DESIGN AND ENGINEERING OF BIOBASED MATERIALS – PROCESS ENGINEERING & THERMAL RECYCLING OF POLY(LACTIC ACID), AND STUDIES IN FUNCTIONAL SILANE AND SILOXANES By Xiangke Shi A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemical Engineering-Doctor of Philosophy 2014 ABSTRACT DESIGN AND ENGINEERING OF BIOBASED MATERIALS – PROCESS ENGINEERING & THERMAL RECYCLING OF POLY(LACTIC ACID), AND STUDIES IN FUNCTIONAL SILANE AND SILOXANES By Xiangke Shi Poly (lactic acid) (PLA) is a biobased and biodegradable polymer that is manufactured from plant-biomass resources. It is one of the few biobased polymers with excellent mechanical properties and clarity that can compete successfully with current fossil-based polymers in the marketplace. PLA can be recycled or composted (biodegradable under composting conditions) to provide for an environmentally responsible end-of-life option. Therefore, PLA is finding increasing use in food and single use disposable packaging and in industrial products where it offers significant value from economical as well as environmental perspectives. PLA melt is characterized by a relatively low melt viscosity that prevents it from being readily processed in blown film unit operations. In general, blown film processing requires relatively high melt viscosity as well as nonlinear viscoelastic properties (strain hardening). In this study, we have synthesized a new modified PLA molecule containing reactive epoxy groups by reacting PLA with chain extenders (CE). Addition of this new reactive PLA molecule at low levels (5% to 10%) in to base PLA resin significantly improved the melt strength and processability of PLA into blown films. The reaction mechanism operating in this process was investigated. The rheological properties of the CE/PLA products were studies as a function of temperatures, process conditions and CE concentrations to provide fundamental data and processing parameters for successful blown films operation. A viable end-of-life option for PLA is chemical recycling back to monomer – a virtual cycle of monomer to polymer and back to monomer – a circular biobased economy. Today’s industrial processes are based on the ring opening polymerization of the lactide monomer. Current approaches to PLA recycle is to hydrolyze it to lactic acid, purify it and then reform into lactide which can then enter into the polymerization step. However, we have shown that the polymerization of lactide to PLA follows a reversible kinetic model. We have used this reversible polymerization to recycle PLA to lactide monomer using catalytic thermal depolymerization with success. The mechanism and the rate of lactide formation from PLA as a function of time, temperature, and catalyst concentration using thermogravimetric analysis (TGA) were studied. The non-equilibrium depolymerization process leads to high yields of the desired lactide from PLA. The experimental depolymerization data fit a two-step reaction mechanism described by the Avrami equation. The model thus obtained provided all the critical parameters affecting this recycling process. Based on these data, a pilot plant of PLA recycling via thermal depolymerization is proposed. The sizing, energy balance, material balance and economic analysis were considered and included in this model recycling. Finally, water soluble hydroxyl alkyl terminated silanes and polysiloxanes were prepared from the reaction of ethylene carbonates (or glycerol carbonate) and 3-aminopropylalkoxylsilanes. The kinetics of the hydrolysis reaction and subsequent condensation reactions under acidic conditions were studied in details using 29Si NMR. Conventional polymerization of these hydroxyl alkyl terminated silanes yielded linear, branched or resinous siloxy polymers via selfpolycondensation through exchange of the ethoxy groups attached to the silicon atom with the terminal hydroxyl groups. The synthesis of the silane monomers as well as the structure and key properties of the hydroxyl alkyl terminated siloxy polymers were investigated by DSC, 1H NMR, FTIR, TGA and GPC. ACKNOWLEDGEMENTS I would like to thank my advisor Dr. Ramani Narayan for supporting me with my projects, for providing me the opportunity to work on many interesting projects, and for giving me the freedom to discover my interests during my PhD. I would like to thank Dr. Daniel Graiver, who has been advising and supporting me throughout my PhD. A special thanks to him for guiding me not only in the scientific field but also in career development and daily life. I would also like to thank Mr. Ken Farminer, for his valuable support and encouragement and the frequent chats about world history. I would also like to thank Dr. Yuya Tachibana, for his valuable guidance and advices. I would like to thank the members of Biobased Materials Research Group for helping me in the many ways they did. Thanks for every member of the group, I had a wonderful time working, learning and growing during my Ph.D. Also, my thanks to Northern Technologies International Corporation Inc, especially, Dr. Shilpa Manjure for the supporting of my projects. I am deeply thankful for the advice and support offered by my Ph.D. committee: Dr. Gregory Baker, Dr. Andre Lee, Dr. Krishnamurthy Jayaraman, and Dr. Laurent Matuana during the course of my work. I am grateful for all the help and support from my family: my parents, my grandparents, and especially my wife, Cheng Lu. I could not imagine my life if my wife did not travel with me to United States. iv It is impossible to name all in this tiny space. Many people I have come across during my work have taught me one thing or the other. I would like to thank all those people who have enriched my life and knowledge. v TABLE OF CONTENTS LIST OF TABLES........................................................................................................ ix LIST OF FIGURES...................................................................................................... xi KEY TO SYMBOLS OR ABBREVIATIONS ........................................................ xiv Chapter 1. Introduction and Background .................................................................. 1 1.1 Fossil resources and their impact ........................................................................... 1 1.2 Biobased and Biodegradable Materials ................................................................. 2 1.3 Organization of this dissertation ............................................................................ 4 Chapter 2. PLA melt strength enhancement for blown film application ................. 7 2.1 Introduction ............................................................................................................ 7 2.2 Experimental .......................................................................................................... 8 2.3 Results and Discussion ........................................................................................ 11 2.3.1 Model compound reaction study................................................................... 11 2.3.2 Epoxy functionalized PLA and gel content analysis .................................... 13 2.3.3 Molecular Weight Characterizations ............................................................ 14 2.3.4 Rheological Properties .................................................................................. 17 2.4 Conclusions .......................................................................................................... 22 2.5 Future work .......................................................................................................... 23 Chapter3. PLA blown film and plasticization with Poly(ethylene glycol) ............. 24 3.1 Introduction .......................................................................................................... 24 3.2 Materials and experiments ................................................................................... 27 3.3 Results and discussions ........................................................................................ 30 3.3.1 Molecular weight effect on plasticization performance................................ 30 3.3.2 Effect of plasticizer on crystallization rate of PLA ...................................... 33 3.3.3 Chain end trans-esterification reaction and its effect on molecular weight .. 34 3.3.4 Chain end reaction and selectivity reaction with chain extenders ................ 35 3.3.5 Effect of plasticization on blown film mechanical properties ...................... 38 vi 3.4 Conclusions .......................................................................................................... 39 3.5 Future work .......................................................................................................... 40 Chapter 4. High temperature mold induced crystallization for increased Heat Deflection Temperature performance of PLA-Talc compound .............................. 41 4.1 Introduction .......................................................................................................... 41 4.2 Experimental Section ........................................................................................... 45 4.3 Results and Discussion ........................................................................................ 46 4.3.1 Molecular weight and thermal properties of different grades of PLA .......... 46 4.3.2 TGA analysis of compounded product for thermal degradation and filler content determination............................................................................................. 48 4.3.2 DSC analysis of Compounded product for crystallization kinetics .............. 49 4.3.3 HDT analysis and crystallinity of Compounded product for two different mold temperature ............................................................................................................ 50 4.4 Conclusions .......................................................................................................... 53 Chapter 5. Depolymerization Kinetics of Poly(lactic acid) to Lactide ................... 54 5.1 Introduction .......................................................................................................... 54 5.2 Study of depolymerization reaction kinetics using TGA ..................................... 56 5.2.1 Experimental section ..................................................................................... 57 5.2.2 Results and Discussion ................................................................................. 58 5.2.2.1 Equilibrium monomer constant and rate of depolymerization ............... 58 5.2.2.2 Thermogravimetric analysis ................................................................... 63 5.2.2.3 Mechanism of the Depolymerization Reaction ...................................... 65 5.2.2.5 Theoretical Model of Depolymerization ................................................ 67 5.2.3 Conclusions ................................................................................................... 72 5.3 Recycle process design using thermal depolymerization .................................... 73 5.3.1 Sizing ............................................................................................................ 73 5.3.2 Process design ............................................................................................... 73 5.3.3 Materials Balance.......................................................................................... 75 5.3.4 Energy balance .............................................................................................. 76 5.3.5 Conclusions and Future works ...................................................................... 79 5.4 Effect of the present of plastic mixture on the recycling process ........................ 79 5.4.1 Experimental ................................................................................................. 80 vii 5.4.2 Results and Discussion ................................................................................. 81 5.4.3 Conclusions ................................................................................................... 83 Chapter 6. The Use of Glycerol Carbonate in the Preparation of Highly Branched Siloxy Polymers............................................................................................................ 84 6.1 Introduction .......................................................................................................... 84 6.2 Materials and methods/ Experimental Procedure ................................................ 86 6.3 Results and Discussion ........................................................................................ 87 6.3.1 Ring-opening urethane bond formation of aminosilanes with glycerol carbonate ................................................................................................................ 87 6.3.2 Monomer-Polymers Equilibrium by 1H NMR.............................................. 91 6.3.3 Viscosity and Thermal Properties ................................................................. 92 6.3.4 Polymerization by Rearrangement ................................................................ 94 6.4 Conclusions .......................................................................................................... 99 Chapter 7. Hydrolysis and Condensation of Water Soluble Alkoxysilanes Under Acidic Conditions ...................................................................................................... 100 7.1 Introduction ........................................................................................................ 100 7.2 Experimental Section ......................................................................................... 102 7.3 Results and Discussion ...................................................................................... 105 7.3.1 Synthesis and Characterization of AB2 monomer ...................................... 105 7.3.2 Hydrolysis/Condensation under very mild acidic conditions ..................... 109 7.3.3 Hydrolysis/Condensation under acidic conditions...................................... 111 7.3.4 Interdependency of water and acid concentrations on the hydrolysis/ condensation reactions ......................................................................................... 113 7.4 Conclusions ........................................................................................................ 125 BIBLIOGRAPHY ..................................................................................................... 126 viii LIST OF TABLES Table 2. 1 Gel content analysis by Soxhlet Extraction ................................................ 13 Table 2. 2 Effect of CE content on PLA molecular weight ......................................... 14 Table 3. 1 Glass transition temperature of extruded PLA/PEG blends .................... 31 Table 3. 2 Comparison of crystallization rate with and without plasticizers ............ 33 Table 3. 3 Mechanical properties comparison of PLA blown films ........................... 37 Table 4. 1 Molecular weights of PLA samples 3001 D, 3051D and 3251 D ............... 45 Table 4. 2 Molecular weight determination of PLA 3001D, 3051D, and 3251D ....... 46 Table 4. 3 Processing Condition and Talc Content Analysis ...................................... 47 Table 4. 4 DSC study of crystallization behavior of three grades of PLA. ................ 48 Table 4. 5 Effect of mold temperature on the degree of crystallization and HDT .... 51 Table 5. 1 Equilibrium monomer constants and the corresponding rate constants of the depolymerization reaction at different temperatures ........................................... 58 Table 5. 2 Reaction temperature and catalyst concentration of PLA degradation in the TGA experiments ...................................................................................................... 63 Table 5. 3 Vapor pressure of lactides at 220℃ ............................................................ 74 Table 5. 4 Specifications and operating conditions for depolymerization reactor and condensers ........................................................................................................................ 75 Table 5. 5 Summary of materials balance for depolymerization process .................. 76 Table 5. 6 Summary of the energy balance of the process .......................................... 77 Table 5. 7 Recycling yields of the lactide in PLA-PC and PLA-PBAT mixtures ...... 82 ix Table 6. 1 Kinetics, viscosity and yield of hydroxyl terminated polysiloxy branched polymers: MESiGC, DESiGC, and TESiGC ................................................................ 89 Table 7. 1 Sample composition for hydrolysis and condensation reactions ............ 103 x LIST OF FIGURES Figure 2. 1 DSC spectrum of PEG-CE mixture and steric acid-CE mixture ............ 12 Figure 2. 2 Reaction scheme between Epoxy based CE and PLA .............................. 13 Figure 2. 3 (A) Scheme of multiple PLA chains connected to a single CE molecule (B) Scheme of one PLA chain connected to one CE molecule ........................................... 16 Figure 2. 4 GPC spectrum of neat PLA and chain extended PLA with 0.25%, 0.5, 1% CE content prepared from EFPLA. .............................................................................. 17 Figure 2. 5 Comparison of complex viscosity frequency sweep at 180℃ between neat PLA, chain extended PLA with 0.25%, 0.5, 1% CE content from EFPLA and chain extended PLA with 1% CE content from one step processing. .................................. 18 Figure 2. 6 Comparison of complex viscosity frequency sweep at 170℃, 180℃, 190℃, and 200℃ between neat PLA, chain extended PLA prepared from EFPLA, and LDPE. ........................................................................................................................................... 19 Figure 2. 7 Comparison of extensional viscosity 180℃ at a rate of 1rad/s between neat PLA, chain extended PLA prepared from EFPLA, and LDPE. ................................ 20 Figure 2. 8 Strain hardening behavior of modified PLA containg 1% CE at different shear rates (5 rad/s, 2 rad/s, 1 rad/s, and 0.5 rad/s). .................................................... 21 Figure 2. 9 Comparison of strain hardening behavior of Neat PLA and modified PLA containing 1% CE at 5rad/s ........................................................................................... 22 Figure 3. 1 Synthesis of Biobased PEG from biomass ................................................. 27 Figure 3. 2 Screw configuration of the ZSK 30 twin-screw extruder. ....................... 29 Figure 3. 3 Glass transition temperature plot of extruded PLA/PEG blends ........... 32 Figure 3. 4 GPC plot of PLA and plasticized PLA ...................................................... 35 Figure 3. 5 GPC plot of PLA, one step sample, and two-step sample ........................ 37 Figure 4. 1 Serving temperature of food vs HDT temperture of PLA ....................... 42 Figure 4. 2 Stereo Isomers of lactic acid and lactide ................................................... 43 Figure 5. 1 Recycle processes of PLA ............................................................................ 55 xi Figure 5. 2 Arrhenius plot of depolymerization reaction constant kd ........................ 61 Figure 5. 3 Lactide concentration of PLA containing 0.6% catalyst as a function of time at different temperatures: 160℃(Δ), 180℃(○), and 200℃(□). ............................ 62 Figure 5. 4 Depolymerization of PLA containing 0.6% catalyst at different temperatures .................................................................................................................... 63 Figure 5. 5 TGA decomposition of PLA containing different concentrations of catalyst at 200℃............................................................................................................................. 65 Figure 5. 6 Proposed two-step reaction during the depolymerization of PLA in the presence of catalyst ......................................................................................................... 67 Figure 5. 7 Avrami plot of PLA depolymerization TGA data at containing different concentrations of catalyst at 200℃: 1% (■), 2% (▲), 4% (♦), and 10% (●) ............. 69 Figure 5. 8 Correlation of the overall reaction constant ln(K) with the catalyst concentration ln(C1) : 160℃(▲), 180℃(♦), and 200℃(■) ........................................... 70 Figure 5. 9 Correlation of the overall reaction constant ln(K) with the depolymerization reaction constant ln(kd) at 0.6% catalyst concentration .............. 71 Figure 5. 10 Correlation of the Avrami exponent with reaction temperature .......... 72 Figure 5. 11 Scheme of the depolymerization reactor equipped with top mounted distillation column and material flow chart ................................................................. 74 Figure 5. 12 TGA thermograms of PLA-PC blend with different ratio (A: PLA80/PC20, B: PLA70/ PC30, C: PLA60/PC40, D: PLA40, PC60) isotherm at 200℃ with 1% catalyst concentration ..................................................................................... 81 Figure 6. 1 Partial FTIR spectra of the reaction between DESi and GC as a function of time. .............................................................................................................................. 88 Figure 6. 2 GC concentration as a function of time and second-order kinetics (insert) of the urethanization reaction with MESiGC (solid triangle), DESiGC (solid square), or TESiGC (solid circle) at 22 °C. ................................................................................. 89 Figure 6. 3 Preparation of hydroxyl terminated silanes .............................................. 91 Figure 6. 4 1H NMR spectra (500 MHz, DMSO-d6) of hydroxyl terminated silanes 92 Figure 6. 5 TGA of MESiGC (solid line), DESiGC (dotted line), and TESiGC (dashed line) ................................................................................................................................... 94 Figure 6. 6 Condensation polymerization reaction scheme (DESiGC) ...................... 95 xii Figure 6. 7 Viscosity of DESiGC during polymerization at 80°C and 50Pa .............. 96 Figure 6. 8 Glass transition temperature as a function of polymerization time at 80°C and 50Pa ........................................................................................................................... 97 Figure 6. 9 GPC charts of MESiGC (left) and right DESiGC (right) during polymerization at 80°C and 50Pa .................................................................................. 98 Figure 7. 1 Reaction scheme of 3-Aminopropylmethyldiethoxysilane with ethylene carbonate ....................................................................................................................... 105 Figure 7. 2 1H NMR spectrum of the hydroxyl terminated silane AB2 monomer .. 106 Figure 7. 3 Path A – polymerization to hyperbranched polycarbosiloxy (Si-O-C) backbone polymers ....................................................................................................... 107 Figure 7. 4 Path B – polymerization to linear polysiloxane (Si-O-Si) backbone polymers ......................................................................................................................... 107 Figure 7. 5 29Si NMR of the hydroxyl terminated silane AB2 monomer ................. 108 Figure 7. 6 Hydrolysis under close to neutral conditions ([H+]=10−6 mol/L) at different times ............................................................................................................................... 110 Figure 7. 7 Intermediate formed by the reaction between water and the hyperbranched polymer monomer .............................................................................. 110 Figure 7. 8 Effect of [H+] concentration on the hydrolysis reaction. (a: approximately 10−6 mol/L, b: 0.081 mol/L, c:0.16 mol/L) ................................................................... 112 Figure 7. 9 Silanol condensation as monitored by 29Si NMR .................................... 113 Figure 7. 10 Condensation conversion as a function of the reaction time ............... 114 Figure 7. 11 Fit of the polynomials describing the concentration of alkoxysilane (a), silane-diol (b), siloxanes (c) .......................................................................................... 116 Figure 7. 12 Concentration of silane-diol (P2) as a function of time and water content. (a): acid=0.02 mol/L. (b): acid=0.03 mol/L. (c): acid=0.04 mol/L ............................. 118 Figure 7. 13 Concentration of Silane-diol (P2) as a function of time and acid concentration. (a): water=3 mol/L. (b): water=6 mol/L. (c): water=10 mol/L ........ 121 Figure 7. 14 Concentration of the siloxanes as a function of time and water concentration. (a): acid=0.010 mol/L. (b): acid=0.025 mol/L. (c): acid=0.10 mol/L 123 xiii KEY TO SYMBOLS OR ABBREVIATIONS Full Name Abbreviations/Symbol Poly(hydroxylalkanoates) PHA Poly(lactide) PLA Poly(ethylene terephthalate) PET Poly(caprolactone) PCL Poly(butylene adipate-co-terephthalate) PBAT Chain extender CE Reactive extrusion REX Epoxy functionalized PLA EFPLA Poly(ethylene glycol) PEG Heat deflection temperature HDT Thermogravimetric analysis TGA Gel permeation chromatography GPC Differencial Scanning Calorimetry DSC xiv Weight average molecular weight Mw Number average molecular weight Mn Polydispersity index PDI Low density polyethylene LDPE High density polyethylene HDPE Polypropylene PP Poly(butylene succinate) PBS Glass transition temperature Tg Poly(D-lactide) PDLA Polystyrene PS Tetrahydrofuran THF Specific mechanical energy SME Fatty acid methyl esters FAME Glycerol carbonate GC 3-aminopropyldimethylethoxysilane MESi 3-aminopropyldiethoxymethylsilane DESi 3-aminopropyltriethoxysilane TESi xv Fourier transform infrared spectroscopy FTIR xvi Chapter 1. Introduction and Background 1.1 Fossil resources and their impact In the last three centuries, the industrial revolution has greatly improved our daily life. Such rapid industrialization heavily relied on fossil resources as the major source of energy and materials. In the eighteenth century, coal was a major source of energy. Steam engines, which use coal as the energy source, provided a great source of power which animal power could not match. The late nineteenth century saw the use of petroleum as a new source of energy. The ability to refine petroleum provided a range of liquid fuels, from gasoline to diesel. Combustion engines boosted the development of the automobile industry which greatly improved daily life. Even today, petroleum, natural gas and coal still provide almost 70% of the energy in the United States.1 Petroleum is also a great source of chemical materials. From synthesized textiles to food additives, petroleum based materials can be found almost everywhere in our daily life. Polymeric materials (plastic) are also based on petroleum. Comparing with traditional materials such as wood, metal, and ceramics, plastics are light weight, strong, and easy to process. The yearly production of plastics increased from 1.6 million tons in the 1950s to nearly 300 million tons today.2 However, despite the great improvement of materials brought by fossil resources, there also exists a series of challenges. Firstly, enormous amounts of fossil resources, formed over millions of years, are being consumed. Although such resources are still abundant, this high rate of consumption is not sustainable. With a growing global population, alternative resources are becoming increasingly desirable. In addition, the side-effects of using fossil resources, in particular global warming, caused by increased carbon dioxide levels in the atmosphere, is negatively affecting the climate. Such dramatic changes have caused increase of the sea level, agriculture disasters, and extinction of some species.3 “White pollution”, which is a term widely used in Asia, 1 caused by plastic waste due to intensive use and poor management after use, has also become a significant environmental issue. 1.2 Biobased and Biodegradable Materials Biobased materials have been used throughout the history of human society. Biobased materials such as wood, leather, and cotton which are derived from biomass instead of petrochemicals are almost everywhere in our daily life. However, most of the plastics are obtained from petroleum based resources. For example, polyethylene and polypropylene are obtained from the polymerization of petroleum based ethylene and propylene. Because of the intrinsic nature as well as their lower cost, the recycling percentage of these plastics are much smaller than metal materials such as aluminum. Moreover, most of the petroleum based plastics do not degrade in the natural environment. This property can provide the plastic product an effective shelf life of value to its application, but can result in a series of environmental issues at the end of life of these plastics. Increasingly large amounts of plastic waste discarded into the environment have been found drifting in every ocean. 4 The poor management of plastic wastes together with ocean currents have accumulated the plastic drifts into waste islands and these floating waste islands have been a major threat for the ocean environment. Facing such a challenges, much attention has been directed toward manufacturing polymers from biomass feedstock since the replacement of petroleum based chemicals with bio-based chemicals offers many intrinsic value propositions: 1. The use of chemicals from biomass feedstock offers a reduced carbon footprint and sustainable carbon cycle. 2. Bio-based plastics derived from biomass is independent of the petroleum resources. 2 3. The increased need for biomass feedstock can empower the agriculture industry, which is beneficial for the rural area economy. Some biobased polymers such as starch, cellulose and poly hydroxylalkanoates (PHA) can be obtained directly from biomass. Other biobased polymers can be polymerized from biobased monomers such as poly lactide (PLA). There are also polymers that contain a certain portion of biobased materials such as biobased polyethylene terephthalate (PET).5 Biodegradable polymers can be decomposed in bioactive environments by the enzymatic action of microorganisms such as bacteria, fungi, and algae.6 Although most of the biobased polymers are biodegradable, not all bibased materials are biodegradable and not all biodegradable materials are biobased. For example, Braskem – a Brazilian petrochemical company has been producing polyethylene from ethylene derived from biobased ethanol. The ethanol is obtained by fermentation of sugarcane. Although such poly ethylene is biobased, it is still not biodegradable. Also, there are many biodegradable polymers that are not biobased. Ecoflex or poly(butylene adipate-co-terephthalate) (PBAT), which is a biodegradable aliphatic-aromatic polyester introduced by BASF, is 100% petroleum based.7 Poly caprolactone (PCL), is also a good example of biodegradable material from petroleum sources. Thus, an environmentally responsible product should be biobased and biodegrade at the end of its use. PLA is such a polymer that is both bio-based and biodegradable. 3 1.3 Organization of this dissertation This thesis consists of eight chapters and three major segments. The first segment (chapter 1) gives a general introduction of carbon cycle, biobased materials and biodegradable materials. The second segment (Chapter 2-5) is focused on modifications to improve the properties of PLA. Chapter 2 covers the rheological properties of modified PLA for blown film applications. The key modification was the introduction of an epoxy based chain extender (CE) in a reactive extrusion (REX) process. The reaction mechanism of the CE with PLA was investigated. Model compounds were used and indicated that the epoxy functional group on the CE was selectively reacting with only the carboxylic acid chain-ends of the PLA. Epoxy functionalized PLA (EFPLA) containing high percentages of CE was prepared without gel formation. The rheological properties of the CE/PLA products at different temperatures and different compositions were used to provide fundamental data for blown film processing in comparison with blown film operations of commercial grade polyethylene. Chapter 3 deals with the preparation of blown PLA films and improving the toughness of PLA by adding plasticizers. Biobased polyethylene glycol (PEG) was used in this study as the major plasticizer. Specific molecular weights of PEG were investigated. Extrusion compounding was used in this study to obtain samples containing different plasticizers concentrations. The reduction in glass transition temperature as a function of molecular weight of the plasticizer and plasticizer concentrations were studied. The mechanical properties of the blown films with and without plasticizers were compared. Chapter 4 deals with improvements in the crystallization behavior of PLA for high temperature applications. Two major factors: the molecular weight of PLA and the meso content 4 in the PLA were investigated in this chapter. Extrusion compounding was also used to study the effect of talc as nucleating agents. Different molding conditions (temperature, holding time) were investigated and the degree of crystallization was correlated with the thermal deflection temperature (HDT) of the injection molded samples. Chapter 5 discusses the end of life options of PLA. A possible route to recycle PLA by thermal depolymerization was investigated. The fundamental thermodynamic parameters and reaction constants are covered in this chapter. The mechanism and the rate of lactide formation from PLA as a function of time, temperature, and catalyst concentration were obtained using thermogravimetric analysis (TGA). In this study the non-equilibrium depolymerization process leads to high yields of the desired lactide from PLA. The experimental depolymerization data fit a two-step reaction mechanism described by the Avrami equation. The model thus obtained provides all the critical parameters affecting this recycling process. In addition, a process design of recycling PLA via thermal depolymerization is proposed in this chapter. The materials balance and energy balance are discussed and calculated. The third segment (Chapter 6-7) is focused on new hydrophilic silanes and polysiloxanes. Chapter 6 discusses the preparation and characterization of these water soluble hydroxyl alkyl terminated silanes and polysiloxanes that were prepared from the reaction of glycerol carbonate and 3-aminopropylalkoxylsilanes. It was found that this reaction follows a second order rate constant irrespective of the type of aminosilane that were used. Conventional polymerization of these hydroxyl alkyl terminated silanes yielded linear, branched or resinous siloxy polymers via self-polycondensation through exchange of the ethoxy groups attached to the silicon atom with the terminal hydroxyl groups. The synthesis of the silane monomers as well as the structure and key 5 properties of the hydroxyl alkyl terminated siloxy polymers were investigated by DSC, 1H NMR, FTIR, TGA and GPC. Chapter 7 deals with the kinetics of the hydrolysis and subsequent condensation reactions under acidic conditions of hydrophilic silane that was obtained from ethylene carbonate and 3aminodiethoxylsilane. The compositions of the silanol containing hydrolysis intermediates and the siloxanes condensation products were identified under different conditions using 29Si NMR. The concentration of the intermediates species was found to depend on specific combinations of the acid catalyst and the water concentration. Under certain conditions the intermediate silane-diols were stable and did not condense even under mild acidic conditions. 6 Chapter 2. PLA melt strength enhancement for blown film application 2.1 Introduction Poly (lactide) (PLA) is a bio-based and biodegradable polymer. It is obtained by bacterial fermentation of corn starch to lactic acid. However, simply polymerizing lactic acid does not provide high quality products. The lactic acid polymerization reaction follows a condensation mechanism that generates one molecule of water with every step. This reversible reaction causes problems with the polymerization reaction and also leads to high polydispersity. To overcome this problem, lactic acid is first oligomerized and then catalytically dimerized to make the cyclic lactide monomer. While dimerization also generates water, it can be separated by distillation prior to polymerization. High molecular weight PLA is produced from the lactide monomer by ringopening polymerization using metal salt catalysts.8-10 As a high modulus, bio-based material derived from corn starch, PLA is an excellent candidate for packaging film production. The blown film process requires the polymer melts to exhibit sufficient melt viscosity as well as nonlinear viscoelastic properties (strain hardening). However, PLA polymer melts show a relatively low melt viscosity which are generally not suitable for the blown film process. Long chain branching or a small fraction of high molecular weight polymer should be introduced to this linear polymer to improve its melt rheology properties for the blown film process.11,12 Chain extenders have been widely used to increase the rheological properties of polyesters by reacting with either the carboxylic acid or the hydroxyl end groups. Diisocyanates have also been investigated to react with the chain ends of PLA to improve its rheological properties.13,14 However, the toxicity of isocyanates as well as the high possibility of gel formation has limited the further application of such technology. Epoxy based chain extenders were also studied to increase the 7 molecular weight of polyesters and introduce long chain branching. 15-20 Epoxy based chain extenders are available in both solid form and liquid form. They are less toxic and can be introduced via reactive extrusion. The reaction mechanism was concluded to be the either the carboxylic acid or the hydroxyl chain end of PLA to ring open the terminal epoxy functional group in the chain extender to form an ether or ester linkage and a secondary hydroxyl group.21-24 There are very few studies focusing on the reactivity difference of the hydroxyl chain end and the carboxylic acid chain end. Due to lack of understanding of the reaction mechanism, samples containing high concentrations of CE have not been prepared to “avoid” crosslinking of the products. In this study, the reactivity of the carboxylic acid and the hydroxyl end groups with model compounds has been studied. A masterbatch with high concentration of CE was prepared with no gel formation. The detailed rheological properties of the chain extended PLA materials from the masterbatch were studied and compared with the conventional one batch processing. 2.2 Experimental PLA 3051 D with a molecular weight of ~100,000 and 4% D content was purchased from Natureworks, LLC. (NE, USA). Chain Extender (Joncryl 4368F) was purchased from BASF (MI, USA). Blown film grade low density polyethylene (LDPE) (NA 952000) was purchased from LyondellBasell (MI, USA). Stearic acid reagent grade was purchased from Sigma Aldrich (MO, USA). Poly (ethylene glycol) with a molecular weight of 2000 was purchased from Sigma Aldrich (MO, USA). The PLA resin pellets were dried overnight at 70oC prior to compounding. The CE was powdered using a mortar and pestle. The PLA resin pellets and powdered CE were fed into a ZSK 8 30 twin-screw extruder (Werner Pfeiderer) with an L/D ratio of 30 using two separate volumetric feeders. Samples containing 1%, 5% and 10% CE were compounded at a throughput of 4.8 kg/hr. The samples were cooled in a water bath and pelletized. After drying, the 5% CE sample which was epoxy functionalized PLA (EFPLA) was used to further compound with neat PLA resin pellets and extruded again to obtain the diluted specimens at 0.25%, 0.5%, and 1% CE content. The temperature profile for the compounding was 140, 165, 170, 175, 175, and 170oC. The screw speed was set to 125 rpm for all compounding. The compounded melt strand was cooled in a water bath and then pelletized and dried for further use. The molecular weight was determined by gel permeation chromatography (GPC) equipped with a refractive index detector (Shimadzu, Tokyo, Japan, RID-10A) and a combination of three columns (Waters Co., Israel). Tetrahydrofuran was used as the mobile phase with a flow rate of 0.50 mL/min at 40 °C. Polystyrene standards were used to obtain the calibration curve for molecular weight calculation. The model compound reaction kinetics was monitored by DSC. The model compound (stearic acid and PEG) was premixed with powered CE at 1:1 weight ratio. 10 mg of the sample was sealed in a T zero aluminum pan and heated from room temperature to 220oC at a heating rate of at 10oC/min in the furnace of a Differential Scanning Calorimeter (TA Instruments, Q20, USA). The heat flow was recorded and plotted to observe the exothermic heat signal of the ring opening reaction. Soxhlet Extraction was performed for 3 days (72 hours) using dichloromethane as the solvent. The weight of pre-dried cellulose thimble before and after the extraction was recorded for the gel 9 content analysis. For each Soxhlet extraction experiment, 1 gram of the compounded resin was used and each experiment was repeated twice to ensure the results. The complex viscosity was measured by dynamic rotational experiments conducted using a parallel-plate geometry with a plate separation of 1.05 mm. A frequency sweep was performed using a frequency range of 1 – 100 rad/s. For testing with frequencies lower than 1 rad/s (ex. 0.1 rad/s – 1 rad/s), severe degradation of PLA was observed due to the long testing time. Four temperatures 170, 180, 190 and 200 ℃ were tested. The extensional viscosity was measured using an ARES rheometer equipped with EVF fixture at 180 ℃22,25. The test specimens were compression molded into a dimension of 18*10*1 mm3 before testing. Four different Hencky strain rates 0.1 rad/s, 0.5 rad/s, 1 rad/s and 5 rad/s were measured. The following equation was used to calculate the extensional viscosity for PLA melt in steady stain rate.  E  (T )  2  2 R  H A0 ( s ) 3 exp(   H (t )) m  (Eq.2.1)   Where  E is the value of tensile stress growth,  H is the applied Hencky strain rate, T is the torque, R is the drum radius, A0 is initial area of the sample measured in the solid state, and t is time data. The density of PLA in solid state, at 180 ℃,  s , was considered as 1.25 g/cm3, 22 and the value for PLA melt  m , was 1.115 g/cm3 10 2.3 Results and Discussion 2.3.1 Model compound reaction study Several previous research works have discussed the possible reaction between PLA and Joncryl based epoxy chain extenders.15,17,24 However, there is still uncertainty with respect to the reactivity of the two major functional groups present in PLA, hydroxyl and carboxyl groups. In this study, a series of experiments were designed with two different model compounds to study the reaction kinetics and temperatures between the chain extender and the functional groups of PLA. Stearic acid and low molecular weight (Mw=2000) PEG were used as model compounds to represent the carboxylic acid functional group and hydroxyl functional group of PLA, respectively. The ring opening reaction was observed in DSC by monitoring the exothermic signal. 11 Figure 2. 1 DSC spectrum of PEG-CE mixture and steric acid-CE mixture As shown in Figure 2.1, an exothermic signal was observed for stearic acid, indicating reaction had occurred between the CE and stearic acid. The reaction temperature was between 130-200℃ with a peak temperature of 168℃. In contrast, no exothermic signal was observed in the spectra of the PEG/CE mixture indicating there is no reaction between the hydroxyl functional group and the CE within the temperature range tested. The general processing temperature for PLA is between the melting temperature and 200℃. As no exothermic signal was observed in the case of the hydroxyl compound reaction, it is apparent that the hydroxyl chain ends of PLA did not react with the CE. In contrast, the exothermic signal observed in the carboxylic acid chain end model compound indicated the ring opening reaction has taken place below 200℃ (Figure 2.2). 12 Figure 2. 2 Reaction scheme between Epoxy based CE and PLA 2.3.2 Epoxy functionalized PLA and gel content analysis From the DSC study, the results indicated only the carboxylic acid chain end of PLA reacted with the CE in the processing temperature range of PLA. Thus, upon reaction, the PLA will only form a branched structure instead of a crosslinked network regardless of the content of CE. Two samples of high CE content (5% and 10%) in PLA were prepared by reactive extrusion. Soxhlet extraction was performed to study the gel content in these series of PLA samples with high CE content. The experimental results are shown in Table 2.1. No residual weight was observed in the dried thimble indicating no gel was formed during the reactive extrusion process. CE content (%) Thimble (g) Thimble + Sample (g) After Soxhletion （g） Gel content* 1 3.719 4.782 3.661 -0.054 3.431 4.590 3.413 -0.015 3.594 4.892 3.557 -0.028 3.600 4.644 3.557 -0.041 3.630 4.766 3.607 -0.020 3.582 4.672 3.536 -0.042 5 10 Table 2. 1 Gel content analysis by Soxhlet Extraction *The slight negative gel content was due to the intensive drying of thimble. No gel was observed in the thimble after extraction. 13 As discussed previously, PLA samples containing high CE content was prepared without gel formation. Such sample containing high CE content still contains reactive epoxy functional groups. In other words, we have obtained epoxy functionalized PLA (EFPLA) by adding excess amount of CE into the polymer via reactive extrusion. Three different samples (0.25%, 0.5% and 1% CE content) were prepared through further compounding and reacting the EFPLA with neat PLA. 2.3.3 Molecular Weight Characterizations The effect of the CE on the molecular weight distribution of the compounded material was analyzed using GPC. The molecular weight calculation was based on Polystyrene standards and the results are listed in Table 2.2. CE content (%) Mw (103) Mn (103) PDI (103) 0 139 95 1.46 0.25* 184 77 2.38 0.5* 250 102 2.45 1* 361 121 2.99 1 248 143 1.73 5 327 171 1.90 10 146 88 1.66 Table 2. 2 Effect of CE content on PLA molecular weight *These samples were compounded from EFPLA containing 5% CE A molecular weight increase from 1.39*105 to 3.27*105 g/mol was observed as the CE content was increased from 0% to 5%. However, further increasing the CE content from 5% to 10% resulted in a decrease of molecular weight from 3.27*105 to 1.46*105 g/mol. Such a phenomenon 14 can be explained by the chain end availability of PLA. As each CE contains multiple reactive epoxy functional groups, the chain extension can occur by connecting multiple PLA chains to a single CE molecule resulting in an increased weight average molecular weight (Mw) (Figure 2.3 A).With higher concentrations of CE, the number of reactive chain ends became limited. Therefore, instead of connecting multiple PLA chains together, the possibility of only one chain connecting with one CE molecule increases (Figure 2.3 B). The molar ratio of these two reactants can be roughly estimated from the molecular weight of PLA and CE. The molecular weight of PLA is around 100,000 according to our GPC analysis, the molecular weight of CE is around 5,800 from the previous studies26. Thus, with around 5.5% of CE content, we can achieve a 1:1 molar ratio of PLA molecule and CE molecule. Such estimation further explained the decrease of molecular weight of the sample with 10% CE content as the CE molecule is in large excess (1:2 ratio). 15 Figure 2. 3 (A) Scheme of multiple PLA chains connected to a single CE molecule (B) Scheme of one PLA chain connected to one CE molecule Three lower concentrations of CE samples (0.25%, 0.5% and 1%) were prepared by reacting the 5% CE EFPLA with neat PLA via reactive extrusion. A higher efficiency of chain extending was observed by comparing the molecular weight of these samples with the sample prepared from directly compounding PLA with CE (one batch process). Higher polydispersity index (PDI) and Mw were observed in the sample obtained from the EFPLA containing 1% CE (PDI=2.99 Mw=3.61*105 g/mol) as compared to the sample containing 1% CE from one batch process (PDI=1.73 Mw=2.48*105 g/mol). GPC curves of these samples were compared with the original PLA resin used in this study (Figure 2.4). A clear high molecular weight shoulder was observed. Such high molecular weight fraction in the sample is caused by the chain extending reaction connecting multiple PLA chains together. Also, as the concentration of CE increases, the height of the shoulder increases. 16 Figure 2. 4 GPC spectrum of neat PLA and chain extended PLA with 0.25%, 0.5, 1% CE content prepared from EFPLA. 2.3.4 Rheological Properties Complex viscosity of the compounded samples were tested by dynamic rotational experiments with a frequency sweep from 1rad/s to 100 rad/s. Polymers possessing a branched structure tend to behave more pseudoplastically than linear polymers of the same molecular weight due to their more compact nature27. As CE concentration was increased, a more significant shear thinning was observed (Figure 2.5) indicating the branched structure obtained from the chain extending. There was a dramatic difference between the sample containing 1% CE from EFPLA and the one step processing. Even with a higher CE content, the complex viscosity of 1% CE sample from one step processing was lower than the sample containing 0.25% CE prepared from EFPLA. The higher 17 PDI in the sample containing 0.25% CE and the high molecular weight tail might have further increased the viscosity of the compounded product and caused such phenomenon. Figure 2. 5 Comparison of complex viscosity frequency sweep at 180℃ between neat PLA, chain extended PLA with 0.25%, 0.5, 1% CE content from EFPLA and chain extended PLA with 1% CE content from one step processing. A complex viscosity-temperature correlation study was conducted to compare the rheological properties of chain extended PLA with blown film grade LDPE. Four temperatures (170 ℃, 180 ℃, 190 ℃ and 200 ℃) were studied (Figure 2.6). A more significant viscosity decrease was observed for the chain extended PLA as the testing temperature was increased. When the testing temperature was below 180 ℃, the chain extended PLA containing 0.5% to 1% CE showed a higher complex 18 viscosity compared to LDPE. However, when the temperature was further increased to 190 ℃ and 200 ℃, the complex viscosity of PLA decreased significantly while LDPE was less impacted. Figure 2. 6 Comparison of complex viscosity frequency sweep at 170℃, 180℃, 190℃, and 200℃ between neat PLA, chain extended PLA prepared from EFPLA, and LDPE. Extensional viscosity is an important property for polymer processing, especially film blowing and foaming. The extensional viscosity of a polymer can be increased by increasing the molecular weight. The extensional viscosity testing was conducted at 180 ℃ which is common temperature for PLA processing. A cross composition plot (Figure 2.7) was made to compare the extensional viscosity of chain extended PLA sample, unmodified PLA and LDPE at 1 rad/s. Similar to the result of complex viscosity, as CE concentration increased, the extensional viscosity of the polymer melt increased. In fact, the chain extending PLA showed as high as 100 time higher extensional viscosity comparing with the unmodified PLA. The extensional viscosity of LDPE was overlapping with the chain extended PLA sample containing 1%. Strain hardening, which is 19 also important for film blowing bubble stability and foaming cell size control, was also observed in the testing (Figure 2.8, Figure 2.9). Figure 2. 7 Comparison of extensional viscosity 180℃ at a rate of 1rad/s between neat PLA, chain extended PLA prepared from EFPLA, and LDPE. 20 Figure 2. 8 Strain hardening behavior of modified PLA containg 1% CE at different shear rates (5 rad/s, 2 rad/s, 1 rad/s, and 0.5 rad/s). 21 Figure 2. 9 Comparison of strain hardening behavior of Neat PLA and modified PLA containing 1% CE at 5rad/s 2.4 Conclusions The reaction mechanism between epoxy based CE and PLA was examined in this study. The carboxylic acid chain ends of PLA were shown to be reactive with the CE under specific reactive extrusion conditions (below 200 ℃). EFPLA sample containing high concentration of CE (5% and 10%) were prepared without gel formation. The chain extended PLA samples from the EFPLA showed a higher reactive efficiency of CE compared with the one step processing. Complex viscosity and extensional viscosity were characterized for the compounded PLA from EFPLA and compared with a commercial grade LDPE. The results showed with 1% of CE content, the chain extended PLA can exhibit comparable rheological properties with the film grade LDPE. Such properties allow PLA to be a suitable biodegradable replacement for LDPE in blown film applications. 22 2.5 Future work From our experiments, we have characterized the reaction of epoxy with the carboxyl chain end of PLA using GPC. Another potential method is to use titration method to obtain the yield of epoxy functional groups during the extrusion process. As discussed before, we are preparing a reactive epoxy functionalized PLA by reacting CE with PLA. The number of functional groups available can be of great interest. In the future work, researchers can investigate the epoxy value of a compounded resin using titration method and thus obtain the yield of epoxy ring opening reaction from reactive extrusion process. With such understanding, researchers can further study the effect of molar ratio of epoxy with carboxyl chain end (CE content), process temperature and screw design with the reaction yield as the response. Design of experiments can be conducted to obtain the optimal process condition and thus further improve the efficiency of such process. 23 Chapter 3. PLA blown film and plasticization with Poly(ethylene glycol) 3.1 Introduction Plastic film is one of the most widely used materials in the world. Due to its lightweight, flexibility, and excellent moisture barrier performance, it has found many uses from food packaging to electronics packaging. The most commonly used materials for plastic film applications are LDPE, High density polyethylene (HDPE) and polypropylene (PP). The two major film production methods are melt casting and blowing. Blown films can provide better molecular orientation and generally have better properties than cast films. The film blowing process has been intensively studied and widely used over 40 years. The film blowing process is a biaxial stretching of the polymer melt from an annular die to a bubble with certain required dimensions. As discussed in the previous chapter, extensional rheology is a useful technique to determine the non-linear behavior of the polymer melt in the film blowing process.28 In blown film processing, bubble stability is a key factor that determines the properties of the resulting film. Bubble stability is known to be associated with the strain hardening rheology behavior of the polymer melt.29,30 There are two major obstacles in the blown film processing of PLA: low melt viscosity and bubble instability. Both can be modified by introducing branching and a high molecular weight tail to the polymer. In the previous chapter, the rheology modification of PLA with chain extender was discussed. However, as PLA is a brittle material, improvement in its ductility is also important in order to obtain high quality films. There are many different approaches to increase the ductility of PLA. 24 31 In order to increase the elongation at break, many different methods have been described in the literature. These methods can be classified into: copolymerization, melt blending with flexible polymers and plasticization. Copolymerization has long been known as a powerful method to obtain polymers with properties unattainable by homopolymerization. The copolymerization of PLA with other materials is most often conducted via a ring opening reaction in order to obtain a good control over the chemistry. For example, polycaprolactone (PCL), a biodegradable polymer which is obtained from the ring opening reaction of carprolactone, has been used. PCL is characterized by a very low glass transition temperature (-60℃) and is known to have excellent flexibility. Copolymers of PCL with PLA were shown to have very interesting properties. By varying the ratio of the lactide to the caprolactone monomer, a wide range of plastic resins from a soft gel to a tough plastic have been obtained. 32 Another strategy is to melt blend other flexible polymers into PLA. Melt blending of polymers is typically achieved by extrusion, a very economical and efficient processing method. A large variety of biodegradable flexible polymers have been used as toughening modifiers for PLA. For example, poly(butylene succinate) (PBS) has a very low Tg and is a very good modifier for PLA. Although it is immiscible with PLA, it is possible to obtain blends with more than 150% of elongation at break by adding appropriate coupling agents.33 Another type of biodegradable polymer that has been used is poly(butylene adipate-co-terephthalate)(PBAT) which is available commercially as “Ecoflex” from BASF. Similar to PBS, it is also a flexible polyester with a low glass transition temperature (-30℃). By blending PBAT with PLA, it was observed that the ductility of the product increased with increasing the concentration of PBAT.34 25 Plasticizers are widely used additives in the plastics industry. They have been used to improve the processability of polymers (such as cellulose acetate). Generally, by adding sufficient plasticizer into a polymer, the flexibility and ductility of the polymer can be enhanced. However, in order to be efficient, the plasticizer must be miscible with the polymer system so that the mobility of the polymer chains will be improved to significantly lower the glass transition temperature (Tg) and increase the polymer toughness. 35 The depression of Tg is a key factor in the performance evaluation of plasticization. Theoretically, the performance of plasticizers can be predicted by the similarity of the solubility parameters (δ) between the plasticizer and PLA.36 Several non-toxic esters such as triethyl citrate, tributyl citrate, acetyltriethyl citrate, and glycerin triacetate have been incorporated into PLA and an increase in toughness accompanied by a drop in the modulus have been noted. 37,38 However, there are limitations to the use of small molecule plasticizers principally due to evaporation during melt processing and migration toward the surfaces during storage. By increasing the molecular weight of plasticizers, the migration and evaporation of plasticizers can be reduced. 39 Poly (ethylene glycol) (PEG) is a class of hydrophilic polymers obtained from the ring opening polymerization of ethylene oxide. They are nontoxic, water-soluble, and biodegradable. Although traditionally considered as a petroleum based polymer, ethylene production from biobased ethanol indicates that such these class of polymers can be prepared from biobased resources. 40 26 Figure 3. 1 Synthesis of Biobased PEG from biomass40 The miscibility of PEG and PLA depends on the molecular weight.41,42 Low molecular weight PEG exhibits better miscibility with PLA and results in a more pronounced reduction of the glass transition temperature. The Fox equation was used to predict the glass transition temperature of a polymer blend system. Such reduction in the glass transition temperature can lead to significant improvement in ductility of PLA even at lower PEG concentrations. 43 3.2 Materials and experiments PLA 3051 D with a molecular weight of ~100,000 and 4% D content was purchased from Natureworks, LLC. (NE, USA). Chain Extender (CE) (Joncryl 4368F) was purchased from BASF (MI, USA). PEG resin with molecular weight 6000, 4000, and 1000 were purchased from Sigma Aldrich (MO, USA). The PLA resin pellets were dried overnight at 70oC prior to compounding. The PLA resin pellets and plasticizers were fed into a ZSK 30 twin-screw extruder (Werner Pfeiderer) with an L/D ratio of 30 using two separate volumetric feeders (Figure 3.2). All samples were compounded at a total throughput of 4.8 kg/hr. The samples were cooled in a water bath and pelletized. The 27 Temperature profile for the compounding was 140, 165, 170, 175, 175, and 170oC. The screw speed was set to 200 rpm for all compounding. 28 Figure 3. 2 Screw configuration of the ZSK 30 twin-screw extruder. 29 The molecular weight was determined by gel permeation chromatography (GPC) equipped with a refractive index detector (Shimadzu, Tokyo, Japan, RID-10A) and a combination of three columns (Waters Co., Israel). Tetrahydrofuran was used as the mobile phase with a flow rate of 0.50 mL/min at 40 °C. Polystyrene standards were used to obtain the calibration curve for molecular weight calculation. The glass transition temperature of the compounded product was determined using DSC. 10 mg of the sample was sealed in a T zero aluminum pan and heated from room temperature to 220oC at a heating rate of at 10oC/min in the furnace of a Differential Scanning Calorimeter (TA Instruments, Q20, USA). The heat flow was recorded and plotted to glass transition signal of the material. The film blowing experiment was conducted with a Killion Blow Film Extruder (KL-100). The extruder temperature was set at 165℃, 170℃ and 170℃. The die temperature was set at 165 ℃. The film process was conducted with a screw speed of 10RPM, drawing rate of 5RPM and blow up ratio of 2. The blown film mechanical property testing was conducted with United Testing Systems (UTS) model SFM-20 load frame following ASTM D882-09. The thickness of the tested film was 1 mil (0.0254 mm). 3.3 Results and discussions 3.3.1 Molecular weight effect on plasticization performance The Fox equation was used to predict the glass transition temperature of a system the PLA/PEG blends. 30 1 w w  1  2 Tg Tg1 Tg 2 Eq. 3.1 ,where w1 is the weight percentage of the first polymer with a glass transition temperature of Tg1; w2 is the weight percentage of the first polymer with a glass transition temperature of Tg2. The glass transition temperature of PLA is approximately 58℃ and the glass transition temperature of PEG depends on its molecular weight. From the Fox-Flory equation, the glass transition temperature depression of the extruded blend can be predicted as shown in the Table 3.1. Weight Percent Predicted PEG Predicted PEG Predicted PEG (℃) (6000) (℃) (℃) (4000) (℃) (℃) (1000) (℃) Pure PEG -49 - -50 - -53 - 0 61.0 61 61.0 61 61.0 61 5 53.0 54 52.9 52.1 52.6 51.6 10 45.4 47.6 45.2 43.2 44.5 41.4 15 38.1 42.9 37.8 39.1 36.9 36 20 31.1 41.1 30.8 37.7 29.6 32.2 30 18.1 ** 17.6 35.8 16.1 28 40 6.2 ** 5.5 33.5 3.7 * (%) Table 3. 1 Glass transition temperature of extruded PLA/PEG blends * The glass transition temperature was not detectable or the signal is too weak. ** Phase separation occurs, the product is not homogenous. 31 The performance of the PLA resin blends with different molecular weights and concentrations of PEG can be compared from the data shown in Table 3.3 and Figure 3.3. Figure 3. 3 Glass transition temperature plot of extruded PLA/PEG blends The molecular weight of PEG showed a significant impact on the plasticization performance. At low PEG concentration, irrespective of the molecular weight of PEG, the addition of plasticizers decreased the glass transition temperature of the plasticized PLA to approximately 40℃ with 20% PEG concentration. However, when the plasticizer content was further increased, lower molecular weight PEG plasticizers showed better compatibility whereas the high molecular weight PEG (Mw=6000) showed a limited miscibility with PLA and began to phase separate. This observation agrees with the previously reported work44 where it was noted that as the molecular weight of PEG increased, the miscibility between PLA and PEG decreased. 32 Of greatest interest is the minimum required concentration of PEG in PLA that will lower the glass transition temperature of the blend to below room temperature. Under this condition, it is expected that the resin blend will exhibit better ductility and toughness at room temperature. 3.3.2 Effect of plasticizer on crystallization rate of PLA In addition to reduction of the glass transition temperature, plasticizers can also improve the rate of crystallization of PLA by both lowering the system viscosity and providing nucleating sites.44 A series of studies were conducted to compare the effect of plasticizer on the rate of PLA crystallization. Three PLA samples with different compositions were prepared by extrusion. (Table 3.2). PEG grade Talc Content CE PEG Peak Degree of Content content area crystallization (%) (%) (J/g) 1000 1 1 20 23.1 30.8 1000 0 1 20 15.6 20.8 6000 0 1 25 20.2 28.7 6000 1 1 25 22.1 31.4 6000 0 0 25 * * Table 3. 2 Comparison of crystallization rate with and without plasticizers From the data in Table 3.2 it is clearly observed that the addition of plasticizer or nucleating agent alone is ineffective as the PLA showed a relatively slow rate of crystallization. At a cooling rate of 10℃/min, no crystallization peak signal was observed. However, when both plasticizer and nucleating agent were added, the crystallization rate improved significantly and almost 30% of crystallization was obtained in the cooling cycle. 33 As discussed in the previous section, the molecular weight of PEG significantly affected the plasticization performance. However, in this study, the molecular weight did not significant affected the rate of crystallization. 3.3.3 Chain end trans-esterification reaction and its effect on molecular weight As PEG is a hydroxyl functional polymer with a much lower molecular weight compared to PLA, the number of hydroxyl chain ends in PEG can be very large compared with that in a plasticized sample. The PLA grade used in this study was 3051D with a molecular weight of 100,000. In contrast, the molecular weights of PEG plasticizers were in the range of 1000 to 6000. Thus, even with only 20% of plasticizer in the system, the number of PEG chain ends would be much higher than the number of PLA chain ends. A possible trans-esterification side reaction: could occur during the melt processing where the hydroxyl functional group of PEG is inserted into the ester linkage of PLA to decrease the molecular weight of the PLA.45 Such reaction is undesirable since a decrease in the molecular weight of PLA generally causes a decrease in the mechanical properties. The molecular weights of melt extruded plasticized PLA samples and original pure PLA samples were compared using GPC to clarify this possible side reaction (Figure 3.4). 34 Figure 3. 4 GPC plot of PLA and plasticized PLA Figure 3.4 shows the GPC traces of PLA and PLA containing 20% of PEG 1000 prepared by melt extrusion. The signal peak of PLA is observed at approximately 9 minutes elution time and the signal peak of PEG 1000 at approximately 12 minutes. It is apparent that no significant change in the retention time of PLA containing 20 wt% PEG was observed indicating that the trans-esterification reaction did not occur or was very minor and not significant enough to be observed by GPC. 3.3.4 Chain end reaction and selectivity reaction with chain extenders As was discussed in a previous section, the goal of PLA plasticization is to improve the ductility of PLA blown films. For blown film applications, it is necessary to increase the molecular weight of PLA by chain extenders as discussed in Chapter 2. 35 To be effective, it is desired that the chain extenders would react with the chain ends of PLA to yield long chain branched structures. Also, since the number of PEG hydroxyl chain ends is in huge excess compared with the PLA chain ends, it is very important to understand the chain extending reaction mechanism in order to selectively increase the molecular weight of the PLA instead of the PEG fraction. Two different methods were examined to obtain chain extended PLA containing PEG plasticizer. The first method was to chain extend PLA without plasticizer and then add PEG plasticizer in a separate step. Such a method involves two processing steps. In the first step, PLA will react with chain extender to increase its molecular weight. In the second step, PEG is added into the chain extended PLA to introduce the plasticizing effect. Thus, in this method the hydroxyl chain ends of PEG does not interfere with the chain extending reaction. However, the use of two separate steps, is not favorable in industrial productions. The second method consists of a one-step process. In this method, PEG plasticizer, PLA, and chain extender were fed into the extruder simultaneously. Both melt blending of PLA and plasticizer and chain extending of PLA were achieved within one pass of extrusion. As discussed in chapter 2, the carboxylic acid chain ends of PLA are much more reactive to the epoxy based chain extenders at the processing temperature used comparing with the hydroxyl functional groups. Thus, the addition of hydroxyl functional PEG plasticizers will not affect the chain extending reaction. 36 The molecular weights of samples prepared by these two methods were analyzed using GPC and the molecular weight distributions of three different samples are shown in Figure 3.5. Figure 3. 5 GPC plot of PLA, one step sample, and two-step sample The retention time was observed to decrease in both samples containing 1% CE, indicating that the PLA molecular weight increased in these samples. Furthermore, a clear and distinct lower molecular weight PEG peak can be observed in the spectra indicating that no reaction between the chain extender and PEG occurred. Interestingly, a much more significant low molecular weight tail can be observed in the sample prepared by the two step method. The lower molecular weight portion of PLA can be explained by the breakdown of high molecular weight PLA during the extrusion process 37 as two extrusion passes were applied in the two step method and therefore degradation is likely to occur under these conditions. 3.3.5 Effect of plasticization on blown film mechanical properties PLA containing 1% CE as well as samples prepared by the one step method and the two step method were blown into films using a conventional LDPE film blowing equipment. Compared with unmodified PLA, the materials containing 1% CE showed a higher melt strength. Such observations agree with our previously reported data in chapter 2. The films obtained were tested for mechanical properties and the results are shown in Table 3.3. Sample Number 1 2 3 PEG content [%] 0 10 Process 1 step Modulus Modulus Toughness Toughness [MD] [TD] [MD] [TD] (MPa) (MPa)] (MJ/m3) (MJ/m3) 3002±194 2940±561 1.15±0.34 0.49±0.20 2331±61 2713±65 1.64±0.10 0.72±0.27 10 2 steps 2063±15 2225±48 1.36±0.45 0.51±0.10 Table 3. 3 Mechanical properties comparison of PLA blown films The secant modulus of PLA containing 1% CE is around 3000 MPa in both machine direction (MD) and transverse direction (TD) which is close to the previous reported PLA modulus in bulk.46 With the addition of 10% PEG 6000, a significant drop in the modulus of the film can be seen. Around 30% of loss in modulus was observed in our study (Table 3.3). However, as PLA is a high modulus polymer47, such loss of modulus does not impact its application significantly. Interestingly, the samples obtained from two step processing showed slightly lower mechanical properties compared with the one step processing samples. Such 38 observation can be correlated to the GPC analysis and can be explained by the degradation caused by the extra step. The toughness of the film was slightly improved by the addition of 10% PEG, but was still significantly lower compared with other commercial blown films. Similarly films containing 20% PEG also prepared by the one step method demonstrated severe blocking and plasticizer migration in the product films. 3.4 Conclusions PEG was introduced as a plasticizer to modify and improve the physical properties of PLA. The effects of molecular weight and composition of the blend on plasticization were studied. It was observed that low molecular weight PEG exhibited higher miscibility with PLA with no apparent phase separation. Two different methods were used and compared to introduce PEG into chain extended PLA samples. The one step method utilizing the higher reactivity of carboxyl functional group with epoxy showed better results compared with the two step method. The observed differences were directly related to the extent of degradation caused by the extra extrusion step in the two step method. The results of this study allowed us to prepare PLA blown films in a conventional LDPE film blowing equipment using the chain extended PLA. Stable bubble was obtained in this process and the tensile moduli of the films were comparable to that of bulk PLA. The addition of 10% PEG showed a 33% loss in tensile modulus but a slight increase in toughness. However, adding higher concentrations of PEG into the system caused severe blocking of the film blowing process such that high quality samples could not be obtained. 39 3.5 Future work Significant blocking of the films and unstable bubble were observed when 20% of PEG was added into PLA. PEG also showed a high tendency to migrate from the films to the surface. In order to produce films with competitive performance to current polymer films, further work should focus on finding appropriate plasticizer that does not migrate out of the PLA bulk does not cause the blocking issue. 40 Chapter 4. High temperature mold induced crystallization for Increased Heat Deflection Temperature performance of PLA-Talc compound 4.1 Introduction Poly(lactic acid) or Polylactide (PLA) is a thermoplastic aliphatic polyester derived from renewable resources such as corn starch (in the United States), tapioca roots, chips or starch (mostly in Asia), or sugarcane (in the rest of the world). With a current global annual capacity of approximately 150,000 tons and projected increase above 200,000 tons by 2015, PLA is the most widely available biopolymer today despite its shortcomings in processability, impact resistance and thermal resistance.48 For example, in the food packaging and handling industry such as cutlery and beverage cups, non-biodegradable plastic materials such as PE and PP are hard to recycle because of food contamination. In this market as well as in many other applications, especially where higher temperature is required, PLA is severely limited by its low glass transition temperature (56°C). 49 In contrast, the major advantages of PLA are: high modulus, biobased, and, more importantly, biodegradable.50 However, in these real life applications, PLA materials generally show very poor performance when handling hot food and beverages (Figure 4.1). The cutlery and stir bars begin to soften when exposed to high temperatures such as hot coffee and packaging for fresh grilled chickens. This is due to the loss of modulus of the amorphous PLA phase and the significant viscoelastic behavior above 56°C. Furthermore, PLA materials have 41 melting temperatures in the range of 150-180°C but slow crystallization rates, therefore, most of the PLA injection molded products are amorphous or only contain a very low degree of crystallization. Since retention of properties is dependent on crystallization, increasing the degree of crystallization is critical for enhancing the high temperature, dimensional stability and mechanical performance. Crystallized PLA Hot coffee Hot tea Grilled Chicken Soup Amorphous PLA Water Ice Tea 0 30 60 90 120 150 180 210 Temperature [oF] Figure 4. 1 Serving temperature of food vs HDT temperture of PLA51 Some of the factors which researchers have studied to affect the crystallization behavior of PLA are meso content, molecular weight and the addition of modifiers such as nucleating agents. PLAs are obtained from the ring opening polymerization of optically active lactide. Lactide is a cyclic dimer of lactic acid. Due to the presence of a chiral center, there are 42 two optically active isomers: L-lactic acid and D-lactic acid. L-lactic acid is found in milk products, in the muscle tissue of animals and is abundant in nature. In contrast, Dlactic acid is not common in nature. These isomers leads to three lactide isomers as shown in Figure 4.2. Figure 4. 2 Stereo Isomers of lactic acid and lactide In addition, properties such as melting temperature, glass transition temperature, and crystallinity also depend on the ratio of these optical isomers as well as the molecular weight of the polymer. It was shown previously that PLA containing more than 7% of the D-lactic acid isomer is completely amorphous.52 When the L-lactic acid content is higher than 93%, PLA exhibits semi-crystalline properties. This is caused by variation of L and D lactic acid ratios which introduces irregularities in the polymer chains thus affecting crystallization and hence the behavior of the polymer. Depending on the L:D isomer ratio within the polymer, PLA has a glass transition temperature in the range of 55-70 °C, and a melting temperature in the range of 130-180 °C.PLA 43 polymers are mostly produced from the ring opening reaction of S,S lactide with small amount of meso lactide. The optical purity of S,S lactide (e.g. the meso content) significantly affects the properties of the PLA polymer. The degree of crystallization as well as the rate of crystallization decreases as the meso content increases. For example, a 1% increase in the meso content increases the crystallization half time by 40%. 53 The other approach is to introduce nucleating agents to lower the energy barrier towards crystallization. Different nucleating agents such as talc54 and montmorillonite clay55 have shown good performance results. Interestingly, the introduction of poly(D-lactide)(PDLA), which is the homo polymer of R,R lactide, can result in the formation of a stereo complex. The formation of stereo complex between the two helical polymer chains in solution has been reported. 56 The stereo complex formed from 1:1 ratio of PLLA and PDLA has a melting temperature of 230 °C. 57 Such high melting temperature which is over 50 °C above the value of normal PLA as well as the high miscibility due to the same molecular structure, made this material a potential nucleating agent. Studies have shown that addition of high molecular weight PDLA led to the formation of nucleation sites, and much faster rate of crystallization.58 However, the synthesis of PDLA is difficult and the price of PDLA is not competitive for large scale production. Another important factor is the molecular weight of PLA. Generally, a polymer with a lower molecular weight has a higher rate of crystallization due to the lower viscosity and abundance of chain ends that can act as nucleating sites.59 Although such 44 factors have been well accepted, few studies have been reported in the literature taking advantage of this phenomenon. The objective of this study was to study the effect of crystallization on the properties of various PLA grades and to test the performance of different nucleating agents on the crystallization. Different molding conditions (temperature, holding time) were studied in order to obtain the crystallized sample. The degree of crystallization was correlated with the heat deflection temperature (HDT) of the injection molded samples. 4.2 Experimental Section Injection molding grades of PLA 3001D, 3251D and 3051D (now 3052D) were obtained from Nature Works LLC (NE, USA). The Talc grade ABT2500 (average particle size of 2.3 micron) was kindly supplied by Specialty Minerals (PA, USA). The molecular weight was determined by gel permeation chromatography (GPC) equipped with a refractive index detector (Shimadzu, Tokyo, Japan, RID-10A) and a combination of three columns (Waters Co., Israel, Styragel HR1 THF and Styragel HR4E THF). Tetrahydrofuran (THF) was used as the mobile phase with a flow rate of 0.50 mL/min at 40 °C. The GPC data were calibrated with Poly Styrene (PS) standards. Thermal degradation was determined by a Q50 thermo gravimetric analyzer from TA instruments (DE, USA). Thermal properties were analyzed using a Q20 Differential scanning calorimeter (DSC) from TA instruments. Heat distortion temperature or deflection temperature under load (HDT) was determined using a Q800 DMA from TA instruments following ASTM D648.60 Mechanical properties were analyzed with United Testing Systems (UTS) model SFM-20 load frame following ASTM D3039. 45 Three different grades of PLA (3051D, 3251D and 3001D) resin pellets were dried overnight at 75 ºC prior to compounding. PLA resin pellets and talc were dry fed into a 30 mm ZSK30 twin-screw extruder manufactured by Werner Pfleiderer (NJ, USA) with a L/D of 30. The throughput was set to be 1.2 kg/h for talc and 4.8 kg/h for PLA resin pellets. The temperature profile of the extruder was set to be: 140, 165, 170, 175,175 and 165℃. The screw speed was set to be 125 rpm in all cases. The compounded melt strand was cooled in a water bath and pelletized before further use. Prior to molding, the compounded pellets were dried overnight in a convection oven at 75ºC to remove moisture. The injection molding experiments were performed on a 75 ton Milacron, Cincinnati injection molder equipment equipped with an oil heater. Injection molding of both room temperature and high temperature mold (100℃) were carried out. For high temperature mold, the holding time was set to be 180 seconds. 4.3 Results and Discussion 4.3.1 Molecular weight and thermal properties of different grades of PLA The molecular weights of the different PLA grades were determined by GPC, the results are listed in Table 4.1. 3001 D 3051 D 3251 D Grade 9.7*E4 9.9*E4 7.7*E4 Mw 7.03*E4 7.24*E4 5.78*E4 Mn 1.38 1.36 1.33 PDI Table 4. 1 Molecular weights of PLA samples 3001 D, 3051D and 3251 D From Table 4.1, it is apparent that the molecular weight of 3001D and 3051D are similar (Mw =1.0*105, Mn=7*104). However, the molecular weight of PLA grade 3251D is much lower compared with the other two PLA grades. In fact, 3251D has an Mw close to 80000, which is only 80% of the other two grades. In addition, all three 46 grades of PLA showed a similar PDI value which indicated a similar molecular weight distribution. The important point in this study was to compare the effects of the molecular weight on the rate of crystallization. Furthermore, the fact that all these samples had similar molecular weight distribution ruled out other factors such as low molecular weight tail and the present of monomers.61 In addition, melting temperature, the degree of crystallization as well as the crystal structure can be significantly affected by the meso content. It has been previously reported that as the meso content is increased the melting temperature of the final polymer decreases.62 An empirical equation in this study has been well adapted to describe such a relation: Tm (℃)  175 - 300 * Wmeso (Eq. 4.1) ,where Tm is the melting temperature of PLA determined from DSC and wmeso is the meso content in the polymer. In this study the melting temperature of each grade of PLA was determined by DSC and is listed in Table 4.2. And the meso content of each grade of PLA was also calculated from the empirical equation. (Eq. 4.1) 3001 D 3051 D 3251 D Grade 170 151 171 Tm [℃] 55-57 55-57 55-57 Tg [℃] 1-2% 8% 1-2% Meso Content* Table 4. 2 Molecular weight determination of PLA 3001D, 3051D, and 3251D *Calculated from previous empirical equation. It is apparent from the data in Table 4.2 that PLA 3051D showed a much lower melting temperature which corresponds to a higher meso content close to 8%; The other two grades of PLA showed a higher melting temperature, close to 170℃, indicating a much lower meso content. 47 4.3.2 TGA analysis of compounded product for thermal degradation and filler content determination Specific mechanical energy (SME) value was used to determine the energy input during the processing. The SME was calculated using Eq. 4.263: W  Torque SME  m RPM a RPM r (Eq.4.2) ,where RPMa is the speed of screw; RPMr is the factory rated screw speed and the value is 500 for the equipment used; W is the motor power of the extruder (90 KW for the equipment used); m is the processing mass flow rate. The calculation result is shown in Table 4.3. Grade Talc Content Extruder Torque (%) SME** by TGA RPM (KJ/kg) 3051D 20.34% 125 63 648 3001D 20.51% 125 68 715 3251D 19.86% 125 53 543 Table 4. 3 Processing Condition and Talc Content Analysis TGA was used to confirm the talc content of the extruded products. The organic portion was removed by heating the sample at 550℃ in air leaving only the inorganic portion (talc). The analysis of the results are very close to the expected value (Table 4.3). 48 4.3.2 DSC analysis of Compounded product for crystallization kinetics The mold temperature is critical parameter for a polymer to crystallize.64 In the injection molding process, the polymer melt was injected into the mold and held for a certain time at the mold temperature to obtain the crystallinity before it was cooled down. Thus, this process is essentially duplicating the cooling cycle from a high temperature melt to a lower temperature in the DSC experiments. Also, the annealing time, which is the amount of time the polymer remains in the mold, can also be simulated by holding the sample for a predetermined time isothermally in the DSC. It is also important to determine the optimal crystallization temperature because the rate of crystallization is greatly dependent on this factor. 64 DSC was used to determine this value by cooling the polymer from the melt (180℃) to room temperature to observe the peak crystallization temperature. In agreement with previous studies65,66, the peak crystallization temperature of our PLA samples was close to 100℃. After obtaining the value of the peak crystallization temperature, in the range of 90-100℃ (Table 4.4), 100℃ was implied in the DSC study for the rate of crystallization for all samples. The complete crystallization time as well as the peak time and that of the final degree of crystallization are given in Table 4.4. Grade Crystallizat ion Peak area (J/g) PLA degree Temperatu Complete Crystallizati of crystallizati on peak re (℃) crystallizati on (min) (min) on 3051D 21.27 28.6 103.84 3.4 1.03 3001D 44.86 59.8 90.3 1.7 0.59 3251D 38.84 51.8 90.2 1.5 0.36 Table 4. 4 DSC study of crystallization behavior of three grades of PLA. It is clearly observed fromTable 4.4 that the samples prepared from PLA grades with the lowest meso content showed a much faster rate of crystallization. In fact, PLA 49 3001D and 3251D required only half of the time compared with PLA 3051D which contains 8% meso content. Also the degree of crystallization of these samples after annealing was obtained by reheating the sample again in the DSC to observe the melting entropy. This value was calculated using Eq. 5.3: X%   1   100 , (Eq.5.3)  c 1  wt ,where X% is the degree of crystallization, H is the entropy of melting of the sample, Hc is the theoretical heat of fusion of PLA crystals (93.7 J/g), and wt is the weight percentage of talc. As shown in Table 5.4, the degree of crystallization of these samples follows a similar trend where the PLA grades containing less meso content showing a higher degree of crystallization. PLA 3001 D showed the highest value, around 60% whereas PLA 3051 D showed a value of only around 30%. The degree of crystallization values obtained in this study is similar to those reported elsewhere. 67 When PLA samples 3001D and 3251D are compared (having similar low meso content), the molecular weight effect can be observed. Generally, the lower molecular weight samples showed a higher rate of crystallisation and a lower degree of crystallization. The difference is much smaller (10%) when the meso content is compared. 4.3.3 HDT analysis and crystallinity of Compounded product for two different mold temperatures As discussed in a previous section, when the temperature is above the glass transition temperature of a polymer, only the crystalline phase of the material 50 contributes to the modulus whereas the amorphous phase starts to flow. In order to enhance the crystallinity, a longer holding time as well as a higher mold temperature was used. Although some degree of crystallization can be achieved when the mold is kept at room temperature, the cooling rate is generally too high for PLA to obtain a sufficient degree of crystallization.67 In this study, the holding time was set to be 180 seconds due to limitations in the injection molding equipment. For sample to be successfully removed from the mold, it should have a certain level of strength either by cooling the PLA below its glass transition temperature or by increasing the crystallization. Room temperature and 100℃ were tested in the study. The injection molded sample bars were tested for HDT following ASTM standard D648. The three point bending method applies a steady pressure on the sample as the temperature is increased to the formation of the sample. The deformation was recorded and the HDT was determined based on the temperature at which the sample deformation reaches a certain value. The testing results are listed in Table 5.5. It is important to note that even with 3 min of holding time, the samples prepared from PLA 3051D could not be removed from the mold. 51 PLA grade Mold Temperature X% HDT [℃] 3051 D RT 3.0 54.1 3051 D 100℃ 3251 D RT 25.1 54.1 3251 D 100℃ 48.6 138.0 3001 D RT 25.3 54.4 3001 D 100℃ 58.6 130.1 ** ** Table 4. 5 Effect of mold temperature on the degree of crystallization and HDT ** Sample failed to be removed from the mold. It is clear from the data in Table 4.5 that samples prepared from PLA samples 3001D and 3251D showed a higher degree of crystallization as well as a high HDT value when the mold temperature was increased to 100℃. The two samples showed a HDT of higher than 130℃, which is higher than most of the application requirements of PLA resin. Also, although some of the samples obtained a certain degree of crystallization (25%), the HDT was still close to the glass transition temperature of PLA (56℃). Such a dramatic increase in the HDT temperature is simply the result of the physical definition of HDT. Thus, since the sample fails automatically when a certain level of deformation is reached, any sample with a degree of crystallization lower than a certain limit will fail at the Tg of PLA. If follows that only when a certain degree of crystallization is obtained, the sample shows a HDT temperature higher than 56℃. Thus, there is no clear trend with sample regardless of the degree of crystallization. 52 4.4 Conclusions In this study, three different grades of PLA: 3001D, 3251D, and 3051D were analyzed for their meso content and molecular weight. Based on this study, the molecular weight, the meso content and the mold temperature were tested to determine their effects on the crystallization behavior of PLA. It was found that lower molecular weight samples showed a higher rate of crystallization whereby the lower meso content sample exhibited higher rate of crystallization as well as higher degree of crystallization. Room temperature mold holding was determined to be unsuitable to obtain PLA samples with a high degree of crystallization due to the high cooling rate of the melt. With sufficient mold holding time (3 min) at a high mold temperature (100℃), it was possible to obtain sample bars with a glossy finish. The HDT of injection molded bars was determined to be higher than 130℃ according to ASTM 648D. 53 Chapter 5. Depolymerization Kinetics of Poly(lactic acid) to Lactide 5.1 Introduction Polylactic acid (PLA) is well known as a bio-based and biodegradable polymer derived from lactic acid, which is obtained by bacterial fermentation of corn starch.68 However, simply polymerizing lactic acid is not an effective process as it yields low quality polymeric products. Direct polycondensation of lactic acid generates one molecule of water for every condensation step and the reversible nature of this polycondensation leads to products that are characterized by high polydispersity. Therefore, lactic acid is first oligomerized and then catalytically dimerized to produce the cyclic lactide monomer. High molecular weight PLA is produced from this lactide monomer by ring-opening polymerization using various metal salt catalysts. Originally, the manufacturing process was relatively costly, however, the use of modern industrial production techniques pioneered by Cargill Inc., have greatly reduced the cost of production69. As a result of these improvements, PLA is used today in many applications including packaging, electronics, and textiles.70 The most common end-of-life options for PLA are either composting or recycling. As a biodegradable material, PLA decomposes in industrial compost environments or in natural environments.71 Alternatively, PLA can be collected and then recycled or reused as a polymer or hydrolyzed to lactic acid. Unlike polyolefins, the ester groups in PLA easily undergo hydrolysis to yield low molecular weight oligomers and eventually lactic acid. Indeed, the current recycling approach is to fully hydrolyze the 54 polymer to lactic acid using boiling water or steam72 (Figure 5.1). The lactic acid obtained by this process can be re-used in the production of ‘fresh’ PLA or in other applications. This recycling approach is applicable to either a waste material in the manufacturing plant as well as post-consumer materials that contain PLA. Unfortunately, this approach requires significant amounts of water for the hydrolysis step as well as extra steps that are related to purification of the lactic acid and its conversion to lactide before it can be re-used. Figure 5. 1 Recycle processes of PLA A preferred approach is to thermally degrade PLA under controlled conditions and recycle it directly to the lactide (Figure 5.1). 73 Nishida et al. have studied the pyrolysis degradation mechanism of PLA in the presence of a tin catalyst.74 They found that the pyrolysis of Poly (L-lactic acid) (PLLA) in a dynamic heating mode starts through a 55 random degradation process which then shifts to a zero order weight loss. They also examined the effect of different end-groups on the depolymerization reaction and concluded that it was negligible. 75-77 Jamshidi et al. investigated the effect of temperature during the thermo-degradation process of PLA and correlated the overall degradation process with depolymerization, and cyclic oligomerization, as well as intermolecular and intramolecular transesterifications. 78 Generally, PLA is characterized by a chiral center and, hence, two optically active isomers: L-lactic acid and D-lactic acid. L-lactic acid is the only naturally occurring isomer. However, these two lactic acid isomers can lead to three lactide isomers (DD, LL and DL).73 The transformation from one type of lactide to another commonly occurs by racemization during the depolymerization process. Tsukegi et al.79 observed that in the absence of a catalyst the racemization of lactide increased with temperature and time. Furthermore, the effect of magnesium oxide as a catalyst on the racemization of lactide was studied by Motoyama80 who concluded that MgO is both a catalyst for depolymerization and racemization. Shuklov et. al. 81 found that the present of a base catalyst and higher reaction temperatures facilitated the racemization reaction. 5.2 Study of depolymerization reaction kinetics using TGA Thermogravimetric analysis (TGA) has been widely used to study the degradation of various polymers. The advantages of this technique include high sensitivity, easy control of the temperature and the need for small sample size. In this work, the focus was on key parameters affecting the isothermal depolymerization of PLA directly to the lactide. The degradation process in our work is comparable to common industrial 56 batch processing conditions and as such, the results can be used for large scale productions to improve the efficiency of the PLA recycling process. 5.2.1 Experimental section PLA resin (3051D) was obtained from NatureWorks (Blair, NE, USA). Stannous octoate catalyst was purchased from Sigma Aldrich (St. Louis, MO, USA) and used as received. Methylene chloride was purchased from J.T.Baker (Center Valley, PA, USA). All samples were dried at 80℃ overnight before being used. Typically, 1 g of PLA resin was dissolved in 20 g of methylene chloride using different concentrations of catalyst (0.6%, 1%, 2%, 4% and 10%). Each solution was then cast into a glass petri dish to yield about 1 mm thick film. The cast films were then dried under vacuum to remove all the solvent. Q 20 DSC from TA Analysis (Castle point, DE, USA) was used to study the equilibrium monomer concentration at high temperature. The film sample (approximately 10 mg) was sealed in a hermetic aluminum pan and hold at constant temperatures (180, 200, and 220℃) for a certain period of time. The sample was then quenched in liquid nitrogen. The sealed pan was opened after quenching and the sample was dissolved in deuterated chloroform (CDCl3). 1H NMR was conducted on a Unity Plus 500 MHz NMR spectrometer from Varian Inc. (Palo Alto, CA, USA) to determine the lactide content. In a typical analysis, a sample was dissolved in CDCl3 and the spectrum was recorded using the solvent peak as the internal standard. Q50 TGA from TA Analysis was used for the thermo-degradation analysis. A cast film sample (approximately 5 mg) was placed in an aluminum pan under a constant nitrogen flow 57 (50 ml min-1) and held isothermally at different constant temperatures (160, 170, 180, 190, 200, 210, and 220℃). The weight loss was recorded at an interval of 0.0083 s for 1h and the results obtained were used to calculate the Avrami parameters82,83 84. 5.2.2 Results and Discussion 5.2.2.1 Equilibrium monomer constant and rate of depolymerization The polymerization rate constant as well as the depolymerization rate have been extensively discussed in previous work73. The depolymerization reaction follows a zero-order mechanism: the rate of depolymerization only depends on the catalyst concentration. The rate of depolymerization can be expressed as: Rd  [Cat ]  k d (Eq.5.1) ,where Rd is the rate of depolymerization, [Cat] is the concentration of catalyst, kd is the rate constant for depolymerization reaction. On the other hand, the rate of polymerization reaction has been well studied and can be expressed as: R p  [ M ]  [Cat ]  k p (Eq.5.2) ,where [M] is the concentration of lactide monomer, kp is the rate constant for polymerization reaction. From previous work73, we have obtained the fundamental reaction constants for kp. According to Arrhenius equation, the rate of polymerization at a certain temperature can be calculated with the following equation: k p  A448 exp(  Ea 1 1 (  )) (Eq.5.3) R T 448 58 ,where A448 is the preexponential constant and Ea is the activation energy of the polymerization reaction. In our previous work, such value has been reported as A448= 86 h-1 cat mol%-1 and Ea= 70.9 kJmol-1.73 A series of calculation was conducted using such equation and reaction constant to obtain the reaction constant of polymerization reaction at a certain temperature and the results are shown in Table 5.1. Temperature Me wt percent kp kd [℃] [%] h-1 cat mol%-1 h-1 cat mol%1monomer wt% 160 2.15 45.11 0.97 180 3.00 106.51 3.20 190 3.27 159.16 5.20 200 4.91 233.83 11.45 220 5.80 481.63 27.94 Table 5. 1 Equilibrium monomer constants and the corresponding rate constants of the depolymerization reaction at different temperatures The monomer equilibrium concentration, [Me] can be expressed using the equilibrium state assumption: at equilibrium stage, the rate of polymerization Rp matches with the rate of depolymerization Rd. R p  Rd (Eq.5.4) By substituting eq.5.1 and eq.5.2 into eq.5.4, we can obtain the equilibrium monomer concentration: Me  Rd (Eq.5.5). Rp 59 Thus, the equilibrium monomer concentration is independent on the catalyst concentration. Such constant concentration is the ratio between the rate of polymerization and rate of depolymerization at a given temperature. Also, the concentration of catalyst does not affect the equilibrium monomer concentration. Rearrangement of this equation gives the expression for kd which is the reaction constant of the depolymerzation: k d  M e  k p (Eq.5.6) A series of monomer equilibrium concentrations obtained from previous work and the DSC study have been listed in Table 5.1. The calculated value of kd was obtained from eq.5.6 using the value of kp and monomer equilibrium concentration. The Arrhenius plot of the depolymerization constants kd is shown in Figure 5.2. High linear coefficient (a=0.9955) was obtained. From the linear regression of such data, the activation energy of the depolymerization reaction was calculated to be 12.1 kJ mol-1. 60 Figure 5. 2 Arrhenius plot of depolymerization reaction constant kd A monomer concentration versus time plot showing the rate at which the depolymerization reaction occurs was also obtained (Figure 5.3). By heating PLA polymer in the present of catalyst, the rate of depolymerization can be obtained as a function of time and temperature. The monomer equilibrium constant depends on the reaction temperature. At a higher temperature, the equilibrium stage was reached in a much shorter time and the monomer equilibrium concentration was higher. Such phenomena can be explained by the fact that both the polymerisation and depolymerisation reactions proceeded faster at a higher temperature. 61 Figure 5. 3 Lactide concentration of PLA containing 0.6% catalyst as a function of time at different temperatures: 160℃(Δ), 180℃(○), and 200℃(□). Comparing with the process of recycling lactide by thermo depolymerization, two major differences should be noted. First of all, due to the nature of such reversible reaction, the equilibrium monomer concentration is always low. Thus, constant removal of the lactide from the system is required to drive the equilibrium of the system to depolymerization and obtain higher yield of lactide. Secondly, all the previous work was based on low concentration of catalyst. As the depolymerization reaction is a zero order reaction, the depolymerzation reaction will only proceed at a very low rate. To obtain a desirable rate of depolymerization and increase the efficiency of recycling, a higher catalyst concentration is needed. 62 5.2.2.2 Thermogravimetric analysis TGA was used to follow the depolymerization process of PLA. The lactide formed was removed by the continuous nitrogen flow in the TGA such that the weight loss was determined quantitatively as a function of the depolymerization reaction.85 Five different temperatures and four different catalyst concentrations were studied (Table 5.2). Representative TGA spectra of samples containing 0.6 wt% catalyst at different temperatures (160, 170, 180, 190, 200, and 210℃) were shown in Figure 5.4. Figure 5. 4 Depolymerization of PLA containing 0.6% catalyst at different temperatures At lower temperatures (160-180℃), the rate of the depolymerization reaction was relatively slow but the rate of reaction substantially increased at higher temperatures. After 60 minutes the total weight loss was less than 50% at 160℃, but a full weight loss which indicates high yield of depolymerzation was observed at 210℃. It was further 63 apparent from the data that the rate of the depolymerzation reaction (and the formation of the lactide) was directly proportional to the reaction temperature. Catalyst Temperature [wt. %] [°C] 0.6 160 0.6 n R2 ln(K) ln(k2) 1.01 -7.33 0.056 0.9693 170 1.04 -6.50 0.576 0.9877 0.6 180 1.39 -6.51 1.180 0.9989 0.6 190 1.51 -6.10 1.758 0.9981 0.6 200 1.43 -5.13 2.312 0.9813 0.6 210 1.71 -5.04 2.842 0.9983 1.0 190 1.46 -6.97 1.758 0.9581 1.0 200 1.80 -5.49 2.312 0.9966 1.0 210 1.92 -5.22 2.842 0.9899 1.0 220 2.31 -5.04 3.351 0.9914 2.0 160 1.06 -6.81 -0.056 0.9863 2.0 180 1.33 -5.99 1.180 0.9994 2.0 200 1.73 -5.16 2.312 0.9997 4.0 160 1.18 -5.48 -0.056 0.9998 4.0 180 1.47 -4.89 1.180 0.9969 4.0 200 1.39 -3.76 2.312 0.9961 10.0 160 1.17 -4.96 -0.056 0.9954 10.0 180 0.95 -3.25 1.180 0.9995 10.0 200 1.32 -3.07 2.312 0.9994 Table 5. 2 Reaction temperature and catalyst concentration of PLA degradation in the TGA experiments When the temperature was kept constant at 200℃, increasing the catalyst concentration from 0.6% to 4%, led to a significant increase in the rate of weight loss (Figure 5.5). A number of previous studies have addressed the effect of the chain-end 64 functional groups on the degradation of PLA and concluded that the depolymerization reaction was not affected by the nature of the end-groups.85-87 In this study, no attempt was made to vary the terminal chemical groups of PLA. However, it was noted that the rate of the thermal depolymerization reaction was heavily dependent on the catalyst concentration compared with which the concentration of the terminal groups was very small. Figure 5. 5 TGA decomposition of PLA containing different concentrations of catalyst at 200℃. 5.2.2.3 Mechanism of the Depolymerization Reaction It is apparent from the degradation kinetics in Figures 5.4 and 5.5 that the decomposition rate of PLA is directly proportional to the temperature and the catalyst concentration. Unlike the zero order depolymerization reaction under equilibrium conditions that was obtained previously6, where the rate of the depolymerization was only a linear function of the catalyst concentration, a more complex kinetic pattern was 65 observed here. In all our experiments the rate of depolymerization followed sigmoidal shape patterns clearly indicating that the mechanism of this non-equilibrium depolymerization is different than simple zero order rate. It is apparent from the TGA data that the rate of weight loss in all cases initially increased with time suggesting an initiation step in the degradation mechanism. This initiation step is similar to the polymerization reaction of lactide to PLA where the catalyst first forms an “activation site” with the hydroxyl compound73. However, in this case the activation reaction is not significant because of the longer reaction time (8 to 10 h) involved. Based on these observations the following two step mechanism is proposed for the non-equilibrium depolymerization reaction; the initial “activation” step involves the reaction of the catalyst with an ester linkage to form an intermediate complex (Figure 5.6). The formation of this activation complex can occur anywhere along the polymer chain and is not specific to the chain-ends. In the second step, chain scission of the activation complex leads to the formation of the lactide. This depolymerization reaction follows an “unzipping” mechanism (Figure 5.6) where the activated chain end “back bites” a neighboring ester linkage to form the lactide monomer. 66 Figure 5. 6 Proposed two-step reaction during the depolymerization of PLA in the presence of catalyst 5.2.2.5 Theoretical Model of Depolymerization The depolymerization reaction can be described by the Avrami equation. Although this equation was developed to describe nucleation and crystallization kinetics in polymer systems,88,89 it has also been used to analyze various kinetic reactions including thermal decomposition of various solids.90,91 Many studies have also reported simulating linear polymers depolymerization with Avrami equation.92,93 In its linear form it is expressed as: ln( ln( X ))  ln K  n ln(t ) 67 (Eq.5.7) ,where X is the fraction untransformed at time t, K is the overall rate constant and n is the Avrami exponent related to the reaction mechanism and is usually in the range of 1-4. Thus, a plot of ln(-ln(X)) vs ln(t) yields a linear plot where the slope is the Avrami exponent (n) and the intercept is the natural log of the rate constant, ln(K). Plotting our TGA data using the linear form of the Avrami equation (Eq. 5.7) produced a series of linear functions as predicted by this model. Figures 5.7 and 5.8 are representative plots of the data at different catalyst concentrations and different temperatures, respectively. Plots of all our data in this form yield linear relationships (Table 2) and indicate that the Avrami model can be used to describe the depolymerization mechanism. 68 Figure 5. 7 Avrami plot of PLA depolymerization TGA data at containing different concentrations of catalyst at 200℃: 1% (■), 2% (▲), 4% (♦), and 10% (●) An analysis of the parameters of the Avrami equation indicates that the overall rate constant K, derived from the intercept, is greatly affected by, both, the catalyst concentration (Figure 5.7) and the temperature. This is to be expected since the rate at which the unzipping reaction occurs depends on these two variables. However, the value of n, which represents the Avrami exponent related to the reaction mechanism, is primarily affected by the degradation temperature and much less by the catalyst concentration. In the original development of the Avrami equation, the two major steps were the “nucleation” and subsequent crystal growth. Although these two steps have different rate constants, the equation contains only a single rate constant that describes the overall 69 crystallization process. Since our depolymerization reaction also consists of two steps, the overall rate constant, K, contains three terms: the rate of initiation, the rate of depolymerzation kd and the catalyst concentration C1. Thus, we can obtain linear plots with ln(K) and ln (C1) for three different temperatures (160, 180 and 200℃) (Figure 5.8). Figure 5. 8 Correlation of the overall reaction constant ln(K) with the catalyst concentration ln(C1) : 160℃(▲), 180℃(♦), and 200℃(■) C1 is directly related to the overall rate constant K. Thus, increasing the reaction temperature leads to higher K values, indicating that the overall rate is directly proportional to the temperature. The rate constant of depolymerization, kd has been calculated in previous discussion. A represented plot of ln (kd) and ln (K) was shown in Figure 5.9. 70 Figure 5. 9 Correlation of the overall reaction constant ln(K) with the depolymerization reaction constant ln(kd) at 0.6% catalyst concentration A positive correlation can be observed from the plot. Such plots further proved the fit of the Avrami equation and validated the proposed mechanism. The Avrami exponent, n, is in the range between 1 and 2 with somewhat dependent on the temperature (Figure 5.10). Such dependency has also been reported elsewhere in other thermal degradation of polymers.90 Apparently, at high temperatures the mobility of the chain increases and leads to changes in the unzipping rate and possibly the mechanism of reaction. Furthermore, the change in n could also be affected by a chain transfer reaction of the activation site from one chain to another and thus cause the change in order of reaction (n). 71 Figure 5. 10 Correlation of the Avrami exponent with reaction temperature 5.2.3 Conclusions In this work, a feasible route to recycle PLA directly to lactide by thermal depolymerization is described. The kinetics of the depolymerization reaction of PLA under isothermal non-equilibrium conditions was studied. A series of TGA experiments were performed to study the effect of catalyst and temperature on the rate of depolymerization. The data confirmed previous results that the catalyzed depolymerization reaction is not limited by the chain-end functional groups of the PLA chains. A two-step reaction mechanism was proposed in this study; the first step is the formation of activation sites of the catalyst with an ester linkage on the polymer chain; the second step is an “unzipping” reaction of the PLA chain from the activation sites. 72 The Avrami equation appears to fit to the experimental TGA data and provides quantitave information of the fundamental parameters of this reaction mechanism. 5.3 Recycle process design using thermal depolymerization From previous discussion, the effect of temperature, catalyst concentration on the reaction mechanism are studied. Our data indicate that the catalyzed depolymerization reaction is not limited by the chain-end functional groups of the PLA chains. Excess catalyst can be inserted into an ester linkage any place along the polymer chain to form an activated site for the depolymerization. Based on this mechanism the following recycle process is proposed. 5.3.1 Sizing We have chosen a processing capacity of 600 ton/year of recycled PLA as a general example. If a scaling up or scaling down is needed for specific situation, most of the values denoted here can be easily changed. For a plant with a production capacity of 600 ton/year and 300 working days per year, the daily processing capacity should be 2 ton/day. Using the polymer melt density to be 1.25 ton/m3, it follows that the total volume of the reactor has to be 5 m3 using a safe volume of 3 times the polymer content. 5.3.2 Process design In the designed process, the recycle plastics are dried to remove moisture. The dried plastics are then fed into the depolymerization reactor together with the depolymerization catalyst. The reactor tank is held under reduced pressure at 220℃. The reactor is equipped with a distillation column to obtain the liquid lactide from the gas phase of the reactor. A schematic drawing of the reactor equipped with a top mounted distillation column and the material flow chart is shown in Figure 5.11. 73 Vacuum Line Vacuum Pump Motor Cooling Water In Water Pump Lactide Condensor Condenser E-2 E-3 E-4 Cooling Water Out E-5 Liquid Lactide Water Cooling Recycled PLA Catalyst Lactide Container Reactor Waste Draining Valve Polymer Residue Tank Figure 5. 11 Scheme of the depolymerization reactor equipped with top mounted distillation column and material flow chart The operating condition of the reactor is determined based on the reaction condition of the depolymerization reaction. From our previous discussion, the depolymerization 74 reaction proceed much faster at higher temperatures. However, too high temperature is applied (above 230℃), PLA degradation and undesired reactions occur. Thus we have designed the reactor at 220℃. From the previous literatures, the vapor pressures of LL, DD and meso lactides at 220℃are shown in Table 5.394. Lactide Temperature [℃] Vapor Pressure [mmHg] Vapor Pressure [kPa] Meso 220 274.2 36.5 L,L 220 267.4 35.7 D,D 220 267.4 35.7 Table 5. 3 Vapor pressure of lactides at 220℃94 Thus, the reactor and condenser should be operated under reduced pressure of 35.7 kPa. The melting temperature of L,L lactide,D,D lactide, meso lactide and D,L lactide are 56℃, 97℃, 97℃ and 127℃ respectively.94 To prevent the solidification of liquid lactide, we designed the condenser temperature at 130℃, which is above the melting temperature of all the lacides. The operating conditions of the equipment are listed in Table 5.4. Equipment Compacity Temperature [℃] Pressure [kPa] Reactor 5m3 220 35.7 Condenser -- 130 35.7 Table 5. 4 Specifications and operating conditions for depolymerization reactor and condensers 5.3.3 Materials Balance The overall material balance of the process can be expressed in Equation 5.8. 75 W p  Wc  Wl  Ww (Eq.5.8) Where Wp and Wc are the weight of recycled polymer and the weight of catalyst; Wl is the lactide obtained from gas phase and Ww is the weight of residue mixture after the depolymerization reaction. Using 1% weight percent of catalyst dosing for a 2 ton/day capacity and assuming 90% designed yield, the value of these parameters are summarized in the Table 5.5. Material Abbreviation Weight [Kg] Recycled PLA Wp 2000 Catalyst Wc 20 Lactide Wl 1800 Residue waste Ww 220 Table 5. 5 Summary of materials balance for depolymerization process 5.3.4 Energy balance The energy balance of the depolymerization reactor is designed based on the energy required for each reaction or physical phase transformation. The total energy required to heat the reactants in the reactor and to drive the reactions forward can be expressed as: H t  H h  H r  H e (Eq.5.9) , where Ht is the total amount of energy required for the process; Hh is the energy required to heat up the reactants to the desired temperature; Hr is the energy required for the depolymerization reaction and He is the energy required to evaporate the lactide product into the gas phase. 76 For Hh, which is the energy required to heat up the reactants to desired temperature, it can be expressed as: T H h  (  C p (T )dT  H m * X %)Wt (Eq.5.10) To Here, To is the starting temperature, 25℃, T is the final reaction temperature at 220℃. Also since the concentration of the catalyst is very small (1%) it can be neglected in this calculation. In previous studies94, the specific heat capacity of amorphous polylactides is expressed as: C p (T )  1.71  0.00147T ( o C ) (Eq.5.11) The melting enthalpy (ΔHm) of pure PLA is 93.7J/g. However, the degree of crystallization (X%) of the recycled PLA is most commonly given as 30% of this value. Thus, by integration of eq. 5.10 using eq. 5.11 for Cp, and daily processing throughput of 2000kg (Wt) one can get the energy required to heat the reactants to the desire temperature (Hh) from the following equation: 0.00147 * (T 2  T02 )  H m * X %) * Wt (Eq.5.12) 2  396.67kJ / kg * 2000kg  220kWh H h  (1.71 * (T  T0 )  The depolymerization reaction enthalpy value has been reported95 to be ΔH=23.15 kJ/mol=160.76 kJ/kg. With a 90% final conversion, the total energy required would be: 77 H r  H * Wl  160.76kJ / kg *1800kg  80kWh (Eq.5.13) ΔHe is the energy required to evaporate the lactide product into the gas phase and is reported94 to be 56 kJ/mol=.388.9kJ/kg H e  H e * Wl  388.9kJ / kg *1800kg  195kWh (Eq.5.14) Thus, the energy balance for the reactor can be summarized in the following Table 5.6. Item Requited Energy [kJ/kg] Quantity [kg] Energy/operation [kWh] Hh 396.67 2000 220 Hr 160.76 1800 80 He 388.9 1800 195 Ht -- 1800 495 Table 5. 6 Summary of the energy balance of the process Thus, the energy required to produce 1800 kg of lactide is 495kWh or, on weight bases, for each kilogram of lactide recovered, the energy required is 0.275kWh. Similarly, the energy balance of the condenser required to remove the lactide vapor at 220℃ to 130℃ so that a lactide liquid can be obtained from the following equation: Q  H c  C p * T (Eq.5.15) Where Q is the total heat needed to be removed in order to maintain the condenser temperature and ΔHc is the enthalpy of lactide condensation (56kJ/mol =-388.9kJ/kg). Using the specific heat capacity of lactides as given in equation 5.11, Q can be calculated from the following equation: 78 0.00147 * (T 2  T02 ) Q  H c  1.71 * (T  T0 )  (Eq.5.16) 2 Thus, using T=130℃, T0=220℃and ΔHc=-388.9kJ/kg, the heat required by the cooling water system in the condenser (Q) is calculated to be 520kJ/kg. 5.3.5 Conclusions and Future works From the energy balance and materials balance discussed above, recycling PLA directly to lactide is an alternative method with great interest. In our calculation, the energy required to obtain lactide from recycled PLA is 0.245kWh, which is much lower than the reported values96,97 of the lactide people obtain starting from biomass. Here, we should note that, directly comparing the values might not be sufficient to draw the final conclusion because many important factors such as shipping energy use, collecting energy use, have not been considered. Future studies should consider these factors to further study the feasibility for the PLA thermal depolymerization recycle process. 5.4 Effect of the present of plastic mixture on the recycling process In any recycled PLA operation one has to consider other components in the input stream such as other plastic materials, various additives and inorganic fillers. In many of the current processing operations PLA is blended with EcoflexTM - poly (butylene adipate-co-terephthalate (PBAT) and inorganic fillers such as talc that are commonly added into PLA in order to improve the performance of the product. These components can be very difficult to separate from the PLA mainstream. Indeed, in traditional hydrolysis methods, where PLA is hydrolyzed to lactic acid, these components could 79 cause severe contamination to the lactic acid stream. In contrast, one of the advantages of recycling PLA via thermal depolymerization is the ability to obtain high purity lactide in the vapor phase by distillation and avoid the separation step. The distillation of lactide from pure PLA via the depolymerization reaction has been discussed in previous discussions. Here, a series of experiments were conducted to study the presence of other plastic components (PBAT and Polycarbonate) in PLA and their effect on the efficiency of the thermal recycling. 5.4.1 Experimental PLA 3051 D with a molecular weight of ~100,000 and 4% D content was purchased from Natureworks, LLC. (NE, USA). EcoflexTM, which is the trade name of poly (butylene adipate-co-terephthalate) or PBAT was obtained from BASF, LLC. (MI, USA). The physical blend of PLA and PBAT (10%, 25% and 50%) were prepared using melt extrusion. PLA-Polycarbonate (PC) physical blends (20%, 30%, 40% and 60% PC content) were supplied by Bio-Ecocompatible Materials Lab (Pisa University, Italy). 2Ethylhexanoic acid tin(II) salt (a.k.a. tin(II) ethylhexanoate, stannous octoate), 405.11 g/mol, was purchased from Aldrich (MO, USA) and used without further purification. CH2Cl2 was purchased from J.T.Baker (PA, USA). Thermo gravimetric analysis (TGA) was conducted on a TGA (Q50, TA analysis, USA). In a typical experiment, a sample (~5 mg) was placed in aluminum pan under a constant nitrogen flow and the weight loss data were collected at regular time intervals (0.008333s) over time at a constant temperature. 80 A polymer resin (24 g) was dried at 80℃ overnight to remove moisture. The resin and 2-ethylhexanoic acid tin (II) salt catalyst (1.0 g) were added into a one neck flask. The depolymerization experimental was carried out at 210℃ (heated with an oil bath) under vacuum (50 Pa). The volatile product was collected in a cold trap for further analysis. 5.4.2 Results and Discussion The data in Figure 5.12 clearly indicate that the weight loss is directly proportional to the concentration of the lactide in the sample. Figure 5. 12 TGA thermograms of PLA-PC blend with different ratio (A: PLA20/PC800, B: PLA30/ PC70, C: PLA40/PC60, D: PLA60/PC40) isotherm at 200℃ with 1% catalyst concentration 81 Furthermore, at long reaction times, the weight loss curve flattens to a plateau with a final weight that is very close to the amount of PC present in the blend. Such observations indicate that the PLA portion was successfully removed from the system by depolymerization to lactide and only the PC portion remained in the system. Comparing these data with the previous recycling experiments discussed in section 5.2, where pure PLA was used, it appears that the rate of the weight loss in this series of studies is similar and only slightly smaller than observed for pure PLA. This minor discrepancy can be explained by the limited mass transfer and possible side reactions caused by the presence of polycarbonate in these blends. Similar experiments were conducted with PBAT, and similar results were observed. However, large amounts of PBAT showed a more significant impact on the yield of the recycling process. The experimental conditions as well as the final yield are presented in Table 5.7. Polymer PLA Content [%] Catalyst Content [%] Temperature Recycle Yield [℃] [%] PBAT 100 0.1 210 99.6 PBAT 90 0.1 210 99.0 PBAT 75 0.1 210 81.6 PBAT 50 0.1 210 68.2 PC 80 1 200 97.8 PC 70 1 200 100.0 PC 60 1 200 92.3 PC 40 1 200 91.3 Table 5. 7 Recycling yields of the lactide in PLA-PC and PLA-PBAT mixtures 82 5.4.3 Conclusions The recycling of PLA from mixtures of polymers was studied and discussed. PBAT and PC were used as examples to test the feasibility of the thermal depolymerization recycling process. The results indicate that the depolymerization reaction proceeds smoothly in blends containing these polymers (PC and PBAT). High yields (larger than 80%) of the recycle lactide was obtained from these small-scale setups. Although somewhat lower rate of recycling was observed compared with pure PLA, the difference was insignificant and would have only a minor effect on the feasibility of this recycling process. By obtaining the lactide from the gas phase it can be separated from the blends without the need for extensive purification and separation processes. 83 Chapter 6. The Use of Glycerol Carbonate in the Preparation of Highly Branched Siloxy Polymers 6.1 Introduction Glycerol is an important and readily available chemical derived from biomass feedstock. In recent years, it became even more abundant as a by-product from the production of fatty acid methyl esters (FAME). Glycerol has been the subject of many publications related to its use in the production of acrylic acid, propanediol, dihydroxyacetone, succinic acid, polyglycerol esters, etc.98-100 Several publications are available in the literature dealing with the conversion of glycerol to glycerol carbonate (GC). Ochoa-Gómez et al. obtained ((4-hydroxymethyl)-1,3-dioxolan-2-one from the reaction between glycerol and dimethyl carbonate.101 Cho et al. have also reported that GC could be obtained by transesterification of ethylene carbonate with glycerol in ionic liquids.102 It was also noted that GC can be polymerized by anionic polymerization. Anioic polymerization of GC with deprotonated trimethylolpropane as an initiator was used to obtain hyperbranched aliphatic polyether.103 Such hyperbranched polymers are a subclass of dendritic polymers and exhibit properties that can be significantly different than their linear analogs104. Some of their important characteristics include low viscosity even at high molecular weights, lack of crystallinity, and exceptionally high solubility in many solvents. Biomass-based hyperbranched polymers derived from alkyds are known and have been used as air drying, low viscosity resins.105,106 84 Alkoxy silanes are widely used as coupling agents, polymer additives, resin precursors and crosslinking agents. Various 3-aminopropylalkylalkoxylsilanes are commercially available and are commonly used in textile, cosmetic, and food industries. The structure and the properties of the products derived from these aminosilanes depend on the number of alkoxy groups attached to the silicon atom. Various silicones have been used in shampoos and conditioners because of the flexible nature of the siloxanes bonds which imparts softness, silky sense and dry feel to the hair. However, due to the fact that most siloxanes are water insoluble, the products has often been delivered from emulsions. Additionally, attempts have been made to prepare hydrophilic silicone graft and block polymers where water soluble groups (such as polyethylene glycol) have been incorporated to their structures. We have already reported on the preparation of hydrophilic hyperbranched silane polymers derived from 3-aminopropyldiethoxymethylsilane and ethylene carbonate.107,108 It should be emphasized that these reactions proceeded smoothly in the bulk with no need for a solvent or a catalyst and the extent of the polymerization could simply be determined by the amount of ethanol removed as a by-product. Due to the inherent hydrophilic property of hydroxyl functional groups within the silane molecule, the products exhibited unique hydrophilic properties. Herein, we report the synthesis of hydroxyl terminated silanes derived from biomass-based GC with three different aminosilanes: 3- aminopropyldimethylethoxysilane (MESi), 3-aminopropyldiethoxymethylsilane (DESi) 85 and 3-aminopropyltriethoxysilane (TESi). These alkyl hydroxyl terminated silanes were then further polymerized to yield highly branched siloxy polymers. The reaction kinetics, the structure and some key properties of these silanes and siloxy polymers were studied. 6.2 Materials and methods/ Experimental Procedure The silanes 3-aminopropyldimethylethoxysilane (MESi), 3- aminopropyldiethoxymethyl-silane (DESi), and 3-aminopropyltriethoxysilane (TESi) were purchased from GELEST, Inc. (PA, USA). These aminosilanes were kept under nitrogen atmosphere to prevent premature hydrolysis. Glycerol carbonate (GC) was kindly supplied by Huntsman Petrochemical Co. (TX, USA). All of chemicals were reagent grade and were used without further purification unless noted. 1 H NMR spectra were recorded on a 500 MHz NMR spectrometer (Varian Inc., USA, Unity Plus 500MHz) using a solvent peak of DMSO-d6 as internal standard. Fourier transform infrared spectroscopy (FTIR) spectra were obtained on a Shimadsu spectrophotometer (Shimadsu Co., Tokyo, Japan, IRAffinity-1) equipped with a single reflection ATR system (PIKE Technologies, Madison, USA, MIRacle ATR). The molecular weight was determined by gel permeation chromatography (GPC) equipped with a refractive index detector (Shimadzu, Tokyo, Japan, RID-10A) and a combination of two columns (Waters Co., Israel, Styragel HR1 THF and Styragel HR4E THF). Tetrahydrofuran was used as the mobile phase with a flow rate of 0.50 mL/min at 40 °C. Thermal degradation was determined by a thermogravimetric analyzer (TA instruments, New Castle, USA; Hi-Res TGA 2950). Viscosity was measured by a Brookfield viscometer (Brookfield Engineering Laboratories Inc., MA, USA; DV-E viscometer with LV-Spindle Set) at 21 °C. 86 A 500 ml three neck round bottom flask equipped with two glass caps and a rubber septum for flowing nitrogen gas was loaded with 3-aminopropylsilane (1.0 mole) and GC (1.0 mole) under a nitrogen atmosphere. The reaction mixture was stirred at room temperature for 36 hours. A 20 ml vial fitted with a cap was loaded with 3-aminopropylsilane (10 mmole) and GC (10 mmole) under a nitrogen atmosphere. A drop of the reaction mixture was placed directly on the ATR diamond window of the FTIR and then covered with a rubber septum under a nitrogen atmosphere. The reaction was monitored by recording the FTIR spectrum every few minutes. The kinetic parameters were evaluated from the peak area of the carbonyl peaks. A 500 ml three-neck round bottom flask equipped with one glass cap, a rubber septum and a vacuum connection was loaded with the hydroxyl terminated silane obtained in the previous step under nitrogen atmosphere. The reaction mixture was stirred at 80 °C and 50 Pa while ethanol was removed from the system. 6.3 Results and Discussion 6.3.1 Ring-opening urethane bond formation of aminosilanes with glycerol carbonate The ring-opening reaction of GC with the amine functional group of the silanes produced monomers containing a urethane (carbamate) linkage and two terminal hydroxyl alkyl groups. The ring-opening reaction to yield the urethane bond was followed by IR spectrometry. An example showing the changes in the FTIR spectra as a function of the reaction time for DESiGC is shown in Figure 6.1. 87 Figure 6. 1 Partial FTIR spectra of the reaction between DESi and GC as a function of time. It is clearly observed that the carbonyl peak in the GC at 1800 cm-1 decreased as the reaction progressed, while a new carbonyl peak related to the urethane formation at 1776 cm-1 is increased. Furthermore, the intensity of the broad peak near 3300 cm-1 also increased due to the formation of the hydroxyl groups. Clearly, the urethane bond formation proceeded as expected by ring-opening reaction of GC to yield the desired hydroxyl terminated silanes. It is possible to determine the kinetics of this reaction by normalizing the spectra and monitoring the consumption of the GC or the formation of the urethane linkage as a function of time. The reaction kinetics for the formations of MESiGC, DESiGC, and DESiGC are shown in Figure 6.2. 88 Figure 6. 2 GC concentration as a function of time and second-order kinetics (insert) of the urethanization reaction with MESiGC (solid triangle), DESiGC (solid square), or TESiGC (solid circle) at 22 °C. The data indicate that the urethane formation reaction follows a second order rate constant independent of the type of the silanes which can be described by equation 6.1: 1 1  k t  [C ] [C 0 ] (Eq 6.1) , where k is the rate constant, t is the reaction time and C is the concentration of the glycerol carbonate. Plots of 1/[C] as a function of time yield linear correlations indicating second order rate reactions as shown in Figure 6.2. The rate constants and the correlation coefficients of the linear regressions are listed in Table 6.1. It should be noted that the reaction kinetics of the cyclic carbonate 4-(3-butenyl)-1,3-dioxolan-2one with hexylamine109 and ethylene carbonate with MESi, DESi, or TESi107 were also found to follow a second order rate reaction although the reported rate constants were 89 higher than we observed here. The difference could be due to the bulkier carbonate or the effect of intramolecular hydrogen bonding in the GC. It should further be noted that the reactions of GC with all three silanes proceeded in the bulk (no solvent) without a catalyst. The yields are higher than 96% (as shown in Table 1) as determined from 1H NMR. Polysiloxy branched polymers k [g·mol-1·sec-1] R2 Monomer Viscosity [Pa∙sec] Polymer Viscosity [Pa∙sec] Yield [ %]a MESiGC 6.3×10-3 0.992 2.59 - 98 DESiGC 4.8×10-3 0.995 2.32 127b 96 TESiGC 5.2×10-3 0.993 1.49 2650c 96 a: Measured by 1H NMR, b: Polymerization after 19.5 h, c: Polymerization after 1.5 h Table 6. 1 Kinetics, viscosity and yield of hydroxyl terminated polysiloxy branched polymers: MESiGC, DESiGC, and TESiGC As shown in Figure 6.3, the ring opening reaction of GC with amines can yield two regioisomers. One group of isomers is the 1,2-hydroxyl glycerol derivatives noted as MESiGC, DESiGC, and TESiGC. The other group is the 1,3-hydroxyl glycerol derivatives noted as MESiGC’, DESiGC’, and TESiGC’. Furthermore, the 1,2- hydroxyl glycerol derivatives MESiGC, DESiGC, and TESiGC have additional optical isomers since the GC used in this paper contained a racemic mixture and the primary substituted glycerol had a chiral center at the secondary carbon of the glycerol. It has been established that usually the ring opening reactions of substituted 1,2-carbonate such as propylene carbonate and GC produce primarily the 1,2-hydroxyl isomers over the 1,3-hydroxyl isomer due to steric hindrance.110 Thus, one would expect that MESiGC, DESiGC, and TESiGC would be the dominant products over MESiGC’, 90 DESiGC’, and TESiGC’. However, due to the large number of these possible isomers, the 1H NMR spectra were too complicated to assign individual peaks to each isomer. Figure 6. 3 Preparation of hydroxyl terminated silanes 6.3.2 Monomer-Polymers Equilibrium by 1H NMR The 1H NMR spectra of MESiGC, DESiGC, and TESiGC (as shown in Figure 6.4) further reveals the presence of free ethanol at 1.05 ppm indicating that some oligomerization had occurred as a result of the reaction between the ethoxy groups and the terminal alkyl hydroxyl groups. We observed a similar monomer-oligomers equilibrium previously in the reaction of aminosilanes with ethylene carbonate.107 The extent of branching depended on the number of alkoxy groups of the aminosilanes (Figure 6.3). Thus, the reaction of GC with MESi that contains a single ethoxy group (B) yields the silane MESiGC with two hydroxyl groups (A). Upon polymerization, this silane yields A2-B1 type hyperbranched polymers. The reaction of GC with DESi that has two ethoxy groups (B) yields the hydroxyl terminated silane DESiGC. This A2-B2 type silane monomer can be polymerized to produce highly branched polymers. Similarly, the reaction of GC with TESi that contains three ethoxy groups (B) yields 91 the A2-B3 hydroxyl terminated silane TESiGC. Polymerization of this silane yields highly branched polymers and gels at relatively low conversion. Figure 6. 4 1H NMR spectra (500 MHz, DMSO-d6) of hydroxyl terminated silanes The extent of the rearrangement reaction and the monomer-oligomers equilibrium can also be determined by 1H NMR from the ratio of the ethoxy groups to free ethanol. This ratio was found to be 0.36, 0.25 and 0.24 for MESiGC, DESiGC and TESiGC, respectively. Thus, under identical conditions and in the absence of any catalysts, the monomer-oligomers equilibrium appears to be directly related to the number of ethoxy groups attached to the silicon atom 6.3.3 Viscosity and Thermal Properties Initially, the viscosity of the hydroxyl alkyl terminated silanes was inversely proportional to the number of ethoxy groups attached to the silicon atom (2.59, 2.32, 92 and 1.49 cPs for MESiGC, DESiGC, and TESiGC, respectively). Undoubtedly, the viscosity here depends on the molecular weight (e.g. the monomer-oligomers distribution), the density of branching and, possibly, hydrogen bonds involving the terminal hydroxyl groups. It is expected that extensive rearrangement reactions will lead to high molecular weights and higher viscosities. However, the same rearrangement reactions will also produce free ethanol which will act as a diluent and will tend to decrease the viscosity. Furthermore, as the density of branching increases, more compact structures will be obtained, which will tend to lower the viscosity. Irrespective of these individual contributions, the overall viscosity of these silanes is relatively low. Thermal degradation of these silanes as shown in Figure 5 indicated that weight loss occurs as soon as the heating started. This initial weight loss was most likely caused by the evaporation of ethanol. Actual thermal decomposition occurs when the temperature was higher than 200 °C. Unlike common silanes and polysiloxanes that degrade to yield SiO2 residue in proportion to the amount of silicon atoms in their structure, the result of the thermal degradation of these silanes depended on their structures. Essentially no residue was observed from the degradation of silane MESiGC due to the formation of volatile product during the degradation. Whereby, significant residue was obtained even at 500 °C in air from DESiGC and TESiGC. The relatively large residues are much more than one would expect from the silicone content in these silanes that degrades to SiO2. 93 Figure 6. 5 TGA of MESiGC (solid line), DESiGC (dotted line), and TESiGC (dashed line) 6.3.5 Polymerization by Rearrangement The hydroxyl alkyl silanes thus obtained can undergo self-rearrangements to yield high molecular weight siloxy polymers simply by an exchange reaction of the terminal hydroxyl groups with the ethoxy groups attached to the silicon atom. Unlike typical hydrolysis – condensation reactions of alkoxy silanes that are initiated by water and require an acid or a base catalyst to yield siloxane polymers, this rearrangement reaction does not require water and takes place even in the absence of a catalyst to yield siloxy polymers. The first few steps of this rearrangement reactions that lead to highly branched polymers are shown in Figure 6.6. 94 Figure 6. 6 Condensation polymerization reaction scheme (DESiGC) The monomer – oligomers equilibrium mentioned previously can be shifted to yield polymers by removing ethanol from the silanes. The polymerization of DESiGC at 80 °C is shown as an example in Figure 6.6. It is apparent that the viscosity increased with the polymerization time as ethanol was removed from the reaction mixture. It was observed that large amounts of ethanol were removed from the reaction mixture initially but the amount of ethanol decreased as the reaction progressed. The viscosity of the reaction mixture continuously increased from 2.32 Pa·sec to 127.4 Pa·sec as listed in Table 1. These observations are typical to condensation polymerization reactions where the initial reaction produces a large number of oligomers with minor changes in the viscosity followed by a significant increase in the viscosity as these oligomers are 95 polymerized to high molecular weight polymers. Due to the A2-B2 structure of DESiGC, the polymer is expected to be a highly branched polymer that will gel at low degree of polymerization. Figure 6. 7 Viscosity of DESiGC during polymerization at 80°C and 50Pa The Glass transition temperature (Tg) increased with polymerization and was found to be directly proportional to the degree of polymerization (as shown in Figure 7.8). The hyperbranched polymers obtained from MESiGC were observed to have the highest Tg in this series and their Tg increased linearly shortly after the beginning of the polymerization as a function of the polymerization. The Tg of the branched polymers prepared from DESiGC increased significantly during the first 4 hours of polymerization followed by a more moderate change as the polymerization progressed further. The Tg of the branched polymers prepared from TESiGC was not observed in the first few hours of polymerization and then remained at lower temperatures 96 compared with MESiGC and DESiGC. These changes in the glass transition temperature are related to the structure of the growing polymers where the free chainlength is inversely proportional to the extent of branching. Figure 6. 8 Glass transition temperature as a function of polymerization time at 80°C and 50Pa 97 Figure 6. 9 GPC charts of MESiGC (left) and right DESiGC (right) during polymerization at 80°C and 50Pa The hyperbranched polymer samples were further characterized with GPC. It should be noted that TESiGC does not fully dissolve in tetrahydrofuran and we were not able to characterize the samples with this method. As shown in Figure 6.9, the monomer – oligmer equilibrium could be observed by the multiple response peaks in the GPC spectra. Moreover, the retention time of the response peaks decreased as we increased the reaction time which indicated the proceeding of polymerization reaction. Different from linear polymers, we were not able to provide a quantitative molecular weight compare to polystyrene standard because the retention time cannot directly reflect the true molecular weight of highly branched polymers. 98 6.4 Conclusions In this study, biobased glycerol carbonate was used to prepare a series of hydroxyl terminated silanes and highly branched siloxy polymers in high yields (>96%), all reactions were run in the bulk in line with the green chemistry principles. The ring opening reaction kinetics was studied using FTIR. The polymerization of hydroxyl terminated silanes to highly branched siloxy polymers was conducted under vacuum at 80℃. Viscosity as well as thermal properties of the polymer were reported. 99 Chapter 7. Hydrolysis and Condensation of Water Soluble Alkoxysilanes Under Acidic Conditions 7.1 Introduction Silicone based hyperbranched polymers are characterized by a densely branched backbone and a large number of reactive end-groups. Like other hyperbranched polymers, the end-group functionality is directly proportional to the molecular weight and their rheological properties do not follow the well defined behavior of linear polymers. In our previous work111 we described the synthesis of ABn siloxy-based hyperbranched polymers having alkoxy terminal groups that were prepared by reacting aminosilanes with cyclocarbonates (Scheme 1). The hyperbranched structure was obtained by rearranging the terminal hydroxyl group with an alkoxide group. Alkoxysilanes have been widely used as coupling agents, polymer additives, resin precursors, reinforcement and crosslinking agents. Thus, the hydrolysis of alkoxy silanes under various conditions and the subsequent condensation of the silanols were extensively studied 112-124. These reactions are generally written as: Si(OR)n + nH2O Si(OR)4-n(OH)n + nROH Hydrolysis Si-OR + HO-Si Si-O-Si + ROH Condensation + alcohol Si-OH + HO-Si Si-O-Si + H2O Condensation + water However, it is well known that the actual sequence of these reactions is much more complex as all these reactions are reversible and include many different intermediates 100 having different reaction rates. A comprehensive review125 describes the effects of the structure of the silane and the reaction conditions (e.g. the number and the size of the alkoxy groups, water concentration, pH, type of catalyst, exchange reactions with the solvent, chemical additives and pressure) on these hydrolysis/condensation reactions. Since many of the condensation products of simple alkoxysilanes were immiscible in the reaction mixture or formed gels shortly after the reaction started, various solvents were used to keep the reactants in solution. Indeed, it has been shown that the hydrolysis rate varied greatly when different solvents were used and this variation was attributed mainly to the polarity of the solvent. Alcohols were the most common solvents in many of these studies and they greatly extended the miscibility of the reaction mixture. However, since the alcohol is also a byproduct of the hydrolysis reactions, it impacted the equilibrium and shifted it toward the reverse alcoholysis reactions distorting the kinetics of the system. Unlike previous reports in the literature, the silane we studied contains a hydroxyl terminated substituent attached to the silicon atom. Furthermore, when it was subjected to water and catalyst, all the hydrolysis and condensation products were soluble in water. Thus, one can study the kinetics of these reactions without the complications resulting from phase separation. However, the hydroxyl group can undergo a rearrangement reaction with an ethoxy group to yield a new siloxy bond. Alternatively, the ethoxy groups can undergo conventional hydrolysis/condensation reactions to yield siloxanes bonds. Both rearrangement and hydrolysis/condensation reactions are possible and depend on the reaction conditions. Both of these reaction schemes will lead to high molecular weight polymers. However, 101 the rearrangement reaction will lead to siloxy AB2 hyperbranched polymers whereas the hydrolysis/condensation reaction pathway will produce linear polysiloxanes. The course of the reactionis determined by the presence of water. Thus, if no water is added, rearrangement and hyperbranched polymers are formed [1] whereas linear polysiloxanes are obtained if water is added. In this study we have investigated the course of the hydrolysis/condensation reactions as a function of the water content and the acid concentration using 29SiNMR as the main tool to follow the reaction products and intermediates. The main advantage of using this technique is the fact that the proton decoupled 29Si NMR spectrum provides direct structural information of all the intermediates as it is sensitive to the environment of the silicon atom and it is not affected by other impurities126. Although the relaxation time of the silicon atom is inherently slow, the addition of chromium (III) cetylacetonate as relaxation agent127 greatly decreased this relaxation time. Using this relaxation agent has allowed us to quantitatively follow the hydrolysis and condensation reactions as well as the concentration of all intermediates. 7.2 Experimental Section 3-aminopropyl diethoxymethylsilane was purchased from GELEST, Inc. (PA, USA). Ethylene carbonate anhydrous (99 %) and Chromium (III) acetylacetonate (97%) were purchased from Sigma-Aldrich Co. (MO, USA). Acetone was purchased from J.T. Baker (USA). All material were used as obtained unless specified otherwise. 29 Si NMR spectra were recorded on a 500 MHz NMR spectrometer (Varian Inc., USA, Unity Plus 500MHz) using TMS as the internal standard. The spectra were 102 acquired with a pulse width of 25 s (corresponding to a 90o tip angle) and a pulse delay of 8 s. We used d1=8 seconds to ensure that all the silicon atoms were relaxed (although the relaxation time with this additive was only 1 s) so that the intensity of all the Si signals are quantitative. Ethylene carbonate (1 mol) and 3-aminopropyl methyl diethoxysilane (1 mol) were added to a 500 ml round bottom flask equipped with a rubber septum. The reaction mixture was stirred at room temperature under nitrogen atmosphere for 36 h. Under these mild conditions the carbonate ring reacted with the amine to yield diethoxymethyl-(propyl carbamic acid 2-hydroxyethyl ester) silane (hydroxy-DEMC silane). The preparation of this hydroxyl terminated urethane derivative silane is shown in Figure 7.1. The hydrolysis/condensation reactions were measured as a function of acid and water content at different times by 29Si NMR. In a typical procedure, a 5 mm NMR tube was filled with 0.35ml silane and 0.35ml acetone to yield 1mol/L solution. Chromium (III) acetylacetonate (1%) (paramagnetic relaxation agent) was then added as a relaxation agent and the reaction mixture was mixed briefly until it became homogenous. Different amounts of acid (HCl) and water were then added to each sample to initiate the hydrolysis and condensation reactions (Table 7.1). The tube containing the sample to be studied was then quickly inserted into the NMR and spectra were recorded at regular time intervals. In all cases, 256 scans were taken over a period of 30 minutes and d1 was set to be 8 seconds to ensure quantitative results. 103 Sample name [H+](mol/L) Water (mol/L) Sample name [H+](mol/L) Water (mol/L) WA a 2.78 2A 0.0081 2.78 WB a 5.56 2B 0.016 5.56 WC a 11.11 2C 0.033 11.11 1A 0.0040 2.78 3A 0.016 2.78 1B 0.0081 5.56 3B 0.033 5.56 1C 0.016 11.11 3C 0.065 11.11 Table 7. 1 Sample composition for hydrolysis and condensation reactions The integration of each NMR peak was used to define the relative concentration of each silicone species under different experimental conditions. The relative concentrations were than normalized such the total concentration was between 0 and 1. Thus, if no hydrolysis occurred and only the hydroxy-DEMC silane was observed, its concentration was recorded as 1. Conversely, if it had undergone complete hydrolysis its concentration was recorded as 0. When multiple species (e.g. silanols and siloxanes) were present, the sum of all the peak integrals was normalized to 1 and the relative concentration of each species was then between 0 and 1. Since the condensation reactions produced multiple siloxanes (disiloxane, trisiloxane, etc.) we grouped all these species into a single group to represent the condensation. We realize that the rate of condensation is not necessarily the same for each of these species. However, we were not trying to determine the rate constant of these individual reactions. Instead, our focus was on the overall reaction products which depend on the particular combination of water and acid concentrations. 104 7.3 Results and Discussion 7.3.1 Synthesis and Characterization of AB2 monomer The reaction of 3-aminopropylmethyldiethoxysilane with ethylene carbonate yields hydroxy-DEMC silane – a silane having a hydroxyl terminated urethane side chain as shown in Figure 7.1. This reaction proceeds to completion under nitrogen atmosphere even at room temperature and leads to a water soluble silane. Figure 7. 1 Reaction scheme of 3-aminopropylmethyldiethoxysilane with ethylene carbonate It is apparent from the 1H NMR that even under these mild conditions all the amines reacted with the carbonates to yield the urethane linkages (Figure 7.2). Unlike the few other water soluble silanes, the water solubility here does not depend on salt formation and it is not limited to a certain pH. Consequently, this silane provides a good opportunity to study the hydrolysis and condensation reactions without complications due to phase separation. 105 Figure 7. 2 1H NMR spectrum of the hydroxyl terminated silane AB2 monomer Furthermore, the hydroxy-DEMC silane is a typical AB2 type monomer (one hydroxyl group and two ethoxide groups) which upon rearrangement of the terminal hydroxyl groups with the ethoxy group can yield hyperbranched polymers containing carbo-siloxy linkages as shown in Figure 7.3. This pathway was thoroughly investigated before and was shown to proceed by simple heating under anhydrous conditions to remove ethanol. The polymerization reaction does not involve hydrolysis to silanols and the extent of this hyperbranch polymerization reaction is directly proportional to the amount of ethanol that is removed by distillation. 106 Figure 7. 3 Path A – polymerization to hyperbranched polycarbosiloxy (Si-O-C) backbone polymers Alternatively, the hydroxy-DEMC silane can also be hydrolyzed at RT and condensed by acid or base catalysis to yield linear polysiloxanes containing hydroxyl side chains as shown in Figure 7.4. Figure 7. 4 Path B – polymerization to linear polysiloxane (Si-O-Si) backbone polymers 107 Close examination of our hydroxy-DEMC silane by 1H NMR (Figure 7.2) indicates that even under the mild conditions used to prepare this monomer, some hyperbranching had occurred and free ethanol was present (e.g. m’ and k’ in Figure 7.2). It is also apparent from the 29Si NMR (Figure 7.5) that the spectrum consists of three characteristic peaks, D11, D12 and D13 related to the degree of rearrangement and the different chemical environment of the silicon atom. Figure 7. 5 29Si NMR of the hydroxyl terminated silane AB2 monomer The D11 peak arises from a complete rearrangement of both the ethoxy groups with the hydroxyl terminated side chains such that the silicon atom has no ethoxy group substituents. D12 is the peak of a silicon atom having only one ethoxide as a result of rearrangement with one hydroxyl terminated side chain. Peak D13 is due to a silicon atom containing both ethoxy groups (no rearrangement) where no hydrolysis or rearrangement occurred as in the original hydroxy-DEMC silane. The integration ratio of these peaks D11:D12;D13 is 10.5:42:47.5 indicating that some rearrangement is 108 unavoidable even in the absence of heat, catalyst or water. Thus, this AB2 monomer is actually composed of about 2 repeat units. The fact that the D11 peak is much smaller than the D12 peak is most likely due to the fact that the terminal hydroxyl group now attached to the silicon atom is much larger than ethanol and thus, hinders further reaction. 7.3.2 Hydrolysis/Condensation under very mild acidic conditions It is well known that the hydrolysis of alkoxysilanes is greatly inhibited when high purity, low ionic conductivity water is used in non-glass containers. Under these conditions, alkoxysilanes bearing no autocatalytic functionality were found to be stable for weeks or months. In contrast, if the water is not highly pure the hydrolysis of alkoxysilanea is “substantial” in a very short period of time. We observed that the addition of relatively small amounts of slightly acidic distilled water ([H+]=10-6 mol/L) only led to partial hydrolysis with no condensation to form any siloxanes linkages (Figure 7.6). 109 Figure 7. 6 Hydrolysis under close to neutral conditions ([H+]=10−6 mol/L) at different times Thus, when 5.56 mol/L water was introduced, clear solution was obtained and peaks corresponding to the interactions of water with the silane and the silanols were observed in the NMR spectrum. We tentatively assigned the intermediate D12’ and D13’ to the corresponding hydrated intermediates and peaks D12” and D13” to the corresponding silanols (Figure 7.7). Figure 7. 7 Intermediate formed by the reaction between water and the hyperbranched polymer monomer It should be noted that under these conditions, an apparent stable equilibrium was established and no changes were observed in the relative concentrations of these 110 intermediates and products over the full time period of the experiment. Furthermore, no apparent hydrolysis of bulky alkoxy side groups was observed as is apparent from the sharp D11 peak that remained unchanged (e.g. no hydrolysis of D11 and the lack of the corresponding D11’ and D11” peaks). This difference in the rate of hydrolysis is in agreement with previous results which concluded that the hydrolysis is greatly affected by the size of the alkoxy groups. In our case, the hydrolysis of the bulky alkoxy urethane side chain was not even observed in the time frame of our experiments compared with the much faster hydrolysis rates of the ethoxy groups, especially under this very low acid concentration. It is also important to note here that no condensation was observed. Apparently, the bulky side chains and the high solubility of this silane in the water were sufficient to stabilize the intermediates in solution and prevent further condensation. 7.3.3 Hydrolysis/Condensation under acidic conditions At higher acid concentrations the formation of a silanol was clearly observed in the 29Si NMR spectra (Figure 7.8). Initially, only intermediates D11’, D11” D13’ and D13” were formed (Figure 7.8a) and no silane-diol was obtained even after 30 min reaction time. However, as the concentration of the acid was increased to 0.08 mol/L (Figure 7.8b) a small amount of silane-diol could be observed. The concentration of this silanol remained very small and did not increase with time under these conditions. As the concentration of the acid was further increased (to 0.16mol/L), a significant amount of silane-diol was observed (Figure 7.8c). The NMR peak of the silane-diol appears as a very sharp singlet indicating that it is relatively stable under these acidic conditions, hence, it remained in the aqueous solution with minimal condensation. 111 Figure 7. 8 Effect of [H+] concentration on the hydrolysis reaction. (a: approximately 10−6 mol/L, b: 0.081 mol/L, c:0.16 mol/L) However, when the acid concentration was higher than 0.016mol/L, the silane-diol was no longer observed as it condensed to yield higher molecular weight siloxanes as shown in Figure 7.9. Here, the progress of the condensation reactions was monitored 112 by 29Si NMR with an initial delay of 220 seconds followed by a series of spectra taken every 193 seconds. Under these conditions all the ethoxy groups hydrolyzed to silanols and the concentration of the silane-diol (P2) at -4.5 ppm continuously decreased with time. Simultaneously, some condensation occurred and the concentration of the disiloxane (P3) at -13ppm and trisiloxane (P4) at -21ppm increased with time. It should be emphasized that no phase separation or gelation was observed and the linear siloxanes condensation species remained soluble in the aqueous solution. Figure 7. 9 Silanol condensation as monitored by 29Si NMR 7.3.4 Interdependency of water and acid concentrations on the hydrolysis/ condensation reactions Since the hydrolysis and condensation reactions are affected by both the acid and the water content, the effect of these variables can be better observed using a 3D plot where both variables are monitored simultaneously. Figure 7.10 is an example of such 113 a plot where the concentration of the silane monomer is plotted as a function of acid and water concentrations. Figure 7. 10 Condensation conversion as a function of the reaction time Here, the silane concentration was derived from the NMR peaks integration which were normalized between 1 (initial concentration) and 0 (no silane in the reaction mixture). It is apparent that the effect of acid concentration is much more pronounced than the effect of the water content under the current experimental conditions. Thus, complete hydrolysis was attained and no traces of the alkoxysilane were observed beyond an acid concentration of 0.3 mol/L at 2.5 mol/L of water or beyond an acid concentration of 0.4 mol/L as the water content increased to 12.5 mol/L. Furthermore, the extent of hydrolysis is somewhat lower at high water content than that observed at 114 low water content. This difference is most likely simply related to the fact that the pH of the system (which determines the rate of hydrolysis) is lower as more water is added to the reaction mixture while maintaining the acid concentration constant. It should also be noted that this 3D plot (as well as the following 3D plots) were generated by a polynomial equation. The goodness of fit of these polynomial equations can best be determined by plotting the predicted values against the experimental values as shown in Figure 7.11. A perfect fit would have all the predicted values equal to the experimental values resulting in a linear line along the diagonal. It is apparent (Figure 7.11A) that the polynomial equation here adequately describes the experimental data. 115 Predicted R2=0.9818 B Actual R2=0.9895 C Figure 7. 11 Fit of the polynomials describing the concentration of alkoxysilane (a), silane-diol (b), siloxanes (c) 116 The hydrolysis reaction can be followed by the appearance of the silane-diol (P2) whereby the condensation reactions can be followed by its disappearance and the formation of new siloxanes species. The rate of disappearance of the silane-diol (P2) as a function of the water content is shown in Figure 7.12 for 3 different acid concentrations. By comparing Figure 7.12A to 7.12B and 7.12C, it is apparent that the rate of condensation is highly dependent and directly proportional to the acid concentration at any given time and/or water content. Furthermore, the rate of condensation is significantly slower as the water content is increased. Thus, under low acid concentration (Figure 7.12A) essentially all the silane-diol condensed when only 3 mol/L water was used but about 60% of it remained unchanged when 10 mol/L was used. Similarly, after a short period of time (3 min) about 45% of the silane had condensed when 3 mol/l water was present but the rate of condensation was much slower when 10 mol/L water was used such that at this region (short time, low acid and high water content) almost no condensation was observed. At higher acid concentration (Figure 7.12B) more condensation took place and no traces of the silane-diol were found provided the water content was less than 5 mol/L after about 10 minutes. At even higher acid concentration (Figure 7.12C) most of the silane-diol had condensed and only traces of it could be found at the high water content region and short time. 117 Figure 7. 12 Concentration of silane-diol (P2) as a function of time and water content. (a): acid=0.02 mol/L. (b): acid=0.03 mol/L. (c): acid=0.04 mol/L 118 Further details highlighting the simultaneous hydrolysis and condensation reaction can be seen by plotting the normalized concentration of the silane-diol (P2) as a function of time and acid concentration (Figure 7.13A-C). It is apparent that at low acid concentration and low water (Figure 7.13A) the silane-diol was not observed within the 30 minutes time frame of the experiment. Under these conditions no hydrolysis took place and the original hydroxy-DEMC silane remained unchanged. Upon increasing the acid concentration, hydrolysis took place and the concentration of this intermediate first increased to a maximum (around 0.015 mol/L) and then decreased. Further increasing the acid concentration beyond the maximum led to a decrease in the amount of this silane-diol as it condensed to higher molecular weight siloxanes. Finally, above an acid concentration of 0.03 mol/L, no silanols could be observed and only siloxanes were observed. A similar maximum was observed as a function of the reaction time of various alkoxysilanes128 where a partial hydrolysis took place simultaneously with the condensation reaction. Although different silanes were used, the NMR data indicated a pronounced maximum of silanol-rich intermediates. Interestingly, these hydrolysis/condensation reactions are fast and the overall shape of the surface contours describing the concentration of the silane-diol along the time axis does not change very much other than the fact that the maximum at short reaction time is more pronounced (e.g. higher concentration of silane-diol) that at long reaction time. Similarly, increasing the water content led to a lower degree of condensation and higher concentrations of the silane-diol (Figures 7.13A-C). It should 119 be mentioned here that an excellent fit of the polynomial equation used to generate the contour plots to the experimental data was obtained as seen in Figure 7.11B. 120 Figure 7. 13 Concentration of Silane-diol (P2) as a function of time and acid concentration. (a): water=3 mol/L. (b): water=6 mol/L. (c): water=10 mol/L 121 The condensation step is known to be a complicated process and involves a cascade of different reactions. Under acid conditions, condensation proceeds between different species (e.g. silanols, alkoxsides, and protonated silanes) as well as reactants having different molecular weight. Furthermore, hydrolysis and condensation reactions occur simultaneously and are reversible. Since we were not interested in specific intermediates or products, we grouped any di and tri-siloxanes (P3 and P4, respectively.) species that were observed in the 29Si NMR to represent the “product of condensation”. Here again the total concentration from the NMR integration was normalized to 1 and was then plotted as a function of time and water concentration for different amounts of acid (Figure 7.14A-C). 122 Figure 7. 14 Concentration of the siloxanes as a function of time and water concentration. (a): acid=0.010 mol/L. (b): acid=0.025 mol/L. (c): acid=0.10 mol/L 123 At low acid concentration (Figure 7.14A) only some condensation was observed and the total amount of siloxanes was low. No siloxanes were observed throughout the timeframe of the experiment if the water concentration was above 5 moll/L. Further, maximum siloxanes concentration was observed at longest reaction time (30 min) and under minimum water content (2.6 mol/L). The inverse relationship with respect to water is related to the previous observation that under these low acid conditions water stabilized the silanols, which then retarded the condensation reactions. Increasing the acid concentration (Figure 7.14B) had a significant effect on the condensation reactions; at low water content and long reaction time only siloxanes were observed whereby at high water content (12.5 mol/L) and short reaction time (3 min), no siloxanes were present. Either decreasing the water concentration or increasing the reaction time led to a higher concentration of siloxanes although the water appeared to have a more pronounced effect. As the acid concentration was increased further (Figure 7.14C) most of the products in the reaction mixture were siloxanes. However, only partial condensation was obtained at the highest water content and shortest reaction time. Here again an excellent fit of the polynomial equation to the experimental data was observed (Figure 7.11C). Our results do not follow the simplistic equation suggested to describe the role of water as dictating the degree of condensation in the hydrolysis of alkoxysilane129. Our data indicate that the role of water cannot be separated from the role of acid, and even then, the degree of condensation under any combination of these variables depends on 124 the reaction time. The difference is most likely due to the unique nature of our hydroxyDEMC silane which, under certain conditions, hydrolizes to a stable silane-diol. 7.4 Conclusions Both water and acid play an important role in the hydrolysis/condensation reactions. We have used 29 Si NMR to follow the silanol intermediate and siloxanes products as a function of water and acid concentrations at different reaction times. This study employed a hydroxy-DEMC silane, which contained a terminal hydroxyl group side chain. Consequently, this silane, the hydrolysis intermediates and the condensation products were all soluble in the aqueous solutions under all experimental conditions. The complex relationships of water and acid were studied using 3D contour plots where the effect of any particular combination of these independent variables could be observed. It was found that at low acid and high dilution complete hydrolysis occurred to yield a silane-diol, which remained stable with no apparent condensation. Upon increasing the acid concentration or decreasing the water content enhanced the condensation reactions. The 3D contour plots provided information on the extent of hydrolysis and condensation for any combination of time, acid and water content. The identification of conditions that lead to complete hydrolysis with no condensation allows one to maximize the availability of silanol-rich species which has an important impact on surface modifications. 125 BIBLIOGRAPHY 126 BIBLIOGRAPHY (1) Administration, U. S. E. I.; Administration, U. S. E. I., Ed. Annual Energy Review, 2011. (2) Narayan, R. Michigan State University 2005, 2010. (3) Root, T. L.; Price, J. T.; Hall, K. R.; Schneider, S. H.; Rosenzweig, C.; Pounds, J. A. Nature 2003, 421, 57. (4) (5) 2010, 4, 25. (6) Barnes, D.; Milner, P. Marine Biology 2005, 146, 815. Shen, L.; Worrell, E.; Patel, M. Biofuels, Bioproducts and Biorefining Gross, R. A.; Kalra, B. Science 2002, 297, 803. (7) Yamamoto, M.; Witt, U.; Skupin, G.; Beimborn, D.; Müller, R. J. Biopolymers Online 2005. (8) Ovitt, T. M.; Coates, G. W. J Am Chem Soc 1999, 121, 4072. (9) Nomura, N.; Ishii, R.; Akakura, M.; Aoi, K. J Am Chem Soc 2002, 124, (10) Albertsson, A.-C.; Varma, I. K. Biomacromolecules 2003, 4, 1466. (11) Stange, J.; Uhl, C.; Munstedt, H. Journal of Rheology 2005, 49, 1059. (12) Polance, R.; Jayaraman, K. Polymer Engineering & Science 1995, 35, (13) Li, B.-H.; Yang, M.-C. Polymers for Advanced Technologies 2006, 17, (14) Chen, B.-K.; Shen, C.-H.; Chen, S.-C.; Chen, A. F. Polymer 2010, 51, (15) Pilla, S.; Kim, S. G.; Auer, G. K.; Gong, S.; Park, C. B. Polymer 5938. 1535. 439. 4667. Engineering & Science 2009, 49, 1653. 127 (16) Pilla, S.; Kramschuster, A.; Yang, L.; Lee, J.; Gong, S.; Turng, L.-S. Materials Science and Engineering: C 2009, 29, 1258. (17) Li, H.; Huneault, M. A. Journal of Applied Polymer Science 2011, 122, (18) Inata, H.; Matsumura, S. Journal of Applied Polymer Science 1985, 30, 134. 3325. (19) Cardi, N.; Po, R.; Giannotta, G.; Occhiello, E.; Garbassi, F.; Messina, G. Journal of Applied Polymer Science 1993, 50, 1501. (20) Lai, M.; Kim, J.-K. Polymer 2005, 46, 4722. (21) Najafi, N.; Heuzey, M. C.; Carreau, P. J. Composites Science and Technology 2012, 72, 608. (22) Corre, Y.-M.; Maazouz, A.; Duchet, J.; Reignier, J. The Journal of Supercritical Fluids 2011, 58, 177. (23) Di Maio, L.; Coccorullo, Macromolecular Symposia 2005, 228, 185. (24) 50, 613. (25) I.; Montesano, S.; Incarnato, L. Corre, Y. M.; Duchet, J.; Reignier, J.; Maazouz, A. Rheol Acta 2011, Sentmanat, M. Rheol Acta 2004, 43, 657. (26) Villalobos, M.; Awojulu, A.; Greeley, T.; Turco, G.; Deeter, G. Energy 2006, 31, 3227. (27) Larson, R. G. Macromolecules 2001, 34, 4556. (28) Micic, P.; Bhattacharya, S. N.; Field, G. Int Polym Proc 1998, 13, 50. (29) Dealy, J. M. Rheol Acta 1990, 29, 519. (30) Dealy, J. M. J Rheol 1990, 34, 1133. (31) Liu, H. Z.; Zhang, J. W. J Polym Sci Pol Phys 2011, 49, 1051. (32) 59, 1281. Hiljanen-Vainio, M.; Karjalainen, T.; Seppälä, J. J Appl Polym Sci 1996, 128 (33) Vannaladsaysy, V.; Todo, M.; Jaafar, M.; Ahmad, Z.; Pasomsouk, K. Polymer Engineering & Science 2010, 50, 1485. (34) Signori, F.; Coltelli, M.-B.; Bronco, S. Polymer Degradation and Stability 2009, 94, 74. (35) Ferry, J. D. Viscoelastic properties of polymers; Wiley New York, 1980; Vol. 3. (36) Murariu, M.; Da Silva Ferreira, A.; Alexandre, M.; Dubois, P. Polymers for Advanced Technologies 2008, 19, 636. (37) Ljungberg, N.; Wesslén, B. J Appl Polym Sci 2002, 86, 1227. (38) Labrecque, L. V.; Kumar, R. A.; Davé, V.; Gross, R. A.; McCarthy, S. P. J Appl Polym Sci 1997, 66, 1507. (39) Castle, L.; Mercer, A. J.; Gilbert, J. Food Additives & Contaminants 1988, 5, 277. (40) Haveren, J. v.; Scott, E. L.; Sanders, J. Biofuels, Bioproducts and Biorefining 2008, 2, 41. (41) Younes, H.; Cohn, D. European Polymer Journal 1988, 24, 765. (42) Sheth, M.; Kumar, R. A.; Davé, V.; Gross, R. A.; McCarthy, S. P. J Appl Polym Sci 1997, 66, 1495. (43) Baiardo, M.; Frisoni, G.; Scandola, M.; Rimelen, M.; Lips, D.; Ruffieux, K.; Wintermantel, E. J Appl Polym Sci 2003, 90, 1731. (44) Pillin, I.; Montrelay, N.; Grohens, Y. Polymer 2006, 47, 4676. (45) Gallardo, A.; San Roman, J.; Dijkstra, P. J.; Feijen, J. Macromolecules 1998, 31, 7187. (46) Auras, R. A.; Harte, B.; Selke, S.; Hernandez, R. Journal of plastic film and sheeting 2003, 19, 123. (47) Gruber, P.; O'Brien, M. Biopolymers Online 2002. (48) Narayan, R.; Wu, W.-m.; Criddle, C. S.; US Patent 20,130,023,674: 2013. (49) Carlson, D.; Dubois, P.; Nie, L.; Narayan, R. Polymer Engineering & Science 1998, 38, 311. 129 (50) Conn, R.; Kolstad, J.; Borzelleca, J.; Dixler, D.; Filer, L.; LaDu, B.; Pariza, M. Food and chemical toxicology 1995, 33, 273. (51) CARDELLO, A. V.; MALLER, O. Journal of Food Science 1982, 47, (52) 19, 123. Auras, R. A.; Harte, B.; Selke, S.; Hernandez, R. J Plast Film Sheet 2003, 1549. (53) Kolstad, J. J. J Appl Polym Sci 1996, 62, 1079. (54) Fowlks, A. C.; Narayan, R. J Appl Polym Sci 2010, 118, 2810. (55) Pluta, M.; Galeski, A.; Alexandre, M.; Paul, M. A.; Dubois, P. J Appl Polym Sci 2002, 86, 1497. (56) Tsuji, H.; Horii, F.; Hyon, S. H.; Ikada, Y. Macromolecules 1991, 24, (57) Tsuji, H.; Tezuka, Y. Biomacromolecules 2004, 5, 1181. (58) Yamane, H.; Sasai, K. Polymer 2003, 44, 2569. (59) Ergoz, E.; Fatou, J.; Mandelkern, L. Macromolecules 1972, 5, 147. (60) Takemori, M. T. Polymer Engineering & Science 1979, 19, 1104. 2719. (61) 40, 6898. (62) (63) 32, 251. Pan, P.; Kai, W.; Zhu, B.; Dong, T.; Inoue, Y. Macromolecules 2007, Witzke, D., Thesis, Michigan State Univ, 1996. Guha, M.; Ali, S. Z.; Bhattacharya, S. Journal of Food Engineering 1997, (64) Nakamura, K.; Watanabe, T.; Katayama, K.; Amano, T. J Appl Polym Sci 1972, 16, 1077. (65) Suryanegara, L.; Nakagaito, A. N.; Yano, H. Composites Science and Technology 2009, 69, 1187. (66) Kulinski, Z.; Piorkowska, E. Polymer 2005, 46, 10290. 130 (67) Harris, A. M.; Lee, E. C. J Appl Polym Sci 2008, 107, 2246. (68) Lunt, J. Polym Degrad Stabil 1998, 59, 145. (69) 1998, 78. Narayan, R. Paradigm for successful utilization of renewable resources (70) Auras, R.; Harte, B.; Selke, S. Macromolecular Bioscience 2004, 4, 835. (71) Kale, G.; Auras, R.; Singh, S. P.; Narayan, R. Polymer Testing 2007, 26, (72) Sinclair, R. G. Journal of Macromolecular Science, Part A 1996, 33, 585. (73) Witzke, D. R.; Narayan, R.; Kolstad, J. J. Macromolecules 1997, 30, 1049. 7075. (74) Nishida, H.; Mori, T.; Hoshihara, S.; Fan, Y.; Shirai, Y.; Endo, T. Polym Degrad Stabil 2003, 81, 515. (75) Fan, Y.; Nishida, H.; Shirai, Y.; Tokiwa, Y.; Endo, T. Polym Degrad Stabil 2004, 86, 197. (76) Fan, Y.; Nishida, H.; Shirai, Y.; Endo, T. Polym Degrad Stabil 2003, 80, (77) Fan, Y.; Nishida, H.; Shirai, Y.; Endo, T. Polym Degrad Stabil 2004, 84, (78) Jamshidi, K.; Hyon, S. H.; Ikada, Y. Polymer 1988, 29, 2229. 503. 143. (79) Tsukegi, T.; Motoyama, T.; Shirai, Y.; Nishida, H.; Endo, T. Polym Degrad Stabil 2007, 92, 552. (80) Motoyama, T.; Tsukegi, T.; Shirai, Y.; Nishida, H.; Endo, T. Polym Degrad Stabil 2007, 92, 1350. (81) Shuklov, I. A.; Jiao, H.; Schulze, J.; Tietz, W.; Kühlein, K.; Börner, A. Tetrahedron Lett 2011, 52, 1027. (82) Avrami, M. J Chem Phys 1939, 7, 1103. (83) Avrami, M. J Chem Phys 1941, 9, 177. 131 (84) Jeziorny, A. Polymer 1978, 19, 1142. (85) Nishida, H.; Mori, T.; Hoshihara, S.; Fan, Y. J.; Shirai, Y.; Endo, T. Polymer Degradation and stability 2003, 81, 515. (86) Fan, Y. J.; Nishida, H.; Hoshihara, S.; Shirai, Y.; Tokiwa, Y.; Endo, T. Polymer Degradation and stability 2003, 79, 547. (87) Fan, Y. J.; Nishida, H.; Shirai, Y.; Endo, T. Polymer Degradation and stability 2003, 80, 503. (88) Kozlowsk.W J Polym Sci Pol Sym 1972, 47. (89) Yu, L.; Liu, H.; Dean, K.; Chen, L. Journal of Polymer Science Part B: Polymer Physics 2008, 46, 2630. (90) Tian, Y.; Li, Y.; Xu, X.; Jin, Z. Carbohydrate Polymers 2011, 84, 1165. (91) Khanna, Y. P.; Taylor, T. J. Polym Eng Sci 1988, 28, 1042. (92) Bystritskaya, E. V.; Karpukhin, O. N.; Kutsenova, A. V. International Journal of Polymer Science 2011, 2011. (93) Bogachev, A.; Korobov, M.; Senyavin, V.; Davydov, V.; Rakhmanina, A. Thermochimica acta 2006, 444, 91. (94) Witzke, D. R. Ph.D., Michigan State University, 1997. (95) Witzke, D. R.; Narayan, R.; Kolstad, J. J. Macromolecules 1997, 30, 7075. (96) Groot, W. J.; Borén, T. The International Journal of Life Cycle Assessment 2010, 15, 970. (97) Vink, E. T.; Rabago, K. R.; Glassner, D. A.; Gruber, P. R. Polymer degradation and stability 2003, 80, 403. (98) Behr, A.; Eilting, J.; Irawadi, K.; Leschinski, J.; Lindner, F. Green Chem 2008, 10, 13. (99) Gu, Y.; Jerome, F. Green Chemistry 2010, 12, 1127. 132 (100) Zhou, C.-H.; Beltramini, J. N.; Fan, Y.-X.; Lu, G. Q. Chemical Society Reviews 2008, 37, 527. (101) Ochoa-Gómez, J. R.; Gómez-Jiménez-Aberasturi, O.; Ramírez-López, C.; Belsué, M. Organic Process Research & Development 2012, 16, 389. (102) Cho, H.-J.; Kwon, H.-M.; Tharun, J.; Park, D.-W. Journal of Industrial and Engineering Chemistry 2010, 16, 679. (103) Rokicki, G.; Rakoczy, P.; Parzuchowski, P.; Sobiecki, M. Green Chemistry 2005, 7, 529. (104) Jikei, M.; Kakimoto, M.-a. Progress in Polymer Science 2001, 26, 1233. (105) van Benthem, R. A. T. M. Progress in Organic Coatings 2000, 40, 203. (106) Mańczyk, K.; Szewczyk, P. Progress in Organic Coatings 2002, 44, 99. (107) Tachibana, Y.; Shi, X.; Graiver, D.; Narayan, R. SILICON 2012, 4, 167. (108) Shi, X.; Graiver, D.; Narayan, R. SILICON 2012, 4, 109. (109) Tomita, H.; Sanda, F.; Endo, T. Journal of Polymer Science Part A: Polymer Chemistry 2001, 39, 162. (110) Rousseau, J.; Rousseau, C.; Lynikaite˙, B.; Šačkus, A.; de Leon, C.; Rollin, P.; Tatibouët, A. Tetrahedron 2009, 65, 8571. (111) Dvornic, P. R.; Owen, M. J. In Silicon-Containing Dendritic Polymers; Springer: 2009, p 1. (112) Plueddemann, WASHINGTON DC: 1984. E. P.; AMERICAN CHEMICAL SOCIETY (113) Baines, J.; Eaborn, C. J Chem Soc 1955, 4023. (114) Mehrotra, R.; Narain, R. Indian J Chem 1967, 5, 444. (115) Sanchez, J.; McCormick, A. The Journal of Physical Chemistry 1992, 96, 8973. (116) McCormick, A.; Sanchez, J. In ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY; AMER CHEMICAL SOC 1155 16TH ST, NW, WASHINGTON, DC 20036: 1990; Vol. 200, p 220. 133 (117) Arkles, B.; Steinmetz, J. R.; Zazyczny, J.; Mehta, P. Journal of Adhesion Science and Technology 1992, 6, 193. (118) Schmidt, H.; Scholze, H.; Kaiser, A. Journal of Non-Crystalline Solids 1984, 63, 1. (119) Hardman-Rhyne, K. A.; Coyle, T. D.; Lewis, E. P.; Spooner, S. MRS Online Proceedings Library 1986, 73, null. (120) Stoffer, J. O.; Oostendorp, D. J. In Proceedings of the Water-borne and Higher-solids Coatings Symposium; University of Southern Mississippi: 1988. (121) Osterholtz, F. D.; Pohl, E. R. Journal of Adhesion Science and Technology 1992, 6, 127. (122) Leyden, D. E.; Atwater, J. B. Journal of Adhesion Science and Technology 1991, 5, 815. (123) Oostendorp, D. J.; Bertrand, G. L.; Stoffer, J. O. Journal of Adhesion Science and Technology 1992, 6, 171. (124) Chambers, R. C.; Jones, W. E.; Haruvy, Y.; Webber, S. E.; Fox, M. A. Chemistry of Materials 1993, 5, 1481. (125) Assink, R. A.; Kay, B. D. Annual Review of Materials Science 1991, 21, 491. (126) Wrackmeyer, B. In Annual Reports on NMR Spectroscopy; Webb, G. A., Ed.; Academic Press: 2006; Vol. Volume 57, p 1. (127) Rankin, S. E.; McCormick, A. V. Magnetic Resonance in Chemistry 1999, 37, S27. (128) Hoebbel, D.; Garzó, G.; Engelhardt, G.; Till, A. Zeitschrift für anorganische und allgemeine Chemie 1979, 450, 5. (129) Konrad, E.; Bächle, O.; Signer, R. Justus Liebigs Annalen der Chemie 1929, 474, 276. 134