Detection of induced paramagnetic moments in Pt on ${\mathrm{Y}}_3{\mathrm{Fe}}_5{\mathrm{O}}_{12}$ via x-ray magnetic circular dichroism

Magnetic moments in an ultrathin Pt film on a ferrimagnetic insulator ${\mathrm{Y}}_3{\mathrm{Fe}}_5{\mathrm{O}}_{12}$ (YIG) have been investigated at high magnetic fields and low temperatures by means of x-ray magnetic circular dichroism (XMCD). We observed an XMCD signal due to the magnetic moments in a Pt film at the Pt $L_3$ and $L_2$ edges. By means of the element-specific magnetometry, we found that the XMCD signal at the Pt $L_3$ edge gradually increases with increasing the magnetic field even when the field is much greater than the saturation field of YIG. Importantly, the observed XMCD intensity was found to be much greater than the intensity expected from the Pauli paramagnetism of Pt when the Pt film is attached to YIG. These results imply the emergence of induced paramagnetic moments in Pt on YIG and explain the characteristics of the unconventional Hall effect in Pt/YIG systems.

Figure 1

(a) $M-H$ curve for the YIG slab when the magnetic field $H$ was applied along the [111] direction. The inset to (a) shows the magnified view of the $M-H$ curve in the low-$H$ range ($\left|H\right|<5\mathrm{kOe}$). The $M-H$ curve was measured with a vibrating sample magnetometer. (b) An AFM image of the polished (111) surface of the YIG slab before the Piranha treatment, where the surface roughness is $R_{\mathrm{a}}\sim 0.2\mathrm{nm}$. Similar AFM image and roughness values were confirmed also for the YIG surface with the Piranha treatment. (c) A cross-sectional TEM image of the Pt/YIG sample. The inset shows a blowup of the Pt/YIG interface.

Figure 2

A schematic illustration of the experimental setup for the XMCD measurements. A magnetic field, $\mathbf{H}$, parallel or antiparallel to the incoming x-ray beam, was applied to the Pt/YIG or Pt/YAG sample at an angle of ${10}^{\circ }$ to the normal direction $\mathbf{n}$ of the sample surface.

Figure 3

(a) The normalized XAS for the Pt $L_3$ and $L_2$ edges of the Pt/YIG and Pt/YAG samples at $T=5.5\mathrm{K}$ and $H=50\mathrm{kOe}$. The XAS edge jump, defined as the difference in the XAS intensity between $11540\mathrm{eV}$ ($13255\mathrm{eV}$) and $11610\mathrm{eV}$ ($13325\mathrm{eV}$), is normalized to 1 ($2.{22}^{-1}$) for the $L_3$ edge ($L_2$ edge) according to Refs. [35] and [56]. The XAS offset value for the $L_3$ edge ($L_2$ edge) of the Pt/YIG sample is set to be 0 (1) [39, 52, 53, 54, 56]. Similarly, the XAS offset value for the $L_3$ edge ($L_2$ edge) of the Pt/YAG sample is set to be 0 (1) and then, for clarity, is shifted along the vertical axis by a constant value of 0.75 (0.75). The triangles indicate the oscillation structures in the EXAFS region. The peak structure marked with an asterisk in the XAS at $11607\mathrm{eV}$ for the Pt/YAG sample appears due to elastic diffraction from the YAG substrate [54]. (b) The XMCD spectra and their integral for the Pt $L_3$ and $L_2$ edges of the Pt/YIG sample. The blue (red) arrow indicates the negative (positive) XMCD signal at the Pt $L_3$ edge ($L_2$ edge). The inset to (b) shows the XMCD spectra for the Pt $L_3$ edge of the Pt/YIG sample measured at $H=\pm 50\mathrm{kOe}$, where the blue arrows indicate the XMCD signal. In (b), the XMCD spectra for the Pt $L_3$ edge of the Pt/YAG sample are also plotted.

Figure 4

The ESM curve at the Pt $L_3$ edge of the Pt/YIG sample at $T=5.5\mathrm{K}$ and the fixed photon energy of $11568$ eV, where the blue dots and blue solid line represent the experimental data and the linear-fitting result, respectively. The XMCD data at the Pt $L_3$ edge of the Pt/YAG sample (gray circle), estimated from Fig. 3, and of a Pt foil at $T=10\mathrm{K}$ (gray dashed line), taken from Ref. [56], are also plotted. The inset shows the magnified view of the ESM curve of the Pt/YIG sample in the low-$H$ range ($\left|H\right|<5\mathrm{kOe}$). The XMCD values of the vertical axes are multiplied by $-1$ for clarity, since the sign of the XMCD signal at the Pt $L_3$ edge is negative [see Fig. 3].

Figure 5

[(a) and (b)] The normalized XAS for the (a) Pt $L_3$ edge and (b) Pt $L_2$ edge of the Pt/YIG sample. [(c) and (d)] The normalized XAS for the (c) Pt $L_3$ edge and (d) Pt $L_2$ edge of the Pt foil. The horizontal axes are shifted with respect to the energy at the inflection point $E_0$ for the Pt spectra. In (a) and (b) [(c) and (d)], the XAS for the Au $L_3$ and $L_2$ edges of the Au foil and difference in the XAS between the Pt/YIG sample (Pt foil) and Au foil are also plotted, where the energy scale of the Au XAS is expanded by a factor of 1.07 to take into account the difference in the lattice constant between Pt and Au [72] and is shifted in energy on the basis of the oscillation structures in the EXAFS region of the Pt spectra.