LAUSR

Abstract Industries such as orthodontics and athletic apparel are adopting vat photopolymerization (VP) to manufacture customized products with performance tailored through geometry. However, vat photopolymerization is limited by low manufacturing speeds and the trade-off between manufacturable part size and feature resolution. Current VP platforms and their optical sub-systems allow for simultaneous maximization of only two of three critical manufacturing metrics: layer fabrication time, fabrication area, and printed feature resolution. The Scanning Mask Projection Vat Photopolymerization (S-MPVP 1 ) system was developed to address this shortcoming. However, models developed to determine S-MPVP process parameters are accurate only for systems with an intensity distribution that can be approximated with a first order Gaussian distribution. Limitations of optical elements and the use of heterogeneous photopolymers result in non-analytic intensity distributions. Modeling the effect of non-analytic intensity distribution on the resultant cure profile is necessary for accurate manufacturing of multiscale products. In this work, a model to predict the shape of cured features using analytic and non-analytic intensity distribution is presented. First, existing modeling techniques developed for laser and mask projection VP processes were leveraged to create a numerical model to relate the process parameters (i.e. scan speed, mask pattern irradiance) of the S-MPVP system with the resulting cure profile. Then, by extracting the actual intensity distribution from the resin surface, we demonstrate the model's ability to use non-analytic intensity distribution for computing the irradiance for any projected pattern. Using a custom S-MPVP system, process parameters required to fabricate test specimens were experimentally determined. These parameters were then input into the S-MPVP model and the resulting cure profiles were simulated. Comparison between the simulated and printed specimens dimensions demonstrates the model’s effectiveness in predicting the dimensions of the cured shape with an error of 2.9%.