TEM Study of Mg-Doped Bulk GaN Crystals
- PDF / 2,732,768 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 28 Downloads / 248 Views
INTRODUCTION Progress in GaN technology has been delayed for a long time because of difficulties in obtaining p-doping. Only recently success of Amano et al [1] followed by Nakamura [2] resulted in p-doping of GaN. GaN p-n junction blue light emitting diodes (LEDs) and lasers have been obtained [3]. Despite this success Mg p-doping is still not fully understood. Originally low-energy-electron-beam irradiation (LEEBI) [1] and recently thermal annealing have been found to activate Mg and result in p-doping [2,3]. There are reports that Mg is not distributed uniformly in the layer and has a high tendency to diffuse to the layer surface [4]. EXPERIMENTAL In this paper structural studies of Mg doped bulk GaN crystals will be described. The GaN crystals have been grown by the High Nitrogen Pressure Solution Method [5] from a solution of liquid gallium containing 0.1-0.5 at.% of Mg [6]. Before the growth conditions are established,
363 Mat. Res. Soc. Symp. Proc. Vol. 572 ©1999 Materials Research Society
the solution was homogenized at a temperature of 1100°C for several hours. Then the temperature was increased up to the growth conditions (T = 1500 - 1600'C) and N2 pressure about 15 kbars was applied. Typical time of growth was in the range of 100 - 150 hours. Crystals in the form of hexagonal platelets (similar to those for undoped crystals) reaching dimensions up to 8 mm were found in the cooler part of the crucible. This shows that the Mg impurity does not influence the relative growth rates to a significant degree and therefore the form of the hexagonal platelets is similar to that for undoped crystals. However, it was observed that the platelets with Mg are generally somewhat thicker suggesting that Mg impurity does accelerate the growth in the c-direction. Observation with the Nomarsky microscope shows that as in undoped samples one surface normal to the c-axis is more rough, with many surface steps while the opposite surface is always mirror like. Convergent beam electron diffraction (CBED) was applied to study polarity of these crystals [7]. Earlier studies of undoped GaN crystals showed that the polarity of the crystals can also be recognized by using mechanochemical polishing with aqueous solution of KOH [8, 9]. Cross-section samples were prepared from crystals directly removed from the crucible and also from those etched chemically. Electron microscopy was applied in order to detect the presence of any structural defects, observation of surface morphology and determination of crystal polarity.
Sim-.6
•
N
Fig. 1: Cross-section TEM micrograph of the rough side of an as-grown GaN:Mg platelet sample. The inset shows the experimental and calculated CBED patterns together with a hardball model showing atom arrangement along the c-axis. Fig. I shows a cross-section electron micrograph from a crystal directly removed from the crucible. It can be seen that one side of the as grown crystal has a rough surface (saw-like) with an amplitude of 10 nrm. The distance between the spikes was about 7-10 nm. The opposite si
Data Loading...
