Deep level impurities in semiconductors for nuclear radiation detection

This thesis describes experiments on the behaviour of deep level defects in the semiconductors Ge, Si and GaAs. High-purity samples of these materials are often used for the fabrication of solid state nuclear radiation detectors; it is within this context that the work has been performed. Deep level impurities may have considerable effect on the performance of solid state devices. By trapping signal carriers (electrons and holes) they may change the effective lifetimes of these entities, and so affect parameters such as switching times in photoconductors, the efficiency of solid state lasers and lightemitting diodes, and the energy resolution of semiconductor radiation detectors. Positive uses of their effects include the increased switching times of certain Si diodes, higher quantum gain in some photoconductors and the production of high resistance compound semiconductor wafers. The deep level defects may be caused by contamination of the material with other elements, by lattice line and point defects, or associations of these imperfections and impurities. They may be present in the as-grown semiconductor crystal, introduced during processing of the material, or during the operation of the final device (e.g. radiation damage). No general formalism exists to explain or predict the properties of these deep level impurities, particularly because of their nonhydrogenic nature. A greater understanding of solid state physics than currently exists will need to be achieved to produce a theoretical treatment of their behaviour. The study of their properties is therefore desirable from both a device and a fundamental point of view. Chapter 1 describes experiments on radiation damage centres in Ge. The γ-irradiation of p-type Ge crystals grown from silica crucibles under an H\(_2\) atmosphere always produces two deep acceptor levels. We detail evidence showing these are most likely due to oxygen-vacancy complexes. Heat treating samples before irradiation appears to reduce the amount of oxygen available for production of the deep levels, and so these samples are hardened to γ-radiation damage. Similarly, Li ions drifted through the crystal cause radiation hardening of this material, possibly by binding oxygen into stable Li-O pairs, and also by direct passivation of the γ-induced defect centres. The thermal and electrical stability of these centres is also discussed. Proton and neutron irradiation of Ge produced acceptor levels only, whereas y-damage produced both donor and acceptor levels. Chapter 2 deals with deep level defects found in single crystal and polycrystalline GaAs. Many different defect species are present - even the high purity epitaxial wafers used to fabricate radiation detectors showed high trap densities. The Poole-Frenkel effect (field enhanced emission) and a magnetic field sensitivity were observed in one of the deep donor levels (E\(_c\) - 0.62 eV). Chapter 3 deals with the fabrication of thin, highly doped contacts to semiconductors by pulsed laser melting of an evaporated dopant layer. The use of Li to produce n\(^+\) layers has met with the most success, and good quality Si and Ge radiation detectors were fabricated in this fashion. The advantage and problems of the technique are discussed. Chapter 4 deals with the measurement of the energy levels and capture cross sections of defects related to over 25 different elemental impurities deliberately introduced into Ge. There appears to be a band of energies at approximately one-third the band gap of the material (E\(_g\) = 0.66 eV) into which many deep metal-related states (particularly donors) fall. The capture cross sections of these states for majority carriers are generally 10\(^{-16}\) to 10\(^{-18}\) cm\(^2\). These facts may be related to the notion that many of the energy levels measured for metal-related centres in Ge may be due to defects of complicated nature, rather than simple subsitutional or interstitial defects. Chapter 5 discusses experiments in which point defects in Ge and GaAs are neutralised by the incorporation of atomic hydrogen. Data on the depth and efficiency o.f the passivation as a function of hydrogen plasma exposure duration and temperature are presented. Significantly in Ge, copper-related defects may be neutralised to a depth of ~ 100 μm for a 3 hour plasma exposure at 300 ° C. Miscellaneous experiments on the behaviour of deep level defects are described in Chapter 6. They fall into an 'interest only' category and no detailed data is presented or conclusions drawn