LiPCon: A Physics-Aware Spatially Resolved Light Penetration Compensation Method for Low-Cost 3D-Printed Microfluidics

Three-dimensional (3D) printing has emerged as a cost-effective alternative for rapid prototyping of complex microfluidic biochip structures. However, fabricating small, multi-layered 3D-printed microfluidic devices remains challenging due to light penetration effects during printing, which cause over-curing and dimensional deviations from the intended design.

We present LiPCon, a novel physics-aware design-for-manufacturing approach that significantly improves the dimensional fidelity of 3D-printed microfluidic structures. We introduce the first dataset quantifying discrepancies between designed and fabricated geometries in 3D-printed microfluidic devices. Using this dataset, we train a convolutional neural network that incorporates both design intent and the underlying physics of light-material interaction to predict localized, spatially-resolved compensation values.

Unlike the state-of-the-art method that applies uniform Z-position-based corrections, our approach performs spatially-resolved corrections to the 3D model. Experimental validation across complex multi-layer designs demonstrates that LiPCon consistently outperforms the state-of-the-art, reducing dimensional errors from 34–79% to just 1–6% in most cases, even when using a hobby-grade printer with off-the-shelf resin.