Proposed Calculations

The basic problems we are confronted with are:

1. The preferred location of the RE ions in Si and III-V materials.
2. The nature of any gap states.
3. The character of the bonding between the s, p, d, f orbitals of the RE and the surrounding ligands.
4. The strength of this bonding and the frequency of the vibrations of the ligands, when these involve light elements such as O, C, F and N.
5. The splitting and shifts of core atomic levels and the overlap of the f states with the ligands.
6. The influence of a surface confining potential to this transition ands its significance for the nanocrystalline effect seen in ZnS:Mn [7].

We propose to investigate these questions through an ab initio method [27] using local density functional techniques. This code has been extended to deal with d and f-electrons and can be run within the framework of a cluster or supercell.

The method has been used successfully to treat TM impurities in Si and diamond. The work on Ni complexed with H was the first to successfully deal with the Jahn-Teller distortion of Ni and deal with the structure of the Ni-H complex. This work has recently submitted to Physical Review Letters.

We plan two types of calculations:

1. To treat InP:Yb using the same technique as Schlüter et al ie keeping the f electrons as part of the core but allowing relaxation of the surrounding P coordination shell. This will yield the character of the two gap states seen in DLTS [12].
2. To investigate the effect of an O-environment on Si:Er where Er sits at a lattice site as well as an interstitial one. Since, there is no information on the location of O atoms, it is important to vary the number and positions of O to develop an understanding of how O affects both the gap levels and the intra-f dipole matrix elements. The latter can only be treated if f-orbitals are treated as part of the valence shell. The coordination of the rare-earth ion will then be varied between 2 to 4 O anions. The clusters/unit-cells are then relaxed, and the wave functions and optical cross -sections found. This approach is both novel and likely to lead to an understanding of the crucial parameters which determine a maximum optical cross section and is an essential first step if we are to suggests ways in which the optical properties are to be optimised.

This approach overcomes a possible objection to the present study in that as the environment to the RE impurity is not known, then theory is unlikely to help in either determining it, or in exploring the intra-f optical transitions. The aim of the present proposal is to investigate the optical properties for a whole range of different coordination shells. The ab initio method is ideal for this as no parameters are assumed and all atoms will be fully relaxed.

There will be a requirement for spin-orbit interaction to be built into the code as this is important for RE impurities. This section of code needs to include PVM and BLACS procedures allowing it to be run on a parallel architecture. The Newcastle group have already parallelised the code through the inclusion of SCALAPACK-2 routines.