Data-Driven Printability Modeling of Hydrogels for Precise Direct Ink Writing Based on Rheological Properties.

Hydrogels are gaining significant attention in soft robotics and electronics due to their favorable mechanical properties and sustainability. While hydrogel inks enable three-dimensional (3D) printing as a key fabrication technique, the relationship between their rheological behavior and printability remains insufficiently understood. This study quantitatively examines this correlation through a rheology-printability database of 150 3D-printed hydrogels analyzed via machine learning. The database includes nonlinear rheological metrics, such as large-amplitude oscillatory shearing (LAOS), which mimic real 3D printing conditions involving repeated flow and stoppage. Printability is quantitatively evaluated in horizontal and vertical directions and inconsistency through image analysis of 3D printed structures. A predictive model for printability is developed using Random Forest regression, achieving reliable predictions within a 10% margin. Permutation importance analysis suggested that horizontal printability is primarily influenced by variables related to post-extrusion recovery and relaxation process, whereas vertical printability is mainly governed by viscous responses under high-strain-rate flow through the nozzle. Overall, this study provides quantitative insights into the intricate relationship between hydrogel rheology and 3D printability, paving the way for the sustainable design of hydrogel inks and their 3D printing processes for the precise fabrication of soft robotics structures and electronics.