Challenges in the field of large-m supermirrors

Polarising and non-polarising supermirrors with angles of reflection θ_c of m ≃ 2 times the critical angle of Ni are now produced in large quantities mostly by using sputtering techniques. The coatings can be deposited with such perfection that the surface morphology of the substrates becomes as important for achieving high reflectivities as the perfection of the layer structure itself. As a consequence, m can be increased to ≃ 4 simply by increasing the number of layers to ≃ 1000. However, this requires that the individual thicknesses of the layers decrease too, approaching the thickness of the interdiffusion zones thus diminishing the reflectivity at large reflection angles. In addition, the total thickness of the coatings ends up to be of the order of typically 5 μm. Therefore, the forces on the substrate increase due to the stress within the layers leading finally to the destruction of the substrates. Clearly, such supermirrors are of no practical use. The stress within the layers can be significantly reduced by properly tuning the partial pressures of the reactive gases during sputtering. Moreover, by using anisotropic sputtering it is possible to induce an anisotropic stress distribution within the plane of the layers. As a consequence, it is possible to fabricate magnetic multilayers that exhibit a well-defined easy axis of magnetisation due to magnetostriction. This effect is particularly pronounced for polarising FeCoV/TiN_x supermirrors, such that remanent mirrors can be produced that allow the polarisation of neutrons in zero magnetic field.