Bridging the Mid-Infrared-to-Telecom Gap with Silicon Nanophotonic Spectral Translation

Abstract

Expanding far beyond traditional applications in optical interconnects at telecommunications wavelengths, the silicon nanophotonic integrated circuit platform has recently proven its merits for working with mid-infrared (mid-IR) optical signals in the 2-8 μm range. Mid-IR integrated optical systems are capable of addressing applications including industrial process and environmental monitoring, threat detection, medical diagnostics, and free-space communication. Rapid progress has led to the demonstration of various silicon components designed for the on-chip processing of mid-IR signals, including waveguides, vertical grating couplers, microcavities, and electrooptic modulators. Even so, a notable obstacle to the continued advancement of chip-scale systems is imposed by the narrow-bandgap semiconductors, such as InSb and HgCdTe, traditionally used to convert mid-IR photons to electrical currents. The cryogenic or multi-stage thermo-electric cooling required to suppress dark current noise, exponentially dependent upon the ratio Eg/kT, can limit the development of small, low-power, and low-cost integrated optical systems for the mid-IR. However, if the mid-IR optical signal could be spectrally translated to shorter wavelengths, for example within the near-infrared telecom band, photodetectors using wider bandgap semiconductors such as InGaAs or Ge could be used to eliminate prohibitive cooling requirements. Moreover, telecom band detectors typically perform with higher detectivity and faster response times when compared with their mid-IR counterparts. Here we address these challenges with a silicon-integrated approach to spectral translation, by employing efficient four-wave mixing (FWM) and large optical parametric gain in silicon nanophotonic wires.