Biaxial stress driven tetragonal symmetry breaking and high-temperature ferromagnetic semiconductor from half-metallic ${\mathrm{CrO}}_2$

It is highly desirable to combine the full spin polarization of carriers with modern semiconductor technology for spintronic applications. For this purpose, one needs good crystalline ferromagnetic (or ferrimagnetic) semiconductors with high Curie temperatures. Rutile ${\mathrm{CrO}}_2$ is a half-metallic spintronic material with Curie temperature 394 K and can have nearly full spin polarization at room temperature. Here, we find through first-principles investigation that when a biaxial compressive stress is applied on rutile ${\mathrm{CrO}}_2$, the density of states at the Fermi level decreases with the in-plane compressive strain, there is a structural phase transition to an orthorhombic phase at the strain of $-5.6\%$, and then appears an electronic phase transition to a semiconductor phase at $-6.1\%$. Further analysis shows that this structural transition, accompanying the tetragonal symmetry breaking, is induced by the stress-driven distortion and rotation of the oxygen octahedron of Cr, and the half-metal-semiconductor transition originates from the enhancement of the crystal field splitting due to the structural change. Importantly, our systematic total-energy comparison indicates the ferromagnetic Curie temperature remains almost independent of the strain, near 400 K. This biaxial stress can be realized by applying biaxial pressure or growing the ${\mathrm{CrO}}_2$ epitaxially on appropriate substrates. These results should be useful for realizing full (100%) spin polarization of controllable carriers as one uses in modern semiconductor technology.

Figure 1

The crystal structure (a) and the Brillouin zone (b) of rutile phase $(P{4}_2/mnm)$, and the crystal structure (c) and the Brillouin zone (d) of the orthorhombic phase (Pnnm) of ${\mathrm{CrO}}_2$.

Figure 2

The $\mathrm{\Delta }d/d_0$ dependence of the structural parameters ($a,b,c,u_x,u_y$) (a) and the bond parameters ($l_{b1},l_{b2},{\alpha }_1,{\alpha }_2$) (b). The in-plane lattice constants and $u$ parameters are different along the $x$ and $y$ axes in the orthorhombic phase, and the structural phase transition can also be seen in the curves of bond lengths and angles.

Figure 3

Spin-resolved total and partial density of states of ${\mathrm{CrO}}_2$ for (a) the rutile equilibrium structure, (b) the orthorhombic structure with $d$ compressed by 6%, and (c) the orthorhombic structure with $d$ compressed by 8%.

Figure 4

Orbital-resolved partial density of states of Cr-$d$ and O-$p$ (a), (b) for the rutile equilibrium structure, (c), (d) for the rutile phase with $d$ compressed by 5%, (e), (f) for the orthorhombic structure with $d$ compressed by 6%, and (g), (h) for the orthorhombic structure with $d$ compressed by 8%.

Figure 5

Spin-polarized band structures of ${\mathrm{CrO}}_2$ (a), (b) for the rutile equilibrium structure, (c), (d) for the rutile phase with $d$ compressed by 5%, (e), (f) for the orthorhombic structure with $d$ compressed by 6%, and (g), (h) for the orthorhombic structure with $d$ compressed by 8%. The left column presents the majority-spin band structure, and the right column the minority-spin one.

Figure 6

The strain-dependent compression ratios of the horizontal bond ($l_{b1}$) and the O-O distance ($l_{\mathrm{O}}$).

Figure 7

The $|\mathrm{\Delta }d/d_0|$ dependence of total energy (a) and stress (b). It can be seen that for $|\mathrm{\Delta }d/d_0|\ge 5.6\%$, the total energy of the orthorhombic phase is only a little smaller than that of the tetragonal phase, but the stress is much smaller.