C&EN: TODAY'S HEADLINES - MICROFLUIDIC CIRCUITS

CELIA HENRY

The microfluidic equivalent of an integrated circuit has been made for the first time by researchers at the California Institute of Technology.

“This is the first demonstration of such a complex microfluidic device with thousands of valves and hundreds of individually addressable chambers,” says Stephen R. Quake, associate professor of applied physics at Caltech. “To put that all together on a 1-inch by 1-inch chip is a notable achievement.” The Caltech team also includes graduate students Todd Thorsen and Sebastian J. Maerkl.

A “multiplexor,” which is a combinatorial array of binary valve patterns, controls the fluid flow on the devices. The number of control channels is logarithmically related to the number of flow channels, allowing a large number of flow channels while maintaining a small number of control channels. One device, for example, contains 1,000 chambers and 3,574 valves, but has only 22 inlets [Science, published online Sept. 26, http://www.sciencemag.org/cgi/content/abstract/1076996].

The devices, made of polydimethylsiloxane (PDMS), are constructed in two layers, with the control layer stacked on top of the flow layer. The control channels have both narrow and wide sections; a valve is formed where a wide section crosses over a flow channel. These valves are opened and closed by applying pressure to the control channels, which deflects the PDMS membrane between the crossed channels, shutting off the flow channel.

One such device functions as a memory device consisting of 1,000 individually addressable compartments. “We can create arbitrary patterns in this array, just like you can in a regular memory,” Quake says. “On the simplest level, it’s analogous to RAM in that you can store bits of fluid and randomly pull them out.” A second device allows reagents to be separately loaded, mixed, and recovered. Quake’s team used this device in an assay to test for the expression of an enzyme.

Quake plans to use these microfluidic devices for a variety of applications, including structural genomics, genetic analysis, environmental microbiology, and high-throughput screening. “We will design a chip specifically for each application, but we’ll make use of components that have been useful in other circumstances,” Quake says.

The devices do have shortcomings. “There is not a perfect material, and PDMS is no exception. There are always going to be issues such as permeability and nonspecific binding,” Quake says. “But with clever design rules and device geometry, you can work within the limitations of any given material and still make very powerful devices.”