Ind. Eng. Chem. Res.

2000, 39, 3069-3075

3069

Fluid-Liquid and Fluid-Solid Transitions of Poly(ethylene-co-octene-1) in Sub- and Supercritical Propane Solutions
Ka Chun Chan, Hertanto Adidharma, and Maciej Radosz*
Department of Chemical Engineering and Macromolecular Studies Group, Louisiana State University, Baton Rouge, Louisiana 70803-7303

Fluid-liquid (FL) and fluid-solid (FS) phase transitions are measured and correlated with two SAFT equations of state for poly(ethylene-co-octene-1) solutions in sub- and supercritical propane. The polymer concentration is found to have a weak effect on both FL and FS transitions in the range of this study (0.01-0.1 polymer weight fraction). Increasing molecular weight is found to increase the FL pressure and FS temperature, whereas increasing branch density is found to decrease the FL pressure and FS temperature.
Introduction The phase behavior of polymer solutions in sub- and supercritical fluids is usually a major unknown in designing new polymerization reactors, heat exchangers, and separators that can operate in an efficient and fouling-free manner.1 This is the case, for example, for ethylene copolymers, such as poly(ethylene-co-octene1) (EO), where we need thermodynamic models and experimental data. Many investigators have reported phase-equilibrium data for polyethylene and ethylene copolymers in hydrocarbons: for example, polyethylene in ethylene;2 polyethylene in propane;3,4 poly(ethylene-co-propylene) with varying propylene concentration, and hence polymer crystallizability, in C5-C9 alkanes;5 and propylene, 1-butene, 1-hexene, and 1-octene copolymers of ethylene in propane.6 Also Whaley reported the phase-equilibrium data for EO in propane.7 Those previous studies concentrated on the effects of molecular weight and branch density on the fluid-liquid (FL) phase behavior, usually characterized in terms of the cloud-point pressure. However, the fluid-solid (FS) phase behavior of ethylene copolymers in sub- and supercritical fluids has not been well characterized yet. The objective of this preliminary study, therefore, is to understand how FL and FS phase transitions of EO depend on the concentration of polymer, its molecular weight, and its branch density in sub- and supercritical propane solutions. The approach is to measure both transitions in a batch optical cell with a transmittedlight probe and then model the experimental data with a polymer equation of state. Experimental Section Apparatus. FL and FS phase equilibria are measured in a batch optical cell with a transmitted-light probe. A simplified schematic of the unit is shown in Figure 1. A detailed description of the equipment and procedures is provided by Chan et al.8 In brief, the phase transitions can be either observed visually on a video
* Corresponding authors new address: Department of Chemical and Petroleum Engineering, University of Wyoming, Laramie, WY 82071-3295. E-mail: radosz@uwyo.edu. Telephone: 307-766-2500. Fax: 307-766-6777.

Figure 1. Schematic of the batch optical cell with the transmittedlight probe: P, pressure transducer; T, temperature probe; D, photodiode; G, Heise gauge.

monitor, via a borescope, or detected using a transmitted-light probe. The accuracy for the pressure transducer is (0.5 bar and is (0.1 C for the temperature probe. A known amount of polymer and solvent is placed in the cell and well mixed until equilibrium is reached. The FL pressure, also referred to as the cloud-point pressure, is measured at constant temperature, and the FS temperature is measured at constant pressure. The cloud-point pressure is also characterized visually as being of either the bubble- or dew-point type.8 The FL pressure is measured by lowering the pressure until the mixture turns cloudy, which sharply decreases the transmitted-light intensity. The result is then checked by raising the pressure until the mixture is completely clear. The difference is found to be insignificant (less than 2 bar). Therefore, the pressure at which the mixture turns cloudy is taken as the FL pressure in this study. The FS temperature is measured isobarically by both cooling and heating; during cooling, the solution turns cloudy, and during heating, the cloudy solution turns clear. Both the cooling and heating rates are kept at less than 1 Cmin-1 to minimize kinetic effects. The difference between the cooling- and heating-induced FS temperature is found to be around 10-15 K. This is expected as suggested, for example, by Kohn et al.9 and Condo et al.3 The reason for this discrepancy between the cooling- and heating-induced FS temperature is a subcooling effect, which inhibits the nucleation of crystals in solution. The superheating effect, which inhibits the dissolution of crystals, is

10.1021/ie990761e CCC: $19.00 2000 American Chemical Society Published on Web 06/22/2000

3070

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000

ares ) aseg + achain

(1)

Figure 2. Generic EO structure. Table 1. Properties of Polymer Samples polymer sample EO-4-80K EO-8-120K EO-14-120K wt % octene 13.9 25.8 38.7 BD 4 8 13.6 Mw/kg mol-1 83 115 120 Mw/Mn 1.62 2.1 1.44 DSC Th/Ca 105.5 87.2 55.5 DSC Tc/Ca 92.3 68.3 41.8

a T ) heating-induced temperature and T ) cooling-induced h c temperature.

believed to be usually less significant and, hence, the heating-induced FS temperature is usually accepted as being closer to a true equilibrium FS temperature. Materials. Three EO samples are used in this work. They all have a generic structure shown in Figure 2; the number of side hexyl branches (shaded segments) per unit backbone (blank segments) varies from sample to sample. These samples are characterized with differential scanning calorimetry (DSC), temperature rising elution fractionation (TREF), and gel permeation chromatography (GPC). Their properties are given in Table 1. These samples are coded in terms of the branch density (BD) and weight-averaged molecular weight (Mw). BD is defined as the number of branches per 100 ethyl units in the polymer backbone, and hence it is a measure of the comonomer incorporation. For example, EO-4-80K has a BD of 4 and a Mw of about 80 000 gmol-1. EO-4-80K is a cross-fractionated sample: first fractionated with respect to crystallinity using preparative TREF in propane and then refractionated with respect to molecular weight using a pressure-increasing mode in supercritical propane. EO-14-120K is fractionated only with respect to molecular weight in supercritical propane, and EO-8-120K is used without fractionation or purification. Propane (99.5% purity) is obtained from Praxair and also used without further purification. SAFT Modeling The experimental data obtained in this work are correlated with two versions of the SAFT equations of state, SAFT10,11 and SAFT1,12 which enable us explicitly to account for variable polymer microstructure because of the variability in comonomer incorporation. SAFT is developed on the basis of an argon equation of state for the segment term and a hard-sphere pair-correlation function for the chain term10 and extended to heterosegmented molecules, such as copolymers.11 SAFT1 is developed on the basis of a perturbation theory of square-well fluids (Barker-Hendersons perturbation theory) for the segment term and a square-well paircorrelation function for the chain term.12 For nonassociating fluids, the dimensionless residual Helmholtz energy of chain molecules ares in the SAFT equation of state is given by

where aseg is the segment term that accounts for nonideality of the reference fluid of nonbonded chain segments (monomers) and achain is the chain term that accounts for covalent bonding. The parameters for both versions of SAFT are the segment number m, the segment molar volume oo in mLmol-1, and the segment energy u/k in K. SAFT1 also requires an extra parameter , which is the reduced range of the potential well. All of the parameter values used in this work are given in Table 2. The detailed equations are not presented here because they can be found in ref 10 and 11 for SAFT and in ref 12 for SAFT1. Although the polymers used in this work are polydisperse, SAFT and SAFT1 model the polymer as a monodisperse component. We model EO as a heterosegmented chain, as shown in Figure 2. It consists of two types of segments, the backbone type and the branch type, which are connected with three types of bonds, backbone-backbone, backbone-branch, and branch-branch. Each segment type has a unique set of SAFT parameters. The parameters for the backbone-type segments are obtained from a correlation developed for long n-alkanes.10-12 Because six-carbon branches are attached to a polyethylene backbone in EO, the branch-type segments are assumed to have the same SAFT segment parameters as nhexane. EO chains are characterized with two structure parameters that can be estimated from the branch density or the comonomer fraction. The first structure parameter is the segment fraction

R,i ) mR,i/mi

(2)

where mR,i is the number of segments of type R in chain i and mi is the number of all segments in chain i. For EO, the following working equations are used to calculate the backbone-type and the branch-type segment fractions:

backbone )

200 200 + 6BD

(3) (4)

branch ) 1 - backbone

The segment fraction R,i is used to calculate the segment fraction of type R in the whole mixture

xR )

R,imiXi i miXi i

(5)

where Xi is the mole fraction of chain i (component i) in the mixture. The second structure parameter is the bond fraction

BR,i )

nR,i

njk,i k>j

(6)

where the term in the numerator is the number of bonds of type R in chain i and the term in the denominator is the total number of all bonds in chain i. For EO, the

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000 3071

Table 2. SAFT and SAFT1 Parameters for EO + Propane components SAFT propane EO-0-10K EO-0-120K EO-4-80K EO-8-120K EO-14-120K propane EO-0-10K EO-0-120K EO-4-80K EO-8-120K EO-14-120K m 2.696 532.480 4688.728 2379.701 2552.295 3872.405 1.667 271.517 2388.800 1212.530 1300.455 1972.948 (oo)bb/mLmol-1 (oo)br/mLmol-1 (u/k)bb/K (u/k)br/K ()bb ()br 13.457a 12.00a 12.00a 12.00 12.00 12.00 16.747a 25.172a 25.207a 25.203 25.203 25.206 21.086 21.086 21.086 12.475 12.475 12.475 210.0 210.0 210.0 193.03a 210.0a 210.0a 202.720 202.720 202.720 193.223a 281.479a 281.985a 281.922 222.578 281.930 222.578 281.971 222.578 n/a n/a n/a n/a n/a n/a 1.6914a 1.6545a 1.6542a 1.6542 1.6885 1.6542 1.6885 1.6542 1.6885

SAFT1

For propane and polymer without branches, there is only one u/k, oo, and for the entire molecule. Table 4. Melting Temperature, Number of Crystal Units, and Parameter c for FS Calculations polymer sample EO-0-10K EO-0-120K EO-4-80K EO-8-120K EO-14-120K
a

Table 3. Bond Fractions and Segment Fractions for Polymers polymer samples EO-0-10Kb EO-0-120Kb EB-5-100Kc EH-5-100Kc EO-4-80K EO-8-120K EO-14-120K
a

bb-br br-br bb br bb-bb bond bond segment segment bond fractiona fractiona fractiona fraction fraction 1 1 0.950 0.915 0.893 0.807 0.710 0 0 0.025 0.021 0.018 0.032 0.048 0 0 0.025 0.064 0.089 0.161 0.242 1 1 0.950 0.915 0.893 0.807 0.710 0 0 0.050 0.085 0.107 0.193 0.290

melting temp/Ca 141.85 141.85 135.07 128.98 117.82

no. of crystal units, u 407 3589 1568 1483 1817

c from SAFT 0.85 0.85 0.69 0.67 0.66

c from SAFT1 1.00 1.00 0.81 0.77 0.75

Melting temperature calculated from eq 11.

bb-bb ) backbone to backbone, bb-br ) backbone to branch, br-br ) branch to branch. b EO-0-10K ) NIST 1482 and EO-0120K ) NIST 1484 (data from Condo et al.3). Mw: EO-0-10K ) 13 600 g/mol with Mw/Mn ) 1.19; EO-0-120K ) 119 600 g/mol with Mw/Mn ) 1.19. c EB ) poly(ethylene-co-butene-1) and EH ) poly(ethylene-co-hexene-1) (data from Han et al.5).

following working equations are used to calculate the three bond fractions:

Bbackbone-backbone ) Bbackbone-branch )

200 200 + 6BD

(7) (8)

BD 200 + 6BD

Tm is the equilibrium melting temperature of pure polymer, Hu is the enthalpy of melting per mole crystal unit at Tm (8.22 kJ/mol of crystal unit),14 v is the molar volume difference between pure liquid polymer and pure solid polymer (4.937 cm3/mol of crystal unit),15 and uc is the effective number of crystal units in a molecule, that is, the units that are actually crystallized at any given set of conditions. The equilibrium melting temperature of pure copolymer Tm depends on the melting point of the homopolymer Tm (415 K),14 Hu, and the mole fraction of the crystal units p according to the following relationship:16

Bbranch-branch ) 1 - Bbackbone-backbone - Bbackbone-branch (9)

The values for all of the segment and bond fractions are given in Table 3. Fluid-Solid Equilibria Calculation SAFT has been applied to calculating the crystallization temperature of crystalline and semicrystalline binary systems,13 where the crystalline phase is assumed to be pure polymer and the density of the pure liquid and pure solid polymer are assumed to be independent of temperature and pressure. In this work, we show that these assumptions are still reasonable for a first pass accurate to represent the FS temperatures for EO + propane systems as well. The final expression for this approximation, developed from the equality of fugacities of the polymer in the solution and solid phase in ref 13, is as follows:

1 R 1 ) ln p Tm Tm Hu

(11)

To determine uc, we define an auxiliary parameter c as follows:

c uc/u

(12)

where u is the total number of crystallizable units in a molecule. c is estimated by fitting to experimental data. The values for Tm, u, and c are reported in Table 4. Following Flory,16 u is taken to be the total number of ethylene units per molecule. Hence, u can be determined from the molecular structure alone; the higher the branch density of the EO, the lower the number of ethylene units. On the other hand, uc may also depend on other factors, such as solvent type and solvent ratio. In this work we use only one solvent, i.e., propane; thus, we only explore the effect of molecular structure on uc. The other factors will be explored in our future work. Results and Discussion

ln

( ) [ ( )

pLxpL Hu Tm P -1 + )u p RTm T RT c

(10)

where pL is the fugacity coefficient of polymer in solution, p is the fugacity coefficient of pure liquid polymer, xpL is the polymer mole fraction in solution,

The experimental phase-transition data obtained in this work are reported in Table 5. These data provide a basis for estimating kijs and c. Because the EO molecules consist of two types of segments, backbone-type and branch-type, three kijs

3072

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000

Table 5. Experimental Data for EO + Propane polymer (weight fraction) EO-4-80K (0.05) T/C 180.0 140.0 120.0 100.0 95.0 90.0 91.1 93.1 97.0 80.5 80.8 83.8 180.0 140.0 100.0 83.3 84.2 86.1 72.9 74.1 76.9 180.0 140.0 100.0 70.0 60.0 47.3 47.8 52.1 37.1 38.2 41.7 P/bar 534 528 528 532 535 537 700 1000 1400 700 1000 1400 485 471 463 700 1000 1400 700 1000 1400 408 384 356 337 336 700 1000 1400 700 1000 1400 transition typea FL FL FL FL FL FL FS* FS* FS* FS FS FS FL FL FL FS* FS* FS* FS FS FS FL FL FL FL FL FS* FS* FS* FS FS FS polymer (weight fraction) EO-14-120K (0.01) T/C 180.0 140.0 100.0 70.0 60.0 40.1 40.8 43.2 29.1 29.3 31.5 180.0 140.0 100.0 53.3 53.9 56.1 42.5 43.1 44.7 P/bar 386 360 332 348 378 700 1000 1400 700 1000 1400 404 379 355 700 1000 1400 700 1000 1400 transition type FL FL FL FL FL FS* FS* FS* FS FS FS FL FL FL FS* FS* FS* FS FS FS

EO-14-120K (0.10)

EO-8-120K (0.049)

EO-14-120K (0.049)

Transition type: FL ) fluid-liquid transition, FS* ) fluid-solid transition upon heating, FS ) fluid-solid transition upon cooling.

about 0.05. We note that this parameter is likely to be somewhat model specific. This is shown in Table 4 and Figure 4, where we observe a systematic difference in c derived from SAFT and SAFT1. We find that c decreases as the percent comonomer, and hence the branch density, increases; c reaches an asymptotic value at about 40 wt % comonomer. This behavior is well represented by the following empirical equation:

c ) c0 + c1e-(wt % comonomer)/c2

(13)

Figure 3. kij as a function of temperature.

must be estimated, i.e., kbackbone-propane, kbranch-propane, and kbackbone-branch for both SAFT and SAFT1. We set kbackbone-branch ) 0 and fit the other kijs to experimental FL pressures; kbackbone-propane is fitted to the experimental data for EO-0-120K + propane taken from ref 3 at a polymer weight fraction of 0.0025-0.075, and kbranch-propane is fitted to the cloud-point data for EO14-120K + propane from this work at a polymer weight fraction of 0.049. These kijs are allowed to be temperature dependent. Figure 3 shows a plot of kij as a function of temperature, and Table 6 gives kij correlations. The parameter c is fitted to the FS transition data for EO-0-120K, EO-4-80K, and EO-14-120K + propane from this work at a polymer weight fraction of

where c0, c1, and c2 are constants listed in Table 7. This equation is illustrated in Figure 4 for SAFT and SAFT1. The kij correlations given in Table 6 and eq 13 are used for all the predictions discussed below without further readjustment. First, however, Figure 5 compares the experimental results of this work for EO in propane with those of Han et al.,6 for poly(ethylene-co-butene-1) (EB) and poly(ethylene-co-hexene-1) (EH) also in propane, and at a similar BD of about 4. Han et al. find that increasing the branch length, i.e., going from the ethyl branch in EB to the butyl branch in EH, decreases the FL pressure. Our results are essentially consistent with this trend: the EO pressure is lower than the EH pressure. For the record, the FL transition observed at the lowest temperature, around 90 C in Figure 5, is less reliable; the FL transitions near the FS boundary are hard to distinguish from the FS transitions. Figure 6 shows a P-T diagram measured in this work for EO-14-120K + propane. The FL points (filled) exhibit a flat U-LCST behavior. The FS points (open) suggest a slightly positive slope, which means that the FS temperature increases with increasing pressure. The

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000 3073

Table 6. Binary Interaction Parameters kij as an Empirical Function of Temperature kij kbackbone-propane kbranch-propane kbackbone-propane kbranch-propane
a

correlationa SAFT 0.0218 0.14249 - (4.45455 10-4)T SAFT1 0.0058 - 6 10-5(T - 393.15) T - 42.49081 -0.03101 - (1.33 10-5) exp 53.41292

system fitted EO-0-120K + propane EO-14-120K + propane EO-0-120K + propane EO-14-120K + propane

T range/C 70-180 70-180 70-180 70-180

T in the correlation is in K.

Figure 4. Parameter c as a function of weight percent octene-1 in EO.

Figure 6. P-T phase boundaries for EO-14-120K + propane at different EO weight fraction.

Figure 5. P-T phase boundaries for EB-5-100K, EH-5-100K, and EO-4-80K + propane at 0.05 polymer weight fraction. Experimental data for EB and EH are from Han et al.6 Table 7. Constants Used To Calculate c from eq 13 c0 SAFT SAFT1 0.660 74 0.744 15 c1 0.189 12 0.255 85 c2 7.375 84 10.241 23

Figure 7. P-X phase boundaries for EO-14-120K + propane. Symbols are experimental data: circles at 180 C and squares at 140 C. Lines are calculated from SAFT.

FL pressures are around 400 bar, whereas the FS temperatures are between around 40 and 50 C. At pressures above the FL pressure and at temperatures higher than the FS temperature, the solution is in the one-phase region. If pressure is decreased below the FL pressure, as long as the temperature is higher than 50 C, the one-fluid solution will split into two liquid phases. On the other hand, if the temperature is increased above the FS temperature, as long as the

pressure is higher than 400 bar, the liquid and solid phases will merge into a one-fluid solution. Figures 7 and 8 illustrate the effect of EO weight fraction on the FL pressure for EO-14-120K + propane. We find that the polymer concentration has a relatively weak effect on the FL pressure in the range of this study (0.01-0.1 polymer weight fraction). The critical polymer concentration for EO-14-120K + propane is experimentally found to be between 0.05 and 0.1 polymer weight fraction. SAFT and SAFT1 underestimate the critical polymer concentration. This is expected because both SAFT and SAFT1 crudely approximate EO as a single monodisperse pseudocomponent and hence do not

3074

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000

Figure 8. P-X phase boundaries for EO-14-120K + propane. Symbols are experimental data: circles at 180 C and squares at 140 C. Lines are calculated from SAFT1.

Figure 10. P-T phase boundaries for EO-0-10K and EO-0120K + propane at 0.05 EO weight fraction. Experimental data are from Condo et al.3

Figure 9. T-X phase boundaries for EO-14-120K + propane.

account for the actual differences in polydispersity. In general, the higher the polydispersity index, the greater the critical point shift toward the higher polymer concentrations;1 the critical point no longer coincides with the maximum of the cloud-point curve. Also it is worthwhile noticing in Figure 7 that the SAFT-calculated isotherms at 60 and 100 C exhibit two crossover points. One is in the dew-point range at around 340 bar and 0.005 polymer weight fraction (A) and the other is in the bubble-point range at around 315 bar and 0.15 polymer weight fraction (B). This means that when the polymer weight fraction is between 0.005 and 0.15, the FL pressures first decrease with decreasing temperature and then increase with decreasing temperature. This is characteristic of ULCST behavior. Our experimental data show the same behavior in Table 5. All of the cloud-point data, measured at 0.01, 0.049, and 0.098 polymer weight fraction, exhibit such a U-LCST behavior. Therefore, the crossover calculated by SAFT is, at least qualitatively, consistent with this behavior. SAFT1, however, does not capture this behavior, at least from 180 down to 60 C, as shown in Figure 8. Figure 9 shows the effect of EO weight fraction on the FS temperature for EO-14-120K + propane. In

Figure 11. P-T phase boundaries for EO-0-120K, EO-4-80K, EO-8-120K, and EO-14-120K + propane at 0.05 EO weight fraction.

general, as expected, the FS temperature increases with increasing EO concentration. Also, as usual, the heating-induced FS temperatures are systematically higher than those induced by cooling (by about 10 C). The difference is due to the subcooling effect of the crystal formation during cooling, as was explained earlier. Moreover, the FS temperature increases with increasing pressure. Both SAFT and SAFT1 are found to approximate the FS temperature dependence on the EO weight fraction and pressure but not to be very accurate in the low EO weight fraction range. Also, we cannot verify the high EO weight fraction predictions. Figure 10 shows the effect of EO molecular weight. As usual, increasing EO molecular weight increases the FL pressure. On the other hand, the EO molecular weight has a small effect on the FS temperature. When the EO molecular weight is increased by a factor of 12, from 10K to 120K gmol-1, the FS temperature is found to increase only by about 3 C. Again, SAFT and SAFT1 at least qualitatively account for this effect. Figure 11 shows the effect of branch density. In general, increasing the branch density decreases the FL pressure and FS temperature. The decrease in the FL

Ind. Eng. Chem. Res., Vol. 39, No. 8, 2000 3075

pressure is due to decreasing the polymer density and, therefore, decreasing the density difference between the polymer and solvent. The decrease in the FS temperature is due to decreasing the number of crystallizable units, u. Both SAFT and SAFT1 capture these trends nearly quantitatively, except for EO-8-120K, which is probably due to the branch density nonuniformity in this sample. Conclusions FL and FS phase-equilibria data are measured and correlated with two SAFT equations of state for EO solutions in sub- and supercritical propane. The polymer concentration is found to have a weak effect on both FL and FS transitions in the range of this study (0.01-0.1 polymer weight fraction). Increasing the molecular weight is found to increase the FL pressure and FS temperature, whereas increasing the branch density is found to decrease the FL pressure and FS temperature. Both SAFT and SAFT1 are found to capture these trends nearly quantitatively. Nomenclature
TREF ) temperature rising elution fractionation BR,i ) bond fraction of type R in chain i BD ) branch density EB ) poly(ethylene-co-butene-1) EH ) poly(ethylene-co-hexene-1) EO ) poly(ethylene-co-octene-1) FL ) fluid-liquid FS ) fluid-solid GPC ) gel permeation chromatography Hu ) enthalpy of melting per mole crystal unit at Tm (8.22 kJ/mol of crystal unit) Mn ) number-averaged molecular weight (gmol-1) Mw ) weight-averaged molecular weight (gmol-1) T ) temperature Tm ) equilibrium melting temperature of pure polymer ares ) dimensionless residual Helmholtz energy of chain molecules aseg ) segment term that accounts for the nonideality of the reference fluid of nonbonded chain segments (monomers) achain ) chain term that accounts for covalent bonding bb ) backbone br ) branch c ) auxiliary parameter defined in equation (12) kij ) binary interaction parameter m ) segment number mi ) number of all segments in chain i mR,i ) number of segments of type R in chain i nR,i ) number of bonds of type R in chain i p ) mole fraction of the crystal units u ) number of crystal units in a molecule uc ) number of crystal units in a molecule that are crystallized u/k ) segment energy (K) xpL ) polymer mole fraction in solution xR ) segment fraction of type R in the mixture pL ) fugacity coefficient of polymer in solution

p ) fugacity coefficient of pure liquid polymer R,i ) segment fraction of type R in chain i ) reduced range of the potential well voo ) segment molar volume (mLmol-1) v ) molar volume difference between pure liquid polymer and pure solid polymer (4.937 cm3/mol of crystal unit)

Literature Cited
(1) Folie, B.; Radosz, M. Phase Equlilibria in High-Pressure Polyethylene Technology. Ind. Eng. Chem. 1995, 34, 1501. (2) de Loos, T. W.; de Graaf, L. J.; de Swaan Arons, J. LiquidLiquid Phase Separation in Linear Low Density PolyethyleneSolvent Systems. Fluid Phase Equilib. 1996, 117, 40. (3) Condo, P. D.; Colman, E. J.; Ehrlich, P. Phase Equilibria of Linear Polyethylene with Supercritical Propane. Macromolecules 1992, 25, 750. (4) Ehrlich, P.; Kurpen J. J. Phase Equilibria of PolymerSolvent Systems at High Pressures Near Their Critical Loci: Polyethylene with n-Alkanes. J. Polym. Sci., Part A 1963, 1, 32173229. (5) Charlet, G.; Delmas, G. Thermodynamic Properties of Polyolefin Solutions at High Temperatures: 1. Lower Critical Solubility Temperatures of Polyethylene, Polypropylene and Ethylene-Propylene Copolymers in Hydrocarbon Solvents. Polymer 1981, 22, 1181-1189. (6) Han, S. J.; Lohse, D. J.; Radosz, M.; Sperling, L. H. Short Chain Branching Effect on the Cloud-Point Pressures of Ethylene Copolymers in Subcritical and Supercritical Propane. Macromolecules 1998, 31, 2533-2538. (7) Whaley, P. D.; Winter, H. H.; Ehrlich, P. Phase Equilibria of Polypropylene with Compressed Propane and Related Systems. 2. Fluid-Phase Equilibria of Polypropylene with Propane Containing Alcohols as Cosolvents and of Some Other Branched Polyolefins with Propane. Macromolecules 1997, 30, 4887-4890. (8) Chan, K. C.; Russo, P. S.; Radosz, M. Fluid-Liquid Equilibria in Poly(ethylene-co-hexene-1) + Propane: A Light-Scattering Probe of Cloud-Point Pressure and Critical Polymer Concentration. Fluid Phase Equilib. 2000, submitted. (9) Kohn, J. P.; Luks, K. D.; Liu, P. H. Three-Phase SolidLiquid-Vapor Equilibria of Binary n-Alkane Systems (Ethanen-Octane, Ethane-n-Decane, Ethane-n-Dodecane). J. Chem. Eng. Data 1976, 21, 360-362. (10) Huang, S. H.; Radosz, M. Equation of State for Small, Large, Polydisperse, and Associating Molecules. Ind. Eng. Chem. Res. 1990, 29, 2284-2294. (11) Banaszak, M.; Chen, C. K.; Radosz, M. Copolymer SAFT Equation of State. Thermodynamic Perturbation Theory Extended to Heterobonded Chains. Macromolecules 1996, 29, 6481-6486. (12) Adidharma, H.; Radosz, M. Prototype of an Engineering Equation of State for Heterosegmented Polymers. Ind. Eng. Chem. Res. 1998, 37, 4453-4462. (13) Pan, C.; Radosz, M. Modeling of Solid-Liquid Equilibria in Naphthalene, Normal-Alkane and Polyethylene Solutions. Fluid Phase Equilib. 1999, 155, 57-73. (14) vanKrevelen, D. W. Properties of Polymer; Elsevier: New York, 1990. (15) Encyclopedia of Polymer Science and Engineering, 2nd ed.; Wiley-Interscience: New York, 1988. (16) Richardson, M. J.; Flory, P. J.; Jackson, J. B. Crystallization and Melting of Copolymers of Polymethylene. Polymer 1963, 4, 221-236.

Received for review October 18, 1999 Revised manuscript received February 7, 2000 Accepted February 8, 2000 IE990761E