Nanoscale MRI

Magnetic resonance imaging offers rich three-dimensional pictures, but with limited resolution. Imaging at the nanometre scale has now become possible using highly sensitive force-detection techniques.

Seeing inside a complex object is an invaluable aid to understanding it, so three-dimensional imaging is a pressing objective in areas ranging from molecular and cell biology to investigations of the electronic and structural properties of materials. The challenge is made more difficult by the desire to see the object without altering it in the process. This requires a delicate touch, involving as weak an interaction as possible with the object. But such an approach often conflicts with the need for the high spatial resolution that makes fine detail visible.

Writing in Proceedings of the National Academy of Sciences1, a group led by Dan Rugar reports the success of a delicate yet effective approach to imaging — one that gently excites the nuclear spins in their test objects, particles of tobacco mosaic virus, and records their locations by listening to the spins' oscillating magnetization. An image is derived by recording these signals from a three-dimensional mesh of locations within the object; achieving high resolution requires the mesh to be very fine, so that the volume sampled at each grid location is as small as possible. The authors' sensitive detection of the feeble signals from these elements of nanometre-scale volume, and subsequent reconstruction of the three-dimensional structure of the virus from these signals, marks the arrival of a powerful tool for non-perturbative imaging of a single copy of an object — be it biological, electronic or magnetic — at the nanometre scale.

Rugar and colleagues' approach is inspired by the impressive three-dimensional views of a material provided by magnetic resonance imaging (MRI). Rather than measuring how energetic particles interact with the object to obtain an image, the technique uses radio waves whose energy is less than a billionth of that of the X-rays used for diffraction studies or the electrons used in an electron microscope. MRI is itself based on nuclear magnetic resonance (NMR), which exploits the intrinsic and plentiful nuclear magnetic spins present in all substances. These nuclear magnets oscillate at a precisely measurable frequency that is determined by fields generated by neighbouring atoms, and by an externally applied field. Hence, these nuclear magnets are embedded, microscopic probes that reveal details of their host's electronic, magnetic and structural properties. Detailed information obtained from NMR has been extensively used for tasks ranging from identifying organic molecules to illuminating subtle features of exotic superconductors.

For imaging, the external field is arranged to vary controllably across the sample, so that the frequency of the nuclear magnetic oscillation will reveal its precise location. This mechanism underlies non-invasive, three-dimensional MRI of regions deep within a sample. Rather than scattering energetic particles, MRI uses low-energy radio waves to excite the nuclear spins so that their oscillation frequency can be measured. A benefit of using magnetic resonance for imaging is that these magnetic resonance signals allow spatially resolved NMR experiments and characterization that enrich the images with detailed microscopic information.

However, the weak interaction that makes MRI so non-invasive is also its Achilles heel: the interaction of the detector with the spin is so small that, in conventional approaches, many spins (1012–1018) are needed to provide a large enough signal to tease out information about the materials. The dimensions of the resolvable volume are limited by the need to detect the weak oscillatory signal of the few spins in the small-volume elements that make up the image. This limits conventional MRI to volumes of several cubic micrometres, and so reduces the usefulness of the technique in solid-state physics, or molecular or cell biology.

In 1991, John Sidles2 proposed a system for mechanically sensing the weak force that a microscopic ferromagnet exerts on the nuclear magnetic moment in a sample. Tiny forces, he suggested, can be measured by placing the sample under investigation on a compliant cantilever. By observing the slight resulting deflection of the cantilever using, for example, an optical interferometer, extraordinarily small forces can be detected3. Force-detected MRI, dubbed magnetic resonance force microscopy (MRFM), has rapidly improved in sensitivity and spatial resolution4,5,6: it has been used to observe a single electron spin7 and for highly sensitive nuclear-spin detection8. MRFM is also a practical materials probe that has been applied to major problems in science9,10 and technology11. Beyond this, it has been shown that techniques used in conventional pulsed NMR are effective for force-detected magnetic resonance12.

Rugar and colleagues' imaging of individual virus particles1 is a striking advance in MRFM capability that demanded exceptional detection sensitivity. In particular, the ferromagnetic probe must be brought within tens of nanometres of the cantilever-mounted virus. At these distances, the cantilever experiences many other forces from the nearby surfaces — including, for example, van der Waals forces that are typically thousands to millions of times larger than the nuclear magnetic forces to be measured, and dissipative, electrostatic cantilever-surface forces that produce noise that obscures the signal.

The authors' success is the fruit of a decade of work developing ultrasensitive force-detection techniques. They include excitation techniques13, which manipulate the spins to produce a distinctive force signal that can be picked out from the background forces, and a nanofabricated antenna14 that produces a strong radiofrequency magnetic excitation field sufficiently localized that it doesn't disturb the cantilever (the nuclear magnetic forces generate cantilever deflections only at the sub-angstrom level). Finally, the work shows that the noisy signals can be deconvolved into images.

The MRFM procedure will not meet all imaging needs. It is a demanding technique that must be performed in a vacuum and at low temperature. This is a limitation that is shared by electron microscopy of biological specimens, which is nonetheless a highly successful imaging tool. The detection sensitivity of MRFM is improving rapidly, and its history indicates that these capabilities, now at the cutting edge, will soon be routine for MRFM practitioners. But it will be some time before those capabilities can be exploited by the wider microscopy community.

That said, the demonstration1 of the imaging of viral particles at a resolution down to 4 nanometres heralds the emergence of a new microscope for investigating native biological specimens that will compete with, and complement, electron microscopy and NMR spectroscopy. It uniquely combines non-destructive imaging with the capability of imaging individual copies of specimens such as proteins. The approach is also likely to find wide application beyond biology, in investigations of the chemical and elemental make-up of nanostructures in the physical and materials sciences.


  1. 1

    Degen, C. L., Poggio, M., Mamin, H. J., Rettner, C. T. & Rugar, D. Proc. Natl Acad. Sci. USA 106, 1313–1317 (2009).

  2. 2

    Sidles, J. A. Appl. Phys. Lett. doi:10.1063/1.104757 (1991).

  3. 3

    Mamin, H. J. & Rugar, D. Appl. Phys. Lett. doi:10.1063/1.1418256 (2001).

  4. 4

    Sidles, J. A. et al. Rev. Mod. Phys. doi:10.1103/RevModPhys.67.249 (1995).

  5. 5

    Hammel, P. C. & Pelekhov, D. V. Handbook of Magnetism and Advanced Magnetic Materials Vol. 5 (Wiley, 2007).

  6. 6

    Kuehn, S., Hickman, S. A. & Marohn, J. A. J. Chem. Phys. doi:10.1063/1.2834737 (2008).

  7. 7

    Rugar, D. et al. Nature 430, 329–332 (2004).

  8. 8

    Mamin, H. J., Poggio, M., Degen, C. L. & Rugar, D. Nature Nanotechnol. doi:10.1038/nnano.2007.105 (2007).

  9. 9

    Obukhov, Y. et al. Phys. Rev. Lett. doi:10.1103/PhysRevLett.100.197601 (2008).

  10. 10

    Klein, O. Phys. Rev. B doi:10.1103/PhysRevB.78.144410 (2008).

  11. 11

    Thurber, K., Harrell, L. & Smith, D. J. Mag. Res. doi:10.1016/S1090-7807(03)00040-5 (2003).

  12. 12

    Lin, Q. et al. Phys. Rev. Lett. doi:10.1103/PhysRevLett.96.137604 (2006).

  13. 13

    Rugar, D., Budakian, R., Mamin, H. J. & Chui, B. W. AIP Conf. Proc. 696, 45 (2003).

  14. 14

    Poggio, M., Degen, C. L., Rettner, C. T., Mamin, H. J. & Rugar, D. Appl. Phys. Lett. doi:10.1063/1.2752536 (2007).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Hammel, P. Nanoscale MRI. Nature 458, 844–845 (2009).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.