Exchange-Coupled Spin-Fluctuation Theory: Application to Fe, Co, and Ni

Abstract

Finite-temperature properties are modeled for the itinerant-electron ferromagnets Fe, Co, and Ni by employing a spin-fluctuation theory where the modes are coupled by interatomic exchange interactions. Our method is based on the density functional theory using the local density approximation. The latter yields all parameters derived from constrained ground-state properties of noncollinear spin configurations to calculate ab initio the Curie temperatures, the magnetic susceptibilities, and, furthermore, the hcp-fcc phase transition of Co. Our results are in fair agreement with experimental data. [S0031-9007(96)00537-6] PACS numbers: 71.15.Mb, 75.10.Lp, 75.30.Cr, The magnetic properties of transition metals at very low temperatures are well described by spin-polarized band theory provided this is based on the density functional formalism, a rather well-tested computational method being the local density approximation [1,2]. In fact, spin-polarized band theory can be seen as the ab initio version of Stoner-Wohlfarth theory [3,4] in which the magnetic moments are brought about by the itinerant d electrons whose spins align because of intra-atomic exchange interactions. Recent reviews can be found, e.g., in [5,6].At finite temperatures, however, Stoner-Wohlfarth theory fails to account for the magnetic properties in most cases, particularly so for Fe, Co, and Ni. The reason for this is that the magnetic moments are supposed to disappear through spin-flip excitations to the Stoner continuum, a process that costs too much energy leading to unphysically high Curie temperatures and a paramagnetic susceptibility that does not describe the experimentally observed Curie-Weiss law.Awareness of low-energy excitations for explaining the magnetic phase transition arose in the seventies predominantly through the pioneering work of Moriya, Hubbard, Hasegawa, Korenman et al., Gyorffy et al.,, and others; see, for example, [13,14]. The broad consensus reached [13] was that orientational fluctuations of the local magnetization represent the essential ingredients to a thermodynamic theory. Still, the detailed approaches seemingly differed considerably; thus, for instance, Gyorffy et al.[11] emphasized a picture of disordered local moments, whereas Korenman, Murray, and Prange used a fluctuating-local-band picture [10]. In spite of the considerable progress in the formulation of the problem, actual first-principles treatments of spin fluctuations that result in hard numbers for itinerant-electron systems are still rare, a notable exception is given by the work of Staunton and Gyorffy [15].In this Letter, using an approach different from that of Staunton and Gyorffy, in particular, not making any explicit assumptions about the local degree of order, we obtain first-principles estimates for magnetic properties of 3d metals on the basis of ground-state properties using the density functional theory. Assuming the adiabatic approximation for magnetic moments, we separate slow and fast motion, the time scale for spin fluctuations being much ...

Anab initioeffective Hamiltonian for magnetism including longitudinal spin fluctuations

“…This means there is no way for the size of the Ni moments to respond to the energy cost of orientational disorder by reducing the exchange splitting. A number of authors have proposed ways to lift this constraint 3,11 . In both of these methods the parameters of the classical model were obtained by the use of the structure inverse method.…”

Ab initioelectron theory for magnetism in Fe: Pressure dependence of spin-wave energies, exchange parameters, and Curie temperature

“…Other types of transverse fluctuations which are neglected in the two approximations, or longitudinal spin fluctuations which are totally neglected in the Heisenberg model, are more relevant. For an ab initio calculation of T c , therefore, the ab initio spin fluctuation theory of Uhl and Kübler 25 which encompasses both transverse and longitudinal fluctuations, is more appropriate, and first test calculations indeed yielded 26 ‫ץ‬T c /‫ץ‬pϽ0 when taking into account longitudinal fluctuations, and ‫ץ‬T c /‫ץ‬pϾ0 when neglecting them. Alternatively, Kakehashi 27 took both types of fluctuations into account by a variational approach to the degenerate-bands Hubbard model, and obtained for reasonable parameters of the model values for ‫ץ‬T c /‫ץ‬p which are larger in magnitude than our value but with a negative sign, again in disagreement with the experiment.…”

The disordered local moment approach to the spin-disorder resistivity of metallic ferromagnets

“…However, electrons are still expected to be scattered by fluctuating local moments [18,19]. In previous supercell LB calculations [8] the local moment in fcc-Ni was used as an adjustable parameter, and it was found that agreement with experiment requires the local moment of about 0.35 µ B .…”

“…In fcc-Ni, the DLM moment collapses to zero while the moment in a real material is expected to persist due to longitudinal spin fluctuations. [18,19] In this case we employ the fixed-spin moment (FSM) approach [20] and treat the local magnetic moment as an adjustable parameter to recover the experimental value of the SDR [9]. In the case of hcp-Gd we used both the open-core and the LDA+U approaches [21].…”