UP3: Femtosecond x-ray diffraction

Bargheer, Zavoronkov, Gritsai, Wörner, Elsässer

Femtosecond X-ray diffraction combines the excellent accuracy in the determination of crystal structures with the time-resolution of modern femtosecond lasers. The short X-ray pulse takes snapshots of the crystal structure in motion, which is triggered by a sufficiently short pump laser pulse. By varying the time-delay between the pump pulse and the X-ray probe pulse we can assemble a molecular movie, and watch atoms, molecules and solids "in action" on their genuine atomic length-scale (Angstrom) and time-scale (Femtoseconds). This task challenges many researchers worldwide to generate ultrashort X-rays with tabletop setups similar to ours or at various large scale facilities. As the source developement proceeds, more and more complex structural changes can be obeserved. We have demonstrated the first X-ray diffraction experiment on a complex nanostructure with femtosecond time resolution.

Short movie: Swinging atoms in a nanostructure

The cartoon shows a small fraction of a semiconductor nanostructure. Gallium, Arsenic and Aluminum atoms form a crystal lattice composed of alternating layers of the semiconductors Galliumarsenid (GaAs) and Aluminumarsenid (AlAs). The layers are so thin that they cannot be seen with a light microscope. In fact, a stack of 2000 layers has a thickness of only 0.03 millimeter. In our experiment we hit this stack with a special laser pulse. Its energy is only absorbed in the Galliumarsenid layers.

All of them promtly start to expand. The Aluminumarsenid layers are squeezed from both sides, but they can swing back to their original size. We take snapshots of this motion back and forth with our x-ray camera. The cartoon exaggerates the motion. In reality we have observed amplitudes of the atoms which is 1000 times smaller than the interatomic distance. The atoms traverse this short distance in an ultrashort time. This is why our camera has to be extremely quick.

Femtosecond X-ray diffraction Setup

The divergent x-rays from the source can be collected by different kinds of x-ray optics, and can be focused onto the sample. This facilitates triggering the motion of the sample with the pump pulse, because only a smaller area has to be irradiated. In our pioneer experiment, however, we let the divergent x-rays directly shine on the sample and monitored the Bragg reflex on an x-ray CCD camera. Fig. B shows the 2D image on the CCD, which is vertically integrated to obtain Fig. C.

This so-called rocking curve of the sample shows the diffracted intensity as a function of the Bragg angle. In the stationary x-ray diffraction this is a standard method to determine the composition of multilayers. For our example the angular difference between maxima corresponds to the inverse superlattice period of 16 nm.

Timeresolved X-ray diffraction data

When we selectively excite a standing wave in the superlattice with a wavevector corresponding to the superlattice period (simply because that's the symmetry of the sample), it will oscillate with the frequency of the zone-folded phonon. The period is given by the time a sound-wave takes to travel through a superlattice unit cell. For our sample with 8 nm wells and 8 nm barriers this time is 3.5 ps. This standing wave has the same wavevector which creates the satellite peak in the rocking curve, and leads to a modulation of the diffracted intensity, which is plotted (Fig. on the left) as a function of the delay time between excitation pulse and x-ray probe pulse.