In earlier experimental studies we have shown that in accordance with the principles of renormalization group theory the spin dynamics of ordered magnets is controlled by a boson guiding field instead by exchange interactions between nearest magnetic neighbors. In particular, thermal decrease of the magnetic order parameter is given by the heat capacity of the boson field. The typical signature of boson dynamics is that the critical power functions either at T=T_{c} or at T=0 hold up to a considerable distance from critical temperature. The critical power functions of the atomistic models hold asymptotically at T=T_{c} or at T=0 only. In contrast to the atomistic magnons field bosons cannot directly be observed using inelastic neutron scattering. However, for some classes of magnets the field bosons seem to have magnetic moment and thus are able to interact directly with magnons. This interaction, although weak in principle, leads to surprisingly strong functional modifications in the magnon dispersions at small q-values. In particular, the magnon excitation gap seems to be due to the magnon-boson interaction. In this communication we want to show that for small q-values the continuous part of the magnon dispersions can be fitted over a finite q-range by a power function of wave vector. The power function can be identified with the dispersion of the field bosons. It appears that for low q-values magnon dispersions get attracted by the boson dispersion and assume the dispersion of the bosons. This allows for an experimental evaluation of the boson dispersions from the known magnon dispersions. Exponent values of 1, 1.25, 1.5, and 2 have been identified. The boson dispersion relations and the associated power functions of temperature for the heat capacity of the boson fields are now empirically known for all dimensions of the field and for magnets with integer and half-integer spin quantum number. These are two 2× 3 exponent schemes.
Optical magnons should occur in magnets containing two in-equivalent magnetic species only. However, Heisenberg interactions between in-equivalent magnetic atoms can be expected to be weak. This is because free exchange of electrons between chemically different magnetic atoms appears not generally possible. To the best of our knowledge optical magnons have never been identified unambiguously experimentally. Confusion is provided by the fact that two magnon branches commonly occur in antiferromagnets with ferromagnetically ordered crystallographic planes and opposite spin orientations from plane to plane. This applies to MnO, EuTe, CoCl₂, Fe₂O₃, K₂FeF₄. Associated with the ferromagnetic planes is a particular low-energy magnon branch. The high-energy magnon branch is the antiferromagnetic branch and not an optical magnon. In Fe₃O₄ (magnetite), weak interactions between the Fe²⁺ moments and the Fe³⁺ moments are evidenced by the fact that the order parameters of the FeO and of the Fe₂O₃ subsystem have different temperature dependencies. The observed two magnon branches can be attributed to the Fe₂O₃ and to the FeO subsystem, respectively. This applies equally to the two observed magnon branches in mixed crystals such as Rb₂Mn_{0.5}Ni_{0.5}F₄, KCo_{0.71}Mn_{0.29}F₃ or Mn_{0.3}Co_{0.7}F₂ that can be understood as modified dispersions of the constituent materials.
Following the principles of renormalization group theory the typical experimental indications are discussed that in ordered magnets with a three-dimensional spin the dynamics of the spins is controlled by a boson guiding field instead by exchange interactions to the neighboring spins. The spins are, so to say, sensors to probe the dynamics of the relevant boson field. It is evident that these findings are not complementary but fundamentally different from atomistic concepts. The bosons are essentially magnetic dipole radiation emitted upon precession of the ordered moments. Within the individual domain the one-dimensional field has the character of a laser field. The field aligns all spins along its axis. In order that in cubic magnets three-dimensional dynamic symmetry can result a vector average over all one-dimensional boson fields of the individual domains is necessary. It is argued that this averaging process does not work in the critical temperature range of cubic EuS and EuO. As a result, the critical behavior of EuS and EuO is that of the one-dimensional boson field of the isolated domain and agrees with the critical behavior of the one-dimensional antiferromagnet MnF₂. For the magnets with 1D boson field and half-integer spin it is found that the rational exponents β = 1/3, γ = 4/3 and ν = 2/3 give an excellent account of the mean exponent values over the most accurately known experimental data. These exponents obey the scaling relation 2β = 3ν - γ .
The different meaning of dimensionality and universality in field theory and in spin wave theory is illustrated on account of experimental examples. In spin wave theory it is distinguished between the dimensionality of the spin and the dimensionality of the exchange interactions. According to Renormalization Group (RG) theory, these atomistic characteristics are unimportant for the critical dynamics. Instead by inter-atomic interactions the dynamics of the ordered state is determined by the excitations of the continuous magnetic medium. These excitations are bosons. Consequently, the dimensionality of ordered magnets has to be assessed to the dimensionality of the relevant boson field. The most serious consequence of RG theory is that the magnetic ordering transition also is executed by the boson field. Typical for boson dynamics is a finite width of the critical range. In the atomistic models universality applies asymptotically at T_{c} only. It is evident that the critical power functions of the field dynamics are different from those of the atomistic dynamics.
Spin wave theory of magnetism reveals two severe shortcomings. First, it considers non quantized classical spins and, second, the predicted temperature power functions for the thermal decrease of the magnetic order parameter hold asymptotically at T = 0 only. As experiments unambiguously show the dynamics is different for magnets with integer and half-integer spin and the "critical" power functions at T = 0 of type ΔM ≈ T^{ε} or at T = T_{c} of type ≈(T_{c}-T)^β hold over a finite temperature range, independent of spin structure. The finite critical range unequivocally indicates that the dynamics of the spins is controlled by a field of freely propagating bosons instead by exchange interactions. Consequently, field theories are necessary for description of the thermodynamics of ordered magnets. The experimental indications will be discussed that the field quanta are essentially magnetic dipole radiation emitted by the precessing magnetic moments. Since integer and half-integer spins precess differently the generated field quanta and the dynamics of the field are correspondingly different.
It is shown that atomistic spin wave theory gives no correct account of the temperature dependence of the magnetic order parameter. The experimentally observed universal temperature dependence can be explained only by a field theory of magnetism. This means that instead by interacting spins (magnons) the dynamics is controlled by a boson field. The field quanta can be supposed to be magnetic density waves with dispersions that are simple power function of wave vector. This results in the observed universality. In three dimensions the field quanta have no mass and linear dispersion and cannot be observed using inelastic neutron scattering. Experiments on standing magnetic waves in thin ferromagnetic films provide direct information on the dispersion of the field quanta. A careful analysis of the available experimental data indicates that the dispersion of the field bosons is ım q, ım q^2, and ım q^{3/2} in three, two, and one dimensions.
JavaScript is turned off in your web browser. Turn it on to take full advantage of this site, then refresh the page.