The role of antiphase domain boundaries in Si epitaxy by ultrahigh vacuum chemical vapor deposition from SiH 4 or SiH 2 Cl 2 on Si(100)-(2×1)

Abstract

The importance of antiphase domain boundaries for epitaxial multilayer growth during ultrahigh vacuum chemical vapor deposition of Si on Si(100)-(2 × 1) from SiH 4 and SiH 2 Cl 2 at intermediate temperatures is illustrated. Under these conditions multilayer growth is governed by heterogeneous nucleation of dimer strings at antiphase domain boundaries of the underlying layer which represent deep potential minima for diffusing species. This is in contrast to the formation of the first epitaxial layer, which nucleates homogeneously on the flat substrate terraces.If silicon chemical vapor deposition (CVD) on Si(100)-(2 × 1) is carried out in an intermediate temperature and pressure regime, epitaxy proceeds by island nucleation and growth [1,2]. The evolution of the surface topography in this intermediate regime depends on the extent of interlayer transport. If interlayer transport is very unfavorable, the roughness increases with deposition time [3]. In the case of perfect layer-by-layer growth each layer is completely filled by deposited adatoms before nucleation of the next layer starts and the surface roughness is oscillating with maximum amplitude [3]. Experimentally, intermediate behavior is frequently observed during growth under MBE as well as under CVD conditions, where interlayer transport is possible to a certain extent. This results in nucleation of new layers before the underlying layers are completely filled [1,2,4]. Under these conditions reflectance high-energy electron diffraction (RHEED) intensity oscillations of reduced amplitude can be observed if the rate of interlayer transport is sufficiently high.For a more detailed picture of the atomic-scale growth process and the resulting mesoscopic structures it is necessary to follow these processes on a microscopic scale. Scanning tunneling microscopy (STM) studies have shown that island nucleation in SiH 4 -based ultrahigh vacuum CVD (UHV-CVD) on Si(100)-(2 × 1) depends very sensitively on

The role of antiphase boundaries during ion sputtering and solid phase epitaxy of Si(0 0 1)

“…After multilayer erosion by 1.5 keV Ar ions as shown in Figs. 1(c) and 1(d), both types of antiphase boundaries are still observed and the Si(001) surface morphology resembles that observed during multilayer epitaxial growth of Si by UHV CVD [9]. This resemblance suggests that AP2 antiphase boundaries play an important role in surface roughening during ion sputtering by offering preferential nucleation sites of adatom islands.…”

“…This asymmetry in attachment kinetics of adatoms at step edges is due to the Ehrlich-Schwoebel barrier [1][2][3][4], and destabilizes the surface morphology toward the formation of mounds during multilayer epitaxial growth [5,6]. Since adatom islands nucleate independently, coalescence of two adatom islands on dimerized Si(001) [7][8][9][10][11][12] and Ge(001) [11,12] surfaces may result in the formation of antiphase boundaries. An AP1 antiphase boundary forms when two adatom islands meet at S A steps, and an AP2 antiphase boundary forms when two adatom islands meet at S B steps.…”

“…An AP1 antiphase boundary forms when two adatom islands meet at S A steps, and an AP2 antiphase boundary forms when two adatom islands meet at S B steps. Only AP2 antiphase boundaries act as preferential nucleation sites of new adatom islands during subsequent epitaxial growth [7][8][9][10][11][12].…”

“…Much like during epitaxial growth, antiphase boundaries form on Si(001) surfaces during ion sputtering [15] and annealing after submonolayer ion sputtering [15,16]. However, these experiments revealed only one type of antiphase boundary, the AP2 antiphase boundary, in contrast to two formed during epitaxial growth [7][8][9][10][11][12], and the role of antiphase boundaries in roughening of crystalline semiconductor surfaces has not been investigated.…”

“…Antiphase boundaries form when two independently nucleated vacancy islands with different dimer row registry coalesce. An AP1 antiphase boundary forms when two vacancy islands meet with two S B steps between them, much like an AP1 antiphase boundary forms when two adatom islands meet at S A steps during epitaxial growth [7][8][9]; dimer rows are parallel to the S A steps and perpendicular to the S B steps [22]. An AP2 antiphase boundary forms when two vacancy islands meet with two S A steps between them; the islands decorating AP2 antiphase boundaries have two S A steps across the AP2 antiphase boundaries, see Fig.…”