a desired 3D part. Typically, stereolithography produces parts in discrete layers. Exposed areas are cured through the full layer height, whereupon the part is repositioned and recoated with resin before the next layer is exposed. Recently, continuous stereolithographic technologies have been developed which increase print speeds by eliminating the time-consuming repositioning and recoating steps. [19,20] Print speed in continuous stereolithography is dependent on the resin absorbance height, with low-absorbance resins allowing extremely high print speeds of up to 2000 mm h −1 at the cost of part fidelity. [20] In stereolithography, the penetration depth of light in the resin limits accuracy along the vertical axis: unaccounted-for light propagation can cause undesired curing, known as cure-through, overcure, [21,22] the back-side effect, [23] or printthrough error. [19,24] This phenomenon can also contribute to cross-linking heterogeneity, introducing internal stresses which can deform the part and further reduce fidelity. [25] The prevalent strategy to mitigate cure-through is to add nonreactive light absorbers to the resin formulation. [21,[25][26][27][28] Highly absorbing resins have been widely adopted despite the slower print speeds needed to ensure fully cured layers. Alternatively, cure-through can be mitigated without sacrificing speed by modifying the projected images, known as slices, based on modeling of the curing process. Optimization-based methods to eliminate cure-through by adjusting model dimensions have been developed for external surfaces and internal voids in traditional stereolithography. [22,24,29] Manual adjustments to account for cure-through have also been reported. [30] Nevertheless, slice correction has not been described for continuous stereolithography, where cure-through is a more significant and complex problem. Furthermore, existing models of continuous stereolithography are not tailored to this application. [31][32][33] Here, we present a curing model and a slice correction algorithm for continuous stereolithography. Previous noncontinuous approaches used iterative and heuristic processes to find optimal corrections and were restricted to black and white pixels; our correction method uses grayscale, which has previously only been used to improve lateral resolution, [34] along with an exact mathematical solution to precisely set the dose profile within a part. We also present experimental validation of our model and correction approach using a recently developed two-color continuous stereolithographic 3D printer. [20] These methods are Continuous stereolithography offers significant speed improvements over traditional layer-by-layer approaches but is more susceptible to cure-through, undesired curing along the axis of exposure. Typically, cure-through is mitigated at the cost of print speed by reducing penetration depth in the photopolymer resin via the addition of nonreactive light absorbers. Here, a mathematical approach is presented to model the dose profile in a part produced ...
