EPITAXIAL-GROWTH AND PHOTOLUMINESCENCE OF SI PURE-GE/SI QUANTUM STRUCTURES ON SI(311) SUBSTRATES/

Citation
K. Amano et al., EPITAXIAL-GROWTH AND PHOTOLUMINESCENCE OF SI PURE-GE/SI QUANTUM STRUCTURES ON SI(311) SUBSTRATES/, Semiconductor science and technology (Print), 13(11), 1998, pp. 1277-1283
Citations number
16
Categorie Soggetti
Engineering, Eletrical & Electronic","Physics, Condensed Matter","Material Science
ISSN journal
02681242
Volume
13
Issue
11
Year of publication
1998
Pages
1277 - 1283
Database
ISI
SICI code
0268-1242(1998)13:11<1277:EAPOSP>2.0.ZU;2-U
Abstract
Si/pure-Ge/Si quantum structures were grown on Si(311) substrates by g as-source molecular beam epitaxy at various growth temperatures (T-g) and their optical properties were investigated by photoluminescence (P L) spectroscopy. At T-g higher than 740 degrees C, uniform steps which align towards the [<(2)over bar 33>] direction were found to have for med after growing a 2000 Angstrom Si buffer layer. A series of Si/pure -Ge/Si layers on the surface steps with various Ge coverage (Q) exhibi ted intense PL and showed systematic energy shift due to the quantum c onfinement effect even with Q less than one monolayer (ML), which indi cates the existence of the lateral confinement effect. With increasing incident laser power, the PL energy of samples with Q larger than 1 M L shifted to higher energies as can usually be seen for indirect quant um wells (QWs). On the other hand, below 1 ML, PL peaks did not show a ny blue shift with increasing laser power. This can be explained by a bunched density of states at the band edge in quantum wires (QWRs) whi ch were realized by preferential growth of Ge atoms at the step edges. At T-g of 700 degrees C, Ge QWRs were not realized even with Q less t han 1 ML as evidenced by a spectral blue-shift with increasing excitat ion power. This is presumably due to the insufficient surface migratio n length and the resultant incorporation of Ge atoms at the terraces, leading to the formation of SiGe alloy QWs with 1 ML thickness. For Q larger than 3.0 ML, the growth mode changed to being three-dimensional and isosceles-triangle-shaped Ge islands were formed. However, PL mai nly came from the wetting layer since the sizes of the islands were to o large and the crystal quality was probably poor.