' o—‘r-q"- — ' THE ,DEPENDEME 0F PHASE EQUILIBRIA ON THE CONFIGURATEON OF POLYF’ROPYLENE Thesis {qr Hm Dog“: of M. S. MICHIGAN STATE UNWERSITY Ritchie A. Wessling 19,59 . , v . ‘ . “-~’ -.———— mfi —. __ - 7- _ A... — ~ L___ _———7 :— __ _M’ , _ "v .‘fb'.!m;ny* ‘...~ .l‘fff.L>'\ ' THE," THE DEPENDENCE OF PHASE EQUILIBRIA ON THE CONFIGURATION OF POLYPROPYLENE by Ritchie A. wessling Submitted to the College of Science and Arts of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1959 VITA The author was graduated from Resurrection High School of Lensing, Michigan in June, 1950. He enlisted in the Uni - ed States Navy three months later. After receiving an honor- able discharge in September, 1954, he entered Ferris Insti- tute in Big Rapids, Xichigan. He remained there for one year before transferring to Michigan State University where he received the Bachelor of Science degree in chemistry in \ June, 1957. ii THE DEPENDENCE OF PHASE EQUILIBRIA ON THE CONFIGURATION OF POLYPROPYLENE by Ritchie A. Wessling AN ABSTRACT Submitted to the College of Science and Arts of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry year - 1959 Approved: ABSTRACT Phase equilibrium studies were made with isotactic and atactic polypropylene to determine the effect of configura- tion on the thermodynamic interaction parameters as defined by Flory's theory for liquid-liquid phase equilibria. Precipitation temperatures were obtained for polyprop- ylene in a number of poor solvents. The isotactic form was found to be more soluble in solvents which gave liquid—liq- uid separation and less soluble in solvents where it separ- ated as a crystalline phase. From the above results, two solvents were chosen for liquid-liquid.phase equilibrium studies. Phase diagrams were obtained for four fractions each of the polymeric isomers. Precipitation temperatures were observed visually. The in- teraction parameters were found to differ for the two poly- mers. For example, for isotactic polypropylene in phenyl ether, W.=1-414 and 9:: 418.40K; for atactic polypropylene, thev are 0.986 and 426.50K, espectively. The difference in the values of the parameters for the two polymers is larger in poorer solvents and appears to converge in better solvents. A unique point was found where the function,£H(l-G/T), is equivalent for isotactic and atactic polypropylene. At tem- peratures above this point, isotactic polypropylene is more soluble, and below it, less soluble than the atactic form. The results are explained on the basis of molecular models. Under these considerations, the differences in the interaction parameters are justified by theory. iv ACKNOWLEDGEMENT The author is indebted to Doctor J. B. Kinsinger for the inspiration and aid he provided during the course of this work. TABLE OF CONTENTS SUBJECT I. II. III. IV. V. VI. VII. Introduction ‘A. Statement of Purpose B. Historical C. Theory Experimental A. Procedure B. Equipment 0. Reagents Results and Discussion A. Solvent Studies B. Phase Diagrams C. Thermodynamic Interaction Parameters D. General Discussion Conclusions Summary ppendix References vi TABLE I. II. III. IV. VI. VII. VIII. IX. LIST OF TABLJS Viscosity data for isotactic polypropylene. Viscosity data for atactic polypropylene. Precipitation temperatures of 1% mixtures of polypropylene in different solvents. Data for isotactic polypropylene in phenyl ether. Data for atactic polypropylene in phenyl ether. Critical data for isotactic polypropylene in phenyl ether. Critical data for atactic polypropylene in phenyl ether. Critical data for polypropylene in the mixed solvent. The thermodynamic interaction parameters. Critical values at the point of intersection of the 9 curves. PAGE l2 l2 18 24 28 28 LIST OF FIGURES FIGURE 1. Sample tube and stirrer. 2. Illustration of the trend in precipitation temperatures. 3. Phase diagrams for isotactic polypropylene in phenyl ether. 4. Phase diagram of isotactic polypropylene in l—octanol ‘ 5. Phase diagrams of atactic polypropylene in phenyl ether. 6. Comparison of experimental binodials to theory in the system, polypropylene- phenyl ether. 7. The deviation ofnéc from theory for polypropylene in phenyl ether. 8. The dependence of To on”x“for the system, polypropylene-phenyl ether. 9. The dependence ofll,on temperature for the system, polypropylene-phenyl ether. 10. The dependence of Tc on'x‘for the system, polypropylene in 60% phenyl ether-40% benzyl ether. 11. The comparison of 7L,andox“at the point of intersection intho solvents. 12. Density of phenyl ether. 13. Specific volume of polypropylene. PAGE 11 17 21 23 25 26 27 29 33 36 46 46 INTRODUCTION I INTRODUCTION Statement o£;Purpose Since the discovery of stereospecific polymerization (l) (2), several vinyl andcx-olefin polymers have been produced with stereoregular structures. Due to different degrees of stereoregularity, these polymers may vary from an amorphous to a highly crystalline material in the solid state. In the disordered state of solution, the difference in prOperties due to configuration is not as apparent. Because interactions between segments of a polymer chain, in solution, are strongest when phase separation takes place, any difference in the properties of a polymer, due to configuration, would be maximized at this point. Therefore, liquid-liquid phase equilibrium was chosen to study the solution properties of isotactic (all d or all 1) and atactic (random d and l) polypropylene. The investiga- tion had a threefold purpose: First, to observe any differ- ences in the thermodynamic interaction parameters; secondly, to interpret the findings in the light of known configura- tional differences; and finally, to evaluate the dilute sol- ution theory for its ability to explain these results. Historical A study of liquid-liquid phase equilibria has been found to be a less difficult method for evaluating the in- teraction parameters because the complex dilute solution theory converges to the general theory of polymer solutions in the critical region (3). In addition, it has been demon- 2 3 strated that this method gives more realistic values for the entropy parameter,tn.(4) No earlier phase studies have been reported for atactic and stereoregular forms of the same polymer. herefore, absolute comparison of polypropylene to other polymers cannot be made. However, polymers which exist only in a regular structure because of symmetry in the monomer, and atactic polymers have been studied. For instance, the work of Richards with polyethylene (5), and Flory with atactic polystyrene (6) can be cited. The polyethylene system exhibited the qualitative fea- tures predicted by theory; 1.9., a region of liquid-liquid phase separation in the dilute range, with the critical point occurring at a very low polymer concentration. Liquid-crystal- line phase separation occurred at higher concentrations; this region extended from a triple point, where three phases are present, to the melting point of pure polyethylene, and is characteristic of crystalline polymers. The size of the liq- uid-liquid region decreased in better solvents; only liquid- crystalline separation was observed in the best solvent used.. In the atactic polystyrene system, which was more thor- oughly investigated, only liquid-liquid phase separation was observed. The results agreed, qualitatively, with theory; binodials were broader and critical concentrations were high- er than predicted. The linear dependence ofaq, the critical thermodynamic interaction parameter, on a function of the ratio of molar volumes was confirmed. The critical miscibility 4 temperature, 9, determined by this method, agreed with the value obtained by osmOtic pressure measurements; 94 values, however, differed by a large factor. - The literature records little dilute solution data for atactic and isotactic forms of the same polymer. Light scat- tering measurements have bee made by Kinsinger on the poly- propylene system (7). He observed small differences in the second virial coefficients of the two forms. In osmotic pres- sure measurements on isotactic and atactic polystyrene, Dan- usso and Moraglio found definite differences in the second virial coefficients (8). They also noted that isotactic poly- styrene was less solublo than the atactic form. These results, for polypropylene and polystyrene, were the first to indicate that solution properties vary with configuration. Theory (9-11) For a closed binary system, regardless of molecular size, the condition for equilibrium is that the free energy be a minimum; i.e., dF 3 0. At constant temperature and pressure, dF aMldnl +792dn2 : O (Eqn.l) where/u is the chemical potential and n, the number of moles, of each component, and dF Z fiddn“ 1» la“ dno‘ +/u-Bdn£ +/u$dn8 : O (Sqn.2) I I a A I I a .1 if two phs-zses, (x and ,6 , are present. Since the total amount of each component is fixed in a closed system, a '5 . dnd + dné = O. (Eqn.3) To satisfy equations 2 and 3 for all possible compositions O( in eacn phase,/€2.-/4Q. 5 Since two compositions with the see me chemical potential occur in a two- -p}1a ase re5ion, a plot of/pfi versus composition must exhibit a minimum and maximum. Stated mathematically with x2 denoting the mole fraction of solute,(§£fl) : O at the mix- imw: or minimum. At the point of inflection, (5%) I 0. These 3 expressions are the criteria for the existence of two phases in a system at equilibrium. Using the total free energy of the system as a criterion, 3 two phases ca n exist whe (92;; <0 at the composition in question. In the case where tge free ener5y decreases with increasing; temeera ture a point is reached where (SW-:3: >0 over the entire com: osition range. When this occurs, the com- ponents are miscible in all proportions. Lowerin5 the exper- ature of the system induces phase separation which first oc- W1 9 F curs 1%(3:G\= O. A critical condition is found when, at a certain temper- ature and composition,(§fl) : 2.21) = O. The composition at 9’9: 2>X3 which it occurs is the critical composition and the tempera- ture is the upper consolute temperature. The critical points can be experimentally evaluated from the maximum in the phase dia5r9 Is. The above theory also applies to polymer solutions, but volume fractions are used as the composition variable instead of mole fractions. (Volume fraction is actually the correct term for describin5 mixtures of molecules of different sizes.) From the general theory for polymer solutions, the change in free ener5y on mixing is: AFm : kT( n1ln~1 + nalnlv; +‘X,n,nra), (Eqn.4) where n is the number of molecules and/aria the volume frac- tion of each component.7L,is the thermodynamic interaction~ parameter. By use of standard thermodynamic operations and changing from molecules to moles, ”1-7:? = RT [1n mg—Hl-l/xbua +"X./U53] (Eqn.5) The quantity, x, equals HhVSp/Vi. Eh is the number average molecular weight of the polymer (component 2); Vsp is its specific volume; and V1 is the molar volume of the solvent (component 1). The critical conditions are obtained by tak- ing the first and second derivatives of equation 5 with re- spect to N." ('35:) = RT -1/(1-~5) +(1-1/x) +27C/fiw (Eqn.6) (39:15:) .1 RTE-l/(l-AS); +2X,]=O (Eqn.7) Solving simultaneously for m;c and 1—,, in terms of x gives: ”,1: 1/(1+ xé) ‘ (Eqn.8) 7.1,: % +1/x%+1/2x (Eqn.9) Other values of m; and 7-, , near the critical point, cannot be obtained from the preceding equations. However, they can be approximated from a relationship, derived by Flory (12). From these, complete theoretical binodials can be calculated. In the dilute solution theory, which allows for the non- uniformity of the solution, two interaction parameters arise. They are {g, the heat of dilution parameter and Q(, the en- tropy of dilution parameter for a polymer of infinite mole- cular weight. By definition, A'fi' : any: and A's": Rim/v33 Zi§f is the difference between the actual partial molal en- tropy of mixing and the ideal value. 1anan be written in terms of these parameters; X’:(%-- W) 4' K, (Eqn.10) The critical miscibility temperature for a polymer of infin- ite molecular weight is defined by 9 : K.T/w‘ (Eqn.ll) Using 9, equation 10 can be rewritten x': (%- k”) + KEG/1" (Eqn.12) Equating equations 9 and 12, and rearranging terms, gives a linear relationship between the critical temperature and x, from which wgand 9 can be evaluated, experimentally. l/Tc = 1/9 + 1M9 (1/x? + l/2x) (Eqn.13) At the 9 temperature, K, : IV, and the net interaction be- tween segments is zero. The polymer molecules are able to penetrate each other freely, and the excluded volume of the molecules is also zero. At this temperature, the polymer molecules assume their unperturbed dimensions,.[§;. The ex- 0 pansion factor,ov, is defined by f?» :: o< F5, ' (30111.14) 4%; is the rms end to end distance of the chain in a perturb- ed state] and is a function of the thermodynamic inter- action parameters; 0<5- 0&3 :: Gm M(1;9/T)H% . (Eqn.15) ogapproaches one as T approaches 9. Cm is a constant for a given molecular weight and is calculated from the expression, 8 Cm : [flgW/FSF (Eqn.16) 1 . The interaction parameters are characteristic of the non-ideal behavior of polymer solutions, but their exact sig- nificance has not been determined, theoretically. Flory as- sumes that the same parameters describe both inter- and intra— molecular interactions (13), although this has been question- ed (14). The dilute solution theory gives different values for \y.. In dilute solution, the simple lattice theory predicts that (Q{-%)$v l/z for intermolecular interactions where 2, as defined, represents the lattice coordination number. Much more complicated expressions for the interaction parameters have been derived by other methods. These methods consider preferential interactions, molecular size and shape, and chain stiffness. The effect of stereoregularity has not been directly. considered. Therefore, changes in the interaction parameters. due to configuration, must be postulated by indirect methods. This can be done by making qualitative predictions about the effect of stereoregularity_on the quantities which determine the values of the interaction parameters. EXE‘ERIIViENTAL 10 II EKPERIHSNTAL W Mixtures of polymer and solvent were prepared as follows: The components were weighed into the sample tube on an analyt— ical balance. The tube was evacuated to remove trapped air from the polymer sample. The vacuum was held until bubbles ceased to form in the liquid. A stream of nitrogen was pass- ed into the tube for two minutes to flush out the remaining oxygen. The stirring rod and cork were fitted into the tube which was then placed in an oil bath, maintained above the precipitation temperature. The sample was stirred until all the polymer had dissolved. The precipitation temperature was determined, visually, by the following procedure: The bath temperature was lowered at the rate of, approximately, 1° per minute until cloudiness appeared in the solution; the temperature was then raised until the sample cleared. The cycle was repeated at a slower rate (about O.2°/min.) to obtain the endpoint. A single mea- surement, including preparation of the sample, required ap- proximately one hour. Equipment The bath was constructed from a three—liter, cylindri- cal "Pyrex" vessel. It was supported inside a ring, six inch- es in diameter, which was clamped to a ring stand. The bath was insulated with three layers of glass wool, separated by sheets of aluminum foil. The outer surface was wrapped with a double thickness of foil and secured with several strands ll of wire. Windows were cut through the insulation at the front and rear. A 150 watt light bulb was positioned at the rear window to illuminate the bath and samples. The bath was filled with Dow-Corning #550 silicone oil and was agitated by a propeller type, stainless steel stir- rer. Heat was supplied by a 250 watt Cenco immersion heater, controlled by a Powerstat. Operational range extended from 80°C to 300°C; below 80°, the oil became too viscous for adequate stirring. The tube rack was constructed from stainless steel gauze and was suspended in the bath behind the front window. Use of gauze held resistance to oil circulation to a minimum. Q Q C) Figure l.- Sample tube and stirrer. The sample tubes (figure 1) were made from 8 mm. "Pyrex" tubing. One end of a five inch piece was drawn out and seal- ed. The Open and was fitted with a cork through which was passed a stainless steel stirring rod. (Copper rods were used, initially, but were found to catalyze degradation of the polymer.) A Reagents CP grade diphenyl ether and dibenzyl ether were purchas- ed from Fischer Scientific Company and were used as received. Their melting points were 26.500 and 1.800, respectively. Other solvents were technical grade. 12 The polypropylene fractions, used in this research, were furnished by Doctor J. B. KinsingeritViscosity relationships and number average molecular weights are listed below. Table I.- Viscosity data for isotactic polypropylene (7). Fraction [*1] En C'2A 4:80 491,000 C'2B 2.35 198,000 03 1.34 97,800 04 0.63 37,200 05 0.21 11.000 [n]: 1.38 x 10.4fin'8 in decalin at 135°C Table II.- Viscosity data for atactic polypropylene (7). Fraction [71] En BIA 7:12 681,000 A3 4.08 324,000 B2 2.90 214,000 B4 1.42 87,200 0.65 52 a :00 .26 [—71]: 1.60 x 10-41-4'n'8 in cyclohexane at 25°C ¢ According to reference 7, the polypropylene fractions were prepared from samples furnished by Hercules Pow- der Company. The atactic sample was a transparent, rubbery, amorphous material at room temperature. The iso actic sample was characterized by melting point, 161 C,and by powder X-ray diffraction patterns. These results corresponded to those reported in the liter- ature for isotactic polypropylene. IR scans indicated little unsaturation in either sample and light scat- tering studies showed no significant chain branching. RE SULT S AND DI SCUSSION III RESULTS AND DISCUSSION SglyentJStudies Various solvents and solvent mixtures were investigated to determine their usefulness for phase equilibrium studies. The following properties were required: The solvent must be sufficiently poor to give liquid-liquid phase separation with isotactic polypropylene; it must be inert, high-boiling, and thermally stable. (Inspection of phase diagrams for other crystalline polymers (5, 15) indicated that the critical temperatures for isotactic polypropylene would occur near its melting point, 161°C. This was the reason for the temper- ature requirements.) Preliminary tests were made with 1% mixtures of unfrac- tionated polymer. The results are listed below. Table III.- Precipitation temperatures of 1% mixtures of polypropylene in different solvents. atactic isotactic Solvent . 8 I mpgoc) T8100) S** Tetralin 8.40 (2 - 0 L-C o—Dichlorobenzene 10.0 425 70-85 L-C Tetralin(.33)*-pheny1 ether(.67) -—- 96 100-125 L-C o-Dichlorobenzene(.33)- phenyl ether(.67) --- 98 100-120 L-c l-Octanol 10.3 - 124 122 L-L Phenyl ether --— 146 142 L—L o-Dichlorobenzene(.33)- phenyl cyanide(.67) --- 160 150 L-L Benzyl ether(.4)-pheny1 ether(.6) ~-- 163 156 L-L Benzyl ether --- 186 175 L-L Phenyl cyanide 11.17 ins 194 L-L Diethyl phthalate 9.99 ins ins Dimethyl phthalate 10.8 ins ins Ethyl cyanoacetate 11.0 ins ins Nitrobenzene 11.32 ins ins *-volume fractions; **-phase separation, L-C liquid- crystalline, L-L liquid-liquid; 5 -solubility parameter 14 15 Phenyl ether and benzyl ether were the only pure sol- vents tested which met the requirements outlined above. 88- cause of their excellent prOperties (both are extremely inert and high-boiling), they should be useful in other work with polypropylene, such as light scattering at the 9 temperature. Phenyl ether was chosen for phase studies because it had the lower precipitation temperature where degradation could be minimized. Several mixed solvents gave liquid-liquid separation, also. The additional complications of treating a three com- ponent system, however, reduced their value for this work. An exception was made for the phenyl ether-benzyl ether mix- ture. It was decided that the mixture could be treated as a single component because of the similarity of these solvents. They have the same functional group, and similar structures and densities. Where liquid-liquid phase separation occured in the solvent tests, the precipitation temperatures were reprodu- cible, but this was not the case for liquid-crystalline sep- aration.(The latter is not an equilibrium process, but is, essentially, a rate phenomenon.) The endpoint for this type of separation was not reproducible because of the dependence of the precipitation temperature on the rate of cooling and stirring. A precipitate formed by cooling would not dissolve until the temperature was raised from 5° to 20° above the endpoint; cooling slightly below the temperature at which the precipitate dissolved would not induce another phase 16 separation. It was also characteristic of liquid-crystalline phase separation that the viscosity of the solution increased near the endpoint. The results of the solvent studies have led to some rather interesting conclusions. In solvents giving liquid- crystalline separation, isotactic polypropylene was the less soluble form. The difference in solubility was dependent on the solvent and/or the temperature. It was greatest in the best solvents tested and decreased as the precipitation tem- peratures increased. ‘ Danusso and Moraglio found that the isotactic form of polystyrene was less soluble . (8) In addition, they dis- covered that 2-butanone was a selective solvent for atactic polystyrene. For polypropylene, both tetralin and o-dichloro- benzene are selective solvents for the atactic configuration at low temperatures. These results are restricted to solvents which give liquid-crystalline separation. However, Danusso and Moraglio generalized that isotactic polystyrene was al- ways the less soluble form. This conclusion does not hold for polypropylene in solvents which give liquid-liquid sep- aration. In solvents which gave liquid-liquid separation, atac- tic polypropylene was less soluble. Again, the difference in precipitation temperatures was dependent on solvent and/or temperature. However, the trend was reversed from that noted in the better solvents. The difference increased as the tem- perature of precipitation increased. A qualitative conclusion 17 can be drawn from this: The dependence of solubility on the solvent and temperature is different for isotactic and atac- tic polypropylene. The functions representing this depend— ence must intersect at some point to account for the rever- sal in trends. This concept is illustrated by a graphical representation of the trend in figure 2. isotactic—+ to poorer solvents Figure 2.- Illustration of the trend in precipitation temperatures. 18 Phase Diagrams As mentioned in the discussion of solvent studies, phenyl ether and a mixed solvent (40% benzyl ether-60% phenyl ether by volume) were chosen for phase studies. Phase diagrams in the dilute range were obtained for the phenyl ether system; only critical points were measured in the mixed solvent. The data for isotactic polypropylene in phenyl ether are compiled in Table IV. Weight fractions, obtained exper- Table IV.- Data for isotactic polypropylene in phenyl ether.’ Fraction‘ Ayg, Tp(°C) C'2A 0.0061 137.8 C'2A 0.0119 139.2 C'ZA 0.0176 140.2 C92A 0.0181 140.3 C'2A 0.0335 139.8 G'2A 0.1073 135.6 C'2B 0.0115 136.2 C’2B 0.0256 137.0 C92B 0.0294 137.2 C'2B 0.0516 136.0 CF28 0.1320 131.8 C’2B 0.187 128.5 0'23 0.249 123.0 03 0.0092 134.0 03 0.0321 134.6 03 0.0405 134.8 03 0.0480 134.5 03 0.1071 132.8 04 0.0046 121.5 04 0.0344 127.2 C4 0.0514 127.4 C4 0.1059 125.2 04 0.1379 124.8 04 0.1970 124.0 05 0.0752 116-119* _ 05 011372 118-121:_ * liquid-crystalline separation. imentally, were converted to volume fractions by assuming negligible volume change on mixing; therefore, N3 = (Wngp)/(Wl/dl + wevsp) where the values of the specific volume and density were taken at the precipitation temperature. (See Appendix.) The precipitation temperatures were plotted against volume frac- tions to obtain the phase diagrams (Figure 3). Critical points were determined from rectilinear diameters. Precipitation temperatures were reproducible to 1.3.0.200 except for fraction C'2A and in concentrated solutions for all fractions. C'2A had an initial cloudpoint at about 10°C above the endpoint, indicating either an atactic content or a wide molecular weight range. The viscosity of the more concentrated solutions made mixing difficult. The resulting non-homogeniety plus apparently more degradation and larger relative solvent loss contributed to greater uncertainty in hese endpoints. In no case, however, was the uncertainty greater than igt°c. Precipitation temperatures were indepen- dent of cooling rate, but heating of turbid samples was de- pendent on both rate of heating and stirring. The endpoint also varied with the viewing angle; all of the endpoints in this work were obtained by sighting through the sample dir— ectly into the light. Near the triple point, the observations made above did not hold. It was difficult to judge whether the separation was liquid-liquid or liquid—crystalline. The supercooling noted in the solvent studies was not as pronounced, and in the already viscous solutions, no large change in viscosity 20 was apparent. these observations apply only to the more con- centrated samples of fractions 0'28 and C4, and do not affect the critical points. The binodials in Figure 3 exhibit the usual character- istics for liquid-liquid phase separation. Compared to theory, the critical volume fractions, except for 04, were too high. The binodials were much broader than predicted by theory for all fractions. The triple point for C'2B is estimated to oc- cur between 11800 and 123°C, and at a volume fraction of ap- proximately 0.25'0.30. Although a study of liquid-crystalline phase separation is not within the scope of this work, it should be noted that some interesting phenomena were observed in and near this region. The phase diagram for 04 has a different shape at higher concentrations than other fractions, indicating that separation here may have been liquid-crystalline. The flat- ness of the curve is evidence that the triple point temper- ature is only a few degrees below the critical temperature. Fraction 05 gave only liquid-crystalline separation which means that the limit of the liquid-liquid region lies be- tween C4 and C5. Therefore, the size of this region depends on molecular weight as well as solvent. Unfortunately, the experimental methods employed limit this to a qualitative conclusion. Specific volume measurements have to be made to achieve quantitative results. As an addition to the preceding discussion, it is of interest to present some studies of the system, isotactic 140 135 1013(06) 130 125 120 _ J J i l I 0.0 0.05 0.10 nr' 0.15 0.20 0.25 ac Figure 3.- Phase diagrams for isotactic polypropylene- diphenyl ether. - - -, rectilinear diameters and | , calculated N36 polypropylene - l-octanol. In the solvent tests, l-octanol appeared to give liquid-liquid separation and a phase dia- gram was determined for fraction C'2A. The data are plotted in Figure 4. The steady increase in Tp with concentration of polymer (represented by the solid curve) was interpreted to mean liquid-crystalline separation. The scatter in the points was attributed to experimental error. Therefore, no further work was done on this system at the time. Reinterpretation of the data indicates that the broken line is the true re- presentation. C'2A in l-octanol, then, is comparable to C4 in phenyl ether; i.e., it is near the lower limit for liq— uid-liquid separation. Because higher molecular weight frac— tions were not available, this conclusion could not be check- ed further. Since atactic polypropylene is an amorphous polymer, only liquid-liquid separation was observed for this form in phenyl ether. The data are collected in Table V. The exper- imental uncertainty was similar to that for isotactic poly- propylene in the liquid-liquid region.(*o,2 in dilute samp- les totojo in more concentrated samples). In addition, the atactic form was more difficult to dissolve and, when swol— len, adhered strongly to the walls of the tube. The phase diagrams are plotted in Figure 5. The binod- ials were again broader than predicted by theory, and the critical volume fractions, higher except for B6. The atac- tic polypropylene curV38,however, are not as broad as those for both isotactic polypropylene and atactic polystyrene. 125 04 TP c 120 #- FRACTION C'2A 115 l fig- 491,000 0.0 2.0 4.0 weight % C'2A Figure 4.- Phase diagram of isotactic polypropylene in l-octanol. - - -, curve for probable liquid- liquid phase separation. 24 Table V.- Data for atactic polypropylene in phenyl ether. Fraction mg; Tp(OC) BIA 0.0064 143.9 "“ BIA 0.0118 146.4 BIA 0.0160 147.2 BIA 0.0219 147.5 BIA 0.0495 146.0 B2 0.0055 138.3 B2 0.0178 141.1 B2 0.0303 142.2 B2 0.0357 141.9 B2 0.0507 141.8 B2 0.0998 138.3 B4 0.0203 134.4 B 0.0271 134.8 B4 0.0383 135.0 B4 0.0521 134.8 34 0.0994 131.9 B6 0.0069 121.4 B6 0.0376 126.0 B6 0.0591 126.1 B6 0.0746 125.4 B6 0.1088 124.3 The experimental binodials for two fractions of polypropyl- ene are compared to theoretical binodials for the respective molecular weights in Figure 6. Deviation near the critical point is similar for atactic and isotactic polypropylene. . At higher concentrations, the isotactic form deviates more; this may be caused by stiffening of the chain near the tri- ple point. The greater deviation of atactic polystyrene can also be attributed to its greater chain stiffness (caused by the bulky phenyl groups). 8 An odd feature ofathe phase diagrams for polypropylene was the deviation of the critical volume fractions from ideality. The experimental values are plotted against the theoretical values in Figure 7. The trend shifts from pos— 145 140 135 00 \ 130 _ 125 4 120 4 l I 070 0.05 m; 0.10 0.15 Figure 5.- Phase diagrams of atactic polypropylene in diphenyl ether. - - - rectilinear diameters. | _ theoretical N50 FRACTION B2 gTACTIC Mn-2l4,000 142 _- 140 - 138 '* \ o L l I TD 0 FRACTION 0'23 ISOTACTIC Mn-198,000 136 '- 134- ' 132 — 0.00 0.05 r03 0.10 0.15 Figure 6.- Comparison of experimental binodials to theory for the system, poly- propylene - diphenyl ether. ~ - —, theoretical binodials 26 .075 “5c .050 .025 l l ‘ L 0.0 0.025 0.050 0.075 1/(1+x%) Figure 7’.- The deviation of Ar“ from theory for polypropylene in diphenyl ether. 0- isotactic Cl- atactic 27 28 itive deviation at high molecular weights to negative de- viation at lower molecular weights. Since the critical vol- ume fraction is dependent on the solvent (16), this may be due to some specific effect of phenyl ether. F3 he Thermodynamicfgnteraction Parameters The critical data for isotactic and atactic polypropyl- ene are summarized in Tables VI and VII. As predicted by Table VI.- Critical data for isotactic polypropylene in phenyl ether. Fraction Tc 3: {Vsc (calc.) (exp.) C'2A A 140.2 55-85 0.0167 0.019 C’2B 137.2 1400 0.0260 0.029 03 134.8 692 0.0366 0.038 04 127.4 263 0.0581 0.050 Table VII.- Critical data for atactic polypropylene in phenyl ether. Fraction To x IU' (calc.) . (exp.) 15—— BIA 147.5 4796 0.0142 0.021 B2 142.2 1511 0.0251 0.031 B4 135.0 617 0.0387 0.039 B6 126,1 230 0.0619 0.057 theory, a plot of l/Tc versus [l/xé + 1/2xJ was linear with- in experimental error for both polymers. Straight lines, drawn through the points in Figure 8, were obtained by a least squares fit. Both the slopes and intercepts are dif- ferent for the two forms; this results in a sizable differ- ence in the thermodynamic interaction parameters. For iso— tactic polypropylene, 0 is 418.4°K and uh: 1.414; for atac- o tic polypropylene, they were 426.5 K and 0.986, respectively. 2055 4- UV 2.50+—~—w- 6‘ e 1 x10: Tc isotactic - atactic 1. 2.45 - I ‘0 0' II 2.40 F 2.35 L 1 1 1 0.0 2.5 5.0 7.5 10.0 (1/x44-1/2x) x102 Figure 8.- The dependence of Tc tem, polypropylene - diphenyl ether. on x for the sys- 29 The curves have a point of intersection at which the crit- ical temperature is the sate for both polymers. This unique point will be discussed further after the results of the mixed solvent system have been presented. According to theory (Eqn.12), the temperature depend- ence of ygcan be obtained from the interaction parameters. For atactic polypropylene,Z§= v0.486 + 420.5/T and for isotactic, ,= -0.914 + 591.6/T. These functions are plotted in Figure 9. Although these values ofjxware for a polymer of infinite molecular weight, they can be used to show the trend for any molecular weight. )L,is a measure of solubil- ity; therefore, some statements can be made about the rel- ative solubilities of the two forms. At temperatures above the point of intersection, isotactic polypropylene is more soluble and below it, less soluble. At the point of inter- section, the solubilities are the same. This is a hypothet- ical situation, of course, because only low molecular weights are soluble below the point of intersection and the solubil- ity at any temperature is governed by the molecular weight. The critical temperatures in the mixed solvent were measured by finding the precipitation temperatures of sev- eral mixtures in the neighborhood of the critical point, and taking the maximum temperature found as the critical temperature, This causes little error ini‘and 0. Critical points were obtained for four fractions of each polymer. The pertinent data are listed in Table vIII. The results were plotted in the same manner as those from the nhenyl 1.00 0.75 _. isotactic - 0025 In- 15) F\ 7 J‘. (\ 0.0 , I l 1 1.0 2.0 3 3.0 (l/T)x10 Figure 9.- The dependence of )L,on temperature for the system, polypropylene-diphenyl ether. 31 Table VIII.- Critical data for polypropylene in the mixed solvent. Fraction . Tc x Wt.% Isotactic polypropylene C'2A 428.2 3199 1.86 0328 425.0 1289 2.68 C3 423.0 636 3.06 04 415.4 241 3.40 Atactic polypropylene A3 434.0 2120 1.81 B2 432.1 1400 2.25 B4 426.3 568 2.84 B6 - 41§.4 208 3.811 ether system. The curves for this system also intersect, but at a higher temperature and lower molecular weight. The values of the interaction parameters, though less precise, show the same trend as in the phenyl ether system. Isotac- tic polypropylene has a lower 9 and a higherlfi. For compar- ison, the parameters are collected in Table IX. The values Table IX.- The thermodynamic interaction parameters. phenyl ether mixture 9 (0;) VL 9L0_K_)__ WL_ isotactic - 41814 1.414 433:1 ‘ 1.57 atactic 426.5 0.986 442.7 1.13 difference 8.1 0.426 9.6 0.44 of the parameters are larger in the poorer solvent which is to be expected. However, not only the absolute values in— creased, but also the magnitude of the differences. This confirms the conclusions from the solvent studies. (See Figure 2.) 33 4:. 2.45 I . Hr 2.44 c 2.40L 1 x154 Tc isotactic - atactic 2.35L ‘ II 0 II 2.30.. I 2.25 . 34 1 0.0 2.5 .0 7.5 10.0 (l/X +1/2x) x102 Figure 10.- The dependence of To on x for the system, polypropylene- 60% diphenyl ether- 40% dibenzyl ether. 54 The experimental results point to the fact that liquid- liquid phase separation is limited by several factors. The solvent studies show that it is dependent on the solvent and occurs only in relatively poorer solvents. In the phenyl ether system, it was limited by molecular weight and the limit occurs between fractions 04 and 05. Since a decrease in molecular weight corresponds to an increase in solubility, these results are comparable. In both cases, however, tem- perature may be a factor. The point of intersection of the 9 curves for the phenyl ether system (Figure 8) has values of TC and molecular weight which lie betwee C4 and 05. This unique point may represent the limit of liquid-liquid phase separation for the system. If this is correct, isotactic k polypropylene of the critical molecular weight has a triple point and critical point which coincide. Also at this point, isotactic and atactic polypropylene have the same X,values and, therefore, the same solubilities. Comparison of the Table X.- Critical values at the point of inter- section of the 9 curves. phenyl ether mixed solvent Tc(°K) 399.8 409.0 Eh 34,200 20,000 X“ 0.566 0.593 intersection points in phenyl ether and the mixed solvent shows that they also depend on solvent and molecular weight. 35 Therefore, both the mclocular weight and the thermodynamic interaction parameters control the size of the liquid-liq- uid region. The only theoretical limitation for this type of separation is that 1“}. g. The values of X“ at the point of intersection are plotted in Figure 11. A relatively snall change in G (or a change to a slightly better solvent) pro— duced a large change in the molecular weight at the point of intersection. This indicates that the molecular weight may approach infinity in better solvents. The preceding discussion suggests the existence of a critical size since molecular size depends on the limiting factors. Because the interaction parameters depend on con- figuration, it follows that the critical size will have a similar dependence. In view of their dependence on configuration, the in- teraction p° ame ers may be a measure of stereoregularity in the chain. 9 can be used as an example. From the data, it is logical to assume that G is a minimum for a perfect iso- tactic chain and is a maximum for a perfect syndiotactic (alternating d and 1) chain. 9 for the atactic chain would lie somewhere between them, depending on the amount and dis- tribution of each configuration. Syndiotactic polypropylene is not available so this conclusion cannot be checked exner- imentally. It is interesting to Speculate, however, that 9 would be a direct empirical measure of stereoregularity if the 9 curves for all configurations had the same point of intersection. 0.600 0.575 7L": 0.550 0.525 00500 41 L l 0.0 2.5 5.0 7.5 10.0 (1/x%+1/2x) x102 Figure ll.- The comparison of“xw and x at the point of intersection in two solvents: o - diphenyl ether 0 - mixed solvent 36 37 General Discussion Since the interaction parameters were different for isotactic and atactic polypropylene, it is safe to assume .that solution properties vary for all degrees of stereoreg- ularity. Therefore, the differences in the parameters can be interpreted in terms of configuration and should indicate the effect of configurational differences on such properties as molecular size and flexibility. The rms*end to end distance cannot oe calculated from these data, out must be determined from light scattering measurements. However, phase equilibrium studies do give an indication of what differences might exist. At the 9 temper- ature, the polymer assumes its unperturbed dimensions. Since G is higher for atactic polypropylene, its expansion factor, 0‘, is smaller at a given temperature; e.g., at 9 for atactic, itS¢x is unity but the o