Papers

AVS1997 Session SS3+NS-TuA: Step Dynamics and Film Growth on Silicon

Tuesday, October 21, 1997 2:00 PM in Room A1/2-A

Tuesday Afternoon

Start	Invited?	Item
2:00 PM		SS3+NS-TuA-1 Kinetics of Step Attachment in Growth Studied by LEEM M.S. Altman (Hong Kong University of Science & Technology) The attachment of atoms to steps at the edges of islands and terraces is a fundamental part of the growth process on surfaces. The kinetics of step attachment may therefore play an important role in determining the growth behaviour. For example, a number of distinct growth instabilities have been predicted to occur if step attachment is asymmetric with respect to atoms approaching from the terraces above and below a step, although this causal relationship has not been explicitly confirmed by experiment. Step passivation has also recently been proposed as an alternative to altered terrace diffusion to explain the function of surfactants in growth. In this work, we have used low energy electron microscopy (LEEM) to determine step attachment kinetics and their relation to growth instabilities and mechanisms of surfactant action. A step bunching instability which was observed on the Si(111) (7x7) surface is shown to be driven by asymmetric step attachment kinetics where step motion occurs predominantly as a result of attachment of adatoms from the terrace trailing an advancing step. Attributing this asymmetry entirely to a step edge diffusion barrier yields an effective barrier of -15 meV, although asymmetry in the diffusion prefactor and the attachment probability after arriving at the step edge site cannot be ruled out. Although asymmetry was found to be isotropic, the attachment probability does depend significantly upon the step orientation. The influence of Sb and In surfactants upon the step attachment asymmetry and the attachment probability was also explored.
2:40 PM		SS3+NS-TuA-3 Step Faceting at the (001) Surface of B-doped Si J.B. Hannon, N.C. Bartelt, B.S. Swartzentruber, J.C. Hamilton, G.L. Kellogg (Sandia National Laboratories) The arrangement of steps on heavily B-doped Si(001) surfaces exhibits a dramatic temperature dependence. When the temperature is lowered below 980 C, the nominally straight S_B step restructures, forming a chain of triangular facets with nearly-uniform apex angle. The apex angle decreases with decreasing temperature, until a densely packed striped phase is formed at 850 C. We have used low-energy electron microscopy (LEEM) to quantify the step energetics and morphology as a function of temperature. We find that a simple, systematic temperature dependence of the spatial anisotropy of the free energy of isolated steps is sufficient to account for much of the observed step restructuring. From quantitative analysis of step fluctuations, we find a systematic decrease in the ratio of the S_A to S_B step free energy as the temperature is lowered to the triangle transition temperature. The decrease in the ratio is confirmed by examination of the temperature dependence of the equilibrium shape of islands, suggesting that the triangles form when the relative S_B step free energy becomes so large that S_B steps are no longer thermodynamically stable. This idea is supported by the fact that the temperature dependence of the apex angle of the triangles follows an extrapolation, using the Wulff construction, of the temperature dependence of the step free energies at higher temperatures. Our picture of the step edge faceting suggests that the proliferation of straight S_A step edges at lower temperature is caused by the free energy of isolated S_A steps becoming very small.
3:00 PM		SS3+NS-TuA-4 Hot STM Studies of Heavily Boron-Doped Si(001). M. Krueger, B. Borovsky (University of Minnesota); D.E. Jones, J.P. Pelz (Ohio State University); E. Ganz (University of Minnesota) We use hot scanning tunneling microscopy (STM) to study the temperature-dependent surface morphology of heavily boron-doped Si(001). The transition from the familiar stepped surface to triangular tiles and then to narrow fingers is driven by the changing surface concentration of boron, a quantity that can be measured in situ by hot STM. Hot STM movies offer an atomic-scale view of step edges during annealing and allow direct observation of the diffusion of boron at the surface. We will discuss the effect of surface boron on Si(001) step edge morphology and dynamics at temperatures up to 900 K.
3:20 PM		SS3+NS-TuA-5 Atomic Configuration of Segregated B on Si(001) Surface T. Komeda, Y. Nishioka (Texas Instruments, Tsukuba R&D Center, Japan) Recently, it has been reported that monoatomic steps are arranged in a comb-shape structure on highly B-doped,¹ and highly P-doped² Si(001) surface. The characteristic stripe step structures are originally predicted by Alerhand et al. on the clean Si(001) surface, and its appearance only on a highly doped sample may be due to the enhancement of the anisotropic stress by the dopant segregation. However, compared to Si(111) surface, the atomic scale bonding configuration of the segregated dopants on the Si(001) and consequently the origin of the stress enhancement is not well understood so far. In this paper, the bonding configuration of segregated B dopants on Si(001) surface is investigated with UHV-STM, STS and first-principle calculation. Experimentally, highly B doped Si(001) wafer (0.001 ω cm) is used for the sample. In the occupied state image, a large number of depleted dimers appear. However, they are depleted only in the occupied state image and not in the unoccupied state image. The depleted dimer can be assigned as a B dimer on the top surface which replaces a Si dimer. It is considered that the lack of dangling bonds in the B dimer and the difference of electronegativity between B and Si is responsible for the electric depletion. The expected short bond length of the B dimer on the top surface can be responsible for the increase of the anisotropic stress and can enhance the appearance of the comb-shape step structure. On the other hand, the unoccupied state image contains characteristic paired protrusions whose density is ~5% of the Si atoms on the top surface. The paired bright spots, separated by 2a (a=0.38 nm), are located on neighboring two troughs of dimer rows and in the node position along dimer rows. They can be clearly distinguished from the well-known c-type defects. The most plausible model is that a B atom is bonded to two neighboring dimers in a dimer row. The appearance of the bright spots are well reproduced by the first-principle calculation. ¹D.E. Jones, C. Ebner, J.P. Pelz, Y. Hong, E.G. Bauer, and I.S.T. Tsong, Phys. Rev. Lett. 77 (1996) 330. ²T. Komeda and Y. Nishioka, Appl. Surf. Sci. to be appeared
3:40 PM		SS3+NS-TuA-6 Twinned Epitaxial Layers Formed on Si(111)√3x√3-B H. Hibino, K. Sumitomo, T. Ogino (NTT Basic Research Laboratories, Japan) B-induced √3x√3 reconstruction plays a special role in the formation of the Si/Si(111) interface. Formation of twinned epitaxial layers on Si(111)√3x√3-B is especially interesting because it may enable us to form novel Si structures. Hence, we investigated in detail the growth process of twinned epitaxial layers and their thermal stability. In the initial stages of Si MBE growth on Si(111)√3x√3-B, Si islands one or two bilayers (BL) high are formed. The ratio of the number of Si atoms contained in the 2BL islands to the total number of deposited Si atoms depends on the surface B concentration. The higher the surface B concentration, the larger the ratio. Furthermore, the 2BL islands are twinned with the substrate, but the BL islands are not twinned. We conclusively demonstrated a growth mode of twinned 2BL-high layer-by-layer growth and a growth mode transition between twinned 2BL islands and untwinned BL islands that depends on the surface B concentration. After the coalescence of the twinned 2BL islands, the domain boundaries with the density much higher than that on the substrate remain on the first 2BL. BL islands, rather than 2BL islands, are formed at the domain boundaries of the √3x√3 reconstruction. The grown layers are totally twinned because the BL islands follow the stacking sequences of the twinned 2BL. If we want to grow epitaxial layers twinned with the already grown twinned layers, post-growth anneal is necessary to increase the surface B concentration and to reduce the domain boundary density. Our investigation of the thermal stability of the twinned epitaxial layers indicated that the annealing temperature at which twinned layers are transformed into untwinned layers increases as the thickness increases. It is possible to form superlattices of twinned layers (polytypes) by precisely controlling the parameters of the growth and post-growth anneal.
4:00 PM		SS3+NS-TuA-7 Influence of Hydrogen on Initial Si /Al Thin Film Structure and Morphology D.P. Adams, T.M. Mayer, B.S. Swartzentruber (Sandia National Laboratories) We investigate the coverage-dependent effects of surface hydrogen on Si/Al thin film structure and morphology. Al films are deposited by PVD at temperatures below 250C onto Si(100) substrates having different H coverages. From scanning tunneling microscopy (STM) we find that small amounts of monohydride (0.15 monolayer) significantly affect Al island structure. For small coverages hydrogen causes the Al island density to increase, possibly by blocking metal atom diffusion on Si. STM also demonstrates that Al grown onto fully-passivated Si:H surfaces exhibits a change in growth mode compared with growth onto clean Si. Deposition of Al onto a monohydride-terminated, θ=1, Si surface results in Volmer-Weber growth (i.e., three-dimensional islanding), and Al grown on clean Si forms with a Stranski-Krastanow growth mode. A change in morphology is observed on Si:H versus Si substrates even though the out-of-plane crystallographic texture of 300Å-thick films is identical <110>. Post-growth nuclear reaction ion-beam analysis and secondary ion mass spectrometry reveal a hydrogen - free interface with no accumulation in thick Al films. This work was supported by the United States Department of Energy under Contract DE-AC04-94AL85000. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy.
4:20 PM		SS3+NS-TuA-8 Existence of Magic/Critical Thicknesses for Ultrathin Al Overlayers on Si(001) J.-H. Cho (Oak Ridge National Laboratory); Q. Niu, C.K. Shih (University of Texas, Austin); Z. Zhang (Oak Ridge National Laboratory) We investigate the stability of the Al thin films on Si(001) surface using first-principles total energy calculations. The calculated total energy with respect to the film thickness shows the existence of magic and critical thicknesses for smooth growth of Al on the Si(001) substrate. Especially the surface relaxation and work function are found to correlate strongly with the magic thickness. The present results confirm our earlier semi-quantitative analysis for the "electronic growth" mechanism in which metal overlayers on semiconductor substrates can be stabilized at some magic/critcal thicknesses due to the interplay between the effects of quantum confinement, charge spilling, and interface-induced Friedel osillations. Comparisons with experimental observations are made[1]. [1] ORNL is managed by Lockheed Martin Energy Research Corp. under U.S. Department of Energy contact DE-AC05-96OR22464
4:40 PM		SS3+NS-TuA-9 In-situ Observation of Gas-Source Molecular Beam Epitaxy of Silicon and Germanium on Si(001) I. Goldfarb (University of Oxford, United Kingdom); J.H.G. Owen (University of California, Santa Barbara); D.R. Bowler (University of Keele, United Kingdom); C.M. Goringe (University of Sydney, Australia); P.T. Hayden (University of Oxford, United Kingdom); K. Miki (Electrotechnical Laboratory, Japan); D.G. Pettifor, G.A.D. Briggs (University of Oxford, United Kingdom) The two dominant variables in the growth of silicon and germanium from gas sources are substrate temperature and gas flux. By using an elevated temperature UHV STM we have observed the development of the surfaces during growth with near atomic resolution under a range of temperatures and fluxes. We have applied atomistic modelling to the structures seen by STM to enable us to give confident interpretation of the results. For the growth of silicon on Si(001) from disilane we can give a rather complete account from the dissociation of the first gas molecule to the formation of a whole monolayer. A key role is played by the surface hydrogen. At temperatures up to 500°C the growth of germanium on Si(001) follows a similar path for the first few monolayers, after which the strain becomes relieved by trenches first in one direction, then in a perpendicular direction, and eventually by a combination of facetted pits and clusters, both of which nucleate heterogeneously at surface defects. Understanding these processes is crucial to controlling the size distribution of self-assembled Ge/Si quantum structures.