🔭 astrophysics

Tri-coupler geometries for achromatic nulling interferometry in the near infrared

This paper investigates and compares various three-waveguide photonic tri-coupler geometries for near-infrared achromatic nulling interferometry, demonstrating that tapered tri-couplers and multimode interference couplers significantly outperform standard designs by achieving high exoplanet throughput and improved fabrication tolerance for future exoplanet detection instruments.

Original authors: Harry-Dean Kenchington Goldsmith, Nemanja Jovanovic, Anusha Pai Asnodkar, Yoo Jung Kim, Ahmed Sanny, Pradip Gatkine, Michael P. Fitzgerald

Published 2026-03-02

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Harry-Dean Kenchington Goldsmith, Nemanja Jovanovic, Anusha Pai Asnodkar, Yoo Jung Kim, Ahmed Sanny, Pradip Gatkine, Michael P. Fitzgerald

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Cosmic "Noise-Canceling" Headphones

Imagine you are trying to hear a tiny, whispering firefly (an exoplanet) buzzing next to a blindingly bright stadium floodlight (its host star). In the vast darkness of space, the light from the star is so overwhelming that it completely drowns out the firefly. If you look directly at the star, you see nothing but white glare.

To find the firefly, astronomers need a special trick: Nulling Interferometry. Think of this like noise-canceling headphones. Just as headphones listen to outside noise and play a "reverse" sound wave to cancel it out, these astronomical instruments take the light from two different telescopes, flip one of them upside down (a 180° phase shift), and smash them together.

The Starlight: Because the light waves are perfectly opposite, they cancel each other out. Silence! (This is the "null").
The Firefly: The firefly is slightly off to the side, so its light doesn't line up perfectly to cancel. It survives the crash and gets through. Whisper heard!

The Problem: The "Rainbow" Glitch

The paper discusses a specific tool used to do this canceling: a Tri-coupler. You can think of this as a tiny, microscopic traffic intersection made of glass channels (waveguides) where light travels.

The old way of building these intersections (the Standard Tri-coupler) worked well, but it had a major flaw: it was chromatic.

The Analogy: Imagine a pair of noise-canceling headphones that work perfectly for a bass drum (red light) but fail miserably for a violin (blue light). Because light of different colors (wavelengths) travels at slightly different speeds through the glass, the "cancellation" only works for one specific color. To see the full spectrum of an exoplanet, you need a device that cancels all colors at once. This is called being achromatic.

The Three Contenders

The researchers tested three different designs for this microscopic traffic intersection to see which one could cancel starlight across a wide range of colors (the near-infrared band) while letting the planet light through.

1. The Standard Tri-Coupler (The "Old Reliable")

How it works: Three straight, parallel glass tubes that are very close together. Light "leaks" (tunnels) from one to the other.
The Result: It's a decent worker, but it gets confused by different colors. At the edges of the color spectrum, it lets too much starlight through and blocks too much planet light.
Verdict: Good, but not great for wide-spectrum viewing.

2. The Tapered Tri-Coupler (The "Shape-Shifter")

How it works: This is a clever redesign. Instead of straight tubes, the outer tubes get narrower and narrower as they approach the center, like a funnel or a tapered pencil.
The Analogy: Imagine a traffic intersection where the lanes slowly merge and shift shape to guide cars perfectly, regardless of how fast they are going. By changing the width of the tubes, the researchers tricked the light into behaving the same way for all colors.
The Result: The Winner for Performance. It lets almost 100% of the planet light through and cancels the starlight beautifully across the whole color band.
The Catch: It's very sensitive. If the glass tubes aren't manufactured with perfect precision (even a tiny scratch or a slight width error), the whole system breaks down. It's like a high-performance race car that runs great on a perfect track but stalls on a bumpy road.

3. The MMI (Multimode Interference Coupler) (The "Bouncy Castle")

How it works: Instead of three thin tubes, this uses one wide, open room (a wide waveguide) where light bounces around like a pinball, interfering with itself to create the cancellation pattern.
The Analogy: Think of a large, open dance floor. Instead of three narrow hallways, everyone runs into a big room, bounces off the walls, and finds their way out.
The Result: The Winner for Durability. It's not quite as perfect at letting light through as the Tapered one, and it loses a bit more light to "friction" (loss). However, it is incredibly tough. If the manufacturing isn't perfect, it still works almost as well as the perfect version. It's like a sturdy SUV that handles potholes and bumps without missing a beat.

The "Fringe Tracking" Puzzle

There is one more job these devices must do: Fringe Tracking.

The Analogy: Imagine you are trying to keep a tightrope walker balanced. You need to feel the wind (phase errors) and adjust the rope instantly. The device needs to tell the computer, "Hey, the wind is pushing left, move the rope right!"
The Issue: Some designs get "degenerate." This means at certain colors, the device gets confused and can't tell which way the wind is blowing (the signal becomes flat).
The Finding: The researchers found that by tweaking the Tapered design further, they could create a version that never gets confused, allowing for constant, real-time adjustments.

The Big Picture: What Does This Mean?

This paper is a blueprint for building the next generation of "space glasses" that will help us find Earth-like planets.

We need better tools: The old tools (Standard couplers) are too picky about color.
We have two new options:
- The Tapered Design: If we can build it perfectly, it's the best performer. It's the "Ferrari" of planet hunters.
- The MMI Design: If we can't guarantee perfect manufacturing (which is often the case), this is the "Toyota" that gets the job done reliably. It's robust and forgiving.
The Goal: The ultimate dream is to combine the best of both worlds: a device that is as sensitive as the Ferrari but as tough as the Toyota, all packed onto a single, tiny silicon chip.

In summary: The researchers have tested three ways to build a microscopic "light-canceling" machine. They found that by reshaping the glass tubes (Tapered) or using a wide bouncy room (MMI), we can see exoplanets more clearly across a wider range of colors. While the reshaped tubes are the most efficient, the bouncy room is the most reliable. Both are huge steps forward in our quest to find other worlds.

1. Problem Statement

Direct detection of exoplanets, particularly Earth-like planets in close orbits, is hindered by the overwhelming brightness of their host stars. The required contrast ratios can reach $10^{-10}$ . Nulling interferometry is a technique used to suppress starlight via destructive interference while allowing off-axis planetary light to pass through constructive interference.

While traditional bulk optics and 3D waveguide technologies (like ultrafast laser inscription) have been used, the next generation of instruments (e.g., for the Giant Magellan Telescope or Habitable Worlds Observatory) requires compact, scalable, 2D photonic integrated circuits (PICs).

The Challenge: Standard 2D evanescent couplers (two-waveguide) are inherently chromatic, meaning their nulling performance degrades rapidly across a broadband spectrum (1.5–1.8 µm).
The Goal: Develop and compare 2D tri-coupler (three-waveguide) configurations that offer deep, broadband (achromatic) starlight attenuation, high exoplanet throughput, and robust sensing capabilities for fringe tracking, all while tolerating fabrication imperfections.

2. Methodology

The authors utilized RSoft BeamPROP (a 3-D semi-vector beam propagation simulator) to model and compare three distinct tri-coupler designs fabricated on a silica-on-silicon platform (Ge-doped silica core, $\Delta n \approx 0.03$ ). The designs were simulated across the 1.5–1.8 µm near-infrared band using Transverse Electric (TE) modes.

The three configurations evaluated were:

Standard Tri-coupler: Three equal-width waveguides coupled evanescently.
Tapered Tri-coupler: A bespoke design where the outer waveguides linearly taper in width along the coupling region (inspired by Hsiao et al.) to achieve partial achromatization.
Multimode Interference (MMI) Coupler: A 3×3 MMI device utilizing constructive and destructive interference of multiple excited modes.

Key Metrics Evaluated:

Exoplanet Throughput ( $\eta_P$ ): The fraction of off-axis light (0° phase difference) transmitted to the central output.
Starlight Attenuation ( $\eta_S$ ): The suppression of on-axis light (180° phase difference) at the central output.
Fringe Tracking (Sensing): The ability to detect phase errors using the differential intensity of the outer output ports (non-degenerate sensing).
Fabrication Tolerance: Sensitivity to waveguide asymmetry (lateral shifts) and uniform width reduction (under-etching).
Loss: Intrinsic material loss, bend radiation, and scattering.

3. Key Contributions

Comparative Analysis of 2D Tri-couplers: This is the first comprehensive comparison of standard evanescent, tapered evanescent, and MMI tri-couplers specifically for nulling interferometry in a high-index-contrast 2D platform.
Novel MMI Design: The authors proposed a specific MMI geometry (utilizing $N=7$ to $N=8$ image formation) that fits within a compact footprint comparable to evanescent couplers while providing a 3-way split suitable for nulling.
Non-Degenerate Sensing Demonstration: The paper demonstrates that by adjusting taper parameters, a tri-coupler can be designed to avoid "degenerate" states (where sensing slope is zero) across the entire bandwidth, enabling continuous fringe tracking.
Fabrication Tolerance Quantification: The study quantifies how specific fabrication errors (asymmetry vs. uniform shrinkage) impact null depth, revealing that MMIs are significantly more robust to asymmetry than evanescent couplers.

4. Results

A. Exoplanet Throughput

Standard Tri-coupler: Highly chromatic. Throughput drops from 97% at 1.55 µm to **50%** at the band edges (1.8 µm).
MMI: Improved achromaticity, maintaining >85% throughput across the band.
Tapered Tri-coupler: Best performance, maintaining ~100% throughput across the full 1.5–1.8 µm band.

B. Starlight Attenuation (Null Depth)

Ideal Conditions: All designs can theoretically achieve deep nulls (>60 dB) limited only by simulation precision.
Scattered Light Limit: In realistic simulations including bend radiation, attenuation is limited to ~30–40 dB.
Fabrication Impact (Asymmetry):
- Standard & Tapered: Highly sensitive. A lateral shift of just 0.1 µm degrades attenuation to 20–30 dB.
- MMI: Most robust. Even with a 0.1 µm shift, it maintains attenuation >40 dB.
Fabrication Impact (Uniform Shrinkage): The MMI shows negligible change in throughput (<1%), whereas the standard tri-coupler suffers up to 8% loss due to beat-length shifts.

C. Sensing and Fringe Tracking

Degeneracy Issue: Standard and MMI designs exhibit "degenerate" states at specific wavelengths where the slope of the outer port intensity vs. phase is zero, making fringe tracking impossible at those wavelengths.
Tapered Solution: A specific tapered design (optimized for sensing rather than throughput) was shown to be non-degenerate across the entire bandwidth, allowing continuous phase error detection.

D. Loss

Evanescent Couplers: Very low intrinsic loss (<0.01 dB), but suffer from bend radiation losses (0.16 dB at 1.55 µm, increasing to 0.62 dB at 1.8 µm).
MMI: Higher intrinsic loss (0.3–1.2 dB) due to mode mismatch and scattering, though this is wavelength-dependent.

5. Significance and Conclusion

This work establishes a critical pathway for integrating nulling interferometry into compact, scalable photonic chips for future exoplanet missions.

Trade-offs Identified: There is no single "perfect" coupler.
- The Tapered Tri-coupler offers the best optical performance (highest throughput, best achromaticity, best sensing) but is highly sensitive to fabrication errors.
- The MMI offers superior fabrication tolerance (maintaining >40 dB attenuation despite errors) and good throughput, making it the preferred choice for current technology where fabrication precision is limited, despite its higher intrinsic loss.
- The Standard Tri-coupler is outperformed by both in broadband applications.
Future Outlook: The authors suggest that while current results are promising, future work should explore sub-wavelength gratings and inverse design (AI/Neural Networks) to create ultra-broadband couplers that combine the high throughput of the tapered design with the robustness of the MMI. Experimental validation is the immediate next step to confirm these simulation predictions.

In summary, the paper provides a rigorous framework for selecting photonic beam combiners, highlighting that MMIs are currently the most practical solution for robust, high-contrast imaging, while tapered designs represent the future target for high-performance, precision-fabricated instruments.