PULSE-DURATION DEPENDENT CAPACITANCE ANALYSIS AND ITS APPLICATION TO COPPER GAAS[0.6]P[0.4]

Abstract

A novel technique known as Pulse-duration Dependent Capacitance Analysis (PDCA), which permits an accurate and simultaneous determination of the values of free energy of ionization and capture cross section of a deep trap in semiconductors at constant temperature is presented. Its formulation is based on the rigorous theoretical model of carrier capture kinetics proposed by Pons. The technique requires an accurate fit of a set of analytical formulae to the experimental data of transient capacitance amplitude obtained as a function of filling pulse duration. The approximations used in the formulation of the technique give rise to two data fitting criteria, one of which is on the filling pulse duration and the other on the range of free energy of ionization. The sensitivity of the fitting of the experimental data to the deep level parameters always ensures an accurate fit to be achieved rapidly. Experimentally, the implementation requires a simpler instrumentation than a conventional DLTS setup as no temperature scanning is required. The validity of the technique is demonstrated by studying the copper-related deep traps found in n-GaAs0.6P0.4. A study is undertaken to investigate the states of copper diffused into VPE grown n-GaAs0.6P0.4 and LEC grown n-GaAs. Substantial difference in the states of copper is found in the two semiconductors. It is conjectured to be due to the formation of two different types of copper-related deep traps, interstitial copper and substitutional copper. In copper-doped n-GaAs0.6P0.4, a dominant electron trap of free energy of ionization EC-0.14 eV is found. The capture cross section of the trap is about 1 x 10-21 cm2 at room temperature and has a weak temperature dependence. These properties are attributed to a non-repulsive center having a capturing mechanism involving multiphonon emission process with hardly any lattice relaxation. The anomalous distribution of deep traps and its evolution of spatial distribution with time under junction electric field reveal that the trap is positively-charged and has a high mobility under electric field. The center is thus identified as interstitial copper. To the best of our knowledge we believe this is the first positive identification of interstitial copper as electron traps in n-GaAs0.6P0.4. In copper-doped n-GaAs, no electron trap is detected in the bulk of the semiconductor. A reduction in the electron concentration after the copper diffusion suggests the existence of copper as substitutional copper. This is further supported by our observation of no measurable change in the trap distribution under junction electric field. The formation of substitutional copper in n-GaAs is tentatively attributed to the presence of vacancies arising from the LEC growth procedure.