TSRB Researchers Solve Layering Issues in Photonic Device Design
A recent paper from TSRB and Georgia Tech School of ECE researchers presents a new design method to improve photonic devices used in multi-layer and multi-etch manufacturing processes. Photonic devices control and manipulate light in various ways and are crucial for telecommunications, data centers, and sensors. However, designing these devices for multi-layer systems poses challenges, such as imperfect layering (planarization defects) and limitations in stacking materials.
To overcome these challenges, the researchers propose a topology optimization (TO) technique that includes a "layer constraint" (LC). This LC ensures that the upper layers of a device are always supported by the layers below, significantly enhancing the reliability of the design and reducing the impact of manufacturing defects. The optimization method is designed to work with both multi-layer and multi-etch processes, ensuring the devices can be efficiently produced while maintaining high performance and providing a sense of reassurance and confidence in the new method.
The authors tested their approach by designing two specific devices: a multi-layer T-splitter and a strip-rib waveguide converter. After optimizing the designs with the LC and other fabrication rules, they built and tested these devices using a commercial silicon photonics process (GlobalFoundries). The T-splitter achieved -4.0 dB transmission, while the strip-rib converter had -0.08 dB transmission, closely matching the simulations and demonstrating an impressive and convincing strong performance across different wavelengths.
Photonic devices are widely used in high-speed internet, data centers, optical computing, sensors, and even quantum computing. This method improves the design process and ensures that photonic devices can be manufactured more reliably, reducing defects and enhancing performance. These advances could lead to faster, more efficient data transfer, more powerful computing technologies, and better sensing capabilities across various industries.