ABSTRACT. The use of supercritical carbon dioxide, scco 2, as a transient plasticizer

Transcription

1 ABSTRACT ROYER, Joseph Robert. Supercritical Fluid Assisted Polymer Processing: Plasticization, Swelling and Rheology. (Under the direction of Drs. Saad A. Khan and Joseph M. DeSimone.) The use of supercritical carbon dioxide, scco 2, as a transient plasticizer eliminates the disadvantages associated with many other industrial plasticizers. Because CO 2 is a gas under atmospheric conditions, it can be used as a processing aid and then easily removed from a polymer through evaporation to obtain the original physical properties of the unplasticized polymer matrix. In addition, CO 2 has been shown to be more environmentally friendly in comparison to many of the traditional organic plasticizers. However, the biggest challenge hindering the widespread use of CO 2 as a plasticizer involves a lack of understanding of and data quantifying its effect on polymer swelling and the concomitant reduction in material viscosity. In this work, a three-step approach is used to investigate and quantify the physical phenomena associated with CO 2 -induced plasticization of polymer melts. First, a novel experimental apparatus was designed and constructed to measure equilibrium swelling, swelling kinetics and diffusion of CO 2 into a polymer melt. It was found that diffusion of CO 2 into PDMS exhibited Fickian behavior up until two-thirds of the equilibrium swelling value was obtained. The CO 2 pressure had a negligible effect on the diffusion coefficient; however, the system temperature directly affected the diffusion coefficient. Increased pressure was found to enhance the extent of swelling whereas a maximum was observed with increasing temperature, at pressures above 15 MPa. The Sanchez-Lacombe equation of state was found to be

2 in good agreement with the experimentally calculated variables, and thus, can be used as a predictive tool to obtain physical properties of the CO 2 -PDMS system. Secondly, a high pressure extrusion slit die rheometer was constructed to measure the viscosity of polymer melts plasticized with low concentrations of CO 2. Polystyrene, poly(methyl methacrylate), polypropylene, low density polyethylene, and poly(vinylidene fluoride) were all investigated. CO 2 was found to be an efficient plasticizer for all of these polymer materials, generally lowering the viscosity of the melt 30-80%, depending on processing conditions. Predictive viscoelastic scaling models based on free-volume principles and a prediction of T g depression from a diluent were developed to quantify the effects of CO 2 concentration, pressure and temperature on viscosity. This unique free-volume approach allows the high pressure polymer/co 2 rheology to be predicted based solely on physical parameters of the polymer melt and CO 2. Therefore, only rheological measurements at ambient pressures are required to predict the high pressure polymer/co 2 solution behavior over the concentration and temperature ranges for which the models are valid. Finally, a novel high pressure magnetically levitated sphere rheometer (MLSR) was developed to further investigate the effects of CO 2 on the viscosity of polymer melts. The MLSR measures the difference in magnetic intensity required to levitate a magnetic sphere in a sample fluid while the fluid is at rest and under shear. The observed change in magnetic intensity is directly proportional to the viscoelastic force imposed on the sphere by the surrounding fluid, and thus is used to calculate the fluid viscosity from a calibration of known viscosity standards. The rheometer eliminates many of the disadvantages associated with other high pressure rheometers

3 and can operate over a wide range of CO 2 concentrations at constant pressure with excellent reproducibility. This rheometer was used to measure the viscosity reduction of poly(dimethyl siloxane) by CO 2. The effects of both system pressure and CO 2 were investigated. The viscosity of the polymer melt could be lowered in excess of 97% of its original value at atmospheric pressure by adding a CO 2 concentration of approximately 30 wt%. Additionally, experimental evidence revealed that the elevated pressure significantly increased the polymer/co 2 viscosity.

4

5 PERSONAL BIOGRAPHY Joseph Robert Royer, is a native of Naperville, Illinois born on February 4 th Joseph is a graduate of Naperville North High School and received a Bachelors of Science in Chemical Engineering from the University of Notre Dame du Lac in South Bend, Indiana in While at the University of Notre Dame, he researched catalytic membrane reactors with Dr. Arvid Varma, and studied supercritical water oxidation under the advisement of Dr. Joan F. Brennecke. During his undergraduate career he received an athletic scholarship as a member of both the varsity Cross-Country and Track & Field Teams. He received the Knute Rockne Student Athlete Award in the spring of In the fall of 1995, he entered the Chemical Engineering graduate program at North Carolina State University located in Raleigh, North Carolina. Joseph s research project focused on the construction of rheological devices to determine the viscosity of polymer melts plasticized. He was jointly advised by Dr Saad A. Khan and Dr. Joseph M. DeSimone. In addition to this research, Joseph has also worked on the design of equipment for and measurement of polymer swelling induced by absorption of high pressure CO 2. As a student member of both the Kenan Center for the Utilization of Carbon Dioxide in Manufacturing Technologies and the National Science Foundation Science and Technology Center for Environmentally Responsible Solvents he has collaborated on several other projects. Specifically, he has worked along side, Yvon Gay, Srinivas Siripurapu, and Teri Walker on several projects, which include miscibility of polymer blends and microcellular polymer foaming. ii

6 ACKNOWLEDGMENTS I would like to express my deepest gratitude to both of my academic advisors, Drs. Saad A. Khan and Joseph M. DeSimone. Thank you for encouragement and faith in my visions for this research endeavor, even when the prospects of high pressure rheology were very bleak. I also thank the Kenan Center for Utilization of Carbon Dioxide in Manufacturing, the National Science Foundation Science and Technology Center for Environmentally Responsible Solvents and Processes, and the Office of Naval Research for financial support of my project. Special thanks go to Yvon Gay, who during his stay in Raleigh was a valuable research partner, and M. Adam without whom the MLSR would never have been completed. I would like to give special thanks to my extended research family, especially the rheology lab including Jenny, Bor-Sen, Srini and Akash who helped get me started, and the CO 2 group specifically Flory, Brian, Srini and Erik for their useful discussions and thoughtful listening to my constant complaining. And to all those with whom I have shared the rigors of graduate school, I extend my best wishes. To my friends with whom I spend my social time, thank you for keeping my spirits up in the times when research progress was slow. I especially thank my roommate and frequent golf partner, Matt, with whom I can always spend a Sunday afternoon watching the boys of the Winston Cup Circuit. Last but not least, I must thank my family for their constant support and unconditional love. The Lord for all my many gifts. My parents for their love, plus the use of their house, pool table, and grill while they lived in the UK! And finally, Julie it is you with whom I have shared so much of my life over the past few years. Thank you for always keeping a smile on my face and a joy in my heart, aa. iii

7 TABLE OF CONTENTS LIST OF TABLES ix LIST OF FIGURES xiv CHAPTER I: INTRODUCTION CHAPTER II: NEAR- AND SUPERCRITICAL FLUID INDUCED PLASTICIZATION OF POLYMER MELTS Properties of Supercritical Fluids Supercritical CO Plasticization of Polymer Melts Applications for CO 2 Plasticized Polymer Melts.. 13 Melt Phase Synthesis Enhanced Polymer Blending References CHAPTER III: HIGH PRESSURE RHEOLOGICAL MEASUREMENTS Rheological Concepts Steady Shear Rheology Dynamic Shear Rheology Design of High Pressure Rheometers High Pressure Falling Body Viscometers Pressure Driven Viscometers High Pressure Rotational Rheometers Free Volume Theory and Experimental Studies Definition of Free Volume Doolittle s Equation WLF Theory for Polymer Melts Analogs to WLF Theory for Pressure and Concentration.. 37 iv

8 3.4 Experimental Studies of High Pressure Rheology Pure Fluids Pure Polymer Melts Polymer Solutions References CHAPTER IV: CARBON DIOXIDE-INDUCED SWELLING OF POLY(DIMETHYLSILOXANE) Abstract Introduction Experimental Materials and Methods Materials Swelling Apparatus Swelling Procedure Results and Discussion Swelling Kinetics Diffusion Coefficients of CO 2 into PDMS Equilibrium Swelling Modeling of Equilibrium Swelling Conclusions References CHAPTER V: HIGH PRESSURE RHEOLOGY OF POLYSTYRENE MELTS PLASTICIZED WITH CO 2 : EXPERIMENTAL MEASUREMENTS & PREDICTIVE SCALING RELATIONSHIPS Abstract Introduction Experimental Materials and Methods Materials Rheometer Design Experimental Procedure v

9 5.4 Results and Discussion Experimental Verification of Rheometer Design Viscosity Reduction with CO Free Volume and Plasticization Viscoelastic Scaling and Data Correction Conclusions References CHAPTER VI: HIGH PRESSURE RHEOLOGY OF PLASTICIZED POLYMER MELTS: EXTENSION OF PREDICTIVE VISCOELASTIC SCALING RELATIONSHIIPS Abstract Introduction Viscoelastic Scaling Theory T g Depression WLF Analogs Arrhenius Analogs Experimental Materials and Methods Materials Rheometer and Experimental Procedure Viscoelastic Scaling Procedure Results and Discussion Poly(methyl-methacrylate), (PMMA) Polypropylene, (PP) Low Density Polyethylene, (LDPE) Poly(Vinylidene Fluoride), (PVDF) Conclusions References vi

10 CHAPTER VII: A NOVEL HIGH PRESSURE MAGNETICALLY LEVITATED SHPERE RHEOMETER: DESIGN, FABRICATION AND VISCOSITY MEASUREMENT Abstract Introduction Theory of Magnetic Levitation Sphere Rheometer Sphere Levitation Viscous Forces Newtonian Fluids Elastic Solids Viscoelastic Fluids Rheometer Design and Construction High Pressure Sample Chamber Rheometer Calibration and Experimental Verification Viscosity Measurments Conclusions References CHAPTER VIII: CONCLUSIONS AND FUTURE DIRECTIONS Conclusions Future Directions APPENDIX A: COMPUTER PROGRAMS A.1 Program To Determine Sanchez-Lacombe EOS Characteristic Parameters A.2 Sample Data File for S-L EOS Characteristic Parameters A.3 Program To Determine Phase Equilibrium from Sanchez-Lacombe EOS A.4 Sample Data File for S-L EOS Mixture Density vii

11 A.5 Program To Determine Density of a Mixture at a Specific Concentration A.6 Sample Data File for S-L Density at Specific Concentration APPENDIX B: SWELLING DATA B.1 Swelling of Viscasil-12M B.2 Swelling of Viscasil-100M B.3 Swelling of PS APPENDIX C: RHEOLOGY DATA C.1 Polystyrene Rheology Data C.2 Poly(methyl-methacrylate) Rheology Data C.3 Polypropylene Rheology Data C.4 Low Density Polyethylene Rheology Data C.5 Poly(Vinylidene Fluoride) Rheology Data APPENDIX D: MLSR DATA D.1 PDMS Rheological Data viii

12 LIST OF TABLES Table 4.1 GPC data poly(dimethylsiloxane) samples relative to polystyrene samples Table 4.2 Sanchez-Lacombe pure component characteristic parameters. 76 Table 4.3 Binary interaction parameters obtained by fitting the experimental swelling data to the Sanchez-Lacombe equation of state Table 5.1 Dimensions of slit dies and nozzles used in extrusion rheometer. 95 Table 5.2 Characteristic Sanchez-Lacombe parameters Table 5.3 Material constants used in the Chow-WLF free volume analysis. 112 Table 6.1 Characteristic Sanchez-Lacombe parameters Table 6.2 WLF analog scaling model parameters Table 6.3 Arrhenius analog scaling model parameters Table B.1 Swelling Data for Viscasil-12 M at K Table B.2 Swelling Data for Viscasil-12 M at K Table B.3 Swelling Data for Viscasil-12 M at K Table B.4 Swelling Data for Viscasil-12 M at MPa Table B.5 Swelling Data for Viscasil-12 M at MPa Table B.6 Swelling Data for Viscasil-100 M at K Table B.7 Swelling Data for Viscasil-100 M at K Table B.8 Swelling Data for Viscasil-100 M at K Table B.9 Swelling Data for PS049.5 at K Table B.9 Swelling Data for PS049.5 at K Table C.1 Viscosity of Styron-685D at K without CO ix

13 Table C.2 Viscosity of Styron-685D at K with CO Table C.3 Viscosity of Styron-685D at K without CO Table C.4 Viscosity of Styron-685D at K with CO Table C.5 Viscosity of Styron-685D at K without CO Table C.6 Viscosity of Styron-685D at K with CO Table C.7 Viscosity of Nova-172 at K without CO Table C.8 Viscosity of Nova-172 at K with CO Table C.9 Viscosity of Nova-172 at K without CO Table C.10 Viscosity of Nova-172 at K with CO Table C.11 Viscosity of Nova-172 at K without CO Table C.12 Viscosity of Nova-172 at K with CO Table C.13 Viscosity of Nova-103 at K without CO Table C.14 Viscosity of Nova-103 at K with CO Table C.15 Viscosity of Nova-103 at K without CO Table C.16 Viscosity of Nova-103 at K with CO Table C.17 Viscosity of Nova-103 at K without CO Table C.18 Viscosity of Nova-103 at K with CO Table C.19 Viscosity of PMMA-VO45 at K without CO Table C.20 Viscosity of PMMA-VO45 at K with CO Table C.21 Viscosity of PMMA-VM100 at K without CO Table C.22 Viscosity of PMMA-VM100 at K with CO Table C.23 Viscosity of PP-4036 at K without CO x

14 Table C.24 Viscosity of PP-4036 at K with CO Table C.25 Viscosity of PP-4036 at K without CO Table C.26 Viscosity of PP-4036 at K with CO Table C.27 Viscosity of PP-4036 at K without CO Table C.28 Viscosity of PP-4036 at K without CO Table C.29 Viscosity of PP-4018 at K without CO Table C.30 Viscosity of PP-4018 at K with CO Table C.31 Viscosity of PP-4018 at K without CO Table C.32 Viscosity of PP-4018 at K with CO Table C.33 Viscosity of PP-4018 at K without CO Table C.34 Viscosity of PP-4018 at K with CO Table C.35 Viscosity of PP-4018 at K without CO Table C.36 Viscosity of PP-4018 at K with CO Table C.37 Viscosity of PP-4018 at K without CO Table C.38 Viscosity of LDPE-640I at K without CO Table C.39 Viscosity of LDPE-640I at K without CO Table C.40 Viscosity of LDPE-640I at K with CO Table C.41 Viscosity of LDPE-640I at K without CO Table C.42 Viscosity of LDPE-640I at K with CO Table C.43 Viscosity of LDPE-640I at K without CO Table C.44 Viscosity of LDPE-640I at K without CO Table C.45 Viscosity of Kynar-740 at K without CO Table C.46 Viscosity of Kynar-740 at K with CO xi

15 Table C.47 Viscosity of Kynar-740 at K without CO Table C.48 Viscosity of Kynar-740 at K with CO Table C.49 Viscosity of Kynar-740 at K without CO Table C.50 Viscosity of Kynar-740 at K without CO Table C.51 Viscosity of Kynar K with CO Table C.52 Viscosity of Kynar K without CO Table C.53 Viscosity of Kynar K with CO Table C.54 Viscosity of Kynar K without CO Table C.55 Viscosity of Kynar K without CO Table C.56 Viscosity of Kynar K with CO Table D.1 Viscasil-100M at 30ºC, 6.9 MPa and wt% CO Table D.2 Viscasil-100M at 30ºC, 6.9 MPa and 3.58 wt% CO Table D.3 Viscasil-100M at 30ºC, 6.9 MPa and 7.16 wt% CO Table D.4 Viscasil-100M at 30ºC, 13.8 MPa and wt% CO Table D.5 Viscasil-100M at 30ºC, 13.8 MPa and 9.79 wt% CO Table D.6 Viscasil-100M at 30ºC, 13.8 MPa and wt% CO Table D.7 Viscasil-100M at 30ºC, 13.8 MPa and 17.2 wt% CO Table D.8 Viscasil-100M at 30ºC, 20.7 MPa and wt% CO Table D.9 Viscasil-100M at 30ºC, 20.7 MPa and wt% CO xii

16 Table D.10 Viscasil-100M at 30ºC, 20.7 MPa and wt% CO Table D.11 Viscasil-100M at 30ºC, 20.7 MPa and wt% CO xiii

17 LIST OF FIGURES Figure 2.1 Pressure-temperature diagram for a pure fluid phase change from (a) to (b) by two paths, across the liquid-vapor equilibrium line and through the supercritical region... 5 Figure 2.2 Carbon dioxide phase diagram Figure 2.3 Figure 2.4 Figure 2.5 Solubility parameter, δ, of CO 2 : At -31ºC the fluid undergoes a first-order phase transition, resulting in a step change in δ from liquid to gas. Above the critical point, 31ºC & 70ºC, a continuum of properties exists, allowing previously unstable values to be obtained Density continuum for supercritical CO 2 : At the subcritical conditions densities between 0.2 and 0.8 are not found because of the first-order phase change. In the supercritical state, all values of density are attainable Proposed model of glass-liquid transition for a polymer -CO 2 mixture Figure 2.6 T g of Polystyrene as a function of CO 2 pressure Figure 2.7 Figure 3.1 Figure 3.2 Variation of T* with scco 2 pressure for PS-PI blends of differing compositions. The upward pointing solid triangles represent 0.1wt% PS, while the downward pointing triangles designate 0.2wt%. The solid circles represent experiments with the PS-PI blend exposed to N 2 at 0.2wt% PS Schematic diagram of falling body viscometers. (a) Falling cylinder, (b) Rolling ball Schematic of a back pressure regulated capillary rheometer. The Rheometer piston is used to drive the test fluid through the downstream capillary. The back pressure piston elevates the downstream pressure sufficiently high so as to maintain a one-phase system during viscosity measurement Figure 3.3 Pressurized Couette-Hastchek rotational viscometer xiv

18 Figure 3.4 Figure 3.5 Schematic illustration of the variation of the specific volume, v, of a polymer with temperature, T. The free volume, v F, is represented by the shaded area Viscosity data measured by Gerhardt et al. on PDMS at 50ºC with increasing CO 2 concentrations (0, 4.84, 9.03, 14.4, and 20.7wt% CO 2 ) at different pressures. (a) The raw viscosity data. The plasticization effect can be seen as viscosity values decrease with increasing CO 2 concentration. (b) The data is compressed using a concentration shift factor a C to obtain a single master curve at this temperature Figure 4.1 Schematic diagram of the high pressure swelling apparatus designed for detecting a one-dimensional volume expansion of a polymer melt Figure 4.2 Figure 4.3 Figure 4.4 Optical monitorin of the swelling behavior of the PDMS sample at 20.7Mpa and 70 o C. Images are shown for various times after exposure to CO 2 : (a) Initial time t=0, (b) 3hrs, (c) 6hrs, (d) 20hrs Swelling kinetics for the 95,000 g/mol PDMS sample exposed to CO 2 at 20.7 MPa and 50 o C. Inset figure demonstrates the linear relationship obtained when the initial 60% of the kinetic data are shown as a function of t 1/ Swelling kinetics for the 160,000 g/mol PDMS sample exposed to CO 2 at 50 o C and various pressures Figure 4.5 Swelling behavior as a function of time for the 95,000 g/mol PDMS sample exposed to CO 2 at 20.7 MPa and various temperatures Figure 4.6 Figure 4.7 Figure 4.8 Swelling kinetics for PDMS exposed to CO 2 at 20.7 MPa and 50 o C for various molecular weights. Inset figure displays a close-up view of the initial portion of the kinetic curves to demonstrate more clearly the effect of molecular weight.. 63 Calculated diffusion coefficients for the 95kg/mol PDMS sample exposed to CO 2 at various pressures and temperatures. 67 Calculated diffusion coefficients for three PDMS samples exposed to CO 2 at various pressures and 70 o C. The inset shows the molecular weight dependence of the calculated diffusion coefficients at various pressures xv

19 Figure 4.9 Isothermal data for the equilibrium swelling ratio, ξ, for the 95,000 g/mol PDMS sample exposed to CO 2 at various pressures (a) shows data as a function of CO 2 density whereas (b) reveals the data as a function of system pressure. Lines are drawn to illustrate trends and not theoretical predictions.. 71 Figure 4.10 Isobaric data for the equilibrium swelling ratio, ξ, for two different molecular weight samples exposed to CO 2 at various temperatures and pressures. Data is shown as both as a function of CO 2 density (a) and as a function of system pressure (b). Lines are drawn to illustrate trends and not theoretical predictions Figure 4.11 Isothermal data for the equilibrium swelling ratio, ξ, for three different molecular weight samples exposed to CO 2 at various pressures and 50 o C as a function of CO 2 density Figure 4.12 Theoretical prediction of the isothermal equilibrium swelling ratio, ξ, using the Sanchez-Lacombe equation of state for three different molecular weight samples exposed to CO 2 at various pressures and temperatures (a) 50 o C, (b) 70 o C Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Schematic diagram of high pressure slit die extrusion rheometer with metered CO 2 injection system Schematic drawing of slit die and nozzle design for viscosity measurement Pressure drop along slit die length during viscosity measurement, at various CO 2 concentrations for Styron-685D at 200 C with different nozzle configurations. Binodal curves are calculated using Sanchez-Lacombe equation of state Comparison between parallel plate and slit die measurements for Styron-684D at 200 C Uncorrected viscosity measurements for three different polystyrene melts in the presence of various CO 2 concentrations at 200 C. (a) Styron-685D, numbers represent the average pressure during measurement in MPa, (b) Nova-103, (c) Nova xvi

20 Figure 5.6 Figure 5.7 Figure 5.8 Figure 5.9 Figure 6.1 Uncorrected viscosity measurements for the Styron-685D sample in the presence of various concentrations of CO 2 at (a) 225 C and (b) 250 C Pressure shift factors calculated using the Chow-WLF free volume analysis and used for data analysis and correction as a function of pressure for several temperatures and CO 2 conditions Concentration shift factors calculated using the Chow-WLF free volume analysis. Solid lines represent the predictions of the free volume model. Predictions made at 150 and 175 C respectively show the agreement of the measurements of Kwag et al Master curves generated using Chow-WLF free volume analysis collapsed to the pure melt viscosity at atmospheric pressure and 200 C for the (a) Styron-685D, (b) Nova-103, (c) Nova Outline of steps required to construct a master curve using either of the two viscoelastic scaling analogs Figure 6.2 Apparent viscosity measurements of PMMA at various concentrations of dissolved carbon dioxide at 210ºC. (a) VO45 resin, average pressure in the slit die during measurement in MPa are provided as representative sample. (b) VM100 resin. 133 Figure 6.3 Figure 6.4 Figure 6.5 Figure 6.6 Master curves generated using the WLF scaling analogs. Each curve is collapsed at 210ºC, atmospheric pressure, and zero concentrations. (a) VO45 resin. (b) VM100 resin Apparent viscosity measurements of PP at various concentrations of dissolved carbon dioxide and temperatures. (d) PP-4018 at 180ºC (b) PP-4018 at 190ºC (c) PP-4018 at 200ºC (d) PP-4036 at Master curves generated using the Arrhenius scaling analogs. Each curve is collapsed to a single temperature, atmospheric pressure, and zero concentrations. (a) PP-4018 at 180ºC. (b) PP-4036 at 200ºC Apparent viscosity measurements of LDPE 640I at various concentrations of dissolved carbon dioxide 200ºC xvii

21 Figure 6.7 Figure 6.8 Figure 6.9 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 7.6 LDPE Master curves generated using the Arrhenius scaling analogs, collapsed to a 200ºC, atmospheric pressure, and zero concentrations Apparent viscosity measurements of PVDF at various concentrations of dissolved carbon dioxide. (a) Kynar 740 at 210ºC. (b) Kynar 740 at 225ºC. (c) Kynar 740 at 250ºC. (d) Kynar 460 at 225ºC. (e) Kynar 460 at 250ºC Master curves generated using the Arrhenius scaling analogs. Each curve is collapsed to a single temperature, atmospheric pressure, and zero concentrations. (a) Kynar 740 at 210ºC. (b) Kynar 460 at 225ºC Schematic diagram of the magnetically levitated sphere rheometer Schematic diagram of the high pressure sample chamber for the MLSR Sample set of experimental calibration data for a 300 Poise standard viscosity silicon oil at 25 C and atmospheric pressure. 166 Relative experimental error presented as a deviation from linearity as a function of tube velocity at 25 C and atmospheric pressure. Results have been amplified by plotting this ratio of the change in magnetic current divide by the tube velocity ((I-I o )/v) as a function of tube velocity Calibration curve for 5 silicon viscosity standards determined at 25 C and atmospheric pressure. inset.) Displays the lack of a hysteresis effect as the measurements are taken in both the positive and negative flow directions Comparison of experimental measurements of PDMS taken on a commercially available couette rheometer and with the high pressure MLSR. The data points were all taken on the MLSR at various temperatures ( 30 ºC, 50 ºC, 70 ºC ), while the lines are the measured values on the couette rheometer Figure 7.7 High pressure PMDS/CO 2 viscosity measurements at 30ºC And various CO 2 concentrations. In all figures 0 wt% CO 2 was measured at atmospheric pressure conditions. (a) 6.9 MPa. (b) 13.8 MPa. (c) 20.7 MPa xviii

22 Figure 7.8 Viscosity relative to measurement at 30ºC, 0 wt% CO 2 and atmospheric pressure as a function of system pressure xix

23 Chapter I INTRODUCTION The industrial use of high molecular weight polymeric materials is widespread in today s world. These uses range from coatings applications to molded parts, textiles and many others. Preparation of these useful articles is virtually impossible without the use of auxiliary materials or additives such as solvents, thinners, stabilizers, antioxidants, flame-retardants, anti-microbials, and plasticizers. Specifically, the use of plasticizers in polymer processing has several unique advantages. Plasticizers allow a polymer material to flow more uniformly under less mechanical shear and at lower temperatures. They often lower a polymer s glass transition temperature, T g, and allow the physical properties of the final product to be tailored from rigid to flexible. Unfortunately, the use of plasticizers, as with any polymer additive, can be disadvantageous. Inherently, separation of small molecules from polymer melts is slow and energy intensive, and many plasticizers are organic compounds under regulatory restrictions because they are known to have a deleterious effect on the environment. In this study, we explore the use of supercritical fluids, SCFs, specifically supercritical carbon dioxide, scco 2, as a plasticizing agent. The use of scco 2 as a transient plasticizer eliminates the disadvantages associated with some of the other industrial plasticizers. Because CO 2 is a gas under atmospheric conditions, it can be used as a processing aid and then easily removed from a polymer by vaporization to obtain the original physical properties of the unplasticized polymer matrix. In addition, CO 2 has been shown to be more environmentally friendly in comparison to many of the traditional organic plasticizers. However, significant challenges preventing widespread use of CO 2 as a transient plasticizer are the formation of non-

24 foamed structures upon depressurization and an understanding and measurement of plasticization. As presented in Chapters II and III, a thorough review of the scientific literature reveals that while CO 2 is known to be an effective plasticizer of several high molecular weight polymers, such as polystyrene and poly(methyl methacrylate), very little is know about the mechanism of CO 2 -induced plasticization and the resulting changes in rheology. In this thesis, we focus on measurements of the physical property changes brought about by the incorporation of near- or super-critical CO 2 into a polymer matrix, and examine these results in light of the mechanisms of plasticization. In Chapter IV an experimental device designed to measure the swelling of polymer melts is presented. This device is used to measure CO 2 -induced swelling of selected polymer melts, and provide data for both the modeling of CO 2 solubility into the melts and the estimation of the diffusion coefficients for CO 2 within the polymer. In Chapter V and VI, we present the design of a high pressure extrusion rheometer. This device is used to measure the CO 2 -induced viscosity reduction of polymer melts under high pressure and typical processing conditions. Results of these experiments are used to validate the development of a theoretical model based on freevolume theory to quantitatively predict the rheological behavior of CO 2 plasticized melts from simple measurable polymer properties. In Chapter VII, the thesis describes the design of a novel high pressure magnetically levitated sphere rheometer to determine constant-pressure rheological properties of polymer melts with incorporated CO 2. This novel device complements the previously presented extrusion slit die rheometer to provide a more complete characterization of the CO 2 -induced plasticization process. In particular the magnetically levitated sphere rheometer is used to measure the rheological properties of a plasticized melt under constant-pressure conditions. These results are then compared in detail to previous work to help in gaining an understanding of the links between polymer swelling, plasticization and rheology. 2

25 The topics covered in this thesis, attempt to address some of the critical unresolved issues facing the widespread implementation of CO 2 as a transient plasticizer for polymeric resins in industrial processes. The main unresolved issues that this research endeavor attempts to answers include: (i) What is an appropriate method to measure the high pressure solubility, diffusivity and swelling of a polymer melt exposed to CO 2? (ii) What are some appropriate techniques for measuring the steady state rheological properties of a plasticized polymer melt containing dissolved CO 2 under elevated pressure conditions? (iii) How can free-volume theory be used to predict the rheological behavior of high pressure polymer melts plasticized by CO 2? (iv) Is it possible to directly relate changes in polymer T g caused by incorporation of a high pressure gas diluent to the rheology of the plasticized polymer melt? (v) How can the rheological properties of the plasticized polymer melt be predicted so that minimal measurements are conducted under pressurized conditions? 3

26 Chapter II NEAR- AND SUPERCRITICAL FLUID INDUCED PLASTICIZATION OF POLYMER MELTS Prior to utilizing SCFs in polymer processing, it is essential to review the properties of SCFs and their effects on polymer processing. In this chapter, we present an overview of the properties of SCFs and more specifically scco 2. The effects of SCFs on polymer melts that have previously been investigated are also discussed. 2.1 Properties of Supercritical Fluids A supercritical fluid, SCF, is a substance elevated above both its critical temperature and pressure. 1 The supercritical state is a distinct phase of matter, separate from the liquid and gas phases. SCFs have properties that often lie between those of liquids and gases. To illustrate this point, consider the compression of a gas to a liquid following two different pathways, as shown in Figure 2.1. The first pathway, (1), begins with isothermal compression of the gas, beginning from point (a) as seen in the figure. When the compression reaches the vapor-liquid equilibrium line, two phases coexist, viz. liquid and gas. As a result of the first order phase change and the discontinuity it creates, the physical properties of these two phases are significantly different. The continued compression of the fluid results in a pure liquid phase, point (b). The same compression can proceed through the supercritical region by a separate three-step pathway, (2). In this case, the first step is isobaric heating above the critical 4

27 temperature, followed by isothermal compression above the critical pressure, resulting in a SCF, (c). After isobaric cooling the same final liquid, point (b), is obtained. During the second process, no distinct phase boundaries are crossed, implying no first order phase transition. Thus, the properties of the fluid must vary smoothly, resulting in a continuum of properties between those of the gas and the liquid. The ability of a SCF to achieve this continuum of fluid properties makes them attractive in many applications, because advantage can be taken of properties not found in either the gas or liquid states. (b) Supercritical Fluid (c) Pressure Solid Liquid 1 (a) Critical Point 2 Gas Triple Point Temperature Figure 2.1 Pressure-temperature diagram for a pure fluid phase change from (a) to (b) by two paths, across the liquid-vapor equilibrium line and through the supercritical region. Because of the direct relationship between density and pressure for a SCF, a continuum of densities ranging from vapor-like to liquid-like can be obtained by varying the pressure of the fluid at constant temperature, while maintaining a one phase system. Near infinite compressibilities at the critical point are possible. 5

28 Because minimal pressure changes result in significant density variations, this behavior produces large deviations in the Hildebrand solubility parameter, δ, which can be tailored to creating a situation where chemicals can be pulled into or dropped out of solution by selectively changing system pressure. This tunability with pressure is not limited to density. Other properties like viscosity, dielectric constant and thermal conductivity are also tunable with pressure and temperature. 2 Supercritical CO 2 A phase diagram of CO 2 is presented in Figure By examination of Figure 2.2, the critical point of CO 2 can be found to occur at a temperature of 31.1 C and a pressure of 7.38 MPa (73.8 bar). This critical temperature is relatively low as compared to the critical temperatures of water (374.2 o C), propane (96.7 o C), or methanol (239.5 o C), and makes the supercritical region of CO 2 easily accessible to thermally unstable solutes, including many polymeric materials. 4 The critical pressure of CO 2 is regularly achieved in some common industrial processes, making some applications ideal candidates for SCF adaptation. The unique physical properties of scco 2 have been exploited in a number of both laboratory and industrial scale applications. Some examples include SCF extractions, 5 homogeneous and heterogeneous polymerizations, 6-8 enhanced oil recovery, 9 and as a solvent medium for fluoropolymer coatings. 10 In addition, CO 2 is non-corrosive, inexpensive and environmentally benign. Other common SCFs, such as sch 2 O or pentane are either extremely corrosive, or flammable. 2 6

29 Figure 2.2 Carbon dioxide phase diagram While many small molecules are soluble in scco 2, most polymeric materials have been found insoluble. Since 1992, DeSimone et al. have shown that many high molecular-weight fluorinated polymers are readily soluble and synthesized in scco 2. 8,11 This has lead to the development of block co-polymer surfactants, which increase the utility of scco 2 as a solvent for common hydrocarbon-based polymers. 12 While most polymeric materials themselves are not readily soluble in scco 2, scco 2 itself is soluble to moderately high weight fractions in a number of engineering polymers under elevated pressures. Examples of these are PDMS, 13,14 PEEK, 15 PC, 16 PET, 17 PMMA, 18,19 polystyrene 20 and polyurethane. 21 An understanding of the relationship between pressure and the solvent quality of scco 2 can be gained by examining the Hildebrand solubility parameter, δ. 22 From thermochemical studies on polymer-solvent interactions, it is expected that when δ 7

30 values of the polymer and solvent are similar, maximum solubility will occur. 22 As the values of δ diverge from one another, interaction is minimized, and eventually the polymer will become immiscible in the solvent. The solubility parameter for gaseous, liquid, and supercritical CO 2 is shown in Figure Below the critical point (-30 C) a jump occurs in δ at the vapor pressure, corresponding to the phase change between liquid and gas. The solubility parameter for gaseous carbon dioxide is essentially zero whereas the value for liquid CO 2 is more comparable with a typical hydrocarbon value, near 9 (cal/cc) (1/2). However, at temperatures at or above the critical point (31 C, 70 C) it is possible to vary δ between gaseous and liquid-like values. 23 Similar tunability of scco 2 density can be seen in Figure The wide range of achievable values for both the solubility parameter, and density allow the solvent quality of scco 2 with respect to a specific solute to vary from good to poor with small changes in pressure. 8

31 Liquid -31 o C 31 o C 70 o C Phase Boundary Supercritical Gas Figure 2.3 Solubility parameter, δ, of CO 2 : At -31ºC the fluid undergoes a firstorder phase transition, resulting in a step change in d from liquid to gas. Above the critical point, 31ºC & 70ºC, a continuum of properties exists, allowing previously unstable values to be obtained. Phase Envelope Subcritical Supercritical Figure 2.4 Density continuum for supercritical CO 2 : At the subcritical conditions densities between 0.2 and 0.8 are not found because of the first-order phase change. In the supercritical state, all values of density are 9

32 2.2 Plasticization of Polymer Melts In 1951 the Council of the International Union of Pure and Applied Chemistry defined a plasticizer in the following manner: 25 A plasticizer or softener is a substance or material incorporated in a material to increase its flexibility, workability or distensibility. A plasticizer may reduce the melt viscosity, lower the temperature of second order transitions or lower the elastic modulus of the product. In general, there are two types of plasticizers, internal and external. When an original polymer undergoes a chemical modification or a copolymerization reaction is used to alter the chemical structure of a base polymer to improve flexibility or low temperature properties, it is referred to as internal plasticization. External plasticization is accomplished by addition of a discrete material into the polymer matrix, which brings about the same result of enhanced flexibility and low temperature properties. Our discussion will be limited to external plasticizers only, as this is the focus of this research endeavor. External plasticizers are most often used either to modify the mechanical properties of end-of-the-line products, such as PVC tubing, or during processing to reduce the large energy costs associated with mixing highly viscous materials. 26 These plasticizers are typically small molecules that are compatible with a host polymer allowing them to penetrate much like a solvent. 27 Compressed fluids like supercritical CO 2, N 2 O, propane, and C 2 H 4, can dissolve to appreciably high levels in polymers at elevated pressures, as previously discussed. 28 The absorption of these fluids modifies the polymer by swelling the 10

33 matrix and increasing free-volume and chain mobility. Free-volume is defined as the unoccupied space inside of a polymer matrix, into which polymer chains can move. 29 Large depressions in T g of a wide variety of polymers have been shown to occur by introduction of a high-pressure fluid, specifically using scco 2 as a penetrant. 18,30-33 ScCO 2 behaves much like a typical low molecular weight plasticizer; however the concentration can be selectively tailored by controlling the pressure. This phenomenon is referred to as gas-induced plasticization. Gas-induced plasticization allows an initially glassy polymer to demonstrate the flexible properties associated with a rubbery or liquid-like state. By lowering the T g and reducing viscosity, new processing windows may be available for the plasticized polymer melts that not only reduce energy costs, but also improve the efficiency of the process. 34 ScCO 2 is an efficient plasticizing agent because of its unique fluid properties. Its liquid-like density, high diffusivity, and low molecular weight make it soluble up to high weight fractions 0f 10-20% in a variety of different polymers The specific effects of the high solubility of CO 2 in the plasticization of polymers is shown in Figure The plasticization of an amorphous polymer by scco 2 has three distinct regions. In region I, the swelling of the polymer matrix by the dissolved gas dramatically lowers T g. This drop in T g can occur even with small concentrations of CO 2 in the matrix (1-5 wt%). 20 In region II, a constant T g is observed at significantly lower temperatures than that of the pure substance, T o g. The large hydrostatic pressures generated to increase the solubility of the CO 2 dominates in region III. Loss of freevolume observed by the compression of the matrix by hydrostatic pressure results in a T g increase. The size of each region and the amount of plasticization depends on the 11

34 degree of swelling of the matrix, the polymer compressibility, and the solubility of CO 2 in the polymer. Similar behavior has been observed for the phase behavior of polymer blends plasticized by CO 2. 38,39 Several studies in recent years have focused on plasticization of polymers induced by the dissolution of scco 2. These studies have generally measured or calculated the changes in T g with pressure and system temperature. For instance, a 70 C drop in T g of PMMA at 6 MPa (60 bar) has been observed. 19,29,33 Similar depressions of T g have been demonstrated for PEG, 40 PPO 31,32 and PS. 20 An example of the experimental data available in the literature is shown in Figure This figure details experimental measurements of polystyrene T g reduction due to high pressure CO 2 incorporation over a range of temperatures and pressures, which is comparable to the theoretical model presented in Figure 2.5. Temperature T g o I II III Rubber Glass CO 2 Pressure Figure 2.5 Proposed model of glass-liquid transition for a polymer-co 2 mixture. 12

35 Figure 2.6 T g of polystyrene as a function of CO 2 pressure. 2.3 Applications for CO 2 -Plasticized Polymer Melts Melt Phase Synthesis When a bulk amorphous polymer is heated above its glass transition temperature, or a semi-crystalline polymer above its melting temperature, it is commonly classified as polymeric melt. In the melt phase, the polymer chains are able to rotate freely and slip past each other, and the polymer behaves like a highly viscous liquid. However as the molecular weight of the polymer increases, the viscosity of the melt can increase significantly, decreasing the mobility of the individual chains. 13

36 Many industrially important polyesters and polycarbonates could be manufactured to high molecular weight by solvent free melt-phase polymerization, specifically poly(ethylene terephthalate), PET, and bisphenol-a-polycarbonate, PC. 41,42 While solvent-free melt-phase polymerization methods avoid the use of environmentally detrimental solvent mixtures, it would be convenient to mediate the high viscosity of the polymer melt associated with the attainment of high molecular weight polymer, especially near the end of the polymerization. 43 It has also been noted previously that the condensate associated with certain step-growth polymerizations can be removed more effectively by swelling or plasticization of the polymer melt. 44 In order to achieve high molecular weight, the polymer must be separated from the condensation products to overcome unfavorable equilibrium constraints. 45 It has been suggested that the high melt viscosity of both PET and PC require hightemperature reactors in the final stages to enhance the diffusion and removal of the condensation product. If the viscosity of the melt could be lowered in the high molecular weight regime it may be possible to increase the rate of condensate removal and lower the temperature requirement of the reactor. This assumes that viscosity is the limiting factor in the condensate diffusion. One possible method for improving the melt synthesis is to plasticize the melt during synthesis. 44 This would allow the polymer to have the same viscoelastic properties at lower temperatures. Running the reaction to high molecular weight at a lower temperature would have several industrial advantages. The lower temperature could avoid the generation of unwanted byproducts, such as color bodies in the case of PC. It would also eliminate the need for 14

37 downstream separation of the by-product and generate substantial savings in the energy costs associated with heating. Enhanced Polymer Blending Polymers, by their chainlike nature, tend to be immiscible due to generally low entropy of mixing and typically unfavorable thermodynamics of mixing The ability to alter, in tunable fashion, the miscibility of polymer blends is highly desirable, since it could expand the window of processability. While miscibility tuning is achievable through the use of organic solvents, this route is not environmentally benign and may require the use of hazardous chemicals. Using scco 2 as a plasticizing agent can alter the miscibility of polymer blends. The pressure tuning of CO 2 solubility in a polymer blend more significantly allows for the miscibility to be varied in a controllable fashion. Since scco 2 serves as a plasticizing agent for many polymer melts, it follows that scco 2 may reduce unfavorable interactions between two polymers. This is especially so if each interacts marginally with scco 2 to increase their free volume. This would subsequently enhance their compatibility and shift their phase boundaries in the direction of greater miscibility, and improve viscosity ratios. Watkins et al. 38 have recently provided encouraging evidence that supercritical carbon dioxide can be used to shift phase/morphology boundaries of block copolymers. In addition Walker et al. 39 has shown that in a fundamental study of low molecular weight polystyrene and polyisoprene blends that scco 2 can be used to effectively lower the temperature at which total miscibility between the two 15

38 components occurs. Figure 2.7 illustrates the decrease in T *, the cloud point depression temperature, that is achieved with dissolved high pressure carbon dioxide T* Pressure (MPa) Figure 2.7 Variation of T* with scco 2 pressure for PS-PI blends of differing compositions. The upward pointing solid triangles represent 0.1wt% PS, while the downward pointing triangles designate 0.2wt%. The solid circles represent experiments with the PS-PI blend exposed to N 2 at 0.2wt% PS. 2.4 References 1. Bruno, T. J.; Ely, J. F. (1991) Supercritical Fluid Technology: Reviews in Modern Theory and Applications, CRC Press: Boston. 2. Kiran, E.; Brennecke, J. F. (1991) Supercritical Fluid Engineering Science, American Chemical Society: Washington DC. 3. Ely, J. F.; Magee, J. W.; Haynes, W. M., Thermophysical Properties for Special High CO2 Content Mixtures, Research Report RR-110 (Gas Processors Association, 1987). 16

39 4. Smith, J. M.; VanNess, H. C. (1987) Introduction to Chemical Engineering Thermodynamics, 4th; MacGraw Hill, Inc.: New York. 5. Coenen, H.; Patzold, R.; Sievers, U. (1985) in Supercritical Fluid Technology, J. M. L. Penninger, M. Radosz, M. A. McHugh, V. J. Krukonis Eds; Elsevier: New York, Canelas, D. A.; Betts, D. E.; DeSimone, J. M. (1996) "Dispersion Polymerization of Styrene in Supercritical Carbon Dioxide: Importance of Effective Surfactants," Macromolecules, 29(8), Quadir, M. A.; Kipp, B. E.; Gilbert, R. G.; DeSimone, J. M. (1997) "Emulsion Polymerization in a Hybrid Carbon Dioxide/Aqueous Medium: Plasticized Latex Particles with a Narrower Molecular Weight Distribution," Macromolecules, 30, DeSimone, J. M.; Guan, Z.; Eisbernd, C. S. (1992) "Synthesis of Fluoropolymers in Supercritical Carbon Dioxide," Science, 257, Magee, J. W. (1991) in Supercritical Fluid Technology, T. J. Bruno, J. F. Ely Eds; CRC Press: Boston, Henon, F. E.; Camaiti, M.; Burke, A. L.; Carbonell, R. G.; DeSimone, J. M.; Piacenti, F. (1999) "Supercritical CO 2 as a solvent for polymeric stone protective materials," Journal of Supercritical Fluids, 15(2), DeSimone, J. M.; Maury, E. E.; Menceloglu, Y. Z.; McClain, J. B.; Romack, T. J.; Combes, J. R. (1994) "Dispersion Polymerizations in Supercritical Carbon Dioxide," Science, 265, McClain, J. B.; Betts, D. E.; Canelas, D. A.; Samulski, E. T.; DeSimone, J. M.; Londono, J. D.; Cochran, H. D.; Wignall, G. D.; Chillura-Martino, D.; Triolo, R. (1996) "Design of Nonionic Surfactants for Supercritical Carbon Dioxide," Science, 13. Gerhardt, L. J.; Garg, A.; Bae, Y. C.; Manke, C. W.; Gulari, E. (1992) in Theoretical and Applied Rheology, P. Moldenaers, R. Keunings Eds; Elsevier Science Publishers: Brussels, Belgium, Gerhardt, L. J.; Manke, C. W.; Gulari, E. (1997) "Rheology of Polydimethylsiloxane Swollen with Supercritical Carbon Dioxide," Journal of Polymer Science; Part B: Polymer Physics, 35,