Possible Bose-Einstein condensate associated with an orbital degree of freedom in the Mott insulator CaCr O3

Whether $\mathrm{CaCr}{\mathrm{O}}_{3}$ is a Mott insulator or a correlated metal is still controversial. We have performed measurements of magnetization, specific heat, and thermal conductivity on $\mathrm{CaCr}{\mathrm{O}}_{3}$ samples selected from many batches of high-pressure synthesis. The single-crystal $\mathrm{CaCr}{\mathrm{O}}_{3}$ sample exhibits an unprecedentedly sharp transition at a N\'eel temperature ${T}_{N}\ensuremath{\approx}90\phantom{\rule{0.28em}{0ex}}\mathrm{K}$. The critical behavior of specific heat cannot be rationalized by the renormalization group theory for a second-order magnetic transition. More surprisingly, the thermal conductivity $\ensuremath{\kappa}$ exhibits an anomalous drop on cooling through ${T}_{N}$, which is opposite to all known influence on $\ensuremath{\kappa}$ from either spin or orbital ordering. We have argued, on the basis of anomalies found in all three measurements and structural data, for the coexistence of itinerant \ensuremath{\pi}-bonding electrons in a $c$-axis band and localized $xy$ electrons in $xy$ orbitals responsible for type-C antiferromagnetic order below ${T}_{N}$ and the occupation of a pure, localized $xy$ orbital undergoing a Bose-Einstein condensate at ${T}_{N}$.