Title:

Crystal fields for Er³⁺ in gold and zerofield splittings for Gd³⁺ in lanthanum ethylsulphate

In the first and dominant part of the thesis, a recently developed theory for Crystal Fields (CFs) in metals has been used, in its first order, to investigate the microscopic origin of CF splittings for the dilute Erɜ+:Au alloy. To first order there are three contributions to the CF splitting of the Rare Earth (RE) 4f electron ground state. One is due to penetration effects with ligand ions, another due to a 5d semi localised state about the RE site, the so called virtual bound state (vbs), and the third arises from the interaction of the 4f(RE) electrons with a discrete band state, denoted by lco >, which is split off from the slike majority conduction band of Gold. The coupling between the 5dvbs and the conduction band has been assumed weak. We find, within the limitations of our calculations, that the model used is quite capable of explaining the CF splitting of the ground level for Erɜ+:Au. In particular we have found that the penetration effects, in the limit of no screening, considerably enhance the Crystal Field Parameters (CFPs) over their Point Charge Model (PCM) values, which consist of a fourth order CFP of the wrong sign, three times smaller than the experimental value and a sixth order CFP of the right sign but five times smaller than the observed. Contributions to CFPs directly from neighbours may be described in terms of a Coulomb interaction using different effective total charges (which we call pseudoPoint Charges) on the ligands for the fourth and sixth order CFPs. The sixth order CFP is much more affected by such effects. When screening is introduced, in an approximate way, we find that it tends to increase (over the unscreened PCM values) the magnitude of the CFPs by as much as 20% principally by screening out the sizable contribution, of opposite sign, from the next nearest neighbours. We also find that the probability of having a reversal in sign for the fourth order CFP, with respect to PCM values, is small since in the regime of the screening constant where we might expect this effect the energy of the 5dvbs is negative corresponding to a bound state. We also estimate the CFPs of Erɜ+:Ag, within the same model. It becomes clear that the model fails to explain the observed CFPs, at least within the approximations made in our first order investigation, unless a basic assumption, concerning the nature of the conduction band, is modified. In the second part of the thesis we study the Zero Field Splitting (ZFS) of Gdɜ+:doped lanthanum ethylsulphate. Onsite excitations of the Gdɜ+: in this particular salt have recently been shown to produce a Spin Correlated Crystal Field (SCCF) which results in a ZFS of the same sign as those from experiment but, approximately, two times too large. Intersite oneelectron excitations have been invoked to produce additional contributions to the SCCF and so produce a ZFS which compensates the ZFS of the onsite mechanisms to finally produce ZFS in good agreement with experiment. The resultant ZFS is thus a delicate balance between a number of onsite mechanisms i.e. from the Gdɜ+ ion itself, and a contribution from the host in which it is embedded.
