Oscillatory interlayer exchange coupling in ${\left[\mathrm{Pt}∕\mathrm{Co}\right]}_n∕\mathrm{NiO}∕{\left[\mathrm{Co}∕\mathrm{Pt}\right]}_n$ multilayers with perpendicular anisotropy: Dependence on $\mathrm{NiO}$ and $\mathrm{Pt}$ layer thicknesses

Interlayer exchange coupling has been studied in a series of ${\left[\mathrm{Pt}\left(t_{\mathrm{Pt}}\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_n∕\mathrm{NiO}\left(t_{\mathrm{NiO}}\right)∕{\left[\mathrm{Co}\left(4\mathrm{\AA }\right)∕\mathrm{Pt}\left(t_{\mathrm{Pt}}\mathrm{\AA }\right)\right]}_n$ multilayers with perpendicular anisotropy. The coupling oscillates between antiferromagnetic and ferromagnetic as a function of $t_{\mathrm{NiO}}$ with a period of $\sim 5\mathrm{\AA }$, and the oscillatory behavior is related to the antiferromagnetic ordering of the $\mathrm{NiO}$ layer. This interlayer coupling between two $\mathrm{Co}∕\mathrm{Pt}$ multilayers is shown to occur domain by domain by magnetic force microscopy imaging. For the strongest antiferromagnetic coupling at $t_{\mathrm{NiO}}=11\mathrm{\AA }$, an oscillation with a period of $\sim 6\mathrm{\AA }$ is superposed onto the exponential decay of the coupling strength as a function of $t_{\mathrm{Pt}}$. The exponential decay with $t_{\mathrm{Pt}}$ is ascribed to the exponential decay of the coupling between the $\mathrm{Co}$ layers across the $\mathrm{Pt}$ layers in each $\mathrm{Co}∕\mathrm{Pt}$ multilayer, and the superposed oscillatory behavior can be attributed to multiple reflections of electron waves at the $\mathrm{Co}∕\mathrm{Pt}$ interfaces and their interference. A linear dependence of the antiferromagnetic coupling strength on $1∕n$, (where $n$ is the number of repeats), is suggestive of a surface interaction for this interlayer coupling.

Figure 1

X-ray diffraction result for glass/$\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(5\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{NiO}\left(11\mathrm{\AA }\right)∕{\left[\mathrm{Co}\left(4\mathrm{\AA }\right)∕\mathrm{Pt}\left(5\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(50\mathrm{\AA }\right)$, which shows that both the $\mathrm{Pt}$ and $\mathrm{NiO}$ layers are highly $\mathrm{fcc}\left(111\right)$ textured, and the $\mathrm{Co}$ layers are highly $\mathrm{hcp}\left(100\right)$ textured.

Figure 2

Hysteresis loops at room temperature for the samples of glass/$\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(5\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{NiO}\left(11\mathrm{\AA }\right)$ (solid line), glass/$\mathrm{NiO}\left(11\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(5\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(20\mathrm{\AA }\right)$ (dotted line), and glass/$\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(5\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(20\mathrm{\AA }\right)$ (dashed line). Measurements are made using AGFM with the field applied perpendicular to the sample surface.

Figure 3

The major and minor hysteresis loops at room temperature for the samples with $\mathrm{NiO}$ thicknesses of 11 and $20\mathrm{\AA }$ in $\mathrm{S}1\left(n=3\right)$ and $\mathrm{S}2\left(n=4\right)$, which were measured using SQUID with the applied field along the out-of-plane easy axis. For the measurement of minor loops, a large field was applied first to saturate the whole sample; the field was then decreased until the upper $\mathrm{Co}∕\mathrm{Pt}$ multilayer had reversed direction. Finally, the field was increased again to finish the measurement. During this process, the bottom $\mathrm{Co}∕\mathrm{Pt}$ multilayer remains saturated due to its larger coercivity. The vertical dashed lines in the hysteresis loops give the shifts of the minor loops.

Figure 4

Variation of the minor loop shift $H_{\mathrm{MLS}}$ with the $\mathrm{NiO}$ thickness $t_{\mathrm{NiO}}$ at room temperature for (a) the samples in $\mathrm{S}1$ with $n=3$ and (b) the samples in $\mathrm{S}2$ with $n=4$. The dotted lines are guides to the eyes. The inset in (b) is the hysteresis loop for a $\mathrm{NiO}$ thickness of $13\mathrm{\AA }$, which displays only one transition. For this sample, a measurement of the minor loop is not possible, and the minor loop shift is simply set to zero, but that does not necessarily indicate that the two $\mathrm{Co}∕\mathrm{Pt}$ multilayers are decoupled.

Figure 5

(Color online) XMCD spectra for (a) an AF coupled sample 1: $\mathrm{Si}∕\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(6\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{NiO}\left(11\mathrm{\AA }\right)∕{\left[\mathrm{Co}\left(4\mathrm{\AA }\right)∕\mathrm{Pt}\left(5\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(24\mathrm{\AA }\right)$ and (b) ferromagnetically coupled sample 2: Glass/$\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(6\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{NiO}\left(11\mathrm{\AA }\right)∕{\left[\mathrm{Co}\left(4\mathrm{\AA }\right)∕\mathrm{Pt}\left(5\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(50\mathrm{\AA }\right)$. The data shown are the difference between right circularly polarized (RCP) and left circularly polarized (LCP) x rays, a measure of net magnetization. The energy is tuned to the ${\mathrm{L}}_3$ edge of $\mathrm{Ni}$ and the beam is at normal incidence; hence we probe only the out-of-plane component of the $\mathrm{Ni}$ magnetization. For sample 2, we show data after both positive and negative saturation; the magnetization clearly displays the opposite sign for positive and negative remanence. The insets are a sketch of the canting of the $\mathrm{NiO}$ spins at the interface.

Figure 6

(Color online) Room temperature MFM images in the as-grown states for samples with $\mathrm{NiO}$ thicknesses of 11, 14, 16, and $18\mathrm{\AA }$ in $\mathrm{S}1$. The imaging area is $10\mu \mathrm{m}\times 10\mu \mathrm{m}$. For $\mathrm{NiO}$ thicknesses of 11 and $16\mathrm{\AA }$, the interlayer coupling is antiferromagnetic and only closed domain walls are observed, while for the $\mathrm{NiO}$ thicknesses of 14 and $18\mathrm{\AA }$, the interlayer coupling is ferromagnetic and up and down domains are observed, similar to the domain pattern for the pure $\mathrm{Co}∕\mathrm{Pt}$ multilayer shown in Fig. 6.

Figure 7

(Color online) Schematic diagrams to show the formation of closed domain walls in antiferromagnetic coupled $\mathrm{Co}∕\mathrm{Pt}$ multilayers separated by $\mathrm{NiO}$ spacers. Figure 7a on the left is a MFM image of the domain pattern for glass/ $\mathrm{Pt}\left(100\mathrm{\AA }\right)∕{\left[\mathrm{Pt}\left(5\mathrm{\AA }\right)∕\mathrm{Co}\left(4\mathrm{\AA }\right)\right]}_3∕\mathrm{Pt}\left(20\mathrm{\AA }\right)$ in the as-grown state. The imaging area is $5\mu \mathrm{m}\times 5\mu \mathrm{m}$. The areas with light and dark contrasts correspond to magnetic domains with the moments pointing up and down, respectively. The schematic diagrams on the top and bottom of the image indicate the existence of domain walls between up and down domains, and on their left are the experimental curves of the phase shifts. The experimental curve of the phase shift in (c) is for the domain wall labeled by the solid line in the image with $t_{\mathrm{NiO}}=11\mathrm{\AA }$ in Fig. 6. The experimental curves in (a) and (b) are from the domain walls labeled by the solid line the $\mathrm{Co}∕\mathrm{Pt}$ layer in the figure on the left.

Figure 8

The major and minor hysteresis loops for samples with differing $\mathrm{Pt}$ thicknesses in $\mathrm{S}3$. The numbers represent the $\mathrm{Pt}$ thickness. A switching field $H_{\mathrm{SW}}$ is defined as shown in the hysteresis loop with $\mathrm{Pt}$ thickness of $6\mathrm{\AA }$, being the field at which the lower $\mathrm{Co}∕\mathrm{Pt}$ multilayer reverses completely.

Figure 9

(Color online) Room temperature MFM images at the as-grown state for the samples with the thicknesses of $t_{\mathrm{Pt}}=6$, 8, and $10\mathrm{\AA }$ in $\mathrm{S}3$. The imaging area is $20\mu \mathrm{m}\times 20\mu \mathrm{m}$. Only closed domain walls are observed in these samples, suggestive of antiferromagnetic interlayer coupling between the top and bottom $\mathrm{Co}∕\mathrm{Pt}$ multilayers. The domain size increases with increasing $\mathrm{Pt}$ thickness.

Figure 10

(Color online) Variations of the (a) switching field $H_{\mathrm{SW}}$. The dotted line is the fit to the data showing the exponential decay of $H_{\mathrm{SW}}$ with $t_{\mathrm{Pt}}$ and (b) the minor loop shift $H_{\mathrm{MLS}}$ as a function of the $\mathrm{Pt}$ thickness $t_{\mathrm{Pt}}$. The solid line is a fit to the data, showing that the variation of $H_{\mathrm{MLS}}$ with $t_{\mathrm{Pt}}$ is a superimposition of an oscillation with period of $\sim 6\mathrm{\AA }$ (dashed line) onto an exponential decay (dotted line).

Figure 11

(Color online) Linear dependence of the minor loop shift $H_{\mathrm{MLS}}$ on $1∕n$ in $\mathrm{S}4$ where $n$ is the number of repetitions of the $\mathrm{Co}∕\mathrm{Pt}$ multilayer. The solid line is the linear fit to the experimental data (solid dots). The inset shows the major and minor hysteresis loops for the sample with $n=5$ in $\mathrm{S}4$.