2025 AIChE Annual Meeting

Automated Flow Engineering for the Controlled Fabrication of Complex Microstructures

Manufacturing complex macroscopic structures with mesoscale detail remains a major challenge in additive manufacturing. Current techniques, such as 3D printing, face limitations related to long fabrication times and material stability under operating conditions. An alternative approach is based on fluid convection, where the fundamental building blocks of a desired structure are assembled by controlling flow fields within microfluidic devices. Previous studies using quasi–two-dimensional microfluidic channels have demonstrated precise control of several non-Brownian particles with micron-scale dimensions. However, achieving accurate control and assembly of complex structures remains an open challenge.

In this work, we present a theory-driven computational framework for the automated formation of microstructural assemblies through fully controlled flow fields, enabled by concepts from model predictive control. The framework is implemented in JAX, a computational environment that supports end-to-end automatic differentiation and enables efficient numerical optimization for the real-time control of multiparticle systems. This approach allows the design and assembly of complex particle structures that go beyond the capabilities of current solvers in the literature.

We demonstrate the framework’s performance in model systems, including simultaneous directed control of multiple non-Brownian particle trajectories and the formation of complex polymer-like assemblies. This work provides a generalizable strategy for feedback-driven particle assembly, offering new opportunities for microscale manufacturing, programmable material design, and potentially drug delivery.