Crystallization Rates of Matched Fractions of MgCl2-Supported Ziegler-Natta and Metallocene Isotactic Poly(propylene)s. 2. Chain Microstructures from a Supercritical Fluid Fractionation

Crystallization Rates of Matched Fractions of MgCl2-Supported ZieglerNatta and Metallocene Isotactic Poly(Propylene)s. 1. The Role of ChainMicrostructure†

Rufina G. Alamo* and Jose A. Blanco

Department of Chemical Engineering, Florida Agricultural and Mechanical University andFlorida State University, Tallahassee, Florida 32310-6046

Pawan K. Agarwal

Exxon-Mobil Co., Baytown Polymers Center, Baytown, Texas 77522-5200

James C. Randall

2710 Ridge Road, Steamboat Springs, Colorado 80487

Received October 3, 2002; Revised Manuscript Received December 24, 2002

ABSTRACT: The microstructures of two poly(propylene)s with matched molar masses and overall defectconcentrations are inferred from the crystallization behavior of their narrow molar mass fractions. Onepoly(propylene) was produced with a MgCl2-supported Ziegler-Natta catalyst and the other with ametallocene catalyst. The fractions obtained from the metallocene isotactic poly(propylene) display a rangein molar masses but each has the same defect concentration indicating a uniform intermolecularconcentration of defects in the parent metallocene isotactic poly(propylene). These fractions provide directevidence of the “single site” character of the metallocene catalyst. The variations of crystallization rateswith molar mass reflect different chain diffusion/transport phenomena that are governed by the remnantentanglement state of the melt during crystallization. The molar mass fractions obtained from theZN-iPP confirm that the interchain distribution of the nonisotactic content is broad in this polymer. Thestereodefects are more concentrated in the low molar mass fractions. Furthermore, the invariance of thelinear growth rates among the ZN fractions and the lack of formation of any significant content of the γpolymorph, even in the most defected fraction, is consistent with a nonrandom, blocky intramoleculardistribution of defects in the ZN-iPP molecules. In contrast to the growth rates, the overall crystallizationrates are a direct function of the primary nucleation density, which varies among the fractions and theunfractionated iPPs. Hence, the measured overall crystallization rates would be correlated with nucleationdensity and not necessarily with the microstructure of the iPP molecules. The crystallization data arealso interpreted in light of results from pentad/heptad distributions predicted by two-state and three-state statistical models. Parameters from the models allow the prediction of sequence distribution curvesthat could be used to evaluate each of the models as to their consistency with the crystallization ratedata.

IntroductionIsotactic poly(propylene)s prepared with “classical”

heterogeneous Ziegler-Natta catalyst systems (ZN-iPP)are complex because a distribution of various types ofcatalyst sites leads to a mixture of poly(propylene)molecules having strongly differing intermolecular ste-reosequence distributions. In contrast, poly(propylene)sprepared with the single-sited metallocene family ofisotactic poly(propylene) catalysts (M-iPP) should haveuniform intermolecular stereosequence distributions.1,2

The identification of types, concentrations and distribu-tions of stereosequences found in various poly(propy-lene)s is customarily carried out by 13C NMR. Suchstructural information is important, not only as a wayto correlate with the physical properties of these ma-terials but also as a means to extract informationconcerning the nature of the catalyst sites and themechanisms of the polymerization processes.

The influence of catalyst type on the microstructureof various ZN-iPPs has often been reevaluated as higherfield NMR spectrometers with higher resolution and

sensitivity became available.3-10 In most of these previ-ous studies, 13C NMR analyses of structural defects haverelied upon analyses of either whole polymers,2-4 poly-(propylene)s where an atactic poly(propylene) compo-nent had been removed after crystallization from xy-lene5-9 or fractions from a parent ZN-iPP.11-14 Somelevel of fractionation is required in the microstructuralanalyses of ZN-iPP to ensure that the NMR character-ization is performed on crystalline chains only withoutinterfering contributions from atactic components. Itwas shown that valuable structural information can beobtained by studying different fractions from a parentpoly(propylene) and analyzing the stereosequence dis-tribution of each of the fractions utilizing 13C NMR.11-14

The crystallization behaviors of ZN-iPPs and M-iPPshave been reported in separate publications15-21 but adirect comparison of the physical properties and crystal-lization behavior of these two types of poly(propylene)s,matched by molar mass and defect concentrations, hasyet to be presented. It is the objective of this work tofractionate such a matched pair of ZN-iPP and M-iPPand to compare the structural and crystallization char-acteristics of the individual fractions with that observedfor the whole, parent polymers. This type of study isneeded to document the nature of inter- and intramo-

* Corresponding author. E-mail: [email protected].† Dedicated to Prof. L. Mandelkern on the occasion of his 80th

birthday.

1559Macromolecules 2003, 36, 1559-1571

10.1021/ma021549g CCC: $25.00 © 2003 American Chemical SocietyPublished on Web 02/14/2003

lecular microstructures that leads to differences incrystallization properties. A highly isotactic M-iPP anda corresponding ZN-iPP of the industrial type wereselected. From this study, it is possible to determine indetail how poly(propylene) microstructures, differing atinter- and intramolecular levels, can affect the crystal-lization behavior of isotactic poly(propylene)s.

The influence of the type of catalyst and polymeriza-tion process on microstructures of industrial ZN poly-(propylene)s has been studied by different investi-gators.11-14,22-23 TREF 14,22 and solvent gradient extrac-tion11-13 led to fractions of increasing isotacticity andincreasing molar mass, a result that clearly indicateda nonuniform interchain composition of defects. Theyalso observed that most of the fractions containedsyndiotactic rrrr pentads and, as in other works,7-13,24-26

it was proposed that some of the propagation errorswere “blocky” in nature. Thus, the presence of stereo-blocks in the ZN-iPP was inferred or suspected by mostprevious investigators and emphasized in Busico et al.’slatest work.25

Since it is important to identify the possibility of anonrandom intramolecular distribution or blocky ste-reosequences in iPP chains, other types of physicalevidence must be utilized in conjunction with 13C NMR.This is the only approach that will satisfy the ultimategoal of correlating molecular microstructures withphysical properties. For this purpose, indirect methodsmust be used to identify the intramolecular microstruc-ture of the iPP molecule. One example is a recentpublication in which a proposed blocky microstructureof highly defective iPPs was deduced by their crystal-lization properties.27

Different stereodefect distributions have been associ-ated with quite different crystallization rates, evenwithin a blocky microstructure. For example, in thecrystallization process, short, isolated randomly distrib-uted defected blocks will act as any other noncrystal-lizable single defect, retarding the crystallization rateand reducing the crystallinity and the melting temper-ature of the defected chain with respect to a defect freeiPP chain. Moreover, noncrystallizable long defectiveblocks will have little or no effect upon the crystalliza-tion properties of long isotactic blocks of the chain. Inthe present work, we document the nature of theintermolecular defect distribution from GPC and NMRanalyses of molecular fractions of ZN and metallocenetypes iPPs. Furthermore, indirect analyses are used toextract the intramolecular microstructure of the ZNpoly(propylene)s. Focus is given to the analysis of thecrystallization rates as a parameter that is a functionof the defect microstructure.

Previous stereochemical analyses of the sequences ofZN fractions obtained by successive extractions withn-alkanes solvents were interpreted on the basis of twoor three state models developed by Chujo,28 Doi,29 andBusico et al.25 to simulate a mixture of poly(propylene)chains produced by multisited Ziegler-Natta cata-lysts.4,5,30 It is important to examine both the averagemicrostructures and microstructural distributions ofZiegler-Natta poly(propylenes) in light of their crystal-lization behavior. Therefore, in a separate part of thiswork,31 13C NMR data will be analyzed utilizing all ofthe current as well as some newly modified two stateand three state statistical models. Each model wasevaluated on how closely the predicted microstructuraldistributions conformed to the observed rates of crystal-

lization.31 The various models were also evaluated onthe consistency of microstructural changes from fractionto fraction for the ZN-iPP where, from fractionationdata, the average stereodefect level was observed todecrease with increasing molar mass. The results of aslight modification to the Busico three-state model,25

utilizing first order Markovian statistics for the sym-metric chain component instead of Bernoullian statis-tics, are discussed in the present paper. These resultswere found to be consistent with both the linear growthrates and fractionation data in addition to predictingthe observed 13C NMR pentad/heptad distributions andaverage sequence lengths.

The possibility of a nonrandom distribution of defectsin the ZN-iPP will be discussed in the present paperfrom an analysis of the crystallization rates of thevarious molecular fractions. In addition, the crystal-lization rates of molecular fractions obtained from aM-iPP, will be compared to those from a ZN-iPP with asimilar overall defect content. The value of comparisonswith the M-iPP lies in the fact that we found the latterto have a random and uniform intra- and intermoleculardistributions of stereodefects.

Experimental SectionMaterials. The unfractionated poly(propylene)s were sup-

plied by ExxonMobil. The ZN-iPP was obtained with one ofthe latest MgCl2-supported industrial catalysts, and the met-allocene iPP was obtained with an Exxpol Kaminsky type ofcatalyst.32 The molecular characterization data, as per molarmass, molar mass distribution, and percentage of stereo- andregiodefects, are listed in Table 1. Only stereodefects werefound in the ZN-iPP sample while both stereo- and 2,1-erythro-regiodefects were found in the metallocene iPP. These poly-(propylene)s were chosen because they both comprise the same0.51 mol % overall concentration of all types of defects. Table1 also lists the peak melting temperatures of rapidly crystal-lized samples and the average isotactic sequence lengthcalculated as the average meso run length (MRL).10

Both poly(propylene)s were fractionated via supercriticalfluid extraction by the Phasex Co.33,34 The fractionation wascarried out isothermally by increasing pressure in supercriticaln-propane at 150 °C. Seven fractions were obtained from themetallocene iPP and five fractions from the ZN-iPP. Datarelevant to the molecular characterization of the fractions arealso listed in Table 1. Sample designations encode, in consecu-tive order, the following items: catalyst (“M” for metalloceneand “Z” for Ziegler-Natta), f for a fraction (no “f” for unfrac-tionated parents), weight-average molar mass in kg/mol fol-lowed by “K”, and the total number of defects per 100monomeric units. For example, Mf86K0.56 stands for a fractionobtained from the metallocene iPP with a weight-averagemolar mass of 86 000 g/mol and with a 0.56 mol % total defectconcentration.

Characterization Methods. The molar masses and theirdistributions were determined by standard gel permeationchromatography using polystyrenes as calibration standards.Melting temperatures and heats of fusion were obtained in adifferential scanning calorimeter (Perkin-Elmer DSC-7) using∼4 mg of sample and a heating rate of 10 °C/min. Statictemperature calibration of the instrument was carried out withindium.

The pentad/heptad sequence distributions and the concen-tration of 2,1 defects were obtained from solution 13C NMRspectra carried out at 125 MHz on 13C using 10 mm o.d. sampletubes. Only methyl resonances were employed in the deter-mination of pentad/heptad sequence distributions. This avoidsunwanted contributions from possible differences among nuclearOverhauser effects and relaxation times. Here 15% solutionsin tetrachloroethane-1,2-d2 were used at 125 °C. The pentad/heptad resonance chemical shifts were based on publishedassignments.7,8,35-38 The stereodefect concentration was taken

1560 Alamo et al. Macromolecules, Vol. 36, No. 5, 2003

as half the fraction of mmmr pentads because any stereodefectsequence will always have two, and only two, associated mmmr()rmmm) pentads located at the beginning and the end of eachisotactic sequence. Thus, it is possible to determine the totalstereodefect population independently of the structures of thevarious stereodefects.

The average meso run length (MRL),10,21 which representsan average over contiguous meso sequences starting at n ) 4for r(m)nr, was selected for the characterization of averagesequence length of the isotactic components. This defini-tion precludes contributions from short meso runs such as∼∼∼mmmrmrmmm∼∼∼ and ∼∼∼mmmrmmrmmm∼∼∼,which are most likely not crystallizable but are part of theoverall stereodefect distribution. MRL values are listed inTable 1.

X-ray diffraction patterns were recorded in reflection modeat room temperature using a Philips X’Pert PW3040 MRDdiffractometer operating at 45 mA and 40 kV. Ni-filtered CuKR radiation was used as the source.

Linear growth rates (G) were measured in an Olympuspolarized optical microscope used in conjunction with a Linkamhot stage, TP-93. The temperature was controlled with aprecision of (0.1 °C, and the precision of the eyepiece used tomeasure the spherulitic growth was 0.15 µm. Photographswere also taken during isothermal growth using an InstantPolaroid camera. At any temperature the linearity of the plotsof spherulite radius vs time and reproducibility of the mea-surements were highly satisfactory with regression coefficientshigher than 0.99. The measured growth rates were alsoindependent of the position of the spherulite in the specimenand the calculated uncertainty in the value of G was low((0.01 × 10-6 cm/s).

Overall crystallization rates were taken as the inverse ofthe time required for 50% of the transformation to take place(similar to the half time rate concept).39 The degree oftransformation, at a fixed isothermal temperature, was fol-lowed by the variation of the heat-flow vs time in the DSC-7.

Sample Preparation and Crystallization Procedures.Films approximately 100 µm thick were prepared by compres-sion molding the initial pellets or powders in a laboratoryCarver press preset at 200 °C and were used for the micro-scopic and DSC measurements. Some of the ZN fractions werecrystallized at temperatures >120 °C in thermostated oil bathsfor analysis of crystallographic polymorphs by WAXS. For thispurpose, plaques of 1 cm × 1 cm × 0.3 mm were molded,sandwiched between two metal plates covered with thin Alfoil, and placed in a vacuum sealed tube to prevent degrada-tion. The tube was immersed in a silicone oil bath at 210 °Cfor 15 min and then rapidly transferred to another oil bathpreset at the required crystallization temperature.

Prior to crystallization in the hot stage or in the DSC, thefilms were melted at 200 °C for 5 min and cooled at 40 °C/minto the isothermal crystallization temperature. To maximize

heat transfer, the DSC was operated in conjunction with anintracooler and under dry nitrogen flow.

Results and Discussion

Comparison of Unfractionated Matched ZN andMetallocene Poly(propylene)s. Before results ob-tained from the fractions are analyzed in detail, it isalso of interest to discuss the differences in melting andcrystallization behavior of the unfractionated poly-(propylene)s. Isothermal linear growth rates and DSCmeltings of the metallocene and ZN parents iPPs aregiven in Figures 1 and 2, respectively. At any crystal-lization temperature, the spherulites of the unfraction-ated ZN-iPP grow at about twice the rate of themetallocene iPP. The DSC melting peak of the ZN-iPPis also considerably broader, multipeaked and shiftedto a temperature about 8 °C higher than the meltingbehavior observed for the metallocene iPP crystallites.

Table 1. Characterization of Parent Polymers and Fractions from Metallocene and ZN-iPP

defects (mol %)

sample Mw (g/mol) Mw/Mn stereo regio total Tm (°C)a MRLb

M203K0.51 parent 203 900 2.00 0.11 0.40 0.51 155.0 194Mf86K0.56 86 000 1.43 0.16 0.40 0.56 156.2 176Mf121K0.50 121 000 1.32 0.11 0.39 0.50 156.3 199Mf143K0.54 143 000 1.24 0.07 0.47 0.54 156.1 184Mf200K0.46 200 000 1.23 0.08 0.38 0.46 155.8 216Mf235K0.45 235 000 1.24 0.07 0.38 0.45 154.6 221Mf358K0.41 358 000 1.34 0.07 0.34 0.41 154.0 243Mf383K0.41 383 000 1.47 nac na na naZ263K0.51parent 262 600 3.19 0.51 0.51 161.4 190Zf97K1.03 97 000 1.31 1.03 1.03 159.0 91Zf157Kxxx 157 000 1.27 na na naZf163K0.60 163 000 1.97 0.60 0.60 161.3 161Zf204K0.41 204 000 2.15 0.41 0.41 160.9 239Zf328K0.36 328 000 1.77 0.36 0.36 162.3 270

a Rapidly quenched samples to 25 °C. b Average meso run length,10,21 MRL ) mmmm/[0.5(mmmr + rmmm) +(2,1 defects)]. c Data notavailable.

Figure 1. Linear growth rates as a function of crystallizationtemperature for parents metallocene iPP (O) and ZN-iPP (b)with matched overall concentration of defects, 0.51 mol %.

Figure 2. Melting temperatures of parents ZN-iPP (top) andmetallocene iPP (bottom). Samples were crystallized at -10°C/min and melted at +10 °C/min.

Macromolecules, Vol. 36, No. 5, 2003 Role of Chain Microstructure 1561

Since the overall concentration of total defects in bothpoly(propylene)s is the same (0.51 mol %) and bothchains are of similar molar mass, the differences incrystallization and melting behavior must reflect asignificant difference in distributions of defects. Theissue of stereodefects only in the ZN-iPP and stereo- +regiodefects in the M-iPP will be addressed shortly.Previous fractionations of similar industrial-typeZN-iPPs have revealed a nonuniform concentration ofdefects from chain to chain in these iPPs.11-14,22-23

Racemic units were found to be more concentrated inthe lower molar mass chains. However, the metallocenecatalyst with its “single site” nature, leads to iPP chainsof narrow molar mass distribution and a narrow dis-tribution of defect concentrations among chains. Oneconsequence of a broad intermolecular distribution ofchain defects in the parent ZN-iPP is that it contains afraction of highly defected molecules that do not par-ticipate in the crystallization process. This behaviorleads to a lower effective overall defect concentrationin the ZN than in the M-iPP. The ZN-iPP containslonger isotactic sequences that will be selected earlierin the crystallization and lead to thicker crystallites, inline with the observed higher growth rates and meltingtemperatures.

We should also consider if the observed differencesin melting behavior between ZN-iPP and M-iPP mightbe a consequence of a different partitioning of regio vsstereodefects between the crystalline and noncrystallineregions. For example, if the stereodefects more easilyenter the crystalline regions than do the regio-2,1-defects, the ZN-iPP chains, with defects only of thestereo-type, will be more crystallizable in line with theresults of Figures 1 and 2. This issue has been previ-ously addressed by analyzing the solid state 13C NMRspectra of the crystalline regions of a series of M-iPPswith varied concentrations of stereo- and 2,1-regiode-fects.40,41 Both types of defects were found to enter thecrystalline lattice at levels that do not differ signifi-cantly. In addition, the similarity of the slopes thatcharacterize the variation of the growth rates withtemperature in Figure 1, are indicative of the formationof crystallites with very similar interfacial free energiesin both types of iPPs. Hence, we exclude any significantdifference in the way stereo- (predominantly a singleinversion in configuration shown by the presence ofmmrrmm) and regio-2,1-defects affect the crystallizationof the iPPs used in this investigation.

A confirmation of the predicted interchain micro-structure, i.e., homogeneous for the metallocene iPP andinhomogeneous for the Z-N type, will be obtained frommicrostructural analyses of molecular fractions fromboth iPPs. The nature of the ZN-iPP inhomogeneousdefect distribution as well as a possible nonrandomintrachain defect distribution will be discussed after adetailed analysis of the crystallization rates of thefractions presented in the next section.

Molecular Fractions from Metallocene iPP andZ-N iPP: Fractionation and Analysis of GrowthRates. The molecular characteristics of the fractionsobtained from both iPPs by isothermal pressure profileare listed in Table l. This technique allows solvent freerecovery of the fractions in adequate quantities for NMRand physical studies.33,34 With increasing pressure inthe fractionation, a systematic increase of the molarmass of the fractions was also observed indicating thatthe fractionation took place preferentially by molar

mass.33,34 The parent metallocene iPP comprises chainsof weight-average molar mass of 204 000 g/mol with themost probable molar mass distribution. From thispolymer, seven fractions were isolated with a range inmolar mass from 86 000 to 383 000 g/mol and withsimilar narrow molar mass distributions of 1.3 ( 0.1 inall these fractions. The data are listed in Table l.42 Ofrelevance are the very similar concentrations of defectsfound in all the metallocene fractions, as listed in thefourth and fifth columns of Table 1. A small variationin total defect concentration from 0.56 to ∼0.40 mol %is observed with increasing molar mass. Even so, all ofthe M-iPP fractions have total defect levels that are veryclose to the 0.51 mol % observed for the parent M-iPPmaterial. These results offer direct evidence that M-iPPchains have very small intermolecular variations amongthe concentration of defects and, thus, provide directexperimental basis to conclude that the interchaincomposition distribution is very narrow in the parentM-iPP. With such a narrow variation in the defectcompositional distribution, the metallocene fractionsoffer an ideal series for analyzing the effects of molarmass on iPP crystallization, independent of the effectsof defect content and defect distribution. This analysiswill be reported in a subsequent publication.43

Well formed spherulites were observed under isother-mal crystallization of all the metallocene fractions andthe spherulitic linear growth rates, measured in atemperature range between 134 and 155 °C, are givenin Figure 3A. Following the dictates of kinetics theoriesfor nucleation and crystallization of polymers,44,45 thegrowth rates of all the M-iPP fractions show a similarvariation with undercooling, i.e., a decrease of the rateby about 100 times with increasing crystallization

Figure 3. Linear growth rates of iPP molecular mass frac-tions as a function of crystallization temperature. (A) Fractionsfrom metallocene iPP. (B) Fractions from ZN-iPP. Designationof the fractions follows that given in Table 1.


temperature in the range analyzed. In addition, theeffect of molar mass is reflected by about a one-halfdecrease of the rate with increasing molar mass at anytemperature. For example, at a crystallization temper-ature of 140 °C growth rates values (in cm/s) of 1.68 ×10-6, 1.23 × 10-6, 1.20 × 10-6, 1.19 × 10-6, 0.98 × 10-6,and 0.89 × 10-6 were obtained for a molar mass changefrom 86 000 to 358 000 g/mol. This difference is notlarge, but it is, however, systematic and much largerthan the estimated experimental error in the measure-ment of the growth rate.

A small molar mass dependence of the growth rate isqualitatively similar to that reported for linear poly-ethylene and for other homopolymers when crystalliza-tion takes place at high undercoolings.45-52 For example,linear polyethylene displays a discrete maximum in thecrystallization rate with molar mass for undercoolingslower than 30°. However, at higher undercoolings, themaximum is not observed, and the change of the ratewith molar mass is less pronounced, becoming verysmall at the highest undercoolings. The crystallizationof the metallocene iPP fractions takes place at under-coolings well above 30°. Hence, the small molar massdependence of the growth rate, which is observed inFigure 3A, fits the pattern of other homopolymers andcopolymers.45-52 The linear growth rates of the metal-locene fractions follow the expected variation for chainsof different lengths with very similar contents of defectsand the same random intramolecular distributions ofdefects.

The natural logarithm of the experimental lineargrowth rates (G) of the ZN-iPP fractions are given inFigure 3B for a range of crystallization temperaturesbetween 135 and 155 °C. The x and y axes of this plotare identical to the axes in the plot of growth rates ofthe metallocene fractions (Figure 3A). This consistencyfacilitates a comparative study. In analyzing the dataof Figure 3, parts A and B, we notice smaller but stillsignificant differences in the growth rates of theZN-iPP fractions, when compared to those of the met-allocenes. No systematic variation of G with eithermolar mass or concentration of defects is found in theZN-iPP fractions. In principle, this is an unusual resultbecause a strong effect of the concentration of defects

in lowering G was reported for metallocene iPPs of afixed molar mass.19 The same variation with defects wasalso found in other M-iPP fractions for which the detailsof the fractionation were not provided.16 Therefore, inthe range of defect concentrations of the ZN-iPP frac-tions, one would have expected a significantly lower Gfor the fraction with 1.03 mol % defects than for thefraction with 0.36 mol % defects. The data of Figure 3Bshow that this is not the case; in fact, there are onlyminor differences between the growth rates of theZN-iPP fractions at a fixed crystallization temperature.The differences in crystallization behavior, when com-pared to the metallocene fractions, likely reflect signifi-cant differences in the distributions of defects.

The molar mass and concentration of defects of thefractions obtained from the ZN-iPP, also listed in Table1, confirm the inhomogeneous interchain concentrationof defects of the parent polymer. As the molar mass ofthe fractions increases from 97 000 to 328 000 g/mol,the total defect content decreases from 1.03 to 0.36 mol% respectively. This inverse relation between chainlength and concentration of defects was also observedin ZN-iPP fractions obtained by a solvent-nonsolventextraction fractionation procedure11,12 and in fractionsobtained by TREF.14,22,23 Hence, quite different typesof fractionation studies reveal an interchain variationof the concentration of defects in the ZN parent thatdecreases as the molar mass increases. Neither thefractionation nor characterization data reveal any de-tails about the intramolecular distribution of defects.The molecular characteristics of a defect distribution areprobably best addressed by indirect studies of propertiesthat relate to an intramolecular defect distribution. Inthis context, different possible scenarios are viable andschematics for two of these models for the crystallinechains are given in Figure 4. A third composite modelthat results from applying a three-state statistical modelto the observed NMR pentad/heptads stereosequenceswill be discussed in detail in the second part of thisseries.31 The results of this model are summarized inthe last section of the present paper.

In Figure 4A, a defect distribution is shown that isnonuniform on an intermolecular basis, but the in-tramolecular defect distribution is random. According

Figure 4. Two possible microstructural models of stereosequence distribution in the parent ZN-iPP. (A) The concentration ofdefects is nonuniform between the molecules and randomly distributed intramolecularly. Increasing concentration of defects iscorrelated with molar mass. Different lengths of crystallizable sequences are found. (B) The concentration of defects is alsononuniform between the molecules but the intramolecular distribution of defects is blocky. Long similar crystallizable sequences,highlighted for visual effect, are found in all the chains.


to this model, the defect concentrations increase con-tinuously with decreasing molar mass. The crystalliza-tion behavior of fractions from this model is expectedto be proportional to the average defect composition inthe melt which changes among the fractions. Therefore,after fractionation, much lower crystallization rates areexpected for the fractions of lower molar mass because,on average, they comprise a higher concentration ofdefects and shorter isotactic sequences than the lessdefected fractions.16,17,19,40 The second possibility (de-picted schematically in Figure 4B) is a microstructuralmodel in which the intermolecular defect compositionis also nonuniform with the shorter chains having ahigher concentration of nonisotactic units. However, inthis model the intramolecular distribution of defects isblocky (nonrandom) with long runs of crystallizableisotactic sequences being present in all chains. Onaverage, the second model has similar types of longcrystallizable sequences present in all of the chains. Itis clear that the crystallization behavior of fractionsfrom this second model will be quite different from thebehavior of fractions in the first, which has a randomintramolecular defect distribution. In the second model,the overall concentration of defects also changes fromchain to chain but the crystallization is led by the longisotactic blocks. The concentration of short, less crys-tallizable isotactic sequences is minimal compared tothe concentration of these sequences in the first model.The crystallization behavior of the fractions from thesecond iPP model should be similar to the crystallizationof a diblock copolymer in which one of the blocks isnoncrystallizable.

If the actual ZN defect distribution follows the mi-crostructural model given in Figure 4A, one wouldexpect, after fractionation, that a similar crystallizationrate should be observed for both the metallocene andthe narrow ZN fraction with matching defect concentra-tions. After fractionation, the overall defect concentra-tion will be much closer to the single molecule concen-tration in both iPPs. A difference in crystallizationbehavior, however, is maintained as seen in Figure 3,parts A and B, where the growth rates of fractionsMf358K0.41 and Zf204K0.41 are compared. The spher-ulites of the Ziegler fraction grow more rapidly thanthose of the metallocene counterpart. This reaffirms theidea that, even after fractionation, the Ziegler fractionscontain longer isotactic sequences; put in another way,the broad defect distribution of the parent ZN-iPP issustained to some extent in the fractions. Moreover, thefact that the growth rates of all the Ziegler fractionsare so similar also indicates that the intramoleculardistribution of the defects in the ZN fractions is differentfrom that of the metallocene fractions. The similarityof the rates in Figure 3B suggests that the samenonrandom intramolecular distribution is propagatedalong all the molecules of the parent ZN-iPP. In otherwords, the defect distribution of the ZN-iPP does notconform with the microstructure modeled in the schemeof Figure 4A.

The ZN fraction with the highest concentration ofdefects is also the fraction of the lowest molar mass.Thus, one could argue that the expected decrease of Gwith increasing defect content in the ZN fractions maybe compensated by an opposite effect of molar mass onG, as demonstrated in the metallocene series. If thiswere the case, model 4A could still apply to describe thedefect distribution of the ZN-iPP. To probe this pos-

sibility, the growth rate data of the lowest molar massZN fraction with 1.03 mol % defects were corrected tothe values that this fraction would have if its molarmass is ∼200 000 g/mol. To make this correction, dataobtained for the metallocene fractions, shown in Figure3A, are appropriate because they purely reflect thevariation of G with molar mass independently from theeffect on the rate of the defect concentration. Thus, thegrowth rate data of the fraction Zf97K1.03 were loweredby the difference in G given by the metallocene frac-tions in a range between 97 000 and 204 000 g/moltaken from Figure 3A. The corrected G data for the 1.03mol % fraction (closed diamonds) and the uncorrectedone for fraction Zf204K0.41 of low defect content (opencircles) are plotted in Figure 5. The data are alsocompared with published growth rate data of a metal-locene iPP with 1.0 mol % of defects,19 given by the solidline in this figure. It is evident that the correction formolar mass has very little effect on the variation of Gamong the ZN fractions. The growth rates of the ZNfractions fall on the same line, which is positioned muchhigher than that expected had the defects been ran-domly distributed as is the case for the data of themetallocene iPP with the same 1.0 mol % defects ofFigure 5.

The similarity of the growth rates of all the ZNfractions, even after correction for different molarmasses, provide evidence that suggests that the in-tramolecular defect microstructure of the parentZN-iPP, as well as the fractions, could be blocky. Evenafter careful fractionation, the ZN-iPP growth rates areconsiderably higher than observed for the metallocenefractions. These observations offer additional evidencefor the possible blocky nature of the ZN-iPP. As will beshown below in this work, the lack of formation of theγ polymorph, which requires short isotactic sequences,also rules against model 4A. Hence, the results of thegrowth rates favor the microstructural model shown inFigure 4B.

Additional evidence, which supports the stereoblockdefect microstructure of the ZN-iPP chains, is found inthe WAXS diffractograms obtained in the two mostdefected ZN fractions. We have previously reported thata metallocene iPP with a concentration of 1.0 mol % ofhomogeneously distributed defects, crystallized at 125°C, develops about half of the crystallites in the γpolymorph and the other half in the R monoclinic form.21

It was also observed in our previous work that thepercentage of γ phase developed by two unfractionated

Figure 5. Linear growth rates of ZN fraction Zf97K1.03corrected for the effect of molar mass on the rate ([) comparedwith growth rates of ZN fraction Zf204K0.41 (O). The continu-ous line indicates the variation of growth rates with temper-ature of a metallocene-type iPP with 1.00 mol % of defectsextracted from ref 19.


ZN-iPPs was, at any temperature, below the level ofdetection of this polymorph by WAXS. The low γ contentwas attributed to the broad defect distribution of theZN-iPPs.21 The WAXS diffractograms of the two mostdefected ZN fractions crystallized at 125 °C are shownin Figure 6A for Zf97K1.03 and in Figure 6B forZf163K0.60. The intensity of the reflection at a 2θ of20°, characteristic of the γ polymorph, is low andcorresponds to less than 10% in both fractions. The peakat 2θ ) 16° is associated with the â polymorph, moreabundant in the lower molar mass fraction. In Figure7, the percentage of the γ phase obtained for theZf97K1.03 fraction is compared to the γ form developedwith increasing temperature by a metallocene iPP withmatched concentration of defects. The content of γ phasedeveloped by the ZN fraction is four times lower than

the value of the metallocene iPP with its narrow defectdistribution. It is basically constant in the range oftemperatures studied. Hence, we find that at a matchedoverall concentration of defects of 1.0 mol %, thestructural requirements for the formation of the γpolymorph are present in the metallocene iPP, but theyare basically absent in the ZN fraction. These require-ments were described in detail in our previous work21

and are summarized in the following two major points:1. Chains whose microstructure leads to thin crystallitesmust be present; i.e., there must be short crystallizableisotactic sequences available. 2. There must be a struc-ture that prevents folding in the crystal amorphousinterfacial region and requires tilted ordered chains topropagate a lamellar crystallite. The lamellar propaga-tion is favored by a tilted antiparallel molecular ar-rangement of the γ crystallographic phase.53,54 As seenin Figures 6 and 7, even after crystallization in atemperature range that is most favorable for the forma-tion of the γ phase,21 the content of γ phase obtained inthe ZN fractions is much lower than that expected fora chain with a random defect distribution. The mostcharacteristic reflection of the γ form, that at a 2θ of20°, is almost absent from both diffractograms. Thisresult is expected for a molecule where long isotacticsequences join blocks of poorly stereoregular sequences,such as those of the ZN stereoblocky type describedschematically in Figure 4B.

From the relation between the maximum content ofγ form and the concentration of defects in narrowlydistributed metallocene iPPs, given in Figures 8 and 9of ref 21, we find that 12% of the γ polymorph corre-sponds to an average meso run length (MRL) of 220repeat units. This value is considerably larger than theaverage MRL of 91 units for the ZN-iPP fraction with1.03 mol % defects and suggests that contiguous iso-tactic sequences of at least 220 units must be presentin the Zf97K1.0 fraction in a relatively high concentra-tion. As presented in next section, a modified Busicothree-state model leads to a distribution of sequencelengths for this fraction consistent with these data. Longisotactic runs favor the formation of folded-chain crys-tallites in the ZN fractions and, hence, they crystallizepreferentially in the R polymorph. In a recent publica-tion,27 the analysis of the polymorphic behavior was alsotaken as an indirect measure of the degree to which thedefects in the iPP chain deviate from a random distri-bution.

Overall Crystallization Rates. Previous parallel orindependent studies of the linear growth rates and theoverall crystallization rates measured by dilatometry orby differential scanning calorimetry of linear polyeth-ylene fractions and other types of homopolymers, haveresulted in an identical temperature coefficient.44,46,55,56

In addition, these methods of measurement give identi-cal variation of the rates with molar mass or withconcentration of defects. Hence, data from either growthrates or overall rates have been used to test forconformity with kinetic theories of polymer nucleationand growth.44,46,51,52 We were also interested in testingif the overall crystallization rates measured by DSCfollow the behavior observed by the growth rates of theparent iPPs and their fractions. Furthermore, theoverall crystallization rates apply more directly to thesolidification rates used in industrial processes becausethe data that characterize these rates are mass orvolume-based quantities, and the measurements are not

Figure 6. WAXS diffractograms of ZN-iPP fractions Zf97K1.03and Zf163K0.60 collected at room temperature after 3 days ofcrystallization at 125 °C.

Figure 7. Variation of the concentration of the γ polymorphwith crystallization temperature for different iPPs with matched1.00 mol % overall defect concentration: ([) metallocene typeiPP (data from ref 21); (O) ZN-iPP fraction Zf97K1.03.


limited to having to observe a specific morphologicalfeature as in microscopic methods.

The natural log of the inverse of the time required toobtain 50% of the total transformation (τ0.5), taken atthe peak of the exotherms, is plotted vs crystallizationtemperature in Figure 8A for the metallocene fractions,and in Figure 8B for the ZN fractions. Different featuresare apparent from these plots. At a fixed crystallizationtemperature, the variation of the overall rate among themetallocene fractions is not systematic with molar massas was observed with the growth rates. The variationof the overall rates of the ZN fractions is also differentfrom the behavior observed for the growth rates inFigure 3B. At any temperature, the overall crystalliza-tion rate of the less defected ZN fraction (Zf328K0.36)is about three times faster than the rate of any of theZN fractions. The difference in overall rates betweenthe more defected ZN fractions, although small, isdirectly related to their concentration of defects. Forexample, it is the fraction with the highest concentrationof defects (of the lowest molar mass) that shows thelowest overall rate.

These results can be explained by variations amongthe nucleation densities found in the fractions andparent iPPs. The number of spherulites that develop ineach specimen were measured at selected crystallizationtemperatures. The data of the parent metallocene iPPand its fractions will be reported first. Figure 9 showsmicrographs of the parent metallocene and selectedfractions crystallized at 134 °C for approximately thesame lengths of time. The spherulites are not uniformlydistributed in most of the fractions and their rapidappearance and high numbers are indicative of anucleation enhanced by some external agent. As seen

in Figure 10, parts A and B, for five representativetemperatures, the number of nuclei in a fixed volumefollow the molar mass with the same random patternas the overall crystallization rate. The latter is governedby the rate of nucleation, as demonstrated in thecrystallization of many polymeric systems.44,45,55,57,58

Consequently, the fractions with a higher nucleationdensity will yield a fixed level of crystallinity faster thanfractions in which fewer nuclei are formed, regardlessof the magnitude of the spherulitic linear growth rate.The solid lines in these figures follow the experimentaldata and are only intended as a guide to the eye. Theydo not represent any theoretical functionality.

Nucleants were not added to the parent metalloceneor ZN polymer, inferring that what causes enhancednucleation must be acquired either in the polymeriza-tion or the fractionation process. About 0.1% of theantioxidant Irganox 1010 was added to the n-propanebefore fractionation. However, as seen in Figure 10A,the parent metallocene iPP, free of this antioxidant,shows as high a nucleation density as the highestobserved value among the fractions. From this, weconclude that the antioxidant, which could have beenrandomly distributed during fractionation is not nucle-ating the metallocene fractions. Since significant amountsof MAO are used as cocatalysts during polymerization,this agent or residues from the metallocene catalyst mayenhance the nucleation rate of the metallocene iPP andits fractions.

Significant differences in the nucleation density of theZN-iPP fractions are also observed. Representativemicrographs of the lowest and highest molar mass ZNfractions are given in Figure 11. The nucleation densityof the highest molar mass fraction is over 1 order ofmagnitude higher than any other ZN fraction. Nucle-ation densities and overall crystallization rates for theZN fractions are shown in Figure 12, parts A and B,respectively, as a function of molar mass, for represen-tative crystallization temperatures. Following the nucle-ation density pattern, the overall crystallization ratesshow a small increase with molar mass up to 200 000g/mol and a large increase for the higher molar massfraction. For example, at a crystallization temperatureof 123 °C, the rate changed from 0.34 to 0.46 min-1 fordefects changing from 1.03 to 0.41 mol % and to 1.16min-1 for the less defected, highest molar mass fraction.Thus, as seen in Figure 12B, the variation of thelogarithm of the rate is not linear with increasing molarmass or decreasing concentration of defects. The highestmolar mass, low defected fraction shows a significantlyhigher rate. We ruled out that catalyst residues maybe causing the high rates of the ZN fraction. Largecontents of cocatalyst residues are not an issue in thepolymerization with ZN catalysts. Moreover, statisticalmodels27 predict a fraction of very long, highly isotacticmolecules in ZN-iPP, which naturally will be concen-trated in the higher molar mass fraction. It is reasonablethat these very long, highly isotactic molecules (absentin the metallocene iPP) may persist in the melt asaggregates and lead to a significant enhancement of thenucleation and crystallization rates. Melt memory ef-fects have been observed previously in the crystalliza-tion of iPP59 and were also tested in the ZN fractionZf328K0.36. A small decrease of the overall crystalliza-tion rate at 124 °C was found by increasing the meltresidence time at 200 °C from 5 to 30 min, but highermelt temperatures, such as 250 °C, could not be tested

Figure 8. Overall crystallization rates of iPP molecularfractions as a function of crystallization temperature. (A)Fractions from metallocene iPP. (B) Fractions from ZN-iPP.Designation of the fractions follows that given in Table 1.


because the sample degraded. The concentration of verylong, highly isotactic molecules is more diluted in theparent ZN-iPP and the lower molar mass fractions inagreement with their observed lower nucleation rates.Thus, the overall crystallization rate of the parentZN-iPP and all the fractions, except the one with thehigher molar mass, are very similar, in line with thesimilar growth rates observed in these fractions and inagreement with their blocky defect microstructure.

From the crystallization behavior of the two mostgeneral types of iPPs, we can postulate two differenttypes of factors acting as “precursors” of the overallcrystallization, i.e., catalyst residues and the very longchains present in the ZN-iPPs. These factors affectprimary nucleation and, therefore, the overall crystal-lization rates characteristic of each fraction. Thus, it isnot surprising that a correlation between the overall

crystallization and microstructural variables is notfound in the ZN-iPPs of this study. On the other hand,the linear growth rates which are independent of theprimary nucleation rate and led by a process of second-ary nucleation,44 offer a better tool to correlate with thestructural variables of the poly(propylene) chain. Theseresults indicate that caution must be exercised whenusing overall crystallization rates to correlate physicalproperties with molecular microstructure. Any othermass or volume based measurement of the crystalliza-tion of iPP, such as optical or X-ray scattering may besubjected to the same type of enhanced nucleationdescribed in this work.

Experimental NMR Stereosequences. ModifiedBusico Three-State Statistical Model. The defectedpentads and heptads observed by 13C NMR for theZN-iPP fractions and the parent ZN-iPP are listed in

Figure 9. Polarized optical micrographs of the parent metallocene iPP and selected fractions isothermally crystallized at 134 °Cfor Mf86K0.56 (7 min), Mf143K0.54 (7 min), Mf200K0.46 (9 min), Mf235K0.45 (7 min), and M203K0.51 (7 min).


Table 2. In agreement with previous studies,7-11,22-26

we notice a nonnegligible concentration of purely syn-diotactic sequences (rrrrrr) in all the ZN fractions,including the fraction with the highest molar mass. Inaddition, stereoerrors of the type, rmrrrm or rmrrrr arealso not detected.25 The presence of the mrrrmm heptadindicates that isotactic blocks of opposite handednessare present. The most abundant defect consists of asingle opposite configuration, ∼0001000∼, and is de-tected by the mmrrmm heptad. There are also moredefective (mmrrmr) defects connecting isotactic blocksin all the ZN fractions as indicated by the normalizedintegrals corresponding to the steric pentads and hep-tads listed in Table 2.

The experimental pentad/heptads stereosequenceswere fit with two and three state statistical models. Thepredicted sequence length distributions will be discussedin detail for conformity to the observed fractionation andcrystallization rates in the subsequent paper.31 Thevalue in fitting an observed pentad/heptad distributionwith any statistical model is that the results can be usedto identify the molecular component that gives rise to

the various types of stereodefects. None of the modelsled to a defect distribution resembling the schematicmicrostructural model represented in Figure 4B, whichmost closely adheres to the experimental data. Thisblocky distribution of defects is difficult to propose withthe classical two or three state statistical modelsbecause conceptually it would require switches betweenenantiomorphic (Es) and chain end control (CE) sites.The results obtained with the modified Busico three-state model are of relevance; therefore, they are alsopresented here. In the original work by Busico et al.25

the ZN-iPP chains are modeled as a mixture of threetypes of molecules using a three-state statistical model.For a late generation, but weakly isotactic, Ziegler-Natta poly(propylene), the result was a mixture of veryhigh isotactic molecules produced by an Es state, poorlyisotactic or atactic molecules produced by a CE stateand isotactic molecules produced by switching betweentwo Es control sites (C1), each leading to differentisotacticity levels.

As mentioned earlier, the only change made in Busi-co’s original three-state model was to employ first-orderMarkovian statistics for the symmetric chain (termedCE in Busico’s work, now referred to as CE1) componentas opposed to the Bernoullian statistics used in theoriginal Busico three-state model. The enantiomorphicEs and switching C1 states were used exactly asrepresented by Busico and co-workers.25 This modifiedmodel presents the unfractionated ZN-iPP as a com-posite of chains of different lengths in which moleculesof the same length can have different concentrations ofrandomly distributed defects in two components (fromEs and C1 chains) and nonrandom defects in the third,CE1 chain component. The composite leads to a resultwhere the defect concentration will decrease withincreasing molar mass in agreement with the fraction-ation data. The most significant result obtained fromfitting the modified Busico three-state model to theobserved ZN-iPP pentad/heptad fraction data is that thethree states are found to have similar compositions andsimilar sequence distribution curves in all four ZN-iPPfractions. It is only the relative amounts of eachcomponent that is predicted to change from fraction tofraction with increasing molar mass. The presence ofvery long isotactic sequences, predicted in the nonran-dom CE1 component, is also important. This result maylead to similar growth rates as observed for the ZN-iPPfractions. During slow crystallization, highly isotacticEs molecules and long isotactic sequences from the CE1component will be selected first and at about the samerate for all the fractions. This is inferred from the

Figure 10. A. Variation of the nucleation density withmolecular mass for fractions (b) and unfractionated metal-locene iPP (O) crystallized at 134 °C for ∼7 min. B. Overallcrystallization rate as a function of molecular mass formetallocene iPP fractions crystallized at the indicated isother-mal temperatures.

Figure 11. Polarized optical micrographs of fractions Zf97K1.03 and Zf328K0.36 isothermally crystallized at 134 °C (3 min)


>0.994 isotactic sequence probability of the Es compo-nent predicted for all the fractions.31 The concentrationof the less isotactic CE1 component was found toincrease in the fractions of lower molar mass. However,the effect of this dilution on the crystallization of theEs component is predicted to be small and possiblycompensated by differences in molar masses betweenthe fractions. Therefore, the predicted crystallizationbehavior conforms to the experimental observations. Inaddition, most of the crystallizable sequence lengthspredicted by the ES/C1/CE1 model are over 200 isotacticunits in agreement with the small contents of the γpolymorph that are observed. Homogeneously distrib-uted metallocene iPPs with average meso run lengthsof ∼200 units and higher, led to contents of the γpolymorph of less than 20%.21 Thus, it is expected thata nonrandom, “stereoblocky”, defect distribution of themost populated CE1 type molecules of the lowest molarmass fractions (Zf328K0.36 and Zf163K0.60), wouldcrystallize with an even smaller content of this poly-morph, in agreement with the experimental observations.

A similar pentad/heptad analysis of the stereose-quences in the parent M-iPP and its fractions revealedonly one principal stereodefect and that was mmrrmm,which is the defect that consists of a single oppositeconfiguration. At the same level of NMR sensitivity usedfor the ZN-iPP fractions, the rmmr, mmrm + rrmr, andrmrm pentads and the various rr centered heptads,other than mmrrmm, were below the experimentalNMR noise level in the metallocene iPPs. This result ispredicted by a single state, enantiomorphic site control(Es) model10 and is consistent with the behavior of asingle sited catalyst. In addition to mmmm, the onlyother significant methyl resonances observed are mmmr+ rmmm, mmrr + rrmm, and mmrrmm, which areexpected in a 2:2:1 ratio, as observed for the M-iPPparent and its fractions. This stereodefect, which arisesfrom a single configurational error, accounts for virtu-ally all of the stereodefects in M-iPP and comprisesabout 20% of the total defect distribution. The stereo-defect distribution is, therefore, Bernoullian1 and con-sistent with the linear growth rate data obtained fromthe metallocene series, which does not support anysignificant grouping of regio- or stereodefects. In fact,iPP prepared with “single sited” metallocene catalystshave been shown in independent studies to displayuniformly distributed stereo- and regiodefects.1,60

Concluding RemarksThe crystallization behavior of a metallocene and a

ZN-iPP with the same molar mass and same overallconcentration of defects was studied in relation to theirdefect microstructure. Conclusions with respect to inter-and intramolecular distribution of the defects in eachtype of iPP were possible from the analysis of thecrystallization data of their molecular fractions, whichwere obtained by isothermal pressure profile in super-critical n-propane.

The fractions from the metallocene iPP provide direct,supporting evidence of the “single sited” nature of thecatalyst. They display a range in molar mass but thesame defect concentrations, which indicate the presenceof uniform intermolecular defect concentrations in theparent metallocene iPP. The experimental linear growthrates of the fractions were those expected for chains witha random intramolecular distribution of defects. Thevariations of these rates with molar mass reflect theirdifferent number of entanglements per chain in themelt.

The overall crystallization rates of the parent metal-locene iPP and its fractions are affected by an enhanced

Figure 12. A. Variation of the nucleation density withmolecular mass for fractions (b) and unfractionated ZN-iPP(O) crystallized at 134 °C for 3 min. B. Overall crystallizationrate as a function of molecular mass for ZN-iPP fractionscrystallized at the indicated isothermal temperatures.

Table 2. Stereosequence Distribution of Unfractionated ZN-iPP, and Its Fractions

distribution

sequenceassignment,ppm (TMS)

Z263K0.51parent Zf97K1.03 Zf163K0.60 Zf204K0.41 Zf328K0.36

mmmm 21.8 0.9632 0.9357 0.9683 0.9766 0.9802rmmm 21.55 0.0102 0.0205 0.0120 0.0082 0.0073rmmr 21.33 0.0029 0.0016 0.0016 0.0010 0.0008mmrr 21.02 0.0113 0.0200 0.0099 0.0073 0.0057mmrm + rrmr 20.79 0.0034 0.0036 0.0013 0.0014 0.0011rmrm 20.55 0.0009 0.0008 0.0001 0.0005 0.0004mrrrrm 20.33 0.0006 0.0013 0.0003 0.0003 0.0003mrrrrr 20.3 0.0007 0.0016 0.0003 0.0002 0.0004rrrrrr 20.25 0.0014 0.0036 0.0007 0.0008 0.0012rmrrrm 20.17 0.0000 0.0000 0.0000 0.0000 0.0000rrrrmr 20.13 0.0012 0.0016 0.0006 0.0005 0.0004mmrrrr + mmrrrm 20.06 0.0009 0.0023 0.0007 0.0005 0.0003rmrrmr 20 0.0000 0.0000 0.0000 0.0000 0.0000mmrrmr 19.93 0.0008 0.0014 0.0006 0.0004 0.0003mmrrmm 19.85 0.0035 0.0062 0.0034 0.0026 0.0019


nucleation, probably caused by catalysts or cocatalystresidues. Thus, the rates follow the variation of nucle-ation density in each fraction, and a correlation betweenthese measurements and the defect microstructure ofthe metallocene iPP, or its fractions, is not possible.

The molar fractions obtained from the ZN-iPP con-firmed that the defect composition is broadly distributedon an intermolecular basis. Thus, while the molar massof the fractions varied from 97 000 to 328 000 g/mol, theconcentration of defects decreased from 1.03 to 0.36 mol%. In addition, the basically identical growth ratesobtained in all the ZN fractions and the lack of forma-tion of any significant content of the γ polymorph, evenin the most defected fraction, is consistent with a blockyintramolecular nature of the defects in the ZN-iPPmolecules. The data are also consistent with a micro-structural model in which most of the crystalline chainscomprise long runs of basically defect free isotactic unitsfollowed by runs of other less isotactic or atacticsequences.

Very similar overall crystallization rates were alsofound for most of the ZN fractions, which suggests ablocky microstructure. However, the highest molar massfraction displayed much higher nucleation density and,thus, higher overall crystallization rates than the restof the fractions. This could be attributed to the presenceof aggregates in the melt from the higher concentrationof molecules with very high molar mass in this fraction.These long, highly isotactic molecules are not found inthe narrowly distributed metallocene iPP.

As a consequence of the enhanced nucleation causedeither by foreign particles (possible catalyst residues)in the metallocene iPP, or by possible aggregates in themelt induced by the long chains of the ZN-iPP, acorrelation between the overall crystallization and theirmicrostructural variables may not be found. The lineargrowth rates, which do not depend on primary nucle-ation, were adequately correlated with molar mass infractions from the metallocene iPP, with uniform dis-tribution. The growth rates also served as a useful toolto infer an intramolecular, nonrandom distribution ofdefects in the ZN-iPP chain.

A multistate statistical model of the ZN-iPP micro-structure predicts in all molar mass fractions the samethree types of molecules, almost defect free isotacticchains, mainly isotactic molecules with non-Bernoulliandefect distribution, and poorly isotactic molecules. Thedefect concentration and sequence distribution of thethree types of molecules is the same in all ZN-iPPfractions, differing only in their relative amounts. Thismodel was found to be consistent with the fractionationresults and growth rates of the ZN fractions.

Acknowledgment. This work was supported by theNational Science Foundation Polymer Program (DMR-0094485). The authors acknowledge useful commentson this work by Prof. L. Mandelkern. Eric Ritchson, aREU chemical engineering student, and Dongsheng Liare also acknowledged for helping with the DSC experi-ments and collecting the WAXS diffractograms. We arealso grateful to Charles Ruff of the ExxonMobil Co. forNMR data. J.A.B. acknowledges partial support fromthe Fundacion Repsol YPF.

References and Notes

(1) Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.;Waymouth, R. M. Angew. Chem., Int. Ed. Engl. 1995, 34,1143.

(2) Fischer, D.; Mulhaupt, R. Macromol. Chem. Phys. 1994, 195,1433.

(3) Zambelli, A.; Locatelli, P.; Bajo, G.; Bovey, F. A. Macromol-ecules 1975, 8, 687.

(4) Zambelli, A.; Locatelli, P.; Provasoli, A.; Ferro, D. R. Macro-molecules 1980, 13, 267.

(5) Kawamura, H.; Hayashi, T.; Inoue, Y.; Chujo, R. Macromol-ecules 1989, 22, 2181.

(6) Busico, V.; Corradini, P.; De Martino, L.; Graziano, F.;Iadicicco, A. M. Makromal. Chem. 1991, 192, 49.

(7) Busico, V.; Corradini, P.; De Biasio, R.; Landriani, L.; Segre,A. I. Macromolecules 1994, 27, 4521.

(8) Busico, V.; Cipullo, R.; Corradini, P.; Landriani, L.; Vacatello,M. Macromolecules 1995, 28, 1887.

(9) Busico, V.; Cipullo, R.; Talarico, G.; Segre, A. L.; Chadwick,J. C. Macromolecules 1997, 30, 4786.

(10) Randall, J. C. Macromolecules 1997, 30, 803.(11) Paukkeri, R.; Vaananen, T.; Lehtinen, A. Polymer 1993, 34,

2488.(12) Paukkeri, R.; Lehtinen, A. Polymer 1994, 35, 1673.(13) Lehtinen, A.; Paukkeri, R. Macromol. Chem. Phys. 1994, 195,

1539.(14) Morini, G.; Albizzati, E.; Balbontin, G.; Mingizzi, I.; Sacchi,

M. C.; Forlini, F.; Tritto, I. Macromolecules 1996, 29, 5770.(15) Cheng, S. Z. D.; Janimak, J. J.; Zhang, A.; Hsieh, E. T.

Polymer 1991, 32, 648.(16) Janimak, J. J.; Cheng, S. Z. D.; Giusti, P. A.; Hsieh, E. T.

Macromolecules 1991, 24, 2253.(17) Janimak, J. J.; Cheng, S. Z. D.; Zhang, A.; Hsieh, E. T.

Polymer 1992, 33, 728.(18) Monasse, B.; Haudin, J. M. Colloid Polym. Sci. 1985, 263,

822.(19) Alamo, R. G.; Chi, C. Crystallization Behavior and Properties

of Polyolefins. In Molecular Interactions and Time-SpaceOrganization in Macromoleculer Systems; Morishima, Y.,Norisuye, T., Tashiro, K., Eds.; Springer: Berlin, 1999; p 29.

(20) Alamo, R. G.; Galante, M. J.; Mandelkern, L.; Lehtinen, A.;Paukkeri, R. J. Therm. Anal. 1996, 47, 913.

(21) Alamo, R.; Kim, M.-H; Galante, M. J.; Isasi, J. R.; Mandelk-ern, L. Macromolecules 1999, 32, 4050.

(22) Viville, P.; Daoust, D.; Jonas, A. M.; Nysten., B.; Legras, R.;Dupire, M.; Michel, J.; Debras, G. Polymer 2001, 42, 1953.

(23) Xu, J.; Feng, L.; Yang, S.; Yang, Y.; Kong, X. Eur. Polym. J.1998, 34, 431.

(24) Busico, V.; Cipullo, R.; Corradini, P.; De Biasio, R. Macromol.Chem. Phys. 1995, 196, 491.

(25) Busico, V.; Cipullo, R.; Monaco, G.; Talarico, G.; Vacatello,M.; Chadwick, J. C.; Segre, A. L.; Sudmeijer, O. Macromol-ecules 1999, 32, 4173.

(26) Chadwick, J. C.; Miedema, A.; Ruiseh, B. J.; Sudmeijer, O.Makromol. Chem. 1992 193, 1463.

(27) De Rosa, C.; Auriemma, F.; Circelli, T.; Waymouth, R. M.Macromolecules 2002, 35, 3622.

(28) Chujo, R. Kagaku 1981, 36, 420.(29) Doi, Y. Makromol. Chem., Rapid Commun. 1982, 3, 635.(30) Hayashi, T.; Inoue, Y.; Chujo, R.; Asakura, T. Polymer 1988,

29, 138.(31) Randall, R. C.; Alamo, R. G.; Agarwal, P. K.; Ruff, C. J.

Macromolecules 2003, 36, 1572.(32) Kaminsky, W.; Hahnsen, H.; Kulper, K.; Woldt, R. US Patent

4,542,199, 1985.(33) Watkins, J. J.; Krukonis, V. J.; Condo, P. D., Jr.; Pradhan,

D.; Ehrlich, P. J. Supercrit. Fluids 1991, 4, 24-31.(34) Britto, L. J. D.; Soares, J. B. P.; Peulidis, A.; Krokonis, V. J.

Polym. Sci., Polym. Phys. Ed. 1999 37, 553.(35) Gassi, A.; Zambelli, A.; Resconi, L.; Albizzati, E.; Mazzocchi,

R. Macromolecules 1988, 21, 617.(36) Busico, V.; Cipullo, R.; Monaco, G.; Vacatello, M.; Segre, A.

L. Macromolecules 1997, 30, 6251.(37) Soga, K.; Shiono, T. Makromol. Chem., Rapid Commun. 1987,

8, 305.(38) Busico, V.; Cipullo, R.; Monaco, G.; Vacatello, M.; Bella, J.;

Segre, A. L. Macromolecules 1998, 31, 8713.(39) This definition of the rate of crystallization (1/ι0.5 ) d(Xc)/dt)

assumes constant volume during the transformation and aconstant final degree of crystallinity, Xc, at any crystallizationtemperature.

(40) VanderHart, D. L.; Alamo, R. G.; Nyden, M. R.; Kim, M. H.;Mandelkern, L. Macromolecules 2000, 22, 6078.

(41) Nyden, M. R.; VanderHart, D. L.; Alamo, R. G. J. Comput.Theor. Polym. Sci. 2002, 11, 175.

(42) The last fraction, obtained at the highest pressure was notavailable for 13CNMR analysis.


(43) Alamo, R. G.; Blanco, J. A.; Agarwal, P. Manuscript inpreparation.

(44) Hoffman, J. D.; Frolen, L. J.; Ross, G. S.; Lauritzen, J. I. J.Res. Natl. Bus. Standars 1975, 79A, 671.

(45) Ergoz, E.; Fatou, J. G.; Mandelkern, L. Macromolecules 1972,5, 147.

(46) Fatou, J. G.; Marco, C.; Mandelkern, L. Polymer 1990, 31,1685.

(47) Magill, J. H.; Li, H. M. Polymer 1978, 19, 416.(48) Hoffman, J. D. Polymer 1982, 23, 656.(49) Lopez, L. C.; Wilkes, G. L. Polymer 1988, 29, 106.(50) Alamo, R. G.; Fatou, J. G.; Guzman, J. Polymer 1982, 23,

379.(51) Alamo, R. G.; Mandelkern, L. Macromolecules 1991, 24, 6480.(52) Chen, H.-L.; Li, L.-J.; Ou-Yang, W.-C.; Hwang, J. C.; Wong,

W.-Y. Macromolecules 1997, 30, 1718.

(53) Bruckner, S.; Meille, S. U.; Petraccone, U.; Pirozzi, B. Prog.Polym. Sci. 1991, 16, 361.

(54) Lotz, B.; Graff, S.; Straupe, S.; Wittmann, J. C. Polymer 1991,32, 2902.

(55) Fatou, J. G. Encyclopedia of Polymer Science and Engineer-ing, 2nd ed.; John Wiley & Sons: New York, 1989; Suppl.Vol., p 231.

(56) Allen, R. C.; Mandelkern, L. Polym. Bull. (Berlin)1987, 17,473.

(57) Alamo, R.; Fatou, J. G.; Guzman, J. Polymer 1982, 23, 374.(58) Wagner, J.; Phillips, P. J. Polymer 2001, 42, 8999.(59) Ziabicki, A.; Alfonso, G. C. Colloid Polym. Sci. 1995, 273, 317.(60) Kaminsky, W. Metal Organic Catalyst for Synthesis and

Polymerization; Springer: Berlin, 1999; p 601.

MA021549G


Crystallization Rates of Matched Fractions of MgCl2-Supported Ziegler-Natta and Metallocene Isotactic Poly(propylene)s. 2. Chain Microstructures from a Supercritical Fluid Fractionation

Documents