Effects of Surface Treatments on the Photoluminescence of Porous Si and a Suggested Mechanism for the Photoluminescence
- PDF / 1,135,339 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 12 Downloads / 216 Views
Effects of Surface Treatments on the Photoluminescence of Porous Si and A Suggested Mechanism for the Photoluminescence Masao YAMADA, Kazuaki KONDO, Takae SASAKI,
Akira TAKAZAWA,
and Tetsuro TAMURA
Basic Process Development Division, Fujitsu LTD., 1015 Kamikodanaka, Nakahara-ku,
Kawasaki 211, Japan
ABSThIACT
We report how vacuum annealing and dry oxidation affect the photoluminescence (PL) properties of porous Si. Vacuum annealing weakens the intensity of as-prepared porous Si, whereas, oxidation using dry oxygen at 5 Torr increases the intensity remarkably above 800 1C. Blue shifts as large as 100 nm are also observed for temperatures between 800 *C and 1000 *C.The photoluminescence decay patterns of both as-prepared and dry-oxidized porous Si excited by a nitrogen laser pulse are not exponential, but are fitted well by two exponential decays with lifetimes ranging from a few nanoseconds to over 100 nanoseconds. Further, those decay patterns do not change with changing excitation power by more than three orders of magnitude. These studies suggest that (1) hydrogen termination (with some fluorine termination) on the porous Si surface is not always necessary to get photoluminescence, (2) low surface defect densities are essential to increase the photoluminescence intensity, and (3) the radiative recombination path is not a direct interband transition, but, due to quantum size effects, it may go through some luminescence centers lying in the widened band gap. INTRODUCTION
Surface structures formed by anodic oxidation on porous Si layers are very complex. These surfaces consist of something like a thin a-Si:H layer and/or SiH , (x=1-3) termination with some fluorine terminations. 1 - 71 Because of these complex surface structures, it was difficult to get rid of some ambiguity in determining what structure is mainly responsible for the visible photoluminescence. Some groups say that even if quantum size effects 8 ) are fundamentally important, it, they would be due to siloxene derivatives 31 , and others say that it might be due to a-Si:H4-8) or (SiH 2 ) ,n In order to make discussion simpler, we tried to modify the porous Si surface by dry oxidation,'°0 because we expected that the oxidation would decompose the hydride-compounds formed on the porous Si surface and provide fairly good interfaces for low state defect densities (as is expected from a hydrogen-terminated surface). In the experiment, we observed blue shifts as large as 100 nm between 800 t and 1000 r-.There was also marked enhancement in the PL intensity above 800 *C.Prior to our study, Ito et al." 1 and Shlh et al. 1 ) reported nearly the same phenomena. Therefore, we can say that the photoluminescence of porous Si is not unique to porous Si covered with hydridecompounds, but is a more general characteristic of semiconductors with superfine structures. . In this paper, we focus on structure analyses of dry-oxidized porous Si using Fourier transform infrared spectroscopy (FTIR) and Transmission electron microscope (TEM). We also discuss if there is
Data Loading...
