Layout of a two-color pump–probe experiment at FLASH. The FEL and IR laser pulses hit the sample with an adjustable time delay Δt. A sample (shown is a area of a magnetic multilayer sample imaged by magnetic force microscopy) produces small-angle scattering, which is detected by a CCD camera.
Schematic of the SAXS setup for use at FLASH beamlines BL2 and BL3. The FEL beam direction is along the x direction and the IR laser beam enters along negative y direction. For clarity, the turbo pump attached at the bottom of the sample chamber, the vacuum gauge attached to the timing cross, and the x and y motors for the absorber and sample translators are omitted. Dimensions are given in millimeters. The measures denote the lengths of the components (lower) and (upper) the distance of the sample from the FEL exit-flange as well as the sample–detector distance. See Fig. 3 and text for details.
Schematic of the chamber. The FLASH beam (fine line) enters from the lower right in x direction and the IR-Laser, represented by the red cone enters from the front. In FEL beam direction the first cross contains the solid-state absorbers (Table I ). The second cross houses a metal mirror to reflect the IR laser beam quasi-collinearly to the FEL beam onto the sample. The third cross contains the timing device and the vacuum gauge (not shown for clarity). The following adapter part contains a pneumatic shutter to protect the CCD camera during timing measurements and general alignment. The sample chamber (CF100 cube) contains besides the sample holder a guard hole and the beamstop. The following tube houses the CCD camera. Its chip is positioned from the sample in the shown configuration. See text for details.
Complete setup as used at BL3. Here, the chamber is shown mounted onto the base frame with all additional parts including the optical incoupling setup. See Figs. 2 and 3 to identify the various parts of the chamber.
Photograph of a sample holder with one multi-window sample (left), one adapter plate for smaller samples (right), and a fluorescence screen (center).
Magnetic scattering image from a cobalt/platinum multilayer sample as described in the text. For the exposure, 3 single FEL shots were summed at low repetition rate ( ) and no pump laser was used. Quantitative analysis is done by radial integration after masking the beamstop and charge scattering areas as indicated. From the resulting S(Q) distribution real space parameters of the magnetic system can be extracted. See text for details.
Results from a pump–probe experiment with interleaved unpumped references measurements averaging 100 FEL shots at a repetition rate of for each data point. Evaluation has been done according to Fig. 6 and some typical error bars are given. Unpumped reference and pumped data have been taken in alternating order and with increasing delay time (top axis) as indicated by arrows in (a). In terms of the dynamics in question negative delay times imply an effectively unpumped sample, as the probe pulse impinges on the sample before the pump pulse. Even at negative time delay a clear difference in the scattering peak intensity I is observed. This can only be explained by a static heating effect due to the repeated energy deposition by both, the pump and probe pulses. (b) The peak width w undergoes an ageing effect particularly during the first 4 measurements, i.e., the domain-width distribution becomes narrower. See text for details.
At a reduced repetition frequency of , the heating effect observed when using the full repetition rate of (Fig. 7(a) ) is no longer significant. Data points have also been taken in an alternating unpumped/pumped order.
Magnetic scattering data from the same sample in (a) unpumped and (b) pumped condition. The delay time between pump and probe laser was set to and the pump fluence was . In both cases, three FEL shots were averaged at repetition rate to minimize the heating effect (Fig. 7 ). Speckles cannot be resolved due to the large illuminated area and the short sample-detector distance.
Pump flux dependent ultrafast demagnetization measured at FLASH. The height of the magnetic scattering peak has been determined by radial integration over the CCD images (Fig. 9 ) each of which in an average over three FEL shots. The solid lines represent fits using Eq. (4) . The shift in time zero of about between the measurements is caused by the drift in the synchronization of FLASH (see Sec. II ). The momentary value of time zero has been accounted for while fitting via a simple shift in t.
CCD image (a) and line integration along the Q x and Q y directions (b) and (c) for a strongly pumped sample (120 mJ cm−2 not taking into account the lateral variation of the intensity). The structured pump laser has induced changes in the sample which are aligned in a grid. Therefore, the scattering also shows a grid-like structure and the periodicity yields the pump laser's wavelength. The high-intensity streaks intersecting at are scattering from the membrane window. At higher magnification, fine dark lines become visible which are an artefact from the Ni mesh supporting the aluminum filter used in this experiment. See text for details. (d) The sample is irradiated from the groove side such that the pump-laser radiation is structured on the sample surface.
Aluminum absorber thicknesses used for a photon energy of . The absorbers have an opening of in diameter. Values are calculated from data in Ref. 27
Article metrics loading...
Full text loading...