Electronic transport properties of Ti-supersaturated Si processed by rapid thermal annealing or pulsed-laser melting

Abstract

In the scope of supersaturated semiconductors for infrared detectors, we implanted Si samples with Ti at high doses and processed them by Rapid Thermal Annealing (RTA) to recover the crystal quality. Also, for comparative purposes, some samples were processed by Pulsed-Laser Melting (PLM). We measured the electronic transport properties at variable temperature and analysed the results. Results indicate that for RTA samples the surface layer with a high Ti concentration has a negligible conductivity due to defects. On the contrary, the implantation tail region, which has a measurable conductivity due to a very high electron mobility. This region shows the activation of a very shallow donor and of a deep donor level. While the deep level has been previously reported for Ti in Si, such a shallow level has never been measured, and we suggest that it originates from Ti-Si complexes. Finally, a decoupling effect between the implanted layer and the substrate seems to be present, and a bilayer model is applied to fit the measured properties. The fitted parameters follow the Meyer-Neldel rule. The role of the implantation tails in Si supersaturated with Ti is revealed in this work.

Ti supersaturated Si by microwave annealing processes

“…Regarding the Hall mobility (µ), the values of the Si:Ti sample are remarkably similar to the ones of the substrate, despite the enormous Ti dose implanted. These features are similar to the ones shown for RTA fabricated samples and would justify the application of the bilayer model used in [29] in the next section. This bilayer model and the meaning of the symbols used is described in the supplementary material.…”

“…In summary, the implanted layer mobility is just the measured Hall mobility in the whole temperature range. Finally, as we suggested in [29], the mobility behaviour can be fitted by a traditional potential equation due to phonon scattering at high temperatures (green dashed line), and by a temperature independent value due to a metallic behaviour at low temperatures (blue dashed line). Matthiessen's rule was applied for the whole temperature range.…”

“…The sample processed at 3 kW during 300 s yielded similar properties but much more noisy, probably due to a worse contact to the defective surface layer. All [28] and used in [29]. measurements yielded n-type conduction at all temperatures for both samples.…”

“…The activation energy of this level is unknown since measurements at temperatures below 60 K in the freeze-out region were not reliable. As described in [29], the origin of this shallow level would be complexes formed by Ti impurities and defects. Although Ti is always regarded as a deep level, no deep levels were needed to fit the measurements.…”

“…The electromagnet field (0.9 T) was fed with a bipolar Kepco BOP 50-20 MG power supply. To analyse the results, we will use the bilayer model developed in [28] which was also recently applied for RTA fabricated samples analysed in [29]. This bilayer model includes a temperature dependent decoupling factor (F) that takes into account a conduction limitation effect (G t ) between the implanted layer and the substrate (see figure 1).…”