Kondo Volume Collapse and theγ→αTransition in Cerium

Untitled

“…However, it is in-triguing that the infinite dimensional one-band Hubbard model reduces formally to the impurity Anderson model and exhibits a similar central resonance, which might be associated with f electrons screening their own moments [83]. This blurs the distinction between the Mott transition picture [38], traditionally associated with the Hubbard model, and the Kondo volume collapse [39,40]. The distinction is further complicated in any mean field approach which would pick up only half (e.g., f ↑ d ↓ but not f ↓ d ↑ ) of a Kondo like singlet, and might therefore still look like f magnetic order.…”

“…The central resonance is the well known signature of the Kondo effect for the PAM, and in its impurity Anderson model context, the critical aspect of the Kondo volume collapse model for Ce [39,40]. However, it is in-triguing that the infinite dimensional one-band Hubbard model reduces formally to the impurity Anderson model and exhibits a similar central resonance, which might be associated with f electrons screening their own moments [83].…”

“…At low temperatures there is a much sharper peak which, given the data on the magnetic correlations, we believe is associated with antiferromagnetic ordering. We do not see any sharp feature associated with singlet formation, which drives the volume collapse transition according to the Kondo picture [39,40]. However, it is well known that the signature of singlet formation in the specific heat becomes sharp only for relatively small t f d , where, in this symmetric half-filled model, antiferromagnetism strongly competes [87].…”

“…(1) in the latter regime, by splitting the occupied and empty states to create a new entirely full band for which n(N − n) = 0. The major current debate concerns how the experimentally observed moment is actually first lost with increasing pressure (or W f /U f ) for the 4f and 5f metals, whether the f moment is first quenched similar to the manner discussed here and by the Mott transition model [38], or whether while still in a robust state it is first screened away by a surrounding cloud of valence electrons as in the Kondo volume collapse model [39,40].…”

“…Note that the loss of moment as detected by, e.g., magnetic susceptibility measurements may be due either to quenching of the f -shell moment itself or to its screening by the surrounding valence electrons. These two possibilities underly the Mott transition model of Johansson [38], and the Kondo volume collapse model of Allen and Martin [39] and Lavagna et. al.…”