Abstract
Complex oxides offer rich magnetic and electronic behavior intimately tied to the composition and arrangement of cations within the structure. Rare earth iron garnet films exhibit an anisotropy along the growth direction which has long been theorized to originate from the ordering of different cations on the same crystallographic site. Here, we directly demonstrate the three-dimensional ordering of rare earth ions in pulsed laser deposited $(Eu_xTm_{1-x})_3Fe_5O_{12}$ garnet thin films using both atomically-resolved elemental mapping to visualize cation ordering and X-ray diffraction to detect the resulting order superlattice reflection. We quantify the resulting ordering-induced 'magnetotaxial' anisotropy as a function of Eu:Tm ratio using transport measurements, showing an overwhelmingly dominant contribution from magnetotaxial anisotropy that reaches 30 kJ m$^{-3}$ for garnets with $x = 0.5$. Control of cation ordering on inequivalent sites provides a strategy to control matter on the atomic level and to engineer the magnetic properties of complex oxides.
Introduction
Rare earth iron garnets (REIGs) are pivotal materials in the development of high-speed magnetic memory and logic devices. Their utility stems largely from their ability to support perpendicular magnetic anisotropy (PMA), where the easy axis of magnetization is oriented out of the film plane. For decades, a "growth-induced" anisotropy was observed in these films that could not be fully accounted for by conventional mechanisms like epitaxial strain or simple magnetocrystalline effects.
A theory proposed in the 1970s suggested that this phenomenon arises from the preferential occupancy of specific sites within the complex garnet unit cell by different rare earth ions (e.g., Europium and Thulium) during the growth process. This subtle atomic ordering lowers the overall symmetry of the crystal. However, direct observation of this "atomic seating chart" has remained elusive due to the chemical similarity of the sites and the sub-nanometer scale of the ordering. This study provides the first direct visualization of this 3D atomic arrangement and quantifies its impact on magnetic anisotropy.
Methods and Structural Analysis
The researchers synthesized a composition series of $(Eu_xTm_{1-x})_3Fe_5O_{12}$ thin films using Pulsed Laser Deposition (PLD) on (111) and (110) gadolinium gallium garnet (GGG) substrates. By varying the $x$ value from 0 (pure TmIG) to 1 (pure EuIG), the internal chemistry and resulting anisotropy could be systematically tuned.
The primary evidence for atomic ordering was obtained through Scanning Transmission Electron Microscopy (STEM) combined with Energy Dispersive Spectroscopy (EDS). This technique allowed for the mapping of individual Eu and Tm atoms within the garnet lattice with sub-nanometer precision. Furthermore, high-resolution X-ray diffraction (HRXRD) was employed to identify "forbidden" superlattice reflections, which serve as a structural fingerprint of a repeating 3D ordered pattern.
As shown in the structural models, the garnet unit cell is complex, with rare earth ions occupying dodecahedral sites. The ordering observed is not a simple layering but a sophisticated 3D distribution where specific ions prefer certain inequivalent sites (labeled as $\alpha$, $\beta$, and $\gamma$ in (110) films), effectively breaking the cubic symmetry of the parent material.
Magnetic Anisotropy Quantification
To isolate the contribution of atomic ordering to the total magnetic energy, the team utilized Spin-Hall Magnetoresistance (SMR) measurements. This allowed them to decouple various energy terms, including magnetostatic (shape) anisotropy ($K_{MS}$), magnetoelastic (strain) anisotropy ($K_{ME}$), and the newly defined magnetotaxial anisotropy ($K_{MT}$).
The results demonstrate that $K_{MT}$ is the overwhelmingly dominant contribution at intermediate compositions. For the $x = 0.5$ film, $K_{MT}$ reached approximately 30 kJ m$^{-3}$. Notably, this ordering-induced anisotropy remains significant even when the film is grown under conditions where strain-induced anisotropy ($K_{ME}$) is minimized or zero. This confirms that the "magnetotaxial" effect—a term derived from the Greek taxis for order—is the primary driver of PMA in these complex oxides.
Key Insights and Discussion
The study introduces "magnetotaxial anisotropy" as a precise term to describe what was previously labeled "growth-induced anisotropy." This shift emphasizes that the magnetic properties are a direct consequence of the 3D atomic arrangement rather than just a byproduct of growth kinetics.
The transition from cubic to trigonal symmetry "unlocks" magnetic behaviors that are otherwise forbidden. Interestingly, films grown via PLD exhibited magnetotaxial anisotropy values nearly twice as high as those reported for films grown by Liquid Phase Epitaxy (LPE). This suggests that the high-energy nature of the PLD process may provide superior control over atomic placement, allowing for more pronounced site selectivity.
Conclusion and Future Outlook
This work provides definitive visual and quantitative proof of the long-standing theory regarding cation ordering in REIGs. By demonstrating that magnetic properties can be engineered through the precise 3D arrangement of ions, it opens a new path for atomic-level material design.
However, challenges remain. The observed atomic ordering is metastable; high-temperature annealing (e.g., 1250°C) causes the ions to redistribute randomly, leading to the disappearance of magnetotaxial anisotropy. Additionally, while PLD is an excellent tool for research, scaling this level of atomic control to industrial manufacturing processes like MOCVD will require significant engineering innovation. Future research will likely explore how different growth orientations and substrate choices can further tune these magnetotaxial effects.
Summary updated