Printer Friendly

Blends of a bottle-grade polyethylene terephthalate copolymer with a liquid crystalline polymer. Part II: thermal and transport properties.

INTRODUCTION

In the first part of this work (1), the injection molding morphology and the morphological properties of blends made of a bottle-grade polyethylene terephthalate copolymer, PET, and a liquid crystalline polymer, LCP, were studied. These blends were developed to use in the injection blow molding of bottles. The objectives of that first part were: i) to predict the resultant injection molding morphology from the pure components rheological behavior, ii) to study the interfacial interactions that occurred between these components during the processing, and iii) to measure the level of residual stresses obtained at the blow molding step. In this second part, the thermal and transport properties of dichloromethane in these blends were measured.

The transport properties of dichloromethane in PET were studied by Ruvolo et al. (2) and Vittoria et al. (3). Ruvolo et al. (2), using wide angle X-ray diffraction, WAXD, and sorption properties showed that there was a fraction of impermeable phase in the cold-crystallized PET samples that was not crystalline. This phase was considered to be a mesomorphic phase, characterized by conformationally ordered regions, impermeable to the dichloromethane at low solvent activity. Vittoria et al. (3) found in glassy amorphous PET samples the presence of ordered domains, impermeable to the dichloromethane, that became permeable at higher activities. They also observed a solvent induced crystallization, SING, to occur in the PET after an activity of 0.4, in the aged samples, that was not observed in the fresh samples. They attributed the occurrence of SING in the aged samples to the existence of a more ordered structure.

The transport properties of acetaldehyde in PET were also investigated (4). It was found that at 45[degrees]C and acetaldehyde pressures of 5.33 X [10.sup.4] Pa and above, the acetaldehyde penetrant triggers significant crystallization of the PET.

The transport properties of acetone in PET have also been studied (5). By measurements of WAXD, it has been found that the amount of thermal induced crystallinity, TINC, is higher than that of SING, and that the thermal induced crystallites are more dense than the solvent induced crystallites; therefore the annealed untreated PET is more impermeable to acetone than the acetone-treated PET.

Attempts have also been made to correlate the morphology with the barrier properties of PET. One study (6), for example, found that more numerous and smaller PET spherulites allowed greater oxygen permeability than samples of equivalent crystallinity but larger size spherulites; that is, larger spherulites were more effective in creating increased tortuosity in the path of the oxygen permeant molecules.

Regarding the LCP, their transport properties were also extensively studied by Paul and associates (7-10) and de Candia and associates (11, 12). Chiou and Paul (7) found very low permeability coefficients for these materials; these low coefficients were attributed to low gas solubility, rather than low mobility for penetrant diffusion. The LC order and chain packing led to lower gas solubility; also the gas molecules were assumed to be soluble only in the defects or boundaries between the LC domains. Their conclusion was confirmed in a subsequent study (9) in which the volume fraction of the regions with LC order was increased; a reduction of the gas solubility was observed while the gas diffusivity was only slightly affected. As those authors pointed out, their results were in contrast with what it is commonly observed in semicrystalline polymers where the increase in crystallinity affects equally both the solubility and the diffusivity. Another study (10) by the same authors showed that increased orientatio n and annealing of these materials caused reductions in gas permeabilities; however, the magnitude of the reduction in permeability caused by orientation was not as severe as it was observed in highly oriented amorphous and semicrystalline polymers. The effect of increased crystallinity on the permeability coefficients was smaller than would be expected for similar changes in crystallinity in conventional polymers. It was also assumed that the gas solubility occurred in the boundary regions and that the nematic phase did not sorb gas at all.

Studies using methylene chloride (11) and carbon dioxide (12) as permeants gave evidence of the existence of a disordered phase in the LCP, which behaved as a permeable component. This phase could also be within the liquid crystalline domains: the higher the fraction of this phase, the higher the sorption properties.

Regarding the correlation between blend morphology and barrier properties, it was shown in an early study (13) of blends of a high density polyethylene, HDPE, with a polyamide that the optimum morphology to achieve outstanding banter properties for this particular system was the formation of multiple layers or platelets of the polyamide along the wall thickness. Thus, the disperse phase had to have a high aspect ratio. Also a good compatibilization between both components was desirable. Another study (14) of blends of PET with ethylene-vinyl alcohol, EVOH, showed that high aspect ratio EVOH particles (thinner particles) were more effective in reducing the oxygen permeability than low aspect ratio particles (thicker ones). A study (15) of blends of PET with a LCP showed that the permeability to a mixture of [CO.sub.2], [O.sub.2] and [N.sub.2] decreased as the LCP volume fraction in the blend increased. In these blends, the LCP was distributed in the PET in the form of spherical domains and short fibers, with g ood adhesion between both polymers.

Other studies (16, 17) of the transport properties of [CO.sub.2] in blends of polyetherimide, PEI, with a LCP have found that the permeability of the blends to this gas decreased with the increase in the amount of LCP; a good theoretical fit to the experimental data was found when it was assumed that the blends formed a composite membrane made of fibers and laminates (16). It was also found that as the amount of LCP in the blend increased, the water diffusion coefficient decreased. Also a decrease of the water solubility coefficient with an increase in the amount of LCP in the blends was observed. As a result, the increase of LCP in the blend induced a decrease in the water permeability.

In blends of polyethylene, PE, with an LCP, a slight decrease in oxygen permeability was observed (18); the authors concluded that this decrease was an indication that the LCP did not form a continuous phase.

From the analyses of these results we can conclude that many structural parameters will influence the transport properties; however, some of them can be classified as critical:

a. Degree of crystallinity of the PET and LCP in the blends.

b. Morphology of the disperse phase in the blends.

c. Orientation of the PET and the LCP in the blends.

d. Adhesion or compatibilization between PET and LCP in the blends.

In this work, we will limit ourselves to study the influence of the first three parameters on the transport properties of these blends.

EXPERIMENTAL:

Materials

The bottle-grade PET resin was a copolymer of terephthalic acid, isophthalic acid (3.5 mol%) and diethylene glycol (3.0 mol%), with bulk density [rho] = 1.3435 X [10.sup.3] kg/[m.sup.3] and intrinsic viscosity [[eta].sub.v] = 0.0890 [m.sup.3]/kg. kindly donated by Rhodia Ster of Brazil. The LCP was Vectra A950, a random copolyester made of 75 mol % of 4-hydroxybenzoic acid (HBA) and 25 mol % of 6-hydroxy-2-naphthoic acid (HNA), from Hoechst Celanese.

Blend Preparation

Blends were prepared, on a weight percent basis, by injection molding. The lends are referred to as B20 through B80, where the figures indicate the wt% of LCP.

Before injection molding, the pure components and the blends were dried at 150[degrees]C, for 15 h, under nitrogen atmosphere, to avoid PET degradation by hydrolysis. The blends were then injection molded in an Arburg 270v Injection molding machine, in the starved feed mode, with the temperature of the last zone being equal to 290[degrees]C and the mold temperature equal to room temperature. The injection pressures varied between 600 and 700 X [10.sup.7] Pa, decreasing with the increase in the amount of LCP. The injection molded samples were opaque, rectangular plaques with an area of 4.46 X [10.sup.-4] [m.sup.2] and thickness of (400 [+ or -] 20) X [10.sup.-6] m.

Before all the morphological, dynamic mechanical, calorimetric, X-rays and transport measurements, the blends were annealed at 150[degrees]C for 15 h to obtain more thermodynamically stable samples. This last temperature was chosen because it was close to the PET maximum crystallization temperature, which was found to be approximately 180[degrees]C (19).

Thermal Transitions by Dynamic Mechanical Thermal Analysis (DMTA)

The transport properties of a polymer depend on the free volume within it and on the segmental mobility of the polymer chains (20); then knowledge of the long and short-range segmental mobility of the polymer is necessary. Therefore, the glass transition temperature, [T.sub.g], of the pure components and of the blends, which can be correlated to long range segmental mobility, was measured by dynamic mechanical thermal analysis, DMTA, in an ARES rheometer, from Rheometrics, in the torsion mode, at a heating rate of 4[degrees]C/min, frequency [omega] = 1 Hz and strain of 0.2%, between -100[degrees]C and 250[degrees]C. The [T.sub.g] was the temperature where tan [delta] had a maximum; the transition below [T.sub.g] was named [T.sub.[beta]] and the transition below [T.sub.[beta]] was named [T.sub.[gamma]]. These sub-[T.sub.g] transitions can be related to short-range segmental mobility.

Degree of Cystallinity by Modulated Differential Scanning Calorimetry, MDSC

The degree of crystallinity of the PET and of the LCP in the blends, [[chi].sub.(PET)] and [[chi].sub.c(LCP)] respectively, was determined from measurements of the PET and LCP heats of fusion by using a modulated differential scanning calorimeter, MDSC, from TA Instruments, model 2920. The calibration was made with indium; the heating rate was 50[degrees]C/min. from 150[degrees]C to 320[degrees]C, under a nitrogen atmosphere.

Four well-defined peaks of fusion were observed; two, at around 172.2[degrees]C and 246.3[degrees]C, were attributed to the PET, and two at around 190.6[degrees]C and 288.1[degrees]C were attributed to the LCP.

The theoretical heat of fusion of 100% crystalline PET, [([DELTA][H.sub.m.sup.o]).sub.PET], was taken as 144 kJ/kg (21); therefore [[chi].sub.c(PET)] was calculated from the following equation, which already was corrected for the equipment calculations:

[[chi].sub.c(PET)] = [[([DELTA][H.sub.1]).sub.PET] + [([DELTA][H.sub.2]).sub.PET]]/[[alpha][([DELTA][H.sup.o.sub.m]).sub.P ET]] (1)

where

[([DELTA][H.sub.1]).sub.PET] = PET heat of fusion at around 172.2[degrees]C;

[([DELTA][H.sub.2]).sub.PET] = PET heat of fusion at around 246.3[degrees]C

and

[alpha] = weight fraction of PET in the blend.

The theoretical heat of fusion of 100% crystalline LCP, [([DELTA][H.sub.m.sup.o]).sub.LCP] was taken as 28 kJ/kg (10); therefore [[chi].sub.c(LCP)] was calculated from the equation:

[[chi].sub.c(LCP)] = [([DELTA][H.sub.1]).sub.LCP] + [([DELTA][H.sub.2]).sub.LCP]]/[(1 - [alpha])[([DELTA][H.sup.o.sub.m]).sub.LCP]] (2)

where

[([DELTA][H.sub.1]).sub.LCP] = LCP heat of fusion at around 190.6[degrees]C;

[([DELTA][H.sub.2]).sub.LCP] = LCP heat of fusion at around 288.1[degrees]C

WAXD Measurements

Crystallinity measurements by WAXD are routinely made in polymers; however, with LCPs, it is observed that the values of crystallinity obtained from the scattering data are too high when compared with other methods like DSC, for example. This discrepancy is due to the high orientation that the LCP organized domains have, and also to the belief that the order in these polymers is not truly crystalline, but one-dimensional. The "crystallites" of these LCPs are slab-like lamellar structures (22), having larger lateral dimensions than longitudinal dimensions. Therefore, a real degree of crystallinity cannot be calculated from the WAXD data of LCPs.

Biswas (22) has also pointed out that despite the random sequences in a LCP, diffraction data from oriented samples reveals the presence of some kind of order, that a PCL (paracrystalline lattice) model better describes these measurements, and that the collection of an amorphous halo in these materials is very difficult because of the persistence of nematic order up to the decomposition temperature. In order to measure crystalline orientation in LCPs, the Herman orientation function f must be calculated. In a recent work (23) on blends cf polypropylene and Vectra A950, this orientation function was calculated for this LCP and the PP. It was observed that the pure Vectra A950 had f = 0.751, while in the blends this value was between 0.2 and 0.1, showing that the mixing of the LCP with the PP diminished the Vectra A950 orientation. On the other hand, the PP showed negligible orientation in both, the pure state and in the blends.

Because we lack the WAXD accessories necessary to measure f in our labs, we ran WAXD experiments to analyze qualitatively the bulk orientation of the LCP in our blends. These measurements were made in a Siemens Diffrac D5000 equipment, with Cu[K.sub.[alpha]] radiation, and a rotator sample holder rotating at 120 rpm.

2[theta] scans of the blends, as plaques, and of the pure LCP, as a plaque and as a powder, were made. The LCP powder was obtained from the milling of the plaque.

Morphology of the Samples

The morphology of the injection molded blends was analyzed by using a scanning electron microscope, SEM, from Zeiss, model DSM 960. The samples were immersed in liquid nitrogen and broken after 10 min of immersion. The fracture surfaces were then gold sputtered under vacuum, in all the samples, the analyzed surface was parallel to the main flow direction.

Transport Properties

For these measurements, the injection molded plaques were used. However, before these measurements, the thickness of the plaques was measured again. The LCP and blend plaques maintained almost the same thickness; however, the PET plaque had a thickness increase to (620 [+ or -] 20) x [10.sup.-6] m , not displaying a planar surface, probably due to the annealing promoted by the drying conditions.

The sorption properties were studied by measuring the weight change of the films after immersion in a liquid solvent mixture. This solvent mixture consisted of an active penetrant, dichloromethane, and an inactive solvent, hexane (24). Both solvents were from Labsynth Ltda, Brazil. This solvent mixture was chosen based on the PET phase behavior. As the mixture of the solvents is ideal, the molar fraction of dichloromethane is equal to its activity. The dichloromethane activity was 0.8. To confirm the absence of interactions between hexane and both PET and LCP, samples of both polymers were immersed for 20 days in pure hexane; no change in weight was observed.

The samples were immersed in the liquid solvent mixture and were placed in a controlled temperature bath, at 25[degrees]C. To prevent evaporation during the removal of the samples, a constant volume of the solvent mixture was maintained by adding more solvent from time to time. When the samples were removed from the solvent mixture, they were quickly wiped and weighted. The penetrant weight gain of the samples, [M.sub.t], was therefore measured over a period of 15 days. The equilibrium weight gain of the samples, [M.sub.eq], was taken as the value where no further change in weight gain was observed.

RESULTS AND DISCUSSION

DMTA Results

Figure 1 shows the dynamic mechanical behavior of the blends and of the pure components, and Fig. 2 shows the values of the thermal transitions as observed by DMTA.

From Fig. 1, it can be seen that the blends had only one [T.sub.g] as observed in Part I of this work (1) and also similar to those values. The presence of only one [T.sub.g] would suggest miscibility of the amorphous parts of the LCP and of the PET if the original [T.sub.g]s were far apart. However, as already explained in Part I of this work, owing to the closeness of the [T.sub.g]s, peak resolution is very difficult; therefore, we cannot conclude, based solely on the [T.sub.g] analysis, that miscibility occurred.

Figure 2a shows a plot of the [T.sub.g] of the blends as a function of the amount of LCP; it can be observed that the [T.sub.g] of the blends, as measured by DMTA, up to 60 wt% LCP, are higher than the ones predicted by the additivity rule. This behavior would suggest, in a miscible blend, that strong long-range interactions between the components have occurred; therefore, it can be concluded that the B20 and B40 blends are the strongest interacting blends. On the other hand, the B80 blend has very weak long-range interactions between the components.

Analysis of the [T.sub.[beta],LCP] transition could also help provide understanding of the nature of the interactions between the PET and the LCP. This transition represents the relaxation of the naphthyl groups of the HNA comonomer (9); its magnitude and value were observed (10) to increase with an increase in the number of HNA units in the HBA/HNA copolymers. It has also been observed (10) that the higher this transition for the pure LCP, the better its barrier properties.

In our blends, it can be observed that this transition was very broad; however, its value slightly increased with the increase in PET concentration, as observed in Fig. 2b; that is, the increasing concentration of the PET molecules in the blend slightly hindered the short range movements of the naphthyl groups of the LCP. Thus, short-range interactions between both polymers occurred. The B20 blend had the highest which indicates that in this blend, the naphthyl groups of the LCP were strongly hindered by the presence of the PET molecules.

If we also analyze the [T.sub.[beta].PET], we will observe that this transition is also broad, but it increased with the increase in LCP concentration, as also observed in Fig. 2b. We can attribute this transition to the local motions of the PET phenyl groups; therefore, the increasing concentration of the LCP in the blend strongly hindered the local motions of the PET phenyl groups. The B80 blend had the highest [T.sub.[beta],PET], which indicates that in this blend, the PET phenyl groups were strongly hindered by the presence of the LCP molecules.

It is known that in the blending of polyesters, an "in-situ" compatibilization between both polymers can occur as a result of transesterification reactions (25) with the formation of block, random or graft copolymers (26); therefore, we cannot attribute the increase of [T.sub.[beta].LCP] and solely to the occurrence of some strong purely physical short-range interactions between the respective groups; they can also be the result of transesterification reactions between both polymers.

It is also well known that the depression of the equilibrium melting point [T.sub.m][degrees] (27), due to the mixing of a crystalline polymer and an amorphous one, can be an indication of miscibility. Isothermal crystallization studies of these blends made in our labs (19) has shown that the [T.sub.m][degrees] of the PET, calculated by the standard Hoffmann and Weeks procedure (28) assuming that the thickening factor is constant, decreased very slightly with the increase in the LCP concentration, as shown in Table 1, except for the B80 blend, in which an increase was observed.

This same behavior has also been observed in other blends of thermoplastics with LCPs (29), while in others, no variation has been detected (30). Thus, no conclusion on miscibility can be drawn from the data of [T.sub.m][degrees].

Based on the DMTA analyses, however, we can conclude that strong long and short-range interactions occurred between the PET and the LCP; if transesterification or grafting reactions occurred, these reactions contributed to these strong interactions. The B20 and B40 blends had the strongest long-range interactions between the PET and the LCP of all the blends; the B80 blend, on the other hand, had the weakest long-range interactions of all the blends. In the B20 blend, the LCP segmental mobility was strongly hindered by the PET molecules, while in the B80 blend, the PET segmental mobility was strongly hindered by the LCP molecules.

MDSC Results

Figure 3 shows the differential scanning calorimetry of the blends, while Table 2 displays the results. Two types of analyses can be made: i) the influence of each component in the perfection of the crystallites of the other; and ii) the influence of each component in the degree of crystallinity of the other.

The presence of two and sometimes three melting endotherms in the PET after experiments of isothermal crystallization has been widely discussed (31). In this latter study, the first melting endotherm, which appeared at low crystallization temperatures, was associated with the last steps of secondary crystallization; the second endotherm was associated with most of the secondary crystallization process, originated from rejected species of the main crystals, and the third one was associated with the isothermally grown crystals. In our case, the blends and the pure components were non-isothermally crystallized within a very thin mold; the mold temperature was 25[degrees]C, so therefore, this crystallization occurred under a high degree of undercooling. This high degree of undercooling probably produced only two populations of PET crystallites: one, in small quantity, responsible for the first melting endotherm. at 172.2[degrees]C, with many defects, low melting temperature and formed by secondary crystallization : the other, in large quantity, responsible for the second melting endotherm, at 246.3[degrees]C, made of chain-folded, more perfect crystallites, with high melting temperature, and formed by primary crystallization.

The addition of the LCP to the PET changed the perfection of the first population of PET crystallites; in the B20 and B80 blends, these crystallites were more perfect. while in the B40 and B60 they were more imperfect. The B80 blend showed the most perfect first population of PET crystallites, even more perfect than in the pure PET. On the other hand, the increasing addition of the LCP to the PET diminished the perfection of the second population of PET crystallites: some LCP molecules could have been trapped intraspherutically into the PET spherulites. The B20 blend had the most perfect second population of PET crystallites, similar to the pure PET, while the B80 blend showed the most imperfect second population of PET crystallites.

The two melting endotherms of the LCP, at 190.6[degrees]C and 288.1[degrees]C, can be related to the fast and slow crystallization processes (32) that can occur in HBA/HNA copolymers. The fast process occurs when the polymer is cooled down rapidly from the nematic melt. As the cooling rate decreases, slow crystallization also occurs. These two processes have independent crystallization kinetics. Therefore, the first endotherm arises from the melting of a population of less perfect domains while the second one would arise from the melting of more perfect domains. The low values of the [DELTA]H indicate that the LCP crystallizes to a small degree. Analyzing the first population of LCP domains (first endotherm), we can observe that the addition of 20 wt% of PET to the LCP increased the perfection of these LCP domains in the B80 blend; however, in the other blends, this melting temperature was not detected. Regarding the second population of LCP domains (second endotherm), it is observed that the increasing addit ion of 20 and 60 wt% of PET to the LCP diminished the perfection of the LCP domains of this population, but the addition of 40 wt% of PET increased slightly the perfection of the LCP domains. The B60 blend had the most perfect LCP domains of all the blends.

Regarding the crystallinity of each component, it was observed that the increasing addition of LCP to the PET affected the PET total amount of crystallinity: in some blends, like the B20 and the B60, the amount of crystallinity of the PET increased, while in others, like the B80, the presence of the LCP did not affect the PET amount of crystallinity. In the B40 blend, this crystallinity decreased. Now, the increasing addition of the PET to the LCP decreased the amount of crystallinity of the LCP, that is, the presence of the PET molecules inhibited the formation of crystalline domains in the LCP. The B60 blend had the highest percent of PET crystallinity, while the B80 blend had the highest percent of LCP crystallinity.

The total amount of crystallinity in the blends ([[chi].sub.c,PET] + [[chi].sub.c,LCP]was higher in the blends, except for the B40 blend, The B60 blend had the highest total amount of crystallinity.

We can conclude from these measurements that in the blends, the B20 composition had the most perfect PET crystals, and in larger quantities, while the B60 composition had the most perfect LCP domains; on the other hand, the B60 blend had the highest percent of PET crystallinity, while the B80 blend had the highest percent of LCP crystallinity. The B60 blend had the highest total amount of crystallinity.

WAXD Results

Figure 4 shows the WAXD scans of the LCP in the plaque and powder forms, while Fig. 5 shows the scans of all the blends and the pure components, in the plaque form. From Fig. 4, it can be observed that the LCP in the powder form, has only one peak around 2[theta] = 20[degrees], while the LCP, in the plaque form, has two peaks, one around 2[theta] = 20[degrees] and the other around 2[theta] = 27.7[degrees]. On the other hand, from Fig. 5, it can be observed that the pure PET displayed only one peak, around 2[theta] = 26[degrees], while the blends displayed two peaks, one around 2[theta] = 20[degrees] and the other around 2[theta] = 26[degrees]. Table 3 shows the heights of these peaks in mm. The 26[degrees] peak can be attributed to the presence of the PET crystallites. The 27.7[degrees] peak is present only in the plaque, while the 20[degrees] peak is present in all the samples with LCP. If we assume that the LCP powder has no orientation, then the peak at 20[degrees], in the powder, represents only the cryst alline planes, its intensity increasing with the increase in orientation. If a Gaussian amorphous halo is drawn, the percent of crystallinity of the LCP in the powder can be calculated. A value of 19% was found. This value is low, but low values of percent of crystallinity are characteristic of LCPs, as our previous DSC measurements showed. As crystallinity and orientation cannot be separated by this method, we calculated a rough amount of LCP orientation or an approximated LCP "bulk orientation," B, in these blends, given by B = {[[h.sub.20]-12]/[[h.sub.20(plaque]) - 12]} x 100, where [h.sub.20] = height of the peak at 20[degrees] in the blends and [h.sub.20(plaque)] = height of the peak at 20[degrees] in the LCP plaque. It can be observed that the increasing addition of the PET to the LCP decreased the value of B in the blends, as already observed in blends of this particular LCP with PP (23). The B20 blend had the less oriented LCP domains, while the B80 had the most oriented LCP domains of all the blends.

Morphology Results

Figure 6 shows the SEM micrographs of the blends and the LCP. Because of the small thickness of the samples, the expected "skin-core" morphology was not obtained, as was observed in Part I of this work (1); instead a fibrillar morphology was observed through all the thickness. It was not possible to calculate the aspect ratio of the fibrils because of the heterogeneity of the lengths and diameters of the fibrils; however, we calculated a "mean" diameter d, based on five fibril-diameter measurements, as shown in Table 4.

It can be observed that the fibrils are much coarser in the blends than in the pure LCP. The B20 blend had the thickest fibrils, while the B80 blend had the thinnest LCP fibrils.

Transport Properties

Figures 7a, 7b and 7c show the so-called sorption reduced curves of the pure polymers and of the blends, where the ratio [M.sub.t]/[M.sub.eq] is represented as a function of [square root of (t)]. Before analyzing these results we will make some theoretical considerations on the transport properties through polymeric materials (33). The sigmoid and two-stage forms of the curves of Fig. 7 represent the typical anomalous or non-Fickian behavior of polymers at temperatures below their glass transition and of polymers with long relaxation times, especially with solvents that may swell them. Polymeric materials have a wide spectrum of relaxation times related to many relaxation modes. All relaxation times decrease with an increase in temperature or sorbed concentration due to enhanced segmental motion; however, some relaxation times will decrease more rapidly than others. The overall sorption process in a swelling medium will reflect those characteristic relaxation motions of the network chains that occur within a time scale comparable to or greater than the time scale of the concurrent diffusion process. Generally, in polymers at temperatures below [T.sub.g], deviations from Fickian behavior are considered to appear as a consequence of the finite rates by which changes in polymer structure occur in response to stresses imposed upon the medium before and during the sorption-diffusion process. Stresses and strains initially present in a polymer sample are the result of the processing history, crystallization conditions, mechanical deformation, etc. Therefore the anomalous sigmoid or two-stage character of the sorption curves becomes more prominent as the solvent power and concentration of solvent increase and as the crystallinity or density of the polymer also increases.

Recently, Ouyang and Shore (34) studied acetone transport in PET samples. Based on Harmon's results (35), which account for Case I, II, and mixed transport conditions, the authors showed that the mass transport in PET is accompanied by a large-scale structural rearrangement, which leads to the induced crystallization of the original amorphous phase. During this process, the matrix is under compressive stresses.

Taking into account the profile of the sorption curves of Fig. 7, and the theoretical considerations given above, our experimental results will be analyzed by using a generalized diffusion equation (36, 37), previously derived to describe the combination of Fickian and Case II mechanisms. In this case, the expression for diffusion in one dimension can be given by the following equation:

[partial]C/[partial]t = [partial]/[partial]x[D[partial]C/[partial]x - VC] (3)

where C = concentration of the penetrant,

D = diffusion coefficient

and V = velocity of solvent penetration.

The second term in the square brackets describes the contribution of internal stresses arising from the swelling of the polymer, and the quantity V represents the velocity of solvent penetration as a direct consequence of the internal stress effect. Both parameters D and V are assumed to be constant in Eq 3.

For diffusion in a semi-infinite polymer sample with a constant penetrant concentration [C.sub.0] at the surface, from Eq 3, the concentration profile of the penetrant in the swollen polymer can be calculated as:

C(x, t) = [C.sub.0]/2 {exp(xV/D) erfc

[(x + Vt)/2 [square root of (D)]t] + erfc[(x - Vt)/2 [square root of (D)t]} (4)

and the total amount of the penetrant per unit area entering the polymer sample, which is similar to [M.sub.t], will be:

[M.sub.t] [C.sub.0] {D/V erf[V/2 [(t/D).sup.1/2]] + V t/2 erfc

[-V/2 [(t/D).sup.1/2]] + [(Dt/[pi]).sup.1/2] exp (-[V.sup.2]t/4D)} (5)

which at large times reduces to:

[M.sub.t] = [C.sub.0] (D/V + Vt) (6)

It is possible to assume that the equilibrium concentration of the penetrant in the swollen polymer, [M.sub.eq], is similar to the surface concentration [C.sub.o] or [M.sub.eq] = [C.sub.o] L, where L is the sample thickness; then Eq 6 reduces to:

[M.sub.t]/[M.sub.eq] = l/L (D/V + Vt) (7)

Figure 8 shows the experimental results reported as [M.sub.t]/[M.sub.eq] as a function of t. With exception of sample B80, which will have its sorption curve, Fig. 7c, analyzed later, the relationship between [M.sub.t]/[M.sub.eq] and t of the pure PET and the B20, B40 and B60 blends is linear at large times (t> 25 hours), and is in agreement with Eq 7.

From the slope and intersection of this linear region, V and D were calculated; Fig. 9 shows these results as a function of the PET content. It can be observed that the parameter V decreases exponentially with the increase of the LCP content in the blends.

The LCP "bulk orientation," by WAXD, was found to increase with the increase of the amount of LCP in the blends. The B20 blends were also found to have the less-oriented LCP domains, while the B80 blend had the most-oriented LCP domains of all the blends. Thus, it is possible to postulate that the higher the bulk orientation, the lower and most hindered is the solvent penetration and consequently the internal stress effect.

If the increase in the LCP content hinders the swollen effect of the penetrant, it can be expected a monotonic decrease in the equilibrium bulk concentration of the solvent, [C.sub.eq], as the LCP content increases. This behavior was observed, and it is shown in Fig. 10 where the experimental results are reported as [C.sub.eq] vs. percent LCP.

Regarding the D parameter, it can be observed that the blends B20 through B60 had lower diffusion coefficients than the pure PET; these diffusion coefficients were very similar for the blends B20 and B40 and slightly higher for the B60 blend.

From MDSC and DMTA, we found that the blend B20 had the most perfect PET crystals and the strongest long-range and short-range interactions between PET and LCP. The B60 blend, on the other side, had the most perfect LCP domains and the highest percent of PET crystallinity. As a consequence, in the B20 blends, a more accentuated increase of the impermeable regions can be expected, which could explain the accentuated decrease in the parameter D with respect to the pure PET. According to the model proposed by Peterlin (38) for the diffusion in semicrystalline polymers, proportionality between the overall, D, and amorphous, Da, diffusivities can be represented by the equation:

D = ([PSI]/B) Da (8)

where [PSI] is a tortuosity factor describing the physical obstruction of the crystallites and B is a blocking factor related to amorphous immobilization or to restriction of chain mobility in the amorphous phase. in our case, m the blends with PET-rich phase, blends B20 and B40, the increase of LCP content up to 40% can be considered responsible for the increase of the B value, if it is assumed that this parameter is related to regions that contain LCP, PET spherulites and to regions where the strongest long-range and short-range interactions between PET and LCP predominate. As a consequence, the diffusion coefficient in these blends is lower than the diffusion coefficient of the pure PET samples.

On the other hand, regarding the bulk orientation and the fibril diameters of the LCP domains, we observed that the B60 blend had the most perfect LCP domains and, in comparison with the blends B20 and B40, the lowest fibril diameters. Thus, if it is assumed that the diffusion through the polymer plaques is only one-dimensional (normal to the plaque), this means that the more perfect. thinner, and oriented LCP fibrils of the B60 blend formed a rigid region, which increased the B value in Eq 8. The B60 blend also had the highest percent of PET crystallinity, which gives an increase of the [PSI] value of Eq 8. Therefore, the slightly higher value of the observed diffusion coefficient in the B60 blend could be explained by the fact that for this blend, the ratio [PSI]/B is higher than unity, i.e., the physical obstruction of the crystallites predominates over the tortuous path formed by the oriented LCP fibrils.

In the case of the B80 blend, Its sorption curve shows a typical two-stage behavior. In this case the diffusion may be considered to occur through a LCP matrix having PET as a disperse phase. Also, as can be seen in Fig. 10, the sorption in the B80 blend is very low, and practically null for the LCP, so this means that the sorption occurs only in the PET disperse phase. As a consequence, the amorphous regions of the PET disperse phase could swell. The change in the chemical potential of sorbed solvent during sorption may be separated into free changes because of the restricted swelling against internal cohesive forces and mixing with the system fully relaxed. At the state of quasi-equilibrium (see first plateau in Fig. 7c at [M.sub.t]/[M.sub.eq] [congruent to] 0.58). a temporary balance is achieved between osmotic forces, tending to swell the PET disperse phase and elastic forces opposing this expansion. These phenomena are time-dependent, and thus a second plateau is observed. Unlike the pure PET samples, in this case, the presence of LCP changed the smaller relaxation times to higher values. Therefore, it is not possible to calculate a diffusion coefficient for this blend. It could be postulated that its value through the PET relaxed regions is the same as the calculated value for the pure PET samples.

CONCLUSIONS

From all these studies we can preliminarily conclude the following:

1. By DMTA, only one [T.sub.g] was observed; however, up to 60 wt% of LOP the [T.sub.g]s of the blends were much higher than the ones predicted by the additivity rule, suggesting that strong long-range interactions between the PET and the LCP occurred. The analysis of the sub-transitions and [T.sub.[beta],PET] showed that both sub-transitions were higher in the blends than in the pure components; thus, hindrance of the local motions of the naphthyl groups in the LCP and of the phenyl groups in the PET probably occurred. All these results were taken as evidence of the probable existence of strong long-range and short-range interactions between the LCP and PET functional groups or, if it occurred, to the existence of transesterifications or grafting reactions between both components. The B20 and B40 compositions had the strongest long-range interaction between the PET and LCP; in the B20 blend, the LCP naphthyl groups were strongly hindered by the PET, while in the B80 blend, the PET phenyl groups were strongly hindered by the LCP molecules.

2. The B20 blend had the most perfect PET crystals, while the B60 blend had the most perfect LCP domains; the B60 blend also had the highest percent of PET crystallinity, while the B80 had the highest percent of LCP crystallinity.

3. The LCP "bulk orientation" was found to increase with the increase of the amount of LCP in the blends. The B20 blend was found to have the less-oriented LCP domains while the B80 blend had the most-oriented LCP domains of all the blends.

4. No "skin-core" morphology was observed; instead a fibrillar one was obtained through all the thickness. The fibril diameter, in the blends, decreased with the increase in LCP.

5. The sorption curves of the pure PET and of the blends had sigmoid and two-stage forms, showing the non-Fickian character of the diffusion through these materials. All the blends had lower diffusivity coefficients than the pure PET; the B20 and B40 blends had the lowest diffusivity coefficient of all, while the B60 had a slightly higher diffusivity coefficient than the B20 and B40 compositions. The B80 blend displayed a two-stage sorption curve; thus it was not possible to calculate its diffusivity coefficient.

6. The low D values of the B20 and B40 blends were attributed to the occurrence of strong long-range and short-range interactions between the PET and LCP in these blends and to the perfection of the PET crystals of the B20 blend. The slightly higher D value of the B60 blend was attributed to the formation of more perfect, thinner, and oriented LCP fibrils, which contribute differently to the tortuosity and blocking factors of the Peterlin equation.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]
Table 1

PET Equilibrium Melting Point, [T.sub.m][degrees], as Measured by DSC
(19).

 [T.sub.m][degrees]
Sample ([degrees]C)

 PET 261.12
 B20 259.98
 B40 259.64
 B60 255.45
 B80 268.51
Table 2

Thermal Analysis of the Blends by MDSC.


 [([T.sub.m]).sub.1,PET] [([T.sub.m]).sub.2,PET]
Sample ([degrees]C) ([degrees]C)

PET 172.2 246.3
B20 177.4 246.0
B40 171.5 244.2
B60 165.0 243.9
B80 179.2 242.3
LCP -- --


 [([T.sub.m]).sub.1,LCP] [([T.sub.m]).sub.2,LCP]
Sample ([degrees]C) ([degrees]C)

PET -- --
B20 -- --
B40 -- 280.5
B60 -- 290.5
B80 195.5 284.2
LCP 190.6 288.1


 [([DELTA][H.sub.1]).sub.PET] [([DELTA][H.sub.2]).sub.PET]
Sample (J/g) (J/g)

PET 5.5 34.1
B20 5.3 29.1
B40 3.9 14.6
B60 2.7 16.7
B80 0.6 7.3
LCP -- --


 [([DELTA][H.sub.1]).sub.LCP] [([DELTA][H.sub.2]).sub.LCP]
Sample (J/g) (J/g)

PET -- --
B20 -- --
B40 -- 0.08
B60 -- 0.66
B80 0.09 1.14
LCP 0.8 1.6

 [[chi].sub.c,PET] +
 [[chi].sub.c,PET] [[chi].sub.c,LCP] [[chi].sub.c,LCP]
Sample (%) (%) (%)

PET 27.5 -- 27.5
B20 29.9 -- 29.9
B40 21.4 0.7 22.1
B60 33.6 3.9 37.5
B80 27.4 5.5 32.9
LCP -- 8.6 8.6

[([T.sub.m]).sub.1,PET] = PET first melting peak

[([T.sub.m]).sub.2,PET] = PET second melting peak

[([T.sub.m]).sub.1,LCP] = LCP first melting peak

[([T.sub.m]).sub.2,LCP] = LCP second melting peak.
Table 3

Heights of the Peaks, in mm, and Approximated LCP "Bulk Orientation," B.

 Sample [h.sub.20] [h.sub.26] [h.sub.27.7] B (%)

LCP (powder) 12 -- -- 0
LCP (plaque) 152 -- 22 100
B80 70 20 -- 41.4
B60 33 13 -- 15.0
B40 16 10 -- 2.8
B20 8 4 -- 0
PET (plaque) 0 7 -- 0
Table 4

Mean Diameter d of the Fibrils in the Blends as Calculated by SEM.

Sample d ([micro]m)

 PET --
 B20 2.22 (0.00)
 B40 1.85 (0.34)
 B60 1.55 (0.44)
 B80 1.03 (0.13)
 LCP 0.059 (0.012)

The number in parentheses represents the standard deviation.


ACKNOWLEDGMENTS

The authors wish to express their gratitude to PRONEX-FINEP and FAPESP for the financial aid and to Rhodia Ster for the PET samples.

REFERENCES

(1.) L. B. da Silva and R. E. S Bretas, Polym. Eng. Sci., 40, 1414 (2000).

(2.) A. Ruvolo-Filho and G. M. de Carvalho, J. Macrom. Sci.-Phys., B35, 255 (1996).

(3.) L. Bove, C. D'Aniello, G. Gorrasi, L. Guadagno, and V. Vittoria, J. Appl. Polym. Sci., 62, 1035 (1996).

(4.) G. E. Serad, B. D. Freeman, M. E. Stewart, and A. J. Hill, Polymer, 42, 6929 (2001).

(5.) H. Ouyang, C. C. Chen, S. Lee, and H. Yang, J. Polym. Sci. Part B: Polym. Phys., 36, 163 (1998).

(6.) A. Natu, E. A. Lofgren, W. Kollen, and S. A. Jabarin, SPE ANTEC '98, 1642 (1998).

(7.) J. S. Chiou and D. R. Paul, J. Polym. Sci.: Part B: Polym. Phys., 25, 1699 (1987)

(8.) D. H. Weinkauf and D. R. Paul, Barrier Polymers and Structures, Chap. 3, p. 60-91, W. I. Koros, ed., American Chemical Society, Washington, D.C. (1990).

(9.) D. H. Weinkauf and D. R. Paul, J. Polym. Sci: Part B: Polym. Phys., 29, 329 (1991).

(10.) D. H. Weinkauf and D. R. Paul, J. Polym. Sci.: Part B: Polym. Phys., 30, 837 (1992).

(11.) F. de Candia, A. Renzulli, V. Vittoria, A. Roviello, and A. Sirigu, J. Polym. Sci. Phys., 28, 203 (1990).

(12.) F. de Candia, V. Capodanno, A. Renzulli, and V. Vittoria, J. Appl. Polym. Sci., 42, 2959 (1991).

(13.) P. M. Subramanian, Polym. Eng. Sci., 25, 483 (1985).

(14.) R. Gopalakrishnan, J. M. Schultz, and R. M. Gohil, J. Appl. Polym. Sci., 56, 1749 (1995).

(15.) O. Motta, L. di Malo. L. Incarnato, and D. Acierno, Polymer. 37, 12, 2373 (1996).

(16.) C. Uriarte, J. A. Alfageme, and J. J. Iruin, Eur. Polym.. J., 34, 10, 1405 (1998).

(17.) J. A. Alfageme, M. Iriarte, J. J. Iruin. A. Etxeberria, and C. Urlarte, J. Appl. Polym. Sci., 73, 323 (1999).

(18.) G. Flodberg, A. Heilman, M. S. Hedenqvist, E. R. Sadiku, and U. W. Gedde, Polym. Eng. Sci., 40. 1969 (2000).

(19.) H. B. de Araujo and R. E. S. Bretas, personal notes, (2000).

(20.) S. C. George and S. Thomas, Prog. Polym. Sci., 26, 985 (2001).

(21.) J. Brandrup and E. H. Immergret. Handbook of Polymers, 2nd ed; John Wiley and Sons. New York, NY. p. V-73 (1974).

(22.) A. Biswas, J. Polym. Sci.: Part B: Polym. Phys., 30. 1375 (1992).

(23.) R. E. S. Bretas, G. Cardoso, and A. Kaito, Proc. 15 Annual Meeting of the Polymer Processing Society, PPS-15, Hertogenbosch. The Netherlands, May (1999).

(24.) A. Ruvolo-Filho and G. A. da Silva, J. Macrom. Sci-Phys., B37, 349 (1998).

(25.) K. H. Wei and J. C. Ho, Macromolecules, 30, 1587 (1997).

(26.) R. S. Porter and L. H. Wang. Polymer, 33. 10, 2019, 1992.

(27.) T. Nishi and T. T. Wang, Macromolecules, 8, 2, 227 (1975)

(28.) J. D. Hoffman and J. J. Weeks, J. Res. A. Phys. Chem, 66A, 1, 13 (1962).

(29.) B. de Carvalho and R. E. S. Bretas, J. Appl, Polym. Sci., 55, 233 (1995).

(30.) G. Gabellini and R. E. S. Bretas, J. Appl. Polym. Sci., 61, 1803 (1996).

(31.) F. J. Medellin-Rodriguez, P. J. Phillips, J. S. Lin, and R. Campos. J. Polym. Sci. B: Polym. Phys., 35, 1757 (1997).

(32.) H. C. Langelaan and A. Postherma de Boer, Polymer, 37, 25, 5667 (1996).

(33.) C. E. Rogers, Physics and Chemistry of the Organic Solid State, Vol. 2, Chap. 6, pp. 509-635. D. Fox, M. M. Labes, and A. Weissberg. Eds., Interscience Pub., New York (1965).

(34.) H. Ouyang and S. H. Shore. Polymer, 40, 5401 (1999).

(35.) J. P. Harmon, S. Lee, and J. C. M. Li, Polygm. Sci: Part A: Polym. Chem., 25, 3215 (1985).

(36.) H. L. Frish. T. T. Wang. and T. K. Kwei, J. Polym. Sci. Part A-2, 7, 879 (1969).

(37.) T. T. Wang, T. K. Kwei, and H. L. Frish, J. Polym. Sci., Part A-2, 7, 2019 (1969).

(38.) A. Peterlin, J. Macromol. Sci.-Phys., B28, 57 (1975).

L.B. DA SILVA (1)

A.L. MARINELLI (1)

A. RUVOLO-FILHO (2)

R.E.S. BRETAS (1) *

* To whom correspondence should be addressed: bretas@power.ufscax.br

(1.) Department of Materials Engineering

(2.) Department of Chemistry Universidade Federal de Sao Carlos 13565-905 Sao Carlos, SP, Brasil
COPYRIGHT 2002 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Silva, L.B. Da; Marinelli, A.L.; Ruvolo-Fileho, A.; Bretas, R.E.S.
Publication:Polymer Engineering and Science
Geographic Code:1USA
Date:Aug 1, 2002
Words:7969
Previous Article:Structure and mechanical properties of blends of two thormotropic copolyesters.
Next Article:Formation of a polyetheramide triblock copolymer by reactive extrusion; process and properties.
Topics:


Related Articles
Weld lines and mechanical properties of injection molded polyethylene/polystyrene/copolymer blends.
Morphology and mechanical properties of liquid crystalline copolyester and polyester elastomer blends.
Transesterifications in a polyblend of poly(butylene terephthalate) and a liquid crystalline polyester.
Solid state and mechanical properties of injection molded poly(ether ether ketone)/polyarylate blends.
Thermally Stimulated Depolarization Currents in Poly (Ethyleneterephthalate-ran-p-Hydroxybenzoates.
Effect of thermotropic copolyesteramide on the properties of polyamide-66/liquid crystalline copolyester composites.
Effect of a compatibilizer on the structural development of a thermotropic liquid crystalline polymer/polystyrene blend.
Compatibilizers for thermotropic liquid crystalline polymer/polyethylene blends prepared by reactive mixing.
A lithium lonomer of poly(ethylene-co-methacrylic acid) copolymer as compatibilizer for blends of poly(ethylene terephthalate) and high density...
Hot-fillable containers containing PET/PEN copolymers and blends.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |