*3.1. Configurations and Integration*

To date, the most common configuration for optical sensing involves free-space coupling between the laser, waveguide, and detector by using bulk optics or microlenses. While conceptually simple, free-spacing coupling suffers from low coupling efficiency due to mode mismatch and is sensitive to vibrations and misalignment, thus affecting the robustness and compactness of the sensor. The usage of spot size converters or couplers has addressed some of the aforementioned problems [147]. At the same time, fiber-based coupling methods have also been used to improve the robustness and stability of the setup. This includes fiber pigtailed lasers or the direct coupling of a free-space beam into a fiber through a fiber collimator. Nonetheless, such techniques still fall short of an idealized compact and integrated version of the sensor that would fit completely on a single photonic chip and be suitable for mass production at a low cost.

First approaches for integration include hybrid integration, heterogeneous integration and monolithic integration. Even in integrated setups, inefficient coupling has been the major problem. This typically stems from the mode mismatch and misalignment between the active and passive sections of the chips, i.e., the laser cavity and the passive waveguide. For improving the coupling efficiency between the waveguide and the light sources, the usage of distributed Bragg reflectors (DBR) or appropriate mode profile engineering has been proposed [148]. Separate prototypes of integrations of the laser and the waveguide, the waveguide and the detector, the laser and detector are found in the literature; however, complete integration of all the three components in a single chip is still an active and ongoing pursuit, with very few actual demonstrations. The following sections describe recent efforts in this area.
