Magnetic anisotropy in epitaxial Py/FeMn/Cu(001) and Py/FeMn/Ni/Cu(001) films

Py/FeMn/Cu(001) and Py/FeMn/Ni/Cu(001) films were grown and studied as a function of both the FeMn and the Ni thicknesses using rotating magneto-optic Kerr effect and x-ray magnetic circular dichroism. For Py/FeMn/Cu(001), we find that the FeMn antiferromagnetic order switches the sign of the Py fourfold magnetic anisotropy but has little effect on the step-induced uniaxial magnetic anisotropy. For Py/FeMn/Ni/Cu(001), we find that out-of-plane Ni magnetization has little effect on the Py magnetic anisotropy, but in-plane Ni magnetization enhances the Py magnetic anisotropy in the region just above antiferromagnetic transition thickness. The underlying mechanism could be attributed to the FeMn 3Q spin structure.

Figure 1

LEED patterns of (a) Cu(001) substrate at 130 eV, (b) 10-ML Ni layer at 130 eV, (d) 10-ML FeMn layer at 128 eV, and (d) 15-ML Py overlayer at 127 eV.

Figure 2

Hysteresis loops of Py(15ML)/FeMn/Ni(18ML)/Cu(001) at different FeMn thicknesses. Loops are collected using MOKE in (a) longitudinal and (b) polar configurations. Loops are collected using XMCD effect at Fe L${}_3$ peak for (c) longitudinal and (d) polar configurations. The XMCD loops represent the Py hysteresis loops.

Figure 3

Polar MOKE remanence and polar Py saturation filed of Py(15ML)/FeMn/Ni(18 ML)/Cu(001) as a function of FeMn thickness. Both curves show a “pseudo” spin reorientation transition at $d_{\mathrm{FeMn}}\sim$5 ML. The Ni magnetization is forced to be more in-plane by its strong coupling to the in-plane Py at $d_{\mathrm{FeMn}}<$5 ML and switches to out-of-plane direction as its coupling to the Py becomes weaker at $d_{\mathrm{FeMn}}>$5 ML.

Figure 4

(a) Schematic of the ROTMOKE experiments. The optics are set such that the Kerr signal measures the projection of magnetization in the [100] direction ($\propto \cos {\theta }_M$). The magnetic torque [$l\left({\theta }_M\right)\equiv H_{\mathrm{ROT}}\sin \left({\theta }_H-{\theta }_M\right)$] as a function of the magnetization angle (${\theta }_M$) for Py(15ML)/FeMn/Cu(001) at (b) $d_{\mathrm{FeMn}}$ $=$ 6 ML, (c) 8 ML, (d) 10 ML, (e) 12 ML, (f) 14 ML, and (g) 16 ML. The red lines are theoretical fittings using Eq. (2).

Figure 5

(a) $H_C$, $H_2$, and $H_4$ as a function of FeMn thickness for Py(15ML)/FeMn/Cu(001). (b) Schematic drawing of the in-plane component of FeMn 3Q spin-structure on (001) surface. Red and blue arrows are in-plane projection of spins from different layers. Black dashed lines are [100] steps. The circled spins are uncompensated spins as a result of the [100] steps. The inset in (b) shows the three-dimensional drawing of the 3Q structure.

Figure 6

Magnetic torque vs magnetization angle of Py in Py(15 ML)/FeMn/Ni/Cu(001) for (a)–(c) paramagnetic Ni ($d_{\mathrm{Ni}}$ $=$ 2 ML), (d)–(f) in-plane ferromagnetic Ni ($d_{\mathrm{Ni}}$ $=$ 10 ML), and (g)–(i) out-of-plane ferromagnetic Ni ($d_{\mathrm{Ni}}$ $=$ 18 ML). (a), (d), and (g) are taken at paramagnetic FeMn ($d_{\mathrm{FeMn}}$ $=$ 8 ML); (b), (e) and (h) are taken at $d_{\mathrm{FeMn}}$ $=$ 12 ML where the FeMn is just AFM ordered; and (c), (f) and (i) are taken at $d_{\mathrm{FeMn}}$ $=$ 16 ML where the FeMn is well AFM ordered. The anisotropy value ($H_4$) from the fitting is shown in (j). The $H_4$ is enhanced for in-plane Ni in the range of 10 ML < $d_{\mathrm{FeMn}}<$15 ML.

Figure 7

(a) $H_4$ and (b) $H_2$ of Py(15ML)/FeMn/Ni/Cu(001) as a function of Ni thickness at paramagnetic ($d_{\mathrm{FeMn}}$ $=$ 5 ML), “soft-AFM” ($d_{\mathrm{FeMn}}$ $=$ 12 ML), and well-ordered AFM ($d_{\mathrm{FeMn}}$ $=$ 18 ML) FeMn film. While $H_4$ has little change in the $d_{\mathrm{FeMn}}$ $=$ 5 and 18 ML samples, $H_4$ is enhanced by in-plane Ni magnetization (5 ML < $d_{\mathrm{Ni}}<$15 ML) in the $d_{\mathrm{FeMn}}$ $=$ 12 ML sample. Here, $H_2$ shows no significant changes as a function of Ni for the three FeMn thicknesses considered.

Figure 8

Map of (a) $H_4$ and (b) $H_2$ of Py(15ML)/FeMn/Ni/Cu(001) as a function of both Ni and FeMn thicknesses. Both (a) and (b) uses the same color-map to compare the variations in $H_4$ and $H_4$ in absolute values. In the PM-AFM transition region (10 ML < $d_{\mathrm{FeMn}}<$15 ML), only in-plane Ni magnetization (5 ML < $d_{\mathrm{Ni}}<$15 ML) enhances the $H_4$ value as compared to Py(15ML)/FeMn/Cu(001).