A terahertz nanoscope

Article metrics

Replacing the 'micro' in microscope with 'nano', and using invisible light instead of visible, won't give scientists an instrument that can image nanostructures — unless they first beat the system's diffraction limit.

It requires a bit of imagination to think about using invisible light to look at an object of nanometre dimensions. Yet this is precisely what Huber and co-workers have done in a study described in Nano Letters1.

Ordinary microscopes cannot see things smaller than half the size of the characteristic wavelengths present in visible light, typically a few hundred nanometres. This limitation, called the diffraction limit, becomes a great obstacle to imaging microscopically small objects with invisible, far-infrared light (also called terahertz light; 1 THz is 1012 Hz), because the light can have wavelengths of hundreds of micrometres — a little bigger than the thickness of a human hair. However, using techniques that were originally developed at near- and mid-infrared wavelengths to circumvent this diffraction limit, the authors have demonstrated conclusively that, even at these long wavelengths, the properties of materials can be probed on a nanometre scale. With their technique, which they call terahertz near-field nanoscopy, Huber et al.1 have measured the concentrations of mobile electrons in a nanoscale transistor, resolving details as small as 40 nm, which is around 3,000 times smaller than the wavelength of the light used in their experiment.

Finding a way around the diffraction limit has kept scientists busy for the past few decades. To understand what the diffraction limit entails, imagine an object suspended in the air, and illuminated from behind by a light source. The presence of the object is betrayed by the shadow it casts on a wall some distance away, in the so-called far-field. It's tempting to think that if the object shrinks, the shadow on the wall will shrink in proportion to the cross-sectional area of the object. However, when the object becomes smaller than half the wavelength of the light, the shadow on the wall, in fact, disappears completely. This can be understood by remembering that light can be described as a set of waves propagating in free space. Whereas the waves are mostly blocked by large objects, they can easily bend around small ones and continue to propagate almost as if the object weren't there. This doesn't mean that waves are completely unaffected by a small object. They do, after all, have to bend around it and are thus strongly perturbed in its vicinity, just not far away from it.

In general, therefore, to see things that are smaller than half a wavelength one has to measure the light really close to the sample — that is, in the near-field region — using a very small detector. An alternative is to detect the light in the far-field while the light source, which in that case must be extremely small, is scanned very close to the sample. The latter is the approach taken by Huber et al. to break the diffraction limit at terahertz frequencies.

Their light source is a tiny metal tip, which tapers down to a diameter of about 30 nm, and is illuminated with a terahertz laser. This metal tip, acting as a lightning rod, collects some of the incident light and scatters it in all directions. A large portion of the scattered light is collected and measured with a sensitive detector. With some clever modulation and measurement tricks, the authors ensure that they observe light emitted only by the very end of the tip, which can thus effectively be viewed as a terahertz light source of nanometre dimensions. As it turns out, the amount of terahertz light scattered towards the detector by this source is strongly influenced by the physical properties of the sample underneath it. By scanning the tip across the sample, these physical properties, such as the concentration of mobile electrons, can be measured in unprecedented spatial detail.

Similar experiments at mid-infrared2 and microwave3 frequencies have been performed recently — indeed, every frequency region has its own appeal. With visible light, one can probe electronic transitions in atoms and molecules; with mid-infrared light, one can observe absorption by molecular vibrations and high-frequency lattice vibrations. So what is the added value of doing experiments at terahertz frequencies? Well, as Huber et al. demonstrate1, for the typical concentrations of mobile electrons in semiconductor devices, the near-field optical contrast is actually largest in the terahertz frequency domain (Fig. 1), allowing the authors to probe as few as 100 electrons. Needless to say, this technique will find applications in anything that shows a significant response to terahertz waves, such as semiconductors, superconductors and perhaps even the low-frequency vibrations of biological molecules. With further improvements, it might even be used to characterize single molecules or electrons.

Figure 1: Sharp eyes of the terahertz nanoscope.


Huber et al.1 image a section through a semiconductor device comprising multiple nanoscale transistor structures. a, Terahertz image obtained using light with a wavelength of 118 µm. b, Infrared image obtained using light with a wavelength of 11 µm. Seven transistors (peaks) are clearly identified in both images. However, below the transistors, the terahertz image shows bands that correspond to the different mobile-electron densities (n) present in the device. These bands, which are absent in the infrared image, show the gain in sensitivity obtained at terahertz frequencies. (Modified from ref. 1.)

Huber and colleagues' experiment is performed at a single terahertz frequency. This is a limitation, not of the method, but rather of the terahertz source used to illuminate the tip. It's more than likely that new terahertz sources, such as the terahertz quantum-cascade laser4 or broadband terahertz emitters5, which can deliver radiation in a range of terahertz frequencies, will at some point be used for terahertz 'nano-spectroscopy'.

The smallest feature that can be seen with terahertz near-field nanoscopy is currently determined by the dimensions of the tip apex. However, as tips become smaller, the non-zero penetration depth of terahertz light into the metal of which the tip is made will render the tip partially transparent at its thinnest end. It will be interesting to observe how this will affect the ultimate spatial resolution of the 'terahertz near-field nanoscope'.


  1. 1

    Huber, A. J., Keilmann, F., Wittborn, J., Aizpurua, J. & Hillenbrand, R. Nano Lett. 8, 3766–3770 (2008).

  2. 2

    Brehm, M. et al. Nano Lett. 6, 1307–1310 (2006).

  3. 3

    Ma, H. & Levy, J. Nano Lett. 6, 341–344 (2006).

  4. 4

    Köhler, R. et al. Nature 417, 156–159 (2002).

  5. 5

    van Exter, M. & Grischkowsky, D. R. IEEE Trans. Microwave Theory Tech. 38, 1684–1691 (1990).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Planken, P. A terahertz nanoscope. Nature 456, 454–455 (2008) doi:10.1038/456454a

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.