Silicon Carbide (SiC) has several favorable physical properties for the fabrication of highpower, high-temperature and high-frequency devices. Devices in SiC can operate at high temperatures due to the wide band gap and the high thermal stability of the material. Heat produced by power losses are efficiently dissipated since the thermal conductivity of SiC is high(better than for copper). The high electrical breakdown field strength and the indirect band gap make it possible to design high voltage devices with low power losses. The high saturation drift velocity makes SiC attractive for high frequencies devices. In addition, SiC is chemically inert which allows the device to operate in very harsh environment.

In order to control the carrier lifetime in SiC for bipolar device applications, it is necessary to intentionally introduce defects in the material. This can be done by diffusion of impurities or irradiation by high energetic particles such as electrons or ions. Electron irradiation of SiC introduces intrinsic defects such as vacancies, interstitials and antisites. Intrinsic defects are also introduced during the growth. For SiC, electron irradiation is an attractive technique since diffusion of impurities is difficult in SiC. Consequently, a detailed knowledge of the electronic properties of the introduced defects is necessary for utilizing the electron irradiation technique and to improve the carrier lifetime in unirradiated SiC.

In the thesis, defects associated with intrinsic defects in electron irradiated and in as-grown 6H and 4H SiC have been studied. The electronic properties of the defects have been characterized using various forms of diode capacitance transient techniques. Paper I reports an investigation of electron irradiation induced defects in 4H SiC. Annealing behaviour and dose dependence for these defects were investigated. For two of these traps, the temperature dependence of the capture cross section was measured. Capture cross sections of electron irradiation induced defects in 6H SiC were investigated in paper II. The temperature dependence of the capture cross section for the so-called Ei, El and E2 centers was measured and it was shown that the electron capturing process to El and E2 is governed by a multiphonon capturing process. Further, it is shown that the El and E2 centers act as recombination channels.

Many defects in semiconductors have a stable configuration. However, for some defects the configuration is altered when the charge state changes or if energy is supplied to the defect system. This gives rise to many interesting physical phenomena such as metastability and negative-U behavior of the defects. Both phenomena have been observed in SiC. In paper III observations of two complex metastable defects in 6H SiC are reported. These two defects always appear as an entity with similar binding energies and they are not possible to resolve by deeplevel transient spectroscopy (DLTS). However, one of the defect has three configurations labeled C₁, C₂ and C₃, while the other one has four configurations, labeled C₁, C₂, C₃ and C_4. The common notations C₁, C₂, and C₃ are used since their electronic behaviors are very similar. The configuration C₄ is only observed after hole capture to the defect or after annealing with bias applied to the diode at temperatures above 610 K. The defects with three configurations and with four configurations differ by a factor of two in concentration. It indicates a correlation with the quasi-cubic (2) and hexagonal (1) sites of the 6H SiC lattice. Two of the configurations, the C₂ and C₃ were shown to be stable only when they were occupied by an electron. The electron capture rate was measured for these configurations and it was shown that the electron capturing process to C₁ was responsible for the transformation process to these configurations. In paper IV a more detailed investigation of the configuration transition processes C₁ ➔ C₂ and C₁ ➔ C₃ is presented. An electron capturing and configuration transition model is presented and a transformation probability between the configurations has been defined and measured. In the introduction to this thesis additional photoionisation measurements are presented. The difference between the thermal position in the band gap and the optical ionization energy of C₂ is large. It indicates that a large lattice relaxation occurs when the defect is transformed to C_2. Configuration C₃ was not possible to ionize with light. It suggests that a very large lattice relaxation occur.

Electron irradiation of 6H and 4H SiC give rise to the so-called DLTS peaks E1/E2 and Z1, respectively. In paper V and VI it is demonstrated that both these peaks are associated to the two-electron emission from the acceptor levels of two negative-U centers. In both polytypes, two closely spaced negative-U centers are observed and it was suggested that the two centers may be associated to a defect residing at the cubic and hexagonal lattice sites, respectively. The electronic properties of the negative-U centers in 6H and 4H SiC are very similar. It indicates that the two negative-U defect systems in 4H and 6H SiC have a similar structure.