From Surface to Sheet and Back Again: A Computational Pipeline for Unfolding 3D Geometries, Strain-Guided Refolding, and Fabrication-Ready Self-Folding Structures

Abstract

Many bioelectronic devices, whether conformal, scaffolds, or otherwise, require three-dimensional geometries that cannot be realized through standard planar microfabrication alone. Self-folding thin-film structures, in which a flat precursor autonomously assembles into a target shape upon actuation, are a potential solution to this issue. However, it necessitates a quantitative design pipeline that can translate 3D geometry into fabrication-ready layouts and validate mechanical feasibility prior to fabrication. This thesis presents the development and demonstration of such a pipeline for SU-8 thin-film structures. The pipeline takes in a 3D geometry and then conducts automated surface unrolling, crease-level strain estimation, and finite element simulation of the refolding process. Outputs include fabrication-ready mask files, simulation-ready geometry, and quantitative strain maps identifying regions of elevated bending demand. Folding simulations are conducted in Abaqus using shell element models, with results validated against known SU-8 material limits. Successful folding trajectories are demonstrated for convex polyhedral geometries, and the limitations of the approach for curved surfaces are characterized and discussed. Together, these contributions establish the computational and simulation infrastructure required for predictive, geometry-driven design of SU-8 self-folding microdevices.