Half Semimetallic Antiferromagnetism in the Sr$_2$CrTO$_6$ System, T=Os, Ru

W. E. Pickett,K.-W. Lee
DOI: https://doi.org/10.48550/arXiv.0712.4308
2007-12-28
Materials Science
Abstract:Double perovskite Sr$_2$CrOsO$_6$ is (or is very close to) a realization of a spin-asymmetric semimetallic compensated ferrimagnet, according to first principles calculations. This type of near-half metallic antiferromagnet is an unusual occurrence, and more so in this compound because the zero gap is accidental rather than being symmetry determined. The large spin-orbit coupling (SOC) of osmium upsets the spin balance (no net spin moment without SOC): it reduces the Os spin moment by 0.27 $\mu_B$ and induces an Os orbital moment of 0.17 $\mu_B$ in the opposite direction. The effects combine (with small oxygen contributions) to give a net total moment of 0.54 $\mu_B$ per cell in \scoo, reflecting a large impact of SOC in this compound. This value is in moderately good agreement with the measured saturation moment of 0.75 $\mu_B$. The value of the net moment on the Os ion obtained from neutron diffraction (0.73 $\mu_B$ at low temperature) differs from the calculated value (1.14 $\mu_B$). Rather surprisingly, in isovalent Sr$_2$CrRuO$_6$ the smaller SOC-induced spin changes and orbital moments (mostly on Ru) almost exactly cancel. This makes Sr$_2$CrRuO$_6$ a "half (semi)metallic antiferromagnet" (practically vanishing net total moment) even when SOC is included, with the metallic channel being a small-band-overlap semimetal. Fixed spin moment (FSM) calculations are presented for each compound, illustrating how they provide different information than in the case of a nonmagnetic material. These FSM results indicate that the Cr moment is an order of magnitude stiffer against longitudinal fluctuations than is the Os moment.
What problem does this paper attempt to address?