Barrier properties of injection molded blends of liquid crystalline polyesters (Vectra) and high-density polyethylene.
Thermotropic liquid crystalline polymers (LCPs) are extremely good barriers against permanent gases and larger penetrant molecules. This is the reason for the emerging interest in these materials in packaging technology. Remarkable features of these polymers are their low shear viscosity in the nematic molten state and their high molecular alignment in the solid state after thermoplastic processing (1). The oxygen permeability of liquid crystalline copolyesters such as Vectra is in the same range as that of dry poly(ethylene-co-vinyl alcohol) and the water vapor transmission rate is lower than that of polyethylene (2-5). The low oxygen permeability of Vectra is due to the dense packing of the rigid polymer chains which leads to a very low oxygen solubility (2-4). The excellent barrier properties and the potential of these LCPs to withstand sterilization have resulted in a growing interest in their use in protecting pharmaceuticals (5), food (6), and beverages (7, 8). These quality properties in combination wi th reasonably low dielectric permittivity. a high dielectric strength and low moisture absorption make these materials also suitable for protecting electronic components (9-12).
Polymer blends based on LCPs and conventional thermoplastics may be a possible way to achieve more balanced mechanical properties in the end product and also a way to reduce the material costs. Continuity and lamella shape are the preferred features of the LCP phase in the blends in order to achieve optimum barrier properties. James et al. (13) reported that the oxygen permeability decreased strongly with increasing LCP content in polyethersulfone because of the formation of a thin, almost continuous surface layer of the LCP. Wiberg et al. (14) observed that 5% Vectra in a Vectra/polyethersulfone blend was sufficient to create an almost continuous Vectra phase, and that this gave a very low methanol diffusivity. Incarnato et al. (15) reported improved barrier properties and improved fracture toughness in blends based on maleic-anhydride grafted polypropylene (compatibilizer), polypropylene and LCP.
This paper is a continuation of previously reported studies (16, 17) on the barrier properties of compression-molded and film-blown blends of polyethylene and two Vectra LCPs, and of poly(ethylene terephthalate) and two Vectra LCPs. The earlier papers (16, 17) reported that the barrier properties were improved with the incorporation of the LCPs, but the positive effects were counteracted by the presence of voids in the polymer blends produced by the compression molding and film blowing. It was believed that the high hydrostatic pressure acting on the melt during the final stage of mold filling (injection molding) would reduce the void content and/or reduce the size of the voids. The data obtained by oxygen permeability measurements are related to the blend morphology, particular attention being given to the size and shape of the LCP 'particles.'
One series of blends was based on Vectra A950 (poly(p-hydroxy benzoic acid-co-2-hydroxy-6-naphtoic acid with 73 mol% p-hydroxy benzoic acid; Ticona AG, Germany) and polyethylene. Data for Vectra A950: melting point [approximately equal to] 280[degrees]C; glass transition temperature = 113[degrees]C; density (23[degrees]C) = 1400 kg/[m.sup.3]. The polyethylene (PE) was an extrusion-coating grade CG 8410 (Borealis AB, Sweden) with [M.sub.w] = 85,000 g/mol, density (23[degrees]C) = 941 kg/[m.sup.3], melting peak temperature = 130[degrees]C, and melt flow index = 7.5 g/10 min (according to ISO 1133). A second LCP, Vectra RD501 (Ticona AG, Germany] was also blended with PE. Characteristics of Vectra RD501: melting point = 220[degrees]C; density (23[degrees]C) = 1400 kg/[m.sup.3]. The compatibilizer used was poly(ethylene-stat-methacrylic acid) with 12 mol% methacrylic acid], Nucrel 1202HC (DuPont Co.). Characteristics of Nucrel 1202HC: melting point = 99[degrees]C, melt flow index = 1.5 g/l0 mm (ASTM D-3418) and d ensity = 940 kg/[m.sup.3].
2.2. Blend Preparation
The LCPs were dried at 120[degrees]C during 48 h prior to melt blending. Pelletized blends of LCP, PE and Nucrel were produced in a Brabender Twin Screw Compounder DSK 35/9 (Extruder type: [+ or -] 35/9D) mounted to a Plasti-Corder drive unit and an interface PL2000. The die used processed four strands during each run. The following temperature gradients were used: 305-285[degrees]C (Vectra A950/PE) and 240-250[degrees]C (Vectra RD501/PE) The die temperatures used were 210[degrees]C (Vectra A950/PE) and 260[degrees]C (Vectra RD501/PE). The screw speeds used were 20 rpm (Vectra A950/PE) and 20-25 rpm (Vectra RD501/PE). A Brabender pelletizing unit was used for the strands produced, but for the blends containing 27 and 47 vol% LCP a Moretto mill MLl8/10 pelletizer equipment was used because of the high stiffness of the strands. Blends of LOP contents between 0 and 47 vol% were prepared. The content of Nucrel added to each blend was 10% by weight of LCP.
2.3. Injection Molding
The injection molding machine used was a Battenfeld BA500/125 CDK UNILOG TC140 scl equipped with a three-zone-screw (outer diameter = 25 mm; L/D ratio = 21.6; compression ratio = 1.8). The average shear rates in the screw channel were between 87 and 157 [s.sup.-1] (18), the maximum shear rate was in the metering section. The mold consisted of two stainless steel blocks. The diameter of the circular cavity in the movable mold part was 135 mm and the thickness was 1 mm. The gate with a diameter of 4.5 mm was located in the center of the circular cavity. To keep the melt in the sprue hot, it was necessary to provide the die with a hot channel attachment (Z101/38 X 75/2) that could be regulated up to 300[degrees]C. The following process parameter settings were used: temperature gradient: [T.sub.1] = 280[degrees]C, [T.sub.2] = 290[degrees]C, [T.sub.3] 290[degrees]C; nozzle temperature = 270-290[degrees]C, hot channel attachment temperature = 290[degrees]C (Vectra A950/PE); [T.sub.1] = 200[degrees]C, [T.sub.2] = 22 0[degrees]C, [T.sub.3] = 220[degrees]C; nozzle temperature = 230[degrees]C, hot channel attachment temperature = 240[degrees]C (PE/ Vectra RD501). Injection pressure 180 MPa during 0.4 s, holding pressure 140 MPa during 0.5 s and injection rate 80 c[m.sup.3]/s. Back pressure: 0 Pa, except during molding of the blends with 47 vol% LCP: 30 MPa. Screw speed 300 rpm. The mold temperature was 90[degrees]C. The clamping force was set to 0.5 MN, which was increased to 0.75 MN during injection. The cooling period was 25-30 s (Vectra A950/PE) and 18 s (Vectra RD501/PE). The shear rate during mold filling was calculated according to Rosato et al. (18) and the average shear rate for pure Vectra A950 was 4000 [s.sup.-1]. Similar shear rate values are expected for the blends.
2.4. Scanning Electron Microscopy
Freeze-polished samples were examined in a JEOL JSM-5400 scanning electron microscope after Au-Pd coating in a Desk II sputter (Denton Vacuum). The freeze polishing was carried out using a glass knife in a PMC MT-XL ultramicrotome equipped with cryochamber CR-X (PMC Microtomy and Cryobiology Products). The glass knives were made in a LKB Knife Maker Type 7801B, LKB-Produkter AB, Sweden.
2.5. Solvent Extraction
Hot p-xylene was used to remove PE from the blends to reveal the dimensions and shapes of the LCP particles. Samples (diameter = 10 mm) were punched at a distance of half the radius from the center of the circular discs and treated with hot p-xylene in three steps: 105[degrees]C followed by 120[degrees]C and 130[degrees]C. It was possible to release more layers by performing this etching procedure in three steps. This resulted in separated layers free from PE, revealed by IR The extraction time at each temperature was one week. The thin layers obtained after etching were dried in oven at 105[degrees]C for 1 h. A very sharp and thin knife was used to help to release the layers from each other in the swollen samples if they were not already separated. The mechanical force applied when punching a sample from the injection molded disc was in some. cases ([greater than or equal to]27 vol% LCP) enough to separate a sample in two halves.
2.6. Density Measurements
The densities of the 10-mm-diameter samples punched from the circular discs was obtained using the Archimedes principle, e.g. by comparing the weights of the sample in air and in ethanol at 2200. A density kit provided by Mettler-Toledo was used for this purpose together with an AT26 1 balance with a resolution of [10.sup.-5] g. For each blend, four samples were taken along lines (at half the radius) at an angle of 900 angle to the adjacent sample.
2.7. Permeability Measurements
The permeability to oxygen was determined at 2300 and 0% RH using a Mocon Ox-Than Twin apparatus. Specimens with a thickness of 1 mm and a diameter of 135 mm were mounted in an isolated diffusion cell and were subsequently surrounded by flowing nitrogen gas (with 2% hydrogen) in order to remove sorbed oxygen from the samples. The blends with 47 vol% LCP had interfering knit lines, and the circular active area therefore had to be reduced to 5 [cm.sup.2], the rest of the specimen being covered with an impermeable aluminum foil. After establishment of the conditions above, one side of the sample was exposed to flowing oxygen (99.95%) at atmospheric pressure and the oxygen concentration was maintained at zero on the other side of the specimen. The oxygen transmission rate (Q) through the specimen was measured during the transition period until the steady state flow rate (Q) was obtained. The mean of systematic error of the oxygen permeability measurements at this laboratory was -10.9% according to results obtaine d by participating in 11 round-robin tests. The diffusivity of oxygen ([D.sub.O2]) was obtained by fitting the following equation to the experimental data for Q/[Q.sub.[infinity]] versus time using the Nelder-Mead Simplex method in Matlab (19, 20):
Q/[Q.sub.[infinity]] = 4l/[square root of (4[pi][D.sub.O2]t)]. [e.sup.-[l.sup.2]/4 [D.sub.O2]t] (1)
where l is the specimen thickness. The solubility of oxygen ([S.sub.O2]) was obtained assuming the validity of Henry's law:
[S.sub.O2] = [Q.sub.[infinity]]/[D.sub.O2]p (2)
where p is the partial pressure of oxygen on the high oxygen pressure side of the specimen, p 1 atm.
2.8. Infrared (IR) Spectroscopy
ATR-IR spectra were obtained in a Perkin-Elmer 1760-X Infrared Fourier Transform Spectrometer from 50 < 20 mm pieces cut from the injection molded disks. The blank surface of the disk was facing the ATR crystal. The incidence angle of the IR radiation to the thallium bromide-iodide crystal (KRS-5, Spectra Tech, USA) was 48[degrees]. The number of scans was 30. The resolution was set to 4 [cm.sup.-1] and the sampling of data was set to be one data point/cm'.
Golden gate single reflection ATR-IR spectra were taken using a Perkin-Elmer Spectmm 2000. The second FTIR instrument used was equipped with a Golden Gate Single Reflection Diamond ATR, Graceby Specac Inc. The moment applied to press the diamond with a diameter of 2 mm into contact with the sample was 80 cNm. The number of scans was 8, 20 or 30. Both the diamond and the sapphire were cleaned before starting the analysis of each new sample. The resolution was set to 4 [cm.sup.-1] and data were sampled at one data point/[cm.sup.-1]. The sampling for this measurement was done at half the radius on the discs using a 10 mm punch and at an angle of 90[degrees] angle to the adjacent sample (n = 4).
Cross sections of the injection molded specimens were studied with infrared microscopy using the Perkin-Elmer Spectrum 2000 instrument equipped with an AutoIMAGE Microscope. A square 1 mm X 1 mm in size was scanned. This map of the cross section contained about 100 spectra. The aperture size was thus set to be 100 [micro]m. Each spectrum taken at a certain position in the cross section was based on 20 scans. Samples for scanning were taken at half the radius of the circular discs.
3. RESULTS AND DISCUSSION
3.1. Blend Morphology
Figure 1 presents an overview of the morphologies of the injection molded blends. The morphology showed a detectable variation with depth for blends with >18 vol% LOP. In these blends, a relatively thin core layer was observed (Fig. 1). which consisted of LOP particles of anisotropic shape slightly different from the structure observed in the outer regions of the cross section. The thickness of the core layer with respect to the total thickness of the specimen at a position 30 mm from the center of the injection molded disc was as follows: Vectra A950 series- 10% (18 vol% LOP), 19% (27 vol% LOP) and 12% (47 vol% LOP); Vectra RD501 series: 18% (18 vol% LCP), 18% (27 vol% LOP) and 12% (47 vol% LCP). The blend morphologies of the core were different in the two blend series: the Vectra A950 blends showed morphological co-continuity, whereas the Vectra RD501 blends displayed a very fine fibrous morphology. A very thin surface layer, not detectable on the micrograph shown in Fig. 1, was present on all the samples. It appeared sometimes after the freeze-polishing as a thin rip-off, and its thickness was approximately 15 [micro]m. A thick layer with oriented LOP (referred to as the shear zone) was located between the skin and the central core layer. It constituted the major part of the specimen: 99% in the case of blends with [greater than or equal to] 9 vol% LOP and 80%-90% for the blends with [less than or equal to] 18 vol% LOP. The shear zone morphology was therefore characterized in detail, particularly with regard to the shape of the LOP particles.
The following notations for directions and planes in the circular specimens are used in this paper: r = radial direction; c = circumferential direction, and t = thickness direction (Fig. 2).
Figure 3 presents scanning electron micrographs of freeze-polished specimens. The samples were taken at half the radial distance from the center of the circular disc and they display the blend morphology of the shear zone. Fortunately, the adhesion between LCP and PE was relatively poor despite the use of the Nucrel compatibilizer. The mechanical forces acting on the surface induced by the glass knife were obviously sufficient to open up the interfaces and hence reveal the phase morphology. For each of the blends in Fig. 3, one view is presented on the micrographs (planes with axes c and t). The blends with low LOP content, [greater than or equal to] 4 vol%, showed ellipsoidal particles with the long axis in a radial direction, i.e. the flow direction (Figs. 3a and b). It is evident from a close examination of Figs. 3a and 3b that there is a wide distribution in LCP particle size and aspect ratio. The smaller particles (1-2 [micro]m) are essentially spherical whereas the larger particles have a pronounced ani sotropic shape (aspect ratio ~ 10). The sample with 9 vol% Vectra P1)501 (Fig. 3c) showed a similar tendency: the larger LCP particles were more anisotropic than the smaller particles. Occasional spherical LCP particles were found in this particular sample. Very wide LOP sheets (of considerable volume) were evident in this sample.
The blends with LCP contents of 18 vol% or higher showed a continuous blend morphology (Fig. 3d). An interesting feature seen in Fig. 3d was that the cut surface of the LOP sheets in these LOP-rich blends were rough after freeze polishing. The toughness of the LOPs at these low temperatures was also evident from attempts to freeze-polish the pure LOPs: Vectra A950 could not be polished; the glass knife fractured. Vectra P1)501 could be polished only with great difficulty. Figure 3e shows the structure of the core layer in the ct plane for 47 vol% Vectra RD501.
Table 1 presents a summary of the dimensions of the LOP particles in the shear zone of the injection molded blends. The dimensions of the LOP particles (classified as spheres, ellipsoids or lamellae) along the three reference directions r (radial), c (circumferential) and t (thickness) are presented. The blends with 4 vol% LOP showed ellipsoids with an aspect ratio of 2 and smaller spheres (diameter: 1-2 [micro]m). The ellipsoid had similar dimensions along r and c and they were thinnest along t. The blends with 4 vol% LOP showed a mixed morphology with small spheres (1-2 [micro]m). ellipsoids [aspect ratios = 5 (Vectra A950) and 50(r):3(c):l(t) (Vectra RD501)] and lamellae. For example, the blend with 4 vol% Vectra RD501 showed bands with the long axis along r. The lamellae found in the blend with 4 vol% Vectra A950 had the aspect ratio 29(r):16(c):1(t); the longest dimension being basically the same as that in the corresponding Vectra P1)501 blend, 200 [micro]m. The blends with 9 vol% LOP showed lamellae wi th the approximate aspect ratios 40(r):10-20(c):1(t), with the largest dimension near 500-600 [micro]m, and smaller ellipsoids with aspect ratios between 2 and 4 (dimensions along r and c were similar). The blends with LCP contents [greater than or equal to] 18 vol% displayed only lamellae in the mm range in the r and c directions and with a thickness between 20 and 50 [micro]m. The aspect ratio of these lamellae were thus in the range of 200-500. The LCP morphology for the blends with LCP content 18 vol% may thus be considered to be continuous.
The small LCP spheres (1-2 [micro]m) present in the blends with low LCP content (1-9 vol% LCP) are significantly smaller than the LCP spheres (5-20 [micro]m) present in compression-molded specimens based on blends of the same LCP compositions (16). The high shear forces acting on the material during the mold filling may disintegrate larger particles (fibers) into these very small, 1-2 [micro]m spheres.
An examination of the weight loss of the samples by hot p-xylene extraction further revealed the continuity of the LCP component. It was confirmed by IR spectroscopy of the insoluble fraction that p-xylene extraction removed all PE from the samples. Hence, the remaining mass of sample should be the continuous portion of LCP. Figure 4 shows the mass fraction of the insoluble recoverable sample as a function of the LCP mass fraction of blend. No solid residue could be recovered from the blends with 1 and 4 vol% LCP after extraction, which suggested that the LCP component in these samples formed a discrete phase. The LCP particles were in these cases microscopic and not isolated by the method used. The blend with 9 vol% Vectra A950 showed a small solid residue after extraction, indicating that ~30% of the LCP formed a continuous phase in this blend, the rest (~70%) being in the form of discrete particles. The blend with 9 vol% Vectra RD501 showed no recoverable solid residue after extraction. A comparison betwee n these data and the data presented in Table 1 suggest that smaller lamellae (~ 500 [micro]m along r) do not form continuous structures (agglomerates) in the mm range that can be recovered as a solid residue after extraction. The blends with [greater than or equal to] 18 vol% LCP displayed remaining mass fractions very similar to the initial mass fractions of LCP in the blends, indicating that almost 100% of the LCP formed a continuous phase. The numbers shown adjacent to each data point in Fig. 4 indicate the number of layers recovered after extraction: the number of layers showed an increase with increasing LCP content. These layers had the full diameter of the injection molded discs (Fig. 5). The shape of the layers of LCP-rich material observed suggests that the continuity of the LCP phase was only 'partial' and that it was confined to the rc plane.
Visual inspection of the injection molded discs revealed the presence of knit lines, which are formed where flow fronts meet. The knit lines were more readily distinguished in the blends with a higher LCP content. The blend samples with 47 vol% LCP had knit lines that seriously deteriorated the permeability measurements (overflow of gas through the knit lines) and specimens with smaller surface area without knit lines had to be used for the measurements. Engberg et al. (21, 22) reported that the mechanical strength was seriously low in cold knit lines (co-linear impingement of the melt fronts) and that it was significantly reduced even in so-called warm knit lines. The lack of material continuity in the knit line regions of the blends with 47 vol% LCP is thus not unexpected. The numbers of knit lines detected in the injection molded specimens were as follows: Vectra A950 series; 2 (4 vol% LCP), 4 (9 vol% LCP), 2 (18 vol% LCP), 2 (27 vol% LCP) and 4-6 (47 vol% LCP); Vectra RD501 series: 2 (18 vol% LCP), 2 (27 vol% LCP) and 3-8 (47 vol% LCP).
Reflection IR spectra were taken with the ATR and Golden gate techniques in order to determine the composition of the outermost 0.5 [micro]m of the injection molded sheets. FTIR was used as a surface analysis method. It is a useful technique, since it samples from the top micrometer of the sample (1). Two absorption bands were used for the analysis: 1733 [cm.sup.-1] (carbonyl stretching vibration; the Vectra band) and 2915 [cm.sup.-1] (antisymmetric stretching vibration; the PE band). The penetration depths (i.e. 'sampling' depths) associated with the measurements of these two absorption bands were 0.58 [micro]m (1733 [cm.sup.-1]) and 0.34 [micro]m (2912 [cm.sup.-1]) assuming that these quantities are inversely proportional to the wavelengths (1). Figure 6 presents a concise summary. The surface concentration of Vectra, a quantity that is proportional to the absorbance ratio ([A.sub.1733]/[A.sub.2915]), was essentially proportional to the overall composition. Hence, these data suggest that the surface layer w as not significantly enriched in LCP.
Using IR-microscopy, measuring the ratio of the absorbances associated with the 1733 [cm.sup.-1] and 2915 [cm.sup.-1] bands, the LCP content in the tc plane was assessed. These results should only be considered as semi-quantitative. The general tendency was assessed for the surface regions to be slightly enriched in LOP (Fig. 7a), although some samples showed a more uniform LCP distribution through the cross section (Fig. 7b). The significant variation in composition on the 50 [mu]m level, present in both displayed diagrams, is consistent with the coarse two-phase morphology revealed by SEM.
A very important general question is why the LOP lamellae are formed. The injection of a melt flow through a central gate will spread out the flow in a disc shape. The deformation of the LOP particles will Therefore occur in two orthogonal directions (r and c). The main flow direction is along r and the shear deformation along r, will produce a radial orientation. The melt flow is however, also subjected to a stretching flow in the circumferential direction resulting in a transverse orientation. Finally, the orientations in directions r and c has to be balanced in order to avoid deformation caused by shrinkage after molding (23).
3.2. Void Content in Blends
Figure 8 shows The density at 2200 of samples taken from The injection molded discs as a function of LCP content for the two blend series. The data are averages from samples taken at different positions in four different discs.
Since The polymers of The blends were completely immiscible, a linear relationship between density and volume fraction of LOP is expected:
[rho] = [[rho].sub.PE] + [v.sub.LCP] ([[rho].sub.LCP] - [[rho].sub.PE]) +
[v.sub.Nu]([[rho].sub.Nu] - [[rho].sub.PE]) [approximately equal to] [[rho].sub.PE] + [v.sub.LCP] ([[rho].sub.LCP] - [[rho].sub.PE]) (3)
where [rho] is the density of the blend, [[rho].sub.PE] is the density of PE (= 941 kg/[m.sup.3]), [[rho].sub.LCP] is the density of LCP (= 1400 kg/[m.sup.3]), [[rho].sub.Nu] is the density of Nucrel (= 940 kg/[m.sup.3]), [v.sub.LCP] and [v.sub.Nu] are The volume fractions of LOP and Nucrel in the blend. The data for the blends fall well below The line representing Eq 3, indicating that the injection molded discs contained voids. The void content was calculated as follows:
[v.sub.void] = 1 - [[rho].sub.PE][v.sub.PE] + [[rho].sub.LCP][v.sub.LCP] + [[rho].sub.Nu][v.sub.Nu]/[rho] (4)
where [v.sub.PE] is the volume fraction of PE in the blend. The void contents in The different blends were as follows: Vectra A950 blends: 0.6% (1 vol% LOP), 2.1% (4 vol% LOP), 2.4% (9 vol% LCP), 4.9% (18 vol% LOP), 4.3% (27 vol% LOP) and 8.8% (47 vol% LOP); Vectra RD501 series: 0.7% (1 vol% LOP), 1.9% (4 vol% LOP), 2.8% (9 vol% LOP), 4.3% (18 vol% LOP), 4.8% (27 vol% LOP) and 9.9% (47 vol% LOP). These data are somewhat lower than the void content data earlier presented for samples obtained by compression molding based on The same polymers (16). Film-blown blends displayed even higher void contents (16): the blend with 50 vol% Vectra A950 had more Than 20% voids and The blend with 50 vol% Vectra RD501 had 10%-18% voids. The obvious reason for void formation is the poor adhesion between LOP and PE. The compatibilizer (Nucrel) overcomes this problem to some extent, but not sufficiently to avoid the void formation. It is therefore very probable that The voids are confined to the interfaces between LCP and PE. In the injection molded discs, these interfaces are primarily confined to the rc plane, which is less serious for the transport of gases along t. The LCP-PE interfaces in the compression molded specimens should be randomly distributed on all planes and hence also along planes with t, which would permit gas transport through the voids along the concentration gradient. Weiss et al. (24-26) recently reported about very effective reactive compatibilization of Vectra polymers and polyethylene using the sodium salt of poly(ethylene-ran-acrylic acid). It may be possible to decrease the void content in LCP/PE blends using this new type of compatibilizer.
3.3. Transport Properties
The oxygen transmission rate measurements permitted the determination of oxygen permeability ([P.sub.2]), diffusivity ([D.sub.02]) and solubility ([S.sub.02]). The solubility data obtained from input of the oxygen transmission rate at steady state and the diffusivity obtained from the analysis of the transient period in Eq 2 are associated with very considerable scatter (Fig. 9). The trend is, however, clear: the solubility decreased monotonically with increasing LCP content. For the Vectra A950 series (0-27 vol%) the mean values for the solubility varied between 0.036-0.022 [cm.sup.3]/([cm.sup.3] atm), in relative terms a 39% reduction, and for the Vectra RD501 blends the solubility ranged between 0.029 and 0.007 [cm.sup.3]/([cm.sup.3] atm), which corresponds to a relative decrease of 76%. The solubilities shown in Fig. 9 are one order of magnitude lower than the values earlier reported by Flodberg et al. (16) for compression molded films based on the same polymers.
The data obtained show essentially the expected trend: a linear decrease in oxygen solubility with increasing LCP content. The solubility of oxygen is significantly higher in PE than in LCP and the overall solubility should be proportional to the LCP content according to:
[S.sub.O2] = [S.sub.O2, LCP] * [V.sub.LCP] + [S.sub.O2, PE] [v.sub.PE] =
[S.sub.O2, PE] - [v.sub.LCP]([S.sub.O2, PE] - [S.sub.O2, LCP]) (5)
where [S.sub.O2, LCP] and [S.sub.O2, PE] denote the oxygen solubilities in LCP and PE respectively. It is evident that the data obtained for Vectra RD501 are inconsistent with Eq 5. Extrapolation of the solubility data for the Vectra RD501 blend series would result in a negative value for [S.sub.O2,LCP]. The solubility data for the Vectra A950 seems more reliable [S.sub.O2,LCP] [approximately equal to] 0.005 [cm.sup.3]/([cm.sup.3] atm).
The oxygen diffusivity data presented in Fig. 10 was obtained by analyzing the transient period according to Eq 1. The diffusivity decreased very strongly with increasing LCP content in a similar fashion for both blend series. In the LCP content range of 0-27 vol%, there is a ten-fold decrease in diffusivity in both blends series (Fig. 10).
The change in diffusivity with increasing LCP content resembles to some degree the change in diffusivity with increasing crystallinity for semicrystalline polymers. There is a close analogy between the LOP particles and the crystallites in semicrystalline polymers in that both are almost impenetrable for oxygen. The oxygen molecules have to make a very significant detour in both cases, and according to the classic work of Michaels and Bixler (27), this can be expressed by a geometrical impedance factor ([tau]):
[D.sub.O2] = [D.sub.O2] ([v.sub.LCP] = 0)/[tau][beta] (6)
where [D.sub.O2] is the oxygen diffusivity of the blend sample with LCP, [D.sub.O2] ([v.sub.LCP] = 0) is the oxygen diffusivity in the pure matrix polymer, [beta] is a factor taking into account the constraining effect of the LOP phase on the penetrable PE phase. It is evident that the two polymers are completely immiscible and hence [beta] = 1. Michaels and Bixler (27) proposed the following relationship between [tau] and the volume fraction of the penetrable phase (in this case, 1-[v.sub.LCP]):
[tau] = [(1 - [v.sub.LCP]).sup.-n] (7)
where n is a constant that for semicrystalline polymers takes values between 1.00 (PETP) and 1.88 (linear and branched PE) (28). The best fit of Eqs 6 and 7 to the experimental data (considering data of LCP contents [greater than or equal to] 4 vol%) presented in Fig. 10 yields n = 6, which suggests that the detour must be very long indeed in the LCP blends. Further information on this matter is presented in the section dealing with the analysis of the permeability data.
Figure 11 shows the oxygen permeability (determined from the steady state oxygen transmission rate) plotted as a function of the volume fraction of LCP. The oxygen permeability showed a decreasing trend with increasing LCP content except for the blends with 1 vol% LCP. The permeability decreased by one to two orders of magnitude between 0 and 27 vol% LCP: the larger decrease occurred for the Vectra RD501 blends. The permeability data were analyzed with the model of Fricke (29); the latter was originally derived to describe electrical conductivity of suspensions of particles (oblate spheroids) in a continuous medium. It provides results that are consonant with results obtained by Monte Carlo simulation (30). The Fricke model predicted the correct crystallinity dependence of the geometrical impedance factor for linear polyethylene (31). The input data used for the calculations based on the Fricke model were the oxygen permeabilities of the two polymers (LCP and PE): [P.sub.O2,LCP] = 0.001 [cm.sup.3] mm/([m.sup. 2] d atm); [P.sub.O2,PE] = 70.9 [cm.sup.3] mm/([m.sup.2] d atm). The LCP phase is in the form of oblate spheroids with a certain aspect ratio (width/ thickness) (w/t) surrounded by a continuous PE phase. The equations used in the analysis are as follows:
[P.sub.O2] = [P.sub.O2,PE] + [P.sub.O2,LCP] C / 1 + C (8)
where C is given by:
C = 1 / 3
[1 / 1 + ([P.sub.O2,LCP]/[P.sub.O2,PE] - 1) * (1 - M)] * [v.sub.LCP] / 1 - [v.sub.LCP] (9)
where [v.sub.LCP] is the volume fraction of LCP in the blend and M is equal to:
M (t < w) = ([phi] - sin2[phi]/2)/[sin.sup.3][phi] * cos [phi] (10)
cos [phi] = t / w (11)
where [phi] is an angle given in radians.
The dashed curves in Fig. 11 are theoretical curves based on the Fricke model assuming certain constant values for the aspect ratio (wIt): 35, 50 and 500. None of the experimental permeability data curves follows a single theoretical curve (Fig. 11). The aspect ratios of the oblate LCP spheroids or lamellae for the two blend series increased significantly with increasing LOP content according to the data obtained by SEM (Table 1), which is in qualitative agreement with the predictions based on the Fricke model (Fig. 11). For the Vectra A950 blends the aspects ratios obtained by SEM were: 4 vol% LOP: 29(r) X 16(c) X 1(t) for lamellae and 5(r) X 4(c) X 1(t) for ellipsoids; 9 vol% LOP: 43(r) X 19(c) X 1(t); 18 vol% LOP: > 430(r) X > 430(c) X 1(t); 27 vol% LOP: > 400(r) X > 400(c) X 1(t): 47 vol% LOP > 200(r) x > 200(c) x 1(t). For the Vectra P1)501 blends, the following aspect ratios were obtained: 4 vol% LOP: 50(r) X 3(c) X 1(t); 9 vol% LOP: 40(r) X 10(c) X 1(t); 18 vol% LOP: > 500(r) X > 400(c) x 1(t); 27 vol% LOP: > 300(r) X > 200(c) x 1(t); 47 vol% LOP> 300 (r) > x 200 (c) X 1(t). For the Vectra RD501 blend series the blend morphologies revealed by SEM are in accordance with the data of the aspect ratio of the LOP phase calculated by applying the Fricke model on the oxygen permeability data (Fig. 11). The deviations between theoretical predictions and experimental SEM data for the Vectra RD501 blends were small in view of the accuracy of the methods used. Hence, it seems that the presence of voids in these blend samples (see section 3.2) has only negligible effect on the oxygen permeability. It is suggested the voids are primarily confined to the LOP-PE interfaces and in the injection molded samples these would be placed primarily in the rc plane, which has less consequence for the oxygen transport along the t direction. The Vectra A950 blends showed higher oxygen permeabilities than the Vectra RD501 blends at LOP contents [greater or equal to] 9 vol% despite the fact that no significant morphological difference s were observed between the two blend series. It may be that the voids in the Vectra A950 blends form more continuous paths along t than in the Vectra RD501 blends.
It was impossible to assess the permeability for the blends with 47 vol% LOP with the Mocon apparatus; the oxygen transmission rates were lower than the detection limit 0.02 [cm.sup.3]/([m.sup.2]d) (32). This finding led us to the value for [P.sub.O2.LCP], 0.001 [cm.sup.3] mm/([m.sup.2] d atm), used in the theoretical calculations.
Injection molding of 1-mm-thick discs of binary blends of LCP and PE with LCP contents [greater than or equal to] 18 vol% resulted in the formation of LCP lamellae of macroscopic width (mm). SEM confirmed the two-dimensional continuity of the LCP domains in the radial-circumferential plane, which was due to the presence of radial shear deformation and circumferential stretching of the melt coming from the central gate. The oxygen permeability showed a significant decrease with increasing LCP content. This effect was primarily due to a significant reduction in the diffusivity. The decrease in solubility was also important at higher content of LCP, especially for the Vectra RD501 blend series. The presence of LCP lamellae with very high aspect ratio (at least several hundred) gave these systems (LCP contents [greater than or equal to] 18 vol%) very high geometrical impedance factors. The aspect ratios obtained by analyzing the permeability according to the Fricke model were for Vectra RD501 blend series in acco rdance with those obtained by SEM. The Vectra A950 blends showed higher oxygen permeabilities than predicted by the Fricke model using the blend morphology data obtained by SEM. The injection molded specimens contained microvoids but to a lesser extent than specimens prepared by compression molding or film blowing from similar blends. It is possible that the comparatively high oxygen permeabilities obtained for the Vectra A950 blends were due to presence of more continuous paths of voids along the thickness direction of the disc-shaped samples.
[FIGURE 4 OMITTED]
[FIGURE 6 OMITTED]
[FIGURE 8 OMITTED]
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
Table 1 Shape of LCP Particles in Shear Zone. Blend [L.sub.r] ([micro]m) Vectra A950 1 vol% 15 [+ or -] 3 (a) 1 - 2 (b) 4 vol% 200 [+ or -] 168 (c) 21 [+ or -] 3 (a) 1 - 2 (b) 9 vol% 600 [+ or -] 340 (c) ~ 13 (a) 1 - 2 (b) 18 vol% >10,000 (c) 27 vo1% >10,000 (c) 47 vo1% >10,000 (c) Vectra RD501 1 vol% 1 - 5 (b) 4 vol% 195 [+ or -] 56 (a) 1 - 2 (b) 9 vol% 500 [+ or -] 340 (c) 23 [+ or -] 2 (a) 18 vol% >8000 (c) 27 vol% >12,000 (c) 47 vol% >11,000 (c) Blend [L.sub.c] ([micro]m) Vectra A950 1 vol% 12 [+ or -] 4 (a) 1 - 2 (b) 4 vol% 110 [+ or -] 33 (c) 16 [+ or -] 16 (a) 1 - 2 (b) 9 vol% 270 [+ or -] 60 (c) 11 [+ or -] 7 (a) 1 - 2 (b) 18 vol% >10,000 (c) 27 vo1% >10,000 47 vo1% >10,000 Vectra RD501 1 vol% 17 [+ or -] 16 (a) 1 - 2 (b) 4 vol% 13 [+ or -] 5 (a) 1 - 2 (b) 9 vol% 130 [+ or -] 60 (c) 36 [+ or -] 17 (a) 18 vol% >7000 (c) 27 vol% >7000 (c) 47 vol% >7000 (c) Blend [L.sub.t] ([micro]m) Vectra A950 1 vol% 7 [+ or -] 6 (a) 1 - 2 (b) 4 vol% 7 [+ or -] 2 (c) 4 [+ or -] 2 (a) 1 - 2 (b) 9 vol% 14 [+ or -] 5 (c) 6 [+ or -] 3 (a) 1 - 2 (b) 18 vol% ~ 23 (c) 27 vo1% ~ 25 (c) 47 vo1% ~ 50 (c) Vectra RD501 1 vol% 5 [+ or -] 2 (a) 1 - 2 (b) 4 vol% 4 [+ or -] 1 (a) 1 - 2 (b) 9 vol% 13 [+ or -] 4 (c) 9 [+ or -] 5 (a) 18 vol% ~ 16 (c) 27 vol% ~ 42 (c) 47 vol% ~ 35 (c) (a)Dimension of ellipsoid along r, c or t, average [+ or -] standard deviation. (b)Dimension of sphere along r, c or t, average. (c)Dimension of lamellae along r, c or t, average [+ or -] standard deviation.
The assistance during permeability measurements by L. Hojvall, Packforsk, Sweden is gratefully acknowledged. O. Hansson, Mollers Verktygsmakeri An, Ekero, Sweden is gratefully acknowledged for help during the mold design and for assistance during the injection molding.
(1.) U. W. Gedde, Polymer Physics, Chapman & Hall, London (1995).
(2.) B. D. Freeman and A. J. Hill., in Structure and Properties of Glassy Polymers, M. R. Tant, A. J. Hill, eds., American Chemical Society, Washington, D.C. (1998).
(3.) J. S. Chiou and D. R Paul, J. Polym. Sc., Polym. Phys. Ed., 25, 1699 (1987).
(4.) D. H. Weinkauf and D. R. Paul, in Barrier Polymers and Barrier Structures, p. 60, W. J. Koros, ed., American Chemical Society, Washington, D.C. (1990).
(5.) G. Flodberg, L. Axelson-Larsson, M. S. Hedenqvist, and U. W. Gedde. Packag. Techn. Sci., 14, 159 (2001).
(6.) G. Flodberg, L. Hojvall, M. S. Hedenqvist, and U. W. Gedde, Food Additives and Contaminants, 19, 492 (2002).
(7.) R. W. Lusignea, in Nova-Pack Americas (1999).
(8.) G. Plester and S. Vandekerckhove, U.S. Patent No. 5,707,691 (1998).
(9.) R. Lusignea and J. Perdikoulias, "Extrusion of Oriented LCP Film," in SPEANTEC Tech. Papers (1997).
(10.) S. Kamoi and Y. Matsuki, Liquid-crystal display device containing gas barrier layer, Application No. JP 89125897. Date 19890518: Japan (1990).
(11.) K. E. Wetzel, Low cost ccd packaging, No. U.S. Patent 6,011,294 (2000).
(12.) S. Kato, K. Nakamura, K. Ono, H. Umeyama, and A. Harasawa, Gas-barrier articles and packaging materials thereof Patent No. JP 10202780, Japan (1998).
(13.) S. G. James, A. M. Donald, I. S. Miles, and W. A. MacDonald, High Perform. Polym., 4, 3 (1992).
(14.) G. Wiberg, M. S. Hedenqvist, R. H. Boyd, and U. W. Gedde, Polym. Eng. Sc., 38, 1640 (1998).
(15.) L. Incarnato, P. Scarfato, O. Motta. and D. Acierno, Polym. Comp., 21. 354 (2000).
(16.) G. Flodberg, A. Hellman, M. S. Hedenqvist, E. R Sadiku, and U. W. Gedde. Polym. Eng. Sci., 40, 1969 (2000).
(17.) G. Flodberg, L. Hojvall, M. S. Hedenqvist, E. R. Sadiku, and U. W. Gedde, Int. J. Polym. Mater., 49, 157 (2001).
(18.) D. V. Rosato, D. V. Rosato, and M. G. Rosato, eds., Injection Molding Handbook, 3rd Ed., Kluwer, Norwell, USA (2000).
(19.) R. A. Pasternak, J. F. Schimscheimer, and J. Heller, J. Polym. Sc., 8, 467 (1970).
(20.) M. S. Hedenqvist, Optimisation of D from flux measurements, Department of Fibre- and Polymer Technology, Royal Institute of Technology, Sweden.
(21.) K. Engberg and U. W. Gedde, Progr. Coll. Polym. Sci., 87, 57 (1992).
(22.) K. Engberg, A. Knutsson, P.-E. Werner, and U. W. Gedde, Polym. Eng. Sci., 30, 1620 (1990).
(23.) R. S. Leak, Polymer Rheology, Chapter 12, p. 137, Applied Science Publishers, London (1978).
(24.) H. Zhang, R. A. Weiss, J. E. Kuder, and D. Cangiano, Polymer, 41, 3069 (2000).
(25.) Y. Son and R. A. Weiss, Polym. Eng. Sci., 41, 329 (2001).
(26.) Y. Son and R. A. Weiss, Polym. Eng. Sc., 42, 1322 (2002).
(27.) A. S. Michaels and H. J. Bixler, J. Polym. Sci., 5, 413 (1961).
(28.) M. Hedenqvist and U. W. Gedde, Progr. Polym. Sc., 21, 299 (1996).
(29.) H. Fricke, Phys. Review, 24, 575 (1924).
(30.) P. M. Hadgett, G. Goldbeck-Wood, and A. H. Windle, Polymer, 41, 6151 (2000).
(31.) M. Hedenqvist, A. Angelstok, L. Edsberg, P. T. Larsson, and U. W. Gedde, Polymer, 37, 2887 (1996).
(32.) Operating Manual for the Ox-Tran Twin Oxygen Permeability Tester. Mocon Controls, Inc., Minneapolis (1986).
U. W. GEDDE *
* Corresponding author, E-mail: email@example.com
|Printer friendly Cite/link Email Feedback|
|Author:||Flodberg, G.; Hedenqvist, M.S.; Gedde, U.W.|
|Publication:||Polymer Engineering and Science|
|Date:||May 1, 2003|
|Previous Article:||Ternary composites based on PP-EPDM blends reinforced with flax fibers. Part II: mechanical properties/morphology relationship.|
|Next Article:||A method to control the dispersion of carbon black in an immiscible polymer blend.|