More than a hundred years ago, Guillaume discovered that the alloy Fe65Ni35, at equilibrium under normal pressure, was invar. Indeed, its thermal expansion coefficient was anomalously small over a wide temperature interval. It is now known that the invar effect in this alloy is related to magnetism.
Recently, the ground state spin configuration was computed using spin-density functional theory and supercell methods. According to these calculations, the ground state spin configuration was ferromagnetic at atomic volumes bigger than a specific critical atomic volume. At the critical atomic volume, a few spins flipped. The aim of this licenciate project was to study the spin-flip transition which happens in the random fcc Fe50Ni50, using computer codes based on spin-density functional theory and supercell methods.
During the first part of the project, I characterized the chemical environment of the Fe sites where the spins flipped. These sites were in an Fe-rich environment. I computed the effective exchange parameters of the classical Heisenberg hamiltonian of the random fcc Fe50Ni50. The sign of the effective exchange parameter at any site where a spin flipped changed at the critical atomic volume, whereas the sign of the exchange parameter at any other site remained constant. The dispersion in the effective exchange parameters increased, as the atomic volume decreased from a specific atomic volume towards the critical atomic volume.
In the second part of the project, I computed the relaxed ionic positions. For any type of bond, the dispersion in the nearest neighbour bond length was much bigger than the change, due to the ionic positions relaxation, in the average value of the nearest neighbour bond length. Also, I investigated the effect of the ionic positions relaxation on the local magnetic moments. This effect was found to be relatively small.
After I found these results, I wondered what the effect of the ionic positions relaxation on the dispersion of the effective exchange parameters was and if the change of the dispersion in the effective exchange parameters with the atomic volume could provide an explanation for the invar effect in Fe-Ni invar alloys.
Linköping: Linköpings universitet , 2006. , 32 p.