The Fe-Ni alloy is simulated by four ordered structures, whose total energies are obtained as a function of volume and magnetic moment by band-structure calculations employing the fixed-spin-moment (FSM) method. An analytic fit of these E(M,V) surfaces is made and permits an interpolation for varying Ni concentration. These parametrized surfaces allow the introduction of the thermodynamics of spin fluctuations. Among others, the following characteristic Invar properties are calculated: the magnetic contribution to the thermal expansion coefficient ${\mathrm{\ensuremath{\alpha}}}_{\mathit{m}}$, the critical pressure for the disappearance of magnetism ${\mathit{P}}_{\mathit{c}}$, the pressure dependence of the Curie temperature ${\mathit{dT}}_{\mathit{C}}$/dP, and the high-field susceptibility. These quantities agree well with experiment, especially their variation with Ni content. The key quantity is ${\mathrm{\ensuremath{\alpha}}}_{\mathit{m}}$, which shows a narrow minimum as a function of the Ni concentration before the \ensuremath{\alpha}\ensuremath{\rightarrow}\ensuremath{\gamma} (bcc\ensuremath{\rightarrow}fcc) phase transition occurs. Near the Invar composition the large negative contribution ${\mathrm{\ensuremath{\alpha}}}_{\mathit{m}}$ compensates the positive phonon part so that the total thermal expansion nearly vanishes near room temperature. The present combination of models provides new insight into the nature of the Invar effect.
Read full abstract