Multichannel spinpolarisation detector for electron spectrometers

M. Kolbe, B. Petereit, P. Lushchyk, G. Schönhense
Johannes Gutenberg-Universität Mainz, Institut für Physik, D-55099 Mainz

This project explores a way to a highly efficient spin-resolved photoemission technique. All present spin polarisation detectors work “single channel”, i.e. monoenergeticly and at a single detection angle. Thus, only a small fraction of the electrons behind the spectrometer are spin analysed, although modern hemispherical analysers provide a large energy and k- (i.e. angular) interval in their dispersion plane. Spin polarisation spectroscopy is thus characterised by a drastic reduction of signal intensity. Many problems, e.g. in the area of magnetic materials, need a combination of bulk sensitivity (HAXPES Hard X-ray Photoelectron Spectroscopy) and/or high energy resolution and spin detection. The corresponding losses in detection efficiency of about 4 orders of magnitude (figure of merit P2I) lead to unreasonably high data accumulation times.

In the present project we develop a new generation of spin polarisation detector that works “display-like”, i.e. energy dispersive and, in a second detection plane also angular dispersive. This facilitates parallel analysis of the spin polarisation distribution in the dispersive plane. In turn, the measuring time can be reduced by 4 orders of magnitude. Fig. 1 shows a sketch of the principal set up (schematic). The bundle of electron rays in the dispersive plane is focused to the spin-analyser crystal (transfer optics not shown in Fig. 1) and the scattered rays are imaged onto a spatially resolving detector, here a Delayline detector (Surface Concept GmbH). In the non-dispersive plane an angular-dispersed bundle of rays is scattered and imaged in the same way. The conservation of the electron momentum component parallel to the crystal surface, inherent to the (0,0) beam, guarantees a transfer of the two-dimensional lateral image information that is encoded in the scattering coordinates and angle.

 

geometry of the multichannel spin polarisation detector

Fig. 1: Schematic geometry of the multichannel spin polarisation detector behind a hemispherical energy analyser. The rays in the dispersive (left) and non-dispersive plane (right) are imaged in a “display-like” mode onto the 2D electron detector (Delayline detector).

 

Photograph of the multichannel spin polarisation detector during assemblyFig. 2: Photograph of the multichannel spin polarisation detector during assembly. The unit contains the transfer optics before and behind the analyser crystal. The crystal is positioned on a Cardan mount.


 

The momentum scheme for the scattering process is shown in Fig. 3(b), where we have indicated the interference structure along k. Assuming an inner potential of 9 eV and accounting for a compression of the top layer spacing by 6% [1] yields a position of the first interference maximum at an impact energy of 16 eV. This means that the observed maximum at 25 eV results from multiple scattering. Because of spin-orbit coupling, the majority and minority electrons see different potentials, leading to different scattering amplitudes which govern the asymmetry function [2]. The figure reveals that a variation of E at fixed Θ leads to a variation of the size of the Ewald sphere and hence, to a modification of the interference condition (intensity and asymmetry) as observed experimentally. An analogous argument holds for a variation of the angle. The optimum working point for W(100) is at a scattering energy E = 26 eV. There, we find a large reflectivity of 0.012 and an asymmetry of S = 0.43. The energy dependent asymmetry function and intensity response were measured at the Max-Planck-Institute for Micro Structure Physics (Halle)[3].

 

Figure 3: (a) Geometry of the scattering process for simultaneous acquisition of 16 data points for the idealized case of a parallel beam and (b) corresponding momentum scheme (Ewald construction) for specular reflection [(0,0) beam] for W(100) at 45° and 26 eV. A variation of E by 10% modifies the scattering condition (red).

 

 

Figure 4: (a) Two-dimensional electron intensity distribution (false color map) in the exit field of the electron spectrometer, imaged after reflection at the spin filter crystal. The spectrum corresponds to the 2D photoelectron distribution of an iron film near the Fermi edge (hv = 21.23 eV). (b) Corresponding asymmetry pattern calculated from two measurements with opposite magnetization directions of the iron film. The background was corrected linearly, the intensity variation in the dependence of Θ was not corrected. (c) Line scans of the asymmetry for different angular coordinates Θj.

 

Figure 4 shows first experimental results of photoelectron spectra recorded from a clean Fe (100) sample excited by He I (hν = 21.23 eV) radiation. Figure 4(a) shows the lateral electron distribution pattern of photoelectrons emitted from the Fe(100) surface after being dispersed by the hemispherical energy analyzer and after scattering at the spin filter crystal. The energy-dispersive and angular resolving directions correspond to the axes Ei and Θj in Fig. 3(a), where indices i and j number the resolvable data points. The DLD registers each counting event with respect to its coordinate (Ei, Θj) yielding the accumulated count rate Iij. The intensity cutoff on the right-hand side of Fig. 4(a) represents the Fermi edge (marked by the vertical dotted line) which could be displayed sharply, validating that no significant distortion of the ray bundles occurs during the diffraction process. Figure 4(b) shows a typical asymmetry distribution pattern of Fe(100). It is calculated pixel-by-pixel via

,

where Iij± are the accumulated count rates for the sample magnetized up/downwards. The asymmetry is negative at the Fermi edge, turning to positive values at higher binding energies. The horizontal stripe pattern originates from a fine stripe mesh placed in the beam path of the analyzer optics. It is possible to resolve 16 bright and 16 dark (shadow) stripes with 2 image points each, leading to a minimum number of 64 resolvable angular intervals presently. Together with 15 resolvable energy points, this leads to a very high figure of merit (2D) of

.

 

Results are published in:

Significant spin polarization of Co2MnGa Heusler thin films on MgO(100) measured by ultraviolet photoemission spectroscopy
M. Hahn, G. Schönhense, E. Arbelo Jorge, and M. Jourdan,
Appl. Phys. Lett. 98, 232503 (2011), DOI

Highly Efficient Multichannel Spin-Polarization Detection
M. Kolbe, P. Lushchyk, B. Petereit, H. J. Elmers, G. Schönhense, A. Oelsner, C. Tusche, and J. Kirschner,
Phys. Rev. Lett. 107, 207601, DOI

Test of band structure calculations for Heusler compounds by spin-resolved photoemission spectroscopy
M. Kolbe, S. Chadov, E. Arbelo Jorge, G. Schönhense, C. Felser, H.-J. Elmers, M. Kläui, and M. Jourdan,
PRB  86 (2012) 024422  DOI

References:

[1] R. Yu, H. Krakauer, and D. Singh, Phys. Rev. B 45, 8671 (1992)
[2] J. Kessler, Polarized Electrons (Springer, Berlin, 1985), 2nd ed.
[3] C. Tusche, M. Ellguth, A. Ünal, C.-T. Chiang, A. Winkelmann, A. Krasyuk, M. Hahn, G. Schönhense, and J. Kirschner, Appl. Phys. Lett. 99, 032505 (2011)

 

 


 

Last update: Tuesday, 24-Sep-2013 13:53:47 CEST Email D. Panzer Impressum