ORNL Review Home Page
Featured in This Edition
Previous Article
Next Article
Search the ORNL Review Site
Comment on this article

Incredible Shrinking Labs:
Weighing a Move to the Nanoscale

Imagine a device the size of a deck of cards that could help a physician quickly determine why you are feeling so sick and tired. By rapid analysis of the DNA and proteins in a drop of your blood, a "nanofluidic lab on a chip" could indicate to your doctor a million to a billion times faster than current devices whether a bacterium, virus, or a cancerous tumor is making you ill. At least, that's one application seen by scientists involved in the new field of nanofluidics: the active transport of material through a channel or conduit whose diameter is less than 100 nanometers (nm). A nanometer is a billionth of a meter.

Basic research in nanofluidics is being conducted by ORNL and University of Tennessee (UT) researchers, using internal funding from the Laboratory Directed Research and Development (LDRD) Program at ORNL. The LDRD project is headed by Mike Ramsey, an ORNL corporate fellow in the Laboratory's Chemical and Analytical Sciences Division and the inventor of the lab on a chip.

These commercialized versions of ORNL's lab on a chip are products of Caliper Technologies, Inc., and Agilent Technologies, Inc. (Photo courtesy of Caliper Technologies, Inc.)

The lab on a chip is a microfluidic structure first built by Ramsey 10 years ago to demonstrate the separation of chemicals in very small volumes. Six years ago, this technology was licensed to Caliper Technologies, Inc. Ramsey and his group in CASD are conducting considerable research on developing improved lab-on-a-chip technologies for biological, environmental, forensic, and defense applications. The lab on a chip has been honored by R&D magazine as one of the 40 top innovations it has recognized since beginning its R&D 100 competition in 1963. Additionally, a panel of citizens has named the lab on a chip one of the top 23 technologies developed using Department of Energy funding.

In a lab on a chip, several channels, each the size of a hair in width and one-fifth of a hair in depth, connect reservoirs containing chemicals. The channels and reservoirs are carved, or etched, in a glass chip (such as a microscope slide) using microfabrication techniques. Liquids are "pumped" through each channel, using electric fields or pressure differences, at a rate of one millionth of a drop per second.

"In a nanofluidic device," says Ramsey, "the typical channel would be 1000 to 10,000 times smaller than a hair. We expect that the interactions between the material in the channel and the channel walls—what we call solid-liquid interactions—will be much more dominant in a nanofluidic device from those in a microfluidic device. We expect fluids to change their behavior as we shrink the sizes of the channels through which they flow."

In the first nanofluidic experiment conducted at ORNL by Ramsey, Steve Jacobson, and Chris Culbertson, it was observed that electrokinetic transport is reduced as dimensions are scaled down from the micrometer range to less than 100 nm. These results are the first experimental verification of statistical mechanical theories developed more than 30 years ago for electrically driven fluid transport through small channels. "We fabricated an 80-nanometer channel that is 20 microns wide and 80 nanometers deep," Ramsey says. "We forced an electrolyte—a sodium tetraboride buffer solution—through it. We confirmed the theory that the fluid velocity would be reduced in such a small channel."

Above: Scanning electron microscope image of four nanotrenches vertically connecting two microtrenches. The nanotrenches are only 100 x 100 nm2 in cross section but more than 40 microns long. Below: Perspective of a 900 x 80 nm2 trench cut in silicon for nanofluidic studies at ORNL. The trench was created using a focused ion beam at Vanderbilt University and imaged with an atomic force microscope at ORNL.

The reduction in fluid velocity at these small scales is related to the fact that the channel dimensions are similar to the electrical double-layer dimensions. The electrical double layer is the region of fluid near the solid-liquid interface where positive and negative charges may be separated so that the solution is not electrically neutral at a given location, giving rise to electrically driven fluid transport. As the channel approaches the double-layer thickness, the "pumping regions" begin to overlap and become less efficient in imparting momentum to the fluid.

"In our LDRD project," Ramsey says, "we are trying to show that we have the ability to fabricate nanofluidics devices with nanoscale features, that we can do experiments with them, and that we can understand their fluid transport characteristics experimentally and theoretically, using computational simulation.

"The ability to fabricate fluidic structures with dimensions at the molecular scale will allow fundamental studies of fluid transport at the smallest possible dimensions. In addition, practical tools for the analysis of biopolymer molecules, such as DNA and proteins, could well result from nanofluidic studies."

Ramsey's group is trying to learn how to fabricate nanoscale channels by shrinking the width and depth of channels etched in glass slides for microfluidic lab-on-a-chip devices. In ORNL's Solid State Division, Dave Zehner, Tony Haynes, and Arthur Baddorf are working with Len Feldman and his colleagues at Vanderbilt University to apply thin-film and ion-milling technologies to creating nanoscale channels for nanofluidic devices.

Sorting proteins by size and sequencing DNA a million to a billion times faster than current technologies are possible, practical applications of nanofluidic devices. But first, proof-of-principle experiments must be done. Ramsey's group plans to do some of these experiments.

One problem in using a lab on a chip to separate DNA molecules and proteins by size is that a sieving polymer must be added. When DNA molecules and proteins flow through this chemical sieve, the smaller biopolymers move faster than the larger ones, thus causing the separation.

"But it's a nuisance to get sieving polymer into a lab on a chip," Ramsey says. "It takes time and adds to the cost of making the chip.”

Ramsey's group plans to do an experiment with a one-dimensional nanoscale channel to determine if biomolecules such as proteins or DNA strands can be separated by size, based on their mobility through the channel. "We believe that the larger molecules will move faster than the smaller ones as a result of hydrodynamic effects," Ramsey says. "The bigger molecules flowing along at the center of the fluid won't get as close to the channel wall as the little ones. Thus, the larger molecules experience, on average, a greater velocity and travel through a molecular-size channel faster. This is speculation, but we will do experiments to determine whether this hypothesis is valid."

Another experiment the ORNL group hopes to do is to show that a two-dimensional nanoscale channel structure can sequence a single strand of DNA by obtaining its electrical signature. Evidence obtained at Harvard University and theory done there by Dan Branton and colleagues suggest that each of four types of DNA bases—adenine (A), cytosine (C), guanine (G), and thymine (T)—can produce its own distinctive electrical conductivity through a nanoscale channel.

"If we could send a strand of DNA bases single file along a nanoscale channel with electrodes at each end," Ramsey says, "we might be able to interrogate the strand by measuring a current. The current variations indicate the order of bases in the DNA strand."

A snapshot from a molecular dynamics simulation of sodium chloride aqueous solution in a 4-nm cylindrical pore: red (oxygen); green (hydrogen); yellow (sodium); and purple (chloride). Some of the sodium and chloride ions are hidden from view. ORNL and UT researchers are simulating the behavior and transport of water containing salt and DNA or proteins in ultra-small channels (e.g., through a 4-nm pore) subjected to electric fields.

To help guide and interpret the results of experiments by Ramsey's group, Hank Cochran of the Chemical Technology Division and Shengteng Cui of UT are using computational molecular simulation. They are modeling the movements of individual atoms and molecules in fluids confined in nanoscale channels. They are also simulating the behavior and transport of water containing salt and DNA or proteins in ultra-small channels (e.g., through a 4-nm pore) subjected to electric fields. In employing these models, they take into account the electric-field and surface forces that extend through the liquid tightly confined inside a nanoscale channel. Modeling fluids in nanoscale channels under experimental conditions is challenging because the number of atoms or molecules simulated must be larger than is currently feasible by molecular methods, yet the models from continuum fluid dynamics become inaccurate at the nanoscale. Cui and Cochran are developing a new methodology that uses continuum equations, with local fluid properties determined from molecular simulation. This new approach should enable an understanding of the results of nanofluidic experiments and guide the design of nanofluidic devices.

In time, nanofluidics research could lead to very small devices that may have a very large impact.

Beginning of Article

Related Web sites

ORNL Chemical and Analytical Sciences Division
R&D Magazine Online
ORNL Solid State Division