MIRCOMB for « nonlinear integrated Mid-IR frequency COMB » is aiming at creating novel integrated nonlinear sources on CMOS for applications in the mid infrared (mid-IR, wavelength range between 2 and 10 um). The mid-IR wavelength domain is of great interest for a huge range of applications that affect almost every aspects of our society, from highly sensitive biological and chemical sensors, homeland security applications, industrial and environmental monitoring, defence, astronomy... The reason is that most chemical and biological compounds have unique fingerprints in this range that can be exploited in detection schemes with unprecedented reliability, sensitivity and precision. The surge in interest in the mid-IR has been recently driven by breakthroughs in novel sources in this wavelength range. QCLs generating light between 4 μm and 250 μm are commercially available but remain extremely bulky. The availability of integrated components for the mid-infrared remains actually extremely limited. Devices operating in the mid-IR are largely restricted to stand-alone and narrow-band sources and simple passive optical light guiding, typically based on highly multimoded fiber in chalcogenide glass, silver halide fibers, and hollow core fibers. However, over the past three years, there have been calls for the migration of techniques developed by the near-IR community to the mid-IR. In particular, Soref and Jalali have pointed out that group IV materials like silicon and Ge are ideal candidates for applications far beyond the window of telecom applications – right up to 5 μm and 10 μm respectively. This project addresses this vision and will benefit many sectors and priority areas listed by this P2N call. More specifically MIRCOMB will target highly innovative i) supercontinuum and ii) frequency-COMB sources on a compact SiGe-based chip-scale platform.