Structural investigations were also carried out using TEM on eight different single nanowires taken from two samples. Figure 4a displays a TEM image of a whole nanowire, while Figure 4b shows a high-resolution picture of the

nanowire revealing find more its check details silicon lattice. No defects were detected in the crystalline matrix of any of the observed nanowires which give evidence of their very good crystallinity. Fast Fourier transform (FFT) of TEM pictures (inset of Figure 4b) of all observed nanowires show that the (111) planes of silicon are oriented perpendicular to the growth axis. The observed nanowires therefore grew along the [111] direction, which is different from the ones characterized by GIXD and from the substrate orientation Daporinad in vivo (100). In this case, there is no epitaxial relation between the nanowires and their substrate. The monocrystalline quality of the observed [111] nanowires in spite of their nonepitaxial growth is an important feature for the possible future use of this technique on noncrystalline substrates

such as stainless steel or glass. It ensures that semiconductor nanowires can be grown on universal substrates with a very good crystalline quality. We also notice on the TEM pictures that the nanowires’ surface presents low-contrast clusters. Energy dispersive X-ray microanalysis of these areas did not allow any detection of contamination materials such as aluminum (unshown results). This feature could be actually caused by topography effects due to the roughness of the nanowires’ surface as described in Figure 2e. Figure 4 Transmission electron microscopy. TEM view of a silicon nanowire which grew in the AAO template. (a) Low-resolution view of the nanowire. (b) High-resolution picture near the apex of the nanowire. Flucloronide Upper inset is an FFT of the image showing the periodicity along the growth axis corresponding to the (111) planes of silicon. Lower inset presents a high-resolution view clearly displaying the (111) planes. Two types of nanowires

therefore grew in the AAO template, one in epitaxy with the (100) substrate and another one with no crystalline relation with it, each type being clearly detected with a separate technique. Using SEM pictures such as the one of Figure 2e, it is not possible to visually differentiate between the two types of wires since they are all well individualized and fully guided in the nanopores. The most likely cause for the nonepitaxial nanowire growth is a partial deoxidation of the silicon substrate during the vapor HF step before catalyst electrodeposition. If the silicon surface at the bottom of a pore is only partially deoxidized, the remaining native oxide would disturb the initial growth steps by screening the substrate and therefore preventing a good epitaxy. This effect is known and described in the case of copper electrodeposition in nanoporous alumina [27].