Shape from Release: Inverse Design and Fabrication of Controlled Release Structures

We propose a gradient-based inverse design framework for the problem of architected release. The backbone of our solution is a simple forward model, based on a geometric abstraction, that is governed by the Eikonal equation. We formulate our problem as a PDE-constrained topology optimization whose objective is to fit the release curve computed from the Eikonal equation solution to the target release profile. We present an exact sensitivity analysis of the discrete fast marching solution and obtain a robust and efficient simulation model with analytical derivatives. Our formulation enables gradient-based optimization for discovering shapes achieving novel release behaviors and exploring trade-offs between release curve closeness and fabricability. A novel image-based measurement setup for acquiring the physical release curves of 3D printed objects is used for testing.

• Suitable for 3D printed pharmaceuticals
• Produces freeform 3D printable geometries
• No additional material needed, no multi-lamination
• Replaces trial & error process with a topology optimization

PCT patent application PCT/EP2023/077032 filed in September 2023, published under WO2024/074417