Near Band Gap Photoluminescence Broadening In n-Gan Films
- PDF / 386,327 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 98 Downloads / 236 Views
655 Mat. Res. Soc. Symp. Proc. Vol. 482 © 1998 Materials Research Society
EXPERIMENTAL METHODS The GaN films were grown by plasma assisted molecular beam epitaxy using methods which were described in detail previously [10]. In this method GaN is formed by the reaction of Ga evaporated from a conventional effusion cell with molecular nitrogen, activated first by an electron cyclotron resonance assisted microwave plasma. The deposition system consists of a Varian Gen II MBE unit with an ASTEX compact ECR source incorporated in one of the effusion cell ports. The films were deposited on (0001) sapphire substrates using a three step process [11]. First the surface of the substrate was converted from A120 3 to AIN by exposing the substrate to an ECR nitrogen plasma at 800 0C. The second step involves the deposition of a low temperature (550 0 GaN-buffer, approximately 300 A thick. The third step involves the deposition of about I to 2 pim thick film at 700-800 °C. The photoluminescence of the investigated GaN films was excited with a 10 mW He-Cd laser (X= 325 nm) and the spectra were collected with a 0.5m spectrophotometer equipped with a holographic grating blazed at 250 nm and detected with a photomultiplier tube. The carrier concentration in the films was determined at room temperature by Hall effect measurements using the Van der Pauw method. The full width at half maximum (FWHM) of the NBG photoluminescence line of the spectra was calculated by linear interpolation of the succesive experimental points bracketing the half of the maximum intensity points. The values calculated this way were in very good agreement with the ones measured by fitting the emission peak with a Voigt function [12]. Nevertheless, we consider the experimental error to be the one calculated by the error propagation of the uncertainty in the wavelength given by the spacing ( A)=0.2 nm) of the adjacent points in the spectrum: 1240 AE = V/AE2 + AE> Vi2.-1240-- A2 = 2.75meV
(1)
EXPERIMENTAL RESULTS AND DISCUSSION Typical photoluminescence spectra, measured at 77K, for three samples whose carrier concentration are 1.0.10 2, 1.5.1019 and 5.3. 10 " cm-3 are shown in figure 1. As seen from these spectra, the main recombination transition occurs at 3.47 eV with no detectable transitions in the yellow part of the spectrum. For the heavily doped samples the near bandgap peak is broadened towards the Stokes part of the line, characteristic of bound electron states. The main difference between these samples is the width of the photolumineseence line, which in this case varies from 18 meV for the lightly doped sample to 99 meV for the heaviest doped sample. Figure 2 shows the full width at half maximum, of the 3.47 eV photoluminescence peak, versus the net carrier concentration ( n=ND-NA ) in a log-log scale. It is apparent from these data that the FWHM for samples with n1018 cm-3 )increase as a function of doping approximately as a power of net carrier concentration. The value of dopants concentration in semiconductors above which the impurity electron
Data Loading...
