91³Ô¹ÏÍø

Streamlining Photonic Integrated Circuit Design with PDKs

Twan Korthorst

Nov 14, 2022 / 3 min read

Calls for chips to deliver increasingly high levels of performance with low power consumption have never been louder. Electronic components in traditional semiconductors have their limitations when it comes to applications like high-speed data communications, imaging, and advanced sensing, which are becoming more ubiquitous each year. Photonic integrated circuits (ICs), on the other hand, tap into the power of light for substantially greater speed and capacity, along with advantages such as miniaturization, lower thermal effects, and large integration capability.

But, compared to traditional silicon chips, how easy is it to develop a photonic IC?

Photonic integrated circuit design is generally not a straightforward process, given that the device¡¯s performance is linked to material as well as optical properties that are tied to geometrical shapes. Foundries use different material platforms and their own process flows that, based on the material and optical properties, influence the performance of the photonic ICs at the physical level. In many cases, experience is the best indicator of how well a designer will be able to efficiently model a performant photonic integrated circuit.

In the semiconductor industry, process design kits (PDKs) have long been used to model the fabrication process for the design tools used for an electrical IC. The same approach can be applied to photonic integrated circuits, bridging the gap between the foundry technology and design requirements. PDKs can also accelerate the process toward first-pass chip success. Read on for more insights into how PDKs can help you accelerate your photonic IC design process.

Neon Colorful Tunnel Door Abstract Speed Blurred Motion Rotor Long Exposure Swirl Spiral Circle Wave Pattern Distorted Macro Photography

Building Blocks for Your Photonic IC Design

So, what does a PDK provide? Think of it as a plug-in library for your design environment, mainly providing a set of building blocks. Typical semiconductor PDK components include a device library (with symbols and layouts), verification decks (such as design rule checks and layout versus schematic), technology data (such as layers, colors, and process constraints), and simulation models of primitive devices (such as transistors, capacitors, resistors, and inductors), and a design rule manual.

IC designers can use these components to simplify the process of building different types of complex photonic circuits for different applications. Because these building blocks have been predefined and tested, designers can proceed with the assurance of meeting their quality, cost, and time-to-market targets.

Photonic building blocks found in a PDK typically include several types of waveguides, passive devices like splitters, combiners, and filters and, in addition, can contain active devices such as phase shifters, detectors, semiconductor optical amplifiers, and lasers.

Designers can also create their own building blocks, provided they follow the foundry¡¯s design and fabrication rules. In recent years, most of the foundries offering photonic IC manufacturing have added formal physical design rule verification decks to their PDKs. These design rule checking (DRC) decks contain rules to check whether a design is actually manufacturable and performs checks on minimum distances between layers, minimum overlap, acute angles, and much more. Especially in silicon photonics manufacturing, this is common practice and rule decks can contain up to hundreds of checks.

Modeling of Photonic Components

Accurate modeling of a photonic component requires consideration of its geometry, materials, and electro-optical properties. PDK model libraries can typically contain:

  • The refractive index of the material stack versus the wavelength
  • Values of pre-characterized components or validated components
  • Analytic models, which can be used to predict performance when real data isn¡¯t available

Signal-level simulation tools for photonic ICs use simplified or compact models in contrast to solving EM-equations in full 3D fashion at the device level. The function of many photonic devices can be represented by an S-matrix, which describes the signal transfer, given certain dependencies on wavelength, temperature, or geometry between the ports of the component. For components like a laser or optical amplifier, rate-equations or complex curve fitted multi-dimensional matrices are often used as a model to represent the behavior.

Supporting Custom Components in the IC Design Flow

When you¡¯re designing a photonic integrated circuit, the PDK from the foundry becomes an integrated part of your design environment and provides the components and design rules as described above. Often a designer needs additional application-specific components to complement the PDK. There are solutions available that support PDK-driven and custom simulation. Synopsys OptSim electro-optic co-simulation solution is such an option. Synopsys OptSim provides support for foundry-qualified model libraries and custom device models in a unified, schematic-driven layout photonic IC design flow. Learn more in our technical paper about electro-optic co-simulation for photonic ICs.

Synopsys also provides the industry¡¯s most comprehensive photonic IC foundry support, with PDKs available from foundries around the world for photonic processes such as:

  • Silicon
  • Silicon nitride
  • Indium phosphide
  • Polymers
  • Silica-on-glass

Synopsys offers engineering services to help foundries, IDMs, and design teams set up foundry-specific compact models, building block definitions, layouts, and physical verification rules.

Designing photonic ICs is complex, involving specific steps based on the application as well as the foundry. By using a PDK, design teams can accelerate their process to develop designs with hundreds of photonic components, taking advantage of the power of light for high-speed communications, LiDAR for automotive systems, imaging for aerospace and aeronautics, and sensing for healthcare.

Continue Reading