A temperature rising elution fractionation study of short chain branching behavior in ultra low density polyethylene.
Ultralow-density polyethylenes (ULDPE) also known as very low-density polyethylenes (VLDPE), are an important and emerging class of linear polyethylenes. They have shown great promise because of superior properties such as flexibility, environmental stress-crack resistance, cold impact, and toughness (1). Important commercial applications include packaging and films. They are differentiated from linear low-density polyethylenes (LLDPE) mainly through density, which in the case of LLDPE's, is [greater than or equal to]0.915 g/cc, while ULDPE's possess densities below 0.915 g/cc. ULDPE's are manufactured by copolymerizing ethylene with alpha-olefins such as butene, hexene, octene, and 4-methylepentene or a combination of these comonomers, utilizing transition metal catalysts (e.g., Ziegler-Natta) while using either gas-phase or solution processes.
ULDPE is characterized by having little or no long chain branches, but the comonomers employed create a significant number of short chain branches (SCB) on the main backbone chain. The number and type of SCB's influence the morphology and the solid state properties of ULDPE. The existence of short chain branches makes ULDPE amenable to analysis by temperature rising elution fractionation (TREF). In this present work TREF was employed in both the analytical and preparative modes to determine SCB distribution and carry out fractionation of the polymer into large enough fractions for a series of further tests. TREF fractions were subsequently characterized by 13C-nuclear magnetic resonance (NMR), differential scanning calorimetry (DSC), dynamic mechanical thermal analysis (DMTA), density gradient, gel permeation chromatography (GPC), analytical TREF, and simple tensile tests.
The resin analyzed in this work is a commercially available (Dow Attane 4001) ultralow-density polyethylene (ULDPE), which is a copolymer of ethylene and 1-octene. Density was measured with a Tecam T.M. density gradient column using water and isopropanol and was found to be 0.912 g/[cm.sup.3]. Additionally, the resin had a nominal melt index [approximately equal to] 1, with Mw [approximately equal to] 120,000, and Mw/Mn [approximately equal to] 6.32. All film samples were molded between (DuPont) Teflon-coated aluminum sheets using a hot press at 200 [degrees] C, and applying a pressure of 10,000 psi for 1 min. Subsequently, samples were quenched into ice water. This procedure produced samples that are approximately 0.1 to 0.3 mm thick.
Preparative Temperature Rising Elution Fractionation (TREF)
Temperature rising elution fractionation (TREF) is the method of choice for fractionation of "crystallizable" polymers. In polyethylenes, particularly ULDPE, crystallizability is dependent on short chain branching distribution and/or copolymer composition. In a typical TREF experiment the polymer is first dissolved in a hot solvent (140 [degrees] C) such as 1,2,4-trichlorobenzene (TCB) at a concentration of 0.5%. An antioxidant such as BHT at a concentration of 0.1%, is added to the solution to prevent oxidative degradation of the polymer. A hot inert support such as Chromasorb P was then added to the hot polymer solution. The resulting support/polymer/solvent mixture is subjected to programmed crystallization at a cooling rate of 1.5 [degrees] C/h from 135 to 30 [degrees] C. Fractionation of polymer chains as a function of short chain branching distribution occurs during this crystallization cycle. Chains with a low number of branches are deposited first on the support, followed by chains which have increasing amounts of short chain branches.
The crystallized support/polymer/solvent system is loaded onto a stainless-steel column (3/4 by 2 inches) in a programmable oven. There elution is carried out using TCB pumped at 4 ml/min while a concurrent temperature rise of 25 [degrees] C/h is applied. The concentration of eluting polymer is detected with a MIRAN 1A CVF infrared detector set at 3.41 [[micro]meter] (the C-H stretch frequency). IR and temperature data are also simultaneously saved on a computer for future evaluation. Eluting polymer solution was collected as six fractions for further analysis. Fraction codes and the corresponding temperature ranges over which they were collected are: U2 = 30 to 60 [degrees] C, U3 = 60 to 70 [degrees] C, U4 = 70 to 75 [degrees] C, U5 = 75 to 85 [degrees] C, U6 = 85 to 100 [degrees] C, and U7 = 100 to 105 [degrees] C. Analytical amounts of these fractions were "re-TREFed." Additional information on the operation of TREF systems is well described in a number of publications (2-6).
ULDPE and its fractions were evaluated for their thermal behavior with a Perkin-Elmer System 7 Differential Scanning Calorimeter (DSC). ULDPE whole polymer was run "as received," while fractions were filtered from solution and vacuum dried at room temperature for three days then run "as prepared." Samples that were 5 mg in weight were subjected to a heating rate of 10 [degrees] C/min over a 0 to 150 [degrees] C temperature range.
Dynamic Mechanical Analysis
An investigation of the dynamic mechanical behavior of whole polymer and its fractions was undertaken using a Polymer Labs dynamic mechanical thermal analyzer (DMTA). Thin rectangular films having a length of 15 mm and a thickness of 0.1 to 0.3 mm were used. These were subjected to a tensile force of 0.01 N at 3 Hz while the sample was concurrently given a temperature ramp of 2 [degrees] C/min from -140 to 120 [degrees] C.
Mechanical behavior was studied with a computer controlled Instron 4201 tensile testing unit. The standard procedure employed for this work was ASTM method D 1708 (7). A nominal crosshead speed of 50 mm/min was applied to dogbone specimens. Toughness, for the purposes of this work, is defined as the area under the stress-strain curve.
Nuclear Magnetic Resonance (NMR) Spectroscopy
13C nuclear magnetic resonance (NMR) was performed to determine compositional variation using proposed ASTM method XX70-8605-2 (8). NMR spectra were obtained on a Bruker 3000; samples were run in a 10 mm tube using 1,2-dichlorobenzene and xylene-d as the lock solvent with [approximately equal to]1 gm of polymer. Spectra were obtained with 90 [degrees] pulse width and a pulse repetition time of 10 s, at 130 [degrees] C.
Gel Permeation Chromatography
Gel permeation chromatography was employed to determine molecular weight and molecular weight distribution of the fractions and whole polymer. GPC was performed at 130 [degrees] C on a Waters 150 C chromatograph.
ULDPE whole copolymer examined with analytical TREF yielded an essentially bimodal distribution of short chain branches as illustrated in Fig. 1. The low temperature peak is broad, covering the range from 30 to 100 [degrees] C, and is centered at 70 to 75 [degrees] C. As part of this broad peak there is a second characteristic peak centered at 37 [degrees] C. In contrast, the high temperature peak centered at [approximately equal to] 101 [degrees] C is very sharp and only extends over a five degree range. The wide temperature range of elution is a clear indication of a heterogeneous system. In analogy with LLDPE materials the low temperature peak is often referred to as "branched," while the high temperature peak is ascribed to "linear" polymer chains. The existence of a bimodal distribution of short chain branching in alphaolefins has been widely reported and has been attributed to the existence of two types of active sites on the catalyst.
Figure I is a qualitative illustration of the comonomer distribution in ULDPE and is basically the weight fraction of eluting polymer as a function of temperature. Therefore to arrive at a quantitative analysis it is necessary to utilize a calibration curve as suggested by Wild. This curve which is unique for each resin, is generated by collecting narrow fractions of the polymer in question and then determining each fractions comonomer content (SCB/1000 C) with NMR. Branching determined with NMR is the sum of short chain branches and chain ends; this branching number can be corrected for chain end effects with a knowledge of the molecular weight (8). Figure 2 illustrates the calibration curve (methyl concentration vs. elution temperature) for the ethylene/octene copolymer.
Preparative TREF of the copolymer was carried out, and the polymer was split into six different fractions. Figure 3 shows analytical TREF's of the preparative fractions, demonstrating that these preparative fractions are relatively homogeneous in character with respect to their short chain branching distribution. While re-TREFed fractionated polymers eluted over a relatively narrow temperature range, compared to the whole polymer a small overlap between fractions does occur in each case.
In order to clarify the relationship between molecular weight and short chain branching distribution, ULDPE and its six TREF fractions were cross-fractionated with respect to molecular weight. Superimposed molecular weight distributions of the these TREF fractions are presented in Fig. 4A. It is evident that increasing comonomer level is also consistent with a decrease of molecular weight in these materials. Similar behavior has been reported in studies of LLDPE's (9,10). The relationship between weight average and number average molecular weight and short chain branching is further clarified in Fig. 4B. Here it is observed that a significant transition occurs in molecular weight (both number and weight) between the "linear" and "branched" fractions of the polymer. Additionally, the polydispersity of the fractions as listed in Table 1 decreases, but not linearly, with a decrease in branching content.
Analysis of the densities of the TREF generated fractions revealed that they varied from [approximately equal to]0.902 to 0.942 g/[cm.sup.3], a spread of [approximately equal to]0.04 g/[cm.sup.3]. The lowest density is of the most branched fraction U2, while "linear" high temperature fraction U7 had the highest density. The trend of increasing density with decreased branching reveals differences in crystallizability and crystallinity of the fractions.
Comonomer levels determined by the 13C-NMR method of De Pooter, et al. (9) were another indicator of the heterogeneous character of the polymer. Whole polymer had a comonomer content of 3.63 mole % octene, while fractions varied from a low of 0.7% for the high density fraction to 6.09% octene for the lowest density fraction, or a 9 fold variation. Figure 5 presents a graphical illustration of density variation with comonomer content. Changes in melting point determined by differential scanning calorimetry as a function of comonomer level, are shown in Fig. 6. Melting points ranged from 90 to 132 [degrees] C. Figure 7B clearly illustrates the unimodal nature of the homogeneous fractions. This is in contrast to Fig. 7A, the thermal behavior of ULDPE whole polymer which exhibits a heterogeneous character. There is a clear difference of [approximately equal to] 10 [degrees] C between the peak maximum of whole polymer and the highest melting fraction U7.
Table 1. Molecular Weight and Short Chain Branching.
SCB/ (*)SCB/ Sample Mw Mn Mw/Mn 1000 C 1000 C
ULDPE 120,000 19,000 6.32 18.1 17.6 U2 73,100 14,400 5.08 30.5 29.9 U3 102,000 22,700 4.49 21.2 20.8 U4 118,000 27,600 4.28 17.9 17.6 U5 130,000 30,400 4.54 13.2 12.9 U6 161,000 34,900 4.61 6.3 6.0 U7 252,000 75,300 3.35 3.5 3.4
* Corrected # of short chain branches = SCB - (6 x 1400)/Mn corrected for chain ends.
An important goal of this work was to carry out mechanical testing on the ULDPE whole polymer and its fractions. Included among the measured mechanical properties are; Young's modulus, toughness, max strain, and maximum load; results are listed in Table 2. Overall, whole polymer exhibited properties such as toughness, break stress, and max strain values which were greater than any of the fractions. An exception to this trend is Young's modulus, which while lower for the whole polymer and the more branched fractions (U2-U5) was noticeably greater for the more linear fractions (U6 and U7). Another noticeable difference exists between branched fractions (U2 through U6) and the linear fraction U7. Fraction U7 is characterized by having a lower elongation to break, and a corresponding low toughness value; it generally behaves as a "brittle" material compared with U2 through U6.
Additionally, dynamic mechanical behavior and in particular, the [Alpha] and [Beta] transitions were examined. As seen in Fig. 8 the [Alpha] transition temperature increases with decreasing comonomer content. The [Alpha]-relaxation is a transition whose position depends on lamellar thickness of the polymer. Corroborating evidence is provided by DSC data shown in Fig. 7B where fraction U7 has a higher melting temperature implying a larger stem length.
The [Beta]-relaxation has been reported by Popli et al. (12) to be due to interfacial material and hence the [TABULAR DATA FOR TABLE 2 OMITTED] existence of a crystalline phase is necessary for the existence of a [Beta] transition. Others (14,15) have attributed it to branch points and in particular, the intensity of the transition was related to branch content. However, more recent studies have questioned these assignments and have supported the concept that the [Beta]-relaxation is the glass transition temperature for this polymer type (13). The T-transition is a low temperature relaxation process found at [approximately equal to] -120 [degrees] C. It was observed in all fractions and the whole ULDPE polymer.
[TABULAR DATA FOR TABLE 3 OMITTED]
13C-Nuclear magnetic resonance was performed on ULDPE and its fractions. The spectrum of the whole polymer is shown in Fig. 9 together with resonance assignments. These follow the commonly accepted terminology for an ethylene/octene copolymer shown below:
The polymer is characterized by having hexyl branches (32.2 ppm). In addition to the previously determined comonomer content, average sequence lengths together with the number of SCB/1000 C were calculated and are shown in Table 3. As anticipated, while comonomer content decreased the average sequence lengths of ethylene increased. These compared favorably to those calculated from Bernoullian statistics.
The ethylene/octene copolymer investigated in this paper was characterized with a number of techniques. From this series of experiments a number of conclusions can be drawn.
1) Fractionation with respect to short chain branching revealed a heterogeneous polymer having a bimodal SCB distribution and an elution range of approximately 75 [degrees] C. Individual fractions were characteristically unimodal in character and eluted over a narrow temperature range. The branching calibration curve was a straight line with a negative slope.
2) Fractionation by TREF followed by molecular weight analysis disclosed a noticeable decrease in molecular weight with an increase in the number of short chain branches. In addition, individual fractions had a lower molecular weight distribution compared to the whole polymer.
3) DSC analysis of ULDPE and its fractions provided additional evidence regarding the heterogeneous character of the resin. Interestingly, fraction U7 had a significantly higher melting point than the whole polymer.
4) Mechanical measurements such as toughness provided additional insight into the synergistic behavior of fractions visa vis whole polymer. Accordingly ductile behavior was noted for the "branched" (U2-U6) fractions in contrast to the brittle nature of the "linear" U7 sample.
5) Density which is an accurate indicator of structural irregularities in a polymer was noted to decrease with increasing short chain branching and comonomer content.
6) The 13C NMR experiments showed unequivocally the exclusive existence of hexyl type branches in the whole polymer and its fractions. Average sequence lengths of ethylene were noted to vary 15 fold from the highly branched to the linear fraction (U7).
The authors wish to thank the donors of the Petroleum Research Fund administered by the American Chemical Society for support of this work. Dr. W. Knight of Dow Chemical-Texas kindly performed the GPC measurements. L. Wild of Quantum Chemical, S. L. Peterson of PSU and H2K of State College gave us invaluable assistance in the setup of our original TREF apparatus.
1. L. C. Rundloff, in Modern Plastics Encyclopedia, p. 53, McGraw-Hill, New York (1990).
2. L. Wild, T. R. Ryle, D. C. Knobeloch, and I. R. Peat, J. Polym. Sci. Polym. Phys. Ed., 20, 441 (1982).
3. L. Wild, T. R. Ryle, and D. C. Knobeloch, Am. Chem. Soc., Div. Polym. Chem. Polym. Prepr., 23, 133 (1982).
4. D. C. Knobeloch and L. Wild, presentation at Polyolefins IV Conference, p. 427 Society of Plastics Engineers (February 27, 1984).
5. L. Wild and T. Ryle, Am. Chem. Soc., Div. Polym. Chem. Polym. Prepr., 18, 182 (1977).
6. F. M. Mirabella, International GPC Symposium '87, p. 180, Itasca, Ill. (1987).
7. Annual Book of Standards, 8.02, D1708-84, (1991).
8. M. De Pooter, P. B. Smith, K. K. Dohrer, K. F. Bennett, M. D. Meadows, C. G. Smith, H. P. Schouwenaars, and R. A. Geerards, J. Appl. Polym. Sci., 42, 399 (1991).
9. P. Schouterden, G. Groenickx, and B. Van der Heijden, F. Jansen, Polymer, 2099 11987).
10. F. M. Mirabella and E. A. Ford, J. Polym. Sci. Polym. Phys. Ed., 25, 777 (1987).
11. T. Usami, Y. Gotoh, and S. Takayama, Macromolecules, 19, 2722 (1986).
12. R. Popli, M. Glotin, and L. Mandelkern, J. Polym. Sci., Polym. Phys. Ed., 22, 40 (1984).
13. H. Sha, X. Zhang, and I. R. Harrison, Thermochemica Acta, 192, 233 (1991).
14. W. G. Oakes and D. W. Robinson, J. Polym. Sci., 14, 505 (1954).
15. D. E. Kline, J. A. Sauer, and A. E. Woodward, J. Polym. Sci., 22, 455 (1956).
16. L. D. Cady, Plast. Eng., January 1987, p. 25.
|Printer friendly Cite/link Email Feedback|
|Author:||Karoglanian, Serop A.; Harrison, Ian R.|
|Publication:||Polymer Engineering and Science|
|Date:||Mar 15, 1996|
|Previous Article:||Competition between textural transitions and pressure effects on the viscosity of thermotropic liquid-crystal polymers.|
|Next Article:||A review of positron annihilation lifetime spectroscopy as applied to the physical aging of polymers.|