Rate of termination. Unperturbed polymer dimensions in the absence of long range interactions. sx Characteristics for calculation of chain length distribution using the extended method of moments. T Temperature. T Time. Glass transition temperature. Ui Repeating unit of kind i in polymer chains. Volume fraction of acrylic acid in the aqueous phase. Volume fraction of vinyl monomer in the solution. volume of titrant spend in blank sample. volume of titrant spend in polymer sample. Weight fraction of EGDMA in the total monomer content. Weight fraction of gel. List of Symbols xl Initial fraction of MAA weight. X Stable radical. X coordinate in a Debye plot. Initial mole fraction of crosslinker in the total monomer mixture. Y coordinate in a Debye plot. Mole fraction of initial double bonds belonging to the crosslinker. Initial mole fraction of divinylbenznene in the total monomer mixture. Yi Primary radical of the kind i. Initial mole ratio between initiator and total amount of monomers double bonds. Initial mole ratio between RAFT agent and initiator. Initial mass fraction of monomers in the organic phase. Initial volumetric fraction of monomers in the organic phase. Initial volumetric fraction of organic phase in the total liquid content (organic phase plus water). Initial mole ratio between monomer and RAFT agent. Initial mole ratio between NMRP mediator (TEMPO) and initiator. Greek Symbols α Functionality of the crosslinker (number of active double bonds). Relative rate of radical termination by combination. Relative rate of radical termination by disproportionation. Krönecker symbols. Krönecker symbols. Specific viscosity of the polymer in the solvent. xli Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers Specific viscosity of pure solvent. Specific viscosity of solution. Wavelength of the light. Vacuum Wavelength of the incident light in MALLS. θ Scattering angle in MALLS. Θ Good solvent conditions. Propagation probability. Exponent in the scaling law versus . List of Abbreviations A RAFT agent. AA Acrylic acid. AAM Acrylamide. AIBN 2,2´-azobis(2-methylpropionitrile). APS Ammonium persulfate. ATR Attenuated total reflectance. ATRP Atom Transfer Radical Polymerisation. BEDA Bisphenol-A ethoxylate diacrylate. BPO Benzoyl peroxide. BS Branching site. CDT Cyanomethyl dodecyl trithiocarbonate. CL Cross-linker. [CL]0 Initial concentration of cross-linker. CLD Chain length distribution. CM Transition metal/ligand complex. CPA 4-cyano-4-phenyl carbonothioylthio-pentanoic acid. CRP Controlled radical polymerisation. List of Symbols xlii CS Crosslinking site. CTA Chain transfer agent. CTP CH3 transfer to polymer center. CX Oxidized transition metal/deactivator. D Dimer. DDMAT 2-(dodecylthiocarbonothioylthio)-2-methylpropionic acid. DMA N,N-dimethylacrylamide. DMAEMA 2- (dimethylamino)ethyl methacrylate. DMF N,N-dimethylformamide. DNA Deoxyribonucleic acid. DT Degenerative chain transfer. DTLRP Degenerative chain living radical polymerisation. DVB Divinylbenzene. [DVB]0 Initial concentration of divinylbenzene. EBrP Ethyl 2-bromopropionate. EGDMA Ethylene glycol dimethacrylate. [EGDMA]0 Initial concentration of ethylene glycol dimethacrylate. EPR Electron paramagnetic resonance. FRP Free-radical polymerisation. FTIR Fourier transform infra-red. 5Fu 5-fluorouracil. GPC Gel Permeation chromatography. HDDA 1,6-hexanediol diacrylate. HHU Head-head units from termination by combination. HMTETA 1,1,4,7,10,10-hexamethyltriethylenetetramine. HTP Chain transfer to polymer center (methine group). I Initiator. [I]0 Initial concentration of initiator. IDB Non-reactive internal double bond. IP Inlet pressure through the bridge top to bottom. IPDB Index of pendant double bonds. IPN Interpenetrating polymer networks. xliii Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers ITP Iodine transfer polymerisation. IUPAC International union of pure and applied chemistry. IVDP Intrinsic viscosity differential pressure. LCB Long chain branching. LCST Lower critical solubility temperature. LS Light scattering. M Monomer. [M]0 Initial concentration of monomer. MAA Methacrylic acid. [MAA]0 Initial concentration of methacrylic acid. MALLS Multi angle laser light scattering. MBAm N,N´-methylenebisacrylamide. MBPA Methyl α-bromophenylacetate. [MBPA]0 initial concentration of methyl α-bromophenylacetate. MCT Mercury cadmium telluride. m-DVB meta-divinylbenzene. MEQH Monomethyl ether hydroquinone. MIPs Molecular imprinted polymers. MMA Methyl methacrylate. MW Molecular weight. MWD Molecular weight distribution. nBuA n-butyl acrylate. NIPA N-isopropylacrylamide. NMR Nuclear magnetic resonance. NMRP Nitroxide Mediated Radical Polymerisation. PAA Poly(acrylic acid). PBAA Poly(butyl acrylic acid). PDB Pendant double bonds. PDI Polydispersity index. PDMA Poly(dimethyl acrylamide). PDMAEMA Poly (2-(dimethylamino)ethyl methacrylate). PDVB Poly(divinylbenzene). List of Symbols xliv p-DVB para-divinylbenzene. PEAA Poly(ethyl acrylic acid). PLP Pulsed laser polymerisation. PMAA Poly(methacrylic acid). PMDETA N,N,N´,N´´,N´´,-pentamethyldiethylenetriamine. PMMA Poly(methyl methacrylate). PnBuA Poly(n-butyl acrylate). PNIPA Poly(N-isopropylacrylamide). PPAA Poly(propyl acrylic acid). PS Polystyrene. PVA Poly(vinyl alcohol). PVC Poly(vinyl chloride). [R] Concentration of radicals. RAFT Reversible addition-fragmentation chain transfer polymerisation. RALS Right angle light scattering. RI Refractive index. RX ATRP initiator. STY Styrene. [STY]0 Initial concentration of styrene. SAP Superabsorbent polymers. SCB Short chain branching. SEC Size Exclusion Chromatography. SEM Scanning Electron Microscopy. SET Outer sphere single electron transfer. SG Saturated end group in polymer chains. SM Dormant radical from the monomer. SP Single pulse. Span 60 Sorbitan monostearate. SRa Swelling ratio. SR Dormant radical from the initiator. T Solvent. [T]0 Initial concentration of solvent. xlv Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers TAM Transfer to acrylate monomer reaction. TAO Tetrallyloxyethane. TBP Theory of branching process. TDBD Terminal double bond from termination by disproportionation. TDBM Terminal double bond from chain transfer to monomer. TDM Transfer to diacrylate monomer reaction. TEMED N,N,N’,N’-tetramethylethylenediamine. TEMPO 2,2,6,6-tetramethylpiperidinyl-1-oxy. THF Tetrahydrofuran. TMPTA Trimethylolpropane triacrylate. TPBVP Two point boundary value problems. TS Saturated nitroxyl radicals. UCST Upper critical solubility temperature. UV Ultra violet. V50 2,2´-azobis(2-methylpropionamidine) dihydrochloride. Vac Vinyl acetate. VC Vinyl chloride. CHAPTER 1 INTRODUCTION Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 1-2 1.1 Introduction 1.1.1 Free-Radical (Radical) Polymerisation Here, a short discussion on chemical terminology will be introduced on how to cope with IUPAC recommendations. Thus, the terms “radical” or “free-radical” are restricted to those radicals which do not form parts of radical pairs. In the past the term “radical” was used to designate a substituent group bound to a molecular entity, as opposed to “free-radical”, which nowadays is simply called radical. The bound entities should no longer be called radicals (McNaughton and Wilkinson, 2012). The different kinds of polymers can be classified by their molecular architecture, namely whether they can be linear, branched or crosslinked polymers. The molecular structure of a linear polymer (Figure 1.1(a)) is a continuous chain of multiple repeating units. These units are added to the main chain following continuously without any side chain, besides the ones that are part of monomer structure. A branched polymer (Figure 1.1(b)) is composed by polymeric segments joint to the main chain. These segments are called branches, lateral chains or pending chains. At last, a crosslinked polymer (Figure 1.1(c)) is the result of the bond of polymer chains, forming a network of interconnected chains. The formation of multiple interconnections could lead to a single giant molecule consisting of all polymer chains of the system connected by at list one link. Thus, a polymer network can be considered as one or many polymer chains of extremely high molecular mass or infinite (Carothers, 1936). The radical polymerisation reaction involving a monomer with a double bond with another monomer with two double bonds is a typical example of a copolymerisation with simultaneous occurrence of branching/crosslinking. The divinyl monomer besides its function as a co-monomer is also a branching agent. With the polymerisation process advancement it is formed a tridimensional network. In this system can be found at the same time linear chains, branched chains or main chains connected by crosslinking points. In recent years the synthesis and characterization of branched/crosslinked polymers has been a subject with growing research activity. Indeed, there is much room for improving several properties of these materials as compared with their linear counterparts. Higher functional group densities and appropriated solubility and viscosity ranges can be exploited by producing hyperbranched polymers with different applications fields, such as drug or gene 1-9 CHAPTER 1. Introduction 1.1.2.3 Reversible Addition-Fragmentation Polymerisation Radical polymerisations involving a reversible chain transfer step for chain equilibration and also displaying the characteristics of living polymerisations were first reported in 1995 (Krstina et al. and Matyjaszewsky et al.). The mechanism of the reversible chain transfer step may involve homolytic substitution or addition-fragmentation. An essential feature is that the product of chain transfer is also a chain transfer agent with similar activity to the precursor transfer agent. The process has also been termed degenerate or degenerative chain transfer since the polymeric starting materials and products have equivalent properties and differ only in molecular weight (Moad and Solomon, 2006). Depiction of the main polymer coreequilibration step in RAFT polymerisation is carried out in Figure 1.2. M M Pm Pn ka S S kf kf Pn S Pm Pn ka Z Z Pm S Z S S Figure 1.2. Depiction of main polymer equilibration step in a RAFT polymerisation. 1.1.2.4 Other Controlled Polymerisation Techniques There are other controlled/living polymerisation techniques with scientific and industrial importance, namely, a combination of competitive outer sphere single electron transfer (SET) and degenerative chain transfer mechanisms (DT). Iodine-transfer, telluride-mediated and stibine-mediated polymerisations are the degenerative chain transfer most used. Here, will be only reported iodine transfer polymerisation (ITP) as illustration as none of this techniques as been performed in this work. Iodine transfer polymerisation as a method of living radical polymerisation was reported firstly by Tatemoto in 1992. The process involves conducting a polymerisation with a conventional initiator (e.g. AIBN) in the presence of an activated alkyl iodide. This controlled technique has been used for styrene (Gaynor et al., 1995; Goto et al., 1998), acrylates (Gaynor et al., 1995) and fluoro-olefins (Ameduri and Boutevin 1999; Tatemoto, 1992) polymerisations. The occurring of various side reactions in polymerisation of vinyl acetate like head addition during propagation and the formation of an aldehyde end group by acid catalyzed decomposition of end group (Iovu and Matyjaszewski, 2003) are drawbacks of Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 1-10 Iodine-transfer polymerisation. Despite this side reactions, relatively low dispersities <1.4 are observed for molecular weights less than 20000 (Moad and Solomon, 2006). Outer sphere single electron transfer has a significant role in organic chemistry. This process provides an intermediate radical anion or cation. The living polymerisation of vinyl chloride (VC) with alkyl iodide initiators and nascent Cu(0) catalyst is considered to involve an SET process (Percec et al., 2002 and 2003). SET does not require a metal catalyst and can involve other single electron reducing agents as dithionite (Percec et al., 2005). The Percec group developed for the first time, a polymerisation process able to produce PVC with controlled features (Percec et al., 2002 and 2003). The further discovery of the single electron transferdegenerative chain living radical polymerisation (SET-DTLRP) (Percec et al., 2004a), made possible the use in synthesis of activated (Coelho et al., 2008 and 2009) and non-activated monomers (Percec et al, 2004a and 2004b), using a method which should hopefully be possible to implement at industrial scale (Coelho et al., 2011a and 2011b). The main interest of SET-DTLRP is the production of poly(vinyl chloride) block copolymers. By controlling the composition and the molecular weight of this block copolymer it is possible to produce a material with similar mechanical and thermal properties to conventional plasticized poly(vinyl chloride) (Coelho et al., 2006a, 2006b and 2006c) without requiring the use of extractable and possibly toxic external plasticizers. 1.2 Intramolecular Cyclization Reactions The issue of intramolecular cyclization reactions was revealed with the first studies about copolymerisation with crosslinking and gel formation. The cyclizations reactions occur between a radical site and a pendant double bond in the same polymer molecule and may be classified in two types: primary cyclization (Figure 1.3), which occurs in primary molecules (without branching or crosslinking) and secondary cyclization (Figure 1.4), which occurs in molecules containing two or more primary chains. These intramolecular reactions have a huge influence on the homogeneity of the network and gel formation. Thus, the understanding of these reactions is of great importance to ensure the product quality and to improve the polymer properties. The effect of cyclization is more pronounced in FRP with more diluted systems (Gonçalves et al., 2007; Trigo et al., 2008). Some significant improvements are achieved with controlled systems, ATRP (Gonçalves et al., 2010a, 2010b, 2010c and 2010d), NMRP (Gonçalves et al., 2013a) and especially with RAFT, where the Luo group has 1-11 CHAPTER 1. Introduction achieved some important results on particle size by decreasing the level of the amphiphilic RAFT agent (Ye et al., 2011).   Figure 1.3. Schematic of a general primary cyclization reaction.   Figure 1.4. Schematic of a general secondary cyclization reaction. 1.3 Relevance and Motivation The free-radical polymerisation of vinyl monomers is responsible for more than 50 % of the industrial production of synthetic polymers. The economic impact of product and process engineering in this area is obviously very important even arising from small improvements. Besides commodity polymers, new materials for special technology applications are also produced through this class of chemical processes. Development of tools allowing the design of polymers with tailored properties is especially important in this context owing to the strong relation between the molecular architecture and the end-use properties of such advanced materials. The synthesis of hyperbranched polymers through step-growth polymerisation is a well studied subject. However, some materials are produced by free-radical polymerisation. One of the major characteristics of this method is the gel formation at relatively low monomer conversion, which can harm the properties of the final product. Thus, to avoid the gelation at low conversion is necessary to implement new synthesis techniques such as the use of a chain transfer agent, semi-batch reactor (Gonçalves et al., 2007). In the last decades, many synthetic polymers have been produced by free-radical polymerisation. The most used synthesis Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 1-12 techniques are bulk, solution, emulsion and suspension. In the last years, new technologies have been employed like miniemulsion and supercritical carbon dioxide. This work will focus on the “living” controlled radical polymerisation which allows the production of polymers with high homogeneity comparatively to free-radical polymerisation technique. There are various kinds of controlled techniques being ATRP, NMRP and RAFT the most used. The ATRP technique is very versatile, and may be used in several chemical systems but has the inconvenience of generate huge quantities of metals in the final product. The NMRP technique it is based on nitroxide radicals which are very easy to handle but are almost restricted to styrenic monomers. The RAFT technique can be used with a wide range of monomers and in different chemical systems but the RAFT transfer agents are not easy to obtain. In the industrial area, the advances in the polymer production occur in a rapid way. This material have a lot of properties to be analysed which are modified in function of the process variables (feed concentration, temperature, etc.). Many polymers are commercialized without knowing all the reactions involving on its production. Indeed, CRP techniques have the need for a definition of the mechanism that is responsible for the controlled copolymerisation process, with presence of secondary reactions; with effects of the controlling agent on the gel point and the formation of intramolecular cyclizations with effects on properties of the final product. The present work also includes the issues above described. The experimental data obtained associated with the modeling tool used allow the improvement of the molecular architecture of the produced materials and its future implementation at industrial level. Besides that, the results obtained in this work can be used to obtain useful tools in the project of hyperbranched and polymer networks with tailored properties. 1.4 Objectives and Outline The scope of the present is the synthesis and properties of the products from Free-Radical Polymerisation (FRP) and Controlled Radical Polymerisation (CRP) of multivinyl monomers. Therefore, it is experimentally assessed the impact of the synthesis conditions on the formation mechanism and on the properties of hyperbranched polymers, gels and hydrogels. Namely, the following parameters were changed along the experimental program:  Different synthesis techniques are used (FRP, ATRP, NMRP and RAFT) with various types of mediator agents, initiators, solvents, degree of neutralization (in acrylic and methacrylic acids) and temperatures. 1-13 CHAPTER 1. Introduction  Different types of monomers families (acrylates, methacrylates, styrenics and acrylamides).  Solution, suspension and inverse-suspension are used as polymerisation systems.  Two bench polymerisation reactors are used, one atmospheric with a semi-industrial capacity of 2.5 L and a pressurized reactor with capacity of 1 L to allow the performance of aqueous reactions near and above 100 °C.  The partial and the final products are characterized through a set of techniques to assess their properties and the molecular architecture. Namely are evaluated the following properties:  Dynamic of conversion is performed by gravimetry, size exclusion chromatography with refractive index detection and in some case by in-line FTIRATR.  Dynamic of product formation through the determination of average molecular weights, z-average radius of gyration and absolute molecular weight distribution by gel permeation chromatography with different detectors (light scattering, refractive index, intrinsic viscosity and ultra violet (both in the aqueous case)).  Also by gravimetry is determined the swelling ratio capacity, gel and sol fractions.  Iodine chloride derivatization method is used to determine the concentration of pendant double bonds of the networks. This variable is also estimated less accurately by FTIR in off-line mode.  Using the ultraviolet and refractive index detection is determined the dynamics of drug releasing by the hydrogels produced (aqueous system). With “smart polymers” it is assessed the response of these hydrogels to environmental changes on pH or/and temperature As a complement to this work, these experimental studies are supported by the development of kinetic models with the capacity to describe these non-linear polymerisations systems (Costa and Dias, 1994, 2003, 2005, 2006 and 2007; Dias and Costa, 2003, 2005a, 2005b, 2006, 2007 and 2010). It is used a computational package that allows the automatic inclusion into the kinetic schemes of different kinds of branching and crosslinking phenomena. Thus, using the experimental runs performed in this work has leaded to the improvement of a tool that could help in the inception of synthesis conditions to obtain materials with tailored properties. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 1-14 Chapter 2 deals with conventional free-radical copolymerisation of multivinyl monomers. Chapter 3 is focused on the study of atom transfer radical copolymerisation of acrylates. Chapter 4 shows the results obtained with nitroxide mediated radical copolymerisation of STY/DVB. Chapter 5 presents the reversible addition-fragmentation polymerisation of styrene and copolymerisation of styrene/divinylbenzene. In Chapter 6 it is assessed the conventional free-radical copolymerisation synthesis of hydrogels from water compatible monomers. Chapter 7 shows the importance of RAFT to the production of water compatible polymers. Chapter 8 describes the conclusions and suggests some recommendations for future work. CHAPTER 2 CONVENTIONAL FREE-RADICAL COPOLYMERISATION OF MULTIVINYL MONOMERS Abstract. In the present chapter the impact of the synthesis conditions on the formation mechanism and on the properties of hyperbranched polymers is assessed through experiments on styrene/divinylbenzenes and of methyl methacrylate/ethylene glycol dimethacrylate polymerisations. Gel formation and some gel properties were also studied with the former chemical system. More specifically, the experimental program has carried out measurements on:  Solution polymerisation at batch and semi-batch operation. Suspension polymerisation at batch operation.  Dynamics of conversion using gravimetry, size exclusion chromatography (SEC) with refractive index (RI) detection and in some cases exploiting in-line FTIR-ATR.  The molecular architecture of the products through the determination of average molecular weights, z-average radius of gyration and absolute molecular weight distribution by size exclusion chromatography coupled with light scattering and refractive index detectors.  The influence of key polymerisation parameters on the dynamics of gelation, namely the initial relative amount of crosslinker on gel formation through the determination of gel and sol fraction along polymerisation time and the assessment of the influence of initial proportions between monomers and inert diluents.  The morphology of the produced gel beads of STY/DVB by SEM with different kinds and amounts of the diluent. Intramolecular cyclization was mostly neglected in the complementary studies on modelling using the experimental data obtained below cited. This chapter is based on the following publications: M.A.D. Gonçalves, R.C.S. Dias, M.R.P.F.N. Costa, Macromol Symp. 259 (2007) 124-134. I.M.R Trigo, M.A.D. Gonçalves, R.C.S. Dias, M.R.P.F.N. Costa, Macromol. Symp. 271 (2008) 107-119. M.A.D. Gonçalves, V.D. Pinto, R.C.S. Dias, M.R.P.F.N. Costa, Macromol Symp. 302 (2011) 179-190. 2-9 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers chromatographic traces of the SEC analysis (the full procedure to obtain these results is described in appendix A). 2.2.6 In-line FTIR-ATR Monitoring These polymerisations were monitored in-line using an Attenuated Total Reflection immersion probe coupled to a Fourier Transform Infra-Red spectrophotometer. The following instruments were used: Axiom analytical immersion probe model DPR 207 (ZnSe element with spectral cut-off at 600 cm-1, maximum pressure and temperature operation 60 bar and 280 °C, respectively) and an ABB Bomem Fourier Transform Infra-Red spectrophotometer, model FTLA2000-104. The probe and the spectrophotometer are connected by a three arms light guide and an ABB Bomem Mercury-Cadmium-Telluride (MCT) detector (model D10B), cooled with liquid nitrogen. The spectrometer is equipped with this kind of the detector in order to increase the sensitivity of the analysis. These FTIR-ATR measurements were performed using the spectrum of air taken at room temperature as the reference background being the optical system continuously flushed with argon. A resolution of 4 cm-1 was used for the spectra that were taken over the full MIR range from 600 cm-1 to 4000 cm-1. Each spectrum was calculated from 128 interferograms (the full procedure is described in appendix A). 2.2.7 Measurement of the Weight Fraction of Gel Polymer samples (soluble and insoluble fractions) obtained by precipitation in methanol were filtrated and dried in vacuum overnight. Afterwards, they were washed several times, during one week, in large amounts of THF in order to collect the insoluble network. The time evolution of the insoluble weight fraction of polymer during the polymerisation was thus measured (the full procedure is described in appendix A). 2.2.8 Swelling Ratio Measurements For each run, the final suspension, correspondent to six hours polymerisation reaction time, was processed as above described in order to isolate the produced gel beads (section 2.4.1.5). The swelling ratio of the dried gel beads was estimated by weighing around 1 g of material which was afterwards immersed in a large amount of THF. After 24 hours the swollen gel beds were weighed again and the ratio between swollen and dried weights was used to estimate the swelling ration of these materials (the full procedure is described in appendix A). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-10 2.2.9 Scanning Electron Microscopy The dried gel beads above described were also analysed by Scanning Electron Microscopy (SEM) in the Centro de Microscopia da Universidade do Porto (CEMUP). The influence of the synthesis conditions in the structure (macroporous formation) of the produced materials could thus be assessed. 2.3 Solution copolymerisation of Styrene with Divinylbenzene 2.3.1 Kinetic Modelling The modelling of the present case study was carried out using the general kinetic approach allowing the prediction of molecular weight as well as z-average molecular radius of gyration for non-linear irreversible polymerisations system (Costa and Dias, 2005 and 2007; Dias and Costa, 2005a and 2006). 2.3.1.1 Chemical Species Twenty nine chemical species are considered in the kinetic modelling for the copolymerisation of styrene with divinylbenzene initiated by AIBN (Table 2.4). In Figures 2.2 and 2.3 are represented the monomers and the pendant double bonds present in this system: styrene monomer, divinylbenzene monomers consisting of two isomers (metaand para-) with different reactivities. Hence, the two pendant double bonds arising from DVB isomers are treated as two additional monomers. Some other chemical groups such as initiator, solvent and chain transfer agent as well as their primary radicals are represented in Figures 2.4 and 2.5. H2C CH CH H2C CH2 CH CH2 H2C CH HC (a) (b) (c) Figure 2.2. Schematic representations of (a) styrene (b) m-divinylbenzene and (c) p-divinylbenzene. 2-11 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers Table 2.4. Chemical groups for the modelling of radical copolymerisation of styrene/ divinylbenzene. Group description Alias Styrene (STY) M1 m-Divinylbenzene (m-DVB) p-Divinylbenzene (p-DVB) M2 M3 Initiator (I) Solvent (T) Chain transfer Agent (CTA) Inhibitor (Z) Retarder (R) Primary Radical from initiator (PRI) Primary radical from solvent (PRS) Primary radical from chain transfer agent (PRCTA) Radical from styrene (RS) Radical from m-divinylbenzene (RmDVB) Radical from p-divinylbenzene (RpDVB) Radical from m-pendant double bonds (RmPDB) Radical from p-pendant double bonds (RpPDB) Radical from retarder (RR) m-pendant double bonds (m-PDB) p-pendant double bonds (p-PDB) Polymerised Styrene Polymerised m-Divinylbenzene Polymerised p-Divinylbenzene Crosslinking site from m-divinylbenzene Crosslinking site from p-divinylbenzene Fragments from initiator, solvent, chain transfer agent, inhibitor and retarder C1 C2 C3 C4 C5 R1 R2 R3 A1 A2 A3 A4 A5 A6 M4 M5 U1 U2 U3 U4 U5 F1-F5 Five different kinds of polymer radicals are shown in Figures 2.6-2.8 since they present different structures and therefore different reactivities are also expected (Okay, 2000; Moad and Solomon, 2006; Nyhus et al., 1999; Hecker, 2000). Pendant double bonds (PDB) arising from commercial divinylbenzene, which are akin of two additional monomers in the present analysis, are also mutually distinguished as well as from the double bonds in the monomers as they are known to show different reactivities (Hecker, 2000). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-12 CH CH2 CH H2C (a) (b) Figure 2.3. Schematic representations of (a) p-pendant double bond and (b) m-pendant double bond. CH3 C CH3 CN CH3 C CN CH3 H3C N N (a) (b) (c) Figure 2.4. Schematic representation of (a) toluene (b) AIBN and (c) carbon tetrabromide. CN CH3 H3C H H (a) (b) (c) Figure 2.5. Schematic representations of primary radicals from (a) initiator (b) solvent and (c) chain transfer agent. H C CN CH3 CH2 H3C Figure 2.6. Schematic representation of the radical from styrene. H CH2 C HC CN CH3 CH2 H3C H C CN CH2 CH CH3 CH2 H3C (a) (b) Figure 2.7. Schematic representations of radical from (a) p-divinylbenzene and (b) radical from mdivinylbenzene. 2-13 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers H C CN CH3 CH2 H3C H C CN CH3 CH2 H3C (a) (b) Figure 2.8. Schematic representation of the radicals arising from (a) m-PDB and (b) p-PDB. 2.2.1.2 Chemical Reactions The kinetic mechanism considered in the polymerisation/copolymerisation of a monovinyl monomer (styrene) with a divinyl monomer (divinylbenzene) initiated by a thermal initiator (AIBN) will be now discussed. In this case, as previously stated, one must take into account the two isomers of divinylbenzene, metaand para-, as they show different reactivities. In Table 2.5 the kinetic scheme considered for this polymerisation system is presented. A total count of 114 chemical reactions is supposed to exist: initiator decomposition (1), initiation of monomers and pendant double bonds from primary radicals (15), propagation of monomers and pendant double bonds with different kinds of polymer radicals (30), chain transfers to solvent (6), chain transfer to chain transfer agent (6), inhibition of polymer and primary radicals (9), retardation of polymer centered radicals (21), termination by disproportionation of polymer radicals (21), termination by combination of polymer radicals (21). Despite the often difficult distinction between inhibitors (such as 4-tert-butylcatechol) and retarders (deactivation of primary radicalsdeactivation/slowing of polymer radicals), the kinetic steps here considered involving these two species take into account the deactivation of all kinds of radicals and the existence of a polymer radical site with a lower reactivity coming from propagation with a retarder (such as oxygen). This radical is supposed to polymerise with monomers, as commonly accepted for oxygen centered radicals in styrene polymerisation (Moad and Solomon, 2006). The leading chemical reactions present in this system are below described. As in nearly every free-radical polymerisations, it is convenient to consider the initiator decomposition as its first step. In Figure 2.9 it is represented that process leading here to the Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-14 formation of two primary radicals and nitrogen, where f represents the initiator efficiency factor. Table 2.5. Kinetic scheme of radical copolymerisation of styrene/divinylbenzene. Kinetic Step Chemical Equation Initiator Decomposition Initiation of monomers and PDBs Styrene Propagation m-DVB and p-DVB propagations m-PDB and p-PDB propagations Chain transfer to solvent Chain transfer to agent Inhibition of polymer radicals Inhibition of primary radicals Retardation of polymer radicals Termination by combination Head-Head Unit Termination by disproportionation Saturated + Unsat. Units 2f C CN CH3 CH3 CN H3C CH3 C CN CH3 H3C N2 N N Figure 2.9. AIBN decomposition leading to primary radicals. The initiation stage consists in the set of chemical reactions of primary radicals, arising from the initiator decomposition with the carbon-carbon double bonds present in the system. The reactivities of the double bonds depend on the monomer where they belong and therefore the initiation reactions present different kinetic constants. In Figures 2.10-2.12 are presented the initiation reaction by primary radicals of styrene monomer and divinylbenzene isomers. 2-15 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers The reactivity of the double bonds of the same monomer varies when one of them has been converted in a macromolecular chain. This fact is an example of a substitution effect and thus, as we have seen above, is necessary to distinguish the pendant double bonds from both DVB isomers. In Figures 2.13 and 2.14 are depicted the initiation reactions of the pendant double bonds from m-DVB and p-DVB, respectively. H2C CN H C CH CH3 CN H3C CH3 CH2 H3C Figure 2.10. Initiation reaction of the styrene monomer by primary radicals. CN H CH2 CH H2C CH3 C H3C CH H2C CH CN CH3 CH2 H3C Figure 2.11. Initiation reaction of the m-divinylbenzene monomer by primary radicals. CN H CH2 CH2 H2C CH3 C H3C CH HC HC CN CH3 CH2 H3C Figure 2.12. Initiation reaction of the p-divinylbenzene monomer by primary radicals. The reactivities of double bonds depend on the monomer type and by the effect of substitution inside the divinyl monomer; also the macromolecular radicals present different reactivities depending on the parent monomer and their chain position. Thus, the reactivity of the freeradical depends on the substituent of the carbon atom where the radical is located. Thus, is Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-16 necessary to distinguish the propagation reactions according to the nature of monomer, kinds of double bonds and macromolecular radicals involved. It was already stated that it is necessary to consider five different kinds of double bonds and it will be also necessary to consider five kinds of macromolecular radicals. Therefore in the propagation reaction stage it is required to describe twenty five different processes. CH H2C H C CN CH3 CH2 H3C CN CH3 H3C Figure 2.13. Initiation reaction of the m-divinylbenzene pendant double bond by primary radicals. CN CH3 H3C H CH C CH2 CN CH3 CH2 H3C Figure 2.14. Initiation reaction of the p-divinylbenzene pendant double bond by primary radicals. In Figure 2.15 it is represented the propagation stage of the styrene monomer with a macromolecular radical of the same monomer (homopropagation). The new formed radical is derived from the styrene monomer. H2C CH2 CH CH2 CH2 Figure 2.15. Propagation of styrene monomer with a radical formed by styrene monomer. 2-17 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers A radical formed by styrene can also react with one of the double bonds of the metaor paradivinylbenzene being formed by this process a radical derived from metaor paradivinylbenzene as shown in Figures 2.16 and 2.17, respectively. At last, a radical from styrene reacts with a pendant double bond of metaor paradivinylbenzene by the process shown on Figures 2.18 and 2.19, respectively. The macromolecular radical resulting from this process is different from the previous because it belongs to the internal structure of the macromolecule. Hence it is called internal radical or radical from the pendant double bond. Furthermore, this process creates a tetrafunctional branched (crosslinking) point of the polymer and therefore leads to the formation of a branched structure. CH2 CH2 HC CH2 HC H2C CH2 CH CH2 Figure 2.16. Propagation of m-DVB monomer with a radical formed by styrene monomer. CH2 CH2 H2C HC CH2 CH CH2 HC CH2 Figure 2.17. Propagation of p-DVB monomer with a radical formed by styrene monomer. CH2 HC H2C CH2 CH2 Figure 2.18. Propagation of m-DVB pendant double bond with a radical formed by styrene monomer. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-18 CH2 HC CH2 CH2 CH2 Figure 2.19. Propagation of p-DVB pendant double bond with a radical formed by styrene monomer. When a radical derived from metaor paradivinylbenzene propagates with a styrene unit as shown in Figures 2.20 and 2.21 a radical of styrene and a pendant double bond (with position depending on the DVB isomer) are formed. This is called a cross-propagation. The propagation of a metaor paradivinylbenzene with a radical formed by metaor paradivinylbenzene monomer is represented in Figures 2.22-2.25. From this reaction is obtained a new radical of metaor paradivinylbenzene and also a pendant double bond. H2C CH2 CH2 HC CH CH2 H CH2 CH2 HC Figure 2.20. Propagation of styrene monomer with a radical formed by m-DVB monomer. H2C H H2C H2C CH2 CH CH CH CH2 CH2 Figure 2.21. Propagation of styrene monomer with a radical formed by p-DVB monomer. 2-25 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers produces a dead polymer chain and a radical that is less reactive than the propagating radical but still capable of reinitiating the polymerisation is called retardation or degradative chain transfer. CH2 CH2 CH2 CH2 Figure 2.42. Termination by combination of two radicals derived from m-divinylbenzene pendant double bonds. In Figure 2.43 the formation of a primary radical of toluene (solvent used in this system) with a radical derived from styrene is depicted. In Figure 2.44 it is shown the formation of a primary radical of the chain transfer agent (CBr4) with a radical derived from styrene. Each macromolecular radical is capable of forming a primary radical of toluene and CBr4 resulting on eight more chemical reactions due to chain transfer reactions. The transfer agents can be used industrially in a radicalar polymerisation with the purpose of controlling the molecular mass of the polymer, the polymerisation rate or the nature of the end groups in the polymer. Intramolecular cyclization reactions occur when a radical site reacts with a pendant double bond in the same polymer molecule. When that reaction takes place in a primary chain of the polymer molecule it is called a primary cyclization and is depicted in Figure 2.45, with the copolymerisation of styrene with divinylbenzene as an example. If the intramolecular reaction occurs between a radical located in a different chain from the pendant double bond it is called a secondary cyclization. For the sake of a good comprehension of the scheme, it is used now a less detailed representation in Figure 2.46. Table 2.6 and 2.7 present the numerical values of the 115 kinetic parameters used in the simulations. Past experimental works concerning the homopolymerisation of styrene (Moad and Solomon, 2006), pure m-divinylbenzene or pure p-divinylbenzene (Nyhus et al., 1999) and the crosslinking of styrene with both pure DVB isomers and their mixture (Hecker, 2000) were used for estimating most of the needed parameters. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-26 H H CH2 CH2 CH3 H H Figure 2.43. Transfer to solvent reaction between a radical from styrene and toluene. H H CH2 CH2 Figure 2.44. Transfer to chain transfer agent reaction between a radical from styrene and CBr4. H2C H CH2 C CH2 H2C HC CH2 CH2 Figure 2.45. Schematic representation of a primary intramolecular cyclization reaction in the copolymerisation of styrene and divinylbenzene. H R CH CH2 H CH2 Figure 2.46. Schematic representation of a secondary intramolecular cyclization reaction. 2-27 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers Table 2.6. Basic set of kinetic parameters considered in the modelling of the radical copolymerisation of styrene with divinylbenzene at 60 °C. Kinetic Step Kinetic parameter involved a) AIBN decomposition b) Styrene homopropagation c) Styrene/m-and p-DVB propagations d) Styrene/m-and p-PDB propagations d) mand p-DVB homopropagations e) Polymer radicals termination f) Chain transfer to solvent g) Chain transfer to agent h) Inhibition i) Reaction radical/retarder j) a) All the kinetic parameters are expressed in dm3mol-1s-1, unless otherwise stated. b) Collected from Moad and Solomon, (2006) (p.71) and from Bevington (1955). c) IUPAC benchmark value (Moad and Solomon, 2006) collected from Buback et al., (1995). d )Reactivity ratios and were collected from Hecker (2000). e) Propagation constants and are based on the experimental evaluation of the parameter for the homopolymerisation of m-divinylbenzene and pdivinylbenzene (Nyhus et al., 1999) at 70 °C. Activation energies of propagation ( ) Moad and Solomon (2006) and for termination ( ) (Brandrup et al., 1999) were used to estimate this parameter at 60 °C. An overall average termination constant ( ) was considered in these calculations. f) is an average termination constant in the framework of the classical kinetics. Scattered values in the range 0.01 to 0.04 for the parameter can be found on the literature (Okay, 2000; Brandrup et al., 1999; Odian, 2004; Matheson et al., 1951; Beuermann and Buback, 2002). The value was estimated from time/conversion data (prior to noticeable Norrish-Trommsdorff effect) as measured in the present work. Termination of styrene is generally accepted to occur by combination ( ). Nevertheless, values of are also mentioned in the literature (Moad and Solomon, 2006). Experimental data of the present work are also consistent with . g) Correspondent to the pair styrene/toluene (Moad and Solomon, 2006). h) Scattered values for the pair styrene/CBr4 ( ) are reported in Brandrup et al., (1999): the experimental results of the present system are consistent with transfer constant close to ideality ( ) as also reported in the literature (Moad and Solomon, 2006). i) Correspondent to the pair styrene/p-benzoquinone (Moad and Solomon, 2006). j) In the range of known rate constants for reaction of carbon-centered radicals with oxygen (Moad and Solomon, 2006). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-28 Table 2.7. Assumptions used in the present system for some kinetic parameters considered in the radical copolymerisation of styrene with divinylbenzene at 60 °C. Kinetic Step Kinetic parameter involved Propagation of STY, DVB and PDB with radicals a) A2 to A5 Geometric decay Propagations with radicals RR b) Initiations c) Chain transfer to agent and to solvent d) Inhibition d)e) Reaction of radicals with retarder d) Termination f) a) It is considered that a decrease of reactivity of RmDVB relatively to RS occurs as inferred from experimental data leading to . In the same conditions a lower reactivity ratio for RpDVB ( ) is estimated. It is postulated that RmPDB and RpPDB present a reactivity drop in a geometric sequence relatively to RmDVB and RpDVB ( and , respectively). These reactivity drops are plausible owing to steric factors. It has been previously shown (Costa and Dias, 2003) that only with a strong deviation from ideality a noticeable influence of these parameters in the gelation of the system is expected. b) The propagation for the monomers with non-carbon centered radicals (oxygen centered) is considered to be much slower than with the correspondent normal propagation (Moad and Solomon, 2006). c) The rate constant for the initiation of the different monomers (j=1,...,5) with different kinds of primary radicals (k=1,...,3) are considered to take the same values as in the corresponding propagations with RS. d) The different propagating radical sites are considered to have similar reactivity decays for reactions with CTA, solvent, inhibitor and retarder. e) The inhibition of the primary radicals is supposed to occur with RS at the same relative extent. f) It is considered that the kinetic constants for all termination reactions take the same values as the average termination rate constant. 2.3.2 Results and Discussion Table 2.1 describes the set of experiments on the radical copolymerisation of STY with DVB in toluene solution at 60 °C. Linear polystyrene (run 1) and STY/DVB without (run 2) and with CTA (run 3) have been produced in a batch reactor. Semi-batch runs using different feeding policies of DVB are described in experiments 4, 5 and 6. A more or less constant global mole fraction of DVB was used in all batch and semi-batch experiments. In Figure 2.47 are presented the experimentally measured and predicted values of global monomer conversion. Those data were used to estimate the parameter 2-29 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers considering kinetic controlled polymerisation (Okay et al., 1999) for STY/DVB copolymerisation prior to gelation. In Figure 2.48 are compared the experimental measurements of for the non-linear polymerisations of STY/DVB in batch reactor (run 2) with the predictions obtained considering the set of kinetic parameters presented in Table 2.6 and 2.7. It can be observed that in these conditions ( ) a huge discrepancy between predictions and experimentally values of (and gelation time) occurs. Some of this discrepancy can be explained by the influence of the reaction medium, namely the effect of the solvent (Nyhus et al., 1999) on the kinetics of these polymerisation systems (toluene in this case versus nearly bulk polymerisation in (Hecker, 2000)). However, the most important cause of the delay of the gelation is likely to be the occurrence of cyclization (intramolecular) reactions (Okay, 2000; Dias and Costa, 2005). For the sake of simplicity, in the present work, these combined effects are taken into account by fitting apparent values of the reactivity of pendant double bonds, as they are the major parameters controlling the crosslinking process. It is also considered that the reactivity of mand ppendant double bonds is affected in the same proportion by these phenomena. Apparent values of reactivity ratios and were therefore estimated using the experimental information in run 2 (see Figure 2.48). Note that these apparent reactivity ratios are consistent with a decrease of reactivity of PDB ( as reported in other works in this field (Okay, 2000). This modified set of kinetic parameters was used in the remaining predictions here presented. In this system, the priority was the description of semi-batch operation, with a main goal being the prediction of the polymerisation behaviour using the same set of kinetic parameters as for batch operation and this was achieved, as shown in Figure 2.49. Figure 2.50 shows the comparison of measured and predicted values of, , and for a semi-batch run. A good agreement is obtained for (which is the molecular mass directly measured by MALLS) but significant deviations are observed for and close to gel point when a high dispersion of molecular sizes occurs. This is a consequence of the approximations involved in the indirect estimation of e which considers a homogeneous polymer population inside each SEC slice. Inaccurate measurements for and are thus obtained using SEC/RI/MALLS for highly polydispersed polymers. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-30 0.0 0.2 0.4 0.6 0.8 1.0 0 4 8 12 16 Run 1 2 3 4 5 6 Predictions Monomer conversion Time (h) 6 1, 2, 3, 4, 5 Figure 2.47. Time evolution of the measured overall conversion for linear and non-linear polymerisation systems also considering different operation conditions. 104 105 106 107 0 2 4 6 8 10 Weight average molecular weight Time (h) r14= 0.92, r15= 0.5 rap 14= 2.72, rap 15= 1.48 Run 2 Figure 2.48. The influence of the reactivity of the pendant double bonds on the predicted (run 2) and its comparison with the correspondent measured values. 2-31 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers 104 105 106 107 0 4 8 12 16 Run 1 2 4 5 Predictions Weight average molecular weight Time (h) 1 245 Figure 2.49. Predicted and measured for linear and non-linear polymerisation systems in batch and semi-batch reactor. 103 104 105 106 107 0 4 8 12 Weight Average Number Average z - Average Predictions Average molecular weight Time (h) Run 4 Mn Mw Mz Figure 2.50. Time evolution of , and during the radical copolymerisation of STY/DVB in semi-batch reactor using different feed policies. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-32 Figure 2.51 shows the comparison of batch operation including a CTA and different feed policies of DVB to a semi-batch reactor. These results show that operation in semi-batch reactor can be used as an alternative to the inclusion of CTA in batch operation, namely for producing soluble branched polymers at higher monomer conversions. The time evolution of the experimental z-average mean square radius of gyration ( ) in THF solution as obtained by MALLS is presented in Figure 2.52. For the smaller molecular dimensions approaching 10 nm, higher experimental errors are observed as the lower limit of the instrument detections is attained. The theoretical predictions of are only valid for Gaussian chains, described as a set of beads connected separated by length b=0.692 nm from the experimental nm with a theta solvent (trans-decaline at 22 °C from Terao and Mays (2004)). For polystyrene in THF, nm at 25 °C (Terao and Mays, 2004). Neglecting the small temperature differences, we estimate an expansion factor of the gyration radius for linear polydispersed polystyrene in THF at 30 °C as . Since an extension of this method for taking into account the presence of the excluded volume effect has not yet been developed (Costa and Dias, 2007) it was assumed the equality of the expansion factors for branched and linear polymer molecules for obtaining the predictions in Figure 2.52. This assumption has been previously used by several researchers (Dobkowski, 1985) and has not been questioned in more recent works with similar goals. Note that Monte Carlo simulation and molecular dynamics have gone a long way to describe real chains and even branched ones (Steinhauser, 2005), but the complexity of kinetic schemes such as the one here discussed precludes the direct use of the approaches which have successfully tackled simpler structures such as comb and star polymers. A polymer population with large dimensions but at a low concentration is detected by SEC/RI/MALLS close to gel point, as presented in Figures 2.53 and 2.54. Figure 2.55 shows the chromatogram of an STY/DVB sample collected close to gel point. A molecular fraction with strong light scattering signal but weak refractive index response is easily indentified. Figure 2.56 compares chromatograms of samples synthesized at different conditions. In the presence of CTA (run 3) a high polydispersity of molecular sizes is observed. In both case, an inversion in the relation of MW versus elution volume is observed, confirming the erroneous interpretation of chromatograms of branched polymers which would be caused from using a 2-33 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers calibration with linear polymers, as seen in Figure 2.57. Similar observations have been reported in the literature (Bannister et al., 2006). 104 105 106 107 0.0 0.2 0.4 0.6 0.8 0 5 10 15 Weight average molecular weight Feed Rate (mL/min) Time (h) 4 4 5 6 6 53 Figure 2.51. Time evolution of in the presence of CTA or using different feed policies in semibatch reactor. 0 30 60 90 0 5 10 15 z-average mean square radius of gyration (nm) Time (h) 5 4 2 Figure 2.52. Time evolution of in THF solution during the radical copolymerisation of STY/DVB in batch or semi-batch reactor. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-34 0 200 400 600 800 1000 16 18 20 22 24 26 Normalized Response Elution volume (ml) Refractive Index Signal t=0.5 h 3.57 12.2 10.82 7.55 Figure 2.53. Measured refractive index signal in the SEC chromatograms of samples of STY/DVB corresponding to different polymerisation times in a semi-batch reactor. 0.00 0.05 0.10 0.15 0.20 0.25 16 17 18 19 20 21 22 23 24 Absolute Response Elution volume (ml) Light Scattering Signal t=0.5 h 12.2 10.15 7.55 9.2 11.75 11.12 10.82 Figure 2.54. Measured 90° light scattering signal in the SEC chromatograms of samples of STY/DVB corresponding to different polymerisation times in a semi-batch reactor. Note that in the interpretations of chromatograms such as the one presented in Figure 2.55 the lower limit of light scattering detection must be accounted for and a small region at the right 2-41 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers Table 2.11. Basic set of kinetic parameters considered in the modelling of radical copolymerisation of MMA with EGDMA at 60 °C. Kinetic Step Kinetic parameter involved a) Initiator decomposition b)c) (AIBN) (BPO) Methyl methacrylate homopropagation d) Polymer radicals termination e) Chain transfer to solvent f) Chain transfer to agent g) Inhibition h) Reaction radical/retarder i) Chain transfer to monomer j) Chain transfer to initiator k) (BPO);=0 (AIBN) a) All the kinetic parameters are expressed in dm3mol-1s-1, unless otherwise stated. b) For AIBN, was collected from Moad and Solomon, (2006) at p.71 and from Bevington (1955). c) For BPO, was taken to be the value for a benzene solution measured by inhibition with DPPH (Bevington et al., 2003) which should be more accurate than the previously reported (Barson and Bevington, 1997) and is known to be above 0.9 (Moad et al., 1982). d) is a IUPAC benchmark value (Moad and Solomon, 2006) collected from Beuermann et al., 1997. e ) is an average termination constant in the framework of classical kinetics. Scattered values in the range 0.12 to 0.27 for the parameter can be found in the literature (Matheson et al., 1949; Fernández-Garcia et al., 1998; Hutchinson, 2005; Brandrup et al., 2009). was estimated from time versus conversion data (prior to noticeable Norrish-Trommsdorff effect) as measured in the present work. Termination of MMA occurs predominantly by disproportionation ( ) but note that values of from 0.16 to 0.7 are also mentioned in the literature (Moad and Solomon, 2006). f) The parameter was collected from Hutchinson, 2005 at p.168. g) The parameter was collected from Li et al., (1989a) and Li et al., (1989b). h) parameter in the range of the inhibition reactions by some quinones (Moad and Solomon, 2006). i) in the range of the rate constant for the reaction of carbon centered radicals with oxygen. j) The parameter was collected form Hutchinson (2005) at p.169. k) The parameter from Brandrup et al., (1999). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-42 Table 2.12. Propagation rate coefficients ( ) considered in the radical copolymerisation of MMA with EGDMA at 60 °C. Radical/monomer M1 M2 M3 A1 820.0 a) 1223.9 b) 237.7 c) A2 1122.7 b) 1672.8 b) 324.9 c) A3 598.9 d) 895.5 d) 173.9 d) a) IUPAC benchmark value and can be found in Moad and Solomon, 2006 at p.219. b) The rate coefficients , and are based on the reactivity ratios , and and were collected from Li et al., (1989a) and Li et al., (1989b). c) The rate coefficients are based on the reactivity ratio obtained in this work from experimental data and on the ratios as suggested in Li et al., (1989a and 1989b). d) It is considered in this work that a geometric decay relation holds for the reactivity of radical A3 with the different kind of monomers: . These reactivity drops are plausible owing to steric factors. The effect of this assumption is also discussed in the current work. 2.4.2 Results and Discussion In Figure 2.63 are presented the experimentally measured and predicted values of the overall monomer conversion. Those data were used to estimate the parameter considering valid the classical polymerisation kinetics, which was estimated to hold up to monomer conversion . A decrease of owing to a different regimen of diffusion control was observed for higher monomer conversions and this effect was quantified using an empirical correlation of the decrease of the termination rate constant as proposed by Tobita and Hamielec (1989). These solution data do not show the slight influence of the presence of EGDMA on which is nevertheless known to occur at bulk polymerisation when is increased from 0 to 0.01 (Li et al., 1989b). In Figures 2.64 and 2.65 are compared the experimental measurements and theoretical predictions for in the non-linear polymerisation of MMA+EGDMA initiated by AIBN and BPO. Error bars presented in those two Figures for the experimental measurements of this work are those directly indicated by the software (ASTRA) of the SEC/RI/MALLS system. The experimental data on lead to an estimation of the reactivity ratio of PDB, which is . With this apparent reactivity ratio, good agreement between measurements and predictions for runs carried out with different amounts of EGDMA was achieved except when the reaction conditions are favourable to the occurrence of 2-43 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers intramolecular reactions (high monomer conversion with non-linear systems). Is important to note that this reactivity ratio is consistent with a decrease of reactivity of PDB ( ), as reported in other works on this subject (Landin et al., 1988; Li et al., 1989a and 1989b). Further studies in order to clarify the values of PDB reactivities (which actually should be higher than these apparent values) should include experiments at higher dilutions in order to assess the intramolecular cyclization effects (Landin et al., 1988; Dias and Costa, 2005a). 0.0 0.2 0.4 0.6 0.8 1.0 0 3 6 9 12 15 Run 1 (AIBN) Run 2 (AIBN) Runs 3-6 (BPO) Monomer conversion Time (h) Figure 2.63. Time evolution of the measured and predicted overall monomer conversion for different polymerisation runs performed with AIBN and BPO as initiators. Note that the kinetic approach used lumps all isomers with same counts of groups into the same conventional chemical species. Therefore, an accurate consideration of intramolecular reactions is possible only for the smallest sized loops (Dias and Costa, 2005a). It is nevertheless conceivable that models using empirical pseudo-rate constants of cyclization might prove to be useful. They present some mathematical difficulties in their implementation namely when generating functions are introduced and this has delayed their development. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-44 105 106 107 0 3 6 9 Weight average molecular weight Time (h) Run 2 0.460 % EGDMA Run 1 0.310 % EGDMA Figure 2.64. Time evolution of the measured and predicted for polymerisations runs performed with AIBN. 105 106 107 0 3 6 9 12 15 Weight average molecular weight Time (h) Run 3 4 5 6 Figure 2.65. Time evolution of the measured and predicted for polymerisations runs performed with BPO. 2-45 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers The most innovative results of this work are presented in Figures 2.66 and 2.67, where experimental z-average mean-square molecular radius of gyration ( ) of the synthesised copolymers are compared with the predictions of the present kinetic approach (Costa and Dias, 2007). This method is valid for tree-like polymers with Gaussian chains considered as a set of beads connected by massless freely rotating rods. These predictions are therefore valid at Θ conditions. Molecular expansion in good solvents (the excluded volume effect) should be taken into account because the experimental measurements were performed in THF at 30 °C. The length of the Kuhn segment connecting centers of mass repeating units nm was estimated using the relation valid for poly(methyl methacrylate) in a in Θ solvent (Ioan et al., 1995; Búrdalo et al., 2000). Using the published value for poly(methyl methacrylate) in THF, at 30 °C (Búrdalo et al., 2000) we estimate that also in THF at 30 °C for linear polydispersed poly(methyl methacrylate) . An extension of this method for taking into account the presence of the excluded volume effect has not yet been developed (Costa and Dias, 2007) and in this system we have assumed equality of the expansion factors for branched and linear polymer molecules. Several other researchers (Dobkowvski, 1985) have also considered this assumption in more recent works with similar goals, but a more exact estimation of is needed. In spite of these assumptions, a good agreement is often observed between the predictions and the experimental measurements of when intramolecular reactions are less important and therefore some reliable information concerning the molecular architecture of these polymers can be obtained using this approach. The exceptions are, as it could be expected, the high discrepancies between predictions and measurements with runs 1 and 5 where higher monomer conversions at gel point are observed. The complexity of kinetic schemes such as the one here discussed precludes the direct use of other approaches (such as Monte Carlo simulation or molecular dynamics) which should be improved order to efficiently describe real linear or branched chains (Steinhauser, 2005). A major advantage of this general kinetic approach is the possibility of considering more complex kinetic schemes whenever desired. In the present case study, radicals from pendant double bonds (A3) were distinguished from radicals of MMA or EGDMA because it is plausible that they present different reactivities (see Table 2.10). Figure 2.68 shows the Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-46 predicted effect of the reactivity of RPDB on the time evolution of , using run 2 as an example. 0 20 40 60 80 100 120 0 3 6 9 _ Rg (nm) Time (h) Run 2 0.460 % EGDMA Run 1 0.310 % EGDMA Figure 2.66. Predicted and measured time evolution of in the copolymerisation system MMA/EGDMA initiated by AIBN. 0 20 40 60 80 100 120 0 3 6 9 12 15 _ Rg (nm) Time (h) Run 3 4 5 6 Figure 2.67. Predicted and measured time evolution of in the copolymerisation system MMA/EGDMA initiated by BPO. 2-47 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers Different values of the reactivities of RPDB were considered in the simulations using the values ( ) present in Table 2.11 as a reference. It can be observed that these parameters can have a major effect on the predictions only when these radicals are reacting much more slowly ( ) than the reference situation we have considered. However, this is a plausible situation due to lower mobility of these radicals and therefore the estimation apparent reactivity ratios of PDB can also be affected by this phenomenon. The reliability of the predictions of the present kinetic approach was also assessed using experimental data previously obtained for bulk polymerisation of MMA/EGDMA (Li et al., 1989a and 1989b). Considering the polymerisation conditions and the kinetic parameters used in these works, our predictions of the monomer conversion at gel point were compared with the experimental values observed in the presence of CBr4 as CTA. This comparison is presented in Figure 2.69 and the good agreement between the predictions and measurements confirms the correctness of the foundations of this method. Figure 2.70 shows the relation between measured average radius of gyration and average molecular weight for linear MMA and different non-linear samples of MMA/EGDMA synthesized in this case. The well known decrease of the size of non-linear polymers when compared with the linear analogous is here confirmed. Nevertheless, it is important to recall that these polymer samples are highly dispersed in molecular mass and the correct comparison is made using as a reference instead of . Indeed, for a population of linear polymer molecules a power relation (with exponent 0.5 for Gaussian chains) between and should be observed (Dias and Costa, 2007). Note in the same Figure the high estimated errors in the measurements of by SEC/RI/MALLS (only can be directly measured) when highly dispersed polymer samples are analysed, as previously presented in the system styrene/divinylbenzene (Gonçalves et al., 2007). Figure 2.71 shows the build-up of a polymer population with large dimensions but at low concentration (forming a kind of cluster) as the gel point is approached. The chromatogram presented in Figure 2.72 of a MMA/EGDMA sample collected close to gel point shows in detail a molecular fraction with a strong light scattering signal but a weak refractive index. Figure 2.73 shows an inversion in the relation molecular mass versus elution volume of two highly crosslinked samples of MMA/EGDMA confirming again the existence in the samples of polymer species with same molecular weight but very different molecular sizes. Similar observations have also been recently reported in the literature (Bannister et al., 2006). Figure 2.74 presents the time evolution of the chromatograms of MMA/EGDMA Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-48 samples showing the formation of a long end tail as the gel point is approached, in agreement with previously theoretical results (Dias and Costa, 2005b). 105 106 107 0 1 2 3 4 5 6 7 Weight average molecular weight Time (h) Run 2 0.460% EGDMA CR=0.005 0.01 0.1 CR=1 and 200 Figure 2.68. The predicted effect of the reactivity of radicals of pendant double bonds on the time evolution of . 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Predicted Gel Conversion Experimental Gel Conversion 2% 1% 0.5% 0.3% EGDMA 1% CTA Figure 2.69. Predicted and measured gel conversion for the system MMA/EGDMA in the presence of CTA. Experimental data collect from Li et al. (1989a and 1989b) and predictions from the present kinetic model. 2-49 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers 101 102 105106107 z-average radius of gyration (nm) Average Molecular Weight Linear Cross-Linked z - average Weight average Figure 2.70. Relation between average radius of gyration and average molecular weight for linear MMA and several non-linear samples of MMA/EGDMA. 0.00 0.02 0.04 0.06 0.08 0.10 16 17 18 19 20 21 22 23 24 Absolute response Elution volume (ml) Light Scattering Signal t=0.5 h 1.5 2.98 4.27 3.98 Run 6 Figure 2.71. Observed 90° light scattering signal in the SEC chromatograms of samples of MMA/EGDMA corresponding to different polymerisation times. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-50 0.0 0.2 0.4 0.6 0.8 1.0 104 105 106 107 108 109 16 17 18 19 20 21 22 23 24 Relative response (RI and LS) Molecular weight (MW) Elution volume (ml) MW RI LS Run 6 (4.27 h) Figure 2.72. Molecular weight along the SEC chromatogram for a sample of MMA/EGDMA. 104 105 106 107 108 16 17 18 19 20 21 22 23 24 Molecular weight (MW) Elution volume (ml) Linear PMMA Crosslinked MMA/EGDMA Polymers Figure 2.73. Observed relations molecular weight versus elution volume for MMA/EGDMA copolymers. 2-57 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers instance in the in-line monitoring of butyl acrylates/vinyl acetate emulsion copolymerisation (Jovanović and Dubé, 2003). With emulsion processes, this phenomenon is due to unstable polymerisation conditions which seem also to occur at the gel point for the suspension crosslinking polymerisation performed in the present chapter. For the operation conditions used in the present chapter, gelation occurs at low monomer conversion (around 10 % with as estimated by SEC/RI/MALLS) and by consequence “catastrophic coagulation” precluded the obtainment by in-line FTIR-ATR of reliable quantitative information concerning the reactivity of the different monomers/pendant double bonds involved. Besides this clogging phenomenon, it is possible that the spectra observed with inline FTIR-ATR can also be a result of coating of the ATR crystal during the polymerisations. Further experiments with different operating conditions (e.g changing the kind of stabilizer, its concentration and agitation speed) must be performed to elucidate this issue with suspension polymerisation. On other hand, in-line FTIR-ATR monitoring of emulsion polymerisation of vinyl monomers is reported is previous research works without apparent occurrence of coating of ATR crystal (Chatzi et al., 1997; Storey et al., 1998; Hua and Dubé, 2001; Jovanović and Dubé, 2001; Hua and Dubé, 2002; Ouzineb et al., 2003; Jovanović and Dubé, 2003). Figure 2.79. FTIR-ATR spectra observed in the in-line monitoring of the suspension copolymerisation of STY/DVB using n-heptane and toluene as diluents of the organic phase. 0 2000 4000 6000 8000 10000 Seconds 0 .05 .1 .15 .2 Absorbance 1800 1700 1600 1500 1400 1300 1200 1100 1000 Wavenumber (cm-1) Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 2-58 However, given its different initiation mechanism and likely occurrence of space separation of radicals, emulsion polymerisation is completely different from the suspension polymerisation operation here studied and major modelling changes would have to be discussed. Use of confocal Raman spectroscopy (with a higher penetration depth) should overcome the coating of ATR crystal in both circumstances (suspension/emulsion), but this requires major changes on the experimental set-up here used. 2.6 Conclusions The conventional radical copolymerisation (FRP) of monoand divinyl monomers has been experimentally studied in batch and semi-batch reactors using STY/DVB as model system. It was shown that the production of hyperbranched polymers with an improved control of molecular architecture can be carried out by adjusting the feed policy of the divinyl monomer (semi-batch operation). These results can be especially useful to obtain soluble branched polymers at higher conversions than with a batch operation. Experimental results here presented concerning the semi-batch solution polymerisation of styrene and divinylbenzene showed the possibility of control of gelation through the design of appropriated feeding policies. Nevertheless, some shortcomings of this approach should be here stressed. Lack of reproducibility of the polymerisations was found during this research. This problem is probably a consequence of non-constant feeding rates due to the peristaltic pumps used. Note that even a small error in the feeding flow rates has a high effect on crosslinking process, especially when divinyl monomer is to be feed to the polymerisation mixture. In principle, this shortcoming can be avoided using higher precision pumps (e.g. syringe pumps/GPC pumps). Semi-batch STY/DVB here studied cannot also be extended to the production of polymer particles in a process similar to suspension polymerisation (process later on explored in this research). Emulsion semi-batch polymerisation of vinyl/multivinyl monomers is an alternative to the production of these kinds of particles but much more complex mechanistic issues are expected in the framework of such heterogeneous processes. Inter-phase transport of both monomers (vinyl/multivinyl) and radical compartmentalization are some phenomenon leading to the need of development of a much more complex analysis when emulsion semi-batch crosslinking polymerisation is considered. 2-59 CHAPTER 2. Conventional Free-Radical Copolymerisation of Multivinyl Monomers For the produced STY/DVB and MMA/EGDMA hyperbranched copolymers their molecular architecture was also assessed through the used SEC/RI/MALLS system. It was shown that a complicated interpretation of SEC chromatograms of branched polymers arises due to the possible change in the relation between molecular weight and elution time (or hydrodynamic radius), owing to the existence of copolymer chains with the same molecular weight but quite different molecular sizes (elution volumes). Suspension copolymerisation of styrene/divinylbenzene with gel formation was also experimentally studied by performing a set of different runs in a batch reactor. The dynamics of product molecular properties was measured by SEC/RI/MALLS. These reactions were inline monitored by FTIR-ATR. The produced gel beads were analysed by scanning electron microscopy (SEM) and the impact on the morphology of these materials of parameters such as the proportion of thermodynamically “bad/good” solvents in the diluent was studied. The dynamics of the gel fraction in batch reactor was followed and the swelling ratio of the resulting gel beads was also quantified. SEM characterization of these gel beads confirmed the formation of macroporous structures if appropriated synthesis conditions are used, namely concerning the thermodynamic affinity of the diluent mixture. In the present work, different proportions of n-heptane/toluene (bad/good solvents) were considered for this purpose. Measured swelling ratios of these gel beads are also in agreement with previous works and a maximum swelling ratio of 11 was here observed. In-line FTIR-ATR monitoring of the crosslinking process here performed has shown the occurrence of “catastrophic coagulation” at the gel point thus precluding the intended in-line measurement of the building parameters of the polymer networks. A new design of operating conditions should be carried out in order to avoid this phenomenon so that this spectroscopic technique can be fully exploited. Further experimental studies with more emphasis on describing cyclization effects should be undertaken in order to clarify these issues. CHAPTER 3 ATOM TRANSFER RADICAL POLYMERISATION OF ACRYLATES Abstract. This chapter reports experimental and modelling studies concerning the conventional (FRP) and atom transfer radical polymerisation (ATRP) of hyperbranched acrylates and methacrylates. A set of experiments was performed in solution polymerisation in a batch reactor using n-butyl acrylate, methyl acrylate and methyl methacrylate as monomers, 1,6hexanediol diacrylate, bisphenol A ethoxylate diacrylate and ethylene glycol dimethacrylate as crosslinkers. Some variables have been changed and their effect is assessed along the experimental program:  Parameters changed:  Polymerisation temperature.  Monomer dilution.  Kind/amount of crosslinker used.  ATRP mediation system.  Products characterization in terms of:  Monomer conversion through GPC and gravimetry.  Molecular architecture of the products through SEC/RI/MALLS allowing the determination of average molecular weights, z-average radius of gyration and absolute molecular weight distribution.  In-line FTIR-ATR to observe polymer formation and quantitatively analyse the reactivity of PDB and the formation of intramolecular cyclizations.  This experimental analysis is complemented with modelling studies including branching and crosslinking in the absence of cyclization. This chapter is based on the following publications: M.A.D. Gonçalves, R.C.S. Dias, M.R.P.F.N. Costa, Macromol. Symp. 289 (2010) 1-17. M.A.D. Gonçalves, V.D. Pinto, R.C.S. Dias, M.R.P.F.N. Costa, Macromol. Symp. 296 (2010) 210228. M.A.D. Gonçalves, I.V.R Trigo, R.C.S. Dias, M.R.P.F.N. Costa, Macromol. Symp. 291-292 (2010) 239-250 M.A.D. Gonçalves, R.C.S. Dias, M.R.P.F.N. Costa, Chem. Eng. Technol. 33 (2010) 1797-1813 3-3 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates 3.1 Introduction Non-linear radical polymerisations are used to produce soluble and insoluble crosslinked materials with important applications in several domains, such as biomedicine, pharmaceutics, biotechnology, environment and microelectronics. Conventional free radical polymerisation (FRP) leads to microgels (very high molecular weight soluble non-linear polymers) and gels (insoluble materials) with inhomogeneous structures due to the combination of slow initiation, fast propagation and termination. Formation of loops due to intramolecular cyclization reactions is another mechanism with negative impact in the properties of FRP synthesized non-linear polymers (Matsumoto, 1995). Controlled radical polymerisation (CRP) has been recently exploited to increase the homogeneity and therefore to obtain products with higher performances in their end-use applications. The use of Atom Transfer Radical Polymerisation (ATRP) to obtain acrylate/diacrylate copolymers is an example of such efforts (Yu et al., 2007; Gao et al., 2007 and 2008). In Figure 3.1 is depicted the ATRP main equilibrium reaction of a polymer radical from n-butyl acrylate mediated by CuBr using PMDETA or HMTETA as ligand. Figure 3.1. Schematic of an ATRP equilibrium reaction of a polymer radical from n-butyl acrylate. Nowadays, there are still some important open issues in this field, namely concerning the impact of intramolecular cyclizations/unequal functional group reactivity in the structure of these materials. The design of new operation conditions to manipulate the molecular architecture (e.g. operation in semi-batch reactor (Gonçalves et al., 2007)) or the control of gelation and/or post gel properties are other important aspects concerning the polymer reaction engineering of these kinds of polymerisation systems. This work reports experimental and theoretical studies concerning the FRP and ATRP production of acrylate/diacrylate microgels. Some important features of the molecular architecture of these materials are investigated and differences between the two kinds of polymerisation systems are studied. The impact of the synthesis conditions on the structure of the products is assessed in order to develop tools for the design of materials with improved end use properties. Hyperbranched polyacrylates (only soluble polymers will be here Br Ligand-1 Cu+2 Br-1 CH3 CH3 Ligand-1 CH2 Cu+1 deactivation CH2 H H (CH2)3OCO (CH2)3OCO activation Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-4 discussed) were synthesized in a laboratory batch reactor and the products were characterized by size exclusion chromatography with detection of refractive index and multi-angle laser light scattering (SEC/RI/MALLS) signals. Chemical properties describing the molecular architecture of these materials were measured, namely molecular weights and z-average radius of gyration. The influence of the synthesis technique (ATRP versus FRP) on the structure of these materials was also investigated. The FTIR-ATR in-line monitoring of these non-linear copolymerisations was also performed in order to assess the ability of this technique to provide information about specific features of the formation mechanisms and structure of hyperbranched polyacrylates (e.g. reactivity of pendent double bonds and cyclizations). FTIR-ATR in-line monitoring also has an important potential use for the specification of feed programs with semi-batch reactors which have impact on the molecular architecture of the produced hyperbranched polymers, as recently shown for the system styrene/divinylbenzene (Gonçalves et al., 2007). 3.2 Experimental Details 3.2.1 Materials In ATRP experiments of acrylates/diacrylates and MMA/EGDMA, N,N-dimethylformamide (DMF) of 99.8 % purity, ethyl 2-bromopropionate (EBrP) of 99 % purity, Cu(I)Br of 98 % purity, N,N,N′,N″,N″-pentamethyldiethylenetriamine (PMDETA) of 99 % purity, n-butyl acrylate (nBuA) of 99 % purity stabilized with 10 to 55 ppm monomethyl ether hydroquinone (MEHQ), methyl acrylate (MA) of 99 % purity stabilized with 100 ppm MEHQ, 1,6Hexanediol diacrylate (HDDA) of 80 % purity stabilized with 100 ppm MEHQ, bisphenol A ethoxylate diacrylate (BEDA) with 688 and 99 % purity stabilized with 250 ppm MEHQ, anisole of 99 % purity, methyl α-bromophenylacetate (MBPA) of 97 % purity, 1,1,4,7,10,10-hexamethyltriethylenetetramine (HMTETA) of 97 % purity methyl methacrylate of 99 % purity stabilized with 10 to 100 ppm monomethyl ether hydroquinone and ethylene glycol dimethacrylate (EGDMA) of 98 % purity stabilized with 100 ppm monomethyl ether hydroquinone were purchased from Sigma Aldrich and used as received. In FRP experiments the same monomers and crosslinkers were used and AIBN of 98 % purity and toluene of 99.7 % purity were also purchased from Sigma Aldrich and used as received. Monomers were used as received to mimic the industrial practice, and the presence of inhibition/retardation of polymerisation is taken into account in kinetic modelling. 3-5 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates 3.2.2 Polymerisation Set-up All experiments were carried out using the experimental set-up described in chapter 2. In ATRP experiments with acrylates/diacrylates (see Table 3.1), DMF, acrylate monomer, diacrylate monomer, PMDETA and CuBr were premixed at 60 °C for at least 30 min in a volumetric flask. Good solubility of CuBr in the polymerisation system was observed. That mixture was afterwards charged to the reactor, which had previously been purged with argon at a flow rate 40 cm3/min, and brought up to the polymerisation temperature (60 °C). When the temperature set-point was attained, the bubbling process was maintained for one hour before initiation (as well as for the whole polymerisation). Then, the initiator (EBrP) was added to the system defining t = 0. At prescribed polymerisation times, samples of polymer were withdrawn from the reactor and analysed by SEC/RI/MALLS. In these set of experiments the monomer concentration is expressed in mol/dm3. Similar procedures were performed in FRP experiments (see Table 3.2) with exception of the pre-mixing period which is absent. In these experiments the monomer and initiator concentrations are expressed in mol/dm3. Table 3.1. Description of a set of experiments performed in the study of the ATRP copolymerisation of acrylate/diacrylate monomers at 60 °C. Run M [M]0(mol/L) CL M/EBrP/CuBr/PMDETA VM(%) 1 MA 2.44 0 50/1/0.45/0.5 22 2 nBuA 2.44 0 50/1/0.45/0.5 35 3 nBuA 2.41 HDDA 2 50/1/0.45/0.5 35 4 nBuA 2.28 HDDA 10 50/1/0.45/0.5 35 5 nBuA 2.18 HDDA 16.8 50/1/0.45/0.5 35 6 nBuA 2.24 BEDA 5 50/1/0.45/0.5 35 7 nBuA 1.64 BEDA 5 50/1/0.45/0.5 25 8 nBuA 1.04 BEDA 5 50/1/0.45/0.5 15 9 nBuA 2.26 BEDA 5 200/1/0.45/0.5 35 In ATRP experiments (see Table 3.3) with MMA/EGDMA, the solvents, monomers, CuBr and HMTETA were premixed at 60 °C for at least 30 min in a volumetric flask. Solubility problems with copper specie were avoided by using anisole (Xia and Matyjaszewski, 1997) plus DMF (only a small amount is needed) as a co-solvent (Pascual et al., 1999). That mixture was afterwards charged to the reactor, which had previously been purged with argon at a flow Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-12 of (-CH(R)-) groups (HTP) while (-CH3) groups (CTP) are introduced in the polymer chains only via the propagation of nBuA. Radicals derived from HTP and CTP are also distinguished due to their expected different reactivity and named R6 and R7, respectively. These radicals can also undergo propagation reactions with monomers/macromonomers as depicted in Eq. (3.5). The main distinctive features of ATRP compared to conventional radical polymerisation are due to activation (Ri)/deactivation(Di) equilibrium of all kinds of polymer radicals described by Eq. (3.6). The same activation/deactivation equilibrium is considered to hold for the primary radicals, as described by Eq. (3.7). Although they transport polymer structures such as terminal double bonds or polymerisation repeating units, radicals derived by chain transfer to monomers are here named primary radicals due to their small size. The formation of three-dimensional branched architectures (BS) is possible due to reactions consisting of the intermolecular chain transfer to polymers involving the 7 different kinds of polymer radicals and the 2 kinds of chain transfer to polymer sites (HTP and CTP), as represented in Eqs. (3.8) and (3.9). Saturated chains ends (SG) are formed when monomer/macromonomer derived growing radicals are involved and HTP and CTP are reformed with the correspondent growing radicals undergo such process. Chain transfer to monomers and chain transfer to solvent reactions are described by Eqs. (3.10-3.12). Besides the formation of the aforementioned kinds of primary radicals, saturation of chain ends (SG) results when monomer/macromonomers polymer radicals are involved while reformation of HTP and CTP is expected when R6 and R7 undergo chain transfer. Terminal double bonds (M4) are formed in mechanism involving chain transfer to monomers. Termination by combination reactions described by Eq. (3.13) promotes the formation of HTP centres when monomer/macromonomers polymer radicals are involved or else they merely connect 2 polymer molecules. Termination by disproportionation, depicted by Eq. (3.14), leads to the saturation (SG) or the formation of HTP/CTP centres in one chain end. Conversely, in the other chain end, reactive terminal double bonds (M5) or non-reactive internal double bonds (IDB) are formed. Reversible activation/deactivation of initiator: (3.1) 3-13 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates Initiation of monomers, PDBs and TDBs by primary radicals of initiator: (3.2) Initiation of monomers, PDBs and TDBs by primary radicals derived by transfer to monomers and to solvent: (3.3) Propagations involving growing radicals of monomers, PDBs and TDBs: (3.4) Propagations involving radicals of HTP and CTP: ; (3.5) Reversible activation/deactivation of growing radicals: (3.6) Reversible activation/deactivation of primary radicals: (3.7) Chain transfer to HTP polymer centre: (3.8) Chain transfer to CTP polymer centre: (3.9) Chain transfer to an acrylate monomer: (3.10) Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-14 Chain transfer to a diacrylate monomer: (3.11) Chain transfer to solvent: (3.12) Termination by combination: (3.13) Termination by disproportionation: (3.14) Modelling studies of FRP and ATRP acrylate/diacrylate were carried out considering basic sets of rate coefficients collected from previous works (see also Table 3.8)), namely for initiator thermal decomposition (Moad and Solomon, 2006), nBuA propagation (Asúa et al., 2004), chain transfer to monomer (Maeder and Gilbert, 1998), chain transfer to solvent (Moad and Solomon, 2006; McKenna et al., 1999), intramolecular chain transfer to polymer (Brandrup et al., 1999), termination (McKenna et al., 1999; Fernández-Garcia et al., 2004). Rate coefficients for reversible kinetic steps of ATRP, namely, radical activation, radical deactivation or equilibrium constant activation/deactivation, measured/estimated in previous works (Tang and Matyjaszewsky, 2007; Matyjaszewski et al., 2001; Tang et al., 2006), were also here considered. Sensitivity analysis of the branching and crosslinking processes to kinetic parameters was performed by considering perturbations of the rate coefficients governing different reaction steps, namely: propagation of pendent double bonds, propagation of the different kinds of terminal double bonds, intermolecular chain transfer to different polymer centres, reactivity of the different kinds of polymer radicals and probabilities of radical termination by combination/disproportionation. Under the experimental conditions here used, simulation results show a particular sensitivity to the reactivity of pendent double bonds of diacrylate monomer and therefore only this parameter was used in the fitting of experimental results, as discussed below. 3-15 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates Table 3.8. Basic set of rate coefficients considered in the modelling studies of the FRP and ATRP copolymerisation of acrylate/diacrylate monomers. Kinetic Step Rate Coefficient Expression a) Initiator thermal decomposition nBuA propagation Initiation by primary radicals b) Chain transfer to monomer Chain transfer to solvent Intermolecular chain transfer to polymer Radical termination Combination/disproportionation ATRP equilibrium constant Radical deactivation Propagation of terminal double bonds b) Propagation of pendent double bonds b) a) Kinetic parameters expressed in dm3 mol-1s-1, unless otherwise stated. R= 8.314 Jmol-1K-1. b) This work. Intramolecular chain transfer to polymer (backbiting) is also an important source of short chain branching (SCB) in acrylates polymerisation. It has been studied in several recent works, namely by Nikitin, Hutchinson and co-workers (Nikitin and Hutchinson, 2005 and 2006; Nikitin et al., 2009) aiming to obtain kinetic data and theoretical models of this phenomenon (Plessis et al., 2000a; Nikitin et al., 2007; Wang and Hutchinson, 2008; Willemse et al., 2005; Barner-Kowollik et al., 2008, Buback et al., 2008a). This mechanism involves the formation of a mid-chain radical due to the intramolecular abstraction of an Hatom from a neighbour methine group. Additional propagation of this radical leads to the occurrence of SCB in the main polymer chain. These mid-chain radicals can also undergo β-scission reactions which either reform terminal radicals belonging to short polymer chains or generate terminal double bonds later Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-16 producing LCB. Propagation, termination and chain transfer reactions involving these species were theoretically and experimentally analysed and evidence for branching in nBuA polymerisation up to high conversions, due to intramolecular chain transfer to polymer, were thus obtained (Nikitin et al., 2009). Nowadays it is consensual in the scientific community that mid-chain radicals formed in acrylates polymerisation strongly differ from terminal radicals, namely due to their slow propagation rate. A more complex kinetics is therefore expected with polymerisation conditions where these effects are dominant. Branching mechanisms (intramolecular/intermolecular) in acrylates polymerisation should be especially important at high temperatures (e.g. in the range 75 to 100 °C) and at high polymer concentration (e.g. emulsion/suspension polymerisation at high monomer conversion) leading eventually to gelation (Plessis et al., 2000b; González et al., 2006). Simulations here performed, considering the experimentally used polymerisation conditions show that the contribution of intermolecular chain transfer to polymer to the formation of non-linear connections is negligible as compared with the main crosslinking process. Therefore, it is plausible that intramolecular chain transfer to polymer can be neglected under the experimental conditions (rather low temperature) considered in the present research and thus this mechanism was not included in the modelling studies. Past experimental/theoretical works using the same simulation method (Gonçalves et al., 2007; Trigo et al., 2008) considered the effect of inhibition/retardation reactions in FRP and CRP polymerisation process. Kinetic models accounting for the presence of an inhibitor and/or a retarder in the polymerisation system were developed and it was concluded that, for those experimental conditions, the effect of inhibition on average molecular weights should only be noticeable above around 500 ppm. Results thus obtained showed that this possible issue was avoided with the monomers and experimental set-up used. Transient behaviour studies (before and after AIBN addition) due to thermal initiation of acrylates concerning the importance of the induction period were obtained with the present chemical system. Note that the high proportion initiator/monomer used (namely in ATRP experiments) and relatively low polymerisation temperature makes this issue negligible. However, this issue is relevant in the case of FRP and NMRP of styrene at high temperatures (130 °C) as we will see in chapter 4. 3-17 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates 3.3.2 ATRP of MMA/EGDMA 3.3.2.1 Chemical Species and Chemical Reactions In the ATRP of MMA/EGDMA modelling studies are used 22 chemical groups. 15 active species and 7 non-active groups are taking into account in order to refine the description of the non-linear copolymer structure in the present analysis. The equations (3.15) to (3.23) describe the kinetic scheme considered in the ATRP of MMA/EGMA comprising a total of 44 chemical reactions. Once more, each reversible process contributes two different elementary steps. In these equations, M1 and M2 represent the two monomers (MMA and EGDMA), RX the initiator (MBPA), R the radical from the initiator, CX the deactivator (CuBr2/HMTETA) and CM the complex transition metal/ligand (CuBr/HMTETA). On the other hand, T represents the solvent and RT the radical from the solvent. In this instance, only one macromonomer is considered, which corresponds to the PDB resulting from EGDMA initiation/propagation (M3). Three different growing radicals (Rk) resulting from the polymerisation of monomers/macromonomers are distinguished and the correspondent deactivated chain ends are named Dk. therefore this are the 15 active species presents in this system. The seven inactive species included in the present analysis are MMA and EGDMA repeating units (U1 and U2, respectively), crosslinking sites resulting from the polymerisation of PDB (CS), polymerized fragments from initiator and solvent (F1 and F2, respectively), saturated terminal units (SG) and head-head units from termination by combination reactions (HHU). In this set of 44 chemical reactions depicted in Eqs. (3.153.23), the following main classes are present: 2 reactions of activation/deactivation of initiator, 6 of initiations, 9 propagations, 6 activation/deactivation of polymer radicals, 9 chain transfers and 12 terminations. Reversible activation/deactivation of initiator: (3.15) Initiation of monomers and PDBs by primary radicals of initiator: (3.16) Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-18 Initiation of monomers and PDBs by primary radicals of solvent: (3.17) Propagations involving growing radicals of monomers and PDBs: (3.18) Reversible activation/deactivation of growing radicals: (3.19) Chain transfer to monomers: (3.20) Chain transfer to solvent: (3.21) Termination by combination: (3.22) Termination by disproportionation: (3.23) Table 3.9 presents a basic set of rate parameters considered in the modelling of the ATRP of MMA/EGDMA. All these parameters have been collected from previous works concerning mostly the linear FRP and ATRP of methacrylate monomers. Initiator and radical activation/deactivation and initiation of MMA with primary radicals were collected from Ohno et al. (1998); Tang et al. (2008); Shipp and Matyjaszewski, (2000). The propagation constant of MMA have been taken from Beuermann et al., (1997). Rate coefficients for chain transfer to monomer and to solvent were collected from Hutchinson, (2005). Radical termination constants were retired from (Fernández-Garcia et al., (1998); Moad and Solomon (2006). The extension to the non-linear case studies here considered is not straightforward because of the large number of kinetic parameters involved (notice the number of radicals and double bonds with different reactivities) and little information can be found in the literature concerning the kinetic of the ATRP of similar systems. The calculations presented here are 3-19 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates based on some postulated reactivity ratios of the different kinds of radicals and double bonds which have been previously discussed in the framework of FRP of MMA/EGDMA (Trigo et al., 2008 and references therein). Nevertheless, it should be pointed out that the major parameters governing the crosslinking process are the reactivities of the pendant double bonds and these parameters will be estimated here using the available experimental data. These estimates have been obtained by neglecting intramolecular cyclizations. However, this can be an important issue even at 50 % dilution, as it will be discussed below. Particularities of the activation/deactivation equilibrium are also addressed in this chapter. Table 3.9. Basic set of rate parameters considered in the modelling studies of the ATRP of MMA/EGDMA. Kinetic Step Rate Coefficient Expression a) Initiator and radicals activation (EBPA at 22 °C in Acetonitrile) (PS-Br at 110 °C in BPH) Initiator and radicals deactivation MMA propagation Initiator of MMA by primary radicals , Chain transfer to monomer Chain transfer to solvent Radical termination a) Kinetic parameters expressed in dm3 mol-1s-1, unless otherwise stated. R= 8.314 Jmol-1K-1. 3.3.2.2 Mathematical Modelling Including Branching and Crosslinking It is known (Hutchinson and Penlidis, 2007) that acrylate monomers (among others) are susceptible to undergo chain transfer to monomer reactions because they contain aliphatic tertiary hydrogens which can be involved in an H-atom abstraction, as presented in Figure 3.2. A new polymer chain created by the resulting radical will contain an unsaturated end group (terminal double bond) capable of undergoing a further reaction with concomitant formation of non-linear structures via Long Chain Branching (LCB). Other sources of LCB in acrylates polymerisation are intermolecular chain transfer to polymer mechanisms. These Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 3-20 reactions usually involve the H-atom abstraction of a methine group (-CH(R)-) with formation of a mid-chain radical as presented in Figure 3.3 and/or the H-atom abstraction in a -CH3group, as depicted in Figure 3.4. Intermolecular chain transfer to polymer centres (-CH(R)- and-CH3-) are created via the incorporation of monomer units in polymer chains (propagation reactions). LCB in acrylates via intermolecular chain transfer to polymer is usually considered to be result of H-atom abstraction of a methine group (Figure 3.3). For the sake of generality, here both mechanisms will be considered. Termination by combination and termination by disproportionation are also a source of chain transfer to polymer centres (methine groups) and terminal double bonds, respectively. This means that termination mechanisms can also have an indirect contribution for LCB. NMR spectroscopy has been applied to quantify the incidence of the different long chain branching mechanisms (Britton et al., 1998; Ahmad et al., 1998) in different polymerisation systems, namely involving nBuA and vinyl acetate. Gelation due to simultaneous intermolecular chain transfer to polymer and termination by combination was also experimentally observed and theoretically studied in a several works, namely concerning emulsion polymerisation of nBuA (Arzamendi et al., 1994; Plessis et al., 2000a and 2001; González et al., 2006 and 2007). The importance of intermolecular chain transfer to polymer was also recently shown through the synthesis of Z-RAFT star acrylate polymers (Boschmann and Vana, 2007). Besides LCB, non-linear structures are formed in acrylate/diacrylate copolymerisation due to the polymerisation of pendent double bonds, as depicted in Figure 3.5. This crosslinking process is intentionally promoted through the addition of the diacrylate monomer to the polymerisation system. The amount of divinyl monomer is used in practice to manipulate the extent of the crosslinking process. Reactivity of pendent double bonds and/or incidence of intramolecular cyclization reactions are parameters with a strong influence in the efficiency of crosslinking polymerisations. C CH3 CH CH3 CH2 CH3 CH2 CH2 H CH2 H H (CH2)3OCO H (CH2)3OCO (CH2)3OCO (CH2)3OCO . CH2 Figure 3.2. Depiction of a chain transfer reaction involving an acrylate monomer (n-butyl acrylate used for illustration) with creation of a terminal double bond that can lead to long chain branching. 3-21 CHAPTER 3. Atom Transfer Radical Polymerisation of Acrylates Figure 3.3. Depiction of intermolecular chain transfer to polymer in acrylates due to the H-atom abstraction of methine hydrogen. Figure 3.4. Depiction of intermolecular chain transfer to polymer in acrylates due to the H-atom abstraction of a –CH3 group. Figure 3.5. Depiction of pendant double bonds propagation (crosslinking) in acrylate/diacrylate copolymerisation. Note that even if it is desirable that as many as possible of these kinetic parameters can be measured from experimental data, or indirectly estimated by more or less fundamental models of chemical reactions (an area which is expected to yield more and more accurate predictions in the not so distant future), only a few of them need be accurately known for the range of initial compositions and temperatures in our experiments and the actual number of fitted parameters is very small as discussed in the next section. The present modelling approach is able to accommodate mechanisms such as intramolecular chain transfer to polymer (backbiting) leading to SCB. Much more complex kinetic schemes must be considered in these circumstances, as previously shown for a case study including primary cyclizations (Dias and Costa, 2005a). Results of the present work show that, under the polymerisation conditions here considered (diluted solution polymerisation at relatively low monomer conversion), crosslinking is dominant in comparison to intermolecular chain C CH2 CH3 CH2 CH2 CH3 CH2 CH CH3 CH3 CH2 H CH2 H (CH2)3OCO H (CH2)3OCO (CH2)3OCO (CH2)3OCO . CH2 C CH H C CH2 CH3 CH3 CH2 CH2 (CH2)3OCO CH2 CH2 CH3 H H CH2 H (CH2)3OCO (CH2)3OCO (CH2)3OCO CH CH2 . CH CHCH2 CH3 (CH2)3OCO CH2 C C OCO CH2 H (CH2)6OCO CH3 CH2 OCO H CH2 H (CH2)3OCO (CH2)6OCO Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 4-58 appropriate for obtaining information concerning the consumption of C=C bonds and the presence of pendant double bonds in the networks. Note that the commercial crosslinker used in this research contains two isomers (mand pdivinylbenzene), and also ethylvinylbenzene. The IR spectrum corresponding to this mixture is much more difficult to analyse, particularly when compared with the spectrum of the purified isomers. In fact, for fundamental studies, the use of purified isomers of the divinylbenzene crosslinker would be advantageous, as reported before for the FRP of styrene/divinylbenzene (Hecker, 2000). Figure 4.55. FTIR spectra of the styrene and divinylbenzene monomers. Assignments corresponding to C=C bonds in the monomers cab ne identified at around 992, 1019, 1410, 1452 and 1630 cm-1. Peaks at 992 and 1630 cm−1 are especially useful in this context due to their strong intensities and the low interference of other chemical groups in this region of the spectra. Figure 4.56 shows the FTIR spectrum observed for the isolated NMRP networks produced in run 9 where 50 % DVB was present in the initial monomer mixture. The peaks at 992 and 1630 cm−1 can be clearly observed in the spectrum of these products, showing the effective C=C functionalization (high concentration of pendant double bonds in the network) of the final materials obtained in this run. Figures 4.57 and 4.58 present the FTIR spectra observed for the isolated NMRP networks produced in runs 8 and 10, respectively. Note that in run 8 only DVB was used, whereas in run 10 the initial monomer mixture composition was 95/5 styrene/divinylbenzene. Strong IR absorbances can be observed at 992 and 1630 cm−1 for the 4-59 CHAPTER 4. Nitroxide Mediated Radical Polymerisation of Styrene/Divinylbenzene products of run 8 (see Figure 4.57), identifying a high concentration of pendant double bonds in these networks. Conversely, weak signals of these functional groups are observed in the spectrum corresponding to the products of run 10 (see Figure 4.58), showing (as expected) a lower presence of PDB in these networks. The comparison of results presented in Figures 4.56, 4.57 and 4.58 illustrates the usefulness of FTIR analysis of the isolated networks in order to obtain information concerning the presence of PDB in the synthesized materials. The qualitative effect of the synthesis conditions (initial composition, temperature, etc.) on the process of double bond incorporation in the networks is thus readily obtained through the comparison of the IR spectra of the different products. Figure 4.56. FTIR spectrum observed for the isolated NMRP network produced in run 9 (50 % DVB in the initial monomer mixture). Peaks at 992 and 1630 cm−1 can be clearly observed in the spectrum of the product, showing the C=C functionalization of the materials. In principle, FTIR analysis can also be used to quantify the concentration of PDBs in the synthesized products (Hecker, 2000). Nevertheless, the reliability of the results requires the identification of an internal reference peak in the IR spectra with a constant proportion in the monomer and polymer. Alternatively, the use of model molecules to perform a group response calibration (e.g., based on 4-isopropyl styrene) seems to be a possibility to increase the reliability of the quantitative results in the IR-spectroscopy of PDBs (Hecker, 2000). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 4-60 Figure 4.57. FTIR spectrum observed for the isolated NMRP network produced in run 8 (100 % DVB in the initial monomer mixture). Strong IR absorbances observed at 992 and 1630 cm−1 identify a high concentration of pendant double bonds in the network. Figure 4.58. FTIR spectrum observed for the isolated NMRP network produced in run 10 (5 % DVB in the initial monomer mixture). 4-61 CHAPTER 4. Nitroxide Mediated Radical Polymerisation of Styrene/Divinylbenzene Another shortcoming associated with the use of this technique for PDB quantification occurs when the concentration of these functional groups in the material is low. Specific peaks corresponding to the C=C bonds present a weak intensity, particularly when compared with the responses of other chemical groups (see Figure 4.58), which hinders their rigorous quantification. Chemical titration, extensively used in this research, is a technique with a broader range of applicability concerning the observed concentration of PDBs in the networks. 4.6 Conclusions The solution polymerisation system was experimentally studied through the synthesis and characterization by SEC/RI/MALLS of these non-linear copolymers. Model predictions were compared with the measured dynamics in a batch reactor of key polymer properties such as monomer conversion, molecular weights and z-average radius of gyration. Different experimental methods were used to obtain insight on the molecular architecture of the networks, such as: the analysis by SEC/RI/MALLS proved to be especially useful in this research: reaction time evolution of the crosslinking process was, therefore, studied and the effect of the different polymerisation conditions on the molecular structure of products was compared. This technique showed that, when compared with FRP, NMRP allows only limited control over the crosslinking process. NMRP non-linear polymerisation proceeds in a similar way to a random FRP process and, therefore, a very broad polymer population is formed. And the analysis by SEM of the different products synthesized did not yield clear conclusions on the effect of operation conditions on materials morphology. Nanoand micro-sized structures were identified in linear/non-linear products obtained by FRP and also by NMRP. These structures seem to be a consequence of the synthesis process (e.g., operation at reaction temperature above the glass transition temperature and/or secondary nucleation (Yuyama et al., 2000; Ma et al., 2001 and 2003) and post-polymerisation treatment of products (precipitation/drying). Design of new operation conditions is needed (e.g., operation of a micro-channel suspension reactor) in order to elucidate the actual effect of nanoand microgelation on products morphology. Manipulation (isolation) of the produced gel beads (Tanaka et al., 2007) and minimization of secondary nucleation during polymerisation (Yuyama et al., 2000; Ma et al., 2001 and 2003) are also important issues on this subject. This work has confirmed that NMRP technique allows the synthesis of hyperbranched polymers with improved homogeneity, owing to less intramolecular cyclizations as compared Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 4-62 to conventional radical polymerisations. Moreover, it was also shown that NMRP allows operation with higher divinyl monomer content without gelation, which is an important issue in the production of hyperbranched polymers. In addition, this work has allowed the assessment of NMRP capabilities for synthesizing networks and gel with controlled molecular architecture. An ideal crosslinking process has been predicted to lead to polymer populations with different numbers of crosslinking points and narrow size distributions (Dias and Costa, 2010). However, the results here presented show high deviations from this behaviour with NMRP of STY/DVB, in contrast with the recent observation for living anionic polymerisation of 1,4-divinylbenzene (Hirao et al., 2011). Compared with FRP, NMRP does indeed allow some control on the crosslinking process, but intramolecular cyclizations cannot be suppressed and remain a major feature. The design of operation conditions combining CRP with variable concentration of crosslinker along the reaction time (e.g. in a semi-batch process) can eventually lead to an even higher degree of control of network formation. The results presented here can also be used to carry out kinetic modelling studieSty/model development on NMRP crosslinking of STY/DVB, which should be useful in the design of such operation conditions. Recent modelling studies based on the use of the method of moments and the Flory-Stockmayer theory of gelation (Hernández-Ortiz et al., 2009 and 2012) showed that NMRP seems to lead to more homogeneous polymer networks, but this effect is less important when the initial content of crosslinker is increased. Combination of the new experimental results presented here with a general kinetic modelling approach based on generating functions (Costa and Dias, 1994, 2005, 2006 and 2007; Dias and Costa, 2003, 2005, 2006, 2007 and 2010) will hopefully improve this discussion. Detailed kinetic models were developed for the NMRP of STY/DVB including intramolecular cyclization reactions, as a complement to this work. A comprehensive experimental program concerning the solution and suspension NMRP of STY/DVB was carried out. For comparison purposes, some FRP experiments with the same polymerisations systems were also performed. CHAPTER 5 REVERSIBLE ADDITION-FRAGMENTATION CHAIN-TRANSFER POLYMERISATION OF STYRENE AND COPOLYMERISATION OF STYRENE/DIVINYLBENZENE Abstract. Experimental studies concerning RAFT homopolymerisation of styrene (linear) and copolymerisation (non-linear) of styrene and commercial divinylbenzene were performed in aqueous suspension (leading to gel formation) to assess the use of controlled radical polymerisation for the production of non-conventional polymer networks. The experiments were carried out in stirred batch reactor and changes of a few operation parameters were tried in order to assess their effect in key properties:  Kind of polymerisation reactor:  2.5 L atmospheric reactor.  1 L pressurized reactor.  Explored parameters:  Temperature and monomer dilution.  Amount of crosslinker (DVB).  Kind of RAFT agent and the initial ratio of RAFT agent with monomer and initiator.  Products characterization:  Monomer conversion determined by GPC and gravimetry.  Molecular architecture of the products, namely:  Average molecular weights.  Average radius of gyration.  In-line and off-line FTIR-ATR.  Iodine chloride titration for determine PDBs concentration.  Gel fraction  SEM for observation of polymer particle morphology.  Comparison between FRP/NMRP/RAFT of the chemical system STY/DVB. This chapter is based on the following publication: M.A.D. Gonçalves, V.D. Pinto, R.C.S. Dias, J.C. Hernández-Ortiz, M.R.P.F.N. Costa, Macromol. Symp. 333 (2013) 273-285. 5-3 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 5.1 Introduction Among controlled radical polymerisation techniques, reversible addition-fragmentation Chain-Transfer (RAFT) polymerisation presents some advantages comparatively with other CRP techniques. It is especially versatile because it can be used with different classes of monomers (e.g. with organic and water compatible monomers) and considering a wide range of operation conditions (e.g. using a broad range of polymerisation temperatures) (Moad et al., 2012; Braunecker and Matyjaszewski, 2007). Due to these features, RAFT presents promising industrial applicability (Zhang e Ray, 2011; Russum et al., 2005 and 2006; Rivera et al., 2005; Smulders et al., 2005; Moad et al., 2005, 2009 and 2012) and its use with dispersed systems was recently assessed in many important works (Cunningham, 2008; Zetterlund et al., 2008). In fact, the “green” polymerisation offered by aqueous dispersed systems also provides high heat dissipation rates and low viscosity products that are easily handled when compared with solution or bulk polymers. It was stated that polymerisation techniques that do not change the number of radicals per particle (such RAFT) should be considered (Butté et al, 2000 and 2001). Also, minemulsion is preferable because monomer droplets become the main loci of particle nucleation when an oil-soluble initiator is used, and in these conditions particles behave as bulk reactors with large productivity due to radical compartmentalization. Typical RAFT agents are quite water insoluble and therefore, when operating with emulsion polymerisation, there is no efficient transport between droplets and particles as it would be needed to achieve a controlled polymerisation. Conversely, with water mobile RAFT agents, kinetics of emulsion polymerisation is strongly affected (lowering of reaction rates) because the aqueous phase radicals involved with particle formation become also deactivated (Russum et al., 2005). As above described, these difficulties can be circumvented on considering the operation with RAFT in miniemulsion, a combination offering two additional advantages: radical segregation leading to the decrease of bimolecular termination and improved retention of the control agent into the polymer particles due to permanent bounding polymer/RAFT CTA across the polymerisation (Russum et al., 2005; Butté et al, 2000 and 2001). Avoiding a large excess of metal complexes in final product (characteristic of ATRP) and high reaction temperatures to achieve a fairly high polymerisation rates (usually higher than 100 °C with NMRP) are other key advantages of RAFT when industrial implementation is considered. Undesirable higher incidence of termination mechanisms associated with RAFT (which can be minimized by increasing the initial mole ratio between RAFT agent and Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-10 5.3 RAFT Homopolymerisation of Styrene 5.3.1 Chemical Species and Chemical Reactions Different possibilities for kinetic schemes describing the RAFT polymerisation of styrene are traduced by Equations (5.1) to (5.25) complemented by Tables 5.3 and 5.4 where are represented the set of chemical species and the basic set of rate coefficients considered in the modelling studies of the RAFT polymerisation of styrene. Equations (5.1) to (5.13) describe the conventional steps included in radical polymerisation, namely initiator decomposition, monomer initiation, monomer propagation, irreversible chain transfers to monomer, solvent initiator, termination by combination and disproportionation. In order to assess the effect of thermal initiation of styrene (possible at high polymerisation temperatures such in some runs performed in this work) the Mayo dimerization of styrene was also included in the kinetic scheme. Equations (5.14) to (5.17) represent the reaction scheme describing RAFT steps involving primary radicals. Finally, Equations (5.18) to (5.25) describe the reaction scheme considering macro-radical intermediate (adduct) and cross-termination in a RAFT polymerisation. In Figure 5.3 is depicted the main chain equilibration step in RAFT polymerisation. R0 ka Addition-fragmentation with primary radicals M R0 S Polymer core-equilibrium Polymer pre-equilibrium S Pn kf kf RT M R0 RT ka ka M Pm S Pn RT Z Z S Z S S ka S RT S S kf S kf Pn kf S kf Pn RT S Pn Pm ka Pn Z Z ka RT Z S Z Pm Z S S S Z S S Figure 5.3. Depiction of the main chain equilibration step in RAFT polymerisation. 5-11 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Initiator Decomposition: (5.1) Mayo dimerization of styrene: (5.2) Thermal initiation of styrene: (5.3) Monomer initiation by primary radicals of the initiator: (5.4) Monomer initiation by primary radicals of the monomer: (5.5) Monomer initiation by primary radicals of the dimer: (5.6) Monomer propagation: (5.7) Irreversible chain transfer to agent: (5.8) Irreversible chain transfer to monomer: (5.9) Irreversible chain transfer to dimer: (5.10) Irreversible chain transfer to initiator: (5.11) Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-12 Termination by combination: (5.12) Termination by disproportionation: (5.13) RAFT addition and fragmentation of different kinds of primary radicals: (5.14) (5.15) (5.16) Monomer initiation by primary radicals from RAFT: (5.17) RAFT addition with CTA (pre-equilibrium): (5.18) RAFT addition and fragmentation: (5.19) Cross-termination: (5.20) Cross-oligomer termination: (5.21) (5.22) 5-13 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Cross-primary radicals termination: (5.23) (5.24) (5.25) Table 5.3. Set of chemical species considered in the RAFT polymerisation of styrene. Group description Alias Styrene monomer M Thermal initiator (AIBN) I Dimer from styrene MM Solvent T Primary radical from initiator R0 Primary radical from monomer RM Primary radical from dimer RD Primary radical from RAFT agent RA Active polymer chain with DP= n 1 Pn Dead polymer chain with DP= n 1 Dn RAFT agent A Dormant polymer chain with DP= n 1 APn Macro-radical intermediate (adduct) PnA*Pm This kinetic scheme is based on the RAFT modelling work recently published by SaldívarGuerra (Zapata-González et al., 2011). It is important to stress that modelling of RAFT process was a controversial subject in the last decade, as reported in some important papers (Barner-Kowollik et al., 2001 and 2003; Wang and Zhu, 2003; Wang et al., 2003; Wulkow et al., 2004). Definition of at least the following four fundamental types of chains was considered in most past works: propagating radical chains , adduct radical chains , dormant chains and dead chains (see e.g. Sun et al., 2008; Wang and Zhu, 2003; Zhang and Ray, 2001). An additional definition of primary intermediate radicals (formed in the pre-equilibrium) can also be found in these papers (Sun et al., 2008; Zhang and Ray, 2001). Works dealing simultaneously with one-, twoand three-arm adducts were also recently published (see e.g. Hernández-Ortiz et al., 2010). Principles here described were also Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-14 recently used by Zapata-González et al., (2011) to calculate the full molecular weight distribution in RAFT polymerisation. Table 5.4. Basic set of rate coefficients considered in the modelling studies of the NMRP copolymerisation of STY/DVB monomers. Group description a) Rate coefficient expression b) Initiator thermal decomposition Styrene propagation Initiation of styrene by primary radicals Chain transfer to monomer Chain transfer to solvent Chain transfer to dimer Mayo dimerization Dimer decomposition (s-1) Thermal initiation Radical termination a) Kinetic parameters expressed in dm3mol-1s-1, unless otherwise stated. b) References of the rate coefficients are the same as in Table 4.4 of Chapter 4. 5.3.2 Results and Discussion Main experimental results obtained in this chapter for styrene RAFT polymerisation are graphically summarized in Figures (5.4)-(5.12). In those figures, for each run described in Table 5.1, the following information is provided:  Dynamics of monomer conversion.  Dynamics of average molecular weights ( and ) and of molecular weight dispersity (DM).  Observed Refractive Index (RI) signal in SEC chromatographic traces correspondent to RAFT polystyrene samples collected at different reaction times. Change of the concentration and molecular weight of the polymer population along the reaction time is thus showed. 5-15 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB  Observed Multi-Angle Laser Light Scattering (MALLS) signal in SEC chromatographic traces correspondent to RAFT polystyrene samples also collected at different reaction times. Reaction time growth of the molecular size of the polymer population can be here observed.  Comparison of the observed RI and MALLS signals correspondent to samples with t = 12 hr reaction time. With these comparisons, the effect of operation conditions on the molecular architecture of the RAFT produced polystyrene (namely polymer population homogeneity) is highlighted. Besides insights on the kinetics of monomer consumption and product formation with aqueous RAFT suspension, important features of product molecular structure are also provided by the results presented in Figures (5.4)-(5.12). Note that crucial details of the molecular of the produced RAFT polystyrene are not discernible when only the average properties of the polymer are observed. In fact, the formation of a higher size secondary polymer population (SP) is observed in these SEC traces when DDMAT or CDT are used as RAFT agents but not when TBTGA is considered. This is a relevant result with these commercially available CTAs, confirming the importance of the choice of the RAFT agent to achieve a higher control of the polymer molecular architecture. Formation of higher sized secondary polymer populations is almost only residually observed when the RI signal is considered but its importance becomes evident with the MALLS signal (see graphics (c), (d) and (e) along Figures (5.4)-(5.12)). Using multiple detection is therefore an important advantage when a rigorous characterization of the molecular architecture of the products is sought. For instance, the estimation of the molecular weight dispersity (DM) is presented in graphics (b) of Figures (5.4)-(5.12) considering the inclusion of the secondary population and the exclusion of these chains (using only the main RI peak). In spite of some experimental uncertainties (effect of noisy signals with diluted samples, etc.) huge differences in DM are observed considering the presence/absence of this secondary population when DDMAT or CDT are used as RAFT agents. With TBTGA (see Figure 5.4), in spite of the lower polymerisation rate, a polymer with a reasonable molecular homogeneity is produced (DM around 1.4). Very low values of DM (around 1.2) are estimated with DDMAT or CDT when the secondary population is not included but very high dispersity of polymer population (DM ~ 2 or even higher) results when these chains are accounted. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-16 0.0 0.2 0.4 0.6 0.8 1.0 0 120 240 360 480 600 720 T=70 oC Monomer Conversion Time (min) RAFT Polymerization Styrene 50% v/v (~4.4 mol/L) AIBN/Styrene=0.208% (mol/mol) TBTGA/AIBN=1.841 (mol/mol) 103 104 105 1 2 3 0 120 240 360 480 600 720 DM Mn Mw Average Molecular Weight DM Time(min) T=70 oC (a) (b) 0.0 0.2 0.4 0.6 0.8 1.0 21 22 23 24 25 26 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC RI t=12 h t=2 h 6 0.0 0.2 0.4 0.6 0.8 1.0 20 21 22 23 24 25 26 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC LS t=12 h t=2 h 6 (c) (d) 0.0 0.2 0.4 0.6 0.8 1.0 19 20 21 22 23 24 25 26 27 Normalized Response Elution Volume (mL) t=12 h RI Styrene RAFT Polymerization at 70 oC LS 0,0 0,2 0,4 0,6 0,8 1,0 103104105106 Normalized Response Molecular Weight (g/mol) t=12 h RI Styrene RAFT Polymerization at 70 oC LS (e) (f) Figure 5.4. Main results for RAFT polystyrene synthesis in aqueous suspension at 70 °C using TBTGA CTA (run 1 in Table 5.1). (a) Measured reaction time evolution of monomer conversion. (b) Measured reaction time evolution of molecular weight dispersity (DM) and average molecular weights ( and ). (c) Normalized RI signal of polystyrene samples with different polymerisation time. (d) Normalized MALLS signal (90° detector) for the same samples described in (c). (e) Comparison of the normalized RI and MALLS signals of a polystyrene sample with polymerisation time t = 12 h. (f) Change of the RI and LS signals along the molecular weight for the same sample described in (e). 5-17 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 0.0 0.2 0.4 0.6 0.8 1.0 0 120 240 360 480 600 720 T=70 oC Monomer Conversion Time (min) RAFT Polymerization Styrene 50% v/v (~4.4 mol/L) AIBN/Styrene=0.216% (mol/mol) DDMAT/AIBN=1.869 (mol/mol) 103 104 1 2 3 4 0 120 240 360 480 600 720 DM (including secondary population) DM (excluding secondary population) Mn Mw Average Molecular Weight DM Time(min) T=70 oC (a) (b) 0.0 0.2 0.4 0.6 0.8 1.0 19 20 21 22 23 24 25 26 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC RI t=12 h t=2 h 8 0.0 0.2 0.4 0.6 0.8 1.0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) LS Signal Styrene Polymerization at 70 oC t=12 h t=3 h 8 h (c) (d) 0.0 0.2 0.4 0.6 0.8 1.0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) LS Signal Styrene Polymerization at 70 oC t=12 h RI Signal 0,0 0,2 0,4 0,6 0,8 1,0 103104105106 Normalized Response Molecular Weight (g/mol) t=12 h RI Styrene RAFT Polymerization at 70 oC LS (e) (f) Figure 5.5. Main results for RAFT polystyrene synthesis in aqueous suspension at 70 °C using DDMAT CTA (run 2 in Table 5.1). (a) Measured reaction time evolution of monomer conversion. (b) Measured reaction time evolution of molecular weight dispersity (DM) and average molecular weights ( and ). (c) Normalized RI signal of polystyrene samples with different polymerisation time. (d) Normalized MALLS signal (90° detector) for the same samples described in (c). (e) Comparison of the normalized RI and MALLS signals of a polystyrene sample with polymerisation time t = 12 h. (f) Change of the RI and LS signals along the molecular weight for the same sample described in (e). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-18 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 T=70 oC Monomer Conversion Time (min) RAFT Polymerization Styrene 50% v/v (~4.4 mol/L) AIBN/Styrene=0.213% (mol/mol) CDT/AIBN=1.884 (mol/mol) 103 104 1 2 3 4 5 0 120 240 360 480 600 720 DM (with SP) DM (without SP) Mn Mw Average Molecular Weight DM Time(min) T=70 oC (a) (b) 0,0 0,2 0,4 0,6 0,8 1,0 19 21 23 25 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC RI t=12 h t=2 h 0,0 0,2 0,4 0,6 0,8 1,0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) LS Signal Styrene Polymerization at 70oC t=12 h t=4 h (c) (d) 0,0 0,2 0,4 0,6 0,8 1,0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oCt=12 h RI LS 0,0 0,2 0,4 0,6 0,8 1,0 103104105106 Normalized Response Molecular Weight (g/mol) t=12 h RI LS Styrene RAFT Polymerization at 70 oC (e) (f) Figure 5.6. Main results for RAFT polystyrene synthesis in aqueous suspension at 70 °C using CDT CTA (run 3 in Table 5.1). (a) Measured reaction time evolution of monomer conversion. (b) Measured reaction time evolution of molecular weight dispersity (DM) and average molecular weights ( and ). (c) Normalized RI signal of polystyrene samples with different polymerisation time. (d) Normalized MALLS signal (90° detector) for the same samples described in (c). (e) Comparison of the normalized RI and MALLS signals of a polystyrene sample with polymerisation time t = 12 h. (f) Change of the RI and LS signals along the molecular weight for the same sample described in (e). 5-19 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 T=70 oC Monomer Conversion Time (min) RAFT Polymerization Styrene 100% v/v (~8.7 mol/L) AIBN/Styrene=0.105% (mol/mol) CDT/AIBN=1.910 (mol/mol) 103 104 105 1 2 3 4 5 0 120 240 360 480 600 720 DM (with SP) DM (without SP) Mn Mw Average Molecular Weight DM Time(min) T=70 oC (a) (b) 0,0 0,2 0,4 0,6 0,8 1,0 18 20 22 24 26 28 Normalized Response Elution Volume (mL) RI t=12 h t=1 h Styrene RAFT Polymerization at 70 oC 100% Monomer in the organic phase 0,0 0,2 0,4 0,6 0,8 1,0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) LS t=12 h t=1 h Styrene RAFT Polymerization at 70 oC 100% Monomer in the organic phase (c) (d) 0,0 0,2 0,4 0,6 0,8 1,0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC 100% Monomer in the organic phase t=12 h RI LS 0,0 0,2 0,4 0,6 0,8 1,0 103104105106 Normalized Response Molecular Weight (g/mol) t=12 h RI LS Styrene RAFT Polymerization at 70 oC 100% Monomer in the organic phase (e) (f) Figure 5.7. Main results for RAFT polystyrene synthesis in aqueous suspension at 70 °C using CDT CTA (run 4 in Table 5.1). (a) Measured reaction time evolution of monomer conversion. (b) Measured reaction time evolution of molecular weight dispersity (DM) and average molecular weights ( and ). (c) Normalized RI signal of polystyrene samples with different polymerisation time. (d) Normalized MALLS signal (90° detector) for the same samples described in (c). (e) Comparison of the normalized RI and MALLS signals of a polystyrene sample with polymerisation time t = 12 h. (f) Change of the RI and LS signals along the molecular weight for the same sample described in (e). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-26 Some of these issues are further illustrated in Figures 5.13 and 5.14 where the effect of the polymerisation temperature on the observed RI and MALLS signals of the final samples (t=12 hr) are directly compared. Measured effects of the reaction temperature and of initial mole ratio between RAFT agent and initiator on the dynamics of monomer conversion are presented in Figures 5.15 and 5.16. Figure 5.13. Effect of the polymerisation temperature (in the range 70 to 150 °C) on the observed RI signals of the final samples (t=12 hr). The effect on the kinetics of polymerisation of the change of the RAFT agent is illustrated in Figures (5.17)-(5.20) where three experimental runs at 70 °C using TBTGA, DDMAT and CDT are compared. Lowering of polymerisation rate when using the dithiobenzoate TBTGA instead of trithiocarbonates (DDMAT or CDT) is here clearly observed (Figure 5.17). Different conversion profiles are observed and the retardation effect associated to dithiobenzoates (TBTGA) becomes clear when compared with trithiocarbonates (DDMAT and CDT). Some scattering observed for the measured values of monomer conversion is possibly due to non-ideal sampling of the reactor along polymerisation time due to the special heterogeneity of the suspension (adhesion of organic phase to the reactor walls was evident with some polymerisation runs). This issue is an important shortcoming of the polymerisation 0.0 0.2 0.4 0.6 0.8 1.0 19 20 21 22 23 24 25 26 27 Normalized Response Elution Volume (mL) RI Signal 90 oC t=12 h 130 oC T=70 oC 150 oC 110 oC 5-27 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB conditions here explored. Observed dynamics of and for these same three experiments are presented in Figure 5.18 and 5.19. 0.0 0.2 0.4 0.6 0.8 1.0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) LS Signal 90 oC t=12 h T=130 oC 70 oC 150 oC 110 oC Figure 5.14. Similar comparison described in Figure 5.13 considering the MALLS signal. In spite of differences in the earlier polymerisation stages (with measurements affected by the noisy signal due to low polymer concentration), similar trends are observed with final average molecular weights in the range 7000 to 11000 for and 15000 to 30000 for . Nevertheless, important differences are observed in the product molecular size homogeneity due to the influence of the secondary population that is formed when DDMAT or CDT are used. In fact, with TBTGA reasonable low values of the molecular weight dispersity (DM 1.2) are achieved, as can be observed in Figure 5.20. With DDMAT or CDT, very high values of DM are observed when this high size set of chains is considered in the polymer population although dispersity values close to 1 are estimated in the absence of these secondary chains. It is also enlightning to compare the synthesis of polystyrenes using different polymerisation techniques, as illustrated in Figures (5.21)-(5.25). Results for styrene polymerisation at 130 °C using FRP, NMRP and RAFT are presented in these Figures, including the dynamics of product formation and some details of the polymers molecular architectures. Similar reaction rates are observed when these FRP and RAFT runs are compared and lower monomer conversions were measured with NMRP in the described conditions. Similar dynamics for Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-28 were measured with RAFT and NMRP (final values in the range of 20000) but nearly five times higher values of result with FRP (Figure 5.22). 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 T=90 oC 110 130 150 Monomer Conversion Time (min) RAFT Polymerization Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) DDMAT/AIBN~2 (mol/mol) Figure 5.15. Measured dynamics of monomer conversion for RAFT styrene polymerisation in aqueous suspension at different temperatures and using DDMAT as CTA agent. 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 RAFT/AIBN=1.882 RAFT/AIBN=3.931 Monomer Conversion Time (min) RAFT Polymerization with DDMAT Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) T=130 oC Figure 5.16. Measured dynamics of monomer conversion for RAFT styrene polymerisation in aqueous suspension at 130 °C and using DDMAT as CTA agent. 5-29 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 TBTGA DDMAT CDT Monomer Conversion Time (min) RAFT Polymerization Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) RAFT/AIBN~2 (mol/mol) T=70 oC Figure 5.17. Comparison of the observed time evolution of monomer conversion for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 102 103 104 105 0 120 240 360 480 600 720 TBTGA DDMAT CDT Number Average Molecular Weight Time (min) T=70 oC RAFT Polymerization Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) RAFT/AIBN~2 (mol/mol) Figure 5.18. Comparison of the observed time evolution of for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-30 102 103 104 105 0 120 240 360 480 600 720 TBTGA DDMAT CDT Weight Average Molecular Weight Time (min) T=70 oC RAFT Polymerization Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) RAFT/AIBN~2 (mol/mol) Figure 5.19. Comparison of the observed time evolution of for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 1 2 3 4 5 0 120 240 360 480 600 720 TBTGA DDMAT (including SP) DDMAT (excluding SP) CDT (including SP) CDT (excluding SP) Molecular Weight Dispersity (DM) Time (min) T=70 oC RAFT Polymerization Styrene~4.4 mol/L AIBN/Styrene~0.2% (mol/mol) RAFT/AIBN~2 (mol/mol) Figure 5.20. Comparison of the observed time evolution of DM for aqueous suspension of RAFT synthesis of polystyrene at 70 °C using the CTA agents DDMAT, TBTGA and CDT. 5-31 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 0,0 0,2 0,4 0,6 0,8 1,0 0 120 240 360 480 600 720 RAFT with DDMAT/AIBN~2 NMRP with TEMPO/AIBN~1 FRP Monomer Conversion Time (min) T=130 oC Styrene/Xylene=50/50 AIBN/Styrene~0.2% (mol/mol) Figure 5.21. Measured dynamics of monomer conversion for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 102 103 104 105 106 0 120 240 360 480 600 720 RAFT with DDMAT/AIBN~2 NMRP with TEMPO/AIBN~1 FRP Weight Average Molecular Weight Time (min) T=130 oC Styrene/Xylene=50/50 AIBN/Styrene~0.2% (mol/mol) Figure 5.22. Measured dynamics of for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-32 Molecular size distribution, as seen by SEC, is showed in Figures 5.23 and 5.24 making evident the high size dispersion associated with FRP synthesis. Higher homogeneity of polymer population is observed when NMRP or RAFT are considered but the latter system (is these particular conditions) is affected by the formation of a bimodal size distribution. These issues are reflected in the molecular weight dispersity of the products, as presented in Figure 5.25. Only in the absence of this secondary population similar values of DM are obtained using NMRP or RAFT. 5.4 RAFT Copolymerisation of Styrene with Divinylbenzene 5.4.1 Results and Discussion In Table 5.5 are presented some vibrational assignments correspondent to styrene and divinylbenzene monomers (Choi and Kertesz, 1997; Colthup et al., 1990; Hecker, 2000). These features can also be observed in Figure 5.26 where the off-line IR spectra collected for STY, DVB RAFT polystyrene and RAFT STY/DVB (95/5) are compared. Especially important for the study of the kinetics of crosslinking are the assignments correspondent to C=C bonds that can be identified at around 992, 1019, 1410, 1452 and 1630 cm-1. For this chemical group, strong absorptions and well defined peaks (minimizing the interference with other structures) are observed at 992 and 1630 cm-1. These characteristics frequencies are therefore good candidates to obtain information concerning the monomers carbon-carbon double bonds consumptions and also concerning the presence of pendant double bonds in the polymer/network. In fact, these bands are not present in isolated RAFT polystyrene and very low responses are observed in these regions for isolated RAFT STY/DVB (95/5) networks due to the relative small amount of DVB used (see Figure 5.26). Potentialities of off-line FTIR analysis of isolated products are further enhanced in Figures (5.27)-(5.29) where RAFT materials with high pendant double bonds content (e.g. resulting from DVB homopolymerisation) are considered. Clear qualitative information about the presence of the PDBs in the networks can be obtained and the comparison of products with different reaction times and/or resulting from different initial compositions can also be performed. 5-33 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB 0.0 0.2 0.4 0.6 0.8 1.0 19 20 21 22 23 24 25 26 27 Normalized Response Elution Volume (mL) RI Signal RAFT with DDMAT/AIBN=2 Styrene Polymerization at 130 oC NMRP with TEMPO/AIBN=1.1 FRP Figure 5.23. Observed RI SEC traces for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). 0.0 0.2 0.4 0.6 0.8 1.0 17 19 21 23 25 Normalized Response Elution Volume (mL) LS Signal RAFT with DDMAT/AIBN=2 Styrene Polymerization at 130 oC NMRP with TEMPO/AIBN=1.1 FRP Figure 5.24. Observed MALLS SEC traces for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-34 1 2 3 0 120 240 360 480 600 720 RAFT including SP RAFT excluding SP NMRP FRP Molecular weight dispersity (DM) Time (min) T=130 oCStyrene/Xylene=50/50 AIBN/Styrene~0.2% (mol/mol) Figure 5.25. Observed DM for polystyrene synthesis in aqueous suspension at 130 °C considering different reaction techniques (FRP, NMRP and RAFT). Figure 5.26. Observed off-line FTIR spectra for styrene and divinylbenzene monomers and polystyrene (run 7 in Table 5.1) and poly(STY/DVB) with 5 % DVB (run 1 in Table 5.2). 5-35 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Figure 5.27. Observed off-line FTIR spectra for DVB networks correspondent to different polymerisation times for run 2 in Table 5.2. Figure 5.28. Observed off-line FTIR spectra for DVB networks correspondent to different polymerisation times for run 3 in Table 5.2. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-42 To obtain this calibration, the final samples correspondent to runs 1, 2 and 3 in Table 5.2 were considered. Measurements using both methods (FTIR and ICl titration) were performed for these networks. The following relation between the normalized FTIR 1630 cm-1 peak intensity ( and the concentration of PDBs in the network (expressed in mol/g of polymer) was thus estimated: (5.27) This calibration was used to estimate the PDB concentration for networks with a single (and simpler) FTIR measurement. These results are presented in Figure 5.37. Dynamics of PDBs concentration for runs 2 and 3 in Table 5.2 were obtained and compared with the reaction time evolution of the same variable for run 1 which was fully measured using ICl titration. Note that a much smaller amount of DVB was used in run 1 and, under these circumstances, chemical titration is preferable to FTIR spectroscopy due to the low response of PDBs observed (when compared with that correspondent to other groups). 0,0 0,2 0,4 0,6 0,8 1,0 1,2 060 120 180 240 300 360 420 Normalized intensity ratio Network/Monomer of the 1630 cm-1 peak Reaction Time (min) RAFT polymerization of DVB at 70 oC DDMAT CTA Run 2 Run 3 Figure 5.36. Observed ratios between network and DVB considering the normalized FTIR 1630 cm-1 peak (using the 1494 cm-1 as internal reference). Measurements are correspondent to samples with different polymerisation times in runs 2 and 3 in Table 5.2. (see Eq. (5.26)). 5-43 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Results presented in Figure 5.37 show the (small) effect of intramolecular cyclization that is probably caused when monomer is diluted from bulk to a 50 % monomer/solvent solution (see comparison between runs 2 and 3 in Table 5.2). Strategy here presented should be extended to more diluted polymerisation systems in order to assess the possible effect of cyclization on network formation, even with RAFT polymerisation. Comparison of the incidence of these intramolecular mechanisms in FRP, NMRP (Gonçalves et al., 2013a; Aguiar, 2013; Aguiar et al., 2013a and 2013b) and RAFT of STY/DVB is an expected result of this research line. Results here presented (and expected extensions) can be used to develop kinetic modelling studies including intramolecular cyclizations with RAFT STY/DVB polymerisation, as recently performed with NMRP of the same chemical system (Aguiar, 2013; Aguiar et al., 2013a and 2013b). SEC/RI/MALLS proved to be a valuable technique to obtain insights on the RAFT crosslinking polymerisation. Figure 5.38 shows measured dynamics of , and for STY/DVB (95/5) RAFT copolymerisations at 130 °C (run 1 in Table 5.2). Gelation at around 360 min identified with these conditions. Dissimilitudes between linear (STY) and non-linear (STY/DVB) RAFT polymerisation are highlighted in Figure 5.39 where the measured dynamics of for run 7 in Table 5.1 and run 1 in Table 5.2 are compared. 0 1 x 10-3 2 x 10-3 3 x 10-3 4 x 10-3 5 x 10-3 6 x 10-3 7 x 10-3 0 120 240 360 480 600 720 Concentration of Pendant Double Bonds (mol/g polymer) Reaction Time (min) Run 2 - Estimated using FTIR/Titration Calibration Run 3 - Estimated using FTIR/Titration Calibration Measured by ICl Titration Run 1 - Measured by ICl Titration Figure 5.37. Dynamics of the PDB concentration (mol/g polymer) measured by ICL titration (run 1 in Table 5.2) and considering also the calibration between the normalized FTIR 1630 cm-1 peak intensity ( ) and PDB concentration. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-44 103 104 105 106 0120 240 360 480 600 720 Number Average Weight Average Z Average Average Molecular Weight Time (min) RAFT Styrene/Divinylbenzene Copolymerization at 130 oC (5% DVB) Figure 5.38. Measured dynamics of , and in aqueous suspension RAFT copolymerisation of STY/DVB (run 1 in Table 5.2) 103 104 105 106 0120 240 360 480 600 720 Styrene/Divinylbenzene RAFT copolymerization (5% DVB) Styrene RAFT polymerization Weight average Molecular Weight Time (min) Figure 5.39. Comparison of the observed dynamics of in aqueous RAFT suspension polymerisation of styrene (run 7 in Table 5.1) and RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). 5-45 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB SEC estimated monomer conversion for these same two runs is showed in Figure 5.40 and similar kinetics for the linear and non-linear cases is observed under these circumstances. Measured time evolution of PDBs concentration for STY/DVB (95/5) RAFT copolymerisation (run 1 in Table 5.2) is showed in Figure 5.41. These concentrations were measured through the ICl titration method above described (Gonçalves et al., 2013a). The impact of some operation conditions on the dynamics of STY/DVB RAFT products formation was studied through the change of particular parameters, such as the kind of RAFT CTA considered. The effect of the initial organic phase composition (e.g. STY, DVB, and diluent amounts) on the dynamics of global monomer conversion was briefly assessed through comparisons of runs 2 in Table 5.1 and runs 2 and 3 in Table 5.2 (see Figure 5.42) where the same reaction temperature and RAFT CTA were considered (70 °C and DDMAT). Similar dynamics of monomer conversions where measured within the ascribed RAFT polymerisation conditions. 0.0 0.2 0.4 0.6 0.8 1.0 0120 240 360 480 600 720 Run 1 in Table 5.2 Run 7 in Table 5.1 Monomer Conversion Time (min) Figure 5.40. Measured dynamics of monomer conversion in aqueous suspension RAFT polymerisation of styrene (run in Table 5.1) and RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-46 0,0 5,0x10-4 1,0x10-3 1,5x10-3 2,0x10-3 0 120 240 360 480 600 720 Time (min) Concentration of total PDB (mol/g polymer) Measured Dynamics of Pendant Double Bonds Concentration RAFT Styrene/Divinylbenzene Copolymerization at 130 oC (5% DVB) Figure 5.41. Measured dynamics of PDB concentration in RAFT copolymerisation of STY/DVB (run 1 in Table 5.2). 0,0 0,2 0,4 0,6 0,8 1,0 0120 240 360 480 600 720 RAFT Styrene RAFT Divinylbenzene (50% DVB/Toluene) RAFT Divinylbenzene (100% DVB) Monomer Conversion Time (min) Figure 5.42. Comparison of the measured dynamics of monomer conversion in aqueous suspension RAFT polymerisation of styrene (run 2 in Table 5.1) and DVB (runs 2 and 3 in Table 5.2) at 70 °C. 5-47 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Key features of the RAFT crosslinking mechanism become evident when the SEC traces of products with different polymerisation times are compared. Observation of these SEC traces for linear (STY) and non-linear (STY/DVB) runs highlights central issues of network formation. In Figure 5.43 are showed the SEC RI signals of polystyrene samples correspondent to different polymerisation times. Besides the growth of with reaction time of the main polymer population, a secondary polymer population with higher size and low concentration can be identified in these chromatograms. This feature is enhanced in Figure 5.44 where the LS and RI signals of the final sample correspondent to the same run (run 3 in Table 5.1) are compared. The huge molecular size of the secondary population becomes evident through the respective LS signal. This means that, even in the linear case, non-ideal mechanisms can be involved in RAFT polymerisation leading to an increase of the products dispersity. Slow fragmentation mechanisms in RAFT leading to bimodal distributions formation are a possible justification for these observations (Zapata-González et al., 2011). Differences/similarities between STY and DVB RAFT polymerisations are illustrated in Figure 5.45 where the observed time evolution of for runs 2 in Table 5.1 and runs 2 and 3 in Table 5.2 are compared. 0,0 0,2 0,4 0,6 0,8 1,0 19 21 23 25 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oC RI t=12 h t=2 h Figure 5.43. Normalized RI signal of polystyrene samples (run 3 in Table 5.1) with different polymerisation times. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-48 0,0 0,2 0,4 0,6 0,8 1,0 17 19 21 23 25 27 Normalized Response Elution Volume (mL) Styrene RAFT Polymerization at 70 oCt=12 h RI LS Figure 5.44. Comparison of the normalized RI and MALLS signals of a polystyrene sample (run 3 in Table 5.1) with polymerisation time t= 12 h. 103 104 105 106 0120 240 360 480 600 720 RAFT Styrene RAFT Divinylbenzene (100% DVB) RAFT Divinylbenzene (50% DVB/Toluene) Weight average Molecular Weight Time (min) Insoluble network after this polymerization time Figure 5.45. Measured dynamics of for RAFT styrene polymerisation (run 2 in Table 5.1) and RAFT DVB (runs 2 and 3 in Table 5.2). 5-49 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB For the DVB RAFT polymerisation runs (2 and 3), the values presented are correspondent to the soluble phase which, in contrast to pure STY polymerisation, vanishes after some polymerisation time. Coexistence of sol and gel is observed when STY/DVB RAFT polymerisation is promoted, even with low amount of DVB (e.g. run 1 in Table 5.2). In spite of these important differences, some similarities in the SEC traces of RAFT polystyrene and soluble RAFT poly (STY/DVB) can be identified, as shown in Figure 5.46. RI signal shows a bimodal population for poly(STY/DVB) and the very high molecular size of the secondary (crosslinked) set of chains is highlighted by the correspondent LS signal. Molecular size of the secondary population developed with S RAFT polymerisation (as above discussed) is located in a region close to that observed with non-linear RAFT polymerisation but the correspondent concentration is significantly lower comparatively to the latter. 0,0 0,2 0,4 0,6 0,8 1,0 16 18 20 22 24 26 28 LS - STY RI - STY LS - STY/DVB RI - STY/DVB Normalized Response Elution Volume (mL) RAFT Polymerization at 130 oC t= 12 hr LS - STY RI - STY LS - STY/DVB RI - STY/DVB Figure 5.46. Normalized RI and LS signals observed for polystyrene (run 7 in Table 5.1) and soluble poly (STY/DVB) network (run 1 in Table 5.2) synthesized RAFT polymerisation at 130 °C. Comparisons for the molecular architecture of the different RAFT products synthesized in this work are also illustrated in Figures 5.47 and 5.48. SEC observed structural dissimilitudes of RAFT poly(STY) and soluble poly(DVB)s are showed in Figure 5.47. Even analysing only the soluble phase, the highly crosslinked nature of the latter systems show a clear contrast with the linear case (see also Figure 5.46 for comparison with poly(STY/DVB) with low amount of DVB). Figure 5.48 compares the SEC traces of poly(DVB) samples, prepared with 50 % of toluene, and collected at different reaction times. Dynamics of two different sub- Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 5-50 populations can be observed and differences between samples correspondent to 2 and 3 hours of reaction time are a consequence of gelation. 0,0 0,2 0,4 0,6 0,8 1,0 19 20 21 22 23 24 25 26 27 28 Normalized Response Elution Volume (mL) RAFT Polymerization at 70 oC Sty t=2 h RI Signal DVB (50% TOL) t=2 h DVB t=1 h Figure 5.47. Comparison of the observed RI signal for RAFT synthesized polystyrene (run 2 in Table 5.1) and soluble poly(DVB) (runs 2 and 3 in Table 5.2). 0,0 0,2 0,4 0,6 0,8 1,0 20 21 22 23 24 25 26 27 28 29 t=1 hr t=2 hr t=3 hr Normalized Response Elution Volume (mL) Divinylbenzene RAFT Polymerization at 70 oC (50% Toluene) RI t=1 hr t=2 hr t=3 hr Figure 5.48. Observed RI signal for RAFT synthesized soluble poly(DVB) samples. Different stages of the crosslinking process (reaction times 1, 2 and 3 hr) are compared. 5-51 CHAPTER 5. Reversible Addition-Fragmentation Chain-Transfer Polymerisation of STY and Copolymerisation of STY/DVB Results present in Figures 5.47 and 5.48 also show structural differences between RAFT poly(DVB) products synthesized using different dilutions (bulk and 50 % toluene), evidencing the possible effect of cyclization due to intramolecular propagation. In Figure 5.49(a)-(d) are shown examples of dried products obtained analysed by SEM. It was possible to observe the influence of synthesis conditions on the morphology of the produced materials. (a) (b) (c) (d) Figure 5.49. SEM micrographs of different polystyrene samples synthesized by RAFT. 5.5 Conclusions Aqueous suspension RAFT polymerisations of styrene and styrene/divinylbenzene were experimentally studied using three different commercially available RAFT agents. An experimental program was performed changing (besides the chemical nature of the RAFT agent), the amount of DVB in the initial monomer mixture (0 to 100 %), the polymerisation temperature (in the range 70 to 150 °C), dilution in the organic phase and the initial composition (monomer, initiator and RAFT agent initial concentrations in the organic phase). A strong influence of the operation conditions on the kinetics of polymerisation and polymer 6-3 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation 6.1 Introduction Hydrogels are three dimensional networks of hydrophilic polymers holding a large amount of water while maintaining the solid state. Hydrogels can swell significantly in the presence of an aqueous solution and can also deswell upon certain stimulations. Thus, hydrophilic gels are able to retain, at a high absorption rate, huge amounts of water (up to 1000 g/g relatively to the dry weight). Thanks to these unique properties, superabsorbent polymers (SAP) have many important applications in hygienic/sanitary industries, agriculture, environment, separation process and other chemical engineering operations. According to the provenience of the materials used in the synthesis, SAPs can be divided in two groups: synthetic SAP that are obtained through the copolymerisation of petrochemical-based monomers and those SAP obtained from the grafting of natural polymers such as polysaccharides (e.g., starch) or polypeptides (Buchholz and Graham, 1997; Buchholz and Peppas, 1994; Zohuriaan-Mehr and Kabiri, 2008). Synthetic SAPs are mainly obtained through the polymerisation of acrylic acid (AA) with different kinds of crosslinkers such as N,N′-methylenebisacrylamide (MBAm), trimethylolpropane triacrylate (TMPTA) or tetraallyloxyethane (TAO). Water-phase polymerisations are generally carried out and therefore the low solubility of some crosslinkers in aqueous medium should be accounted for (Arriola et al., 1997). This problem can be overcome by using an organic solvent and conventional crosslinkers (e.g. divinylbenzene or ethylene glycol dimethacrylate) at the expenses of introducing in the products chemicals hindering the direct use of the materials for some applications. In practice, SAPs are produced using the aforementioned kinds of crosslinkers at levels in the range 0.01 to 2 wt % (considering TMPTA as a model crosslinker) in order to obtain materials with a high swelling ratio (which is attained at a low crosslinker content) and a low soluble fraction (which is more easily fulfilled using a higher crosslinker content). The mole fraction of crosslinker ( ) used in the production of such materials can therefore be as lower as 0.0025 % to 0.5 %. Thermal, redox or even photo initiation are used to carry out the free radical copolymerisation of acrylic acid aqueous solutions in the presence of crosslinkers. Aqueous solutions of AA salts (obtained through the neutralization of AA with a base such as NaOH) are also usually considered in SAP production. Reaction temperatures in the range 40 to 70 °C are often considered in these synthesis processes. Note that in several kinetic studies concerning the polymerisation of acrylic and methacrylic acid (namely using PLP), much lower temperature Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-4 ranges have been considered (e.g. 2 °C to ambient temperature). Nevertheless, in order to assure high reaction rates (and almost complete monomer conversion) in industrial processes, higher temperatures must be used. Other parameters such as the degree of neutralization and monomer concentration have a strong impact on the polymerisation of these ionic monomers (Anseth et al., 1996; Beuermann et al., 2006, 2007a, 2007b, 2007c and 2008; Buback and Junkers, 2006; Buback et al., 2008b; Cutié et al., 1997a and 1997b; Henton et al., 1997; Kabanov et al., 1973 and 1975; Kuchta et al., 2000; Kurland, 1980; Lacík et al., 2001, 2003, 2004 and 2009; Li and Schork, 2006; Renard and McKenna, 2000). Lack of reaction control often associated with aqueous solution production of SAPs (fast exothermic reaction and high viscosity with a concomitant temperature raise) can be overcome using inverse suspension polymerisation which also allows the direct synthesis of powder products or microspheres (Bahaj et al., 2010; Bajpai et al., 2007; Bodugöz and Güven, 2002; Chen et al., 2004; Kiatkamjornwong and Phunchareon, 1999; Omidian et al., 2003; Wang et al., 1997). Porous hydrogels composites with improved properties can also be synthesized changing the conventional procedure through the introduction of inorganic fillers and a porogen (Kabiri et al., 2004). Smart hydrogels are water compatible polymer networks undergoing fast and reversible changes in their hydrophilic/hydrophobic microstructure. These hydrogels are also capable of responding to changes in parameters such as temperature, pH, electric or magnetic fields and are often designated as ‘stimuli-responsive’ polymers or ‘smart’ polymers. Changes at microscopic level become macroscopically apparent and include a huge shrinking/swelling of the network. In Figure 6.1 is depicted a model of a stimuli-responsive drug-delivery system. Figure 6.1. A model depicting a stimuli-responsive drug-delivery system (adapted from Bajpai et al., 2010). 6-5 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation These reversibility transitions are driven by neutralization of charged groups due to pH shift, changes in efficiency of hydrogen bonding or in ionic strength (Galaev and Mattiasson, 2008). Owing to these unique features, hydrogels have found many applications as biomaterials for drug and protein delivery, tissue engineering, nanotechnology and microfabrication techniques. pH sensitive hydrogels are polyelectrolytes containing ionisable groups (pendant acidic or basic groups) which are able to release or accept protons. These mechanisms can be stimulated by changes in the environmental pH where the polymer network is located. Hydrogels containing weakly acid groups swell when the pH is increased and, conversely, those containing weakly basic groups collapse by increasing the pH. These phase transitions (swelling/deswelling) can be explored to synthesize useful materials for drug delivery systems (Bajpai et al., 2010). Nevertheless, the performance of these smart hydrogels (e.g. swelling kinetics) is strongly dependent on the molecular architecture of the networks (e.g. composition or crosslinking density) and on the synthesis conditions used in their production. Temperature of preparation, polymerisation medium (e.g. solution, suspension or emulsion) and/or the presence of a porogen leading to porous hydrogel are some parameters with influence in the end-use properties of such advanced materials (Gemeinhart et al., 2000). The size and shape of the synthesized gel particles, conditioning the water diffusion process, also have a strong impact on the response time of pH sensitive hydrogels. Ionic hydrogels, known as polyelectrolytes, are prepared through the polymerisation of cationic or anionic monomers. A combination of positive and negative charges leads to amphoteric macromolecules (Sen and Guven, 1999; Sutani et al., 2002). Anionic hydrogels networks are known as acid homopolymers charged negatively or copolymers of anionic monomer. Anionic hydrogels are known to exhibit a marked increase in the swelling ratio with increase in the environmental pH (Ende et al., 1995; Lee and Chiu, 2002). Some anionic hydrogels are presented in Figure 6.2. As cationic hydrogel networks are commonly referred the homopolymers of positively charged basic, cationic monomers, copolymers of cationic and natural monomers. Some cationic hydrogels as PDMAEMA and Poly(vinyl pyridine) are shown in Figure 6.3 as example. As schematized in Figures 6.4 and 6.5, the response to pH of cationic pendent groups is contrary to that of anionic pendant groups. The cationic hydrogel remain collapsed in the basic environment and swollen in the acid environment due to electrostatic repulsion between the positively charged groups (Ende et al., 1995; Lee and Chiu, 2002). For the anionic Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-6 hydrogel the opposite situation takes place. In Figure 6.6 is represented the behaviour of a cationic and an anionic hydrogel in acid and basic medium. Figure 6.2. Structural representation of some anionic hydrogels. PDMAEMA Poly(vinyl pyridine) CH3 CH2 C CH N C O CH2 CH3 (CH2)2 N O CH3 Figure 6.3. Structural representation of some cationic hydrogels. (a) (b) Figure 6.4. (a) Schematic of the response of a cationic hydrogel in basic medium. (b) Schematic of the response of a cationic hydrogel in acid medium. (a) (b) Figure 6.5. (a)Schematic of the response of an anionic hydrogel in acid medium. (b) Schematic of the response of an anionic hydrogel in basic medium. 6-7 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation Figure 6.6. Schematic of a pH-sensitive hydrogel in acidic and basic medium. (A) Anionic (B) cationic. Amphoteric hydrogels networks are macromolecules that have positive and negative charges in the entire polymer network. The presence of ionic species along the polymeric chain has distinct effects on the properties of the amphoteric hydrogels as they are in solution or in solid state. In Figure 6.7 is a schematic representation of an amphoteric hydrogel. Neutral or nonionic hydrogels are homopolymeric or copolymeric networks which do not have any charged group in their structure. Neutral hydrogels swell to equilibrium when the osmotic pressure of the solvent is balanced with the sub-chain stretching energy (Bajpai et al., 2010). The swelling and collapse of neutral hydrogel networks occur as a result of a change in the environmental temperature (Ostroha et al., 2004). Some neutral hydrogels networks are shown in Figure 6.8. OH CH2 C C CH3 CH2 CH3 C O C O CH3 (CH2)2 N O CH3 Figure 6.7 Schematic representation of a polyamphoteric hydrogel of MAA/DMAEMA. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-8 N N Poly(vinyl caprolactam) Poly(vinyl pyrrolidinone) O O CH2 CH2 CH CH Figure 6.8 . Schematic representation of some neutral hydrogels. Temperature-sensitive hydrogels have the capability of collapsing and swelling with temperature shifts in the surrounding medium. they can be classified as negative and positive thermosensitive hydrogels as well as thermally reversible gels (Peppas et al., 2000; Qiu and Park, 2001). This class is probably the most commonly studied of environmental-sensitive polymer systems in drug-delivery research. Thermoresponsive hydrogels may be prepared by crosslinking polymers which exhibit a lower critical solubility temperature (LCST) (Wu et al., 2003). The common characteristic of temperature-sensitive polymers is the presence of hydrophobic groups, such as methyl, ethyl and propyl (Qiu and Park, 2001). Some structures of these polymers are shown in Figure 6.9. NH PNIPA N PDMA CH3 CH3 CH3 CH3 CH CH CH C C O O CH2 CH2 Figure 6.9. Structures of some temperature-sensitive polymers. Negative temperature-sensitive hydrogels have a LCST and collapse when heated above it. These hydrogels show an on-off drug release (Soppimath et al., 2002) with on at a low temperature and off at high temperature allowing pulsatile drug release. In Figure 6.10 is represented the response of a negative temperature-sensitive hydrogel when heated/cooled. Positive temperature-sensitive hydrogels have an upper critical temperature (UCST). These hydrogels collapse upon cooling below the UCST. Poly(acrylic acid) and poly(acrylamide) have a positive temperature dependence of swelling (Qiu and Park, 2001). There are several 6-9 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation studies with the goal of synthesizing hydrogels with double sensitivity. This is accomplished with the copolymerisation of a temperature-sensitive monomer, usually NIPA, with a pHsensitive monomer such AA or MAA (Brazel and Peppas, 1995; Chen and Hoffman, 1995; Feil et al., 1992; Liang et al., 2004; Vakkalanka et al., 1996; Yong-Hee et al., 1994). These new classes of polymers are able to react to environmental changes and have found utility in intracellular drug delivery in which subtle pH differences across the endosomal membrane trigger the delivery of protein or DNA (Bajpai et al., 2010). Figure 6.10. Illustration of the physical response of a negative temperature-sensitive hydrogel when submitted to heat or cool processes. This chapter reports the synthesis/characterization of different classes of hydrogels and also the kinetics of their formation process. Time-evolution, in a batch reactor, of key properties, namely monomer conversion, weight fraction of gel, molecular weights and average radius of gyration of the soluble phase are assessed. The performance of the resulting hydrogels was also assessed (e.g. swelling ratio in different environmental conditions) in order to relate the synthesis conditions with the end-use properties of such materials. A general kinetic approach is used to predict the dynamics of hydrogel formation. Comparison between experimental measurements and these predictions allow the validation of a computational tool useful in the design of processes for production of tailored hydrogels. 6.2 Experimental Part 6.2.1 Materials N-isopropylacrylamide (NIPA) of 99 % purity, N,N-dimethylacrylamide (DMA) of 99 % purity stabilized with 500 ppm monomethyl ether hydroquinone (MEHQ), acrylic acid (AA) of 99 % purity stabilized with 180-200 ppm MEHQ, methacrylic acid (MAA) of 99 % purity stabilized with 250 ppm MEHQ, methylene bisacrylamide (MBAm) of 99 % purity, 1,6- Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-10 hexanediol diacrylate (HDDA) of 80 % purity stabilized with 100 ppm MEHQ, trimethylpropane triacrylate (TMPTA) stabilized with 100 ppm methylethylhydroquinone, tetraallyoxyethane (TAO), hydroquinone of 99 % purity, cyclohexane of 99 % purity, 2,2azobis(2-methylpropionamidine) dihydrochloride (V50) of 98 % purity, and the redox system ammonium persulfate (APS) of 98 % purity, tetramethylethylenediamine (TEMED) of 99 % purity, thioglycolic acid (CTA) with 99 % purity. All of them were purchased from SigmaAldrich. Acrylamide (AAM) with 98 % purity and ethyl cellulose (stabilizer) were purchased from Fluka. Span 60 and Span 80 (stabilizers) were purchased from Panreac. A current grade of liquid paraffin, toluene and sodium hydroxide (NaOH) were used when needed. All the products were used as received. 6.2.2 Polymerisation Set-up The aqueous solution homopolymerisation of acrylic acid was performed in batch reactor in order to have some insight concerning the kinetics of polymerisation of this monomer in conditions similar to those used to synthesize SAP materials. These polymerisations were inline monitored using a FTIR-ATR probe which allows the measurement of monomer conversion. The same experimental set-up, a reactor with maximum capacity of 2.5 L for which a detailed description has been presented in section chapter 2 was used to synthesize SAP materials based on acrylic acid/trimethylolpropane triacrylate copolymerisation. In order to extend the process to the post-gelation period, maintaining good agitation and heat dissipation conditions, these experiments were performed in inverse suspension. This technique is especially useful for gel production at isothermal conditions since keeping a good stirring of the reaction vessel is a key feature. In the continuous phase, toluene and/or cyclohexane were used as organic diluents and Span 60 and/or ethylcellulose as oilsuspending agents. The ratio oil-phase/water-phase in the reactor was 5/1 and the agitation speed of 400 rpm. Volume fraction of acrylic acid in the aqueous phase was VAA = 15 %. A similar feed composition was used in the solution homopolymerisation of acrylic acid. Polymerisations of smart hydrogels were performed in a batch reactor using also the inverse suspension process. Reactions were performed at low scale in a 200 mL reaction vessel and at higher scale in a 2.5 L reaction device (description in chapter 2). Stable suspensions were produced with a volumetric ratio aqueous/organic phases=0.2, 1 % (w/w) of surfactant in the continuous phase and agitation speed at 300 rpm. Argon sweeping was used to prevent inhibition of the polymerisation by oxygen. NIPA, DMA, AAM, AA and MAA were chose as 6-11 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation the main water soluble monomers. MBAm, TMPTA and HDDA were selected as crosslinkers. The influence of the kind of crosslinker (e.g. bifunctional/trifunctional) on the gelation process could therefore be observed. The pair APS/TEMED was used for redox initiation at low polymerisation temperature (e.g. 20 °C) whereas V50 was selected as a thermal initiator (polymerisations at 50 °C). Cyclohexane was chosen for the continuous phase in the inverse suspension polymerisations with Span 80 as stabilizer. Partial neutralization of AA/MAA was carried out using sodium hydroxide. Thioglycolic acid was chosen as a chain transfer agent (CTA). A few solution polymerisations were in-line monitored using an immersion probe exploiting FTIR-ATR detection. 6.2.3 Product Analysis by SEC/RI/MALLS At prescribed polymerisation times, reaction samples were collected from the reactor, quenched at low temperature in a solution containing hydroquinone to stop the reactions, and afterward prepared for injection of the soluble polymer in the SEC/RI/MALLS system. Dynamics of monomer depletion, weight fraction of gel, molecular weights and average radius of gyration were thus measured. The same equipment and procedures as in the previous chapters were used changing the eluent for water (with 200 ppm of sodium azide and 0.1 M of Na2HPO4 with pH=9) at T = 50 °C, flow rate of 0.5 ml/min and changing the size exclusion chromatography system. For this analysis a train of 3 SEC aqueous columns (Viskotek A2000 + Viskotek A3000 + Viskotek A6000) were considered to fractionate the polymers by size (different configurations were also used in order to not exceed the recommended maximum column pressures). Polymerisation samples collected from the inverse suspension reaction set-up at different polymerisation time were treated as above described and the products were characterized by SEC/RI/MALLS. Important features of the crosslinking process can be obtained by such analysis, as depicted in Figures 6.11-6.16. Measurements of the following properties were thus carried out:  Monomer conversions of the main water soluble vinyl monomers. Low crosslinker concentrations were used in this experimental program and the dynamics of their consumption could not be estimated.  Molecular weight distributions (and their averages) before and after gelation (soluble phase).  z-average radius of gyration before and after gelation (soluble phase). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-18 Table 6.2. Kinetic schemes considered in the modelling studies with the generic steps of radical polymerisation. Kinetic Step Chemical Equation Initiator Decomposition Acrylic Acid (M1) initiation Acrylamide (M2) initiation MBAm (M3) initiation Pendent Double Bonds (PDB) initiation Acrylic acid propagation Acrylamide Propagation MBAm propagation Pendant double bond propagation Chain transfer to acrylic acid Chain transfer to acrylamide Chain transfer to agent Termination by combination Termination by disproportionation Inhibition Inactive Products 6.3.3 Kinetic Parameters Despite the technical importance of poly(acrylic acid) and poly(methacrylic acid), scarce information concerning the kinetics of polymerisation of these monomers in aqueous phase could be found in the literature up to about one decade ago. In the last years, this challenging problem was studied by a few research groups aiming at measuring propagation and termination rate coefficients of these monomers, especially in aqueous phase and considering different synthesis conditions (showing the effect of temperature, concentration, pH, ionic strength, etc.). A short review concerning this issue is presented in Tables 6.3-6.6 where the 6-19 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation correspondent bibliographic sources and main remarks concerning the measurements are also described. Among other important issues, PLP-SEC measurements for water soluble monomers are difficult because SEC analysis cannot be carried out using THF as eluent and this well established technique is not directly feasible. This issue can be partially circumvented by performing a prior modification of poly(acrylic acid) generating the related methyl ester which can be afterwards analysed using the conventional PLP-SEC with THF as eluent. Table 6.3. Some values of the propagation rate constant ( ) of acrylic acid (AA) in water. (L mol-1s-1) T (°C) pH [M] Remarks ~500 to ~2000 a) Room 2.2-2.9 30 %wt AA Minimum at pH ~6.5 with conversion with ionic strength ~30000 to ~60000 b) ~40000 to ~70000 b) 20 25 N.N. N:N. 0.83-0.90 mol/L 0.90-1.37 mol/L ( ) c) 2.3-25 N.N. 30 %wt AA ( ) c) 2.3-28.5 N.N 40 %wt AA ( ) d) 2.5-19.5 N.N 1 %wt AA 2%, 5% AA also included ( ) d) 2.8-24.7 N.N. 3 %wt AA Maximum at ~3% AA ( ) d) 2.1-20.1 N.N. 10 %wt AA Solvent equality effects; Association of AA with polymer/radicals 112000 e) 82200 e) 16700 e) 59900 e) 6 6 6 6 10 %N. 50 %N. 95 %N. 110%N. 0.69 mol/L 0.69 mol/L 0.69 mol/L 0.69 mol/L [M] corresponds to ~5%wt AA. A model for the dependence of with neutralization is presented. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-20 Table 6.3. Continuation. Without estimation of , important kinetic data concerning the aqueous polymerisation of AA or AA/TMPTA is presented (Arriola et al., 1997; Cutié et al., 1997a and 1997b; Henton et al., 1997). The effect of neutralization of AA and solids content (dilution related) was extensively studied in these works in the temperature range 55 to 85 °C. R= 8.314 J mol-1K-1. a) Anseth et al., 1996; b) Kuchta et al., 2000; c) Lacík et al., 2001; d) Lacík et al., 2003; e) Lacík et al., 2004; f) Li and Schork, 2006; g) Cutié et al., 1997a; h) Kabanov et al., 1973 and 1975; i) Kurland, 1980. A similar procedure can be used with poly(methacrylic acid). Nevertheless, polymer modification can introduce non negligible errors in molecular weight measurements and so to rate coefficient estimates. A detailed discussion concerning this issue and other important features associated with measurements of rate coefficients of water soluble monomers can be found in the references presented in Tables 6.3-6.6. To sum up, propagation and termination rate coefficients, besides temperature, also depend (at least) on:  Monomer/solvent concentration ratio with non-ionic systems. A decrease of about one order of magnitude in was observed upon increasing monomer concentration.  Degree of ionization. At low monomer concentration, a decrease in of about one order of magnitude was measured when the degree of ionization was changed from 0 to 100 %.  Opposite variations were observed when the two effects (monomer concentration and ionization) are present: a weaker drop of with monomer concentration was found when the monomer is partially ionized. For a fully ionized monomer, increases when monomer concentration is also increased.  Occurrence of Trommsdorff effect is another issue complicating the kinetics of these polymerisation systems. (L mol-1s-1) T (°C) pH [M] Remarks 38667 f) 55 65 %N. 33% solids (AA) (~25 %wt AA) Estimated from experimental data in g) 650 h) 6600 h) 2500 h) 23 23 23 7.9 11 13.6 1.2 mol/L AA 1.2 mol/L AA 1.2 mol/L AA 5000 i) 50 N:N. bulk 6-21 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation It is therefore difficult to establish a fully reliable set of kinetic parameters valid for the different conditions to be considered in the synthesis of water soluble homopolymers based on acrylic or methacrylic acids or related superabsorbent hydrogels. The aim of the present work is the development of a simple kinetic model being able to capture the essential features of SAP production considering particular operation conditions with an associated set of kinetic parameters. Conversely this method can be used to predict the influence of the change of operating conditions (change in temperature, pH, and so on, and the concomitant change in the kinetic parameters) on the dynamics of gelation. These aspects are illustrated in the next section where simulations performed with different sets of kinetic parameters are presented. Table 6.4. Some reported values of the propagation rate constant ( of methacrylic acid (MMA) in water. Increase of with increasing MAA concentration for fully monomer ionization and T =40 °C is reported in Beuermann et al., 2007a. Ionic and hydrogen-bonded intermolecular interactions between the activated state and the molecular environment can be at the source of the variation of with monomer concentration and degree of ionization. Increase of with conversion was observed in Beuermann et al., 2007b. R=8.314 J mol-1k-1. a) Kuchta et al., 2000; b) Beuermann et al., 2006; c) Beuermann et al., 2007c; d) Buback et al., 2008b; e) Lacík et al., 2009; (L mol-1s-1) T (°C) pH [M] Remarks ( ) a) 18-89 N.N. 15 %wt MAA ( ) b) ( ) b) ( ) b) 20-80 23-80 25-60 N.N. N.N. N.N. 5 %wt MAA 45 %wt MAA 100 %wt MAA 15%, 30%, 60% MAA also included Intermolecular interactions Between H2O, MAA and transition state structures ( ) c) 18-89 N.N. 15 %wt MAA IUPAC benchmark 7700 d) 4900 d) 3300 d) 50 50 50 N.N. N.N. N.N. 10 % MAA 20 % MAA 30 % MAA conversion 6602 e) 4229 e) 2702 e) 2477 e) 60 60 60 80 70 % N. 70 % N. 70 % N. 100%N. 5 % MAA 20 % MAA 40 % MAA 40 % MAA Other measurements were performed. Variation of with T, [M] and N. Is available Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-22 Table 6.5. Some reported values of the termination rate constant ( ) of acrylic acid (AA) in water. a) Anseth et al., 1996; b) Renard and McKenna, 2000; c) Li and Shork, 2006; d) Cutié et al., 1997a; Table 6.6. Some reported values of the termination rate constant ( ) of methacrylic acid (MAA) in water. a) Beuermann et al., 2008; b) Buback et al., 2008b; 6.4 Results and Discussion 6.4.1 Simulation Results Using the Proposed Kinetic Approach Figure 6.17 depicts the predicted dynamics of the weight fraction of gel ( ) and monomer conversion ( ) during a batch SAP production. Simulations were carried out considering the synthesis with a trifunctional crosslinker (TMPTA used as case study) at an initial mole fraction in the monomer mixture (around the lower limit used in practice (Arriola et al., 2007)). In these simulations three different values of the rate coefficient for the homopropagation of acrylic acid ( ) were considered, in a range that is plausible for this (L mol-1s-1) T (°C) pH [M] Remarks to a) Very low values ( to 1) at high conversion (~0.9) and pH=6.5 Room 2.2-9.9 30 %wt AA Minimum at pH~6.5 with conversion with ionic strength to 40 b) 70 75 %N. Inverse suspension c) 55 65 %N. 33 % solids (AA) (~25 %wt AA) Estimated from experimental data ind) (L mol-1s-1) T (°C) pH [M] Remarks Decrease from to a) 50 N.N. 30%wt MAA with conversion P= 200 bar Decrease from to a) 50 N.N. 60 %wt MAA with conversion P= 200 bar a) , , a) N.N. Change of with conversion (p) for different initial fraction of MAA ( weight b) b) b) b) 50 N.N. Same expression for change of with p used in a) was considered 6-23 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation monomer in these particular conditions (see Table 6.3 and Li and Schork, 2006). An initial mole ratio between initiator and monomer and a rate coefficient for initiator unimolecular decomposition valid for V50 (2,2′-Azobis(2methylpropionamidine)dihydrochloride) at T = 50 °C have been assumed. Initial volumetric fraction of acrylic acid in the aqueous phase was set as . The reactivity ratio for acrylic acid/TMPTA was considered to be as previously measured for this chemical system in similar conditions (Arriola et al., 2007). Note that, for the sake of simplicity, polymer radicals derived from the two monomers were not distinguished in the present analysis and by consequence corresponds to an ideal copolymerisation (note that the reactivity ratio for TMPTA is very difficult to measure (Arriola et al., 2007)). The reactivity of pendant double bonds is another parameter of a very difficult experimental estimation (Arriola, et al., 2007). In simulations presented in Figure 6.17 the value was chosen, which corresponds to an equal reactivity of PDB and acrylic acid (under these conditions, reactivity of PDBs is higher than individual double bonds of TMPTA: . To the termination rate coefficient was assigned the value which is also in the range of the values considered in other works concerning also the kinetics of acrylic acid polymerisation (Li and Schork, 2006). Equality between initiation and propagation rate coefficients ( ) was also considered along this work. Results presented in Figure 6.17 show the important impact of the rate coefficient of acrylic acid (i.e. temperature, pH, concentration, ionic strength) in gel formation during superabsorbent hydrogels production. With higher values of , the primary chain length increases with concomitant higher gel formation. Figure 6.18 shows the predicted dynamics of the weight average molecular weight ( ) for the the same systems described in Figure 6.17. The ability of the present method to predict polymer properties before and after gelation becomes here evident. Besides MWD and its averages, z-average radius of gyration (Θ state) and sequences distributions can also be computed before and after gelation as before shown (Costa and Dias, 1994, 2003, 2005, 2006 and 2007; Dias and Costa, 2003, 2005a, 2005b, 2006, 2007 and 2010; Gonçalves et al., 2007, 2010a, 2010b, 2010c, 2010d and 2011a; Trigo et al., 2008). The effect of the reactivity of pendant double bonds on the predicted dynamics of the weight fraction of gel ( ) and monomer conversion ( ) for batch SAP production using a trifunctional crosslinker is presented in Figure 6.19. The following parameters Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-24 (according to the above discussion) were now fixed: , , = 15 %, , , 20000 L mol−1 s−1, = 5 × 107 L mol−1s−1. Figure 6.17. Predicted dynamics of the weight fraction of gel and monomer conversion in batch SAP production using a trifunctional crosslinker with . Figure 6.18. Predicted dynamics of the weight average molecular weight in batch SAP production using a trifunctional crosslinker with . 6-25 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation Figure 6.19. Prediction dynamics of the weight fraction of gel and monomer conversion in batch SAP production using a trifunctional crosslinker with . Figure 6.20. Predicted dynamics of the weight fraction of gel in batch SAP production considering different initial mole fraction of trifunctional crosslinker ( . Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-26 Figure 6.21. Predicted time evolution of monomer conversion and weight gel fraction during acrylic acid/triacrylate copolymerisation with . Other parameters considered in the simulations: , (T=50 °C for V50), = 15 %, , , 30000 L mol−1 s−1, = 5 × 107 L mol−1s−1, . Under these conditions, the reactivity of pendant double bonds (quantified by the parameter r* or equivalently by ) has a huge effect on the dynamics of gel formation. As expected, low gel content is predicted if the reactivity of PDBs is much lower than for the initial double bonds of the crosslinker (e.g. = 0.58) and increases with the value of this parameter. Due to the low content of crosslinker in the polymerisation system, the overall monomer conversion is almost insensitive to this parameter. Simulations like those presented in Figure 6.19 can be used to estimate the reactivity of PDBs using experimental measurements of the dynamics of gel formation. The effect of the initial mole fraction of crosslinker on the dynamics of the weight fraction of gel is illustrated in Figure 6.20. This parameter can be readily used to manipulate the properties of the final products, as depicted in that Figure. Simulations for ranging from the lower limit used in practice (around 0.0025 %) to ten times this value show the change of from around 0.4 to 1. Remaining parameters fixed in these simulations are: , (T=50 °C for V50), = 15 %, , , 20000 L mol−1 s−1, = 5 × 107 L mol−1s−1, . Operation with the higher limit of (around 0.25 %) is illustrated in Figure 6.21. Under these conditions gelation is 6-27 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation predicted to occur within some hundredths of seconds and the weight fraction of gel in the polymer rises very fast to around 1. However, in practice, polymerisation must be prolonged in order to reach high monomer conversion. The ability to numerical calculate the characteristics equations presented in Gonçalves et al., 2011b is a crucial step for the prediction of gel properties using generating functions of population balance equations, as proposed by Dias and Costa in the same paper (Gonçalves et al., 2011b). Development of very sharp numerical boundary layers is a special feature of these problems making very difficult the treatment of the associated two point boundary value problems (TPBVP), as discussed before (Costa and Dias, 2003 and 2005; Dias and Costa, 2003 and 2005b). These aspects are illustrated for the present chemical systems in Figure 6.22 and Figure 6.23. Note that the solution of this kind of problems in the context of this theory plays a role equivalent to the calculation of the extinction probabilities in the framework of the Theory of the Branching Processes presented in Gonçalves et al., 2011b. Figure 6.22. Numerical solution of the characteristics (vector s) correspondent to the SAP synthesis using the conditions described for Figure 6.21. For illustration purposes the chosen polymerisation time was t = 203.6 s. Another possible way to manipulate the dynamics of gelation is the choice of the functionality of the crosslinker, as depicted in Figure 6.24. Three different synthesis processes, correspondent to the use of crosslinkers with different functionalities (2, 3 and 4) by Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-34 behaviour was experimentally confirmed. Main features of this polymerisation system could therefore be captured by the simple kinetic model here considered. Other parameters considered in the simulations: (T=50 °C for V50), Lmol1s-1, , , , Lmol-1s-1. A comparison between the predictions of the above described kinetic models and experimental measurements are depicted in Figures 6.31 and 6.32. Good agreement between predictions/measurements could only be obtained if a very low reactivity of the pendant double bonds of the crosslinker is considered. In Figure 6.31 a good agreement between predicted and measured molecular weights is observed if a low reactivity of PDB is considered ( . In Figure 6.32 it was not possible to have a good description of the dynamics of weight fraction of gel for the same parameters. These discrepancies are likely to arise from the heterogeneity of network formation process, highlighting the effects of intramolecular cyclization, phase separation during the formation of the hydrogel or excluded volume effects, as previously pointed out (Dušek, 1996). 104 105 106 107 108 040 80 120 160 200 240 Predictions Before Gelation Predictions After Gelation Measurements Weight Average Molecular Weight Time (min) T= 50 oC 80 % Neutralization with NaOH [Acrylic Acid]=2.2 mol/L (~15 v %) [V50]=4.3x10-3 mol/L (~0.2 mol %) [MBAm]=0.1 mol% RPDB=0.1 Figure 6.31. Predicted and measured time evolution of the molecular weight of the soluble phase during inverse suspension hydrogels formation. Synthesis of AA/MBAm superabsorbent hydrogel was here considered for illustration purposes. 6-35 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation 0.0 0.2 0.4 0.6 0.8 1.0 040 80 120 160 200 240 Measurements Weight Fraction of Gel (wg) Time (min) T= 50 oC 80 % Neutralization with NaOH [Acrylic Acid]=2.2 mol/L (~15 v %) [V50]=4.3x10-3 mol/L (~0.2 mol %) [MBAm]=0.1 mol% Predictions with RPDB=0.1 Predictions with RPDB=0.025 Figure 6.32. Predicted and measured time evolution of the weight fraction of gel for the same system described in Figure 6.31. 6.4.3 Characterization of the Obtained Hydrogels The kinetics of polymerisation has a huge impact on the dynamics of gel formation. Moreover with the same chemical system (e.g. polymerisation involving AA or MAA), besides temperature or initial concentrations (monomers, initiators, etc.), the pH of the reaction media has an enormous influence on the dynamics of crosslinking process (Gonçalves et al., 2011b). Measurement of kinetics of polymerisation at different operation conditions is thereafter a key issue in the study of hydrogel synthesis. The typical synthesis conditions of FRP of these hydrogels are in Table 7.1 of chapter 7. A relation between the synthesis conditions and hydrogel performances can be assessed as depicted in Figures 6.33-6.36. Different pH-sensitive (Figures 6.33 and 6.35), temperaturesensitive (Figure 6.34) and superabsorbent hydrogels (6.36) were tested. Monomer chemical structures, initial proportion between monomers, initial amount of crosslinker, degree of neutralization (e.g. with AA/MAA) and polymerisation temperature are some parameters which were identified to have an important influence on the materials swelling ratio. Thus in Figure 6.33 is shown the usefulness of hydrogels as pH sensitive materials (e.g. drug delivery applications). The initial composition of the polymerisation system has an impact on the Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-36 performance of the stimuli-responsive gel beads. Figure 6.34 highlight the high response to temperature of PNIPA hydrogel in comparison with a non-temperature-responsive hydrogel (PDMA). In figure 6.35 can be observed the stimulation by the pH of the environmental media, namely when comparing acidic media with basic media. A swelling increase by around 8 times triggered by pH change from 1.2 to 7.5 was measured for this hydrogel. Equilibrium swelling ratio of around 900 can be reached with these particular synthesis conditions as presented in Figure 6.36. 0 5 10 15 20 25 30 35 0 1 2 3 4 5 6 Weight Swelling Ratio (SR) Time (hr) pH= 1.2 7.5 1.2 7.5 1.2 7.5 Figure 6.33. Repeated swelling (pH=7.5 and collapsing pH=1.2) of inverse suspension synthesized AA/AAM/MBAm hydrogels. Actually the issue of network heterogeneity is important in different chemical systems and several different experimental techniques have been used to try to elucidate this phenomenon (Dušek, 1996; Duchet and Pascault, 2003). Analysis by SEM is a useful technique here illustrated in Figure 6.37. Despite the care to avoid artefacts from interaction with electron beam (Dušek, 1996; Duchet and Pascault, 2003), formation of particle populations with very different sizes are discernible (at microand nano-scales), leading to believe into the heterogeneity of the formation process of the hydrogels here studied. 6-37 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation 0 1 2 3 4 5 6 15 20 25 30 35 40 NIPA/MBAm HYDROGEL DMA/MBAm HYDROGEL Hydrogel Volume (cm3) Temperature (oC) Figure 6.34. Observed volume change of NIPA/MBAm and DMA/MBAm inverse suspension synthesized hydrogels stimulated by temperature. 0 5 10 15 20 25 30 35 40 010 20 30 40 50 60 pH=1.2 pH=7.5 pH=10.1 pH=13 Weight Swelling Ratio (SR) Time (min) High Gel stimulation by pH Figure 6.35. Measured dynamics of swelling of an inverse suspension synthesized pH-responsive hydrogel (AA/AAM/MBAm) in buffer solutions of different values of pH. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 6-38 0 200 400 600 800 1000 030 60 90 120 150 180 Super-absorbent Hydrogel 1 Super-absorbent Hydrogel 2 Super-absorbent Hydrogel 3 Weight Swelling Ratio (SR) Time (min) Figure 6.36. Examples of the measured swelling kinetics of different superabsorbent hydrogels showing the effect of synthesis conditions on the materials performance. (a) (b) Figure 6.37. Gel micrographs of inverse suspension synthesized hydrogel beads. (a) AA/MBAm hydrogel. (b) DMA/MBAm hydrogel. 6.5 Conclusions Comparison between predictions of the proposed kinetic method and measurements performed in the acrylic acid/trimethylolpropane triacrylate copolymerisation show that this simple model is able to capture the main features of this polymerisation system. The 6-39 CHAPTER 6. Inverse-Suspension Free-Radical Polymerisation Leading to hydrogels Formation dynamics of hydrogel formation by terpolymerisation of (AA/AAM, MBAm) was investigated through their production in a batch reactor operating with inverse-suspension. A general kinetic approach was applied to the modelling studies concerning the synthesis of these materials. Comparison of experimental data with predictions of the developed kinetic models put into evidence the likely heterogeneity of the network formation process (e.g. loop formation reactions and phase segregation). Additional experimental work with a much lower initial mole fraction of crosslinker (and also different kinds of crosslinkers) should be performed in order to confirm the good foundations of this theory. Estimates of the reactivity of pendant double bonds of the crosslinkers can also eventually be obtained by comparison of these new experimental results with model predictions. CHAPTER 7 INVERSE-SUSPENSION REVERSIBLE ADDITIONFRAGMENTATION RADICAL POLYMERISATION LEADING TO HYDROGELS FORMATION Abstract. Temperature and pH stimuli–responsive hydrogel particles were synthesized using inverse-suspension polymerisation (with a few runs also carried out with solution polymerisation) in a batch stirred reactor. Different water soluble co-monomers were present in the initial mixture (e.g. N-isopropylacrylamide and acrylic acid) as well as crosslinkers with different functionalities. Commercially available RAFT agents 4-cyano-4-phenyl carbonothioylthio-pentanoic acid (CPA), 2-(dodecylthiocarbonothioylthio)-2-methylpropionic acid (DDMAT) and cyanomethyl dodecyl trithiocarbonate (CDT) were alternatively used. Thus along the experimental program were changed/assessed the following variables:  Parameters changed:  Kind and/or monomers concentrations.  Kind and/or crosslinker concentrations.  Kind and/or RAFT agents concentrations.  Kind of solvent and monomer dilution (aqueous system/organic solvent).  Initiation system.  Degree of neutralization.  Products characterization:  SEC with a tetra-detector array (RI+UV+IVDP+LS) running with aqueous eluent allowing the determination of average molecular weights, conversion, zaverage radius of gyration and viscosity.  In-line FTIR-ATR for estimation some kinetic parameters.  Swelling ratio changes of hydrogels with pH and/or temperature.  Drug delivery testing of hydrogels triggered by environmental changes. This chapter is based on the following publication: M.A.D. Gonçalves, V.D. Pinto, R.A.S. Costa, R.C.S. Dias, J.C. Hernández-Ortiz, M.R.P.F.N. Costa, Macromol. Symp. 333 (2013) 41-54. 7-9 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation isolated networks samples collected at different polymerisation times seems to lead to a better description of the crosslinking process, namely concerning the pendant double bonds reactivity (Hecker, 2000). Crosslinker amount used in hydrogels preparation is very low (a few percent) and even with off-line FTIR monitoring the study of the crosslinking process is a difficult task. 13C labelling of the crosslinker (as before performed with trimethylolpropane triacrylate (TMPTA) in the framework of network in superabsorbent gels (Arriola et al., 1997)) should be a better option in this context. 7.2.5 Swelling Ratio Sensitivity Measurements Synthesized hydrogel particles, after isolation, were tested in order to assess their sensitivity to stimulations triggered by changes in the surrounding media. In particular, was measured the variation of the hydrogels weight swelling ratio, in water solutions, at different temperatures and/or pH values. Stimulation of the networks by changes in these parameters was thus observed. Typical results obtained are presented in Figure 7.7 where sensitivity of NIPA/MBAm hydrogels to temperature changes is used as illustration example. 0.0 0.2 0.4 0.6 0.8 1.0 5 6 7 8 9 10 RI Signal RALS Signal IVDP Signal Normalized Response Elution Volume (mL) FRP synthetized Poly(acrylic acid) RALS Signal RI Signal IVDP Signal Figure 7.3. Refractive index (RI), right angle light scattering (RALS) and intrinsic viscositydifferential pressure (IVDP) signals simultaneously observed in the SEC analysis of a water soluble PAA sample. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7-10 0.0 0.2 0.4 0.6 0.8 1.0 6 7 8 9 10 11 RAFT PAA FRP PAA RAFT PMAA Normalized Response Elution Volume (mL) IVDP Signal FRP PAA RAFT PAA RAFT PMAA Figure 7.4. IVDP signals observed in the SEC analysis of different water soluble polymers synthesized, highlighting the influence of operation conditions (FRP/RAFT) on the products molecular structure and properties. Figure 7.5. In-line FTIR-ATR spectra observed during DMAEMA/EGDMA FRP polymerisation (run 1 in Table 7.3). Absorption peak at around 935 cm-1 was considered to estimate the double bonds conversion, using also the peak at 1720 cm-1 as internal reference. 7-11 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation 0.0 0.2 0.4 0.6 0.8 1.0 010 20 30 40 50 60 In-Line FTIR-ATR Measurement Gravimetric Final Monomer Conversion Double Bond Conversion Time (min) DMAEMA/EGDMA polymerization in Paraffin at 60 0C Figure 7.6. FTIR-ATR estimated dynamics of monomer conversion for DMAEMA/EGDMA polymerisation (run 1 in Table 7.3). Similar measurements were performed with runs 2 and 3 in Table 7.3 but even lower monomer conversions were obtained in these experiments (almost negligible after 8 hours of polymerisation in run 3). 0 5 10 15 20 25 30 15 20 25 30 35 40 45 NIPA/MBAm Hydrogel with 1% (mol/mol) MBAm NIPA/MBAm Hydrogel with 2% (mol/mol) MBAm Weight Swelling Ratio (SR) Temperature (oC) Figure 7.7. Measured equilibrium swelling ratio of NIPA/MBAm hydrogels in aqueous solutions at different temperatures illustrating networks sensitivity to changes in this parameter. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7-12 7.2.6. Drug Release Testing The performance of the produced hydrogels was also assessed by considering drug delivery applications. Different model drugs (caffeine, 5-fluorouracil, isoniazid and ibuprofen) were considered in these studies which were carried out pouring pre-incubated network particles in aqueous solutions at different conditions (changing pH/temperature). Drug release was measured by UV detection in aqueous samples collected at different elapsed times. Typical results are presented in Figure 7.8 using the caffeine release from a NIPA/MAA/MBAm hydrogel at different pH/temperature conditions as illustration example. Two different surrounding water solutions were considered: pH=1/T=37 °C (collapsed particles) and pH=7/T=22 °C (swollen particles). Hydrogel beads were pre-loaded with caffeine during 48 hr in a 2.25 mg/ml drug water solution. In spite of the complexities associated with the mathematical modelling for drug delivery (Bajpai et al., 2010; Galaev and Matiasson, 2008), a good agreement is observed by fitting the experimental data to exponential rise laws (the effect of different stimulations on the drug release profiles observed is also here highlighted). 30 40 50 60 70 80 90 100 060 120 180 Release at 22 oC and pH=7 Release at 37 oC and pH=1 Caffeine Released (mg caffeine / g of dry hydrogel) Time (min) NIPA/MAA/MBAm Hydrogel (67/33/1 mol/mol/mol) Figure 7.8. Dynamics of caffeine release from a pH/temperature sensitive synthesized hydrogel (NIPA/AA/MBAm) measured by UV detection at 270 nm. Two different surrounding water solutions were considered: pH=1/T=37 °C (collapsed particles). And pH=7/T= 22 °C (swollen particles). Drug loading was performed by swelling the hydrogel beads in caffeine aqueous solution during 48 hours. 7-13 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation Applications of the different classes of "smart" hydrogels here studied are further depicted in Figures 7.9-7.14. In Figure 7.9 is presented the measured equilibrium weight swelling ratio of anionic (AA based) and cationic (DMAEMA based) hydrogels in aqueous solutions at different pH values. These results illustrate the networks sensitivity to changes of this parameter. Note that an inverse effect of the pH on the swelling ratio of these hydrogels can be explored to trigger different macroscopic effects, as for instance the transition between shrunk to swollen networks by changing the pH from 1 to 8 (e.g. resembling the stomach/intestine pH change in human body) with AA hydrogels and the opposite with DMAEMA hydrogels. In Figure 7.10 is showed the comparison for the change of the equilibrium weight swelling ratio with pH considering FRP and RAFT synthesized AA hydrogels. These results illustrate the high effect of the synthesis technique used on the swelling properties of the hydrogels. In fact, the primary chain length of the networks is strongly affected when FRP is replaced by RAFT. This effect can eventually be used to tune the swelling properties of the hydrogels (e.g. designing the initial ratios between monomers/RAFT agent/initiator). It is worth to note that measurements presented in Figures 7.9 and 7.10 were obtained using buffer aqueous solutions at different pH values. 0 5 10 15 20 25 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 AA Hydrogel DMAEMA Hydrogel Weight Swelling Ratio (SR) pH Figure 7.9. Measured equilibrium weight swelling ration of anionic (AA based) and cationic (DMAEMA based) hydrogels in aqueous solutions at different pH values illustrating networks sensitivity to changes in this parameter. Inverse effect of the pH on the swelling ratio of these hydrogels can be exploited to trigger different macroscopic effects (e.g. shrunk to swollen networks by changing the pH from 1 to 8 with AA hydrogels and the opposite with DMAEMA hydrogels). Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7-14 These buffer solutions were prepared using the proper amounts of HCl, NaOH, KCl, KHP (potassium hydrogen phthalate - C8H5KO4), KH2PO4 (potassium dihydrogen phosphate), Na2B4O710H2O (borax) and Na2HPO4 (disodium hydrogen phosphate). Nevertheless, the swelling ratio of hydrogels is also strongly dependent on the ionic strength and size of the ions and counter-ions present in the used aqueous solutions. Accordingly, a different dependence of the hydrogels swelling ratio on pH changes can be observed if other aqueous solutions at the same pH values are considered (e.g. changing the used salts and/or using just HCl and NaOH to prepare the aqueous solutions with the desired pH values). 0 5 10 15 20 25 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 FRP AA/MBAm Hydrogel RAFT AA/MBAm Hydrogel Weight Swelling Ratio (SR) pH Figure 7.10. Comparison of the change of the equilibrium weight swelling ratio with pH for FRP and RAFT synthesized AA hydrogels. High effect of the synthesis technique used on this parameter is observed. The primary chain length of the networks is strongly affected when FRP is replaced by RAFT and this effect can eventually be used to tune the swelling properties of the hydrogels. Note that results presented in Figures 7.9 and 7.10 were obtained using buffer aqueous solutions at different pH values. The swelling ratio of hydrogels is also strongly dependent on the ionic strength and size of the ions/counter-ions present in the solutions. A different dependence of SR can be observed if other aqueous solutions at the same pH values are considered (e.g. changing the used salts). Measured dynamics of release of different drugs from cationic and anionic hydrogels is illustrated in Figures 7.11-7.14. In Figure 7.11 and 7.12 is showed the dynamics of release of 5-fluorouracil from DMAEMA based (cationic) and AA based (anionic) hydrogels, respectively. In both cases, the release of the drug was measured in acidic (pH=1) and alkaline (pH=10) aqueous solutions. In spite of the differences between the two hydrogels, 7-15 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation slightly higher steady state release of the drug was always observed with the alkaline environment. 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0240 480 720 960 1200 1440 pH=1 pH=10 Concentration of 5Fu in the aqueous solution (mg/g) Time (min) Dynamics of release of 5Fu from a DMAEMA hydrogel placed in aqueous solutions at different pH values (acidic (pH=1) and alkaline (pH=10) conditions) Figure 7.11. Dynamics of release of 5-fluoruracil from a pH sensitive hydrogel (cationic hydrogel based on DMAEMA) measured by UV detection at 270 nm. 0.00 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0240 480 720 960 1200 1440 pH=1 pH=10 Concentration of 5Fu in the aqueous solution (mg/g) Time (min) Dynamics of release of 5Fu from a AA hydrogel placed in aqueous solutions at different pH values (acidic (pH=1) and alkaline (pH=10) conditions) Figure 7.12. Dynamics of release of 5-fluoruracil from a pH sensitive hydrogel (anionic hydrogel based on AA) measured by UV detection at 270 nm. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7-16 0.00 0.02 0.04 0.06 0.08 0.10 0240 480 720 960 1200 1440 DMAEMA hydrogel AA hydrogel Concentration of ibuprofen in the aqueous solution (mg/g) Time (min) Dynamics of release of ibuprofen from DMAEMA and AA hydrogels placed in aqueous solution at alkaline conditions (pH=10) Figure 7.13. Comparison of the dynamics release of ibuprofen from cationic (DMAEMA based) and anionic (AA based) hydrogels, both placed in aqueous solution at pH=10 (release measured by UV detection at 223 nm). 0 10 20 30 40 50 60 70 80 90 100 0240 480 720 960 1200 1440 FRP Hydrogel, pH=1 FRP Hydrogel, pH=10 RAFT Hydrogel, pH=1 RAFT Hydrogel, pH=10 Fraction of 5Fu released to the aqueous solution (%) Time (min) Dynamics of release of 5Fu from FRP and RAFT AA Hydrogels Figure 7.14. Dynamics of release of 5-fluorouracil from FRP and RAFT synthesized pH sensitive hydrogels illustrating the effect of the molecular architecture of the networks on their performance. Amount of drug released is here expressed as the fraction of drug loaded in the hydrogel that is transferred to the aqueous solution (release measured by UV detection at 270 nm). 7-17 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation Note that some other complex effects such as hydrogel/drug interaction (Bajpai et al., 2010; Galaev and Matiasson, 2008) (e.g. see discussions about diffusion and chemically controlled delivery systems in chapter 11 of Galaev and Mattiasson, 2008) should be taken into account when drug release studies are performed, as for instance formation of complexes between drugs and polymer networks. These issues also have a strong effect on the amount of a specific drug that is possible to load in a hydrogel. The effect of the combination between specific drugs and hydrogels is illustrated in Figure 7.13 where the dynamics of release of ibuprofen from DMAEMA and AA polymer networks, both placed in aqueous solution at pH=10 (release measured by UV detection at 223 nm), is showed. Comparison of the dynamics of release of 5-fluorouracil from FRP and RAFT synthesized pH sensitive hydrogels is illustrated in Figure 7.14. The amount of drug released in this case is expressed as the fraction of drug loaded in the hydrogel that is transferred to the aqueous solution (release measured by UV detection at 270 nm). Note that much lower release fractions were observed (both at pH=1 and 10) when RAFT hydrogels were considered. In all cases presented in Figures 7.11-7.14, drugs loadings were performed by swelling the hydrogels in 5-fluorouracil or ibuprofen aqueous solutions during 48 hr. These results should be a consequence of the different molecular architectures associated with FRP and RAFT networks (affecting namely their swelling ratio, as showed in Figure 7.10) and highlight the relation between structure and end use properties of these materials. 7.3 Results and Discussion Reversible addition-fragmentation chain transfer polymerisation is probably the most versatile CRP technique allowing the polymerisation of different classes of monomers. Nevertheless, the degree of control of polymerisation that is attained with RAFT is strongly dependent on the reaction conditions used. Specific combination between monomer, RAFT agent, initiator and solvent used in the polymerisation is a central issue to obtain tailored products with RAFT polymerisation. Temperature and initial proportions monomer/RAFT agent/initiator/solvent also have a huge effect on the kinetics of formation and on the control of the molar masses of RAFT polymers. Some other issues arise when RAFT is directly performed in water, namely the low solubility of most RAFT agents in pure water (forcing the use organic co-solvents) and their potential hydrolysis (pH dependent) with loss of control on the polymerisation process (see Chaduc et al., 2012a and references therein). When aqueous dispersed systems are considered (e.g. the Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 7-18 industrially important emulsion/miniemulsion/suspension processes and their inverse counterparts) some other aspects like the transport of reactants (monomers, initiators, RAFT agents) between organic and aqueous phases become also of crucial importance. In this context, the RAFT inverse miniemulsion of acrylamide and acrylic acid were recently reported (Ouyang et al., 2011; Qi et al., 2007) and the effect of pH on the hydrolysis of the RAFT agent and polymerisation in the continuous phase (eventually in the absence of RAFT agent) were identified as phenomena potentially involved in some loss of control observed with particular conditions. A secondary peak was observed in the RI curve (see discussion below in the context of the results here presented) which was attributed to different polymer populations formed in both phases (aqueous and organic). These aspects get an additional importance in the framework of RAFT dispersed systems that has been very recently explored to produce amphiphilic copolymers and nanoparticles/nano-objects with different morphologies like spheres, fibres and vesicles (Chaduc et al., 2012b; Zhang et al., 2011, 2012a and 2012b). Below are discussed some of the findings involving probably related mechanisms that are present in the RAFT inverse suspension formation of hydrogels or their linear counterparts. Very fast reactions are generally involved in the FRP synthesis of hydrogels, as before showed with different classes of monomers (Gonçalves et al., 2011b and 2013b). Fast gelation is observed with a few percent of crosslinker agent and parameters such as polymerisation temperature, monomer concentration and neutralization have a very strong influence on the dynamics of gelation (see results presented in Gonçalves et al., 2011b and 2013b). Similar behaviour was observed in the FRP polymerisation runs detailed in Table 7.1, even considering a low polymerisation temperature (T=20 °C). It is known that more amenable kinetics of polymerisation can be achieved on replacing FRP by RAFT polymerisation. Design of operation conditions, namely the initial proportions between initiator/RAFT agent/monomer, can be used to manipulate reaction rates and also to design the degree of polymerisation. With network formation, these parameters can be used to try the manipulation of the primary chain length (thus affecting gelation) and the minimization of intramolecular reactions (cyclizations) leading to the decrease of crosslinking efficiency. If a direct aqueous polymerisation is intended, as in many cases involving hydrogels, a major problem to be faced with RAFT polymerisation is the low water solubility of usual RAFT agents. This issue applies to DDMAT and therefore the aqueous RAFT polymerisation is not 7-25 CHAPTER 7. Inverse-Suspension Reversible Addition-Fragmentation Radical Polymerisation Leading to Hydrogels Formation Hydrolysis of the RAFT agent and/or polymerisation both in aqueous and organic phases (with formation of controlled and non-controlled polymer populations) should be the key mechanisms, also with a potential effect on the loss of control in these polymerisation systems, as discussed above. Additional theoretical/experimental studies should help in the interpretation of the experimental observations here presented. 7.4 Conclusions A comprehensive experimental program concerning the synthesis, characterization and testing (e.g. for drug delivery applications) of smart hydrogels was performed. Combining different aqueous compatible monomers, pH and temperature sensitive materials were obtained. Anionic, cationic and amphoteric network gel beads were produced using the inversesuspension technique. The kinetics of their building process was studied by SEC with a tetra detection array and also using in-line FTIR-ATR. Only limited information on the crosslinking process was possible to obtain with in-line FTIR-ATR monitoring, even using optimized conditions for IR measurements (bulk monomer polymerisation in a dispersed media with low IR absorbance). The Possibility of probe coating during the synthesis process and the low crosslinker content associated to hydrogels production are factors affecting negatively the use of such technique in this context. Use of off-line FTIR analysis of isolated polymers (Gonçalves et al., 2013c; Hecker, 2000), chemical analysis of pendant double bonds (Gonçalves et al., 2013a; Hecker, 2000) or 13C labelling of the crosslinker (Arriola et al., 1997) should provide improved information concerning the crosslinking process. Additional work in this research should hopefully also elucidate if really important gains in the sensitivity of these hydrogels can be achieved through replacement of FRP by RAFT polymerisation. Complementation of the experimental work here reported with new theoretical developments on the kinetic description of gels formation by the RAFT process (Gonçalves et al., 2013c) should also be explored, in this way, it is expected to improve simulation tools with the purpose of helping the specification of synthesis conditions leading to tailored advanced materials. Several issues concerning RAFT polymerisation were discussed in the conclusions of chapter 5 and are equally applied when using aqueous polymerisation systems. CHAPTER 8 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 8-2 This work has carried out a study on the synthesis and characterization of hyperbranched polymers, gels and hydrogels arising from the radical polymerisation of multivinyl monomers. The experimental program has considered the following aspects:  Radical polymerisation techniques used:  Free radical polymerisation (FRP).  Atom transfer radical polymerisation (ATRP).  Nitroxide mediated radical polymerisation (NMRP).  Reversible addition-fragmentation radical polymerisation (RAFT).  Different kinds of monomer/products families:  For organic system:  Styrenics.  Acrylates.  Methacrylates.  For aqueous system:  Acrylic and methacrylic acids copolymers.  Acrylamides.  Reaction systems:  Solution/bulk.  Suspension.  Inverse-suspension.  Polymerisation reactors:  Batch (atmospheric and pressurized).  Semi-batch.  Tested parameters in the polymerisations:  Temperature.  Monomers concentration.  Crosslinker concentration and functionalization.  Initiation system (thermal decomposition and redox).  Concentration/kind of polymerisation controlling agents.  Products characterization:  Molecular architecture of the products by:  GPC/RI/MALLS. 8-3 CHAPTER 8. Conclusions and Suggestions for Future Work  GPC/RI/MALLS/IVDP/UV.  Gravimetric analysis (conversion and gel fraction).  Titrimetric analysis of pendant double bonds.  In-line and off-line FTIR-ATR.  Products tested in terms of:  Swelling ratio.  Controlled drug release. These experimental results were used to test simulation tools to predict:  Gelation and relative amount of gel.  Molecular masses before and after gelation.  Radius of gyration before and after gelation.  In the absence/presence of intramolecular cyclizations. The obtained products were characterized through size exclusion chromatography systems. Comprising RI and MALLS detectors only (non-aqueous polymers) and with a tetra-detector array for hydrophilic polymers. With these systems it was possible to obtain detailed information about the molecular architecture of the synthesized materials. Namely, the number, weight and z average molecular weights ( , , ) as well as the average radius of gyration ( ) of the products were measured. Besides these average values, molecular weights distributions and the variation of absolute molecular weight and radius of gyration with the elution volume were also measured. A differential refractometer was used to measure the required refractive index increment (dn/dc) of polymers, monomers and solvents. The gravimetric method was used to determine the monomer conversion and the gel fraction. The use of in-line FTIR-ATR has also allowed the determination of the monomer conversion (with some limitations) and has confirmed the formation of different kinds of polymer populations. In batch mode, FTIR allowed qualitative information of pendant double bonds. The concentration of pendant double bonds was also measured by the iodine chloride addition method. The morphology of the gel beads was observed by scanning electron microscopy. The networks structure has a strong influence on the final product properties. The materials obtained by conventional radical polymerisation (FRP) have high molecular weight dispersion indices leading to a network with a broad distribution of molecular weights of its internal chains. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 8-4 Important improvements were made by changing the feed policy of the reactants, using semibatch reactors and chain transfer agents (Gonçalves et al., 2007). Controlled radical polymerisation techniques (ATRP, NMRP and RAFT) lead to a higher homogeneity of the networks, decreasing the effect of intramolecular cyclizations and permitting the increase of the amount of effective crosslinks. However, as assessed in this work, the actual networks are far from ideality, and further studies must still be performed to better understand their structure from a quantitative point of view. The use of water compatible monomers to produce superabsorbent polymers and smart hydrogels is known to be useful for applications in biomedical, pharmaceutical and environmental industries. It was shown that hydrogel networks are potentially useful in controlled drug delivery by taking advantage of the unique properties of the produced polymers. Such “smart” hydrogels can respond to stimulation on pH, temperature, ionic strength, among others leading to changes of retention of the active encapsulated species. For the conventional radical polymerisation (FRP) presented in chapter 2 it was shown that the experimental results here presented concerning the semi-batch solution polymerisation of styrene and divinylbenzene showed the possibility of control gelation through the design of appropriated feeding policies (Gonçalves et al., 2007). Nevertheless, some shortcomings of this approach should be here pointed out. Lack of reproducibility of the polymerisations was found during this research. This problem is likely a consequence of non-constant feeding rates due to the peristaltic pumps used. Note that even a small error in the feeding flow rates has a high effect on crosslinking process, especially when divinyl monomer is to be feed to the polymerisation mixture. In principle, this shortcoming can be avoided using higher precision pumps (e.g. syringe pumps/GPC pumps). Semi-batch STY/DVB as here studied cannot also be extended to the production of polymer particles in a process similar to suspension polymerisation (a process later on explored in this research). Emulsion semi-batch polymerisation of vinyl/multivinyl monomers is an alternative to the production of these kinds of particles but much more complex mechanistic issues are expected in the framework of such heterogeneous processes. Inter-phase transport of both monomers (vinyl/multivinyl) and radical compartmentalization are some phenomena leading to the need of development of a much more complex analysis when emulsion semi-batch crosslinking polymerisation is to be considered. 8-5 CHAPTER 8. Conclusions and Suggestions for Future Work A detailed kinetic model taking into account some complexities of these polymerisation systems has been developed in the absence of intramolecular reactions. The modelling studies can be used to design new feed policies with impact on the properties of the synthesized materials. For the copolymerisation of MMA with EGDMA in toluene at 60 °C using batch operation a general kinetic approach was applied. The decrease of the reactivity of pendant double bonds measured in this work is consistent with earlier studies with the same chemical system (Landin et al., 1988; Li et al., 1989a and 1989b). For experiments at low monomer conversion (less to 0.5) a good agreement between predictions and experimental measurements is observed. Predicted weight average molecular weights and z-average radius of gyration are too low at higher monomer conversions with nonlinear systems (Trigo et al., 2008). In chapter 2 the suspension copolymerisation of styrene/divinylbenzene with gel formation was also described. It was shown that the main features of this crosslinking process can be captured by the kinetic model developed using as a single fitting parameter the relative reactivity of the pendant double bonds. Under these circumstances, with the same set of kinetic parameters, it is possible obtain good predictions of the dynamics of before and also after gelation, which is a major contribution of this work for the polymer reaction engineering of such processes. Nevertheless, the unrealistic low values of reactivity estimated (around 6 % of styrene reactivity) should be mostly a consequence of neglecting intramolecular cyclizations in the model used (Gonçalves et al., 2011a). Important differences were observed in the molecular architecture of non-linear products synthesized either by FRP or by ATRP as shown in chapter 3. Improved microgel homogeneity can be obtained by ATRP of acrylate/diacrylate monomers. Comparatively to FRP, ATRP also allows the operation with higher crosslinker mole fraction without gel formation. These results can be especially useful to obtain hyperbranched polymers at higher conversions than with FRP. The effect of cyclizations on the molecular architecture of these products was also detected: ATRP allows the production of more homogeneous hyperbranched polyacrylates (less intramolecular cyclizations) than FRP but, even with ATRP, this mechanism becomes increasingly important with higher dilution (Gonçalves et al., 2010a, 2010b, 2010c and 2010d). It was also shown that, besides molecular weights, the z-average radius of gyration can also be predicted with reliability in the framework of the present kinetic approach. The good foundations of this simulation tool make it useful to design hyperbranched polymers with tailored properties. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 8-6 It was shown in chapter 4 that an unified theoretical kinetic approach is capable of dealing with NMRP and captures the most important aspects of this system. The sensitivity analysis studies performed show that the reactivity of pendant double bonds is the controlling factor for the formation of non-linear structures. This single parameter was used in the fitting studies. For the NMRP of STY/DVB in solution polymerisation system, reduced reactivities of PDB were estimated in comparison to the vinyl monomers. This technique showed that, when compared with FRP, NMRP allows only limited control over the crosslinking process. NMRP non-linear polymerisation proceeds in a similar way to a random FRP process and, therefore, a polymer population with a very broad distribution of sizes is formed (Gonçalves et al., 2010c and 2010d). Moreover, the occurrence of intramolecular cyclizations was detected by SEC/RI/MALLS of NMRP STY/DVB products by comparing experiments at different monomer dilutions. Therefore, such reactions cannot be completely avoided by using NMRP, especially at a low starting monomer concentration. The incidence of intramolecular cyclization reactions with NMRP of STY/DVB was also detected through the quantification of PDBs polymer concentration, which was measured using chemical analysis. Hence this work has confirmed that the NMRP technique allows the synthesis of hyperbranched polymers with improved homogeneity, owing to less intramolecular cyclizations as compared to conventional radical polymerisations. Moreover, it was also shown that NMRP allows operation with higher divinyl monomer content without gelation, which is an important issue in the production of hyperbranched polymers. Polymerisations runs were extended past the gel point also as discussed in chapter 4 (in aqueous suspension system) and so the dynamics of gel formation could be studied. The impact of certain operation parameters on the crosslinking process was assessed: the initial amount of crosslinker (DVB) content was changed between 0 and 100 %, initial monomer dilution, global monomer concentration in the organic phase was changed between 20 and 100 % v/v, polymerisation temperature, experiments at 90 and 130 °C were carried out. Previous results for STY/DVB gel formation at 60 °C were also considered for comparison purposes and polymerisation mechanism: different runs comparing FRP with NMRP were carried out through the change of the initial ratios of initiator/monomer and mediator/initiator (Gonçalves et al., 2013a). 8-7 CHAPTER 8. Conclusions and Suggestions for Future Work Results obtained in the chapter 4 are especially useful in the study of the effect of intramolecular cyclization reactions on the crosslinking process. Experimental data obtained (Gonçalves et al., 2013a) were explored to develop mathematical models including the effect of primary cyclization based on the development of balance of sequences connecting radical centers and pendant double bonds present in the same polymer chain (Aguiar et al., 2013a). The rate constant for cyclization was considered a function of the sequence length. Studies included the preand post-gelation periods and comparisons between experimental measurements and model predictions for pendant double bond concentration, average molecular weights and weight fraction of gel were thus performed (Aguiar et al., 2013b). It was possible the estimation of the rate constant of cyclization for the smallest ring (3 monomeric units). A value in the order of 500 s-1 was estimated for this parameter at 90 C rather high when compared with the intermolecular propagation (as [M+PDB] lies between 1000 and 10000 s-1). Jacobsen-Stockmayer theory predicts that pairs of radical site + PDB in the same molecule with longer distances between them show a decreasing rate of cyclization, but their impact stays strong even at bulk conditions and is a major factor in all multivinyl radical polymerisations. Influence of key polymerisation parameters (e.g. crosslinker content, monomer dilution, reaction temperature) on cyclization was also identified through the comparison of experimental data with different kinds of mathematical models (Gonçalves et al., 2013a; Aguiar et al., 2013a and 2013b). In these studies a much higher relative crosslinking reactivity was observed at 130 C as compared to 90 C, likely as an effect of the chain mobility (Aguiar et al., 2013b). Despite the efforts here reported for the clarification of these kinds of crosslinking processes, this is an open issue in the scientific community, as testified by other very recent publications on this field (Scott et al., 2014; Nikitin et al., 2013; Hamzehlou et al., 2013). The impact of the use of different kinds of polymerisation mechanisms (e.g. FRP/NMRP) on crosslinking (trying to avoid/decrease cyclizations) and development of mechanistic models able to describe such complex polymerisations are some research lines also explored in these publications which will be probably enhanced in future works. Experimental work of chapter 5 has been focused in RAFT polymerisation with the trithiocarbonates DDMAT and CDT lead to the formation of a secondary polymer population with low concentration but high molecular size. This secondary population is only residually discernible when the RI signal is observed but becomes evident when the MALLS signal is inspected. Formation of this side population is not observed when RAFT agent dithiobenzoate Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 8-8 TBTGA is used, confirming the influence of the chemical structure of the RAFT agent on the degree of control of the polymerisation that is achieved in RAFT processes (Gonçalves et al., 2013c). Recent developments on pulsed-laser assisted techniques, namely Single–Pulse/Pulsed– Laser–Polymerisation/Time-Resolved Electron-Paramagnetic-Resonance (SP-PLP-EPR) (Barth and Buback, 2010) should also be considered in the discussion of the RAFT results obtained in chapters 5 and 7. In fact, using this technique it is possible the experimental study of polymerisation systems with more than one kind of radicals as for instance acrylates with secondary radicals or tertiary midchain radicals (appearing due to backbiting of secondary radicals). Moreover, this technique can be used to enlighten the problem concerning the different types of radicals involved in RAFT polymerisation and associated kinetic mechanisms. In fact, using this technique, rate coefficients of addition and fragmentation (and the correspondent equilibrium constants) were directly measured for butyl acrylate RAFT polymerisation using a trithiocarbonate (Meiser et al., 2013). Other issues of RAFT polymerisation were also very recently addressed by the same research group when studying the kinetics of dithiobenzoate-mediated methyl methacrylate polymerisation (Sidoruk et al., 2013). In this latter paper, different explanations for retardation with thiobenzoates were analysed, namely the following four mechanisms debated by scientific community:  Slow fragmentation of the RAFT intermediate radical.  Intermediate radical termination.  Intermediate radical termination followed by “missing step”  Cross-termination restricted to small radicals. It was found that the concentration of the intermediate radicals is reduced compared with retarded systems and that cross-termination occurs to a weaker extent in the framework of the chemical system studied. Significant amounts of “missing step” products were also observed (Sidoruk et al., 2013). It was also concluded that slow addition and “missing step” reactions contribute to the absence of rate retardation in dithiobenzoate mediated MMA polymerisation (Sidoruk et al., 2013). Thus, experimental results of Buback group seem to show that the effect of slow fragmentation mechanisms should be small in RAFT polymerisation. Possibly, the secondary peak observed in the RAFT experiments performed in this work should be a consequence of the formation of 3-arm bridges during the polymerisation process, leading to 8-9 CHAPTER 8. Conclusions and Suggestions for Future Work the formation of high size polymer species. New theoretical developments allowing the modeling of such complex mechanism should be sought in order to validate this hypothesis. In chapter 6 it is shown that the impact of intramolecular cyclizations in superabsorbent hydrogels production should be weak (in contrast to styrene/divinylbenzene copolymers and similar) due to the very low content of crosslinker used and to the high values of the primary chain lengths involved. Polymerisation temperature, degree of neutralization of monomers (when AA or MAA are used), initial amount of crosslinker and other features of initial composition of the polymerisation system (e.g. ratios between main vinyl monomers) are some operation parameters with a strong effect on end-use properties of the synthesized hydrogels (Gonçalves et al., 2011b and 2013b). Gelation kinetics in aqueous systems is faster when compared with organic systems. The produced materials were characterized in terms of swelling ratio and their environmental response to temperature and/or pH changes. These studies here presented are hopefully improvements for the development of tools concerning the design of production processes of tailored hydrogels. In chapter 7 a comparison between the synthesis processes of these kinds of networks using FRP and RAFT polymerisation (considering three different commercially available RAFT agents) was experimentally performed. Important dissimilitudes in the dynamics of gel formation were identified when FRP mechanism is replaced by the RAFT polymerisation. Some peculiarities of the RAFT synthesis were studied (e.g. solubility of the used RAFT agents in water and monomers considered, effect of initial composition/kinetics on gelation) in order to find proper conditions to produce the sought hydrogels using this CRP technique. The use of multiple detection (especially including light scattering) in SEC analysis proved to be very important in order to obtain a rigorous characterization of products molecular architecture. In fact, the formation of an unexpected higher size secondary polymer population was detected by light scattering when RAFT polymerisation is used, even with mono-vinyl monomers (Gonçalves et al., 2013d). It would seem that these results could be be explained in the framework of recent theoretical findings describing the RAFT slow fragmentation model leading to bimodal distributions formation (Zapata-González et al., 2011), but this is in contradiction with other experimental evidence as above discussed. Therefore, other mechanisms causing loss of control such as hydrolysis of the RAFT agent and/or simultaneous polymerisation in aqueous and organic phases should also be considered in further developments on this research line. 9-5 References Beuermann, S., Buback, M., Davis, T.P., Gilbert, R.G., Hutchinson, R.A., Olaj, O.F., Russel, G.T., Schweer, J., van Herk, A.M., 1997. Critically evaluated rate coefficients for free-radical polymerization .2. Propagation rate coefficients for methyl methacrylate. Macromol. Chem. Physic. 198, 1545-1560. Beuermann, S., Buback, M., 2002. Rate coefficients of free-radical polymerization deduced from pulsed laser experiments. Prog. Polym. Sci. 27, 191-254. Beuermann, S., Buback, M., Hesse, P., Lacík, I., 2006. Free-radical propagation rate coefficient of non-ionized methacrylic acid in aqueous solution from low monomer concentrations to bulk polymerisation. Macromolecules 39, 184-193. Beuermann, S., Buback, M., Hesse, P., Kukučková, S., Lacík, I., 2007a. Propagation kinetics of free-radical methacrylic acid polymerisation in aqueous solution. The effect of concentration and degree of ionization. Macromol. Symp. 248, 23-32. Beuermann, S., Buback, M., Hesse, P., Kukučková, S., Lacík, I., 2007b. Propagation rate coefficient of non-ionized methacrylic acid radical polymerisation in aqueous solution. The effect of monomer conversion. Macromol. Symp. 248, 41-49 Beuermann, S., Buback, M., Hesse, P., Gilbert, R.G., Kuchta, F-D., Lacík, I., van Herk, A.M., 2007c. Critically evaluated rate coefficients for free-radical polymerisation. Part 6: Propagation rate coefficient of methacrylic acid in aqueous solution. Pure and Applied Chemistry 79, 1463-1469. Beuermann, S., Buback, M., Hesse, P., Hutchinson, R.A., Kukučková, S., Lacík, I., 2008. Termination kinetics of the free-radical polymerisation of non-ionized methacrylic acid in aqueous solution. Macromolecules 41, 3513-3520. Bevington, J.C., 1955. The sensitized polymerization of styrenethe rate and efficiency of initiation. Transactions Faraday Society 51, 1392-1397. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 9-6 Biasutti, J.D., Davis, T.P., Lucien, F.P., Heuts, J.P.A., 2005. Reversible additionfragmentation chain transfer polymerization of methyl methacrylate in suspension. J. Polym. Sci. Pol. Chem. 43, 2001-2012. Bodugöz, H., Güven, O., 2002. The synthesis of nonporous poly(isobutyl methacrylate) microspheres by suspension polymerization technique and investigation of their swelling properties. J. Appl. Polym. Sci. 83, 349-356. Boschmann, D., Vana, P., 2007. Z-RAFT star polymerizations of acrylates: Star coupling via intermolecular chain transfer to polymer. Macromolecules 40, 2683-2693. Brandrup, J., Immergut, E.H., Grulke, E.A., Abe, A., Bloch, D.R., 1999. Polymer Handbook, 4th edition. John Wiley & Sons. Braunecker, W. A., Matyjaszewski, K., 2007. Controlled/Living radical polymerization: Features, developments, and perspectives. Prog. Polym. Sci. 32, 93-146. Brazel, C.S., Peppas, N.A., 1995. Synthesis and characterization of thermomechanically and chemomechanically responsive poly(N-isopropylacrylamide-co-methacrylic acid) hydrogels. Macromolecules 28, 8016-8020. Britton, D., Heatley, F., Lovell, P.A., 1998. Chain transfer to polymer in free-radical bulk and emulsion polymerisation of vinyl acetate studied by NMR spectroscopy. Macromolecules 31, 2828-2837. Buback, M., Gilbert, R.G., Hutchinson, R.A., Klumperman, B., Kuchta, F.D., Manders, B.G., O’Driscoll, K.F., Russel, G.T., Schweer, J., 1995. Critically Evaluated Rate Coefficients for Free-Radical Polymerization .1. Propagation Rate Coefficient for Styrene. Macromol. Chem. Physic. 196, 3267-3280. Buback, M., Junkers, T., 2006. Termination kinetics of tert-butyl methacrylate and of nbutyl methacrylate free-radical bulk homopolymerizations. Macromol. Chem. Physic. 207, 1640-1650. 9-7 References Buback, M., Hesse, P., Junkers, T., Sergeeva, T., Theis, T., 2008a. PLP labelling in ESR spectroscopy analysis of secondary and tertiary acrylate propagating radicals. Macromolecules 41, 288-291. Buback, M., Hesse, P., Hutchinson, R.A., Kasák, P., Lacík, I., Stach, M., Utz, I., 2008b. Kinetics and modeling of free-radical batch polymerization of nonionized methacrylic acid in aqueous solution. Ind. Eng. Chem. Res. 47, 8197-8204. Buchholz, F.L., Graham, A.T. (Eds.), 1997. Modern superabsorbent polymer technology. New York: Wiley.VCH, ISBN 978-0-471-19411-8. Buchholz, F.L., Peppas, N.A., 1994. Superabsorbent Polymers: Science and Technology. ACS symposium Series, Washington. Búrdalo. J., Medrano. R., Saiz. E., Tarazona. M.P., 2000. A simple method to determine unperturbed dimensions of polymers using size exclusion chromatography and multi angle light scattering. Polymer 41. 1615-1620. Butté, A., Storti, G., Morbidelli, M., 2000. Miniemulsion living free radical polymerization of styrene. Macromolecules 33, 3485-3487. Butté, A., Storti, G., Morbidelli, M., 2001. Miniemulsion living free radical polymerization by RAFT. Macromolecules 34, 5885-5896. Carothers, W.H., 1936. Polymers and polyfunctionality. Transactions Faraday Society 32, 39-49. Chaduc, I., Lansalot, M., D’Agosto, F., Charleux, B., 2012a. RAFT polymerisation of methacrylic acid in water. Macromolecules 45, 1241-1247. Chaduc, I., Girod, M., Antoine, R., Charleux, B., D’Agosto, F., Lansalot, M., 2012b. Batch emulsion polymerisation mediated by poly(methacrylic acid) macro RAFT agents: One-pot synthesis of self-stabilized particles. Macromolecules 45, 5881-5893. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 9-8 Chatzi, E.G., Kammona, O., Kiparissides, C., 1997. Use of midrange infrared optical-fiber probe for the on-line monitoring of 2-ethylhexyl acrylate/styrene emulsion copolymerization. J. Appl. Polym. Sci. 63,799-809. Chen, G.H., Hoffman, A.S., 1995. Graft-copolymers that exhibit temperature-induced phase-transitions over a wide-range of pH. Nature 373, 49-52. Chen, X., Shan, G., Huang, J., Huang, Z., Weng, Z., 2004. Synthesis and properties of acrylic-based superabsorbent. J. Appl. Polym. Sci. 92, 619-624. Cheyney, L.V.E., Kelley, E.J., 1942. Unsaturation of synthetic rubberlike materials. Ind. Eng. Chem. 34, 1323-1326. Choi, C.H., Kertesz, M., 1997. Conformational information from vibrational spectra of styrene, trans-stilbene, and cis-stilbene. J. Phys. Chem. A 101, 3823-3831. Coelho, J.F.J, Silva, A.M.F.P, Popov, A.V., Persec V., Abreu, M.V., Gonçalves, P.M.F.O., Gil, M.H., 2006a. Syntehsis of poly(vinyl chloride)-b-poly(n-butyl acrylate)-b-poly(vinyl chloride) by the competitive single-electron-transfer/degenerative-chain-transfer-mediated living radical polymerisation in water. J. Polym. Sci. Pol. Chem. 44, 3001-3008. Coelho, J.F.J, Carreira, M., Popov, A.V., Gonçalves, P.M.F.O., Gil, M.H., 2006b. termal and mechanical characterization of poly(vinyl chloride)-b-poly(n-butyl acrylate)-bpoly(vinyl chloride) by single electron transfer-degenerative chain transfer mediated living radical polymerisation in water. Eur. Polym. J. 42, 2313-2319. Coelho, J.F.J, Carreira, M., Gonçalves, P.M.F.O., Popov, A.V., Gil, M.H., 2006c. Processability and characterization of poly(vinyl chloride)-b-poly(n-butyl acrylate)-bpoly(vinyl chloride) prepared by living radical polymerisation of vinyl chloride. Comparison with a flexible commercial resin formulation prepared with PVC and dioctyl phthalate. Journal of. Vinyl & Additive Technology 12, 159-165. 9-9 References Coelho, J.F.J, Carvalho, E.Y., Marques, D.S., Popov, A.V., Persec., V., Gil, M.H, 2008. Influence of the isomeric structures of butyl acrylate on its single-electron transferdegenerative chain transfer living radical polymerisation in water catalyzed by Na2S2O4. J. Polym. Sci. Pol. Chem. 46, 6542-6551. Coelho, J.F.J, Gois, J.R., Fonseca, A.C., Carvalho, R.A., Popov, A.V., Persec., V., Gil, M.H., 2009. Synthesis of poly(2-methoxyethyl acrylate) by single electron transferdegenerative chain transfer living radical polymerisation catalyzed by Na2S2O4 in water. J. Polym. Sci. Pol. Chem. 47, 4454-4463. Coelho, J.F.J, Fonseca, A.C., Gois, J.R., Gonçalves, P.M.F.O., Popov, A.V., Gil, M.H., 2011a. Scaling-up of poly(vinyl chloride) prepared by single electron transfer degenerative chain transfer mediated living radical polymerisation in water media: Low molecular weight. Kinetic analysis. Chem. Eng. J. 169, 399-413. Coelho, J.F.J, Fonseca, A.C., Gonçalves, P.M.F.O., Popov, A.V., Gil, M.H., 2011b. Particle features and morphology of poly(vinyl chloride) prepared by living radical polymerisation in aqueous media. Insight about particle formation mechanism. Polymer 52, 2998-3010. Colthup, N.B., Daly, L.H., Wiberley, S.E., 1990. Introduction to infrared and raman spectroscopy, third ed., Academic Press, Boston. Cooper, A.I., Hems, W.P., Holmes, A.B., 1998. Synthesis of cross-linked polymer microspheres in supercritical carbon dioxide. Macromol. Rapid Comm. 19, 353-357. Cooper, A.I., Hems, W.P., Holmes, A.B., 1999. Synthesis of highly cross-linked polymers in supercritical carbon dioxide by heterogeneous polymerization. Macromolecules 32, 2156-2166. Costa, M.R.P.F.N., Dias, R.C.S., 1994. A general kinetic analysis of non-linear irreversible copolymerizations. Chem. Eng. Sci. 49, 491-516. Synthesis and Properties of Products of Radical Polymerisation of Multivinyl Monomers 9-10 Costa, M.R.P.F.N., Dias, R.C.S., 2003. Prediction of Sol Fraction and Average Molecular Weights after Gelation for Non-Linear Free Radical Polymerizations Using a Kinetic Approach. Macromol. Theor. Simul. 12, 560-572. Costa, M.R.P.F.N., Dias, R.C.S., 2005. An improved General Kinetic Analysis of NonLinear Irreversible Polymerizations. Chem. Eng. Sci. 60, 423-446. Costa, M.R.P.F.N., Dias, R.C.S., 2006. Kinetic modeling of non-linear polymerization. Macromol. Symp. 242, 72-82. Costa, M.R.P.F.N., Dias, R.C.S., 2007. Prediction of Mean-Square Radius of Gyration of Tree-Like Polymers by a General Kinetic Approach. Polymer 48, 1785-1801. Coutinho, F.M.B., Rabelo, D., 1992. Scanning electron-microscopy study of styrene divinylbenzene copolymers. Eur. Polym. J. 28, 1553-1557. Crompton, T.R., Reid, V.W., 1963. Unsaturation in high impact polystyrene: analytical studies. J. Polym. Sci. Part A 1, 347-355. Cunningham, M., 2008. Controlled/living radical polymerization in aqueous dispersed systems. Prog. Polym. Sci. 33, 365-398. Cutié, S.S., Henton, D.E., Powell, C., Reim, R.E., Smith, P.B., Staples, T.L., 1997a. The effects of MEHQ on the polymerization of acrylic acid in the preparation of superabsorbent gels. J. Appl. Polym. Sci. 64, 577-589. Cutié, S.S., Smith, P.B., Henton, D.E., Staples, T.L., Powell, C., 1997b. Acrylic acid polymerisation kinetics. J. Appl. Polym. Sci. Pol. Phys 35, 2029-2047. Destarac, M., 2010. Controlled radical polymerization: industrial stakes, obstacles and achievements. Macromolecules 4, 165-179. [Document text truncated for crawler view.]