This Small Business Innovation Research (SBIR) Phase I project aims to develop software for an accurate simulation of polymer co-extrusion. During co-extrusion, different polymers are extruded simultaneously through a single die. Coextruded plastic parts combine the functionalities and benefits of several polymers into a single multi-functional product. Since the polymer layers get redistributed as they flow through a die, control of polymer distribution in coextruded products is difficult. Even though use of software is critical for optimizing the process, die designers in plastic industry rarely use commercial co-extrusion software because these packages cannot simulate multilayer flow in complicated co-extrusion die geometries, and also fail to capture complex material behavior of polymers. A unique algorithm, called mesh partitioning technique, which will be employed in the new co-extrusion software, will allow simulation of any complex co-extrusion system. A viscoelastic equation will be used to simulate co-extrusion. Incorporation of viscoelastic effects in co-extrusion simulation is important to accurately capture the rearrangement of polymer layers as different polymers flow together in a co-extrusion die. Accuracy of the software will be verified using the experimental data in the literature. New experiments will be conducted to identify the mechanism behind polymer layer rearrangement during co-extrusion.
The commercial potential of this project is complete elimination of the trial-and-error approach currently being used to design co-extrusion systems, which will cut the time-to-market of coextruded products by over 30%. Many different types of companies in plastic industry, including material suppliers, plastic part manufacturers and die manufacturers, will be able to cut cost and increase revenues using this software. The software, which will be developed in this project, will also enhance the scientific understanding of the root cause behind various complexities, such as encapsulation of higher viscosity polymer by less viscous polymer, and instability of the interface between adjacent polymer layers during co-extrusion. Even though such complexities are commonly observed in co-extrusion, the driving mechanisms behind these complex phenomena are still not understood completely.
Coextrusion, which involves simultaneous extrusion of several different polymers through a die to form a single multi-layered product, combines the functionalities and benefits of several polymers into a single product. Depending upon the rheology of polymers used for coextrusion, the polymers in various layers may get redistributed as they flow through the die such that the distribution of various polymers at the inlet and at the exit of the die may be quite different. Because of this redistribution of polymer layers, finding a die geometry which will give the required layer distribution in the final product can be extremely difficult. Therefore, a coextrusion simulation software, which predicts the shape of the interface in the die, can be an excellent design aide for a coextrusion die designer. Flow simulations based upon a generalized Newtonian constitutive theory, which does not include viscoelastic effects, do not capture the coextrusion layer rearrangement due to secondary flows caused by the second normal stress difference in viscoelastic fluids. Also, the coextrusion simulations available in the literature using purely viscous formulation of the flow, have failed to capture the encapsulation of a polymer by another polymer, which is often observed in coextrusion. Therefore, the goal of this project was to include the viscoelastic effects in a coextrusion simulation. The proposed coextrusion software including viscoelastic effects was successfully developed in the SBIR Phase I project. To capture the viscoelastic behavior of polymers during coextrusion, the Giesekus model was employed in the current work. Besides the Giesekus constitutive equation, the mass and momentum conservation equations were solved using the finite element method. Tetrahedral finite elements were used for the coextrusion simulation. The mesh of tetrahedral finite elements was not modified or regenerated during coextrusion simulation. Instead, the interface between adjacent layers of different polymers was represented by a surface mesh of triangular finite elements. The tetrahedral elements which were intersected by the mesh of triangular elements on the interface were partitioned into two tetrahedral, prismatic, or pyramidal finite elements. The software developed in this project used to simulate viscoelastic flow of two polystyrenes (Styron 472 and Styron 678) during coextrusion in a square channel. Experimental data available in the literature shows that starting with a straight interface shape at the contact line, where the two polymers meet for the first time, Styron 678 progressively encapsulated Styron 472 as the two polymers flowed side-by-side in the square channel. In the present work, no such encapsulation of Styron 472 by Styron 678 was observed in the simulation using the viscoelastic formulation. In the viscoelastic simulation of the bi-layer flow, starting with a straight line interface shape at the contact line, where the two polymers meet for the first time, as shown in Figure 1, the interface started to wave upwards near the middle and near the two ends, with troughs in between, resulting in the W shaped interface at the die exit. This wavy shape of the interface is caused by the secondary flow vortices in the bi-layer flow, which are shown in Figure 2. Viscoelastic simulation of the bi-layer coextrusion converged up to Deborah number of 1. The simulation started to diverge at higher Deborah numbers. In order to understand the mechanisms behind the encapsulation observed in polymer coextrusion, the coextrusion experiments in this work were conducted using Newtonian fluids (glycerol, motor oil, and silicon oils), as well using Boger fluids, which are viscoelastic in nature. The coextrusion channel was constructed from transparent polycarbonate blocks to allow visualization of the entire channel. In the experiments with glycerol and the silicone oils no encapsulation was observed regardless of the viscosity grade of silicone oil. In experiments with motor oil and silicon oil, motor oil encapsulated the silicone oil with 50 cST viscosity, as well as the silicone oil with 200 cST viscosity, even though the viscosity of the 50 cST silicone was less than that of the motor oil, and the viscosity of the 200 cST silicone oil was more than that of the motor oil. No encapsulation was observed in the experiments with two Boger fluids, or with combinations of Newtonian and Boger fluids. Based upon the simulations and experiments in this project it is concluded that encapsulation observed in coextrusion is not caused by viscoelastic effects. The encapsulation phenomenon is also not caused solely by the difference in the viscosity of the fluid. We believe that polymer encapsulation in coextrusion is a surface phenomenon, which is caused by the difference in the wettability (spreading tendency of the polymer on die surface) and surface tension of the two polymers used. The polymer with higher wettability is expected to encapsulate the other polymer during coextrusion.
