FMR and SMOKE studies of bilinear and biquadratic exchange coupling in bcc 
Fe/Cu, Ag/Fe trilayers. The effect of in-plane four-fold anisotropies on 
magnetization reversal.

B. Heinrich, Z. Celinski and J.F. Cochran in Nanomagnetism ed.: A. 
Hernando (1993, Kluwer) pp. 33-48.

Surface Magneto-Optical Effect (SMOKE) and Ferromagnetic Resonance (FMR) 
studies of the exchange coupling in bcc Fe/Cu/Fe(001) and Fe/Ag/Fe(001) 
structures are presented. It is shown that the interfaces in bcc 
Fe/Cu/Fe(001) and Fe/Ag/Fe(001) trilayers can be signifcantly imporved by 
choosing an apropriate growth procedure. The exchange coupling in bcc 
Fe/Cu/Fe and Fe/Ag/Fe trilayers was studied as a function of interlayer 
thickness. The intepretation of magnetization loops for Fe/Cu/Fe and 
Fe/Ag/Fe trilayers requires simultaneous presence of bilinear and 
biqadratic exchange coupling between magnetic layers. The measured values 
of the bilinear biquadratic exchange coupling in Fe/Cu/Fe were compared 
with a model recently proposed by Slonczewski, which trats the exhanger 
coupling in trilayers with imperfect interfaces. It is shown that the 
exchange coupling unobscured by interface roughness exhibits a strong 
short wavelength oscillatory behavior which is in agreement with recent 
first principles band calculations. The imporved interfaces in the 
Fe/Ag/Fe system revealed new features in the exchange coupling which were 
absent in samples grown entirely at room temperature. FMR and SMOKE 
measurements showed symmetric long wavelength oscillation (lambda~5 ML) in 
the exchange coupling. Magnetization reversal was studied in Fe/Ag/Fe/Ni 
trilayers. One can grow structures in which the magnetization reversal 
proceedsi n two steps only. A samll negative field of ~-15 Oe was needed 
to reverse the magnetization in the Fe film. A significantly larger field 
of several hundred Oe corresponds to switching the magnetization of the Fe 
film from parallel to the antiparallel configuration with respect to the 
magnetic moment of the Fe/Ni film.