A critical evaluation of Ag- and Ti-hyperdoped Si for Si-based infrared light detection

Abstract

Following recent successful demonstrations of enhanced infrared absorption in Au-hyperdoped Si, there has been strong interest in fabricating other metal-hyperdoped Si systems as a highly attractive approach for Si-based infrared photodetection. In this work, we address the somewhat contentious issue in the literature as to whether it is possible, using ion implantation and nanosecond pulsed-laser melting, to achieve hyperdoping of Si with Ag and Ti at concentrations exceeding that required to form an intermediate impurity band within the Si bandgap (NIB similar to 6 X 1019 cm-3). A wide range of characterization techniques were used to investigate these material systems, especially the quality of liquid-phase epitaxy, impurity concentration distribution both in depth and laterally, and impurity lattice location. Our results indicate that the high concentrations of opto-electrically active Ag or Ti in monocrystalline Si required to form an impurity band are not achieved. In particular, the usual behavior during rapid solidification is for near-complete surface segregation of the impurity, or for it to be trapped within a highly defective subsurface layer due to filamentary breakdown. Although our measurements showed that the maximum concentration of impurities outside metal-rich filaments is comparable to NIB for both Ag and Ti, there is no preferential Ag or Ti lattice location after pulsed-laser melting anywhere in the material. Thus, the concentration of opto-electrically active Ag and Ti that can be homogeneously incorporated into Si is expected to be well below NIB, leaving Au as the only viable impurity to date for achieving the required level of hyperdoping in Si.