In this webinar, we showcase the latest updates to our photonic inverse design technology. Photonic inverse design uses advanced computational methods to find optimal photonic device geometries. In a previous release, we added feature size constraints for parametric or shape-based optimization. With feature size constraints, designers can generate optimized devices that can be manufactured with standard photolithography techniques. In this webinar, we first demonstrate how feature size constraints can be applied to a shape-based optimization problem. In collaboration with foundry partner CompoundTek, we demonstrate a grating coupler design developed with inverse design, resulting in a significant improvement of the coupling efficiency as well as a 20x reduction in footprint. Second, we discuss our extension of our optimization toolbox by adding feature size constraints to our topology-based inverse design package.
Using inverse design techniques based on the adjoint method, many parameters can be simultaneously optimized for a set of different manufacturing and operating conditions. This ensures that the resulting design is both optimal and robust against variability in fabrication and packaging. In the webinar, we will demonstrate using the new topology optimization features to generate designs that are compatible with standard photolithography fabrication processes. We will show you how feature size constraints can generate high-performance, non-intuitive designs in radically reduced footprints, creating new opportunities for component designers. We will illustrate the workflow with the generation and optimization of a broadband O-band/C-band demultiplexer and share our results from an ultra-small footprint (<100um2) 4 channel O-band CWDM. We also show how to easily extract the GDS layout for the resulting structures.
The webinar begins with a review of the adjoint method and corresponding techniques for parametric (or shape-based) and topological optimization. Topology optimization is a new extension to Lumerical’s optimization toolbox, enabling the generation of optimal designs where intuition of the initial shape is limited. Next, we describe how to use photonic inverse design to develop designs that are robust to manufacturing. We will demonstrate the power of these advanced optimization techniques using both the varFDTD and FDTD solvers driven by the Automation API as applied to typical structures and design goals encountered in components engineered for integrated photonics.
- Introduction to Photonic Inverse Design (PID) and the Adjoint Method
- Topology-based vs shape-based PID
- How PID generates optimized and manufacturable devices
- CompoundTek’s Grating Coupler Design
- Topology PID workflow for generating photolithography compatible devices
- Examples of topology PID generated structures
Explore our current inverse design examples in our application gallery and knowledge exchange (KX), including
- Y-branch design,
- Waveguide crossing design,
- Grating coupler optimization, and
- O-Band/C-Band Demultiplexer.
These examples can be accessed directly from the most recent version of our products, which are available for a free 30-day trial.
Photonic Inverse Design for Manufacturable Devices
April 2, 2020 01:00 am
(UTC +00:00 GMT)