Magnetic first-order phase transition and crossover associated with random anisotropy in crystalline DyxY1-xAl2

Magnetic phase diagram of superantiferromagnetic TbCu₂ nanoparticles

The structural state and static and dynamic magnetic properties of TbCu2 nanoparticles are reported to be produced by mechanical milling under inert atmosphere. The randomly dispersed nanoparticles as detected by TEM retain the bulk symmetry with an orthorhombic Imma lattice and Tb and Cu in the 4e and 8h positions, respectively. Rietveld refinements confirm that the milling produces a controlled reduction of particle sizes reaching ≃6 nm and an increase of the microstrain up to ≃0.6%. The electrical resistivity indicates a metallic behavior and the presence of a magnetic contribution to the electronic scattering which decreases with milling times. The dc-susceptibility shows a reduction of the Néel transition (from 49 K to 43 K) and a progressive increase of a peak (from 9 K to 15 K) in the zero-field-cooled magnetization with size reduction. The exchange anisotropy is very weak (a bias field of ≃30 Oe) and is due to the presence of a disordered (thin) shell coupled to the antiferromagnetic core. The dynamic susceptibility evidences a critical slowing down in the spin-disordered state for the lowest temperature peak associated with a spin glass-like freezing with a tendency of zv and β exponents to increase when the size becomes 6 nm (zv ≃ 6.6 and β ≃ 0.85). A Rietveld analysis of the neutron diffraction patterns 1.8 ≤ T ≤ 60 K, including the magnetic structure determination, reveals that there is a reduction of the expected moment (≃80%), which must be connected to the presence of the disordered particle shell. The core magnetic structure retains the bulk antiferromagnetic arrangement. The overall interpretation is based on a superantiferromagnetic behavior which at low temperatures coexists with a canting of surface moments and a mismatch of the antiferromagnetic sublattices of the nanoparticles. We propose a novel magnetic phase diagram where changes are provoked by a combination of the decrease of size and the increase of microstrain.

A neutron polarization analysis study of moment correlations in (Dy(0.4)Y(0.6))T(2) (T = Mn, Al)

We present a study of the magnetic moment correlations of two pseudo-binary C15 Laves phase compounds, (Dy(0.4)Y(0.6))Mn(2) and (Dy(0.4)Y(0.6))Al(2), both of which have spin-glass-like magnetic ground states at low temperature. We use neutron powder diffraction with polarization analysis to isolate the diffuse scattering associated with the correlated spin-glass ground state, and compare and contrast the two systems. Despite there being differences of correlation length scale, we discover that the moment-moment correlations of these two disordered magnets are quite similar over a short range, and hence conjecture that the Mn ions in (Dy(0.4)Y(0.6))Mn(2) have little influence on the ground-state magnetic properties.

Spin models with random anisotropy and reflection symmetry

We study the critical behavior of a general class of cubic-symmetric spin systems in which disorder preserves the reflection symmetry s(a) --> -s(a) , s(b) --> s(b) for b not equala . This includes spin models in the presence of random cubic-symmetric anisotropy with probability distribution vanishing outside the lattice axes. Using nonperturbative arguments we show the existence of a stable fixed point corresponding to the random-exchange Ising universality class. The field-theoretical renormalization-group flow is investigated in the framework of a fixed-dimension expansion in powers of appropriate quartic couplings, computing the corresponding beta functions to five loops. This analysis shows that the random Ising fixed point is the only stable fixed point that is accessible from the relevant parameter region. Therefore, if the system undergoes a continuous transition, it belongs to the random-exchange Ising universality class. The approach to the asymptotic critical behavior is controlled by scaling corrections with exponent Delta= -alpha(r) , where alpha(r) approximately -0.05 is the specific-heat exponent of the random-exchange Ising model.