Silicon Photonics-based Heterodyne Interferometric Imager for free-space imaging

This paper presents the design, fabrication, and demonstration of a silicon photonics-based heterodyne interferometric imaging system capable of performing one-dimensional spectroscopy and two-dimensional image reconstruction using a polarization-diversifying photonic integrated circuit with 91 baselines.

The Gist

The Big Idea: Turning a Giant Telescope into a Microchip

Imagine you want to take a picture of a star or the sun. Traditionally, you need a massive telescope with a giant mirror (like a giant eye) to catch enough light. These telescopes are huge, heavy, and require complex systems of cooling fans and motors just to keep the mirrors from warping due to heat. They are like trying to take a photo with a camera the size of a house.

This paper introduces a new way to do this: The "Microscope" Telescope.

The researchers built a tiny chip (about the size of a fingernail) made of silicon that can do the same job as a giant telescope, but it's small enough to fit in your pocket. They call this system MICRO.

How It Works: The Orchestra Analogy

To understand how this chip works, imagine a symphony orchestra trying to figure out what a single instrument is playing without seeing it.

1. The Problem: If you only have one microphone (one telescope mirror), you can hear the sound, but you can't tell exactly where the instrument is or what it looks like. You need many microphones spread out to figure out the shape of the sound.
2. The Old Way (Direct Imaging): You build a giant wall of microphones. It's expensive and hard to manage.
3. The New Way (Interferometry): Instead of one giant wall, you use many small microphones scattered around a room. You record the sound at each one and then mix them together mathematically. The "mixing" creates a pattern (like ripples in a pond) that reveals the shape of the source.

The MICRO Chip is the ultimate mixer.
Instead of having 14 separate microphones scattered across a room, this chip has 14 tiny "ears" (apertures) etched directly onto a piece of glass. It catches light from the sun or stars at these 14 spots, mixes them together inside the chip, and tells a computer what the image looks like.

The Secret Sauce: Heterodyne Mixing (The "Radio Tuner")

Here is the clever part. The light coming from the sun is very weak and quiet. To hear it clearly, the chip uses a technique called Heterodyne Interferometry.

• The Analogy: Imagine you are trying to hear a whisper in a noisy room. If you shout a specific note (a reference tone) at the same time the person whispers, the two sounds mix together to create a "beat" or a rhythm that is much easier for your ear to detect.
• In the Chip: The researchers shine a very strong, tunable laser (the "Local Oscillator" or LO) onto the chip. This laser acts like the shout. When the weak starlight hits the chip, it mixes with the strong laser. This boosts the signal, making the faint starlight loud enough to be measured, even if it's billions of miles away.

What Did They Actually Do?

The team built this chip and tested it in two ways:

1. Listening to the "Colors" (Spectroscopy):
   They shone different types of light onto the chip. The chip successfully identified the specific "notes" (wavelengths) in the light.

   • Real-world use: This helps scientists measure the temperature of stars or detect magnetic fields on the sun (by seeing how the light splits, known as the Zeeman effect).

2. Taking a "Picture" (Imaging):
   They shone a laser beam at the chip to simulate a star. By combining the data from the 14 different "ears" on the chip, the computer reconstructed an image of where the light was coming from.

   • The Result: They successfully created a blurry but recognizable picture of a single point of light and even a cluster of three points. It's like looking at a low-resolution photo, but the chip proved the concept works.

Why Is This a Big Deal? (SWaP)

The paper talks a lot about SWaP (Size, Weight, and Power).

• Current Telescopes: Like a moving truck. They need huge mirrors, heavy cooling systems, and lots of electricity.
• The MICRO Chip: Like a smartphone. It is tiny, uses very little power, and is light enough to be put on a satellite or a drone.

Because it is so small, we could potentially put many of these chips on a satellite. If you have one chip, you get a small picture. If you have a fleet of chips working together, you could build a virtual telescope the size of a city, giving us incredibly sharp images of the universe without building a massive physical structure.

The Hiccups and Future Plans

The chip isn't perfect yet.

• The "Traffic Jam": Inside the chip, the light has to travel through tiny paths (waveguides). Because the paths are so crowded, some light gets lost along the way (like traffic jams on a highway). This makes the final image a bit fuzzy.
• The Fix: The researchers plan to build better chips with more layers to reduce traffic jams and add more "ears" to get a clearer picture.

Summary

This paper is about shrinking the world's most powerful telescopes down to the size of a microchip. By using a clever mixing technique (heterodyne) and a chip with 14 tiny light-catchers, they proved we can take pictures of stars and analyze their light without needing a giant, heavy telescope. It's a major step toward putting high-tech astronomy on a satellite that fits in a suitcase.

Technical Summary

1. Problem Statement

Traditional high-resolution stellar and solar imaging systems (e.g., DKIST, Hinode SOT) rely on large optical trains, fold mirrors, and bulky stabilization instruments to achieve high spatial resolution and perform spectropolarimetry (measuring Zeeman splitting and Doppler shifts). These systems suffer from:

• High SWaP (Size, Weight, and Power): They require massive optics, active cooling (to 0°C), and complex electronics for stabilization.
• Scalability Issues: Expanding the aperture size or adding more baselines for better image reconstruction significantly increases system complexity and cost.
• Measurement Limitations: Traditional systems often capture only one polarization or wavelength at a time, requiring sequential measurements and recalibration, which increases observation time.
• Noise Sensitivity: Electronic systems are susceptible to thermal (Johnson) noise and electromagnetic interference (EMI).

The authors propose a solution using Silicon Photonics (SiPh) to create a compact, scalable, and low-SWaP interferometric imaging system capable of heterodyne detection for both spectroscopy and 2-D imaging.

2. Methodology

The paper presents the design, fabrication, and demonstration of the MICRO (Micro Interferometric Coherent Receiver for Optical) system, a Silicon Nitride (Si$_3$N$_4$) based Photonic Integrated Circuit (PIC).

• Architecture:

  • Heterodyne Detection: The system uses a strong Local Oscillator (LO) laser mixed with incoming light signals. This down-converts optical signals to radio frequencies (RF), amplifying weak signals (like starlight) and enabling phase-sensitive measurements.
  • Aperture Array: The PIC features 14 apertures arranged irregularly along a semicircle (22 mm diameter) to maximize unique baseline pairs for Fourier space (UV-plane) coverage.
  • Polarization Diversity: A 3-layer Si$_3$N$_4$ waveguide stack (150 nm thick layers separated by 50 nm SiO$_2$) is used to create polarization-diversifying gratings. These gratings separate incoming light into S- and P-polarizations, allowing simultaneous measurement of both.
  • On-Chip Components:
    • 2x4 Optical Hybrids: Combine input signals with the LO to generate In-phase (I) and Quadrature (Q) components.
    • Thermal Phase Shifters: Used to tune phase relationships and generate fringe data for image reconstruction.
    • Balanced Photodetectors: Located off-chip, they detect the mixed signals to extract visibility amplitude and phase.

• Fabrication:

  • Utilized ASML PAS 5500 deep-UV lithography.
  • Process involved LPCVD deposition of stoichiometric Si$_3$N$_4$, ICP etching, and CMP planarization.
  • Titanium/Gold heaters were deposited for thermal phase shifting.
  • The design includes waveguide crossings (simulated loss ~1.1 dB) to route signals between layers.

• Experimental Setup:

  • Inputs: Free-space collimated beams (simulating stellar sources) coupled vertically into the gratings.
  • LO: A tunable laser amplified by an EDFA, edge-coupled into the PIC and split to feed all 14 apertures.
  • Signal Processing: RF signals from balanced receivers are mixed using analog frequency mixers and lock-in amplifiers to extract visibility data, which is then processed via the van Cittert-Zernike theorem for image reconstruction.

3. Key Contributions

• First Free-Space SiPh Heterodyne Imager: Unlike previous fiber-coupled demonstrations, this system accepts free-space inputs, allowing the PIC to utilize its full aperture array for imaging.
• Polarization-Resolved Interferometry: The 3-layer Si$_3$N$_4$ design successfully separates and processes two distinct polarizations simultaneously, crucial for distinguishing Zeeman splitting (magnetic fields) from Doppler shifts (velocity).
• Compact 2-D Imaging: Demonstrated the ability to reconstruct 2-D images of point sources using a single chip with 14 apertures, replacing the need for large mechanical interferometer arrays.
• Low-SWaP Scalability: Proved that complex interferometric tasks (spectroscopy + imaging) can be performed on a chip-scale device, significantly reducing the footprint compared to traditional telescope instruments.

4. Results

• Waveguide Characterization:
  • Propagation loss: 0.672 dB/cm.
  • Bending loss: 0.6 dB/$\pi$ (at 100 $\mu$m radius).
  • Grating coupling efficiency: ~25% (6.34 dB for S-pol, 7.29 dB for P-pol).
  • Total theoretical system loss (grating to output): 16 dB, with crossing losses being a significant factor (10 dB).
• 1-D Spectroscopy:
  • Successfully reconstructed the spectrum of a narrow-linewidth laser (30 kHz) and a broadband source filtered by a Fiber-Bragg Grating (FBG).
  • The system resolved the FBG notch (0.1 nm FWHM) with an extinction ratio of ~5 dB (limited by system noise floor).
• 2-D Image Reconstruction:
  • Reconstructed images of single, double, and triple point sources (simulating stars).
  • Single Source: Clear peak intensity at the center.
  • Multiple Sources: Successfully resolved two and three distinct point sources at varying distances and power levels.
  • Limitations: Images showed sparse sampling artifacts due to the limited number of baselines and the "mirroring" effect of complex visibility. Image quality was also affected by system noise and optical losses.

5. Significance and Future Outlook

This work validates the feasibility of using Silicon Photonics for next-generation astronomical instruments, particularly for solar magnetography and stellar imaging.

• Scientific Impact: Enables the measurement of magnetic fields (via Zeeman splitting) and plasma dynamics (via Doppler shifts) with a compact, robust system that could be deployed on satellites or small telescopes where large optics are impractical.
• Technological Advancement: Demonstrates a path toward "enclosed" systems that perform real-time image reconstruction without bulky external optics.
• Future Work:
  • Loss Reduction: Increasing layer counts (e.g., to 5 layers) to minimize waveguide crossing losses and improve signal-to-noise ratio.
  • UV-Plane Coverage: Implementing rotation stages to rotate the PIC, filling gaps in the Fourier space for higher resolution.
  • Integration: Integrating on-chip photodetectors and electronic-photonic interposers to eliminate off-chip detection bottlenecks.
  • Solar Application: Adding quarter-waveplates to the front end to specifically target circularly polarized light for solar magnetic field mapping.

In conclusion, the MICRO system represents a significant step toward miniaturizing high-performance interferometric imaging, offering a scalable alternative to traditional bulky optical systems for space and ground-based astronomy.