Constructing a magnetic handle for antiferromagnetic manganites

An intrinsic property of antiferromagnetic materials is the compensation of the magnetic moments from the individual atoms that prohibits the direct interaction of the spin lattice with an external magnetic field. To overcome this limitation we have created artificial spin structures by heteroepitaxy between two bulk antiferromagnets ${\mathrm{SrMnO}}_3$ and ${\mathrm{NdMnO}}_3$. Here, we demonstrate that charge transfer at the interface results in the creation of thin ferromagnetic layers adjacent to $A$-type antiferromagnetism in thick ${\mathrm{NdMnO}}_3$ layers. A novel interference based neutron diffraction technique and polarized neutron reflectometry are used to confirm the presence of ferromagnetism in the ${\mathrm{SrMnO}}_3$ layers and to probe the relative alignment of antiferromagnetic spins induced by the coupling at the ferro- to antiferromagnet interface. A density functional theory analysis of the driving forces for the exchange reveals strong ferromagnetic interfacial coupling through quantifiable short range charge transfer. These results confirm a layer-by-layer control of magnetic arrangements that constitutes a promising step on a path towards isothermal magnetic control of antiferromagnetic arrangements as would be necessary in spin-based heterostructures like multiferroic devices.

Figure 1

Sketch of multilayer systems of ${\mathrm{NdMnO}}_3$ and ${\mathrm{SrMnO}}_3$ with different ${\mathrm{SrMnO}}_3$ thicknesses. Gray arrows indicate the spin orientation in a purely A-type AFM system, green and purple the ${\mathrm{NdMnO}}_3$ spins that align parallel and antiparallel with it. Every second AFM layer in the 2 ML sample has spins antiparallel to the continuous AFM structure, leading to destructive interference in neutron diffraction.

Figure 2

[11 ML-${\mathrm{NdMnO}}_3$/1|2|3 ML-${\mathrm{SrMnO}}_3{]}_{x40}$ superlattices on ${\mathrm{TbScO}}_3$ measured with polarized neutron reflectometry at 20 K with a 100 mT field. Spin-up neutron polarization is shown in brighter colors than the corresponding spin-down measurements. Solid lines are fits to the data.

Figure 3

Neutron diffraction $\mathrm{\Theta }-2\mathrm{\Theta }$ scan around AFM (0 0 1) peak from [11 ML-${\mathrm{NdMnO}}_3$/1 ML-${\mathrm{SrMnO}}_3{]}_{x40}$. The gray data and shaded area indicate the 100 K measurement of the structural substrate reflection that was subtracted from the lower T data presented here. Solid lines through the low T data indicate fits to two Voigt functions using the same Gaussian widths, relative peak intensity, and peak separation obtained from first fitting the substrate reflection.

Figure 4

Temperature dependent intensity at (0 0 1) from [11 ML-${\mathrm{NdMnO}}_3$/1|2|3 ML-${\mathrm{SrMnO}}_3{]}_{x40}$ measured with neutron diffraction.

Figure 5

Results of density functional theory calculations for ${\mathrm{NMO}}_7/{\mathrm{SMO}}_m$ superlattices using GGA. The $c$-lattice direction is shown horizontally. (a) Sketch of the interface region of the 4 simulations with effective exchange along the $c$-axis (${\mathrm{J}}_{\mathrm{C}}$) and the magnetic moments on the Mn-ions furthest away from the interface ${\mu }_{\mathrm{SMO}}/{\mu }_{\mathrm{NMO}}$. (b) Mn charge occupancy (blue squares) and magnetic moment (red circles) for the $m=2$ simulation (top) with a sketch of the interface atoms and orbitals (bottom).