Uncertainty aware design of photonic integrated circuits in presence of correlated manufacturing uncertainties

Fabrication technologies have to face the unavoidable reality of manufacturing uncertainties limiting the sustainable complexity and poses major challenges in achieving high production yield. This is particularly true for high-index-contrast photonic devices that are much longer than the operation wavelength; a small change in device dimensions can cause a dramatic phase error in the propagating light. While for compact photonic devices, the assumption of perfectly correlated variables can be a rather accurate description of the real variability, this condition might not be completely fulfilled when large circuits are considered. In this paper, we report the use of standard Polynomial Chaos Expansion technique for the stochastic analysis of photonics circuits in presence of multiple correlated random variables by exploiting the Karhunen-Loeve (KL) transformation.