Spin canting and orbital order in spinel vanadate thin films

We report on the epitaxial film growth and characterization of ${\mathrm{CoV}}_2{\mathrm{O}}_4$, a near-itinerant spinel vanadate, grown on (001) ${\mathrm{SrTiO}}_3$. The symmetry lowering of the unit cell from cubic in the bulk to orthorhombic in the films results in dramatic differences in the magnetic anisotropy compared to bulk, as determined from structural and magnetic characterization. Bulk cubic ${\mathrm{CoV}}_2{\mathrm{O}}_4$ has been found to defy predictions by showing orbital degeneracy seemingly lasting to very low temperatures, with only small anomalies in magnetization and neutron experiments signaling a possible spin/orbital glass transition at $T=90$ K. In epitaxial thin films presented in this paper, structurally tuning the ${\mathrm{CoV}}_2{\mathrm{O}}_4$ away from cubic symmetry leads to a completely different low temperature noncollinear ground state. Via magnetization and neutron scattering measurements we show that the 90-K transition is associated with a major spin reorientation away from the ferrimagnetic easy axis [001] to the [110] direction. Furthermore, the V spins cant away from this direction with extracted perpendicular moments providing evidence of a larger canting angle compared to bulk. This result indicates that compressive strain pushes the system deeper into the insulating state, i.e., away from the localized-itinerant crossover regime.

Figure 1

XRD scan of the 00L peaks of a 100-nm-thick ${\mathrm{CoV}}_2{\mathrm{O}}_4$ film and the ${\mathrm{SrTiO}}_3$ substrate. The film and substrate peaks are indicated with CVO and STO, respectively.

Figure 2

(a) $M$ vs $T$ for a 55-nm orthorhombic film (blue curve), a 100-nm orthorhombic film (red curve), and for a 300-nm orthorhombic film (black curve); each are measured while warming in $H=1000$ Oe after cooling the sample in zero field (ZFC, dotted line), and after cooling the sample in the measurement field (FC, solid line). The field is applied in the film plane, along $a$ or $b$. (Inset) $M$ vs $T$ for the 300-nm thick cubic polymorph film. (b) $M$ vs $T$ for the 300-nm thick orthorhombic film measured in $H=1000$ Oe (red triangles, $H$ applied out of plane of the film, along the $c$ axis; black squares, $H$ applied in the film plane along $a$ or $b)$, ZFC open symbols, and FC closed symbols.

Figure 3

Elastic neutron scattering of a 300-nm-thick ${\mathrm{CoV}}_2{\mathrm{O}}_4$ film grown on a $10\times 10{\mathrm{mm}}^2$ (001) ${\mathrm{SrTiO}}_3$ substrate. (a) (111) Bragg peak intensity as a function of temperature. The red line is a power-law fit. (b) $\left(\overline{2}\overline{2}0\right)$ (black squares), and (202) (red open circles) Bragg peak intensities as a function of temperature.

Figure 4

Elastic neutron scattering of a 300-nm-thick ${\mathrm{CoV}}_2{\mathrm{O}}_4$ film grown on a $10\times 10{\mathrm{mm}}^2$ (001) ${\mathrm{SrTiO}}_3$ substrate. (002) Bragg peak intensity as a function of temperature. (Inset) Radial scan of the (002) Bragg peak at 6 K (black squares) and at 120 K (red circles).