Imaging Spin Noise

Noise—small thermal or quantum fluctuations of particles that generate electric or magnetic signals—is generally thought of as undesirable interference, obscuring the true signals scientists are trying to detect.

Noise isn't necessarily bad, however. The random spin fluctuations of electrons or protons, for example, can reveal much about collections of atoms or molecules. The signals from spin noise are so weak compared with those induced by outside forces, such as lasers and magnets, that until about 20 years ago, detectors weren't sensitive enough to measure them.

In recent years, however, technological advances have allowed scientists to begin exploring spin noise as a source of information about atomic and molecular systems. In particular, spectra from nuclear spin noise can yield the same information about molecular structure as those from common nuclear magnetic resonance measurements. A major difference is that unlike conventional NMR spectroscopy, spin noise spectroscopy doesn't require exposing the sample to radio waves and a strong magnetic field.

Now, Alexej Jerschow, assistant chemistry professor at New York University and Norbert Müller, chemistry professor at Johannes Kepler University in Linz, Austria, have taken spin noise spectroscopy to its next logical step: spin noise imaging. Demonstrating a technique roughly analogous to magnetic resonance imaging, they unveiled the first spin noise images of room-temperature samples of water and deuterium oxide inside a container (Proc. Natl. Acad. Sci. USA 2006, 103, 6790).

The work has generated buzz at recent NMR meetings, including the 47th Experimental Nuclear Magnetic Resonance Conference held last month in Pacific Grove, Calif. That such faint signals could be captured in the first place "is astonishing to a lot of people," comments Warren S. Warren, chemistry professor at Duke University and director of Duke's Center for Molecular & Biomolecular Imaging.

Spin noise stems from the spin fluctuations of an ensemble of particles. Protons, for example, can have spins of either -½ or +½. The individual spins randomly fluctuate, and at any moment, an ensemble of these spins can assume a magnetization that points slightly in one direction or another. The fluctuating magnetic moment precesses and decays just like the ensemble moment in ordinary NMR.

Developments such as the greatly improved sensitivity of detectors and cryogenically cooled probes "allow us to detect extremely weak signals," Jerschow tells C&EN. To assemble a two-dimensional image, Jerschow and Müller combined spin noise measurements with MRI techniques for obtaining images from NMR signals. They pieced together a two-dimensional image from a series of 30 image slices.

Because the signals from spin noise are so faint, Warren notes, the technique isn't likely to replace MRI in conventional medical settings.

So what might spin noise imaging be good for? The lack of dependence on an external radio-frequency source could make it safer to study delicate samples such as explosives. Spin noise imaging machines could also be combined with portable NMR technologies now in development, such as ex situ NMR and microtesla NMR—useful for placing inside the bore of an oil well, for example.

Spin noise imaging could be of particular use, however, in mesoscale systems having a small number of spins, such as microfluidic chips, notes chemistry professor Alexander Pines of the University of California, Berkeley.

As the number of spins decreases, the spin noise signal becomes the dominant effect, stronger than what could be obtained with conventional NMR or MRI. Also, some materials have extremely long relaxation times—up to two years, in some cases—making conventional MRI difficult. But noise is always there, and always constant.

"I think this is the most promising field of application," Jerschow says.