N I S T Center for Neutron Research
Accomplishments and Opportunities 2001
Magnetic Semiconductor Superlattices
Currently a great deal of attention is being focused on spintronics, a new area of solid-state electronics. In spintronics not only the electric current but also its spin state is controlled. Spin valves and spin injectors are the first practical applications of spintronics. Further progress in developing new devices hinges critically on the availability of suitable materials. Such materials need to be “good” semiconductors, easy to integrate in typical integrated circuits, and their electronic properties should exhibit strong sensitivity to the carrier’s spin, ferromagnetism being an especially desirable property.
EuS is one of the very few natural ferromagnetic (F M) semiconductors. Since it becomes F M at a low temperature (Tc = 16.6 K) it is an unlikely choice for applications. However, studying the properties of heterostructures made on its base may give an important insight into fundamental processes taking place in all classes of materials under consideration.
GaMnAs is a man-made F M semiconductor. It is an example of a diluted magnetic semiconductor (D M S) in which a fraction of nonmagnetic cations (Ga) is substituted with magnetic ions (Mn). Such a material can readily be incorporated into modern GaAs-based semiconductor devices. Its Tc is still below room temperature, but this limitation may be lifted in other materials of this class (Refer to Reference 1).
Interlayer exchange coupling (I E C) in superlattices (S L), composed of ferromagnetic and nonmagnetic layers, is a crucial element of all spin-valve type devices that utilize the giant magnetoresistance effect. In metallic S L's currently being used, conduction electrons transfer the interlayer interactions through nonmagnetic spacers (Refer to Reference 2). Here we address the question whether I E C phenomena are possible in all-semiconductor superlattices, like EuS/PbS and GaMnAs/GaAs, where the carrier concentrations are many orders of magnitude lower than in metals.
The nonmagnetic spacer in EuS/PbS S L's is a narrow gap semiconductor with electron concentration of the order of 1017 cm-3 to 1018 cm-3. For thin PbS layers (dPbS < 70 Å) neutron reflectivity spectra, shown in Figure 1, have revealed a pronounced maximum of magnetic origin at the position corresponding to the doubled structural S L periodicity, thus indicating the existence of antiferromagnetic (A F M) interlayer arrangements (Refer to Reference 3).
Graphics Caption FIGURE 1. Unpolarized neutron reflectivity spectra for EuS/PbS S L with thin (23 Å) PbS spacer. Antiferromagnetic interlayer exchange coupling below Tc and at zero external field is clearly visible (blue curve). Applying a strong enough magnetic field (185 G in this case) parallel to the S L surface forces all the EuS layer’s magnetizations to ferromagnetic configuration (green curve). Above Tc the system is nonmagnetic, the only Bragg peak comes from the chemical S L periodicity.
Graphics Caption FIGURE 2. The sample with thick (135 Å) PbS layers is almost ferromagnetically coupled. Application of an external magnetic field enhances the F M Bragg peaks and lowers the intensity between them (at the A F M peak position).
Graphics Caption FIGURE 3. Polarized neutron reflectivity spectra for GaMnAs/GaAs superlattice.
For much thicker PbS spacers (dPbS > 120 Å) the only magnetic peaks visible in the reflectivity profiles, see example in Figure 2, coincide with the chemical ones, thus leading to the conclusion that the magnetization vectors in adjacent EuS layers are parallel, which indicates F M I E C.
In the intermediate PbS thickness range (70 Å < dPbS < 120 Å), both A F M and F M peaks are present. Polarized neutron analysis of these maxima gives evidence that the magnetization vectors of adjacent EuS layers are not colinear. Hence, the I E C found in EuS/PbS S L's has an oscillatory character similar to that occurring in metallic S L's, although the oscillation period is much longer than the one in metallic systems.
In order to confirm that the free carriers, present in the PbS layer in such a small amount, are the cause of the observed oscillatory I E C, a series of analogous measurements have been carried out on EuS/YbSe S L's. The structure and lattice constant of YbSe are the same as those of PbS. In contrast to PbS, YbSe is a semi-insulator with a negligible carrier concentration. Neutron reflectivity profiles have shown no evidence of any interlayer coupling in the all investigated samples. That finding, together with the oscillatory character of coupling in S L's with PbS spacer, strongly points to the leading role of PbS free electrons in providing the necessary I E C mechanism, similar to that discovered in metallic multilayers.
Ferromagnetic ordering in GaMnAs is carrier (holes) induced; the Mn atoms, apart from being the magnetic element in the system, act also as acceptors providing the holes responsible for transferring exchange interactions between them. The details of the magnetic ordering, in particular its range, are still being disputed.
To address the latter issue, polarized neutron reflectometry has been performed on a number of GaMnAs/GaAs superlattices. Figure 3 shows an example of the obtained reflectivity profile in the vicinity of the first S L Bragg peak, for one of the samples. The very presence of the magnetic contribution to the structural S L Bragg peak is a strong confirmation of the F M I E C between consecutive GaMnAs layers. The absence of any spin-flip scattering shows that the sample is in a one-domain state, i.e., the F M ordering in GaMnAs is long range, and the sample is spontaneously saturated. The peak in (--) cross section, and its absence in the (++), is proof that the magnetization is directed oppositely to the external magnetic guide field, hence the long range ordering has formed spontaneously, without the influence of the external field. More details can be found in Reference 4.
References
[1] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287,1018 (2000).
[2] P. Bruno, Phys. Rev. B52, 411 (1995).
[3] H. Kepa, J. Kutner-Pielaszek, J. Blinowski, A. Twardowski, C. F. Majkrzak, T. Story, P. Kacman, R. R. Galazka, K. Ha, H. J. M. Swagten, W. J. M de Jonge, A. Yu. Sipatov, V. Volobuev, T. M. Giebultowicz, Europhys. Lett. 56, 54 (2001).
[4] H. Kepa, J. Kutner-Pielaszek, A. Twardowski, C. F. Majkrzak, J. Sadowski, T. Story, T. M. Giebultowicz, Phys. Rev. B64, 121302 (2001).
Authors
A. Yu. Sipatov, V. Volobuev
Kharkov State Polytechnical University
Kharkov, Ukraine
H. Kepa, J. Kutner-Pielaszek, A. Twardowski
Institute of Experimental Physics
Warsaw University
Warsaw, Poland
T. Story, J. Sadowski
Institute of Physics
Polish Academy of Sciences
Warsaw, Poland
T. M. Giebultowicz
Physics Department
Oregon State University
Corvallis, OR 97331
C. F. Majkrzak
N I S T Center for Neutron Research
National Institute of Standards and Technology
Gaithersburg, MD 20899-8562