Dynamic Flow Control Through Active Matter Programming Language

Dynamic networks of cytoskeleton and motor proteins can generate force that is essential in many cellular functions. In this talk, we show how to use biological active matter, which consumes chemical energy and generates force at molecular scales, to drive microfluidics towards constructing a single programmable device that can solve various micron-scale transport problems. Here, using optically-controlled motor-microtubule systems, we introduce a programming strategy for microfluidic control where flow fields are assembled through linear superposition of a set of fundamental flows generated by predefined programming modules. In general, the active matter is highly non-linear and will break down the linearity of Stokes flows. Combining experiments and theories, we identify a critical length for the spacing among the composition of optical signals, over which the flows created by different signals can be linearly superposed, and below which the superposition fails due to transport of active networks. Based on superposition, we define a modular active matter programming language that can spatiotemporally sculpt and composite complex flow fields. We build a coarse-grained model that quantitatively predicts the active fluid dynamics under arbitrary optical input. Model-driven programming design and optimization are realized in experiments for particle transport, extensional rheology of polymers and micron-scale manipulation tasks of human cells.