Mid-Term Review

Doping Si with appropriate deep level impurities at sufficiently high concentrations can lead to the formation of intermediate bands within the band gap.  This permits absorption in Si at sub-band gap wavelengths (beyond the band-to-band absorption limit at 1100 nm). Such a material opens up the possibility of fabricating Si-based optoelectronic devices that operate in the near- to mid-infrared region, which are substantially cheaper and more CMOS-compatible than their compound semiconductor counterparts.

Since deep level dopants have very low solubility in Si, highly non-equilibrium conditions are required to trap them inside the Si lattice. A well-established method for fabricating such hyperdoped materials is ion implantation followed by pulsed laser melting (PLM). After PLM, the steep temperature gradient induced by the laser pulse results in rapid resolidification of the Si crystal from a dopant-rich melt, allowing the dopants to be incorporated at concentrations which exceed the equilibrium solubility by several orders of magnitude. Indeed, an infrared photodetector has been previously demonstrated using Au hyperdoped Si fabricated by ion implantation and pulsed laser melting [1]. However, the external quantum efficiency of the demonstration device is poor (~2.8E-4 at 1310 nm) with room for improvement in both device architecture and optical absorption.

In the current work, we optimize both the ion implantation and PLM conditions in an attempt to further increase the substitutional Au concentration. RBS/C, SEM, TEM, EDS were employed to characterize both the structure and composition of the Au hyperdoped Si for further optimization of the optical absorption.  We show a ten-fold increase in sub-band gap absorption compared with previous attempts when the peak Au concentrations exceeds ~1 atom. %- a result that is particularly interesting given that these Au concentrations transcend the threshold concentration for the so-called “cellular breakdown”, a morphological feature previously thought to impede the sub-band gap absorption, to occur.

[1]        Mailoa, J.P., et al., Nat Comm, 2014. 5: p. 3011.