Sudden shifts sharpen X-ray pulses


Nuclear resonance techniques use only a narrow fraction of the X-ray spectrum. By squeezing off-resonant photons onto resonance via the precisely controlled motion of a resonant target, scientists have significantly increased the number of resonant photons available in an X-ray pulse. The enhanced intensity permits experiments that were previously unfeasible and further optimisation should lead to new experimental possibilities.

  • Share

Narrow X-ray resonances such as those of Mössbauer nuclei form the basis for a broad range of experiments based on high-precision spectroscopy. However, the spectra of the short pulses delivered by modern X-ray sources are orders of magnitude broader than the narrow resonances and only a tiny fraction of the available photons interacts resonantly with the sample, while the vast majority of the photons produce an off-resonant background. This restricts the achievable spectral, spatial and temporal resolution for commonly-used narrow resonances, and is a decisive factor against experiments with the sharpest available resonances.

To overcome this restriction, a novel method to amplify the spectrally-broad X-ray pulses of state-of-the-art X-ray sources in a narrow spectral region has been developed and implemented. Unlike monochromators, this method does not simply remove the undesired off-resonant background. Instead, off-resonant photons are converted into resonant ones, such that the number of resonant photons in the pulse is increased.

Schematic setup and X-ray pulse spectra

Figure 1. (a) Schematic setup. The incoming short X-ray pulses are sharpened via an iron absorber foil mounted on a piezo drive, and subsequently their spectrum is detected via a standard Mössbauer drive setup with a single-line analyser foil. (b) Experimental data together with theoretical fits. The figure shows the number of photons in an integration window of 60-168 ns after the excitation as function of the Mössbauer detuning of the analyser foil. (c) Actual piezo motion as a function of time, recovered from the experimental data. The indicated displacement of 0.43 Å is half the resonant X-ray wavelength. (d) X-ray pulse spectra behind the piezo “photon shovel” recovered from the experimental data, normalised to the input pulse spectrum. The black line shows the usual absorption dips of the magnetically-split nuclear resonance without motion. The blue curve shows the spectrum for the motion in (c), and clearly exhibits the conversion of absorption into enhancement on resonance. The dashed arrows indicate the spectral redistribution, which can be controlled via the piezo motion. Grey shaded areas indicate spectral regions from which photons were redistributed onto resonance.

Pictorially, the method works like a mechanical digger that shapes a hill out of a flat landscape by removing earth at the sides and piling it up. Analogously, the new method removes photons in off-resonant spectral regions, and piles them up in a narrow resonant spectral region. In the actual experiments performed at ID18, the Nuclear Resonance Beamline (ESRF) and P01, the High Resolution Dynamics Beamline (PETRA III, DESY), the role of the digger is taken by a piezoelectric material which performs specific motions upon precisely characterised and controlled electric signals, see Figure 1a. The “photon shovel” is formed by a thin iron absorber foil on the piezo, which temporarily stores X-ray photons while it performs its motion. The operation of the piezoelectric photon shovel exploits the Mössbauer effect: nuclei such as 57Fe used in this experiment, when embedded in a solid-state environment, can absorb and emit X-ray photons essentially without recoil.

The working principle of the photon shovel is based on interference effects: incoming resonant X-ray photons either pass through the sample without interaction, or are temporarily absorbed by the 57Fe foil (see Figure 1a). Without motion of the sample, the interaction with the 57Fe foil leads to a phase shift, such that the scattered photons destructively interfere with the non-interacting part. As a consequence, a dip due to absorption is observed in the spectrum. If instead the absorber foil is displaced by half the resonant wavelength after the non-interacting part has passed the sample, but before most absorbed photons have been reemitted, the destructive interference can be converted into constructive interference. In this way, the number of resonant photons is increased. The motion not only affects the resonant photons, but conversely also attenuates the X-ray pulse in a spectral range around the narrow resonance. Overall, the enhancement of the resonant photon number is thus achieved by shovelling (or squeezing) off-resonant photons onto the resonance; see Figure 1d.

In these experiments, the displacement of the absorber foil had to be controlled to less than a tenth of a nanometre, and took place within a few nanoseconds (Figure 1c). An enhancement by a factor of up to 4 was observed for the number of photons on resonance as compared to the incoming photon number. Theoretical calculations suggest that an enhancement by a factor of 10 should be feasible with an optimised absorber foil and corresponding motion, limited only by the additional electronic attenuation due to the photon-shovel material. In a subsequent experiment, the possibility for further enhancements with multiple shovels was successfully demonstrated.

In the future, this new technique could be deployed in the routine operation of beamlines at synchrotrons and free-electron lasers. The increased intensity results in shorter measurement times and enables measurements with presently too low signal rates. Also, the higher signal rates translate into better energy and spatial resolution. Furthermore, the mechanical control of X-ray-matter interactions could become a valuable tool in X-ray quantum optics, alleviating the need for additional X-ray control fields in certain settings. The technique also opens the possibility to track motions on atomic length scales in a wide range of scientific and technological applications.


Principal publication and authors
Spectral narrowing of X-ray pulses for precision spectroscopy with nuclear resonances, K.P. Heeg (a), A. Kaldun (a), C. Strohm (b), P. Reiser (a), C. Ott (a), R. Subramanian (a), D. Lentrodt (a), J. Haber (b), H.-C. Wille (b), S. Goerttler (a), R. Rüffer (c), C.H. Keitel (a), R. Röhlsberger (b), T. Pfeifer (a), J. Evers (a), Science 357, 375-378 (2017); doi: 10.1126/science.aan3512.
(a) Max-Planck-Institut für Kernphysik, Heidelberg (Germany)
(b) Deutsches Elektronen-Synchrotron DESY, Hamburg (Germany)
(c) ESRF