Lab-on-chip design automation takes cue from EDA

Thus far, most microfluid chips have been designed for specific applications, but researchers at Duke and the University of California at Los Angeles (UCLA) are aiming at general-purpose microfluid chips akin to the first GP microprocessors invented by Intel.

A Duke University spin-off, Advanced Liquid Logic Inc. (Research Triangle Park, N.C.), is readying a GP microfluidic processor for announcement later this year and delivery by early 2009. To make its GP microfluidics technology digital, Advanced Liquid Logic has standardized on a droplet size of 300 nanoliters, instead of piping continuous streams of fluids. I/O lines and reservoirs pump droplets onto and off of the microfluidic chips using electrical pulses, which move the droplets along channels segmented into slots just big enough for a single droplet. Under each slot is an electrode. By sequencing electrical pulses to the electrodes, electrical attraction and repulsion can be used to move the droplets one slot at a time.

"Our first product will be a system, since we believe we have to prove to the world that microfluidics is not just feasible, but practical and economical," said Richard West, Advanced Liquid Logic's CEO. The system will include a microfluidic "chip" whose channels and reservoirs are cast in FR-4 printed-circuit board material. The general-purpose processor will pipe fluids around its surface using algorithms defined on an attached personal computer.

"You place our digital microfluidics chip into our compact benchtop analyzer, then the software takes over to perform sample preparation, clinical chemistry, assays, PCRs [polymerase chain reactions] and many other standard laboratory operations in minutes instead of hours—all on the same platform, even on the same chip, and with the sensitivity and precision of state-of-the-art laboratory instruments," West said.

University of Texas (Austin) professor David Pan, ISPD general chair, has been working on algorithms to automate the operation of such digital microfluidics chips. "We are designing a high-performance droplet router for digital microfluidics that achieves much better routability," said doctoral candidate Minsik Cho. "Our digital droplet-transport mechanism controls operations according to a preprogrammed schedule."

According to Cho, preventing collisions between droplets at intersecting channels on a microfluidic chip requires nondeterministic polynomial time (refered to as "NP complete"), meaning that the solution is very difficult to optimize. But when the optimal solution is found, it should work for similar NP-complete problems, Cho said.