We propose an analysis method for the accurate estimation of the hole trap (H1, EV + 0.85 eV) concentration in n-type GaN via minority carrier transient spectroscopy (MCTS). The proposed method considers both the hole occupation during a filling (current injection) period and the quick carrier recombination via the hole traps near the depletion layer edge immediately after a reverse bias is applied. The reverse bias voltage dependence of the MCTS spectrum indicates that an accurate trap concentration, as well as the hole diffusion length and electron capture cross section of the hole trap, can be determined.

We studied deep levels in quartz-free hydride-vapor-phase epitaxy (QF-HVPE)-grown homoepitaxial n-type GaN layers within which three electron and eight hole traps were detected. The dominant electron and hole traps observed in the QF-HVPE-grown GaN layers were E3 (EC − 0.60 eV) and H1 (EV + 0.87 eV), respectively. We found that the E3 trap density of QF-HVPE-grown GaN (∼1014 cm−3) was comparable with that of MOVPE-grown GaN layers, whereas the H1 trap density of QF-HVPE-grown GaN (∼1014 cm−3) was much smaller than that of an MOVPE-grown GaN layer with a low-residual-carbon growth condition. A detailed analysis of the QF-HVPE-grown GaN layers revealed that the H1 trap density is almost equal to the carbon impurity concentration and other impurities that compensate the Si donors besides the carbon impurity were hardly detected in the QF-HVPE-grown GaN layers.

No abstract

Electron traps generated during the reactive ion etching (RIE) process in n-type 4H-SiC are investigated using the deep-level transient spectroscopy technique and isothermal capacitance transient spectroscopy (ICTS) technique. Two electron traps of the Z1/2 center (EC−0.64eV) and the EH3 center (EC−0.74eV) are detected in the RIE-etched sample by ICTS measurement at 300 K. A method is proposed to determine the depth profiles of the electron traps that are localized near the etched surface, whereby a depth profile is extracted from the dependence of averaged trap density on the depletion layer width. An exponential distribution is assumed as the depth profile of the electron traps generated during the RIE process. The extracted depth profile was confirmed to be consistent with that determined by the double-correlation method. An appropriate function for the depth profile of carrier traps is assumed and the dependence of the averaged trap density on the depletion layer width is analyzed, which enables the extraction of a depth profile that has both higher depth resolution and higher resolution in the carrier trap density with the proposed method than that with the double-correlation method.

Gamma-ray irradiations of up to 500 kGy on homoepitaxial n-type GaN layers were carried out, and the formation of electron traps was investigated by deep-level transient spectroscopy (DLTS) using Ni Schottky barrier diodes (SBDs). Before performing DLTS, current–voltage (I–V) and capacitance–voltage (C–V) measurements of the SBDs were performed and it was found that there was no change in the net donor concentration, ideality factor, and Schottky barrier height after irradiation. In the DLTS measurements, two new peaks, labeled G1 and G2, were observed after irradiation. The filling pulse width dependence of G1 revealed that the peak consists of two electron trap levels, labeled G1a (EC − 0.13 eV) and G1b (EC − 0.14 eV). Isothermal capacitance transient spectroscopy measurements of samples with different Schottky barrier heights showed that the G2 peak is a complex peak consisting of at least three electron traps, labeled G2a (EC − 0.80 eV), G2b (EC − 0.98 eV), and G2c (EC − 1.08 eV). The production rates (formation rates of traps by gamma-ray irradiation) for each trap were obtained. Finally, we investigated the annealing behavior of each trap and found that G1b and G2b decreased by the same amount with increasing annealing temperature, suggesting that the behavior originates from a recombination of vacancy–interstitial (Frenkel) pairs.