Adaptive Integrated Microfluidic System with Real-Time Dielectric Feedback for Deterministic Micro-Manipulation

Abstract

Abstract Conventional microfluidic particle sorting systems predominantly rely on static, open-loop operational principles, rendering them highly vulnerable to performance deterioration induced by cellular phenotypic heterogeneity and localized hydrodynamic fluctuations. Here, we introduce an adaptive integrated microfluidic system (AIMS) that implements a closed-loop sensing−decision−actuation−verification control framework via on-chip integration of dual-frequency impedance cytometry (IFC) and tunable traveling surface acoustic waves (T-TSAW). An upstream IFC array (1#IFC) performs real-time, label-free electrical interrogation of single-cell dielectric phenotypes. These transient signals are processed in real time by a custom finite-state machine (FSM) algorithm to dynamically modulate the downstream T-TSAW field. Concurrently, a downstream IFC array (2#IFC) performs post-sorting verification and feeds real-time state metrics back to the FSM core. This initiates an adaptive hill-climbing optimization loop to enable continuous acoustic frequency calibration. The system performance was rigorously evaluated using monodisperse polystyrene microspheres and biologically relevant colloidal models. The closed-loop architecture successfully isolated target MCF-7 cells from an erythrocyte-dominated simplified blood model, achieving an acoustic sorting efficiency of 95 ± 1.5% and a purity of 94 ± 2.5%. Furthermore, continuous adaptive sorting of primary mouse immune cell populations (PBMCs and neutrophils) was demonstrated, achieving a sorting efficiency of 92 ± 1.8% and a purity of 93 ± 6.7%. Ultimately, the seamless integration of transient electronic sensing algorithms with multiphysics mechanical actuation establishes a highly robust architectural framework for autonomous microelectromechanical manipulation platforms.