Conditions for Magnetism in Pu Systems

While light actinide elements are Pauli paramagnets with the itinerant 5f states at the Fermi level, the localization for actinides behind Am leads to magnetic ordering analogous to lanthanides. The specific position of Pu just before the localization threshold should make it very sensitive to external variables, but various Pu allotropic phases have surprisingly nearly identical weak magnetic susceptibility, despite a large volume expansion exceeding 20% for the fcc -Pu comparing to monoclinic -Pu. This situation poses a problem for theoretical description, because LDA or GGA calculations indicate the formation of magnetic moments for the expanded phases1. Our recent LDA+U calculations2 (around mean field version), which yield a non-magnetic ground state (S = 0. L = 0) and correct cohesion properties, suggest that the key, at least for Pu, can be the 5f-count approaching the nonmagnetic 5f 6 state, i.e. much higher than the originally expected value close to 5.0. The calculations also indicate3 that the magnetic order is not established, when the lattice is expanded by Am doping, or if the dimensionality is reduced. For the terminal concentration of pure Am the LDA+U calculations3 correctly reproduce the non-magnetic 5f 6 state. The main characteristics of the Pu-5f states is the density of one–electron states concentrated around 1 eV binding energy, while the empty states of the 5f7/2 character are about 4 eV above EF. Variations with increasing Am concentration in the fcc Pu-Am solid solution were found negligible. If this picture is credible, the high value of the g-coefficient of low-temperature electronic specific heat, as well as high-intensity f-emission observed close to EF by PES, should have been due to many-body states, as suggested in Refs.4,5. One way to test the adequacy of our LDA+U calculations is to study experimentally the behaviour of expanded Pu. For this purpose, the doping by Am is the most efficient tool. Unlike the common -phase stabilizing dopings by Al or Ga, the doping by Am expands the lattice, and the expansion is larger than e.g. for the Ce doping. Considering the alternative scenario, at which a narrow 5f band at EF is the key ingredient bringing -Pu to the verge of magnetism, such expansion should lead to increase of g, presumably to onset on magnetic order, and all properties would undergo a dramatic development due to the changing density of one-electron states at EF, increasing up until it is suddenly forced to leave the Fermi level due to incipient localization.

HAVELA Ladislav;  JAVORSKY Pavel;  WASTIN Franck;  COLINEAU Eric;  GOUDER Thomas;  SHICK A.B.;  DRCHAL V.; 

2007-01-10

U.S. Department of Energy by University of California, Lawrence Livermore National Laboratory

JRC35876