This article spotlights a breakthrough from Stanford researchers: they’ve built an integrated photonic circuit on a CMOS-compatible chip that can both generate and analyze any polarization state of coherent light. By combining a polarization-splitting grating coupler with a two-stage binary-tree mesh of Mach–Zehnder interferometers, the team achieves on-chip polarization control that preserves the optical signal for further processing.
That means you get a compact, scalable alternative to those big, finicky external optics.
Integrated photonic polarization control on a CMOS chip
The device sits on a silicon-on-insulator platform. It uses a 70 nm partial etch to route horizontal and vertical polarizations into separate quasi-TE ports.
It doesn’t just separate polarizations—it also acts as a polarization rotator and beam splitter. That makes it a compact building block for on-chip polarization manipulation.
The integration works with standard CMOS manufacturing, so it’s aimed at scalable deployment in larger photonic systems.
At the heart of the setup is a two-stage binary-tree photonic mesh made of Mach–Zehnder interferometers (MZIs). This mesh can self-configure to steer light through different interference paths, giving precise control over the polarization state.
The chip measures complex amplitudes at the PSGC ports and uses adaptive, self-configuring algorithms to reconstruct the input Stokes vectors—without destroying the signal. So, it can both read and set polarization states entirely on-chip, mapping the whole Poincaré sphere in a reversible, programmable way.
The PSGC and the on-chip MZI mesh
The polarization-splitting grating coupler (PSGC) is designed to couple horizontal and vertical polarizations into separate quasi-TE waveguide channels. It also acts as a polarization rotator and beam splitter, letting you handle polarization compactly without external optics.
Finite-difference time-domain (FDTD) simulations at the standard telecommunication wavelength of 1.55 µm show a solid mode overlap of 86.4% with a 10.4 µm Gaussian mode. The insertion loss hovers around −2.9 dB overall, with a minimum of −2.6 dB near 1.535 µm.
About half the light ends up lost to the substrate because of near-vertical symmetry in the device geometry. That’s a common trade-off in compact on-chip polarization components.
By measuring the complex amplitudes at the PSGC ports and applying self-configuring control loops to the MZI mesh, the system non-destructively characterizes the input polarization state. This approach preserves optical power and phase information for downstream processing, which is crucial for integrated photonic systems that need continued signal integrity for routing, modulation, or detection.
Non-destructive analysis and on-chip synthesis
One of the most compelling features is the ability to run the photonic mesh in reverse. That lets you programmatically synthesize any polarization state right on the chip, covering the full Poincaré sphere.
In practice, researchers can analyze and generate polarization states without needing external polarization optics. This drastically cuts down system size, cost, and fragility.
The result is a stable, reconfigurable polarization platform that preserves phase coherence and directs light to a single circuit output. That’s essential for coherent communications and high-precision polarization sensing.
Why this matters for industry and science
Swapping out bulky, delicate external polarization components for an integrated, CMOS-compatible solution has big implications. The on-chip approach reduces alignment sensitivity and environmental drift, making field deployments more robust.
CMOS-compatible fabrication supports scalable production, reproducibility across wafers, and easier integration with other photonic and electronic subsystems. This kind of integration drives next-generation coherent communication links, advanced polarimetric sensing, and any application that needs precise, controllable light polarization on a compact footprint.
Performance highlights and potential applications
- On-chip Stokes vector reconstruction: These measurements don’t destroy the signal, so you can use it for more processing later.
- Reverse-operation synthesis: You can create any polarization state right on the chip, which means full Poincaré sphere access.
- Single-chip polarization control: This setup does away with bulky external optics and those polarization measurements that eat up your signal.
- CMOS-compatible fabrication: It’s all built on silicon-on-insulator, so you get scalable manufacturing in regular commercial foundries.
- Coherence preservation: The phase information sticks around, which is great for whatever coherent processing comes next.
This platform actually makes stable, scalable polarization control possible in real-world systems. It’s especially promising for coherent communications, polarimetric sensing, and integrated photonic networks.
Researchers are still working out the kinks—like boosting coupling efficiency and cutting down on substrate losses. But honestly, this approach seems likely to become a staple in future photonic chips. Maybe it’s finally the bridge between lab demos and something you can use out in the field.
Here is the source article for this story: Light’s Polarisation Fully Controlled On A Single Chip