Using x ray magnetic linear dichroism (XMLD) for magnetometry requires detailed knowledge of its dependence on the relative orientation of polarization, magnetic moments, and crystallographic axes. We show that Ni{sup 2+} L{sub 2,3} XMLD in cubic lattices has to be described as linear combination of two fundamental spectra - not one as previously assumed. The spectra are calculated using atomic multiplet theory and the angular dependence is derived from crystal field symmetry. Applying our results to Co/NiO(001) interfaces, we find perpendicular coupling between Ni and Co moments.

Systematic error in the Long Trace Profiler (LTP) has become the major error source as measurement accuracy enters the nanoradian and nanometer regime. Great efforts have been made to reduce the systematic error at a number of synchrotron radiation laboratories around the world. Generally, the LTP reference beam has to be tilted away from the optical axis in order to avoid fringe overlap between the sample and reference beams. However, a tilted reference beam will result in considerable systematic error due to optical system imperfections, which is difficult to correct. Six methods of implementing a non-tilted reference beam in the LTP are introduced: (1) application of an external precision angle device to measure and remove slide pitch error without a reference beam, (2) independent slide pitch test by use of not tilted reference beam, (3) non-tilted reference test combined with tilted sample, (4) penta-prism scanning mode without a reference beam correction, (5) non-tilted reference using a second optical head, and (6) alternate switching of data acquisition between the sample and reference beams. With a non-tilted reference method, the measurement accuracy can be improved significantly. Some measurement results are presented. Systematic error in the sample beam arm is not addressed in …