Lattice site and thermal stability of transition metals in germanium

Although the first transistor was based on germanium, current chip technology mainly uses silicon due to its larger abundance, a lower price and higher quality silicon-oxide. However, a very important goal in microelectronics is to obtain faster integrated circuits. The advantages of germanium compared to silicon (e.g. a higher mobility of the charge carriers) motivates further research on germanium based materials. Semiconductor doping (e.g. introducing impurities into silicon and germanium in order to alter - and control - their properties) can be done by ion implantation or by in situ doping, whereby the host material is doped during growth. This thesis focuses on introducing dopants by ion implantation. The implantation as well as the subsequent measurements were performed in ISOLDE (CERN) using the emission channeling technique. Although ion implantation generates undesired defects in the host material (e.g. vacancies), such damage can be reduced by performing the implantation at an elevated temperature. In this thesis, the implanted ions are transition metals. In contrast to elements from group III (e.g. boron) and V (e.g. phosphorus), which act as electrical dopants, transition metal impurities in germanium introduce deep levels around the middle of the band gap, and therefore can act as compensating electrical defects. On the other hand, transition metals may be incorporated in non-magnetic semiconductors such as germanium to induce magnetic behaviour. Such materials, e.g. manganese doped germanium, are known as dilute magnetic semiconductors (DMS). Both electric and magnetic properties of transition metal impurities strongly depend on the occupied lattice sites. Although the behaviour of substitutional (S) impurities (i.e. transition metals on an originally occupied germanium position) is relatively well established, interstitial transition metals are poorly understood. In this work, we studied the lattice site of manganese and nickel implanted in germanium. It was observed that apart from substitutional locations, also interstitial positions were occupied. All the investigated 3d transition metals in germanium had at least a fraction of the impurities located at a near-bond-centered (BC) position. This BC site is located between two nearest neighbouring S sites (along the direction) and is related to an impurity-vacancy-complex or the split-vacancy configuration. When an impurity on the substitutional site traps one vacancy which is available due to the implantation process, the impurity spontaneously evolves to the BC site. The fraction of transition metals on the S site and on the near-BC location varied from 20% to 40% and from 20% to 30% respectively. In as-implanted nickel in germanium, 19$\pm$3% of the nickel atoms occupied an additional interstitial site: a T site with a large root-mean-square (rms) displacement. This tetrahedral site was interpreted to be related to defect complexes. The lattice sites were determined for different annealing temperatures. The S site was found to be stable up to at least 350C and the near-BC site up to 300°C for the investigated 3d transition metals. For manganese in germanium, the fitted fractions decrease as a function of increasing annealing temperature. This is likely due to the diffusion of defects during annealing, which are able to trap manganese atoms in disordered regions. This causes a decreasing amount of manganese atoms on high symmetry sites. When nickel was implanted in germanium, the fractions initially increased up to an annealing temperature of 350°C, whereby the annealing causes a recovery of the lattice structure. Afterwards, the fractions decreased, which we interpret as being due to clustering or inward long range diffusion. In conclusion, apart from substitutional locations, also near-BC sites are occupied for the already investigated 3d transition metals (manganese, iron, nickel and copper) in germanium. For manganese in germanium, these results are particularly interesting in the context of dilute magnetic semiconductors, since the magnetic behaviour of magnetic dopants strongly depends on the occupied lattice sites. The experiment on nickel in germanium, which constituted the first emission channeling measurement making use of a nickel isotope, showed the occupation of an unexpected interstitial site (the T site with a large rms displacement)