Physics Colloquium: Predicting the Spreading of Filaments in Direct Ink Writing Additive Manufacturing of High Performance Thermoset Polymers

Prof. Nicolas Alvarez, Drexel University

Abstract: Themoset resins offer significant advantages over thermoplastics, especially in the high performance categories of aerospace and automobile parts. Direct Ink Writing (DIW) and Ambient Reactive Extrusion (ARE) are two scalable additive manufacturing meth- ods used to print/manufacture parts from thermoset resins. Both methods consist of extruding a thermoset resin from a nozzle to deposit filaments (a.k.a. rows) to build up layers, similar to the well-known Fused Deposition Modeling (FDM) process. Un- like FDM, thermosets are relatively low viscosity Newtonian fluids, which makes their depostion challenging. Printability of thermoset resins requires either a rheological mod- ifier, such as colloidal particles, or fast curing reactions during deposition. In the case of fast curing systems, the extruded filament can undergo significant spreading during the curing process. Successful printing requires the anticipation of such shape change to ensure accurate print dimensions and successful cohesion between extruded rows. In the case of rheological modifiers, the thermoset is formulated into a colloidal organogel whose modulus is intended to hold the filament shape during curing. However, the organogels undergo significant shear thinning during the extrusion process, and therefore also expe- rience significant spreading before gelation of the thermoset network. In this work, we develop a numerical simulation model for predicting the shape of filaments during the printing process. Significant emphasis is placed on the dynamic contact angle boundary condition and the rheological properties of the resin formulations. The model is devel- oped and validated with the aid of experimental results on both thermal and photo curing resins. Most importantly, we present a generalized spreading theory using scaling analysis to predict the shape of arbitrary resin systems with a small set of measured parameters. Ultimately, we aim to include these results and predictions in slicing software to predict printing parameters and avoid the Edisonian approach of print optimization.