No abstract

The magnetic and magnetotransport properties of several series of sandwiches consisting of two ferromagnetic layers (Ni, Co, Ni80Fe20) separated by a noble metal (Cu, Ag, Au) are described. In order to vary the relative orientation of the magnetizations of the two ferromagnets, one of them was constrained by exchange anisotropy (e.g., NiFe/Fe50Mn50). The ferromagnetic layers are magnetically soft and not coupled antiparallel, giving very large changes of resistance at low fields. At room temperature relative changes ΔR/R of 4.1% in 10 Oe for Si/Ta 50 Å/NiFe 62 Å/Cu 22 Å/NiFe 40 Å/FeMn 70 Å/Ta 50 Å and 8.7% in 20 Oe has been obtained for a structure based on Co/Cu/Co layers. The magnetoresistance versus the thickness of the ferromagnetic layer shows a broad peak near 80 Å for Ni, Co and NiFe, demonstrating the importance of bulk rather than interfacial spin-dependent scattering, in contrast to Fe/Cr multilayers. The magnetoresistance decreases exponentially with increasing interlayer (Cu and Au) thickness, indicating that the magnetoresistance is due to the exchange of polarized electrons from one ferromagnetic layer to the other. The variation with Ag interlayer thickness is different for structural reasons.

No abstract

No abstract

We report an accurate method for the measurement of the saturation magnetostriction λs of thin conducting ferromagnetic films using spin valve structures. We describe an alternative to the cantilever beam method, called the bending method, which utilizes the inverse magnetostrictive effect. Typical soft ferromagnetic films exhibit hysteretic magnetization loops, which make the measurements of the magnetoelastic energy (hence λs) imprecise or erroneous. We show how to resolve the hysteresis problem by application of a transverse field. We also demonstrate the quantitative connection between the cantilever and bending method in a comparative study of films spanning a large magnetostriction range. The sensitivity of measurement of magnetostriction using the bending method is better than 1×10−7.