Computer Science Research Profile:
Expanding the Domain of Computing

From the Fall 2011 issue of Threads

Professor Chris Dwyer aligning a laser to examine a self-assembled nanostructure

Chris Dwyer views his work as expanding the domain of computing. Twenty years ago, it would have been hard to guess why anyone would want to carry a computer in a pocket, said the Associate Professor of Electrical and Computer Engineering, with a secondary appointment in Computer Science.

“What my group is studying with collaborators is the next sort of expansion,” Dwyer said, “where the technology’s improved to the point where we can take miniaturization of computers to the ultimate limit, down to a molecular scale.”

Using DNA as the substrate for building self-assembling circuits, Dwyer’s group is investigating a technology that can bypass the need for semiconductor foundries and the $40 million to $60 million entry fee they charge for making a conventional, silicon-integrated circuit.

Schematic of molecular logic gates on a DNA substrate designed as a multiplexer to detect an array of biomolecules

“What we’re basically taking advantage of when we’re building these structures is biopharmaceutical infrastructure,” Dwyer said. “And that infrastructure is larger than the semiconductor infrastructure. But that’s not the end of the story. The beautiful part of this is that any one of my students can manufacture 1,000 trillion of these nanostructures in a few hours. And that rivals the output of all of these foundries combined. That’s one student working alone in a lab that’s not optimized for anything but clutter.

“So that’s really where the beauty is — that we have precision and manufacturing scale and without a $10 billion foundry. The resolution and scale that we can pattern things at are clearly far beyond what the semiconductor industry can do today and, in the foreseeable future, will be able to do. We’re well beyond the physical limits that will prevent them from building smaller structures.”

Since joining Duke in mid-2004, Dwyer, who received tenure last year, has been using DNA to build logic gates — the primitive building blocks for computing — piecing them together and studying the systems to better understand what they theoretically could be used for. His devices have electrical properties similar to a diode, which allows current to flow in one direction. To enable more complex applications, his group is trying to develop a class of device that they call Diffusive Exciton Valves — DEVs. These devices would be similar to the transistors that are fundamental to creating conventional computers.

By making the devices so small, Dwyer — who received the Presidential Early Career Award for Scientists and Engineers in 2009 — sees much opportunity for expanding computing in both the biological and pure materials contexts. Through collaborations, his group already has demonstrated some simple steps toward injecting computers into blood serum and other fluids, allowing for computations beyond simple detection of biomarkers. In the pure materials domain, they see the potential to imbed computers into building materials for smart materials computing.

“If you could make self-healing materials as the chemists envision it, that would solve a lot of the problems building with very thin, lightweight materials,” Dwyer said. “Most of the bulk of a material is there to make it resilient to all forms of dynamic change — damage being one of them.”

A molecular pattern organized on a DNA substrate, was the first demonstration of a fully addressable, self-assembled patterning process

As a member of a Defense Advanced Research Projects Agency study group and others organized by the Office of the Secretary of Defense, Dwyer also is working on several potential, security-related applications. Using the DNA technology, his group has developed a stronger Physically Unclonable Function, embedded in small physical objects and used for authentication. Current PUFs cannot be shared. Their device — the Resonance Energy Transfer Physically Unclonable Function — can be controllably cloned by the manufacturer, making it more practical for use by multiple parties and yet difficult to reverse engineer because of its molecular scale. His group also has found that a fair amount of data can be encoded into a pollen-size particle. Using that, they are working on a contact tag tracer that could be used to tag currency or other items and monitor how the items are transferred or who is in physical contact with them.

Using similar technology, Dwyer’s group also is working on enhancing optically active identification tags, similar to radio frequency identification tags that are used in theft deterrent systems at stores and libraries. Using a beam of light, data in optically active tags can be read at a greater distance than radio frequency allows. Dwyer’s group is working on how to miniaturize and build in signatures so that information like a product’s UPC can be embedded around a bag of potato chips, allowing for even quicker data retrieval and store checkout. They also are working on increasing the storage density of optical media using DNA. In simple proof-of-principle experiments, they have demonstrated the ability to achieve between four and six times greater density than a traditional optics-based approach.

“But in the limit, we’ve demonstrated theoretically that we can get several thousand-fold density,” Dwyer said. “That’s basically enough to take a snapshot of most of the Web right now and store it.”

Dwyer, who was elected a Kavli fellow of the U.S. National Academy of Sciences last year, noted potentially many more applications for the DNA technology exist: “This is just scratching the surface.”

Return to Reseach Profiles home page