Printer Friendly

Structure formation in the cure of low-shrinkage unsaturated polyesters.

Structure Formation in the Cure of Low-Shrinkage Unsaturated Polyesters

Unsaturated polyester resins account for about 60% of the thermoset polymers used in the Unied States and are the most widely used thermosets in polymer composites. They offer a good balance of properties, are relatively inexpensive, and are amenable to most fabrication processes. With the addition of fillers, reinforcements, and other additives, they are available in the form of compounds for compression molding (sheet molding compounds), injection molding (bulk molding compounds), resin transfer molding, and pultrusion.

When traditional unsaturated polyester resins were first used in fiber-reinforced materials, their applications were limited by a number of problems, the most important being poor surface quality. The fibers often showed through, necessitating costly prepaint operations, such as dry sanding of the surface. Other problems included warpage or "sinks" on the surface of molded parts and inability to mold to exact dimensions. These problems were attributed to the high polymerization shrinkage during cure. As the polyester crosslinks with styrene, it shrinks away from the mold walls, creating internal stresses that lead to surface and structural flaws.

The addition of certain thermoplastic materials to the polyester resin was found to compensate for and even eliminate shrinkage, thereby solving most of the problems. These thermoplastic materials are called "low-shrink" or "low-profile" additives (LPA). Commercial PLAs include vinyl acetate polymers, acrylic polymers, polystyrene, and polycaprolactone. Although some research has been conducted on the function of LPA, the exact mechanism of LPA efficacy remains obscure. Most commercial LPAs are developed empirically, and they still cannot solve all surface-related problems.

The reaction of unsaturated polyester resin and styrene is a free-radical, chain-growth, crosslinking copolymerization. It includes three major reactions: styrene-polyester vinylene, styrene-styrene, and polyester vinylene-polyester vinylene. The polyester/styrene mixture may be pictured as a system of many coiled polyester chains swelled in styrene. The coil size depends on molecular weight, chain stiffness, the thermodynamic compatibility, and chain concentration. Chemical reaction may occur inside, outside, and at the surface of the coils.

During the styrene-unsaturated polyester copolymerization, decomposition of the initiator creates free radicals that grow and form long-chain molecules by connecting styrene monomers and polyester molecules in both inter-and intramolecular reactions. These long-chain molecules tend to form spherical-type structures because of the intramolecular crosslinking between the pendant C=C bonds of the polyester molecules. Dusek first pointed out such structures in the products of monovinyl-divinyl reactions at medium or high divinyl monomer concentration and called these spheres "microgel particles." Primary microgel particles were shown experimentally to form at a very early stage of the reaction.

The interparticle crosslinking, forming a macroscopic network structure, may occur through C=C bonds at or near the surface of the microgel, with styrene monomers serving as chain extenders. Depending upon the concentrations of styrene monomers and microgels, the morphologies of the reacted resins can be quite different. At high styrene concentration, i.e., low microgel concentration, the individual microgel particle can be easily identified, and the size is quite large because of the styrene swelling effect. The particles are externally connected to form a "dumbbell" shape with particles at the ends and styrene chains in the middle. The overall network built up by the dumbbell shapes is a "tree-like" or "coral-like" structure.

At low styrene concentration, the unsaturated polyester resins tend to entangle and overlap with each other, so that no individual particle or dumbbel shape can be observed. The structure is "flake" like.

Microgel formation is an intrinsic feature of unsaturated polyester resins. The presence of LPA tends to enhance the formation of individual microgel particles in the LPA-rich region. In this article, integrated rheokinetic-morphological measurements on a low-shrinkage unsaturated polyester resin are described. The mechanism of microstructure formation during reaction and its effects on reaction kinetics and rheological changes are discussed.

Materials

The unsaturated polyester resin used was a 1:1 mixture of maleic anhydride and propylene glycol containing 35% by weight styrene with an average of 10.13 vinylene groups per molecule. The lowfrofile additive was polyvinyl acetate (PVAc), 40% by weight in styrene. All polyester resins were used as received, without removing the inhibitor. Methyl ethyl ketone peroxide (MEKP) with 25% cobalt naphthanate promoter (COB) was used as the initiator for reaction temperatures from 35[degrees]C to 55[degrees]C, and t/butyl peroxy-2-ethyl hexanoate (PDO) was used for reaction temperatures from 75[degrees]C to 85[degrees]C. Initiator concentration was 1.1% by weight of total resin for all samples. PDO and MEKP were added to a preweighed mixture of polyester resin, LPA, and styrene monomer, and after stirring, the mixture was stored below 5[degrees]C for future use. COB was mixed into the solution at room temperature just before the kinetic and rheological experiments.

Instrumentation and

Procedures

A Rheometrics Dynamic Analyzer was used in the oscillatory mode to measure the storage shear modulus, G', and the loss shear modulus, G", during reaction. Samples were held in serrated, 25-mm-diameter disposable aluminum parallel plates. The gap between the two plates was set at 1.1 mm and the frequency was 10 rad/sec for all experiments. Strain ratio was set at 10% at low moduli, and, in order to measure the moduli at high conversions, it was changed to 1% after [G'>10.sup.6] [dynes/cm.sup.2].

Reaction kinetics were measured by differential scanning calorimetry (DSC.) Reactions were conducted in volatile aluminum sample pans capable of withstanding at least 2-atm internal pressure after sealing. Sample weights ranged from 10 to 20 mg with an empty pan as reference. Isothermal runs were ended when there was no further exotherm. Reactions were also carried out in the scanning mode from 320K to 520K at a heating rate of 10[degrees]C/min. the total heat of reaction was determined from the area under the scanning curve.

For the morphological study, the resin was moolded into a disk shape in a sealed glass dish set in a constant temperature water bath. Reaction was stopped at a preset time by removing the glass dish from the water bath and quenching with ice. One piece of the cooled sample was tested by DSC in the scanning mode to determine the residue reactivity. Other pieces were etched in dichloromethane--a good solvent for the reactants--with 1% benzoquinone for 10 hrs to dissolve any soluble materials. The undissolved samples were then dried and gold-coated. The fracture surfaces were viewed in a scanning electron microscope (SEM) under magnifications ranging from 800X to 4000X.

The change in opacity of the resin during reaction was monitored by detecting the light transmission through a thin sample bounded by two pieces of cover glass. The arrangement was mounted on a sample holder located in a controlled-temperature chamber. The amount of light transmitted was sensed by a photo-dial.

Conversion and Structure

Formation

The conversion profile, rheological changes, and change of sample opacity during a reaction of the low-shrinkage unsaturated polyester resin at 80[degrees]C are shown in Fig. 1. The conversion profile indicates that significant reaction did not occur until 30 min at the given conditions. Rheological measurements, however, show a crossover of the G' and G" curves at about 10 min. The resin conversion at this point was less than 1% (table). Many researchers have shown that the crossover of G' and G" is a good indication of gel point. Both G' and G" reach their plateau values at around 30 min--at less than 5% conversion. The shear moduli remained fairly constant during the rest of the reation, although DSC showed that the reaction progressed to greater than 80% conversion. The transmitted light intensity indicates that the resin turned from transparent to opaque very early in the reaction. The change of opacity follows that of the shear moduli. These results suggest that structure formation occurs at the very beginning of the reaction for low-shrinkag unsaturated polyester resins.

To verify this typothesis, SEM micrographs (Fig. 2) were taken of samples cured to different conversions. Sample 2a--1% conversion--was a soft and translucent solid, while samples 2b, 2c, and 2d--5% to 10%, 60%, and 82% conversion, respectively--were rigid and opaque solids. Solvent etching dissolved 10% to 15% of sample 2a and slightly changed its shape. Samples 2b, 2c, and 2d were relatively inert to solvent etching, with only slight weight loss and little shape change. All the micrographs show small, compact globules, 0.5- to 1.0-[micrometer] in diameter, throughout the sample surfaces. The only obvious difference in surface morphology is that the globules are more highly packed in samples taken at h igher conversions. Apparently, the microstructure of this compound was formed at a conversion much below 1%.

The same compound near the gel point, i.e., crossover of the G' and G" curves at 0.25% conversion, appeared as a nonuniform, translucent, viscous fluid or soft gel. Solvent etching dissolved nearly 80% of the sample. From the SEM micrographs of different locations on the remaining thin film, (Fig. 3), it is clear that the remaining materials formed a loose network structure that was swollen by the solvent. Globules were not completely formed in Figs. 3a-3c but are easily identified in Fig. 3d. These micrographs suggest that polymer chains, with or without globule formation, are chemically linked to form a network at very low conversions.

LPA and Structure

Formation

Figure 4 shows SEM micrographs of the same resin without any LPA at different conversions: Figs. 4a and 4b from samples cured before the gel point at 0.1% and 0.23% conversion, respectively; and Fig. 4c from a sample cured to 52% conversion. Solvent etching dissolved most of samples 4a and 4b, but had little effect on sample 4c. The micrographs of the remaining material show that samples 4a and 4b formed a loose network without any globule structure, and sample 4c formed a flake-type structure. Comparison of Figs. 3 and 4 suggest that the presence of LPA enhances the formation of globules--an important feature for the cure of low-shrinkage polyester resins.

The effect of LPA concentration on sample morphology is confirmed in Fig. 5. All samples were reacted at 80[degrees]C for 3 hrs. The cured resin with 1% PVAc is transparent and shows a flake-type structure. Its surface is not as smooth as that of the sample without LPA. Apparently, there was not enough LPA for phase separation. At 2% PVAc addition, the cured sample became translucent, and the etched surface showed a heterogeneous structure. At 3% PVAc addition, very small globules (0.1-[micrometer] diameter) started to form, which appear to be fused together. At 15% PVAc concentration, individual globules of much larger size are easily seen. Samples 5c and 5d are both opaque after being cured.

LPA, Reaction Kinetics,

and Rheology

Although the presence of LPA greatly affects the microstructure of unsaturated polysters, its influence on reaction kinetics and rheological changes appears to be minor. Comparison of conversion profiles at 80[degrees]C of resins with and without LPA (Fig. 6) shows the resin with LPA to have a slightly lower reaction rate--expected because of the dilution effect of the LPA. Changes of shear module (Fig. 6) during reaction are similar for both resins, except for the higher initial G" of the resin without LPA. For both resins, the major rheological changes occurred in the first 30 min, during which resin conversion was less than 5%.

Comparison of the degree of conversion at G' = G" for the various formulations, given in the Table, shows that for all reactions, the presence of LPA slightly delayed the gelation. Reaction temperature and initiator did have a profound effect on gel conversion. At the higher temperatures and with PDO as the initiator, gel conversion was always below 1%. A similar result has been reported at 90[degrees]C to 130[degrees]C with both TBP and PDO as initiators. Increasing reaction temperature tends to increase the gel conversion.

At the lower temperatures and with MEKP/COB as the initiator, gel conversion is higher than 1%. Similar results have been reported by others. Increasing reaction temperature tends to decrease the gel conversion. Because of differences in initiator reactivity and experimental limitations, it is difficult to carry out reactions covering a broad temperature range using the same initiator. Therefore, the relative importance of initiator type and reaction temperature cannot be verified from the data given in the table.

A possible explanation for the observed difference in the gel conversion is described schematically in Fig. 7. (The "=" symbol stands for styrene.) At higher temperature, unsaturated polyester resin is more compatible with styrene and occupies a l arger space in the styrene solution. This results in a shorter distance, or an overlap and chain entanglement, between adjacent polyester molecules. A small number of intermolecular reactions are capable of linking many polyester chains and forming a macrogelation. At low temperatures, the coil size of the polyester molecules in styrene becomes smaller, which requires more intermolecular reactions to form the macro-gelation.

The final structure of the resin is determined very early in the reaction. After that, further reaction is predominantly intramolecular cyclization, which increases the overall conversion and the local crosslinking density, but, as indicated by the G' and G" curves, does not seem to much enhance the mechanical properties of the whole sample.

Th relatively small differences in reaction kinetics, rheological changes, and gel conversion for resins with and without LPA is unexpected. As shown in the SEM micrographs, their morphologies are very different. Apparently, the formation of globules does not affect the reaction mechanism very much, although it provides a means for the reduction of polymerization shrinkage.
COPYRIGHT 1989 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Hsu, C.P.; Lee, L. James
Publication:Plastics Engineering
Date:Nov 1, 1989
Words:2267
Previous Article:Interlocking additive effects forcing total-system approach.
Next Article:Plastics East showcases new products and services.
Topics:

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters