Exchange bias systems studied by high resolution quantitative magnetic force microscopy

Abstract

It is generally believed that exchange bias (EB) implies the presence of pinned uncompen- sated moments pin-UCS in the antiferromagnet (AF) layer that are coupled to the ferromagnet (F) layer. An obstacle to understanding the EB e�ect is that only a subset of the UCS (those pinned and coupled to the F) are responsible for the EB-e�ect. The materials used, but also the experimental method and preparation may a�ect these subsets of UCS in distinct ways [19], and an interpretation of UCS measurements must take this into account. Moreover, the materials morphology, texture, defect density and nature of grain boundaries in uence the density and spatial distribution of the pin-UCS. Experimental methods that measure the pin-UCS density distribution with spatial resolution comparable to the materials' grain size are needed. Here we study F/AF heterostructure-samples by VSM and quantitative, high resolution MFM. MFM works in magnetic �elds (up to several T) but is not element speci�c. Analyzing data acquired with the F-layer in the saturated state and with di�erent magnetization states of the tip allows the separation of the di�erent sources of MFM contrast. Using quantitative MFM we measure the local areal density of pinned uncompensated moments (pin-UCS) in the antifer- romagnetic (AF) CoO layer and correlate the F-domain structure in a perpendicular anisotropy CoPt multilayer with the pin-UCS density [15]. Larger applied �elds drive the receding domains to areas of proportionally higher pin-UCS aligned antiparallel to F-moments. This con�rms our prior results [19] that these antiparallel pin-UCS are responsible for the EB-e�ect, while parallel UCS coexist. The data con�rm that the evolution of the F-domains is determined by the pin-UCS in the AF-layer, and also present examples of frustration in the system. This frus- tration and the inhomogeneous spatial distribution of the pin-UCS also have a major e�ect on the coercivity of the EB-systems that has not yet been acounted for. Moreover, grain-boundary engineering can be used to decouple the AF grains leading to a stronger EB-e�ect but a smaller coercivity. New approaches with rare-earth-ferrimagnet/ferromagnet bilayers to increase unidirectional anisotropy provided by the EB-e�ect will be discussed.