Passive Silicon Nitride On-Chip Polarimetry: Precise Polarization Detection with Imperfect Components

TL;DR

The paper tackles the challenge of measuring polarization of visible free-space light with compact, integrable devices. It proposes a fully passive silicon nitride PIC that maps polarization into on-chip intensity measurements using a compact 2D grating coupler and passive interferometers, enabling single-shot polarization reconstruction. A calibration framework estimates imperfect parameters, including $\alpha$, $x_i$, and $t_{ij}$, to correct for non-ideal polarization splitting and phase behavior. Experimental results across 50 polarization states yield strong agreement with theory, achieving $\Delta S_{RMS} = 0.028$, and the approach is compatible with CMOS scaling and potential array implementations for spatially resolved polarimetry.

Abstract

Polarization is a fundamental property of light that carries distinct and valuable information. Consequently, its precise measurement is crucial for numerous applications, including biomedical imaging, remote sensing, and optical communication. Since polarization cannot be measured directly, it is typically inferred by converting it into intensity signals using dedicated optical elements. Conventional approaches, however, predominantly rely on bulky optical components, leading to considerably high fabrication costs and limited integration density. Here, we introduce a passive photonic integrated circuit capable of precisely determining the polarization state of visible free-space light. An silicon nitride on-chip architecture employing a compact polarization-splitting grating coupler and a set of passive interferometers encodes the polarization information into intensity signals, allowing conventional detectors to accurately reconstruct the polarization state. With increasing compactness of photonic components, however, susceptibility to fabrication tolerances as well as intrinsic design constraints increases, potentially leading to non\-/ideal behaviour. To address this, we introduce a robust calibration procedure that enables precise measurements even in the presence of imperfections. The chip design, combined with the calibration procedure, offers a robust, small-footprint, and high-speed approach to polarimetry, enabling a wide range of applications.

Figures (10)

• Figure 1: Design of a compact two-dimensional grating coupler with a focusing grating design. (a) Schematic representation of different types of grating couplers. (b) SEM image of the curved two-/dimensional grating coupler. (c) Measured (solid line) and ideal (dashed line) normalized transmission for the different ports of the grating coupler, with incident light linearly polarized at various orientations. The polarization direction was adjusted using the 2D polarization state generator shown in Fig. \ref{['fig: setup']}.
• Figure 2: Artistic illustration of the chip design (a) and optical microscopy image of the fabricated chip (b). A focusing 2D grating coupler serves as the interface that couples free-space light into the waveguides on the chip. This coupler splits the light into two linearly polarized components and directs each into separate waveguides. The two polarization components are then analyzed for their relative phase and amplitude using passive on-chip interferometers, providing complete information about the 2D polarization state of the incident light. Finally, the output signals of the on-chip architecture are coupled back into free space by means of standard grating couplers. To simplify experiments, the chip layout was intentionally designed with certain distances increased, resulting in a total footprint of 3295 × 325 µm.
• Figure 3: a) Illustration of the experimental setup: A collimated Gaussian beam first passes through a polarization state generator, which includes a horizontally aligned linear polarizer and two liquid crystal variable retarders (LCVRs) with their slow axes aligned at +45° and 0°, respectively. The beam then enters the chip's input region. The chip's output signals are monitored by means of an imaging system, consisting of a D-/shaped mirror, an objective and a camera. b) Stokes parameters produced by the polarization state generator depending on the retardance set at the two liquid crystal variable retarders. The polarization state generator allows for the controlled creation of any possible polarization state by adjusting the retardance of LCVR$_1$ between 0 and $\pi$ and LCVR$_2$ between 0 and $2 \pi$.
• Figure 4: Results from the calibration procedure, illustrating the output intensities of the chip for various input polarization states. The input polarization states are produced by applying specific retardances to the liquid crystals of the polarization state generator, as depicted in Fig. \ref{['fig: setup']}. Consequently, the intensities are presented as a function of the applied retardances. (a) displays the experimentally measured output intensities, while (b) presents the theoretically calculated values obtained using the fitted theoretical model.
• Figure 5: Measured and theoretically expected Stokes parameters for 50 different polarization states, forming a closed loop on the surface of the Poincaré sphere. The distinct polarization states were probed by the input interface of the chip. The chip’s output intensities were used to reconstruct the polarization state of the input light. An alternative representation of the measurement results, with the measurement points depicted as dots on the surface of the Poincaré sphere, is displayed in the inset. The theoretical and experimental values show strong agreement, with a root mean square deviation of $\Delta S_{RMS} = 0.028$.