Silicon photonics has emerged as a reliable, cost-effective, and CMOS-compatible technology. Silicon exhibits transparency at telecom wavelengths (1.31 and 1.55 µm) and, due to its high refractive index contrast with silicon oxide, enables the fabrication of high-performance optical components. However, its indirect bandgap hinders the monolithic integration of laser sources, leading to increasing reliance on direct-bandgap III–V materials. Several approaches for integrating lasers on silicon have been developed, but their environmental impacts and the extensive use of III–V materials raise new concerns. This work presents a cradle-to-gate comparative Life Cycle Assessment (LCA) across 16 impact categories of laser fabrication in the CEA-LETI cleanrooms. Variations in the native substrate used for epitaxial growth of the active region are investigated: InP or GaAs substrates—as well as differences in laser architecture: conventional hybrid lasers or membrane lasers and finally: two different types of active region quantum dot and quantum well.  Given the exponential growth of AI-driven datacenters and the massive deployment of III-V lasers for optical interconnects, this study provides insights for guiding the industry toward more sustainable large-scale integration of III-V lasers for co-packaged optics.

Speaker : Elona Fidon : Research Engineer III/V Materials Growth & Sustainable Electronics
LTM, Univ. Grenoble Alpes, CNRS