The initial stages of homoepitaxy on GaAs(001) are studied with atomic-resolution scanning tunneling microscopy and Monte Carlo simulations that include the zinc blende structure of GaAs, the (2 3 4) reconstruction of the (001) surface, and the kinetics of As 2 incorporation. The reconstruction is found to favor nucleation on the top-layer arsenic dimers and to cause small islands to be unstable until they adopt the local (2 3 4) structure. [S0031-9007(98)06564-8] 02.70.Lq, 61.16.Ch, 68.35.Fx Heterostructures composed of III-V semiconductors are central to meeting the expanding requirements of optoelectronics and mobile telecommunications. The performance of devices based on such structures, in turn, depends crucially on the ability to fabricate high-quality heterogeneous interfaces. However, the growth kinetics on III-V semiconductor surfaces are poorly understood at the atomistic level even for homoepitaxy. Thus, to provide a framework for understanding and controlling the formation of heterogeneous III-V interfaces, we report in this Letter a study of the initial stages of homoepitaxy on the most intensively studied III-V surface, GaAs(001).The GaAs(001) surface presents a particular challenge because of the presence, under normal (As-rich) growth conditions, of the As-terminated (2 3 4) reconstruction [1], which substantially modifies the three topmost atomic layers of the surface (Fig. 1). Although the energetics of static surface reconstructions on this surface have been studied at length [2,3], much less is known about how reconstructions affect the kinetics during epitaxial growth from molecular beams [4][5][6][7].Our study is based on the combined use of atomicresolution scanning tunneling microscopy (STM) and kinetic Monte Carlo (KMC) simulations that include the zinc blende structure of GaAs, the (2 3 4) reconstruction of GaAs(001), and the kinetics of As 2 incorporation. The situation with an As 4 source is substantially different and will be presented elsewhere. By working in the low-coverage regime and making systematic comparisons between measured and simulated island morphologies, we find that the (2 3 4) reconstruction influences the growth both by favoring particular sites for island nucleation and in determining the stability of small islands.The experiments were carried out in a combined ultra-high-vacuum STM (Omicron GmbH, Germany) and molecular-beam epitaxy (MBE) system (DCA, Finland). Atomically clean GaAs(001)-(2 3 4) surfaces were prepared by standard methods using singular (to within 0.2 ± ) epiready substrates, which produced surfaces with terraces typically 1000 Å wide separated by 2.83 Å high Ga-As bilayer steps. These surfaces were held at 580 ± C in an As 2 flux of ͑2.5 6 1.0͒ 3 10 14 molecules cm 22 s 21 , after which Ga was codeposited at a rate of 0.1 monolayer (ML) s 21 (6 3 10 13 atoms cm 22 s 21 ) for no more than 1 s. The samples were then immediately transferred to the STM chamber and quenched to room temperature. The transfer process takes only a few seconds, and the sa...

No abstract

No abstract

No abstract

No abstract

No abstract

No abstract

We have performed a series of simulations of clusters of galaxies on the basis of the smoothed particle hydrodynamics technique in a spatially-flat cold dark matter universe with Ω = 0.3, λ = 0.7, and H 0 = 70km/s/Mpc as one of the most successful representative cosmological scenarios. In particular, we focus on the Sunyaev-Zel'dovich effect in submm and mm bands, and estimate the reliability of the estimates of the global Hubble constant H 0 and the peculiar velocity of clusters v r . Our simulations indicate that fractional uncertainties of the estimates of H 0 amount to ∼ 20% mainly due to the departure from the isothermal and spherical gas density distribution. We find a systematic underestimate bias of H 0 by ∼ 20% for clusters z ≈ 1, but not at z ≈ 0. The gas temperature drop in the central regions of our simulated clusters leads to the underestimate bias of v r by ∼ 5% at z ≈ 0 and by ∼ 15% at z ≈ 1 in addition to the statistical errors of the comparable amount due to the non-spherical gas profile.