Effect of Interface Manipulation for MBE Growth of AlN on 6H-SiC

  • PDF / 107,961 Bytes
  • 6 Pages / 612 x 792 pts (letter) Page_size
  • 38 Downloads / 204 Views

DOWNLOAD

REPORT


Effect of Interface Manipulation for MBE Growth of AlN on 6H-SiC

Koichi Naniwae1,* , Jeff Hartman2, Chris Petrich1, Robert F. Davis2 and Robert J. Nemanich1,2 1 Dept. of Physics, North Carolina State Univ, Raleigh, NC 27695-8202 2 Dept. of Material Science and Engineering, North Carolina State Univ, Raleigh, NC * Present Address: Photonic and Wireless Devices Research Labs., NEC Corporation, 34 Miyukigaoka, Tsukuba, Ibaraki 305-8501, Japan

ABSTRACT AlN layers were grown on 6H-SiC(0001) by molecular beam epitaxy using ammonia as the nitrogen source. Clean (√3x√3)R30° SiC surfaces was prepared by in-situ annealing alone and also by in situ annealing consisted of followed by Si deposition and subsequent annealing. The surface morphology of the AlN films observed by AFM was significantly changed by the nucleation procedure. When the AlN growth was initiated with Al flux exposure on a SiC surface prepared by thermal annealing, the surface roughness of the AlN was significantly reduced. Two-dimensional growth of AlN was observed with reflection high-energy electron diffraction from the very beginning. Atomically flat AlN surfaces with a RMS-roughness of ~0.3 nm were obtained. On the other hand, when film growth was initiated with an ammonia flux exposure on a Si rich SiC surface, a high density of bumps was observed. The bumps seemed to originate from SiNx formation at the heteroepitaxial interface. It was found that control of the Si composition and the V/III ratio at the growth interface is crucial for the AlN film quality.

INTRODUCTION Because single crystal III-nitride substrates are not commercially available, heteroepitaxial film growth is required to obtain large area surfaces for optoelectronic devices based on these materials. The control of the hetero-epitaxial interface between the epitaxial layer and the substrate is one of the key issues to achieve device quality group-III nitride films. Sapphire has been the most commonly used substrate - despite the fact that its lattice constant and thermal expansion coefficient are quite different from those of any of the nitrides. On the other hand, SiC has several advantages over the sapphire as a nitride growth substrate. These include a small lattice mismatch, similar thermal expansion coefficients, a large thermal conductivity, the feasibility of similar cleavage planes, and the availability of a conductive substrate. In spite of these potential advantages, the quality of III-nitride films grown on SiC substrates is not significantly better than

T5.6.1

those grown on sapphire substrates [1-2]. One aspect that may contribute to the poor quality of the films is the properties of the SiC substrate surface. As-received commercial SiC substrates usually contain not only dislocation and micropipes, which originate in the bulk crystal of the substrate but also a high density of ridges and scratches which result from the lapping and polishing process. Obtaining polished surfaces with small roughness is difficult at least in part because of the physical hardness and che