MBE Growth of Epitaxial Calcium Fluoride on Silicon

  • PDF / 1,718,031 Bytes
  • 10 Pages / 417.6 x 639 pts Page_size
  • 0 Downloads / 214 Views

DOWNLOAD

REPORT


INTROIXJCTION

Molecular beam epitaxy (MBE) has been used successfully to explore heteroepitaxial systems because of its low growth temperature, molecular control of interfaces, and ease in depositing new materials. In this paper, the use of KBE for heteroepitaxial growth of insulators on semiconductors is discussed. In particular, the growth of CaF 2 on Si and its overgrowth with Si and Ge is described. Epitaxial insulators have a number 1 of potential applications in the semiconductor industry. These include: semiconductor-on-insulator (SOI) structures, including use as a lattice matching material to go from Si substrates to semiconductors desired for optoelectronic, microwave, or very high speed applications, three-dimensional (3-D) device structures; and improved gate insulators. The advantages of an an epitaxial insulator for these application is detailed below. The ability to use the epitaxial insulators as an intermediary to go from a Si substrate to another semiconductor such as GaAs would allow the semiconductor industry to take advantage of the developed VLSI technology in Si while creating optoelectronic, microwave, or extremely high speed interface devices on the same chip. In addition, Si substrates are much cheaper than any other semiconductor and generally are substantially more robust than most of the III-V and II-VI semiconductors. An epitaxial insulator not only provides a means for lattice matching two different semiconductors but, in addition, provides a means to electrically isolate devices in the two semiconductors. At the present time, all VLSI circuits must be laid out in twodimensions (2-D). This obviously limits the density of active devices to much lower levels than would be possible if three-dimensional circuits could be laid out with comparable line widths. Perhaps a more important advantage of 3-D ICs is the increase of circuit complexity that may aid in the development of advanced technologies such as artificial intelligence and supercomputers. In addition, 3-D ICs will allow a greater separation of interconnect lines, which could reduce capacitive coupling between rapidlv switching lines.

Mat. Res. Soc. Symp. Proc. Vol. 54. , 1986 Materials Research Society

286

Epitaxial CaF 2 may be also desirable for use as an improved gate insulator. Presently, MOS technology depends on thin (-25 nm) oxide layers to However, the yield from isolate the gate electrode from the semiconductor. oxidation processes pushed below -15 nm of oxide thickness (which will be required as VLSI starts to be scaled below I pm dimensions) is expected to be poor due to pinholes and other defects in the oxide. In addition, it is difficult to completely eliminate some roughness (on an atomic scale) at the Si/SiO interface. This roughness results in a degradation of the mobility o? carriers in the device, which is expected to become more severe for thinner oxides. An epitaxial insulator, which can be grown by a twodimensional growth process which proceeds layer by layer such as MBE, will be atomically smooth. In ad