Semiconductor III–V photonic crystal (PC) laser is regarded as a promising ultra-compact light source with unique advantages of ultralow energy consumption and small footprint for the next generation of Si-based on-chip optical interconnects. However, the significant material dissimilarities between III-V materials and Si are the fundamental roadblock for conventional monolithic III-V-on-silicon integration technology. Here, we demonstrate ultrasmall III-V PC membrane lasers monolithically grown on CMOS-compatible on-axis Si (001) substrates by using III-V quantum dots. The optically pumped InAs/GaAs quantum-dot PC lasers exhibit single-mode operation with an ultra-low threshold of ~0.6 μW and a large spontaneous emission coupling efficiency up to 18% under continuous-wave condition at room temperature. This work establishes a new route to form the basis of future monolithic light sources for high-density optical interconnects in future large-scale silicon electronic and photonic integrated circuits.

A continuously tunable optofluidic distributed feedback (DFB) dye laser was demonstrated on a monolithic replica molded poly(dimethylsiloxane) (PDMS) chip. The optical feedback was provided by a phase-shifted higher order Bragg grating embedded in the liquid core of a single mode buried channel waveguide. Due to the soft elastomeric nature of PDMS, the laser frequency could be tuned by mechanically stretching the grating period. In principle, the mechanical tuning range is only limited by the gain bandwidth. A tuning range of nearly 60nm was demonstrated from a single dye laser chip by combining two common dye molecules Rhodamine 6G and Rhodamine 101. Single-mode operation was maintained with less than 0.1nm linewidth. Because of the higher order grating, a single laser, when operated with different dye solutions, can provide tunable light output covering the entire spectrum from near UV to near IR in which efficient laser dyes are available. An array of five DFB dye lasers with different grating periods was also demonstrated on a chip. Such tunable integrated laser arrays are expected to become key components in inexpensive advanced spectroscopy chips.

Photonics inverse design relies on human experts to search for a design topology that satisfies certain optical specifications with their experience and intuitions, which is relatively labor‐intensive, slow, and sub‐optimal. Machine learning has emerged as a powerful tool to automate this inverse design process. However, supervised or semi‐supervised deep learning is unsuitable for this task due to: (1) a severe shortage of available training data due to the high computational complexity of physics‐based simulations along with a lack of open‐source datasets and/or the need for a pre‐trained neural network model; (2) the issue of one‐to‐many mapping or non‐unique solutions; and (3) the inability to perform optimization of the photonic structure beyond inverse designing. Reinforcement Learning (RL) has the potential to overcome the above three challenges. Here, we propose Learning to Design Optical‐Resonators (L2DO) to leverage RL that learns to autonomously inverse design nanophotonic laser cavities without any prior knowledge while retrieving unique design solutions. L2DO incorporates two different algorithms – Deep Q‐learning and Proximal Policy Optimization. We evaluate L2DO on two laser cavities: a long photonic crystal (PC) nanobeam and a PC nanobeam with an L3 cavity, both popular structures for semiconductor lasers. Trained for less than 152 hours on limited hardware resources, L2DO has improved state‐of‐the‐art results in the literature by over 2 orders of magnitude and obtained 10 times better performance than a human expert working the same task for over a month. L2DO first learned to meet the required maxima of Q ‐factors (>50 million) and then proceeded to optimize some additional good‐to‐have features (e.g., resonance frequency, modal volume). Compared with iterative human designs and inverse design via supervised learning, L2DO can achieve over two orders of magnitude higher sample‐efficiency without suffering from the three issues above. This work confirms the potential of deep RL algorithms to surpass human designs and marks a solid step towards a fully automated AI framework for photonics inverse design.