Using first principles techniques and Monte Carlo simulations, a team of European researchers has shown that adding a layer of graphene to the antiferromagnetic material BaMnO₃ changes its magnetic state. As reported in a recent issue of Physical Review B, the resulting system displays the properties of both metallic and topological insulating states and could suggest a path toward low-power spintronics components.

Spintronic devices, which utilize electron spin as an additional degree of freedom, have the potential to store and transfer data more efficiently than existing electronics. However, these devices require materials with properties that do not naturally coexist, such as long spin coherence length and low spin-orbit coupling. This has motivated researchers to consider combinations of materials whose interactions at the interface could produce the desired properties.  

In this new work, led by Zeila Zanolli at RWTH Aachen University (now at the Catalan Institute of Nanoscience and Nanotechnology, ICN2), researchers analyzed the magnetic properties of a graphene-BaMnO₃ system. Graphene is a versatile and attractive component for spintronics materials due to its long spin coherence length, while oxide interfaces produce spin-orbit effects that lead to long carrier lifetimes and other desirable properties.

Although graphene is nonmagnetic, experiments have recently shown that magnetism can be induced in graphene by proximity to a magnetic insulator. This inspired the research team to explore whether graphene can, in turn, be used to control the magnetic configuration of a BaMnO₃ substrate. If so, they reasoned, the graphene-BaMnO₃ system could display the key properties required for memory and spin filter applications.

First principles calculations showed that bulk BaMnO₃ is characterized by spins that exhibit a triangular in-plane arrangement, in agreement with experimental results. Calculations for slabs of BaMnO₃ several layers thick revealed similar results. However, the addition of graphene led to overall magnetic softening in the substrate and changed the spin alignment at the interface to mostly out-of-plane and ferromagnetic. Monte Carlo simulations showed consistent results with the first principles calculations.

“This is a huge change for a magnetic system,” says Zanolli. “[These results show] that it is possible to use graphene to dramatically affect the properties of a substrate,” she says.

The researchers calculated the electronic band structure of the system using an approach that incorporates spin-orbit coupling, recently implemented in the first-principles code SIESTA. Their results show a large electronic gap (~10 meV) opening due to spin-orbit interaction at the graphene-BaMnO₃ interface. This gap is more than 300 times larger than the gap associated with spin-orbit coupling in pristine graphene.

To further characterize the system, the researchers calculated its anomalous Hall conductivity (AHC) as a function of chemical potential. The ACH is a result of the spin-orbit coupling in a ferromagnetic material. “We found that it is quantized, which is a clear signature of topological nontrivial properties,” Zanolli says. Furthermore, the results showed a background of trivial metallic bands, suggesting coexistence of metallic and topological insulating states that the researchers call a hybrid quantum anomalous Hall effect.

Zanolli expects that this result can be extended from graphene to other two-dimensional layered materials, and possibly from BaMnO₃ to other transition metal oxides. If so, this work suggests a path toward designing new materials with magnetic and topological properties ideally suited for specific applications.

Zhenhua Qiao, who leads the low-dimensional electronic transport group at the University of Science and Technology of China, is encouraged by this study and commends the approach. Dissipationless electronics based on graphene are highly anticipated, he says, and while theory has been proposed to realize the quantum anomalous Hall effect to this end, it cannot be achieved in pure graphene. “The most hopeful approach becomes considering charge-compensated n-p co-doping or utilizing a magnetic insulator,” says Qiao, who was not affiliated with this work. “To avoid the influence of the magnetic field, the ferromagnetic plane of an ABO₃-type ferromagnetic insulator is an ideal choice,” he says.

Read the abstract in Physical Review B.