The phenomenon known as giant magnetoresistance (GMR), which has won the Nobel prize of physics in 2007, is based on the effect of antiferromagnetic coupling (AFC) between adjacent layers of magnetic materials. This effect has become fundamental for the manufacturing of powerful magnetic data storage devices, such as present day computer hard disks. Until recently, AFC was known only for ferromagnetic materials. Now, a team of scientists from the Department of Physics, Korea University, Seoul, the National Institute of Standards and Technology, Gaithersburg, and the Department of Physics, University of Notre Dame have demonstrated for the first time the existence of this property in semiconductor multilayers [Chung et al., Phys. Rev. Lett. (2008) 101, 237202].

The drawback of metallic materials capable of AFC is that they can only hold data, not process it. Thus, data has to be moved from the storage to some semiconductor-based device where it is processed, slowing down the whole system. Semiconductor materials with magnetic properties would be capable of processing and storing data at the same time, making the slow data transfer unnecessary. “For the realization of GMR-like spin memory devices with ferromagnetic semiconductors, spontaneous AFC must be obtained in respective multilayer structures. Our discovery is exactly this spontaneous anitiferromagnetic spin alignment in GaMnAs multilayers, separated by a non-magnetic material”, explains J.-H. Chung, the corresponding author. “This combination would provide a significant advantage over metallic systems, because the memory function could be operated by gate voltage instead of current as in the case of metallic ferromagnetic systems.” Chung and his group were able to develop this promising material by introducing extra charge carriers (Be) in the insulating, non-magnetic GaAs spacers that separate the GaMnAs layers. In other words, layers of magnetic doped semiconductors are spaced by layers of carrier-doped semiconductors. The researchers proved this multilayer arrangement by studying the stacks by means of a method known as polarized neutron reflectometry.

Although the fabrication of the material is relatively inexpensive, GaMnAs-based multilayers cannot be utilized practically for spintronic devices so far because of their rather low Curie temperature (Tc). “Up to now, about 170 K in the single GaMnAs layer is the highest Tc we observed,” says Chung. “Therefore, further effort must be made to find similar materials with higher Tc.” Once respective semiconductor-based multilayer systems operating at ambient temperatures are developed, they will have substantial advantages over the existing metallic ferromagnetic systems due to the possibility of easy electrical control.