L Journal of Alloys and Compounds 271–273 (1998) 831–836 Optical and magneto-optical spectroscopy of uranium and plutonium compounds: recent theoretical progress a, 1 ,a 1 ,a 1 ,a a b * P.M. Oppeneer , A.Ya. Perlov , V.N. Antonov , A.N. Yaresko , T. Kraft , M.S.S. Brooks a Max-Planck Research Group ‘ Theory of Complex and Correlated Electron Systems’, University of Technology, D-01062 Dresden, Germany b European Commission, Joint Research Center, Institute for Transuranium Elements, D-76125 Karlsruhe, Germany Abstract First-principles calculations are reported which illustrate that, for those actinide compounds where the 5f electrons are sufficiently delocalized, energy band theory based upon the local spin-density approximation (LSDA) describes the optical and MO spectra reasonably well. Examples which we examine in detail are URhAl and UFe . The delocalized LSDA approach meets limitations for those actinide 2 compounds, where the electrons in the correlated 5f shell are nearly localized. Just as in the case of lanthanide compounds having localized 4f electrons, a satisfactory description of the optical spectra could be obtained by using a generalization of the LSDA, in which explicitly f electron Coulomb correlations are taken into account (LSDA1U approach). A third group consists of compounds in which the 5f electrons are neither fully delocalized nor localized, but have experimentally been classified as quasilocalized. The suitable theoretical approach to such compounds is yet to be resolved. We further consider the Pu monochalcogenides, the unusual physical properties of which were previously treated with different models, and discuss the optical spectrum of PuTe.  1998 Elsevier Science S.A. Keywords: First-principles calculations; Actinide; URhAl; UFe ; PuTe; Optical spectrum 2 1. Introduction more detail, and see how many compounds have actually been classified as having localized 5f electrons, then it One of the most intriguing aspects of actinide com- appears that only a few compounds could be classified as pounds is the great variability in the localization degree of such. The clearest example is UPd , where experimentally 3 the 5f electrons. Varying from one actinide compound to the 5f electrons were detected at 1 eV below the Fermi another, the 5f electrons may range from being nearly energy ( E ) [4], and a sharp crystal field (CF) transition F localized to being practically itinerant. The 5f localization was observed [5]. The latter is, in analogy to the CF tendency and concomitant physical properties have been transitions in lanthanides, a fingerprint of localized f extensively investigated for uranium intermetallic com- electrons. Very recently, an inelastic neutron peak was pounds (see, e.g. Refs. [1,2]). The 5f localization tendency observed for URhAl, which could be due to an inter- may be looked upon in terms of the 5f band width, which multiplet transition [6]. This observation thus raises the is narrower than the 3d band width, yet broader than that interesting question if URhAl could possibly be a localized of 4f [3]. These two enveloping bounds mark the different f material. For quite a number of other U compounds it has approaches to treating the 5f electrons that have become been found that the 5f electrons are not really localized, customary: either a model of localized f electrons is neither are they fully delocalized. This sort of intermediate adopted, which has proven to be applicable to explaining case has been termed ‘semilocalized’ or ‘quasilocalized’. many properties of lanthanides, or a delocalized band An experimental definition for semilocalized 5f electrons is model is adopted. The latter has proven to be the valid that the occupied 5f electrons are detected at 0.5–0.8 eV approach for transition metals. below E . This is found to be the case for, e.g. USe [7], F If we now consider the group of U intermetallics in UTe [8], and UNiSn and UPtSn [9]. In various other U compounds the 5f electrons are considered to be delocal- ized. Well-known examples are UN [10] and UFe [11]. 2 * Corresponding author. Tel.: 149 351 4634282; fax: 149 351 This (de)localization classification is somewhat simplified, 4637029; e-mail: peter@tmps08.mpg.tu-dresden.de 1 because the origin of 5f delocalization (due to direct f–f Permanent address: Institute of Metal Physics, Academy of Sciences of Ukraine, 252680 Kiev, Ukraine. overlap or f ligand hybridization) is not considered. 0925-8388 / 98 / $19.00  1998 Elsevier Science S.A. All rights reserved. PII: S0925-8388(98)00228-X