Programmable 2D to 3D Shape Morphing of Homogeneous Alginate Hydrogels Enabled by Laser Engraving and Data‐Driven Prediction

ABSTRACT

Shape morphing in hydrogels has attracted significant interest for soft robotics, biomedical devices, tissue engineering, and biomimetic systems, where programmable structural transformation is essential. Such behaviour typically arises from differential strain within the material, and conventional strategies rely on multilayer or rigid–soft bilayer architectures to induce mismatch‐driven deformation. Although effective, these require precise assembly, careful interfacial control, and complex fabrication to prevent delamination. Here, a laser‐engraving strategy achieves programmable shape transformation in a single homogeneous stimuli‐responsive sodium alginate–poly(vinyl alcohol) (SA–PVA) hydrogel. Micro‐scale grooves are patterned onto a planar dehydrated substrate in one maskless carbon dioxide (CO ₂ ) laser step. Subsequent ionic crosslinking drives spatially asymmetric swelling within the patterned regions, modulated by the accompanying geometric stiffness contrast, generating controlled three‐dimensional morphing from an initially flat sheet. Because the resulting calcium‐alginate network is water‐insoluble, the morphed geometry is retained. This eliminates multilayer assembly while enabling localized programming of deformation, providing a simplified, scalable route to morphable hydrogel systems. To accelerate design, a physics‐informed machine learning surrogate is trained on synthetic data from a Timoshenko‐based curvature pipeline. Complementing analytical theory, it predicts full‐field three‐dimensional morphologies from arbitrary two‐dimensional engraved patterns, enabling data‐driven forward and inverse design of programmable hydrogel actuators.