5f delocalization

The B1-B2 phase transition is the prominent feature in the high-pressure structural behavior of these compounds. Although the interactinide distances reach the Hill transition zone (Hill 1970) under pressure in some of these compounds, the 5f electrons do not seem to have any link to this phase transition, since it is also the typical pressure-induced structural transition for the alkali halides. In addition, for a structural phase transition linked to 5f delocalization, one would in general expect that the high-pressure structure has lower symmetry than cubic. This assumption is based on an empirical correlation between itinerant f-states and low-symmetry structures, and is thought to be connected with strong directionality of the hybridized (s,p,d,f) orbitals. 

From the Mossbauer hyperfine spectra of NpAl2 under pressures up to 8.3 GPa (Litterst et al. 1986) another interesting feature has been observed the 5f delocalization induced by reduction of atomic volume is linked to dynamical processes. It is evident from fig. 32 that the spectra at higher pressures cannot be derived simply by a reduction of hyperfine field. This would, for example, not explain the growth of the... 

Unlike 1,3-dithiepin anion 144a, the evidence for the instability of 145a and for the lack of aromaticity associated with lOn-electron delocalization through the sulfur atom has been reported 91,92). The reaction of the disodium salt of c/s-dimercaptoethylene (155) with either l,2-dibromo-3-propanol or l,3-dibromo-2-propanol yielded 6,7-dihydro-5f/-l,4-dithiepin-6-ol (156). Treatment of the methoxy derivative 157 derived from 156 with two equivalents of lithium dicyclohexylamide resulted in an effective elimination of methanol to give 5//-l,4-dithiepin (145) as a colorless liquid. Lithiation of 145 with n-butyllithium in tetrahydrofuran at —70 °C... 

In a) and P) the non bonding-hypothesis for 5 f electrons is retained, differences in cohesive energy being only due to promotion of outer electrons from one to another orbital state and ionization energies (or electron affinities) due to the different valence states attained. Therefore, any further discrepancy found with experimental values, is indicative of the metallic bonding introduced by delocalization of the 5f electrons (point y). 

As an example NpOs2 and even NpAl2 display a proportional decrease of the hyper-fine field and isomer shift, indicative of increasing 5 f delocalization while no change is detected in NpCo2Si2 where 5f electrons are expected to be well localized ... 

This chapter is devoted to photoemission spectroscopy and the related inverse photo-emission spectroscopy, which are well developed experimental tools to study occupied and empty electronic levels, respectively. Special emphasis is given to the 5f electrons and their localized or delocalized character. 

In most instances, the magnetic structure of a compound can be understood to be based on interacting localized spin centers, such as classical 3d/4d/5d transition metal ions and 4f lanthanide or 5f actinide cations with unpaired electrons. Note that while the assumption of localized moments is valid for many compounds comprising such spin centers, even partial electron delocalization in mixed-valence coordination compounds renders many localized spin models inapplicable. 

The situation improves for the 5f elemental metals due to the greater delocalization of the 5f as compared to the 4f electrons. In fact, the equilibrium lattice constants of the 5f elemental metals are well reproduced even with the f electrons treated as band states, as long as spin polarization is taken into account (12). The spin polarization "turns on" at Am, and correlates with an inferred localization of the f electrons. Even for such strongly correlated metals as the heavy fermion superconductors UPt3 and UBe-ia, the equilibrium lattice constant and bulk modulus are well-given oy LDA f band calculations 0M4). 

The physicochemical properties of actinide metals confirm the presence of band-like 5f electrons up to Pu. The participation of these 5f electrons in the metallic bond is assumed to begin with Pa. In the first half of the actinide series, 5f electrons are similar to d electrons in typical transition metals the 5f electron orbitals are more extended than 4f orbitals for the light actinides, 5f electrons are "delocalized" and hybridized in a rather large band with 6d and/or 7s electrons. Starting with Am, the 5f electrons are localized again, like 4f electrons in the lanthanides. 

It is generally accepted nowadays that the sequentially increasing occupation of 5f states dominates the electronic properties in the series of actinide elements (see table 2.1). The analogy with lanthanides, in which the 4f states are gradually filled, is not complete. The 4f electronic states are confined deeply in the core of the lanthanide ion and can be treated in most cases as localized. On the other hand, a non-negligi-ble overlap of the more extended 5f wave functions belonging to neighbouring actinide atoms in a solid leads to the delocalization of the 5f states which resembles the formation of the d band in transition metals. The question about the localized versus itinerant 5f electron behaviour has become one of the most central ones within electronic structure considerations. This controversial behaviour is quite well... 

The heavier actinides with localized 5f-electrons adopt the dhcp structure (similar to some lanthanides) at ambient pressure and temperature. At high pressures they undergo structural transitions to lower symmetry structures, marking the pressures at which 5f-electron delocalization takes place 23 GPa for Am (Benedict et al. 1985a), 43 GPa for Cm (Benedict et al. 1985b), 32 GPa for Bk (Benedict et al. 1984) and 41 GPa for Cf (Benedict et al. 1984). 

The An-An spacing in the Laves phases of the light actinides, with exception of AnAl2 compounds, is far below the Hill limit. Therefore the overlap of the 5f wave functions must be regarded as the main delocalizing mechanism. This provides the justification for the application of the band approach, at least in a first approximation. 

Discussion Since dv v exceeds the Hill limit in these materials the hybridization with ligand states dominates the mechanisms of the 5f electron delocalization. The d-states are situated well below EF in AnUX3 compounds and therefore they do not hybridize with the 5f states of the actinide atoms. If d-states with relatively high N(E) are present in the valence band (which is undoubtedly the case in UT3 compounds) the strength of the 5f-d hybridization becomes a very critical parameter. It depends mostly on details of the mutual position (overlap) of the d-states and the 5f-states which remain pinned at (or near) EF. 

The NpT2X2 compounds are usually treated as materials with localized 5f states, and crystalline electric-field effects are held responsible for the depression of the magnetic moments (2.4/tB is expected for the free-ion moment of Np). However, the heavy-fermion behaviour (found in NpCu2Si2) is able to introduce an instability of the localized 5f states. In this sense the reduction in hyperfine field at the Np nucleus could be understood as being due to a partial loss of the orbital moments resulting from 5f-electron delocalization. 

Similar to hydrogen also the addition of boron can be used to stimulate the expansion of the crystal lattice with a consequent 5f localization. This effect was illustrated in the system UB4-YB4, where magnetic order appears at intermediate concentrations, while both terminal phases are paramagnetic. In UB4 this is due to 5f-5f overlap and a considerable delocalization, whereas a weakening of the interatomic exchange interactions between the localized U-moments play a role for low U concentrations (Wallash et al. 1987 and further references quoted therein). 

The X-ray and neutron diffraction data mentioned previously have been used in conjunction with the technique of polarized neutron diffraction (at 4.2 K) to deduce spin-density distributions in [MnPc]. In further investigations it proved possible to determine individual 3d and 4j orbital populations on the manganese ion together with an estimate of 24% for the <5f-orbital electron density delocalized in the macrocyclic ring. From these studies it appears that the charge on the manganese is approximately -I-1 this charge appears to be achieved primarily by the loss of a 3d electron rather than a 4 electron. 

The choice between inner- and outer-sphere mechanisms is difficult to ascertain. In Table 12.14, the outer-sphere mechanism is imposed by the reducing agent. [Ru(NH3)5f + is an inert species and does not allow formation of bridging species. Although [Cr(bipy)3] is formally labile, the chelate effect may predispose this complex to an outer-sphere mechanism. The delocalized tt systems of the bipy ligands of [Cr(bipy)3] may lower the barrier for outer-sphere electron transfer relative to [Ru(NH3)g] MLCT in [Cr(bipy)3] may facilitate the electron transfer. 

Bulk magnetic data under applied pressure (Fournier et al. 1980) gave strong evidence for itinerant magnetism in UN, in contrast to UAs (and UP), where f electron itinerancy is weak, if present at all. The xSR data for the comparable AFM structure in UAs is not fundamentally different from that of UN, which lends direct evidence to the statement made earlier that xSR is not particularly sensitive to 5f electron delocalization. 

---