There are many different ways to construct complex structures from small particles. One is to use optical traps or tweezers to grab hold of and manipulate individual particles (Lab Chip 8, 2174–2181; 2008). Another is to engineer the interactions between particles in some way, such as by chemically functionalizing their surfaces, so that they self-assemble to form the desired structure (Science 295, 2418–2421; 2002). The former approach enables almost any structure to be built, but is too slow for bulk fabrication of multiple copies. And although the latter is ideally suited to bulk synthesis, exerting sufficient control over interparticle interactions to get them to organize themselves into arbitrary shapes without direction is challenging.

Credit: © 2010 APS

As an alternative that has the ease and control of techniques based on directed assembly and the high throughput of those based on self-assembly, Tobias Schneider and colleagues have explored a variety of algorithms for building complex structures in a microfluidic assembly line (Phys. Rev. Lett. in the press; preprint at http://arXiv.org/abs/1101.3791). The idea is to use the flow fields generated in a microfluidic chamber, which is fed by a series of inlets and outlets distributed around it, to steer particles suspended in the chamber into position. By controlling individually the rate of flow in or out of each inlet or outlet, the shape of these fields can be intricately controlled. This in turn can be used to guide the flow of suspended particles, and, by using particles that stick together when they come into contact, could be used to guide the formation of complex structures.

The authors first considered the conditions needed to guide simultaneously the movement of N particles confined in two dimensions in a circular Hele–Shaw cell. The number of inlet/outlet ports needed to achieve this is 2N+1, placing a modest limit on the number of particles that can be manipulated in a cell of given size. But they found that a more stringent limit is imposed by the fact that, as the number of particles is increased, the rate of flow needed through each port rises sharply. In simulations, this prevented them from being able to manipulate more than six particles at a time with moderate flows.

To overcome this, rather than manipulate many particles simultaneously, Schneider et al. instead built structures sequentially, one particle at a time. To do this in two dimensions, they find that only seven ports are needed — which they demonstrate by simulating the sequence of flows needed to construct letters of the English alphabet (pictured). And their analysis suggests that extending to the building of structures in three dimensions could be possible with just 11 ports.