In 2006, Caltech scientist Paul Rothemund invented a DNA origami technique to make different structures on the nano scale. The technique exploits the well-known self-assembly properties of single-stranded DNA to make almost limitless shapes and patterns that can be programmed by synthesising appropriate DNA sequences. Amusingly, Rothemund produced nanoscopic smiley faces, snow flakes and even a map of the Americas.

Hareem Maune, at the time a graduate student in applied physics working under Marc Bockrath on carbon nanotubes, Si-ping Han, a material science graduate student working for William Goddard III, and then undergraduate Robert Barish, working with Erik Winfree, reasoned that DNA origami might provide a neat way to construct nanoscale devices. [Maune et al., Nature Nanotech. (2009) 10.1038/nnano.2009.311]

Fundamentally, the researchers figured that DNA origami could be used to make two-dimensional “breadboards”, the patterned base layer for building a circuit from components, in this case carbon nanotubes. Sophisticated devices might be built by attaching the components to a template pattern on a breadboard assembled with the right DNA sequences.

“For nanoscale devices to become commercially useful,” says Goddard, “it is essential to have a technology that is scalable, that is, allows thousands, millions, billions to be manufactured simultaneously.” Until now, attempts at making nanoscale devices incorporating multiple carbon nanotubes have been hampered by the difficulty in placing the nanotubes into the correct geometry in a scalable way.

One challenge is to provide molecular attachment points on the featureless carbon nanotube surface that could both anchor the nanotube and identify different nanotubes. “This difficulty with chemically grabbing a nanotube at a well-defined ‘handle’ is the essence of the problem when you're trying to place nanotubes where you want them,” Winfree explains. DNA could provide the needed hook.

“DNA is the perfect molecule for recognizing other strands of DNA, and single-stranded DNA also just happens to like sticking to carbon nanotubes,” explains Han, “so we mix bare nanotubes with DNA molecules in salt water, and they stick all over the nanotubes' surfaces. However, we make sure that a little bit of each DNA molecule is protected, so that that little portion doesn't stick to the nanotube, and we can use it to recognize the DNA attached to the DNA origami instead.”

By creating DNA-coated nanotubes with two different labelling sequences, the team gave the nanotubes the ability to sit on one of two different positions demarcated by complementary DNA sequences on their breadboard. Thus, they can essentially “paint” a device pattern in two colours. “A standard DNA origami is a rectangle about 100 nm in size, with over 200 ‘pixel’ positions where arbitrary DNA strands can be attached,” Winfree says. By “painting” a stripe of one DNA sequence perpendicular to the other, they were able to get the carbon nanotubes to attach and form perpendicular crosses, resulting in field-effect transistors (FET), one of the most basic devices for building semiconductor circuits. Indeed this approach can be extended to many different colours to enable the assembly of complex devices.

Although there are challenging problems to overcome, the approach is, in principle, highly scalable. A DNA origami template could one day be used to self-assemble complex logic units, and to do this for billions or trillions of units that self-assemble in parallel, the researchers point out.