Co/Ni multilayers for spintronics: High spin polarization and tunable magnetic anisotropy

In this paper we analyze, in detail, the magnetic properties of (Co/Ni) multilayers, a widely used system for spintronics devices. We use spin-polarized photoemission spectroscopy, magneto-optical Kerr effect, x-ray magnetic circular dichroism (XMCD), and anomalous surface diffraction experiments to investigate the electronic, magnetic, and structural properties in [Co($x$)/Ni($y$)] single-crystalline stacks grown by molecular-beam epitaxy. The spin polarization depends sensitively on the surface termination and for Co terminated stacks is found to be much larger than bulk Co, reaching at least 90% for 2 Co atomic planes. We observe a magnetization transition from in plane to out of plane when varying the Ni coverage on a Co layer in the submonolayer range, confirming the interface origin of the perpendicular magnetic anisotropy in this system. Angle-dependent XMCD using strong applied magnetic field allows us to show that the orbital magnetic moment anisotropy in Co is responsible for the anisotropy and that our results are consistent with Bruno's model. Surface x-ray diffraction shows that fcc stacking is preferred for 1-monolayer Co-based superlattices, whereas the hcp stacking dominates for larger Co thicknesses. We finally explored the role of the stacking sequence on the Co and Ni magnetic moments by ab initio calculations.

Figure 1

Top left: majority and minority spin PES spectra measured on bulk Ni and Co films and a series of Co/Ni superlattices. Top right: spin polarization spectra; bottom: corresponding spin polarization (SP) at $E_F$. The arrows show the increase of the minority spin PES near $E_F$ responsible for the SP increase.

Figure 2

Effect of the covering on the spin polarization of the stack for 3-ML Ni capping (top) and for 1-ML Au capping (bottom). This shows that Ni covering decreases the SP whereas an almost full polarized injection may be achieved in Au.

Figure 3

Co magnetic-moment orientation depending on Ni capping measured by using Kerr magnetometry. (a) Sample stack, (b) hysteresis loops for 3-ML Co with different Ni capping, (c) Ni cap thickness leading to in-plane (IP) to (OOP) out-of-plane magnetization transition for different Co thicknesses. The red curve is a fit using Eq. (2).

Figure 4

Examples of XMCD raw data observed at the Co and Ni $L$ edges varying the angle γ between the x-ray beam and the surface normal. The magnetic field to get the XMCD is applied along the x-ray beam.

Figure 5

Top: orbital and spin effective Co magnetic moments measured on the 1-ML Co sample for different γ angles and temperatures. Bottom: $m_L$ and $m_s^{\mathrm{eff}}$ slopes plotted vs temperature.

Figure 6

Similar experiments as in Fig. 5 but for Ni.

Figure 7

Co orbital and spin effective moment anisotropies measured at room temperature for three samples with 1-, 2-, and 3-ML Co. The Co PMA and the $T_z$ contribution are observed to increase when decreasing the Co thickness. This is consistent with the interfacial origins of both anisotropies.

Figure 8

Top: total Co and Ni magnetic moment variation with temperature for the 1-ML Co (left) and 3-ML Co (right) samples. Bottom: the $g$ factors extracted from these data..

Figure 9

Left: hysteresis loops measured by XMCD at the Co edge on a 1-ML Co sample for out-of-plane and in-plane applied fields. The area between OOP and IP loops gives the effective anisotropy.Right: resulting macroscopic magnetic anisotropy vs the atomic-orbital moment anisotropy. A linear law is found in agreement with Bruno's model (dashed red line). The blue line comes from results on sputtered samples reported in Ref. [19].

Figure 10

LEED patterns obtained on (a) the Ni(111) surface and (b) after the growth of 2-ML-thick Co film. The unit cell chosen for rod simulation is indicated in (c). Some truncation rods indexation is shown in (d). Six rods were measured for each sample.

Figure 11

Experimental structure factors (blue open circles, together with error bars) and the best-fit structure factors (solid red line) truncations rods and simulation for the six samples, from top to bottom: Ni(111) substrate, Ni + 1-ML Co, Ni + 1-ML Co+ 1-ML Ni, Ni + 2-ML Co, Ni + 2-ML Co + 1-ML Ni, Ni + 3-ML Co. The schematic representation of the models of stacks obtained after rod simulations is shown on the right.