The kinetics and mechanism of arsine adsorption on the (4 × 2) surface of gallium arsenide (001) has been studied by scanning tunneling microscopy, infrared spectroscopy, and ab initio quantum chemistry calculations. Arsine forms a dative bond to a gallium dimer. Then, this species either desorbs from the surface or decomposes to an AsH 2 or AsH fragment with hydrogen transfer to an arsenic site. Finally, desorption of hydrogen leaves arsenic dimers on the surface. The energy barriers for arsine desorption and dissociation into AsH 2 are estimated to be 9.3 and 16.5 kcal/mol, respectively. Gallium hydride is not produced upon dissociation of AsH 3 because this process is not energetically favorable.

The relationship between the reflectance difference spectra and the atomic structure of arsenic-rich reconstructions of GaAs͑001͒ has been investigated. Scanning tunneling micrographs reveal that a roughening process occurs as the surface structure changes with decreasing arsenic coverage from 1.75 to 0.75 monolayers ͑ML͒. At 1.65 ML As, small pits, one bilayer in depth and having the same c(4ϫ4) reconstruction as the top layer, form in the terraces. At the same time, gallium atoms are liberated to the surface, disrupting the c(4 ϫ4) ordering. At about 1.4 ML As, (2ϫ4) domains nucleate and grow on top of the c(4ϫ4). Further desorption of arsenic causes the underlying layer to gradually decompose into a metastable (2ϫn) phase (n ϭ2, 3, or 4͒, and finally into the (2ϫ4). In the reflectance difference spectra, negative peaks at 2.25 and 2.8 eV correlate with the c(4ϫ4)-type arsenic dimers. However, the intensity of the latter feature strongly depends on the presence of adsorbates, such as alkyl groups and gallium adatoms. By contrast, the intensity of the positive peak at 2.9 eV is directly proportional to the density of (2ϫ4)-type dimers.

The surface roughness of gallium arsenide ͑001͒ films produced by metalorganic vapor-phase epitaxy has been studied as a function of temperature and growth rate by in situ scanning tunneling microscopy. Height-height correlation analysis reveals that the root-mean-height difference follows a power-law dependence on lateral separation, i.e., ⌫(L)ϭkL a , up to a critical distance L c , after which it remains constant. For layer-by-layer growth, the roughness exponent, ␣, equals 0.25 Ϯ0.05, whereas the critical distance increases from 50 to 150 nm as the substrate temperature increases from 825 to 900 K. The roughness exponent jumps to 0.65Ϯ0.1 upon transitioning to three-dimensional island growth. By relating the height-height correlation analysis to the Einstein diffusivity relationship, the activation energy for gallium surface diffusion has been estimated: E d ϭ1.35Ϯ0.1 eV.

Germanium ͑100͒ crystals, 9°off-axis towards the ͓011͔ were exposed to 2.0 Torr of tertiarybutylarsine and 99.0 Torr of hydrogen at 650°C, then heated to between 450 and 600°C in vacuum or H 2. The resulting surfaces consist of narrow dimer-terminated terraces, with ͑1ϫ2͒ and ͑2ϫ1͒ domains, that are separated by steps between one and eight atomic layers in height. The distribution of ͑1ϫ2͒ and ͑2ϫ1͒ domains changes with temperature, exhibiting a pronounced maximum in the ͑1ϫ2͒ fraction at 510°C. These results suggest that the arsenic passivation of germanium is a critical step in gallium arsenide heteroepitaxy.