A new type of magnetic semiconductor?

In recent years, calcium hexaboride (CaB6) doped with lanthanum has puzzled magnetism researchers, in large part because it retains a modest ferromagnetism even at 900 K - a surprisingly high Curie temperature for a compound which doesn’t contain traditional magnetic metals such as nickel or iron. All of the explanations proposed so far have been based upon the assumption that the parent material, CaB6, is a semimetal, a conclusion drawn from calculations performed within the framework of Density Functional Theory (whose importance was recognized with the award of a Nobel prize two years ago). However, this DFT framework was designed for calculating ground state properties and is known to systematically underestimate band gaps – the well-known semiconductor germanium is predicted by DFT to be a semi-metal. A better framework for calculating excitations is the so-called "GW approximation" which has been shown to predict the fundamental band gaps of anorganic semiconductors extremely well. In the Computational Materials Science group, such GW calculations were performed for CaB6 and it was found to have a bandgap of 0.8 eV, comparable to that of germanium. If this prediction is confirmed by experiment then important applications await the compound in the field of spintronics, where an electron's spin and not just its charge carries information. So far, it has been difficult to combine semiconductors with magnetic metals because of a conductivity mismatch which kills the spin polarized current injected into a semiconductor. Unlike doped CaB6, most conventional magnetic semiconductors have low Curie temperatures, that is, they typically cease to be magnetic at room temperature. In addition, it is frequently difficult to dope such semiconductors both p-type and n-type. This should not be a problem for CaB6. Entirely new applications such as reprogrammable logic can be envisaged if a true spin-transistor exhibiting gain could be made with CaB6 in addition to magnetic sensors and memory elements.

H.J. Tromp, P. van Gelderen, P.J. Kelly, G. Brocks, and P.A. Bobbert, CaB6: A New Semiconducting Material for Spin Electronics, Phys. Rev. Lett. 87, 016401 (2001).

Macintosh HD:Users:KellyPJ:Documents:Admin:MESA+:Jaarverslag:2001:CaB6.gif

Figure 1. Crystal structure of CaB6.