Fermi Level Pinning at GaN-interfaces: Correlation of electrical admittance and transient spectroscopy
- PDF / 63,862 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 61 Downloads / 201 Views
1
F99W11.82
bath and etched with HF and HCL. The ohmic contacts were formed by evaporating a 100nm thick Al layer followed by an annealing step at 500oC in nitrogen athmosphere. A 100nm thick Pt layer deposited by magnetron sputtering was commonly used as Schottky contact. Both contacts were arranged on the GaN layer and have a diameter of 750µm. The distance between the metal layers was 1mm. We carried out C-V measurements, admittance spectroscopy (AS) and thermal admittance spectroscopy in the frequency range between 20Hz and 1MHz and at temperatures between 80K and 400K using a high precision LCR meter HP4284A and a liquid-nitrogen flow cryostat. For both, AS and TAS measurements the capacitance and the conductance were measured as a function of frequency and temperature (fore more details see /7/). If the resistance of the neutral bulk region of the samples is not negligible a serial circuit model of the complex impedance was used for analysis (see /6/). In addition, we performed DLTS measurements at a frequency of 1MHz using the capacitance meter Boonton 7200 and a pulse generator HP8110A. The capacitance transients were recorded via a fast A/D converter with a time resolution of 1µs and analyzed by the boxcar method. DLTS measurements at lower frequencies were made at Schottky contacts which break down at frequencies below 1MHz. In this case, the transients were measured using the LCR meter HP4284A with a minimal time step of 300ms in the boxcare method. Results and Discussion The Pt Schottky contacts on GaN layers show barrier heights up to 0.95eV determined by C-V-spectroscopy at frequencies between 50kHz and 1MHz. This barrier is slightly lower than the 1.04eV found in /8/. The net donor concentrations were in the range of 5x1016 to 5x1017 cm-3 determined by C-V-characteristics and by Hall effect measurements at room temperature. In generally, we can distinguish two different categories of Schottky contacts as represented by the samples #1 and #2 in Fig. 1. Sample #1 shows only a little variation in the conductance spectrum for different bias voltages which is caused by a small Schottky barrier. Furthermore, the contact breaks down at relatively low frequencies of about 5x104 s-1 resulting in a completely bias independent conductivity and capacitance above this frequency. The other group shows larger barrier heights resulting in a stronger influence of the bias on the conductance (sample #2 in Fig.1). The rectifying behavior of this contact vanished at higher frequencies of about 106 s-1. These different properties of the contacts were correlated with the surface roughness measured by atomic force microscopy (AFM),which is shown on the right side of Fig.1. The significant difference between both samples becomes visible. Sample #2 exhibits a sharp distribution of low height spikes equivalent to a nearly perfect flat surface. In contrast, the AFM measurements of sample #1 revealed a rough surface with a broad distribution of heights. Obviously Schottky contacts on a smooth surface exhibit better rectifying
Data Loading...
