LQWER CRETICAL SOLUNCN TEMPERATURES FOR POLY- 3.0LEHNS The“: {or flu: Dogma o§ M. S. MICHEGAI‘! STATE UNEVERSETY Roland Joseph Tetreauit 1963 .r _M_-. - . “BR/“213: Michigan State ‘ University «J THESIS c 72/ MICHIGAN STATE UNIVERSITY ZHIGAN ABSTRACT LOWER CRITICAL SOLUTION TED'E’ERATURES FOR POLY-a —OLEFINS by Roland J. Tetreault Lower critical solution temperatures (LCST) were determined for several fractions of five polymers: isotactic polypropylene, atactic polypropylene, isotactic polybutene-l, atactie polybutene—l, and polyoctene-l. The LCST of the first four polymers were determined in n~pentane and that for the last polymer in n-butane. One polyoctene-l fraction was studied in four hydrocarbon solvents. Finally phase sep- aration temperatures were determined for a polyoctene-l fraction dis- solved in varying mixtures of n~butane and n~pentane. It was shown that a linear relationship exists between 13%? vs #, where H is the molecular weight of the fraction. This relation- ship was anticipated from Flory‘s upper critical solution temperature theory although it does not specifically predict a LCST. A linear relationship also exists between the critical temperature of the sol- vent and the LOST of the polymer solution. Phase separation teameratwres were determined for a three comaonent system. A positive deviation from ideal behavior was observed. No theory as yet exists for such a system; in fact this is the first three component system (2 solvents, l polymer) ever studied for s LCST. LOWER CRITICAL SOLUTION TBMPWTWES FOR POLY-a—C-LEFINS By Roland Joseph Tetreault A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1963 ACKNOWLEDGMENTS The author is indebted to Dr. J. B. Kinsinger for the guidance and helpful ruggestions offered during this investigation. He also expresses his appreciation to the Dow Chemical Company who nude possible the completion of this study through their financial support. ii DJTRODUCTION . . . . . . History Theory . O O C O C O EXPERII‘ENTAL . . . . . . EQUipmnt o o o o 0 Reagents . . . . Preparation of Tubes Procedure RESULTS AND DISCUSSION . TABLE OF CONTENTS .006 PhaseDiagrams....... Molecular Weight Dependence The Dependence of LCST on Solvent 0 Q C O Three—Component System . . . . . . Analysis Based on the Cell Mbdel . REFERENCES APPENDICES iii O O O O O O O O O D O .0. O O O O O O I I O O O O O C 0 O C O . I O O C C O O O O Q 0.00. Table I. II. III. IV. V. VII. VIII. LIST OF TABLES Page Gas chromatograph analysis of solvents . . . . . . . . . . 8 Viscosity data for polymer fractions . . . . . . . . . . . 13 LCST data for polyad-olefins . . . . . . . . . . . . . . . 21 LCST for polymer of infinite molecular weight . . . . . . 27 LCST data for one polyoctene-l fraction (F-éA) in several solvents . . . . . . . . . . . . . . . . . . . 31 Delmas, Patterson, and Somqynsky constants calculated by least squares fit of experimental data . . . . . . 36 Comparison of calculated and experimentally determined LCST for the polyoctene-l (F-6A) in different solvents ho Comparison of calculated and experimentally determined SI for all the polymers studied . . . . . . . . . . . ho iv Figure 1. Chemical potential as a function of x; . . . . . . 2. Gas chromatogram of research grade napentane . . . 3. Gas chromatogram of’pure grade n-butane . . . . . . 1;. Gas chromatogram of pure grade neOpentane . . . . . 5. Phase diagram for atactic polypropylene fractions . 6. Phase diagram for atactic polybutene-l fractions . 7. Phase diagram for isotactic polypropylene fractions 8. Phase diagram for isotactic polybutene-l fractions 9. LCST for polymer of infinite molecular weight . . . lo. LCST for infinite molecular weight atactic polypropylene. ll. LCST for infinite molecular weight polyoctene-l in n-pentaneandinn~butane ............ 12. LCST for isotactic polypropylene of infinite molecular weight . . . . . . . . . . . . . . . . . . . . . . 13. LCST for infinite molecular weight polyethyleneoxide 1h. Phase diagram for a polyoctene—l fraction in different solvents . . . . . . . . . . . . . . . . . . . . . 15. LCST for two polymer fractions vs solvent critical temperature .................... 16. Phase separation temperatures for solutions of polyoctene-l (Ii-711) in mixtures of pentane and butane . . . . . 1?. Plot of Delmas, Patterson, and Somcynsky equation for LIST’OF FIGURES LCST using constants calculated by least squares fit of experimental data . . . . . . . . . . . . . Page 5 9 10 ll 16 17 18 19 22 23 2h 25 29 3O 33 35 38 LIST OF APPENDICES Page I. Phase separation termeratures for polymer fractions in n“p€ntaneeeeeeeeeeeeeeeeeeeseeuh II. Phase separation temperatures for polyoctene-l fmCtionSInn-bummeeeeeeeeeeeeeeo 1L8 III. Phase separation temperatures for the polyoctene-l fraction, F-GA, in different solvents . . . . . . . 50 IV} Phase separation temperatures for the polyoctene-l _ fraction, F—7A, in mixtures of n~pentane and n'bummeeeeeeeseeeeeeeeeeeeee 5]. vi INTRODUCTION Histgx It has been known for some time tint certain nixed systems: exist where the mutual solubility of a pair of liquids decreases with in- creasing temperature. The minim temperature at which imiscibility occurs is called the lower critical solution temperature (LCST or TCL). All the early data refer to systems where both components were highly polar and the LCST was related to the increase in entropy associated with the rupture of hydrogen bonds. Only a few years ago Freeman and Rowlinson1 observed this same behavior for hydrocarbon polymers in hydrocarbon solvents, a system which is notoriously nonpolar in character. This observation which was not predicted by the Hildebrand-Scatchard solubility theory aroused e great deal of interest. The authors of this initial report associated the decreasing solubility of the polymer with increasing temperature with the expansion of the solvent as it approached its critical temper- ature and a rapid decrease of its solubility parameter relative to that of the polymer. Rowlinson and Freeman2 published simultaneously with the above work their results with ethane solutions of pure liquid hydrocarbons with between 21; and 37 carbon atoms. Their results establish beyond - a doubt that MST are found in mixtures of nonpolar molecules of the same chemical type if the molecular sizes and energies of interaction of the two components are different. They showed that the LCST de- creased with increasing molecular weight of the solute and that solute 2 molecules'with saturated rings or with.unsaturation gave lower LCST than their corresponding saturated hydrocarbons. Baker and.his coworkers3 studied the phase equilibria for unfrac- tionated.polyisobutene of mean molecular weight from 250 to 2g500,000 in nepentane. The high.molecu1ar weight polymers were precipitated at temperatures slightly above the normal boiling point of the solvent. They showed that the thermodynamic properties change profoundly with molecular weight. In agreement with Freeman andRowlinson1 they point out that negative excess heats and excess entropies of mixing are thermodynamic necessities in a binary solution that is close to a LCST, and that these properties are incompatible with the Flory-Huggins equation. . Delmas, Patterson, and.Somcynsky4 used the solubility parameter theory and.molecular theory of polymer solutions developed by Prigogine and collaborators to treat quantitatively the negative (exothermic) heats of'mixing occurring in some nonpolar polymer—solvent systems and the LCST. Heats of mixing were obtained calorimetrically for unfrac- tionated polyisobutylene (FIB) with solvents in the n-alkane series. Their experimental data are in good agreement with their cell model theory. The development of their theory leads to the equation: T . r CL R1 Mfg.) + 305;") (1) where: r1 - (n + l)/§3 n - no. of carbon atoms of the solvent. .A and B - constants. Subscripts l and 2 refer to solvent and.polymer respectively. T ' LCST in WK; R.= gas constant in ca1./deg.-mole. CL X.= Flory interaction parameter. 3 A was evaluated from calorimetric heats of mixing of FIB in the n-paraf- fins and B was chosen to give the best fit with the critical solution temperatures. They used equation (1) to predict a LCST and their calcu- lated values were in fair agreement with the experimental data of Freeman and Roviinson’. However, they assured a value for 7» corres- ponding to infinite molecular weight polymer, thus eliminating the important molecular weight dependence from their eqmtion. &llard5 determined the LCST for four fractions of polyoctene-l in n—pentane. Solving equation (1) for TCL/r 1, he obtained 3% _ Rx} [annular/z . (2) 1‘; fi Then he kept the molecular weight dependence in the equation by using the following value for X, . .1 1 1 X-‘§+;;7z'+-2-; (3) iv whe re .973 X ' 3 En - number average molecular weight vsp - Specific volume of polymer V3 - molar volume of solvent. The parameter X follows from the Flory—Huggins theory which is useful for upper critical solution temperatures. Using the A and B parameters determined by Delms, Patterson, and Somcynslqy“, he calculated the LCST for P13 in n-pentane and tr: calculated values agreed well with Baker's3 observed data; Also a plot of '1‘ vs l/Scl/Z gave a straight line CL which when extrapolated to infinite molecular weight yielded a temper- ature whichuas called 91. analogous to the familiar Flory theta temperature for upper critical solution h temperatures (UCST). Also, 6L correSponds to the maximum temperature at which solvent and polymer of infinite molecular weight can coexist in a single phase. Equation (2) was found useful for calculating 6L. In the present work an attempt was made to further elucidate the dependence of LOST on molecular weight by determining the LCST for a number of fractionated poly-a~olel‘ins. £29.11 The stability of a binary phase can be characterized in terms of the chemical potential, 1.1!, of the components. If we consider a binary system whose two components are in equilibrium, thermocb'namic arguments6 show that for equilibrium with respect to diffusion for a two component system . < O a n2 Tn; (u) where n - number of moles. Making use of the Gibbs-Dulles relation, it can be shown that An . x 9;; at . x Du and equation (1;) is equivalent to .31.: <0 and Ti: <0 (6) These conditions are illustrated in Figure 1, which shows the dependence of the mole fraction on the chemical potential. Below TCL (curve 1) a single phase exists and the conditions of equation (6) are always satis- fied. However, at a temperature T2, the system consists of three parts (curve 3): one rich in solvent, a second rich in solute, and the third portion, the simultaneous presence of two phases. At TCL (curve 2) there S is a transition between the two states. The horizontal portion of ' curve 3 is reduced to a single point of inflection at C which mathemati- cally satisfies the restrictions 591 I. 52“]. I O (7) Ti? 3 x22 and . ‘55“1 <0 ' (8) x23 Pu X2 —> ' Fig. 1. 'Chemical potential as a function of x2 The conditions for stability (1;) can be written also in terms of the free energy of mixing since we can write (Jul) ‘ Fm <9) 3:27,? 3x22 T,P Therefore, for stability it follows from (h) and (9) that for a stable phase in equilibrium .52? . ( 3 x122) > O (10) and at the critical point (323M - o (11) (5 X} c It can be shown 63‘ that ()xz/ ()T depends essentially on the change in partial molar enthalpy for both components. At a LCST sz/JT is negative then A 2:: (fizz c > O (12) and at an UCST dx'Z/J T is positive then a 213 75-33296 < 0 (13) If we consider the relation PM - HM - ran (in) and (11), we can write 323M :- T 35M (15) 23‘522 3X22 c c Fran (11), (12), (13), (1i), and (15) the curvature of the partial solar heat content and the partial molar entropy must have the same sign at the critical point. This defines these added conditions for a critical point: (gait-fl) > O for a LOST (l6) and c) E . (31-?) < O for an UCST (17) As a first approximation, assume that the LOST has some similarity to the well known UCST. Therefore it seems reasonable that the Flory 7 equation7 for dilute polymer solutions 1 1 1 1 l -—- - l --— -—- 18 where: TCU I UCST in 0K 9 U - Flory theta temperature 1y, - entropy parameter which predicts an.UCST, may apply also for LCST at least with respect to the molecular weight dependence. This equation (18) is the basis for the plots of l/TCL.V‘ l/Hi/E, where H.is the molecular weight. leas used to plot the data rather than x since the molar volume, V1, of the solvent was not available above its boiling point. Also, the term l/fiM was dropped because this factor was negligible compared to l/iiiil/2 for the molecular weight species used in this work. EXPERIMENTAL i ment .A variable temperature bath with Dow-Corning #550 silicone oil was used for all phase separation determinations. The polymer solutions were sealed in "Pyrex" capillary tubes (3 mm. i.d.3 11 mm. o.d.). (It was found that smaller capillaries prevented good mixing which was critical for observing uniform end.points.) Four sample tubes were suspended in the bath at one time by means of a wire screen support. Reagents Research grade normal pentane, propane and.pure grade butane, neo- pentane, and isobutane were purchased from Phillips Petroleum Co. The supplier claims e.purity of 99.8h mol per cent for its research grade and 99 mol per cent for its pure grade. Gas chromatograms were obtained for these reagents to verity their purity. ‘An F. and.N; Scient. Corp. model 609 Flame Ionization Chrom. was used isothermally at 200°C. Table I summarizes the results and Figures 2 through h are representations of the chromatograms. Hence it appears according to the chromatographic analysis that the reagents have a purity better than that claimed by the supplier, Table 1. Gas chromatograph analysis of solvents. (Alumina column) Reagent Percent impurity based on relative peak heights Probable impurity propane 0.097 butane butane 0.037 propane isobutane 0.1h some isomer pentane 0.12 some isomer neopentane 0.h8 prepane, butane and others Figure 2. Gas chromatogram of research grade n-pentane. 'T I I I I I I I I 5'67 8910111213 Time (min.) 10 Figure 3. Gas chromatogram of pure grade n-butane. IIIIII W I I I I' . 0123145-678910111213 Time (min. ) 11 Figure h. Gas chromatogram of pure grade neOpentane. 1 KM _ I I I' I I I I I '6 7 8 9 10 ll 12 13 WC Time (min.) 12 The polymer fractions used in this research were furnished by Dr. J. B. Kinsinger. ‘Viscosity relationships and number average molecular weights are listed in.Table II. The isotactic polypropylene and the polyoctene-l fractions were used as received. The atactic polypropylene was dissolved in qyclohexane then filtered through a coarse sintered glass funnel. The isotactic polybutene-l fractions were dissolved in hot tetralin then filtered through a heated coarse sintered glass funnel into methanol. The precipitated polymer was washed three times in.methanol and finally dried in a vacuum oven to constant weight. The atactic polybutene-l fractions were dissolved in cyclohexane then fil- tered through 3 coarse sintered glass funnel into methanol. The pre- cipitated.polymer was treated similarly to the isotactic fractions. Preparation of Tubes The polymer and solvent were added to the capillary tubes in one of four ways: 1. The crystalline fractions were weighed directly into the tubes on a micro balance with a precision of i 0.03 mg. 2. Since atactic polybutene-l is soluble in nrpentane at room temperature, solutions of known concentrations were prepared by succes- sive dilutions and a sample or each concentration was added to a tube ‘with a hypodermic syringe. 3. The atactic polypropylene (insoluble in nepentane at room temperature) and the polyoctene-l (soluble in n—pentane but used only with gaseous solvents) fractions were dissolved in qyclohexane and this solution was added to the tubes. This latter solvent was removed from the polymer under vacuum and the tubes were brought to constant weight by heating at 50°C in a vacuum oven. 13 Table II. Viscosity data for polymer fractions . Sample No. Polymer deciillijte r f _ (T) (X 10 5) C-S Isotactic Polypropylene° 0.27 0.11 0.14 I ' 0.63 0.37 6-3 " " 1.314 0.97 E-ZA " " I..80 h.91 JK-h Atactic PolyprOpylene’ 0.099 0.031 JK-6 ' " 0.133 0.01416 B-Z " " 0.300 0.123 JK-S " " 0.9h8 0.520 MI-6 Isotactic Polybutene-l’ 0.550 6 0.926 PEI-h - a 0.710 1.3h 141-7 " " 1.200 2.h6 MI-S I I 1.690 3.77 A-3 Atactic Polybutene-lm 0.357 1.22 A~2 " ' 0.5M; 3.06 A—121 " " 1.55 23.0 F-llA Polyoctene-l5 0.fi 0.60 F-IOA ' 1.00 2.50 F- 9A " 1.75 6.07 F- 711 " 11.19 16.8 1?- 6A .. 5.71 25.0 F- 511 ' 8.60 140.0 fl.Isot.'=u:tic PP‘ [I1] I 1.38 x 10"4 En... in decalin at 135°C. Atactic PPa [Y1] - 1.60 x 10'4 Eff" in cyclohemne at 25°C. Isotactic PB“[T\] . 5.85 x 10" FL“. in n-nonane at 80°C. Atactic P311 In] - 5.85 x 10"” Rum“ in n—nonane at 80°C. Polyoctene-l5 [N] v 5.75 x 10-5 Ewe." in cyclohexane at 30°C. 1h In all of the above three methods, after the solvent or solution was added, the tube Opening was covered with a rubber cap and the tube was frozen in liquid.Bz until ready for sealing. The tubes were removed from the liquid B; one at a time, attached to a vacuum line and sealed. There was no measurable loss of solvent in the sealing process. h. When the solvent was gaseous at room temperature, the polymer or solution was added to the tube by one of the above methods then the tube was cooled, evacuated, and the gas was condensed into it. The tube was then sealed and weighed to obtain the amount of gas added. Procedure After the tubes were immersed in the thermostat and allowed to heat sufficiently to dissolve the polymer, they were manually agitated to insure homogeneity. The temperature of the bath was raised about one degree per minute to determine the approximate temperature range for phase separation. The thermostat was then cooled and the tubes agitated again. This time the bath temperature was raised at a slower rate (about 0.2 degrees per minute) to obtain the endpoint, T9, which is defined for this work as that temperature where a sharp increase in the solution cloudiness was observed. This endpoint must not be mis- taken with that temperature at which the heavier phase starts to settle. At the lower temperatures the solution is clear. As the temperature increases an apalescence gradually appears, then the solution cloudi- ness increases rapidly, and finally the heavier phase settles. There can.be as little as 0.2 degrees or as much as 5 degrees between the sudden increase in cloudiness and the settling of the heavier phase. RESULTS AED DISCUSSION Phase Diagrams Figures 5 through 8 summarize the phase separation data for the polypropylene and polybutene-l fractions. It was impractical to show the data for the polyoctene-l fractions in one figure because the curves lie too close to each other and some points overlap. There are three striking differences between these phase dia- grams for the LCST and those corresponding to an UCST7a for polymer- solvent binary systems. First, as observed for an UCST, the drift of the critical temperature toward lower weight fraction of polymer as the molecular weight is increased does not appear. In fact the critical weight fraction appears invariant with molecular weight within experi— mental error. Second, for a LCST the polymer molecular weight dependence of the critical temperature is inverted from that found for UCST, that is, the latter rises with molecular weight whereas the former decreases. Third, the shape of the phase separation curves for the LCST are much more uniform with each.molecular weight than is generally found for UCST. Finally it is noted that the temperature range over which the opales- cence occurs is much narrower for a LCST than for an UCST and hence the precipitation temperatures are more precise and reproducible. Some scatter of the data will be noticed in these plots. Some of this results from the dependence of the phase separation temperature on the rate of heating. Since the phase separation temperature is sharp, slight changes in the heating rate cause TD to change slightly. This change can.be traced, at least in part, to temperature gradients in the thermostat. 15 16 Figure 5. Phase diagram for atactic polypropylene fractions. . 185 — * . O. 180 " JK-b /’ O ‘ . Ogfl/JK-é 17S - - £3 (03) £70 - C)" 0 04’5/ ‘1 {3 160 - JK-S 155 " M/ W A m '. \J, ~41 193 -' l 1 l 41 I l 1 O .01 .02 .03 .0h .05 .06 .07 Weight ratio 160 158 156 (°C) 15h ‘ 152 150 lh8 Figure 6. 17 Phase diagram for atactic polybutene-l fractions. d) A-3 .01 .02 .03 .Oh Weight ratio .06 .07 Figure 7. Phase 18 diagram for isotactic polypropylene fractions. Weight ratio 180 175 — ‘ C-S M 170 TP (°C) 165 C-h _ 160 O . o O 0 C-3 ’ E-ZA S / 1 0 Q A O O \ur_(3 C) I 1 1 l I I I 0 .01 .02 .03 .0h .05 .06 .07 160 158 156 (°C) 15h 152 150 19 Figure 8. Phase diagram for isotactic polybutene-l fractions. MI-o rx l/ - :m-I. l l I I I I I .0h .05 .06 . .07 Weight ratio 20 The LCST and the molecular weights for all polymer fractions studied are listed in Table III. An indication of the reproducibility of these LCST's is shown from the data for isotactic polybutene~l fraction, HI—Y, The author obtained h2h.9°K as compared to Ballard's h2h.h°K$5 Molecular Weight Dependence of TCL If it is assumed there is a similarity in the molecular weight dependence for a LCST and.an UCST, equation (18) implies that a plot of l/TbL‘vs l/i'il/z should give e straight line. Figures 9 through ll in- dicate that this is the case. These results then confirm the idea that some aspects of the UCST theory may apply, within our error limits, to LCST. 0n the other hand, a combination of the Delmas, Patterson, and Somcynsky and the Flory theories, equation (2), gives the expression, assuming A - 0 r A plot of TCL vs 1/I11/z should give a straight line also and Figure 12 confirms the usefulness of this relationship. It was on the assumption that the parameter.A is zero that Ballard5 found fair agreement between calculated and observed TCL values. Therefore we cannot distinguish between the two treatments on the basis of this experimental data be- cause we are probably on a linear portion of both theoretical curves. Figure 17 is a plot of the Delmas, Patterson, and Somcynshy equation (1) without elimination of.A, and it can be seen that the data fall in the linear portion of the curve. The upper critical miscibility temperature fer polymer of infinite molecular weight, now called the "Flory temperature", and given the 21 Table III. LCST data for polyao-olefins. Polymer Fraction Moleezlag;§;ight kgg; isotactic polypropylene 0-5 0.11 hh5.0 " ' C-h 0.37 h30.8 u a c-3 0.97 h26.0 ' ' E-B 1.2h b2h.95 ' u E~2A h.91 h21.5 Atactic polypropylene JK-L 0.0309 h50.2 ' ' JK-6 0.0hh6 hh8.3 " ' 3—2 0.123 h36.0 n a JK-S 0.520 126.8 isotactic polybutene-l MI-h 1.79 h26.h ' “ MI-S 3.76 h2h.h ' ' MI-é h.33 h26.9 ' ' MI-7 5.00 hZh.9 " ' mm 5.00 hunt-5 atactic polybutene-l A-3 1.22 h25.6 " ' A-2 3.06 h23-9 " ” A-12l 23.0 h20.9 polyoctene-l* F-llA 0.60 393.3 " F-lOA 2.50 388.3 .. F— 9A 6.07 386.8 I F- 7A 16.0 386.5 I. F- 6A 25.0 386.6 ' F- SA h0.0 386.3 i"In butane . -22 Figure 9. LCST for polymer of infinite molecular weight, in n-pentane 2.140 - 2.38 A. \V 2036 ‘- ‘ g -‘ ~2 _ o 2.31; - °\ 2900 ‘ T 2 PM 1 2.32 - t‘ 2.30 "' 2'28"<) Isotactic polypropylene Z3 Atactic polybutene-l C] Isotactic polybutene-l 2.26— O 2.2).; l I l i 1 o 2 h 8 10 23 Figure 10. LCST for infinite molecular weight atactic polypropylene. 2.38 .' 2.36 - 2.3h - 2.32 - 2.30 ~ $900 (°K)‘l 2.28- 2.26- n-pentane 2.21;- 2.22~ O Figure 11. LCST for infinite molecular weight polyoctene-l in n- 2h pentane and in n-butane. 2.59 _ 2.58 2.57 2.55 2-5b Fit—4 ,0! 3AA (°K)"1 2.28‘ 2.271 2.26 2.25 2.2h -4 O in n-butane in n—pentane 25 Figure 12. LCST for isotactic polypropylene of infinite molecular weight. ME- 0 M0 h35 L30 * CL (010 825 26 symbol 6, is the solution ana10gue of the Boyle temperature for a gas. That is, at 0 the intermolecular forces which cause the polvner to ex- pand are exactly counterbalanced.hy the intra-molecular segment-segment forces which cause the polymer to contract. At this special tempera- ture, polymer solutions become ideal in their behavior and the second virial coefficient vanishes. He BOW'prOpOSS the Flory temperature be symbolized.by 6U.and the new temperature extrapolated to infinite molecular weight for a LCST (Figures 9 through 11) be symbolized by 9L5 and defined as the maximum temperature at which solvent and polymer of infinite molecular weight can be maintained in a single phase. At this temperature we suggest the intra-molecular forces which cause the sol— vent to expand are Just counterbalanced.by the intermolecular forces which.prevent this expansion. That is, if these solutions have a negative ARM, the solvent-polymer interactions are extremely favorable and should oppose the general expansion in the solvent as it approaches its critical temperature. The temperature, 9 and the slopes for the curves in Figures 9 L through ll are listed in Table IV. The slopes of these curves are quite significant for ucsr in that they permit the calculation of L4) 1, the solvent entropy of interaction parameter [see equation (18)]. How- ever, the slopes in Table IV cannot be definitively interpreted since the change in the molar volume of the solvent in this temperature range is unknown. However, it is interesting that the sign of the slopes is constant and their values fall within a narrow range. It is also significant that the slopes andeL for atactic polypropylene is greater than the corresponding values for isotactic polypropylene whereas these are reversed for the atactic and isotactic polybutene-l system. This 27 Table IV. LCST for polymer of infinite molecular weight Po :- Solvent 51 ‘ - intercept e e W «firzzoxm [fin] vii) j’:_ Isotactic P.P. pentane -0.01hh 2.3985 h17.8 0.889 Atactic P.P. pentane -0.0lOl 2.388 1.18.8 0.891 Isotactic P.B. pentane 43.00862 2.370 1421.9 0.898 Atactic P.B. pentane -0.0117 2.3815 hl9.9 0.893 Polyoctene-l pentane -0.0lh2 2.285 h37.6 0.931 Polyoctene-l butane ~o.0112 2.595 385.1; 0.907 28 behavior has also been observed in UCST studiesl3sl4 and give additional evidence that the thermodynamic interaction between solvent and.polymer depends on the chain geometry but the mgnitude of the effect is wholely solvent dependent. Although these polymer-solvent systems are distinctly nonpolar, the molecular weight dependence gives a reasonable fit to the data12 for poly- ethyleneoxide in water solutions also. .Figure 13 illustrates that aqueous solutions of polar polyethyleneoxide also obey equation (18). The LCST found in this system.undoubtedly involve hydrOQen-bond rupture. The Dependence of LCST oanolvent To study the dependence of the LCST on solvent, phase diagrams were obtained for a single polyoctene-l fraction in four different hydrocar- bon solvents. This data is illustrated in Figure it and the LCST are summarized in Table V} We have seen that the LCST is invariant with polymer weight fraction (Figures 5 through 8) for different molecular weight fractions. Figure 1h shows that the LCST for the same polymer fraction in different solvents is also invariant with the weight frac- tion of the polymer. Therefore the LCST appear somewhat insensitive to solvent also. .An attempt was made to find a relationship between the LCST and the critical temperature, Tc, of the solvent (see Tables IV and V) but a constant ratio does not exist, aaLthough the solvent ap- pears to be within 9/10 of its critical temperature before the phase break. “It is presumably the decreasing configurational energy and increas— ing molar volume of the pure solvent as it approaches its own gas-liquid critical point that makes it a 'poorer' solvent for the polymer.'1 The 29 Figure 13. LCST for infinite molecular weight polyethyleneoxide. 2.80 '- 2.70 CL (°K) 2.60 2.50 2.1.LO i l I I E 1 l 6 8 10 12 11; 1000 /M1/ 2 30 C Figure 1b. Phase diagram for a polyoctene-l fractiofisin diiferezt solvents. Butane 11h -— \GNG—O W G) 112 - c) Ne0pentane W ISObUtane 0 \AA r V”V\ PrOpane 38 ”err—6.0. Molecular weight = 2,500,000 I I 1 n . 0 .01 .02 .03 .0h .05 .06 .07 Weight ratio 31 Table V. LCST data of one polyoctene-l fraction (F-6A) in several solvents. 50mm; (515 (20‘ (£101). :91; Tc 1‘ c .. Propm ~ 6.3 370 309 0.835 61 Isobutane 6.25 hO8 357 0.875 51 n—butane 6.7 1.25 387 0.911 38 chpentane 6.2 hBh 38h 0.835 50 Pentane 7.1 1.70 1:395 0.931; 31 32 LCST of polyoctene-l does not follow the Hildebrandv »: Smxwzanm om ©m3~.sa as; accuse. gown Humowmo: om vcwmzn 36 Table VI. Delmas, Patterson, and.SomcynsKy constants calculated by least squares fit of the experimental data. A B Polymer Solvent (cal/base mole) (cal/degz base mole) polyoctene-l n-butane h0.h3 0.00h7h atactic polypropylene n-pentane 69.28 0.00357 isotactic polypropylene n-pentane 67.h5 0.00365 atactic polybutene-l napentane 67.92 0.00363 isotactic polybutene-l n~pentane 66.h2 0.00371 N N Y’ NA-oBZ-i—znZ-f- (21) i=1 1 1-1 1 where; 1.1 x3...— 701. y‘R% N - number of data points. Figure 17 is I plot of equation (1) using the constants A and B calcu- lated by a least squares fit of the polyoctene—l data in nebutane. All the experimental data points lie'within the small rectangle on the curve. Notice in.Table VI that the values arch and B for atactic and iso- tactic polypropylene follow the same trend as the atactic and isotactic polybutene-l. .A reverse trend was noted earlier (Table IV). Also the values of.A and B are surprisingly constant for the range of polymers that were studied. According to the Delmas‘ theory A . zef; J: m/B (22) B - 10.5 (kl/éeg:)N (23) where: 6: a minimum potential energy of interaction of 2 1’J segments of type i and J. z - coordination number N -,Avogadro's number k - Boltzman constant or aw" €11 Therefore, A is the parameter that takes into account the solvent-sol“ vent interactions while B takes care of the solvent-polymer interactions. Ballard had some success predicting the LCST and EL by assuming A was 38 Figure 1?. Plot of Delmas, Patterson, and S mcynsky equation" for LCST using constants calculated by least squares fit of exper- imental data. ' u ' T "Rx. = mung-:1.) +0.00h7h (ii-Ii 39 zero. This assumption is reasonable according to the authors“1 of the theory, when the polymer and solvent differ only in chain lengths. First of all, a polyoctene~l chain with its six~carbon pendant group on every other carbon may not be very similar to n~pentane. Secondly, a fairly large (Delmas'q values range from 10 to 22) value of.A was obtained by” the least squares fit of the experimental data for the same polymer in n—butane (Table VI). Thirdly, an attempt was made to predict 6L with A equal to zero and those calculated values are approximately twice the observed values (see Table VIII). Therefore, Ballard apparently was not justified in letting A c O. The reason he found fair agreement with experimental values is probably because he determined B from one of his experimental points and this lat- ter value corrected for the null A value. The constants A and B in Table VI were used in equation (1) to calculate the LCST andeL for some of the systems studied. The results of these calculations are in Tables VII and VIII. The agreement for the LCST of fraction F-éA in.propane is excellent while the agreement of the same fraction in n-pentane is within 6%. A comparison of the observed and calculated 6 is presented in L Table VIII.‘ Here also the worst agreement is for polyoctene~1 in n- pentane but even then the error is less than 6%. BL is an interesting value in that it represents a combination of the Flory theory, equation (18) and the Delmas theory, equation (1), i.e., 6 is determined from L a.plot resulting fron.the Flory equation and it is calculated from the Delmas theory. The agreement appears very good. ,A calculated 6L for atactic polypropylene is not found in Table VIII because the calculation produced imaginary roots. This resulted to Comparison of calculated and experimentally'diermined LCST for polyoctene-l (F~6A) in different solvents. Table VII. Solvent TCL(°K) UCST (0K) calc. exp. calc. propane 309.h 309.2 110.2 napentane h6h.2 h39.05 165.6 Table VIII. Comparison of the calculated and experimentally determined 0 for all the polymers studied. L o Polymer Solvent calc. 92b:.K) calc v. $g:::3) A80 polyoctene-l propane 308.5 - -- 110.5 polyoctene-l butane 385.6 385.1. 523.8 138.2 polyoctene-l pentane h62.7 h37.6 -- 165.8 isotactic polypropylene pentane h22.h L17.8 815.6 398.3 atactic polypropylene pentane .~- h18.8 835.6 -- isotactic polyhtene-l pentane h22.1 h21.9 80h.§ 382.1 atactic polybutene-l pentane hi7.7 h19.9 820.7 h03.o hi £rom.too large a magnitude in either A or B. For example, subtracting the small quantity 0.27? from.A; 69.277; gives a real root very close to the experimental value. Tables VII and VIII also list calculated UCST and 911' Apparently these calculated.values do not predict experimental fact. For example, the calculated.values predict that atactic polybutene-l or infinite molecular weight will go into solution only above 130°C, yet it is soluble at room.temperature. The calculated values, 9U, could not be checked experimentally for the isotactic fractions because the theory calls for liquid-liquid separation and the isotactic polymers separate as a crystalline phase at a higher temperature than that predicted. A.solution of atactic polybutene—l and one of atactic polypropylene each in napentane were cooled to the freezing point of the solvent (IhZQK) without observing any precipitation. For some reason the Delmas theory*predicts very well LCST and GL (within 6%) but does not agree with experiment for UCST. From the above then; it can.be seen that the existing theories of polymer solutions are inadequate to treat accurately LCST and fail completely to predict both an UCST and a LCST. Obviously this is an area where experimental work is ahead of theoretical development. 10. 11. 12. 13. 1h. 15. P. J. C. G. L. I. P. J. I. S. H. REFERENCES Freeman and J. S. Rowlinson, Polmr _1_, 20 (1960). Rowlinson and P. I. Freeman, Pure Appl. Chem, _2_, 329 (1961). Baker, at 11., Pagans-r 2, 215 (1962). Delmas, D. Patterson, and T. Somcynslqy, J. Pol. Sci. £1, 79 (1962). Ballard, "Dilute Solution Properties of Polyoctene-l“, Ph.D. Thesis, Michigan State University, 1963. Prigogine and R. Defay, Chemical Thermodynamics, London: J. B. Longmans Green and Co., 1951;, p. 2min.) p. 285. Flory, Principles of Polymer Chemistry, Ithaca, N.Y.; Cornell University Press, 1933, p. 5145; (a) p. 51.6. Kinsinger and R. E. Hughes, J. Phys. Chem. 61, 2002 (1959). Intrinsic viscosities were determined by Mr. Donal Streeter of the Dow Chemical Company. Curtis Wilkins, “Solution Properties of Atactic and Isotactic W. F. R. J. J. R. E. A. P. Polybutene-1", Ph.D. Thesis, Michigan State University, 1963. iérigbgum, J. E. Kurz, and P. Smith, J. Phys. Chem. 65, 1981; 1961 . Bailey, Jr. and R. w. Callard, J. Appl. Pol. Sci. 1, 56 (1959). Uessling, "The Dependence of Phase Equilibria on the Config- uration of Polypropylem,‘ 11.3. Thesis, Michigan State University, 1959. Mullooly, "The Dependence of Phase Equilibria on the Configur- ation of Polyhutene-l," 11.5. Thesis, Michigan State University, 1961. Hildebrand and R. L. Scott, Regular Solutions, Englewood Cliffs, New Jersey: Prentice-Hall, 1962, p. 171. he APPENDICES 113 APPENDIX I Phase separation teaperatures for polymer fractions in n—pentane. hh Ugt. of P01. Ugt. of Solv. 1‘ (mg) (mg) (.2, __ Fraction C53 8.30 265.b5 15h.h 11.21 58.81 1511.0 209; 149-20 15307 1.20 59.10 153.h 0.95 72.20 151. 0.95 65.85 151.2 2.95 19.20 151.3 1.20 59.10 155.1 5.56 130.36 153.8 1.10 101.50 153.7 0.15 89.5h 15h.2 2.66 106.17 153.1 6.60 199.27 153.1 2.98 180.20 153.1 2.10 59.00 153.5 0.97 96.63 153.5 Fraction C—h 0.10 232.55 163.8 0.61 101.69 160.8 0.77 83.92 160.1 0.111 35-53 158-9 0.98 30.12 158.h 1.19 12.55 158.3 7.13 33.62 158.9 3.17 79.57 158.0 0.98 30.12 158.9 3017 79-57 15803 0.08 111.82 163.5 0050 814.91 15903 2.52 16.90 159.0 2.12 57.79 159.0 1.76 62.71 158.2 0.61 101.69 159.8 Fraction C-S 2.15 121.30 175.3 2.15 90.90 172.3 0.93 117.36 176.8 1.17 101.25 172.7 3.39 121.11 172.0 15 Appendix I (cont.) Hgt. of P01. Wgt. of Solv. Tp (mg) (mg) (cc) Fraction C-S (c0nt.) 1.21 101.31 172.7 3.58 71.19 173.1 1.00 106.10 173.5 -0.58 127.70 176.5 Fraction E-ZA 0.19 151.10 151.9 1.28 162.11 150.9 0.13 135.37 153. 2 1.89 132.13 119. 2 2.68 * 119.16 119.1 3.50 116.10 118. 6 5.21 151.36 118 3 1.35 73.69 118. 9 1.60 52.25 119 3 2.60 59.92 119.7 2.99 63.20 118.2 5.18 80.00 150.9 Fraction JK-h 1.62 218.5 186.2 1.98 263.6 178.8 5.17 218.1 178.3 2.88 257.7 183.5 5.21 105.7 177.5 5.10 125.0 177.3 1.11 111.5 177.7 6.96 130.3 177.5 Fraction JK-S 6.31 115.8 151.1 1.91 150.0 153. 8 1.25 188.0 151.0 3.30 239.1 151. 2 6.39 131.1 151 h 7.16 116.6 151.1 3.71 188.0 153. 8 2.87 221.0 151.3 10.21’ 311.6 153.7 11.97 273.6 151.1 11.95 239.1 151.2 9.66 288.1 151.5 16 Appendix I (cont.) Wgt. of P01. 291. of 361v. 1 (mg) (mg) (08) Fraction JK~6 3.85 237.5 178.3 1.52 230.0 182.0 2.21 208.5 179.1 2.11 92.7 171.1 1.00 - 110.1 176.0 3.97 93.1 175.1 5.16 92.0 176.6 7.99 83.2 178.0 1.85 129.0 182.5 3.71 158.2 179.0 1.13 138.5 175.5 5.76 135.1 175. Fraction B—2 1.76 170.1 161.3 2.20 113.0 163.5 2.61 128.7 163.0 3.07 133.1 163.3 1.12 138.3 163.1 1.28 130.9 163.2 5.76 120.5 163.5 Fraction MI-1 1.32 112.80 153.7 1.13 103.22 153.8 0.82 80.93 153.6 1.17 71.30 153.5 3.05 119.85 153.3 1.12 251.71 153.9 5.53 73.92 157.3 0.36 171.06 151.9 0.55 296.70 155.6 0.22 285.51 160.9 2.76 112.35 153. 5.68 91.69 155.5 Fraction MI-S 0.53 115.67 151.7 2.82 190.10 151.5 2.55 215.25 151.6 2.76 137.76 151.5 2.90 111.80 151.5 3.23 108.85 151.5 5.12 131.63 152.1 17 Appendix I (cont.) Hgt. of P01. Ugt. of‘Solv. 1P (m9) (m9) (°C) Fraction MI-6 0.10 121.38 159.1 0.29 100.31 151.0 1.65 120.55 153.9 0.80 89.55 151.0 2.92 750.98 157.1 2.10 88.50 155.3 Fraction MI-Y 0.19 161.16 152.3 0.12 150.65 156.0 2.06 116.89 151.9 2.22 97.23 152.1 2.86 97.91 152.2 3.59 102.21 152.6 6.93 101.75 153.9 Fraction A-121 63.5 1016.1 150.3 31.3 851.7 118.3 16.3 519.5 118. 1001‘ 1118.3 1148.6 6.9 158.6 117. 3.1 387.6 118.0 1.7 575.1 118.1 1.0 1216.1 119. Fraction A-Z 212.2 3516.7 152.0 207.1 5219.5 151.1 193.3 6113.5 151.2 12.9 869.2 151.2 9.5 1191.6 151.9 7.1 2368.0 153.2 5.6 7022.9 156.6 170.2 6811.3 151.0 ‘ Fraction A-B 257.2 2119.6 155.0 17.8 225.2 151.8 65.1 1096.0 153.5 11.3 1101.5 152.9 27.6 1382.5 152. 16.1 1610.1 151.0 7.9 1578.7 156.7 1.3 1282.3 161.2 APPENDIX II Phase separation temperatures for potyoctcnc-l fractions in n-butane. Hgt. of‘Pol. Hgt. of 501v. Tp (mg) (mg) (00) Fraction F-SA 1.26 133.1 113.9 1.50 101.1 113.6 2.16 99.7 113.5 2.89 111.3 113.6 3.07 111.3 113.1 1.70 113.7 113.5 5.37 88.0 113.8 1.90 190.1 113.6 3.00 158.1 113.7 3.13 129.6 113.3 1.00 108.5 113.1 Fraction F-éA 1.50 151.5 113.9 2.19 168.7 113.8 2.26 136.0 113.7 2.55 107.5 113.6 3.56 83.6 111.2 3.55 117.0 111.1 1.21 98.3 111.1 3.32 111.5 113.7 Fraction F—7A 1.13 206.5 113.5 5.05 169.1 113.7 5.91 183.1 113.1 7.59 155.1 113.7 7.03 190.0 113.8 1.27 255.6 111.1 2.11 259.1 113.9 3.07 152.3 113.0 3.71 132.1 113.8 Fraction F-9A 1.26 139.7 111.1 1.55 109.0 113.8 1.97 113.6 113.8 3.10 96.9 113.9 3009 she} Inch 2.70 83.5 111. 18 . 19 Appendix II (cont.) Hgt. of P01. Hgt. of Solv. TP (”19) (mg) (Dc) Fraction F-lOA 1.31 133.6 116.1 1.85 119.2 115.3 3.02 118.7 115.2 1.06 117.0 115.2 1.05 126.5 115.3 6.18 111.3 115.1 2.95 151.2 115.7 3.75 151.3 115.5 5.27 153.9 115.1 Fraction F-llA 1.88 195.1 121.9 2.29 119.1 121.2 2.56 133.9 120.7 3.57 122.1 120.1 3.75 116.9 120.3 5.20 125.2 120.5 5.61 111.1 120.7 APPENDIX III Phase separation temperatures for the polyoctenc—l fraction, F-éA, in different solvents. Hgt. of P01. Wgt. of 801v. Tp (mg) (mg) (°C) Neopentans 2.10 257.9 111.8 2.61 230.2 111.7 3.81 210.7 110.6 3.30 212.3 112.2 3.86 197.1 111.1 3.87 160.2 111.1 6.50 203.0 111.7 1.85 216.1 111.1 Isobutanc 2.01 201.7 81.9 2.09 135.9 81.6 3.06 156.3 81.6 b.1117 156.0 8’40 1.08 159.9 81.3 1.53 137.3 81.6 1.52 126.0 81.6 5.31 139.2 81.5 Propane 1.11 158.1 36.7 2.12 135.9 37.0 2.35 112.0 36.2 2.97 119.0 36.3 1.72 135.9 37.3 1.70 127.1 37.1 See Appendix II for F-éA in n-butanc. SO APPENDIX IV Phase separation temperatures for the poxyoctene-l fraction, F-7A, in mixtures of n—pentane and n-butane. W96. 01' POI Wgt. of Pent. Wgt. of But. '1: (mg) - (mg) (mg) op ( C) 0.95 17.1 51.6 110.9 1.89 60.5 59.1 112.1 2.85 83.3 78.1 112.7 1.19 62.3 71.1 110.3 3.32 68.1 58.1 113.2 3.32 56.1 59.5 110.8 5.10 58.1 75.5 137.0 1.17 112.9 39.1 156.9 2.51 125.1 28.6 155.8 2.52 90.1 31.7 153.1 2.61 78.0 15.0 158. 3.30 82.2 21.2 156.5 3.71 71.3 22.6 152.9 3.58 71.5 11.8 158.1 1.77 77.6 11.6 158.7 1.56 11.2 127.9 126.1 1.66 38.1 100.3 128.1 2075 3502 71101 132.1 3.08 10.5 108.8" 127.9 1.25 27.1 73.2 129.2 3.55 26.1 80.1 127.8 3007 21105 52.5 13300 1.67 23.1 62.5 130.0 3.99 130.1 2.2 165.2 1.18 21.3 231.1 118.9 CnL'MkSTRY Liam MICHIGAN STATE UNIVER ITY L I IN lllllllillllllll 0 1 3 1293 3 75 098 IBRARIES 1