Low-noise Fourier Transform Spectroscopy Enabled by Superconducting On-Chip Filterbank Spectrometers

This paper proposes a hybrid spectroscopic architecture that couples a medium-resolution Fourier transform spectrometer with a low-resolution on-chip filterbank spectrometer to significantly reduce photon noise and enable high-sensitivity, large-field line intensity mapping at resolutions of R~1000 for ground-based and balloon-borne astronomy.

The Gist

The Big Picture: Listening to the Universe's Whisper

Imagine you are trying to listen to a specific instrument in a massive, chaotic orchestra. The orchestra is the universe, and the instruments are galaxies. To understand how they are formed and how they move, astronomers need to "listen" to the specific notes (frequencies) they emit. This is called spectroscopy.

The problem is that the "music" of the universe (in millimeter and sub-millimeter waves) is very faint, and the "hall" (our atmosphere or the telescope) is very noisy.

The Problem: The "Blind" vs. The "Overwhelmed"

Astronomers currently have two main ways to listen to this music, but both have flaws:

1. The "Super-Listener" (Fourier Transform Spectrometer - FTS):

   • How it works: Imagine a detective who listens to the entire orchestra at once to figure out who is playing what. It's incredibly efficient and can hear a huge range of notes at the same time.
   • The Flaw: Because it listens to everything at once, the background noise is overwhelming. It's like trying to hear a single violin in a stadium full of people shouting. The signal gets lost in the "photon noise" (static).

2. The "Filter Bank" (On-Chip Filterbank Spectrometer - FBS):

   • How it works: Imagine a choir where every singer has a specific note they are allowed to sing. You put a filter in front of each singer so they only hear their own note. This is great for reducing noise because you aren't listening to the whole orchestra, just one note at a time.
   • The Flaw: To get high-quality sound (high resolution), you need thousands of singers (detectors). Building a choir that big is incredibly expensive, difficult to manufacture, and prone to errors. It's like trying to build a stadium-sized choir where every single singer must be perfect.

The Solution: The "Smart Filter" (FBDFTS)

The authors of this paper propose a brilliant hybrid solution: The Filterbank-Dispersed Fourier Transform Spectrometer (FBDFTS).

Think of it as a two-stage security checkpoint for light.

1. Stage 1: The Wide Net (The FTS).
   First, the light goes through the "Super-Listener" (the FTS). This device separates the light based on how far the waves travel, giving us a medium-resolution map of the whole scene. It's like a wide-angle camera that takes a blurry but very detailed photo of the whole orchestra.

2. Stage 2: The Noise Canceller (The Filterbank).
   Instead of sending that blurry photo directly to a detector, we pass it through a "Smart Filter" (the low-resolution filterbank).

   • The Analogy: Imagine you have a bucket of muddy water (the noisy light from the FTS). Instead of trying to drink the whole bucket, you pour it through a series of fine sieves (the filterbank). Each sieve catches a tiny, specific color of the mud.
   • The Result: The detector only sees a tiny, clean slice of the light. Because the slice is so small, the "noise" (the static) drops dramatically—by more than 10 times!

Why This is a Game-Changer

• Best of Both Worlds: You get the wide field of view and efficiency of the "Super-Listener" (FTS) but with the low noise of the "Filter Bank."
• No Massive Choir Needed: You don't need thousands of expensive detectors. You only need enough to cover the small slices created by the filterbank. This makes the technology much cheaper and easier to build.
• Sharper Vision: This setup allows astronomers to build instruments that can see details 1,000 times sharper than before, without the noise getting in the way.

What Can We Do With This?

The paper predicts that if we build this instrument (specifically using the James Clerk Maxwell Telescope as a test bed), we could:

• Map the "First Galaxies": We could see the very first galaxies forming in the universe, which are currently too faint and blurry to see clearly.
• Listen to the "CO Song": Carbon Monoxide (CO) is like a universal tracer for gas clouds where stars are born. This new instrument could map the "power spectrum" (the rhythm and pattern) of these gas clouds across the universe with incredible precision.
• Speed: It would be able to map the sky 10 to 100 times faster than current methods, turning a project that takes a lifetime into one that takes a few years.

The Bottom Line

The authors are essentially saying: "We can't build the perfect, massive choir of detectors yet, and the current 'Super-Listener' is too noisy. So, let's combine them. Let's use the Super-Listener to do the heavy lifting, and then use a small, smart filter to clean up the noise. This gives us a super-powerful, low-noise telescope that can finally hear the faint whispers of the early universe."

Technical Summary

1. Problem Statement

The paper addresses the critical limitations facing current spectroscopic architectures for millimeter (mm) and sub-millimeter (sub-mm) astronomy, specifically regarding medium-resolution ($R \sim 1000$) line intensity mapping (LIM) and large field-of-view (FoV) galaxy surveys.

• Fourier Transform Spectrometers (FTS): While FTS offer high throughput and broad spectral coverage, their inherent "multiplexing advantage" (where every detector measures the full spectrum simultaneously) results in a significant increase in photon noise. This noise scales with the total spectral bandwidth, severely limiting mapping speeds for high-resolution applications.
• On-Chip Filterbank Spectrometers (FBS): These superconducting devices offer compact spectral dispersion but face three major hurdles for scaling to $R \sim 1000$:
  1. Detector Count: Resolution scales linearly with the number of detectors ($\sim 2R$ per polarization). Achieving $R \sim 1000$ requires thousands of additional detectors per pixel, which is currently technologically unfeasible for large arrays.
  2. Efficiency Loss: Dielectric losses in the superconducting circuit materials (e.g., SiN) increase significantly as the filter bandwidth narrows (higher $R$), reducing overall efficiency.
  3. Fabrication Tolerances: High resolution requires feature variations $< 50$ nm, which is difficult to maintain across large arrays.
• Grating Spectrometers: These struggle to scale to $R \ge 1000$ in the mm/sub-mm regime due to throughput losses (Jacquinot's advantage) and the difficulty of coupling them to large FoVs without sacrificing spatial dimensions.

2. Methodology

The authors propose a hybrid architecture called the Filterbank-Dispersed Fourier Transform Spectrometer (FBDFTS).

• Core Concept: Coupling a medium-resolution FTS ($R \sim 1000$) with a low-resolution on-chip filterbank spectrometer ($R \sim 200$) acting as a post-dispersion element.
• Operational Mechanism:
  1. The FTS performs the primary spectral resolution ($R \sim 1000$).
  2. The light exiting the FTS is dispersed by the FBS, which splits the light into narrow spectral bands before it reaches the detectors.
  3. By narrowing the bandwidth incident on each detector, the FBS drastically reduces the photon noise inherent to the FTS's broadband measurement.
• Advantages over Alternatives:
  • Compactness: Unlike grating dispersers which require large optics ($\sim 1$ m scale for mm waves), the FBS disperses light over a footprint of $\sim 1$ cm.
  • Imaging Preservation: Unlike grating spectrometers that sacrifice a spatial dimension for spectral dispersion, the FBDFTS maintains the full 2D imaging capability of the FTS focal plane.
  • Polarization: The FBS handles polarization at the focal plane using standard orthomode transducer designs, avoiding the need for complex dual-dispersion optics.

3. Key Contributions

• Architectural Innovation: The proposal of the FBDFTS as a viable solution to achieve $R \sim 1000$ mapping spectroscopy without requiring the immediate development of massive, high-resolution FBS arrays.
• Noise Reduction Analysis: Theoretical demonstration that a post-dispersed FTS can reduce photon noise by more than an order of magnitude compared to a nominal FTS, while retaining the FTS's high throughput.
• Mapping Speed Projections: Calculation of projected mapping speeds for an FBDFTS from two platforms:
  • Ground-based: CCAT-prime/FYST observatory.
  • Balloon-borne: Similar to the BLAST observatory.
• LIM Feasibility Study: A detailed performance prediction for a Line Intensity Mapping (LIM) experiment using the James Clerk Maxwell Telescope (JCMT). The study models a photon-noise-limited system with $R=1000$ (FTS) and $R=200$ (FBS).

4. Results

• Photon Noise Reduction: Simulations (Fig. 3) show that using a low-resolution ($R=200$) FBS as a post-disperser reduces the Noise Equivalent Power (NEP) of the photon noise significantly compared to a standard FTS.
• Mapping Speed:
  • The FBDFTS achieves mapping speeds comparable to low-resolution instruments but with $R \sim 1000$ resolution.
  • It offers an order-of-magnitude improvement in mapping speed over a nominal FTS.
  • It can achieve mapping speeds similar to the EoR-Spec instrument (which uses a Fabry-Perot) but with significantly higher spectral resolution.
• JCMT LIM Experiment:
  • The proposed instrument would require $\sim 4.5 \times 10^4$ detectors and a moving mirror travel of $\sim 450$ mm.
  • Performance: It is predicted to measure CO power spectra (transitions CO(2-1), CO(3-2), CO(4-3)) at redshifts $z=0.5, 1.3, 2.1$ with a Signal-to-Noise Ratio (SNR) of 10–100.
  • Survey Time: These measurements are achievable within $10^5$ to $10^6$ spectrometer-hours (roughly 1,000 hours of integration).
  • Comparison: This represents a factor of $\sim 15$ improvement over the projected performance of the CONCERTO experiment (which uses $R \sim 300$).

5. Significance

• Enabling Next-Gen Surveys: The FBDFTS architecture bridges the gap between current technology and the requirements for future large-scale surveys like AtLAST (Atacama Large Aperture Submillimeter Telescope), which targets $R \sim 2000$ and large FoVs to study the first forming galaxies ($z > 10$).
• Technological Feasibility: By leveraging existing, mature technologies (medium-resolution FTS and low-resolution FBS), the FBDFTS avoids the "valley of death" where scaling FBS to $R \sim 1000$ directly is currently impossible due to detector count and fabrication limits.
• Scientific Impact: This approach enables the first near-term $R \sim 1000$ sub-mm/mm LIM experiments. This resolution is crucial for:
  • Recovering higher spatial modes in power spectra.
  • Resolving galaxy sources along the line of sight.
  • Effectively removing contaminant spectral features (interlopers) in cosmological studies.
• Systematic Mitigation: The paper notes that the FTS component can serve as a self-calibrator for the FBS, leveraging the extensive heritage of FTS calibration to mitigate systematic errors.

In conclusion, the paper presents a pragmatic and highly promising pathway to achieving high-resolution, large-field-of-view spectroscopy for cosmology and galaxy evolution, overcoming the specific noise and scalability bottlenecks of current standalone technologies.