Falconieri, Mauro, Borsella, Elisabetta, D'Amato, Rosaria, Fabbri, Fabio, Trave, Enrico, Bello, Valentina, Mattei, Giovanni, Nie, Yaru, and Wang, Dayang
The process responsible for the visible-near infrared photoluminescence (PL) emission in Si nanostructures has been generating significant controversy for years. Models based on band to band recombination of electrons and holes in Si nanostructures, or on recombination at defects located at the surface are currently proposed, but it is experimentally difficult to distinguish the two mechanisms since both are size-dependent. Here we compare the spectroscopic properties of Si nanocrystals (Si-nc), as prepared by laser pyrolysis and after complete conversion to amorphous silica by alkali etching-assisted oxidation. The strong resemblance of the spectral and time-behavior of the red PL emission in both systems, suggests that this emission is dominated by oxide-related defect states. We show that the non-exponential time decay of the PL emission in both systems (Si-nc and oxidised amorphous sample), can be modelled as the sum of exponential decays from four emitting centres thus ruling out the interpretation in terms of the so-called 'stretched exponential' function. By analysing their emission properties, we show the four emitting centres are actually non-bridging oxygen defects. Changes in the emission energies of these defects, due to size-dependent strain and chemical inhomogeneity, can be erroneously attributed to quantum confinement effect in the Si-nc. [ABSTRACT FROM AUTHOR]