Possible Mechanisms for Photoluminescence in Spark-Processed Si

  • PDF / 868,085 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 2 Downloads / 179 Views

DOWNLOAD

REPORT


ABSTRACT Possible mechanisms which may explain the strong blue/violet (410 nm) photoluminescence (PL) in spark-processed Si (sp-Si) are critically reviewed and discussed. They include involvement of silicon-oxides, silicon-hydrides, silanol groups, siloxene, and silicon carbide. It is concluded from the experimental evidence that none of the above mechanisms can satisfactorily explain the major contribution(s) to PL in sp-Si.

INTRODUCTION The origin of the strong blue/violet photoluminescence (PL) in spark-processed silicon (sp-Si) has been unresolved since its discovery in 1992.1 Several models have been proposed in the meantime. This paper critically reviews some of the suggested mechanisms without repeating the entire body of experimental results. The reader is referred for this to the quoted literature.

RESULTS AND DISCUSSION The experimental details (parts of which have been confirmed by others2-4) are briefly summarized below. For preparing sp-Si, a high frequency (several KHz), high voltage (several thousand volts) and low current (several mA) electric pulse is applied between a silicon substrate (or any other semiconductor or semimetal5, 6) and a counter electrode. 1 The highest PL intensity and the highest PL peak energy (3 eV/410 rm) is achieved when a stream of dry air is directed during spark-processing towards the front surface of silicon. 7 A less intense and green (2.368 eV/525 rum) PL is observed when the spark-processing is conducted in stagnant air or nitrogen. In both cases, the resulting PL is remarkably stable towards UV radiation, HF-etching, and high temperature annealing. The resulting substance has decay times in the nanosecond region. They are essentially independent of the detection wavelength (including the red wavelength). 9 For interpretation of the above results, it has been suggested 10 that the photoluminescence in sp-Si stems from SiO2 , or SiOx or radiation-damaged silicon oxide which may be formed 11 12 during spark-processing. Indeed, luminescent centers have been found in silicate glasses. , However, this luminescence was mostly traced to the presence of impurities, in particular to rare earth and transition metals or to Ge. 13 -15 High purity synthetic materials such as Coming 7940 or Suprasil, instead, show essentially no room-temperature fluorescence in the spectra range of 16 17 interest here. , In order to clarify this point, we have conducted a series of PL measurements on stoichiometric as well as on non-stoichiometric Si-O compounds which were produced by sputtering, vapor deposition, or ion implantations involving Si or silica substrates. The PL intensities of all these substances are observed to be more than three orders of magnitude smaller than those for sp-Si, as Figure 1 clearly demonstrates. High purity SiO 2 shows no detectable PL signal under the present experimental conditions. Moreover, the PL spectra depicted in 151 Mat. Res. Soc. Symp. Proc. Vol. 358 01995 Materials Research Society

Figure 1 are mostly different in shape and peak wavelength compared to