Hostname: page-component-7b7567c8c-wxvwp Total loading time: 0 Render date: 2026-11-17T06:09:49.841Z Has data issue: false hasContentIssue false

Fermi Level Pinning at GaN-interfaces: Correlation of electrical admittance and transient spectroscopy

Published online by Cambridge University Press:  13 June 2014

H. Witte
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
A. Krtschil
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
M. Lisker
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
D. Rudloff
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
J. Christen
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
A. Krost
Affiliation:
Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, D-39016 Magdeburg, Germany
M. Stutzmann
Affiliation:
Walter Schottky Institute, Technical University Munich, D-85748 Garching, Germany
F. Scholz
Affiliation:
4th Institute of Physics, University Stuttgart, D-70550 Stuttgart, Germany

Abstract

In GaN layers grown by molecular beam epitaxy as well as metal organic vapor phase epitaxy significant differences were found in the appearance of deep defects detected by thermal admittance spectroscopy as compared for deep level transient spectroscopy measurements. While, thermal admittance spectroscopy measurements which were made under zero bias conditions only show thermal emissions at activation energies between 130 and 170 meV, further deep levels existing in these GaN layers were evidenced by transient spectrocopy. This discrepancy is explained by a pinning effect of the Fermi level at the metal / GaN interface induced by high a concentration of the deep levels showing up in thermal admittance spectroscopy. We compare our results with a GaAs:Te Schottky- diode as a refernec sample. Here, both spectroscopic methods give exactly the same deep level emissions.

Information

Type
Research Article
Copyright
Copyright © 1996 Materials Research Society
Figure 0

Fig. 1:(Left) Admittance spectra of two MOVPE-grown GaN samples at different bias voltages. (Right) Histogram of the GaN surface morphology measured by atomic force microscopy for both samples #1 and #2.

Figure 1

Fig. 2:(Left) TAS spectra of different GaN samples grown by MOVPE (#1 and #2) and by MBE (#3 and #4), respectively, at a modulation frequency of 10kHz and under zero bias condition. The resulting activation energies are 130-140meV for T1 and 150-170meV for T2, respectively. (Right) DLTS-spectra of the MBE samples #3 and #4 at 100kHz (LF-DLTS) and at 1MHz (HF-DLTS). The corresponding emission rates are en = 36.1 s−1 for the HF-DLTS spectrum and en = 0.4 s−1 for the LF-DLTS, respectively.

Figure 2

Fig. 4:(Left) Arrhenius plots of deep defects in n-type GaN layers obtained by HF-DLTS, LF-DLTS and TAS. (Right) Arrhenius plots of three deep defect levels obtained by the spectroscopical methods DLTS, TAS and low frequency DLTS in a GaAs:Te sample with an electron concentration of 2×1015 cm−3. The activation energies of the deep traps are: G1: 350±30 meV; G2: 670±40 meV and G3: 740±50 meV, respectively.