Ultrafast X-ray photography

Article metrics

A super-fast, lensless microscope has been developed that works by decoding the diffraction patterns of bright, laser-like flashes of X-rays. This advance should enable ultrafast events at the nanoscale to be recorded.

Stroboscopic photography has profoundly improved our understanding of both nature and technology, as illustrated by the beautiful snapshots of a milk droplet splashing, or a bullet piercing an apple, produced by the pioneer of the technique, Harold Edgerton. In Physical Review Letters, Ravasio et al.1 report that they have captured an image of a nano-object using a single burst of X-rays from the fastest strobe light in existence.

The most rapid natural events occur at the attosecond (10−18 seconds) or femtosecond (10−15 seconds) timescale, and must therefore occur in the nanoworld, as even light travels a distance of only 300 nanometres in 1 femtosecond. Because the spatial resolution of any camera or microscope is limited by the wavelength of the illuminating light, short bursts of light at very small wavelengths are needed to see ultrafast motion directly at the nanoscale. Pulses of X-rays would be ideal, for example for directly visualizing how the electrons that bind molecules together adjust as the molecule's structure changes during a chemical reaction or a conformational change; how the motion of a biological molecule relates to its function; or how rapidly magnetic materials can switch state in data-storage devices. But producing bright bursts of X-rays is technically demanding, so most ultrafast studies have used indirect approaches, such as spectroscopy, to study the fastest events. Furthermore, rapid bursts of electrons in an electron microscope have already been successfully used for ultrafast imaging at the nanoscale2,3.

The past decade, however, has witnessed an exciting convergence of advances that could transform our ability to visualize nanoscale events. First, 'lensless' diffractive imaging enables scientists to use coherent (laser-like and directed) X-ray beams to take pictures of an object, simply by shining a beam on the object and collecting the diffracted photons4; sophisticated computational algorithms then process the diffraction pattern to reveal the image. And second, new sources of light can generate coherent, ultrashort pulses of X-rays. For example, X-ray free-electron lasers (XFELs) produce high-energy X-ray beams5 that have pulse durations of 100 to 200 femtoseconds. But these stadium-sized facilities are expensive and exclusive: access is limited to one or a few users at a time, and they exist in only a few places worldwide. Moreover, the current pulse repetition rate of XFELs is limited to about 10 to 120 hertz.

On the other hand, table-top lasers can produce femtosecond pulses throughout the visible and infrared regions of the spectrum, and these can be directly converted into ultraviolet and low-energy X-rays using an extreme nonlinear process called high-harmonic generation6. In this process, an electron is first ripped from an atom by the laser field, then accelerated away from the resulting ion and back again. The electron accumulates kinetic energy during its journey, which is liberated as an X-ray photon when the electron recombines with the ion. The resulting X-ray beams can be fully coherent under certain conditions, and the pulse duration can span from tens of femtoseconds to less than 100 attoseconds. Although the energy per pulse is a thousand to a million times less than that produced by XFELs, the pulse repetition rate can be very high, ranging from 10 Hz to more than 50 kHz.

Lensless diffractive imaging has succeeded in reconstructing snapshots of virions, cells, nanostructures and plasmas at good resolutions using high-harmonic7,8, synchrotron-radiation9,10 and XFEL sources11. With high-harmonic sources7, images at a resolution of 50 nanometres have been obtained by exposing a nano-object to a pulsed beam of X-rays for 80 minutes, whereas images at a resolution of 120 nanometres required an exposure of only 30 seconds. In such experiments, data from many pulses must be combined to obtain a single image, because the intensity in each pulse is too low to provide an image on its own. Nevertheless, this approach can be used to follow nanoscale dynamics simply by varying the delay between a 'pump' laser pulse (which induces the dynamic behaviour) and probe pulses of X-rays.

In some situations, however, it is better to capture diffraction patterns in a single, high-intensity shot of X-rays — for example, if the object under investigation is destroyed by the experimental conditions. To date, single-shot X-ray imaging has been the exclusive realm of large XFEL facilities11,12, because only these sources produce pulses of sufficient intensity. XFELs have been used to capture dynamic motion11: a nano-object is first perturbed by a pump laser pulse, and then probed by single pulses of X-rays at different times afterwards. The resulting series of images, obtained at a resolution of 50 nanometres, show how the object changes over time in response to the pump.

Ravasio et al.1 now report that lensless diffractive imaging can be used to take a picture of a nano-object by means of a single, bright, ultrafast (20 femtoseconds) burst of X-rays generated using high harmonics (Fig. 1). The authors used a low-energy X-ray beam — which had a photon energy of around 40 electronvolts, corresponding to a wavelength of 32 nanometres — to obtain an image that had a resolution of 120 nanometres. The key to their success was using a bigger laser to generate sufficiently bright X-rays: the lens in the X-ray-production system had a focal length of 5.5 metres. The laser apparatus was therefore relatively large, but it was still much smaller than an XFEL. Furthermore, because the X-ray pulses produced by high harmonics are generated by an optical laser, they are perfectly synchronized to that laser. This makes it much easier to capture dynamically changing images at femtosecond time resolution using a pump–probe procedure than when using an XFEL source, in which the pump laser and the X-rays are not perfectly synchronized.

Figure 1: Decoding the notes.

Ravasio et al.1 report a method for obtaining stroboscopic flash images of nanoscale objects. a, An ultrafast pulse (exposure time, 20 femtoseconds) of laser-like X-ray beams generates a diffraction pattern when shone on a nanoscale object. b, The diffraction pattern is decoded, using computer algorithms, to produce an image of the object — in this case, micrometre-sized 'harmonic notes'. Each pixel is 59 nanometres wide. (Images taken from ref. 1.)

Ravasio et al.1 have so far obtained pictures only of test objects, but lensless, diffractive X-ray imaging holds much promise as a window into the fast-moving nanoworld. Indeed, ultrafast high-harmonic beams have already been used, for example, to follow indirectly how electrons and atoms couple together as a molecule changes shape13, how heat flows in nanostructures, and how electrons move between a catalytic surface and molecules adsorbed to that surface14.

There are many reasons to expect an explosion in the number of applications of ultrafast X-ray imaging. The energy range over which bright harmonic beams could be generated was previously limited to photon energies of less than 100 electronvolts (wavelengths greater than 13 nanometres). This range severely restricted the variety of objects that could be imaged. Fortunately, recent advances in high-harmonic light sources make it possible to create bright beams that span from the ultraviolet into the high-energy X-ray region of the electromagnetic spectrum (that is, to wavelengths of less than 1 nanometre)15. This will make it possible not only to substantially increase the variety of objects that can be imaged, but also to obtain much higher spatial resolution (potentially below 10 nanometres), essentially creating a super-resolution microscope that can image thick samples in three dimensions.


  1. 1

    Ravasio, A. et al. Phys. Rev. Lett. 103, 028104 (2009).

  2. 2

    Barwick, B. et al. Science 322, 1227–1231 (2008).

  3. 3

    Siwick, B. J., Dwyer, J. R., Jordan, R. E. & Miller, R. J. D. Science 302, 1382–1385 (2003).

  4. 4

    Miao, J., Charalambous, P., Kirz, J. & Sayre, D. Nature 400, 342–344 (1999).

  5. 5

    Arthur, J. et al. Linac Coherent Light Source (LCLS) Conceptual Design Rep. SLAC-R593 (Stanford, 2002).

  6. 6

    Kapteyn, H. C., Murnane, M. M. & Christov, I. P. Phys. Today 58 (3), 39–44 (2005).

  7. 7

    Sandberg, R. L. et al. Opt. Lett. 34, 1618–1620 (2009).

  8. 8

    Sandberg, R. L. et al. Proc. Natl Acad. Sci. USA 105, 24–27 (2008).

  9. 9

    Song, C. et al. Phys. Rev. Lett. 101, 158101 (2008).

  10. 10

    Shapiro, D. et al. Proc. Natl Acad. Sci. USA 102, 15343–15346 (2005).

  11. 11

    Barty, A. et al. Nature Photon. 2, 415–419 (2008).

  12. 12

    Chapman, H. N. et al. Nature 448, 676–679 (2007).

  13. 13

    Li, W. et al. Science 322, 1207–1211 (2008).

  14. 14

    Miaja-Avila, L. et al. Phys. Rev. Lett. 101, 046101 (2008).

  15. 15

    Popmintchev, T. et al. Proc. Natl Acad. Sci. USA 106, 10516–10521 (2009).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Murnane, M., Miao, J. Ultrafast X-ray photography. Nature 460, 1088–1090 (2009) doi:10.1038/4601088a

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.