Myung M. Sung, Sung Yong Lee, Chang G. Kim, Sun Sook Lee, Young Kuk Lee, Eun Ju Nah, Chuleui Kim, Sang Heon Lee, and Yunsoo Kim Advanced Materials Division Korea Research Institute of Chemical Technology Yusong P. O. Box 107, Taejon 305-600, Korea Vapor phase epitaxy of magnesium oxide has been accomplished on both Si(001) and Si(111) substrates using the precursors methylmagnesium tert-butoxide and methylmagnesium isopropoxide. The deposition was carried out, mainly with methylmagnesium tert-butoxide, in the temperature range 400-850 C under high vacuum conditions. From x-ray crystallographic analysis, the precursor methylmagnesium tert-butoxide was found to exist as a tetramer in the crystalline state. It had to be heated at 140 C to give a reasonable vapor pressure, whereas the precursor methylmagnesium isopropoxide was heated at 100 C, indicating that the two precursors have the same degree of association in the gas phase. The films were characterized by reflection high energy electron diffraction (RHEED) in situ in the growth chamber, and x-ray diffraction (XRD), x-ray pole figure analysis, scanning electron microscopy (SEM), and x-ray photoelectron spectroscopy (XPS) after the growth. Epitaxial growth of magnesium oxide on Si(111) was achieved by raising the substrate temperature only, however, on Si(001), it was necessary to grow a silicon carbide buffer layer prior to the deposition of MgO. Magnesium oxide is thermodynamically very stable, has a low dielectric constant and a low refractive index, and has been widely used as substrate for growing various thin film materials, particularly oxides of the perovskite structure. There has been a considerable interest in integrating the physical properties of these oxides with semiconductor materials such as GaAs and Si.1-8 In this regard, it is considered very important to be able to grow MgO buffer layers epitaxially on the semiconductors. Various oxide films can then be grown on such buffer layers, eliminating the need for using MgO single crystal substrates. On the other hand, magnesium oxide is used as a protective coating in plasma display panels (PDPs). To be useful for the manufacturing of PDPs, the growth rate must be improved. There have been several reports on chemical vapor deposition (CVD) of MgO using various magnesium compounds such as diethylmagnesium, bis(2,2,6,6-tetramethyl-3,5-heptanedionato)magnesium, bis(acetylacetonato)magnesium, magnesium acetate, magnesium 2-ethylhexanoate, bis(cyclopentadienyl)magnesium, and methylmagnesium tert-butoxide. These compounds must be vaporized at temperatures higher than 100 C, and except for methylmagnesium tert-butoxide, generally need an additional oxygen source such as H2O or O2. In our experiments, stainless-steel ultrahighvacuum (UHV) chambers were employed for the growth of MgO thin films. The precursors were synthesized according to the method described previously.9 The base pressures of the deposition chambers were below 1 x 10-8 Torr, and the chamber pressures increased to 3-5 x 10-6
Torr after the introduction of the precursors.
Thin films of MgO tend to grow preferentially in the [111] direction by CVD of metal organic compounds, and therefore its growth in that direction is anticipated on Si(111). The MgO films grown at low temperatures using methylmagnesium tert-butoxide are polycrystalline, and epitaxy on this substrate could only be achieved around 800 C. XRD and pole figure analysis of the films clearly revealed their epitaxial nature. X-ray photoelectron spectra of the films indicated formation of stoichiometric films and were almost identical with those of MgO single crystals. SEM plan-view images of the films showed flat, featureless surfaces. To elucidate the pyrolysis mechanism of the precursor methylmagnesium tertbutoxide, it was adsorbed onto a Si substrate and flash evaporated toward a quadrupole mass analyzer attached to the UHV chamber. The mass spectrum indicated that the precursor is pyrolyzed by -hydrogen elimination between the methyl group and the tert-butyl group. On clean Si(001) substrates on the other hand, all efforts to grow epitaxial MgO films have failed. The lattice mismatch of MgO with respect to Si is -22.4 %. Since magnesium oxide has a lattice mismatch of -3.3 % with respect to cubic silicon carbide and the latter grows epitaxially on Si(001) relatively easily, it was decided to grow a silicon carbide buffer layer on Si(001) prior to the deposition of MgO. SiC buffer layers were deposited on clean Si(001) substrates using 1,3-disilabutane as the source compound. After the formation of a SiC buffer layer was confirmed by RHEED, a magnesium oxide film was deposited on it. A corresponding RHEED pattern indicated epitaxial growth of MgO, although a certain degree of twinning was evident judging from the appearance of extra diffraction spots. This twinning phenomenon became much less pronounced when 4-off Si(001) substrates were used. XRD, rocking curve measurement, and pole figure analysis also indicated that the films were epitaxially grown in the [100] direction at the growth temperatures of 700-850 C. When the growth temperature was lowered, below 600 C for example, there always appeared ring structures in the RHEED patterns indicating polycrystalline growth of the films. References 1. D. K. Fork, F. A. Ponce, J. C. Tramontana, and T. H. Geballe, Appl. Phys. Lett. 58, 2294 (1991). 2. K. Saiki, K. Nishita, and A. Koma, Jpn. J. Appl. Phys. 37, L1427 (1998). 3. J.-G. Yoon and K. Kim, Appl. Phys. Lett. 68, 2523 (1996). 4. L. S. Hung, J. A. Agostinelli, J. M. Mir, and L. R. Zheng, Appl. Phys. Lett. 62, 3071 (1993). 5. D. K. Fork and G. B. Anderson, Appl. Phys. Lett. 63, 1029 (1993). 6. L. S. Hung, L. R. Zheng, and T. N. Blanton, Appl. Phys. Lett. 60, 3129 (1992). 7. K. Nashimoto, D. K. Fork, and T. H. Geballe, Appl. Phys. Lett. 60, 1199 (1992). 8. Y. Kaneko, N. Mikorshibar, and T. Yamashito, Jpn. J. Appl. Phys. 30, 1091 (1991). 9. W. Koh, S.-J. Ku, and Y. Kim, Bull. Korean Chem. Soc. 19, 281 (1998).