A deep-level spectroscopic technique for determining capture cross-section activation energy of Si-related DX centers in AlxGa1−xAs

We find the DX centers in Si-doped AlAs for the first time. The activation energy is measured as 0.56 eV from deep level transient spectroscopy (DLTS). The DX centers in n-AlAs exhibit a large capture energy 0.5 eV and a persistent photoconductivity. These properties are similar to those of the DX centers in AlxGa1−xAs with x∼0.3. However, the carrier concentration in the DX centers revealed by DLTS is not linearly proportional to Si donor concentration. This result is interpreted by the band structure that the DX center level lies at 30 meV above the X-conduction band (CB) minima and at 150 meV below the L-CB minima. The DX center is found not to be associated with the X-CB minima, but the L-CB minima.

Several deep level defects were observed by conventional deep level transient spectroscopy (DLTS) and high-resolution Laplace DLTS (LDLTS) in n-type 4H-SiC junction barrier Schottky diodes. We have shown that the broad DLTS peak labeled Z1/2 has, in fact, two components, Z1 and Z2, with activation enthalpies for electron emission of 0.63 eV and 0.68 eV, respectively. The reorientation process between these two components was observed. A combination of double-correlated DLTS and LDLTS demonstrated an anomalous reduction of the emission rate and an increase of the activation enthalpy of Z2 with an increase of the reverse bias applied to the diode. The possible explanation of this phenomenon could be correlated with a tensile stress in epitaxial SiC layers. The results observed are discussed in the frame of the model that correlates Z1 and Z2 with carbon vacancies (VC), located at hexagonal (h) and cubic (k) lattice sites, respectively. We also discussed the origin of other traps E0–E5 with particular emphasis on a N-related shallow donor level located at 0.04 eV below the conduction band, which has never been previously reported by DLTS studies.

Deep-level transient spectroscopy (DLTS), which is widely used to characterize deep impurity centers in semiconductors, assumes a single exponential wave form for the transient junction capacitance. When there are several closely spaced energy levels this assumption is no more valid, and the conventional DLTS may lead to errorneous results. To overcome this difficulty we propose here a novel method which we call the multi-exponential DLTS(MEDLTS). The transient wave form of the junction capacitance is directly analysed into multi-exponential compouents using the nonlinear least-squares analysis program “DISCRETE” developed by Provencher. The resolved time constants τ of these components are then displayed in the form of aT2τ−1/T plot. According to the results of simulation with various parameters MEDLTS is shown quite effective to resolve closely spaced energy levels which can not be resolved by the conventional DLTS. As an example of the application of this method deep levels in Si: Au were investigated. The results have shown that a single peak in conventional DLTS actually consists of two adjacent levels with activation energies and capture cross-sectionsEB1=0.49 eV,σB1=1.1×10−14cm2 andEB2=0.46 eV,σB2=1.3×10−15 cm2 and with amplitude ratio 1∶1.

Si doped GaN grown by molecular beam epitaxy on sapphire substrates were characterized by capacitance transient spectroscopy. Conventional deep level transient spectroscopy (DLTS) measurements displayed six deep level defects, labeled A1, A, B, C1, C, and D, with activation energy ranging from 0.20 to 0.82 eV below the conduction band. Based on the logarithmic dependence of the DLTS spectral peaks on the filling pulse width, it is deduced that the defects A, B, C, and D are concentrated in the vicinity of line dislocations. Double-correlation DLTS (DDLTS) measurements, on the other hand, showed that only defects A (0.82 eV) and D (0.22 eV) exhibited deep donor-like characteristics. Following a 1.0 MeV electron irradiation of the GaN sample, one radiation-induced peak, E, with activation energy less than 0.20 eV was observed in the DLTS spectrum. However, after annealing at 350 °C, this DLTS peak intensity was found to diminish significantly.

A new method of differentiating the deep level transient spectroscopy (DLTS) signal is used to increase the resolution of conventional DLTS. Using this methods, more than one single deep level with small differences in activation energy or capture cross section, which are often hard to determine by conventional DLTS, can be distinguished. A series of lattice-mismatched InxGa1-xP samples are measured by improved DLTS to determine accurately the activation energy of a lattice-mismatch-induced deep level. This level cannot be clearly determined using conventional DLTS because the two signals partly overlap each other. Both the signals are thought to originate from a phosphorus vacancy and lattice-mismatch-induced defect.

Evidence for two Si-related DX like centers in Al xGa 1− xAs and GaAs

The defects formed in a 6H-SiC p <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">+</sup> n diode by irradiation with 1 MeV electrons have been studied by both single-alpha-particle-induced charge transient spectroscopy and conventional deep level transient spectroscopy (DLTS). The charge collection efficiency was significantly degraded by the electron irradiation. A radiation-induced defect (X) was observed by charge transient spectroscopy. We assign this defect to the electron trap Ei already known in literature and observed by us with DLTS, as its activation energy, 0.50 eV, and annealing behavior, are similar. Moreover, as peaks related to X/Ei disappear after annealing at 250°C and charge collection efficiency also significantly recover after annealing at 250°C, we conclude that this defect is mainly responsible for the decreased charge collection efficiency.

Si-doped, n-type heteroepitaxial layers of Al0.12Ga0.88N grown by metalorganic chemical vapor deposition on SiC substrates were characterized by capacitance transient spectroscopies. Conventional deep level transient spectroscopy (DLTS) reveals the presence of a dominant deep level with an activation energy for electron emission to the conduction band of (0.61±0.02) eV. The activation energy of this deep level displays a pronounced field dependence as determined from double-correlation DLTS (DDLTS), which is indicative of a deep donor level in n-type semiconductors. A deep level is observed by optical-DLTS (O-DLTS) with a threshold energy for electron photoemission to the conduction band of 0.77 eV, which appears to be of identical origin as the dominant deep level detected by DLTS. Two additional deep levels are detected with O-DLTS in the upper half of the band gap of our Al0.12Ga0.88N sample with threshold energies of 0.83 and 1.01 eV.

In conventional Deep Level Transient Spectroscopy (DLTS) measurements, the analysis of the results is based upon the assumption of an exponential current or capacitance transient. We present experimental and computational results on a novel experimental method for determining when the assumption of exponentiality is not satisfied by the sample under study. The measurement may be performed without any changes in the conventional double-boxcar DLTS system.

We have carried out high resolution Laplace deep level transient Spectroscopy (DLTS) and conventional DLTS on silicon implanted with very low doses of either silicon, germanium, erbium, or ytterbium, and compared the results to those from electron-irradiated silicon. DLTS spectra of all the samples initially look very similar, and a peak at 95 K appears in all spectra which may be due to the vacancy-oxygen (VO) defect. We have carried out detailed measurements of the capture cross section and activation energy of this defect using Laplace DLTS. We show that, when the mass of the implanted ion exceeds that of silicon, the defect has a much smaller electron capture cross section than that expected for the VO defect, and a smaller activation energy. Hydrogen has been introduced, either by wet chemical processing or plasma, to all samples to observe the hydrogen–VO interactions resulting in VOH. By using high resolution DLTS we are able to establish that, after hydrogenation, the VOH defect exists with an identical emission rate in the silicon-implanted silicon and the electron-irradiated silicon, but not in the silicon implanted with heavier ions. We conclude that the peak at 95 K in the DLTS spectra in the case of the heavier ions is due to a different defect, confirming earlier reports in the literature. This defect is negatively charged, unlike VO, which is acceptor-like. We are also able to observe VOH in samples where VO is not present, after these samples have been annealed. We attribute this to release of V and H atoms from other defects during annealing.

Recently there has been clear evidence that local alloy disorder splits the DX center in multiple levels. This effect is observed by deep-level transient spectroscopy (DLTS) from different thermal emission rates for the multiple levels in AlxGa1−xAs. We report for the first time the simultaneous measurement of two capture barrier and two ionization entropies for the DX center in Se-doped AlxGa1−xAs. The AlxGa1−xAs was grown by metalorganic chemical vapor deposition at two different alloy compositions (x=0.19 and 0.23). We obtained the capture rates from a DLTS experiment by simultaneously monitoring the two transient signals while changing the filling pulse width. The capture rates show exponential temperature dependence from which the thermal capture barriers are extracted. Together with the emission rate values the ionization entropy is calculated after modifying the appropriate equations for a degenerate semiconductor (ND&amp;gt;1×1017 for AlGaAs). The results are discussed in the context of other published values.

A simple, low-cost, and flexible, automatic, deep-level transient spectroscopy (DLTS) system with an Apple II microcomputer for semiconductor devices analysis is described. By using an interactive Basic software program, all the instrument parameters including sampling aperture locations, temperatures, excitation pulse frequency, width and amplitude can be controlled automatically. The transient signals are also averaged by the software program. Furthermore, a linear square regression method is used to calculate the activation energy. The trap density, capture cross section, activation energy of n-type Sidoped GaAs, grown by molecular beam epitaxy (MBE) system, was measured and compared with conventional boxcar DLTS sytem. In addition, the DLTS signal spectra and the Arrhenius plots can be plotted by a line printer and displayed on a CRT monitor during one thermal scan. An IEEE-488 interface bus was designed for communication between personal computers and instruments. This configuration provides the advantages of ease of operation and rapid set up, especially for the purpose of data acquisition and processing.

Characterization of Si-doped Ga 0.52In 0.48P grown by solid source molecular beam epitaxy using deep level transient spectroscopy

We have studied electrically active defects in buried layers, produced by heavy ion implantation in silicon, using both conventional deep level transient spectroscopy (DLTS) and an isothermal spectroscopic technique called time analyzed transient spectroscopy operated in constant capacitance mode (CC-TATS). We show that CC-TATS is a more reliable method than DLTS for characterization of the heavily damaged buried layers. The major trap produced in the buried layers in p-type Si by MeV Ar+ implantation is found to have an energy level at Ev+0.52 eV. This trap, believed to be responsible for compensation in the damaged layer, shows exponential capture dynamics. We observed an unusually high thermal activation energy for capture, which is attributed to a macroscopic energy barrier for carriers to reach the buried layer. We observe two other majority carrier traps, and also a minority carrier trap possibly due to inversion within the depletion layer.

We report on the capture barrier for the gallium related DX center in Cd0.99Mn0.01Te. In order to determine the barrier height, two methods were applied: an analysis of the persistent photoconductivity decay and the optical deep level transient spectroscopy technique. Over a range of temperatures varying from 77to105K, the capture barrier height, deduced from the decay time constants of photoconductivity, has been found to be equal to 0.22eV. An apparent hole trap, observed with the optical deep level transient spectroscopy, was attributed to the thermally activated capture cross section of a DX center with a 0.23eV capture barrier. The obtained data are close to 0.21eV, the value of the capture barrier which we determined earlier with the help of the deep level transient spectroscopy method.