The time-resolved hard X-ray diffraction endstation KMC-3 XPP at BESSY II

Abstract

The time-resolved hard X-ray diffraction endstation KMC-3 XPP for optical pump/X-ray probe experiments at the electron storage ring BESSY II is dedicated to investigating the structural response of thin film samples and heterostructures after their excitation with ultrashort laser pulses and/or electric field pulses. It enables experiments with access to symmetric and asymmetric Bragg reflections via a four-circle diffractometer and it is possible to keep the sample in high vacuum and vary the sample temperature between ∼15 K and 350 K. The femtosecond laser system permanently installed at the beamline allows for optical excitation of the sample at 1028 nm. A non-linear optical setup enables the sample excitation also at 514 nm and 343 nm. A time-resolution of 17 ps is achieved with the `low-α' operation mode of the storage ring and an electronic variation of the delay between optical pump and hard X-ray probe pulse conveniently accesses picosecond to microsecond timescales. Direct time-resolved detection of the diffracted hard X-ray synchrotron pulses use a gated area pixel detector or a fast point detector in single photon counting mode. The range of experiments that are reliably conducted at the endstation and that detect structural dynamics of samples excited by laser pulses or electric fields are presented.

Keywords: beamline instrumentation; ferroelectric switching; optical excitation; thermal transport; time-resolved X-ray diffraction.

Figure 1

Layout of the KMC-3 beamline and the XPP endstation at BESSY II. (a) The separated single bunch (indicated by the enlarged dark cyan circle within the ion clearing gap) circulates around the BESSY II storage ring and emits X-ray pulses sketched in dark blue. The divergence of the emitted X-ray beam is parallelized by mirror 1, the wavelength is selected by a Si (111) double-crystal monochromator, and the X-ray beam is finally focused onto the sample by mirror 2. The X-rays diffracted from the sample are counted by a gated pixel area detector or scintillator with photomultiplier. The bunch marker signal at 1.25 MHz from the storage ring trigger is used to synchronize the pump laser pulses (shown in dark red). A function generator (FG) triggered by the laser provides a voltage pulse sequences for sample excitation (indicated in orange), and the gate pulse for the detector synchronized to the synchrotron frequency, ν. An electronic delay unit allows shifting of the laser timing τ_L with respect to ν; the same is possible for the delay τ_V of the voltage sequence for ferroelectric switching studies as indicated by the orange voltage pattern. (b) Top view of the experimental hutch for the time-resolved diffraction experiments. The inset shows the alternative generation of second and third harmonic laser light. ‘POL & λ/2 WP’ indicates the polarizer motorized half-wave plate combination that changes the incident laser pump power on the sample.

Figure 2

Side view and laser beam path in the diffraction chamber. The sample is mounted inside a vacuum chamber on a cryogenically cooled sample holder attached to the three-circle goniometer; the detector is mounted on the 2θ arm moving outside of the evacuated chamber (not shown here). The pump laser is coupled into the chamber through a quartz window attached to the evacuated beam tube. The last lens is motorized and used for the optimization of the spatial overlap of X-ray and laser spot on the sample (lenv, lenh); the laser spot size on the sample is adjusted by the movement of the second lens (lenf). A power meter and beam profiler allow to monitor the laser beam parameters. The laser and the X-ray beams have an angular offset of ∼2°.

Figure 3

Setup for the application of voltage pulse sequences to FE test devices. The sample is glued on a conducting metal holder that is connected to an oscilloscope for the measurement of the FE switching current. A W needle establishes the electrical contact to a metallic top electrode (see also the inset). In order to change between different top electrodes, the needle can remotely lifted up and the sample moved relatively to the needle with the x, y stages. The plastic needle holder movement is remotely operated and its lowering motion is damped with an induction coil.

Figure 4

Time-resolved optical pump/X-ray probe experiments of a 157 nm-thick SrRuO₃ layer on a SrTiO₃ substrate at T = 300 K. (a) Integration of the reciprocal space maps along q _x showing the temporal evolution of the out-of-plane lattice vector q _z for the SrRuO₃ 002 lattice vector after excitation with λ = 514 nm at the repetition rate ν = 104 kHz with fluence F = 5 mJ cm⁻². The vertical gray arrow indicates the peak position without laser heating. The inset of (a) is the RSM of the measurement at τ_L = +140 ps for λ = 514 nm. (b) Strain η as function of the delay τ_L induced by the optical excitation of the SrRuO₃ layer with ultrashort laser pulses with 1028 nm wavelength (red circles) and 514 nm (open blue squares).

Figure 5

Time-resolved measurement of a SrRuO₃ thin film after laser excitation in low-α operation mode of BESSY II. (a) Time-resolved strain η (filled blue symbols) in a 157 nm-thick layer after optical excitation at room temperature with 1028 nm using a fluence F = 2.5 mJ cm⁻² as a function of delay τ_L. A simulation of the transient strain using the udkm1Dsim toolbox (Schick et al., 2014a ▸) is shown by the dashed blue line. Its convolution with a time-resolution of 17 ps (FWHM) is shown by the blue solid line. The pulse length of the X-ray pulses is indicated by the filled red Gaussian pulse at τ_L = 0. (b) RSM of the 002 Bragg reflection of SrRuO₃. The color plot is a snapshot at τ_L = −40 ps before laser excitation and the thin lines on top are measured at τ_L = +40 ps.

Figure 6

(a) Room temperature data of the strain measured in Dy in low-α mode reveal the acoustic oscillations of the Dy film. The inset shows a simulation of the strain in the heterostructure where the color represents the strain: blue encodes compressive strain, red expansive strain. Whenever the average strain in the Dy layer is large, the measurement shows a maximum. (b) Measured strain in the Dy layer as a function of the sample temperature. For temperatures below the Néel temperature a remarkable transient negative thermal expansion is observed. (c) Transient peak width change of the dysprosium peak as a function of the sample temperature.

Figure 7

(a) ω scan of the 111 Bragg reflection of Au nanotriangles on a bare Si substrate (filled blue symbols) and a Si substrate that has been functionalized prior to the NT deposition (open red symbols). In both cases, the Laue oscillations yield a thickness of the Au NTs of ∼5 nm. (b) Temporal evolution of the strain in the Au NTs after laser excitation with λ = 1028 nm and F = 0.1 mJ cm⁻².

Figure 8

Simultaneously measured electrical and structural response of a PZT thin film as a function of the delay between onset of the electric field and the X-ray probe, τ_V. In (a) we show the switching current and overlaid with gray shading the applied PUND voltage sequence that indicates when the field is applied to the sample. The strain η obtained from the 002 Bragg reflection is shown in (b) and the intensity of the Bragg reflection is shown in (c). The width as function of the delay τ_V is plotted in (d) and (e) for the in-plane and out-of-plane components, respectively. The dashed gray vertical lines indicate the beginning of the PUND pulses.