The oscillating field is precisely tuned so that only the nuclei in a very small sliver of the sample, called the resonant slice, respond. By scanning the magnet in a three-dimensional pattern, the researchers can move the resonant slice throughout the sample. It is this precision that allows the device to create such a high-resolution image.

Other forms of high-resolution imaging, such as scanning tunneling microscopy and atomic force microscopy, can only see the surface of a substance. Because of the resonant slice, MRFM can penetrate deep into the sample, building a 3-D picture of its internal structure.

MRFM first emerged in the early 1990s, and IBM has been a consistent leader in the field. In a landmark experiment in 2004, Rugar and his colleagues used the technology to detect the spin of a single electron. More recently, they generated images of a nonbiological sample with resolution as good as 90 nanometers--far better than conventional MRI, but not nearly sensitive enough to model individual biological structures.

Now, after years of painstaking incremental advances, Rugar's team has achieved nanometer-resolution imaging of a biological sample. The team chose to use the hardy, well-understood tobacco mosaic virus as a proof of concept and saw details as small as four nanometers. .

"This is actually the first time this technique has been used on a biological sample," says Rugar. "We wanted to show it really could do biology, because after all, that's our overall goal."

This successful experiment opens the door to a wide range of biological applications, says Rugar. In particular, he would like to be able to image individual proteins in order to determine their internal three-dimensional structure.

"You have thousands of proteins in your body that have no known structure, because there's no technique to determine their structure," says Rugar. Right now, the gold standard for solving protein structure is x-ray crystallography, which is limited to proteins that can be crystallized.

The nano-MRI scanner would not be subject to that constraint. In theory, with further improvements in resolution, it would be possible to examine proteins in their native state by rapidly freezing them. MRFM must be carried out at a very low temperature--barely above absolute zero--to minimize the noise created by thermal vibrations.

"The real significance of this is it shows that the limits of MRFM haven't been reached, and they're still on the way to doing an atomic imager," says Jonathan Jacky, a research scientist at the University of Washington. "An atomic-scale imager would be one of the most significant scientific instruments ever. It would be on the same level as the telescope or the light microscope. That's what's really exciting about this."