Enhancing the coercivity of ferrimagnetic NiCo2O4 films via vicinal substrate growth
Abstract
Ferrimagnetic spinel oxides with high Curie temperature and perpendicular magnetic anisotropy are attractive for spintronic memory applications, yet optimizing their coercivity and thermal stability for robust device operation remains a challenge. Here, we report a substrate engineering approach that substantially improves the magnetic properties of epitaxial NiCo2O4 (NCO) films. On conventional flat MgAl2O4 (001) substrates, NCO films exhibit a coercivity of only ∼50 Oe at 14 nm thickness at 300 K and suffer from rapid performance degradation under high current density. We identify surface Ni segregation as the underlying cause through polarized neutron reflectometry, which reveals a ∼15 Å magnetically inhomogeneous surface layer, and X-ray photoelectron spectroscopy depth profiling, which quantifies 4.1% excess Ni at the surface relative to stoichiometric composition. By introducing 3° miscut vicinal substrates, we induce step-flow growth that kinetically suppresses cation segregation. The resulting films show a coercivity of ∼150 Oe at 14 nm at 300 K and achieve a >5-fold enhancement (∼260 Oe vs. ∼50 Oe) in 7 nm ultrathin films at 300 K, which also exhibit the largest anomalous Hall resistance favorable for device readout. Critically, the optimized films maintain square hysteresis loops under current densities up to 2000 µA, whereas flat-substrate films exhibit nearly complete magnetic collapse under the same conditions. Temperature-dependent anomalous Hall effect scaling analysis reveals that the chemically uniform films exhibit a finite scattering-independent Hall conductivity component, σ(0)xy = 24.3 Ω−1 cm−1, consistent with the recovery of the intrinsic Berry-phase contribution to anomalous Hall transport. These results demonstrate that controlling epitaxial growth kinetics through substrate engineering provides an effective route to high-coercivity oxide films for spintronic device applications.

Please wait while we load your content...