Inclusion of Inverse Terms in Design of Experiments for Mixtures Applied to Wood Plastic Composites

Abstract

Wood–plastic composites (WPCs) are materials that combine a polymeric phase with a filler, typically wood powder or fiber. This common practice significantly enhances certain mechanical properties of WPCs, such as tensile, compressive, and flexural strength, especially when coupling agents are added. Polyethylene terephthalate (PET) is a thermoplastic commonly used due to its high stiffness, mechanical strength, and chemical resistance. However, the increasing demand for PET has led to a rise in plastic waste accumulation, posing an environmental challenge. Therefore, recycling and promoting a circular economy are considered key strategies to reduce environmental impact and achieve sustainable development. The main challenge in WPC production lies in its formulation—specifically, determining the appropriate proportion of thermoplastic polymer, wood filler, and coupling agent. To optimize the mechanical properties of WPCs, it is necessary to vary these component proportions and understand how these changes affect the composite properties. This article describes a statistical approach for optimizing the mechanical properties of a WPC with PET as the polymer matrix. It is assumed that the mechanical properties of the WPC are a function of the proportions of the components that make up the mixture, so the optimization consists of determining these proportions in a way that maximizes the mechanical properties. With the experimental design, three models with inverse terms were fitted, which modeled the relationship between the component proportions and the flexural, compression, and tensile strength. The optimal formulation for maximizing flexural strength was found to be 58% PET, 39% wood, and 3% E-GMA. The optimal formulation for maximizing compressive strength was found to be 58% PET, 40% wood, and 2% E-GMA. The optimal formulation for maximizing tensile strength was found to be 86% PET, 11% wood, and 3% E-GMA. The results are validated through analysis of variance of the mixture experimental design for mixtures. Finally, three response surface plots are presented for each mechanical property, which graphically describe the relationship between the different proportions of the mixture components and the mechanical properties.