THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF TEE COLLEGE OF ENGINEERING CRYSTALLIZATION KINETICS AND MORPHOLOGY OF ISOTACTIC POLYSTYRENE BLENDS Steven L. a5'rt A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Chemical Engineering) in The University of Michigan November, 1970 IP-835

ACKNOWLEDGMENTS The author wishes to express his appreciation to many people who made this study possible: To Associate Professor Gregory S. Y. Yeh, chairman of the doctoral committee, who contributed unselfishly of his time, advice, and assistance during the course of this work. To Professor G. B. Williams whose wise counsel, patient advice, and encouragement during the critical stages of this study greatly contributed to its successful conclusion. To the members of the doctoral committee: Professors W. C. Bigelow and L. O. Brockway, and Associate Professor Orville F. Kimball, who have offered helpful criticisms, suggestions, and advice throughout this study. To the National Science Foundation, Rackham Graduate School, the Dow Chemical Company, the Whirlpool Corporation, and the Department of Chemical and Metallurgical Engineering for financial assistance. To Dan Luch for the many hours of helpful and interesting discussions and suggestions during the years of our graduate studies. To Jens and Leah Pedersen for many years of common Joy and mutual suffering as neighbors and fellow graduate students. ii

To the folks in Warsaw for care packages and for caring. To my parents for their quiet encouragement, high standards, patient love, and unselfish support. To my family, Bamse, Tinker, and Mano who provided many a joyous and fun-loving moment during the past several years and above all to my wife, Jackie, who with her constant love, patience, and confidence has made the truly great sacrifices necessary for the successful completion of this dissertation. iii

TABLE OF CONTENTS ACKNOWLEDGMENTS................ ii LIST OF TABLES vii LIST OF FIGURES........................ viii LIST OF SYMBOLS USED IN THIS STUDY..................... xiv CHAPTER PAGE I. GENERAL INTRODUCTION... 1 II. MATERIALS, EQUIPMENT, AND ELECTRON MICROSCOPY TECHNIQUES......... 4 A. Materials................................... B. Equipment........................ 4 C. Electron Microscopy Techniques.............. 6 III. CRYSTALLIZATION OF ISOTACTIC/ATACTIC POLYSTYRENE BLENDS................ *...................... 8 A. Introduction.............................. 8 1. Technical Review........................ 8 2. Kinetic Theory.................... 12 B. Experimental................................ 18 1. Preparation............................. 18 2. Optical Microscopy...................... 20 3. Measurement of Growth Rate.............. 21 4. Melting Temperature Determination....... 22 5. X-ray Studies.......................... 24 C. Results..... 24 1. Morphology....................... 24 2. Effect of Temperature.................. 27 3. Effect of Diluent Concentration......... 32 4. Effect of Diluent Molecular Weight.. 33 5. Optical Melting Temperatures............ 38 iv

CHAPTER PAGE 6. X-ray Examination of Lattice Parameters and Defects........................... 42 D. Discussion.................................... 43 1. Kinetics of Spherulitic Growth Rate...... 43 2. Growth Rate Phenomena.....46............ E. Conclusions........... *.....to......... 52 IV. CRYSTALLIZATION OF ISOTACTIC POLYSTYRENE/PLASTICIZER MIXTURES.........................'D. 54 A. Introduction........................... 54 B. Experimental.............. 58 1. Preparation................ 58 2. Measurement of Growth Rate. 5 3. Measurement of Melting Temperature....... 61 C. Results.......... *................ 61 1. Morphology 61 2. Effect of Temperature............... 62 3. Effect of Diluent Concentration.......... 62 4. Effect of Diluent Type... 63 5. Melting Temperature Determination....... 66 6. Threshold Crystallization Study........... 73 D. Discussion................. 75 1. Growth Rate Phenomena............... 75 2. Activation Energy for Viscous Transport.. 82 E. Conclusions........................... 90 V. STRAIN INDUCED CRYSTALLIZATION................... 93 A. Introduction................................. 93 B. Experimental................................. 98 C. Results.............. *............... 101 1. Morphology of Holey Films............... 101 2. Gold Decoration of Holey Films........... 103 3. Morphology of Films Stretched on Mylar... 107 4. Morphology of Films Stretched on Water... 113 D. Discussion............................... 131 V

CHAPTER PAGE 1. Characterization of Structure........... 131 2. Mechanism of Strain Induced Crystallization',.............................. 135 E. Conclusions............1...................... 137 VI. MISCELLANEOUS STUDIES......................... 139 A. Introduction................................. 139 B. Experimental............................ 142 C. Results................................ 143 1. Amorphous Structure................ 143 2. Crystallization of Isotactic Polystyrene from the Glassy Amorphous State.......... 150 D. Discussion................................. 156 E. Conclusions................................. 157 VII. GENERAL CONCLUSIONS AND MAJOR FINDINGS........... 159 VIII. RECOMMENDATIONS FOR FUTURE STUDY........... 164 APPENDIXo.........................................o.........a166 BIBLIOGRAPHY............................ 178 vi

LIST OF TABLES TABLE PAGE I. Materials Used for Crystallization Study........ 5 II. Half width of the (110) X-ray Peak for Mixtures of IPS/APS..................3............ III. Threshold Crystallization Temperatures for Mixtures of IPS/Benzophenone and IPS/APS........... 75 IV. Relationship between Tmax and Tm for Mixtures of IPS/Plasticizers...................... 77 V. Threshold Growth Rates for Mixtures of IPS/Benzophenone and PS/APS................... 84 VI. Threshold Crystallization TemperaturesExperimental and Theoretical........................ 87 VII. Parameters of Gold Decorated Fibers............. 106 VIIIA. Strain Induced Crystallization of Thin Films of IPS/Benzophenone 60/40 Stretched on Water at Room Temperature, Platinum Shadowed.......... 115 VIIIB. Same Film as in Table VIIIA, Gold Decorated..... 116 VIIIC. Same Film as in Table VIIIA, Amyl Acetate Etched. 116 IX. Spherulitic Growth Rate Data for Mixtures of Isotactic/Atactic Polystyrene.................. 166 X. Spherulitic Growth Rate for Mixtures of Isotactic Polystyrene and Di-methyl Phthalate............17 XI. Spherulitic Growth Rate for Mixtures of Isotactic Polystyrene and Di-decyl Phthalate. e..... 73 XII. Unshadowed Microstructure.................. 176 XIII. Platinum Shadowed Microstructure............ 177 vii

LIST OF FIGURES FIGURE PAGE 1. Circuit diagram for the hot stage of the optical microscope.................. 19 2. Light micrograph (through crossed polarizers) of isotactic polystyrene partially melted and recrystallized at 180~C......... 28 3. Electron micrograph of isotactic polystyrene crystallized from the glassy amorphous state at 140~C. Platinum shadowed at 300.......................... 25 4. Light micrograph (through crossed polarizers) of isotactic/atd&tic polystyrene 40/60 (APS Mw = 19,800) crystallized from the melt at 180~C. Arrows indicate serrated edge of spherulite....... 28 5. Light micrograph (through crossed polarizers) of isotactic polystyrene crystallized from the melt at 180~C. Arrows indicate smooth edge of spherulite....... 29 6. Spherulitic growth rate of isotactic polystyrene.. 30 7. Spherulitic growth rate for mixtures of IPS/APS mw= 4,800.... 31 M - 4,800..................... 31 8. Spherulitic growth rate for mixtures of IPS/APS @ 1800~C............................... 34 9. Data from literature for spherulitic growth rate for mixtures of IPS/APS......35 10. "Maximum" spherulitic growth rate for mixtures of IPS/APS............................... 37 11. Optical melting temperature for IPS/APS mixtures vs. concentration of APS.......................... 40 12. Optical melting temperature for IPS/APS mixtures vs. molecular weight of APS........... 41 13. Comparison of experimental data to theory......... 45 14. Melting temperature vs. concentration of chain ends.. * *. 51

FIGURE PAGE 15. Isotactic polystyrene/benzophenone 50/50 crystallized from the glassy amorphous state at 40~C. Platinum shadowed at 30~......................... 74 16. Spherulitic growth of mixtures of isotactic polystyrene/di-methyl phthalate 90/10................ 64 17. Spherulitic growth rates for mixtures of isotactic polystyrene and di-methyl phthalate............... 65 18. Maximum spherulitic growth rates for mixtures of isotactic polystyrene and b-enzophenone............ 61 19. Spherulitic growth rates for mixtures of isotactic polystyrene and di-decyl phthalate................ 68 20. Maximum spherulitic growth rates for mixtures IPS/APS, IPS/benzophenone, IPS/DMP, IPS/DDP....... 69 21. Melting temperatures and glass transition temperatures for mixtures of isotactic polystyrene and di-methyl phthalate............................... 71 22. Melting temperatures and glass transition temperatures for mixtures of isotactic polystyrene and di-decyl phthalate.................7.............. 72 23. Isotactic polystyrene/benzophenone 80/20, crystallized from the glassy amorphous state at 82~C. Gold decorated at 90~................... 74 24. Comparative spherulitic growth rates for mixtures of IPS/APS and IPS/plasticizer.................... 81 25. Isotactic polystyrene holey film crystallized at 140~C for 5 minutes. Platinum shadowed at 300~.... 102 26. Isotactic polystyrene holey film crystallized on glass slide at 140~C for 10 minutes. Gold decorated at 90~...................................... 102 27. Isotactic polystyrene holey film crystallized at 1400~C for 10 minutes. Gold decorated at 90~. Arrows indicate tri-layer decorative pattern...... 104 28. Isotactic polystyrene thin film crystallized at 2100C for 5 hours. Platinum shadowed at 30~...... 104 29. Isotactic polystyrene thin film crystallized at 210~C for 5 hours. Gold decorated................ 105 ix

FIGURE PAGE 30. Isotactic polystyrene holey film crystallized at 140~C for 10 minutes. Platinum shadowed on one side and gold decorated on reverse side. Arrows indicate area where platinum stacks up............ 105 31. Isotactic polystyrene holey film crystallized at 140~C for 10 minutes. Gold decorated at 900. Dense crystalline regions are indicated by arrows.. 108 32. Isotactic polystyrene stretched 100% on mylar, annealed at 1750C for 20 minutes. Platinum shadowed at 30~. Stretch direction is verticle and crystalline fibers are indicated by small arrows......... 108 33. Isotactic polystyrene stretched 100% on mylar, annealed at 175~C for 20 minutes. Electron diffraction pattern. Sketch indicates designations for crystalline reflections.....1..................... 111 34. Isotactic polystyrene stretched 100% on mylar, annealed at 175~C for 5 minutes. Gold decorated. Stretch direction is verticle. Small arrows indicate tri-layer decorative pattern................. 111 35. Isotactic polystyrene stretched 100% on mylar, annealed at 1750C for 20 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30~. Stretch direction is verticle............................ 112 36. IPS/APS/benzophenone 10/80/10 stretched 100% on mylar, annealed at 1550C for 20 minutes, amyl acetate etched for 10 seconds. Platinum shadowed at 300. Stretch direction is indicated by large arrow.........a.............................. 112 37. Comparative growth rates for row structures and spherulites, IPS/APS Mw = 900 40/60............... 114 38. Isotactic polystyrene/benzophenone 60/40 stretched 100%, annealed at 1550C for 2 minutes. Platinum shadowed at 30~. Stretch direction is indicated by arrow.................. 118 39. Isotactic polystyrene/benzophenone 60/40 stretched 100%, annealed at 1550C for 2 minutes. Electron diffraction pattern.. 118 40. Isotactic polystyrene/benzophenone 60/40 stretched 300%, annealed at 1550~C for 6 minutes. Electron diffraction pattern.. 118 x

FIGURE PAGE 52. Isotactic polystyrene/benzophenone 60/40 stretched 400%, annealed at 155~C for 6 minutes, amyl acetate etched for 30 seconds. Platinum shadowed at 30~. Stretch direction is indicated by large arrow... 129 53. Isotactic polystyrene/benzophenone 60/40 stretched 300%, annealed at 155~C for 6 minutes. Platinum shadowed at 30~. Stretch direction is horizontal. 130 54. Isotactic polystyrene/benzophenone 60/40 stretched 500%, annealed at 155~C for 6 minutes. Platinum shadowed at 30~. Dark field of figure 53 using (102) reflection. Stretch direction is horizontal, as indicated by electron diffraction pattern...... 130 55. Isotactic polystyrene Mw = 1,200,000 amorphous film cast from dichlorobenzene solution. Platinum shadowed at 30~................................. 145 56. Atactic polystyrene Mw = 1,800,000 stretched 100% on mylar. Platinum shadowed at 30~. Stretch direction is horizontal......................... 145 57. Atactic polystyrene Mw = 1,800,000 stretched 100% on mylar. Gold decorated. Stretch direction is indicated by large arrow. Small arrows show alignment of gold particles...................... 146 58. Atactic polystyrene annealed 30 minutes at 2200C, ice water quench. Platinum shadowed at 30~....... 146 59. Atactic polystyrene annealed 30 minutes at 220~C, ethanol-dry ice quench. Platinum shadowed at 30~. 148 60. Atactic polystyrene annealed 30 minutes at 1900C, ice water quench. Platinum shadowed at 30~....... 148 61. Atactic polystyrene thin section.. Platinum shadowed at 30~. Cutting direction is verticle....... 149 62. Isotactic polystyrene crystallized at 1400C for 20 minutes. Platinum shadowed at 30............. 149 63. Isotactic polystyrene crystallized at 1400~C for 20 minutes. Electron diffraction pattern. Prints show pattern developed at different intensities... 152 64. Isotactic/atactic polystyrene 50/50 crystallized at 125~C for 60 minutes. Gold decorated.......... 152 65. Isotactic/atactic polystyrene 50/50 crystallized at 125~C for 120 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30~0............. 153 xii!

FIGURE PAGE 66. Isotactic/atactic polystyrene 50/50 crystallized at 125~C for 120 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30~......... 153 67. Isotactic/atactic polystyrene 50/50 crystallized at 1250C for 60 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30~............. 154 68. Isotactic/atactic polystyrene 10/90 crystallized at 145~C for 60 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30............. 154 xiii

LIST OF SYMBOLS USED IN THIS STUDY bo = thickness of monomolecular layer Ci = interfacial concentration of rejected impurity C6(T) - concentration of the crystallizable polymer far removed from the crystallization front C1 = WLF constant, C1 = 4120 cal./mole C2 = WFL constant, C2 = 750K (Boon and Azcue); C2 = 30.7~K (Suzuki and Kovacs) AC = change in heat capacity ED = free energy of activation for transport across the liquid-nucleus interface AF = Gibbs free energy difference per unit volume between the supercooled phase and the crystalline phase AF* = free energy required to form a nucleus of critical size Afu = free energy of fusion per molecule G = spherulitic growth rate, p/minute G(undil) = spherulitic growth rate for homopolymer systems G(dil) = spherulitic growth rate for diluted polymer systems Go = constant which is independent of temperature GT = threshold spherulitic growth rate AH = Heat of fusion k = Boltzmans Constant Mn = number average molecular weight Mw = weight average molecular weight R = gas constant, 1.98 cal./mole-~K T = glass transition temperature g T max= temperature for maximum spherulitic growth rate Tm = melting temperature xiv

T = melting temperature of homopolymer m T =threshold crystallization temperature Tx crystallization temperature V1 = molar volume per repeat unit of diluent Vu = molar volume per repeat unit of crystallizable polymer U2 = volume fraction of crystallizable polymer X1 = polymer-diluent interaction free energy y = exponential variable n = melt viscosity p = cross sectional area of nucleus 5 = length of molecular segment in nucleus ae = end interfacial energy per molecule a = lateral interfacial free energy per molecule I = -2080/(C2+T-Tg) II = -279Tm/T(T -T) III = 0.2Tmnu,2/(Tm-T) xv

CHAPTER I GENERAL INTRODUCTION Polymeric materials are of commercial importance primarily because their long chain molecular structure generates a unique set of physical properties not found in comparable monomeric materials. The behaviocr of this long chain structure is greatly influenced by the chemical arrangement of the monomeric units along the chain backbone and the physical conformation of the molecules or morphology. Previous studies [17, 21] on polymer morphology and related crystallization phenomena have generally concentrated on homopolymer or undiluted pure polymer systems. However, commercial polymers are often diluted with either impurities or by plasticizers which have been purposely added, and these diluents are known to influence both the crystallization kinetics and the resulting morphology of diluted polymer systems [30]. Previous studies on mixtures of isotactic/atactic polystyrene [30] and isotactic polystyrene/benzophenone [5] have indicated that the concentration and molecular weight of the added diluent strongly affects the morphology and crystallization kinetics of isotactic polystyrene. However, the mechanism of the polymer-diluent interaction is not clearly understood, since only a limited range of diluent 1

molecular weights and plasticizers were considered. It was therefore decided to make a more detailed study of this polymer-diluent interaction phenomenon by considering a wider range of atactic polystyrene molecular weights and several different types of low molecular weight commercial plasticizers. In this study we used optical and electron microscopy, wide angle x-ray and optical melting temperature measurements to analyze the morphology and crystallization kinetics of isotactic/atactic polystyrene, as well as isotactic polystyrene/di-methy.l or di-decyl phthalate over a wide range of concentrations and crystallization temperatures. Another mode of polymer crystallization which is of great importance is that of crystallization under strain. Melt extrusion, fiber spinning, and injection molding are perhaps the most common fabrication techniques used for polymer processing, and each of these produces a strain field in the polymer melt prior to crystallization. As a result, most studies have generally concentrated on strain induced crystallization from the melt, and studies of this type are currently being carried on in our laboratory and in others [1, 34]. However, there are few comparable studies of strain induced crystallization from the glassy amorphous or rubbery states of a pure or diluted polymer system. Since the basic mechanism of strain induced crystallization still remains a subject of major speculation, we decided to examine in detail strain induced crystallization behavior of isotactic polystyrene and isotactic polystyrene/benzophenone systems under a variety of conditions from the glassy amorphous and rubbery

states. In this study we used bright and dark field electron microscopy and electron diffraction to examine the morphology of thin films stretched to various elongations and annealed at several different temperatures. Finally, recent studies [12, 62] have indicated that the microstructure of amorphous polymers might play an important role in the mechanism of crystallization from the glassy amorphous state. Since the mechanism and mode of crystallization are of primary importance as far as crystalline morphology is concerned, the morphology of amorphous thin films of isotactic and atactic polystyrene was also examined. The combined study of crystallization kinetics and morphology of isotactic polystyrene blends is organized into five major chapters. Chapter II contains a listing of the common materials and primary techniques used in Chapters III, IV, V, and VI. These chapters deal respectively with: the crystallization of mixtures of isotactic/atactic polystyrene; the crystallization of mixtures of isotactic polystyrene/plasticizer; strain induced crystallization; and miscellaneous studies dealing with the morphology of amorphous polystyrene and related topics. Each of these chapters (III, IV, V, VI) has a technical review section, as well as sections dealing with experimental procedures, results, discussion, and conclusions. General conclusions and major findings are indicated in Chapter VII.

CHAPTER II MATERIALS, EQUIPMENT, AND ELECTRON MICROSCOPY TECHNIQUES The purpose of this chapter is to combine together a common listing of the materials, equipment, and some of the electron microscopy techniques used in the kinetic and morphology sections of this work. Other preparation techniques having to do with optical microscopy and thin film preparation for electron microscopy will be considered under the experimental heading of the appropriate chapter in which the experimental techniques are discussed. A. Materials All the materials used for the following study are characterized in Table I with regard to molecular weight, structure, and source. B. Equipment A JEM-6A electron microscope was used to examine thin films of crystalline and amorphous polystyrene indicated in Chapters IV, V, and VI. The electron microscope was operated for this study at 80 killivolts accelerating potential with magnifications ranging from 5,000 to 40,000X. These magnifications were carefully calibrated and periodically rechecked

Table I Materials Used for Crystallization Study Material M. W. Structtre Source H H H HH isotactic 550,000 -C-C-C-C-C- Dow Chemical polystyrene* (viscosity) H H Company H H H H H Isotactic 1,200,000 -C-C-C-C — Dow Chemical polystyrene** (viscosity) 0 H H Company H H H atactic - -C - Pressure polystyrene***900(light scat.) H H HH ) Chemical 2,300(light scat.) Company 4,800(light scat.) 10,000(light scat.) 19,800(light scat.) 51,000(light scat.) 411,000(light scat.) 1,800,000(light scat.) atactic H H H H polystyrene -C-C-C-C-~ - Shell Chemical (commercial) H HH H Company benzophenone 182 (C6H5)2 Eastman Chemical Company di-methyl 194 C6H4(COQCH3)2 Aldrich Chemphthalate ical Company di-decyl 446 C6H4(COOC10H21)2 Eastman Chemphthalate ical Company benzene 78 C6H Baker Chemical Company chloroform 118 CHC13 Baker Chemical Company dichloro- 146 C6H4C12 Baker Chemical benzene Company cyclohexanone 98 C6H100 Baker Chemical Company *According to x-ray analysis, the relative crystallinity of this material as received is 28%, and it contains 15% atactic polymer which was removable by extraction in boiling methyl ethyl ketone. The crystalline melting point was measured to be 2350C by Dr. F.L. Saunders, Dow Chemical Company. **This material also contained 15% atactic polymer which was removed by extraction in boiling methyl ethyl ketone. ***The ratio of the weight average (light scattering) molecular weight to the number average molecular weight is 1.06.

using a 54,864 line per inch grating replica supplied by William Ladd Inc. A Phillips EM300 electron microscope was also employed for a portion of this study. The transmission x-ray scatter from bulk samples of crystalline isotactic polystyrene (Chapter III) was measured by a Norelco vertical diffractometer using nickel filtered Cu Ka radiation at 35 killivolts and 15 milliamperes. Samples of crystallized polystyrene were scanned through 26 angles ranging from 2.00 to 18.00 at a rate of 0.50 per minute and strip chart recorded at a speed of one inch per minute. 10 disperse and scatter slits and.006" receiving slits were used. C. Electron Microscopy Techniques Thin films prepared for examination in the electron microscope were usually shadowed with platinum or decorated with gold in order to enhance the contrast or to bring out surface features. Fine platinum shadowing on the specimen surface was produced by positioning the specimen grids at an angle of 30~ and twelve centimeters away from a platinum wire-carbon rod assembly in a vacuum chamber. The assembly consists of 4 millimeters of 0.008 inch diameter platinum wire wrapped around a thin carbon rod connected to two electrodes. Upon reaching a vacuum of approximately 10-5 torr, the voltage across the carbon rod was slowly increased until the platinum wire melted and formed a bead. The vacuum chamber was opened and the carbon rod rotated until the bead was directed at the specimen grids. Upon re-establishing

7 the vacuum and slowly increasing the voltage, a point was reached where the platinum bead vaporized, thereby shadowing the surface features or surface contours of the thin films mounted on the specimen grids. Gold decoration was applied to the specimen surface by placing the grids thirteen centimeters directly beneath a V-shaped thin tungsten wire filament with three millimeters of 0.008 inch diameter gold wire wrapped around the bottom of the V. After evacuating the chamber to approximately 10-5 torr, voltage was slowly increased across the filament until the gold wire melted and formed a bead. By further increasing the voltage, a point was reached where the gold bead uniformly wet the tungsten wire and vaporized, thereby decorating the surface of the thin films mounted on the specimen grids. Although the exact mechanism of gold decoration is not understood, it is known [55] that gold decoration provides a surface mapping which is sensitive to crystalline fibers of polymers.

CHAPTER III CRYSTALLIZATION OF ISOTACTIC/ATACTIC POLYSTYRENE BLENDS A. Introduction Spherulitic crystallization from mixtures of isotactic/atactic polystyrene is strongly influenced by the nature of the noncrystallizable impurity —-atactic polystyrene. Published results [30] indicate that the morphology and crystallization kinetics are related to both the concentration and molecular weight of the impurity, as well as the conditions of crystallization. However, the "interaction" of polymer-diluent molecules is not clearly understood since only a very limited range of diluent molecular weights were considered by Keith and Padden [30,31]. In addition, there is an apparent discrepancy between the data of Keith and Padden [31] and Boon and Azcue [5]. We therefore decided to make a more detailed examination using a wider range of atactic polystyrene molecular weights (900 to 1,800,000). In this study we used optical microscopy to measure the crystallization kinetics of various blends and optical and electron microscopy as well as wide angle x-ray diffraction and melting temperature measurements to examine the morphology of crystallized mixtures. In particular, x-ray line broadening and melting temperature depression were used to detect possible lattice defects caused by atactic polystyrene impurity.

9 1. Technical Review Previous studies by Hay [21], Keith and Padden [30], and Boon, Challa, and Krevelen [7] have established that at a fixed temperature, the radius of a growing spherulite of pure isotactic polystyrene increases linearly with time over a wide range of crystallization temperatures. As the concentration of available amorphous material is depleted, the diffusion of polymer segments to the interface between the crystal and the melt becomes rate controlling, and the radius of a growing crystal correspondingly becomes proportional to the square root of time, retarding the crystallization process [41]. According to the data of Boon, Challa, and Krevelen [7] and Keith and Padden [31], the linear growth rate of pure isotactic polystyrene reaches a maximum at 175~C and 178.50C respectively, and is uniformly depressed at both higher and lower temperatures, forming a generally symmetrical gaussian curve over a wide range of crystallization temperatures. Keith and Padden [31] reported that at a given temperature, dilution of isotactic polystyrene with noncrystallizable atactic polystyrene depresses the spherulitic growth rate in proportion to the amount of diluent added. Under these conditions the growth rate is also influenced by the molecular weight of the diluent in that higher molecular weight atactic polystyrene, 247,000, depresses the growth rate more than lower molecular weight atactic diluent, 41,700, at similar concentrations. They also reported that the position of the maximas in the growth rate curves tends to shift toward'190~C for high concentrations of atactic diluent molecular weight 247,000. Finally, Keith and Padden

10 observed that the optical melting temperature of pure isotactic polystyrene is only slightly depressed —20C to 3~C —by the addition of up to 75% atactic polystyrene, molecular weight 41,700 or 247,000. These results are in qualitative agreement with those subsequently reported by Boon and Azcue [5], who also noted that the spherulitic growth rate of isotactic polystyrene is depressed by the addition of atactic polystyrene, while the melting temperature is essentially unaffected. Boon and Azcue, however, observed that at the optimum growth rate, the amount of depression is approximately proportional to the concentration of isotactic polystyrene to the 2.58 power, rather than to the first power, as reported by Keith and Padden [31]. Keith and Padden [30] also investigated the effect of impurity segregation on the morphology of spherulitic crystallization of isotactic/atactic polystyrene. Their results show that impurities are rejected preferentially by the growing crystals and that impurity diffusion plays an important role in governing the overall crystalline morphology. In particular, the openness of the texture of the resulting spherulites is a direct function of the amount of slowly diffusing impurity entraped within the growing crystal. Based on these results, they proposed a phenomenological theory of spherulitic crystallization which states that the polycrystalline character of this type of crystallization arises principally because the entrapment or accumulation of impurities rejected at the growing crystal front causes the radically growing fibers of the spherulite to branch

11 noncrystallographically. Keith [29] also observed this type of noncrystallographic branching in small hexagonal hedrite nuclei generated At high temperatures in thin films (2-20p) of mixtures of isotactic/atactic polystyrene. Upon reaching a critical diameter, the hexagonal edges of these hedrite nuclei become unstable and develop multibranched radial fibrous structures characteristic of spherulitic crystallization. Keith interprets this behavior in terms of the phenomenological theory mentioned earlier. Based on all of these results, Keith and Padden [311 proposed an approach to absolute theory of spherulitic growth kinetics which takes into account the interfacial concentration of the crystallizable constituent, Ci. According to their model, the interfacial concentration, Ci, is a general function of the spherulitic growth rate, G, the molecular diffusivity, D, and the concentration of the crystallizable polymer at a distance far removed from the crystallization front, C (T), i.e. Ci = f[G,D,C (T)]. The resulting growth rate G does not appear as an explicit function of known parameters of the system, since it is also dependent on the interfacial concentration, Ci. Therefore, according to their model, the overall growth rate cannot be directly calculated when the effect of rejected impurity on the interfacial concentration is considered. In addition, Keith and Padden reported that a parameter of major significance is 6 = D/G, where D is the diffusivity of the impurity in the melt and G is the spherulitic growth rate. They suggest

12 that 6 should be the equilibrium width of the crystalline fibers and that noncrystallographic branching occurs when the fibers reach this width. In a related study, Magill [36] found that, in addition to other factors, the spherulitic growth rate G is dependent on the molecular weight of the crystallizable material, P-Silphenylene Siloxane (TMPS), crystallized from the amorphous melt of a homopolymer system. In this case he was working with a one component system rather than the two component system studied by Keith and Padden. Results show that the growth rate G% (1/Mn)Y where Mn is the number average molecular weight and y is a factor which usually varies from 0 to 1.0. In the high molecular weight range y=O, but as the molecular weight decreases below to the critical molecular weight for chain entanglements, y suddenly assumes values approaching 1.0. Fox and Flory [15] also observed significant changes over the range of the critical molecular weight for the melt viscosity and glass transition temperature of atactic polystyrene. In the case of the melt viscosity, they interpret these changes in terms of the molecular chain entanglements which occur above the critical molecular weight, Mc=36,000. 2. Kinetic Theory As mentioned earlier, numerous studies show that at a fixed temperature, the radius of a growing spherulite increases linearly with time over a wide range of crystallization temperatures. An analysis of the temperature

13 coefficients, as well as the rate of formation and growth of spherulitic centers suggests that nucleation theory may be applied to the crystallization process. Utilizing transition state theory, Turnbull and Fisher [58] have developed an expression for the rate of nucleation in condensed systems. The solid state nucleation rate per unit volume is given by: -ED-AF* N=Noexp( RT ) (1) where ED is the free energy of activation for transport across the liquid-nucleus interface, AF* is the free energy required to form a nucleus of critical size, and No is related to the number of molecules per unit volume in the liquid. For homopolymers the rate of secondary nucleation or the radial growth rate G of spherulites follows a relation of the form: -ED-AF* G=Goexp RT (2) where Go is thought to be a constant independent of temperature, although some experimental results suggest that Go is dependent upon molecular weight [23]. Other investigators [56] have found that the value of Go is also dependent upon the type of analysis used to evaluate the empirical constants found in the ED and AF* terms. As a means of evaluating the term ED describing the segmental jump rate across the liquidD nucleus interface, Hoffman and Weeks [23] suggested the following approximation based on the empirical expression

14 of Williams, Landel, and Ferry [601: ED-=CT/(C2+T-Tg)= 4.12X103T/(C2+T-Tg) (3) where Tg is the glass transition temperature and C1 and C2 are empirical constants usually taking values of C1 = 4.12 Kcal/mole and C2 = 51.6~K for most polymers. This expression describes the segmental jump rate required for the initiation of viscous flow in the bulk. Substitution of equation 3 into 2 yields: G=Goexp[-4.12X03 /R(C2+T-Tg)] exp[AF*/RT] (4) The expression of the free energy required to form a nucleus of critical size, AF*, is presented by several authors. For example Mandelkern [41] uses the following expression: AF=2aePbo+2aub o-pboAfu (5) where Afu is the free energy of fusion per molecule, b0 the thickness of one molecule, vu the lateral interfacial free energy per molecule, and ae the end interfacial free energy per molecule. The further development of this expression is identical to the development to be used for the diluted system, and therefore will not be discussed here. The final form of equation 4 is as follows: G(undil)=Goexp[-4.12x103/R(C2+T-Tg) ] -4Ib a ar Tm(6) O ~'U e exp [ ":T(Tm-T

15 This equation describes the theoretical spherulitic growth rate for homopolymers and is based on a development using monomeric nucleation theory and an empirical approximation for the interfacial jump rate term. In addition, it is assumed that the heat (AH) and entropy (AS) of fusion are independent of temperature. The addition of noncrystallizable diluent to a crystallizable polymer affects the spherulitic growth rate by diluting the concentration of crystallizable molecules. According to the analysis of Gornick and Mandelkern [19], the dilution effect is accounted for by altering the free energy of formation for a critical size nucleus, AF*, as well as the pre-exponential term. The free energy of formation for a two-dimensional nucleus from a polymer-diluent mixture is given by [19]: AF=2o'. Pb o +2O' b -p~b Af -RT(p/b )Znu (7) e o u o o u 2 Except for the last term, equation (7) is identical to the free energy expression used for the homopolymer system, equation (5). The term RT(p/bo)Znu2 is an entropic contribution to the free energy because RRnu2 represents the probability of selecting a number of p/b0 crystalline sequences of length C from the mixture [5]. The free energy surface represented by equation (7) contains a saddle point. The coordinates of the saddle point are obtained by setting (aAF/))p and (3AF/ap)S equal to zero. It is then found that: p = 2a,/Afu

16. =- (2 e/Afu) - (RTnu2/b 2Afu) At the saddle point: AF=AF*=4bo aue/Afu 2aURTknu2/b Af (8) The free energy of fusion, Afu, can be approximated by: Af. = AH(Tm-T) (9) where AH is the heat of fusion and Tm is the equilibrium melting temperature. This approximation implicitely assumes that the heat (AH) and entropy (AS) of fusion are independent of temperature and equal to AHf and ASf at Tm respectively. Other approximations [56] for Afu considering ACp=constant or d(ACp)/dT=constant, result in lower values of AfU than found with the first approximation, particularly for temperatures close to Tg. However, Suzuki and Kovacs [56] indicate that the first approximation (equation 9), although being of simplest form, provides a best fit with the experimentally measured spherulitic growth rates of pure isotactic polystyrene over crystallization temperatures ranging from near Tg to Tm. Substitution of equations (8) and (9) into (4) gives: G(dil)=u G exp-l4.12X10 (] (10) -4b O a Tm 2a Tm9nu exp[ o -+ e RTAH- bAHTm- b'H(T)] where the pre-exponential term Go is multiplied by the concentration u2 of the crystallizable constituent because the rate of nucleation is directly proportional to the

17 concentration of crystallizable units [19]. This equation describes the theoretical spherulitic growth rate for polymer-diluent mixtures based on the assumptions inherent in the development of equation 6 and the dilution analysis by Gornick and Mandelkern [19]. The relationship between the spherulitic growth rate G(dil) of a polymer-diluent mixture and the spherulitic growth rate G(undil) of the pure homopolymer can be derived by combining equations 6 and 10, assuming that the melting temperature Tm, the glass transition temperature Tg and the pre-exponential constant Go are the same for the pure and diluted systems. G(dil) = exp[ am Tn U21G(undil) (11) The assumption of constant glass transition temperature Tg for polystyrene of molecular weights greater than 30,000 is verified by the experimental data of Fox and FAory [15]. The assumption of constant melting temperature for both the pure and diluted isotactic/atactic polystyrene systems, however, is valid only over certain concentration ranges and molecular weights of the atactic diluent, according to the experimental results to be presented in Chapter III. In estimating the term 2a /b AH, Boon and Azcue [5] u 0 used the empirical relation Ca =ab AH with a=0.l for the u o lateral surface of a chain type crystal. This gives 2a /b0AH = 0.2. Substitution of this value into equation 11 gives: 0 2Tm Zn u(2 G(dil) = u~ exp[ T.m-T- G(undil) (12)

18 which is the equation used to compare theoretical predictions to the experimentally determined spherulitic growth rates for the isotactic/atactic polystyrene system. B. Experimental 1. Preparation Solutions containing 0.4% polystyrene in benzene were made by completely dissolving a weighed amount of either isotactic or atactic polystyrene in a measured volume of boiling benzene. Upon cooling to room temperature, predetermined ratios of clear 0.4% atactic and 0.4% isotactic solutions were pipetted together to make up a clear 0.4% mixed solution of isotactic/atactic polystyrene in benzene. We have also mixed hot solutions, but there is no difference. Transparent thin films approximately 60 microns thick, as determined by density calculations based on a polystyrene density of 1.0 gram/cm'3 5], were prepared by casting from the 0.4% mixed solution onto thin cover glasses placed on a hot plate set at 500C. These thin films were then dried under vacuum at 50~C for 24 hours to insure removal of any residual solvent. Samples suitable for x-ray analysis were prepared by mixing the proper ratio of finely ground powders of isotactic and atactic polystyrene. The mixed powders were melted on a hot plate at 260~C for one half hour in a 1/8 inch thick aluminum mold, sandwiched between cover glass in order to protect the melted polymner from oxidative degradation. After complete melting, the mold was quickly transferred to a

Thermist s tor Controler ~ ~ ~~~eaer-NW KoflerMicroHotm TemperaFigure 1 Controler Kofler Micro Hot Stage CIRCUIT DIA"*X 10OF -T HE -HOT.ft~Bt~~.:Z%*~:eif~c l MICROSCOPEi~ Figure 1

20 silicone oil temperature control bath set at 180~C for subsequent isothermal crystallization. 2. Optical Microscopy The radial increase in size of the growing spherulitic crystals of isotactic polystyrene were measured through a 400 power Unitron optical polarizing microscope equipped with a Kofler micro hot stage and a filar micrometer eyepiece which was calibrated with a Unitron precision etched slide with markings at 0.01 millimeters. The Kofler micro hot stage was controlled by an Athena model 51T proportional temperature controller equipped with a 0.04 inch diameter glass bead Fenwal thermistor (GA52J16). The controller is capable of full scale response at temperature fluctuations on the order of +.150C. The accompanying diagram, figure 1, illustrates the hot stage circuitry. Because of the high temperature environment, the glass bead thermistor was connected to the controller lead wires by high melting point solder. The thermistor was positioned under a set screw and in direct contact with the heated surface. The surface temperature of the Kofler hot stage was intermittently measured by a 22 gauge chromel-alumel thermocouple sandwiched between two cover glasses and sealed with epoxy resin. This thermocouple sandwich was placed directly on the surface of the hot stage and maintained in a flat position by a small ring weight. Lead wires from the thermocouple and thermistor, sheathed in ceramic insulation, extended out from the hot stage through a smail giove in the ring spacer pLaced

21 on top of the stage. Additional temperature control was afforded by placing a ground glass lid over the ring spacer on the top of the hot stage assembly. The thermocouple response was measured on a Leeds and Northrup potentiometer capable of accurately measuring temperatures to within +.5~C, and it was calibrated against three mercury thermometers at 970C, 175.C, 210~C, and 2200C in a controlled temperature silicone oil bath. Overall temperature control on the Kofler hot stage was better than +.50C, since the thermocouple could not detect any temperature fluctuations once the isothermal crystallization temperature had been reached. In similar studies, Keith and Padden [311 claimed a Kofler hot stage temperature control of +.20C, although no specific details were advanced as to how they achieved this. 3. Measurement of Growth Rate Thin films of well blended isotactic and atactic polystyrene, sandwiched between two cover glasses, were heated on the microscope hot stage to 250~C (10~C above the melting point of isotactic polystyrene) for five minutes, then cooled at a rate of about 20~C per minute for isothermal crystallizing at temperatures ranging from 100~C to 210~C. According to growth rate measurements extrapolated back to zero time, most spherulitic crystals generally did not start to nucleate and grow until some time after the hot stage had cooled down to the isothermal crystallization temperature. This result indicates that many of the nuclei were destroyed by the thermal treatment at 250~C.

22 The spherulites within a given field of view were very uniform in size, suggesting that they were all generated at the same time. The spherulitic growth rate at a given temperature was determined by measuring the size of a number of spherulites within a field as a function of time, by means of the filar eyepiece micrometer. Spherulitic growth rates were measured for thin films having various concentrations of atactic polystyrene, molecular weights ranging from 900 to 1,800,000, added to isotactic polystyrene of molecular weight 550,000. Thus, the growth rate, G, is the dependent variable effected by the concentration and molecular weight of tihe atactic impurity, as well as the temperature of crystallization. Data scatter from duplicate measurements indicates that the growth rate is reproducible for identical conditions to within +5%. The crystals were also photographed by a 35 millimeter Exacta single lens reflex.: camera directly mounted on the optical microscope. 4. Melting Temperature Determination Spherulitic crystals of isotactic polystyrene are birefringent under crossed polarizers in the optical microscope, and the temperature at which this birefringence totally disappears is designated as the optical melting point. Consequently, the optical melting point represents the temperature at which the onset of crystalline disorganization is observed. Melting points measured by careful dilatometry yield results 3-5~C higher than those measured optically [2?]. However, optical measurements are a relatively

23 convenient means of establishing the dependence of the crystalline melting temperature on the concentration and molecular weight of the impurity diluent. Care was taken when making such measurements, since the crystalline melting point is affected by the heating rate [31]. Thin films of isotactic/atactic polystyrene samples crystallized at 180~C were individually heated on the Kofler hot stage at rates ranging from 0.50C to 1.0~C per minute until the birefringence totally disappeared. These films were then cooled to 180~C and recrystallized, a process which generally occurred within ten to twenty minutes. The morphology of recrystallized spherulites was quite different from the fibrous morphology of the original spherulites. Within the former boundaries of the original spherulites, the recrystallized structure appears to consist of tiny birefrigent beads closely impinging upon one another. However, unrestricted growth from these same beads out into the melt caused small spherulitic structures to develop along the original boundaries, figure 2 (page 24). Banks, Gordon, and Sharples [23 noted similar structural changes in partially melted spherulites of polyethylene and attributed it to residual seed nuclei which act as centers for further growth on subsequent cooling. The optical melting points of these spherulites were measured a second time to insure that the measured melting point was not a function of the heating rate. In general, it was observed that heating rates less than 0.7~C per minute do not influence the ultimate optical melting point. This

24 rate is similar to the heating rate of 0.50C per minute used by Keith and Padden [301 for their determinations of spherulitic melting point. The accuracy of the melting temperature values is judged to'be within + 1.5~C, as indicated by the data scatter (see Figure 11). 5. X-ray Studies X-ray studies were also carried out on spherulites of isotactic polystyrene isothermally crystallized from mixtures with a high concentration (80%) of noncrystallizable atactic polystyrene impurity. The x-ray scatter was measured by a Norelco diffractometer as described in Chapter II. The range of 29 angles scanned includes the (110), (220), and (211) crystalline peaks of isotactic polystyrene, but background scatter from noncrystallizable amorphous material tended to 0 obscure all of these peaks except the (110) d10 =11.02A between 2G of 4.50 and 6.5~. Accordingly, the (110) peak was scanned between four and six times for each specimen, and the resulting values of the peak half-width were averaged. C. Results 1. Morphology Electron microscopy examination of typical crystallized thin films of pure isotactic polystyrene shows that spherulitic aggregates are made of radiating bundles of fine O ribbon fibers, approximately 150A thick, as illustrated in Figure 3. Since these fibers grow at a uniform rate in the

25 Figure 3: Electron micrograph of isotactic polystyrene crystallized from the glassy amorphous state at 140IoC. Platinum shadowed at 30~

26 radial direction, the spherulite crystals are usually symmetrically round in shape. Under crossed polarizers in the optical microscope, the crystals have a positive birefringence withthe normal maltese cross. At these crystallization temperatures there is no evidence of the ring type or zig-zag extinction contours observed in polyethylene and polyamide spherulites [38]. The spherulites in thin films (approximately 60O thick) develop from either 1 to 2V fibrous nuclei or from hexagonal hedrite nuclei. Generation of fibrous nuclei is favored by high molecular weight noncrystalline impurity (above 51,000) and crystallization temperatures below 180~C, while hexagonal nuclei appear for low molecular weight impurity (below 19,800) and crystallization temperatures above 1800C. Both of these results are in qualitative agreement with Keith [291 who noted such hexagonal aggregates in mixtures of low molecular weight atactic (4,500) in isotactic polystyrene, particularly at temperatures above 200~C in thin films ranging from 2 to 20p thick. He found that after reaching a certain diameter, the edges of the hexagonal hedrite nuclei became serrated and started to develop a radial fibrous conformation which later developed in the normal spherulitic mode. Examination of Figure 4, however, will show that for the present system, the spherulitic mode tends to initiate at the center of the hexagonal hedrites, rather than at the edge, as indicated by Keith. Subsequent growth follows a normal fibrous morphology and hexagonal nuclei become overgrown with radial spherulitic fibers. The

27 reason for the difference between our observations and Keith's probably is related to film thickness. Keith worked with very- thin films (2-20p) which tended to inhibit the initiation of spherulitic growth on the lateral surface, whereas the present work was based on films approximately 60p thick. Qualitatively, high concentrations of impurity cause the fibrous texture of the resulting spherulites to be more "open", (note the edges of the spherulites, Figure 4), than spherulites grown from pure isotactic polystyrene, Figure 5. This characteristic was also noted by Keith and Padden [30], who observed a direct correlation between the "openness" of the fibrous texture and the concentration of atactic polypropylene in the isotactic/atactic polypropylene system. 2. Effect of Temperature The spherulitic growth rate of pure isotactic polystyrene is linear with respect to time over a wide range of temperatures, as shown in Figure 6. In addition, the growth rate reaches a clear maximum at approximately 178~C (+ 20C) and is depressed at higher and lower crystallization temperatures, Figure 7. These results are in good agreement with previous studies by Keith and Padden [311 and Boon, Challa, and Krevelen [71, who noted the growth rate maxima for pure isotactic polystyrene at 178.5~C and 175.00C respectively. In general, the growth rate becomes nonlinear only under the conditions in which the diffusion of the crystallizing species becomes rate controlling because of the exhaustion

28 Figure 2: Light micrograph (through crossed polarizers) of isotactic polystyrene partially melted and recrystallized at 180CC. Figure 4: Light micrograph (through crossed polarizers) of lsotactic/atactic polystyrene 40/60 (APS Mw=19,800) crystallized from the melt at 180C. Arrows indicate serrated edge of spherulite.

29 Figure 5: Light micrograph (through crossed polarizers) of isotactic polystyrene crystallized from the melt at 180~C. Arrows indicate smooth edge of spherulite.

SPHERULITIC GROWTH RATE OF ISOTACTIC POLYSTYRENE 60 - Mw 550,000 1800C 7 50 1600C ~ 1900C,0 4r-I 3 0~~~~~~~~~ (D~ CO ~ 30. 9 0 - 20 4 0 60 80 100.120 140 160 180 200 Tire, minutes Figure 6

SPHERULITIC GROWTH RATE FOR MIXTURES OF IPS/APS MW = 4,800 TPS Mw = 550,000 0.3 APS = 4,800 * IPS/APS 100/0 eIPS/APS 80/20 OIPS/APS 60/40 DIPS/APS 40/60 IPS/APS 20/80 0.2 4-, 0.. 0 110 120 130 I1o 150 16 170 10 190 200 0 Temperature, 0C Figure 7

32 or depletion of the crystallizable constituent or when spherulites impinge upon one another. 3. Effect of Diluent Concentration Figure 7 shows that dilution with noncrystallizable atactic polystyrene, molecular weight 4,800, causes a general depression of the spherulitic growth rate, with the maximas still occurring in about the same temperature range (178~C + 2~C), as for undiluted isotactic polystyrene. Dilution with the highest molecular weight atactic polystyrene (1,800,000) does not cause the temperature range for the maximum growth rate to shift to higher temperatures, while similar dilution with very low molecular weight atactic polystyrene, 900 and 2,030, only causes a shift to lower temperatures for the 900 molecular diluent, Table IX. In this case, the maximum growth rate temperature range shifts to about 1750C (+ 20C). Therefore, for the system being studied, the maximum growth rate occurs at about 1780C (+ 2~C) for diluent molecular weights ranging from 2,030 to 1,800,000. In the system studied by Keith and Padden [31] high dilutions with atactic polystyrene, moleculCr weight 247,000, caused the growth maximas of isotactic polystyrene, molecular weight 60,000, to shift to higher temperatures (1900C). There is no clear explanation as to why the shift should definitely occur in one case but not appear to in the other, although the difference in molecular weight of the isotactic component, 550,000 vs. 60,000, most likely has an important effect.

33 Figure 8 presents the growth rate data at 180~C for atactic diluents of different concentrations and molecular weights. This temperature was chosen because the maximum growth rates for pure and diluted isotactic polystyrene all occur at about 180~C. In addition, the change in growth rate as a function of temperature, AG/AT, is minimized at this temperature, thus reducing the effect of temperature fluctuations on the accuracy of the growth rate measurement. At this temperature (180~C), dilution with atactic impurity results in a more or less linear depression of the growth rate, in proportion to the amount of diluent added.- This result is in general agreement with data by Keith and Padden [31] for two different molecular weight atactic polystyrene diluents at 178.5~C, but disagrees with the scattered data taken by Boon and Azcue [5] for mixtures of isotactic (Mw 185,000) and atactic (Mw 260,000) polystyrene, as shown in Figure 9. There is a clear discrepancy between the results of Keith and Padden and those of Boon and Azcue which cannot be accounted for by the slight difference in crystallization temperature [5]. 4. Effect of Diluent Molecular Weight Figure 10 shows a cross plot of Figure 8 using diluent concentration as the parameter. Data for atactic polystyrene, molecular weight 900, with a maximum crystallization temperature at about 1750C (+ 20~C), is also included. Both the uncertainty of the exact maximum growth rate temperature (+ 2~0C), as'well as the reproducibility of the data in

cn SPHERULITIC GROWTH RATE FOR MIXTURES OF IPS/APS @ 1800C 0, 1. 1 SIPS/APS Mt = 2,030 e)IPS/APS MW = 14,800 o 1.. DIPs/APS Mw 10,000. 0.9. IPS/APS Mw 19,800 flIPS/APS Mw 51,000 GIPS/APS Mw = 411,000 XIPS/APS Mw =,8o00,000 mm -- — Linear Growth Rate Depression o 0.7' x Temperature = 1800C 07 06 0,5 04 0 rH~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r ~0.2 rH 0'H 0 20 46 0 0 % Atactic Polystyrene (APS) Figure 8

C/I 4) DATA FROM LITERATURE HERULITIC GROWTH RATE FOR o MIXTURES OF IFS/APS 1.0 mi.wrnu.*4IPS/APS MW=41,7Q0 @ 178.50C [31] 0 0.9 IPs/APS MW=247,000 @ 178.50C [31] -- -— eSIPS/A/PS M =260,ooo @ 1750C [511 ~ o.8 ~~~~~~~~ 0 0 a~~~~~~~~~~~~~~ t~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~c, C~~~~~~~~~~~~~~~~~~~~~~~~~A) Q) 0 5) ~ 0.5 4 0. 22**1 0 1 0. % Atactic Polystyrene (APS%.'- 0,3 91 ~Fu Cd~ ~ ~ ~ ~ 2 Co 608010 % Atactic Polystyrene (APS)~ilC C ~ ~ ~ ~ ~ ~ ~ ~~Fgr 9r

36 Figures 7 and 8, suggests that the confidence limit on the data points in Figure 10 ought to be about + 5%. The results show that the spherulitic growth rate for low dilutions (20%) of noncrystallizable impurity is nearly independent of the impurity molecular weight, except for very high molecular weight diluent. This finding suggests that in this concentration range (20%) the diffusional characteristics of the impurity play only a very minor role in the determination of the overall growth kinetics. Apparently, low concentrations of rejected impurity can be efficiently accommodated between the fibrils of the growing crystal and prevented from diffusing radially. This is in agreement with the morphological observations by Keith and Padden [30] who noted that significant amounts of rejected impurity are entrapped within the growing spherulite and do not diffuse away from the growing crystal front. At intermediate diluent concentrations (40% and 60%), quite noticible effects are observed over the entire molecular weight range. Of particular importance is the sudden 30% increase in the radial growth rate as the diluent molecular weight is increased from 19,800 to 51,000. This is surprising because one would normally expect the growth rate kinetics to be more or less uniformly depressed with increasing diluent molecular weight, similar to the trends noted by Keith and Padden [30] (Figure 9) and by Magill [36]. This result cannot be explained by macroscopic (> 2000 A) phase separation since the films were quite transparent for all mixtures prior to crystallization. In regions from 4,800

SPHERJLITIC GROWTH RATE FOR MIXTURES OF IPS/APS AT'TAXIMUM1' RATE QIPS/APS Mw = 900 @1750C 01PS/APS Mw = 10,000 @1800C X-P3/APS Mw= *IPS/APS Mw 2030@180oc GIPS/APS Mw 19,800 @18000 1,800,000 @18000 $DIPS/APS Mw 4800@1800C oxPS/APS Mw = 51,000 @18000 0.3 $IPS/APS 41w P 1i,00 @18000 80/20 4-) c'J~~~ w 40060 4r3 Tb 0.1 Cd ( 20/80 x~~~~~~~~~~~~~~~~~~ 0 AI 5 6 7 8 9 10 11 12 13 14 15 Ln Molecular Weight APS Figure 10

38 to 19,800 and above 51,000 molecular weight, the curves for intermediate concentrations (40% and 60%) have a similar negative slope, although being displaced. However, in the region for very low molecular weight diluents (4,800 to 900), the growth rate shows a maximum at 4,800 and is depressed with decreasing molecular weight. This growth rate trend for lower molecular weights was also observed by Keith and Padden [31] for a 40% concentration of atactic (molecular weights 540; 2,600) in isotactic polypropylene at 125~C and 1300C. These effects will be discussed in more detail in a later section. At high diluent concentrations (80%), the effect of impurity molecular weight is essentially damped out over the range of particular interest: 19,800 to 51,000. There were no observable changes in spherulitic morphology which could be associated with the effects noted in Figure 10. However, gradual changes in general morphology for spherulites grown from these mixtures are discussed on page 26. 5. Optical Melting Temperatures Figure 11 shows the relationship between optical melting temperature and the concentration and molecular weight of atactic polystyrene diluent in thin films crystallized at 180~C. In general, the melting temperature is depressed with increasing concentration and decreasing molecular weight (below 19,800) of the added atactic impurity. Neither Keith and Padden [30] nor Boon and Azcue [5] observed any significant melting point depression, probably because

39 they used only two atactic polystyrene diluents and both were high molecular weight fractions, 178,000 and 247,000, respectively. In somewhat similar studies, Mandelkern [41] noted that low molecular weight diluents, particularly a-chloronaphthalene, significantly depressed the melting temperature of linear polyethylene. He concluded that the melting point of linear polyethylene was depressed by increasing the solubility and the concentration of the diluent. However, in cases where very poor solvents were used as diluents, he observed that the melting temperature was relatively uneffected. He was able to correlate this result with the appearance of macroscopic (> 2000A) phase separation ~n the melt, thus suggesting that according to the Gibbs phase rule, the melting temperature should remain invariant. The melting point-composition relation is, in effect, an expression of the temperature limit of the solubility of the crystalline polymer in the given solvent. Figure 12 shows a cross plot of Figure 11 with the diluent concentration being the parameter. The crystalline melting temperature is relatively constant above molecular weight 51,000, but is considerably depressed for lower molecular weight diluents. This depression is particularly noticable for atactic diluent molecular weight 900, the effect being most severe for the highest diluent concentration. This result indicates that 900 molecular weight diluent is a much better solvent for crystalline isotactic polystyrene than is higher molecular weight atactic diluent. The data scatter, particularly for diluent molecular weight 411,000, can be directly related

OPTICAL MELTING TEMPEFRATURE FOR IPS/APS MIXTURES VS. CONCENTRATION OF APS 230 *...........,.....s; X h 220 4-)'J | QIPS/APS Mw = 900 0) | (DIPS/APS Mw - 4,800 I O~GIPS/APS Mw = 19,800') FL SIPS/APS Mw = 51,a000 $IPS/APS Mw = 411o000 cQ r XIPS/APS Mw = 1,800,000 by Keith and Padden 30 ]' —-.. Data by Keith and Padden [30] 210 for APS M = 41,700; 247,000 w 4 0 200 0 20 40 60 80 00 % Atactic Polystyrene (APS) Figure 11

11~~~~~~x, 2 WOO, XX 3 5 40% 2124 220 - 2 / 60% 0 0 ~,, (I) a*(' 1,1 a)^ I / a)<D<I 1,2,2 OPTICAL MELTING TEMPERATURES FOR IPS/APS Cd| / 80 t MIXTURES VS. MOLECULAR WEIGHT OF APS Q / IPS/APS Mw = 900 |IPS/APS Mw = 4,800 210 G JoIPS/APS Mw = 19,800 BIPS/APS Mw = 51,000 2,.. eSIPS/APS Mw = 411,000 XIPS/APS Mw 1,800,000 Parameter: % APS Heating Rates: 1. 0.60C/mnin and less Cd 12. 0.60C/min-O..7C/oC/ min ~ | / 3. 0.70C/min-0.80C/min a - 1 4. 0.8~C/min and higher 200 I I 6 7 8'9 - ~) -'10 11 12 13 14 15 16 Ln Molecular Weight APS Figure 12

42 to the faster heating rate used. Various heating rates were used to check the effect of rate on the resulting value of the melting temperature. 6. X-ray Examination of Lattice Parameters and Defects In order to determine whether or not melting point depression is caused by incorporation of noncrystalline impurity in the crystalline lattice of isotactic polystyrene, x-ray line broadening measurements were made on bulk samples of isotactic/atactic polystyrene. The average values of the (110) peak half width and 26 Bragg angles for specimens of isotactic/atactic polystyrene are recorded in Table II. The average of the standard deviation for all specimens measured was 0.11~, which exceeds the maximum variation in the average values recorded for each sample. In another line broadening study, Buchanan and Miller [9] noted a 100% increase in the x-ray half width for the (110) peak upon stretching a bulk specimen of isotactic polystyrene at 1050C some 400% prior to isothermal crystallization at 1650C. Buchanan and Miller [9] used line broadening analysis of the (102) reflection to 0 calculate a crystallite size of 85-100A in the direction parallel to stretch.

43 Table II Half Width of the (110) X-ray Peak for Mixtures of IPS/APS Crystallization Tempera- Half Width Concentration ture 20 (110)** IPS 180~C 4.5-6.5 1. 67 IPS/APS Mw=4800*** 1800C 4.5-6.5 1.65~ IPS/APS Mw=411,000*** 1800~C 4.5-6.5 1.65~ IPS/APS Mw=900*** 175~C 4.5-6.5 1.72~ IPS/APS Mw=4800*** 180~C 4.5-6.5 1.75~ IPS/APS Mw=411,000*** 180~C 4.5-6.5 1.71~ IPS/APS Mw=900*** 1750C 4.5-6.5 1.69~ *Specimen partially melted and recrystallized for 24 hours. **The estimated experimental error based on a scanning rate of 0.5~ per minute and a chart speed of 1 inch per minute (10 boxes) per minute is approximately +.050 (+ 1 box). ***IPS/APS 20/80. D. Discussion 1. Kinetics of Spherulitic Growth Rate The experimental data for the spherulitic growth rate of the diluted isotactic/atactic polystyrene system can be used to evaluate the theoretical kinetics equation for diluted systems as developed by Gornick and Mandelkern [19] and Boon and Azcue [5]. A relationship between the growth rate, G(dil), of the diluted isotactic/atactic polystyrene mixtures, and the growth rate, G(undil), of the pure isotactic polystyrene system was developed in the introduction following the analysis of Boon and Azcue Cs5):

44 G(dil)u2exp[0.2 (Tm-T kn t2] G(undil) (13) where u2 is the volumetric fraction of isotactic polystyrene and Tm is the melting temperature of the diluted mixture. The assumptions upon which equation 13 is based are that the glass transition temperature, Tg, the melting temperature, Tm, and the pre-exponential constant, Go, are the same for both the diluted and undiluted systems. The experimental data was taken at the maximum crystallization temperature, Tx, of 180~C, and this temperature will be used in the evaluation of equation 13. Since it is known that the melting temperatures determined by optical microscopy are 3~ to 5~C below the values found by dilatometry, a value of 5130K, as determined by Dedeurwaerder and Oth [14] using dilatometric techniques will be used as the melting temperature of pure isotactic polystyrene. Substitution of these two temperatures into equation 13 gives: 2.71 G(dil)=u2 G(undil) @1800C (14) If we had used the melting temperature as determined by optical microscopy, the exponent of the u2 term in equation 14 would equal 2.935, a decidedly poorer value, as will be shortly indicated. Figure 13 shows a graphical comparison between the experimental data and equation 14. The experimental results clearly do not agree with the theoretical equation. Refinement of the melting temperature approximation to account for the maximum difference of the melting temperature between pure and diluted isotactic polystyrene (16~C) results in the following form of equation (14):

crf COMPARISON OF EXPERIMENTAL DATA TO THEORY Q)1.1 General Trend of Experimental Data m. I ve Equation:l4 1.0 Equation 15 TeTmperature. 1800 C 4)09 4C $ Direction of Trend at Lower Temperatures 0o.8 4 —) 0 0.7 0.5 0,23 ro 0.3 0.1~~~~~~~~~~~~~~~~~~~~~c r-t 0. 0 20 ~40 68 8010 0 ~~~~~~~~% Atactilc Polystyrene (APS) Figure 13

46 G(dil)=0.645 u271 GG(undil) (15) The factor 0.645 arises from separate evaluation of equations 6 and 10 by substituting Tm = 5130K and Tm = 497~K respectively for pure and diluted polystyrene with 80% atactic diluent molecular weight 4,800. As Figure 13 clearly shows, such a melting temperature refinement makes a correction in the wrong direction. At crystallization temperatures lower than 1800C, equation 13 and the experimental data of Keith and Padden [31] have opposite temperature coefficients, as indicated by the arrows in Figure 13. These results clearly demonstrate that the theoretical kinetics equation developed for diluted systems does not agree with the experimental data obtained for the isotactic/atactic polystyrene system over a wide range of diluent concentrations and molecular weights. Although Boon and Azcue [5] pointed out the effect of opposite temperature coefficients, they did not have sufficient data to conclude that the theoretical equation does not agree with experimental results. Data from the present study, as well as that from the work of Keith and Padden [31], clearly show the discrepancy between predictions and experimental data. 2. Growth Rate Phenomena The dependence of spherulitic growth rate on the average viscosity or molecular weight of the polymer system has taken the following form, according to Keith and Padden [31], and Hoffman and Weeks: G=(1/Mn)Y where M is the average number average molecular weight with y normally

47 assuming values O<y<l. If the behavior of the present system can be adequately described on the basis of normal viscosity or diffusional effects, then the appropriate values of y are 0.16 and 0.70 for 40/60 and 60/40 mixtures between atactic polystyrene, molecular weights 4,800 and 19,800, Figure 10. At the critical molecular weight for chain entanglements in polystyrene, 36,000 [15], y should become increasingly positive in proportion to the increase noted for the melt viscosity over the same molecular weight range, since aMW y However, we find that in the region of the critical molecular weight (between 19,800 and 51,000), y suddenly becomes negative: y = -3.0. This surprising growth rate behavior cannot be explained in terms of normal viscosity or diffusional effects. The increase in growth rate over this molecular weight range (19,800 to 51,000), nevertheless, can be explained by a model based on the phenomenological theory of Keith and Padden 130]. Morphological evidence [30] supporting this theory suggests that significant amounts of noncrystallizable rejected impurity are entrapped within the growing spherulites, as indicated by the increased "openness" of the texture of high impurity concentrations. The means by which entrapment occurs is not considered, although slow radial diffusion is the implied basis for impurity entrapment. At molecular weights above the critical molecular weight, chain entanglements are known to influence the viscous transport properties of a polymer melt. Thus, it would be reasonable to suspect that such entanglements might effectively tend to entrap rejected impurity molecules and inhibit radial diffusion

48 away from the growing crystal front. Such entrapment of rejected impurity might also cause crystalline lattice defects, although x-ray line broadening (Table II) and melting temperature depression (Figure 12), do not indicate such defects. Alternately, impurity entrapment might primarily involve self-entanglement rather than entanglement with crystallizing molecules. In either case, as the molecular weight of the rejected impurity is increased above the critical molecular weight, chain entanglements, according to the proposed model, would start to entrap impurity molecules and inhibit radial diffusion. As a result of this entrapment, the effective interfacial concentration of the crystallizable molecules would tend to increase, because the diffusion barrier generated by interfacial accumulation of rejected impurity would be reduced correspondingly by impurity entrapment within the growing spherulites. Since the growth rate is directly proportional to the interfacial concentration of the crystallizable constituent (equation 12), the growth rate would therefore be expected to increase over the range of the critical molecular weight (19,800 to 51,000). In the molecular weight range above 51,000 (Figure 10), the growth rates for the 60/40 and 40/60 mixtures are more or less uniformly depressed with increasing molecular weight. The slopes of the growth rate curves over this region are strikingly similar to the slopes of the same curves between molecular weights 4,800 and 19,800. This similarity suggests that in these two regions (4,800 to 19,800 and 51,000 to 1,800,000), the decreasing growth rate might be explained

49 in terms of the viscosity or diffusional effects of increasing dtluent molecular weights, although. the exact growth mechanism is not clearly understood. Depression of the growth rate for very low molecular weight diluents (Figure 10) is clearly not the result of increased molecular viscosity. Similar depression of the melting temperature, Tm, over the same molecular weight range (Figure 12), however, suggests that there might be a correlation between these two phenomena. At the maximum crystallization temperature, Tmax, a reduced temperature difference, Tm-Tmax, directly corresponds to a reduced rate of nucleation, and as a result, a reduced growth rate (equation 12). Since a reduced temperature difference, Tm-T max occurs -for 60/40 and 40/60 mixtures of atactic polystyrene diluent, molecular weight 900 (Figures 10, 12), one would also normally expect the growth rate to be reduced. In general, significant melting temperature depression can be caused by introduction of crystalline lattice defects, increased concentration of chain ends, or decreased polymerdiluent interaction free energy (increased solubility). X-ray analysis suggests that the addition of noncrystalline atactic polystyrene diluent affects neither the amount of line broadening nor the crystalline lattice spacings within the spherulitic structures. Apparently, the noncrystalline impurity is substantially rejected from the growing crystal lattice, since no lattice defects resulting from the incorporation of such impurity molecules can be experimentally detected by relative line broadening. This

50 result suggests that impurity entrapment occurs primarily by self-entanglement of impurity molecules and/or by entrapment in such a manner that the impurities do not affect the ordering of the chains in the crystal lattice. Melting temperature depression caused by the concentration of chain ends would suggest that there is a linear relationship between the inverse of molecular weight (concentration of chain ends), and the amount of melting temperature depression. Figure 14 shows quite clearly that there is no linear relationship between these two parameters. Flory and Huggins [41] proposed the following equation, derived from basic thermodynamics, for the melting temperature depression caused by the dilution of a polymer: 11 -..2 T- A AHuV1 [(l 1 —2) Xl(l (16) m m u where u2 is the volume fraction of isotactic polystyrene, Vu and V1 are the molar volumes of the repeat unit for isotactic and atactic polystyrene respectively; AHu is the heat of fusion per mole of repeat unit, and X1 is the isotacticatactic polystyrene interaction free energy. Other quantities being equal, a larger depression of the melting temperature should be noted with a good solvent (small X1) than with a poor one. The depression should also be greater for diluents of smaller molar volume. In the case of isotactic/atactic polystyrene mixtures, parameters R, Vu, V1, and AHu would all have constant values. Therefore, the amount of melting temperature depression would be dependent upon values of u2 and X1. Since similar concentrations of atactic

MELTING TEMPERATURE VS. CONCENTRATION OF CHAIN ENDS 40 20/80 *IPS/APS 20/80 (OIPS/APS 40/60 GIPS/APS 60/40 Parameter: Ratio of IPS/APS Concentration 40/6 30 0CJO3 II I / I: rH H l o20Figure 10-7 10-6 10-5 10-4 lo-, 1 0-2 1/MW IMW = Molecular Weight of APS) Figure 114

52 polystyrene have widely different temperature depressions (Figure 12), values of X1 must be dependent upon the molecular weight of atactic polystyrene diluent. Values of the optical melting point are not sufficiently accurate to give meaningful values of the X1 parameter, but they are accurate enough to indicate that the interaction free energy is considerably lower for atactic polystyrene molecular weight 900, than for higher molecular weight diluents. This result indicates that very low molecular weight atactic polystyrene diluents are better solvents than higher molecular weight atactic polystyrene. Thus, the observed melting point depression appears to be related to the solubility of isotactic molecules in atactic polystyrene diluent. E. Conclusions 1. For the isotactic/atactic polystyrene system, the dependence of growth rate on temperature is very similar to that for the undiluted polymer. 2. Dilution of isotactic polystyrene (molecular weight 550,000) with noncrystallizable atactic polystyrene (for particular molecular weights ranging from 2,030 to 1,800,000) causes a uniform depression of the resulting spherulitic growth rate with the optimum growth rates occurring in about the same temperature range as for undiluted isotactic polystyrene (178~ + 2~C). 3. For a 20% dilution with atactic polystyrene, the growth rate is independent of impurity molecular weights up to 411,000.

53 4. For 40% and 60% dilutions, the dependence of the growth rate on diluent molecular weight can be described by normal viscosity effects for impurity molecular weights ranging from 4,800 to 19,800 and from 51,000 to 1,800,000. In between these two ranges, anomalous growth rate behavior suggests that molecular chain entanglements play an important role in determining the interfacial concentration of rejected impurity. For very low molecular weight diluents (900; 2,030), the growth rate is depressed by the corresponding depression of the crystalline melting temperature over the same molecular weight range. 5. The optical melting temperature range for mixtures of isotactic and atactic polystyrene is unaffected by high concentrations of impurity having molecular weights greater than 19,800. However, the melting temperature is rapidly depressed for high concentrations of very low molecular weight atactic diluent. 6. The measured x-ray line broadening for the (110) peak of crystalline isotactic polystyrene is not affected by the addition of high concentrations of various molecular weight atactic polystyrene diluents. 7. The theoretical growth rate equation derived for diluted systems does not describe the experimental data for the isotactic/atactic polystyrene system at 1800C.

CHAPTER IV CRYSTALLIZATION OF ISOTACTIC POLYSTYRENE/PLASTICIZER BLENDS A. Introduction Spherulitic crystallization from mixtures of crystallizable polymer and low molecular weight plasticizer has received very limited attention, compared to the numerous studies on spherulitic crystallization of homopolymers. Mixtures of polymer and low molecular weight plasticizer are often commercially important systems because the mechanical properties of some polymers can be favorably altered by the addition of plasticizer [45]. Therefore, it was decided to examine the influence of di-methyl and di-decyl phthalate, as well as benzophenone on the spherulitic crystallization kinetics of isotactic polystyrene and to compare these results to those obtained in Chapter III. We also determined the threshold crystallization temperatures for mixtures of isotactic/atactic polystyrene as well as isotactic polystyrene/benzophenone. The threshold crystallization temperature (TT) is the lowest temperature at which the onset of crystallization can be detected. This temperature gives us a relative measure of the magnitude of the viscous transport term in the theoretical growth rate equation. 54

55 Boon and Azcue [5] recently reported that at a fixed temperature, the spherulitic growth rate from mixtures of isotactic polystyrene and benzophenone is linear over a wide range of crystallization temperatures. The overall linear growth rate reaches a maximum at a crystallization temperature which is dependent on the benzophenone concentration, and is uniformly depressed at temperatures above and below the optimum crystallization temperature. The resulting growth rate curves are generally gaussian in shape and similar to the growth rate curve found for the undiluted homopolymer system. The growth rate maximas for each concentration ratio shift to lower crystallization temperatures, the amount of shift being proportional to the amount of benzophenone added to the mixture. Furthermore, the maximas of the growth rate curve first increase with the increasing benzophenone content up to about 20% dilution and are gradually depressed at higher dilutions. Boon and Azcue [51 established that there.is a qualitative agreement between the experimentally measured growth rate and the theoretical rates predicted by equation 12 in the introduction of Chapter III particularly for benzophenone concentrations less than 20%. The agreement is increasingly poorer for higher concentrations of benzophenone. The shift of the crystallization range of isotactic polystyrene/benzophenone mixtures to lower temperatures is primarily the consequence of the depression of the melting temperature T and the glass transition temperature T with increasing benzophenone concentration. Boon and Azcue [51]

56 correlate this shift along the temperature axis with a reduced temperature parameter 9 which is defined as: O = (T-Tg)/(Tm-Tg). The growth rate maximas for all concentrations of benzophenone occur at about 9 = 0.57. Mandelkern [42] studied the crystallization kinetics of mixtures of linear polyethylene and a-chloronaphthlene over a wide range of concentrations near the melting temperature by means of dilatometric techniques. As the pure polymer is diluted, the shapes of the isotherms change and are no longer superimposable over the complete transformation range. In the range of very dilute concentrations of polyethylene, 1% to 0.01%, the shapes of the isotherms no longer change, indicating a constant free energy for the formation of a critical size nucleus over the range. However, over the same concentration range, 1% to 0.01%, the crystallization rate is still dependent on concentration, since it continues to be depressed with decreasing polymer concentration. Other studies on the crystallization kinetics of polymer diluent mixtures were reported by Lunak and Bohdanecky [35] and Holland and Lindenmeyer [24]. These studies will not be reviewed in detail because they considered different systems than the one of present interest. Mandelkern [41] also studied the effect of low molecular weight plasticizer on the melting temperature depression of diluted linear polyethylene. Melting temperature depression can be quantitatively described by the Flory-Huggins equation (equation 16), which is based on the thermodynamics

57 of ideal solution behavior using a correlation factor X1 called the polymer-diluent interaction parameter. Other quantities being equal, a larger depression of the melting temperature, Tm, should be observed with a good solvent (smaller values of X1), than with a poor one. The size of the diluent molecule should also affect Tm, the depression being predicted to be greater for diluents of smaller molar volume.. Experimental results for the melting temperature depression of linear polyethylene with n-butyl phthalate, O-nitrotoluene, a-chloronaphth4an~f) and tetralin show good agreement with the predictions of the Flory-Huggins equation. Poorer solvents such as n-butyl phthalate and O-nitrotoluene depress Tm much less than similar concentrations of better solvents, a-chloronaphthaenee- and tetralin. Other important properties of polymer-diluent systems were examined by Taeger and Suvorova [57], who studied the effect of di-methyl and di-decyl phthalate on the viscosity, glass transition temperature, and phase separation temperature of plasticized polystyrene mixtures. The phase separation temperatures and solution viscosity are uniformly higher for mixtures of polystyrene and di-decyl phthalate than for similar mixtures using di-methyl phthalate, because di-decyl phthalate is both a poorer solvent and has a higher molar volume (molecular size) than does di-methyl phthalate. This behavior is consistent with the results usually found when comparing similar diluents having widely different molecular weights. Polystyrene mixtures with di-decyl phthalate, however, have a surprisingly lower glass transition temperature

58 than comparable mixtures with di-methyl phthalate. Taeger and Suvorova [57] explain this behavioral contradiction in terms of a proposed model based on the existence of a supermolecular structure in the polystyrene phase. According to their explanation, a rather poor high molecular weight solvent such as di-decyl phthalate, which cannot successfully penetrate and dissolve these structures, absorbs onto the surface and lubricates the movement of one structure with respect to its neighbors. Good solvents such as dimethyl phthalate, however, penetrate and partially dissolve the polystyrene structure, thereby inhibiting the translational movement of these structures and causing a higher glass transition temperature. In viscosity studies, these structures are apparently destroyed by concentrated shear stresses, and the molecular volume of the diluent, rather than the diluent solubility, has the predominate influence on the resulting properties. B. Experimental 1. Preparation The characteristics of all materials used in this study are discussed in Chapter II. Dilute 0.4% solutions of both isotactic polystyrene and low molecular weight plasticizer in chloroform were made by completely dissolving a weighed amount of polymer or plasticizer in a measured volume of boiling chloroform. After cooling to room temperature, predetermined ratios of clear 0.4% isotactic and 0.4% plasticizer solutions were pipetted together to make

59 up a clear 0.4% mixed solution of isotactic polystyrene/plasticizer in chloroform. Thin films approximately 601Q thick, as determined by density calculations based on a polystyrene/plasticizer density of ~l.0 gram/cm3, were prepared by casting from the 0.4% mixed solution onto thin cover glasses at room temperature under a cool air fan. These films were then exposed to the ambient atmosphere at room temperature for 12 hours to allow complete removal of residual chloroform solvent. Prolonged exposure for several days under these conditions would result in significant evaporation of volatil benzophenone and di-methyl phthalate plasticizers. Samples suitable for a threshold crystallization study in the electron microscope were prepared by casting thin 0o films (~ 500A thick) from 0.2% benzene solutions of isotactic polystyrene/benzophenone (ranging from 100/0 to 50/50) onto surfaces of freshly cleaved mica. Similar thin films of isotactic/atactic polystyrene molecular weight 4,800 were prepared in exactly the same way. The pieces of mica were then sealed in pyrex glass tubes and annealed at various temperatures in silicone oil temperature control baths (+.50C) for 24 hours. After thermal treatment, the mica pieces were removed from the pyrex tubes and the crystallized thin films were floated off the mica onto a water surface and picked up on 200 mesh copper grids for subsequent study in the electron microscope. 2. Measurement of Growth Rate The optical microscope assembly used in this study is

60 the same as described in Chapter III. Thin films of isotactic polystyrene/plasticizer, sandwiched between two cover glasses, were heated on the microscope hot stage to 250~C for five minutes before cooling at the desired isothermal crystallization temperature, ranging from 1030C to 1850C. Spherulitic crystals generally started to nucleate and grow before the hot stage had finished cooling to the isothermal crystallization temperature, cooling at a rate of 20~C per minute. Accordingly, it was particularly important to view just one field of growing spherulites in order to measure the growth rate of crystals, most of which are generated at about the same time. Within this given field of view, the crystals were very uniform in size and the resulting growth rate was determined by measuring the size of a number of spherulites as a function of time. Data scatter caused by variations in concentration due to excessive loss of volatile plasticizer were minimized by using fresh films for each growth rate measurement. Spherulitic growth rates were measured in thin films having various concentrations of di-methyl phthalate or di-decyl phthalate added to isotactic polystyrene of molecular weight 550,000. Thus the growth rate, G, depends on the type and the concentration of the added plasticizer as well as the temperature of crystallization. Data scatter from duplicate measurements indicates that the growth rate under normal conditions is reproducible to within + 5%.

61 3. Melting Temperature Measurement Thin films of isotactic polystyrene/plasticizer, crystallized at a variety of temperatures, were individually heated on a Kofler hot stage at a rate between 0.50C and 1.0~C per minute until the crystalline birefrigence totally disappeared. The temperature at which the birefringence disappeared was designated as the optical melting temperature. The same films were not recrystallized for a second melting point determination because excessive loss of volatile plasticizer significantly affected the concentration ratio and hence the observed melting temperature. Thin films of isotactic polystyrene/diluent, prepared for the threshold crystallization study, were either shadowed with platinum at 300 or vertically decorated with gold at 90~ and then examined in a JEM-6A electron microscope for the first signs of crystalline structure. The characteristics of the electron microscope and discussions of the shadowing and decoration techniques are recorded in Chapter II. C. Results 1. Morphology Electron microscope examination of crystallized thin films of isotactic polystyrene/plasticizer shows that the spherulites are made up of fibrous structures similar to those obseived in spherulites grown from mixtures of isotactic/atactic polystyrene, Figure 15 on page 74. Under crossed polarizers in the optical microscope, the crystals are birefringent

62 with a normal maltese cross extinction pattern. The spherulites develop from small 1 to 2 micron symmetrically round fibrous nuclei. In the plasticized system, there is no evidence of the hexagonal hedrite nuclei observed under certain conditions in mixtures of isotactic/atactic polystyrene. Spherulites grown from plasticized mixtures have the same fine fibrous texture as spherulites of. pure isotactic polystyrene, This is in agreement with the observations of Boon and Azcue [5] for spherulites grown from mixtures of isotactic polystyrene/benzophenone. 2. Effect of Temperature The spherulitic growth rate of isotactic polystyrene/plasticizer is linear with respect to time over a wide range of temperatures, as indicated in an example shown in Figure 16. The growth rate curve for this particular mixture reaches a clear maximum at approximately 160~C and is depressed at both higher and lower temperatures, Figure 17. In general, the growth rate becomes non-linear only under the conditions in which the diffusion of the crystallizable constituent becomes rate controlling, which starts to occur when spherulites impinge upon one another. There is no evidence of a non-linear growth rate caused by increased plasticizer concentration at the growing crystal interface. 3. Effect of Diluent Concentration As noted in Figure 17, when the concentration of di-methyl phthalate is increased, the maximas in the growth

63 rate curves shift to lower temperatures, the amount of shift being proportional to the amount of plasticizer added. This result is in general agreement with the effects reported by Boon and Azcue [5] in the isotactic polystyrene/benzophenone system. In addition, it should be noted that the growth rate maximas for the diluted di-methyl phthalate system are substantially higher than the maxima for the pure system and reach an overall optimum growth rate at about 20 to 30% dilution, Figure 20. This optimum is also observed when the maximum growth rates are plotted as a function of concentration in the isotactic polystyrene/benzophenone system, Figure 18. In this case, the overall optimum growth rate, according to the data of Boon and Azcue [5], occurs at 20% dilution and is in general agreement (better than 10%) with data taken on the present equipment at the same temperatures using mixtures of isotactic polystyrene (molecular weight 550,000) and benzophenone. 4. Effect of Diluent Type When di-decyl phthalate is added to isotactic polystyrene, the resulting radial growth rate curves do not have the same characteristics as observed when benzophenone and di-methyl phthalate are used as the added diluent, Figure 19. As the concentration of di-decyl phthalate increases, the growth rate maximas shift to lower temperatures in accordance with the amount of plasticizer added. However, above 20% dilution the growth rate maximas remain fixed at about 145~C for dilutions up to 80% plasticizer, while in

SPHERULITIC GROWTH OF MIXTURES OF ISOTACTIC POLYSTYRENE/PI4METHYL ~~30 1- ~~PHTHALATE 90/10 30 25 G) cY \=160~ C * H2 0 15|1300C C/D 10 0 5 10C 15 20 25 30 35 4'0 45 50 Time, minutes Figure 16

SPHERULITIC GROWTH RATES FOR MIXTURES OF ISOTACTIC POLYSTYRENE AND DI-METHYL *HTHALATE 0.7 *.I PS OK 0PS/DMP 90/10 0.6 4 a IPS/DMP 80,2 OIPS/DMP 60/40 0.5)- e II S\ ISIIPS/DMP ~0/4' 0.e 4-3 0.3 a 0C 0.2 100 110 120 130 14o 150 160 170 180 190 200 Temperature, OC Figure 17

66 the case of benzophenone and di-methyl phthalate, they continue to shift to lower temperatures, Figure 17. The magnitude of the overall optimum growth rate at comparable concentrations (20%) is considerably higher for the di-decyl phthalate system than for the benzophenone and di-methyl phthalate systems; 1.1 microns per minute as compared to 0.76 and 0.73 microns per minute respectively, Figure 20. In addition, it should be noted that the sharp overall maximum at 20% dilution for the benzophenone system is considerably more diffuse for the di-methyl phthalate system, and is actually quite dispersed between 20% and 40% dilution for the di-decyl phthalate system. These observations will be examined in more detail in the discussion section. 5. Melting Temperature Determination Figure 21 shows the relationship between the optical melting temperature and the concentrations of plasticizer in crystallized thin films of isotactic polystyrene/di-methyl phthalate. In general, the melting temperature is depressed with increasing concentration of added impurity. Also indicated in Figure 21 is the effect of added plasticizer on the glass transition temperature of the mixture, based on the dynamometric balance data of Taeger and Suvorova ~-573. The temperature range between the optical melting temperature and the glass transition temperature increases with increasing plasticizer concentration because the glass transition temperature is depressed more rapidly

MAXIMUJM SPHERUJLITIC GROWTH{ RATES FOR MIXTURES OF ISOTACTIC POLYSTYRENE AND DENZOPHENONE 2.0','H rd 1.6C OH'H rd O~~~~~~~~(IPS/Benzophenone IPS M 55010-00 ~~7 1~~~~d~~ B~IPS/Ben'Zophenone Boon- and'Azcu~e E51 Cd.'H ro 2=, ~F~~ ~~ 10 20 30 4 % Benzophenone Figure 18

*;. SPHERULTTIC GROWTH RATE FOR-MIXTURES OF ISOTACTIG PaLYST TE NE AND DI-DECYL PHTHALATE 1. * IPS 1.d- \`O IPS/DDP 90/10 IPS/DDP 80/20 g0. 0 *IPS/DDP 70/30 rOIPS/DDP 60/40 QIPS/DDP 4110/ 0. 3 2!-8 @PIFS/DDP 20/80 4-)0 0. o0 o5 0. 0, rOl 0, 0.2 r 0. A joG 110 120 130 140 150 160 170 180 190 200 Temperature, 0C Figure 19

MAXIMUM SPHERULITIC GROWTH RATES OVR MIXTURES OF IPS/APS, IPS/BENZOPHENONE, IPS DMP, AND IPS/DDP eIPS/Di-Methyl Phthalate -1 01. 0 IPS/Di-Decyl Phthalate EIPS/Benzophenone m:IPS/APS M=4,800oo 0.9* OIPS/APS Mw=19,800'.BC 7 g IPS/APS Mw=51,001 4 0.7 o.] 0 6_ 0.2 3_ \ O. 0 a - 0.0 0.2 0.4 0.6 0.8 1.o % Diluent Figure 20

70 than the melting temperature. This trend was also noted by Boon and Azcue [5] for the isotactic polystyrene/benzophenone system and served as an important basis for their explanation of the overall growth rate maxima observed in Figure 18. The relationship between di-decyl phthalate concentration and the optical melting temperature-of crystallized thin films is shown in Figure 22. Also indicated in this figure is the effect of added plasticizer on the glass transition temperature of the mixture, based on the data of Taeger and Suvorova [57]. Although the melting temperature is consistently depressed with increasing plasticizer concentration, the glass transition temperature appears to have a minimum around 20% to 30% dilution and actually increases slightly for higher dilutions. Taeger and Suvorova attribute this anomalous behavior to macroscopic (> 2000A).phase separation which starts to occur at about 10% dilution. For this system, the temperature range between the melting temperature and the glass transition temperature sharply increases up to 20% dilution where it levels off and slightly diminishes for higher dilutions. In the two systems studied (di-methyl and di-decyl phthalate) the greater melting temperature depression (at comparable concentrations) is observed for di-methyl phthalate, which also happens to be the better solvent, according to the phase separation temperatures reported by Taeger and Suvorova [57]. This result is consistent with the findings of Mandelkern [42] who noted that (at comparable concentrations)

25" "-4:3 T MELTING TEMPERARTURES AND GLASS URNITO 4-) Cd 2d0 m ~~~~~~~TEM:PERATURtS FOR MIXTURES OF ISCACI kP~~~~O L Y S T Y R E M EPLYTYEN AND DI-MVETHYL PHTHAIAAT EF: 150 QOptical Melting Temperatures 4-) e Glass Transition* Temperature Data cei by Taeger and Suvorova [571 Fe100.r 50 0) 4.) Tg E) E-50 -H 4-3 0 20 4 0 60 80o % Di-Methyl Phthalate Figure 21

0 0 k 200 S MELTING TEMPERATURES AND GLASS TRANSITION TEMPERATURES FOR MIXTURES 0E: OF ISOTACTIC POLYSTYRENE AND DI-DECYL PHTHALATE'Ht 150 O Q 0Optical Melting Temperatures e Tg Data by Taeger and Suvorova [57] c-i H 100 C) F-i 50 0 8 ~~~~~~~~~~~Tg F —i ho: -50 4-3r 0 20 4 0 60 8o 100 % D!-Decyl Phthalate F4gure 22

73 low molecular weight diluents which are good solvents depress the melting temperature of linear polyethylene more than diluents which are poor solvents. 6. Threshold Crystallization Study The threshold crystallization temperature is characterized as the annealing temperature at which the first crystalline structures appear from the glassy amorphous state after a long annealing time. For the present study, a 24 hour annealing time was judged to be sufficiently long to induce crystallization at threshold temperatures. At this characteristic temperature molecular chain segments start to orient and pack together in such a way as to form a few discrete bundles of crystalline fibrils fanning out from a central nucleus region, Figure 23. Less than 2~C above this temperature, these discrete bundles develop into extensive spherulitic structures while slightly below this threshold temperature (~ 20C), there are no signs of crystalline morphology whatsoever. The first appearance of fibril structure characteristic of the onset of polymer crystallization is most easily detected by electron microscope examination of the surfaces of platinum shadowed or gold decorated thin films, rather than by electron diffraction, which generally shows an -amorphous pattern. Values of the threshold crystallization temperature for mixtures of isotactic/atactic polystyrene are recorded in Table III. The glass transition temperature, Tg, and melting temperature,.Tm, values are based on the

74 r:i~~................!ii'::': Figure 15: Isotactic polystyrene/benzophenone 50/50, crystallized from the glassy amorphous state at 406. Platinum shadowed at 30. Figure 23: Isotactic polystyrene/benzophenone 80/20, crystallized from the glassy amorphous state at 82C. Gold decorated at 90.

75 results of Boon and Azcue [5], while the threshold crystallization temperature, TT, values were experimentally measured to within + 20C. Analysis of these results in terms of the equations describing the theoretical growth rate kinetics will be considered in the discussion section. Table III Threshold Crystallization Temperatures for Mixtures of IPS/Benzophenone and IPS/APS Concentration Tg, TmOC T, OC TT-Tg IPS/benzo. 50/50 -15 180 30 46 IPS/benzo. 60/40 - 2 190 42 46 IPS/benzo. 65/35 8 196 52 44 IPS/benzo. 70/30 16 201 63 47 IPS/benzo. 75/35 26 208 75 49 IPS/benzo. 80/20 36 213 82 46 IPS/benzo. 90/10 58 227 98 40 IPS/APS Mw 4,800 10/90 - - 104 - IPS/APS Mw 4,800 50/50 85 230 105 20 w IPS/APS Mw 4,800 100/0 85 240 113 28 *Data from Boon and Azcue [5] **Data fromDedeurwaerder and Oth [14] D. Discussion 1. Growth Rate Phenomena The spherulitic growth rate in diluted systems is dependent upon the melting temperature, Tm, and the glass transition temperature, Tg, of the mixture as well as the

76 concentration of the crystallizable constituent, u2, according to equation 10. Dilution of isotactic polystyrene with di-methyl or di-decyl phthalate causes a substantial depression of both the melting and glass transition temperatures, Figures 21 and 22. At a given crystallization temperature, depression of Tg increases the molecular mobility which leads to a more rapid spherulitic growth rate. However, simultaneous depression of the melting temperature reduces the driving force for secondary nucleation (or growth rate), thus tending to reduce the increase of growth rate expected due to depression of the glass transition temperature. The third effect of dilution is to depress the rate of secondary nucleation (or growth rate) by reducing the probability of interaction between crystallizable polystyrene molecules. At a given temperature and concentration, the interaction of these three effects (Tm, Tg, and u2) principally determines the magnitude of the spherulitic growth rate, resulting in the behavior noted in Figures 17 and 19. Shifting of the maximas of the growth rate curves, Tmax, to lower temperatures with increasing plasticizer concentration can be correlated with melting temperature,, Tm, depression over the same range of concentrations, si-1: the maximas occur at temperatures ranging from 0.84 Tm to 0.89 Tm, Table IV. This correlation is in general agreement with the results iound for most crystallizable polymers [41]. The fact that the growth rate maximas remain fixed at about 145~C (~+ 2~C) for di-decyl phthalate concentrations above

77 20% (Figure 18) is reflected in the levelling off of the melting temperature curve over the same range of concentrations, Table IV. Table IV Relationship Between T and Tm max for Mixtures of IPS/Plasticizer Concentration T rC T,IC T /Tr, (~K/~K) IPS 233.0 178.0 0.890 IPS/DMP 90/10 219.0 157.0 0.875 IPS/DMP 80/20 205.0 145.0 0.875 IPS/DMP 70/30 194.5 122.0 0.844 IPS/DDP 90/10 218.0 160.0 0.872 IPS/DDP 80/20 212.0 143.0 0.858 IPS/DDP 70/30 208.0 147.0 0.871 IPS/DDP 60/40 208.0 147.0 0.871 IPS/DDP 40/60 206.0 145.0 0.870 IPS/DDP 20/80 202.5 145.0 0.878 For the isotactic polystyrene/benzophenone system, Boon and Azcue [5] attribute the increase in spherulitic growth rate for dilutions below 20% to the widening temperature difference between the melting and glass transition temperatures, i.e., the effect due to the increased molecular mobility outweighs the depressive effect due to the reduced rate of nucleation. This growth rate reaches a maximum at 20% dilution (Figure 18) and is depressed at higher benzophenone concentrations because the effect of the reduced

78 probability of molecular interaction overwhelms other growth factors, thereby causing a marked depression in the overall spherulitic growth rate. Presumably similar forces are operative in the isotactic polystyrene/di-methyl or di-decyl phthalate systems, although in the case of di-decyl phthalate, the curve shows a characteristic flattening of the growth rate over a wide diluent concentration range (20% to 40%), Figure 20. This tendency for curve flattening at the maximum growth rate cannot be accounted for by the behavior of the glass transition and melting temperatures because these parameters remain essentially invariant for diluent concentrations exceeding 20%, Figure 22. Under these conditions (invariant T and Tg), if the isotactic m g polystyrene/di-decyl phthalate system forms a "homogeneous" phase, then the probability effect of dilution would tend to cause a monotonic (not necessarily linear) depression of the growth rate in direct proportion to the increasing diluent concentration. This type of monotonic depression is observed for increasing concentrations of diluent in the isotactic/atactic polystyrene system, Figure 8. For the present system, however, it appears that bhisotactic polystyrene is not being uniformly diluted over the concentration range in question (20% to 40% di-decyl phthalate), Figure 20, since the growth rate is not being depressed. This result suggests that the crystallizing mrelt might actually form a microscopic two phase mixture over this concentration range. A model based on phase separation between an isotactic polystyrene rich phase and a di-decyl

79 phthalate phase is consistent with the observation that the crystalline phase is apparently not being diluted by the addition of di-decyl phthalate over this particular concentration range. For diluent concentrations exceeding 40%, it is most likely that some phase separation still exists, as indicated by the lack of melting temperature depression (Figure 22, also see page 72). There is no evidence suggesting macroscopic (> 2000A) phase separation in any of the thin films studied, since no signs of optical turbidity were observed in films diluted with either atactic polystyrene or low molecular weight plasticizer at the isothermal crystallization temperature prior to crystallization. The crystallization temperature for mixtures diluted with either di-methyl or di-decyl phthalate were at least 1000C above the macroscopic phase separation temperatures (ranging from 0~ to 80~C) reported for these respective mixtures by Taeger and Suvorova [57]. However, there are some indications that a microscopic phase separation might exist in these systems, particularly the di-decyl phthalate system. As mentioned earlier, Taeger and Suvorova [57] proposed the existence of a structured polystyrene phase as being a reasonable model to explain why polystyrene/di-decyl phthalate mixtures have a lower glass transition temperature than comparable concentrations of polystyrene/di-methyl phthalate. Structures which might be regarded as phase separation structure have actually been observed by Gezovich and Geil [18] in poly(vinyl chloride) plasticized with di-octyl phthalate and chlorinated parafin.

80o They reported spherical structures on the order of 0.1 to 5.0 microns diameter in bulk samples brittle fractured under vacuum at liquid nitrogen temperatures and then ion etched. In addition, they also detected such structures by small angle x-ray scattering methods. A direct comparison of the spherulitic growth rates for isotactic polystyrene diluted with a 20% concentration of either atactic polystyrene or low molecular weight plasticizer suggests that there might be a difference in the microscopic structure of these two systems, Figure 24. The growth rate for the atactic polystyrene system is only slightly affected by the molecular weight of the atactic diluent. However, the system similarly diluted with comparable concentrations of plasticizer, particularly di-decyl phthalate, shows a spectacular 100% to 200% increase in growth rate, Figure 24. At this point it might be argued that the difference in the growth rate for these two systems is essentially due to the difference in the molecular weight of the diluent. In order to clarify this issue, it is necessary to extrapolate the behavior of the isotactic/atactic polystyrene system to very low diluent molecular weights (- 450). An extrapolation based on the Hoffman and Weeks [23] relationship (Ga(l/MK)Y), giving y the maximum value of 1.0, results in a 20% increase in growth rate between diluent molecular weights 900 and 45o.

1.2 -- COMPARATIVE SPHERULITIC GROWTH RATES FOR MIXTURES OF IPS/APS AND IPS/PLASTICIZER 1.1 - U OIPS/Benzophenone Mw = 182 eIPS/Di-Methyl Phthalate Mw = 194 1.0 CNIPS/Di-Decyl Phthalate Mw = 446 QIPS/APS M = 900 O.91- *IPS/APS Mw = 2,030 |)IPS/APS Mw = 4,800 0.8 L OGIPS/APS Mw = 19,800' r.8 o EIPS/APS Mw = 51,000 > 0.7L e XIPS/APS Mw 1,800,000 0.7 IPS/Diluent are in ratio of 80/20 ( 0.6 o 0.5 0.4 *\ Gm(1/Mn)1'O ~ -0.3 -- - - - " 0.2 0.1 E 0.1 4 5 6 7 8 9 10 11 12 13 14 15 L n Diluent Molecular Weight Figure 2 4

82 This extrapolation obviously does not explain the spectacular increases in growth rate observed for the mixture diluted with di-decyl phthalate. It is now apparent that the difference in the behavior of these two systems (isotactic/atactic polystyrene and isotactic polystyrene/plasticizer) is related to the microstructure of the mixture. As mentioned earlier, the system diluted with atactic polystyrene apparently forms a "homogeneous" phase, since the resulting growth rate is monotonically depressed with increasing diluent concentration. However, the system diluted with di-decyl phthalate, at least over certain concentration ranges, apparently forms a two phase mixture. If we assume that in the latter case we actually do have a two phase system, then how can we explain the spectacular increase in growth rate associated with the plasticized system? If we consider the terms in equation 2, it is apparent that upon the creation of a two phase system either the viscous transport term, ED, or the free energy of nucleus formation, AF*, must be significantly altered. We will try to distinguish between these two possibilities in the following discussion of the activation energy for viscous transport. B. Activation Energy for Viscous Transport At threshold crystallization conditions, the spherulitic growth rate is principally controlled by the activation energy required for viscous transport across the liquidnucleus interface, ED, equation (2). Thus, the results of

83 the threshold crystallization study for mixtures of isotactic polystyrene/benzophenone and isotactic/atactic polystyrene maybe analyzed by the theoretical growth rate equation 10 developed from equation 2 in the introduction of Chapter III with specific terms evaluated by Boon and Azcue [5]: G(diL) = 9.1 106 -208 0' G(dil) = 9.1x106u2exp[75+T-Tg] (14) exp-279Tm + 0.2Tmknup /min. T(Tm-T) (Tm-T) where u2 and T are the volumetric concentration of isotacm tic polystyrene and the melting temperature of the mixture respectively. In this form they assume that the empirical constant, C2 = 750K, in the viscous transport activation energy term (I) has the same value for both pure and diluted isotactic polystyrene systems. We also assume that the value of G is constant (page 13) and that the first approximation 0 of the Af term (equation 9) is valid for the temperature u range being considered [56]. The spherulitic growth rate for threshold crystallization, GT, can be experimentally measured from electron micrographs of thin films crystallized at the threshold temperature for 24 hours, Figure 23. The values of the threshold growth rates, GT, for all mixtures are recorded in Table V.

84 Table V Threshold Growth Rates for Mixtures of IPS/Benzophenone and IPS/APS Spherulite Concentration Sizes,. GT, i/min G/9.1X106u2 (I+II+III) IPS/Benzo. 90/10 1.09 7.57X10-4 9.25X10'-1 -23.00 IPS/Benzo. 80/20.59 4.10X10-4 5.63X10-ll -23.60 IPS/Benzo. 75/25.71 4.95X10-4 7.26X10-11 -23.35 IPS/Benzo. 70/30.87 6.04X10-4 9.48X10-1' -23.05 IPS/Benzo. 65/35.84 5.83X10-4 9.86X10-11 -23.00 IPS/Benzo. 50/50.68 4.72X10-4 10.41X10-11 -22.95 IPS/APS 100/0.75 5.18X10-4 5.72X10-1 -23.55 IPS/APS 50/50.62 4.34Xl0-4 9.52X10-1' -23.05 The following analysis will be based on the threshold growth rate for the isotactic polystyrene/benzophenone 80/20 mixture, GT = 4.10X10-4/min (Table V). If the value of the growth rate, GT, for this particular mixture is substituted into equation 14, we have the following result: 4.10X10-4=9.1X106 (0.8)exp(I)exp(II+III) (15) 5.63X10-1 =exp(I+II+III) -23.60= (I+II+III) (16) The complete solution of equation 16 requires a trial and error procedure based on the assumption of various crystallization temperatures. The following table indicates the approximate values of the exponential terms of equation 14 as a function of various assumed threshold crystallization temperature, TT, using Tm = 213~C and Tg = 36~C (Table III).

Assumed Threshold Temperature (OK),TT I II III (I+II+III) 325 -22.85 -2.60 -.135 -25.58 330 -21.85 -2.63 -.139 -24. 42 335 -20.60 -2.68 -.145 -23.42 A threshold crystallization temperature of 334~K(610C) satisfies equation 16 for the isotactic polystyrene/benzophenone 80/20 mixture. However, the experimentally measured threshold temperature, TT, for this particular mixture is 820~C (Table III), some 210C higher than indicated by an analysis based on the theoretical kinetics equation. This discrepancy can be resolved by adjusting the value of C2 while leaving C1 invariant in the activation energy viscous transport term (I) in equation 14. This constant can be reevaluated by recalculating the values of exponential terms II and III for a crystallization temperature of 82~C and substituting the results into equation 16: I-2.920-.165=-23.60 (17) I=-20.520 -2080 -20.52 (18) C2+T-Tg According to equation (18), when TT = 82~C and Tg = 36~C (Table III), the recalculated value is C2 = 55.50K. If we use the third approximation for the Afu term (d(AC )/dT=constant) [56] in equation 14, the resulting value of C2 following the same type of analysis would be C2 = 57.7~0C. Thus, we can see that the calculated value of C2 is relatively independent of the assumed form of the free energy of fusion.

86 A question might be raised at this point as to whether this new value of C2 is also valid for other concentrations of benzophenone in isotactic polystyrene. Calculated values of the threshold crystallization temperature using equation 18, with C2 = 55.50K, and data in Table III are compared to the experimentally measured values in Table VI. Examination of Table VI shows that the value of C2 = 55.5~K, is independent of the benzophenone concentrations exceeding 20%. This result is also reflected in the relative independence of the temperature difference between the threshold crystallization temperature and the glass transition temperature, Tx-Tg, (Table III), as well as the threshold growth rate, GT, (Table V) to the respective benzophenone concentrations. The value of the exponential term (I) for the activation energy for viscous transport at threshold crystallization may also be calculated for the pure isotactic polystyrene system (Table V), using values of Tm and Tg as recorded in Table III. I-2.53-0.00=-23.55 (19) I=-21.02 If we use the appropriate value of the activation energy term (I) together with the experimentally measured threshold crystallization temperatures (Table III), we can calculate the values of the activation energy empirical constant, C2, using Tg = 85~C. C+-208 = ~-21.02 (20) C2+T-Tg

87 Table VI Threshold Crystallization Temperature, Experimental and Calculated Concentration C Cal **C2 ~0K CK IPS/Benzo. 50/50 30 31.0 55.5 6.5 IPS/Benzo. 60/40 42 43.8 55.5 6.5 IPS/Benzo. 65/35 52 53.7 55.5 6.5 IPS/Benzo.70/30 63 62.2 55.5 6.5 IPS/Benzo. 75/25 75 72.0 55.5 6.5 IPS/Benzo. 80/20 82 82.0 55.5 6.5 IPS/Benzo. 90/10 98 98.0 62.0 12.8 IPS/APS M =4,800 10/90 104 w IPS/APS M =4,800 50/50 105 105.0 79.2 33.2 w IPS/APS Mw=4,800 100/0 113 113.0 71.2 25.4 * +20C. **Calculated according to the values of Go and C used by Boon and Azcue [5]. ***Calculated for values of G, C and K reported by Suzuki and Kovacs [56]. 0

88 The values of the empirical constant, C2, are 71.2~K and 79,20K for the IPS/APS 100/0 and IPS/APS (Mw=4,800) 50/50 mixtures respectively. These values are in general agreement with the value of C2 = 75.0~K, reported by Boon and Azcue [5] for the pure isotactic polystyrene system. If we follow a similar analysis using the values of the empirical constants G0, Ci=l.5cal./mole, and K = 4boauae/kAHf in equation 14 as determined by Suzuki and Kovacs [56] for the isotactic polystyrene system, then C2 = 6.50K for the benzophenone mixtures and C2 = 25.4~K and 33.2~K for IPS/APS 100/0 and IPS/APS (M =4,800) 50/50 mixtures respectively, Table VI. For pure isotactic polystyrene, Suzuki and Kovacs [56] found that C2 = 30.70K. The difference in the magnitude of the values of C2 for the methods of Boon and Azcue [5] and Suzuki and Kovacs [56] is related to the type of analysis used to evaluate the other empirical constants in the growth rate equation. Suzuki and Kovacs [56] also reported that according to their analysis, C2 = fg/af where fg is the fractional free volume at the glass transition temperature and af is the coefficient of thermal expansion. Bueche [10] indicates that fg = 0.025 for almost all glass forming substances and cf = 10-4 for most polymers. In addition Bueche [10] found that af is approximately 10-3 for low molecular weight plasticizers. Thus, by using the values of the free volume parameters (fg, af) reported by Bueche [10], we can see that the value of C2 should be depressed by the addition of low molecular weight plasticizer, a result experimentally verified for the benzophenone system.

89 Regardless of the analysis technique used, it is apparent that the empirical constant C2 in the activation energy for viscous transport term (I) is significantly lower for the system diluted with benzophenone than for the isotactic/atactic polystyrene system. In effect this means that for equivalent values of TT-Tg (crystallization temperature-glass transition temperature) in the range near Tg, where the growth rate is principally controlled by the activation energy for viscous transport (term I), the resulting growth rate for the isotactic/atactic polystyrene system will actually be higher than the rate for benzophenone system. However, in the range closer to the maximum growth rate where the free energy term exerts a stronger influence, we know that the benzophenone system has a growth rate substantially higher than the equivalent rate for the isotactic/atactic polystyrene system, Figure 20. The only way we can reconcile these experimental results with the derived form of the spherulitic growth rate equation (4) is to suggest that the free energy of critical nucleus formation is uniformly lower for the benzophenone system than for the atactic polystyrene system at equivalent dilutions. Apparently, the creation of a two phase plasticized mixture tends to lower the free energy-of critical nucleus formation AF* by decreasing the interfacial free energies au and ae (equation 7) associated with the formation of a critical nucleus. Although the exact mechanismof nucleus formation is presently unknown, it would seem reasonable to suggest that in the case of isotactic polystyreneplasticized with benzophenone and perhaps with di-decyl phthalate, the growth mechanism involves

90 the addition of whole structural units (with a lower free energy of critical nucleus formation) rather than segments of single molecules, onto the crystalline growth front. For the isotactic/atactic polystyrene system, however, the "homogeneous" phase growth rate behavior and growth rate magnitude would suggest a non-structured nucleation process or one of smaller magnitude, which perhaps involves the addition of single molecules or small groups of loosely bound molecules onto the growth front. E. Conclusions 1. For the isotactic polystyrene/di-methyl or di-decyl phthalate systems, the dependence of growth rate on temperature is very similar to that for the undiluted polymer, i.e., the growth rate curves are generally gaussian in shape with a sharp maximum at the optimum crystallization temperature. 2. As the concentration of di-methyl phthalate is increased, the crystallization range shifts to lower temperatures, the amount of shift being proportional to the amount of plasticizer added. 3. The overall maximum growth rate for the isotactic polystyrene/di-methyl phthalate system occurs between 20 and 30% dilution. 4. As the concentration of di-decyl phthalate is increased, the crystallization range shifts to lower temperatures for up to 20% dilution. Above 20% dilution, the maximum growth rates remain fixed at about 145~C (+ 2~C) for dilutions up to 80% plasticizer. Apparently, the curves

91 no longer shift to lower temperatures because the melting temperature over this concentration range is nearly constant (Figure 22). 5. The overall maximum growth rate for mixtures of isotactic polystyrene and di-decyl phthalate is relatively flat between 20% and 40% dilution, suggesting a microscopic phase separation between an isotactic polystyrene rich phase and a di-decyl phthalate phase over this concentration range. 6. The oDtical meltin tcmnera.tLrs fnr mixt.res of isotactic polystyrene/di-methyl or di-decyl phthalate are uniformly depressed with increasing plasticizer concentration, the amount of depression being greatest for comparable concentrations of di-methyl phthalate. 7. A direct comparison of the spherulitic growth rates for mixtures of isotactic/atactic polystyrene and isotactic polystyrene/di-decyl phthalate over the same diluent concentration and molecular weight range suggests that there is a basic difference in the microstructure of these two blends. It is suggested that the former mixture behaves as a "homogeneous" phase since the growth rate (concentration of isotactic polystyrene) is monotonically depressed with increasing concentrations of atactic polystyrene. The latter mixture, however, appears to behave as a two phase system, at least over certain concentration ranges. Thus, since the optimum spherulitic growth rate for the plasticized mixture is much higher than the comparable rate for the atactlc mixture, the creationrof a two phase system apparently effects either the viscous transport mechanism or the free energy of critical nucleus formation

92 in the spherulite growth process. 8. According to an analysis of the threshold crystallization temperatures for mixtures of isotactic/atactic polystyrene and isotactic polystyrene/benzophenone, the empirical constants C1 and C2 in the WLF form of the viscous transport term in the growth rate equation take on different values for these two mixtures. If we consider C to be invariant for both systems, then C2 is significantly lower for the benzophenone system, according to the methods of Boon and Azcue [5] and Suzuki and Kovacs [56]. The magnitude of this constant (C2), however, is dependent upon which method is used. If the derived form of the spherulitic growth rate equation (equation 4) is correct, then the free energy of critical nucleus formation must also be lower for the benzophenone system in order to explain the higher maximum growth rates observed. Thus, it is suggested that the spherulitic growth mechanism for plasticized systems involves the addition of whole structural units (with a lower free energy of critical nucleus formation) to the growth front, while the mechanism for the atactic polystyrene system is a nonstructural nucleation process involving the addition of single molecules or small groups of loosely bound molecules to the growth front.

CHAPTER V STRAIN INDUCED CRYSTALLIZATION A. Introduction Melt extrusion or injection molding of a crystallizable polymer often introduces an oriented crystalline morphology which is quite different than the type of morphology normally found in random spherulitic crystallization. Recent studies designed to elucidate the nature of this structure have concentrated on strain induced crystallization from the melt state. However, some crystalline polymers, particularly those having a slow spherulitic growth rate, can also be crystallized from the glassy or rubbery states. Therefore, in order to more fully understand the nature of strain induced crystallization, a comprehensive study was undertaken to examine the influence of extension ratio, substrate, annealing temperature, plasticizer, and annealing time on the morphology of strain induced crystallization of isotactic polystyrene from the glassy and rubbery amorphous states. In the work most directly related to the present study, Yeh and Geil [61] observed strain induced crystallization from the glassy amorphous state of poly(ethylene terephthalate). They reported that strain induced crystallization occurs by rotation, alignment, and perfection of 93

94 0 the internal order of 75A paracrystalline ball-like structures originally present in the amorphous material. Thin films of amorphous poly(ethylene terephthalate) became two dimensionally ordered when drawn 500% even at temperatures below the nominal glass transition Tg. This appearance of crystallinity below Tg coincides with the alignment of ball-like structures at angles ranging from 45~ to 50~ to the stretch direction. Subsequent heat treatment for 15 minutes at temperatures ranging up to 2600C increased the amount of crystallinity as well as the internal ordering of the ball-like structures, as indicated by small angle x-ray diffraction. Thin films annealed while in contact with a substrate (glass slide) developed row type structures oriented perpendicular to the stretch direction, with the crystalline fibers (rows of ball-like structures) prominently rising above the surface of the film. Dark field electron micrographs of thin films heat set at 2400C and 260~C, 0 using the (010) and (110) reflections, show 75A diffracting regions, but no evidence of any extended chain nuclei oriented parallel to the stretch direction. In experiments involving strain-induced crystallization of natural rubber from the melt state, Andrews [1] observed the change from spherulitic morphology in unstrained films to fibrillar morphology in highly stretched films. Thin films stretched to low elongations at room temperature and subsequently cooled to -26~0C developed fibrous row structures oriented perpendicular to the stretch direction, while similar films stretched to high elongations, 400% to

95 700%, develop fibril type structures aligned parallel to the stretch direction. Recent work by Luch [34] on the strain induced crystallization of natural rubber from the melt state shows structures similar to those observed by Andrews [1], although they appear to differ in detail. In addition, Luch noted that the fibillar structures in highly stretched thin films can be reversibly transformed to perpendicular structures by suitable thermal treatment. Keller and Machin [33] have examined strain induced crystallization of lightly crosslinked polyethylene from the rubbery melt state. In highly stretched films, they observed very fine structures or lamallae arranged perpendicular to the draw direction. In films stretched to lower elongations prior to crystallization, however, the perpendicular lamallae assumed a twisting conformation similar to those often observed in normal spherulitic crystallization. Their explanation for this behavior is that when the stress is low, the crystalline ribbons can freely twist as they grow, but as the stress is increased, the twisting becomes progressively more difficult because an increasing overall crystallization rate generates a higher fiber density. Keller and Machin [33] also suggest that line nuclei are probably present in the observed row structures and that the line nuclei are most likely made up of extended chain crystals. This would indicate that the resulting growth mechanism for row structures involves epitaxial growth of folded chain fibers perpendicular to these extended chain nuclei. Pennings and Kiel [47], among others, have produced

96 Shish-Kebab type row structures by shear induced crystallization of dilute solutions of polyethylene in xylene. These structures appear to be made up of platelet type lamallae superimposed on a filamentary ribbon backbone. The structure of the filamentary ribbon backbone is presently under intensive research by numerous investigators. For polyethylene shear crystallized at small supercoolings from dilute solution, Wikjord and Manley [59] succeeded in removing the lamallar row structures by nitric acid oxidation and selective toluene dissolution, leaving naked filamentary ribbons. Differential thermal analysis of these ribbon structures suggests a dual molecular conformation of both folded and extended chains. Thermal and oxidative behavior of polyethylene prepared at lower crystallization temperatures suggests a greater content of chain folds in the central ribbon as the degree of supercooling is increased. Row structures are also frequently found in bulk material prepared by melt extrusion or by injection molding [49]. Clark and Garber [13] observed that row structures are characteristic of the highly oriented surface of molded bars of polyoxymethylene while random spherulitic crystallization occurs in the unoriented interior. Some evidence suggesting the presence of fibril nuclei in blown films of polyoxymethylene was also reported by Clark and Garber [13]. 0 They observed 300A diameter fibrils oriented parallel to the stretch direction in the same field as crystalline row structures oriented perpendicular to the stretch direction. They suggest that the fibrils, supposedly generated by high

97 stress, are responsible for the row structure morphology, but the chain conformation within these fibrils has not been established. From the above discussion, we can see that the exact nature of line nucleation and the mechanism for the growth of row structures is not clearly understood at the present time. The results of Wikjord and Manley [59] indicate that the nature of the filamentary ribbon backbone in shear crystallized polyethylene is quite complex, consisting of both folded and extended chains. Dark field micrographs of strain crystallized poly(ethylene terephthalate) by Yeh and Geil [61] also indicate a segmented rather than an extended type structure. Thus, the backbone structures aligned parallel to the orientation direction —-if they exist —-are certainly more complex than the simple extended chain model suggested by Keller [33]. In addition, it is not clear whether the mechanism for the growth of row structures involves epitaxial growth on some type of line nucleus or a rearrangement of previously existing crystalline structures. In the present study, we have tried to expand our knowledge of strain induced crystallization by examining the detailed behavior of isotactic polystyrene strain crystallized from the amorphous glassy and rubbery states under a wide variety of conditions. In particular, we have examined thin films stretched up to 500% elongation by adding 40% benzophenone to isotactic polystyrene in order to reduce the glass transition temperature of the mixture to -2~C, in effect making the isotactic polystyrene/benzophenone 60/40

98 thin films rubbery at room temperature. We have used bright and dark field electron microscopy and electron diffraction to study stretched thin films which were either shadowed with platinum, decorated with gold, or mildly etched with amyl acetate, in order to remove the noncrystallizable diluent and expose the underlying crystalline structure. Since the gold decoration technique in particular is not well understood, we have carried out additional studies on crystallized holey films of isotactic polystyrene in order to elucidate the mechanism of gold decoration with respect to fiber orientation. B. Experimental Thin films of amorphous isotactic polystyrene with a large number of small holes were made by preparing dilute 0.2% solutions of isotactic polystyrene in diethylene chloride. After complete dissolution of polystyrene, several drops of water were added, and the combined solution was vigorously shaken until a finely dispersed emulsion was formed. Drops of the emulsion were cast onto glass slides coated with sodium hexa-meta phosphate, a water soluble replica releasing agent, and drained in such a way as to cause the resulting film to have a graduated thickness. While the emulsion thin films were drying, moist breathing on the film surface helped to create large numbers of tiny holes 0.1 to 10 i in diameter. After complete drying, the amorphous holey films of isotactic polystyrene were floated off onto a water surface and picked up on specimen grids for

99 subsequent annealing at 140~C in a controlled temperature hot air oven (+ 2~C) for 10 to 15 minutes. The crystallized holey films were prepared for electron microscope examination by either shadowing the surface with platinum, or decorating it with gold. Rubbery thin films capable of being stretched 500% on a water surface were cast from dilute 0.4% solutions of isotactic polystyrene/benzophenone 60/40 in benzene made by pipetting together proper ratios of 0.4% isotactic polystyrene and 0.4% benzophenone in benzene solutions. Thin films o approximately 1,000 to 1,200A thick, as judged by light reflections, were cast onto glass slides coated with sodium hexa-meta phosphate. After air drying at room temperature for 20 minutes, these films were floated off onto a clean water surface. Insufficient drying caused these films to shrink and thicken upon contacting the water surface. On the other hand prolonged drying tended to remove excess amounts of volatile benzophenone, making the films brittle at room temperature. Thin films having yellowish-orange light reflections, 0o 1200A, were most successfully stretched to elongations ranging from 100% to 500% by means of a pair of modified draftsman dividers, described by Yeh and Geil [62]. Specimen grids were placed on the surface of the stretched film and picked up by plunging a finger down through the water surface and scooping up the grids. These stretched films were annealed at 125~C, 155~C, or 175~0C in a temperature controlled hot air oven (+ 20C) from 1 to 60 minutes and

100 then prepared for electron microscope examination by either shadowing with platinum or decorating with gold. Films of unplasticized isotactic polystyrene were cast directly onto a Mylar substrate and stretched 50 to 100% in a small mechanical stretching device especially designed for this purpose, Yeh and Geil [62]. After stretching, these films were annealed at 1550C or 175~C for 20 minutes and shadowed with platinum or decorated with gold. The stretched films were then removed from the Mylar substrate by stripping with a layer of poly(acrylic acid) dried down from a 10% aqueous solution. The strripped films were O 0O backed with 200A-300A of carbon and the poly(acrylic acid) was redissolved in water, leaving the extraction replica to be picked up on specimen grids. The method for selective etching, similar to that used by Padden [46], involved mounting stretched thin films of isotactic polystyrene/benzophenone 60/40 on pieces of freshly cleaved mica and crystallizing the films in the same manner as before. After crystallization, the pieces of mica were plunged into a bath of amyl acetate, at 23~C, for times normally ranging from 20 to 30 seconds, drained, and then air dried. Other solvents such as xylene, acetone, and methyl ethyl ketone were also tried at various temperatures and etching times, but room temperature amyl acetate was found to be the most suitable because it did not cause solvent induced crystallization in noncrystalline amorphous thin films. After selective etching, the stretched films were shadowed with platinum, coated with carbon, and floated

101 off onto a water surface to be picked up on grids. Films of isotactic/atactic polystyrene stretched on M.ylar substrate were also selectively etched. After crystallization, the thin films mounted in the stretching device were plunged into a bath of amyl acetate, at 230C, for times normally ranging from 20 to 30 seconds. They were then drained and air dried. The etched films were shadowed with platinum and then removed from the mylar substrate by stripping with poly(acrylic acid.). C. Results 1. Morphology of Holey Films Electron microscopy examination of thin holey films of isotactic polystyrene, crystallized while being supported on 200 mesh copper grids, shows radial fibers uniformly distributed around the edges of 0.1 to 10.0 p diameter holes, as illustrated in Figure 25. The measured width of radial fibers, as well as the non-oriented fibers, is 0 110-120A in platinum shadowed holey films crystallized at 140~C. This result is in general agreement with the 127A small angle x-ray long period reported by Manley and Blais [43].in bulk isotactic polystyrene crystallized at 140~C. Similar holey films crystallized while being supported on a substrate (glass slide) do not develop the radial fiber conformation, Figure 26. Andrews [1], upon observing similar radial fiber formation in thin films of natural rubber, attributed the radial orientation to the concentric stress field which is locally generated around the edge of

102 Oi 25/ Figure 25: Isotactic polystyrene holey film crystallized at 1400C for 5 minutes. Platinum shadowed at 30* Figure 26: Isotactic polystyrene holey film crystallized on glass slide at 140C for 10 minutes. Gold decorated at 90

103 each hole. Apparently, the effect of the substrate in Figure 26 is to relieve this local stress field, thus causing a random fiber orientation. 2. Gold Decoration of Holey Films Gold decoration of crystallized isotactic polystyrene holey films shows a tri-layer decorative pattern which appears to coincide with the radial fiber orientation, as illustrated in Figure 27. Examination of Figure 27 will clearly show that the tri-layer decorative pattern consists o of two lines of gold particles, 45A in diameter, on either 0o side of a completely vacated region, measuring 60A between the edges of the aligned gold particles. It is of interest to note that our value of 60A is about the same as the value o of 65A reported by Haller and Magill [20] for the vacated width of gold decorated fibers of polysiloxane crystallized at 1300C from the melt state. It is well known that the equilibrium width of crystalline fibers is a strong function of crystallization temperature. Accordingly, thin films of isotactic polystyrene were crystallized at three different temperatures and the resulting average dimensions of the platinum shadowed and gold decorated fibers are compared in Table IV (see Figures 25, 27, 28, 29).

104 Figure 27: Isotactic polystyrene holey film crystallized at 140C for 10 minutes. Gold decorated at 90. Arrows indicate tri-layer decorative Dattern. Figure 28: Isotactlc polystyrene thin film crystallized at 210C for 5 hours. Platinum shadowed at 30$

105 2,>~ ~ ~~~V Figure 29: Isotactic polystyrene thin film crystallized at 210C for 5 hours. Gold decorated. ~,:,'~ ~",T': ~".~'~:. ~' ~' <, "t:~"" ~..~.:'~. i.~'~~':~.:~':.......'..'~'.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............,q,,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.... w., F ~ ~ ~ (j, ~~~~~~I"~~~~f': f ~4'~ Zt*''~B~'9 Figure 3 90 Isotactic polystyrene holey film crystallized at 1 0C for 10 minuts. Polatinu.'"::"~':'~i:'"'~' ~': ":'""* "....' side. ~Arrows indicate are'wer platinum sc up. sr.0. 0./.... ~ O ~.. Figur 30: sotci oytreehlyfl crystallized::: at10 0miue.Paiu shadowed on ne side and gold decratdonrves side. ~rrows indic~~~~~~~ ateae hr ltnmsak

106 Table VII Parameters of Gold Decorated Fibers Au vacant Au Temperature ~C Pt. Area particle X-ray* 0 0 0 0 140 110-120A 60-65A 50-55A 127A 175 150-165A 55-60A 45-50A 160A 0 0 0 0 210 200-210A 55-60A 45-50A 215A *Small angle x-ray data, Manley and Blais [43]. Although the dimensions of the platinum shadowed fibers are in general agreement with the small angle x-ray data of Manley and Blais [43], the width of the vacated region in similar films decorated with gold appears to be insensitive to lamallar thickness or crystallization temperature. This result appears to be in conflict with the observations of Spit [55], who noted that in thin films of crystallized nylon 6, the gold particles and phosphotungstic acid stain seem to be sensitive to one and the same region, presumably the amorphous layer adjacent to the crystalline fiber. In a separate experiment, we tried to establish whether the gold decoration particles stack up along the bottom edge of raised crystalline fibers, as does platinum, or perhaps are sensitive to surface structure on top of the crystalline fibers. Figure 30 shows a crystallized holey film of isotactic polystyrene shadowed with platinum on one side and decorated with gold on the reverse side. Careful examination will show that the gold decoration particles tend to stack up in about the same region as the platinum

107 material (+ 25X) and perhaps a little toward the center of the crystalline fibers. Thus, with this type of experiment it cannot be clearly established exactly where the gold particles reside, i.e., in the middle or on the edge of lamallae. Figure 31 shows a high resolution micrograph of a very thin holely film of crystallized isotactic polystyrene decorated with gold. The dark radial fibers can be associated with crystalline fibrous regions oriented edge on for maximum electron scattering density. Adjacent regions are either completely amorphous or only partially crystalline, and in any case have a lower electron density or mass thickness. In most areas, the tri-layer decorative pattern appears to coincide directly with the most dense regions in the thin film, the crystalline fibers. However, in some areas, particularly where the radial fibers are closely packed together, the line of gold particles actually appears to reside on top of the dense crystalline fibers. Although these results are not conclusive, it appears that the trilayer decorative pattern may in some instances be sensitive to the surface structure of the fibers, perhaps even along the crystalline core regions, as indicated by these experiments and the data in Table VII. 3. Morphology of Films Stretched on Mylar Thin films of isotactic polystyrene stretched 100% on a Mylar substrate at room temperature are amorphous according to surface morphology and corresponding electron diffraction patterns. However, when these stretched films

108;~.........~.!,. Figure 31: Isotactic polystyrene holey film crystallized at 140C for 10 minutes. Gold decorated at 90. Dense crystalline regions are indicated by arrows. Figure 32: Isotactic polystyrene stretched 1.00% on mylar, annealed at 175FC for 20 minutes. Platinum shadowed. Stretch direction is verticle and crystalline fibers are indicated by small arrows.

109 are annealed at 175~C, a uniform field of crystalline fibers grow perpendicular to the original draw direction, Figure 32. These platinum shadowed fibers range in width from 170 0 to 190A, which is somewhat larger than the small angle x-ray o long period of 160A reported by Manley and Blais [43] for bulk unoriented isotactic polystyrene crystallized at 175~0C. Examination of Figure 32 indicates that the fibrous structures are somewhat discontinuous, perpendicular to the stretch direction, as in PET [61], and evidently twist into the plane of the film at various random locations. The selected area; electron diffraction pattern corresponding to the area in Figure 32 shows a strong (102) reflection (within 9~ of the (001) chain axis [9]) parallel to the draw direction and strong equitorial reflections which correspond to the (300), and (220) planes, Figure 33. This result indicates that the molecular chain C axis is oriented parallel to the original stretch direction, which is consistent with the molecular.rientations found by Natta [44] in stretched fibers of isotactic polystyrene. This molecular orientation 0 and the width of the fibers (170A along the stretch direction) would tend to suggest that the fibrous row structures have 0o a folded chain conformation across the 170A width of the fiber. The folded chain conformation is also found in crystalline structures of other polymers [17]. Finally, Miller and Buchanan [9] reported that the average theoretical crystallite size, based on (102) x-ray line broadening data, O o is about 85A to 100A in direction parallel to stretch for oriented isotactic polystyrene.

110 Gold decoration of similarly crystallized films of oriented isotactic polystyrene shows fibrous row structures mostly aligned perpendicular to the original stretch direction, Figure 34. The width of the vacated region in the tri-layer 0 pattern measures 60-65A. The fact that the decorated row structures appear to be somewhat more continuous perpendicular to the stretch direction is probably related to the lower fiber density in the decorated film (compared to Figure 32). Figure 35 illustrates the effect of selective amyl acetate etching of a film of isotactic polystyrene stretched 100% and annealed at 175~0C. The clearly defined row structures actually appear to be stacks of lamallar ribbon structures oriented edge on perpendicular to the original draw direction. These etched fibers range in width 0 from 160 to 180A, which is in good agreement with the results noted for unetched films crystallized under similar conditions. A finely distinguished Shish-Kebab structure is found in thin films of isotactic/atactic polystyrene/benzophenone 10/80/10 cast from solution onto Mylar substrate and crystallized under stress at 1550C from the glassy amorphous state, Figure 36. We have selectively etched this thin film (Figure 36) to preferentially remove the noncrystalline impurity. In many areas of this film the crystalline structure appears to develop quite independently of any common extended line nuclei. rlThe long shish-kebab structures, oriented parallel to the stretch direction, have iamallae overgrowths attached perpendicular to some type of common

111 d_~esignations for crystalline.reflections.:,i0:/, i: y: E E f::.:i..........:igE.000... St....:.:'.' Figure 3: Isotactic c polystyre e stretched 100% on mylar, annealed at 175C for 5 minutes. Gold decorated. Stretch direction is verticle. Small arrows indicate tri-layer decorative pattern. 51~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:::ii~i:i'-i::;'::::::i::::i:::::::::::::.:;ii. i:ii~ii: -00~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:iiii~:iii:iii~i'I:~'i~~~i: iilsi-ji;i::l'i::i:::':::'::

112 Figure 35: Isotactic polystyrene stretched 100% on mylar, annealed at 175C for 20 minutes, amyl acetate etched for 20 seconds. Platinum shadowed at 30. Stretch direction is verticle. Figure 36: IPS/APS/benzophenone 10/80/10 stretched 100% on mylar, annealed at 15~C for 20 minutes, amyl acetate etched for 10 seconds. Platinum shadowed. Stretch direction indicated by large arrow.

113 o backbone. These lamallae overgrowths measure 140 to 160A in thickness, which is somewhat larger than the small angle o x-ray long period of 135A reported by Manley and Blais [43] for bulk unoriented isotactic polystyrene crystallized at 155~C. The ends of strongly etched-shish-kebab structures often appear to curl around some 180~. The reason for this behavior is unknown, but might be related to redeposition after etching. in a separate experiment, we optically measured the comparative growth rates at 1750C for oriented lamallae in shish-kebab type structures and randomly oriented lamallae in spherulites of isotactic/atactic polystyrene molecular weight 900. The growth rates for the two structures are essentially the same in a thin film (60 i thick) which has an indeterminate stress field. 4. Morphology of Films Stretched on Water Unfortunately, thin films stretched on Mylar substrates are limited to extensions less than 100%. In order to investigate the effect of greater extension, up to 500%, as well as annealing temperature and time, thin films of isotactic polystyrene plasticized with a 40% concentration of benzophenone (Tg = -20C) were stretched on a water surface, mounted on grids, and annealed at different temperatures and times, Tables VIIIA, VIIIB, VIIIC. Unannealed films stretched up to 500% show no signs of crystallinity, either with regard to the appearance of surface structure or a discrete electron diffraction pattern. However, after annealing a film stretched 100% for 2 minutes

COMPARATIVE GROWTH RATE FOR ROW STRUCTURES AND SPHERULITES, IPS/APS M = 900 40/60 @ 175%C w 25 e Spherulites ORow Structures 220 4 —) I o 15 0 0 1 10 N -C) cI~~~~~~~00 0 ~10 20 30 1050 0 70 0 0 Time, Minutes Figure 37

115 Table VIIIA Strain Induced Crystallization of Thin Films of IPS/Benzophenone 60/40 Stretched on Water at Room Temperature, Platinum Shadowed I struc- i fibture, O rils, Lateral, Length, TaOC Stretch ta,min. {(dq) ~i a ) _ _ ii 125 200% 4 120 -.69.13-2.8 125 200% 5 110 -.75.10-2.4 125 200% 10 120 - - 125 200% 20 115 -.72.16-2.2 125 300% 2 --- - 125 300% 2 120 -.62.15-2.0 125 300% 4 115 -.69.09-1.4 125 300% 6 120 -.77.12-2.3 155 100% 2 145 - - 155 100% 2 140 - - - 155 100% 4 --- - 155 100% 6 145 -.46.15-2.7 155 200% 2 134 -.66.11-1.8 155 200% 4 137 -.39.09-1.8 155 200% 6 144 - - - 155 300% 2 137 -.48.07-1.4 155 300% 6 135 -.74 - 155 300% 6 136 -.61.17-2.3 155 300% 8 140 -.54.15-2.9 155 400% 2 137 -.34' - 155 400% 6 135 160 - - 155 400% 30 138 - - - 155 400% 60 140 - - - 175 100% 6 148 - - - 175 200% 4 151 152 - - 175 500% 2 152 145.45 3.0+ 175 500% 6 --- 143 - 175 500% 6 --- 147

118 Figure 38: Isotactic polystyrene/benzophenone 6 0/40 stretched 100~, annealed at 155t for 2 minutes. Platinum shadowed at 30. Stretch direction indicated by arrow. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Figure 39: Figure 40; Electron diffraction pattern Electron diffraction pattern corresponding to Figure 38. corresponding to Figure 44.

11,9 Figure 41: Figure 42: IPS/benzophenone 60/40 IPS/benzophenone 60/40 stretched 3009, annealed at str.tched 300%{, annealed at 125C for 4 minutes. Platinum 12C for 6 minutes. Platinum shadowed at 30 Stretch shadowed at 30~ Stretch direction is verticle. direction is verticle.

120U Figures 32 and 38. Instead, groups of oriented fibers appear to be arranged into column structures aligned parallel to the stretch direction. After 6 minutes annealing time under these conditions, Figure 42, the crystalline surface structure is now more extensive with perpendicularly oriented fibers being arranged, not only in small groups extending from 0.1 to 0.5 p along the stretch direction, but also along column structures extending 2 to 3 p or more. Even small groups of fibers, however, appear to reach a more or less uniform lateral width of 0.6 to 0.8 p (independent of annealing time) before impinging upon neighboring structures, Table VillA. Earlier results suggest that the spherulitic growth rate for these mixtures should be about 0.6 p per minute, which would indicate that impingement probably occurs within the first minute of growth. Apparently the nucleation density for these films is too high to correlate the increase in lateral width with the growth rate kinetics. The measured thickness of individual fibers in 0 0 films annealed at 125~C ranges from llOA to 120A, as shown in Figure 43. This value is in good agreement with the 0 small angle x-ray long period of 120A reported by Manley and Blais [43] for unoriented bulk isotactic polystyrene crystallized at 125~C. When a film stretched 300% is annealed at 155~C for 6 minutes, the surface crystallization appears to be complete with the entire surface of the film being covered with fibrous structures, Figure 44. This surface morphology does not appear to be altered by extended annealing time up to 60

121 Figure 43: Figure 44: IPS/benzophenone 60/40 IPS/benzophenone 60/40 stretched 300%, annealed at stretched 300%, annealed at 12~C for 6 minutes. Platinum 155C for 6 minutes. Platinum shadowed at 30' shadowed at 30. Stretch direction is verticle, as indicated by electron diffraction pattern.

1 22 minutes once the crystallization is complete. Most of' the surface structure in Figure 44 can be associated with extended column structures 2 to 3 P long aligned parallel to the stretch direction, but interspersed among these columnns are small independently developed groups of oriented structures, not necessarily related to any common line nucleus. The typical electron diffraction pattern corresponding to this area, Figure 40, has a sharper and more intense (102) reflection parallel to the draw direction, indicating a higher degree of molecular alignment than observed in films stretched 100%, Figure 39. A directcomparison of the surface morphology of films stretched 100% and 300% indicates that the higher elongations also tend to generate more extended column type structures than found in films stretched 100%, compare figures 32, 38 and 44. The lateral width of oriented structures in stretched films annealed at 155~ C reaches a more or less uniform value of 0.3 to 0.6 p (independent of annealing time) before neighboring structures impinge upon one another, Table VIIIA. The measured thickness of individual fibers in films annealed 0 at 155~C ranges from 135 to 145A, which is in general agreement with the small angle x-ray long period of 137A reported by Manley and Blais [43] for unoriented bulk isotactic polystyrene crystallized at 1550C. Apparently elongations ranging from 100% to 400% and annealing times up to 60 minutes have no significant influence on the resulting thickness of perpendicular structures, Table VIIIA.

123 Selective etching of a thin film stretched 300% and annealed at 1550C for 6 minutes reveals the underlying crystalline structure beneath the surface, Figure 45. Within the same field of view, we can see extended column structures (I),small groups or bundles of oriented fibers (II), and stacks of lamallae lying in the plane of the film (III). The extended column structures seem to be made up of long rows of perpendicularly oriented fibers which appear to have a common line nucleus. In nearby regions we can see small bundles of fibers mostly oriented perpendicular to the stretch direction, but not necessarily having a common line orientation. In between these fibrous structures, we can see stacks of lamallae lying in the plane of the film. This type of lamallar structure is commonly observed in thin films crystallized while not under stress. This result would suggest that extended column structures might be associated with areas of high stress while bundle structures, and particularly lamallae structures might correspond to low stress areas. When thin films stretched 500% are annealed at 175~C for 2 minutes, fibril type structures aligned parallel to the stretch direction are usually observed, Figure 46. Similar fibril structures have been found by Luch [34] in highly stretched thin films of natural rubber crystallized from the melt state. The fibril structure in isotactic 0 polystyrene (140 to 150A in diameter) is particularly prominent at higher extensions and higher annealing temperatures (Table VIIIA). Occasionally, they are found at

124 Figure 45: Isotactic polystyrene/benzophenone 60/40 stretched 300~, annealed at 155C for 6 minutes, amyl acetate etched for 30 seconds. Platinum shadowed. Stretch direction is verticle. Small arrows indicate important struct ural features. Figure 46: Isotactic polystyrene/benzophenone 60/40 stretched 5009, annealed at 175~ for 2 minutes. Platinum shadowed. Stretch direction is nearly horizontal, as indicated by electron diffraction pattern.

125 elongations as low as 200%, when stretched films are crystallized at 175~C, Figure 47. Otherwise, the annealing temperature apparently has little discernible effect on the type of morphology (other than fiber thickness) observed in annealed films having the same elongation prior to crystallization. Higher magnifications of Figure 46 shows that fibril structures are generally discontinuous on the surface for distances greater than 0.5p, Figure 48. It should also be noted that along the border of the extended column structures, the parallel' fibrils appear to closely intermingle with the perpendicularly oriented fibers. The typical electron diffraction pattern corresponding to Figure 46 shows a splitting of the very intense (102) reflection, indicating a very high degree of molecular alignment parallel to the stretch direction, Figure 49. The angle between the (102) reflection and the C axis measures 90, in good agreement with the angular relationship found by Buchanan and Miller [9] in stretched bulk films of isotactic polystyrene. The spacing of the layer lines along the stretch 0 direction equals 6.54A, which agrees with the C axis unit cell parameter, as determined by Natta [44]. Gold decoration of a thin film stretched 500% and annealed at 1550C for 4 minutes also indicates parallel fibril structures, as suggested by the vacated areas aligned parallel to the stretch direction (see arrows), Figure 50. The vacated areas vary in length from 0.07 to 0.5 p parallel to the stretch direction and have widths ranging from 70 to

126 Figure 4L7: Isotactic polystyrene/benzophenone 60/40 stretched 200%, annealed at 175~ for 4 minutes. Platinum shadowed at 30~ Stretch direction is indicated by large arrow. Figure 48: Isotactic polystyrene/henzophenone 60/40 stretched 500%, annealed at 17?C for 2 minutes. Platinum shadowed at 30~ Stretch direction is hori zont l.

127 Figure 49: Isotsctic polystyren&/benzophenone 60/40 stretched 500%, annealed at 155C for 4 minutes. Electron diffraction pattern. Sketch shows lattice designations for crystalline reflections. Figure 50: Isotactic polystyrene/benzophenone 60/40 stretched 500%, annealed at 1550C for 4 minutes. Gold decorated. Stretch direction is indicated by large h~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~rTOW ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~iiii~~l

128 0 l00A between parallel lines of gold particles. Careful examination of Figure 50 will also show that the parallel fibrils and perpendicular fibers (indicated by vacated areas perpendicular to the stretch direction), although growing in separate regions, closely intermingle on the border separating them. Selective etching of a thin film stretched 400% and annealed at 1550~C for 6 minutes reveals the inlternal structure of the extended columns, as illustrated in Figures 51 and 52. Along the backbone of the extended column structures we can see common line nuclei measuring 100 to 110A in diameter (see arrows). The lamallar overgrowths attached perpendicularly to the line 0 nuclei measure 110 to 120A in width, somewhat less than unetched row structures prepared under similar conditions. Apparently, in this case, the etching solvent also attacks the regions in between the lamallar overgrowths. Figures 53 and 54 show a bright field-dark field pair of micrographs from the same region of a thin film stretched 500% and crystallized at 155~C for 4 minutes. The dark field uses the (102) reflection which should show the areas of the crystalline structure in which the molecules are most strongly aligned in the stretch direction. Although the resolution of the dark field micrograph is marginal, it is still possible to distinguish individual diffracting areas, indicated by 0 the dark spots in the negative print, some 150 to 200A in diameter. The individual diffracting regions do not appear to be elongated in the stretch direction, although in some cases they appear to align (see arrow A). In the corresponding bright field area (Figure 53)

129 Figure 51: Isotactic polystyrene/benzophenone 60/40 stretched 400%, annealed at 155C for 6 minutes, amyl acetate etched for 30 seconds. Platinum shadowed. Stretch direction is indicated by large arrow. Figure 52: I sotacti c polystyrene/benzophenone 60/40 stretched 400%, annealed at 155C for 6 minutes, amyl acetate etched for 30 seconds. Platinum shadowed. Stretch direction is indicated by large arrow.

130 Figure 53: Isotactic polystyrene/benzophenone 60/40 stretched 500%, annealed at 155b for 6 minutes. Platinum shadowed at 30~ Stretch direction is horizontal. Figure 54: Isotactlic polystyrene/benzophenone 60/40 stretched 500I, Innealed at 15C00 for 6 minutes. Platinum shadowed. Dark field of figure 53 using (102) reflection. Stretch direction is horizontal, as indicated by electron diffraction pattern.

131 indistinct fibril structures aligned parallel to the stretch O direction measure 150A in width. Theoretical calculations based on the line broadening of the (102) reflection by Buchanan and Miller [9] indicate a crystallite size ranging 0 from 85-100A, considerably smaller than the diffracting areas observed in the dark field print. Perhaps the discrepancy here is partially due to the relatively poor resolution obtained by the apperture technique in dark field microscopy. In any case, it is apparent that the fibril structures have segmented diffracting regions, indicating that the molecular structure is predominately folded chain rather than extended chain. This result is in agreement with the findings of Yeh 0 and Geil [61] who observed 75A ball-like diffracting regions, but no extended chain nuclei, in strain crystallized PET from the glassy state. D. Discussion 1. Characterization of Structure Bassette [3] and Spit [55] both indicated that gold decoration of spherulitic structure in crystallized polymer films reveals a layered decorative effect which appears to be sensitive to crystalline fibers. In addition, Spit [55] demonstrates in crystallized thin films of nylon 6, that gold particles and phosphotungstic acid stain seem to be sensitive to one and the same region, presumably the amorphous layer adjacent to the crystalline fiber. In the present study, gold decoration of dispersed

132 crystalline fibers of isotactic polystyrene reveals a trilayer decorative pattern which consists of two lines of gold particles on either side of a completely vacated region, Figure 27. In the region near the hole where the fibers impinge on one another the gold particles appear to lie on top of the fibers. Experiments involving both platinum shadowing and gold decoration of opposite sides of the same thin film suggest that the decorative pattern is sensitive to the crystalline fibers, but not necessarily the raised edges of these fibers. Other results show that the decorative pattern coincides with the most dense regions in a crystallized film, the crystalline fibers oriented edge on to the plane of the film. Finally, we found that the width of the vacated 0 region, 60A, is apparently insensitive to crystallization temperature over a range of temperatures where the crystalline 0 fibers increase in width from 127 to 210A. Although these results are not conclusive, it appears that the tri-layer decorative pattern may in some instances be sensitive to the surface structure of the fibers, perhaps even along the crystalline core, as indicated by these experiments and data in Table VII. In all strain induced crystallization experiments, the molecular orientation, as determined by electron diffraction, is found to be more or less parallel to the stretch direction, while the most common morphological structure seems to be arranged perpendicular to the stretch direction. The folded chain model reconciles the orientation of these two structural features by specifying that molecular chains, approximately 24,000A long, must have chain folds in order

133 0 0 to fit into crystalline fibers ranging from 110A to 160A thick. Other models such as the fringed micelle, in which most chains tend to extend from one fiber to the next with few fold backs, do not apply in this case because selective etching experiments have clearly established the separate lamallar identity of each perpendicular fiber, Figure 36. If the molecular structure in these fibers is somehow different than the folded chain structure normally found in lamallar ribbons, then the resulting crystalline growth rates for row structures and spherulites should reflect this difference, since the growth rates are controlled by the rate of secondary nucleation at the growing interface. Figure 37 shows that the growth rates for row structures and spherulites are equal under the same conditions, thus implying that their molecular structures are also identical. We should point out that in this case we may be comparing epitaxial growth in row structures to normal spherulitic growth. Growth rates for strain induced crystallization of polymers are seldom obtainable. In any case, all of these results suggest that the crystalline fibers aligned perpendicular to the stretch direction in strain crystallized films have a folded chain molecular structure. The microstructures generated as a result of strain induced crystallization may be separated into three loosely defined categories. Type I structure is made up of small groups or bundles of fibers growing more or less perpendicular to the orientation direction, but developing independently of any common extended line nuclei, Figure 45, Such

1314 structures are most commonly found in films stretched to low elongations (100%), but also are found in films at higher elongations (200% to 300%). The electron diffraction pattern corresponding to this structure indicates a preferential molecular alignment more or less parallel to the stretch direction. Extended column type structures found in Figures 44 and 46 are designated as being of type II. This structure appears to be made up of rows of crystalline fibers growing perpendicular to some type of common line nucleus, Figures 51 and 52. Column structures are usually observed at elongations greater than 200%, while the corresponding electron diffraction pattern indicates a somewhat higher degree of molecular alignment than found in type I structure. Fibril structures aligned parallel to the stretch direction are designated as type III structures. These fibrils, 140 to 150A wide and extending 0.07 to 0.5 i in the stretch direction, are generated at high elongations (400 to 500%) and high annealing temperatures, Table VIII. The electron diffraction corresponding to these structures suggests a high degree of molecular alignment parallel to the stretch direction, while dark field studies using the (102) reflection indicate that the internal structure of these fibrils is segmented. This latter result suggests that the molecular conformation of these fibrils is predominately folded chain rather than extended chain. Gold decoration and selective etching experiments indicate that type III fibrils often occur in regions closely adjacent to type II structures.

135 2. Mechanism of Strain Induced Crystallization Keller [49] has proposed that the mechanism of strain induced crystallization basically involves the epitaxial growth of folded chain fibers on an extended chain nucleus backbone as depicted in the following diagram. Experimental Row Structures evidence indicating the existence of these proposed nuclei was recently presented by Wikjord and Manley [59], who observed that filamentary ribbon structures remain after the selective dissolution of oxidated polyethylene row structures produced by shear induced crystallization from dilute xylene solutions. Differential thermal analysis and selective oxidation behavior suggests that the ribbon structures observed in polyethylene have a dual molecular conformation of both folded and extended chains. Wikjord and Manley also established that the folded chain character of these filamentary ribbons increases as the crystallization temperature is lowered. Fibril type structures have also been observed in other studies by Andrews [1] and Luch E34], but there is no clear evidence relating these structures to the extended chain nuclei proposed by Keller. In the present study, fibril (type III) structures,

136 aligned parallel to the stretch direction, were observed in highly stretched thin films of isotactic polystyrene. Dark field studies using the (102) reflection indicate that the internal structure of these fibrils is segmental, suggesting a predominately fold chain molecular conformation.) In the study on strain induced crystallization of isotactic polystyrene, the degree of supercooling for fibril structures grown at 1550~C was approximately AT = 350C, considerably higher than the supercooling used by Wikjord and Manley [59] for shear induced crystallization of polyethylene row structures from dilute xylene solution. These results suggest that a folded chain molecular conformation for fibril structures grown from the amorphous rubbery state is not inconsistent with the experimental evidence on solution grown row structures by Wikjord and Manley. We also observed a filamentary backbone in mildly etched column (type II) structures of strain crystallized isotactic polystyrene, Figures 51 and 52. The dimensions and the orientation of this backbone structure suggest that it might be the line nucleus responsible for the generation of extended column structures, a mechanism proposed by Keller [49]. However, there is no evidence at the present time indicating what type of internal structure the filamentary backbone might have and no evidence indicating the mechanism for the growth of row structures along the backbone.

137 E. Conclusions 1. The dimensions of the gold tri-layer decorative pattern for crystalline fibers of isotactic polystyrene appear to be insensitive to crystallization temperature over the range of temperatures where the fibers increase in width from 127 to 210A. This result suggests that gold decoration may be sensitive to the surface structure of these fibers, perhaps even the crystalline core region. 2. Crystalline fibers aligned perpendicular to the stretch direction in strain crystallized thin films appear to have a folded chain molecular structure. 0 3. Fibril type structures 140 to 150A wide and aligned parallel to the stretch direction are found in thin films of isotactic polystyrene/benzophenone 60/40 stretched 400 to 500% and annealed at 1750C from the rubbery amorphous state. The electron diffraction pattern corresponding to these films indicates a high degree of molecular alignment parallel to the stretch direction while dark field std'les using the (102) reflection indicate that the internal structure of these fibrils is segmented. This latter result suggests that the molecular conformation in the fibrils is predominately folded rather than extended chain. 4. In thin films stretched to lower elongations (300%) prior to crystallization from the glassy amorphous state, extended column type structures are often observed. These column structures are made up of rows of crystalline fibers 120 to 160A thick, growing perpendicular to a common

138 filamentary backbone. 5. In thin films strain crystallized at very low elongations (100%), small groups or bundles of fibers aligned more or less perpendicular to the orientation direction appear to develop independently of any common line nucleus.

CHAPTER IV CRYSTALLIZATION OF ISOTACTIC POLYSTYRENE/PLASTICIZER BLENDS A. Introduction Spherulitic crystallization from mixtures of crystallizable polymer and low molecular weight plasticizer has received very limited attention, compared to the numerous studies on spherulitic crystallization of homopolymers. Mixtures of polymer and low molecular weight plasticizer are often commercially important systems because the mechanical properties of some polymers can be favorably altered by the addition of plasticizer [45]. Therefore, it was decided to examine the influence of di-methyl and di-decyl phthalate, as well as benzophenone on the spherulitic crystallization kinetics of isotactic polystyrene and to compare these results to those obtained in Chapter III. We also determined the threshold crystallization temperatures for mixtures of isotactic/atactic polystyrene as well as isotactic polystyrene/benzophenone. The threshold crystallization temperature (TT) is the lowest temperature at which the onset of crystallization can be detected. This temperature gives us a relative measure of the magnitude of the viscous transport term in the theoretical growth rate equation. 54

140 styrene butadiene rubber. In the case of polisobutylene, they observed that ultra-sonic treatment,while reducing the molecular weight of the material, does not appreciably affect the size of the resulting microstructure. In closely related work, Yeh and Geil [62] observed 0 a 75A nodular structure in bulk and thin films of amorphous poly(ethylene terephthalate),PET. They found that upon drawing thin films some 500%, these nodular structures tend to align approximately 50~ to the draw direction. Careful dark field microscopy using the inner amorphous ring suggests that these structures have a paracrystalline-type order. Crystallization at temperatures near Tg occurs by movement, aggregation, and alignment of these structures into spherical patches which then seem to form tree-like branches. In later stages of development, small bundles of spherulitic fibers in which the nodules are aligned perpendicular to the fiber axis are commonly observed. The small angle x-ray long period is found to increase as expected with annealing temperature, suggesting a regularization of the irregular folds originally present in the ball-like nodules. 0 Frank and Stuart [16] reported a 100A nodular microstructure in bulk amorphous polycarbonate from bisphenol A. This structure, unobservable under normal conditions, was detected by ion etching of bulk films which had been annealed just below Tg. They were able to correlate the presence of this structure with a slight change in the temperature dependence of the mechanical loss angle in amorphous polycarbonate.

141 O Similar 125A nodular structures have been observed in thin films of polycarbonate from bisphenol A by Carr and Geil [12]. They noted that upon annealing just below Tg, these nodules enlarge to 250A. Tensile deformation of a cast 0 thin film causes the nodules to break down into 60A units which align perpendicular to the stretch direction. Annealing of the film prior to deformation, however, causes these nodules to align along shear lines. In an extension of this work, Siegman and Geil [53] examined the crystallization of polycarbonate A from the glassy amorphous state. They found that crystallization near Tg occurs by aggregation of the enlarged nodules into irregularly shaped lamallae single crystals which, in later stages, develop fibril type structures characteristic of spherulitic crystallization. They indicate that subsequent growth of these fibrils occurs by interfacial addition of nodules. In the study of a plasticized system, Gezcvich and O Geil [18] observed a 200A nodular structure in mixtures of poly(vinyl chloride) and di-octyl phthalate, by means of brittle fracturing at liquid nitrogen temperatures followed by ion etching under vacuum. Large structures on the order of 0.1,u and 5 to 10 p were observed when chlorinated parafin was used as the plasticizer. The small, angle x-ray long period was found to increase with increasing plasticizer concentration, suggesting that the structures were becoming separated by a greater distance as the plasticizer content increased.

142 B. Experimental Solutions containing 0.2% polystyrene in appropriate solvent (benzene, dichlorobenzene, or cyclohexanone) were made by dissolving a weighed amount of either isotactic or atactic polystyrene in a measured volume of boiling solvent. Once the solution had cooled to room temperature, amorphous o 0 thin films 400A to 600A thick, as judged by light reflection, were cast onto glass slides coated with sodium hexa-meta phosphate, a water soluble releasing agent. Annealing and crystallization experiments were carried out on some of these films in a temperature controlled (+ 2~C) hot air oven at temperatures up to 140~C for periods of time ranging from 1 to 60 minutes. These thin films were then floated off onto a water surface and picked up on copper grids for subsequent study in the electron microscope. Most films were either shadowed with platinum or decorated with gold. However, a number of unshadowed films were also examined. Occasionally, crystallized thin films of isotactic polystyrene were selectively etched in room temperature amyl acetate. The etching procedure, similar to the one used by Padden [46], involves dipping the film covered glass slide into a bath of etching solvent for times ranging from 10 to 120 seconds, draining the slide, and allowing it to air dry. The etched films were then shadowed with platinum, coated 0 with about 200A of carbon, and floated off onto a water surface to be picked up on specimen grids. Oriented thin films of amorphous polystyrene were

14 3 prepared by stretching on Mylar substrate in the same manner as discussed in Chapter V. These films were either shadowed with platinum or decorated with gold and then stripped off the Mylar with a 10% aqueous solution of poly(acrylic acid). Bulk samples of atactic polystyrene were examined by surface replication and ultra-microtoming. Pellets of commercial atactic polystyrene were melted on a glass slide and annealed at 190~C-, 2050C, and 2200C in a hot air oven for 30 minutes before quenching in either a -70~C mixture of ethanol-dry ice, ice water, or the ambient atmosphere. One stage surface replicas of these quenched bulk samples were made by shadowing the surface with platinum and stripping the replica with a 10% aqueous solution of poly (acrylic acid). 0 The replica was then coated with about 200A of carbon backing and floated onto a water surface in order to redissolve the poly(acrylic acid). A Cambridge Ultramicrotome device was 0 used to shave ultra thin 700-900A sections from the same commercial polystyrene which were then shadowed with platinum and examined in the electron microscope. Specimen preparation and microtoming techniques are fully discussed in Sjostrand [54] and will not be dealt with here. C. Results 1. Amorphous Structure 0 An unshadowed 20A grainy microstructure was observed in all of the thin films of amorphous polystyrene examined in a JEM-6A electron microscope. The average diameters and center to center distance between these structures for a

14 4 variety of conditions is recorded in Table XII. According to these results, none of the major conditions studied (variation of molecular weight, solvent, annealing temperature) has a significant influence on the size of the observed microstructure. When similar unshadowed thin films were examined in a Phillips EM300 electron microscope with a much higher resolution, smaller structures on the 0 0 order of 7A to 8A were observed in atactic polystyrene. Because of the difficulties involved in the interpretation of phase contrast images and the need for. a high resolution microscope (which was not always available) these structures were not investigated further. When amorphous thin films of isotactic polystyrene were shadowed with platinum, sharp 30A grainy microstructures were observed, Figure 55. The average diameter and the center to center distance between these structures in shadowed amorphous thin of several different polymers prepared under different conditions is recorded in Table XIII. It appears, based on these results, that the size of the platinum shadowed microstructure is essentially invariant with respect to the conditions examined. Orientation of amorphous thin films should cause the resulting microstructure to align into some type of preferred orientation. Accordingly, thin films of polystyrene were stretched 50 to 100% on a Mylar substrate and either shadowed with platinum or decorated with gold, Figures 56 and 57. Because of the effect of residual astigmatism at high magnification (30,000) on the alignment

145 Figure 55: Isotactic polystyrene Mw-1,200,000 amorphous film cast from dichlorobenzene solution. Platinum snadowed at 30~. Figure 56: Atactic polystyrene Mw=1,800,000 stretched 100% on mylar. Platinum shadowed at 30. Stretch direction is horizontal.v

Figure 57: Atactic polystyrene Mw=1,800,000 stretched 100% on mylar. Gold decorated. Stretch direction indicated by large arrow. Small arrows.hnw alignment of gold particles. Fiagure 58: At actic polystyrene annealed 30 minutes o at 220C, ice water quench. Platinum shadowed at 30

147 of closely spaced particles, it was difficult to clearly establish with any degree of certainty whether or not platinum shadowed microstructures tended to align preferentially, Figure 56. In stretched films decorated with gold, however, the gold particles did show a definite tendency to align at angles ranging from 30~ to 70~ to the stretch direction, Figure 57. It was established, by means of through focus micrographs, that this alignment of gold particles is independent of astigmatism. In bulk samples of atactic polystyrene annealed at 2200C for 30 minutes and quenched in ice water, a 0 nodular structure on the order of 150A is observed, Figure 58. Similarly prepared samples annealed under the same conditions and quenched in ethanol-dry ice mixtures (-700C), however, do not show this structure. Figure 59. In other experiments using similar samples annealed at a lower temperature (190~C), ice water quenching does not generate this same nodular structure observed at higher annealing temperatures, Figure 60. Thus, the appearance of an enlarged microstructure under one set of conditions, suggests that isolation and examination of an amorphous microstructure in general is highly dependent upon the experimental conditions used. Thin sections of commercial atactic polystyrene (sectioned at room temperature) were shadowed with platinum parallel to the cutting direction and examined in the electron microscope, Figure 61. The column of segmented dark areas (parallel to the arrow) is thought to be an artifact generated

148 FIgureo59: Atactic polystyrene annealed 30 minutes at 220C, ethanol-dry ice quench. Platinum shadowed o-t 30. Figure 60: Atactic polystyrene annealed 30 minutes at 190C, ice water quench. Platinum shadowed at 30.

i149 Fi gure 61 Atactic polystyrene thin section* Platinum shadowed at 30. Cutting direction is verticle "~~ ~~~~, 4 ~ ~~~~~~~~~~~~~s 4 *44~~~~~~~~~~~~~~~~~~~~~ Figure 62: Istactic polystyrene crynsetalize a 10 fr2miue.Platinum shadowed at 30~ ~1 V:-~ ui;r 4'x~i p~~~u~:"i~; L~.~*.:~~rr~~~R U. N~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~: Nt~~~~~~~-ti6s:ir:c~- rcpt~ r~ X~''=Pis~r~l":~~.V 8" Wh " i"Stvs ~ 3a ~ ~ ~ ~ ~~~~~:,?~ ~ic: *~ 2~~:,~e~isl a~"ri~~,~~lt*~.;b$ ~ ~ ~S~h:'g 1. 71.i::%~*, ~ ~ ~.~*/:~, Q~:t~ t~n";:l~~' " ~ gIM dr? ~~~;~ ""~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~o ZI~-i~~i ~ ~,r~~~00, 9,;~~r~~~ ~t*-""~~-.,a~f7 A;~: *:;: 6 ~~:Z?i'k 79~~~~~~~~~~~~& s* X1-~ r~~dM. Li-;h-*?., -~a?~ ~~~~~~~t,,.~~~;~~;v *. -.4-A ht7~

150 by the glass knife blade as it cuts through the sample. Dark striations perpendicular to the knife blade artifact are probably caused by compressive deformation of the thin section, as indicated, by Sjostrand's [54] discussion on the mechanism of thin sectioning. The background platinum 0 shadowed structure is on the order of 30A, approximately the same size as observed in shadowed thin films of atactic polystyrene, Table XIII. 2. Crystallization of Unoriented Isotactic Polystyrene from the Glassy Amorphous State When thin films of amorphous isotactic polystyrene are annealed 35~C to 40~C above the glass transition temperature Tg for 10 to 15 minutes, dark fibrous structures characteristic of spherulitic crystallization start to form on the film surface. Crystalline thin films of isotactic polystyrene shadowed with platinum show the splaying fibrous bundle structure observed in spherulitic crystallization of other polymers [17], Figure 3. Examination of these fibers at high magnification indicate that the platinum shadowed microstructure does not preferentially align itself into a crystallograph orientation, Figure 62. The thickness of the fibers in thin films crystallized at 1400C ranges from O 110 to 120A, which is in general agreement with the small 0 angle x-ray long period of 127A reported by Manley and Blais [43] for bulk isotactic polystyrene crystallized at 140~0C. The typical electron diffraction pattern for the area

151 0 in Figure 3 has concentric rings with Bragg spacings of 11.02A o' 0 0 0 0 6.41A, 5.56A, 4.80A, 4.12A, and 4.04A, which correspond to the (110), (300), (220), (211), (410) and (311) planes of isotactic polystyrene respectively, Figure 63. Gold decoration of films partially crystallized at 125~C shows the splaying fibrous structure with a tri-layer vacated area measuring 55A, Figure 64. A more detailed discussion of the effects of gold decoration can be found in Chapter V. Quite often it is difficult to observe fine crystalline structure in thin films because the overlying uncrystallized amorphous material partially obscures it. Therefore, efforts were made to use selective etching techniques to preferentially remove this overlying amorphous material. Figures 65 and 66 show a thin film of isotactic/atactic polystyrene 50/50 crystallized at 1250C and selectively etched with amyl acetate before shadowing with platinum. The irregularly shaped structures appearing on the surface 0 of crystalline lamallae measure 120-150A diameter, while the background regions adjacent to the lamallae do not have any nodular structure. In other crystalline thin films prepared under similar conditions, nodular structures 0 ranging from 140 to 170A primarily appear along the backbone edge of crystalline fibers, Figure 67. These nodular structures along the edge or the surface of crystalline lamallae, however, are not observed in all etched films. When more dilute mixtures of isotactic/atactic polystyrene 10/90 are crystallized at 145~C and selectively etched with amyl acetate, the resulting crystalline fibers

152 Figure 63: Isotactic polystyrene crystallized at 140C for 20 minutes. Electron diffraction pattern. Prints show pattern developed at different intensities. Figure 64: Isotactic polystyrene/at actic polystyrene 50/50 crystallized at 125C for 60 minutes. Gold decorat ed.

153 RER Fi gure 65: Isotactic polystyrene/atactic polystyrene 50/50 crystallized at 125C for 120 minutes, amyl acetate etch for 20 seconds. Platinum shadowed at 30*. Figure 66: Isotactic polystyrene/atactic polystyrene 50/50 crystallized at 12~C for 120 minutes, amyl acetate etch for 20 seconds. Platinum shadowed at 30.

154 Figure 67: Isotactic polystyrene/atactic polystyrene 50/50 crystallized at 125C for 60 minutes, amyl o acetate etch for 20 seconds. Platinum shadowed at 30. Figure 68: Isotactic polystyrene/atactic polystyrene 10/90 crystallized at 14~C for 60 minutes, amyl acetate etch for 20 seconds. Platinum shadowed at 30.

155 appear to be made up of stacks of lamallar ribbon structures which have no surface nodules, Figure 68. The thickness of O these ribbon lamallae measures 65-75A, which is somewhat 0 larger than the value of 60A reported by Keith [29] for the thickness of single crystals of isotactic polystyrene grown from mixtures of isotactic/atactic polystyrene at 1100C. It is uncertain at the present time why we find nodular structure in some etched films, but not in others. The nodules themselves might be caused either by preferential solvent attack in between primary structural units or perhaps by reprecipitation of partially dissolved polymer molecules. D. Discussion Amorphous thin films of atactic and isotactic polystyrene have a platinum shadowed microstructure on the order 0 of 30A, which appears to be essentially invariant with respect to the experimental conditions examined (molecular weight, solvent type, annealing temperature), Table XIII. In subsequent orientation studies on these thin films, it could not be clearly established whether these structures tend to preferentially align under tensile stress. In addition, these microstructures do not appear to align crystallographically in spherulitic fibers of crystallized 0 isotactic polystyrene. These results suggest that the 30A platinum shadowed microstructure observed in amorphous polystyrene might not be comparable to the 75A paracrystalline ball-like structure observed by Yeh and Geil [62] in

156 0 poly(ethylene terephthalate) and the 125A structures found by Carr and Geil [12] in polycarbonate A. Furthermore, there is no clear evidence linking this 30A microstructure to the O0 60A structure observed by Schoon and Teichman [50] in amorphous polystyrene and poly(methyl methacrylate). There0 fore, the proper interpretation of this 30A structure is still uncertain and is further complicated by the fact that the ultimate resolution of the platinum-carbon shadowing method, according to the results of Bradley [8], is also 0 in the 20 to 30A range. Larger nodular structures, however, are observed in bulk samples of commercial atactic polystyrene annealed at 2200C and quenched in ice water. Similarly prepared samples quenched in ethanol-dry ice mixtures did not show this structure. Other samples of bulk atactic polystyrene annealed at 1900C and quenched in ice water likewise did not show this structure. At the present time, we have no reasonable explanation of why we should see this nodular structure in bulk samples annealed at higher temperatures (220~C) but not in samples annealed at lower temperatures (190~C). As mentioned earlier, these results suggest that successful isolation and examination of amorphous microstructure is highly dependent upon experimental conditions used. In amorphous thin films of isotactic polystyrene crystallized from the glassy amorphous state, no surface aggregation of structural units is observed prior to the formation of spherulitic fibers. In fully crystallized

157 thin films, the only identifable structural units are the 0 spherulitic fibers and the 30A platinum shadowed microstructure. When these films are selectively etched, however, 0 surface nodular structures on the order of 120-170A are often observed. These nodules are quite irregular in shape and only occur on the surfaces of crystalline structures. The origin of these nodular structures is uncertain, but it might be related either to the presence of primary block units in crystalline fibers or the reprecipitation of partially dissolved molecules. E. Conclusions 1. A platinum shadowed microstructure on the order of 30A is found in all amorphous thin films of isotactic and atactic polystyrene examined. The size of this structure appears to be essentially unchanged with respect to the various conditions studied. In addition, these microstructures do not appear to align crystallographically in spherulitic fibers of crystallized isotactic polystyrene. Thus, the origin of these structures cannot be clearly established at the present time. 2. Spherulitic fibers of isotactic polystyrene have the same type of lamallar ribbon structure found in spherulitic fibers of other crystallizable polymers. 3. After annealing at 2200C for 30 minutes, nodular structures on the order of 150A are observed on the surface of bulk samples of ice water quenched atactic polystyrene. 4. Irregularly shaped nodular structures, ranging from

158 0 120-170A diameter, are often found in crystalline films of isotactic/atactic polystyrene which have been selectively etched with amyl acetate. Such nodules are thought to be either primarily structural units within the crystalline fibers or else the reprecipitation of partially dissolved molecules.

CHAPTER VII GENERAL CONCLUSIONS AND MAJOR FINDINGS The purpose of this study is to investigate the effect of a noncrystallizable diluent on the spherulitic crystallization kinetics and morphology of isotactic polystyrene and to examine strain induced crystallization from the glassy amorphous and rubbery states of similar blends. The significant findings and major conclusions are discussed in the following section. In the kinetics study for mixtures of isotactic/atactic polystyrene, the measured spherulitic growth rates and morphology are generally quite similar to those reported by Keith and Padden [31]. However, for 40% and 60% dilutions with atactic polystyrene, the growth rate shows an unexpected, yet significant rise, as the diluent molecular weight increases from 19,800 to 51,000. This anomalous growth rate behavior cannot be explained on the basis of normal viscosity or diffusional effects, but can be reasonably explained in terms of the phenomenological theory of Keith and Padden [30] and the chain entanglement effect noted by Fox and Flory [15]. Since the critical molecular weight for chain entanglements in polystyrene is 36,000, we suggest that atactic impurity molecules having molecular weights 159

160 greater than 36,000, are substantially entrapped within the growing spherulite by means of these chain entanglements. As a result, the interfacial concentration of rejected impurity Would be proportionally reduced, thus causing a noticible rise in the spherulitic growth rate. In the kinetics study for the plasticized mixtures, a direct comparison of the spherulitic growth rate for mixtures of isotactic/atactic polystyrene and isotactic polystyrene/di-decyl phthalate over the same concentration and diluent molecular weight range indicates that there is a basic difference in the microstructure of these two systems. It is suggested that the former mixture behaves as a "homogeneous" phase since the growth rate is montonically depressed with increasing concentrations of atactic polystyrene. The latter mixture, however, appears to behave as a two phase system, both with regard to the measured spherulitic growth rate and the melting temperature depression. Thus, since the maximum spherulitic growth rate for the plasticized mixture is much higher than the comparable rate for the atactic mixture, the creation of a two phase system apparently affects either the viscous transport mechanism or the free energy of critical nucleus formation in the spherulitic growth process. According to an analysis of the threshold crystallization temperatures for mixtures of isotactic/atactic polystyrene and isotactic polystyrene/benzophenone, the empirical constants C1 and C2 in the WLF form of the viscous transport term in the growth rate equation takes on different values

161 for these two mixtures. If we consider C to be invariant 1 for both systems, then C2 is significantly lower for the benzophenone system, according to the methods of Boon and Azcue [5] and Suzuki and Kovacs [56]. The magnitude of this constant (C2), however, is dependent upon which method is used. If the derived form of the spherulitic growth rate equation (equation 4) is correct, then the free energy of critical nucleus formation must also be lower for the benzophenone system in order to explain the higher maximum growth rates observed. Thus, it is suggested that the spherulitic growth mechanism for plasticized systems involves the addition of whole structural units (with a lower free energy of critical nucleus formation) to the growth front, while the mechanism for the atactic polystyrene system is a nonstructural nucleation process involving the addition of single molecules or small groups of loosely bound molecules to the growth front. This is the first time that an analysis of this type has ever been made on a polymer/diluent system. In the study of strain induced crystallization of isotactic polystyrene, three different types of morphology O O were observed. Fibril type structures 130A to 150A wide and aligned parallel to the stretch direction were found in thin films of isotactic polystyrene/benzophenone 60/40 stretched 400% to 500% and annealed at 1750C from the rubbery amorphous state. The electron diffraction pattern corresponding to these fibrils indicates a high degree of molecular alignment parallel to the stretch direction, while dark field studies

162 using the (102) reflection indicate that the internal structure of these fibrils is segmented. This latter result suggests that the molecular conformation in the fibrils is predominately folded rather than extended chain. The segmented internal structure of these fibrils agrees with the experimental data of Yeh and Geil [61] and Wikjord and Manley [59] for different polymer systems strain crystallized under different conditions. These combined results suggest that the extended chain line nuclei model proposed by Keller [49] needs to be modified to account for a substantial number of chain folds in the proposed line nuclei. This is the first time that fibril type structures have been 0brtvedc in strain induced crystallization from the rubbery amorphous state near Tg. In thin films stretched to lower elongations (300%) prior to crystallization from the rubbery amorphous state, extended column type structures were often observed. These column structures are made up of rows of crystalline fibers 0 0 120A to 160A thick, oriented perpendicular to a common line type nucleus. A central filamentary backbone type structure (100 to llOA diameter) has been found in mildly etched column type structures. This finding suggests that line type nuclei might be responsible for the generation of column type structures, a model proposed by Keller [49]. Filamentary backbone structures of this type have never been found before in column type structures produced by strain induced crystallization of a polymer from the rubbery amorphous state.

163 In thin films strain crystallized at very low elongations (100%), small groups or bundles of fibers aligned more or less perpendicular to the orientation direction appear to develop independently of any clearly distinguishable common line nucleus. This result suggests that there might be a mechanism for strain induced crystallization (oriented crystallization) other than the line nucleation model proposed by Keller [49].

CHAPTER VIII, RECOMMENDATIONS FOR FUTURE STUDY 1. The microscopic phase separation observed in plasticized mixtures can be independently verified by using small angle x-ray scatter to determine whether or not isotactic polystyrene and di-decyl phthalate form separate phases over concentrations ranging from 20% to 40% plasticizer. In addition, small angle x-ray techniques can be used to determine if phase separation is also present when other plasticizers are used. If this should be the case, then the proposed difference in the nucleation mechanism for crystallization also could be independently established, if it is verified by small angle x-ray that the isotactic/atactic polystyrene system is "homogeneous". 2. The proposed difference in the microstructure of the isotactic/atactic polystyrene and the isotactic polystyrene/plasticizer systems can be partially verified by using either an atactic polystyrene diluent in the 400 molecular weight range or a phthalic acid ester having a 900 molecular weight as diluents in the comparative spherulitic growth rate studies. 3. The differences in crystalline morphology caused by the addition of a noncrystallizable diluent to a crystallizable polymer could be studied by using crystalline thin 164

165 films which are selectively etched. The impurity should be much more sensitive to the etching solvent than is the crystalline polymer. 4. The effect of chain entanglements on the interfacial concentration of rejected impurity could be partially verified by using poly a-methyl styrene as the impurity diluent. The slower radial diffusion of poly a-methyl styrene should cause a greater depression of the radial growth rate over the critical molecular weight range than was noted for atactic polystyrene diluent. 5. Dark field microscopy using a beam tilting device might provide us with a better understanding of the internal structure of the fibrils observed in highly stressed thin films of isotactic polystyrene. In addition, comparable strain induced crystallization studies using amorphous polycarbonate A would help to show a) whether or not such fibril structures are found in other polymers, strain crystallized from the glassy amorphous state; and b) whether or not such fibril structures are formed by alignment of the nodules found in amorphous polycarbonate A, thus causing a segmented structure. 6. Measures of the spherulitic growth rate for mixtures of isotactic/atactic polystyrene of various molecular weights at temperatures other than 1800C would help to provide us with a better basis for suggesting modifications to the theoretical growth rate equation.

APPENDIX Tab le IX Spherulitic Growth Rate Data for Mixtures of Isotactic/Atactic Polystyrene Tx, ~ C AM R G Time, min./Size, p 180 100/0 0.383 48 38 121 145 164 198 18.2 32.3 47.7 56.0 62.9 76.2 180 100/0 0.389 48 73 93 125 16.3 26.8 34.5 45.1 190 100/0 0.284 59 95 133 160 194 16.7 26.2 37.3 45.9 54.7 190 100/0 0.295 78 103 126 150 180 25.6 33.4 42.2 47.8 55.1 200 100/0 0.157 214 249 281 315 357 9.3 14.6 19.5 24.1 32.0 170 100/0 0.339 67 95 134 146 166 189 22.4 32.1 45.8 50.7 57.0 64.6 160 100/0 0.291 103 130 148 176 30.5 38.2 43.7 52.9 140 100/0 0.131 60 95 140 183 221 8.1 12.7 19.4 24.0 29.7 180 4800 80/20 0.293 54 78 98 154 14.6 21.1 27.2 44.5 180 4800 80/20 0.293 96 121 148 181 211 235 264 12.9 19.7 29.3 38.4 46.7 53.3 61.9 190 4800 80/20 0.187 71 99 119 142 176 191 215 13.2 19.7 22.4 25.1 32.4 35.8 41.1 200 4800 80/20 0.081 123 158 186 216 246 7.0 9.9 12.4 15.3 17.1

167 Tx, ~C Mw R G Time, min./size,' 170 4800 80/20 0.298 75 101 129 151 172 196 7.8 15.7 24.2 28.9 47.6 46.4 160 4800 80/20 0.249 107 130 155 183 200 24.3 30.1 36.5 41.3 48.7 140 4800 80/20 0.108 163 202 238 280 16.7 20.5 25.2 29.9 180 4800 60/40 0.211 102 137 171 200 233 19.6 27.4 35.3 41.6 48.1 190 4800 60/40 0.160 214 285 333 340 360 15.8 27.0 35.9 36.5 40.2 170 4800 60/40 0.192 73 101 128 152 188 211 14.1 18.5 23.4 29.7 38.2 43.9 160 4800 60/40 0.165 75 114 146 181 212 250 6.2 14.9 20.2 25.1 30.9 35.7 140 4800 60/40 0.051 98 132 174 202 5.2 7.1 9.5 10.3 180 4800 40/60 0.151 108 143 179 207 243 12.2 16.4 22.2 26.1 31.8 180 4800 40/60 0.163 30 42 53 67 79 94 5.1 7.3 8.6 11.3 13.2 16.8 190 4800 40/60 0.083 212 248 282 320 352 5.4 8.7 11.2 15.1 17.4 190 4800 40/60 0.099 199 233 285 328 14.3 15.7 22.1 27.0 170 4800 40/60 0.111 80 108 145 171 200 230 9.2 12.1 15.7 19.2 22.0 26.8 160 4800 40/60 0.089 94 112 159 204 240 6.9 8.1 10.4 14.2 18.1 190 4800 20/80 0.038 54 77 121 152 1.9 2.7 4.4 5.9 180 4800 20/80 0.092 77 105 111 133 157 170 7.2 9.9 10.7 12.2 14.9 16.3 170 4800 20/80 0.065 55 82 101 117 3.5 5.4 6.7 7.6

168 Tk, ~C Mw R G Time min./size 160 4800 20/80 0.036 115 162 207 243 6.4 8.7 9.5 10.8 180 411,000 60/40 0.213 33 58 85 112 127 136 5.3 10.1 14.0 20.8 22.4 25.7 180 411,000 60/40 0.217 39 52 69 81 97 110 6.3 9.6 13.4 15.9 18.7 21.2 180 411,000 40/60 0.148 47 65 87 116 129 6.1 8.4 12.8 16.1 18.3 180 411,000 20/80 0.090 61 93 119 140 161 5.4 8.7 10.3 12.1 13.9 180 51,000 80/,0 0.293 25 41 57 71 91 101 8.4 12.7 16.9 21.7 24.5 27.1 180 51,000 60/40 0.250 23 37 51 63 73 85 6.3 9.7 12.9 16.7 19.4 20.2 180 51,000 60/40 0.248 18 30 40 50 60 70 4.1 7.8 10.7 12.1 15.9 17.7 180 51,000 40/60 0.182 22 33 47 60 80 5.7 7.3 10.3 13.1 16.9 180 51,000 li0/60 0.180 25 38 48 58 70 80 4.3 6.5 8.4 10.2 11.4 14.1 180 51,000 20/80 0.106 35 58 66 97 136 3.5 6.6 6.1 10.7 11.3 180 51,000 20/80 0.098 56 78 100 136 174 6.4 8.7 11.0 14.3 15.6 180 19,800 80/20 0.290 16 30 45 62 78 5.3 9.2 13.9 18.4 22.5 180 19,800 60/40 0.180 17 30 42 52 62 72 4.2 6.1 8.7 10.3 12.4 14.1 180 19,800 60/40 0.178 25 37 47 57 67 77 6.2 8.1 10.9 12.4 13.3 15.4 180 19,800 40/60 0.143 26 39 49 59 69 80 5.8 7.2 8.9 10.7 11.5 13.6 180 19,800 40/60 0.143 18 28 38 48 58 68 3.6 4.2 6.9 7.4 9.5 10.6

169 Tx, ~C Mw R G Time, min./size, p 180 19,800 20/80 0.075 40 59 64 76 3.7 4.3 4.3 4.4 180 10,000 60/40 0.193 19 30 40 50 60 70 4.2 6.7 9.8 11.2 12.9 14.7 180 10,000 40/60 0.148 27 36 46 56 66 3.8 5.4 6.7 8.1 9.5 180 2,030 80/20 0.288 20 30 41 50 60 6.2 9.8 12.7 15.3 17.9 180 2,030 60/40 0.195 20 30 40 50 60 70 5.2 7.3 9.8 11.2 13.4 15.7 175 2,030 60/40 0.193 25 35 45 56 6.3 8.0 10.2 12.8 180 2,030 40/60 0.118 40 50 61 70 80 5.7 6.1 8.7 9.5 9.5 180; 900 80/20 0.285 16 30 38 49 54 4.6 7.3 10.7 12.2 13.8 170 900 80/20 0.290 26 36 46 60 76 8.2 11.9 14.3 18.8 23.6 175 900 80/20 0.293 83 93 108 119 130 22.7 25.4 29.3 32.8 36.4 175 900 60/40 0.150 19 29 48 60 70 80 14.1 5.3 8.5 10.2 11.8 13.7 180 900 60/40 0.110 100 110 120 16.2 17.1 18.0 175 900 40/60 0.097 71 82 95 106 5.0 6.3 7.5 7.7 170 900 40/60 0.092 19 38 53 80 2.1 4.9 5.4 7.8 175 900 20/80 0.021 62 86 121 1.3 1.8 2.5 180 1o00,000 80/20 0.268 18 28 38 50 60 70 6.1 8.7 11.3 15.4 17.3 19.8 190 1~00,000 60/40 0.160 24 34 46 69 69 4.1 6.4 7.7 9.8

170 Tx oC M_ R G Time, min./size_ 180 1,800,000 60/40 0.178 47 65 85 109 132 9.1 11.3 15.5 19.8 23.4 190 1,800,000 40/60 0.092 26 42 59 70 85 3.1 3.8 6.2 7.5 7.8 180 1,800,000 40/60 0.147 57 78 94 118 137 160 9.1 12.7 14.8 17.9 20.3 24.5 180 1,800,000 20/80 0.075 36 53 72 89 2.2 3.8 5.1 6.2 Nomenclature T,~C = Crystallization temperature, OC M = Molecular weight (number average) RW = Ratio of IPS/APS G = Spherulitic growth rate, i/minute Time = Observed crystallization time, minutes Size = Observed size of spherulites at any given time, I

171 Table X Spherulitic Growth Rate Data for Mixtures of Isotactic Polystyrene and Di-methyl Phthalate Tx, ~C R G Time, min./size,,u 130 90/10 0.187 10 15 25 36 4.1 5.1 7.0 9.0 140 90/10 0.417 9 14 21 26 32 5.9 8.1 10.9 12.9 15.5 150 90/10 0.580 12 17 23 29 34 9.8 12.2 15.9 19.2 22.2 160 90/10 0.595 6 13 18 23 28 33 4.2 8.1 11.5 14.3 17.4 20.3 170 90/10 0.438 9 14 19 24 29 4.2 6.0 8.3 10.3 12.9 120 80/20 0.270 5 10 15 20 31 40 5.8 6.9 8.4 9.6 12.9 15.7 130 80/20 0.537 10 15 20 25 30 35 8.8 11.5 14.5 16.9 19.3 22.1 140 80/20 0.725 10 15 20 25 30 36 7.8 12.1 15.7 19.5 23.0 26.9 150 80/20 0.695 4 9 14 19 25 30 4.6 8.5 11.9 15.2 18.7 21.9 160 80/20 0.560 6 11 16 21 26 4.1 7.0 9.5 12.5 15.4 110 70/30 0.498 6 11 16 21 26 33 7.8 9.4 11.0 14.6 17.1 20.4 120 70/30 0.725 7 12 17 22 28 8.4 12.4 15.8 19.7 24.1 130 70/30 0.716 7 12 17 22 19 7.3 11.2 1 8.1 23.4 140 70/30 0.604 7 12 18 23 28 37 7.6 11.7 14.1 17.5 20.7 26.4 150 70/30 0.287 16 23 28 33 46 7.1 9.4 10.8 12.2 13.5 16.2

172 Tx, oC R G Time, min./Size, ~ 103 60/40 0.635 3 8 13 18 24 6.6 9.6 12.9 16.0 19.6 110 60/40 0.648 10 15 20 26 32 10.2 13.4 15.8 20.9 25.0 120 60/40 0.560 4 9 14 19 24 29 7.3 10.2 10.9 13.8 16.7 19.6 Nomenclature Tx,C = Crystallization temperature, ~C R = Ratio IPS/DMP G = Spherulitic growth rates,P/minute Time = Observed Crystallization Time, minutes Size = Observed size of spherulites at any given time, i

173 Table XI Spherulitic Growth Rate Data for Mixtures of Isotactic Polystyrene and Di-decyl Phthalate Tx, oC R G Time min./Size 140 90/10 0.459 5 11 17 29 5.2 7.9 10.4 16.3 150 90/10 0.726 6 11 16 21 26 31 8.8 12.3 15.8 19.5 23.0 27.2 160 90/10 0.847 4 9 14 19 24 29 4.8 9.2 12.9 17.4 21.8 26.1 170 90/10 0.720 7 12 17 22 6.4 9.6 13.6 16.7 170 90/10 0.720 5 10 15 20 25 30 4.6 8.4 11.6 15.4 18.6 21.5 185 90/10 0.263 15 25 35 45 4.5 6.8 10.1 12.1 120 80/20 0.780 3 8 13 18 23 9.1 12.2 16.4 20.6 24.6 130 80/20 1.000 7 12 17 23 29 34 14.4 19.7 24.7 31.2 37.1 41.7 140 80/20 1.100 5 10 15 20 26 32 9.7 14.7 19.8 26.4 33.0 38.8 140 80/20 1.105 6 11 16 20 25 30 35 7.8 10.9 14.6 24.5 29.9 35.4 40.2 150 80/20 1.092 8 13 19 24 30 9.5 15.6 21.4 26.9 33.5 160 80/20 0.905 7 12 18 24 31 36 6.9 11.1 15.7 21.0 26.3 29.8 175 80/20 0.400 14 \20 25 31 9.5 10.7 12.5 15.3 110 70/30 0.440 16 21 27 33 40 47 8.5 10.4 12.6 16.2 19.4 23.1 120 70/30 0.710 5 10 15 20 25 n 8.3 12.2 15.7 19.1 22.5

174 Tx ~C R G Time, min./Size, 2 130 70/30 0.855 4 9 14 19 25 6.0 10.6 14.4 17.7 21.4 140 70/30 1.070 5 10 15 20 25 30 7.6 14.2 20.0 24.9 30.9 36.3 140 70/30 1.050 6 11 16 22 29 35 8.6 13.8 17.6 21.0 29.6 35.5 140 70/30 1.020 11 16 22 27 33 10.7 15.8 21.2 26.7 32.4 150 70/30 1.100 4 9 14 19 24 8.4 13.7 19.0 26.1 30.7 175 70/30 0.638 11 16 21 26 7.2 10.6 13.7 16.7 110 60/40 0.430 2 7 12 17 22 27 5.8 7.7 10.5 11.9 13.9 16.1 140 60/40 1.030 7 12 17 22 27 10.6 15.2 20.7 26.0 30.9 150 60/40 1.100 5 10 15 20 25 30 8.1 13.9 19.3 24.6 30.6 35.9 160 60/40 1.015 7 12 17 24 29 6.5 11.6 15.2 20.2 23.2 175 60/40 0.640 11 16 22 27 32 6.4 9.7 13.4 16.6 20.2 140 40/60 0.873 8 13 18 23 7.5 12.3 16.4 20.9 150 40/60 0.870 27 32 37 42 47 18.4 22.4 26.9 30.2 33.5 160 40/60 0.650 7 12 17 22 4.5 7.8 11.0 14.2 130 20/80 0.417 16 21 27 32 12.4 14.8 17.8 -- 140 20/80 0.570 10 15 20 26 5.7 8.6 10.5 11.5

175rTx, _~C R G me, r.?,.) ze, n/ 150 20/80 0. 595 38 43 50 24.0 26.9 29.5 160 20/80 0. 487 i 6 21 26 31 9.8 13.0 14.6 17.2 19.7 Tx,"~C= Crystal.iz.Atj ct..)-, temperat-ure, 0C R = Ratio of IPS/D)DP G = Spherulitic Growth Rate,Wi/minute Time = Observed Crystall.izat-izon Time, minutes Size = Observed Size of SphleruLites at any given time, V1

176 Table XII Unshadowed Microstructure Material MW Solvent T t S C 1-4 trans - benzene 26 - 18 25 isoprene syndiotactic 108,000 xylene 26 - 16 23 PMMA polystyrene 4,800 cyclohexanone 26 - 20 25 polystyrene 4,800 cyclohexanone 110 0 20 26 polystyrene 4,800 cyclohexanone 110 60 19 26 polystyrene 4,800 dichlorobenzene 26 - 19 24 polystyrene 1,800,000 cyclohexanone 26 - 21 25 polystyrene 1,800,000 cyclohexanone 110 60 20 26 polystyrene 1,800,000 dichlorobenzene 110 0 18 24 polystyrene 1,800,000 dichlorobenzene 110 60 19 24 polystyrene 1,800,000 dichlorobenzene 10 120 20 25 1-4 cis - benzene 26 - 18 22 isoprene isotactic 550,000 dichlorobenzene 26 - 18 24 polystyrene isotactic 550,000 dichlorobenzene 140 60 21 22 polystyrene poly(vinyl - methyl ethyl 26 - 19 26 chloride) ketone Nomenclature M = molecular weight TW = casting temperature,~C t = annealing time, minuteB S - diameter of particle, A C = center to center distance between particles, A

Table XI.Ti Platinum Snhad.owed Miicrostructure Material Y Solventf T t S C 1-4 trans - benzene 26 - 25 30 isoprene polystyrene,8O00 cyclohexanone 26 24 32 polystyrene 1,80C0000 cyclohexanone 26 25 32 1-4 cis - benzene 26 - 26 33 isoprene isotactic 550,000 dlchlorobenzene 1.40 0 24 30 polystyrene isotactic 550,000 dichilorobenzene 140 60 28 33 polystyrene polystyrene 26 25 32 thin section N' orenric lature M = molecular weight w oC T = casting temperature, ~C t = annealing time, minutes S = diameter of particle, 0 o C = center to center distance between particles, A

BIBLIOGRAPHY 1. Andrews, E.H., "Crystalline Morphology in Thin Films of Natural Rubber, II Crystallization under Strain," Proceeding of the Royal Society, A Vol. 277, pp. 562-570 (1964). 2. Banks, W.; Gordon, M.; Sharples, A., "The Crystallization of Polyethylene after Partial Melting," Polymer, 4(3), pp. 289-302, (1963). 3. Bassette, G.A.; Blundell, D.J.; Keller, A., "Surface Structure of Polyethylene Crystals as Revealed by Surface Decoration, I. Preliminary Survey," J. of MacromoZecuZar Science (Physics), Bl(l), pp. 161-184, (1967). 4. Blais, P.; Manley, R. St. John, "Crystallization of Isotactic Polystyrene from Solution," Technical Report No. 460, Pulp and Paper Research Institute of Canada, Montreal, Canada. 5. Boon, J.; Azcue, J.M., "Crystallization Kinetics of Polymer-Diluent Mixtures. Influence of Benzophenone on the Spherulitic Growth Rate of Isotactic Polystyrene," J. of Polymer Science, Part A-2, Vol. 6, pp. 885-894 (1968). 6. Boon, J.; Challa, G.; Van Krevelen, D.W., "Crystallization Kinetics of Isotactic Polystyrene. II. Influence of Thermal History on Number of Nuclei," J. of Polymer Science, Part A-2, Vol. 6, pp. 18351851 (1968). 7. Boon, J.; Challa, G.: Van Krevelen, D.W., "Crystallization Kinetics of Isotactic Polystyrene. I. Spherulitic Growth Rate," J. of PoZymer Science, Part A-2, Vol. 6, pp. 1791-1801 (1968). 8. Bradley, D.E., "Study of Background Structure in Platinum/Carbon Shadowing Deposits," British J. of Applied Physics, Vol. 11, p. 506 (1960). 9. Buchanan, D. R.; Miller, R.L., "X-ray Line Broadening in Isotactic Polystyrene," J. of AppZied Physics, Vol. 37, No. 11, pp. 4003-4012 (1966). 178

179 10. Bueche, F., PhysicalZ Properties of PoZymers, Interscience Publishers, New 1York (1962). 11. Busse, W.F., "Mechanical Structure in Polymer Melts. II. Roles of Entanglements in Viscosity - and Elastic Turbulence," J. of Polymer Science, A-2, Vol. 5, pp. 1261-1281 (1967). 12. Carr, S.H.; Geil, PF.; Baer, E., "The Development of Spherulites from Structural Units in Glassy Poiy(bisphenol-A-carbonate)," J. of Macromolecular Science (Physics), B2(l), 13, (1968). 13. Clark, E.S.; &arber, C.A., "The Effect of Industrial Processing on the Morphology of Crystalline Polymers," A.I.Ch.E. Meeting,.Atlanta, Georgia, February (1970). 14. Dedeurwaerder, J.F.; Oth, M., "Properties of Mixtures of Polystyrene and Benzophenone," J. of Chim. Physics, Vol. 56, p. 940 (1959). 15. Fox, T.G.; i'ory, P. J.,,Second-Order Transition Temperatures and Related Properties of Polystyrene. I. Influence of Molecular Weight," J. of Applied Physics, Vol. 21, PP. 581-591 (1950). 16. Frank, W.; Goddar, H.; Stuartg H.A., "Electron Microscopic Investigations on Amorphous Polycarbonate," J. of Polymer Science, Vol. B, pp. 711-713 (1967). 17. Geil, P.H., PoZlymer Single Crystals, Interscience Publishers, New York, (1963). 18. Gezovich, D.M.; Geil, P.H., "Morphology of Plasticized Polyvinylchloride," Division of Macromolecular Science, Case Western Reserve University, Cleveland, Ohio. 19. Gornick, K.; Mandelkern, L., "Effect of Noncrystallizing Component on the Crystallization Kinetics of Polymers," J. of Applied Physics, Vol. 33, No. 3, p. 907 (1962t). 20. Haller, M.H.; Magill, J.H., "Morphology of Polysiloxane Crystallized from the Melt," J. of Applied Physics, Vol. 40, No. 11, p. 4261 (1969). 21. Hay, J.N., "Crystallization Kinetics of Isotactic Polystyrene," J. of PoZymer Science, Vol. A, No. 3, p. 433 (1965). 22. Hay, J.N.; Sabir, M., "Crystallization Kinetics of High Polymers. Polyethylene Oxide-Part I," PoZymer, Vol. 10, pp. 187 (1969).

180 23. Hoffman, J.D.; Weeks, J.J., "Rate Spherulitic Crystallization with Chain Folds in Polychlorotrifluorethylene," J. of AppZied Physics, Vol. 37, p. 1723 (1962). 24. Holland, V.; Lindenmeyer, P., "The Morphology and Crystal Growth Rate of Polyethylene Complexes," J. of Polymer Science, Vol. 57, p. 589 (1962). 25. Jenckel, E.; Teege, E.; Hinrichs, W., "Transkristallisation in hockmolekularen Stoffen," KoZZoid, Vol. 129, p. 19 (1952). 26. Kargin, V.A.; Koretskaya, T.A.; Bogayevskaya, T.A., "Crystallization of Isotactic Polystyrene," Polymer Science USSR, Vol. 6, No. 3, pp. 489-494 (1964). 27. Kargin, V.A.; Sogolava, T. I.; Talipor, G.,'Supermolecular Structures of Plasticized and Nonplasticized Crystallin.e Polystyrene," PoZymer Science USSR, Vol. 5, pp. 937-945 (1964). 28. Kargin, VT.AA; Koretskaya, T.A.; Markova, G.S.; Ovchinnikov, Yu. K., "Electron Microscopy and Electron Diffraction Studies of the Spherulites of Isotactic Polystyrene," PolZymer Science USSR, Vol. 8, No. 7, pp. 14'i[0-1417 (1966). 29. Keith, i.D., "On the Relation between Different Morphological Forms in High Polymers," J. of Polymer Science, A, Vol. 2, pp. 4339-4360 (1964). 30. Keith, H.D.; Padden, F.J. Jr., "Spherulitic Crystallization from the Melt. I. Fractionation and Impurity Segregation and their Influence on Crystalline Morphology," J. of AppZlied Physics, Vol. 35, No. 4, pp. 1270-1285 (1964). 31. Keith, H.D.; Padden, F.J. Jr., "Spherulitic Crystallization from the Melt. II. Influence of Fractionation and Impurity Segregation on the Kinetics of Crystallization," J. of AppZied Physics, Vol. 35, No. 4, pp. 1286-1296 (1964). 32. Keller, A., "The Spherulitic Structure of Crystalline Polymers. Part I. Investigations with the Polarizing Microscope," J. of Polymer Science, Vol. 17, pp. 291307 (1955). 33. Keller, A.; Machin, M.J., "Oriented Crystallization in Polymers," J. of Macromolecular Science (Physics), Bl(1), pp. 41-61 (1967). 34. Luch, D., "Strain Induced Crystallization of Rubber," Private Communications.

181 35. Lunak, S.; Bohdanecky, M., "Crystallization Kinetics of Polycarbonate in Solution," J. of PoZymer Science, C, No. 16, p. 103 (1965). 36. Magill, J.H., "Crystallization Kinetics of' Polytetramethyl-P-Silphenylene Siloxane (TMPS) Fractions," J. of Polymer Science, Polymer Letters, Vol. 6, pp. 853-857 (1968). 37. Magill, J.H. "Spherulitic Crystallization StUdies of Poly (tetramethyl-p-silphenylene)-siloxane (TMPS). Part III,' J. of Polymer Science, Part A-2, Vol. 7, pp. 1187-1195 (1969). 38. Magill, J.H., "Spherulitic Crystallization. Part I.'Odd Even' Polyamides: Nylon 56 and Nylon 96," J. of PoZymer Science, Part A, Vol. 3, pp. 1195-1219 (1965). 39. Magill, J.H., "Crystallization of Polyamides. Part II: Nylon 6 and Nylon 66," PoZymer, Vol. 6, pp. 367-371 (1965). 40. Magill, J.H., "Formation of Spherulites in Polyamide Melts: Part III. Even-Even Polyamides," J. of PoZymer Science, Part A-2, Vol. 4, pp. 243-265 (1966). 41. Mandelkern, L., Crystallization of PoZymers, McGrawFill, New York, (1964). 42. Mandelkern, L., "The Crystallization Kinetics of Polymer-Diluent Mixtures: The Temperature Coefficient of the Process," PolZymer, Vol. 5, No. 12, pp. 637-648 (1964). 43. Manley, R. St. John; Blais, J.J., "X-ray Long Periods in Bulk Crystallization Polymers," J. of MacromoZecular Science (Physics), B1(3), pp. 525-566 (1967). 44. Natta, G.; Corradini, R.; Bassi, I.W., "Crystal Structure of Isotactic Polystyrene," Nuovo Cimento SuppZement, Vol. 15, p. 68, (1960). 45. Nielsen, L.E., MechanicaZ Properties of Polymers, Reinhold Publishing Corp., New York, p. 55 (1962). 46. Padden, F.J.; Keith, H.D., "Crystallization in Thin Films of Isotactic Polypropylene," J. of AppZied Physics, Vol. 37, No. 11, pp. 4013-4020 (1964).

182 47. Pennings, A.J.; Kiel, A.M., "Fractionation of Polymers by Crystallization from Solution. III. On the Morphology of Fibrillar Polyethylene Crystals Grown in Solution," KoZZoid Z., Vol. 205, p. 160 (1966). 48. Price, F.P., "Nucleation in Polymer Crystallization," Report No. 67-C-411, General Electric Research and Development Center, Schenectady, New York. 49. Price, F.P., "Polymer Crystallization During Flow," Report No. 67-C-097, General Electric Research and Development Center, Schenectady, New York. 50. Schoon, G.F.; Teichman, 0., "Elektronenmikroskoposche Untersuchung des Quellungsvorgangs bei Makromolekulen," KoZZoid Z., Vol. 157, pp. 35-45 (1964). 51. Schoon, G.F.; Kretschmer, R., "Elektronmikroskopische Untersuchlwn.igen des Quellungsvorgangs bei Makromolekulen 7I Teil: Des System Polymethyl-methancrylat," KoZZoid Z. Vol. 197, pp. 45-50 (1964). 52. Schoon,'1. F.; Kretschmer, R., "Interpretation of Micromorphological Structure," KoZZlloid Z, Polymer Z, Vol. 211, Nos. 1 & 2, pp. 53-62 (1966). 53. Siegman, A.; Geil, P.H., "Crystallization of Polycarbonate from the Glassy State, Part I. Thin Films Cast from Solution," T.R. 127, T.R. 128, Division of Macromolecular Science, Case Western Reserve University. 54. Sjostrand, F.S., Electron Microscopy of Cells and Tissues, Vol. 1, Instrumentation and Techniques, p. 232, Academic Press, New York (1967). 55. Spit, B.J., "Gold Decoration Applied to Polymers," J. of MacromoZecuZar Science (Physics), B2(1), pp. 45-54 (1968). 56. Suzuki, T.; Kovacs, A.J., "Temperature Dependence of Spherulitic Growth Rate of Isotactic Polystyrene," Polymer Journal, Vol. 1, No. 1, pp. 82-100 (1970). 57. Taeger, A.A.; Suvorova, A.I.; Dreval, V. Ye.; Gakova, N.P.; Lutskaya, S.P., "Newtonian Viscosity. and Glass Temperature of Polystyrene Plasticized with Aromatic Acid Esters," Polymer Science USSR, Vol. 10, No. 10, p. 2649 (1968). 58. Turnbull, D.; Fisher, J.C., "The Rate of Nucleation in Condensed Systems," J. of Chemical Physics, Vol. 17, p. 71 (1949).

183 59. Wikjord, A.G.; Manley, R. St. John, "The Morphology of Fibrillar Polyethylene Crystals," J. of Macromolecular Science (Physics), B2(3), pp. 501-537 (1968). 60. Williams, M.L.; Landel, H.F.; Ferry, J.D., "The Temperature Dependence of the Relaxation Mechanism in Amorphous Polymers and Other Glass Forming Liquids," J. of American ChemicaZ Society, Vol. 77, p. 3701 (1955). 61. Yen, G.S.Y.; Geil, P.H., "Strain Induced Crystallization of Polyethylene Terephthalate," J. of Macromolecular Science, (Physics), Bl1(2), pp. 251-277 (1967). 62. Yeh, G.S.Y.; Geil, P.H., "Crystallization of PoLyethylene Terephthalate from the Glassy Ariorphous State," J. of MacromolecuZar Science (Physics), B1(2), pp. 235-249 (1967).

UNIVERSITY OF MICHIGAN 3 90 15 03023 028111111 3II 9015 03023 0281