Abstract

The $f$-electrons in the lanthanides form an open shell with localized atomic-like orbitals. The traditional approach to study such systems uses atomic multiplet theory, crystal field theory and accounts well for the magnetic moments and shape of X-ray absorption and emission spectra.[1,2] However, it does not tell us anything about how these electrons affect the other electrons in the system, such as the lanthanide $d$-orbitals or the nitrogen $p$ orbitals in their nitrides. These are important to understand whether these materials are semiconductors or semimetals, to understand the hybridization effects of $f$-electrons with delocalized electrons in the system. The problem of how to reconcile band structure theory with the localized treatment of the $f$-electrons in a many-electron framework is not trivial. We review some of the existing approaches with a critical hindsight and discuss their successes and failures. One approach is to treat the open $f$-shell as core electrons [3,4] but which induce a spin-polarization on the other orbitals via the exchange effect. Relatedly, there are model approaches, in which the localized magnetic moments of the $f$-electrons are treated simply as localized spins with exchange interactions to the conduction electrons. [5] The self-interaction correction (SIC) approach is one attempt to make an orbital dependent Hamiltonian with the framework or density functional theory. [6] Another is the L(S)DA+U approach [7] in which the strong Coulomb interactions U and exchange interactions J of f-electrons are treated in a parametrized Hartree-Fock mean field treatment. But in this approach,[8-10] is the solution unique or does it depend on the initial choices of the density matrix for the f-electrons? Yet another approach is the quasi-particle self-consistent GW method, [11-15] which supposedly should describe the U and J effects via the self-energy and has no free parameters. Dynamic mean field theory using the Hubbard-I approximation[16-19] is perhaps the most sophisticated approach to bring together the atomic and band physics used so far. We will review these methods and illustrate some of their successes and shortcomings.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call