Papers by Keyword: Z_1/2

Abstract: Germanium (Ge) doping of 4H silicon carbide (SiC) has recently attracted attention because a conductivity-enhancing effect was reported. In this work, we report on an experimental and theoretical approach to elucidate this effect. Ge and tin (Sn) – a second candidate of group IV elements – have been implanted into n-type 4H-SiC. Despite the expected isoelectric nature of Ge and Sn, a more efficient annealing of implantation-induced defects was observed compared to noble gas implantation with identical simulated initial implantation damage. In particular, a strong reduction of the prominent Z_1/2 defect was observed. Density functional theory calculations under equilibrium conditions show that Ge is mainly incorporated on a substitutional silicon lattice site without creating new charge transition levels in the bandgap. The low abundance of other Ge-related defects suggests that kinetic mechanisms should be responsible for the observed effect of group IV doping.

Abstract: In this paper the impact of high temperature annealing on the formation of intrinsic defects in 4H-SiC such as Z_1/2 and EH_6/7 was examined. Therefore, three epitaxial layers with various initial concentrations of the Z_1/2- and EH_6/7-centers (10¹¹ – 10¹³ cm^-3) were investigated. It turns out that depending on the initial defect concentration the high temperature annealing leads to a monotone increase of the Z_1/2- and EH_6/7-concentration in a temperature range from 1600 to 1750°C. For a defined temperature above these values, the resulting defect concentration is independent of the sample’s initial values. Beside the growth conditions themselves such as C/Si ratio the thermal post-growth processing has a severe impact on the carrier lifetime which must be taken into account during device fabrication.

Abstract: It has been clarified that Z_1/2 center, a well known deep level as a lifetime killer, can be reduced to the concentration below 10¹¹ cm^-3 by thermal oxidation or C⁺ implantation plus Ar annealing. In this study, the authors investigate the trap-reduction phenomena systematically (experimentally), and propose a model to analyze the phenomena. Furthermore, prediction of the defect distributions is realized by solving a diffusion equation in accordance with the trap reduction model. This analytical model can explain almost all experimental data: oxidation-temperature dependence, oxidation-time dependence, and initial-Z_1/2-concentration dependence of the defect reduction. Based on these results, the authors accomplish to eliminate the Z_1/2 center to a depth of 100 μm in the sample with a relatively high initial-Z_1/2-concentration of 10¹³ cm^-3 by thermal oxidation at 1400°C for 16.5 h.

Abstract: Deep Level Transient Spectroscopy (DLTS) and Double-correlated DLTS (DDLTS) measurements have been conducted on Schottky contacts fabricated on n-type 4H-SiC epilayers using different contact metals in order to separate the EH₆- and EH₇-centers, which usually appear as a broad double peak in DLTS spectra. The activation energy of EH₆ (E_C - E_T(EH₆) = 1.203 eV) turns out to be independent of the electric field. As a consequence, EH₆ is acceptor-like according to the missing Poole-Frenkel effect. Therefore, it can be excluded that the EH₆-center and the prominent acceptor-like Z_1/2-center belong to different charge states of the same microscopic defect as theoretically suggested. It is proposed that EH₆ is a complex containing a carbon vacancy and another component available at high concentrations. The activation energy of EH₇ (E_C - E_T(EH₇) = 1.58 eV) has been evaluated indirectly by fitting the DLTS spectra of the EH_6/7 double peak taking the previously determined parameters of EH₆ into account.

Abstract: Annealing of the Z1/2 and EH6/7 has been studied by DLTS after ion implantation of MeV Si ions and subsequent annealing in either N2 or O2 at 1150 °C, in the dose range 1 - 4 × 108 Si / cm2. It is found that the annealing rate of these prominent defects is greatly enhanced after thermal oxidation, and in particular close to the surface area, due to injection of a defect species which annihilates with both Z1/2 and EH6/7. The migration part of the diffusion coefficient of the injected defect is established to be in the range 1 – 2 × 10-8 cm2/s, and the measured concentration versus depth profiles of both Z1/2 and EH6/7 are accurately simulated by a simple model.