We present a comprehensive computational and experimental examination of the $\mathrm{C}{\mathrm{r}}_{1-x}{\mathrm{V}}_x{\mathrm{O}}_2 \left(0\le x\le 0.5\right)$ system. The entire series crystallizes in the rutile structure, but the compounds exhibit significantly different magnetic properties depending on $x$. Lattice parameter $a$ increases linearly with $x$, but the $c$ parameter is slightly reduced due to vanadium-vanadium bonding. The V-for-Cr substitution creates $\mathrm{C}{\mathrm{r}}^{3+}\text{−}{\mathrm{V}}^{5+}$ pairs; this leads to competition between ferromagnetic $\left(\mathrm{C}{\mathrm{r}}^{4+}\text{−}\mathrm{C}{\mathrm{r}}^{4+}\right)$ and antiferromagnetic $\left(\mathrm{C}{\mathrm{r}}^{3+}\text{−}\mathrm{C}{\mathrm{r}}^{3+}\right)$ interactions such that the materials change from ferromagnetic to antiferromagnetic with increasing $x$. Weak ferromagnetic interactions arising from $\mathrm{C}{\mathrm{r}}^{4+}$ are observed even in the seemingly antiferromagnetic phases with the exception of $x=0.5$, which contains only $\mathrm{C}{\mathrm{r}}^{3+}$. Density functional theory calculations are performed, but they incorrectly predict the $x=0.5$ phase to be a half-metal. This is caused by an incorrect prediction of the oxidation states of chromium and vanadium.

• Received 5 October 2015
• Revised 15 November 2015

DOI:https://doi.org/10.1103/PhysRevB.93.014405