Insight into Non-collinear Exchange Coupling of Fe/Mn/Fe from STM
We used STM to study the film growth and morphology of epitaxial Mn films grown on Fe(001) single crystal whiskers in order to gain insight into the interlayer exchange coupling that was observed in SEMPA measurements of Fe/Mn/Fe(001) trilayers. For noble metal spacer layers, the magnetic coupling is well described by quantum well models where the coupling is determined by the Fermi surface properties of the spacer layer material and the reflection amplitudes for electrons scattering at the interfaces between the spacer layer and the ferromagnetic layer. In contrast, if the spacer layer is an antiferromagnetically ordered material such as Cr or Mn, it can no longer be considered a passive medium transmitting the indirect exchange coupling as in the quantum well picture. In such cases, the exchange coupling of the antiferromagnet to the ferromagnet at the interfaces, as well as the internal exchange coupling within the antiferromagnet, must be considered. We have intensively investigated the coupling through Cr, which is a somewhat special case complicated by the spin density wave nature of antiferromagnetism in Cr. Mn is expected to be a simpler antiferromagnetic spacer if it is grown on a ferromagnet as a bcc or bct film where it has alternating planes of moments parallel and antiparallel to the magnetization of the ferromagnet.
Figure 1. SEMPA images of the two components of magnetization in the top Fe layer of
a Fe/Mn(wedge)/Fe(001). The scale shows the thickness of the Mn spacer layer in ML.
From these images the variation in the direction of magnetization can be determined as
shown schematically in the blow-up of the 9-12 ML region below.
Whereas SEMPA measurements showed predominantly collinear parallel or antiparallel
coupling through Ag, Au and even Cr spacer layers, the coupling through Mn in a
Fe/Mn(wedge)/Fe(100) whisker was no longer collinear. For Mn thicknesses greater
than 9 ML, the coupling angle oscillates with a two-layer period between 90°-
φ and 90°+ φ as seen in Figure 1. The value of φ
was found to be sample dependent and to range between approximately 10 and 30°.
The proximity model was developed by Slonczewski to describe the exchange coupling
through an antiferromagnet where there is a strong coupling across the interface
between the interface spins in the antiferromagnet and the ferromagnet. The
ferromagnetic layers want to be ferromagnetically (antiferromagnetically) coupled,
0° (180°) coupling angle, for a spacer that is an odd (even) number of ML
thick. In the presence of thickness fluctuations in the spacer layer, the proximity
model predicts that the coupling angle between the ferromagnetic layers will vary
around a mean value of 90° with the extremes depending on the size of σ
as shown in Figure 2. For a very narrow thickness distribution the magnetization
varies between 0 and 180°, while for σ greater than about 1, the
ferromagnetic layers are coupled at 90° to each other as has been observed
in previous experiments. Optimized growth on an Fe whisker leads to a narrow
enough thickness distribution so that the magnetization oscillates about 90°
as seen in SEMPA.
Figure 2. The amplitude of the coupling angle variations predicted by the proximity
model for a given σ of the thickness
distribution is shown for regions of the antiferromagnetic spacer that are an odd
(even) number of ML thick by the solid (dashed) line.
If we measure the thickness distribution of the spacer layer with the STM, we
can test the predictions of the proximity model. Because the Fe whiskers are very
flat with large terraces, the measured the root-mean-square height variation or rms roughness,
corresponds
to the standard deviation of the thickness distribution. We can thus measure
σ with the STM for Mn films prepared on
Fe whiskers at substrate temperatures and deposition rates like those used in the
SEMPA investigation. Figure 3 shows an STM topography image of a 9.81 ML thick Mn
film grown on a Fe whisker substrate temperature of 175°C.
Figure 3. An STM image of a 750 x 1000 nm region of a 9.81 ML Mn film where changes
in gray levels correspond to single atom height differences.
The fraction of each layer exposed in the topography image is measured and plotted for each layer as shown in Figure 4. The growth front is fit well by a Gaussian where σ is the rms roughness and hence in this flat region of the whisker the standard deviation of the thickness distribution. For this example, with σ=0.6, the Slonczewski proximity model predicts a variation in the coupling angle of 30° which is in reasonable agreement with the actual coupling angle variations of 23° measured with SEMPA for a Mn film deposited at the same temperature. STM studies were carried out on Mn films grown at three different substrate temperatures near to the growth temperatures for the SEMPA measurements. RHEED intensity oscillation measurements of both the SEMPA and STM samples indicated similar growth. From our STM measurements of the standard deviation of the thickness distribution, we conclude that the predictions of Slonczewski’s proximity model are consistent with the coupling angle variations measured by SEMPA.
Figure 4. The fraction of each layer exposed in Fig. 3 is measured and plotted for each layer.
Growth & Magnetic Oscillatory Exchange Coupling of Mn/Fe (001) and Fe/Mn/Fe (001)
Non-Collinear Exchange Coupling in Fe/Mn/Fe(001): Insight From a Scanning Tunneling Microscopy
Joseph A. Stroscio
Daniel T. Pierce
Robert J. Celotta
Angela Davies - University of North Carolina - Charlotte
D. A. Tulchinsky - Naval Research Laboratory
Supported in part by the Office of Naval Research
Online: July 2002
Last Updated: February 2008
egpwebmaster@nist.gov