Au-catalyzed growth processes and luminescence properties of ZnO nanopillars on Si

Abstract

ZnO nanopillars are often grown on various substrates by catalytic growth processes through a vapor-liquid-solid (VLS) mechanism. However, on silicon substrates, even with the catalyzed growth processes, it is still very difficult to obtain highly oriented ZnO nanopillar arrays. In this work, it was found that in most cases the actual growth process of ZnO on Si catalyzed by Au was not of real VLS character. In the initial growth stage, the substrate surface is partially melted and then oxidized into a very thin layer of SiO2. Zn-rich alloys instead of ZnO are first deposited on the SiO2∕Si substrates and form polycrystalline hillocks in an atmosphere with low O2 partial pressure. The difficulty for ZnO to nucleate on SiO2∕Si is another reason preventing ZnO nanopillars from growing epitaxially on the substrates. Defects, steps, and∕or stress on the substrate surfaces may support the nucleation process and thus may influence the initial growth stage and the control of the growth orientation of the pillars. With the help of scratches on Si substrates and by avoiding the formation of hillocks, well-aligned ZnO nanopillars were obtained both on Si (111) and (100) surfaces. Sharp photoluminescence (PL) peaks of bound exciton emissions and a free exciton emission were detected at low temperature. An unreported PL peak at about 3.367eV and in some cases a donor-acceptor-pair transition at about 3.308eV were also found. At higher temperature, a band-acceptor recombination process together with strong LO-phonon replicas occurred. A donor dopant concentration ND of about 5.7×1018cm−3 and an acceptor ionization energy of 126±2meV were evaluated from the spectra.

Growth of ZnO nanostructures on Au-coated Si: Influence of growth temperature on growth mechanism and morphology

“…EDX data ͑not shown͒ of Au tips on VLS NR grown on thick SiO 2 layers also display no evidence of Si. These data in their entirety indicate that the growth model involving alloying between Si atoms in the substrate and the Au catalyst 11,41 is not consistent with our data. The data shown in Ref.…”

“…The same inconsistency has been highlighted recently by a number of other reports in the literature. 11,12 The growth at these lower temperatures in our system is due entirely to VS growth, and this is demonstrated by the hexagonally faceted morphology and flat NR tops observed in SEM, which are also seen for growth on ZnO buffer layers with no Au catalyst. 5 The absence of VLS growth at this temperature may be explained by considering the phase diagram for a Au-Zn alloy.…”

“…More recent reports have also commented on the inconsistencies in this uniform attribution in a similar vein to the current work. 11 The report in Ref. 11 has suggested growth mechanisms relying on alloying between Si atoms in the substrate and the Au catalyst.…”

“…10 However, some more recent research work reports Au-catalyzed ZnO growth with features inconsistent with the VLS mechanism. 11,12 These reports indicate that the growth mechanism is vapor-solid ͑VS͒ growth.…”

Microscopic origins of the surface exciton photoluminescence in ZnO nanostructures

“…3,2,6,10,15 However, a number of reports of low temperature PL from ZnO nanostructures with varying aspect ratios show no consistent correlation from one report to another [22][23][24] and within the same report 2,15 between the SX peak relative intensity and the nanostructure aspect ratio. Based on the entirety of these reported data we question whether the SX peak's relative intensity is solely or mainly determined only by the nanostructure surface-to-volume ratio.…”