Abstract
In threshold photoemission electron microscopy (threshold PEEM), photoelectrons are excited by UV photons with an energy just above the photoemission threshold. The lateral intensity distribution of these electrons is then imaged by an electrostatic lens system.
In this thesis, the possibilities of threshold PEEM as a tool for magnetic domain
... read more
imaging are studied. The goal is the development of threshold PEEM as a technique for studying both ferromagnetic and anti-ferromagnetic domains.
The idea behind magnetic contrast as a primary contrast mechanism in threshold PEEM is the coupling between the angular distribution of photoelectrons inside the sample and the electromagnetic field of the transmitted light. Through magneto-optical effects, this electromagnetic field is influenced by the magnetic structure of the material. Threshold PEEM therefore in potential combines the well-studied magnetic contrast mechanisms from optical microscopy with the much higher resolution of an electron microscope.
A model for the magnetic contrast in threshold PEEM is developed, based on a combination of magnetically induced optical anisotropy and the three step model for photoemission. The model predicts the electron emission for a given geometry and polarization of the light from a number of material parameters. The model is applied to a non-magnetic, isotropic material with the parameters of Fe and to two magnetic materials. For the isotropic material, a large difference in electron emission is observed between excitation with p- and s-polarized incident light.
The first magnetic material modelled is anti-ferromagnetic NiO (001), which has magnetic linear birefringence as a result of a magnetostrictive lattice contraction. The second material is Fe0.97Si0.03 (001), which is ferromagnetic and has magnetic circular birefringence from spin-orbit coupling. In both cases, the effect of the birefringence on the optical transmission is relatively small. The main effect of the anisotropy is on the collection efficiency for the excited photoelectrons. The model predicts a value for the change in electron collection efficiency as a function of the orientation of the plane of incidence with respect to the magnetic ordering direction. This can then be translated into an expected asymmetry between different domains. For NiO the maximum expected contrast in a p/s ratio image is 0.7% for 90 degree rotated domains and 0.9% for 180 degree rotated domains. For FeSi it is 15% for domains with perpendicular magnetizations and 6% for domains with opposite magnetizations.
PEEM experiments are also performed on polished crystals of both materials. In
the case of NiO (001) strong contrast is observed, but a magnetic origin can be ruled
out. We ascribe this contrast to differences in O deficiency as a result of sample preparation.
This was supported by the behaviour upon oxygen exposure, and the disappearance
of the contrast after prolonged sputtering. An upper limit for the anti-ferromagnetic contrast
in NiO (001) of about 0:5% could be determined, based on the detection limit of our
setup. For FeSi (001) we find regular patterns in the p/s ratio images with an asymmetry
of (0.75 +/- 0.10)%, which are very likely magnetic domains.
show less