Effects of hydrogen/deuterium absorption on the magnetic properties of Co/Pd multilayers

The effects of hydrogen (H${}_2$) and deuterium (D${}_2$) absorption were studied in two Co/Pd multilayers with perpendicular magnetic anisotropy (PMA) using polarized neutron reflectivity (PNR). PNR was measured in an external magnetic field $H$ applied in the plane of the sample with the magnetization $M$ confined in the plane for ${\mu }_{o}H=$ 6.0 T and partially out of plane at 0.65 T. Nominal thicknesses of the Co and Pd layers were 2.5 and 21 Å, respectively. Because of these small values, the actual layer chemical composition, thickness, and interface roughness parameters were determined from the nuclear scattering length density profile $({\rho }_n)$ and its derivative obtained from both x-ray reflectivity and PNR, and uncertainties were determined using Monte Carlo analysis. The PNR ${\rho }_n$ showed that although ${\mathrm{D}}_2$ absorption occurred throughout the samples, absorption in the multilayer stack was modest (0.02 D per Pd atom) and thus did not expand. Direct magnetometry showed that ${\mathrm{H}}_2$ absorption decreased the total $M$ at saturation and increased the component of $M$ in the plane of the sample when not at saturation. The PNR magnetic scattering length density $({\rho }_m)$ revealed that the Pd layers in the multilayer stack were magnetized and that their magnetization was preferentially modified upon ${\mathrm{D}}_2$ absorption. In one sample, a modulation of $M$ with twice the multilayer period was observed at ${\mu }_{o}H=$ 0.65 T, which increased upon ${\mathrm{D}}_2$ absorption. These results indicate that ${\mathrm{H}}_2$ or ${\mathrm{D}}_2$ absorption decreases both the PMA and total magnetization of the samples. The lack of measurable expansion during absorption indicates that these changes are primarily governed by modification of the electronic structure of the material.

Figure 1

Diagram showing the scattering geometry for the PNR experiment. The magnetic field $\mathit{H}$ is applied in the plane of the sample along the $x$ axis. For specular reflectivity, the angle of incidence of the neutrons is identical to the angle of reflection $\alpha$. The scattering wave-vector transfer $\mathit{Q}={\mathit{k}}_{\mathit{f}}-{\mathit{k}}_{\mathit{i}}$ is parallel to the $z$ axis and perpendicular to the sample surface. Since magnetic neutron scattering is sensitive to the components of $M$ perpendicular to $\mathit{Q}$, only the components of $M$ in the plane of the sample (x-y plane) are probed by PNR.

Figure 2

(a) and (b): High angle x-ray diffraction of the Co/Pd multilayer for samples A and B, respectively. The expected positions of the sapphire substrate peaks and Pd bulk buffer layer peaks are indicated. Multilayer peaks are indicated by a red dot. (c) and (d): X-ray reflectivity measurements of samples A and B, respectively. The black dots in the reflectivity graphs are the data and the red lines are the fit to the data.

Figure 3

Sketch of (a) sample A and (b) sample B used in XRR and PNR models. The location of the interface roughness parameters σ and the layers used as fitting parameters are illustrated. The dashed red arrows indicate the magnetization used in the PNR model only.

Figure 4

Magnetic moment measurements in 1 atm of He (blue dashed curves) and ${\mathrm{H}}_2$ (red solid curves) with the magnetic field applied perpendicular ($H$⊥) and parallel ($H$∥) to the sample surface. (a) Data for sample A, (b) data for sample B. Top left and bottom right insets in (b) are close-up views of the data in (b) for the $H$⊥ and $H$∥ configurations, respectively.

Figure 5

MFM image (5 μm × 5 μm) of sample B performed at H $=$ 0 at room temperature after magnetizing it out of the plane of the sample.

Figure 6

PNR using neutrons with ($--$) and ($+$ $+$) incoming and outgoing polarization states with the sample in helium [(a) and (b)] and deuterium [(c) and (d)] in a 0.65-T magnetic field applied in the plane of sample A. Experimental data are black dots and the model fit is the red line.

Figure 7

PNR using neutrons with ($--$) and ($+$ $+$) incoming and outgoing polarization states with the sample in helium [(a) and (b)] and deuterium [(c) and (d)] in a 6-T magnetic field applied in the plane of sample A. Experimental data are black dots and the model fit is the red line.

Figure 8

PNR using neutrons in the high-$Q$ regime ($Q>0.1$ Å${}^{-1}$) with ($-$) and ($+$) incoming polarization states with the sample in air [(a) and (b)] and deuterium [(c) and (d)] in a 0.65-T magnetic field applied in the plane of sample B. The positions of the single order ($Q_1$) and half order magnetic peaks ($Q_{1/2}$) are indicated. Experimental data are black dots and model fit is the red line.

Figure 9

PNR using neutrons in the low $Q$ regime ($Q$ < 0.1 Å${}^{-1}$) with ($-$) and ($+$) incoming polarization states with the sample in air [(a) and (b)] and deuterium [(c) and (d)] in a 0.65-T magnetic field applied in the plane of sample B in the low $Q$ regime. Experimental data are black dots and model fit is the red line.

Figure 10

Nuclear SLD profiles (blue dashed curve) and its derivative (green solid curve) for sample B (a) in air and (b) for the sample in deuterium. The vertical dotted lines indicate the positions of the interfaces. The corresponding sample profile is shown.

Figure 11

Nuclear SLD profile in air (helium) (blue dashed curve) and in deuterium (red solid curve) for (a) sample A and (b) sample B at 0.65 T.

Figure 12

Magnetic SLD in air (blue dashed curve) and in ${\mathrm{D}}_2$ (red solid curve) for sample A at (a) 6 T and (b) 0.65 T for two Co/Pd bilayers in the stack. The film-substrate interface is set at a thickness of zero. The equivalent magnetization, calculated by dividing ${\rho }_m$ by $2.853\times {10}^{-9}$ Å${}^{-2}$/(${10}^{-3}$ A/m).

Figure 13

Magnetic SLD in air (blue dashed curve) and in ${\mathrm{D}}_2$ (red solid curve) for sample B at 0.65 T for one Co/Pd bilayer in the stack. The film-substrate interface is set at a thickness of zero. The equivalent magnetization, calculated by dividing ${\rho }_m$ by $2.853\times {10}^{-9}$ Å${}^{-2}$/(${10}^3$ A/m), is shown in the scale on the right.