Magnetically sensitive X-ray measurement techniques
X-ray Magnetic Circular Dichroism (XMCD)XMCD is the X-ray equivalent of the magneto-optical Kerr effect; the quantity measured is the dependence of the absorption on X-ray helicity close to a magneticallysensitive resonant absorption edge. An example is the spin-orbit-split 2p→3d transitions (L2 and L3 edges) of the transition-metal ions. The origin of X-ray magnetic dichroism is explained for a simple example in the Infobox. In contrast to the magneto-optical Kerr effect, XMCD involves a localized initial state, and hence provides chemical and orbital selectivity. If many-body effects can be neglected, XMCD sum-rules can yield both the orbital and spin magnetic moments µL,µB,of the resonant ion (see Fig. I.2). Note that the XMCD contrast is a strong effect: at resonance, the dichroic X-ray interaction per atom can exceed that of 40 electrons! Application of the XMCD technique at the SwissFEL will profit from flexible wavelength tuning, and ultrafast pumpprobe measurements will follow the individual time evolutions of the spin and orbital moments in the sample.
Magnetic-contrast holographyWith the strong XMCD contrast of resonant magnetic scattering and the high coherence of the SwissFEL, it is possible to per form time-resolved lensless magnetic imaging (see also Chapter III). An elegant X-ray holographic method for thin magnetic films has been developed by Eisebitt et al.  (Fig. I.3). Since the experiment is per formed at a synchrotron, spatial-filtering is required to produce transverse coherence, and a monochromator is used to provide sufficient longitudinal coherence.
The magnetic thin-film sample, which spans a hole in the sample support, transmits the resonant, circularly-polarized object beam, and close to the sample is a smaller hole, through which an undisturbed reference beam is transmitted. The two beams inter fere on their way to the CCD area detector, where they produce a holographic image, and a simple Fourier transform is sufficient to reconstruct a real-space image of the sample. The spatial resolution of the technique is presently limited by the diameter of the reference beam aper ture (100 nm). But with enhanced data analysis techniques, it should be possible to approach the diffraction limit given by the soft X-ray wavelength (1.6 nm at 775 eV). Due to the limited coherent flux currently available from 3rd generation synchrotrons, collection of a single image requires several minutes of exposure time. With the high peak intensity of the SwissFEL, it will be possible to collect a holographic image in a single shot. Considering that the sample is a thin membrane, this may be a destructive experiment, requiring a fresh sample for each successive measurement. One could then envisage repeated, single-shot pump-probe investigations, on similarly- prepared samples, of sub-ps magnetic behavior triggered by laser or THz pump pulses. Although fine sample-dependent details may vary from shot to shot, the basic switching process remains the same.
Fig. I.3. The magnetic holography method of Eisebitt et al. [4,5]. As described in the text, synchrotron light, filtered to introduce coherence, passes through either the sample or the reference pinhole and inter feres to produce a holographic image. This hologram is Fourier-transformed to reconstruct the real-space object, in this case a collection of magnetic squares.
Spin-polarized time-of-flight (TOF) photoelectron spectroscopyA direct measurement of the electron polarization can be per formed by using photoemission, where electrons are ejected from the sample sur face, and where spin analysis is per formed, e.g., via Mott scattering. For sufficiently high photon energies, band-structure-dependent final state effects can be neglected.
The photoelectric cross-sections for the magnetic sub-shells of magnetic elements peak in the photon energy range 30–100 eV, which will be accessible by the SwissFEL. Spin-analysis using Mott detectors is highly inefficient, and laser-induced magnetic phenomena can be extremely fast (sub-ps). Thus the high peak intensity and the shor t pulse duration of the SwissFEL make it ver y attractive for such measurements. Fur thermore, with a pulsed photon source, a simple electron time-of-flight (TOF) detector can be used to provide energy resolution (see Fig. I.4). One concern is that the high peak intensity of the SwissFEL will briefly produce a high density of low-energy photoelectrons near the sample sur face, causing space-charge effects, which may distor t the photoelectron energy distribution. However, since the interaction is to a good approximation electrostatic, the electron spin-polarization will not be affected.
Fig. I.4. Schematic plan of a spin-polarized time-of-flight (TOF) photoelectron spectrometer. Energy resolution is provided by TOF detection. Because the Mott-scattering spin analyzer can simultaneously detect both spin orientations, true single-shot operation is possible
The three-temperature model of magnetismThe three-temperature model (Fig. I.i2) for magnetizaton dy-namics  has been invoked to help explain the sub-ps demagnetization observed by Beaurepaire et al. . The model assumes that the electron motion (el), the electron spins (mag) and lattice phonons (lat) represent three mutually-interacting thermal reservoirs. The internal equilibration of the resevoirs occurs at the plasma frequency (el), the spin-wave frequency (mag) and the phonon scattering rate (lat). Furthermore, the reservoirs interact with one another via the electron-phonon interaction (el-lat), the spinlattice relaxation rate (mag-lat) and the spin-orbit interaction (el-mag).
Note that the 3-temperature model is purely thermodynamic in nature and hence fails to account for the wavelength of the excitation. With a tunable source such as the SwissFEL, the depence of bath temperature on wavelength can be investigated .
Fig. I.i2. A schematic diagram of the three-temperature model, showing the time constants for equilibration and interaction of the reservoirs. Fundamental criticism of the model may be clarified with ultrafast pump-probe experiments at the Swiss- FEL.