Low-profile mechanisms in composites with thermoplastic/polyester blend matrices
Unsaturated Polyester (UP) resins are extensively used as a matrix for the production of composites due to their low cost and versatility. Their significant polymerization shrinkage has however a negative impact on the dimensional stability of the molded parts and on their surface quality. Low-Profile Additives (LPAs), that are thermoplastics, are usually added to the formulation of these materials to reduce shrinkage during curing. However, the exact role of the LPA is still controversial. The goal of this thesis is to gain a better understanding of the mechanism of shrinkage control, also called Low-Profile Effect. The approach used in the study consists of correlating the different stages observed during curing of the system UP/LPA/Styrene to the low-profile effect. The results obtained with the blend are compared to those obtained with the neat UP resin in order to highlight the role of the additive. Two types of system are investigated: incompatible, based on Poly(dimethyl siloxane) and Poly(methyl methacrylate), and compatible, based on Poly (vinyl acetate). Emphasis is placed on the blend containing PVAc. The study considers the changes taking place in the systems in the bulk, in the absence of inorganic materials, and at the surface of glass fibers. It has been shown that phase separation, which occurs during the crosslinking reaction, plays a determinant role in the mechanism of shrinkage control. Phase separation leads to the formation of two phases, a polyester-rich phase and an LPA-rich phase. Gelation that occurs early in these systems locks the morphology. Depending on the type of additive, its concentration and the cure temperature, different kinds of morphology have been observed. The low-profile effect is intimately linked to the morphology composed of interconnected globules of polyester. The formation of this particular structure strongly modifies the cure kinetics and the rheokinetics of the blends, compared to those of the neat UP resin. The most significant effects that have been observed are an increase in the gel conversion and in the gel time, and a slow down in the cure kinetics. An increase of the gel time up to 50% has been reported in the presence of PVAc depending on the cure temperature. Work was then carried out in order to check whether the surface treatment of the glass fibers could modify the adjacent morphology of the two-phase matrix. The surface treatment showed a spectacular influence under specific conditions only, i.e. for a low temperature and a low concentration in LPA. The conditions leading to the formation of interconnected globules of polyester are, however, different and in this case, the morphology remains unchanged at the surface of the fibers. It is to be noted that in these reactive systems, the morphology at the surface of the fiber can be correlated to the surface tension of the phases in the case of incompatible additives only. Regarding the low-profile effect, the role of glass fibers is thus not so effective on the morphology. On the other hand, it has been shown that depending on the composition of their sizing, glass fibers can strongly affect the cure kinetics. By the end of the study, it was possible to identify the main points associated with the mechanism of shrinkage control. This mechanism is based on an increase in the gel conversion in presence of an LPA, and expansion of the LPA at the early stages of curing. At the later stages of curing, the mechanism is based on the formation of microvoids that develop in the LPA-rich phase and compensate for further shrinkage. Until now, the low-profile effect has been almost entirely attributed to the formation of these microvoids. Furthermore, it was shown that microvoiding could be associated with the formation of macroscopic voids that can be present at the surface of the parts in the form of pinholes. The entire course of reaction has thus to be considered to explain the low-profile effect.
EPFL_TH3077.pdf
restricted
11.82 MB
Adobe PDF
7d0b8b0fd5c1cf8e613a366ba7019cf2