The temperature dependence of the 969 meV 'G' optical transition in silicon

Abstract

The temperature dependence of the intensity of the G vibronic band is reported. Absorption data are given for the range 2 to 200 K, and luminescence data for the range 2 to 90 K. The data show that the molecular structure of the G centre is invariant in these ranges. From 2 to 30 K the luminescence intensity undergoes a specimen-dependent growth. We suggest the growth is caused by the thermal ionisation of competing shallow exciton traps. Above 30 K the luminescence intensity undergoes an intrinsic decrease, and is quenched to 0.2% of the low-temperature value by 77 K.

On the temperature dependence of point-defect-mediated luminescence in silicon

“…This is accurate for Ref. 6 but the active region of the W line LED in Ref. 2 was lightly p-doped to a resistivity of 5 ⍀ cm.…”

“…1 shows three least-squares fits ͑for three different carrier concentrations͒ of this model to the data of Bao et al 2 for the temperature dependence of electroluminescence intensity from a W line light-emitting diode ͑LED͒. The inset shows a least-squares fit to the data of Davies et al 6 for the temperature dependence of G line photoluminescence intensity for an assumed carrier concentration of 10 15 / cm. This data series is truncated below 20 K to exclude points influenced by a trap state not contained in this model.…”

“…2,3,5 The most detailed phenomenological model of this temperature dependence, developed by Davies et al 6 to describe G line photoluminescence, is not based on specific physical mechanisms. It relies on two unknown energies: an unspecified trap level and an energy barrier for a general nonradiative process.…”

“…While this model was able to fit experimental observations, it does not provide the detail necessary to consider the feasibility of this type of device. This letter presents an initial attempt to build on the work of Davies et al 6 by describing the observed luminescence using familiar concepts from semiconductor physics. Employing this approach is an initial step toward addressing questions of materials design for point-defect luminescence in silicon.…”

“…6 These levels likely arise from bound exciton formation at the neutral carbon complex responsible for the G line. Because the data are qualitatively similar, we hypothesize that the same is true of the neutral self-interstitial complex responsible for the W line.…”