Figure 3
(a) SEMPA (Refs.
19,
34) measurement of a stripe section at low temperature (10 K). The spin polarization is encoded by a gray scale. The measured image is 4000 pixels
wide and 5 pixels (21 nm) high. For a better inspection, the image has been stretched by a factor of 70 along the vertical direction. Within the spatial resolution of the experiment, this mesurement corresponds to five scans of the same scan line, which are displayed as successive lines in one image. Thermal drift causes some displacement between the scan lines. (b) SEMPA measurement of a stripe section at a high temperature (330 K). The measured image is 400 pixels
wide and 35 pixels (370 nm) high. This image is displayed in its real proportions. (c) Experimental spin profiles across one stripe at a low temperature (10 K,
, empty circles) and close to the transition temperature (330 K,
, black squares), as extracted from images (a) and (b), respectively. To obtain the profile, the scan lines are aligned to compensate for the thermal drift. Then, the aligned scan lines are averaged. To further reduce the noise level, these raw profiles are resampled by averaging a number of adjacent points. Both profiles have been normalized to the same width and amplitude. The horizontal error bars represent the spatial resolution
of the experiment. It is determined from the topographic profile of a sharp edge, which is obtained by the same procedure from a topographic image measured with exactly the same parameters as the magnetic image in (b). For the low-
profile, the horizontal error is on the order of the size of the data points. The vertical error bars are due to the statistical noise of the secondary electrons counted in SEMPA (Refs.
19,
34). Note that no traces of an in-plane component of the spin polarization can be found in simultaneously recorded images of the in-plane spin polarization. Details of the experiment will be published in a more extended review and are available on request.
Reuse & Permissions