Much of the recent scientific interest in magnetic multilayers has been driven by their potential use in practical device applications such as magneto-optical storage media. Following the discovery of large magneto-optical Kerr effects (MOKE) in U compounds by Reim and Schoenes[1], researchers at IBM attempted to make a room temperature ferromagnet film using uranium. They started to work on amorphous USb and UAs, which they found to be ferromagnetic with a Curie temperature, TC 140 K. When other elements, such as Fe, Mn, Cu were mixed into the film, TC did not increase. They then turned to the production of sputtered multilayer films of Co and UAs layers with the aim of inducing polarisation of the U through exchange coupling to the Co layers [2]. Among the limited initial studies, magneto-optical experiments showed that there was a moment on the uranium site at low temperature in all samples and even at room temperature for one multilayer. The difficulty to confirm these results unambiguously arises from the fact that the main part of the magnetic signal arises from the large Co moment, both within the multilayer and in the buffer material.

To further investigate these multilayers, X-ray magnetic circular dichroism (XMCD) measurements at the M4 (3726 eV) and M5 (3551 eV) edges of uranium have been performed, since such a study allows one to probe the 5f magnetism separately. XMCD measurements were made on three of the UAs/Co samples from IBM, namely 12 x [UAs(80 Å)/Co(20 Å)], 15 x [UAs(60 Å)/Co(20 Å)] and 20 x [UAs(40 Å)/Co(20 Å)] at beamline ID12.

The 5f3 magnetic moment deduced from the XMCD measurements, and the magnetisation value for UAs amorphous films, reported by Fumagalli et al. [2], of 0.7 ~ mB are in good agreement. Furthermore, this result is consistent with previous studies performed on uranium monopnictides strongly suggesting a U3+ (5f3) configuration. As depicted in Figure 122, we stress the direct relationship between the maximum polar Kerr rotation measured by Fumagalli et al. [2] at a temperature of 10 K and under a 3 T applied magnetic field and the 5f magnetic moment deduced from our XMCD experiment, performed at 35 K and 4 T. We note the linear relation between the two sets of data, confirming that the MOKE results are directly related to the magnetic contribution of the uranium.



Fig. 122: The 5f magnetic moment deduced from our XMCD measurements plotted versus the maximum polar Kerr rotation [2]for the three multilayers.



The aim of our study was to see whether the uranium moment contributes to the room temperature Kerr effect. For that, we concentrated on the 12 bilayered sample which exhibits the largest uranium magnetic moment and as a consequence the largest polar Kerr rotation. The U-M4 edge XMCD signal area has been measured as a function of temperature and field (Figure 123). Assuming the exchange field due to cobalt to be constant in the region of interest here, the variation of the 5f moment has been calculated within the mean field approximation. This model oversimplifies the physics of these multilayers, but shows that the applied field alone is not sufficient for the polarisation of the uranium; the role of the cobalt is clearly demonstrated. Saturation is achieved above ~ 2 T as shown by the field dependent measurement; a fortiori, 4 T applied along the direction used is enough to saturate the sample. The saturated uranium moment reaches ~ 0.85(15) µB.


Fig. 123: Temperature dependence for a 4 T applied field (left panel) and field dependence at 35 K (right panel) of the 5f magnetic contribution. The 5f moment calculated in a mean field scheme is shown as lines (both panels).



Finally, at room temperature we confirmed the presence of a small XMCD signal at the U-M4 edge. This corresponds to an induced moment on the uranium of ~ 0.05 µB, considerably lower than the 0.14 mB predicted by the IBM group [2]. Therefore, due to the strong spin-orbit coupling, even a small uranium contribution can lead to a sizeable magneto-optical effect.

[1] W. Reim and J. Schoenes in Ferromagnetic materials, K.H.J. Buschow, E.P. Wolfarth (Eds.), North Holland, Amsterdam, 5, 133 (1990).
[2] P. Fumagalli, T.S. Plaskett, D. Weller, T.R. McGuire and R.J. Gambino, Phys. Rev. Lett., 70, 230 (1993).

Principal Publications and Authors
N. Kernavanois (a,b), D. Mannix (c,d), P. Dalmas de Réotier (b), J.-P. Sanchez (b), A. Yaouanc (b), A. Rogalev (c), G.H. Lander (e), W.G. Stirling (c,d), Phys. Rev. B, 69, 054405 (2004).
(a) ILL
(b) CEA, DRFMC/SPSMS, Grenoble (France)
(c) ESRF
(d) Department of Physics, The University of Liverpool (UK)
(e) European Commission, Joint Research Centre, Institute for Transuranium Elements, Karlsruhe (Germany)