What is the longitudinal magneto-optical Kerr effect?

We explore the commonly used classification scheme for the magneto-optical Kerr effect (MOKE), which essentially utilizes a dual definition based simultaneously on the Cartesian coordinate components of the magnetization vector with respect to the plane of incidence reference frame and specific elements of the reflection matrix, which describes light reflection from a ferromagnetic surface. We find that an unambiguous correspondence in between reflection matrix elements and magnetization components is valid only in special cases, while in more general cases, it leads to inconsistencies due to an intermixing of the presumed separate effects of longitudinal, transverse and polar MOKE. As an example, we investigate in this work both theoretically and experimentally a material that possesses anisotropic magneto-optical properties in accordance with its crystal symmetry. The derived equations, which specifically predict a so-far unknown polarization effect for the transverse magnetization component, are confirmed by detailed experiments on epitaxial hcp Co films. The results indicate that magneto-optical anisotropy causes significant deviations from the commonly employed MOKE data interpretation. Our work addresses the associated anomalies, provides a suitable analysis route for reliable MOKE magnetometry procedures, and proposes a revised MOKE terminology scheme.