At the moment it is not possible to tell why the in-plane ordering of the films grown on Al2O3(1120) is inferior to that grown on the (001) MgO and Ge substrates. The difficulty with the Al2O3 growth may have to do with the 6° miscut of the substrates, or it may be due to an intrinsic difficulty with (011) growth of the Fe/Si multilayers. Previous work has shown that AF coupling in Fe/Si multilayers is dependent upon formation of a metastable iron silicide spacer layer phase.[4] The possibility exists that the spacer silicide does not grow well on Fe in the (011) orientation. This question can be answered only by further growth studies on better (011) substrates and careful structural characterizations.
A related question is whether the larger remanent magnet moment in the (011)-textured films might be due to a fundamental difference in magnetic properties from the (001)-textured films. Because a 46-repeat Fe/Si multilayer grown on Al2O3 has a remanence of only about 10%, this is unlikely. Undoubtedly the trilayer on Al2O3 has a higher remanence than the multilayer because the thinner film is more greatly impacted by the poor substrate surface quality. The staircase morphology caused by the 6° miscut of this Al2O3 substrate may lead to wavy interfaces between the Fe and iron silicide films or to pinholes through the silicide layers. Wavy interfaces can cause increased magnetostatic coupling or even biquadratic coupling,[11] both of which would tend to increase the remanence. The large remanence of the multilayers grown on glass substrates is likely also due to pinholes or magnetostatic coupling.
Pinhole-induced may explain the unusual temperature dependence of the remanence. (Magnetostatic coupling is expected to be approximately temperature-independent.) Fe atoms in bridges through the silicide spacer layers are expected have a reduced Curie temperature. A larger remanence at low temperature therefore makes sense if the remanence is derived from pinhole coupling and is not an intrinsic effect. Low Curie-temperature material may also be present in the iron silicide spacer layer or in at the iron/iron silicide interfaces.
By growing on a number of substrate materials and by using different deposition conditions, Fe/Si multilayers have been prepared with a varying degree of ordering. A large amount of accumulated evidence demonstrates that high remanence of the magnetization curves in Fe/Si multilayers is associated with interface roughness. The remanence is therefore not likely to be related to unusual exchange coupling but instead to originate from defects, perhaps pinholes through the silicide spacer layer. Since the remanent magnetization is caused by extrinsic effects, future studies should concentrate instead on measurements of the saturation field of the magnetization curves in order to learn more about the interlayer coupling.
We would like to thank E.E. Fullerton, J.A. Borchers, R.M. Osgood III and Y. Huai for helpful discussions, and B.H. O'Dell and S. Torres for technical assistance. Part of this work was performed under the auspices of the U.S. Department of Energy by LLNL under contract No. W-7405-ENG-48.