Self-assembled GaAs nanowires have been grown on Si by molecular beam epitaxy without the use of any outside metal catalyst. The growth occurs on Si facets obtained by the cleavage of Si(100) substrates. The growth has been obtained with or without Ga pre-deposition. In both cases two kinds of nanowires have been obtained. The wires of the first type clearly present a Ga droplet at their free end and have a lattice structure that is wurtzite for wide regions beneath the Ga droplet. The second type, in contrast, ends with pyramidally shaped GaAs and has a crystal lattice that is mainly zincblende with only a few and small wurtzite regions, if any. The Ga-ended nanowires are longer than the others and thinner on average. The experimental findings suggest that the two types of nanowires grow after different growth processes.

The electron effective mass, $m_{\mathrm{e}}$, has been determined by magnetophotoluminescence in as-grown and hydrogenated $\mathrm{Ga}{\mathrm{As}}_{1-x}{\mathrm{N}}_x$ samples for a wide range of nitrogen concentrations (from $x<0.01\%$ to $x=1.78\%$). A modified $\mathbf{k}\mathbf{\bullet }\mathbf{p}$ model, which takes into account hybridization effects between N cluster states and the conduction band edge, reproduces quantitatively the experimental $m_{\mathrm{e}}$ values up to $x⩽0.6\%$. Experimental and theoretical evidence is provided for the N complexes responsible for the nonmonotonic and initially puzzling compositional dependence of the electron mass.

Single-crystal ZnSe nanowires are grown on a prepatterned gold catalyst by molecular-beam epitaxy. Optimum selectivity and maximum nanowire densities are obtained for growth temperatures in the range $400–450°\mathrm{C}$ , but gold-assisted growth is demonstrated for temperatures as low as $300°\mathrm{C}$ . This suggests a diffusion process on/through the catalyst particle in the solid state, in contrast to the commonly assumed liquid phase growth models. Straight wires, as thin as $10\mathrm{nm}$ , nucleate together with thicker and saw-like structures. A gold particle is always found at the tip in both cases.

We study the martensitic phase transformation in ${\mathrm{Ni}}_{\mathit{x}}$ ${\mathrm{Al}}_{1\mathrm{-}\mathit{x}}$ in the composition range 0.60<x<0.66 with the quasiharmonic approximation and molecular dynamics. These alloys present a high-temperature bcc phase (austenite) transforming at low temperature into a compact 3M structure (martensite). In our computation the potential energy of the system is approximated by an embedded-atom potential calibrated to reproduce selected properties of the pure elements in the fcc structure and of the Ni-Al intermetallic compound. We determine the transition temperature ${\mathit{T}}_0$ as a function of composition, pressure, and uniaxial stress. At zero pressure the computed ${\mathit{T}}_0$ is in qualitative agreement with the experimental results, and reproduces the strong composition dependence of the transition temperature. Again in agreement with experiment, the computation of thermodynamic properties shows that the transition is weakly first order, entropy driven, and closely matches the Zener picture. These conclusions are supported by the analysis of several microscopic quantities, including the mean-square displacement, the structure factor, and the phonon modes. Computations performed with applied uniaxial stress show that it always favors the 3M structure. The effect of pressure is more subtle: the 3M phase is favored by a moderate pressure, while higher pressures stabilize the bcc structure and decrease ${\mathit{T}}_0$. We analyze the effect of disorder on the transition.