Postdoctoral Fellow The University of Texas at Austin
Vat photopolymerization (VPP) processes, such as mask-projection stereolithography, work by selectively projecting light onto a vat of photo-resin which cures over time. However, oxygen dissolved within the photo-resin acts as an inhibitor and delays the photocuring. Since oxygen can diffuse from areas without light to areas receiving light, the boundaries of lighted shapes are slower to cure, resulting in an ‘inhibitive’ blurring effect. In contrast, the incoming light can scatter when traveling through the photo-resin and locally cure areas where light should not be present, resulting in a ‘facilitative’ blurring effect. Both blurring effects lower the accuracy of the final print. In this work, we first gain an understanding on blurring by conducting VPP on a single layer using various shapes. We then code a pixel-by-pixel, time-discretized model in Python to simulate the curing observed in our experiments. Finally, a machine learning algorithm is set up via PyTorch to counteract the blurring effects and produce a more accurate print.
Learning Objectives:
Upon completion, participants will be able to describe how, in vat photopolymerization, oxygen dissolved within the resin will inhibit curing and can diffuse to blur the resultant print.
Upon completion, participants will be able to describe how, in vat photopolymerization, incoming light can scatter which also acts to blur the resultant print.
Upon completion, participants will be able to describe how both blurring mechanisms can be modeled and how a machine learning algorithm can be used to counteract the blurring.
