Photonic integrated circuits (PICs) are considered as the way to make photonic systems or subsystems cheap and ubiquitous. PICs still are several orders of magnitude more expensive than their microelectronic counterparts, which has restricted their application to a few niche markets.

Nanophotonics finds ever broadening applications requiring complex components with many parameters to be simultaneously designed. Recent methodologies employing optimization algorithms commonly focus on a single performance objective, provide isolated designs, and do not describe how the design parameters influence the device behaviour. Here we propose and demonstrate a machine-learning-based approach to map and characterize the multi-parameter design space of nanophotonic components. Pattern recognition is used to reveal the relationship between an initial sparse set of optimized designs through a significant reduction in the number of characterizing parameters. This defines a design sub-space of lower dimensionality that can be mapped faster by orders of magnitude than the original design space. The behavior for multiple performance criteria is visualized, revealing the interplay of the design parameters, highlighting performance and structural limitations, and inspiring new design ideas. This global perspective on high-dimensional design problems represents a major shift in modern nanophotonic design and provides a powerful tool to explore complexity in next-generation devices.

Real photonic waveguides are affected by structural imperfections due to fabrication tolerances that cause scattering phenomena when the light propagates through. These effects result in extrinsic propagation losses associated with the excitation of radiation and backscattering modes. In this work, we present a comprehensive review on the extrinsic loss mechanisms occurring in optical waveguides, identifying the main origins of scattering loss and pointing out the relationships between the loss and the geometrical and physical parameters of the waveguides. Theoretical models and experimental results, supported by statistical analysis, are presented for two widespread classes of waveguides: waveguides based on total internal reflection (TIR) affected by surface roughness, and disordered photonic crystal slab waveguides (PhCWs). In both structures extrinsic losses are strongly related to the waveguide group index, but the mode shape and its interaction with waveguide imperfections must also be considered to accurately model the scattering loss process. It is shown that as long as the group index of PhCWs is relatively low (ng\<30), many analogies exist in the radiation and backscattering loss mechanisms with TIR waveguides; conversely, in the high ng regime, multiple scattering and localization effects arise in PhCWs that dramatically modify the waveguide behavior. The presented results enable the development of reliable circuit models of photonic waveguides, which can be used for a realistic performance evaluation of optical circuits

Sidewall roughness in optical waveguides represents a severe impairment for the proper functionality of photonic integrated circuits. The interaction between the propagating mode and the roughness is responsible for both radiative losses and distributed backscattering. In this paper, a unified vision on these extrinsic loss phenomena is discussed, highlighting the fundamental role played by the sensitivity of the effective index neff of the optical mode to waveguide width variations. The nw model presented applies to both 2D slab waveguides and 3D laterally confined waveguides and is in very good agreement with existing models that individually describe radiative loss or backscattering only. Experimental results are presented, demonstrating the validity of the nw model for arbitrary waveguide geometries and technologies. This approach enables an accurate description of realistic optical waveguides and provides simple design rules for optimization of the waveguide geometry in order to reduce the propagation losses generated by sidewall roughness.

Manufacturing variations are becoming an unavoidable issue in modern fabrication processes; therefore, it is crucial to be able to include stochastic uncertainties in the design phase. In this paper, integrated photonic coupled ring resonator filters are considered as an example of significant interest. The sparsity structure in photonic circuits is exploited to construct a sparse combined generalized polynomial chaos model, which is then used to analyze related statistics and perform robust design optimization. Simulation results show that the optimized circuits are more robust to fabrication process variations and achieve a reduction of 11%-35% in the mean square errors of the 3 dB bandwidth compared to unoptimized nominal designs.