Epitaxial growth of MgO.Al2O3 on Si(100) by MBE
Moldovan, Antoniu; Norga, Gerd J.
Italy

Epitaxial oxides on Si are interesting for integration of oxide/silicon/oxide resonant tunnel diode (RTD) with Si CMOS. To achieve high quality re-epitaxy of Si on the epi-oxide, thermal stability of the Si/epi-oxide stack is an important criterion. Several recent reports suggest that the Si/ gamma-Al2O3 hetero-epitaxy system has excellent potential for CMOS-compatible RTDs because of its intrinsic thermodynamic stability [1, 2].
The crystal structure of gamma-Al2O3 is usually considered to be a cubic spinel, ascribed to Fd-3m space group symmetry (a=7.89Ǻ), with tetrahedral and octahedral sites for aluminium atoms. Gamma-Al2O3/Si(100) represents a interesting example of pseudomorphic hetero-epitaxy: the epitaxial constraint imposed by the Si substrate forces alumina to adopt the cubic Gamma-Al2O3 structure, which in bulk form is a metastable polymorph of the thermodynamically preferred phase, hexagonal alpha-Al2O3 (sapphire). A considerable disadvantage of gamma-alumina is the relatively large lattice mismatch (2.7 %) with Si and its inherent tendency to transform to the thermodynamically stable hexagonal alpha-Al2O3 phase (corundum) as the layer thickness is increase above a few nm.
In this paper we report for the first time the alloying of hetero-epitaxial gamma-alumina layers on Si(100) with MgO, to obtain stoichiometric spinel (MgO. Al2O3). In fact, alloying with MgO has a twofold beneficial effect: (1) reduction of the lattice mismatch with Si from 2.7 % for gamma-Al2O3 to 0.7 % for MgO.Al2O3; (2) stabilization of the cubic spinel phase over the gamma-Al2O3 phase. Furthermore, MgO, like Al2O3, features excellent thermodynamic stability in contact with Si, and in addition MgO.Al2O3 is also an excellent insulator. We present an adaptation of the oxygen exchange process for gamma-Al2O3/Si(111) of Jung et al. [3] to the MgO. Al2O3/Si(100) system, along with structural (AFM, RHEED, TEM) and electrical (CV, IV) characterization of the layers.
[1] M. Ishida, H. Hori, F. Kondo, D. Akai, K. Sawada, Jpn. J. Appl. Phys. 39, 2078 (2000).
[2] T. Kimura, M. Ishida, Jpn. J. Appl. Phys. 38, 853 (2000).
[3] Y. C. Jung, H. Miura , M. Ishida, Jpn. J. Appl. Phys. 38, 2333 (1999).
back